diff options
Diffstat (limited to 'fs/btrfs')
61 files changed, 28688 insertions, 5792 deletions
diff --git a/fs/btrfs/Kconfig b/fs/btrfs/Kconfig index ecb9fd3be143..d33f01c08b60 100644 --- a/fs/btrfs/Kconfig +++ b/fs/btrfs/Kconfig @@ -31,3 +31,22 @@ config BTRFS_FS_POSIX_ACL Linux website <http://acl.bestbits.at/>. If you don't know what Access Control Lists are, say N + +config BTRFS_FS_CHECK_INTEGRITY + bool "Btrfs with integrity check tool compiled in (DANGEROUS)" + depends on BTRFS_FS + help + Adds code that examines all block write requests (including + writes of the super block). The goal is to verify that the + state of the filesystem on disk is always consistent, i.e., + after a power-loss or kernel panic event the filesystem is + in a consistent state. + + If the integrity check tool is included and activated in + the mount options, plenty of kernel memory is used, and + plenty of additional CPU cycles are spent. Enabling this + functionality is not intended for normal use. + + In most cases, unless you are a btrfs developer who needs + to verify the integrity of (super)-block write requests + during the run of a regression test, say N diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile index 40e6ac08c21f..d7fcdba141a2 100644 --- a/fs/btrfs/Makefile +++ b/fs/btrfs/Makefile @@ -7,6 +7,8 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \ extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \ extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \ export.o tree-log.o free-space-cache.o zlib.o lzo.o \ - compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o + compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \ + reada.o backref.o ulist.o qgroup.o send.o btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o +btrfs-$(CONFIG_BTRFS_FS_CHECK_INTEGRITY) += check-integrity.o diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c index eb159aaa5a11..761e2cd8fed1 100644 --- a/fs/btrfs/acl.c +++ b/fs/btrfs/acl.c @@ -59,22 +59,19 @@ struct posix_acl *btrfs_get_acl(struct inode *inode, int type) if (!value) return ERR_PTR(-ENOMEM); size = __btrfs_getxattr(inode, name, value, size); - if (size > 0) { - acl = posix_acl_from_xattr(value, size); - if (IS_ERR(acl)) { - kfree(value); - return acl; - } - set_cached_acl(inode, type, acl); - } - kfree(value); + } + if (size > 0) { + acl = posix_acl_from_xattr(value, size); } else if (size == -ENOENT || size == -ENODATA || size == 0) { /* FIXME, who returns -ENOENT? I think nobody */ acl = NULL; - set_cached_acl(inode, type, acl); } else { acl = ERR_PTR(-EIO); } + kfree(value); + + if (!IS_ERR(acl)) + set_cached_acl(inode, type, acl); return acl; } @@ -230,7 +227,11 @@ int btrfs_init_acl(struct btrfs_trans_handle *trans, if (ret > 0) { /* we need an acl */ ret = btrfs_set_acl(trans, inode, acl, ACL_TYPE_ACCESS); + } else { + cache_no_acl(inode); } + } else { + cache_no_acl(inode); } failed: posix_acl_release(acl); diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c index 7ec14097fef1..58b7d14b08ee 100644 --- a/fs/btrfs/async-thread.c +++ b/fs/btrfs/async-thread.c @@ -64,6 +64,8 @@ struct btrfs_worker_thread { int idle; }; +static int __btrfs_start_workers(struct btrfs_workers *workers); + /* * btrfs_start_workers uses kthread_run, which can block waiting for memory * for a very long time. It will actually throttle on page writeback, @@ -88,27 +90,10 @@ static void start_new_worker_func(struct btrfs_work *work) { struct worker_start *start; start = container_of(work, struct worker_start, work); - btrfs_start_workers(start->queue, 1); + __btrfs_start_workers(start->queue); kfree(start); } -static int start_new_worker(struct btrfs_workers *queue) -{ - struct worker_start *start; - int ret; - - start = kzalloc(sizeof(*start), GFP_NOFS); - if (!start) - return -ENOMEM; - - start->work.func = start_new_worker_func; - start->queue = queue; - ret = btrfs_queue_worker(queue->atomic_worker_start, &start->work); - if (ret) - kfree(start); - return ret; -} - /* * helper function to move a thread onto the idle list after it * has finished some requests. @@ -153,12 +138,20 @@ static void check_busy_worker(struct btrfs_worker_thread *worker) static void check_pending_worker_creates(struct btrfs_worker_thread *worker) { struct btrfs_workers *workers = worker->workers; + struct worker_start *start; unsigned long flags; rmb(); if (!workers->atomic_start_pending) return; + start = kzalloc(sizeof(*start), GFP_NOFS); + if (!start) + return; + + start->work.func = start_new_worker_func; + start->queue = workers; + spin_lock_irqsave(&workers->lock, flags); if (!workers->atomic_start_pending) goto out; @@ -170,18 +163,19 @@ static void check_pending_worker_creates(struct btrfs_worker_thread *worker) workers->num_workers_starting += 1; spin_unlock_irqrestore(&workers->lock, flags); - start_new_worker(workers); + btrfs_queue_worker(workers->atomic_worker_start, &start->work); return; out: + kfree(start); spin_unlock_irqrestore(&workers->lock, flags); } -static noinline int run_ordered_completions(struct btrfs_workers *workers, +static noinline void run_ordered_completions(struct btrfs_workers *workers, struct btrfs_work *work) { if (!workers->ordered) - return 0; + return; set_bit(WORK_DONE_BIT, &work->flags); @@ -212,14 +206,20 @@ static noinline int run_ordered_completions(struct btrfs_workers *workers, work->ordered_func(work); - /* now take the lock again and call the freeing code */ + /* now take the lock again and drop our item from the list */ spin_lock(&workers->order_lock); list_del(&work->order_list); + spin_unlock(&workers->order_lock); + + /* + * we don't want to call the ordered free functions + * with the lock held though + */ work->ordered_free(work); + spin_lock(&workers->order_lock); } spin_unlock(&workers->order_lock); - return 0; } static void put_worker(struct btrfs_worker_thread *worker) @@ -331,7 +331,7 @@ again: run_ordered_completions(worker->workers, work); check_pending_worker_creates(worker); - + cond_resched(); } spin_lock_irq(&worker->lock); @@ -340,7 +340,7 @@ again: if (freezing(current)) { worker->working = 0; spin_unlock_irq(&worker->lock); - refrigerator(); + try_to_freeze(); } else { spin_unlock_irq(&worker->lock); if (!kthread_should_stop()) { @@ -405,7 +405,7 @@ again: /* * this will wait for all the worker threads to shutdown */ -int btrfs_stop_workers(struct btrfs_workers *workers) +void btrfs_stop_workers(struct btrfs_workers *workers) { struct list_head *cur; struct btrfs_worker_thread *worker; @@ -433,7 +433,6 @@ int btrfs_stop_workers(struct btrfs_workers *workers) put_worker(worker); } spin_unlock_irq(&workers->lock); - return 0; } /* @@ -462,56 +461,55 @@ void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max, * starts new worker threads. This does not enforce the max worker * count in case you need to temporarily go past it. */ -static int __btrfs_start_workers(struct btrfs_workers *workers, - int num_workers) +static int __btrfs_start_workers(struct btrfs_workers *workers) { struct btrfs_worker_thread *worker; int ret = 0; - int i; - for (i = 0; i < num_workers; i++) { - worker = kzalloc(sizeof(*worker), GFP_NOFS); - if (!worker) { - ret = -ENOMEM; - goto fail; - } + worker = kzalloc(sizeof(*worker), GFP_NOFS); + if (!worker) { + ret = -ENOMEM; + goto fail; + } - INIT_LIST_HEAD(&worker->pending); - INIT_LIST_HEAD(&worker->prio_pending); - INIT_LIST_HEAD(&worker->worker_list); - spin_lock_init(&worker->lock); - - atomic_set(&worker->num_pending, 0); - atomic_set(&worker->refs, 1); - worker->workers = workers; - worker->task = kthread_run(worker_loop, worker, - "btrfs-%s-%d", workers->name, - workers->num_workers + i); - if (IS_ERR(worker->task)) { - ret = PTR_ERR(worker->task); - kfree(worker); - goto fail; - } - spin_lock_irq(&workers->lock); - list_add_tail(&worker->worker_list, &workers->idle_list); - worker->idle = 1; - workers->num_workers++; - workers->num_workers_starting--; - WARN_ON(workers->num_workers_starting < 0); - spin_unlock_irq(&workers->lock); + INIT_LIST_HEAD(&worker->pending); + INIT_LIST_HEAD(&worker->prio_pending); + INIT_LIST_HEAD(&worker->worker_list); + spin_lock_init(&worker->lock); + + atomic_set(&worker->num_pending, 0); + atomic_set(&worker->refs, 1); + worker->workers = workers; + worker->task = kthread_run(worker_loop, worker, + "btrfs-%s-%d", workers->name, + workers->num_workers + 1); + if (IS_ERR(worker->task)) { + ret = PTR_ERR(worker->task); + kfree(worker); + goto fail; } + spin_lock_irq(&workers->lock); + list_add_tail(&worker->worker_list, &workers->idle_list); + worker->idle = 1; + workers->num_workers++; + workers->num_workers_starting--; + WARN_ON(workers->num_workers_starting < 0); + spin_unlock_irq(&workers->lock); + return 0; fail: - btrfs_stop_workers(workers); + spin_lock_irq(&workers->lock); + workers->num_workers_starting--; + spin_unlock_irq(&workers->lock); return ret; } -int btrfs_start_workers(struct btrfs_workers *workers, int num_workers) +int btrfs_start_workers(struct btrfs_workers *workers) { spin_lock_irq(&workers->lock); - workers->num_workers_starting += num_workers; + workers->num_workers_starting++; spin_unlock_irq(&workers->lock); - return __btrfs_start_workers(workers, num_workers); + return __btrfs_start_workers(workers); } /* @@ -568,9 +566,10 @@ static struct btrfs_worker_thread *find_worker(struct btrfs_workers *workers) struct btrfs_worker_thread *worker; unsigned long flags; struct list_head *fallback; + int ret; -again: spin_lock_irqsave(&workers->lock, flags); +again: worker = next_worker(workers); if (!worker) { @@ -584,7 +583,10 @@ again: workers->num_workers_starting++; spin_unlock_irqrestore(&workers->lock, flags); /* we're below the limit, start another worker */ - __btrfs_start_workers(workers, 1); + ret = __btrfs_start_workers(workers); + spin_lock_irqsave(&workers->lock, flags); + if (ret) + goto fallback; goto again; } } @@ -618,14 +620,14 @@ found: * it was taken from. It is intended for use with long running work functions * that make some progress and want to give the cpu up for others. */ -int btrfs_requeue_work(struct btrfs_work *work) +void btrfs_requeue_work(struct btrfs_work *work) { struct btrfs_worker_thread *worker = work->worker; unsigned long flags; int wake = 0; if (test_and_set_bit(WORK_QUEUED_BIT, &work->flags)) - goto out; + return; spin_lock_irqsave(&worker->lock, flags); if (test_bit(WORK_HIGH_PRIO_BIT, &work->flags)) @@ -652,9 +654,6 @@ int btrfs_requeue_work(struct btrfs_work *work) if (wake) wake_up_process(worker->task); spin_unlock_irqrestore(&worker->lock, flags); -out: - - return 0; } void btrfs_set_work_high_prio(struct btrfs_work *work) @@ -665,7 +664,7 @@ void btrfs_set_work_high_prio(struct btrfs_work *work) /* * places a struct btrfs_work into the pending queue of one of the kthreads */ -int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) +void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) { struct btrfs_worker_thread *worker; unsigned long flags; @@ -673,7 +672,7 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) /* don't requeue something already on a list */ if (test_and_set_bit(WORK_QUEUED_BIT, &work->flags)) - goto out; + return; worker = find_worker(workers); if (workers->ordered) { @@ -712,7 +711,4 @@ int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work) if (wake) wake_up_process(worker->task); spin_unlock_irqrestore(&worker->lock, flags); - -out: - return 0; } diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h index 5077746cf85e..063698b90ce2 100644 --- a/fs/btrfs/async-thread.h +++ b/fs/btrfs/async-thread.h @@ -109,11 +109,11 @@ struct btrfs_workers { char *name; }; -int btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work); -int btrfs_start_workers(struct btrfs_workers *workers, int num_workers); -int btrfs_stop_workers(struct btrfs_workers *workers); +void btrfs_queue_worker(struct btrfs_workers *workers, struct btrfs_work *work); +int btrfs_start_workers(struct btrfs_workers *workers); +void btrfs_stop_workers(struct btrfs_workers *workers); void btrfs_init_workers(struct btrfs_workers *workers, char *name, int max, struct btrfs_workers *async_starter); -int btrfs_requeue_work(struct btrfs_work *work); +void btrfs_requeue_work(struct btrfs_work *work); void btrfs_set_work_high_prio(struct btrfs_work *work); #endif diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c new file mode 100644 index 000000000000..ff6475f409d6 --- /dev/null +++ b/fs/btrfs/backref.c @@ -0,0 +1,1631 @@ +/* + * Copyright (C) 2011 STRATO. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include "ctree.h" +#include "disk-io.h" +#include "backref.h" +#include "ulist.h" +#include "transaction.h" +#include "delayed-ref.h" +#include "locking.h" + +struct extent_inode_elem { + u64 inum; + u64 offset; + struct extent_inode_elem *next; +}; + +static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb, + struct btrfs_file_extent_item *fi, + u64 extent_item_pos, + struct extent_inode_elem **eie) +{ + u64 data_offset; + u64 data_len; + struct extent_inode_elem *e; + + data_offset = btrfs_file_extent_offset(eb, fi); + data_len = btrfs_file_extent_num_bytes(eb, fi); + + if (extent_item_pos < data_offset || + extent_item_pos >= data_offset + data_len) + return 1; + + e = kmalloc(sizeof(*e), GFP_NOFS); + if (!e) + return -ENOMEM; + + e->next = *eie; + e->inum = key->objectid; + e->offset = key->offset + (extent_item_pos - data_offset); + *eie = e; + + return 0; +} + +static int find_extent_in_eb(struct extent_buffer *eb, u64 wanted_disk_byte, + u64 extent_item_pos, + struct extent_inode_elem **eie) +{ + u64 disk_byte; + struct btrfs_key key; + struct btrfs_file_extent_item *fi; + int slot; + int nritems; + int extent_type; + int ret; + + /* + * from the shared data ref, we only have the leaf but we need + * the key. thus, we must look into all items and see that we + * find one (some) with a reference to our extent item. + */ + nritems = btrfs_header_nritems(eb); + for (slot = 0; slot < nritems; ++slot) { + btrfs_item_key_to_cpu(eb, &key, slot); + if (key.type != BTRFS_EXTENT_DATA_KEY) + continue; + fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); + extent_type = btrfs_file_extent_type(eb, fi); + if (extent_type == BTRFS_FILE_EXTENT_INLINE) + continue; + /* don't skip BTRFS_FILE_EXTENT_PREALLOC, we can handle that */ + disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); + if (disk_byte != wanted_disk_byte) + continue; + + ret = check_extent_in_eb(&key, eb, fi, extent_item_pos, eie); + if (ret < 0) + return ret; + } + + return 0; +} + +/* + * this structure records all encountered refs on the way up to the root + */ +struct __prelim_ref { + struct list_head list; + u64 root_id; + struct btrfs_key key_for_search; + int level; + int count; + struct extent_inode_elem *inode_list; + u64 parent; + u64 wanted_disk_byte; +}; + +/* + * the rules for all callers of this function are: + * - obtaining the parent is the goal + * - if you add a key, you must know that it is a correct key + * - if you cannot add the parent or a correct key, then we will look into the + * block later to set a correct key + * + * delayed refs + * ============ + * backref type | shared | indirect | shared | indirect + * information | tree | tree | data | data + * --------------------+--------+----------+--------+---------- + * parent logical | y | - | - | - + * key to resolve | - | y | y | y + * tree block logical | - | - | - | - + * root for resolving | y | y | y | y + * + * - column 1: we've the parent -> done + * - column 2, 3, 4: we use the key to find the parent + * + * on disk refs (inline or keyed) + * ============================== + * backref type | shared | indirect | shared | indirect + * information | tree | tree | data | data + * --------------------+--------+----------+--------+---------- + * parent logical | y | - | y | - + * key to resolve | - | - | - | y + * tree block logical | y | y | y | y + * root for resolving | - | y | y | y + * + * - column 1, 3: we've the parent -> done + * - column 2: we take the first key from the block to find the parent + * (see __add_missing_keys) + * - column 4: we use the key to find the parent + * + * additional information that's available but not required to find the parent + * block might help in merging entries to gain some speed. + */ + +static int __add_prelim_ref(struct list_head *head, u64 root_id, + struct btrfs_key *key, int level, + u64 parent, u64 wanted_disk_byte, int count) +{ + struct __prelim_ref *ref; + + /* in case we're adding delayed refs, we're holding the refs spinlock */ + ref = kmalloc(sizeof(*ref), GFP_ATOMIC); + if (!ref) + return -ENOMEM; + + ref->root_id = root_id; + if (key) + ref->key_for_search = *key; + else + memset(&ref->key_for_search, 0, sizeof(ref->key_for_search)); + + ref->inode_list = NULL; + ref->level = level; + ref->count = count; + ref->parent = parent; + ref->wanted_disk_byte = wanted_disk_byte; + list_add_tail(&ref->list, head); + + return 0; +} + +static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path, + struct ulist *parents, int level, + struct btrfs_key *key_for_search, u64 time_seq, + u64 wanted_disk_byte, + const u64 *extent_item_pos) +{ + int ret = 0; + int slot; + struct extent_buffer *eb; + struct btrfs_key key; + struct btrfs_file_extent_item *fi; + struct extent_inode_elem *eie = NULL; + u64 disk_byte; + + if (level != 0) { + eb = path->nodes[level]; + ret = ulist_add(parents, eb->start, 0, GFP_NOFS); + if (ret < 0) + return ret; + return 0; + } + + /* + * We normally enter this function with the path already pointing to + * the first item to check. But sometimes, we may enter it with + * slot==nritems. In that case, go to the next leaf before we continue. + */ + if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) + ret = btrfs_next_old_leaf(root, path, time_seq); + + while (!ret) { + eb = path->nodes[0]; + slot = path->slots[0]; + + btrfs_item_key_to_cpu(eb, &key, slot); + + if (key.objectid != key_for_search->objectid || + key.type != BTRFS_EXTENT_DATA_KEY) + break; + + fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); + disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); + + if (disk_byte == wanted_disk_byte) { + eie = NULL; + if (extent_item_pos) { + ret = check_extent_in_eb(&key, eb, fi, + *extent_item_pos, + &eie); + if (ret < 0) + break; + } + if (!ret) { + ret = ulist_add(parents, eb->start, + (unsigned long)eie, GFP_NOFS); + if (ret < 0) + break; + if (!extent_item_pos) { + ret = btrfs_next_old_leaf(root, path, + time_seq); + continue; + } + } + } + ret = btrfs_next_old_item(root, path, time_seq); + } + + if (ret > 0) + ret = 0; + return ret; +} + +/* + * resolve an indirect backref in the form (root_id, key, level) + * to a logical address + */ +static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info, + int search_commit_root, + u64 time_seq, + struct __prelim_ref *ref, + struct ulist *parents, + const u64 *extent_item_pos) +{ + struct btrfs_path *path; + struct btrfs_root *root; + struct btrfs_key root_key; + struct extent_buffer *eb; + int ret = 0; + int root_level; + int level = ref->level; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + path->search_commit_root = !!search_commit_root; + + root_key.objectid = ref->root_id; + root_key.type = BTRFS_ROOT_ITEM_KEY; + root_key.offset = (u64)-1; + root = btrfs_read_fs_root_no_name(fs_info, &root_key); + if (IS_ERR(root)) { + ret = PTR_ERR(root); + goto out; + } + + rcu_read_lock(); + root_level = btrfs_header_level(root->node); + rcu_read_unlock(); + + if (root_level + 1 == level) + goto out; + + path->lowest_level = level; + ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq); + pr_debug("search slot in root %llu (level %d, ref count %d) returned " + "%d for key (%llu %u %llu)\n", + (unsigned long long)ref->root_id, level, ref->count, ret, + (unsigned long long)ref->key_for_search.objectid, + ref->key_for_search.type, + (unsigned long long)ref->key_for_search.offset); + if (ret < 0) + goto out; + + eb = path->nodes[level]; + while (!eb) { + if (!level) { + WARN_ON(1); + ret = 1; + goto out; + } + level--; + eb = path->nodes[level]; + } + + ret = add_all_parents(root, path, parents, level, &ref->key_for_search, + time_seq, ref->wanted_disk_byte, + extent_item_pos); +out: + btrfs_free_path(path); + return ret; +} + +/* + * resolve all indirect backrefs from the list + */ +static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info, + int search_commit_root, u64 time_seq, + struct list_head *head, + const u64 *extent_item_pos) +{ + int err; + int ret = 0; + struct __prelim_ref *ref; + struct __prelim_ref *ref_safe; + struct __prelim_ref *new_ref; + struct ulist *parents; + struct ulist_node *node; + struct ulist_iterator uiter; + + parents = ulist_alloc(GFP_NOFS); + if (!parents) + return -ENOMEM; + + /* + * _safe allows us to insert directly after the current item without + * iterating over the newly inserted items. + * we're also allowed to re-assign ref during iteration. + */ + list_for_each_entry_safe(ref, ref_safe, head, list) { + if (ref->parent) /* already direct */ + continue; + if (ref->count == 0) + continue; + err = __resolve_indirect_ref(fs_info, search_commit_root, + time_seq, ref, parents, + extent_item_pos); + if (err) { + if (ret == 0) + ret = err; + continue; + } + + /* we put the first parent into the ref at hand */ + ULIST_ITER_INIT(&uiter); + node = ulist_next(parents, &uiter); + ref->parent = node ? node->val : 0; + ref->inode_list = + node ? (struct extent_inode_elem *)node->aux : 0; + + /* additional parents require new refs being added here */ + while ((node = ulist_next(parents, &uiter))) { + new_ref = kmalloc(sizeof(*new_ref), GFP_NOFS); + if (!new_ref) { + ret = -ENOMEM; + break; + } + memcpy(new_ref, ref, sizeof(*ref)); + new_ref->parent = node->val; + new_ref->inode_list = + (struct extent_inode_elem *)node->aux; + list_add(&new_ref->list, &ref->list); + } + ulist_reinit(parents); + } + + ulist_free(parents); + return ret; +} + +static inline int ref_for_same_block(struct __prelim_ref *ref1, + struct __prelim_ref *ref2) +{ + if (ref1->level != ref2->level) + return 0; + if (ref1->root_id != ref2->root_id) + return 0; + if (ref1->key_for_search.type != ref2->key_for_search.type) + return 0; + if (ref1->key_for_search.objectid != ref2->key_for_search.objectid) + return 0; + if (ref1->key_for_search.offset != ref2->key_for_search.offset) + return 0; + if (ref1->parent != ref2->parent) + return 0; + + return 1; +} + +/* + * read tree blocks and add keys where required. + */ +static int __add_missing_keys(struct btrfs_fs_info *fs_info, + struct list_head *head) +{ + struct list_head *pos; + struct extent_buffer *eb; + + list_for_each(pos, head) { + struct __prelim_ref *ref; + ref = list_entry(pos, struct __prelim_ref, list); + + if (ref->parent) + continue; + if (ref->key_for_search.type) + continue; + BUG_ON(!ref->wanted_disk_byte); + eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte, + fs_info->tree_root->leafsize, 0); + BUG_ON(!eb); + btrfs_tree_read_lock(eb); + if (btrfs_header_level(eb) == 0) + btrfs_item_key_to_cpu(eb, &ref->key_for_search, 0); + else + btrfs_node_key_to_cpu(eb, &ref->key_for_search, 0); + btrfs_tree_read_unlock(eb); + free_extent_buffer(eb); + } + return 0; +} + +/* + * merge two lists of backrefs and adjust counts accordingly + * + * mode = 1: merge identical keys, if key is set + * FIXME: if we add more keys in __add_prelim_ref, we can merge more here. + * additionally, we could even add a key range for the blocks we + * looked into to merge even more (-> replace unresolved refs by those + * having a parent). + * mode = 2: merge identical parents + */ +static int __merge_refs(struct list_head *head, int mode) +{ + struct list_head *pos1; + + list_for_each(pos1, head) { + struct list_head *n2; + struct list_head *pos2; + struct __prelim_ref *ref1; + + ref1 = list_entry(pos1, struct __prelim_ref, list); + + for (pos2 = pos1->next, n2 = pos2->next; pos2 != head; + pos2 = n2, n2 = pos2->next) { + struct __prelim_ref *ref2; + struct __prelim_ref *xchg; + + ref2 = list_entry(pos2, struct __prelim_ref, list); + + if (mode == 1) { + if (!ref_for_same_block(ref1, ref2)) + continue; + if (!ref1->parent && ref2->parent) { + xchg = ref1; + ref1 = ref2; + ref2 = xchg; + } + ref1->count += ref2->count; + } else { + if (ref1->parent != ref2->parent) + continue; + ref1->count += ref2->count; + } + list_del(&ref2->list); + kfree(ref2); + } + + } + return 0; +} + +/* + * add all currently queued delayed refs from this head whose seq nr is + * smaller or equal that seq to the list + */ +static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq, + struct list_head *prefs) +{ + struct btrfs_delayed_extent_op *extent_op = head->extent_op; + struct rb_node *n = &head->node.rb_node; + struct btrfs_key key; + struct btrfs_key op_key = {0}; + int sgn; + int ret = 0; + + if (extent_op && extent_op->update_key) + btrfs_disk_key_to_cpu(&op_key, &extent_op->key); + + while ((n = rb_prev(n))) { + struct btrfs_delayed_ref_node *node; + node = rb_entry(n, struct btrfs_delayed_ref_node, + rb_node); + if (node->bytenr != head->node.bytenr) + break; + WARN_ON(node->is_head); + + if (node->seq > seq) + continue; + + switch (node->action) { + case BTRFS_ADD_DELAYED_EXTENT: + case BTRFS_UPDATE_DELAYED_HEAD: + WARN_ON(1); + continue; + case BTRFS_ADD_DELAYED_REF: + sgn = 1; + break; + case BTRFS_DROP_DELAYED_REF: + sgn = -1; + break; + default: + BUG_ON(1); + } + switch (node->type) { + case BTRFS_TREE_BLOCK_REF_KEY: { + struct btrfs_delayed_tree_ref *ref; + + ref = btrfs_delayed_node_to_tree_ref(node); + ret = __add_prelim_ref(prefs, ref->root, &op_key, + ref->level + 1, 0, node->bytenr, + node->ref_mod * sgn); + break; + } + case BTRFS_SHARED_BLOCK_REF_KEY: { + struct btrfs_delayed_tree_ref *ref; + + ref = btrfs_delayed_node_to_tree_ref(node); + ret = __add_prelim_ref(prefs, ref->root, NULL, + ref->level + 1, ref->parent, + node->bytenr, + node->ref_mod * sgn); + break; + } + case BTRFS_EXTENT_DATA_REF_KEY: { + struct btrfs_delayed_data_ref *ref; + ref = btrfs_delayed_node_to_data_ref(node); + + key.objectid = ref->objectid; + key.type = BTRFS_EXTENT_DATA_KEY; + key.offset = ref->offset; + ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0, + node->bytenr, + node->ref_mod * sgn); + break; + } + case BTRFS_SHARED_DATA_REF_KEY: { + struct btrfs_delayed_data_ref *ref; + + ref = btrfs_delayed_node_to_data_ref(node); + + key.objectid = ref->objectid; + key.type = BTRFS_EXTENT_DATA_KEY; + key.offset = ref->offset; + ret = __add_prelim_ref(prefs, ref->root, &key, 0, + ref->parent, node->bytenr, + node->ref_mod * sgn); + break; + } + default: + WARN_ON(1); + } + BUG_ON(ret); + } + + return 0; +} + +/* + * add all inline backrefs for bytenr to the list + */ +static int __add_inline_refs(struct btrfs_fs_info *fs_info, + struct btrfs_path *path, u64 bytenr, + int *info_level, struct list_head *prefs) +{ + int ret = 0; + int slot; + struct extent_buffer *leaf; + struct btrfs_key key; + unsigned long ptr; + unsigned long end; + struct btrfs_extent_item *ei; + u64 flags; + u64 item_size; + + /* + * enumerate all inline refs + */ + leaf = path->nodes[0]; + slot = path->slots[0]; + + item_size = btrfs_item_size_nr(leaf, slot); + BUG_ON(item_size < sizeof(*ei)); + + ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item); + flags = btrfs_extent_flags(leaf, ei); + + ptr = (unsigned long)(ei + 1); + end = (unsigned long)ei + item_size; + + if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + struct btrfs_tree_block_info *info; + + info = (struct btrfs_tree_block_info *)ptr; + *info_level = btrfs_tree_block_level(leaf, info); + ptr += sizeof(struct btrfs_tree_block_info); + BUG_ON(ptr > end); + } else { + BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA)); + } + + while (ptr < end) { + struct btrfs_extent_inline_ref *iref; + u64 offset; + int type; + + iref = (struct btrfs_extent_inline_ref *)ptr; + type = btrfs_extent_inline_ref_type(leaf, iref); + offset = btrfs_extent_inline_ref_offset(leaf, iref); + + switch (type) { + case BTRFS_SHARED_BLOCK_REF_KEY: + ret = __add_prelim_ref(prefs, 0, NULL, + *info_level + 1, offset, + bytenr, 1); + break; + case BTRFS_SHARED_DATA_REF_KEY: { + struct btrfs_shared_data_ref *sdref; + int count; + + sdref = (struct btrfs_shared_data_ref *)(iref + 1); + count = btrfs_shared_data_ref_count(leaf, sdref); + ret = __add_prelim_ref(prefs, 0, NULL, 0, offset, + bytenr, count); + break; + } + case BTRFS_TREE_BLOCK_REF_KEY: + ret = __add_prelim_ref(prefs, offset, NULL, + *info_level + 1, 0, + bytenr, 1); + break; + case BTRFS_EXTENT_DATA_REF_KEY: { + struct btrfs_extent_data_ref *dref; + int count; + u64 root; + + dref = (struct btrfs_extent_data_ref *)(&iref->offset); + count = btrfs_extent_data_ref_count(leaf, dref); + key.objectid = btrfs_extent_data_ref_objectid(leaf, + dref); + key.type = BTRFS_EXTENT_DATA_KEY; + key.offset = btrfs_extent_data_ref_offset(leaf, dref); + root = btrfs_extent_data_ref_root(leaf, dref); + ret = __add_prelim_ref(prefs, root, &key, 0, 0, + bytenr, count); + break; + } + default: + WARN_ON(1); + } + BUG_ON(ret); + ptr += btrfs_extent_inline_ref_size(type); + } + + return 0; +} + +/* + * add all non-inline backrefs for bytenr to the list + */ +static int __add_keyed_refs(struct btrfs_fs_info *fs_info, + struct btrfs_path *path, u64 bytenr, + int info_level, struct list_head *prefs) +{ + struct btrfs_root *extent_root = fs_info->extent_root; + int ret; + int slot; + struct extent_buffer *leaf; + struct btrfs_key key; + + while (1) { + ret = btrfs_next_item(extent_root, path); + if (ret < 0) + break; + if (ret) { + ret = 0; + break; + } + + slot = path->slots[0]; + leaf = path->nodes[0]; + btrfs_item_key_to_cpu(leaf, &key, slot); + + if (key.objectid != bytenr) + break; + if (key.type < BTRFS_TREE_BLOCK_REF_KEY) + continue; + if (key.type > BTRFS_SHARED_DATA_REF_KEY) + break; + + switch (key.type) { + case BTRFS_SHARED_BLOCK_REF_KEY: + ret = __add_prelim_ref(prefs, 0, NULL, + info_level + 1, key.offset, + bytenr, 1); + break; + case BTRFS_SHARED_DATA_REF_KEY: { + struct btrfs_shared_data_ref *sdref; + int count; + + sdref = btrfs_item_ptr(leaf, slot, + struct btrfs_shared_data_ref); + count = btrfs_shared_data_ref_count(leaf, sdref); + ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset, + bytenr, count); + break; + } + case BTRFS_TREE_BLOCK_REF_KEY: + ret = __add_prelim_ref(prefs, key.offset, NULL, + info_level + 1, 0, + bytenr, 1); + break; + case BTRFS_EXTENT_DATA_REF_KEY: { + struct btrfs_extent_data_ref *dref; + int count; + u64 root; + + dref = btrfs_item_ptr(leaf, slot, + struct btrfs_extent_data_ref); + count = btrfs_extent_data_ref_count(leaf, dref); + key.objectid = btrfs_extent_data_ref_objectid(leaf, + dref); + key.type = BTRFS_EXTENT_DATA_KEY; + key.offset = btrfs_extent_data_ref_offset(leaf, dref); + root = btrfs_extent_data_ref_root(leaf, dref); + ret = __add_prelim_ref(prefs, root, &key, 0, 0, + bytenr, count); + break; + } + default: + WARN_ON(1); + } + BUG_ON(ret); + } + + return ret; +} + +/* + * this adds all existing backrefs (inline backrefs, backrefs and delayed + * refs) for the given bytenr to the refs list, merges duplicates and resolves + * indirect refs to their parent bytenr. + * When roots are found, they're added to the roots list + * + * FIXME some caching might speed things up + */ +static int find_parent_nodes(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 bytenr, + u64 time_seq, struct ulist *refs, + struct ulist *roots, const u64 *extent_item_pos) +{ + struct btrfs_key key; + struct btrfs_path *path; + struct btrfs_delayed_ref_root *delayed_refs = NULL; + struct btrfs_delayed_ref_head *head; + int info_level = 0; + int ret; + int search_commit_root = (trans == BTRFS_BACKREF_SEARCH_COMMIT_ROOT); + struct list_head prefs_delayed; + struct list_head prefs; + struct __prelim_ref *ref; + + INIT_LIST_HEAD(&prefs); + INIT_LIST_HEAD(&prefs_delayed); + + key.objectid = bytenr; + key.type = BTRFS_EXTENT_ITEM_KEY; + key.offset = (u64)-1; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + path->search_commit_root = !!search_commit_root; + + /* + * grab both a lock on the path and a lock on the delayed ref head. + * We need both to get a consistent picture of how the refs look + * at a specified point in time + */ +again: + head = NULL; + + ret = btrfs_search_slot(trans, fs_info->extent_root, &key, path, 0, 0); + if (ret < 0) + goto out; + BUG_ON(ret == 0); + + if (trans != BTRFS_BACKREF_SEARCH_COMMIT_ROOT) { + /* + * look if there are updates for this ref queued and lock the + * head + */ + delayed_refs = &trans->transaction->delayed_refs; + spin_lock(&delayed_refs->lock); + head = btrfs_find_delayed_ref_head(trans, bytenr); + if (head) { + if (!mutex_trylock(&head->mutex)) { + atomic_inc(&head->node.refs); + spin_unlock(&delayed_refs->lock); + + btrfs_release_path(path); + + /* + * Mutex was contended, block until it's + * released and try again + */ + mutex_lock(&head->mutex); + mutex_unlock(&head->mutex); + btrfs_put_delayed_ref(&head->node); + goto again; + } + ret = __add_delayed_refs(head, time_seq, + &prefs_delayed); + mutex_unlock(&head->mutex); + if (ret) { + spin_unlock(&delayed_refs->lock); + goto out; + } + } + spin_unlock(&delayed_refs->lock); + } + + if (path->slots[0]) { + struct extent_buffer *leaf; + int slot; + + path->slots[0]--; + leaf = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(leaf, &key, slot); + if (key.objectid == bytenr && + key.type == BTRFS_EXTENT_ITEM_KEY) { + ret = __add_inline_refs(fs_info, path, bytenr, + &info_level, &prefs); + if (ret) + goto out; + ret = __add_keyed_refs(fs_info, path, bytenr, + info_level, &prefs); + if (ret) + goto out; + } + } + btrfs_release_path(path); + + list_splice_init(&prefs_delayed, &prefs); + + ret = __add_missing_keys(fs_info, &prefs); + if (ret) + goto out; + + ret = __merge_refs(&prefs, 1); + if (ret) + goto out; + + ret = __resolve_indirect_refs(fs_info, search_commit_root, time_seq, + &prefs, extent_item_pos); + if (ret) + goto out; + + ret = __merge_refs(&prefs, 2); + if (ret) + goto out; + + while (!list_empty(&prefs)) { + ref = list_first_entry(&prefs, struct __prelim_ref, list); + list_del(&ref->list); + if (ref->count < 0) + WARN_ON(1); + if (ref->count && ref->root_id && ref->parent == 0) { + /* no parent == root of tree */ + ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS); + BUG_ON(ret < 0); + } + if (ref->count && ref->parent) { + struct extent_inode_elem *eie = NULL; + if (extent_item_pos && !ref->inode_list) { + u32 bsz; + struct extent_buffer *eb; + bsz = btrfs_level_size(fs_info->extent_root, + info_level); + eb = read_tree_block(fs_info->extent_root, + ref->parent, bsz, 0); + BUG_ON(!eb); + ret = find_extent_in_eb(eb, bytenr, + *extent_item_pos, &eie); + ref->inode_list = eie; + free_extent_buffer(eb); + } + ret = ulist_add_merge(refs, ref->parent, + (unsigned long)ref->inode_list, + (unsigned long *)&eie, GFP_NOFS); + if (!ret && extent_item_pos) { + /* + * we've recorded that parent, so we must extend + * its inode list here + */ + BUG_ON(!eie); + while (eie->next) + eie = eie->next; + eie->next = ref->inode_list; + } + BUG_ON(ret < 0); + } + kfree(ref); + } + +out: + btrfs_free_path(path); + while (!list_empty(&prefs)) { + ref = list_first_entry(&prefs, struct __prelim_ref, list); + list_del(&ref->list); + kfree(ref); + } + while (!list_empty(&prefs_delayed)) { + ref = list_first_entry(&prefs_delayed, struct __prelim_ref, + list); + list_del(&ref->list); + kfree(ref); + } + + return ret; +} + +static void free_leaf_list(struct ulist *blocks) +{ + struct ulist_node *node = NULL; + struct extent_inode_elem *eie; + struct extent_inode_elem *eie_next; + struct ulist_iterator uiter; + + ULIST_ITER_INIT(&uiter); + while ((node = ulist_next(blocks, &uiter))) { + if (!node->aux) + continue; + eie = (struct extent_inode_elem *)node->aux; + for (; eie; eie = eie_next) { + eie_next = eie->next; + kfree(eie); + } + node->aux = 0; + } + + ulist_free(blocks); +} + +/* + * Finds all leafs with a reference to the specified combination of bytenr and + * offset. key_list_head will point to a list of corresponding keys (caller must + * free each list element). The leafs will be stored in the leafs ulist, which + * must be freed with ulist_free. + * + * returns 0 on success, <0 on error + */ +static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 bytenr, + u64 time_seq, struct ulist **leafs, + const u64 *extent_item_pos) +{ + struct ulist *tmp; + int ret; + + tmp = ulist_alloc(GFP_NOFS); + if (!tmp) + return -ENOMEM; + *leafs = ulist_alloc(GFP_NOFS); + if (!*leafs) { + ulist_free(tmp); + return -ENOMEM; + } + + ret = find_parent_nodes(trans, fs_info, bytenr, + time_seq, *leafs, tmp, extent_item_pos); + ulist_free(tmp); + + if (ret < 0 && ret != -ENOENT) { + free_leaf_list(*leafs); + return ret; + } + + return 0; +} + +/* + * walk all backrefs for a given extent to find all roots that reference this + * extent. Walking a backref means finding all extents that reference this + * extent and in turn walk the backrefs of those, too. Naturally this is a + * recursive process, but here it is implemented in an iterative fashion: We + * find all referencing extents for the extent in question and put them on a + * list. In turn, we find all referencing extents for those, further appending + * to the list. The way we iterate the list allows adding more elements after + * the current while iterating. The process stops when we reach the end of the + * list. Found roots are added to the roots list. + * + * returns 0 on success, < 0 on error. + */ +int btrfs_find_all_roots(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 bytenr, + u64 time_seq, struct ulist **roots) +{ + struct ulist *tmp; + struct ulist_node *node = NULL; + struct ulist_iterator uiter; + int ret; + + tmp = ulist_alloc(GFP_NOFS); + if (!tmp) + return -ENOMEM; + *roots = ulist_alloc(GFP_NOFS); + if (!*roots) { + ulist_free(tmp); + return -ENOMEM; + } + + ULIST_ITER_INIT(&uiter); + while (1) { + ret = find_parent_nodes(trans, fs_info, bytenr, + time_seq, tmp, *roots, NULL); + if (ret < 0 && ret != -ENOENT) { + ulist_free(tmp); + ulist_free(*roots); + return ret; + } + node = ulist_next(tmp, &uiter); + if (!node) + break; + bytenr = node->val; + } + + ulist_free(tmp); + return 0; +} + + +static int __inode_info(u64 inum, u64 ioff, u8 key_type, + struct btrfs_root *fs_root, struct btrfs_path *path, + struct btrfs_key *found_key) +{ + int ret; + struct btrfs_key key; + struct extent_buffer *eb; + + key.type = key_type; + key.objectid = inum; + key.offset = ioff; + + ret = btrfs_search_slot(NULL, fs_root, &key, path, 0, 0); + if (ret < 0) + return ret; + + eb = path->nodes[0]; + if (ret && path->slots[0] >= btrfs_header_nritems(eb)) { + ret = btrfs_next_leaf(fs_root, path); + if (ret) + return ret; + eb = path->nodes[0]; + } + + btrfs_item_key_to_cpu(eb, found_key, path->slots[0]); + if (found_key->type != key.type || found_key->objectid != key.objectid) + return 1; + + return 0; +} + +/* + * this makes the path point to (inum INODE_ITEM ioff) + */ +int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root, + struct btrfs_path *path) +{ + struct btrfs_key key; + return __inode_info(inum, ioff, BTRFS_INODE_ITEM_KEY, fs_root, path, + &key); +} + +static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root, + struct btrfs_path *path, + struct btrfs_key *found_key) +{ + return __inode_info(inum, ioff, BTRFS_INODE_REF_KEY, fs_root, path, + found_key); +} + +/* + * this iterates to turn a btrfs_inode_ref into a full filesystem path. elements + * of the path are separated by '/' and the path is guaranteed to be + * 0-terminated. the path is only given within the current file system. + * Therefore, it never starts with a '/'. the caller is responsible to provide + * "size" bytes in "dest". the dest buffer will be filled backwards. finally, + * the start point of the resulting string is returned. this pointer is within + * dest, normally. + * in case the path buffer would overflow, the pointer is decremented further + * as if output was written to the buffer, though no more output is actually + * generated. that way, the caller can determine how much space would be + * required for the path to fit into the buffer. in that case, the returned + * value will be smaller than dest. callers must check this! + */ +char *btrfs_iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path, + struct btrfs_inode_ref *iref, + struct extent_buffer *eb_in, u64 parent, + char *dest, u32 size) +{ + u32 len; + int slot; + u64 next_inum; + int ret; + s64 bytes_left = size - 1; + struct extent_buffer *eb = eb_in; + struct btrfs_key found_key; + int leave_spinning = path->leave_spinning; + + if (bytes_left >= 0) + dest[bytes_left] = '\0'; + + path->leave_spinning = 1; + while (1) { + len = btrfs_inode_ref_name_len(eb, iref); + bytes_left -= len; + if (bytes_left >= 0) + read_extent_buffer(eb, dest + bytes_left, + (unsigned long)(iref + 1), len); + if (eb != eb_in) { + btrfs_tree_read_unlock_blocking(eb); + free_extent_buffer(eb); + } + ret = inode_ref_info(parent, 0, fs_root, path, &found_key); + if (ret > 0) + ret = -ENOENT; + if (ret) + break; + next_inum = found_key.offset; + + /* regular exit ahead */ + if (parent == next_inum) + break; + + slot = path->slots[0]; + eb = path->nodes[0]; + /* make sure we can use eb after releasing the path */ + if (eb != eb_in) { + atomic_inc(&eb->refs); + btrfs_tree_read_lock(eb); + btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); + } + btrfs_release_path(path); + + iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref); + parent = next_inum; + --bytes_left; + if (bytes_left >= 0) + dest[bytes_left] = '/'; + } + + btrfs_release_path(path); + path->leave_spinning = leave_spinning; + + if (ret) + return ERR_PTR(ret); + + return dest + bytes_left; +} + +/* + * this makes the path point to (logical EXTENT_ITEM *) + * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for + * tree blocks and <0 on error. + */ +int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical, + struct btrfs_path *path, struct btrfs_key *found_key) +{ + int ret; + u64 flags; + u32 item_size; + struct extent_buffer *eb; + struct btrfs_extent_item *ei; + struct btrfs_key key; + + key.type = BTRFS_EXTENT_ITEM_KEY; + key.objectid = logical; + key.offset = (u64)-1; + + ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0); + if (ret < 0) + return ret; + ret = btrfs_previous_item(fs_info->extent_root, path, + 0, BTRFS_EXTENT_ITEM_KEY); + if (ret < 0) + return ret; + + btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]); + if (found_key->type != BTRFS_EXTENT_ITEM_KEY || + found_key->objectid > logical || + found_key->objectid + found_key->offset <= logical) { + pr_debug("logical %llu is not within any extent\n", + (unsigned long long)logical); + return -ENOENT; + } + + eb = path->nodes[0]; + item_size = btrfs_item_size_nr(eb, path->slots[0]); + BUG_ON(item_size < sizeof(*ei)); + + ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item); + flags = btrfs_extent_flags(eb, ei); + + pr_debug("logical %llu is at position %llu within the extent (%llu " + "EXTENT_ITEM %llu) flags %#llx size %u\n", + (unsigned long long)logical, + (unsigned long long)(logical - found_key->objectid), + (unsigned long long)found_key->objectid, + (unsigned long long)found_key->offset, + (unsigned long long)flags, item_size); + if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) + return BTRFS_EXTENT_FLAG_TREE_BLOCK; + if (flags & BTRFS_EXTENT_FLAG_DATA) + return BTRFS_EXTENT_FLAG_DATA; + + return -EIO; +} + +/* + * helper function to iterate extent inline refs. ptr must point to a 0 value + * for the first call and may be modified. it is used to track state. + * if more refs exist, 0 is returned and the next call to + * __get_extent_inline_ref must pass the modified ptr parameter to get the + * next ref. after the last ref was processed, 1 is returned. + * returns <0 on error + */ +static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb, + struct btrfs_extent_item *ei, u32 item_size, + struct btrfs_extent_inline_ref **out_eiref, + int *out_type) +{ + unsigned long end; + u64 flags; + struct btrfs_tree_block_info *info; + + if (!*ptr) { + /* first call */ + flags = btrfs_extent_flags(eb, ei); + if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + info = (struct btrfs_tree_block_info *)(ei + 1); + *out_eiref = + (struct btrfs_extent_inline_ref *)(info + 1); + } else { + *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1); + } + *ptr = (unsigned long)*out_eiref; + if ((void *)*ptr >= (void *)ei + item_size) + return -ENOENT; + } + + end = (unsigned long)ei + item_size; + *out_eiref = (struct btrfs_extent_inline_ref *)*ptr; + *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref); + + *ptr += btrfs_extent_inline_ref_size(*out_type); + WARN_ON(*ptr > end); + if (*ptr == end) + return 1; /* last */ + + return 0; +} + +/* + * reads the tree block backref for an extent. tree level and root are returned + * through out_level and out_root. ptr must point to a 0 value for the first + * call and may be modified (see __get_extent_inline_ref comment). + * returns 0 if data was provided, 1 if there was no more data to provide or + * <0 on error. + */ +int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb, + struct btrfs_extent_item *ei, u32 item_size, + u64 *out_root, u8 *out_level) +{ + int ret; + int type; + struct btrfs_tree_block_info *info; + struct btrfs_extent_inline_ref *eiref; + + if (*ptr == (unsigned long)-1) + return 1; + + while (1) { + ret = __get_extent_inline_ref(ptr, eb, ei, item_size, + &eiref, &type); + if (ret < 0) + return ret; + + if (type == BTRFS_TREE_BLOCK_REF_KEY || + type == BTRFS_SHARED_BLOCK_REF_KEY) + break; + + if (ret == 1) + return 1; + } + + /* we can treat both ref types equally here */ + info = (struct btrfs_tree_block_info *)(ei + 1); + *out_root = btrfs_extent_inline_ref_offset(eb, eiref); + *out_level = btrfs_tree_block_level(eb, info); + + if (ret == 1) + *ptr = (unsigned long)-1; + + return 0; +} + +static int iterate_leaf_refs(struct extent_inode_elem *inode_list, + u64 root, u64 extent_item_objectid, + iterate_extent_inodes_t *iterate, void *ctx) +{ + struct extent_inode_elem *eie; + int ret = 0; + + for (eie = inode_list; eie; eie = eie->next) { + pr_debug("ref for %llu resolved, key (%llu EXTEND_DATA %llu), " + "root %llu\n", extent_item_objectid, + eie->inum, eie->offset, root); + ret = iterate(eie->inum, eie->offset, root, ctx); + if (ret) { + pr_debug("stopping iteration for %llu due to ret=%d\n", + extent_item_objectid, ret); + break; + } + } + + return ret; +} + +/* + * calls iterate() for every inode that references the extent identified by + * the given parameters. + * when the iterator function returns a non-zero value, iteration stops. + */ +int iterate_extent_inodes(struct btrfs_fs_info *fs_info, + u64 extent_item_objectid, u64 extent_item_pos, + int search_commit_root, + iterate_extent_inodes_t *iterate, void *ctx) +{ + int ret; + struct list_head data_refs = LIST_HEAD_INIT(data_refs); + struct list_head shared_refs = LIST_HEAD_INIT(shared_refs); + struct btrfs_trans_handle *trans; + struct ulist *refs = NULL; + struct ulist *roots = NULL; + struct ulist_node *ref_node = NULL; + struct ulist_node *root_node = NULL; + struct seq_list tree_mod_seq_elem = {}; + struct ulist_iterator ref_uiter; + struct ulist_iterator root_uiter; + + pr_debug("resolving all inodes for extent %llu\n", + extent_item_objectid); + + if (search_commit_root) { + trans = BTRFS_BACKREF_SEARCH_COMMIT_ROOT; + } else { + trans = btrfs_join_transaction(fs_info->extent_root); + if (IS_ERR(trans)) + return PTR_ERR(trans); + btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem); + } + + ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid, + tree_mod_seq_elem.seq, &refs, + &extent_item_pos); + if (ret) + goto out; + + ULIST_ITER_INIT(&ref_uiter); + while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) { + ret = btrfs_find_all_roots(trans, fs_info, ref_node->val, + tree_mod_seq_elem.seq, &roots); + if (ret) + break; + ULIST_ITER_INIT(&root_uiter); + while (!ret && (root_node = ulist_next(roots, &root_uiter))) { + pr_debug("root %llu references leaf %llu, data list " + "%#lx\n", root_node->val, ref_node->val, + ref_node->aux); + ret = iterate_leaf_refs( + (struct extent_inode_elem *)ref_node->aux, + root_node->val, extent_item_objectid, + iterate, ctx); + } + ulist_free(roots); + roots = NULL; + } + + free_leaf_list(refs); + ulist_free(roots); +out: + if (!search_commit_root) { + btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem); + btrfs_end_transaction(trans, fs_info->extent_root); + } + + return ret; +} + +int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info, + struct btrfs_path *path, + iterate_extent_inodes_t *iterate, void *ctx) +{ + int ret; + u64 extent_item_pos; + struct btrfs_key found_key; + int search_commit_root = path->search_commit_root; + + ret = extent_from_logical(fs_info, logical, path, + &found_key); + btrfs_release_path(path); + if (ret < 0) + return ret; + if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK) + return -EINVAL; + + extent_item_pos = logical - found_key.objectid; + ret = iterate_extent_inodes(fs_info, found_key.objectid, + extent_item_pos, search_commit_root, + iterate, ctx); + + return ret; +} + +static int iterate_irefs(u64 inum, struct btrfs_root *fs_root, + struct btrfs_path *path, + iterate_irefs_t *iterate, void *ctx) +{ + int ret = 0; + int slot; + u32 cur; + u32 len; + u32 name_len; + u64 parent = 0; + int found = 0; + struct extent_buffer *eb; + struct btrfs_item *item; + struct btrfs_inode_ref *iref; + struct btrfs_key found_key; + + while (!ret) { + path->leave_spinning = 1; + ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path, + &found_key); + if (ret < 0) + break; + if (ret) { + ret = found ? 0 : -ENOENT; + break; + } + ++found; + + parent = found_key.offset; + slot = path->slots[0]; + eb = path->nodes[0]; + /* make sure we can use eb after releasing the path */ + atomic_inc(&eb->refs); + btrfs_tree_read_lock(eb); + btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); + btrfs_release_path(path); + + item = btrfs_item_nr(eb, slot); + iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref); + + for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) { + name_len = btrfs_inode_ref_name_len(eb, iref); + /* path must be released before calling iterate()! */ + pr_debug("following ref at offset %u for inode %llu in " + "tree %llu\n", cur, + (unsigned long long)found_key.objectid, + (unsigned long long)fs_root->objectid); + ret = iterate(parent, iref, eb, ctx); + if (ret) + break; + len = sizeof(*iref) + name_len; + iref = (struct btrfs_inode_ref *)((char *)iref + len); + } + btrfs_tree_read_unlock_blocking(eb); + free_extent_buffer(eb); + } + + btrfs_release_path(path); + + return ret; +} + +/* + * returns 0 if the path could be dumped (probably truncated) + * returns <0 in case of an error + */ +static int inode_to_path(u64 inum, struct btrfs_inode_ref *iref, + struct extent_buffer *eb, void *ctx) +{ + struct inode_fs_paths *ipath = ctx; + char *fspath; + char *fspath_min; + int i = ipath->fspath->elem_cnt; + const int s_ptr = sizeof(char *); + u32 bytes_left; + + bytes_left = ipath->fspath->bytes_left > s_ptr ? + ipath->fspath->bytes_left - s_ptr : 0; + + fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr; + fspath = btrfs_iref_to_path(ipath->fs_root, ipath->btrfs_path, iref, eb, + inum, fspath_min, bytes_left); + if (IS_ERR(fspath)) + return PTR_ERR(fspath); + + if (fspath > fspath_min) { + pr_debug("path resolved: %s\n", fspath); + ipath->fspath->val[i] = (u64)(unsigned long)fspath; + ++ipath->fspath->elem_cnt; + ipath->fspath->bytes_left = fspath - fspath_min; + } else { + pr_debug("missed path, not enough space. missing bytes: %lu, " + "constructed so far: %s\n", + (unsigned long)(fspath_min - fspath), fspath_min); + ++ipath->fspath->elem_missed; + ipath->fspath->bytes_missing += fspath_min - fspath; + ipath->fspath->bytes_left = 0; + } + + return 0; +} + +/* + * this dumps all file system paths to the inode into the ipath struct, provided + * is has been created large enough. each path is zero-terminated and accessed + * from ipath->fspath->val[i]. + * when it returns, there are ipath->fspath->elem_cnt number of paths available + * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the + * number of missed paths in recored in ipath->fspath->elem_missed, otherwise, + * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would + * have been needed to return all paths. + */ +int paths_from_inode(u64 inum, struct inode_fs_paths *ipath) +{ + return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path, + inode_to_path, ipath); +} + +struct btrfs_data_container *init_data_container(u32 total_bytes) +{ + struct btrfs_data_container *data; + size_t alloc_bytes; + + alloc_bytes = max_t(size_t, total_bytes, sizeof(*data)); + data = kmalloc(alloc_bytes, GFP_NOFS); + if (!data) + return ERR_PTR(-ENOMEM); + + if (total_bytes >= sizeof(*data)) { + data->bytes_left = total_bytes - sizeof(*data); + data->bytes_missing = 0; + } else { + data->bytes_missing = sizeof(*data) - total_bytes; + data->bytes_left = 0; + } + + data->elem_cnt = 0; + data->elem_missed = 0; + + return data; +} + +/* + * allocates space to return multiple file system paths for an inode. + * total_bytes to allocate are passed, note that space usable for actual path + * information will be total_bytes - sizeof(struct inode_fs_paths). + * the returned pointer must be freed with free_ipath() in the end. + */ +struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root, + struct btrfs_path *path) +{ + struct inode_fs_paths *ifp; + struct btrfs_data_container *fspath; + + fspath = init_data_container(total_bytes); + if (IS_ERR(fspath)) + return (void *)fspath; + + ifp = kmalloc(sizeof(*ifp), GFP_NOFS); + if (!ifp) { + kfree(fspath); + return ERR_PTR(-ENOMEM); + } + + ifp->btrfs_path = path; + ifp->fspath = fspath; + ifp->fs_root = fs_root; + + return ifp; +} + +void free_ipath(struct inode_fs_paths *ipath) +{ + if (!ipath) + return; + kfree(ipath->fspath); + kfree(ipath); +} diff --git a/fs/btrfs/backref.h b/fs/btrfs/backref.h new file mode 100644 index 000000000000..032f4dc7eab8 --- /dev/null +++ b/fs/btrfs/backref.h @@ -0,0 +1,72 @@ +/* + * Copyright (C) 2011 STRATO. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#ifndef __BTRFS_BACKREF__ +#define __BTRFS_BACKREF__ + +#include "ioctl.h" +#include "ulist.h" +#include "extent_io.h" + +#define BTRFS_BACKREF_SEARCH_COMMIT_ROOT ((struct btrfs_trans_handle *)0) + +struct inode_fs_paths { + struct btrfs_path *btrfs_path; + struct btrfs_root *fs_root; + struct btrfs_data_container *fspath; +}; + +typedef int (iterate_extent_inodes_t)(u64 inum, u64 offset, u64 root, + void *ctx); +typedef int (iterate_irefs_t)(u64 parent, struct btrfs_inode_ref *iref, + struct extent_buffer *eb, void *ctx); + +int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root, + struct btrfs_path *path); + +int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical, + struct btrfs_path *path, struct btrfs_key *found_key); + +int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb, + struct btrfs_extent_item *ei, u32 item_size, + u64 *out_root, u8 *out_level); + +int iterate_extent_inodes(struct btrfs_fs_info *fs_info, + u64 extent_item_objectid, + u64 extent_offset, int search_commit_root, + iterate_extent_inodes_t *iterate, void *ctx); + +int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info, + struct btrfs_path *path, + iterate_extent_inodes_t *iterate, void *ctx); + +int paths_from_inode(u64 inum, struct inode_fs_paths *ipath); + +int btrfs_find_all_roots(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 bytenr, + u64 time_seq, struct ulist **roots); +char *btrfs_iref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path, + struct btrfs_inode_ref *iref, struct extent_buffer *eb, + u64 parent, char *dest, u32 size); + +struct btrfs_data_container *init_data_container(u32 total_bytes); +struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root, + struct btrfs_path *path); +void free_ipath(struct inode_fs_paths *ipath); + +#endif diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index d9f99a16edd6..5b2ad6bc4fe7 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -24,6 +24,21 @@ #include "ordered-data.h" #include "delayed-inode.h" +/* + * ordered_data_close is set by truncate when a file that used + * to have good data has been truncated to zero. When it is set + * the btrfs file release call will add this inode to the + * ordered operations list so that we make sure to flush out any + * new data the application may have written before commit. + */ +#define BTRFS_INODE_ORDERED_DATA_CLOSE 0 +#define BTRFS_INODE_ORPHAN_META_RESERVED 1 +#define BTRFS_INODE_DUMMY 2 +#define BTRFS_INODE_IN_DEFRAG 3 +#define BTRFS_INODE_DELALLOC_META_RESERVED 4 +#define BTRFS_INODE_HAS_ORPHAN_ITEM 5 +#define BTRFS_INODE_HAS_ASYNC_EXTENT 6 + /* in memory btrfs inode */ struct btrfs_inode { /* which subvolume this inode belongs to */ @@ -51,12 +66,12 @@ struct btrfs_inode { /* held while logging the inode in tree-log.c */ struct mutex log_mutex; + /* held while doing delalloc reservations */ + struct mutex delalloc_mutex; + /* used to order data wrt metadata */ struct btrfs_ordered_inode_tree ordered_tree; - /* for keeping track of orphaned inodes */ - struct list_head i_orphan; - /* list of all the delalloc inodes in the FS. There are times we need * to write all the delalloc pages to disk, and this list is used * to walk them all. @@ -72,17 +87,13 @@ struct btrfs_inode { /* node for the red-black tree that links inodes in subvolume root */ struct rb_node rb_node; - /* the space_info for where this inode's data allocations are done */ - struct btrfs_space_info *space_info; + unsigned long runtime_flags; /* full 64 bit generation number, struct vfs_inode doesn't have a big * enough field for this. */ u64 generation; - /* sequence number for NFS changes */ - u64 sequence; - /* * transid of the trans_handle that last modified this inode */ @@ -103,11 +114,6 @@ struct btrfs_inode { */ u64 delalloc_bytes; - /* total number of bytes that may be used for this inode for - * delalloc - */ - u64 reserved_bytes; - /* * the size of the file stored in the metadata on disk. data=ordered * means the in-memory i_size might be larger than the size on disk @@ -115,9 +121,6 @@ struct btrfs_inode { */ u64 disk_i_size; - /* flags field from the on disk inode */ - u32 flags; - /* * if this is a directory then index_cnt is the counter for the index * number for new files that are created @@ -132,6 +135,15 @@ struct btrfs_inode { u64 last_unlink_trans; /* + * Number of bytes outstanding that are going to need csums. This is + * used in ENOSPC accounting. + */ + u64 csum_bytes; + + /* flags field from the on disk inode */ + u32 flags; + + /* * Counters to keep track of the number of extent item's we may use due * to delalloc and such. outstanding_extents is the number of extent * items we think we'll end up using, and reserved_extents is the number @@ -141,24 +153,9 @@ struct btrfs_inode { unsigned reserved_extents; /* - * ordered_data_close is set by truncate when a file that used - * to have good data has been truncated to zero. When it is set - * the btrfs file release call will add this inode to the - * ordered operations list so that we make sure to flush out any - * new data the application may have written before commit. - * - * yes, its silly to have a single bitflag, but we might grow more - * of these. - */ - unsigned ordered_data_close:1; - unsigned orphan_meta_reserved:1; - unsigned dummy_inode:1; - unsigned in_defrag:1; - - /* * always compress this one file */ - unsigned force_compress:4; + unsigned force_compress; struct btrfs_delayed_node *delayed_node; @@ -191,13 +188,29 @@ static inline void btrfs_i_size_write(struct inode *inode, u64 size) BTRFS_I(inode)->disk_i_size = size; } -static inline bool btrfs_is_free_space_inode(struct btrfs_root *root, - struct inode *inode) +static inline bool btrfs_is_free_space_inode(struct inode *inode) { - if (root == root->fs_info->tree_root || - BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) + struct btrfs_root *root = BTRFS_I(inode)->root; + + if (root == root->fs_info->tree_root && + btrfs_ino(inode) != BTRFS_BTREE_INODE_OBJECTID) + return true; + if (BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) return true; return false; } +static inline int btrfs_inode_in_log(struct inode *inode, u64 generation) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + int ret = 0; + + mutex_lock(&root->log_mutex); + if (BTRFS_I(inode)->logged_trans == generation && + BTRFS_I(inode)->last_sub_trans <= root->last_log_commit) + ret = 1; + mutex_unlock(&root->log_mutex); + return ret; +} + #endif diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c new file mode 100644 index 000000000000..9197e2e33407 --- /dev/null +++ b/fs/btrfs/check-integrity.c @@ -0,0 +1,3359 @@ +/* + * Copyright (C) STRATO AG 2011. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +/* + * This module can be used to catch cases when the btrfs kernel + * code executes write requests to the disk that bring the file + * system in an inconsistent state. In such a state, a power-loss + * or kernel panic event would cause that the data on disk is + * lost or at least damaged. + * + * Code is added that examines all block write requests during + * runtime (including writes of the super block). Three rules + * are verified and an error is printed on violation of the + * rules: + * 1. It is not allowed to write a disk block which is + * currently referenced by the super block (either directly + * or indirectly). + * 2. When a super block is written, it is verified that all + * referenced (directly or indirectly) blocks fulfill the + * following requirements: + * 2a. All referenced blocks have either been present when + * the file system was mounted, (i.e., they have been + * referenced by the super block) or they have been + * written since then and the write completion callback + * was called and a FLUSH request to the device where + * these blocks are located was received and completed. + * 2b. All referenced blocks need to have a generation + * number which is equal to the parent's number. + * + * One issue that was found using this module was that the log + * tree on disk became temporarily corrupted because disk blocks + * that had been in use for the log tree had been freed and + * reused too early, while being referenced by the written super + * block. + * + * The search term in the kernel log that can be used to filter + * on the existence of detected integrity issues is + * "btrfs: attempt". + * + * The integrity check is enabled via mount options. These + * mount options are only supported if the integrity check + * tool is compiled by defining BTRFS_FS_CHECK_INTEGRITY. + * + * Example #1, apply integrity checks to all metadata: + * mount /dev/sdb1 /mnt -o check_int + * + * Example #2, apply integrity checks to all metadata and + * to data extents: + * mount /dev/sdb1 /mnt -o check_int_data + * + * Example #3, apply integrity checks to all metadata and dump + * the tree that the super block references to kernel messages + * each time after a super block was written: + * mount /dev/sdb1 /mnt -o check_int,check_int_print_mask=263 + * + * If the integrity check tool is included and activated in + * the mount options, plenty of kernel memory is used, and + * plenty of additional CPU cycles are spent. Enabling this + * functionality is not intended for normal use. In most + * cases, unless you are a btrfs developer who needs to verify + * the integrity of (super)-block write requests, do not + * enable the config option BTRFS_FS_CHECK_INTEGRITY to + * include and compile the integrity check tool. + */ + +#include <linux/sched.h> +#include <linux/slab.h> +#include <linux/buffer_head.h> +#include <linux/mutex.h> +#include <linux/crc32c.h> +#include <linux/genhd.h> +#include <linux/blkdev.h> +#include "ctree.h" +#include "disk-io.h" +#include "transaction.h" +#include "extent_io.h" +#include "volumes.h" +#include "print-tree.h" +#include "locking.h" +#include "check-integrity.h" +#include "rcu-string.h" + +#define BTRFSIC_BLOCK_HASHTABLE_SIZE 0x10000 +#define BTRFSIC_BLOCK_LINK_HASHTABLE_SIZE 0x10000 +#define BTRFSIC_DEV2STATE_HASHTABLE_SIZE 0x100 +#define BTRFSIC_BLOCK_MAGIC_NUMBER 0x14491051 +#define BTRFSIC_BLOCK_LINK_MAGIC_NUMBER 0x11070807 +#define BTRFSIC_DEV2STATE_MAGIC_NUMBER 0x20111530 +#define BTRFSIC_BLOCK_STACK_FRAME_MAGIC_NUMBER 20111300 +#define BTRFSIC_TREE_DUMP_MAX_INDENT_LEVEL (200 - 6) /* in characters, + * excluding " [...]" */ +#define BTRFSIC_GENERATION_UNKNOWN ((u64)-1) + +/* + * The definition of the bitmask fields for the print_mask. + * They are specified with the mount option check_integrity_print_mask. + */ +#define BTRFSIC_PRINT_MASK_SUPERBLOCK_WRITE 0x00000001 +#define BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION 0x00000002 +#define BTRFSIC_PRINT_MASK_TREE_AFTER_SB_WRITE 0x00000004 +#define BTRFSIC_PRINT_MASK_TREE_BEFORE_SB_WRITE 0x00000008 +#define BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH 0x00000010 +#define BTRFSIC_PRINT_MASK_END_IO_BIO_BH 0x00000020 +#define BTRFSIC_PRINT_MASK_VERBOSE 0x00000040 +#define BTRFSIC_PRINT_MASK_VERY_VERBOSE 0x00000080 +#define BTRFSIC_PRINT_MASK_INITIAL_TREE 0x00000100 +#define BTRFSIC_PRINT_MASK_INITIAL_ALL_TREES 0x00000200 +#define BTRFSIC_PRINT_MASK_INITIAL_DATABASE 0x00000400 +#define BTRFSIC_PRINT_MASK_NUM_COPIES 0x00000800 +#define BTRFSIC_PRINT_MASK_TREE_WITH_ALL_MIRRORS 0x00001000 + +struct btrfsic_dev_state; +struct btrfsic_state; + +struct btrfsic_block { + u32 magic_num; /* only used for debug purposes */ + unsigned int is_metadata:1; /* if it is meta-data, not data-data */ + unsigned int is_superblock:1; /* if it is one of the superblocks */ + unsigned int is_iodone:1; /* if is done by lower subsystem */ + unsigned int iodone_w_error:1; /* error was indicated to endio */ + unsigned int never_written:1; /* block was added because it was + * referenced, not because it was + * written */ + unsigned int mirror_num:2; /* large enough to hold + * BTRFS_SUPER_MIRROR_MAX */ + struct btrfsic_dev_state *dev_state; + u64 dev_bytenr; /* key, physical byte num on disk */ + u64 logical_bytenr; /* logical byte num on disk */ + u64 generation; + struct btrfs_disk_key disk_key; /* extra info to print in case of + * issues, will not always be correct */ + struct list_head collision_resolving_node; /* list node */ + struct list_head all_blocks_node; /* list node */ + + /* the following two lists contain block_link items */ + struct list_head ref_to_list; /* list */ + struct list_head ref_from_list; /* list */ + struct btrfsic_block *next_in_same_bio; + void *orig_bio_bh_private; + union { + bio_end_io_t *bio; + bh_end_io_t *bh; + } orig_bio_bh_end_io; + int submit_bio_bh_rw; + u64 flush_gen; /* only valid if !never_written */ +}; + +/* + * Elements of this type are allocated dynamically and required because + * each block object can refer to and can be ref from multiple blocks. + * The key to lookup them in the hashtable is the dev_bytenr of + * the block ref to plus the one from the block refered from. + * The fact that they are searchable via a hashtable and that a + * ref_cnt is maintained is not required for the btrfs integrity + * check algorithm itself, it is only used to make the output more + * beautiful in case that an error is detected (an error is defined + * as a write operation to a block while that block is still referenced). + */ +struct btrfsic_block_link { + u32 magic_num; /* only used for debug purposes */ + u32 ref_cnt; + struct list_head node_ref_to; /* list node */ + struct list_head node_ref_from; /* list node */ + struct list_head collision_resolving_node; /* list node */ + struct btrfsic_block *block_ref_to; + struct btrfsic_block *block_ref_from; + u64 parent_generation; +}; + +struct btrfsic_dev_state { + u32 magic_num; /* only used for debug purposes */ + struct block_device *bdev; + struct btrfsic_state *state; + struct list_head collision_resolving_node; /* list node */ + struct btrfsic_block dummy_block_for_bio_bh_flush; + u64 last_flush_gen; + char name[BDEVNAME_SIZE]; +}; + +struct btrfsic_block_hashtable { + struct list_head table[BTRFSIC_BLOCK_HASHTABLE_SIZE]; +}; + +struct btrfsic_block_link_hashtable { + struct list_head table[BTRFSIC_BLOCK_LINK_HASHTABLE_SIZE]; +}; + +struct btrfsic_dev_state_hashtable { + struct list_head table[BTRFSIC_DEV2STATE_HASHTABLE_SIZE]; +}; + +struct btrfsic_block_data_ctx { + u64 start; /* virtual bytenr */ + u64 dev_bytenr; /* physical bytenr on device */ + u32 len; + struct btrfsic_dev_state *dev; + char **datav; + struct page **pagev; + void *mem_to_free; +}; + +/* This structure is used to implement recursion without occupying + * any stack space, refer to btrfsic_process_metablock() */ +struct btrfsic_stack_frame { + u32 magic; + u32 nr; + int error; + int i; + int limit_nesting; + int num_copies; + int mirror_num; + struct btrfsic_block *block; + struct btrfsic_block_data_ctx *block_ctx; + struct btrfsic_block *next_block; + struct btrfsic_block_data_ctx next_block_ctx; + struct btrfs_header *hdr; + struct btrfsic_stack_frame *prev; +}; + +/* Some state per mounted filesystem */ +struct btrfsic_state { + u32 print_mask; + int include_extent_data; + int csum_size; + struct list_head all_blocks_list; + struct btrfsic_block_hashtable block_hashtable; + struct btrfsic_block_link_hashtable block_link_hashtable; + struct btrfs_root *root; + u64 max_superblock_generation; + struct btrfsic_block *latest_superblock; + u32 metablock_size; + u32 datablock_size; +}; + +static void btrfsic_block_init(struct btrfsic_block *b); +static struct btrfsic_block *btrfsic_block_alloc(void); +static void btrfsic_block_free(struct btrfsic_block *b); +static void btrfsic_block_link_init(struct btrfsic_block_link *n); +static struct btrfsic_block_link *btrfsic_block_link_alloc(void); +static void btrfsic_block_link_free(struct btrfsic_block_link *n); +static void btrfsic_dev_state_init(struct btrfsic_dev_state *ds); +static struct btrfsic_dev_state *btrfsic_dev_state_alloc(void); +static void btrfsic_dev_state_free(struct btrfsic_dev_state *ds); +static void btrfsic_block_hashtable_init(struct btrfsic_block_hashtable *h); +static void btrfsic_block_hashtable_add(struct btrfsic_block *b, + struct btrfsic_block_hashtable *h); +static void btrfsic_block_hashtable_remove(struct btrfsic_block *b); +static struct btrfsic_block *btrfsic_block_hashtable_lookup( + struct block_device *bdev, + u64 dev_bytenr, + struct btrfsic_block_hashtable *h); +static void btrfsic_block_link_hashtable_init( + struct btrfsic_block_link_hashtable *h); +static void btrfsic_block_link_hashtable_add( + struct btrfsic_block_link *l, + struct btrfsic_block_link_hashtable *h); +static void btrfsic_block_link_hashtable_remove(struct btrfsic_block_link *l); +static struct btrfsic_block_link *btrfsic_block_link_hashtable_lookup( + struct block_device *bdev_ref_to, + u64 dev_bytenr_ref_to, + struct block_device *bdev_ref_from, + u64 dev_bytenr_ref_from, + struct btrfsic_block_link_hashtable *h); +static void btrfsic_dev_state_hashtable_init( + struct btrfsic_dev_state_hashtable *h); +static void btrfsic_dev_state_hashtable_add( + struct btrfsic_dev_state *ds, + struct btrfsic_dev_state_hashtable *h); +static void btrfsic_dev_state_hashtable_remove(struct btrfsic_dev_state *ds); +static struct btrfsic_dev_state *btrfsic_dev_state_hashtable_lookup( + struct block_device *bdev, + struct btrfsic_dev_state_hashtable *h); +static struct btrfsic_stack_frame *btrfsic_stack_frame_alloc(void); +static void btrfsic_stack_frame_free(struct btrfsic_stack_frame *sf); +static int btrfsic_process_superblock(struct btrfsic_state *state, + struct btrfs_fs_devices *fs_devices); +static int btrfsic_process_metablock(struct btrfsic_state *state, + struct btrfsic_block *block, + struct btrfsic_block_data_ctx *block_ctx, + int limit_nesting, int force_iodone_flag); +static void btrfsic_read_from_block_data( + struct btrfsic_block_data_ctx *block_ctx, + void *dst, u32 offset, size_t len); +static int btrfsic_create_link_to_next_block( + struct btrfsic_state *state, + struct btrfsic_block *block, + struct btrfsic_block_data_ctx + *block_ctx, u64 next_bytenr, + int limit_nesting, + struct btrfsic_block_data_ctx *next_block_ctx, + struct btrfsic_block **next_blockp, + int force_iodone_flag, + int *num_copiesp, int *mirror_nump, + struct btrfs_disk_key *disk_key, + u64 parent_generation); +static int btrfsic_handle_extent_data(struct btrfsic_state *state, + struct btrfsic_block *block, + struct btrfsic_block_data_ctx *block_ctx, + u32 item_offset, int force_iodone_flag); +static int btrfsic_map_block(struct btrfsic_state *state, u64 bytenr, u32 len, + struct btrfsic_block_data_ctx *block_ctx_out, + int mirror_num); +static int btrfsic_map_superblock(struct btrfsic_state *state, u64 bytenr, + u32 len, struct block_device *bdev, + struct btrfsic_block_data_ctx *block_ctx_out); +static void btrfsic_release_block_ctx(struct btrfsic_block_data_ctx *block_ctx); +static int btrfsic_read_block(struct btrfsic_state *state, + struct btrfsic_block_data_ctx *block_ctx); +static void btrfsic_dump_database(struct btrfsic_state *state); +static void btrfsic_complete_bio_end_io(struct bio *bio, int err); +static int btrfsic_test_for_metadata(struct btrfsic_state *state, + char **datav, unsigned int num_pages); +static void btrfsic_process_written_block(struct btrfsic_dev_state *dev_state, + u64 dev_bytenr, char **mapped_datav, + unsigned int num_pages, + struct bio *bio, int *bio_is_patched, + struct buffer_head *bh, + int submit_bio_bh_rw); +static int btrfsic_process_written_superblock( + struct btrfsic_state *state, + struct btrfsic_block *const block, + struct btrfs_super_block *const super_hdr); +static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status); +static void btrfsic_bh_end_io(struct buffer_head *bh, int uptodate); +static int btrfsic_is_block_ref_by_superblock(const struct btrfsic_state *state, + const struct btrfsic_block *block, + int recursion_level); +static int btrfsic_check_all_ref_blocks(struct btrfsic_state *state, + struct btrfsic_block *const block, + int recursion_level); +static void btrfsic_print_add_link(const struct btrfsic_state *state, + const struct btrfsic_block_link *l); +static void btrfsic_print_rem_link(const struct btrfsic_state *state, + const struct btrfsic_block_link *l); +static char btrfsic_get_block_type(const struct btrfsic_state *state, + const struct btrfsic_block *block); +static void btrfsic_dump_tree(const struct btrfsic_state *state); +static void btrfsic_dump_tree_sub(const struct btrfsic_state *state, + const struct btrfsic_block *block, + int indent_level); +static struct btrfsic_block_link *btrfsic_block_link_lookup_or_add( + struct btrfsic_state *state, + struct btrfsic_block_data_ctx *next_block_ctx, + struct btrfsic_block *next_block, + struct btrfsic_block *from_block, + u64 parent_generation); +static struct btrfsic_block *btrfsic_block_lookup_or_add( + struct btrfsic_state *state, + struct btrfsic_block_data_ctx *block_ctx, + const char *additional_string, + int is_metadata, + int is_iodone, + int never_written, + int mirror_num, + int *was_created); +static int btrfsic_process_superblock_dev_mirror( + struct btrfsic_state *state, + struct btrfsic_dev_state *dev_state, + struct btrfs_device *device, + int superblock_mirror_num, + struct btrfsic_dev_state **selected_dev_state, + struct btrfs_super_block *selected_super); +static struct btrfsic_dev_state *btrfsic_dev_state_lookup( + struct block_device *bdev); +static void btrfsic_cmp_log_and_dev_bytenr(struct btrfsic_state *state, + u64 bytenr, + struct btrfsic_dev_state *dev_state, + u64 dev_bytenr); + +static struct mutex btrfsic_mutex; +static int btrfsic_is_initialized; +static struct btrfsic_dev_state_hashtable btrfsic_dev_state_hashtable; + + +static void btrfsic_block_init(struct btrfsic_block *b) +{ + b->magic_num = BTRFSIC_BLOCK_MAGIC_NUMBER; + b->dev_state = NULL; + b->dev_bytenr = 0; + b->logical_bytenr = 0; + b->generation = BTRFSIC_GENERATION_UNKNOWN; + b->disk_key.objectid = 0; + b->disk_key.type = 0; + b->disk_key.offset = 0; + b->is_metadata = 0; + b->is_superblock = 0; + b->is_iodone = 0; + b->iodone_w_error = 0; + b->never_written = 0; + b->mirror_num = 0; + b->next_in_same_bio = NULL; + b->orig_bio_bh_private = NULL; + b->orig_bio_bh_end_io.bio = NULL; + INIT_LIST_HEAD(&b->collision_resolving_node); + INIT_LIST_HEAD(&b->all_blocks_node); + INIT_LIST_HEAD(&b->ref_to_list); + INIT_LIST_HEAD(&b->ref_from_list); + b->submit_bio_bh_rw = 0; + b->flush_gen = 0; +} + +static struct btrfsic_block *btrfsic_block_alloc(void) +{ + struct btrfsic_block *b; + + b = kzalloc(sizeof(*b), GFP_NOFS); + if (NULL != b) + btrfsic_block_init(b); + + return b; +} + +static void btrfsic_block_free(struct btrfsic_block *b) +{ + BUG_ON(!(NULL == b || BTRFSIC_BLOCK_MAGIC_NUMBER == b->magic_num)); + kfree(b); +} + +static void btrfsic_block_link_init(struct btrfsic_block_link *l) +{ + l->magic_num = BTRFSIC_BLOCK_LINK_MAGIC_NUMBER; + l->ref_cnt = 1; + INIT_LIST_HEAD(&l->node_ref_to); + INIT_LIST_HEAD(&l->node_ref_from); + INIT_LIST_HEAD(&l->collision_resolving_node); + l->block_ref_to = NULL; + l->block_ref_from = NULL; +} + +static struct btrfsic_block_link *btrfsic_block_link_alloc(void) +{ + struct btrfsic_block_link *l; + + l = kzalloc(sizeof(*l), GFP_NOFS); + if (NULL != l) + btrfsic_block_link_init(l); + + return l; +} + +static void btrfsic_block_link_free(struct btrfsic_block_link *l) +{ + BUG_ON(!(NULL == l || BTRFSIC_BLOCK_LINK_MAGIC_NUMBER == l->magic_num)); + kfree(l); +} + +static void btrfsic_dev_state_init(struct btrfsic_dev_state *ds) +{ + ds->magic_num = BTRFSIC_DEV2STATE_MAGIC_NUMBER; + ds->bdev = NULL; + ds->state = NULL; + ds->name[0] = '\0'; + INIT_LIST_HEAD(&ds->collision_resolving_node); + ds->last_flush_gen = 0; + btrfsic_block_init(&ds->dummy_block_for_bio_bh_flush); + ds->dummy_block_for_bio_bh_flush.is_iodone = 1; + ds->dummy_block_for_bio_bh_flush.dev_state = ds; +} + +static struct btrfsic_dev_state *btrfsic_dev_state_alloc(void) +{ + struct btrfsic_dev_state *ds; + + ds = kzalloc(sizeof(*ds), GFP_NOFS); + if (NULL != ds) + btrfsic_dev_state_init(ds); + + return ds; +} + +static void btrfsic_dev_state_free(struct btrfsic_dev_state *ds) +{ + BUG_ON(!(NULL == ds || + BTRFSIC_DEV2STATE_MAGIC_NUMBER == ds->magic_num)); + kfree(ds); +} + +static void btrfsic_block_hashtable_init(struct btrfsic_block_hashtable *h) +{ + int i; + + for (i = 0; i < BTRFSIC_BLOCK_HASHTABLE_SIZE; i++) + INIT_LIST_HEAD(h->table + i); +} + +static void btrfsic_block_hashtable_add(struct btrfsic_block *b, + struct btrfsic_block_hashtable *h) +{ + const unsigned int hashval = + (((unsigned int)(b->dev_bytenr >> 16)) ^ + ((unsigned int)((uintptr_t)b->dev_state->bdev))) & + (BTRFSIC_BLOCK_HASHTABLE_SIZE - 1); + + list_add(&b->collision_resolving_node, h->table + hashval); +} + +static void btrfsic_block_hashtable_remove(struct btrfsic_block *b) +{ + list_del(&b->collision_resolving_node); +} + +static struct btrfsic_block *btrfsic_block_hashtable_lookup( + struct block_device *bdev, + u64 dev_bytenr, + struct btrfsic_block_hashtable *h) +{ + const unsigned int hashval = + (((unsigned int)(dev_bytenr >> 16)) ^ + ((unsigned int)((uintptr_t)bdev))) & + (BTRFSIC_BLOCK_HASHTABLE_SIZE - 1); + struct list_head *elem; + + list_for_each(elem, h->table + hashval) { + struct btrfsic_block *const b = + list_entry(elem, struct btrfsic_block, + collision_resolving_node); + + if (b->dev_state->bdev == bdev && b->dev_bytenr == dev_bytenr) + return b; + } + + return NULL; +} + +static void btrfsic_block_link_hashtable_init( + struct btrfsic_block_link_hashtable *h) +{ + int i; + + for (i = 0; i < BTRFSIC_BLOCK_LINK_HASHTABLE_SIZE; i++) + INIT_LIST_HEAD(h->table + i); +} + +static void btrfsic_block_link_hashtable_add( + struct btrfsic_block_link *l, + struct btrfsic_block_link_hashtable *h) +{ + const unsigned int hashval = + (((unsigned int)(l->block_ref_to->dev_bytenr >> 16)) ^ + ((unsigned int)(l->block_ref_from->dev_bytenr >> 16)) ^ + ((unsigned int)((uintptr_t)l->block_ref_to->dev_state->bdev)) ^ + ((unsigned int)((uintptr_t)l->block_ref_from->dev_state->bdev))) + & (BTRFSIC_BLOCK_LINK_HASHTABLE_SIZE - 1); + + BUG_ON(NULL == l->block_ref_to); + BUG_ON(NULL == l->block_ref_from); + list_add(&l->collision_resolving_node, h->table + hashval); +} + +static void btrfsic_block_link_hashtable_remove(struct btrfsic_block_link *l) +{ + list_del(&l->collision_resolving_node); +} + +static struct btrfsic_block_link *btrfsic_block_link_hashtable_lookup( + struct block_device *bdev_ref_to, + u64 dev_bytenr_ref_to, + struct block_device *bdev_ref_from, + u64 dev_bytenr_ref_from, + struct btrfsic_block_link_hashtable *h) +{ + const unsigned int hashval = + (((unsigned int)(dev_bytenr_ref_to >> 16)) ^ + ((unsigned int)(dev_bytenr_ref_from >> 16)) ^ + ((unsigned int)((uintptr_t)bdev_ref_to)) ^ + ((unsigned int)((uintptr_t)bdev_ref_from))) & + (BTRFSIC_BLOCK_LINK_HASHTABLE_SIZE - 1); + struct list_head *elem; + + list_for_each(elem, h->table + hashval) { + struct btrfsic_block_link *const l = + list_entry(elem, struct btrfsic_block_link, + collision_resolving_node); + + BUG_ON(NULL == l->block_ref_to); + BUG_ON(NULL == l->block_ref_from); + if (l->block_ref_to->dev_state->bdev == bdev_ref_to && + l->block_ref_to->dev_bytenr == dev_bytenr_ref_to && + l->block_ref_from->dev_state->bdev == bdev_ref_from && + l->block_ref_from->dev_bytenr == dev_bytenr_ref_from) + return l; + } + + return NULL; +} + +static void btrfsic_dev_state_hashtable_init( + struct btrfsic_dev_state_hashtable *h) +{ + int i; + + for (i = 0; i < BTRFSIC_DEV2STATE_HASHTABLE_SIZE; i++) + INIT_LIST_HEAD(h->table + i); +} + +static void btrfsic_dev_state_hashtable_add( + struct btrfsic_dev_state *ds, + struct btrfsic_dev_state_hashtable *h) +{ + const unsigned int hashval = + (((unsigned int)((uintptr_t)ds->bdev)) & + (BTRFSIC_DEV2STATE_HASHTABLE_SIZE - 1)); + + list_add(&ds->collision_resolving_node, h->table + hashval); +} + +static void btrfsic_dev_state_hashtable_remove(struct btrfsic_dev_state *ds) +{ + list_del(&ds->collision_resolving_node); +} + +static struct btrfsic_dev_state *btrfsic_dev_state_hashtable_lookup( + struct block_device *bdev, + struct btrfsic_dev_state_hashtable *h) +{ + const unsigned int hashval = + (((unsigned int)((uintptr_t)bdev)) & + (BTRFSIC_DEV2STATE_HASHTABLE_SIZE - 1)); + struct list_head *elem; + + list_for_each(elem, h->table + hashval) { + struct btrfsic_dev_state *const ds = + list_entry(elem, struct btrfsic_dev_state, + collision_resolving_node); + + if (ds->bdev == bdev) + return ds; + } + + return NULL; +} + +static int btrfsic_process_superblock(struct btrfsic_state *state, + struct btrfs_fs_devices *fs_devices) +{ + int ret = 0; + struct btrfs_super_block *selected_super; + struct list_head *dev_head = &fs_devices->devices; + struct btrfs_device *device; + struct btrfsic_dev_state *selected_dev_state = NULL; + int pass; + + BUG_ON(NULL == state); + selected_super = kzalloc(sizeof(*selected_super), GFP_NOFS); + if (NULL == selected_super) { + printk(KERN_INFO "btrfsic: error, kmalloc failed!\n"); + return -1; + } + + list_for_each_entry(device, dev_head, dev_list) { + int i; + struct btrfsic_dev_state *dev_state; + + if (!device->bdev || !device->name) + continue; + + dev_state = btrfsic_dev_state_lookup(device->bdev); + BUG_ON(NULL == dev_state); + for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { + ret = btrfsic_process_superblock_dev_mirror( + state, dev_state, device, i, + &selected_dev_state, selected_super); + if (0 != ret && 0 == i) { + kfree(selected_super); + return ret; + } + } + } + + if (NULL == state->latest_superblock) { + printk(KERN_INFO "btrfsic: no superblock found!\n"); + kfree(selected_super); + return -1; + } + + state->csum_size = btrfs_super_csum_size(selected_super); + + for (pass = 0; pass < 3; pass++) { + int num_copies; + int mirror_num; + u64 next_bytenr; + + switch (pass) { + case 0: + next_bytenr = btrfs_super_root(selected_super); + if (state->print_mask & + BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION) + printk(KERN_INFO "root@%llu\n", + (unsigned long long)next_bytenr); + break; + case 1: + next_bytenr = btrfs_super_chunk_root(selected_super); + if (state->print_mask & + BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION) + printk(KERN_INFO "chunk@%llu\n", + (unsigned long long)next_bytenr); + break; + case 2: + next_bytenr = btrfs_super_log_root(selected_super); + if (0 == next_bytenr) + continue; + if (state->print_mask & + BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION) + printk(KERN_INFO "log@%llu\n", + (unsigned long long)next_bytenr); + break; + } + + num_copies = + btrfs_num_copies(&state->root->fs_info->mapping_tree, + next_bytenr, state->metablock_size); + if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) + printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", + (unsigned long long)next_bytenr, num_copies); + + for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) { + struct btrfsic_block *next_block; + struct btrfsic_block_data_ctx tmp_next_block_ctx; + struct btrfsic_block_link *l; + + ret = btrfsic_map_block(state, next_bytenr, + state->metablock_size, + &tmp_next_block_ctx, + mirror_num); + if (ret) { + printk(KERN_INFO "btrfsic:" + " btrfsic_map_block(root @%llu," + " mirror %d) failed!\n", + (unsigned long long)next_bytenr, + mirror_num); + kfree(selected_super); + return -1; + } + + next_block = btrfsic_block_hashtable_lookup( + tmp_next_block_ctx.dev->bdev, + tmp_next_block_ctx.dev_bytenr, + &state->block_hashtable); + BUG_ON(NULL == next_block); + + l = btrfsic_block_link_hashtable_lookup( + tmp_next_block_ctx.dev->bdev, + tmp_next_block_ctx.dev_bytenr, + state->latest_superblock->dev_state-> + bdev, + state->latest_superblock->dev_bytenr, + &state->block_link_hashtable); + BUG_ON(NULL == l); + + ret = btrfsic_read_block(state, &tmp_next_block_ctx); + if (ret < (int)PAGE_CACHE_SIZE) { + printk(KERN_INFO + "btrfsic: read @logical %llu failed!\n", + (unsigned long long) + tmp_next_block_ctx.start); + btrfsic_release_block_ctx(&tmp_next_block_ctx); + kfree(selected_super); + return -1; + } + + ret = btrfsic_process_metablock(state, + next_block, + &tmp_next_block_ctx, + BTRFS_MAX_LEVEL + 3, 1); + btrfsic_release_block_ctx(&tmp_next_block_ctx); + } + } + + kfree(selected_super); + return ret; +} + +static int btrfsic_process_superblock_dev_mirror( + struct btrfsic_state *state, + struct btrfsic_dev_state *dev_state, + struct btrfs_device *device, + int superblock_mirror_num, + struct btrfsic_dev_state **selected_dev_state, + struct btrfs_super_block *selected_super) +{ + struct btrfs_super_block *super_tmp; + u64 dev_bytenr; + struct buffer_head *bh; + struct btrfsic_block *superblock_tmp; + int pass; + struct block_device *const superblock_bdev = device->bdev; + + /* super block bytenr is always the unmapped device bytenr */ + dev_bytenr = btrfs_sb_offset(superblock_mirror_num); + if (dev_bytenr + BTRFS_SUPER_INFO_SIZE > device->total_bytes) + return -1; + bh = __bread(superblock_bdev, dev_bytenr / 4096, + BTRFS_SUPER_INFO_SIZE); + if (NULL == bh) + return -1; + super_tmp = (struct btrfs_super_block *) + (bh->b_data + (dev_bytenr & 4095)); + + if (btrfs_super_bytenr(super_tmp) != dev_bytenr || + strncmp((char *)(&(super_tmp->magic)), BTRFS_MAGIC, + sizeof(super_tmp->magic)) || + memcmp(device->uuid, super_tmp->dev_item.uuid, BTRFS_UUID_SIZE) || + btrfs_super_nodesize(super_tmp) != state->metablock_size || + btrfs_super_leafsize(super_tmp) != state->metablock_size || + btrfs_super_sectorsize(super_tmp) != state->datablock_size) { + brelse(bh); + return 0; + } + + superblock_tmp = + btrfsic_block_hashtable_lookup(superblock_bdev, + dev_bytenr, + &state->block_hashtable); + if (NULL == superblock_tmp) { + superblock_tmp = btrfsic_block_alloc(); + if (NULL == superblock_tmp) { + printk(KERN_INFO "btrfsic: error, kmalloc failed!\n"); + brelse(bh); + return -1; + } + /* for superblock, only the dev_bytenr makes sense */ + superblock_tmp->dev_bytenr = dev_bytenr; + superblock_tmp->dev_state = dev_state; + superblock_tmp->logical_bytenr = dev_bytenr; + superblock_tmp->generation = btrfs_super_generation(super_tmp); + superblock_tmp->is_metadata = 1; + superblock_tmp->is_superblock = 1; + superblock_tmp->is_iodone = 1; + superblock_tmp->never_written = 0; + superblock_tmp->mirror_num = 1 + superblock_mirror_num; + if (state->print_mask & BTRFSIC_PRINT_MASK_SUPERBLOCK_WRITE) + printk_in_rcu(KERN_INFO "New initial S-block (bdev %p, %s)" + " @%llu (%s/%llu/%d)\n", + superblock_bdev, + rcu_str_deref(device->name), + (unsigned long long)dev_bytenr, + dev_state->name, + (unsigned long long)dev_bytenr, + superblock_mirror_num); + list_add(&superblock_tmp->all_blocks_node, + &state->all_blocks_list); + btrfsic_block_hashtable_add(superblock_tmp, + &state->block_hashtable); + } + + /* select the one with the highest generation field */ + if (btrfs_super_generation(super_tmp) > + state->max_superblock_generation || + 0 == state->max_superblock_generation) { + memcpy(selected_super, super_tmp, sizeof(*selected_super)); + *selected_dev_state = dev_state; + state->max_superblock_generation = + btrfs_super_generation(super_tmp); + state->latest_superblock = superblock_tmp; + } + + for (pass = 0; pass < 3; pass++) { + u64 next_bytenr; + int num_copies; + int mirror_num; + const char *additional_string = NULL; + struct btrfs_disk_key tmp_disk_key; + + tmp_disk_key.type = BTRFS_ROOT_ITEM_KEY; + tmp_disk_key.offset = 0; + switch (pass) { + case 0: + tmp_disk_key.objectid = + cpu_to_le64(BTRFS_ROOT_TREE_OBJECTID); + additional_string = "initial root "; + next_bytenr = btrfs_super_root(super_tmp); + break; + case 1: + tmp_disk_key.objectid = + cpu_to_le64(BTRFS_CHUNK_TREE_OBJECTID); + additional_string = "initial chunk "; + next_bytenr = btrfs_super_chunk_root(super_tmp); + break; + case 2: + tmp_disk_key.objectid = + cpu_to_le64(BTRFS_TREE_LOG_OBJECTID); + additional_string = "initial log "; + next_bytenr = btrfs_super_log_root(super_tmp); + if (0 == next_bytenr) + continue; + break; + } + + num_copies = + btrfs_num_copies(&state->root->fs_info->mapping_tree, + next_bytenr, state->metablock_size); + if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) + printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", + (unsigned long long)next_bytenr, num_copies); + for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) { + struct btrfsic_block *next_block; + struct btrfsic_block_data_ctx tmp_next_block_ctx; + struct btrfsic_block_link *l; + + if (btrfsic_map_block(state, next_bytenr, + state->metablock_size, + &tmp_next_block_ctx, + mirror_num)) { + printk(KERN_INFO "btrfsic: btrfsic_map_block(" + "bytenr @%llu, mirror %d) failed!\n", + (unsigned long long)next_bytenr, + mirror_num); + brelse(bh); + return -1; + } + + next_block = btrfsic_block_lookup_or_add( + state, &tmp_next_block_ctx, + additional_string, 1, 1, 0, + mirror_num, NULL); + if (NULL == next_block) { + btrfsic_release_block_ctx(&tmp_next_block_ctx); + brelse(bh); + return -1; + } + + next_block->disk_key = tmp_disk_key; + next_block->generation = BTRFSIC_GENERATION_UNKNOWN; + l = btrfsic_block_link_lookup_or_add( + state, &tmp_next_block_ctx, + next_block, superblock_tmp, + BTRFSIC_GENERATION_UNKNOWN); + btrfsic_release_block_ctx(&tmp_next_block_ctx); + if (NULL == l) { + brelse(bh); + return -1; + } + } + } + if (state->print_mask & BTRFSIC_PRINT_MASK_INITIAL_ALL_TREES) + btrfsic_dump_tree_sub(state, superblock_tmp, 0); + + brelse(bh); + return 0; +} + +static struct btrfsic_stack_frame *btrfsic_stack_frame_alloc(void) +{ + struct btrfsic_stack_frame *sf; + + sf = kzalloc(sizeof(*sf), GFP_NOFS); + if (NULL == sf) + printk(KERN_INFO "btrfsic: alloc memory failed!\n"); + else + sf->magic = BTRFSIC_BLOCK_STACK_FRAME_MAGIC_NUMBER; + return sf; +} + +static void btrfsic_stack_frame_free(struct btrfsic_stack_frame *sf) +{ + BUG_ON(!(NULL == sf || + BTRFSIC_BLOCK_STACK_FRAME_MAGIC_NUMBER == sf->magic)); + kfree(sf); +} + +static int btrfsic_process_metablock( + struct btrfsic_state *state, + struct btrfsic_block *const first_block, + struct btrfsic_block_data_ctx *const first_block_ctx, + int first_limit_nesting, int force_iodone_flag) +{ + struct btrfsic_stack_frame initial_stack_frame = { 0 }; + struct btrfsic_stack_frame *sf; + struct btrfsic_stack_frame *next_stack; + struct btrfs_header *const first_hdr = + (struct btrfs_header *)first_block_ctx->datav[0]; + + BUG_ON(!first_hdr); + sf = &initial_stack_frame; + sf->error = 0; + sf->i = -1; + sf->limit_nesting = first_limit_nesting; + sf->block = first_block; + sf->block_ctx = first_block_ctx; + sf->next_block = NULL; + sf->hdr = first_hdr; + sf->prev = NULL; + +continue_with_new_stack_frame: + sf->block->generation = le64_to_cpu(sf->hdr->generation); + if (0 == sf->hdr->level) { + struct btrfs_leaf *const leafhdr = + (struct btrfs_leaf *)sf->hdr; + + if (-1 == sf->i) { + sf->nr = le32_to_cpu(leafhdr->header.nritems); + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "leaf %llu items %d generation %llu" + " owner %llu\n", + (unsigned long long) + sf->block_ctx->start, + sf->nr, + (unsigned long long) + le64_to_cpu(leafhdr->header.generation), + (unsigned long long) + le64_to_cpu(leafhdr->header.owner)); + } + +continue_with_current_leaf_stack_frame: + if (0 == sf->num_copies || sf->mirror_num > sf->num_copies) { + sf->i++; + sf->num_copies = 0; + } + + if (sf->i < sf->nr) { + struct btrfs_item disk_item; + u32 disk_item_offset = + (uintptr_t)(leafhdr->items + sf->i) - + (uintptr_t)leafhdr; + struct btrfs_disk_key *disk_key; + u8 type; + u32 item_offset; + u32 item_size; + + if (disk_item_offset + sizeof(struct btrfs_item) > + sf->block_ctx->len) { +leaf_item_out_of_bounce_error: + printk(KERN_INFO + "btrfsic: leaf item out of bounce at logical %llu, dev %s\n", + sf->block_ctx->start, + sf->block_ctx->dev->name); + goto one_stack_frame_backwards; + } + btrfsic_read_from_block_data(sf->block_ctx, + &disk_item, + disk_item_offset, + sizeof(struct btrfs_item)); + item_offset = le32_to_cpu(disk_item.offset); + item_size = le32_to_cpu(disk_item.size); + disk_key = &disk_item.key; + type = disk_key->type; + + if (BTRFS_ROOT_ITEM_KEY == type) { + struct btrfs_root_item root_item; + u32 root_item_offset; + u64 next_bytenr; + + root_item_offset = item_offset + + offsetof(struct btrfs_leaf, items); + if (root_item_offset + item_size > + sf->block_ctx->len) + goto leaf_item_out_of_bounce_error; + btrfsic_read_from_block_data( + sf->block_ctx, &root_item, + root_item_offset, + item_size); + next_bytenr = le64_to_cpu(root_item.bytenr); + + sf->error = + btrfsic_create_link_to_next_block( + state, + sf->block, + sf->block_ctx, + next_bytenr, + sf->limit_nesting, + &sf->next_block_ctx, + &sf->next_block, + force_iodone_flag, + &sf->num_copies, + &sf->mirror_num, + disk_key, + le64_to_cpu(root_item. + generation)); + if (sf->error) + goto one_stack_frame_backwards; + + if (NULL != sf->next_block) { + struct btrfs_header *const next_hdr = + (struct btrfs_header *) + sf->next_block_ctx.datav[0]; + + next_stack = + btrfsic_stack_frame_alloc(); + if (NULL == next_stack) { + btrfsic_release_block_ctx( + &sf-> + next_block_ctx); + goto one_stack_frame_backwards; + } + + next_stack->i = -1; + next_stack->block = sf->next_block; + next_stack->block_ctx = + &sf->next_block_ctx; + next_stack->next_block = NULL; + next_stack->hdr = next_hdr; + next_stack->limit_nesting = + sf->limit_nesting - 1; + next_stack->prev = sf; + sf = next_stack; + goto continue_with_new_stack_frame; + } + } else if (BTRFS_EXTENT_DATA_KEY == type && + state->include_extent_data) { + sf->error = btrfsic_handle_extent_data( + state, + sf->block, + sf->block_ctx, + item_offset, + force_iodone_flag); + if (sf->error) + goto one_stack_frame_backwards; + } + + goto continue_with_current_leaf_stack_frame; + } + } else { + struct btrfs_node *const nodehdr = (struct btrfs_node *)sf->hdr; + + if (-1 == sf->i) { + sf->nr = le32_to_cpu(nodehdr->header.nritems); + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO "node %llu level %d items %d" + " generation %llu owner %llu\n", + (unsigned long long) + sf->block_ctx->start, + nodehdr->header.level, sf->nr, + (unsigned long long) + le64_to_cpu(nodehdr->header.generation), + (unsigned long long) + le64_to_cpu(nodehdr->header.owner)); + } + +continue_with_current_node_stack_frame: + if (0 == sf->num_copies || sf->mirror_num > sf->num_copies) { + sf->i++; + sf->num_copies = 0; + } + + if (sf->i < sf->nr) { + struct btrfs_key_ptr key_ptr; + u32 key_ptr_offset; + u64 next_bytenr; + + key_ptr_offset = (uintptr_t)(nodehdr->ptrs + sf->i) - + (uintptr_t)nodehdr; + if (key_ptr_offset + sizeof(struct btrfs_key_ptr) > + sf->block_ctx->len) { + printk(KERN_INFO + "btrfsic: node item out of bounce at logical %llu, dev %s\n", + sf->block_ctx->start, + sf->block_ctx->dev->name); + goto one_stack_frame_backwards; + } + btrfsic_read_from_block_data( + sf->block_ctx, &key_ptr, key_ptr_offset, + sizeof(struct btrfs_key_ptr)); + next_bytenr = le64_to_cpu(key_ptr.blockptr); + + sf->error = btrfsic_create_link_to_next_block( + state, + sf->block, + sf->block_ctx, + next_bytenr, + sf->limit_nesting, + &sf->next_block_ctx, + &sf->next_block, + force_iodone_flag, + &sf->num_copies, + &sf->mirror_num, + &key_ptr.key, + le64_to_cpu(key_ptr.generation)); + if (sf->error) + goto one_stack_frame_backwards; + + if (NULL != sf->next_block) { + struct btrfs_header *const next_hdr = + (struct btrfs_header *) + sf->next_block_ctx.datav[0]; + + next_stack = btrfsic_stack_frame_alloc(); + if (NULL == next_stack) + goto one_stack_frame_backwards; + + next_stack->i = -1; + next_stack->block = sf->next_block; + next_stack->block_ctx = &sf->next_block_ctx; + next_stack->next_block = NULL; + next_stack->hdr = next_hdr; + next_stack->limit_nesting = + sf->limit_nesting - 1; + next_stack->prev = sf; + sf = next_stack; + goto continue_with_new_stack_frame; + } + + goto continue_with_current_node_stack_frame; + } + } + +one_stack_frame_backwards: + if (NULL != sf->prev) { + struct btrfsic_stack_frame *const prev = sf->prev; + + /* the one for the initial block is freed in the caller */ + btrfsic_release_block_ctx(sf->block_ctx); + + if (sf->error) { + prev->error = sf->error; + btrfsic_stack_frame_free(sf); + sf = prev; + goto one_stack_frame_backwards; + } + + btrfsic_stack_frame_free(sf); + sf = prev; + goto continue_with_new_stack_frame; + } else { + BUG_ON(&initial_stack_frame != sf); + } + + return sf->error; +} + +static void btrfsic_read_from_block_data( + struct btrfsic_block_data_ctx *block_ctx, + void *dstv, u32 offset, size_t len) +{ + size_t cur; + size_t offset_in_page; + char *kaddr; + char *dst = (char *)dstv; + size_t start_offset = block_ctx->start & ((u64)PAGE_CACHE_SIZE - 1); + unsigned long i = (start_offset + offset) >> PAGE_CACHE_SHIFT; + + WARN_ON(offset + len > block_ctx->len); + offset_in_page = (start_offset + offset) & + ((unsigned long)PAGE_CACHE_SIZE - 1); + + while (len > 0) { + cur = min(len, ((size_t)PAGE_CACHE_SIZE - offset_in_page)); + BUG_ON(i >= (block_ctx->len + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT); + kaddr = block_ctx->datav[i]; + memcpy(dst, kaddr + offset_in_page, cur); + + dst += cur; + len -= cur; + offset_in_page = 0; + i++; + } +} + +static int btrfsic_create_link_to_next_block( + struct btrfsic_state *state, + struct btrfsic_block *block, + struct btrfsic_block_data_ctx *block_ctx, + u64 next_bytenr, + int limit_nesting, + struct btrfsic_block_data_ctx *next_block_ctx, + struct btrfsic_block **next_blockp, + int force_iodone_flag, + int *num_copiesp, int *mirror_nump, + struct btrfs_disk_key *disk_key, + u64 parent_generation) +{ + struct btrfsic_block *next_block = NULL; + int ret; + struct btrfsic_block_link *l; + int did_alloc_block_link; + int block_was_created; + + *next_blockp = NULL; + if (0 == *num_copiesp) { + *num_copiesp = + btrfs_num_copies(&state->root->fs_info->mapping_tree, + next_bytenr, state->metablock_size); + if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) + printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", + (unsigned long long)next_bytenr, *num_copiesp); + *mirror_nump = 1; + } + + if (*mirror_nump > *num_copiesp) + return 0; + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "btrfsic_create_link_to_next_block(mirror_num=%d)\n", + *mirror_nump); + ret = btrfsic_map_block(state, next_bytenr, + state->metablock_size, + next_block_ctx, *mirror_nump); + if (ret) { + printk(KERN_INFO + "btrfsic: btrfsic_map_block(@%llu, mirror=%d) failed!\n", + (unsigned long long)next_bytenr, *mirror_nump); + btrfsic_release_block_ctx(next_block_ctx); + *next_blockp = NULL; + return -1; + } + + next_block = btrfsic_block_lookup_or_add(state, + next_block_ctx, "referenced ", + 1, force_iodone_flag, + !force_iodone_flag, + *mirror_nump, + &block_was_created); + if (NULL == next_block) { + btrfsic_release_block_ctx(next_block_ctx); + *next_blockp = NULL; + return -1; + } + if (block_was_created) { + l = NULL; + next_block->generation = BTRFSIC_GENERATION_UNKNOWN; + } else { + if (next_block->logical_bytenr != next_bytenr && + !(!next_block->is_metadata && + 0 == next_block->logical_bytenr)) { + printk(KERN_INFO + "Referenced block @%llu (%s/%llu/%d)" + " found in hash table, %c," + " bytenr mismatch (!= stored %llu).\n", + (unsigned long long)next_bytenr, + next_block_ctx->dev->name, + (unsigned long long)next_block_ctx->dev_bytenr, + *mirror_nump, + btrfsic_get_block_type(state, next_block), + (unsigned long long)next_block->logical_bytenr); + } else if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "Referenced block @%llu (%s/%llu/%d)" + " found in hash table, %c.\n", + (unsigned long long)next_bytenr, + next_block_ctx->dev->name, + (unsigned long long)next_block_ctx->dev_bytenr, + *mirror_nump, + btrfsic_get_block_type(state, next_block)); + next_block->logical_bytenr = next_bytenr; + + next_block->mirror_num = *mirror_nump; + l = btrfsic_block_link_hashtable_lookup( + next_block_ctx->dev->bdev, + next_block_ctx->dev_bytenr, + block_ctx->dev->bdev, + block_ctx->dev_bytenr, + &state->block_link_hashtable); + } + + next_block->disk_key = *disk_key; + if (NULL == l) { + l = btrfsic_block_link_alloc(); + if (NULL == l) { + printk(KERN_INFO "btrfsic: error, kmalloc failed!\n"); + btrfsic_release_block_ctx(next_block_ctx); + *next_blockp = NULL; + return -1; + } + + did_alloc_block_link = 1; + l->block_ref_to = next_block; + l->block_ref_from = block; + l->ref_cnt = 1; + l->parent_generation = parent_generation; + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + btrfsic_print_add_link(state, l); + + list_add(&l->node_ref_to, &block->ref_to_list); + list_add(&l->node_ref_from, &next_block->ref_from_list); + + btrfsic_block_link_hashtable_add(l, + &state->block_link_hashtable); + } else { + did_alloc_block_link = 0; + if (0 == limit_nesting) { + l->ref_cnt++; + l->parent_generation = parent_generation; + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + btrfsic_print_add_link(state, l); + } + } + + if (limit_nesting > 0 && did_alloc_block_link) { + ret = btrfsic_read_block(state, next_block_ctx); + if (ret < (int)next_block_ctx->len) { + printk(KERN_INFO + "btrfsic: read block @logical %llu failed!\n", + (unsigned long long)next_bytenr); + btrfsic_release_block_ctx(next_block_ctx); + *next_blockp = NULL; + return -1; + } + + *next_blockp = next_block; + } else { + *next_blockp = NULL; + } + (*mirror_nump)++; + + return 0; +} + +static int btrfsic_handle_extent_data( + struct btrfsic_state *state, + struct btrfsic_block *block, + struct btrfsic_block_data_ctx *block_ctx, + u32 item_offset, int force_iodone_flag) +{ + int ret; + struct btrfs_file_extent_item file_extent_item; + u64 file_extent_item_offset; + u64 next_bytenr; + u64 num_bytes; + u64 generation; + struct btrfsic_block_link *l; + + file_extent_item_offset = offsetof(struct btrfs_leaf, items) + + item_offset; + if (file_extent_item_offset + + offsetof(struct btrfs_file_extent_item, disk_num_bytes) > + block_ctx->len) { + printk(KERN_INFO + "btrfsic: file item out of bounce at logical %llu, dev %s\n", + block_ctx->start, block_ctx->dev->name); + return -1; + } + + btrfsic_read_from_block_data(block_ctx, &file_extent_item, + file_extent_item_offset, + offsetof(struct btrfs_file_extent_item, disk_num_bytes)); + if (BTRFS_FILE_EXTENT_REG != file_extent_item.type || + ((u64)0) == le64_to_cpu(file_extent_item.disk_bytenr)) { + if (state->print_mask & BTRFSIC_PRINT_MASK_VERY_VERBOSE) + printk(KERN_INFO "extent_data: type %u, disk_bytenr = %llu\n", + file_extent_item.type, + (unsigned long long) + le64_to_cpu(file_extent_item.disk_bytenr)); + return 0; + } + + if (file_extent_item_offset + sizeof(struct btrfs_file_extent_item) > + block_ctx->len) { + printk(KERN_INFO + "btrfsic: file item out of bounce at logical %llu, dev %s\n", + block_ctx->start, block_ctx->dev->name); + return -1; + } + btrfsic_read_from_block_data(block_ctx, &file_extent_item, + file_extent_item_offset, + sizeof(struct btrfs_file_extent_item)); + next_bytenr = le64_to_cpu(file_extent_item.disk_bytenr) + + le64_to_cpu(file_extent_item.offset); + generation = le64_to_cpu(file_extent_item.generation); + num_bytes = le64_to_cpu(file_extent_item.num_bytes); + generation = le64_to_cpu(file_extent_item.generation); + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERY_VERBOSE) + printk(KERN_INFO "extent_data: type %u, disk_bytenr = %llu," + " offset = %llu, num_bytes = %llu\n", + file_extent_item.type, + (unsigned long long) + le64_to_cpu(file_extent_item.disk_bytenr), + (unsigned long long)le64_to_cpu(file_extent_item.offset), + (unsigned long long)num_bytes); + while (num_bytes > 0) { + u32 chunk_len; + int num_copies; + int mirror_num; + + if (num_bytes > state->datablock_size) + chunk_len = state->datablock_size; + else + chunk_len = num_bytes; + + num_copies = + btrfs_num_copies(&state->root->fs_info->mapping_tree, + next_bytenr, state->datablock_size); + if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) + printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", + (unsigned long long)next_bytenr, num_copies); + for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) { + struct btrfsic_block_data_ctx next_block_ctx; + struct btrfsic_block *next_block; + int block_was_created; + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO "btrfsic_handle_extent_data(" + "mirror_num=%d)\n", mirror_num); + if (state->print_mask & BTRFSIC_PRINT_MASK_VERY_VERBOSE) + printk(KERN_INFO + "\tdisk_bytenr = %llu, num_bytes %u\n", + (unsigned long long)next_bytenr, + chunk_len); + ret = btrfsic_map_block(state, next_bytenr, + chunk_len, &next_block_ctx, + mirror_num); + if (ret) { + printk(KERN_INFO + "btrfsic: btrfsic_map_block(@%llu," + " mirror=%d) failed!\n", + (unsigned long long)next_bytenr, + mirror_num); + return -1; + } + + next_block = btrfsic_block_lookup_or_add( + state, + &next_block_ctx, + "referenced ", + 0, + force_iodone_flag, + !force_iodone_flag, + mirror_num, + &block_was_created); + if (NULL == next_block) { + printk(KERN_INFO + "btrfsic: error, kmalloc failed!\n"); + btrfsic_release_block_ctx(&next_block_ctx); + return -1; + } + if (!block_was_created) { + if (next_block->logical_bytenr != next_bytenr && + !(!next_block->is_metadata && + 0 == next_block->logical_bytenr)) { + printk(KERN_INFO + "Referenced block" + " @%llu (%s/%llu/%d)" + " found in hash table, D," + " bytenr mismatch" + " (!= stored %llu).\n", + (unsigned long long)next_bytenr, + next_block_ctx.dev->name, + (unsigned long long) + next_block_ctx.dev_bytenr, + mirror_num, + (unsigned long long) + next_block->logical_bytenr); + } + next_block->logical_bytenr = next_bytenr; + next_block->mirror_num = mirror_num; + } + + l = btrfsic_block_link_lookup_or_add(state, + &next_block_ctx, + next_block, block, + generation); + btrfsic_release_block_ctx(&next_block_ctx); + if (NULL == l) + return -1; + } + + next_bytenr += chunk_len; + num_bytes -= chunk_len; + } + + return 0; +} + +static int btrfsic_map_block(struct btrfsic_state *state, u64 bytenr, u32 len, + struct btrfsic_block_data_ctx *block_ctx_out, + int mirror_num) +{ + int ret; + u64 length; + struct btrfs_bio *multi = NULL; + struct btrfs_device *device; + + length = len; + ret = btrfs_map_block(&state->root->fs_info->mapping_tree, READ, + bytenr, &length, &multi, mirror_num); + + device = multi->stripes[0].dev; + block_ctx_out->dev = btrfsic_dev_state_lookup(device->bdev); + block_ctx_out->dev_bytenr = multi->stripes[0].physical; + block_ctx_out->start = bytenr; + block_ctx_out->len = len; + block_ctx_out->datav = NULL; + block_ctx_out->pagev = NULL; + block_ctx_out->mem_to_free = NULL; + + if (0 == ret) + kfree(multi); + if (NULL == block_ctx_out->dev) { + ret = -ENXIO; + printk(KERN_INFO "btrfsic: error, cannot lookup dev (#1)!\n"); + } + + return ret; +} + +static int btrfsic_map_superblock(struct btrfsic_state *state, u64 bytenr, + u32 len, struct block_device *bdev, + struct btrfsic_block_data_ctx *block_ctx_out) +{ + block_ctx_out->dev = btrfsic_dev_state_lookup(bdev); + block_ctx_out->dev_bytenr = bytenr; + block_ctx_out->start = bytenr; + block_ctx_out->len = len; + block_ctx_out->datav = NULL; + block_ctx_out->pagev = NULL; + block_ctx_out->mem_to_free = NULL; + if (NULL != block_ctx_out->dev) { + return 0; + } else { + printk(KERN_INFO "btrfsic: error, cannot lookup dev (#2)!\n"); + return -ENXIO; + } +} + +static void btrfsic_release_block_ctx(struct btrfsic_block_data_ctx *block_ctx) +{ + if (block_ctx->mem_to_free) { + unsigned int num_pages; + + BUG_ON(!block_ctx->datav); + BUG_ON(!block_ctx->pagev); + num_pages = (block_ctx->len + (u64)PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; + while (num_pages > 0) { + num_pages--; + if (block_ctx->datav[num_pages]) { + kunmap(block_ctx->pagev[num_pages]); + block_ctx->datav[num_pages] = NULL; + } + if (block_ctx->pagev[num_pages]) { + __free_page(block_ctx->pagev[num_pages]); + block_ctx->pagev[num_pages] = NULL; + } + } + + kfree(block_ctx->mem_to_free); + block_ctx->mem_to_free = NULL; + block_ctx->pagev = NULL; + block_ctx->datav = NULL; + } +} + +static int btrfsic_read_block(struct btrfsic_state *state, + struct btrfsic_block_data_ctx *block_ctx) +{ + unsigned int num_pages; + unsigned int i; + u64 dev_bytenr; + int ret; + + BUG_ON(block_ctx->datav); + BUG_ON(block_ctx->pagev); + BUG_ON(block_ctx->mem_to_free); + if (block_ctx->dev_bytenr & ((u64)PAGE_CACHE_SIZE - 1)) { + printk(KERN_INFO + "btrfsic: read_block() with unaligned bytenr %llu\n", + (unsigned long long)block_ctx->dev_bytenr); + return -1; + } + + num_pages = (block_ctx->len + (u64)PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; + block_ctx->mem_to_free = kzalloc((sizeof(*block_ctx->datav) + + sizeof(*block_ctx->pagev)) * + num_pages, GFP_NOFS); + if (!block_ctx->mem_to_free) + return -1; + block_ctx->datav = block_ctx->mem_to_free; + block_ctx->pagev = (struct page **)(block_ctx->datav + num_pages); + for (i = 0; i < num_pages; i++) { + block_ctx->pagev[i] = alloc_page(GFP_NOFS); + if (!block_ctx->pagev[i]) + return -1; + } + + dev_bytenr = block_ctx->dev_bytenr; + for (i = 0; i < num_pages;) { + struct bio *bio; + unsigned int j; + DECLARE_COMPLETION_ONSTACK(complete); + + bio = bio_alloc(GFP_NOFS, num_pages - i); + if (!bio) { + printk(KERN_INFO + "btrfsic: bio_alloc() for %u pages failed!\n", + num_pages - i); + return -1; + } + bio->bi_bdev = block_ctx->dev->bdev; + bio->bi_sector = dev_bytenr >> 9; + bio->bi_end_io = btrfsic_complete_bio_end_io; + bio->bi_private = &complete; + + for (j = i; j < num_pages; j++) { + ret = bio_add_page(bio, block_ctx->pagev[j], + PAGE_CACHE_SIZE, 0); + if (PAGE_CACHE_SIZE != ret) + break; + } + if (j == i) { + printk(KERN_INFO + "btrfsic: error, failed to add a single page!\n"); + return -1; + } + submit_bio(READ, bio); + + /* this will also unplug the queue */ + wait_for_completion(&complete); + + if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) { + printk(KERN_INFO + "btrfsic: read error at logical %llu dev %s!\n", + block_ctx->start, block_ctx->dev->name); + bio_put(bio); + return -1; + } + bio_put(bio); + dev_bytenr += (j - i) * PAGE_CACHE_SIZE; + i = j; + } + for (i = 0; i < num_pages; i++) { + block_ctx->datav[i] = kmap(block_ctx->pagev[i]); + if (!block_ctx->datav[i]) { + printk(KERN_INFO "btrfsic: kmap() failed (dev %s)!\n", + block_ctx->dev->name); + return -1; + } + } + + return block_ctx->len; +} + +static void btrfsic_complete_bio_end_io(struct bio *bio, int err) +{ + complete((struct completion *)bio->bi_private); +} + +static void btrfsic_dump_database(struct btrfsic_state *state) +{ + struct list_head *elem_all; + + BUG_ON(NULL == state); + + printk(KERN_INFO "all_blocks_list:\n"); + list_for_each(elem_all, &state->all_blocks_list) { + const struct btrfsic_block *const b_all = + list_entry(elem_all, struct btrfsic_block, + all_blocks_node); + struct list_head *elem_ref_to; + struct list_head *elem_ref_from; + + printk(KERN_INFO "%c-block @%llu (%s/%llu/%d)\n", + btrfsic_get_block_type(state, b_all), + (unsigned long long)b_all->logical_bytenr, + b_all->dev_state->name, + (unsigned long long)b_all->dev_bytenr, + b_all->mirror_num); + + list_for_each(elem_ref_to, &b_all->ref_to_list) { + const struct btrfsic_block_link *const l = + list_entry(elem_ref_to, + struct btrfsic_block_link, + node_ref_to); + + printk(KERN_INFO " %c @%llu (%s/%llu/%d)" + " refers %u* to" + " %c @%llu (%s/%llu/%d)\n", + btrfsic_get_block_type(state, b_all), + (unsigned long long)b_all->logical_bytenr, + b_all->dev_state->name, + (unsigned long long)b_all->dev_bytenr, + b_all->mirror_num, + l->ref_cnt, + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long) + l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num); + } + + list_for_each(elem_ref_from, &b_all->ref_from_list) { + const struct btrfsic_block_link *const l = + list_entry(elem_ref_from, + struct btrfsic_block_link, + node_ref_from); + + printk(KERN_INFO " %c @%llu (%s/%llu/%d)" + " is ref %u* from" + " %c @%llu (%s/%llu/%d)\n", + btrfsic_get_block_type(state, b_all), + (unsigned long long)b_all->logical_bytenr, + b_all->dev_state->name, + (unsigned long long)b_all->dev_bytenr, + b_all->mirror_num, + l->ref_cnt, + btrfsic_get_block_type(state, l->block_ref_from), + (unsigned long long) + l->block_ref_from->logical_bytenr, + l->block_ref_from->dev_state->name, + (unsigned long long) + l->block_ref_from->dev_bytenr, + l->block_ref_from->mirror_num); + } + + printk(KERN_INFO "\n"); + } +} + +/* + * Test whether the disk block contains a tree block (leaf or node) + * (note that this test fails for the super block) + */ +static int btrfsic_test_for_metadata(struct btrfsic_state *state, + char **datav, unsigned int num_pages) +{ + struct btrfs_header *h; + u8 csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + unsigned int i; + + if (num_pages * PAGE_CACHE_SIZE < state->metablock_size) + return 1; /* not metadata */ + num_pages = state->metablock_size >> PAGE_CACHE_SHIFT; + h = (struct btrfs_header *)datav[0]; + + if (memcmp(h->fsid, state->root->fs_info->fsid, BTRFS_UUID_SIZE)) + return 1; + + for (i = 0; i < num_pages; i++) { + u8 *data = i ? datav[i] : (datav[i] + BTRFS_CSUM_SIZE); + size_t sublen = i ? PAGE_CACHE_SIZE : + (PAGE_CACHE_SIZE - BTRFS_CSUM_SIZE); + + crc = crc32c(crc, data, sublen); + } + btrfs_csum_final(crc, csum); + if (memcmp(csum, h->csum, state->csum_size)) + return 1; + + return 0; /* is metadata */ +} + +static void btrfsic_process_written_block(struct btrfsic_dev_state *dev_state, + u64 dev_bytenr, char **mapped_datav, + unsigned int num_pages, + struct bio *bio, int *bio_is_patched, + struct buffer_head *bh, + int submit_bio_bh_rw) +{ + int is_metadata; + struct btrfsic_block *block; + struct btrfsic_block_data_ctx block_ctx; + int ret; + struct btrfsic_state *state = dev_state->state; + struct block_device *bdev = dev_state->bdev; + unsigned int processed_len; + + if (NULL != bio_is_patched) + *bio_is_patched = 0; + +again: + if (num_pages == 0) + return; + + processed_len = 0; + is_metadata = (0 == btrfsic_test_for_metadata(state, mapped_datav, + num_pages)); + + block = btrfsic_block_hashtable_lookup(bdev, dev_bytenr, + &state->block_hashtable); + if (NULL != block) { + u64 bytenr = 0; + struct list_head *elem_ref_to; + struct list_head *tmp_ref_to; + + if (block->is_superblock) { + bytenr = le64_to_cpu(((struct btrfs_super_block *) + mapped_datav[0])->bytenr); + if (num_pages * PAGE_CACHE_SIZE < + BTRFS_SUPER_INFO_SIZE) { + printk(KERN_INFO + "btrfsic: cannot work with too short bios!\n"); + return; + } + is_metadata = 1; + BUG_ON(BTRFS_SUPER_INFO_SIZE & (PAGE_CACHE_SIZE - 1)); + processed_len = BTRFS_SUPER_INFO_SIZE; + if (state->print_mask & + BTRFSIC_PRINT_MASK_TREE_BEFORE_SB_WRITE) { + printk(KERN_INFO + "[before new superblock is written]:\n"); + btrfsic_dump_tree_sub(state, block, 0); + } + } + if (is_metadata) { + if (!block->is_superblock) { + if (num_pages * PAGE_CACHE_SIZE < + state->metablock_size) { + printk(KERN_INFO + "btrfsic: cannot work with too short bios!\n"); + return; + } + processed_len = state->metablock_size; + bytenr = le64_to_cpu(((struct btrfs_header *) + mapped_datav[0])->bytenr); + btrfsic_cmp_log_and_dev_bytenr(state, bytenr, + dev_state, + dev_bytenr); + } + if (block->logical_bytenr != bytenr) { + printk(KERN_INFO + "Written block @%llu (%s/%llu/%d)" + " found in hash table, %c," + " bytenr mismatch" + " (!= stored %llu).\n", + (unsigned long long)bytenr, + dev_state->name, + (unsigned long long)dev_bytenr, + block->mirror_num, + btrfsic_get_block_type(state, block), + (unsigned long long) + block->logical_bytenr); + block->logical_bytenr = bytenr; + } else if (state->print_mask & + BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "Written block @%llu (%s/%llu/%d)" + " found in hash table, %c.\n", + (unsigned long long)bytenr, + dev_state->name, + (unsigned long long)dev_bytenr, + block->mirror_num, + btrfsic_get_block_type(state, block)); + } else { + if (num_pages * PAGE_CACHE_SIZE < + state->datablock_size) { + printk(KERN_INFO + "btrfsic: cannot work with too short bios!\n"); + return; + } + processed_len = state->datablock_size; + bytenr = block->logical_bytenr; + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "Written block @%llu (%s/%llu/%d)" + " found in hash table, %c.\n", + (unsigned long long)bytenr, + dev_state->name, + (unsigned long long)dev_bytenr, + block->mirror_num, + btrfsic_get_block_type(state, block)); + } + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "ref_to_list: %cE, ref_from_list: %cE\n", + list_empty(&block->ref_to_list) ? ' ' : '!', + list_empty(&block->ref_from_list) ? ' ' : '!'); + if (btrfsic_is_block_ref_by_superblock(state, block, 0)) { + printk(KERN_INFO "btrfs: attempt to overwrite %c-block" + " @%llu (%s/%llu/%d), old(gen=%llu," + " objectid=%llu, type=%d, offset=%llu)," + " new(gen=%llu)," + " which is referenced by most recent superblock" + " (superblockgen=%llu)!\n", + btrfsic_get_block_type(state, block), + (unsigned long long)bytenr, + dev_state->name, + (unsigned long long)dev_bytenr, + block->mirror_num, + (unsigned long long)block->generation, + (unsigned long long) + le64_to_cpu(block->disk_key.objectid), + block->disk_key.type, + (unsigned long long) + le64_to_cpu(block->disk_key.offset), + (unsigned long long) + le64_to_cpu(((struct btrfs_header *) + mapped_datav[0])->generation), + (unsigned long long) + state->max_superblock_generation); + btrfsic_dump_tree(state); + } + + if (!block->is_iodone && !block->never_written) { + printk(KERN_INFO "btrfs: attempt to overwrite %c-block" + " @%llu (%s/%llu/%d), oldgen=%llu, newgen=%llu," + " which is not yet iodone!\n", + btrfsic_get_block_type(state, block), + (unsigned long long)bytenr, + dev_state->name, + (unsigned long long)dev_bytenr, + block->mirror_num, + (unsigned long long)block->generation, + (unsigned long long) + le64_to_cpu(((struct btrfs_header *) + mapped_datav[0])->generation)); + /* it would not be safe to go on */ + btrfsic_dump_tree(state); + goto continue_loop; + } + + /* + * Clear all references of this block. Do not free + * the block itself even if is not referenced anymore + * because it still carries valueable information + * like whether it was ever written and IO completed. + */ + list_for_each_safe(elem_ref_to, tmp_ref_to, + &block->ref_to_list) { + struct btrfsic_block_link *const l = + list_entry(elem_ref_to, + struct btrfsic_block_link, + node_ref_to); + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + btrfsic_print_rem_link(state, l); + l->ref_cnt--; + if (0 == l->ref_cnt) { + list_del(&l->node_ref_to); + list_del(&l->node_ref_from); + btrfsic_block_link_hashtable_remove(l); + btrfsic_block_link_free(l); + } + } + + if (block->is_superblock) + ret = btrfsic_map_superblock(state, bytenr, + processed_len, + bdev, &block_ctx); + else + ret = btrfsic_map_block(state, bytenr, processed_len, + &block_ctx, 0); + if (ret) { + printk(KERN_INFO + "btrfsic: btrfsic_map_block(root @%llu)" + " failed!\n", (unsigned long long)bytenr); + goto continue_loop; + } + block_ctx.datav = mapped_datav; + /* the following is required in case of writes to mirrors, + * use the same that was used for the lookup */ + block_ctx.dev = dev_state; + block_ctx.dev_bytenr = dev_bytenr; + + if (is_metadata || state->include_extent_data) { + block->never_written = 0; + block->iodone_w_error = 0; + if (NULL != bio) { + block->is_iodone = 0; + BUG_ON(NULL == bio_is_patched); + if (!*bio_is_patched) { + block->orig_bio_bh_private = + bio->bi_private; + block->orig_bio_bh_end_io.bio = + bio->bi_end_io; + block->next_in_same_bio = NULL; + bio->bi_private = block; + bio->bi_end_io = btrfsic_bio_end_io; + *bio_is_patched = 1; + } else { + struct btrfsic_block *chained_block = + (struct btrfsic_block *) + bio->bi_private; + + BUG_ON(NULL == chained_block); + block->orig_bio_bh_private = + chained_block->orig_bio_bh_private; + block->orig_bio_bh_end_io.bio = + chained_block->orig_bio_bh_end_io. + bio; + block->next_in_same_bio = chained_block; + bio->bi_private = block; + } + } else if (NULL != bh) { + block->is_iodone = 0; + block->orig_bio_bh_private = bh->b_private; + block->orig_bio_bh_end_io.bh = bh->b_end_io; + block->next_in_same_bio = NULL; + bh->b_private = block; + bh->b_end_io = btrfsic_bh_end_io; + } else { + block->is_iodone = 1; + block->orig_bio_bh_private = NULL; + block->orig_bio_bh_end_io.bio = NULL; + block->next_in_same_bio = NULL; + } + } + + block->flush_gen = dev_state->last_flush_gen + 1; + block->submit_bio_bh_rw = submit_bio_bh_rw; + if (is_metadata) { + block->logical_bytenr = bytenr; + block->is_metadata = 1; + if (block->is_superblock) { + BUG_ON(PAGE_CACHE_SIZE != + BTRFS_SUPER_INFO_SIZE); + ret = btrfsic_process_written_superblock( + state, + block, + (struct btrfs_super_block *) + mapped_datav[0]); + if (state->print_mask & + BTRFSIC_PRINT_MASK_TREE_AFTER_SB_WRITE) { + printk(KERN_INFO + "[after new superblock is written]:\n"); + btrfsic_dump_tree_sub(state, block, 0); + } + } else { + block->mirror_num = 0; /* unknown */ + ret = btrfsic_process_metablock( + state, + block, + &block_ctx, + 0, 0); + } + if (ret) + printk(KERN_INFO + "btrfsic: btrfsic_process_metablock" + "(root @%llu) failed!\n", + (unsigned long long)dev_bytenr); + } else { + block->is_metadata = 0; + block->mirror_num = 0; /* unknown */ + block->generation = BTRFSIC_GENERATION_UNKNOWN; + if (!state->include_extent_data + && list_empty(&block->ref_from_list)) { + /* + * disk block is overwritten with extent + * data (not meta data) and we are configured + * to not include extent data: take the + * chance and free the block's memory + */ + btrfsic_block_hashtable_remove(block); + list_del(&block->all_blocks_node); + btrfsic_block_free(block); + } + } + btrfsic_release_block_ctx(&block_ctx); + } else { + /* block has not been found in hash table */ + u64 bytenr; + + if (!is_metadata) { + processed_len = state->datablock_size; + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO "Written block (%s/%llu/?)" + " !found in hash table, D.\n", + dev_state->name, + (unsigned long long)dev_bytenr); + if (!state->include_extent_data) { + /* ignore that written D block */ + goto continue_loop; + } + + /* this is getting ugly for the + * include_extent_data case... */ + bytenr = 0; /* unknown */ + block_ctx.start = bytenr; + block_ctx.len = processed_len; + block_ctx.mem_to_free = NULL; + block_ctx.pagev = NULL; + } else { + processed_len = state->metablock_size; + bytenr = le64_to_cpu(((struct btrfs_header *) + mapped_datav[0])->bytenr); + btrfsic_cmp_log_and_dev_bytenr(state, bytenr, dev_state, + dev_bytenr); + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "Written block @%llu (%s/%llu/?)" + " !found in hash table, M.\n", + (unsigned long long)bytenr, + dev_state->name, + (unsigned long long)dev_bytenr); + + ret = btrfsic_map_block(state, bytenr, processed_len, + &block_ctx, 0); + if (ret) { + printk(KERN_INFO + "btrfsic: btrfsic_map_block(root @%llu)" + " failed!\n", + (unsigned long long)dev_bytenr); + goto continue_loop; + } + } + block_ctx.datav = mapped_datav; + /* the following is required in case of writes to mirrors, + * use the same that was used for the lookup */ + block_ctx.dev = dev_state; + block_ctx.dev_bytenr = dev_bytenr; + + block = btrfsic_block_alloc(); + if (NULL == block) { + printk(KERN_INFO "btrfsic: error, kmalloc failed!\n"); + btrfsic_release_block_ctx(&block_ctx); + goto continue_loop; + } + block->dev_state = dev_state; + block->dev_bytenr = dev_bytenr; + block->logical_bytenr = bytenr; + block->is_metadata = is_metadata; + block->never_written = 0; + block->iodone_w_error = 0; + block->mirror_num = 0; /* unknown */ + block->flush_gen = dev_state->last_flush_gen + 1; + block->submit_bio_bh_rw = submit_bio_bh_rw; + if (NULL != bio) { + block->is_iodone = 0; + BUG_ON(NULL == bio_is_patched); + if (!*bio_is_patched) { + block->orig_bio_bh_private = bio->bi_private; + block->orig_bio_bh_end_io.bio = bio->bi_end_io; + block->next_in_same_bio = NULL; + bio->bi_private = block; + bio->bi_end_io = btrfsic_bio_end_io; + *bio_is_patched = 1; + } else { + struct btrfsic_block *chained_block = + (struct btrfsic_block *) + bio->bi_private; + + BUG_ON(NULL == chained_block); + block->orig_bio_bh_private = + chained_block->orig_bio_bh_private; + block->orig_bio_bh_end_io.bio = + chained_block->orig_bio_bh_end_io.bio; + block->next_in_same_bio = chained_block; + bio->bi_private = block; + } + } else if (NULL != bh) { + block->is_iodone = 0; + block->orig_bio_bh_private = bh->b_private; + block->orig_bio_bh_end_io.bh = bh->b_end_io; + block->next_in_same_bio = NULL; + bh->b_private = block; + bh->b_end_io = btrfsic_bh_end_io; + } else { + block->is_iodone = 1; + block->orig_bio_bh_private = NULL; + block->orig_bio_bh_end_io.bio = NULL; + block->next_in_same_bio = NULL; + } + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "New written %c-block @%llu (%s/%llu/%d)\n", + is_metadata ? 'M' : 'D', + (unsigned long long)block->logical_bytenr, + block->dev_state->name, + (unsigned long long)block->dev_bytenr, + block->mirror_num); + list_add(&block->all_blocks_node, &state->all_blocks_list); + btrfsic_block_hashtable_add(block, &state->block_hashtable); + + if (is_metadata) { + ret = btrfsic_process_metablock(state, block, + &block_ctx, 0, 0); + if (ret) + printk(KERN_INFO + "btrfsic: process_metablock(root @%llu)" + " failed!\n", + (unsigned long long)dev_bytenr); + } + btrfsic_release_block_ctx(&block_ctx); + } + +continue_loop: + BUG_ON(!processed_len); + dev_bytenr += processed_len; + mapped_datav += processed_len >> PAGE_CACHE_SHIFT; + num_pages -= processed_len >> PAGE_CACHE_SHIFT; + goto again; +} + +static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status) +{ + struct btrfsic_block *block = (struct btrfsic_block *)bp->bi_private; + int iodone_w_error; + + /* mutex is not held! This is not save if IO is not yet completed + * on umount */ + iodone_w_error = 0; + if (bio_error_status) + iodone_w_error = 1; + + BUG_ON(NULL == block); + bp->bi_private = block->orig_bio_bh_private; + bp->bi_end_io = block->orig_bio_bh_end_io.bio; + + do { + struct btrfsic_block *next_block; + struct btrfsic_dev_state *const dev_state = block->dev_state; + + if ((dev_state->state->print_mask & + BTRFSIC_PRINT_MASK_END_IO_BIO_BH)) + printk(KERN_INFO + "bio_end_io(err=%d) for %c @%llu (%s/%llu/%d)\n", + bio_error_status, + btrfsic_get_block_type(dev_state->state, block), + (unsigned long long)block->logical_bytenr, + dev_state->name, + (unsigned long long)block->dev_bytenr, + block->mirror_num); + next_block = block->next_in_same_bio; + block->iodone_w_error = iodone_w_error; + if (block->submit_bio_bh_rw & REQ_FLUSH) { + dev_state->last_flush_gen++; + if ((dev_state->state->print_mask & + BTRFSIC_PRINT_MASK_END_IO_BIO_BH)) + printk(KERN_INFO + "bio_end_io() new %s flush_gen=%llu\n", + dev_state->name, + (unsigned long long) + dev_state->last_flush_gen); + } + if (block->submit_bio_bh_rw & REQ_FUA) + block->flush_gen = 0; /* FUA completed means block is + * on disk */ + block->is_iodone = 1; /* for FLUSH, this releases the block */ + block = next_block; + } while (NULL != block); + + bp->bi_end_io(bp, bio_error_status); +} + +static void btrfsic_bh_end_io(struct buffer_head *bh, int uptodate) +{ + struct btrfsic_block *block = (struct btrfsic_block *)bh->b_private; + int iodone_w_error = !uptodate; + struct btrfsic_dev_state *dev_state; + + BUG_ON(NULL == block); + dev_state = block->dev_state; + if ((dev_state->state->print_mask & BTRFSIC_PRINT_MASK_END_IO_BIO_BH)) + printk(KERN_INFO + "bh_end_io(error=%d) for %c @%llu (%s/%llu/%d)\n", + iodone_w_error, + btrfsic_get_block_type(dev_state->state, block), + (unsigned long long)block->logical_bytenr, + block->dev_state->name, + (unsigned long long)block->dev_bytenr, + block->mirror_num); + + block->iodone_w_error = iodone_w_error; + if (block->submit_bio_bh_rw & REQ_FLUSH) { + dev_state->last_flush_gen++; + if ((dev_state->state->print_mask & + BTRFSIC_PRINT_MASK_END_IO_BIO_BH)) + printk(KERN_INFO + "bh_end_io() new %s flush_gen=%llu\n", + dev_state->name, + (unsigned long long)dev_state->last_flush_gen); + } + if (block->submit_bio_bh_rw & REQ_FUA) + block->flush_gen = 0; /* FUA completed means block is on disk */ + + bh->b_private = block->orig_bio_bh_private; + bh->b_end_io = block->orig_bio_bh_end_io.bh; + block->is_iodone = 1; /* for FLUSH, this releases the block */ + bh->b_end_io(bh, uptodate); +} + +static int btrfsic_process_written_superblock( + struct btrfsic_state *state, + struct btrfsic_block *const superblock, + struct btrfs_super_block *const super_hdr) +{ + int pass; + + superblock->generation = btrfs_super_generation(super_hdr); + if (!(superblock->generation > state->max_superblock_generation || + 0 == state->max_superblock_generation)) { + if (state->print_mask & BTRFSIC_PRINT_MASK_SUPERBLOCK_WRITE) + printk(KERN_INFO + "btrfsic: superblock @%llu (%s/%llu/%d)" + " with old gen %llu <= %llu\n", + (unsigned long long)superblock->logical_bytenr, + superblock->dev_state->name, + (unsigned long long)superblock->dev_bytenr, + superblock->mirror_num, + (unsigned long long) + btrfs_super_generation(super_hdr), + (unsigned long long) + state->max_superblock_generation); + } else { + if (state->print_mask & BTRFSIC_PRINT_MASK_SUPERBLOCK_WRITE) + printk(KERN_INFO + "btrfsic: got new superblock @%llu (%s/%llu/%d)" + " with new gen %llu > %llu\n", + (unsigned long long)superblock->logical_bytenr, + superblock->dev_state->name, + (unsigned long long)superblock->dev_bytenr, + superblock->mirror_num, + (unsigned long long) + btrfs_super_generation(super_hdr), + (unsigned long long) + state->max_superblock_generation); + + state->max_superblock_generation = + btrfs_super_generation(super_hdr); + state->latest_superblock = superblock; + } + + for (pass = 0; pass < 3; pass++) { + int ret; + u64 next_bytenr; + struct btrfsic_block *next_block; + struct btrfsic_block_data_ctx tmp_next_block_ctx; + struct btrfsic_block_link *l; + int num_copies; + int mirror_num; + const char *additional_string = NULL; + struct btrfs_disk_key tmp_disk_key; + + tmp_disk_key.type = BTRFS_ROOT_ITEM_KEY; + tmp_disk_key.offset = 0; + + switch (pass) { + case 0: + tmp_disk_key.objectid = + cpu_to_le64(BTRFS_ROOT_TREE_OBJECTID); + additional_string = "root "; + next_bytenr = btrfs_super_root(super_hdr); + if (state->print_mask & + BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION) + printk(KERN_INFO "root@%llu\n", + (unsigned long long)next_bytenr); + break; + case 1: + tmp_disk_key.objectid = + cpu_to_le64(BTRFS_CHUNK_TREE_OBJECTID); + additional_string = "chunk "; + next_bytenr = btrfs_super_chunk_root(super_hdr); + if (state->print_mask & + BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION) + printk(KERN_INFO "chunk@%llu\n", + (unsigned long long)next_bytenr); + break; + case 2: + tmp_disk_key.objectid = + cpu_to_le64(BTRFS_TREE_LOG_OBJECTID); + additional_string = "log "; + next_bytenr = btrfs_super_log_root(super_hdr); + if (0 == next_bytenr) + continue; + if (state->print_mask & + BTRFSIC_PRINT_MASK_ROOT_CHUNK_LOG_TREE_LOCATION) + printk(KERN_INFO "log@%llu\n", + (unsigned long long)next_bytenr); + break; + } + + num_copies = + btrfs_num_copies(&state->root->fs_info->mapping_tree, + next_bytenr, BTRFS_SUPER_INFO_SIZE); + if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) + printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", + (unsigned long long)next_bytenr, num_copies); + for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) { + int was_created; + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "btrfsic_process_written_superblock(" + "mirror_num=%d)\n", mirror_num); + ret = btrfsic_map_block(state, next_bytenr, + BTRFS_SUPER_INFO_SIZE, + &tmp_next_block_ctx, + mirror_num); + if (ret) { + printk(KERN_INFO + "btrfsic: btrfsic_map_block(@%llu," + " mirror=%d) failed!\n", + (unsigned long long)next_bytenr, + mirror_num); + return -1; + } + + next_block = btrfsic_block_lookup_or_add( + state, + &tmp_next_block_ctx, + additional_string, + 1, 0, 1, + mirror_num, + &was_created); + if (NULL == next_block) { + printk(KERN_INFO + "btrfsic: error, kmalloc failed!\n"); + btrfsic_release_block_ctx(&tmp_next_block_ctx); + return -1; + } + + next_block->disk_key = tmp_disk_key; + if (was_created) + next_block->generation = + BTRFSIC_GENERATION_UNKNOWN; + l = btrfsic_block_link_lookup_or_add( + state, + &tmp_next_block_ctx, + next_block, + superblock, + BTRFSIC_GENERATION_UNKNOWN); + btrfsic_release_block_ctx(&tmp_next_block_ctx); + if (NULL == l) + return -1; + } + } + + if (-1 == btrfsic_check_all_ref_blocks(state, superblock, 0)) { + WARN_ON(1); + btrfsic_dump_tree(state); + } + + return 0; +} + +static int btrfsic_check_all_ref_blocks(struct btrfsic_state *state, + struct btrfsic_block *const block, + int recursion_level) +{ + struct list_head *elem_ref_to; + int ret = 0; + + if (recursion_level >= 3 + BTRFS_MAX_LEVEL) { + /* + * Note that this situation can happen and does not + * indicate an error in regular cases. It happens + * when disk blocks are freed and later reused. + * The check-integrity module is not aware of any + * block free operations, it just recognizes block + * write operations. Therefore it keeps the linkage + * information for a block until a block is + * rewritten. This can temporarily cause incorrect + * and even circular linkage informations. This + * causes no harm unless such blocks are referenced + * by the most recent super block. + */ + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "btrfsic: abort cyclic linkage (case 1).\n"); + + return ret; + } + + /* + * This algorithm is recursive because the amount of used stack + * space is very small and the max recursion depth is limited. + */ + list_for_each(elem_ref_to, &block->ref_to_list) { + const struct btrfsic_block_link *const l = + list_entry(elem_ref_to, struct btrfsic_block_link, + node_ref_to); + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "rl=%d, %c @%llu (%s/%llu/%d)" + " %u* refers to %c @%llu (%s/%llu/%d)\n", + recursion_level, + btrfsic_get_block_type(state, block), + (unsigned long long)block->logical_bytenr, + block->dev_state->name, + (unsigned long long)block->dev_bytenr, + block->mirror_num, + l->ref_cnt, + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long) + l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num); + if (l->block_ref_to->never_written) { + printk(KERN_INFO "btrfs: attempt to write superblock" + " which references block %c @%llu (%s/%llu/%d)" + " which is never written!\n", + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long) + l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num); + ret = -1; + } else if (!l->block_ref_to->is_iodone) { + printk(KERN_INFO "btrfs: attempt to write superblock" + " which references block %c @%llu (%s/%llu/%d)" + " which is not yet iodone!\n", + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long) + l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num); + ret = -1; + } else if (l->parent_generation != + l->block_ref_to->generation && + BTRFSIC_GENERATION_UNKNOWN != + l->parent_generation && + BTRFSIC_GENERATION_UNKNOWN != + l->block_ref_to->generation) { + printk(KERN_INFO "btrfs: attempt to write superblock" + " which references block %c @%llu (%s/%llu/%d)" + " with generation %llu !=" + " parent generation %llu!\n", + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long) + l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num, + (unsigned long long)l->block_ref_to->generation, + (unsigned long long)l->parent_generation); + ret = -1; + } else if (l->block_ref_to->flush_gen > + l->block_ref_to->dev_state->last_flush_gen) { + printk(KERN_INFO "btrfs: attempt to write superblock" + " which references block %c @%llu (%s/%llu/%d)" + " which is not flushed out of disk's write cache" + " (block flush_gen=%llu," + " dev->flush_gen=%llu)!\n", + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long) + l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num, + (unsigned long long)block->flush_gen, + (unsigned long long) + l->block_ref_to->dev_state->last_flush_gen); + ret = -1; + } else if (-1 == btrfsic_check_all_ref_blocks(state, + l->block_ref_to, + recursion_level + + 1)) { + ret = -1; + } + } + + return ret; +} + +static int btrfsic_is_block_ref_by_superblock( + const struct btrfsic_state *state, + const struct btrfsic_block *block, + int recursion_level) +{ + struct list_head *elem_ref_from; + + if (recursion_level >= 3 + BTRFS_MAX_LEVEL) { + /* refer to comment at "abort cyclic linkage (case 1)" */ + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "btrfsic: abort cyclic linkage (case 2).\n"); + + return 0; + } + + /* + * This algorithm is recursive because the amount of used stack space + * is very small and the max recursion depth is limited. + */ + list_for_each(elem_ref_from, &block->ref_from_list) { + const struct btrfsic_block_link *const l = + list_entry(elem_ref_from, struct btrfsic_block_link, + node_ref_from); + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "rl=%d, %c @%llu (%s/%llu/%d)" + " is ref %u* from %c @%llu (%s/%llu/%d)\n", + recursion_level, + btrfsic_get_block_type(state, block), + (unsigned long long)block->logical_bytenr, + block->dev_state->name, + (unsigned long long)block->dev_bytenr, + block->mirror_num, + l->ref_cnt, + btrfsic_get_block_type(state, l->block_ref_from), + (unsigned long long) + l->block_ref_from->logical_bytenr, + l->block_ref_from->dev_state->name, + (unsigned long long) + l->block_ref_from->dev_bytenr, + l->block_ref_from->mirror_num); + if (l->block_ref_from->is_superblock && + state->latest_superblock->dev_bytenr == + l->block_ref_from->dev_bytenr && + state->latest_superblock->dev_state->bdev == + l->block_ref_from->dev_state->bdev) + return 1; + else if (btrfsic_is_block_ref_by_superblock(state, + l->block_ref_from, + recursion_level + + 1)) + return 1; + } + + return 0; +} + +static void btrfsic_print_add_link(const struct btrfsic_state *state, + const struct btrfsic_block_link *l) +{ + printk(KERN_INFO + "Add %u* link from %c @%llu (%s/%llu/%d)" + " to %c @%llu (%s/%llu/%d).\n", + l->ref_cnt, + btrfsic_get_block_type(state, l->block_ref_from), + (unsigned long long)l->block_ref_from->logical_bytenr, + l->block_ref_from->dev_state->name, + (unsigned long long)l->block_ref_from->dev_bytenr, + l->block_ref_from->mirror_num, + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long)l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num); +} + +static void btrfsic_print_rem_link(const struct btrfsic_state *state, + const struct btrfsic_block_link *l) +{ + printk(KERN_INFO + "Rem %u* link from %c @%llu (%s/%llu/%d)" + " to %c @%llu (%s/%llu/%d).\n", + l->ref_cnt, + btrfsic_get_block_type(state, l->block_ref_from), + (unsigned long long)l->block_ref_from->logical_bytenr, + l->block_ref_from->dev_state->name, + (unsigned long long)l->block_ref_from->dev_bytenr, + l->block_ref_from->mirror_num, + btrfsic_get_block_type(state, l->block_ref_to), + (unsigned long long)l->block_ref_to->logical_bytenr, + l->block_ref_to->dev_state->name, + (unsigned long long)l->block_ref_to->dev_bytenr, + l->block_ref_to->mirror_num); +} + +static char btrfsic_get_block_type(const struct btrfsic_state *state, + const struct btrfsic_block *block) +{ + if (block->is_superblock && + state->latest_superblock->dev_bytenr == block->dev_bytenr && + state->latest_superblock->dev_state->bdev == block->dev_state->bdev) + return 'S'; + else if (block->is_superblock) + return 's'; + else if (block->is_metadata) + return 'M'; + else + return 'D'; +} + +static void btrfsic_dump_tree(const struct btrfsic_state *state) +{ + btrfsic_dump_tree_sub(state, state->latest_superblock, 0); +} + +static void btrfsic_dump_tree_sub(const struct btrfsic_state *state, + const struct btrfsic_block *block, + int indent_level) +{ + struct list_head *elem_ref_to; + int indent_add; + static char buf[80]; + int cursor_position; + + /* + * Should better fill an on-stack buffer with a complete line and + * dump it at once when it is time to print a newline character. + */ + + /* + * This algorithm is recursive because the amount of used stack space + * is very small and the max recursion depth is limited. + */ + indent_add = sprintf(buf, "%c-%llu(%s/%llu/%d)", + btrfsic_get_block_type(state, block), + (unsigned long long)block->logical_bytenr, + block->dev_state->name, + (unsigned long long)block->dev_bytenr, + block->mirror_num); + if (indent_level + indent_add > BTRFSIC_TREE_DUMP_MAX_INDENT_LEVEL) { + printk("[...]\n"); + return; + } + printk(buf); + indent_level += indent_add; + if (list_empty(&block->ref_to_list)) { + printk("\n"); + return; + } + if (block->mirror_num > 1 && + !(state->print_mask & BTRFSIC_PRINT_MASK_TREE_WITH_ALL_MIRRORS)) { + printk(" [...]\n"); + return; + } + + cursor_position = indent_level; + list_for_each(elem_ref_to, &block->ref_to_list) { + const struct btrfsic_block_link *const l = + list_entry(elem_ref_to, struct btrfsic_block_link, + node_ref_to); + + while (cursor_position < indent_level) { + printk(" "); + cursor_position++; + } + if (l->ref_cnt > 1) + indent_add = sprintf(buf, " %d*--> ", l->ref_cnt); + else + indent_add = sprintf(buf, " --> "); + if (indent_level + indent_add > + BTRFSIC_TREE_DUMP_MAX_INDENT_LEVEL) { + printk("[...]\n"); + cursor_position = 0; + continue; + } + + printk(buf); + + btrfsic_dump_tree_sub(state, l->block_ref_to, + indent_level + indent_add); + cursor_position = 0; + } +} + +static struct btrfsic_block_link *btrfsic_block_link_lookup_or_add( + struct btrfsic_state *state, + struct btrfsic_block_data_ctx *next_block_ctx, + struct btrfsic_block *next_block, + struct btrfsic_block *from_block, + u64 parent_generation) +{ + struct btrfsic_block_link *l; + + l = btrfsic_block_link_hashtable_lookup(next_block_ctx->dev->bdev, + next_block_ctx->dev_bytenr, + from_block->dev_state->bdev, + from_block->dev_bytenr, + &state->block_link_hashtable); + if (NULL == l) { + l = btrfsic_block_link_alloc(); + if (NULL == l) { + printk(KERN_INFO + "btrfsic: error, kmalloc" " failed!\n"); + return NULL; + } + + l->block_ref_to = next_block; + l->block_ref_from = from_block; + l->ref_cnt = 1; + l->parent_generation = parent_generation; + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + btrfsic_print_add_link(state, l); + + list_add(&l->node_ref_to, &from_block->ref_to_list); + list_add(&l->node_ref_from, &next_block->ref_from_list); + + btrfsic_block_link_hashtable_add(l, + &state->block_link_hashtable); + } else { + l->ref_cnt++; + l->parent_generation = parent_generation; + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + btrfsic_print_add_link(state, l); + } + + return l; +} + +static struct btrfsic_block *btrfsic_block_lookup_or_add( + struct btrfsic_state *state, + struct btrfsic_block_data_ctx *block_ctx, + const char *additional_string, + int is_metadata, + int is_iodone, + int never_written, + int mirror_num, + int *was_created) +{ + struct btrfsic_block *block; + + block = btrfsic_block_hashtable_lookup(block_ctx->dev->bdev, + block_ctx->dev_bytenr, + &state->block_hashtable); + if (NULL == block) { + struct btrfsic_dev_state *dev_state; + + block = btrfsic_block_alloc(); + if (NULL == block) { + printk(KERN_INFO "btrfsic: error, kmalloc failed!\n"); + return NULL; + } + dev_state = btrfsic_dev_state_lookup(block_ctx->dev->bdev); + if (NULL == dev_state) { + printk(KERN_INFO + "btrfsic: error, lookup dev_state failed!\n"); + btrfsic_block_free(block); + return NULL; + } + block->dev_state = dev_state; + block->dev_bytenr = block_ctx->dev_bytenr; + block->logical_bytenr = block_ctx->start; + block->is_metadata = is_metadata; + block->is_iodone = is_iodone; + block->never_written = never_written; + block->mirror_num = mirror_num; + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + printk(KERN_INFO + "New %s%c-block @%llu (%s/%llu/%d)\n", + additional_string, + btrfsic_get_block_type(state, block), + (unsigned long long)block->logical_bytenr, + dev_state->name, + (unsigned long long)block->dev_bytenr, + mirror_num); + list_add(&block->all_blocks_node, &state->all_blocks_list); + btrfsic_block_hashtable_add(block, &state->block_hashtable); + if (NULL != was_created) + *was_created = 1; + } else { + if (NULL != was_created) + *was_created = 0; + } + + return block; +} + +static void btrfsic_cmp_log_and_dev_bytenr(struct btrfsic_state *state, + u64 bytenr, + struct btrfsic_dev_state *dev_state, + u64 dev_bytenr) +{ + int num_copies; + int mirror_num; + int ret; + struct btrfsic_block_data_ctx block_ctx; + int match = 0; + + num_copies = btrfs_num_copies(&state->root->fs_info->mapping_tree, + bytenr, state->metablock_size); + + for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) { + ret = btrfsic_map_block(state, bytenr, state->metablock_size, + &block_ctx, mirror_num); + if (ret) { + printk(KERN_INFO "btrfsic:" + " btrfsic_map_block(logical @%llu," + " mirror %d) failed!\n", + (unsigned long long)bytenr, mirror_num); + continue; + } + + if (dev_state->bdev == block_ctx.dev->bdev && + dev_bytenr == block_ctx.dev_bytenr) { + match++; + btrfsic_release_block_ctx(&block_ctx); + break; + } + btrfsic_release_block_ctx(&block_ctx); + } + + if (!match) { + printk(KERN_INFO "btrfs: attempt to write M-block which contains logical bytenr that doesn't map to dev+physical bytenr of submit_bio," + " buffer->log_bytenr=%llu, submit_bio(bdev=%s," + " phys_bytenr=%llu)!\n", + (unsigned long long)bytenr, dev_state->name, + (unsigned long long)dev_bytenr); + for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) { + ret = btrfsic_map_block(state, bytenr, + state->metablock_size, + &block_ctx, mirror_num); + if (ret) + continue; + + printk(KERN_INFO "Read logical bytenr @%llu maps to" + " (%s/%llu/%d)\n", + (unsigned long long)bytenr, + block_ctx.dev->name, + (unsigned long long)block_ctx.dev_bytenr, + mirror_num); + } + WARN_ON(1); + } +} + +static struct btrfsic_dev_state *btrfsic_dev_state_lookup( + struct block_device *bdev) +{ + struct btrfsic_dev_state *ds; + + ds = btrfsic_dev_state_hashtable_lookup(bdev, + &btrfsic_dev_state_hashtable); + return ds; +} + +int btrfsic_submit_bh(int rw, struct buffer_head *bh) +{ + struct btrfsic_dev_state *dev_state; + + if (!btrfsic_is_initialized) + return submit_bh(rw, bh); + + mutex_lock(&btrfsic_mutex); + /* since btrfsic_submit_bh() might also be called before + * btrfsic_mount(), this might return NULL */ + dev_state = btrfsic_dev_state_lookup(bh->b_bdev); + + /* Only called to write the superblock (incl. FLUSH/FUA) */ + if (NULL != dev_state && + (rw & WRITE) && bh->b_size > 0) { + u64 dev_bytenr; + + dev_bytenr = 4096 * bh->b_blocknr; + if (dev_state->state->print_mask & + BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH) + printk(KERN_INFO + "submit_bh(rw=0x%x, blocknr=%lu (bytenr %llu)," + " size=%lu, data=%p, bdev=%p)\n", + rw, (unsigned long)bh->b_blocknr, + (unsigned long long)dev_bytenr, + (unsigned long)bh->b_size, bh->b_data, + bh->b_bdev); + btrfsic_process_written_block(dev_state, dev_bytenr, + &bh->b_data, 1, NULL, + NULL, bh, rw); + } else if (NULL != dev_state && (rw & REQ_FLUSH)) { + if (dev_state->state->print_mask & + BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH) + printk(KERN_INFO + "submit_bh(rw=0x%x FLUSH, bdev=%p)\n", + rw, bh->b_bdev); + if (!dev_state->dummy_block_for_bio_bh_flush.is_iodone) { + if ((dev_state->state->print_mask & + (BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH | + BTRFSIC_PRINT_MASK_VERBOSE))) + printk(KERN_INFO + "btrfsic_submit_bh(%s) with FLUSH" + " but dummy block already in use" + " (ignored)!\n", + dev_state->name); + } else { + struct btrfsic_block *const block = + &dev_state->dummy_block_for_bio_bh_flush; + + block->is_iodone = 0; + block->never_written = 0; + block->iodone_w_error = 0; + block->flush_gen = dev_state->last_flush_gen + 1; + block->submit_bio_bh_rw = rw; + block->orig_bio_bh_private = bh->b_private; + block->orig_bio_bh_end_io.bh = bh->b_end_io; + block->next_in_same_bio = NULL; + bh->b_private = block; + bh->b_end_io = btrfsic_bh_end_io; + } + } + mutex_unlock(&btrfsic_mutex); + return submit_bh(rw, bh); +} + +void btrfsic_submit_bio(int rw, struct bio *bio) +{ + struct btrfsic_dev_state *dev_state; + + if (!btrfsic_is_initialized) { + submit_bio(rw, bio); + return; + } + + mutex_lock(&btrfsic_mutex); + /* since btrfsic_submit_bio() is also called before + * btrfsic_mount(), this might return NULL */ + dev_state = btrfsic_dev_state_lookup(bio->bi_bdev); + if (NULL != dev_state && + (rw & WRITE) && NULL != bio->bi_io_vec) { + unsigned int i; + u64 dev_bytenr; + int bio_is_patched; + char **mapped_datav; + + dev_bytenr = 512 * bio->bi_sector; + bio_is_patched = 0; + if (dev_state->state->print_mask & + BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH) + printk(KERN_INFO + "submit_bio(rw=0x%x, bi_vcnt=%u," + " bi_sector=%lu (bytenr %llu), bi_bdev=%p)\n", + rw, bio->bi_vcnt, (unsigned long)bio->bi_sector, + (unsigned long long)dev_bytenr, + bio->bi_bdev); + + mapped_datav = kmalloc(sizeof(*mapped_datav) * bio->bi_vcnt, + GFP_NOFS); + if (!mapped_datav) + goto leave; + for (i = 0; i < bio->bi_vcnt; i++) { + BUG_ON(bio->bi_io_vec[i].bv_len != PAGE_CACHE_SIZE); + mapped_datav[i] = kmap(bio->bi_io_vec[i].bv_page); + if (!mapped_datav[i]) { + while (i > 0) { + i--; + kunmap(bio->bi_io_vec[i].bv_page); + } + kfree(mapped_datav); + goto leave; + } + if ((BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH | + BTRFSIC_PRINT_MASK_VERBOSE) == + (dev_state->state->print_mask & + (BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH | + BTRFSIC_PRINT_MASK_VERBOSE))) + printk(KERN_INFO + "#%u: page=%p, len=%u, offset=%u\n", + i, bio->bi_io_vec[i].bv_page, + bio->bi_io_vec[i].bv_len, + bio->bi_io_vec[i].bv_offset); + } + btrfsic_process_written_block(dev_state, dev_bytenr, + mapped_datav, bio->bi_vcnt, + bio, &bio_is_patched, + NULL, rw); + while (i > 0) { + i--; + kunmap(bio->bi_io_vec[i].bv_page); + } + kfree(mapped_datav); + } else if (NULL != dev_state && (rw & REQ_FLUSH)) { + if (dev_state->state->print_mask & + BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH) + printk(KERN_INFO + "submit_bio(rw=0x%x FLUSH, bdev=%p)\n", + rw, bio->bi_bdev); + if (!dev_state->dummy_block_for_bio_bh_flush.is_iodone) { + if ((dev_state->state->print_mask & + (BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH | + BTRFSIC_PRINT_MASK_VERBOSE))) + printk(KERN_INFO + "btrfsic_submit_bio(%s) with FLUSH" + " but dummy block already in use" + " (ignored)!\n", + dev_state->name); + } else { + struct btrfsic_block *const block = + &dev_state->dummy_block_for_bio_bh_flush; + + block->is_iodone = 0; + block->never_written = 0; + block->iodone_w_error = 0; + block->flush_gen = dev_state->last_flush_gen + 1; + block->submit_bio_bh_rw = rw; + block->orig_bio_bh_private = bio->bi_private; + block->orig_bio_bh_end_io.bio = bio->bi_end_io; + block->next_in_same_bio = NULL; + bio->bi_private = block; + bio->bi_end_io = btrfsic_bio_end_io; + } + } +leave: + mutex_unlock(&btrfsic_mutex); + + submit_bio(rw, bio); +} + +int btrfsic_mount(struct btrfs_root *root, + struct btrfs_fs_devices *fs_devices, + int including_extent_data, u32 print_mask) +{ + int ret; + struct btrfsic_state *state; + struct list_head *dev_head = &fs_devices->devices; + struct btrfs_device *device; + + if (root->nodesize != root->leafsize) { + printk(KERN_INFO + "btrfsic: cannot handle nodesize %d != leafsize %d!\n", + root->nodesize, root->leafsize); + return -1; + } + if (root->nodesize & ((u64)PAGE_CACHE_SIZE - 1)) { + printk(KERN_INFO + "btrfsic: cannot handle nodesize %d not being a multiple of PAGE_CACHE_SIZE %ld!\n", + root->nodesize, (unsigned long)PAGE_CACHE_SIZE); + return -1; + } + if (root->leafsize & ((u64)PAGE_CACHE_SIZE - 1)) { + printk(KERN_INFO + "btrfsic: cannot handle leafsize %d not being a multiple of PAGE_CACHE_SIZE %ld!\n", + root->leafsize, (unsigned long)PAGE_CACHE_SIZE); + return -1; + } + if (root->sectorsize & ((u64)PAGE_CACHE_SIZE - 1)) { + printk(KERN_INFO + "btrfsic: cannot handle sectorsize %d not being a multiple of PAGE_CACHE_SIZE %ld!\n", + root->sectorsize, (unsigned long)PAGE_CACHE_SIZE); + return -1; + } + state = kzalloc(sizeof(*state), GFP_NOFS); + if (NULL == state) { + printk(KERN_INFO "btrfs check-integrity: kmalloc() failed!\n"); + return -1; + } + + if (!btrfsic_is_initialized) { + mutex_init(&btrfsic_mutex); + btrfsic_dev_state_hashtable_init(&btrfsic_dev_state_hashtable); + btrfsic_is_initialized = 1; + } + mutex_lock(&btrfsic_mutex); + state->root = root; + state->print_mask = print_mask; + state->include_extent_data = including_extent_data; + state->csum_size = 0; + state->metablock_size = root->nodesize; + state->datablock_size = root->sectorsize; + INIT_LIST_HEAD(&state->all_blocks_list); + btrfsic_block_hashtable_init(&state->block_hashtable); + btrfsic_block_link_hashtable_init(&state->block_link_hashtable); + state->max_superblock_generation = 0; + state->latest_superblock = NULL; + + list_for_each_entry(device, dev_head, dev_list) { + struct btrfsic_dev_state *ds; + char *p; + + if (!device->bdev || !device->name) + continue; + + ds = btrfsic_dev_state_alloc(); + if (NULL == ds) { + printk(KERN_INFO + "btrfs check-integrity: kmalloc() failed!\n"); + mutex_unlock(&btrfsic_mutex); + return -1; + } + ds->bdev = device->bdev; + ds->state = state; + bdevname(ds->bdev, ds->name); + ds->name[BDEVNAME_SIZE - 1] = '\0'; + for (p = ds->name; *p != '\0'; p++); + while (p > ds->name && *p != '/') + p--; + if (*p == '/') + p++; + strlcpy(ds->name, p, sizeof(ds->name)); + btrfsic_dev_state_hashtable_add(ds, + &btrfsic_dev_state_hashtable); + } + + ret = btrfsic_process_superblock(state, fs_devices); + if (0 != ret) { + mutex_unlock(&btrfsic_mutex); + btrfsic_unmount(root, fs_devices); + return ret; + } + + if (state->print_mask & BTRFSIC_PRINT_MASK_INITIAL_DATABASE) + btrfsic_dump_database(state); + if (state->print_mask & BTRFSIC_PRINT_MASK_INITIAL_TREE) + btrfsic_dump_tree(state); + + mutex_unlock(&btrfsic_mutex); + return 0; +} + +void btrfsic_unmount(struct btrfs_root *root, + struct btrfs_fs_devices *fs_devices) +{ + struct list_head *elem_all; + struct list_head *tmp_all; + struct btrfsic_state *state; + struct list_head *dev_head = &fs_devices->devices; + struct btrfs_device *device; + + if (!btrfsic_is_initialized) + return; + + mutex_lock(&btrfsic_mutex); + + state = NULL; + list_for_each_entry(device, dev_head, dev_list) { + struct btrfsic_dev_state *ds; + + if (!device->bdev || !device->name) + continue; + + ds = btrfsic_dev_state_hashtable_lookup( + device->bdev, + &btrfsic_dev_state_hashtable); + if (NULL != ds) { + state = ds->state; + btrfsic_dev_state_hashtable_remove(ds); + btrfsic_dev_state_free(ds); + } + } + + if (NULL == state) { + printk(KERN_INFO + "btrfsic: error, cannot find state information" + " on umount!\n"); + mutex_unlock(&btrfsic_mutex); + return; + } + + /* + * Don't care about keeping the lists' state up to date, + * just free all memory that was allocated dynamically. + * Free the blocks and the block_links. + */ + list_for_each_safe(elem_all, tmp_all, &state->all_blocks_list) { + struct btrfsic_block *const b_all = + list_entry(elem_all, struct btrfsic_block, + all_blocks_node); + struct list_head *elem_ref_to; + struct list_head *tmp_ref_to; + + list_for_each_safe(elem_ref_to, tmp_ref_to, + &b_all->ref_to_list) { + struct btrfsic_block_link *const l = + list_entry(elem_ref_to, + struct btrfsic_block_link, + node_ref_to); + + if (state->print_mask & BTRFSIC_PRINT_MASK_VERBOSE) + btrfsic_print_rem_link(state, l); + + l->ref_cnt--; + if (0 == l->ref_cnt) + btrfsic_block_link_free(l); + } + + if (b_all->is_iodone || b_all->never_written) + btrfsic_block_free(b_all); + else + printk(KERN_INFO "btrfs: attempt to free %c-block" + " @%llu (%s/%llu/%d) on umount which is" + " not yet iodone!\n", + btrfsic_get_block_type(state, b_all), + (unsigned long long)b_all->logical_bytenr, + b_all->dev_state->name, + (unsigned long long)b_all->dev_bytenr, + b_all->mirror_num); + } + + mutex_unlock(&btrfsic_mutex); + + kfree(state); +} diff --git a/fs/btrfs/check-integrity.h b/fs/btrfs/check-integrity.h new file mode 100644 index 000000000000..8b59175cc502 --- /dev/null +++ b/fs/btrfs/check-integrity.h @@ -0,0 +1,36 @@ +/* + * Copyright (C) STRATO AG 2011. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#if !defined(__BTRFS_CHECK_INTEGRITY__) +#define __BTRFS_CHECK_INTEGRITY__ + +#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY +int btrfsic_submit_bh(int rw, struct buffer_head *bh); +void btrfsic_submit_bio(int rw, struct bio *bio); +#else +#define btrfsic_submit_bh submit_bh +#define btrfsic_submit_bio submit_bio +#endif + +int btrfsic_mount(struct btrfs_root *root, + struct btrfs_fs_devices *fs_devices, + int including_extent_data, u32 print_mask); +void btrfsic_unmount(struct btrfs_root *root, + struct btrfs_fs_devices *fs_devices); + +#endif diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 8ec5d86f1734..43d1c5a3a030 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -85,7 +85,8 @@ struct compressed_bio { static inline int compressed_bio_size(struct btrfs_root *root, unsigned long disk_size) { - u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); + return sizeof(struct compressed_bio) + ((disk_size + root->sectorsize - 1) / root->sectorsize) * csum_size; @@ -119,10 +120,10 @@ static int check_compressed_csum(struct inode *inode, page = cb->compressed_pages[i]; csum = ~(u32)0; - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); csum = btrfs_csum_data(root, kaddr, csum, PAGE_CACHE_SIZE); btrfs_csum_final(csum, (char *)&csum); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); if (csum != *cb_sum) { printk(KERN_INFO "btrfs csum failed ino %llu " @@ -225,8 +226,8 @@ out: * Clear the writeback bits on all of the file * pages for a compressed write */ -static noinline int end_compressed_writeback(struct inode *inode, u64 start, - unsigned long ram_size) +static noinline void end_compressed_writeback(struct inode *inode, u64 start, + unsigned long ram_size) { unsigned long index = start >> PAGE_CACHE_SHIFT; unsigned long end_index = (start + ram_size - 1) >> PAGE_CACHE_SHIFT; @@ -252,7 +253,6 @@ static noinline int end_compressed_writeback(struct inode *inode, u64 start, index += ret; } /* the inode may be gone now */ - return 0; } /* @@ -391,20 +391,21 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, */ atomic_inc(&cb->pending_bios); ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ if (!skip_sum) { ret = btrfs_csum_one_bio(root, inode, bio, start, 1); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } ret = btrfs_map_bio(root, WRITE, bio, 0, 1); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ bio_put(bio); bio = compressed_bio_alloc(bdev, first_byte, GFP_NOFS); + BUG_ON(!bio); bio->bi_private = cb; bio->bi_end_io = end_compressed_bio_write; bio_add_page(bio, page, PAGE_CACHE_SIZE, 0); @@ -420,15 +421,15 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, bio_get(bio); ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ if (!skip_sum) { ret = btrfs_csum_one_bio(root, inode, bio, start, 1); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } ret = btrfs_map_bio(root, WRITE, bio, 0, 1); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ bio_put(bio); return 0; @@ -496,7 +497,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, * sure they map to this compressed extent on disk. */ set_page_extent_mapped(page); - lock_extent(tree, last_offset, end, GFP_NOFS); + lock_extent(tree, last_offset, end); read_lock(&em_tree->lock); em = lookup_extent_mapping(em_tree, last_offset, PAGE_CACHE_SIZE); @@ -506,7 +507,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, (last_offset + PAGE_CACHE_SIZE > extent_map_end(em)) || (em->block_start >> 9) != cb->orig_bio->bi_sector) { free_extent_map(em); - unlock_extent(tree, last_offset, end, GFP_NOFS); + unlock_extent(tree, last_offset, end); unlock_page(page); page_cache_release(page); break; @@ -520,10 +521,10 @@ static noinline int add_ra_bio_pages(struct inode *inode, if (zero_offset) { int zeros; zeros = PAGE_CACHE_SIZE - zero_offset; - userpage = kmap_atomic(page, KM_USER0); + userpage = kmap_atomic(page); memset(userpage + zero_offset, 0, zeros); flush_dcache_page(page); - kunmap_atomic(userpage, KM_USER0); + kunmap_atomic(userpage); } } @@ -534,7 +535,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, nr_pages++; page_cache_release(page); } else { - unlock_extent(tree, last_offset, end, GFP_NOFS); + unlock_extent(tree, last_offset, end); unlock_page(page); page_cache_release(page); break; @@ -587,6 +588,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, page_offset(bio->bi_io_vec->bv_page), PAGE_CACHE_SIZE); read_unlock(&em_tree->lock); + if (!em) + return -EIO; compressed_len = em->block_len; cb = kmalloc(compressed_bio_size(root, compressed_len), GFP_NOFS); @@ -659,7 +662,7 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, bio_get(comp_bio); ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ /* * inc the count before we submit the bio so @@ -672,19 +675,20 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) { ret = btrfs_lookup_bio_sums(root, inode, comp_bio, sums); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } sums += (comp_bio->bi_size + root->sectorsize - 1) / root->sectorsize; ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ bio_put(comp_bio); comp_bio = compressed_bio_alloc(bdev, cur_disk_byte, GFP_NOFS); + BUG_ON(!comp_bio); comp_bio->bi_private = cb; comp_bio->bi_end_io = end_compressed_bio_read; @@ -695,15 +699,15 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, bio_get(comp_bio); ret = btrfs_bio_wq_end_io(root->fs_info, comp_bio, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM)) { ret = btrfs_lookup_bio_sums(root, inode, comp_bio, sums); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ bio_put(comp_bio); return 0; @@ -731,7 +735,7 @@ struct btrfs_compress_op *btrfs_compress_op[] = { &btrfs_lzo_compress, }; -int __init btrfs_init_compress(void) +void __init btrfs_init_compress(void) { int i; @@ -741,7 +745,6 @@ int __init btrfs_init_compress(void) atomic_set(&comp_alloc_workspace[i], 0); init_waitqueue_head(&comp_workspace_wait[i]); } - return 0; } /* @@ -815,6 +818,7 @@ static void free_workspace(int type, struct list_head *workspace) btrfs_compress_op[idx]->free_workspace(workspace); atomic_dec(alloc_workspace); wake: + smp_mb(); if (waitqueue_active(workspace_wait)) wake_up(workspace_wait); } @@ -990,9 +994,9 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, bytes = min(PAGE_CACHE_SIZE - *pg_offset, PAGE_CACHE_SIZE - buf_offset); bytes = min(bytes, working_bytes); - kaddr = kmap_atomic(page_out, KM_USER0); + kaddr = kmap_atomic(page_out); memcpy(kaddr + *pg_offset, buf + buf_offset, bytes); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); flush_dcache_page(page_out); *pg_offset += bytes; diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h index a12059f4f0fd..9afb0a62ae82 100644 --- a/fs/btrfs/compression.h +++ b/fs/btrfs/compression.h @@ -19,7 +19,7 @@ #ifndef __BTRFS_COMPRESSION_ #define __BTRFS_COMPRESSION_ -int btrfs_init_compress(void); +void btrfs_init_compress(void); void btrfs_exit_compress(void); int btrfs_compress_pages(int type, struct address_space *mapping, diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 011cab3aca8d..6d183f60d63a 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -18,6 +18,7 @@ #include <linux/sched.h> #include <linux/slab.h> +#include <linux/rbtree.h> #include "ctree.h" #include "disk-io.h" #include "transaction.h" @@ -36,8 +37,17 @@ static int balance_node_right(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *dst_buf, struct extent_buffer *src_buf); -static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct btrfs_path *path, int level, int slot); +static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, + struct btrfs_path *path, int level, int slot, + int tree_mod_log); +static void tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, + struct extent_buffer *eb); +struct extent_buffer *read_old_tree_block(struct btrfs_root *root, u64 bytenr, + u32 blocksize, u64 parent_transid, + u64 time_seq); +struct extent_buffer *btrfs_find_old_tree_block(struct btrfs_root *root, + u64 bytenr, u32 blocksize, + u64 time_seq); struct btrfs_path *btrfs_alloc_path(void) { @@ -156,10 +166,23 @@ struct extent_buffer *btrfs_root_node(struct btrfs_root *root) { struct extent_buffer *eb; - rcu_read_lock(); - eb = rcu_dereference(root->node); - extent_buffer_get(eb); - rcu_read_unlock(); + while (1) { + rcu_read_lock(); + eb = rcu_dereference(root->node); + + /* + * RCU really hurts here, we could free up the root node because + * it was cow'ed but we may not get the new root node yet so do + * the inc_not_zero dance and if it doesn't work then + * synchronize_rcu and try again. + */ + if (atomic_inc_not_zero(&eb->refs)) { + rcu_read_unlock(); + break; + } + rcu_read_unlock(); + synchronize_rcu(); + } return eb; } @@ -207,10 +230,12 @@ struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root) */ static void add_root_to_dirty_list(struct btrfs_root *root) { + spin_lock(&root->fs_info->trans_lock); if (root->track_dirty && list_empty(&root->dirty_list)) { list_add(&root->dirty_list, &root->fs_info->dirty_cowonly_roots); } + spin_unlock(&root->fs_info->trans_lock); } /* @@ -261,9 +286,9 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans, WARN_ON(btrfs_header_generation(buf) > trans->transid); if (new_root_objectid == BTRFS_TREE_RELOC_OBJECTID) - ret = btrfs_inc_ref(trans, root, cow, 1); + ret = btrfs_inc_ref(trans, root, cow, 1, 1); else - ret = btrfs_inc_ref(trans, root, cow, 0); + ret = btrfs_inc_ref(trans, root, cow, 0, 1); if (ret) return ret; @@ -273,6 +298,513 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans, return 0; } +enum mod_log_op { + MOD_LOG_KEY_REPLACE, + MOD_LOG_KEY_ADD, + MOD_LOG_KEY_REMOVE, + MOD_LOG_KEY_REMOVE_WHILE_FREEING, + MOD_LOG_KEY_REMOVE_WHILE_MOVING, + MOD_LOG_MOVE_KEYS, + MOD_LOG_ROOT_REPLACE, +}; + +struct tree_mod_move { + int dst_slot; + int nr_items; +}; + +struct tree_mod_root { + u64 logical; + u8 level; +}; + +struct tree_mod_elem { + struct rb_node node; + u64 index; /* shifted logical */ + u64 seq; + enum mod_log_op op; + + /* this is used for MOD_LOG_KEY_* and MOD_LOG_MOVE_KEYS operations */ + int slot; + + /* this is used for MOD_LOG_KEY* and MOD_LOG_ROOT_REPLACE */ + u64 generation; + + /* those are used for op == MOD_LOG_KEY_{REPLACE,REMOVE} */ + struct btrfs_disk_key key; + u64 blockptr; + + /* this is used for op == MOD_LOG_MOVE_KEYS */ + struct tree_mod_move move; + + /* this is used for op == MOD_LOG_ROOT_REPLACE */ + struct tree_mod_root old_root; +}; + +static inline void tree_mod_log_read_lock(struct btrfs_fs_info *fs_info) +{ + read_lock(&fs_info->tree_mod_log_lock); +} + +static inline void tree_mod_log_read_unlock(struct btrfs_fs_info *fs_info) +{ + read_unlock(&fs_info->tree_mod_log_lock); +} + +static inline void tree_mod_log_write_lock(struct btrfs_fs_info *fs_info) +{ + write_lock(&fs_info->tree_mod_log_lock); +} + +static inline void tree_mod_log_write_unlock(struct btrfs_fs_info *fs_info) +{ + write_unlock(&fs_info->tree_mod_log_lock); +} + +/* + * This adds a new blocker to the tree mod log's blocker list if the @elem + * passed does not already have a sequence number set. So when a caller expects + * to record tree modifications, it should ensure to set elem->seq to zero + * before calling btrfs_get_tree_mod_seq. + * Returns a fresh, unused tree log modification sequence number, even if no new + * blocker was added. + */ +u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info, + struct seq_list *elem) +{ + u64 seq; + + tree_mod_log_write_lock(fs_info); + spin_lock(&fs_info->tree_mod_seq_lock); + if (!elem->seq) { + elem->seq = btrfs_inc_tree_mod_seq(fs_info); + list_add_tail(&elem->list, &fs_info->tree_mod_seq_list); + } + seq = btrfs_inc_tree_mod_seq(fs_info); + spin_unlock(&fs_info->tree_mod_seq_lock); + tree_mod_log_write_unlock(fs_info); + + return seq; +} + +void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info, + struct seq_list *elem) +{ + struct rb_root *tm_root; + struct rb_node *node; + struct rb_node *next; + struct seq_list *cur_elem; + struct tree_mod_elem *tm; + u64 min_seq = (u64)-1; + u64 seq_putting = elem->seq; + + if (!seq_putting) + return; + + spin_lock(&fs_info->tree_mod_seq_lock); + list_del(&elem->list); + elem->seq = 0; + + list_for_each_entry(cur_elem, &fs_info->tree_mod_seq_list, list) { + if (cur_elem->seq < min_seq) { + if (seq_putting > cur_elem->seq) { + /* + * blocker with lower sequence number exists, we + * cannot remove anything from the log + */ + spin_unlock(&fs_info->tree_mod_seq_lock); + return; + } + min_seq = cur_elem->seq; + } + } + spin_unlock(&fs_info->tree_mod_seq_lock); + + /* + * anything that's lower than the lowest existing (read: blocked) + * sequence number can be removed from the tree. + */ + tree_mod_log_write_lock(fs_info); + tm_root = &fs_info->tree_mod_log; + for (node = rb_first(tm_root); node; node = next) { + next = rb_next(node); + tm = container_of(node, struct tree_mod_elem, node); + if (tm->seq > min_seq) + continue; + rb_erase(node, tm_root); + kfree(tm); + } + tree_mod_log_write_unlock(fs_info); +} + +/* + * key order of the log: + * index -> sequence + * + * the index is the shifted logical of the *new* root node for root replace + * operations, or the shifted logical of the affected block for all other + * operations. + */ +static noinline int +__tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct tree_mod_elem *tm) +{ + struct rb_root *tm_root; + struct rb_node **new; + struct rb_node *parent = NULL; + struct tree_mod_elem *cur; + + BUG_ON(!tm || !tm->seq); + + tm_root = &fs_info->tree_mod_log; + new = &tm_root->rb_node; + while (*new) { + cur = container_of(*new, struct tree_mod_elem, node); + parent = *new; + if (cur->index < tm->index) + new = &((*new)->rb_left); + else if (cur->index > tm->index) + new = &((*new)->rb_right); + else if (cur->seq < tm->seq) + new = &((*new)->rb_left); + else if (cur->seq > tm->seq) + new = &((*new)->rb_right); + else { + kfree(tm); + return -EEXIST; + } + } + + rb_link_node(&tm->node, parent, new); + rb_insert_color(&tm->node, tm_root); + return 0; +} + +/* + * Determines if logging can be omitted. Returns 1 if it can. Otherwise, it + * returns zero with the tree_mod_log_lock acquired. The caller must hold + * this until all tree mod log insertions are recorded in the rb tree and then + * call tree_mod_log_write_unlock() to release. + */ +static inline int tree_mod_dont_log(struct btrfs_fs_info *fs_info, + struct extent_buffer *eb) { + smp_mb(); + if (list_empty(&(fs_info)->tree_mod_seq_list)) + return 1; + if (eb && btrfs_header_level(eb) == 0) + return 1; + + tree_mod_log_write_lock(fs_info); + if (list_empty(&fs_info->tree_mod_seq_list)) { + /* + * someone emptied the list while we were waiting for the lock. + * we must not add to the list when no blocker exists. + */ + tree_mod_log_write_unlock(fs_info); + return 1; + } + + return 0; +} + +/* + * This allocates memory and gets a tree modification sequence number. + * + * Returns <0 on error. + * Returns >0 (the added sequence number) on success. + */ +static inline int tree_mod_alloc(struct btrfs_fs_info *fs_info, gfp_t flags, + struct tree_mod_elem **tm_ret) +{ + struct tree_mod_elem *tm; + + /* + * once we switch from spin locks to something different, we should + * honor the flags parameter here. + */ + tm = *tm_ret = kzalloc(sizeof(*tm), GFP_ATOMIC); + if (!tm) + return -ENOMEM; + + tm->seq = btrfs_inc_tree_mod_seq(fs_info); + return tm->seq; +} + +static inline int +__tree_mod_log_insert_key(struct btrfs_fs_info *fs_info, + struct extent_buffer *eb, int slot, + enum mod_log_op op, gfp_t flags) +{ + int ret; + struct tree_mod_elem *tm; + + ret = tree_mod_alloc(fs_info, flags, &tm); + if (ret < 0) + return ret; + + tm->index = eb->start >> PAGE_CACHE_SHIFT; + if (op != MOD_LOG_KEY_ADD) { + btrfs_node_key(eb, &tm->key, slot); + tm->blockptr = btrfs_node_blockptr(eb, slot); + } + tm->op = op; + tm->slot = slot; + tm->generation = btrfs_node_ptr_generation(eb, slot); + + return __tree_mod_log_insert(fs_info, tm); +} + +static noinline int +tree_mod_log_insert_key_mask(struct btrfs_fs_info *fs_info, + struct extent_buffer *eb, int slot, + enum mod_log_op op, gfp_t flags) +{ + int ret; + + if (tree_mod_dont_log(fs_info, eb)) + return 0; + + ret = __tree_mod_log_insert_key(fs_info, eb, slot, op, flags); + + tree_mod_log_write_unlock(fs_info); + return ret; +} + +static noinline int +tree_mod_log_insert_key(struct btrfs_fs_info *fs_info, struct extent_buffer *eb, + int slot, enum mod_log_op op) +{ + return tree_mod_log_insert_key_mask(fs_info, eb, slot, op, GFP_NOFS); +} + +static noinline int +tree_mod_log_insert_key_locked(struct btrfs_fs_info *fs_info, + struct extent_buffer *eb, int slot, + enum mod_log_op op) +{ + return __tree_mod_log_insert_key(fs_info, eb, slot, op, GFP_NOFS); +} + +static noinline int +tree_mod_log_insert_move(struct btrfs_fs_info *fs_info, + struct extent_buffer *eb, int dst_slot, int src_slot, + int nr_items, gfp_t flags) +{ + struct tree_mod_elem *tm; + int ret; + int i; + + if (tree_mod_dont_log(fs_info, eb)) + return 0; + + for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) { + ret = tree_mod_log_insert_key_locked(fs_info, eb, i + dst_slot, + MOD_LOG_KEY_REMOVE_WHILE_MOVING); + BUG_ON(ret < 0); + } + + ret = tree_mod_alloc(fs_info, flags, &tm); + if (ret < 0) + goto out; + + tm->index = eb->start >> PAGE_CACHE_SHIFT; + tm->slot = src_slot; + tm->move.dst_slot = dst_slot; + tm->move.nr_items = nr_items; + tm->op = MOD_LOG_MOVE_KEYS; + + ret = __tree_mod_log_insert(fs_info, tm); +out: + tree_mod_log_write_unlock(fs_info); + return ret; +} + +static inline void +__tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb) +{ + int i; + u32 nritems; + int ret; + + if (btrfs_header_level(eb) == 0) + return; + + nritems = btrfs_header_nritems(eb); + for (i = nritems - 1; i >= 0; i--) { + ret = tree_mod_log_insert_key_locked(fs_info, eb, i, + MOD_LOG_KEY_REMOVE_WHILE_FREEING); + BUG_ON(ret < 0); + } +} + +static noinline int +tree_mod_log_insert_root(struct btrfs_fs_info *fs_info, + struct extent_buffer *old_root, + struct extent_buffer *new_root, gfp_t flags) +{ + struct tree_mod_elem *tm; + int ret; + + if (tree_mod_dont_log(fs_info, NULL)) + return 0; + + __tree_mod_log_free_eb(fs_info, old_root); + + ret = tree_mod_alloc(fs_info, flags, &tm); + if (ret < 0) + goto out; + + tm->index = new_root->start >> PAGE_CACHE_SHIFT; + tm->old_root.logical = old_root->start; + tm->old_root.level = btrfs_header_level(old_root); + tm->generation = btrfs_header_generation(old_root); + tm->op = MOD_LOG_ROOT_REPLACE; + + ret = __tree_mod_log_insert(fs_info, tm); +out: + tree_mod_log_write_unlock(fs_info); + return ret; +} + +static struct tree_mod_elem * +__tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq, + int smallest) +{ + struct rb_root *tm_root; + struct rb_node *node; + struct tree_mod_elem *cur = NULL; + struct tree_mod_elem *found = NULL; + u64 index = start >> PAGE_CACHE_SHIFT; + + tree_mod_log_read_lock(fs_info); + tm_root = &fs_info->tree_mod_log; + node = tm_root->rb_node; + while (node) { + cur = container_of(node, struct tree_mod_elem, node); + if (cur->index < index) { + node = node->rb_left; + } else if (cur->index > index) { + node = node->rb_right; + } else if (cur->seq < min_seq) { + node = node->rb_left; + } else if (!smallest) { + /* we want the node with the highest seq */ + if (found) + BUG_ON(found->seq > cur->seq); + found = cur; + node = node->rb_left; + } else if (cur->seq > min_seq) { + /* we want the node with the smallest seq */ + if (found) + BUG_ON(found->seq < cur->seq); + found = cur; + node = node->rb_right; + } else { + found = cur; + break; + } + } + tree_mod_log_read_unlock(fs_info); + + return found; +} + +/* + * this returns the element from the log with the smallest time sequence + * value that's in the log (the oldest log item). any element with a time + * sequence lower than min_seq will be ignored. + */ +static struct tree_mod_elem * +tree_mod_log_search_oldest(struct btrfs_fs_info *fs_info, u64 start, + u64 min_seq) +{ + return __tree_mod_log_search(fs_info, start, min_seq, 1); +} + +/* + * this returns the element from the log with the largest time sequence + * value that's in the log (the most recent log item). any element with + * a time sequence lower than min_seq will be ignored. + */ +static struct tree_mod_elem * +tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq) +{ + return __tree_mod_log_search(fs_info, start, min_seq, 0); +} + +static noinline void +tree_mod_log_eb_copy(struct btrfs_fs_info *fs_info, struct extent_buffer *dst, + struct extent_buffer *src, unsigned long dst_offset, + unsigned long src_offset, int nr_items) +{ + int ret; + int i; + + if (tree_mod_dont_log(fs_info, NULL)) + return; + + if (btrfs_header_level(dst) == 0 && btrfs_header_level(src) == 0) { + tree_mod_log_write_unlock(fs_info); + return; + } + + for (i = 0; i < nr_items; i++) { + ret = tree_mod_log_insert_key_locked(fs_info, src, + i + src_offset, + MOD_LOG_KEY_REMOVE); + BUG_ON(ret < 0); + ret = tree_mod_log_insert_key_locked(fs_info, dst, + i + dst_offset, + MOD_LOG_KEY_ADD); + BUG_ON(ret < 0); + } + + tree_mod_log_write_unlock(fs_info); +} + +static inline void +tree_mod_log_eb_move(struct btrfs_fs_info *fs_info, struct extent_buffer *dst, + int dst_offset, int src_offset, int nr_items) +{ + int ret; + ret = tree_mod_log_insert_move(fs_info, dst, dst_offset, src_offset, + nr_items, GFP_NOFS); + BUG_ON(ret < 0); +} + +static noinline void +tree_mod_log_set_node_key(struct btrfs_fs_info *fs_info, + struct extent_buffer *eb, + struct btrfs_disk_key *disk_key, int slot, int atomic) +{ + int ret; + + ret = tree_mod_log_insert_key_mask(fs_info, eb, slot, + MOD_LOG_KEY_REPLACE, + atomic ? GFP_ATOMIC : GFP_NOFS); + BUG_ON(ret < 0); +} + +static noinline void +tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb) +{ + if (tree_mod_dont_log(fs_info, eb)) + return; + + __tree_mod_log_free_eb(fs_info, eb); + + tree_mod_log_write_unlock(fs_info); +} + +static noinline void +tree_mod_log_set_root_pointer(struct btrfs_root *root, + struct extent_buffer *new_root_node) +{ + int ret; + ret = tree_mod_log_insert_root(root->fs_info, root->node, + new_root_node, GFP_NOFS); + BUG_ON(ret < 0); +} + /* * check if the tree block can be shared by multiple trees */ @@ -331,8 +863,13 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans, if (btrfs_block_can_be_shared(root, buf)) { ret = btrfs_lookup_extent_info(trans, root, buf->start, buf->len, &refs, &flags); - BUG_ON(ret); - BUG_ON(refs == 0); + if (ret) + return ret; + if (refs == 0) { + ret = -EROFS; + btrfs_std_error(root->fs_info, ret); + return ret; + } } else { refs = 1; if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID || @@ -350,44 +887,51 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans, if ((owner == root->root_key.objectid || root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) && !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) { - ret = btrfs_inc_ref(trans, root, buf, 1); - BUG_ON(ret); + ret = btrfs_inc_ref(trans, root, buf, 1, 1); + BUG_ON(ret); /* -ENOMEM */ if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) { - ret = btrfs_dec_ref(trans, root, buf, 0); - BUG_ON(ret); - ret = btrfs_inc_ref(trans, root, cow, 1); - BUG_ON(ret); + ret = btrfs_dec_ref(trans, root, buf, 0, 1); + BUG_ON(ret); /* -ENOMEM */ + ret = btrfs_inc_ref(trans, root, cow, 1, 1); + BUG_ON(ret); /* -ENOMEM */ } new_flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF; } else { if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) - ret = btrfs_inc_ref(trans, root, cow, 1); + ret = btrfs_inc_ref(trans, root, cow, 1, 1); else - ret = btrfs_inc_ref(trans, root, cow, 0); - BUG_ON(ret); + ret = btrfs_inc_ref(trans, root, cow, 0, 1); + BUG_ON(ret); /* -ENOMEM */ } if (new_flags != 0) { ret = btrfs_set_disk_extent_flags(trans, root, buf->start, buf->len, new_flags, 0); - BUG_ON(ret); + if (ret) + return ret; } } else { if (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF) { if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) - ret = btrfs_inc_ref(trans, root, cow, 1); + ret = btrfs_inc_ref(trans, root, cow, 1, 1); else - ret = btrfs_inc_ref(trans, root, cow, 0); - BUG_ON(ret); - ret = btrfs_dec_ref(trans, root, buf, 1); - BUG_ON(ret); + ret = btrfs_inc_ref(trans, root, cow, 0, 1); + BUG_ON(ret); /* -ENOMEM */ + ret = btrfs_dec_ref(trans, root, buf, 1, 1); + BUG_ON(ret); /* -ENOMEM */ } + /* + * don't log freeing in case we're freeing the root node, this + * is done by tree_mod_log_set_root_pointer later + */ + if (buf != root->node && btrfs_header_level(buf) != 0) + tree_mod_log_free_eb(root->fs_info, buf); clean_tree_block(trans, root, buf); *last_ref = 1; } @@ -415,7 +959,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, { struct btrfs_disk_key disk_key; struct extent_buffer *cow; - int level; + int level, ret; int last_ref = 0; int unlock_orig = 0; u64 parent_start; @@ -467,7 +1011,11 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, (unsigned long)btrfs_header_fsid(cow), BTRFS_FSID_SIZE); - update_ref_for_cow(trans, root, buf, cow, &last_ref); + ret = update_ref_for_cow(trans, root, buf, cow, &last_ref); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + return ret; + } if (root->ref_cows) btrfs_reloc_cow_block(trans, root, buf, cow); @@ -481,6 +1029,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, parent_start = 0; extent_buffer_get(cow); + tree_mod_log_set_root_pointer(root, cow); rcu_assign_pointer(root->node, cow); btrfs_free_tree_block(trans, root, buf, parent_start, @@ -494,6 +1043,8 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, parent_start = 0; WARN_ON(trans->transid != btrfs_header_generation(parent)); + tree_mod_log_insert_key(root->fs_info, parent, parent_slot, + MOD_LOG_KEY_REPLACE); btrfs_set_node_blockptr(parent, parent_slot, cow->start); btrfs_set_node_ptr_generation(parent, parent_slot, @@ -504,20 +1055,258 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, } if (unlock_orig) btrfs_tree_unlock(buf); - free_extent_buffer(buf); + free_extent_buffer_stale(buf); btrfs_mark_buffer_dirty(cow); *cow_ret = cow; return 0; } +/* + * returns the logical address of the oldest predecessor of the given root. + * entries older than time_seq are ignored. + */ +static struct tree_mod_elem * +__tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info, + struct btrfs_root *root, u64 time_seq) +{ + struct tree_mod_elem *tm; + struct tree_mod_elem *found = NULL; + u64 root_logical = root->node->start; + int looped = 0; + + if (!time_seq) + return 0; + + /* + * the very last operation that's logged for a root is the replacement + * operation (if it is replaced at all). this has the index of the *new* + * root, making it the very first operation that's logged for this root. + */ + while (1) { + tm = tree_mod_log_search_oldest(fs_info, root_logical, + time_seq); + if (!looped && !tm) + return 0; + /* + * if there are no tree operation for the oldest root, we simply + * return it. this should only happen if that (old) root is at + * level 0. + */ + if (!tm) + break; + + /* + * if there's an operation that's not a root replacement, we + * found the oldest version of our root. normally, we'll find a + * MOD_LOG_KEY_REMOVE_WHILE_FREEING operation here. + */ + if (tm->op != MOD_LOG_ROOT_REPLACE) + break; + + found = tm; + root_logical = tm->old_root.logical; + BUG_ON(root_logical == root->node->start); + looped = 1; + } + + /* if there's no old root to return, return what we found instead */ + if (!found) + found = tm; + + return found; +} + +/* + * tm is a pointer to the first operation to rewind within eb. then, all + * previous operations will be rewinded (until we reach something older than + * time_seq). + */ +static void +__tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq, + struct tree_mod_elem *first_tm) +{ + u32 n; + struct rb_node *next; + struct tree_mod_elem *tm = first_tm; + unsigned long o_dst; + unsigned long o_src; + unsigned long p_size = sizeof(struct btrfs_key_ptr); + + n = btrfs_header_nritems(eb); + while (tm && tm->seq >= time_seq) { + /* + * all the operations are recorded with the operator used for + * the modification. as we're going backwards, we do the + * opposite of each operation here. + */ + switch (tm->op) { + case MOD_LOG_KEY_REMOVE_WHILE_FREEING: + BUG_ON(tm->slot < n); + case MOD_LOG_KEY_REMOVE_WHILE_MOVING: + case MOD_LOG_KEY_REMOVE: + btrfs_set_node_key(eb, &tm->key, tm->slot); + btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr); + btrfs_set_node_ptr_generation(eb, tm->slot, + tm->generation); + n++; + break; + case MOD_LOG_KEY_REPLACE: + BUG_ON(tm->slot >= n); + btrfs_set_node_key(eb, &tm->key, tm->slot); + btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr); + btrfs_set_node_ptr_generation(eb, tm->slot, + tm->generation); + break; + case MOD_LOG_KEY_ADD: + /* if a move operation is needed it's in the log */ + n--; + break; + case MOD_LOG_MOVE_KEYS: + o_dst = btrfs_node_key_ptr_offset(tm->slot); + o_src = btrfs_node_key_ptr_offset(tm->move.dst_slot); + memmove_extent_buffer(eb, o_dst, o_src, + tm->move.nr_items * p_size); + break; + case MOD_LOG_ROOT_REPLACE: + /* + * this operation is special. for roots, this must be + * handled explicitly before rewinding. + * for non-roots, this operation may exist if the node + * was a root: root A -> child B; then A gets empty and + * B is promoted to the new root. in the mod log, we'll + * have a root-replace operation for B, a tree block + * that is no root. we simply ignore that operation. + */ + break; + } + next = rb_next(&tm->node); + if (!next) + break; + tm = container_of(next, struct tree_mod_elem, node); + if (tm->index != first_tm->index) + break; + } + btrfs_set_header_nritems(eb, n); +} + +static struct extent_buffer * +tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb, + u64 time_seq) +{ + struct extent_buffer *eb_rewin; + struct tree_mod_elem *tm; + + if (!time_seq) + return eb; + + if (btrfs_header_level(eb) == 0) + return eb; + + tm = tree_mod_log_search(fs_info, eb->start, time_seq); + if (!tm) + return eb; + + if (tm->op == MOD_LOG_KEY_REMOVE_WHILE_FREEING) { + BUG_ON(tm->slot != 0); + eb_rewin = alloc_dummy_extent_buffer(eb->start, + fs_info->tree_root->nodesize); + BUG_ON(!eb_rewin); + btrfs_set_header_bytenr(eb_rewin, eb->start); + btrfs_set_header_backref_rev(eb_rewin, + btrfs_header_backref_rev(eb)); + btrfs_set_header_owner(eb_rewin, btrfs_header_owner(eb)); + btrfs_set_header_level(eb_rewin, btrfs_header_level(eb)); + } else { + eb_rewin = btrfs_clone_extent_buffer(eb); + BUG_ON(!eb_rewin); + } + + extent_buffer_get(eb_rewin); + free_extent_buffer(eb); + + __tree_mod_log_rewind(eb_rewin, time_seq, tm); + + return eb_rewin; +} + +/* + * get_old_root() rewinds the state of @root's root node to the given @time_seq + * value. If there are no changes, the current root->root_node is returned. If + * anything changed in between, there's a fresh buffer allocated on which the + * rewind operations are done. In any case, the returned buffer is read locked. + * Returns NULL on error (with no locks held). + */ +static inline struct extent_buffer * +get_old_root(struct btrfs_root *root, u64 time_seq) +{ + struct tree_mod_elem *tm; + struct extent_buffer *eb; + struct tree_mod_root *old_root = NULL; + u64 old_generation = 0; + u64 logical; + + eb = btrfs_read_lock_root_node(root); + tm = __tree_mod_log_oldest_root(root->fs_info, root, time_seq); + if (!tm) + return root->node; + + if (tm->op == MOD_LOG_ROOT_REPLACE) { + old_root = &tm->old_root; + old_generation = tm->generation; + logical = old_root->logical; + } else { + logical = root->node->start; + } + + tm = tree_mod_log_search(root->fs_info, logical, time_seq); + if (old_root) + eb = alloc_dummy_extent_buffer(logical, root->nodesize); + else + eb = btrfs_clone_extent_buffer(root->node); + btrfs_tree_read_unlock(root->node); + free_extent_buffer(root->node); + if (!eb) + return NULL; + btrfs_tree_read_lock(eb); + if (old_root) { + btrfs_set_header_bytenr(eb, eb->start); + btrfs_set_header_backref_rev(eb, BTRFS_MIXED_BACKREF_REV); + btrfs_set_header_owner(eb, root->root_key.objectid); + btrfs_set_header_level(eb, old_root->level); + btrfs_set_header_generation(eb, old_generation); + } + if (tm) + __tree_mod_log_rewind(eb, time_seq, tm); + else + WARN_ON(btrfs_header_level(eb) != 0); + extent_buffer_get(eb); + + return eb; +} + static inline int should_cow_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf) { + /* ensure we can see the force_cow */ + smp_rmb(); + + /* + * We do not need to cow a block if + * 1) this block is not created or changed in this transaction; + * 2) this block does not belong to TREE_RELOC tree; + * 3) the root is not forced COW. + * + * What is forced COW: + * when we create snapshot during commiting the transaction, + * after we've finished coping src root, we must COW the shared + * block to ensure the metadata consistency. + */ if (btrfs_header_generation(buf) == trans->transid && !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN) && !(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID && - btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC))) + btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)) && + !root->force_cow) return 0; return 1; } @@ -685,7 +1474,7 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, cur = btrfs_find_tree_block(root, blocknr, blocksize); if (cur) - uptodate = btrfs_buffer_uptodate(cur, gen); + uptodate = btrfs_buffer_uptodate(cur, gen, 0); else uptodate = 0; if (!cur || !uptodate) { @@ -699,7 +1488,11 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, if (!cur) return -EIO; } else if (!uptodate) { - btrfs_read_buffer(cur, gen); + err = btrfs_read_buffer(cur, gen); + if (err) { + free_extent_buffer(cur); + return err; + } } } if (search_start == 0) @@ -814,20 +1607,18 @@ static noinline int generic_bin_search(struct extent_buffer *eb, static int bin_search(struct extent_buffer *eb, struct btrfs_key *key, int level, int *slot) { - if (level == 0) { + if (level == 0) return generic_bin_search(eb, offsetof(struct btrfs_leaf, items), sizeof(struct btrfs_item), key, btrfs_header_nritems(eb), slot); - } else { + else return generic_bin_search(eb, offsetof(struct btrfs_node, ptrs), sizeof(struct btrfs_key_ptr), key, btrfs_header_nritems(eb), slot); - } - return -1; } int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key, @@ -902,9 +1693,10 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, orig_ptr = btrfs_node_blockptr(mid, orig_slot); - if (level < BTRFS_MAX_LEVEL - 1) + if (level < BTRFS_MAX_LEVEL - 1) { parent = path->nodes[level + 1]; - pslot = path->slots[level + 1]; + pslot = path->slots[level + 1]; + } /* * deal with the case where there is only one pointer in the root @@ -918,7 +1710,12 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, /* promote the child to a root */ child = read_node_slot(root, mid, 0); - BUG_ON(!child); + if (!child) { + ret = -EROFS; + btrfs_std_error(root->fs_info, ret); + goto enospc; + } + btrfs_tree_lock(child); btrfs_set_lock_blocking(child); ret = btrfs_cow_block(trans, root, child, mid, 0, &child); @@ -928,6 +1725,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, goto enospc; } + tree_mod_log_set_root_pointer(root, child); rcu_assign_pointer(root->node, child); add_root_to_dirty_list(root); @@ -943,15 +1741,13 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, root_sub_used(root, mid->len); btrfs_free_tree_block(trans, root, mid, 0, 1); /* once for the root ptr */ - free_extent_buffer(mid); + free_extent_buffer_stale(mid); return 0; } if (btrfs_header_nritems(mid) > BTRFS_NODEPTRS_PER_BLOCK(root) / 4) return 0; - btrfs_header_nritems(mid); - left = read_node_slot(root, parent, pslot - 1); if (left) { btrfs_tree_lock(left); @@ -981,7 +1777,6 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, wret = push_node_left(trans, root, left, mid, 1); if (wret < 0) ret = wret; - btrfs_header_nritems(mid); } /* @@ -994,17 +1789,16 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, if (btrfs_header_nritems(right) == 0) { clean_tree_block(trans, root, right); btrfs_tree_unlock(right); - wret = del_ptr(trans, root, path, level + 1, pslot + - 1); - if (wret) - ret = wret; + del_ptr(trans, root, path, level + 1, pslot + 1, 1); root_sub_used(root, right->len); btrfs_free_tree_block(trans, root, right, 0, 1); - free_extent_buffer(right); + free_extent_buffer_stale(right); right = NULL; } else { struct btrfs_disk_key right_key; btrfs_node_key(right, &right_key, 0); + tree_mod_log_set_node_key(root->fs_info, parent, + &right_key, pslot + 1, 0); btrfs_set_node_key(parent, &right_key, pslot + 1); btrfs_mark_buffer_dirty(parent); } @@ -1019,7 +1813,11 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, * otherwise we would have pulled some pointers from the * right */ - BUG_ON(!left); + if (!left) { + ret = -EROFS; + btrfs_std_error(root->fs_info, ret); + goto enospc; + } wret = balance_node_right(trans, root, mid, left); if (wret < 0) { ret = wret; @@ -1035,17 +1833,17 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, if (btrfs_header_nritems(mid) == 0) { clean_tree_block(trans, root, mid); btrfs_tree_unlock(mid); - wret = del_ptr(trans, root, path, level + 1, pslot); - if (wret) - ret = wret; + del_ptr(trans, root, path, level + 1, pslot, 1); root_sub_used(root, mid->len); btrfs_free_tree_block(trans, root, mid, 0, 1); - free_extent_buffer(mid); + free_extent_buffer_stale(mid); mid = NULL; } else { /* update the parent key to reflect our changes */ struct btrfs_disk_key mid_key; btrfs_node_key(mid, &mid_key, 0); + tree_mod_log_set_node_key(root->fs_info, parent, &mid_key, + pslot, 0); btrfs_set_node_key(parent, &mid_key, pslot); btrfs_mark_buffer_dirty(parent); } @@ -1107,9 +1905,10 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans, mid = path->nodes[level]; WARN_ON(btrfs_header_generation(mid) != trans->transid); - if (level < BTRFS_MAX_LEVEL - 1) + if (level < BTRFS_MAX_LEVEL - 1) { parent = path->nodes[level + 1]; - pslot = path->slots[level + 1]; + pslot = path->slots[level + 1]; + } if (!parent) return 1; @@ -1142,6 +1941,8 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans, struct btrfs_disk_key disk_key; orig_slot += left_nr; btrfs_node_key(mid, &disk_key, 0); + tree_mod_log_set_node_key(root->fs_info, parent, + &disk_key, pslot, 0); btrfs_set_node_key(parent, &disk_key, pslot); btrfs_mark_buffer_dirty(parent); if (btrfs_header_nritems(left) > orig_slot) { @@ -1193,6 +1994,8 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans, struct btrfs_disk_key disk_key; btrfs_node_key(right, &disk_key, 0); + tree_mod_log_set_node_key(root->fs_info, parent, + &disk_key, pslot + 1, 0); btrfs_set_node_key(parent, &disk_key, pslot + 1); btrfs_mark_buffer_dirty(parent); @@ -1314,7 +2117,12 @@ static noinline int reada_for_balance(struct btrfs_root *root, block1 = btrfs_node_blockptr(parent, slot - 1); gen = btrfs_node_ptr_generation(parent, slot - 1); eb = btrfs_find_tree_block(root, block1, blocksize); - if (eb && btrfs_buffer_uptodate(eb, gen)) + /* + * if we get -eagain from btrfs_buffer_uptodate, we + * don't want to return eagain here. That will loop + * forever + */ + if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0) block1 = 0; free_extent_buffer(eb); } @@ -1322,7 +2130,7 @@ static noinline int reada_for_balance(struct btrfs_root *root, block2 = btrfs_node_blockptr(parent, slot + 1); gen = btrfs_node_ptr_generation(parent, slot + 1); eb = btrfs_find_tree_block(root, block2, blocksize); - if (eb && btrfs_buffer_uptodate(eb, gen)) + if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0) block2 = 0; free_extent_buffer(eb); } @@ -1365,7 +2173,8 @@ static noinline int reada_for_balance(struct btrfs_root *root, * if lowest_unlock is 1, level 0 won't be unlocked */ static noinline void unlock_up(struct btrfs_path *path, int level, - int lowest_unlock) + int lowest_unlock, int min_write_lock_level, + int *write_lock_level) { int i; int skip_level = level; @@ -1397,6 +2206,11 @@ static noinline void unlock_up(struct btrfs_path *path, int level, if (i >= lowest_unlock && i > skip_level && path->locks[i]) { btrfs_tree_unlock_rw(t, path->locks[i]); path->locks[i] = 0; + if (write_lock_level && + i > min_write_lock_level && + i <= *write_lock_level) { + *write_lock_level = i - 1; + } } } } @@ -1439,7 +2253,7 @@ static int read_block_for_search(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *p, struct extent_buffer **eb_ret, int level, int slot, - struct btrfs_key *key) + struct btrfs_key *key, u64 time_seq) { u64 blocknr; u64 gen; @@ -1454,8 +2268,9 @@ read_block_for_search(struct btrfs_trans_handle *trans, tmp = btrfs_find_tree_block(root, blocknr, blocksize); if (tmp) { - if (btrfs_buffer_uptodate(tmp, 0)) { - if (btrfs_buffer_uptodate(tmp, gen)) { + /* first we do an atomic uptodate check */ + if (btrfs_buffer_uptodate(tmp, 0, 1) > 0) { + if (btrfs_buffer_uptodate(tmp, gen, 1) > 0) { /* * we found an up to date block without * sleeping, return @@ -1473,8 +2288,9 @@ read_block_for_search(struct btrfs_trans_handle *trans, free_extent_buffer(tmp); btrfs_set_path_blocking(p); + /* now we're allowed to do a blocking uptodate check */ tmp = read_tree_block(root, blocknr, blocksize, gen); - if (tmp && btrfs_buffer_uptodate(tmp, gen)) { + if (tmp && btrfs_buffer_uptodate(tmp, gen, 0) > 0) { *eb_ret = tmp; return 0; } @@ -1509,7 +2325,7 @@ read_block_for_search(struct btrfs_trans_handle *trans, * and give up so that our caller doesn't loop forever * on our EAGAINs. */ - if (!btrfs_buffer_uptodate(tmp, 0)) + if (!btrfs_buffer_uptodate(tmp, 0, 0)) ret = -EIO; free_extent_buffer(tmp); } @@ -1620,6 +2436,7 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root /* everything at write_lock_level or lower must be write locked */ int write_lock_level = 0; u8 lowest_level = 0; + int min_write_lock_level; lowest_level = p->lowest_level; WARN_ON(lowest_level && ins_len > 0); @@ -1647,6 +2464,8 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root if (cow && (p->keep_locks || p->lowest_level)) write_lock_level = BTRFS_MAX_LEVEL; + min_write_lock_level = write_lock_level; + again: /* * we try very hard to do read locks on the root @@ -1778,7 +2597,8 @@ cow_done: goto again; } - unlock_up(p, level, lowest_unlock); + unlock_up(p, level, lowest_unlock, + min_write_lock_level, &write_lock_level); if (level == lowest_level) { if (dec) @@ -1787,7 +2607,7 @@ cow_done: } err = read_block_for_search(trans, root, p, - &b, level, slot, key); + &b, level, slot, key, 0); if (err == -EAGAIN) goto again; if (err) { @@ -1840,7 +2660,8 @@ cow_done: } } if (!p->search_for_split) - unlock_up(p, level, lowest_unlock); + unlock_up(p, level, lowest_unlock, + min_write_lock_level, &write_lock_level); goto done; } } @@ -1858,21 +2679,199 @@ done: } /* + * Like btrfs_search_slot, this looks for a key in the given tree. It uses the + * current state of the tree together with the operations recorded in the tree + * modification log to search for the key in a previous version of this tree, as + * denoted by the time_seq parameter. + * + * Naturally, there is no support for insert, delete or cow operations. + * + * The resulting path and return value will be set up as if we called + * btrfs_search_slot at that point in time with ins_len and cow both set to 0. + */ +int btrfs_search_old_slot(struct btrfs_root *root, struct btrfs_key *key, + struct btrfs_path *p, u64 time_seq) +{ + struct extent_buffer *b; + int slot; + int ret; + int err; + int level; + int lowest_unlock = 1; + u8 lowest_level = 0; + + lowest_level = p->lowest_level; + WARN_ON(p->nodes[0] != NULL); + + if (p->search_commit_root) { + BUG_ON(time_seq); + return btrfs_search_slot(NULL, root, key, p, 0, 0); + } + +again: + b = get_old_root(root, time_seq); + level = btrfs_header_level(b); + p->locks[level] = BTRFS_READ_LOCK; + + while (b) { + level = btrfs_header_level(b); + p->nodes[level] = b; + btrfs_clear_path_blocking(p, NULL, 0); + + /* + * we have a lock on b and as long as we aren't changing + * the tree, there is no way to for the items in b to change. + * It is safe to drop the lock on our parent before we + * go through the expensive btree search on b. + */ + btrfs_unlock_up_safe(p, level + 1); + + ret = bin_search(b, key, level, &slot); + + if (level != 0) { + int dec = 0; + if (ret && slot > 0) { + dec = 1; + slot -= 1; + } + p->slots[level] = slot; + unlock_up(p, level, lowest_unlock, 0, NULL); + + if (level == lowest_level) { + if (dec) + p->slots[level]++; + goto done; + } + + err = read_block_for_search(NULL, root, p, &b, level, + slot, key, time_seq); + if (err == -EAGAIN) + goto again; + if (err) { + ret = err; + goto done; + } + + level = btrfs_header_level(b); + err = btrfs_try_tree_read_lock(b); + if (!err) { + btrfs_set_path_blocking(p); + btrfs_tree_read_lock(b); + btrfs_clear_path_blocking(p, b, + BTRFS_READ_LOCK); + } + p->locks[level] = BTRFS_READ_LOCK; + p->nodes[level] = b; + b = tree_mod_log_rewind(root->fs_info, b, time_seq); + if (b != p->nodes[level]) { + btrfs_tree_unlock_rw(p->nodes[level], + p->locks[level]); + p->locks[level] = 0; + p->nodes[level] = b; + } + } else { + p->slots[level] = slot; + unlock_up(p, level, lowest_unlock, 0, NULL); + goto done; + } + } + ret = 1; +done: + if (!p->leave_spinning) + btrfs_set_path_blocking(p); + if (ret < 0) + btrfs_release_path(p); + + return ret; +} + +/* + * helper to use instead of search slot if no exact match is needed but + * instead the next or previous item should be returned. + * When find_higher is true, the next higher item is returned, the next lower + * otherwise. + * When return_any and find_higher are both true, and no higher item is found, + * return the next lower instead. + * When return_any is true and find_higher is false, and no lower item is found, + * return the next higher instead. + * It returns 0 if any item is found, 1 if none is found (tree empty), and + * < 0 on error + */ +int btrfs_search_slot_for_read(struct btrfs_root *root, + struct btrfs_key *key, struct btrfs_path *p, + int find_higher, int return_any) +{ + int ret; + struct extent_buffer *leaf; + +again: + ret = btrfs_search_slot(NULL, root, key, p, 0, 0); + if (ret <= 0) + return ret; + /* + * a return value of 1 means the path is at the position where the + * item should be inserted. Normally this is the next bigger item, + * but in case the previous item is the last in a leaf, path points + * to the first free slot in the previous leaf, i.e. at an invalid + * item. + */ + leaf = p->nodes[0]; + + if (find_higher) { + if (p->slots[0] >= btrfs_header_nritems(leaf)) { + ret = btrfs_next_leaf(root, p); + if (ret <= 0) + return ret; + if (!return_any) + return 1; + /* + * no higher item found, return the next + * lower instead + */ + return_any = 0; + find_higher = 0; + btrfs_release_path(p); + goto again; + } + } else { + if (p->slots[0] == 0) { + ret = btrfs_prev_leaf(root, p); + if (ret < 0) + return ret; + if (!ret) { + p->slots[0] = btrfs_header_nritems(leaf) - 1; + return 0; + } + if (!return_any) + return 1; + /* + * no lower item found, return the next + * higher instead + */ + return_any = 0; + find_higher = 1; + btrfs_release_path(p); + goto again; + } else { + --p->slots[0]; + } + } + return 0; +} + +/* * adjust the pointers going up the tree, starting at level * making sure the right key of each node is points to 'key'. * This is used after shifting pointers to the left, so it stops * fixing up pointers when a given leaf/node is not in slot 0 of the * higher levels * - * If this fails to write a tree block, it returns -1, but continues - * fixing up the blocks in ram so the tree is consistent. */ -static int fixup_low_keys(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_disk_key *key, int level) +static void fixup_low_keys(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_disk_key *key, int level) { int i; - int ret = 0; struct extent_buffer *t; for (i = level; i < BTRFS_MAX_LEVEL; i++) { @@ -1880,12 +2879,12 @@ static int fixup_low_keys(struct btrfs_trans_handle *trans, if (!path->nodes[i]) break; t = path->nodes[i]; + tree_mod_log_set_node_key(root->fs_info, t, key, tslot, 1); btrfs_set_node_key(t, key, tslot); btrfs_mark_buffer_dirty(path->nodes[i]); if (tslot != 0) break; } - return ret; } /* @@ -1894,9 +2893,9 @@ static int fixup_low_keys(struct btrfs_trans_handle *trans, * This function isn't completely safe. It's the caller's responsibility * that the new key won't break the order */ -int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *new_key) +void btrfs_set_item_key_safe(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *new_key) { struct btrfs_disk_key disk_key; struct extent_buffer *eb; @@ -1906,13 +2905,11 @@ int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans, slot = path->slots[0]; if (slot > 0) { btrfs_item_key(eb, &disk_key, slot - 1); - if (comp_keys(&disk_key, new_key) >= 0) - return -1; + BUG_ON(comp_keys(&disk_key, new_key) >= 0); } if (slot < btrfs_header_nritems(eb) - 1) { btrfs_item_key(eb, &disk_key, slot + 1); - if (comp_keys(&disk_key, new_key) <= 0) - return -1; + BUG_ON(comp_keys(&disk_key, new_key) <= 0); } btrfs_cpu_key_to_disk(&disk_key, new_key); @@ -1920,7 +2917,6 @@ int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(eb); if (slot == 0) fixup_low_keys(trans, root, path, &disk_key, 1); - return 0; } /* @@ -1966,12 +2962,16 @@ static int push_node_left(struct btrfs_trans_handle *trans, } else push_items = min(src_nritems - 8, push_items); + tree_mod_log_eb_copy(root->fs_info, dst, src, dst_nritems, 0, + push_items); copy_extent_buffer(dst, src, btrfs_node_key_ptr_offset(dst_nritems), btrfs_node_key_ptr_offset(0), push_items * sizeof(struct btrfs_key_ptr)); if (push_items < src_nritems) { + tree_mod_log_eb_move(root->fs_info, src, 0, push_items, + src_nritems - push_items); memmove_extent_buffer(src, btrfs_node_key_ptr_offset(0), btrfs_node_key_ptr_offset(push_items), (src_nritems - push_items) * @@ -2025,11 +3025,14 @@ static int balance_node_right(struct btrfs_trans_handle *trans, if (max_push < push_items) push_items = max_push; + tree_mod_log_eb_move(root->fs_info, dst, push_items, 0, dst_nritems); memmove_extent_buffer(dst, btrfs_node_key_ptr_offset(push_items), btrfs_node_key_ptr_offset(0), (dst_nritems) * sizeof(struct btrfs_key_ptr)); + tree_mod_log_eb_copy(root->fs_info, dst, src, 0, + src_nritems - push_items, push_items); copy_extent_buffer(dst, src, btrfs_node_key_ptr_offset(0), btrfs_node_key_ptr_offset(src_nritems - push_items), @@ -2104,6 +3107,7 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(c); old = root->node; + tree_mod_log_set_root_pointer(root, c); rcu_assign_pointer(root->node, c); /* the super has an extra ref to root->node */ @@ -2123,36 +3127,42 @@ static noinline int insert_new_root(struct btrfs_trans_handle *trans, * * slot and level indicate where you want the key to go, and * blocknr is the block the key points to. - * - * returns zero on success and < 0 on any error */ -static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root - *root, struct btrfs_path *path, struct btrfs_disk_key - *key, u64 bytenr, int slot, int level) +static void insert_ptr(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_disk_key *key, u64 bytenr, + int slot, int level) { struct extent_buffer *lower; int nritems; + int ret; BUG_ON(!path->nodes[level]); btrfs_assert_tree_locked(path->nodes[level]); lower = path->nodes[level]; nritems = btrfs_header_nritems(lower); BUG_ON(slot > nritems); - if (nritems == BTRFS_NODEPTRS_PER_BLOCK(root)) - BUG(); + BUG_ON(nritems == BTRFS_NODEPTRS_PER_BLOCK(root)); if (slot != nritems) { + if (level) + tree_mod_log_eb_move(root->fs_info, lower, slot + 1, + slot, nritems - slot); memmove_extent_buffer(lower, btrfs_node_key_ptr_offset(slot + 1), btrfs_node_key_ptr_offset(slot), (nritems - slot) * sizeof(struct btrfs_key_ptr)); } + if (level) { + ret = tree_mod_log_insert_key(root->fs_info, lower, slot, + MOD_LOG_KEY_ADD); + BUG_ON(ret < 0); + } btrfs_set_node_key(lower, key, slot); btrfs_set_node_blockptr(lower, slot, bytenr); WARN_ON(trans->transid == 0); btrfs_set_node_ptr_generation(lower, slot, trans->transid); btrfs_set_header_nritems(lower, nritems + 1); btrfs_mark_buffer_dirty(lower); - return 0; } /* @@ -2173,7 +3183,6 @@ static noinline int split_node(struct btrfs_trans_handle *trans, struct btrfs_disk_key disk_key; int mid; int ret; - int wret; u32 c_nritems; c = path->nodes[level]; @@ -2218,7 +3227,7 @@ static noinline int split_node(struct btrfs_trans_handle *trans, (unsigned long)btrfs_header_chunk_tree_uuid(split), BTRFS_UUID_SIZE); - + tree_mod_log_eb_copy(root->fs_info, split, c, 0, mid, c_nritems - mid); copy_extent_buffer(split, c, btrfs_node_key_ptr_offset(0), btrfs_node_key_ptr_offset(mid), @@ -2230,11 +3239,8 @@ static noinline int split_node(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(c); btrfs_mark_buffer_dirty(split); - wret = insert_ptr(trans, root, path, &disk_key, split->start, - path->slots[level + 1] + 1, - level + 1); - if (wret) - ret = wret; + insert_ptr(trans, root, path, &disk_key, split->start, + path->slots[level + 1] + 1, level + 1); if (path->slots[level] >= mid) { path->slots[level] -= mid; @@ -2303,6 +3309,7 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans, { struct extent_buffer *left = path->nodes[0]; struct extent_buffer *upper = path->nodes[1]; + struct btrfs_map_token token; struct btrfs_disk_key disk_key; int slot; u32 i; @@ -2314,6 +3321,8 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans, u32 data_end; u32 this_item_size; + btrfs_init_map_token(&token); + if (empty) nr = 0; else @@ -2391,8 +3400,8 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans, push_space = BTRFS_LEAF_DATA_SIZE(root); for (i = 0; i < right_nritems; i++) { item = btrfs_item_nr(right, i); - push_space -= btrfs_item_size(right, item); - btrfs_set_item_offset(right, item, push_space); + push_space -= btrfs_token_item_size(right, item, &token); + btrfs_set_token_item_offset(right, item, push_space, &token); } left_nritems -= push_items; @@ -2520,9 +3529,11 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, u32 old_left_nritems; u32 nr; int ret = 0; - int wret; u32 this_item_size; u32 old_left_item_size; + struct btrfs_map_token token; + + btrfs_init_map_token(&token); if (empty) nr = min(right_nritems, max_slot); @@ -2583,9 +3594,10 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, item = btrfs_item_nr(left, i); - ioff = btrfs_item_offset(left, item); - btrfs_set_item_offset(left, item, - ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size)); + ioff = btrfs_token_item_offset(left, item, &token); + btrfs_set_token_item_offset(left, item, + ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size), + &token); } btrfs_set_header_nritems(left, old_left_nritems + push_items); @@ -2615,8 +3627,9 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, for (i = 0; i < right_nritems; i++) { item = btrfs_item_nr(right, i); - push_space = push_space - btrfs_item_size(right, item); - btrfs_set_item_offset(right, item, push_space); + push_space = push_space - btrfs_token_item_size(right, + item, &token); + btrfs_set_token_item_offset(right, item, push_space, &token); } btrfs_mark_buffer_dirty(left); @@ -2626,9 +3639,7 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, clean_tree_block(trans, root, right); btrfs_item_key(right, &disk_key, 0); - wret = fixup_low_keys(trans, root, path, &disk_key, 1); - if (wret) - ret = wret; + fixup_low_keys(trans, root, path, &disk_key, 1); /* then fixup the leaf pointer in the path */ if (path->slots[0] < push_items) { @@ -2699,7 +3710,8 @@ static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root path->nodes[1], slot - 1, &left); if (ret) { /* we hit -ENOSPC, but it isn't fatal here */ - ret = 1; + if (ret == -ENOSPC) + ret = 1; goto out; } @@ -2721,22 +3733,21 @@ out: /* * split the path's leaf in two, making sure there is at least data_size * available for the resulting leaf level of the path. - * - * returns 0 if all went well and < 0 on failure. */ -static noinline int copy_for_split(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct extent_buffer *l, - struct extent_buffer *right, - int slot, int mid, int nritems) +static noinline void copy_for_split(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct extent_buffer *l, + struct extent_buffer *right, + int slot, int mid, int nritems) { int data_copy_size; int rt_data_off; int i; - int ret = 0; - int wret; struct btrfs_disk_key disk_key; + struct btrfs_map_token token; + + btrfs_init_map_token(&token); nritems = nritems - mid; btrfs_set_header_nritems(right, nritems); @@ -2758,17 +3769,15 @@ static noinline int copy_for_split(struct btrfs_trans_handle *trans, struct btrfs_item *item = btrfs_item_nr(right, i); u32 ioff; - ioff = btrfs_item_offset(right, item); - btrfs_set_item_offset(right, item, ioff + rt_data_off); + ioff = btrfs_token_item_offset(right, item, &token); + btrfs_set_token_item_offset(right, item, + ioff + rt_data_off, &token); } btrfs_set_header_nritems(l, mid); - ret = 0; btrfs_item_key(right, &disk_key, 0); - wret = insert_ptr(trans, root, path, &disk_key, right->start, - path->slots[1] + 1, 1); - if (wret) - ret = wret; + insert_ptr(trans, root, path, &disk_key, right->start, + path->slots[1] + 1, 1); btrfs_mark_buffer_dirty(right); btrfs_mark_buffer_dirty(l); @@ -2786,8 +3795,6 @@ static noinline int copy_for_split(struct btrfs_trans_handle *trans, } BUG_ON(path->slots[0] < 0); - - return ret; } /* @@ -2976,12 +3983,8 @@ again: if (split == 0) { if (mid <= slot) { btrfs_set_header_nritems(right, 0); - wret = insert_ptr(trans, root, path, - &disk_key, right->start, - path->slots[1] + 1, 1); - if (wret) - ret = wret; - + insert_ptr(trans, root, path, &disk_key, right->start, + path->slots[1] + 1, 1); btrfs_tree_unlock(path->nodes[0]); free_extent_buffer(path->nodes[0]); path->nodes[0] = right; @@ -2989,29 +3992,21 @@ again: path->slots[1] += 1; } else { btrfs_set_header_nritems(right, 0); - wret = insert_ptr(trans, root, path, - &disk_key, - right->start, + insert_ptr(trans, root, path, &disk_key, right->start, path->slots[1], 1); - if (wret) - ret = wret; btrfs_tree_unlock(path->nodes[0]); free_extent_buffer(path->nodes[0]); path->nodes[0] = right; path->slots[0] = 0; - if (path->slots[1] == 0) { - wret = fixup_low_keys(trans, root, - path, &disk_key, 1); - if (wret) - ret = wret; - } + if (path->slots[1] == 0) + fixup_low_keys(trans, root, path, + &disk_key, 1); } btrfs_mark_buffer_dirty(right); return ret; } - ret = copy_for_split(trans, root, path, l, right, slot, mid, nritems); - BUG_ON(ret); + copy_for_split(trans, root, path, l, right, slot, mid, nritems); if (split == 2) { BUG_ON(num_doubles != 0); @@ -3019,7 +4014,7 @@ again: goto again; } - return ret; + return 0; push_for_double: push_for_double_split(trans, root, path, data_size); @@ -3221,11 +4216,9 @@ int btrfs_duplicate_item(struct btrfs_trans_handle *trans, return ret; path->slots[0]++; - ret = setup_items_for_insert(trans, root, path, new_key, &item_size, - item_size, item_size + - sizeof(struct btrfs_item), 1); - BUG_ON(ret); - + setup_items_for_insert(trans, root, path, new_key, &item_size, + item_size, item_size + + sizeof(struct btrfs_item), 1); leaf = path->nodes[0]; memcpy_extent_buffer(leaf, btrfs_item_ptr_offset(leaf, path->slots[0]), @@ -3240,10 +4233,10 @@ int btrfs_duplicate_item(struct btrfs_trans_handle *trans, * off the end of the item or if we shift the item to chop bytes off * the front. */ -int btrfs_truncate_item(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - u32 new_size, int from_end) +void btrfs_truncate_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + u32 new_size, int from_end) { int slot; struct extent_buffer *leaf; @@ -3254,13 +4247,16 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, unsigned int old_size; unsigned int size_diff; int i; + struct btrfs_map_token token; + + btrfs_init_map_token(&token); leaf = path->nodes[0]; slot = path->slots[0]; old_size = btrfs_item_size_nr(leaf, slot); if (old_size == new_size) - return 0; + return; nritems = btrfs_header_nritems(leaf); data_end = leaf_data_end(root, leaf); @@ -3280,8 +4276,9 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, u32 ioff; item = btrfs_item_nr(leaf, i); - ioff = btrfs_item_offset(leaf, item); - btrfs_set_item_offset(leaf, item, ioff + size_diff); + ioff = btrfs_token_item_offset(leaf, item, &token); + btrfs_set_token_item_offset(leaf, item, + ioff + size_diff, &token); } /* shift the data */ @@ -3333,15 +4330,14 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, btrfs_print_leaf(root, leaf); BUG(); } - return 0; } /* * make the item pointed to by the path bigger, data_size is the new size. */ -int btrfs_extend_item(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - u32 data_size) +void btrfs_extend_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + u32 data_size) { int slot; struct extent_buffer *leaf; @@ -3351,6 +4347,9 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, unsigned int old_data; unsigned int old_size; int i; + struct btrfs_map_token token; + + btrfs_init_map_token(&token); leaf = path->nodes[0]; @@ -3380,8 +4379,9 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, u32 ioff; item = btrfs_item_nr(leaf, i); - ioff = btrfs_item_offset(leaf, item); - btrfs_set_item_offset(leaf, item, ioff - data_size); + ioff = btrfs_token_item_offset(leaf, item, &token); + btrfs_set_token_item_offset(leaf, item, + ioff - data_size, &token); } /* shift the data */ @@ -3399,7 +4399,6 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, btrfs_print_leaf(root, leaf); BUG(); } - return 0; } /* @@ -3424,6 +4423,9 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans, unsigned int data_end; struct btrfs_disk_key disk_key; struct btrfs_key found_key; + struct btrfs_map_token token; + + btrfs_init_map_token(&token); for (i = 0; i < nr; i++) { if (total_size + data_size[i] + sizeof(struct btrfs_item) > @@ -3489,8 +4491,9 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans, u32 ioff; item = btrfs_item_nr(leaf, i); - ioff = btrfs_item_offset(leaf, item); - btrfs_set_item_offset(leaf, item, ioff - total_data); + ioff = btrfs_token_item_offset(leaf, item, &token); + btrfs_set_token_item_offset(leaf, item, + ioff - total_data, &token); } /* shift the items */ memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr), @@ -3517,9 +4520,10 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans, btrfs_cpu_key_to_disk(&disk_key, cpu_key + i); btrfs_set_item_key(leaf, &disk_key, slot + i); item = btrfs_item_nr(leaf, slot + i); - btrfs_set_item_offset(leaf, item, data_end - data_size[i]); + btrfs_set_token_item_offset(leaf, item, + data_end - data_size[i], &token); data_end -= data_size[i]; - btrfs_set_item_size(leaf, item, data_size[i]); + btrfs_set_token_item_size(leaf, item, data_size[i], &token); } btrfs_set_header_nritems(leaf, nritems + nr); btrfs_mark_buffer_dirty(leaf); @@ -3527,7 +4531,7 @@ int btrfs_insert_some_items(struct btrfs_trans_handle *trans, ret = 0; if (slot == 0) { btrfs_cpu_key_to_disk(&disk_key, cpu_key); - ret = fixup_low_keys(trans, root, path, &disk_key, 1); + fixup_low_keys(trans, root, path, &disk_key, 1); } if (btrfs_leaf_free_space(root, leaf) < 0) { @@ -3545,19 +4549,21 @@ out: * to save stack depth by doing the bulk of the work in a function * that doesn't call btrfs_search_slot */ -int setup_items_for_insert(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - u32 total_data, u32 total_size, int nr) +void setup_items_for_insert(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *cpu_key, u32 *data_size, + u32 total_data, u32 total_size, int nr) { struct btrfs_item *item; int i; u32 nritems; unsigned int data_end; struct btrfs_disk_key disk_key; - int ret; struct extent_buffer *leaf; int slot; + struct btrfs_map_token token; + + btrfs_init_map_token(&token); leaf = path->nodes[0]; slot = path->slots[0]; @@ -3589,8 +4595,9 @@ int setup_items_for_insert(struct btrfs_trans_handle *trans, u32 ioff; item = btrfs_item_nr(leaf, i); - ioff = btrfs_item_offset(leaf, item); - btrfs_set_item_offset(leaf, item, ioff - total_data); + ioff = btrfs_token_item_offset(leaf, item, &token); + btrfs_set_token_item_offset(leaf, item, + ioff - total_data, &token); } /* shift the items */ memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr), @@ -3609,17 +4616,17 @@ int setup_items_for_insert(struct btrfs_trans_handle *trans, btrfs_cpu_key_to_disk(&disk_key, cpu_key + i); btrfs_set_item_key(leaf, &disk_key, slot + i); item = btrfs_item_nr(leaf, slot + i); - btrfs_set_item_offset(leaf, item, data_end - data_size[i]); + btrfs_set_token_item_offset(leaf, item, + data_end - data_size[i], &token); data_end -= data_size[i]; - btrfs_set_item_size(leaf, item, data_size[i]); + btrfs_set_token_item_size(leaf, item, data_size[i], &token); } btrfs_set_header_nritems(leaf, nritems + nr); - ret = 0; if (slot == 0) { btrfs_cpu_key_to_disk(&disk_key, cpu_key); - ret = fixup_low_keys(trans, root, path, &disk_key, 1); + fixup_low_keys(trans, root, path, &disk_key, 1); } btrfs_unlock_up_safe(path, 1); btrfs_mark_buffer_dirty(leaf); @@ -3628,7 +4635,6 @@ int setup_items_for_insert(struct btrfs_trans_handle *trans, btrfs_print_leaf(root, leaf); BUG(); } - return ret; } /* @@ -3655,16 +4661,14 @@ int btrfs_insert_empty_items(struct btrfs_trans_handle *trans, if (ret == 0) return -EEXIST; if (ret < 0) - goto out; + return ret; slot = path->slots[0]; BUG_ON(slot < 0); - ret = setup_items_for_insert(trans, root, path, cpu_key, data_size, + setup_items_for_insert(trans, root, path, cpu_key, data_size, total_data, total_size, nr); - -out: - return ret; + return 0; } /* @@ -3700,22 +4704,30 @@ int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root * the tree should have been previously balanced so the deletion does not * empty a node. */ -static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct btrfs_path *path, int level, int slot) +static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, + struct btrfs_path *path, int level, int slot, + int tree_mod_log) { struct extent_buffer *parent = path->nodes[level]; u32 nritems; - int ret = 0; - int wret; + int ret; nritems = btrfs_header_nritems(parent); if (slot != nritems - 1) { + if (tree_mod_log && level) + tree_mod_log_eb_move(root->fs_info, parent, slot, + slot + 1, nritems - slot - 1); memmove_extent_buffer(parent, btrfs_node_key_ptr_offset(slot), btrfs_node_key_ptr_offset(slot + 1), sizeof(struct btrfs_key_ptr) * (nritems - slot - 1)); + } else if (tree_mod_log && level) { + ret = tree_mod_log_insert_key(root->fs_info, parent, slot, + MOD_LOG_KEY_REMOVE); + BUG_ON(ret < 0); } + nritems--; btrfs_set_header_nritems(parent, nritems); if (nritems == 0 && parent == root->node) { @@ -3726,12 +4738,9 @@ static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_disk_key disk_key; btrfs_node_key(parent, &disk_key, 0); - wret = fixup_low_keys(trans, root, path, &disk_key, level + 1); - if (wret) - ret = wret; + fixup_low_keys(trans, root, path, &disk_key, level + 1); } btrfs_mark_buffer_dirty(parent); - return ret; } /* @@ -3744,17 +4753,13 @@ static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, * The path must have already been setup for deleting the leaf, including * all the proper balancing. path->nodes[1] must be locked. */ -static noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct extent_buffer *leaf) +static noinline void btrfs_del_leaf(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct extent_buffer *leaf) { - int ret; - WARN_ON(btrfs_header_generation(leaf) != trans->transid); - ret = del_ptr(trans, root, path, 1, path->slots[1]); - if (ret) - return ret; + del_ptr(trans, root, path, 1, path->slots[1], 1); /* * btrfs_free_extent is expensive, we want to make sure we @@ -3764,8 +4769,9 @@ static noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans, root_sub_used(root, leaf->len); + extent_buffer_get(leaf); btrfs_free_tree_block(trans, root, leaf, 0, 1); - return 0; + free_extent_buffer_stale(leaf); } /* * delete the item at the leaf level in path. If that empties @@ -3782,6 +4788,9 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, int wret; int i; u32 nritems; + struct btrfs_map_token token; + + btrfs_init_map_token(&token); leaf = path->nodes[0]; last_off = btrfs_item_offset_nr(leaf, slot + nr - 1); @@ -3803,8 +4812,9 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, u32 ioff; item = btrfs_item_nr(leaf, i); - ioff = btrfs_item_offset(leaf, item); - btrfs_set_item_offset(leaf, item, ioff + dsize); + ioff = btrfs_token_item_offset(leaf, item, &token); + btrfs_set_token_item_offset(leaf, item, + ioff + dsize, &token); } memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot), @@ -3822,8 +4832,7 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, } else { btrfs_set_path_blocking(path); clean_tree_block(trans, root, leaf); - ret = btrfs_del_leaf(trans, root, path, leaf); - BUG_ON(ret); + btrfs_del_leaf(trans, root, path, leaf); } } else { int used = leaf_space_used(leaf, 0, nritems); @@ -3831,10 +4840,7 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_disk_key disk_key; btrfs_item_key(leaf, &disk_key, 0); - wret = fixup_low_keys(trans, root, path, - &disk_key, 1); - if (wret) - ret = wret; + fixup_low_keys(trans, root, path, &disk_key, 1); } /* delete the leaf if it is mostly empty */ @@ -3862,9 +4868,9 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, if (btrfs_header_nritems(leaf) == 0) { path->slots[1] = slot; - ret = btrfs_del_leaf(trans, root, path, leaf); - BUG_ON(ret); + btrfs_del_leaf(trans, root, path, leaf); free_extent_buffer(leaf); + ret = 0; } else { /* if we're still in the path, make sure * we're dirty. Otherwise, one of the @@ -4012,7 +5018,7 @@ again: tmp = btrfs_find_tree_block(root, blockptr, btrfs_level_size(root, level - 1)); - if (tmp && btrfs_buffer_uptodate(tmp, gen)) { + if (tmp && btrfs_buffer_uptodate(tmp, gen, 1) > 0) { free_extent_buffer(tmp); break; } @@ -4042,18 +5048,18 @@ find_next_key: path->slots[level] = slot; if (level == path->lowest_level) { ret = 0; - unlock_up(path, level, 1); + unlock_up(path, level, 1, 0, NULL); goto out; } btrfs_set_path_blocking(path); cur = read_node_slot(root, cur, slot); - BUG_ON(!cur); + BUG_ON(!cur); /* -ENOMEM */ btrfs_tree_read_lock(cur); path->locks[level - 1] = BTRFS_READ_LOCK; path->nodes[level - 1] = cur; - unlock_up(path, level, 1); + unlock_up(path, level, 1, 0, NULL); btrfs_clear_path_blocking(path, NULL, 0); } out: @@ -4063,6 +5069,431 @@ out: return ret; } +static void tree_move_down(struct btrfs_root *root, + struct btrfs_path *path, + int *level, int root_level) +{ + path->nodes[*level - 1] = read_node_slot(root, path->nodes[*level], + path->slots[*level]); + path->slots[*level - 1] = 0; + (*level)--; +} + +static int tree_move_next_or_upnext(struct btrfs_root *root, + struct btrfs_path *path, + int *level, int root_level) +{ + int ret = 0; + int nritems; + nritems = btrfs_header_nritems(path->nodes[*level]); + + path->slots[*level]++; + + while (path->slots[*level] == nritems) { + if (*level == root_level) + return -1; + + /* move upnext */ + path->slots[*level] = 0; + free_extent_buffer(path->nodes[*level]); + path->nodes[*level] = NULL; + (*level)++; + path->slots[*level]++; + + nritems = btrfs_header_nritems(path->nodes[*level]); + ret = 1; + } + return ret; +} + +/* + * Returns 1 if it had to move up and next. 0 is returned if it moved only next + * or down. + */ +static int tree_advance(struct btrfs_root *root, + struct btrfs_path *path, + int *level, int root_level, + int allow_down, + struct btrfs_key *key) +{ + int ret; + + if (*level == 0 || !allow_down) { + ret = tree_move_next_or_upnext(root, path, level, root_level); + } else { + tree_move_down(root, path, level, root_level); + ret = 0; + } + if (ret >= 0) { + if (*level == 0) + btrfs_item_key_to_cpu(path->nodes[*level], key, + path->slots[*level]); + else + btrfs_node_key_to_cpu(path->nodes[*level], key, + path->slots[*level]); + } + return ret; +} + +static int tree_compare_item(struct btrfs_root *left_root, + struct btrfs_path *left_path, + struct btrfs_path *right_path, + char *tmp_buf) +{ + int cmp; + int len1, len2; + unsigned long off1, off2; + + len1 = btrfs_item_size_nr(left_path->nodes[0], left_path->slots[0]); + len2 = btrfs_item_size_nr(right_path->nodes[0], right_path->slots[0]); + if (len1 != len2) + return 1; + + off1 = btrfs_item_ptr_offset(left_path->nodes[0], left_path->slots[0]); + off2 = btrfs_item_ptr_offset(right_path->nodes[0], + right_path->slots[0]); + + read_extent_buffer(left_path->nodes[0], tmp_buf, off1, len1); + + cmp = memcmp_extent_buffer(right_path->nodes[0], tmp_buf, off2, len1); + if (cmp) + return 1; + return 0; +} + +#define ADVANCE 1 +#define ADVANCE_ONLY_NEXT -1 + +/* + * This function compares two trees and calls the provided callback for + * every changed/new/deleted item it finds. + * If shared tree blocks are encountered, whole subtrees are skipped, making + * the compare pretty fast on snapshotted subvolumes. + * + * This currently works on commit roots only. As commit roots are read only, + * we don't do any locking. The commit roots are protected with transactions. + * Transactions are ended and rejoined when a commit is tried in between. + * + * This function checks for modifications done to the trees while comparing. + * If it detects a change, it aborts immediately. + */ +int btrfs_compare_trees(struct btrfs_root *left_root, + struct btrfs_root *right_root, + btrfs_changed_cb_t changed_cb, void *ctx) +{ + int ret; + int cmp; + struct btrfs_trans_handle *trans = NULL; + struct btrfs_path *left_path = NULL; + struct btrfs_path *right_path = NULL; + struct btrfs_key left_key; + struct btrfs_key right_key; + char *tmp_buf = NULL; + int left_root_level; + int right_root_level; + int left_level; + int right_level; + int left_end_reached; + int right_end_reached; + int advance_left; + int advance_right; + u64 left_blockptr; + u64 right_blockptr; + u64 left_start_ctransid; + u64 right_start_ctransid; + u64 ctransid; + + left_path = btrfs_alloc_path(); + if (!left_path) { + ret = -ENOMEM; + goto out; + } + right_path = btrfs_alloc_path(); + if (!right_path) { + ret = -ENOMEM; + goto out; + } + + tmp_buf = kmalloc(left_root->leafsize, GFP_NOFS); + if (!tmp_buf) { + ret = -ENOMEM; + goto out; + } + + left_path->search_commit_root = 1; + left_path->skip_locking = 1; + right_path->search_commit_root = 1; + right_path->skip_locking = 1; + + spin_lock(&left_root->root_times_lock); + left_start_ctransid = btrfs_root_ctransid(&left_root->root_item); + spin_unlock(&left_root->root_times_lock); + + spin_lock(&right_root->root_times_lock); + right_start_ctransid = btrfs_root_ctransid(&right_root->root_item); + spin_unlock(&right_root->root_times_lock); + + trans = btrfs_join_transaction(left_root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto out; + } + + /* + * Strategy: Go to the first items of both trees. Then do + * + * If both trees are at level 0 + * Compare keys of current items + * If left < right treat left item as new, advance left tree + * and repeat + * If left > right treat right item as deleted, advance right tree + * and repeat + * If left == right do deep compare of items, treat as changed if + * needed, advance both trees and repeat + * If both trees are at the same level but not at level 0 + * Compare keys of current nodes/leafs + * If left < right advance left tree and repeat + * If left > right advance right tree and repeat + * If left == right compare blockptrs of the next nodes/leafs + * If they match advance both trees but stay at the same level + * and repeat + * If they don't match advance both trees while allowing to go + * deeper and repeat + * If tree levels are different + * Advance the tree that needs it and repeat + * + * Advancing a tree means: + * If we are at level 0, try to go to the next slot. If that's not + * possible, go one level up and repeat. Stop when we found a level + * where we could go to the next slot. We may at this point be on a + * node or a leaf. + * + * If we are not at level 0 and not on shared tree blocks, go one + * level deeper. + * + * If we are not at level 0 and on shared tree blocks, go one slot to + * the right if possible or go up and right. + */ + + left_level = btrfs_header_level(left_root->commit_root); + left_root_level = left_level; + left_path->nodes[left_level] = left_root->commit_root; + extent_buffer_get(left_path->nodes[left_level]); + + right_level = btrfs_header_level(right_root->commit_root); + right_root_level = right_level; + right_path->nodes[right_level] = right_root->commit_root; + extent_buffer_get(right_path->nodes[right_level]); + + if (left_level == 0) + btrfs_item_key_to_cpu(left_path->nodes[left_level], + &left_key, left_path->slots[left_level]); + else + btrfs_node_key_to_cpu(left_path->nodes[left_level], + &left_key, left_path->slots[left_level]); + if (right_level == 0) + btrfs_item_key_to_cpu(right_path->nodes[right_level], + &right_key, right_path->slots[right_level]); + else + btrfs_node_key_to_cpu(right_path->nodes[right_level], + &right_key, right_path->slots[right_level]); + + left_end_reached = right_end_reached = 0; + advance_left = advance_right = 0; + + while (1) { + /* + * We need to make sure the transaction does not get committed + * while we do anything on commit roots. This means, we need to + * join and leave transactions for every item that we process. + */ + if (trans && btrfs_should_end_transaction(trans, left_root)) { + btrfs_release_path(left_path); + btrfs_release_path(right_path); + + ret = btrfs_end_transaction(trans, left_root); + trans = NULL; + if (ret < 0) + goto out; + } + /* now rejoin the transaction */ + if (!trans) { + trans = btrfs_join_transaction(left_root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto out; + } + + spin_lock(&left_root->root_times_lock); + ctransid = btrfs_root_ctransid(&left_root->root_item); + spin_unlock(&left_root->root_times_lock); + if (ctransid != left_start_ctransid) + left_start_ctransid = 0; + + spin_lock(&right_root->root_times_lock); + ctransid = btrfs_root_ctransid(&right_root->root_item); + spin_unlock(&right_root->root_times_lock); + if (ctransid != right_start_ctransid) + right_start_ctransid = 0; + + if (!left_start_ctransid || !right_start_ctransid) { + WARN(1, KERN_WARNING + "btrfs: btrfs_compare_tree detected " + "a change in one of the trees while " + "iterating. This is probably a " + "bug.\n"); + ret = -EIO; + goto out; + } + + /* + * the commit root may have changed, so start again + * where we stopped + */ + left_path->lowest_level = left_level; + right_path->lowest_level = right_level; + ret = btrfs_search_slot(NULL, left_root, + &left_key, left_path, 0, 0); + if (ret < 0) + goto out; + ret = btrfs_search_slot(NULL, right_root, + &right_key, right_path, 0, 0); + if (ret < 0) + goto out; + } + + if (advance_left && !left_end_reached) { + ret = tree_advance(left_root, left_path, &left_level, + left_root_level, + advance_left != ADVANCE_ONLY_NEXT, + &left_key); + if (ret < 0) + left_end_reached = ADVANCE; + advance_left = 0; + } + if (advance_right && !right_end_reached) { + ret = tree_advance(right_root, right_path, &right_level, + right_root_level, + advance_right != ADVANCE_ONLY_NEXT, + &right_key); + if (ret < 0) + right_end_reached = ADVANCE; + advance_right = 0; + } + + if (left_end_reached && right_end_reached) { + ret = 0; + goto out; + } else if (left_end_reached) { + if (right_level == 0) { + ret = changed_cb(left_root, right_root, + left_path, right_path, + &right_key, + BTRFS_COMPARE_TREE_DELETED, + ctx); + if (ret < 0) + goto out; + } + advance_right = ADVANCE; + continue; + } else if (right_end_reached) { + if (left_level == 0) { + ret = changed_cb(left_root, right_root, + left_path, right_path, + &left_key, + BTRFS_COMPARE_TREE_NEW, + ctx); + if (ret < 0) + goto out; + } + advance_left = ADVANCE; + continue; + } + + if (left_level == 0 && right_level == 0) { + cmp = btrfs_comp_cpu_keys(&left_key, &right_key); + if (cmp < 0) { + ret = changed_cb(left_root, right_root, + left_path, right_path, + &left_key, + BTRFS_COMPARE_TREE_NEW, + ctx); + if (ret < 0) + goto out; + advance_left = ADVANCE; + } else if (cmp > 0) { + ret = changed_cb(left_root, right_root, + left_path, right_path, + &right_key, + BTRFS_COMPARE_TREE_DELETED, + ctx); + if (ret < 0) + goto out; + advance_right = ADVANCE; + } else { + ret = tree_compare_item(left_root, left_path, + right_path, tmp_buf); + if (ret) { + ret = changed_cb(left_root, right_root, + left_path, right_path, + &left_key, + BTRFS_COMPARE_TREE_CHANGED, + ctx); + if (ret < 0) + goto out; + } + advance_left = ADVANCE; + advance_right = ADVANCE; + } + } else if (left_level == right_level) { + cmp = btrfs_comp_cpu_keys(&left_key, &right_key); + if (cmp < 0) { + advance_left = ADVANCE; + } else if (cmp > 0) { + advance_right = ADVANCE; + } else { + left_blockptr = btrfs_node_blockptr( + left_path->nodes[left_level], + left_path->slots[left_level]); + right_blockptr = btrfs_node_blockptr( + right_path->nodes[right_level], + right_path->slots[right_level]); + if (left_blockptr == right_blockptr) { + /* + * As we're on a shared block, don't + * allow to go deeper. + */ + advance_left = ADVANCE_ONLY_NEXT; + advance_right = ADVANCE_ONLY_NEXT; + } else { + advance_left = ADVANCE; + advance_right = ADVANCE; + } + } + } else if (left_level < right_level) { + advance_right = ADVANCE; + } else { + advance_left = ADVANCE; + } + } + +out: + btrfs_free_path(left_path); + btrfs_free_path(right_path); + kfree(tmp_buf); + + if (trans) { + if (!ret) + ret = btrfs_end_transaction(trans, left_root); + else + btrfs_end_transaction(trans, left_root); + } + + return ret; +} + /* * this is similar to btrfs_next_leaf, but does not try to preserve * and fixup the path. It looks for and returns the next key in the @@ -4135,7 +5566,8 @@ next: struct extent_buffer *cur; cur = btrfs_find_tree_block(root, blockptr, btrfs_level_size(root, level - 1)); - if (!cur || !btrfs_buffer_uptodate(cur, gen)) { + if (!cur || + btrfs_buffer_uptodate(cur, gen, 1) <= 0) { slot++; if (cur) free_extent_buffer(cur); @@ -4161,6 +5593,12 @@ next: */ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path) { + return btrfs_next_old_leaf(root, path, 0); +} + +int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path, + u64 time_seq) +{ int slot; int level; struct extent_buffer *c; @@ -4185,7 +5623,10 @@ again: path->keep_locks = 1; path->leave_spinning = 1; - ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (time_seq) + ret = btrfs_search_old_slot(root, &key, path, time_seq); + else + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); path->keep_locks = 0; if (ret < 0) @@ -4230,7 +5671,7 @@ again: next = c; next_rw_lock = path->locks[level]; ret = read_block_for_search(NULL, root, path, &next, level, - slot, &key); + slot, &key, 0); if (ret == -EAGAIN) goto again; @@ -4241,6 +5682,19 @@ again: if (!path->skip_locking) { ret = btrfs_try_tree_read_lock(next); + if (!ret && time_seq) { + /* + * If we don't get the lock, we may be racing + * with push_leaf_left, holding that lock while + * itself waiting for the leaf we've currently + * locked. To solve this situation, we give up + * on our lock and cycle. + */ + free_extent_buffer(next); + btrfs_release_path(path); + cond_resched(); + goto again; + } if (!ret) { btrfs_set_path_blocking(path); btrfs_tree_read_lock(next); @@ -4267,7 +5721,7 @@ again: break; ret = read_block_for_search(NULL, root, path, &next, level, - 0, &key); + 0, &key, 0); if (ret == -EAGAIN) goto again; @@ -4289,7 +5743,7 @@ again: } ret = 0; done: - unlock_up(path, 0, 1); + unlock_up(path, 0, 1, 0, NULL); path->leave_spinning = old_spinning; if (!old_spinning) btrfs_set_path_blocking(path); diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 03912c5c6f49..0d195b507660 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -30,6 +30,7 @@ #include <linux/kobject.h> #include <trace/events/btrfs.h> #include <asm/kmap_types.h> +#include <linux/pagemap.h> #include "extent_io.h" #include "extent_map.h" #include "async-thread.h" @@ -47,6 +48,8 @@ struct btrfs_ordered_sum; #define BTRFS_MAGIC "_BHRfS_M" +#define BTRFS_MAX_MIRRORS 2 + #define BTRFS_MAX_LEVEL 8 #define BTRFS_COMPAT_EXTENT_TREE_V0 @@ -85,6 +88,12 @@ struct btrfs_ordered_sum; /* holds checksums of all the data extents */ #define BTRFS_CSUM_TREE_OBJECTID 7ULL +/* for storing balance parameters in the root tree */ +#define BTRFS_BALANCE_OBJECTID -4ULL + +/* holds quota configuration and tracking */ +#define BTRFS_QUOTA_TREE_OBJECTID 8ULL + /* orhpan objectid for tracking unlinked/truncated files */ #define BTRFS_ORPHAN_OBJECTID -5ULL @@ -134,6 +143,12 @@ struct btrfs_ordered_sum; #define BTRFS_EMPTY_SUBVOL_DIR_OBJECTID 2 /* + * the max metadata block size. This limit is somewhat artificial, + * but the memmove costs go through the roof for larger blocks. + */ +#define BTRFS_MAX_METADATA_BLOCKSIZE 65536 + +/* * we can actually store much bigger names, but lets not confuse the rest * of linux */ @@ -161,6 +176,9 @@ static int btrfs_csum_sizes[] = { 4, 0 }; #define BTRFS_FT_XATTR 8 #define BTRFS_FT_MAX 9 +/* ioprio of readahead is set to idle */ +#define BTRFS_IOPRIO_READA (IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)) + /* * The key defines the order in the tree, and so it also defines (optimal) * block layout. @@ -360,6 +378,47 @@ struct btrfs_header { #define BTRFS_LABEL_SIZE 256 /* + * just in case we somehow lose the roots and are not able to mount, + * we store an array of the roots from previous transactions + * in the super. + */ +#define BTRFS_NUM_BACKUP_ROOTS 4 +struct btrfs_root_backup { + __le64 tree_root; + __le64 tree_root_gen; + + __le64 chunk_root; + __le64 chunk_root_gen; + + __le64 extent_root; + __le64 extent_root_gen; + + __le64 fs_root; + __le64 fs_root_gen; + + __le64 dev_root; + __le64 dev_root_gen; + + __le64 csum_root; + __le64 csum_root_gen; + + __le64 total_bytes; + __le64 bytes_used; + __le64 num_devices; + /* future */ + __le64 unsed_64[4]; + + u8 tree_root_level; + u8 chunk_root_level; + u8 extent_root_level; + u8 fs_root_level; + u8 dev_root_level; + u8 csum_root_level; + /* future and to align */ + u8 unused_8[10]; +} __attribute__ ((__packed__)); + +/* * the super block basically lists the main trees of the FS * it currently lacks any block count etc etc */ @@ -405,6 +464,7 @@ struct btrfs_super_block { /* future expansion */ __le64 reserved[31]; u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE]; + struct btrfs_root_backup super_roots[BTRFS_NUM_BACKUP_ROOTS]; } __attribute__ ((__packed__)); /* @@ -415,6 +475,19 @@ struct btrfs_super_block { #define BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL (1ULL << 1) #define BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS (1ULL << 2) #define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO (1ULL << 3) +/* + * some patches floated around with a second compression method + * lets save that incompat here for when they do get in + * Note we don't actually support it, we're just reserving the + * number + */ +#define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZOv2 (1ULL << 4) + +/* + * older kernels tried to do bigger metadata blocks, but the + * code was pretty buggy. Lets not let them try anymore. + */ +#define BTRFS_FEATURE_INCOMPAT_BIG_METADATA (1ULL << 5) #define BTRFS_FEATURE_COMPAT_SUPP 0ULL #define BTRFS_FEATURE_COMPAT_RO_SUPP 0ULL @@ -422,6 +495,7 @@ struct btrfs_super_block { (BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF | \ BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL | \ BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS | \ + BTRFS_FEATURE_INCOMPAT_BIG_METADATA | \ BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO) /* @@ -638,6 +712,36 @@ struct btrfs_root_item { struct btrfs_disk_key drop_progress; u8 drop_level; u8 level; + + /* + * The following fields appear after subvol_uuids+subvol_times + * were introduced. + */ + + /* + * This generation number is used to test if the new fields are valid + * and up to date while reading the root item. Everytime the root item + * is written out, the "generation" field is copied into this field. If + * anyone ever mounted the fs with an older kernel, we will have + * mismatching generation values here and thus must invalidate the + * new fields. See btrfs_update_root and btrfs_find_last_root for + * details. + * the offset of generation_v2 is also used as the start for the memset + * when invalidating the fields. + */ + __le64 generation_v2; + u8 uuid[BTRFS_UUID_SIZE]; + u8 parent_uuid[BTRFS_UUID_SIZE]; + u8 received_uuid[BTRFS_UUID_SIZE]; + __le64 ctransid; /* updated when an inode changes */ + __le64 otransid; /* trans when created */ + __le64 stransid; /* trans when sent. non-zero for received subvol */ + __le64 rtransid; /* trans when received. non-zero for received subvol */ + struct btrfs_timespec ctime; + struct btrfs_timespec otime; + struct btrfs_timespec stime; + struct btrfs_timespec rtime; + __le64 reserved[8]; /* for future */ } __attribute__ ((__packed__)); /* @@ -649,6 +753,54 @@ struct btrfs_root_ref { __le16 name_len; } __attribute__ ((__packed__)); +struct btrfs_disk_balance_args { + /* + * profiles to operate on, single is denoted by + * BTRFS_AVAIL_ALLOC_BIT_SINGLE + */ + __le64 profiles; + + /* usage filter */ + __le64 usage; + + /* devid filter */ + __le64 devid; + + /* devid subset filter [pstart..pend) */ + __le64 pstart; + __le64 pend; + + /* btrfs virtual address space subset filter [vstart..vend) */ + __le64 vstart; + __le64 vend; + + /* + * profile to convert to, single is denoted by + * BTRFS_AVAIL_ALLOC_BIT_SINGLE + */ + __le64 target; + + /* BTRFS_BALANCE_ARGS_* */ + __le64 flags; + + __le64 unused[8]; +} __attribute__ ((__packed__)); + +/* + * store balance parameters to disk so that balance can be properly + * resumed after crash or unmount + */ +struct btrfs_balance_item { + /* BTRFS_BALANCE_* */ + __le64 flags; + + struct btrfs_disk_balance_args data; + struct btrfs_disk_balance_args meta; + struct btrfs_disk_balance_args sys; + + __le64 unused[4]; +} __attribute__ ((__packed__)); + #define BTRFS_FILE_EXTENT_INLINE 0 #define BTRFS_FILE_EXTENT_REG 1 #define BTRFS_FILE_EXTENT_PREALLOC 2 @@ -707,15 +859,56 @@ struct btrfs_csum_item { u8 csum; } __attribute__ ((__packed__)); +struct btrfs_dev_stats_item { + /* + * grow this item struct at the end for future enhancements and keep + * the existing values unchanged + */ + __le64 values[BTRFS_DEV_STAT_VALUES_MAX]; +} __attribute__ ((__packed__)); + /* different types of block groups (and chunks) */ -#define BTRFS_BLOCK_GROUP_DATA (1 << 0) -#define BTRFS_BLOCK_GROUP_SYSTEM (1 << 1) -#define BTRFS_BLOCK_GROUP_METADATA (1 << 2) -#define BTRFS_BLOCK_GROUP_RAID0 (1 << 3) -#define BTRFS_BLOCK_GROUP_RAID1 (1 << 4) -#define BTRFS_BLOCK_GROUP_DUP (1 << 5) -#define BTRFS_BLOCK_GROUP_RAID10 (1 << 6) -#define BTRFS_NR_RAID_TYPES 5 +#define BTRFS_BLOCK_GROUP_DATA (1ULL << 0) +#define BTRFS_BLOCK_GROUP_SYSTEM (1ULL << 1) +#define BTRFS_BLOCK_GROUP_METADATA (1ULL << 2) +#define BTRFS_BLOCK_GROUP_RAID0 (1ULL << 3) +#define BTRFS_BLOCK_GROUP_RAID1 (1ULL << 4) +#define BTRFS_BLOCK_GROUP_DUP (1ULL << 5) +#define BTRFS_BLOCK_GROUP_RAID10 (1ULL << 6) +#define BTRFS_BLOCK_GROUP_RESERVED BTRFS_AVAIL_ALLOC_BIT_SINGLE +#define BTRFS_NR_RAID_TYPES 5 + +#define BTRFS_BLOCK_GROUP_TYPE_MASK (BTRFS_BLOCK_GROUP_DATA | \ + BTRFS_BLOCK_GROUP_SYSTEM | \ + BTRFS_BLOCK_GROUP_METADATA) + +#define BTRFS_BLOCK_GROUP_PROFILE_MASK (BTRFS_BLOCK_GROUP_RAID0 | \ + BTRFS_BLOCK_GROUP_RAID1 | \ + BTRFS_BLOCK_GROUP_DUP | \ + BTRFS_BLOCK_GROUP_RAID10) +/* + * We need a bit for restriper to be able to tell when chunks of type + * SINGLE are available. This "extended" profile format is used in + * fs_info->avail_*_alloc_bits (in-memory) and balance item fields + * (on-disk). The corresponding on-disk bit in chunk.type is reserved + * to avoid remappings between two formats in future. + */ +#define BTRFS_AVAIL_ALLOC_BIT_SINGLE (1ULL << 48) + +#define BTRFS_EXTENDED_PROFILE_MASK (BTRFS_BLOCK_GROUP_PROFILE_MASK | \ + BTRFS_AVAIL_ALLOC_BIT_SINGLE) + +static inline u64 chunk_to_extended(u64 flags) +{ + if ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0) + flags |= BTRFS_AVAIL_ALLOC_BIT_SINGLE; + + return flags; +} +static inline u64 extended_to_chunk(u64 flags) +{ + return flags & ~BTRFS_AVAIL_ALLOC_BIT_SINGLE; +} struct btrfs_block_group_item { __le64 used; @@ -723,6 +916,72 @@ struct btrfs_block_group_item { __le64 flags; } __attribute__ ((__packed__)); +/* + * is subvolume quota turned on? + */ +#define BTRFS_QGROUP_STATUS_FLAG_ON (1ULL << 0) +/* + * SCANNING is set during the initialization phase + */ +#define BTRFS_QGROUP_STATUS_FLAG_SCANNING (1ULL << 1) +/* + * Some qgroup entries are known to be out of date, + * either because the configuration has changed in a way that + * makes a rescan necessary, or because the fs has been mounted + * with a non-qgroup-aware version. + * Turning qouta off and on again makes it inconsistent, too. + */ +#define BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT (1ULL << 2) + +#define BTRFS_QGROUP_STATUS_VERSION 1 + +struct btrfs_qgroup_status_item { + __le64 version; + /* + * the generation is updated during every commit. As older + * versions of btrfs are not aware of qgroups, it will be + * possible to detect inconsistencies by checking the + * generation on mount time + */ + __le64 generation; + + /* flag definitions see above */ + __le64 flags; + + /* + * only used during scanning to record the progress + * of the scan. It contains a logical address + */ + __le64 scan; +} __attribute__ ((__packed__)); + +struct btrfs_qgroup_info_item { + __le64 generation; + __le64 rfer; + __le64 rfer_cmpr; + __le64 excl; + __le64 excl_cmpr; +} __attribute__ ((__packed__)); + +/* flags definition for qgroup limits */ +#define BTRFS_QGROUP_LIMIT_MAX_RFER (1ULL << 0) +#define BTRFS_QGROUP_LIMIT_MAX_EXCL (1ULL << 1) +#define BTRFS_QGROUP_LIMIT_RSV_RFER (1ULL << 2) +#define BTRFS_QGROUP_LIMIT_RSV_EXCL (1ULL << 3) +#define BTRFS_QGROUP_LIMIT_RFER_CMPR (1ULL << 4) +#define BTRFS_QGROUP_LIMIT_EXCL_CMPR (1ULL << 5) + +struct btrfs_qgroup_limit_item { + /* + * only updated when any of the other values change + */ + __le64 flags; + __le64 max_rfer; + __le64 max_excl; + __le64 rsv_rfer; + __le64 rsv_excl; +} __attribute__ ((__packed__)); + struct btrfs_space_info { u64 flags; @@ -772,15 +1031,9 @@ struct btrfs_space_info { struct btrfs_block_rsv { u64 size; u64 reserved; - u64 freed[2]; struct btrfs_space_info *space_info; - struct list_head list; spinlock_t lock; - atomic_t usage; - unsigned int priority:8; - unsigned int durable:1; - unsigned int refill_used:1; - unsigned int full:1; + unsigned int full; }; /* @@ -811,7 +1064,8 @@ struct btrfs_free_cluster { enum btrfs_caching_type { BTRFS_CACHE_NO = 0, BTRFS_CACHE_STARTED = 1, - BTRFS_CACHE_FINISHED = 2, + BTRFS_CACHE_FAST = 2, + BTRFS_CACHE_FINISHED = 3, }; enum btrfs_disk_cache_state { @@ -840,10 +1094,10 @@ struct btrfs_block_group_cache { spinlock_t lock; u64 pinned; u64 reserved; - u64 reserved_pinned; u64 bytes_super; u64 flags; u64 sectorsize; + u64 cache_generation; unsigned int ro:1; unsigned int dirty:1; unsigned int iref:1; @@ -875,9 +1129,17 @@ struct btrfs_block_group_cache { struct list_head cluster_list; }; +/* delayed seq elem */ +struct seq_list { + struct list_head list; + u64 seq; +}; + +/* fs_info */ struct reloc_control; struct btrfs_device; struct btrfs_fs_devices; +struct btrfs_balance_control; struct btrfs_delayed_root; struct btrfs_fs_info { u8 fsid[BTRFS_FSID_SIZE]; @@ -888,6 +1150,7 @@ struct btrfs_fs_info { struct btrfs_root *dev_root; struct btrfs_root *fs_root; struct btrfs_root *csum_root; + struct btrfs_root *quota_root; /* the log root tree is a directory of all the other log roots */ struct btrfs_root *log_root_tree; @@ -899,6 +1162,10 @@ struct btrfs_fs_info { spinlock_t block_group_cache_lock; struct rb_root block_group_cache_tree; + /* keep track of unallocated space */ + spinlock_t free_chunk_lock; + u64 free_chunk_space; + struct extent_io_tree freed_extents[2]; struct extent_io_tree *pinned_extents; @@ -916,14 +1183,11 @@ struct btrfs_fs_info { struct btrfs_block_rsv trans_block_rsv; /* block reservation for chunk tree */ struct btrfs_block_rsv chunk_block_rsv; + /* block reservation for delayed operations */ + struct btrfs_block_rsv delayed_block_rsv; struct btrfs_block_rsv empty_block_rsv; - /* list of block reservations that cross multiple transactions */ - struct list_head durable_block_rsv_list; - - struct mutex durable_block_rsv_mutex; - u64 generation; u64 last_trans_committed; @@ -932,7 +1196,7 @@ struct btrfs_fs_info { * is required instead of the faster short fsync log commits */ u64 last_trans_log_full_commit; - unsigned long mount_opt:20; + unsigned long mount_opt; unsigned long compress_type:4; u64 max_inline; u64 alloc_start; @@ -942,8 +1206,8 @@ struct btrfs_fs_info { wait_queue_head_t transaction_blocked_wait; wait_queue_head_t async_submit_wait; - struct btrfs_super_block super_copy; - struct btrfs_super_block super_for_commit; + struct btrfs_super_block *super_copy; + struct btrfs_super_block *super_for_commit; struct block_device *__bdev; struct super_block *sb; struct inode *btree_inode; @@ -983,6 +1247,16 @@ struct btrfs_fs_info { spinlock_t delayed_iput_lock; struct list_head delayed_iputs; + /* this protects tree_mod_seq_list */ + spinlock_t tree_mod_seq_lock; + atomic_t tree_mod_seq; + struct list_head tree_mod_seq_list; + struct seq_list tree_mod_seq_elem; + + /* this protects tree_mod_log */ + rwlock_t tree_mod_log_lock; + struct rb_root tree_mod_log; + atomic_t nr_async_submits; atomic_t async_submit_draining; atomic_t nr_async_bios; @@ -1036,6 +1310,7 @@ struct btrfs_fs_info { struct btrfs_workers endio_freespace_worker; struct btrfs_workers submit_workers; struct btrfs_workers caching_workers; + struct btrfs_workers readahead_workers; /* * fixup workers take dirty pages that didn't properly go through @@ -1073,6 +1348,8 @@ struct btrfs_fs_info { */ struct list_head space_info; + struct btrfs_space_info *data_sinfo; + struct reloc_control *reloc_ctl; spinlock_t delalloc_lock; @@ -1092,12 +1369,23 @@ struct btrfs_fs_info { spinlock_t ref_cache_lock; u64 total_ref_cache_size; + /* + * these three are in extended format (availability of single + * chunks is denoted by BTRFS_AVAIL_ALLOC_BIT_SINGLE bit, other + * types are denoted by corresponding BTRFS_BLOCK_GROUP_* bits) + */ u64 avail_data_alloc_bits; u64 avail_metadata_alloc_bits; u64 avail_system_alloc_bits; - u64 data_alloc_profile; - u64 metadata_alloc_profile; - u64 system_alloc_profile; + + /* restriper state */ + spinlock_t balance_lock; + struct mutex balance_mutex; + atomic_t balance_running; + atomic_t balance_pause_req; + atomic_t balance_cancel_req; + struct btrfs_balance_control *balance_ctl; + wait_queue_head_t balance_wait_q; unsigned data_chunk_allocations; unsigned metadata_ratio; @@ -1115,10 +1403,44 @@ struct btrfs_fs_info { int scrub_workers_refcnt; struct btrfs_workers scrub_workers; +#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY + u32 check_integrity_print_mask; +#endif + /* + * quota information + */ + unsigned int quota_enabled:1; + + /* + * quota_enabled only changes state after a commit. This holds the + * next state. + */ + unsigned int pending_quota_state:1; + + /* is qgroup tracking in a consistent state? */ + u64 qgroup_flags; + + /* holds configuration and tracking. Protected by qgroup_lock */ + struct rb_root qgroup_tree; + spinlock_t qgroup_lock; + + /* list of dirty qgroups to be written at next commit */ + struct list_head dirty_qgroups; + + /* used by btrfs_qgroup_record_ref for an efficient tree traversal */ + u64 qgroup_seq; + /* filesystem state */ u64 fs_state; struct btrfs_delayed_root *delayed_root; + + /* readahead tree */ + spinlock_t reada_lock; + struct radix_tree_root reada_tree; + + /* next backup root to be overwritten */ + int backup_root_index; }; /* @@ -1206,7 +1528,7 @@ struct btrfs_root { struct list_head root_list; spinlock_t orphan_lock; - struct list_head orphan_list; + atomic_t orphan_inodes; struct btrfs_block_rsv *orphan_block_rsv; int orphan_item_inserted; int orphan_cleanup_state; @@ -1225,6 +1547,10 @@ struct btrfs_root { * for stat. It may be used for more later */ dev_t anon_dev; + + int force_cow; + + spinlock_t root_times_lock; }; struct btrfs_ioctl_defrag_range_args { @@ -1335,6 +1661,38 @@ struct btrfs_ioctl_defrag_range_args { #define BTRFS_CHUNK_ITEM_KEY 228 /* + * Records the overall state of the qgroups. + * There's only one instance of this key present, + * (0, BTRFS_QGROUP_STATUS_KEY, 0) + */ +#define BTRFS_QGROUP_STATUS_KEY 240 +/* + * Records the currently used space of the qgroup. + * One key per qgroup, (0, BTRFS_QGROUP_INFO_KEY, qgroupid). + */ +#define BTRFS_QGROUP_INFO_KEY 242 +/* + * Contains the user configured limits for the qgroup. + * One key per qgroup, (0, BTRFS_QGROUP_LIMIT_KEY, qgroupid). + */ +#define BTRFS_QGROUP_LIMIT_KEY 244 +/* + * Records the child-parent relationship of qgroups. For + * each relation, 2 keys are present: + * (childid, BTRFS_QGROUP_RELATION_KEY, parentid) + * (parentid, BTRFS_QGROUP_RELATION_KEY, childid) + */ +#define BTRFS_QGROUP_RELATION_KEY 246 + +#define BTRFS_BALANCE_ITEM_KEY 248 + +/* + * Persistantly stores the io stats in the device tree. + * One key for all stats, (0, BTRFS_DEV_STATS_KEY, devid). + */ +#define BTRFS_DEV_STATS_KEY 249 + +/* * string items are for debugging. They just store a short string of * data in the FS */ @@ -1363,6 +1721,11 @@ struct btrfs_ioctl_defrag_range_args { #define BTRFS_MOUNT_ENOSPC_DEBUG (1 << 15) #define BTRFS_MOUNT_AUTO_DEFRAG (1 << 16) #define BTRFS_MOUNT_INODE_MAP_CACHE (1 << 17) +#define BTRFS_MOUNT_RECOVERY (1 << 18) +#define BTRFS_MOUNT_SKIP_BALANCE (1 << 19) +#define BTRFS_MOUNT_CHECK_INTEGRITY (1 << 20) +#define BTRFS_MOUNT_CHECK_INTEGRITY_INCLUDING_EXTENT_DATA (1 << 21) +#define BTRFS_MOUNT_PANIC_ON_FATAL_ERROR (1 << 22) #define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt) #define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt) @@ -1386,6 +1749,17 @@ struct btrfs_ioctl_defrag_range_args { #define BTRFS_INODE_ROOT_ITEM_INIT (1 << 31) +struct btrfs_map_token { + struct extent_buffer *eb; + char *kaddr; + unsigned long offset; +}; + +static inline void btrfs_init_map_token (struct btrfs_map_token *token) +{ + memset(token, 0, sizeof(*token)); +} + /* some macros to generate set/get funcs for the struct fields. This * assumes there is a lefoo_to_cpu for every type, so lets make a simple * one for u8: @@ -1406,23 +1780,66 @@ struct btrfs_ioctl_defrag_range_args { offsetof(type, member), \ sizeof(((type *)0)->member))) -#ifndef BTRFS_SETGET_FUNCS +#define DECLARE_BTRFS_SETGET_BITS(bits) \ +u##bits btrfs_get_token_##bits(struct extent_buffer *eb, void *ptr, \ + unsigned long off, \ + struct btrfs_map_token *token); \ +void btrfs_set_token_##bits(struct extent_buffer *eb, void *ptr, \ + unsigned long off, u##bits val, \ + struct btrfs_map_token *token); \ +static inline u##bits btrfs_get_##bits(struct extent_buffer *eb, void *ptr, \ + unsigned long off) \ +{ \ + return btrfs_get_token_##bits(eb, ptr, off, NULL); \ +} \ +static inline void btrfs_set_##bits(struct extent_buffer *eb, void *ptr, \ + unsigned long off, u##bits val) \ +{ \ + btrfs_set_token_##bits(eb, ptr, off, val, NULL); \ +} + +DECLARE_BTRFS_SETGET_BITS(8) +DECLARE_BTRFS_SETGET_BITS(16) +DECLARE_BTRFS_SETGET_BITS(32) +DECLARE_BTRFS_SETGET_BITS(64) + #define BTRFS_SETGET_FUNCS(name, type, member, bits) \ -u##bits btrfs_##name(struct extent_buffer *eb, type *s); \ -void btrfs_set_##name(struct extent_buffer *eb, type *s, u##bits val); -#endif +static inline u##bits btrfs_##name(struct extent_buffer *eb, type *s) \ +{ \ + BUILD_BUG_ON(sizeof(u##bits) != sizeof(((type *)0))->member); \ + return btrfs_get_##bits(eb, s, offsetof(type, member)); \ +} \ +static inline void btrfs_set_##name(struct extent_buffer *eb, type *s, \ + u##bits val) \ +{ \ + BUILD_BUG_ON(sizeof(u##bits) != sizeof(((type *)0))->member); \ + btrfs_set_##bits(eb, s, offsetof(type, member), val); \ +} \ +static inline u##bits btrfs_token_##name(struct extent_buffer *eb, type *s, \ + struct btrfs_map_token *token) \ +{ \ + BUILD_BUG_ON(sizeof(u##bits) != sizeof(((type *)0))->member); \ + return btrfs_get_token_##bits(eb, s, offsetof(type, member), token); \ +} \ +static inline void btrfs_set_token_##name(struct extent_buffer *eb, \ + type *s, u##bits val, \ + struct btrfs_map_token *token) \ +{ \ + BUILD_BUG_ON(sizeof(u##bits) != sizeof(((type *)0))->member); \ + btrfs_set_token_##bits(eb, s, offsetof(type, member), val, token); \ +} #define BTRFS_SETGET_HEADER_FUNCS(name, type, member, bits) \ static inline u##bits btrfs_##name(struct extent_buffer *eb) \ { \ - type *p = page_address(eb->first_page); \ + type *p = page_address(eb->pages[0]); \ u##bits res = le##bits##_to_cpu(p->member); \ return res; \ } \ static inline void btrfs_set_##name(struct extent_buffer *eb, \ u##bits val) \ { \ - type *p = page_address(eb->first_page); \ + type *p = page_address(eb->pages[0]); \ p->member = cpu_to_le##bits(val); \ } @@ -1972,14 +2389,151 @@ BTRFS_SETGET_STACK_FUNCS(root_used, struct btrfs_root_item, bytes_used, 64); BTRFS_SETGET_STACK_FUNCS(root_limit, struct btrfs_root_item, byte_limit, 64); BTRFS_SETGET_STACK_FUNCS(root_last_snapshot, struct btrfs_root_item, last_snapshot, 64); +BTRFS_SETGET_STACK_FUNCS(root_generation_v2, struct btrfs_root_item, + generation_v2, 64); +BTRFS_SETGET_STACK_FUNCS(root_ctransid, struct btrfs_root_item, + ctransid, 64); +BTRFS_SETGET_STACK_FUNCS(root_otransid, struct btrfs_root_item, + otransid, 64); +BTRFS_SETGET_STACK_FUNCS(root_stransid, struct btrfs_root_item, + stransid, 64); +BTRFS_SETGET_STACK_FUNCS(root_rtransid, struct btrfs_root_item, + rtransid, 64); static inline bool btrfs_root_readonly(struct btrfs_root *root) { - return root->root_item.flags & BTRFS_ROOT_SUBVOL_RDONLY; + return (root->root_item.flags & cpu_to_le64(BTRFS_ROOT_SUBVOL_RDONLY)) != 0; } -/* struct btrfs_super_block */ +/* struct btrfs_root_backup */ +BTRFS_SETGET_STACK_FUNCS(backup_tree_root, struct btrfs_root_backup, + tree_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_tree_root_gen, struct btrfs_root_backup, + tree_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_tree_root_level, struct btrfs_root_backup, + tree_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_chunk_root, struct btrfs_root_backup, + chunk_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_chunk_root_gen, struct btrfs_root_backup, + chunk_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_chunk_root_level, struct btrfs_root_backup, + chunk_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_extent_root, struct btrfs_root_backup, + extent_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_extent_root_gen, struct btrfs_root_backup, + extent_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_extent_root_level, struct btrfs_root_backup, + extent_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_fs_root, struct btrfs_root_backup, + fs_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_fs_root_gen, struct btrfs_root_backup, + fs_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_fs_root_level, struct btrfs_root_backup, + fs_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_dev_root, struct btrfs_root_backup, + dev_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_dev_root_gen, struct btrfs_root_backup, + dev_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_dev_root_level, struct btrfs_root_backup, + dev_root_level, 8); + +BTRFS_SETGET_STACK_FUNCS(backup_csum_root, struct btrfs_root_backup, + csum_root, 64); +BTRFS_SETGET_STACK_FUNCS(backup_csum_root_gen, struct btrfs_root_backup, + csum_root_gen, 64); +BTRFS_SETGET_STACK_FUNCS(backup_csum_root_level, struct btrfs_root_backup, + csum_root_level, 8); +BTRFS_SETGET_STACK_FUNCS(backup_total_bytes, struct btrfs_root_backup, + total_bytes, 64); +BTRFS_SETGET_STACK_FUNCS(backup_bytes_used, struct btrfs_root_backup, + bytes_used, 64); +BTRFS_SETGET_STACK_FUNCS(backup_num_devices, struct btrfs_root_backup, + num_devices, 64); + +/* struct btrfs_balance_item */ +BTRFS_SETGET_FUNCS(balance_flags, struct btrfs_balance_item, flags, 64); +static inline void btrfs_balance_data(struct extent_buffer *eb, + struct btrfs_balance_item *bi, + struct btrfs_disk_balance_args *ba) +{ + read_eb_member(eb, bi, struct btrfs_balance_item, data, ba); +} + +static inline void btrfs_set_balance_data(struct extent_buffer *eb, + struct btrfs_balance_item *bi, + struct btrfs_disk_balance_args *ba) +{ + write_eb_member(eb, bi, struct btrfs_balance_item, data, ba); +} + +static inline void btrfs_balance_meta(struct extent_buffer *eb, + struct btrfs_balance_item *bi, + struct btrfs_disk_balance_args *ba) +{ + read_eb_member(eb, bi, struct btrfs_balance_item, meta, ba); +} + +static inline void btrfs_set_balance_meta(struct extent_buffer *eb, + struct btrfs_balance_item *bi, + struct btrfs_disk_balance_args *ba) +{ + write_eb_member(eb, bi, struct btrfs_balance_item, meta, ba); +} + +static inline void btrfs_balance_sys(struct extent_buffer *eb, + struct btrfs_balance_item *bi, + struct btrfs_disk_balance_args *ba) +{ + read_eb_member(eb, bi, struct btrfs_balance_item, sys, ba); +} + +static inline void btrfs_set_balance_sys(struct extent_buffer *eb, + struct btrfs_balance_item *bi, + struct btrfs_disk_balance_args *ba) +{ + write_eb_member(eb, bi, struct btrfs_balance_item, sys, ba); +} + +static inline void +btrfs_disk_balance_args_to_cpu(struct btrfs_balance_args *cpu, + struct btrfs_disk_balance_args *disk) +{ + memset(cpu, 0, sizeof(*cpu)); + + cpu->profiles = le64_to_cpu(disk->profiles); + cpu->usage = le64_to_cpu(disk->usage); + cpu->devid = le64_to_cpu(disk->devid); + cpu->pstart = le64_to_cpu(disk->pstart); + cpu->pend = le64_to_cpu(disk->pend); + cpu->vstart = le64_to_cpu(disk->vstart); + cpu->vend = le64_to_cpu(disk->vend); + cpu->target = le64_to_cpu(disk->target); + cpu->flags = le64_to_cpu(disk->flags); +} + +static inline void +btrfs_cpu_balance_args_to_disk(struct btrfs_disk_balance_args *disk, + struct btrfs_balance_args *cpu) +{ + memset(disk, 0, sizeof(*disk)); + + disk->profiles = cpu_to_le64(cpu->profiles); + disk->usage = cpu_to_le64(cpu->usage); + disk->devid = cpu_to_le64(cpu->devid); + disk->pstart = cpu_to_le64(cpu->pstart); + disk->pend = cpu_to_le64(cpu->pend); + disk->vstart = cpu_to_le64(cpu->vstart); + disk->vend = cpu_to_le64(cpu->vend); + disk->target = cpu_to_le64(cpu->target); + disk->flags = cpu_to_le64(cpu->flags); +} + +/* struct btrfs_super_block */ BTRFS_SETGET_STACK_FUNCS(super_bytenr, struct btrfs_super_block, bytenr, 64); BTRFS_SETGET_STACK_FUNCS(super_flags, struct btrfs_super_block, flags, 64); BTRFS_SETGET_STACK_FUNCS(super_generation, struct btrfs_super_block, @@ -2097,7 +2651,74 @@ static inline u32 btrfs_file_extent_inline_item_len(struct extent_buffer *eb, return btrfs_item_size(eb, e) - offset; } -static inline struct btrfs_root *btrfs_sb(struct super_block *sb) +/* btrfs_dev_stats_item */ +static inline u64 btrfs_dev_stats_value(struct extent_buffer *eb, + struct btrfs_dev_stats_item *ptr, + int index) +{ + u64 val; + + read_extent_buffer(eb, &val, + offsetof(struct btrfs_dev_stats_item, values) + + ((unsigned long)ptr) + (index * sizeof(u64)), + sizeof(val)); + return val; +} + +static inline void btrfs_set_dev_stats_value(struct extent_buffer *eb, + struct btrfs_dev_stats_item *ptr, + int index, u64 val) +{ + write_extent_buffer(eb, &val, + offsetof(struct btrfs_dev_stats_item, values) + + ((unsigned long)ptr) + (index * sizeof(u64)), + sizeof(val)); +} + +/* btrfs_qgroup_status_item */ +BTRFS_SETGET_FUNCS(qgroup_status_generation, struct btrfs_qgroup_status_item, + generation, 64); +BTRFS_SETGET_FUNCS(qgroup_status_version, struct btrfs_qgroup_status_item, + version, 64); +BTRFS_SETGET_FUNCS(qgroup_status_flags, struct btrfs_qgroup_status_item, + flags, 64); +BTRFS_SETGET_FUNCS(qgroup_status_scan, struct btrfs_qgroup_status_item, + scan, 64); + +/* btrfs_qgroup_info_item */ +BTRFS_SETGET_FUNCS(qgroup_info_generation, struct btrfs_qgroup_info_item, + generation, 64); +BTRFS_SETGET_FUNCS(qgroup_info_rfer, struct btrfs_qgroup_info_item, rfer, 64); +BTRFS_SETGET_FUNCS(qgroup_info_rfer_cmpr, struct btrfs_qgroup_info_item, + rfer_cmpr, 64); +BTRFS_SETGET_FUNCS(qgroup_info_excl, struct btrfs_qgroup_info_item, excl, 64); +BTRFS_SETGET_FUNCS(qgroup_info_excl_cmpr, struct btrfs_qgroup_info_item, + excl_cmpr, 64); + +BTRFS_SETGET_STACK_FUNCS(stack_qgroup_info_generation, + struct btrfs_qgroup_info_item, generation, 64); +BTRFS_SETGET_STACK_FUNCS(stack_qgroup_info_rfer, struct btrfs_qgroup_info_item, + rfer, 64); +BTRFS_SETGET_STACK_FUNCS(stack_qgroup_info_rfer_cmpr, + struct btrfs_qgroup_info_item, rfer_cmpr, 64); +BTRFS_SETGET_STACK_FUNCS(stack_qgroup_info_excl, struct btrfs_qgroup_info_item, + excl, 64); +BTRFS_SETGET_STACK_FUNCS(stack_qgroup_info_excl_cmpr, + struct btrfs_qgroup_info_item, excl_cmpr, 64); + +/* btrfs_qgroup_limit_item */ +BTRFS_SETGET_FUNCS(qgroup_limit_flags, struct btrfs_qgroup_limit_item, + flags, 64); +BTRFS_SETGET_FUNCS(qgroup_limit_max_rfer, struct btrfs_qgroup_limit_item, + max_rfer, 64); +BTRFS_SETGET_FUNCS(qgroup_limit_max_excl, struct btrfs_qgroup_limit_item, + max_excl, 64); +BTRFS_SETGET_FUNCS(qgroup_limit_rsv_rfer, struct btrfs_qgroup_limit_item, + rsv_rfer, 64); +BTRFS_SETGET_FUNCS(qgroup_limit_rsv_excl, struct btrfs_qgroup_limit_item, + rsv_excl, 64); + +static inline struct btrfs_fs_info *btrfs_sb(struct super_block *sb) { return sb->s_fs_info; } @@ -2129,6 +2750,11 @@ static inline bool btrfs_mixed_space_info(struct btrfs_space_info *space_info) (space_info->flags & BTRFS_BLOCK_GROUP_DATA)); } +static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping) +{ + return mapping_gfp_mask(mapping) & ~__GFP_FS; +} + /* extent-tree.c */ static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root, unsigned num_items) @@ -2137,6 +2763,17 @@ static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root, 3 * num_items; } +/* + * Doing a truncate won't result in new nodes or leaves, just what we need for + * COW. + */ +static inline u64 btrfs_calc_trunc_metadata_size(struct btrfs_root *root, + unsigned num_items) +{ + return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) * + num_items; +} + void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); @@ -2146,6 +2783,9 @@ int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, u64 num_bytes, u64 *refs, u64 *flags); int btrfs_pin_extent(struct btrfs_root *root, u64 bytenr, u64 num, int reserved); +int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + u64 bytenr, u64 num_bytes); int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 objectid, u64 offset, u64 bytenr); @@ -2180,32 +2820,31 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 num_bytes, u64 min_alloc_size, u64 empty_size, u64 hint_byte, - u64 search_end, struct btrfs_key *ins, - u64 data); + struct btrfs_key *ins, u64 data); int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct extent_buffer *buf, int full_backref); + struct extent_buffer *buf, int full_backref, int for_cow); int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct extent_buffer *buf, int full_backref); + struct extent_buffer *buf, int full_backref, int for_cow); int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, u64 flags, int is_data); int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, - u64 bytenr, u64 num_bytes, u64 parent, - u64 root_objectid, u64 owner, u64 offset); + u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, + u64 owner, u64 offset, int for_cow); int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len); -int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo); -int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, - struct btrfs_root *root); +int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root, + u64 start, u64 len); +void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, + struct btrfs_root *root); int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, u64 parent, - u64 root_objectid, u64 owner, u64 offset); + u64 root_objectid, u64 owner, u64 offset, int for_cow); int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans, struct btrfs_root *root); @@ -2221,7 +2860,6 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 group_start); u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags); u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data); -void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *ionde); void btrfs_clear_space_info_full(struct btrfs_fs_info *info); int btrfs_check_data_free_space(struct inode *inode, u64 bytes); void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes); @@ -2240,29 +2878,30 @@ void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv); struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root); void btrfs_free_block_rsv(struct btrfs_root *root, struct btrfs_block_rsv *rsv); -void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info, - struct btrfs_block_rsv *rsv); -int btrfs_block_rsv_add(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +int btrfs_block_rsv_add(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 num_bytes); -int btrfs_block_rsv_check(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +int btrfs_block_rsv_add_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes); +int btrfs_block_rsv_check(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, int min_factor); +int btrfs_block_rsv_refill(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, - u64 min_reserved, int min_factor); + u64 min_reserved); +int btrfs_block_rsv_refill_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved); int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, struct btrfs_block_rsv *dst_rsv, u64 num_bytes); void btrfs_block_rsv_release(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 num_bytes); -int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *rsv); int btrfs_set_block_group_ro(struct btrfs_root *root, struct btrfs_block_group_cache *cache); -int btrfs_set_block_group_rw(struct btrfs_root *root, - struct btrfs_block_group_cache *cache); +void btrfs_set_block_group_rw(struct btrfs_root *root, + struct btrfs_block_group_cache *cache); void btrfs_put_block_group_cache(struct btrfs_fs_info *info); u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo); int btrfs_error_unpin_extent_range(struct btrfs_root *root, @@ -2274,6 +2913,8 @@ int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range); int btrfs_init_space_info(struct btrfs_fs_info *fs_info); +int btrfs_delayed_refs_qgroup_accounting(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info); /* ctree.c */ int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key, int level, int *slot); @@ -2281,9 +2922,9 @@ int btrfs_comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2); int btrfs_previous_item(struct btrfs_root *root, struct btrfs_path *path, u64 min_objectid, int type); -int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *new_key); +void btrfs_set_item_key_safe(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *new_key); struct extent_buffer *btrfs_root_node(struct btrfs_root *root); struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root); int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path, @@ -2293,6 +2934,21 @@ int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key, struct btrfs_key *max_key, struct btrfs_path *path, int cache_only, u64 min_trans); +enum btrfs_compare_tree_result { + BTRFS_COMPARE_TREE_NEW, + BTRFS_COMPARE_TREE_DELETED, + BTRFS_COMPARE_TREE_CHANGED, +}; +typedef int (*btrfs_changed_cb_t)(struct btrfs_root *left_root, + struct btrfs_root *right_root, + struct btrfs_path *left_path, + struct btrfs_path *right_path, + struct btrfs_key *key, + enum btrfs_compare_tree_result result, + void *ctx); +int btrfs_compare_trees(struct btrfs_root *left_root, + struct btrfs_root *right_root, + btrfs_changed_cb_t cb, void *ctx); int btrfs_cow_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf, struct extent_buffer *parent, int parent_slot, @@ -2303,12 +2959,13 @@ int btrfs_copy_root(struct btrfs_trans_handle *trans, struct extent_buffer **cow_ret, u64 new_root_objectid); int btrfs_block_can_be_shared(struct btrfs_root *root, struct extent_buffer *buf); -int btrfs_extend_item(struct btrfs_trans_handle *trans, struct btrfs_root - *root, struct btrfs_path *path, u32 data_size); -int btrfs_truncate_item(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - u32 new_size, int from_end); +void btrfs_extend_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + u32 data_size); +void btrfs_truncate_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + u32 new_size, int from_end); int btrfs_split_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, @@ -2321,6 +2978,11 @@ int btrfs_duplicate_item(struct btrfs_trans_handle *trans, int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_key *key, struct btrfs_path *p, int ins_len, int cow); +int btrfs_search_old_slot(struct btrfs_root *root, struct btrfs_key *key, + struct btrfs_path *p, u64 time_seq); +int btrfs_search_slot_for_read(struct btrfs_root *root, + struct btrfs_key *key, struct btrfs_path *p, + int find_higher, int return_any); int btrfs_realloc_node(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *parent, int start_slot, int cache_only, u64 *last_ret, @@ -2342,10 +3004,10 @@ static inline int btrfs_del_item(struct btrfs_trans_handle *trans, return btrfs_del_items(trans, root, path, path->slots[0], 1); } -int setup_items_for_insert(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - u32 total_data, u32 total_size, int nr); +void setup_items_for_insert(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *cpu_key, u32 *data_size, + u32 total_data, u32 total_size, int nr); int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_key *key, void *data, u32 data_size); int btrfs_insert_empty_items(struct btrfs_trans_handle *trans, @@ -2363,10 +3025,25 @@ static inline int btrfs_insert_empty_item(struct btrfs_trans_handle *trans, } int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path); +int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path, + u64 time_seq); +static inline int btrfs_next_old_item(struct btrfs_root *root, + struct btrfs_path *p, u64 time_seq) +{ + ++p->slots[0]; + if (p->slots[0] >= btrfs_header_nritems(p->nodes[0])) + return btrfs_next_old_leaf(root, p, time_seq); + return 0; +} +static inline int btrfs_next_item(struct btrfs_root *root, struct btrfs_path *p) +{ + return btrfs_next_old_item(root, p, 0); +} int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path); int btrfs_leaf_free_space(struct btrfs_root *root, struct extent_buffer *leaf); -void btrfs_drop_snapshot(struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, int update_ref); +int __must_check btrfs_drop_snapshot(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + int update_ref, int for_reloc); int btrfs_drop_subtree(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *node, @@ -2379,6 +3056,30 @@ static inline int btrfs_fs_closing(struct btrfs_fs_info *fs_info) smp_mb(); return fs_info->closing; } +static inline void free_fs_info(struct btrfs_fs_info *fs_info) +{ + kfree(fs_info->balance_ctl); + kfree(fs_info->delayed_root); + kfree(fs_info->extent_root); + kfree(fs_info->tree_root); + kfree(fs_info->chunk_root); + kfree(fs_info->dev_root); + kfree(fs_info->csum_root); + kfree(fs_info->quota_root); + kfree(fs_info->super_copy); + kfree(fs_info->super_for_commit); + kfree(fs_info); +} + +/* tree mod log functions from ctree.c */ +u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info, + struct seq_list *elem); +void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info, + struct seq_list *elem); +static inline u64 btrfs_inc_tree_mod_seq(struct btrfs_fs_info *fs_info) +{ + return atomic_inc_return(&fs_info->tree_mod_seq); +} /* root-item.c */ int btrfs_find_root_ref(struct btrfs_root *tree_root, @@ -2397,9 +3098,13 @@ int btrfs_del_root(struct btrfs_trans_handle *trans, struct btrfs_root *root, int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_key *key, struct btrfs_root_item *item); -int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root - *root, struct btrfs_key *key, struct btrfs_root_item - *item); +int __must_check btrfs_update_root(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_key *key, + struct btrfs_root_item *item); +void btrfs_read_root_item(struct btrfs_root *root, + struct extent_buffer *eb, int slot, + struct btrfs_root_item *item); int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct btrfs_root_item *item, struct btrfs_key *key); int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid); @@ -2407,6 +3112,8 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root); void btrfs_set_root_node(struct btrfs_root_item *item, struct extent_buffer *node); void btrfs_check_and_init_root_item(struct btrfs_root_item *item); +void btrfs_update_root_times(struct btrfs_trans_handle *trans, + struct btrfs_root *root); /* dir-item.c */ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, @@ -2484,7 +3191,7 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode, struct bio *bio, u32 *dst); int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode, - struct bio *bio, u64 logical_offset, u32 *dst); + struct bio *bio, u64 logical_offset); int btrfs_insert_file_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 objectid, u64 pos, @@ -2561,7 +3268,7 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf); int btrfs_readpage(struct file *file, struct page *page); void btrfs_evict_inode(struct inode *inode); int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc); -void btrfs_dirty_inode(struct inode *inode, int flags); +int btrfs_dirty_inode(struct inode *inode); struct inode *btrfs_alloc_inode(struct super_block *sb); void btrfs_destroy_inode(struct inode *inode); int btrfs_drop_inode(struct inode *inode); @@ -2579,15 +3286,10 @@ int btrfs_update_inode(struct btrfs_trans_handle *trans, int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_orphan_cleanup(struct btrfs_root *root); -void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending, - u64 *bytes_to_reserve); -void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending); void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, struct btrfs_root *root); int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size); -int btrfs_invalidate_inodes(struct btrfs_root *root); +void btrfs_invalidate_inodes(struct btrfs_root *root); void btrfs_add_delayed_iput(struct inode *inode); void btrfs_run_delayed_iputs(struct btrfs_root *root); int btrfs_prealloc_file_range(struct inode *inode, int mode, @@ -2639,13 +3341,71 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size); /* super.c */ int btrfs_parse_options(struct btrfs_root *root, char *options); int btrfs_sync_fs(struct super_block *sb, int wait); + +#ifdef CONFIG_PRINTK +__printf(2, 3) +void btrfs_printk(struct btrfs_fs_info *fs_info, const char *fmt, ...); +#else +static inline __printf(2, 3) +void btrfs_printk(struct btrfs_fs_info *fs_info, const char *fmt, ...) +{ +} +#endif + +__printf(5, 6) void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function, - unsigned int line, int errno); + unsigned int line, int errno, const char *fmt, ...); + + +void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, + struct btrfs_root *root, const char *function, + unsigned int line, int errno); + +#define btrfs_set_fs_incompat(__fs_info, opt) \ + __btrfs_set_fs_incompat((__fs_info), BTRFS_FEATURE_INCOMPAT_##opt) + +static inline void __btrfs_set_fs_incompat(struct btrfs_fs_info *fs_info, + u64 flag) +{ + struct btrfs_super_block *disk_super; + u64 features; + + disk_super = fs_info->super_copy; + features = btrfs_super_incompat_flags(disk_super); + if (!(features & flag)) { + features |= flag; + btrfs_set_super_incompat_flags(disk_super, features); + } +} + +#define btrfs_abort_transaction(trans, root, errno) \ +do { \ + __btrfs_abort_transaction(trans, root, __func__, \ + __LINE__, errno); \ +} while (0) #define btrfs_std_error(fs_info, errno) \ do { \ if ((errno)) \ - __btrfs_std_error((fs_info), __func__, __LINE__, (errno));\ + __btrfs_std_error((fs_info), __func__, \ + __LINE__, (errno), NULL); \ +} while (0) + +#define btrfs_error(fs_info, errno, fmt, args...) \ +do { \ + __btrfs_std_error((fs_info), __func__, __LINE__, \ + (errno), fmt, ##args); \ +} while (0) + +__printf(5, 6) +void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function, + unsigned int line, int errno, const char *fmt, ...); + +#define btrfs_panic(fs_info, errno, fmt, args...) \ +do { \ + struct btrfs_fs_info *_i = (fs_info); \ + __btrfs_panic(_i, __func__, __LINE__, errno, fmt, ##args); \ + BUG_ON(!(_i->mount_opt & BTRFS_MOUNT_PANIC_ON_FATAL_ERROR)); \ } while (0) /* acl.c */ @@ -2681,20 +3441,88 @@ void btrfs_reloc_cow_block(struct btrfs_trans_handle *trans, void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending, u64 *bytes_to_reserve); -void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, +int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending); /* scrub.c */ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, struct btrfs_scrub_progress *progress, int readonly); -int btrfs_scrub_pause(struct btrfs_root *root); -int btrfs_scrub_pause_super(struct btrfs_root *root); -int btrfs_scrub_continue(struct btrfs_root *root); -int btrfs_scrub_continue_super(struct btrfs_root *root); +void btrfs_scrub_pause(struct btrfs_root *root); +void btrfs_scrub_pause_super(struct btrfs_root *root); +void btrfs_scrub_continue(struct btrfs_root *root); +void btrfs_scrub_continue_super(struct btrfs_root *root); +int __btrfs_scrub_cancel(struct btrfs_fs_info *info); int btrfs_scrub_cancel(struct btrfs_root *root); int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev); int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid); int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, struct btrfs_scrub_progress *progress); +/* reada.c */ +struct reada_control { + struct btrfs_root *root; /* tree to prefetch */ + struct btrfs_key key_start; + struct btrfs_key key_end; /* exclusive */ + atomic_t elems; + struct kref refcnt; + wait_queue_head_t wait; +}; +struct reada_control *btrfs_reada_add(struct btrfs_root *root, + struct btrfs_key *start, struct btrfs_key *end); +int btrfs_reada_wait(void *handle); +void btrfs_reada_detach(void *handle); +int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb, + u64 start, int err); + +/* qgroup.c */ +struct qgroup_update { + struct list_head list; + struct btrfs_delayed_ref_node *node; + struct btrfs_delayed_extent_op *extent_op; +}; + +int btrfs_quota_enable(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info); +int btrfs_quota_disable(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info); +int btrfs_quota_rescan(struct btrfs_fs_info *fs_info); +int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 src, u64 dst); +int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 src, u64 dst); +int btrfs_create_qgroup(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 qgroupid, + char *name); +int btrfs_remove_qgroup(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 qgroupid); +int btrfs_limit_qgroup(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 qgroupid, + struct btrfs_qgroup_limit *limit); +int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info); +void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info); +struct btrfs_delayed_extent_op; +int btrfs_qgroup_record_ref(struct btrfs_trans_handle *trans, + struct btrfs_delayed_ref_node *node, + struct btrfs_delayed_extent_op *extent_op); +int btrfs_qgroup_account_ref(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *node, + struct btrfs_delayed_extent_op *extent_op); +int btrfs_run_qgroups(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info); +int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 srcid, u64 objectid, + struct btrfs_qgroup_inherit *inherit); +int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes); +void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes); + +void assert_qgroups_uptodate(struct btrfs_trans_handle *trans); + +static inline int is_fstree(u64 rootid) +{ + if (rootid == BTRFS_FS_TREE_OBJECTID || + (s64)rootid >= (s64)BTRFS_FIRST_FREE_OBJECTID) + return 1; + return 0; +} #endif diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index b52c672f4c18..07d5eeb1e6f1 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -62,6 +62,7 @@ static inline void btrfs_init_delayed_node( INIT_LIST_HEAD(&delayed_node->n_list); INIT_LIST_HEAD(&delayed_node->p_list); delayed_node->bytes_reserved = 0; + memset(&delayed_node->inode_item, 0, sizeof(delayed_node->inode_item)); } static inline int btrfs_is_continuous_delayed_item( @@ -115,6 +116,7 @@ static struct btrfs_delayed_node *btrfs_get_delayed_node(struct inode *inode) return NULL; } +/* Will return either the node or PTR_ERR(-ENOMEM) */ static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node( struct inode *inode) { @@ -510,8 +512,8 @@ static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) rb_erase(&delayed_item->rb_node, root); delayed_item->delayed_node->count--; - atomic_dec(&delayed_root->items); - if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND && + if (atomic_dec_return(&delayed_root->items) < + BTRFS_DELAYED_BACKGROUND && waitqueue_active(&delayed_root->wait)) wake_up(&delayed_root->wait); } @@ -591,12 +593,16 @@ static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, return 0; src_rsv = trans->block_rsv; - dst_rsv = &root->fs_info->global_block_rsv; + dst_rsv = &root->fs_info->delayed_block_rsv; num_bytes = btrfs_calc_trans_metadata_size(root, 1); ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); - if (!ret) + if (!ret) { + trace_btrfs_space_reservation(root->fs_info, "delayed_item", + item->key.objectid, + num_bytes, 1); item->bytes_reserved = num_bytes; + } return ret; } @@ -609,7 +615,10 @@ static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, if (!item->bytes_reserved) return; - rsv = &root->fs_info->global_block_rsv; + rsv = &root->fs_info->delayed_block_rsv; + trace_btrfs_space_reservation(root->fs_info, "delayed_item", + item->key.objectid, item->bytes_reserved, + 0); btrfs_block_rsv_release(root, rsv, item->bytes_reserved); } @@ -617,23 +626,112 @@ static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, static int btrfs_delayed_inode_reserve_metadata( struct btrfs_trans_handle *trans, struct btrfs_root *root, + struct inode *inode, struct btrfs_delayed_node *node) { struct btrfs_block_rsv *src_rsv; struct btrfs_block_rsv *dst_rsv; u64 num_bytes; int ret; - - if (!trans->bytes_reserved) - return 0; + bool release = false; src_rsv = trans->block_rsv; - dst_rsv = &root->fs_info->global_block_rsv; + dst_rsv = &root->fs_info->delayed_block_rsv; num_bytes = btrfs_calc_trans_metadata_size(root, 1); + + /* + * btrfs_dirty_inode will update the inode under btrfs_join_transaction + * which doesn't reserve space for speed. This is a problem since we + * still need to reserve space for this update, so try to reserve the + * space. + * + * Now if src_rsv == delalloc_block_rsv we'll let it just steal since + * we're accounted for. + */ + if (!src_rsv || (!trans->bytes_reserved && + src_rsv != &root->fs_info->delalloc_block_rsv)) { + ret = btrfs_block_rsv_add_noflush(root, dst_rsv, num_bytes); + /* + * Since we're under a transaction reserve_metadata_bytes could + * try to commit the transaction which will make it return + * EAGAIN to make us stop the transaction we have, so return + * ENOSPC instead so that btrfs_dirty_inode knows what to do. + */ + if (ret == -EAGAIN) + ret = -ENOSPC; + if (!ret) { + node->bytes_reserved = num_bytes; + trace_btrfs_space_reservation(root->fs_info, + "delayed_inode", + btrfs_ino(inode), + num_bytes, 1); + } + return ret; + } else if (src_rsv == &root->fs_info->delalloc_block_rsv) { + spin_lock(&BTRFS_I(inode)->lock); + if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) { + spin_unlock(&BTRFS_I(inode)->lock); + release = true; + goto migrate; + } + spin_unlock(&BTRFS_I(inode)->lock); + + /* Ok we didn't have space pre-reserved. This shouldn't happen + * too often but it can happen if we do delalloc to an existing + * inode which gets dirtied because of the time update, and then + * isn't touched again until after the transaction commits and + * then we try to write out the data. First try to be nice and + * reserve something strictly for us. If not be a pain and try + * to steal from the delalloc block rsv. + */ + ret = btrfs_block_rsv_add_noflush(root, dst_rsv, num_bytes); + if (!ret) + goto out; + + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + if (!ret) + goto out; + + /* + * Ok this is a problem, let's just steal from the global rsv + * since this really shouldn't happen that often. + */ + WARN_ON(1); + ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv, + dst_rsv, num_bytes); + goto out; + } + +migrate: ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); - if (!ret) + +out: + /* + * Migrate only takes a reservation, it doesn't touch the size of the + * block_rsv. This is to simplify people who don't normally have things + * migrated from their block rsv. If they go to release their + * reservation, that will decrease the size as well, so if migrate + * reduced size we'd end up with a negative size. But for the + * delalloc_meta_reserved stuff we will only know to drop 1 reservation, + * but we could in fact do this reserve/migrate dance several times + * between the time we did the original reservation and we'd clean it + * up. So to take care of this, release the space for the meta + * reservation here. I think it may be time for a documentation page on + * how block rsvs. work. + */ + if (!ret) { + trace_btrfs_space_reservation(root->fs_info, "delayed_inode", + btrfs_ino(inode), num_bytes, 1); node->bytes_reserved = num_bytes; + } + + if (release) { + trace_btrfs_space_reservation(root->fs_info, "delalloc", + btrfs_ino(inode), num_bytes, 0); + btrfs_block_rsv_release(root, src_rsv, num_bytes); + } return ret; } @@ -646,7 +744,9 @@ static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root, if (!node->bytes_reserved) return; - rsv = &root->fs_info->global_block_rsv; + rsv = &root->fs_info->delayed_block_rsv; + trace_btrfs_space_reservation(root->fs_info, "delayed_inode", + node->inode_id, node->bytes_reserved, 0); btrfs_block_rsv_release(root, rsv, node->bytes_reserved); node->bytes_reserved = 0; @@ -738,10 +838,8 @@ static int btrfs_batch_insert_items(struct btrfs_trans_handle *trans, btrfs_clear_path_blocking(path, NULL, 0); /* insert the keys of the items */ - ret = setup_items_for_insert(trans, root, path, keys, data_size, - total_data_size, total_size, nitems); - if (ret) - goto error; + setup_items_for_insert(trans, root, path, keys, data_size, + total_data_size, total_size, nitems); /* insert the dir index items */ slot = path->slots[0]; @@ -930,9 +1028,10 @@ do_again: btrfs_release_delayed_item(prev); ret = 0; btrfs_release_path(path); - if (curr) + if (curr) { + mutex_unlock(&node->mutex); goto do_again; - else + } else goto delete_fail; } @@ -957,8 +1056,7 @@ static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node) delayed_node->count--; delayed_root = delayed_node->root->fs_info->delayed_root; - atomic_dec(&delayed_root->items); - if (atomic_read(&delayed_root->items) < + if (atomic_dec_return(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND && waitqueue_active(&delayed_root->wait)) wake_up(&delayed_root->wait); @@ -1010,15 +1108,25 @@ static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans, return 0; } -/* Called when committing the transaction. */ -int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, - struct btrfs_root *root) +/* + * Called when committing the transaction. + * Returns 0 on success. + * Returns < 0 on error and returns with an aborted transaction with any + * outstanding delayed items cleaned up. + */ +static int __btrfs_run_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root, int nr) { + struct btrfs_root *curr_root = root; struct btrfs_delayed_root *delayed_root; struct btrfs_delayed_node *curr_node, *prev_node; struct btrfs_path *path; struct btrfs_block_rsv *block_rsv; int ret = 0; + bool count = (nr > 0); + + if (trans->aborted) + return -EIO; path = btrfs_alloc_path(); if (!path) @@ -1026,23 +1134,25 @@ int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, path->leave_spinning = 1; block_rsv = trans->block_rsv; - trans->block_rsv = &root->fs_info->global_block_rsv; + trans->block_rsv = &root->fs_info->delayed_block_rsv; delayed_root = btrfs_get_delayed_root(root); curr_node = btrfs_first_delayed_node(delayed_root); - while (curr_node) { - root = curr_node->root; - ret = btrfs_insert_delayed_items(trans, path, root, + while (curr_node && (!count || (count && nr--))) { + curr_root = curr_node->root; + ret = btrfs_insert_delayed_items(trans, path, curr_root, curr_node); if (!ret) - ret = btrfs_delete_delayed_items(trans, path, root, - curr_node); + ret = btrfs_delete_delayed_items(trans, path, + curr_root, curr_node); if (!ret) - ret = btrfs_update_delayed_inode(trans, root, path, - curr_node); + ret = btrfs_update_delayed_inode(trans, curr_root, + path, curr_node); if (ret) { btrfs_release_delayed_node(curr_node); + curr_node = NULL; + btrfs_abort_transaction(trans, root, ret); break; } @@ -1051,11 +1161,26 @@ int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, btrfs_release_delayed_node(prev_node); } + if (curr_node) + btrfs_release_delayed_node(curr_node); btrfs_free_path(path); trans->block_rsv = block_rsv; + return ret; } +int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root) +{ + return __btrfs_run_delayed_items(trans, root, -1); +} + +int btrfs_run_delayed_items_nr(struct btrfs_trans_handle *trans, + struct btrfs_root *root, int nr) +{ + return __btrfs_run_delayed_items(trans, root, nr); +} + static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, struct btrfs_delayed_node *node) { @@ -1069,7 +1194,7 @@ static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, path->leave_spinning = 1; block_rsv = trans->block_rsv; - trans->block_rsv = &node->root->fs_info->global_block_rsv; + trans->block_rsv = &node->root->fs_info->delayed_block_rsv; ret = btrfs_insert_delayed_items(trans, path, node->root, node); if (!ret) @@ -1149,7 +1274,7 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work) goto free_path; block_rsv = trans->block_rsv; - trans->block_rsv = &root->fs_info->global_block_rsv; + trans->block_rsv = &root->fs_info->delayed_block_rsv; ret = btrfs_insert_delayed_items(trans, path, root, delayed_node); if (!ret) @@ -1273,6 +1398,7 @@ void btrfs_balance_delayed_items(struct btrfs_root *root) btrfs_wq_run_delayed_node(delayed_root, root, 0); } +/* Will return 0 or -ENOMEM */ int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, int name_len, struct inode *dir, @@ -1294,13 +1420,6 @@ int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, goto release_node; } - ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item); - /* - * we have reserved enough space when we start a new transaction, - * so reserving metadata failure is impossible - */ - BUG_ON(ret); - delayed_item->key.objectid = btrfs_ino(dir); btrfs_set_key_type(&delayed_item->key, BTRFS_DIR_INDEX_KEY); delayed_item->key.offset = index; @@ -1313,6 +1432,14 @@ int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, dir_item->type = type; memcpy((char *)(dir_item + 1), name, name_len); + ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item); + /* + * we have reserved enough space when we start a new transaction, + * so reserving metadata failure is impossible + */ + BUG_ON(ret); + + mutex_lock(&delayed_node->mutex); ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item); if (unlikely(ret)) { @@ -1596,7 +1723,7 @@ static void fill_stack_inode_item(struct btrfs_trans_handle *trans, btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode)); btrfs_set_stack_inode_generation(inode_item, BTRFS_I(inode)->generation); - btrfs_set_stack_inode_sequence(inode_item, BTRFS_I(inode)->sequence); + btrfs_set_stack_inode_sequence(inode_item, inode->i_version); btrfs_set_stack_inode_transid(inode_item, trans->transid); btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev); btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags); @@ -1641,10 +1768,10 @@ int btrfs_fill_inode(struct inode *inode, u32 *rdev) inode->i_gid = btrfs_stack_inode_gid(inode_item); btrfs_i_size_write(inode, btrfs_stack_inode_size(inode_item)); inode->i_mode = btrfs_stack_inode_mode(inode_item); - inode->i_nlink = btrfs_stack_inode_nlink(inode_item); + set_nlink(inode, btrfs_stack_inode_nlink(inode_item)); inode_set_bytes(inode, btrfs_stack_inode_nbytes(inode_item)); BTRFS_I(inode)->generation = btrfs_stack_inode_generation(inode_item); - BTRFS_I(inode)->sequence = btrfs_stack_inode_sequence(inode_item); + inode->i_version = btrfs_stack_inode_sequence(inode_item); inode->i_rdev = 0; *rdev = btrfs_stack_inode_rdev(inode_item); BTRFS_I(inode)->flags = btrfs_stack_inode_flags(inode_item); @@ -1685,12 +1812,10 @@ int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, goto release_node; } - ret = btrfs_delayed_inode_reserve_metadata(trans, root, delayed_node); - /* - * we must reserve enough space when we start a new transaction, - * so reserving metadata failure is impossible - */ - BUG_ON(ret); + ret = btrfs_delayed_inode_reserve_metadata(trans, root, inode, + delayed_node); + if (ret) + goto release_node; fill_stack_inode_item(trans, &delayed_node->inode_item, inode); delayed_node->inode_dirty = 1; @@ -1771,3 +1896,21 @@ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) } } } + +void btrfs_destroy_delayed_inodes(struct btrfs_root *root) +{ + struct btrfs_delayed_root *delayed_root; + struct btrfs_delayed_node *curr_node, *prev_node; + + delayed_root = btrfs_get_delayed_root(root); + + curr_node = btrfs_first_delayed_node(delayed_root); + while (curr_node) { + __btrfs_kill_delayed_node(curr_node); + + prev_node = curr_node; + curr_node = btrfs_next_delayed_node(curr_node); + btrfs_release_delayed_node(prev_node); + } +} + diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h index 7083d08b2a21..4f808e1baeed 100644 --- a/fs/btrfs/delayed-inode.h +++ b/fs/btrfs/delayed-inode.h @@ -107,6 +107,8 @@ int btrfs_inode_delayed_dir_index_count(struct inode *inode); int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, struct btrfs_root *root); +int btrfs_run_delayed_items_nr(struct btrfs_trans_handle *trans, + struct btrfs_root *root, int nr); void btrfs_balance_delayed_items(struct btrfs_root *root); @@ -124,6 +126,9 @@ int btrfs_fill_inode(struct inode *inode, u32 *rdev); /* Used for drop dead root */ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root); +/* Used for clean the transaction */ +void btrfs_destroy_delayed_inodes(struct btrfs_root *root); + /* Used for readdir() */ void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, struct list_head *del_list); diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index 125cf76fcd08..ae9411773397 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -38,17 +38,14 @@ static int comp_tree_refs(struct btrfs_delayed_tree_ref *ref2, struct btrfs_delayed_tree_ref *ref1) { - if (ref1->node.type == BTRFS_TREE_BLOCK_REF_KEY) { - if (ref1->root < ref2->root) - return -1; - if (ref1->root > ref2->root) - return 1; - } else { - if (ref1->parent < ref2->parent) - return -1; - if (ref1->parent > ref2->parent) - return 1; - } + if (ref1->root < ref2->root) + return -1; + if (ref1->root > ref2->root) + return 1; + if (ref1->parent < ref2->parent) + return -1; + if (ref1->parent > ref2->parent) + return 1; return 0; } @@ -85,7 +82,8 @@ static int comp_data_refs(struct btrfs_delayed_data_ref *ref2, * type of the delayed backrefs and content of delayed backrefs. */ static int comp_entry(struct btrfs_delayed_ref_node *ref2, - struct btrfs_delayed_ref_node *ref1) + struct btrfs_delayed_ref_node *ref1, + bool compare_seq) { if (ref1->bytenr < ref2->bytenr) return -1; @@ -101,6 +99,13 @@ static int comp_entry(struct btrfs_delayed_ref_node *ref2, return -1; if (ref1->type > ref2->type) return 1; + /* merging of sequenced refs is not allowed */ + if (compare_seq) { + if (ref1->seq < ref2->seq) + return -1; + if (ref1->seq > ref2->seq) + return 1; + } if (ref1->type == BTRFS_TREE_BLOCK_REF_KEY || ref1->type == BTRFS_SHARED_BLOCK_REF_KEY) { return comp_tree_refs(btrfs_delayed_node_to_tree_ref(ref2), @@ -134,7 +139,7 @@ static struct btrfs_delayed_ref_node *tree_insert(struct rb_root *root, entry = rb_entry(parent_node, struct btrfs_delayed_ref_node, rb_node); - cmp = comp_entry(entry, ins); + cmp = comp_entry(entry, ins, 1); if (cmp < 0) p = &(*p)->rb_left; else if (cmp > 0) @@ -150,16 +155,22 @@ static struct btrfs_delayed_ref_node *tree_insert(struct rb_root *root, /* * find an head entry based on bytenr. This returns the delayed ref - * head if it was able to find one, or NULL if nothing was in that spot + * head if it was able to find one, or NULL if nothing was in that spot. + * If return_bigger is given, the next bigger entry is returned if no exact + * match is found. */ static struct btrfs_delayed_ref_node *find_ref_head(struct rb_root *root, u64 bytenr, - struct btrfs_delayed_ref_node **last) + struct btrfs_delayed_ref_node **last, + int return_bigger) { - struct rb_node *n = root->rb_node; + struct rb_node *n; struct btrfs_delayed_ref_node *entry; - int cmp; + int cmp = 0; +again: + n = root->rb_node; + entry = NULL; while (n) { entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); WARN_ON(!entry->in_tree); @@ -182,6 +193,19 @@ static struct btrfs_delayed_ref_node *find_ref_head(struct rb_root *root, else return entry; } + if (entry && return_bigger) { + if (cmp > 0) { + n = rb_next(&entry->rb_node); + if (!n) + n = rb_first(root); + entry = rb_entry(n, struct btrfs_delayed_ref_node, + rb_node); + bytenr = entry->bytenr; + return_bigger = 0; + goto again; + } + return entry; + } return NULL; } @@ -209,6 +233,136 @@ int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans, return 0; } +static void inline drop_delayed_ref(struct btrfs_trans_handle *trans, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_delayed_ref_node *ref) +{ + rb_erase(&ref->rb_node, &delayed_refs->root); + ref->in_tree = 0; + btrfs_put_delayed_ref(ref); + delayed_refs->num_entries--; + if (trans->delayed_ref_updates) + trans->delayed_ref_updates--; +} + +static int merge_ref(struct btrfs_trans_handle *trans, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_delayed_ref_node *ref, u64 seq) +{ + struct rb_node *node; + int merged = 0; + int mod = 0; + int done = 0; + + node = rb_prev(&ref->rb_node); + while (node) { + struct btrfs_delayed_ref_node *next; + + next = rb_entry(node, struct btrfs_delayed_ref_node, rb_node); + node = rb_prev(node); + if (next->bytenr != ref->bytenr) + break; + if (seq && next->seq >= seq) + break; + if (comp_entry(ref, next, 0)) + continue; + + if (ref->action == next->action) { + mod = next->ref_mod; + } else { + if (ref->ref_mod < next->ref_mod) { + struct btrfs_delayed_ref_node *tmp; + + tmp = ref; + ref = next; + next = tmp; + done = 1; + } + mod = -next->ref_mod; + } + + merged++; + drop_delayed_ref(trans, delayed_refs, next); + ref->ref_mod += mod; + if (ref->ref_mod == 0) { + drop_delayed_ref(trans, delayed_refs, ref); + break; + } else { + /* + * You can't have multiples of the same ref on a tree + * block. + */ + WARN_ON(ref->type == BTRFS_TREE_BLOCK_REF_KEY || + ref->type == BTRFS_SHARED_BLOCK_REF_KEY); + } + + if (done) + break; + node = rb_prev(&ref->rb_node); + } + + return merged; +} + +void btrfs_merge_delayed_refs(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_delayed_ref_head *head) +{ + struct rb_node *node; + u64 seq = 0; + + spin_lock(&fs_info->tree_mod_seq_lock); + if (!list_empty(&fs_info->tree_mod_seq_list)) { + struct seq_list *elem; + + elem = list_first_entry(&fs_info->tree_mod_seq_list, + struct seq_list, list); + seq = elem->seq; + } + spin_unlock(&fs_info->tree_mod_seq_lock); + + node = rb_prev(&head->node.rb_node); + while (node) { + struct btrfs_delayed_ref_node *ref; + + ref = rb_entry(node, struct btrfs_delayed_ref_node, + rb_node); + if (ref->bytenr != head->node.bytenr) + break; + + /* We can't merge refs that are outside of our seq count */ + if (seq && ref->seq >= seq) + break; + if (merge_ref(trans, delayed_refs, ref, seq)) + node = rb_prev(&head->node.rb_node); + else + node = rb_prev(node); + } +} + +int btrfs_check_delayed_seq(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + u64 seq) +{ + struct seq_list *elem; + int ret = 0; + + spin_lock(&fs_info->tree_mod_seq_lock); + if (!list_empty(&fs_info->tree_mod_seq_list)) { + elem = list_first_entry(&fs_info->tree_mod_seq_list, + struct seq_list, list); + if (seq >= elem->seq) { + pr_debug("holding back delayed_ref %llu, lowest is " + "%llu (%p)\n", seq, elem->seq, delayed_refs); + ret = 1; + } + } + + spin_unlock(&fs_info->tree_mod_seq_lock); + return ret; +} + int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans, struct list_head *cluster, u64 start) { @@ -223,20 +377,8 @@ int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans, node = rb_first(&delayed_refs->root); } else { ref = NULL; - find_ref_head(&delayed_refs->root, start, &ref); + find_ref_head(&delayed_refs->root, start + 1, &ref, 1); if (ref) { - struct btrfs_delayed_ref_node *tmp; - - node = rb_prev(&ref->rb_node); - while (node) { - tmp = rb_entry(node, - struct btrfs_delayed_ref_node, - rb_node); - if (tmp->bytenr < start) - break; - ref = tmp; - node = rb_prev(&ref->rb_node); - } node = &ref->rb_node; } else node = rb_first(&delayed_refs->root); @@ -302,18 +444,11 @@ update_existing_ref(struct btrfs_trans_handle *trans, * every changing the extent allocation tree. */ existing->ref_mod--; - if (existing->ref_mod == 0) { - rb_erase(&existing->rb_node, - &delayed_refs->root); - existing->in_tree = 0; - btrfs_put_delayed_ref(existing); - delayed_refs->num_entries--; - if (trans->delayed_ref_updates) - trans->delayed_ref_updates--; - } else { + if (existing->ref_mod == 0) + drop_delayed_ref(trans, delayed_refs, existing); + else WARN_ON(existing->type == BTRFS_TREE_BLOCK_REF_KEY || existing->type == BTRFS_SHARED_BLOCK_REF_KEY); - } } else { WARN_ON(existing->type == BTRFS_TREE_BLOCK_REF_KEY || existing->type == BTRFS_SHARED_BLOCK_REF_KEY); @@ -390,7 +525,8 @@ update_existing_head_ref(struct btrfs_delayed_ref_node *existing, * this does all the dirty work in terms of maintaining the correct * overall modification count. */ -static noinline int add_delayed_ref_head(struct btrfs_trans_handle *trans, +static noinline void add_delayed_ref_head(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_node *ref, u64 bytenr, u64 num_bytes, int action, int is_data) @@ -437,6 +573,7 @@ static noinline int add_delayed_ref_head(struct btrfs_trans_handle *trans, ref->action = 0; ref->is_head = 1; ref->in_tree = 1; + ref->seq = 0; head_ref = btrfs_delayed_node_to_head(ref); head_ref->must_insert_reserved = must_insert_reserved; @@ -455,27 +592,29 @@ static noinline int add_delayed_ref_head(struct btrfs_trans_handle *trans, * we've updated the existing ref, free the newly * allocated ref */ - kfree(ref); + kfree(head_ref); } else { delayed_refs->num_heads++; delayed_refs->num_heads_ready++; delayed_refs->num_entries++; trans->delayed_ref_updates++; } - return 0; } /* * helper to insert a delayed tree ref into the rbtree. */ -static noinline int add_delayed_tree_ref(struct btrfs_trans_handle *trans, +static noinline void add_delayed_tree_ref(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_node *ref, u64 bytenr, u64 num_bytes, u64 parent, - u64 ref_root, int level, int action) + u64 ref_root, int level, int action, + int for_cow) { struct btrfs_delayed_ref_node *existing; struct btrfs_delayed_tree_ref *full_ref; struct btrfs_delayed_ref_root *delayed_refs; + u64 seq = 0; if (action == BTRFS_ADD_DELAYED_EXTENT) action = BTRFS_ADD_DELAYED_REF; @@ -491,14 +630,17 @@ static noinline int add_delayed_tree_ref(struct btrfs_trans_handle *trans, ref->is_head = 0; ref->in_tree = 1; + if (need_ref_seq(for_cow, ref_root)) + seq = btrfs_get_tree_mod_seq(fs_info, &trans->delayed_ref_elem); + ref->seq = seq; + full_ref = btrfs_delayed_node_to_tree_ref(ref); - if (parent) { - full_ref->parent = parent; + full_ref->parent = parent; + full_ref->root = ref_root; + if (parent) ref->type = BTRFS_SHARED_BLOCK_REF_KEY; - } else { - full_ref->root = ref_root; + else ref->type = BTRFS_TREE_BLOCK_REF_KEY; - } full_ref->level = level; trace_btrfs_delayed_tree_ref(ref, full_ref, action); @@ -511,26 +653,27 @@ static noinline int add_delayed_tree_ref(struct btrfs_trans_handle *trans, * we've updated the existing ref, free the newly * allocated ref */ - kfree(ref); + kfree(full_ref); } else { delayed_refs->num_entries++; trans->delayed_ref_updates++; } - return 0; } /* * helper to insert a delayed data ref into the rbtree. */ -static noinline int add_delayed_data_ref(struct btrfs_trans_handle *trans, +static noinline void add_delayed_data_ref(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_node *ref, u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root, u64 owner, u64 offset, - int action) + int action, int for_cow) { struct btrfs_delayed_ref_node *existing; struct btrfs_delayed_data_ref *full_ref; struct btrfs_delayed_ref_root *delayed_refs; + u64 seq = 0; if (action == BTRFS_ADD_DELAYED_EXTENT) action = BTRFS_ADD_DELAYED_REF; @@ -546,14 +689,18 @@ static noinline int add_delayed_data_ref(struct btrfs_trans_handle *trans, ref->is_head = 0; ref->in_tree = 1; + if (need_ref_seq(for_cow, ref_root)) + seq = btrfs_get_tree_mod_seq(fs_info, &trans->delayed_ref_elem); + ref->seq = seq; + full_ref = btrfs_delayed_node_to_data_ref(ref); - if (parent) { - full_ref->parent = parent; + full_ref->parent = parent; + full_ref->root = ref_root; + if (parent) ref->type = BTRFS_SHARED_DATA_REF_KEY; - } else { - full_ref->root = ref_root; + else ref->type = BTRFS_EXTENT_DATA_REF_KEY; - } + full_ref->objectid = owner; full_ref->offset = offset; @@ -567,12 +714,11 @@ static noinline int add_delayed_data_ref(struct btrfs_trans_handle *trans, * we've updated the existing ref, free the newly * allocated ref */ - kfree(ref); + kfree(full_ref); } else { delayed_refs->num_entries++; trans->delayed_ref_updates++; } - return 0; } /* @@ -580,15 +726,16 @@ static noinline int add_delayed_data_ref(struct btrfs_trans_handle *trans, * to make sure the delayed ref is eventually processed before this * transaction commits. */ -int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans, +int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root, int level, int action, - struct btrfs_delayed_extent_op *extent_op) + struct btrfs_delayed_extent_op *extent_op, + int for_cow) { struct btrfs_delayed_tree_ref *ref; struct btrfs_delayed_ref_head *head_ref; struct btrfs_delayed_ref_root *delayed_refs; - int ret; BUG_ON(extent_op && extent_op->is_data); ref = kmalloc(sizeof(*ref), GFP_NOFS); @@ -610,30 +757,33 @@ int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans, * insert both the head node and the new ref without dropping * the spin lock */ - ret = add_delayed_ref_head(trans, &head_ref->node, bytenr, num_bytes, - action, 0); - BUG_ON(ret); + add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr, + num_bytes, action, 0); - ret = add_delayed_tree_ref(trans, &ref->node, bytenr, num_bytes, - parent, ref_root, level, action); - BUG_ON(ret); + add_delayed_tree_ref(fs_info, trans, &ref->node, bytenr, + num_bytes, parent, ref_root, level, action, + for_cow); spin_unlock(&delayed_refs->lock); + if (need_ref_seq(for_cow, ref_root)) + btrfs_qgroup_record_ref(trans, &ref->node, extent_op); + return 0; } /* * add a delayed data ref. it's similar to btrfs_add_delayed_tree_ref. */ -int btrfs_add_delayed_data_ref(struct btrfs_trans_handle *trans, +int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root, u64 owner, u64 offset, int action, - struct btrfs_delayed_extent_op *extent_op) + struct btrfs_delayed_extent_op *extent_op, + int for_cow) { struct btrfs_delayed_data_ref *ref; struct btrfs_delayed_ref_head *head_ref; struct btrfs_delayed_ref_root *delayed_refs; - int ret; BUG_ON(extent_op && !extent_op->is_data); ref = kmalloc(sizeof(*ref), GFP_NOFS); @@ -655,24 +805,26 @@ int btrfs_add_delayed_data_ref(struct btrfs_trans_handle *trans, * insert both the head node and the new ref without dropping * the spin lock */ - ret = add_delayed_ref_head(trans, &head_ref->node, bytenr, num_bytes, - action, 1); - BUG_ON(ret); + add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr, + num_bytes, action, 1); - ret = add_delayed_data_ref(trans, &ref->node, bytenr, num_bytes, - parent, ref_root, owner, offset, action); - BUG_ON(ret); + add_delayed_data_ref(fs_info, trans, &ref->node, bytenr, + num_bytes, parent, ref_root, owner, offset, + action, for_cow); spin_unlock(&delayed_refs->lock); + if (need_ref_seq(for_cow, ref_root)) + btrfs_qgroup_record_ref(trans, &ref->node, extent_op); + return 0; } -int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans, +int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, struct btrfs_delayed_extent_op *extent_op) { struct btrfs_delayed_ref_head *head_ref; struct btrfs_delayed_ref_root *delayed_refs; - int ret; head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS); if (!head_ref) @@ -683,10 +835,9 @@ int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans, delayed_refs = &trans->transaction->delayed_refs; spin_lock(&delayed_refs->lock); - ret = add_delayed_ref_head(trans, &head_ref->node, bytenr, + add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr, num_bytes, BTRFS_UPDATE_DELAYED_HEAD, extent_op->is_data); - BUG_ON(ret); spin_unlock(&delayed_refs->lock); return 0; @@ -704,7 +855,7 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr) struct btrfs_delayed_ref_root *delayed_refs; delayed_refs = &trans->transaction->delayed_refs; - ref = find_ref_head(&delayed_refs->root, bytenr, NULL); + ref = find_ref_head(&delayed_refs->root, bytenr, NULL, 0); if (ref) return btrfs_delayed_node_to_head(ref); return NULL; diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h index e287e3b0eab0..ab5300595847 100644 --- a/fs/btrfs/delayed-ref.h +++ b/fs/btrfs/delayed-ref.h @@ -33,6 +33,9 @@ struct btrfs_delayed_ref_node { /* the size of the extent */ u64 num_bytes; + /* seq number to keep track of insertion order */ + u64 seq; + /* ref count on this data structure */ atomic_t refs; @@ -98,19 +101,15 @@ struct btrfs_delayed_ref_head { struct btrfs_delayed_tree_ref { struct btrfs_delayed_ref_node node; - union { - u64 root; - u64 parent; - }; + u64 root; + u64 parent; int level; }; struct btrfs_delayed_data_ref { struct btrfs_delayed_ref_node node; - union { - u64 root; - u64 parent; - }; + u64 root; + u64 parent; u64 objectid; u64 offset; }; @@ -151,18 +150,27 @@ static inline void btrfs_put_delayed_ref(struct btrfs_delayed_ref_node *ref) } } -int btrfs_add_delayed_tree_ref(struct btrfs_trans_handle *trans, +int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root, int level, int action, - struct btrfs_delayed_extent_op *extent_op); -int btrfs_add_delayed_data_ref(struct btrfs_trans_handle *trans, + struct btrfs_delayed_extent_op *extent_op, + int for_cow); +int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, u64 parent, u64 ref_root, u64 owner, u64 offset, int action, - struct btrfs_delayed_extent_op *extent_op); -int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans, + struct btrfs_delayed_extent_op *extent_op, + int for_cow); +int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info, + struct btrfs_trans_handle *trans, u64 bytenr, u64 num_bytes, struct btrfs_delayed_extent_op *extent_op); +void btrfs_merge_delayed_refs(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_delayed_ref_head *head); struct btrfs_delayed_ref_head * btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr); @@ -170,6 +178,30 @@ int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_head *head); int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans, struct list_head *cluster, u64 search_start); + +int btrfs_check_delayed_seq(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + u64 seq); + +/* + * delayed refs with a ref_seq > 0 must be held back during backref walking. + * this only applies to items in one of the fs-trees. for_cow items never need + * to be held back, so they won't get a ref_seq number. + */ +static inline int need_ref_seq(int for_cow, u64 rootid) +{ + if (for_cow) + return 0; + + if (rootid == BTRFS_FS_TREE_OBJECTID) + return 1; + + if ((s64)rootid >= (s64)BTRFS_FIRST_FREE_OBJECTID) + return 1; + + return 0; +} + /* * a node might live in a head or a regular ref, this lets you * test for the proper type to use. diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c index 31d84e78129b..c1a074d0696f 100644 --- a/fs/btrfs/dir-item.c +++ b/fs/btrfs/dir-item.c @@ -49,9 +49,8 @@ static struct btrfs_dir_item *insert_with_overflow(struct btrfs_trans_handle di = btrfs_match_dir_item_name(root, path, name, name_len); if (di) return ERR_PTR(-EEXIST); - ret = btrfs_extend_item(trans, root, path, data_size); - } - if (ret < 0) + btrfs_extend_item(trans, root, path, data_size); + } else if (ret < 0) return ERR_PTR(ret); WARN_ON(ret > 0); leaf = path->nodes[0]; @@ -116,6 +115,7 @@ int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans, * 'location' is the key to stuff into the directory item, 'type' is the * type of the inode we're pointing to, and 'index' is the sequence number * to use for the second index (if one is created). + * Will return 0 or -ENOMEM */ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, int name_len, @@ -383,8 +383,8 @@ int btrfs_delete_one_dir_name(struct btrfs_trans_handle *trans, start = btrfs_item_ptr_offset(leaf, path->slots[0]); memmove_extent_buffer(leaf, ptr, ptr + sub_item_len, item_len - (ptr + sub_item_len - start)); - ret = btrfs_truncate_item(trans, root, path, - item_len - sub_item_len, 1); + btrfs_truncate_item(trans, root, path, + item_len - sub_item_len, 1); } return ret; } diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 07b3ac662e19..22e98e04c2ea 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -43,24 +43,25 @@ #include "tree-log.h" #include "free-space-cache.h" #include "inode-map.h" +#include "check-integrity.h" +#include "rcu-string.h" static struct extent_io_ops btree_extent_io_ops; static void end_workqueue_fn(struct btrfs_work *work); static void free_fs_root(struct btrfs_root *root); -static void btrfs_check_super_valid(struct btrfs_fs_info *fs_info, +static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, int read_only); -static int btrfs_destroy_ordered_operations(struct btrfs_root *root); -static int btrfs_destroy_ordered_extents(struct btrfs_root *root); +static void btrfs_destroy_ordered_operations(struct btrfs_root *root); +static void btrfs_destroy_ordered_extents(struct btrfs_root *root); static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, struct btrfs_root *root); -static int btrfs_destroy_pending_snapshots(struct btrfs_transaction *t); -static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root); +static void btrfs_destroy_pending_snapshots(struct btrfs_transaction *t); +static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root); static int btrfs_destroy_marked_extents(struct btrfs_root *root, struct extent_io_tree *dirty_pages, int mark); static int btrfs_destroy_pinned_extent(struct btrfs_root *root, struct extent_io_tree *pinned_extents); -static int btrfs_cleanup_transaction(struct btrfs_root *root); /* * end_io_wq structs are used to do processing in task context when an IO is @@ -98,6 +99,7 @@ struct async_submit_bio { */ u64 bio_offset; struct btrfs_work work; + int error; }; /* @@ -256,8 +258,7 @@ void btrfs_csum_final(u32 crc, char *result) static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf, int verify) { - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); char *result = NULL; unsigned long len; unsigned long cur_len; @@ -323,7 +324,8 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf, * in the wrong place. */ static int verify_parent_transid(struct extent_io_tree *io_tree, - struct extent_buffer *eb, u64 parent_transid) + struct extent_buffer *eb, u64 parent_transid, + int atomic) { struct extent_state *cached_state = NULL; int ret; @@ -331,9 +333,12 @@ static int verify_parent_transid(struct extent_io_tree *io_tree, if (!parent_transid || btrfs_header_generation(eb) == parent_transid) return 0; + if (atomic) + return -EAGAIN; + lock_extent_bits(io_tree, eb->start, eb->start + eb->len - 1, - 0, &cached_state, GFP_NOFS); - if (extent_buffer_uptodate(io_tree, eb, cached_state) && + 0, &cached_state); + if (extent_buffer_uptodate(eb) && btrfs_header_generation(eb) == parent_transid) { ret = 0; goto out; @@ -344,7 +349,7 @@ static int verify_parent_transid(struct extent_io_tree *io_tree, (unsigned long long)parent_transid, (unsigned long long)btrfs_header_generation(eb)); ret = 1; - clear_extent_buffer_uptodate(io_tree, eb, &cached_state); + clear_extent_buffer_uptodate(eb); out: unlock_extent_cached(io_tree, eb->start, eb->start + eb->len - 1, &cached_state, GFP_NOFS); @@ -360,18 +365,25 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, u64 start, u64 parent_transid) { struct extent_io_tree *io_tree; + int failed = 0; int ret; int num_copies = 0; int mirror_num = 0; + int failed_mirror = 0; clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags); io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree; while (1) { - ret = read_extent_buffer_pages(io_tree, eb, start, 1, + ret = read_extent_buffer_pages(io_tree, eb, start, + WAIT_COMPLETE, btree_get_extent, mirror_num); - if (!ret && - !verify_parent_transid(io_tree, eb, parent_transid)) - return ret; + if (!ret) { + if (!verify_parent_transid(io_tree, eb, + parent_transid, 0)) + break; + else + ret = -EIO; + } /* * This buffer's crc is fine, but its contents are corrupted, so @@ -379,18 +391,30 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, * any less wrong. */ if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags)) - return ret; + break; num_copies = btrfs_num_copies(&root->fs_info->mapping_tree, eb->start, eb->len); if (num_copies == 1) - return ret; + break; + + if (!failed_mirror) { + failed = 1; + failed_mirror = eb->read_mirror; + } mirror_num++; + if (mirror_num == failed_mirror) + mirror_num++; + if (mirror_num > num_copies) - return ret; + break; } - return -EIO; + + if (failed && !ret && failed_mirror) + repair_eb_io_failure(root, eb, failed_mirror); + + return ret; } /* @@ -403,50 +427,27 @@ static int csum_dirty_buffer(struct btrfs_root *root, struct page *page) struct extent_io_tree *tree; u64 start = (u64)page->index << PAGE_CACHE_SHIFT; u64 found_start; - unsigned long len; struct extent_buffer *eb; - int ret; tree = &BTRFS_I(page->mapping->host)->io_tree; - if (page->private == EXTENT_PAGE_PRIVATE) { - WARN_ON(1); - goto out; - } - if (!page->private) { - WARN_ON(1); - goto out; - } - len = page->private >> 2; - WARN_ON(len == 0); - - eb = alloc_extent_buffer(tree, start, len, page); - if (eb == NULL) { - WARN_ON(1); - goto out; - } - ret = btree_read_extent_buffer_pages(root, eb, start + PAGE_CACHE_SIZE, - btrfs_header_generation(eb)); - BUG_ON(ret); - WARN_ON(!btrfs_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN)); - + eb = (struct extent_buffer *)page->private; + if (page != eb->pages[0]) + return 0; found_start = btrfs_header_bytenr(eb); if (found_start != start) { WARN_ON(1); - goto err; + return 0; } - if (eb->first_page != page) { + if (eb->pages[0] != page) { WARN_ON(1); - goto err; + return 0; } if (!PageUptodate(page)) { WARN_ON(1); - goto err; + return 0; } csum_tree_block(root, eb, 0); -err: - free_extent_buffer(eb); -out: return 0; } @@ -536,34 +537,75 @@ static noinline int check_leaf(struct btrfs_root *root, return 0; } +struct extent_buffer *find_eb_for_page(struct extent_io_tree *tree, + struct page *page, int max_walk) +{ + struct extent_buffer *eb; + u64 start = page_offset(page); + u64 target = start; + u64 min_start; + + if (start < max_walk) + min_start = 0; + else + min_start = start - max_walk; + + while (start >= min_start) { + eb = find_extent_buffer(tree, start, 0); + if (eb) { + /* + * we found an extent buffer and it contains our page + * horray! + */ + if (eb->start <= target && + eb->start + eb->len > target) + return eb; + + /* we found an extent buffer that wasn't for us */ + free_extent_buffer(eb); + return NULL; + } + if (start == 0) + break; + start -= PAGE_CACHE_SIZE; + } + return NULL; +} + static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, - struct extent_state *state) + struct extent_state *state, int mirror) { struct extent_io_tree *tree; u64 found_start; int found_level; - unsigned long len; struct extent_buffer *eb; struct btrfs_root *root = BTRFS_I(page->mapping->host)->root; int ret = 0; + int reads_done; - tree = &BTRFS_I(page->mapping->host)->io_tree; - if (page->private == EXTENT_PAGE_PRIVATE) - goto out; if (!page->private) goto out; - len = page->private >> 2; - WARN_ON(len == 0); + tree = &BTRFS_I(page->mapping->host)->io_tree; + eb = (struct extent_buffer *)page->private; + + /* the pending IO might have been the only thing that kept this buffer + * in memory. Make sure we have a ref for all this other checks + */ + extent_buffer_get(eb); + + reads_done = atomic_dec_and_test(&eb->io_pages); + if (!reads_done) + goto err; - eb = alloc_extent_buffer(tree, start, len, page); - if (eb == NULL) { + eb->read_mirror = mirror; + if (test_bit(EXTENT_BUFFER_IOERR, &eb->bflags)) { ret = -EIO; - goto out; + goto err; } found_start = btrfs_header_bytenr(eb); - if (found_start != start) { + if (found_start != eb->start) { printk_ratelimited(KERN_INFO "btrfs bad tree block start " "%llu %llu\n", (unsigned long long)found_start, @@ -571,13 +613,6 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, ret = -EIO; goto err; } - if (eb->first_page != page) { - printk(KERN_INFO "btrfs bad first page %lu %lu\n", - eb->first_page->index, page->index); - WARN_ON(1); - ret = -EIO; - goto err; - } if (check_tree_block_fsid(root, eb)) { printk_ratelimited(KERN_INFO "btrfs bad fsid on block %llu\n", (unsigned long long)eb->start); @@ -605,14 +640,34 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, ret = -EIO; } - end = min_t(u64, eb->len, PAGE_CACHE_SIZE); - end = eb->start + end - 1; + if (!ret) + set_extent_buffer_uptodate(eb); err: + if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) { + clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags); + btree_readahead_hook(root, eb, eb->start, ret); + } + + if (ret) + clear_extent_buffer_uptodate(eb); free_extent_buffer(eb); out: return ret; } +static int btree_io_failed_hook(struct page *page, int failed_mirror) +{ + struct extent_buffer *eb; + struct btrfs_root *root = BTRFS_I(page->mapping->host)->root; + + eb = (struct extent_buffer *)page->private; + set_bit(EXTENT_BUFFER_IOERR, &eb->bflags); + eb->read_mirror = failed_mirror; + if (test_and_clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) + btree_readahead_hook(root, eb, eb->start, -EIO); + return -EIO; /* we fixed nothing */ +} + static void end_workqueue_bio(struct bio *bio, int err) { struct end_io_wq *end_io_wq = bio->bi_private; @@ -681,11 +736,14 @@ unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info) static void run_one_async_start(struct btrfs_work *work) { struct async_submit_bio *async; + int ret; async = container_of(work, struct async_submit_bio, work); - async->submit_bio_start(async->inode, async->rw, async->bio, - async->mirror_num, async->bio_flags, - async->bio_offset); + ret = async->submit_bio_start(async->inode, async->rw, async->bio, + async->mirror_num, async->bio_flags, + async->bio_offset); + if (ret) + async->error = ret; } static void run_one_async_done(struct btrfs_work *work) @@ -700,12 +758,16 @@ static void run_one_async_done(struct btrfs_work *work) limit = btrfs_async_submit_limit(fs_info); limit = limit * 2 / 3; - atomic_dec(&fs_info->nr_async_submits); - - if (atomic_read(&fs_info->nr_async_submits) < limit && + if (atomic_dec_return(&fs_info->nr_async_submits) < limit && waitqueue_active(&fs_info->async_submit_wait)) wake_up(&fs_info->async_submit_wait); + /* If an error occured we just want to clean up the bio and move on */ + if (async->error) { + bio_endio(async->bio, async->error); + return; + } + async->submit_bio_done(async->inode, async->rw, async->bio, async->mirror_num, async->bio_flags, async->bio_offset); @@ -747,6 +809,8 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode, async->bio_flags = bio_flags; async->bio_offset = bio_offset; + async->error = 0; + atomic_inc(&fs_info->nr_async_submits); if (rw & REQ_SYNC) @@ -768,15 +832,18 @@ static int btree_csum_one_bio(struct bio *bio) struct bio_vec *bvec = bio->bi_io_vec; int bio_index = 0; struct btrfs_root *root; + int ret = 0; WARN_ON(bio->bi_vcnt <= 0); while (bio_index < bio->bi_vcnt) { root = BTRFS_I(bvec->bv_page->mapping->host)->root; - csum_dirty_buffer(root, bvec->bv_page); + ret = csum_dirty_buffer(root, bvec->bv_page); + if (ret) + break; bio_index++; bvec++; } - return 0; + return ret; } static int __btree_submit_bio_start(struct inode *inode, int rw, @@ -788,8 +855,7 @@ static int __btree_submit_bio_start(struct inode *inode, int rw, * when we're called for a write, we're already in the async * submission context. Just jump into btrfs_map_bio */ - btree_csum_one_bio(bio); - return 0; + return btree_csum_one_bio(bio); } static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio, @@ -809,15 +875,16 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, { int ret; - ret = btrfs_bio_wq_end_io(BTRFS_I(inode)->root->fs_info, - bio, 1); - BUG_ON(ret); - if (!(rw & REQ_WRITE)) { + /* * called for a read, do the setup so that checksum validation * can happen in the async kernel threads */ + ret = btrfs_bio_wq_end_io(BTRFS_I(inode)->root->fs_info, + bio, 1); + if (ret) + return ret; return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, mirror_num, 0); } @@ -835,7 +902,8 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, #ifdef CONFIG_MIGRATION static int btree_migratepage(struct address_space *mapping, - struct page *newpage, struct page *page) + struct page *newpage, struct page *page, + enum migrate_mode mode) { /* * we can't safely write a btree page from here, @@ -850,38 +918,10 @@ static int btree_migratepage(struct address_space *mapping, if (page_has_private(page) && !try_to_release_page(page, GFP_KERNEL)) return -EAGAIN; - return migrate_page(mapping, newpage, page); + return migrate_page(mapping, newpage, page, mode); } #endif -static int btree_writepage(struct page *page, struct writeback_control *wbc) -{ - struct extent_io_tree *tree; - struct btrfs_root *root = BTRFS_I(page->mapping->host)->root; - struct extent_buffer *eb; - int was_dirty; - - tree = &BTRFS_I(page->mapping->host)->io_tree; - if (!(current->flags & PF_MEMALLOC)) { - return extent_write_full_page(tree, page, - btree_get_extent, wbc); - } - - redirty_page_for_writepage(wbc, page); - eb = btrfs_find_tree_block(root, page_offset(page), PAGE_CACHE_SIZE); - WARN_ON(!eb); - - was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags); - if (!was_dirty) { - spin_lock(&root->fs_info->delalloc_lock); - root->fs_info->dirty_metadata_bytes += PAGE_CACHE_SIZE; - spin_unlock(&root->fs_info->delalloc_lock); - } - free_extent_buffer(eb); - - unlock_page(page); - return 0; -} static int btree_writepages(struct address_space *mapping, struct writeback_control *wbc) @@ -901,40 +941,28 @@ static int btree_writepages(struct address_space *mapping, if (num_dirty < thresh) return 0; } - return extent_writepages(tree, mapping, btree_get_extent, wbc); + return btree_write_cache_pages(mapping, wbc); } static int btree_readpage(struct file *file, struct page *page) { struct extent_io_tree *tree; tree = &BTRFS_I(page->mapping->host)->io_tree; - return extent_read_full_page(tree, page, btree_get_extent); + return extent_read_full_page(tree, page, btree_get_extent, 0); } static int btree_releasepage(struct page *page, gfp_t gfp_flags) { - struct extent_io_tree *tree; - struct extent_map_tree *map; - int ret; - if (PageWriteback(page) || PageDirty(page)) return 0; + /* + * We need to mask out eg. __GFP_HIGHMEM and __GFP_DMA32 as we're doing + * slab allocation from alloc_extent_state down the callchain where + * it'd hit a BUG_ON as those flags are not allowed. + */ + gfp_flags &= ~GFP_SLAB_BUG_MASK; - tree = &BTRFS_I(page->mapping->host)->io_tree; - map = &BTRFS_I(page->mapping->host)->extent_tree; - - ret = try_release_extent_state(map, tree, page, gfp_flags); - if (!ret) - return 0; - - ret = try_release_extent_buffer(tree, page); - if (ret == 1) { - ClearPagePrivate(page); - set_page_private(page, 0); - page_cache_release(page); - } - - return ret; + return try_release_extent_buffer(page, gfp_flags); } static void btree_invalidatepage(struct page *page, unsigned long offset) @@ -952,15 +980,28 @@ static void btree_invalidatepage(struct page *page, unsigned long offset) } } +static int btree_set_page_dirty(struct page *page) +{ + struct extent_buffer *eb; + + BUG_ON(!PagePrivate(page)); + eb = (struct extent_buffer *)page->private; + BUG_ON(!eb); + BUG_ON(!test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); + BUG_ON(!atomic_read(&eb->refs)); + btrfs_assert_tree_locked(eb); + return __set_page_dirty_nobuffers(page); +} + static const struct address_space_operations btree_aops = { .readpage = btree_readpage, - .writepage = btree_writepage, .writepages = btree_writepages, .releasepage = btree_releasepage, .invalidatepage = btree_invalidatepage, #ifdef CONFIG_MIGRATION .migratepage = btree_migratepage, #endif + .set_page_dirty = btree_set_page_dirty, }; int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, @@ -974,11 +1015,43 @@ int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, if (!buf) return 0; read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree, - buf, 0, 0, btree_get_extent, 0); + buf, 0, WAIT_NONE, btree_get_extent, 0); free_extent_buffer(buf); return ret; } +int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize, + int mirror_num, struct extent_buffer **eb) +{ + struct extent_buffer *buf = NULL; + struct inode *btree_inode = root->fs_info->btree_inode; + struct extent_io_tree *io_tree = &BTRFS_I(btree_inode)->io_tree; + int ret; + + buf = btrfs_find_create_tree_block(root, bytenr, blocksize); + if (!buf) + return 0; + + set_bit(EXTENT_BUFFER_READAHEAD, &buf->bflags); + + ret = read_extent_buffer_pages(io_tree, buf, 0, WAIT_PAGE_LOCK, + btree_get_extent, mirror_num); + if (ret) { + free_extent_buffer(buf); + return ret; + } + + if (test_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags)) { + free_extent_buffer(buf); + return -EIO; + } else if (extent_buffer_uptodate(buf)) { + *eb = buf; + } else { + free_extent_buffer(buf); + } + return 0; +} + struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize) { @@ -996,20 +1069,20 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, struct extent_buffer *eb; eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree, - bytenr, blocksize, NULL); + bytenr, blocksize); return eb; } int btrfs_write_tree_block(struct extent_buffer *buf) { - return filemap_fdatawrite_range(buf->first_page->mapping, buf->start, + return filemap_fdatawrite_range(buf->pages[0]->mapping, buf->start, buf->start + buf->len - 1); } int btrfs_wait_tree_block_writeback(struct extent_buffer *buf) { - return filemap_fdatawait_range(buf->first_page->mapping, + return filemap_fdatawait_range(buf->pages[0]->mapping, buf->start, buf->start + buf->len - 1); } @@ -1024,17 +1097,13 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr, return NULL; ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid); - - if (ret == 0) - set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags); return buf; } -int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct extent_buffer *buf) +void clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, + struct extent_buffer *buf) { - struct inode *btree_inode = root->fs_info->btree_inode; if (btrfs_header_generation(buf) == root->fs_info->running_transaction->transid) { btrfs_assert_tree_locked(buf); @@ -1043,23 +1112,27 @@ int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, spin_lock(&root->fs_info->delalloc_lock); if (root->fs_info->dirty_metadata_bytes >= buf->len) root->fs_info->dirty_metadata_bytes -= buf->len; - else - WARN_ON(1); + else { + spin_unlock(&root->fs_info->delalloc_lock); + btrfs_panic(root->fs_info, -EOVERFLOW, + "Can't clear %lu bytes from " + " dirty_mdatadata_bytes (%llu)", + buf->len, + root->fs_info->dirty_metadata_bytes); + } spin_unlock(&root->fs_info->delalloc_lock); } /* ugh, clear_extent_buffer_dirty needs to lock the page */ btrfs_set_lock_blocking(buf); - clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, - buf); + clear_extent_buffer_dirty(buf); } - return 0; } -static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, - u32 stripesize, struct btrfs_root *root, - struct btrfs_fs_info *fs_info, - u64 objectid) +static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, + u32 stripesize, struct btrfs_root *root, + struct btrfs_fs_info *fs_info, + u64 objectid) { root->node = NULL; root->commit_root = NULL; @@ -1073,7 +1146,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->orphan_item_inserted = 0; root->orphan_cleanup_state = 0; - root->fs_info = fs_info; root->objectid = objectid; root->last_trans = 0; root->highest_objectid = 0; @@ -1084,7 +1156,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->orphan_block_rsv = NULL; INIT_LIST_HEAD(&root->dirty_list); - INIT_LIST_HEAD(&root->orphan_list); INIT_LIST_HEAD(&root->root_list); spin_lock_init(&root->orphan_lock); spin_lock_init(&root->inode_lock); @@ -1097,6 +1168,7 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, atomic_set(&root->log_commit[0], 0); atomic_set(&root->log_commit[1], 0); atomic_set(&root->log_writers, 0); + atomic_set(&root->orphan_inodes, 0); root->log_batch = 0; root->log_transid = 0; root->last_log_commit = 0; @@ -1112,13 +1184,14 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->defrag_running = 0; root->root_key.objectid = objectid; root->anon_dev = 0; - return 0; + + spin_lock_init(&root->root_times_lock); } -static int find_and_setup_root(struct btrfs_root *tree_root, - struct btrfs_fs_info *fs_info, - u64 objectid, - struct btrfs_root *root) +static int __must_check find_and_setup_root(struct btrfs_root *tree_root, + struct btrfs_fs_info *fs_info, + u64 objectid, + struct btrfs_root *root) { int ret; u32 blocksize; @@ -1131,20 +1204,107 @@ static int find_and_setup_root(struct btrfs_root *tree_root, &root->root_item, &root->root_key); if (ret > 0) return -ENOENT; - BUG_ON(ret); + else if (ret < 0) + return ret; generation = btrfs_root_generation(&root->root_item); blocksize = btrfs_level_size(root, btrfs_root_level(&root->root_item)); + root->commit_root = NULL; root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item), blocksize, generation); - if (!root->node || !btrfs_buffer_uptodate(root->node, generation)) { + if (!root->node || !btrfs_buffer_uptodate(root->node, generation, 0)) { free_extent_buffer(root->node); + root->node = NULL; return -EIO; } root->commit_root = btrfs_root_node(root); return 0; } +static struct btrfs_root *btrfs_alloc_root(struct btrfs_fs_info *fs_info) +{ + struct btrfs_root *root = kzalloc(sizeof(*root), GFP_NOFS); + if (root) + root->fs_info = fs_info; + return root; +} + +struct btrfs_root *btrfs_create_tree(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, + u64 objectid) +{ + struct extent_buffer *leaf; + struct btrfs_root *tree_root = fs_info->tree_root; + struct btrfs_root *root; + struct btrfs_key key; + int ret = 0; + u64 bytenr; + + root = btrfs_alloc_root(fs_info); + if (!root) + return ERR_PTR(-ENOMEM); + + __setup_root(tree_root->nodesize, tree_root->leafsize, + tree_root->sectorsize, tree_root->stripesize, + root, fs_info, objectid); + root->root_key.objectid = objectid; + root->root_key.type = BTRFS_ROOT_ITEM_KEY; + root->root_key.offset = 0; + + leaf = btrfs_alloc_free_block(trans, root, root->leafsize, + 0, objectid, NULL, 0, 0, 0); + if (IS_ERR(leaf)) { + ret = PTR_ERR(leaf); + goto fail; + } + + bytenr = leaf->start; + memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header)); + btrfs_set_header_bytenr(leaf, leaf->start); + btrfs_set_header_generation(leaf, trans->transid); + btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV); + btrfs_set_header_owner(leaf, objectid); + root->node = leaf; + + write_extent_buffer(leaf, fs_info->fsid, + (unsigned long)btrfs_header_fsid(leaf), + BTRFS_FSID_SIZE); + write_extent_buffer(leaf, fs_info->chunk_tree_uuid, + (unsigned long)btrfs_header_chunk_tree_uuid(leaf), + BTRFS_UUID_SIZE); + btrfs_mark_buffer_dirty(leaf); + + root->commit_root = btrfs_root_node(root); + root->track_dirty = 1; + + + root->root_item.flags = 0; + root->root_item.byte_limit = 0; + btrfs_set_root_bytenr(&root->root_item, leaf->start); + btrfs_set_root_generation(&root->root_item, trans->transid); + btrfs_set_root_level(&root->root_item, 0); + btrfs_set_root_refs(&root->root_item, 1); + btrfs_set_root_used(&root->root_item, leaf->len); + btrfs_set_root_last_snapshot(&root->root_item, 0); + btrfs_set_root_dirid(&root->root_item, 0); + root->root_item.drop_level = 0; + + key.objectid = objectid; + key.type = BTRFS_ROOT_ITEM_KEY; + key.offset = 0; + ret = btrfs_insert_root(trans, tree_root, &key, &root->root_item); + if (ret) + goto fail; + + btrfs_tree_unlock(leaf); + +fail: + if (ret) + return ERR_PTR(ret); + + return root; +} + static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info) { @@ -1152,7 +1312,7 @@ static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans, struct btrfs_root *tree_root = fs_info->tree_root; struct extent_buffer *leaf; - root = kzalloc(sizeof(*root), GFP_NOFS); + root = btrfs_alloc_root(fs_info); if (!root) return ERR_PTR(-ENOMEM); @@ -1172,7 +1332,8 @@ static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans, root->ref_cows = 0; leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0, - BTRFS_TREE_LOG_OBJECTID, NULL, 0, 0, 0); + BTRFS_TREE_LOG_OBJECTID, NULL, + 0, 0, 0); if (IS_ERR(leaf)) { kfree(root); return ERR_CAST(leaf); @@ -1245,8 +1406,9 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, u64 generation; u32 blocksize; int ret = 0; + int slot; - root = kzalloc(sizeof(*root), GFP_NOFS); + root = btrfs_alloc_root(fs_info); if (!root) return ERR_PTR(-ENOMEM); if (location->offset == (u64)-1) { @@ -1271,9 +1433,8 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, ret = btrfs_search_slot(NULL, tree_root, location, path, 0, 0); if (ret == 0) { l = path->nodes[0]; - read_extent_buffer(l, &root->root_item, - btrfs_item_ptr_offset(l, path->slots[0]), - sizeof(root->root_item)); + slot = path->slots[0]; + btrfs_read_root_item(tree_root, l, slot, &root->root_item); memcpy(&root->root_key, location, sizeof(*location)); } btrfs_free_path(path); @@ -1289,7 +1450,7 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item), blocksize, generation); root->commit_root = btrfs_root_node(root); - BUG_ON(!root->node); + BUG_ON(!root->node); /* -ENOMEM */ out: if (location->objectid != BTRFS_TREE_LOG_OBJECTID) { root->ref_cows = 1; @@ -1315,6 +1476,9 @@ struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info, return fs_info->dev_root; if (location->objectid == BTRFS_CSUM_TREE_OBJECTID) return fs_info->csum_root; + if (location->objectid == BTRFS_QUOTA_TREE_OBJECTID) + return fs_info->quota_root ? fs_info->quota_root : + ERR_PTR(-ENOENT); again: spin_lock(&fs_info->fs_roots_radix_lock); root = radix_tree_lookup(&fs_info->fs_roots_radix, @@ -1425,41 +1589,6 @@ static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi) return 0; } -static int bio_ready_for_csum(struct bio *bio) -{ - u64 length = 0; - u64 buf_len = 0; - u64 start = 0; - struct page *page; - struct extent_io_tree *io_tree = NULL; - struct bio_vec *bvec; - int i; - int ret; - - bio_for_each_segment(bvec, bio, i) { - page = bvec->bv_page; - if (page->private == EXTENT_PAGE_PRIVATE) { - length += bvec->bv_len; - continue; - } - if (!page->private) { - length += bvec->bv_len; - continue; - } - length = bvec->bv_len; - buf_len = page->private >> 2; - start = page_offset(page) + bvec->bv_offset; - io_tree = &BTRFS_I(page->mapping->host)->io_tree; - } - /* are we fully contained in this bio? */ - if (buf_len <= length) - return 1; - - ret = extent_range_uptodate(io_tree, start + length, - start + buf_len - 1); - return ret; -} - /* * called by the kthread helper functions to finally call the bio end_io * functions. This is where read checksum verification actually happens @@ -1475,17 +1604,6 @@ static void end_workqueue_fn(struct btrfs_work *work) bio = end_io_wq->bio; fs_info = end_io_wq->info; - /* metadata bio reads are special because the whole tree block must - * be checksummed at once. This makes sure the entire block is in - * ram and up to date before trying to verify things. For - * blocksize <= pagesize, it is basically a noop - */ - if (!(bio->bi_rw & REQ_WRITE) && end_io_wq->metadata && - !bio_ready_for_csum(bio)) { - btrfs_queue_worker(&fs_info->endio_meta_workers, - &end_io_wq->work); - return; - } error = end_io_wq->error; bio->bi_private = end_io_wq->private; bio->bi_end_io = end_io_wq->end_io; @@ -1498,8 +1616,6 @@ static int cleaner_kthread(void *arg) struct btrfs_root *root = arg; do { - vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE); - if (!(root->fs_info->sb->s_flags & MS_RDONLY) && mutex_trylock(&root->fs_info->cleaner_mutex)) { btrfs_run_delayed_iputs(root); @@ -1508,9 +1624,7 @@ static int cleaner_kthread(void *arg) btrfs_run_defrag_inodes(root->fs_info); } - if (freezing(current)) { - refrigerator(); - } else { + if (!try_to_freeze()) { set_current_state(TASK_INTERRUPTIBLE); if (!kthread_should_stop()) schedule(); @@ -1528,11 +1642,11 @@ static int transaction_kthread(void *arg) u64 transid; unsigned long now; unsigned long delay; - int ret; + bool cannot_commit; do { + cannot_commit = false; delay = HZ * 30; - vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE); mutex_lock(&root->fs_info->transaction_kthread_mutex); spin_lock(&root->fs_info->trans_lock); @@ -1552,11 +1666,14 @@ static int transaction_kthread(void *arg) transid = cur->transid; spin_unlock(&root->fs_info->trans_lock); + /* If the file system is aborted, this will always fail. */ trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + cannot_commit = true; + goto sleep; + } if (transid == trans->transid) { - ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); + btrfs_commit_transaction(trans, root); } else { btrfs_end_transaction(trans, root); } @@ -1564,12 +1681,11 @@ sleep: wake_up_process(root->fs_info->cleaner_kthread); mutex_unlock(&root->fs_info->transaction_kthread_mutex); - if (freezing(current)) { - refrigerator(); - } else { + if (!try_to_freeze()) { set_current_state(TASK_INTERRUPTIBLE); if (!kthread_should_stop() && - !btrfs_transaction_blocked(root->fs_info)) + (!btrfs_transaction_blocked(root->fs_info) || + cannot_commit)) schedule_timeout(delay); __set_current_state(TASK_RUNNING); } @@ -1577,9 +1693,246 @@ sleep: return 0; } -struct btrfs_root *open_ctree(struct super_block *sb, - struct btrfs_fs_devices *fs_devices, - char *options) +/* + * this will find the highest generation in the array of + * root backups. The index of the highest array is returned, + * or -1 if we can't find anything. + * + * We check to make sure the array is valid by comparing the + * generation of the latest root in the array with the generation + * in the super block. If they don't match we pitch it. + */ +static int find_newest_super_backup(struct btrfs_fs_info *info, u64 newest_gen) +{ + u64 cur; + int newest_index = -1; + struct btrfs_root_backup *root_backup; + int i; + + for (i = 0; i < BTRFS_NUM_BACKUP_ROOTS; i++) { + root_backup = info->super_copy->super_roots + i; + cur = btrfs_backup_tree_root_gen(root_backup); + if (cur == newest_gen) + newest_index = i; + } + + /* check to see if we actually wrapped around */ + if (newest_index == BTRFS_NUM_BACKUP_ROOTS - 1) { + root_backup = info->super_copy->super_roots; + cur = btrfs_backup_tree_root_gen(root_backup); + if (cur == newest_gen) + newest_index = 0; + } + return newest_index; +} + + +/* + * find the oldest backup so we know where to store new entries + * in the backup array. This will set the backup_root_index + * field in the fs_info struct + */ +static void find_oldest_super_backup(struct btrfs_fs_info *info, + u64 newest_gen) +{ + int newest_index = -1; + + newest_index = find_newest_super_backup(info, newest_gen); + /* if there was garbage in there, just move along */ + if (newest_index == -1) { + info->backup_root_index = 0; + } else { + info->backup_root_index = (newest_index + 1) % BTRFS_NUM_BACKUP_ROOTS; + } +} + +/* + * copy all the root pointers into the super backup array. + * this will bump the backup pointer by one when it is + * done + */ +static void backup_super_roots(struct btrfs_fs_info *info) +{ + int next_backup; + struct btrfs_root_backup *root_backup; + int last_backup; + + next_backup = info->backup_root_index; + last_backup = (next_backup + BTRFS_NUM_BACKUP_ROOTS - 1) % + BTRFS_NUM_BACKUP_ROOTS; + + /* + * just overwrite the last backup if we're at the same generation + * this happens only at umount + */ + root_backup = info->super_for_commit->super_roots + last_backup; + if (btrfs_backup_tree_root_gen(root_backup) == + btrfs_header_generation(info->tree_root->node)) + next_backup = last_backup; + + root_backup = info->super_for_commit->super_roots + next_backup; + + /* + * make sure all of our padding and empty slots get zero filled + * regardless of which ones we use today + */ + memset(root_backup, 0, sizeof(*root_backup)); + + info->backup_root_index = (next_backup + 1) % BTRFS_NUM_BACKUP_ROOTS; + + btrfs_set_backup_tree_root(root_backup, info->tree_root->node->start); + btrfs_set_backup_tree_root_gen(root_backup, + btrfs_header_generation(info->tree_root->node)); + + btrfs_set_backup_tree_root_level(root_backup, + btrfs_header_level(info->tree_root->node)); + + btrfs_set_backup_chunk_root(root_backup, info->chunk_root->node->start); + btrfs_set_backup_chunk_root_gen(root_backup, + btrfs_header_generation(info->chunk_root->node)); + btrfs_set_backup_chunk_root_level(root_backup, + btrfs_header_level(info->chunk_root->node)); + + btrfs_set_backup_extent_root(root_backup, info->extent_root->node->start); + btrfs_set_backup_extent_root_gen(root_backup, + btrfs_header_generation(info->extent_root->node)); + btrfs_set_backup_extent_root_level(root_backup, + btrfs_header_level(info->extent_root->node)); + + /* + * we might commit during log recovery, which happens before we set + * the fs_root. Make sure it is valid before we fill it in. + */ + if (info->fs_root && info->fs_root->node) { + btrfs_set_backup_fs_root(root_backup, + info->fs_root->node->start); + btrfs_set_backup_fs_root_gen(root_backup, + btrfs_header_generation(info->fs_root->node)); + btrfs_set_backup_fs_root_level(root_backup, + btrfs_header_level(info->fs_root->node)); + } + + btrfs_set_backup_dev_root(root_backup, info->dev_root->node->start); + btrfs_set_backup_dev_root_gen(root_backup, + btrfs_header_generation(info->dev_root->node)); + btrfs_set_backup_dev_root_level(root_backup, + btrfs_header_level(info->dev_root->node)); + + btrfs_set_backup_csum_root(root_backup, info->csum_root->node->start); + btrfs_set_backup_csum_root_gen(root_backup, + btrfs_header_generation(info->csum_root->node)); + btrfs_set_backup_csum_root_level(root_backup, + btrfs_header_level(info->csum_root->node)); + + btrfs_set_backup_total_bytes(root_backup, + btrfs_super_total_bytes(info->super_copy)); + btrfs_set_backup_bytes_used(root_backup, + btrfs_super_bytes_used(info->super_copy)); + btrfs_set_backup_num_devices(root_backup, + btrfs_super_num_devices(info->super_copy)); + + /* + * if we don't copy this out to the super_copy, it won't get remembered + * for the next commit + */ + memcpy(&info->super_copy->super_roots, + &info->super_for_commit->super_roots, + sizeof(*root_backup) * BTRFS_NUM_BACKUP_ROOTS); +} + +/* + * this copies info out of the root backup array and back into + * the in-memory super block. It is meant to help iterate through + * the array, so you send it the number of backups you've already + * tried and the last backup index you used. + * + * this returns -1 when it has tried all the backups + */ +static noinline int next_root_backup(struct btrfs_fs_info *info, + struct btrfs_super_block *super, + int *num_backups_tried, int *backup_index) +{ + struct btrfs_root_backup *root_backup; + int newest = *backup_index; + + if (*num_backups_tried == 0) { + u64 gen = btrfs_super_generation(super); + + newest = find_newest_super_backup(info, gen); + if (newest == -1) + return -1; + + *backup_index = newest; + *num_backups_tried = 1; + } else if (*num_backups_tried == BTRFS_NUM_BACKUP_ROOTS) { + /* we've tried all the backups, all done */ + return -1; + } else { + /* jump to the next oldest backup */ + newest = (*backup_index + BTRFS_NUM_BACKUP_ROOTS - 1) % + BTRFS_NUM_BACKUP_ROOTS; + *backup_index = newest; + *num_backups_tried += 1; + } + root_backup = super->super_roots + newest; + + btrfs_set_super_generation(super, + btrfs_backup_tree_root_gen(root_backup)); + btrfs_set_super_root(super, btrfs_backup_tree_root(root_backup)); + btrfs_set_super_root_level(super, + btrfs_backup_tree_root_level(root_backup)); + btrfs_set_super_bytes_used(super, btrfs_backup_bytes_used(root_backup)); + + /* + * fixme: the total bytes and num_devices need to match or we should + * need a fsck + */ + btrfs_set_super_total_bytes(super, btrfs_backup_total_bytes(root_backup)); + btrfs_set_super_num_devices(super, btrfs_backup_num_devices(root_backup)); + return 0; +} + +/* helper to cleanup tree roots */ +static void free_root_pointers(struct btrfs_fs_info *info, int chunk_root) +{ + free_extent_buffer(info->tree_root->node); + free_extent_buffer(info->tree_root->commit_root); + free_extent_buffer(info->dev_root->node); + free_extent_buffer(info->dev_root->commit_root); + free_extent_buffer(info->extent_root->node); + free_extent_buffer(info->extent_root->commit_root); + free_extent_buffer(info->csum_root->node); + free_extent_buffer(info->csum_root->commit_root); + if (info->quota_root) { + free_extent_buffer(info->quota_root->node); + free_extent_buffer(info->quota_root->commit_root); + } + + info->tree_root->node = NULL; + info->tree_root->commit_root = NULL; + info->dev_root->node = NULL; + info->dev_root->commit_root = NULL; + info->extent_root->node = NULL; + info->extent_root->commit_root = NULL; + info->csum_root->node = NULL; + info->csum_root->commit_root = NULL; + if (info->quota_root) { + info->quota_root->node = NULL; + info->quota_root->commit_root = NULL; + } + + if (chunk_root) { + free_extent_buffer(info->chunk_root->node); + free_extent_buffer(info->chunk_root->commit_root); + info->chunk_root->node = NULL; + info->chunk_root->commit_root = NULL; + } +} + + +int open_ctree(struct super_block *sb, + struct btrfs_fs_devices *fs_devices, + char *options) { u32 sectorsize; u32 nodesize; @@ -1590,29 +1943,32 @@ struct btrfs_root *open_ctree(struct super_block *sb, u64 features; struct btrfs_key location; struct buffer_head *bh; - struct btrfs_root *extent_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); - struct btrfs_root *csum_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); - struct btrfs_root *tree_root = btrfs_sb(sb); - struct btrfs_fs_info *fs_info = NULL; - struct btrfs_root *chunk_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); - struct btrfs_root *dev_root = kzalloc(sizeof(struct btrfs_root), - GFP_NOFS); + struct btrfs_super_block *disk_super; + struct btrfs_fs_info *fs_info = btrfs_sb(sb); + struct btrfs_root *tree_root; + struct btrfs_root *extent_root; + struct btrfs_root *csum_root; + struct btrfs_root *chunk_root; + struct btrfs_root *dev_root; + struct btrfs_root *quota_root; struct btrfs_root *log_tree_root; - int ret; int err = -EINVAL; - - struct btrfs_super_block *disk_super; - - if (!extent_root || !tree_root || !tree_root->fs_info || - !chunk_root || !dev_root || !csum_root) { + int num_backups_tried = 0; + int backup_index = 0; + + tree_root = fs_info->tree_root = btrfs_alloc_root(fs_info); + extent_root = fs_info->extent_root = btrfs_alloc_root(fs_info); + csum_root = fs_info->csum_root = btrfs_alloc_root(fs_info); + chunk_root = fs_info->chunk_root = btrfs_alloc_root(fs_info); + dev_root = fs_info->dev_root = btrfs_alloc_root(fs_info); + quota_root = fs_info->quota_root = btrfs_alloc_root(fs_info); + + if (!tree_root || !extent_root || !csum_root || + !chunk_root || !dev_root || !quota_root) { err = -ENOMEM; goto fail; } - fs_info = tree_root->fs_info; ret = init_srcu_struct(&fs_info->subvol_srcu); if (ret) { @@ -1648,35 +2004,39 @@ struct btrfs_root *open_ctree(struct super_block *sb, spin_lock_init(&fs_info->fs_roots_radix_lock); spin_lock_init(&fs_info->delayed_iput_lock); spin_lock_init(&fs_info->defrag_inodes_lock); + spin_lock_init(&fs_info->free_chunk_lock); + spin_lock_init(&fs_info->tree_mod_seq_lock); + rwlock_init(&fs_info->tree_mod_log_lock); mutex_init(&fs_info->reloc_mutex); init_completion(&fs_info->kobj_unregister); - fs_info->tree_root = tree_root; - fs_info->extent_root = extent_root; - fs_info->csum_root = csum_root; - fs_info->chunk_root = chunk_root; - fs_info->dev_root = dev_root; - fs_info->fs_devices = fs_devices; INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots); INIT_LIST_HEAD(&fs_info->space_info); + INIT_LIST_HEAD(&fs_info->tree_mod_seq_list); btrfs_mapping_init(&fs_info->mapping_tree); btrfs_init_block_rsv(&fs_info->global_block_rsv); btrfs_init_block_rsv(&fs_info->delalloc_block_rsv); btrfs_init_block_rsv(&fs_info->trans_block_rsv); btrfs_init_block_rsv(&fs_info->chunk_block_rsv); btrfs_init_block_rsv(&fs_info->empty_block_rsv); - INIT_LIST_HEAD(&fs_info->durable_block_rsv_list); - mutex_init(&fs_info->durable_block_rsv_mutex); + btrfs_init_block_rsv(&fs_info->delayed_block_rsv); atomic_set(&fs_info->nr_async_submits, 0); atomic_set(&fs_info->async_delalloc_pages, 0); atomic_set(&fs_info->async_submit_draining, 0); atomic_set(&fs_info->nr_async_bios, 0); atomic_set(&fs_info->defrag_running, 0); + atomic_set(&fs_info->tree_mod_seq, 0); fs_info->sb = sb; fs_info->max_inline = 8192 * 1024; fs_info->metadata_ratio = 0; fs_info->defrag_inodes = RB_ROOT; fs_info->trans_no_join = 0; + fs_info->free_chunk_space = 0; + fs_info->tree_mod_log = RB_ROOT; + + /* readahead state */ + INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_WAIT); + spin_lock_init(&fs_info->reada_lock); fs_info->thread_pool_size = min_t(unsigned long, num_online_cpus() + 2, 8); @@ -1699,13 +2059,24 @@ struct btrfs_root *open_ctree(struct super_block *sb, init_waitqueue_head(&fs_info->scrub_pause_wait); init_rwsem(&fs_info->scrub_super_lock); fs_info->scrub_workers_refcnt = 0; +#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY + fs_info->check_integrity_print_mask = 0; +#endif + + spin_lock_init(&fs_info->balance_lock); + mutex_init(&fs_info->balance_mutex); + atomic_set(&fs_info->balance_running, 0); + atomic_set(&fs_info->balance_pause_req, 0); + atomic_set(&fs_info->balance_cancel_req, 0); + fs_info->balance_ctl = NULL; + init_waitqueue_head(&fs_info->balance_wait_q); sb->s_blocksize = 4096; sb->s_blocksize_bits = blksize_bits(4096); sb->s_bdi = &fs_info->bdi; fs_info->btree_inode->i_ino = BTRFS_BTREE_INODE_OBJECTID; - fs_info->btree_inode->i_nlink = 1; + set_nlink(fs_info->btree_inode, 1); /* * we set the i_size on the btree inode to the max possible int. * the real end of the address space is determined by all of @@ -1718,6 +2089,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node); extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree, fs_info->btree_inode->i_mapping); + BTRFS_I(fs_info->btree_inode)->io_tree.track_uptodate = 0; extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree); BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops; @@ -1725,7 +2097,8 @@ struct btrfs_root *open_ctree(struct super_block *sb, BTRFS_I(fs_info->btree_inode)->root = tree_root; memset(&BTRFS_I(fs_info->btree_inode)->location, 0, sizeof(struct btrfs_key)); - BTRFS_I(fs_info->btree_inode)->dummy_inode = 1; + set_bit(BTRFS_INODE_DUMMY, + &BTRFS_I(fs_info->btree_inode)->runtime_flags); insert_inode_hash(fs_info->btree_inode); spin_lock_init(&fs_info->block_group_cache_lock); @@ -1749,6 +2122,13 @@ struct btrfs_root *open_ctree(struct super_block *sb, init_rwsem(&fs_info->cleanup_work_sem); init_rwsem(&fs_info->subvol_sem); + spin_lock_init(&fs_info->qgroup_lock); + fs_info->qgroup_tree = RB_ROOT; + INIT_LIST_HEAD(&fs_info->dirty_qgroups); + fs_info->qgroup_seq = 1; + fs_info->quota_enabled = 0; + fs_info->pending_quota_state = 0; + btrfs_init_free_cluster(&fs_info->meta_alloc_cluster); btrfs_init_free_cluster(&fs_info->data_alloc_cluster); @@ -1760,27 +2140,40 @@ struct btrfs_root *open_ctree(struct super_block *sb, __setup_root(4096, 4096, 4096, 4096, tree_root, fs_info, BTRFS_ROOT_TREE_OBJECTID); + invalidate_bdev(fs_devices->latest_bdev); bh = btrfs_read_dev_super(fs_devices->latest_bdev); if (!bh) { err = -EINVAL; goto fail_alloc; } - memcpy(&fs_info->super_copy, bh->b_data, sizeof(fs_info->super_copy)); - memcpy(&fs_info->super_for_commit, &fs_info->super_copy, - sizeof(fs_info->super_for_commit)); + memcpy(fs_info->super_copy, bh->b_data, sizeof(*fs_info->super_copy)); + memcpy(fs_info->super_for_commit, fs_info->super_copy, + sizeof(*fs_info->super_for_commit)); brelse(bh); - memcpy(fs_info->fsid, fs_info->super_copy.fsid, BTRFS_FSID_SIZE); + memcpy(fs_info->fsid, fs_info->super_copy->fsid, BTRFS_FSID_SIZE); - disk_super = &fs_info->super_copy; + disk_super = fs_info->super_copy; if (!btrfs_super_root(disk_super)) goto fail_alloc; /* check FS state, whether FS is broken. */ fs_info->fs_state |= btrfs_super_flags(disk_super); - btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY); + ret = btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY); + if (ret) { + printk(KERN_ERR "btrfs: superblock contains fatal errors\n"); + err = ret; + goto fail_alloc; + } + + /* + * run through our array of backup supers and setup + * our ring pointer to the oldest one + */ + generation = btrfs_super_generation(disk_super); + find_oldest_super_backup(fs_info, generation); /* * In the long term, we'll store the compression type in the super @@ -1804,10 +2197,55 @@ struct btrfs_root *open_ctree(struct super_block *sb, goto fail_alloc; } + if (btrfs_super_leafsize(disk_super) != + btrfs_super_nodesize(disk_super)) { + printk(KERN_ERR "BTRFS: couldn't mount because metadata " + "blocksizes don't match. node %d leaf %d\n", + btrfs_super_nodesize(disk_super), + btrfs_super_leafsize(disk_super)); + err = -EINVAL; + goto fail_alloc; + } + if (btrfs_super_leafsize(disk_super) > BTRFS_MAX_METADATA_BLOCKSIZE) { + printk(KERN_ERR "BTRFS: couldn't mount because metadata " + "blocksize (%d) was too large\n", + btrfs_super_leafsize(disk_super)); + err = -EINVAL; + goto fail_alloc; + } + features = btrfs_super_incompat_flags(disk_super); features |= BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF; - if (tree_root->fs_info->compress_type & BTRFS_COMPRESS_LZO) + if (tree_root->fs_info->compress_type == BTRFS_COMPRESS_LZO) features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO; + + /* + * flag our filesystem as having big metadata blocks if + * they are bigger than the page size + */ + if (btrfs_super_leafsize(disk_super) > PAGE_CACHE_SIZE) { + if (!(features & BTRFS_FEATURE_INCOMPAT_BIG_METADATA)) + printk(KERN_INFO "btrfs flagging fs with big metadata feature\n"); + features |= BTRFS_FEATURE_INCOMPAT_BIG_METADATA; + } + + nodesize = btrfs_super_nodesize(disk_super); + leafsize = btrfs_super_leafsize(disk_super); + sectorsize = btrfs_super_sectorsize(disk_super); + stripesize = btrfs_super_stripesize(disk_super); + + /* + * mixed block groups end up with duplicate but slightly offset + * extent buffers for the same range. It leads to corruptions + */ + if ((features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) && + (sectorsize != leafsize)) { + printk(KERN_WARNING "btrfs: unequal leaf/node/sector sizes " + "are not allowed for mixed block groups on %s\n", + sb->s_id); + goto fail_alloc; + } + btrfs_set_super_incompat_flags(disk_super, features); features = btrfs_super_compat_ro_flags(disk_super) & @@ -1870,6 +2308,9 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_init_workers(&fs_info->delayed_workers, "delayed-meta", fs_info->thread_pool_size, &fs_info->generic_worker); + btrfs_init_workers(&fs_info->readahead_workers, "readahead", + fs_info->thread_pool_size, + &fs_info->generic_worker); /* * endios are largely parallel and should have a very @@ -1880,28 +2321,34 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->endio_write_workers.idle_thresh = 2; fs_info->endio_meta_write_workers.idle_thresh = 2; + fs_info->readahead_workers.idle_thresh = 2; - btrfs_start_workers(&fs_info->workers, 1); - btrfs_start_workers(&fs_info->generic_worker, 1); - btrfs_start_workers(&fs_info->submit_workers, 1); - btrfs_start_workers(&fs_info->delalloc_workers, 1); - btrfs_start_workers(&fs_info->fixup_workers, 1); - btrfs_start_workers(&fs_info->endio_workers, 1); - btrfs_start_workers(&fs_info->endio_meta_workers, 1); - btrfs_start_workers(&fs_info->endio_meta_write_workers, 1); - btrfs_start_workers(&fs_info->endio_write_workers, 1); - btrfs_start_workers(&fs_info->endio_freespace_worker, 1); - btrfs_start_workers(&fs_info->delayed_workers, 1); - btrfs_start_workers(&fs_info->caching_workers, 1); + /* + * btrfs_start_workers can really only fail because of ENOMEM so just + * return -ENOMEM if any of these fail. + */ + ret = btrfs_start_workers(&fs_info->workers); + ret |= btrfs_start_workers(&fs_info->generic_worker); + ret |= btrfs_start_workers(&fs_info->submit_workers); + ret |= btrfs_start_workers(&fs_info->delalloc_workers); + ret |= btrfs_start_workers(&fs_info->fixup_workers); + ret |= btrfs_start_workers(&fs_info->endio_workers); + ret |= btrfs_start_workers(&fs_info->endio_meta_workers); + ret |= btrfs_start_workers(&fs_info->endio_meta_write_workers); + ret |= btrfs_start_workers(&fs_info->endio_write_workers); + ret |= btrfs_start_workers(&fs_info->endio_freespace_worker); + ret |= btrfs_start_workers(&fs_info->delayed_workers); + ret |= btrfs_start_workers(&fs_info->caching_workers); + ret |= btrfs_start_workers(&fs_info->readahead_workers); + if (ret) { + err = -ENOMEM; + goto fail_sb_buffer; + } fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super); fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages, 4 * 1024 * 1024 / PAGE_CACHE_SIZE); - nodesize = btrfs_super_nodesize(disk_super); - leafsize = btrfs_super_leafsize(disk_super); - sectorsize = btrfs_super_sectorsize(disk_super); - stripesize = btrfs_super_stripesize(disk_super); tree_root->nodesize = nodesize; tree_root->leafsize = leafsize; tree_root->sectorsize = sectorsize; @@ -1916,6 +2363,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, goto fail_sb_buffer; } + if (sectorsize != PAGE_SIZE) { + printk(KERN_WARNING "btrfs: Incompatible sector size(%lu) " + "found on %s\n", (unsigned long)sectorsize, sb->s_id); + goto fail_sb_buffer; + } + mutex_lock(&fs_info->chunk_mutex); ret = btrfs_read_sys_array(tree_root); mutex_unlock(&fs_info->chunk_mutex); @@ -1935,11 +2388,11 @@ struct btrfs_root *open_ctree(struct super_block *sb, chunk_root->node = read_tree_block(chunk_root, btrfs_super_chunk_root(disk_super), blocksize, generation); - BUG_ON(!chunk_root->node); + BUG_ON(!chunk_root->node); /* -ENOMEM */ if (!test_bit(EXTENT_BUFFER_UPTODATE, &chunk_root->node->bflags)) { printk(KERN_WARNING "btrfs: failed to read chunk root on %s\n", sb->s_id); - goto fail_chunk_root; + goto fail_tree_roots; } btrfs_set_root_node(&chunk_root->root_item, chunk_root->node); chunk_root->commit_root = btrfs_root_node(chunk_root); @@ -1948,17 +2401,22 @@ struct btrfs_root *open_ctree(struct super_block *sb, (unsigned long)btrfs_header_chunk_tree_uuid(chunk_root->node), BTRFS_UUID_SIZE); - mutex_lock(&fs_info->chunk_mutex); ret = btrfs_read_chunk_tree(chunk_root); - mutex_unlock(&fs_info->chunk_mutex); if (ret) { printk(KERN_WARNING "btrfs: failed to read chunk tree on %s\n", sb->s_id); - goto fail_chunk_root; + goto fail_tree_roots; } btrfs_close_extra_devices(fs_devices); + if (!fs_devices->latest_bdev) { + printk(KERN_CRIT "btrfs: failed to read devices on %s\n", + sb->s_id); + goto fail_tree_roots; + } + +retry_root_backup: blocksize = btrfs_level_size(tree_root, btrfs_super_root_level(disk_super)); generation = btrfs_super_generation(disk_super); @@ -1966,40 +2424,61 @@ struct btrfs_root *open_ctree(struct super_block *sb, tree_root->node = read_tree_block(tree_root, btrfs_super_root(disk_super), blocksize, generation); - if (!tree_root->node) - goto fail_chunk_root; - if (!test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) { + if (!tree_root->node || + !test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) { printk(KERN_WARNING "btrfs: failed to read tree root on %s\n", sb->s_id); - goto fail_tree_root; + + goto recovery_tree_root; } + btrfs_set_root_node(&tree_root->root_item, tree_root->node); tree_root->commit_root = btrfs_root_node(tree_root); ret = find_and_setup_root(tree_root, fs_info, BTRFS_EXTENT_TREE_OBJECTID, extent_root); if (ret) - goto fail_tree_root; + goto recovery_tree_root; extent_root->track_dirty = 1; ret = find_and_setup_root(tree_root, fs_info, BTRFS_DEV_TREE_OBJECTID, dev_root); if (ret) - goto fail_extent_root; + goto recovery_tree_root; dev_root->track_dirty = 1; ret = find_and_setup_root(tree_root, fs_info, BTRFS_CSUM_TREE_OBJECTID, csum_root); if (ret) - goto fail_dev_root; - + goto recovery_tree_root; csum_root->track_dirty = 1; + ret = find_and_setup_root(tree_root, fs_info, + BTRFS_QUOTA_TREE_OBJECTID, quota_root); + if (ret) { + kfree(quota_root); + quota_root = fs_info->quota_root = NULL; + } else { + quota_root->track_dirty = 1; + fs_info->quota_enabled = 1; + fs_info->pending_quota_state = 1; + } + fs_info->generation = generation; fs_info->last_trans_committed = generation; - fs_info->data_alloc_profile = (u64)-1; - fs_info->metadata_alloc_profile = (u64)-1; - fs_info->system_alloc_profile = fs_info->metadata_alloc_profile; + + ret = btrfs_recover_balance(fs_info); + if (ret) { + printk(KERN_WARNING "btrfs: failed to recover balance\n"); + goto fail_block_groups; + } + + ret = btrfs_init_dev_stats(fs_info); + if (ret) { + printk(KERN_ERR "btrfs: failed to init dev_stats: %d\n", + ret); + goto fail_block_groups; + } ret = btrfs_init_space_info(fs_info); if (ret) { @@ -2032,25 +2511,40 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_set_opt(fs_info->mount_opt, SSD); } +#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY + if (btrfs_test_opt(tree_root, CHECK_INTEGRITY)) { + ret = btrfsic_mount(tree_root, fs_devices, + btrfs_test_opt(tree_root, + CHECK_INTEGRITY_INCLUDING_EXTENT_DATA) ? + 1 : 0, + fs_info->check_integrity_print_mask); + if (ret) + printk(KERN_WARNING "btrfs: failed to initialize" + " integrity check module %s\n", sb->s_id); + } +#endif + ret = btrfs_read_qgroup_config(fs_info); + if (ret) + goto fail_trans_kthread; + /* do not make disk changes in broken FS */ - if (btrfs_super_log_root(disk_super) != 0 && - !(fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR)) { + if (btrfs_super_log_root(disk_super) != 0) { u64 bytenr = btrfs_super_log_root(disk_super); if (fs_devices->rw_devices == 0) { printk(KERN_WARNING "Btrfs log replay required " "on RO media\n"); err = -EIO; - goto fail_trans_kthread; + goto fail_qgroup; } blocksize = btrfs_level_size(tree_root, btrfs_super_log_root_level(disk_super)); - log_tree_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); + log_tree_root = btrfs_alloc_root(fs_info); if (!log_tree_root) { err = -ENOMEM; - goto fail_trans_kthread; + goto fail_qgroup; } __setup_root(nodesize, leafsize, sectorsize, stripesize, @@ -2059,28 +2553,38 @@ struct btrfs_root *open_ctree(struct super_block *sb, log_tree_root->node = read_tree_block(tree_root, bytenr, blocksize, generation + 1); + /* returns with log_tree_root freed on success */ ret = btrfs_recover_log_trees(log_tree_root); - BUG_ON(ret); + if (ret) { + btrfs_error(tree_root->fs_info, ret, + "Failed to recover log tree"); + free_extent_buffer(log_tree_root->node); + kfree(log_tree_root); + goto fail_trans_kthread; + } if (sb->s_flags & MS_RDONLY) { - ret = btrfs_commit_super(tree_root); - BUG_ON(ret); + ret = btrfs_commit_super(tree_root); + if (ret) + goto fail_trans_kthread; } } ret = btrfs_find_orphan_roots(tree_root); - BUG_ON(ret); + if (ret) + goto fail_trans_kthread; if (!(sb->s_flags & MS_RDONLY)) { ret = btrfs_cleanup_fs_roots(fs_info); - BUG_ON(ret); + if (ret) + goto fail_trans_kthread; ret = btrfs_recover_relocation(tree_root); if (ret < 0) { printk(KERN_WARNING "btrfs: failed to recover relocation\n"); err = -EINVAL; - goto fail_trans_kthread; + goto fail_qgroup; } } @@ -2090,26 +2594,35 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->fs_root = btrfs_read_fs_root_no_name(fs_info, &location); if (!fs_info->fs_root) - goto fail_trans_kthread; + goto fail_qgroup; if (IS_ERR(fs_info->fs_root)) { err = PTR_ERR(fs_info->fs_root); - goto fail_trans_kthread; + goto fail_qgroup; } - if (!(sb->s_flags & MS_RDONLY)) { - down_read(&fs_info->cleanup_work_sem); - err = btrfs_orphan_cleanup(fs_info->fs_root); - if (!err) - err = btrfs_orphan_cleanup(fs_info->tree_root); + if (sb->s_flags & MS_RDONLY) + return 0; + + down_read(&fs_info->cleanup_work_sem); + if ((ret = btrfs_orphan_cleanup(fs_info->fs_root)) || + (ret = btrfs_orphan_cleanup(fs_info->tree_root))) { up_read(&fs_info->cleanup_work_sem); - if (err) { - close_ctree(tree_root); - return ERR_PTR(err); - } + close_ctree(tree_root); + return ret; } + up_read(&fs_info->cleanup_work_sem); - return tree_root; + ret = btrfs_resume_balance_async(fs_info); + if (ret) { + printk(KERN_WARNING "btrfs: failed to resume balance\n"); + close_ctree(tree_root); + return ret; + } + return 0; + +fail_qgroup: + btrfs_free_qgroup_config(fs_info); fail_trans_kthread: kthread_stop(fs_info->transaction_kthread); fail_cleaner: @@ -2124,22 +2637,13 @@ fail_cleaner: fail_block_groups: btrfs_free_block_groups(fs_info); - free_extent_buffer(csum_root->node); - free_extent_buffer(csum_root->commit_root); -fail_dev_root: - free_extent_buffer(dev_root->node); - free_extent_buffer(dev_root->commit_root); -fail_extent_root: - free_extent_buffer(extent_root->node); - free_extent_buffer(extent_root->commit_root); -fail_tree_root: - free_extent_buffer(tree_root->node); - free_extent_buffer(tree_root->commit_root); -fail_chunk_root: - free_extent_buffer(chunk_root->node); - free_extent_buffer(chunk_root->commit_root); + +fail_tree_roots: + free_root_pointers(fs_info, 1); + fail_sb_buffer: btrfs_stop_workers(&fs_info->generic_worker); + btrfs_stop_workers(&fs_info->readahead_workers); btrfs_stop_workers(&fs_info->fixup_workers); btrfs_stop_workers(&fs_info->delalloc_workers); btrfs_stop_workers(&fs_info->workers); @@ -2152,41 +2656,54 @@ fail_sb_buffer: btrfs_stop_workers(&fs_info->delayed_workers); btrfs_stop_workers(&fs_info->caching_workers); fail_alloc: - kfree(fs_info->delayed_root); fail_iput: + btrfs_mapping_tree_free(&fs_info->mapping_tree); + invalidate_inode_pages2(fs_info->btree_inode->i_mapping); iput(fs_info->btree_inode); - - btrfs_close_devices(fs_info->fs_devices); - btrfs_mapping_tree_free(&fs_info->mapping_tree); fail_bdi: bdi_destroy(&fs_info->bdi); fail_srcu: cleanup_srcu_struct(&fs_info->subvol_srcu); fail: - kfree(extent_root); - kfree(tree_root); - kfree(fs_info); - kfree(chunk_root); - kfree(dev_root); - kfree(csum_root); - return ERR_PTR(err); + btrfs_close_devices(fs_info->fs_devices); + return err; + +recovery_tree_root: + if (!btrfs_test_opt(tree_root, RECOVERY)) + goto fail_tree_roots; + + free_root_pointers(fs_info, 0); + + /* don't use the log in recovery mode, it won't be valid */ + btrfs_set_super_log_root(disk_super, 0); + + /* we can't trust the free space cache either */ + btrfs_set_opt(fs_info->mount_opt, CLEAR_CACHE); + + ret = next_root_backup(fs_info, fs_info->super_copy, + &num_backups_tried, &backup_index); + if (ret == -1) + goto fail_block_groups; + goto retry_root_backup; } static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate) { - char b[BDEVNAME_SIZE]; - if (uptodate) { set_buffer_uptodate(bh); } else { - printk_ratelimited(KERN_WARNING "lost page write due to " - "I/O error on %s\n", - bdevname(bh->b_bdev, b)); + struct btrfs_device *device = (struct btrfs_device *) + bh->b_private; + + printk_ratelimited_in_rcu(KERN_WARNING "lost page write due to " + "I/O error on %s\n", + rcu_str_deref(device->name)); /* note, we dont' set_buffer_write_io_error because we have * our own ways of dealing with the IO errors */ clear_buffer_uptodate(bh); + btrfs_dev_stat_inc_and_print(device, BTRFS_DEV_STAT_WRITE_ERRS); } unlock_buffer(bh); put_bh(bh); @@ -2254,22 +2771,10 @@ static int write_dev_supers(struct btrfs_device *device, int errors = 0; u32 crc; u64 bytenr; - int last_barrier = 0; if (max_mirrors == 0) max_mirrors = BTRFS_SUPER_MIRROR_MAX; - /* make sure only the last submit_bh does a barrier */ - if (do_barriers) { - for (i = 0; i < max_mirrors; i++) { - bytenr = btrfs_sb_offset(i); - if (bytenr + BTRFS_SUPER_INFO_SIZE >= - device->total_bytes) - break; - last_barrier = i; - } - } - for (i = 0; i < max_mirrors; i++) { bytenr = btrfs_sb_offset(i); if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes) @@ -2313,19 +2818,142 @@ static int write_dev_supers(struct btrfs_device *device, set_buffer_uptodate(bh); lock_buffer(bh); bh->b_end_io = btrfs_end_buffer_write_sync; + bh->b_private = device; } - if (i == last_barrier && do_barriers) - ret = submit_bh(WRITE_FLUSH_FUA, bh); - else - ret = submit_bh(WRITE_SYNC, bh); - + /* + * we fua the first super. The others we allow + * to go down lazy. + */ + ret = btrfsic_submit_bh(WRITE_FUA, bh); if (ret) errors++; } return errors < i ? 0 : -1; } +/* + * endio for the write_dev_flush, this will wake anyone waiting + * for the barrier when it is done + */ +static void btrfs_end_empty_barrier(struct bio *bio, int err) +{ + if (err) { + if (err == -EOPNOTSUPP) + set_bit(BIO_EOPNOTSUPP, &bio->bi_flags); + clear_bit(BIO_UPTODATE, &bio->bi_flags); + } + if (bio->bi_private) + complete(bio->bi_private); + bio_put(bio); +} + +/* + * trigger flushes for one the devices. If you pass wait == 0, the flushes are + * sent down. With wait == 1, it waits for the previous flush. + * + * any device where the flush fails with eopnotsupp are flagged as not-barrier + * capable + */ +static int write_dev_flush(struct btrfs_device *device, int wait) +{ + struct bio *bio; + int ret = 0; + + if (device->nobarriers) + return 0; + + if (wait) { + bio = device->flush_bio; + if (!bio) + return 0; + + wait_for_completion(&device->flush_wait); + + if (bio_flagged(bio, BIO_EOPNOTSUPP)) { + printk_in_rcu("btrfs: disabling barriers on dev %s\n", + rcu_str_deref(device->name)); + device->nobarriers = 1; + } + if (!bio_flagged(bio, BIO_UPTODATE)) { + ret = -EIO; + if (!bio_flagged(bio, BIO_EOPNOTSUPP)) + btrfs_dev_stat_inc_and_print(device, + BTRFS_DEV_STAT_FLUSH_ERRS); + } + + /* drop the reference from the wait == 0 run */ + bio_put(bio); + device->flush_bio = NULL; + + return ret; + } + + /* + * one reference for us, and we leave it for the + * caller + */ + device->flush_bio = NULL; + bio = bio_alloc(GFP_NOFS, 0); + if (!bio) + return -ENOMEM; + + bio->bi_end_io = btrfs_end_empty_barrier; + bio->bi_bdev = device->bdev; + init_completion(&device->flush_wait); + bio->bi_private = &device->flush_wait; + device->flush_bio = bio; + + bio_get(bio); + btrfsic_submit_bio(WRITE_FLUSH, bio); + + return 0; +} + +/* + * send an empty flush down to each device in parallel, + * then wait for them + */ +static int barrier_all_devices(struct btrfs_fs_info *info) +{ + struct list_head *head; + struct btrfs_device *dev; + int errors = 0; + int ret; + + /* send down all the barriers */ + head = &info->fs_devices->devices; + list_for_each_entry_rcu(dev, head, dev_list) { + if (!dev->bdev) { + errors++; + continue; + } + if (!dev->in_fs_metadata || !dev->writeable) + continue; + + ret = write_dev_flush(dev, 0); + if (ret) + errors++; + } + + /* wait for all the barriers */ + list_for_each_entry_rcu(dev, head, dev_list) { + if (!dev->bdev) { + errors++; + continue; + } + if (!dev->in_fs_metadata || !dev->writeable) + continue; + + ret = write_dev_flush(dev, 1); + if (ret) + errors++; + } + if (errors) + return -EIO; + return 0; +} + int write_all_supers(struct btrfs_root *root, int max_mirrors) { struct list_head *head; @@ -2338,14 +2966,19 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) int total_errors = 0; u64 flags; - max_errors = btrfs_super_num_devices(&root->fs_info->super_copy) - 1; + max_errors = btrfs_super_num_devices(root->fs_info->super_copy) - 1; do_barriers = !btrfs_test_opt(root, NOBARRIER); + backup_super_roots(root->fs_info); - sb = &root->fs_info->super_for_commit; + sb = root->fs_info->super_for_commit; dev_item = &sb->dev_item; mutex_lock(&root->fs_info->fs_devices->device_list_mutex); head = &root->fs_info->fs_devices->devices; + + if (do_barriers) + barrier_all_devices(root->fs_info); + list_for_each_entry_rcu(dev, head, dev_list) { if (!dev->bdev) { total_errors++; @@ -2375,6 +3008,8 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) if (total_errors > max_errors) { printk(KERN_ERR "btrfs: %d errors while writing supers\n", total_errors); + + /* This shouldn't happen. FUA is masked off if unsupported */ BUG(); } @@ -2391,9 +3026,9 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) } mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); if (total_errors > max_errors) { - printk(KERN_ERR "btrfs: %d errors while writing supers\n", - total_errors); - BUG(); + btrfs_error(root->fs_info, -EIO, + "%d errors while writing supers", total_errors); + return -EIO; } return 0; } @@ -2407,7 +3042,7 @@ int write_ctree_super(struct btrfs_trans_handle *trans, return ret; } -int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root) +void btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root) { spin_lock(&fs_info->fs_roots_radix_lock); radix_tree_delete(&fs_info->fs_roots_radix, @@ -2420,7 +3055,6 @@ int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root) __btrfs_remove_free_space_cache(root->free_ino_pinned); __btrfs_remove_free_space_cache(root->free_ino_ctl); free_fs_root(root); - return 0; } static void free_fs_root(struct btrfs_root *root) @@ -2437,7 +3071,7 @@ static void free_fs_root(struct btrfs_root *root) kfree(root); } -static int del_fs_roots(struct btrfs_fs_info *fs_info) +static void del_fs_roots(struct btrfs_fs_info *fs_info) { int ret; struct btrfs_root *gang[8]; @@ -2466,7 +3100,6 @@ static int del_fs_roots(struct btrfs_fs_info *fs_info) for (i = 0; i < ret; i++) btrfs_free_fs_root(fs_info, gang[i]); } - return 0; } int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info) @@ -2515,14 +3148,21 @@ int btrfs_commit_super(struct btrfs_root *root) if (IS_ERR(trans)) return PTR_ERR(trans); ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); + if (ret) + return ret; /* run commit again to drop the original snapshot */ trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_commit_transaction(trans, root); + ret = btrfs_commit_transaction(trans, root); + if (ret) + return ret; ret = btrfs_write_and_wait_transaction(NULL, root); - BUG_ON(ret); + if (ret) { + btrfs_error(root->fs_info, ret, + "Failed to sync btree inode to disk."); + return ret; + } ret = write_ctree_super(NULL, root, 0); return ret; @@ -2536,6 +3176,9 @@ int close_ctree(struct btrfs_root *root) fs_info->closing = 1; smp_mb(); + /* pause restriper - we want to resume on mount */ + btrfs_pause_balance(root->fs_info); + btrfs_scrub_cancel(root); /* wait for any defraggers to finish */ @@ -2543,41 +3186,27 @@ int close_ctree(struct btrfs_root *root) (atomic_read(&fs_info->defrag_running) == 0)); /* clear out the rbtree of defraggable inodes */ - btrfs_run_defrag_inodes(root->fs_info); + btrfs_run_defrag_inodes(fs_info); - btrfs_put_block_group_cache(fs_info); - - /* - * Here come 2 situations when btrfs is broken to flip readonly: - * - * 1. when btrfs flips readonly somewhere else before - * btrfs_commit_super, sb->s_flags has MS_RDONLY flag, - * and btrfs will skip to write sb directly to keep - * ERROR state on disk. - * - * 2. when btrfs flips readonly just in btrfs_commit_super, - * and in such case, btrfs cannot write sb via btrfs_commit_super, - * and since fs_state has been set BTRFS_SUPER_FLAG_ERROR flag, - * btrfs will cleanup all FS resources first and write sb then. - */ if (!(fs_info->sb->s_flags & MS_RDONLY)) { ret = btrfs_commit_super(root); if (ret) printk(KERN_ERR "btrfs: commit super ret %d\n", ret); } - if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { - ret = btrfs_error_commit_super(root); - if (ret) - printk(KERN_ERR "btrfs: commit super ret %d\n", ret); - } + if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) + btrfs_error_commit_super(root); + + btrfs_put_block_group_cache(fs_info); - kthread_stop(root->fs_info->transaction_kthread); - kthread_stop(root->fs_info->cleaner_kthread); + kthread_stop(fs_info->transaction_kthread); + kthread_stop(fs_info->cleaner_kthread); fs_info->closing = 2; smp_mb(); + btrfs_free_qgroup_config(root->fs_info); + if (fs_info->delalloc_bytes) { printk(KERN_INFO "btrfs: at unmount delalloc count %llu\n", (unsigned long long)fs_info->delalloc_bytes); @@ -2591,19 +3220,22 @@ int close_ctree(struct btrfs_root *root) free_extent_buffer(fs_info->extent_root->commit_root); free_extent_buffer(fs_info->tree_root->node); free_extent_buffer(fs_info->tree_root->commit_root); - free_extent_buffer(root->fs_info->chunk_root->node); - free_extent_buffer(root->fs_info->chunk_root->commit_root); - free_extent_buffer(root->fs_info->dev_root->node); - free_extent_buffer(root->fs_info->dev_root->commit_root); - free_extent_buffer(root->fs_info->csum_root->node); - free_extent_buffer(root->fs_info->csum_root->commit_root); + free_extent_buffer(fs_info->chunk_root->node); + free_extent_buffer(fs_info->chunk_root->commit_root); + free_extent_buffer(fs_info->dev_root->node); + free_extent_buffer(fs_info->dev_root->commit_root); + free_extent_buffer(fs_info->csum_root->node); + free_extent_buffer(fs_info->csum_root->commit_root); + if (fs_info->quota_root) { + free_extent_buffer(fs_info->quota_root->node); + free_extent_buffer(fs_info->quota_root->commit_root); + } - btrfs_free_block_groups(root->fs_info); + btrfs_free_block_groups(fs_info); del_fs_roots(fs_info); iput(fs_info->btree_inode); - kfree(fs_info->delayed_root); btrfs_stop_workers(&fs_info->generic_worker); btrfs_stop_workers(&fs_info->fixup_workers); @@ -2617,6 +3249,12 @@ int close_ctree(struct btrfs_root *root) btrfs_stop_workers(&fs_info->submit_workers); btrfs_stop_workers(&fs_info->delayed_workers); btrfs_stop_workers(&fs_info->caching_workers); + btrfs_stop_workers(&fs_info->readahead_workers); + +#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY + if (btrfs_test_opt(root, CHECK_INTEGRITY)) + btrfsic_unmount(root, fs_info->fs_devices); +#endif btrfs_close_devices(fs_info->fs_devices); btrfs_mapping_tree_free(&fs_info->mapping_tree); @@ -2624,43 +3262,35 @@ int close_ctree(struct btrfs_root *root) bdi_destroy(&fs_info->bdi); cleanup_srcu_struct(&fs_info->subvol_srcu); - kfree(fs_info->extent_root); - kfree(fs_info->tree_root); - kfree(fs_info->chunk_root); - kfree(fs_info->dev_root); - kfree(fs_info->csum_root); - kfree(fs_info); - return 0; } -int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid) +int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid, + int atomic) { int ret; - struct inode *btree_inode = buf->first_page->mapping->host; + struct inode *btree_inode = buf->pages[0]->mapping->host; - ret = extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree, buf, - NULL); + ret = extent_buffer_uptodate(buf); if (!ret) return ret; ret = verify_parent_transid(&BTRFS_I(btree_inode)->io_tree, buf, - parent_transid); + parent_transid, atomic); + if (ret == -EAGAIN) + return ret; return !ret; } int btrfs_set_buffer_uptodate(struct extent_buffer *buf) { - struct inode *btree_inode = buf->first_page->mapping->host; - return set_extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree, - buf); + return set_extent_buffer_uptodate(buf); } void btrfs_mark_buffer_dirty(struct extent_buffer *buf) { - struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root; + struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root; u64 transid = btrfs_header_generation(buf); - struct inode *btree_inode = root->fs_info->btree_inode; int was_dirty; btrfs_assert_tree_locked(buf); @@ -2672,8 +3302,7 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf) (unsigned long long)root->fs_info->generation); WARN_ON(1); } - was_dirty = set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, - buf); + was_dirty = set_extent_buffer_dirty(buf); if (!was_dirty) { spin_lock(&root->fs_info->delalloc_lock); root->fs_info->dirty_metadata_bytes += buf->len; @@ -2727,32 +3356,36 @@ void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid) { - struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root; - int ret; - ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid); - if (ret == 0) - set_bit(EXTENT_BUFFER_UPTODATE, &buf->bflags); - return ret; + struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root; + return btree_read_extent_buffer_pages(root, buf, 0, parent_transid); } -int btree_lock_page_hook(struct page *page) +int btree_lock_page_hook(struct page *page, void *data, + void (*flush_fn)(void *)) { struct inode *inode = page->mapping->host; struct btrfs_root *root = BTRFS_I(inode)->root; - struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; struct extent_buffer *eb; - unsigned long len; - u64 bytenr = page_offset(page); - if (page->private == EXTENT_PAGE_PRIVATE) + /* + * We culled this eb but the page is still hanging out on the mapping, + * carry on. + */ + if (!PagePrivate(page)) goto out; - len = page->private >> 2; - eb = find_extent_buffer(io_tree, bytenr, len); - if (!eb) + eb = (struct extent_buffer *)page->private; + if (!eb) { + WARN_ON(1); + goto out; + } + if (page != eb->pages[0]) goto out; - btrfs_tree_lock(eb); + if (!btrfs_try_tree_write_lock(eb)) { + flush_fn(data); + btrfs_tree_lock(eb); + } btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN); if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) { @@ -2765,27 +3398,30 @@ int btree_lock_page_hook(struct page *page) } btrfs_tree_unlock(eb); - free_extent_buffer(eb); out: - lock_page(page); + if (!trylock_page(page)) { + flush_fn(data); + lock_page(page); + } return 0; } -static void btrfs_check_super_valid(struct btrfs_fs_info *fs_info, +static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, int read_only) { + if (btrfs_super_csum_type(fs_info->super_copy) >= ARRAY_SIZE(btrfs_csum_sizes)) { + printk(KERN_ERR "btrfs: unsupported checksum algorithm\n"); + return -EINVAL; + } + if (read_only) - return; + return 0; - if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) - printk(KERN_WARNING "warning: mount fs with errors, " - "running btrfsck is recommended\n"); + return 0; } -int btrfs_error_commit_super(struct btrfs_root *root) +void btrfs_error_commit_super(struct btrfs_root *root) { - int ret; - mutex_lock(&root->fs_info->cleaner_mutex); btrfs_run_delayed_iputs(root); mutex_unlock(&root->fs_info->cleaner_mutex); @@ -2795,13 +3431,9 @@ int btrfs_error_commit_super(struct btrfs_root *root) /* cleanup FS via transaction */ btrfs_cleanup_transaction(root); - - ret = write_ctree_super(NULL, root, 0); - - return ret; } -static int btrfs_destroy_ordered_operations(struct btrfs_root *root) +static void btrfs_destroy_ordered_operations(struct btrfs_root *root) { struct btrfs_inode *btrfs_inode; struct list_head splice; @@ -2823,11 +3455,9 @@ static int btrfs_destroy_ordered_operations(struct btrfs_root *root) spin_unlock(&root->fs_info->ordered_extent_lock); mutex_unlock(&root->fs_info->ordered_operations_mutex); - - return 0; } -static int btrfs_destroy_ordered_extents(struct btrfs_root *root) +static void btrfs_destroy_ordered_extents(struct btrfs_root *root) { struct list_head splice; struct btrfs_ordered_extent *ordered; @@ -2859,12 +3489,10 @@ static int btrfs_destroy_ordered_extents(struct btrfs_root *root) } spin_unlock(&root->fs_info->ordered_extent_lock); - - return 0; } -static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, - struct btrfs_root *root) +int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, + struct btrfs_root *root) { struct rb_node *node; struct btrfs_delayed_ref_root *delayed_refs; @@ -2880,28 +3508,36 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, return ret; } - node = rb_first(&delayed_refs->root); - while (node) { + while ((node = rb_first(&delayed_refs->root)) != NULL) { ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node); - node = rb_next(node); - - ref->in_tree = 0; - rb_erase(&ref->rb_node, &delayed_refs->root); - delayed_refs->num_entries--; atomic_set(&ref->refs, 1); if (btrfs_delayed_ref_is_head(ref)) { struct btrfs_delayed_ref_head *head; head = btrfs_delayed_node_to_head(ref); - mutex_lock(&head->mutex); + if (!mutex_trylock(&head->mutex)) { + atomic_inc(&ref->refs); + spin_unlock(&delayed_refs->lock); + + /* Need to wait for the delayed ref to run */ + mutex_lock(&head->mutex); + mutex_unlock(&head->mutex); + btrfs_put_delayed_ref(ref); + + spin_lock(&delayed_refs->lock); + continue; + } + kfree(head->extent_op); delayed_refs->num_heads--; if (list_empty(&head->cluster)) delayed_refs->num_heads_ready--; list_del_init(&head->cluster); - mutex_unlock(&head->mutex); } + ref->in_tree = 0; + rb_erase(&ref->rb_node, &delayed_refs->root); + delayed_refs->num_entries--; spin_unlock(&delayed_refs->lock); btrfs_put_delayed_ref(ref); @@ -2915,7 +3551,7 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans, return ret; } -static int btrfs_destroy_pending_snapshots(struct btrfs_transaction *t) +static void btrfs_destroy_pending_snapshots(struct btrfs_transaction *t) { struct btrfs_pending_snapshot *snapshot; struct list_head splice; @@ -2933,11 +3569,9 @@ static int btrfs_destroy_pending_snapshots(struct btrfs_transaction *t) kfree(snapshot); } - - return 0; } -static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root) +static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root) { struct btrfs_inode *btrfs_inode; struct list_head splice; @@ -2957,8 +3591,6 @@ static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root) } spin_unlock(&root->fs_info->delalloc_lock); - - return 0; } static int btrfs_destroy_marked_extents(struct btrfs_root *root, @@ -2994,11 +3626,9 @@ static int btrfs_destroy_marked_extents(struct btrfs_root *root, &(&BTRFS_I(page->mapping->host)->io_tree)->buffer, offset >> PAGE_CACHE_SHIFT); spin_unlock(&dirty_pages->buffer_lock); - if (eb) { + if (eb) ret = test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags); - atomic_set(&eb->refs, 1); - } if (PageWriteback(page)) end_page_writeback(page); @@ -3012,8 +3642,8 @@ static int btrfs_destroy_marked_extents(struct btrfs_root *root, spin_unlock_irq(&page->mapping->tree_lock); } - page->mapping->a_ops->invalidatepage(page, 0); unlock_page(page); + page_cache_release(page); } } @@ -3027,8 +3657,10 @@ static int btrfs_destroy_pinned_extent(struct btrfs_root *root, u64 start; u64 end; int ret; + bool loop = true; unpin = pinned_extents; +again: while (1) { ret = find_first_extent_bit(unpin, 0, &start, &end, EXTENT_DIRTY); @@ -3046,16 +3678,57 @@ static int btrfs_destroy_pinned_extent(struct btrfs_root *root, cond_resched(); } + if (loop) { + if (unpin == &root->fs_info->freed_extents[0]) + unpin = &root->fs_info->freed_extents[1]; + else + unpin = &root->fs_info->freed_extents[0]; + loop = false; + goto again; + } + return 0; } -static int btrfs_cleanup_transaction(struct btrfs_root *root) +void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans, + struct btrfs_root *root) +{ + btrfs_destroy_delayed_refs(cur_trans, root); + btrfs_block_rsv_release(root, &root->fs_info->trans_block_rsv, + cur_trans->dirty_pages.dirty_bytes); + + /* FIXME: cleanup wait for commit */ + cur_trans->in_commit = 1; + cur_trans->blocked = 1; + wake_up(&root->fs_info->transaction_blocked_wait); + + cur_trans->blocked = 0; + wake_up(&root->fs_info->transaction_wait); + + cur_trans->commit_done = 1; + wake_up(&cur_trans->commit_wait); + + btrfs_destroy_delayed_inodes(root); + btrfs_assert_delayed_root_empty(root); + + btrfs_destroy_pending_snapshots(cur_trans); + + btrfs_destroy_marked_extents(root, &cur_trans->dirty_pages, + EXTENT_DIRTY); + btrfs_destroy_pinned_extent(root, + root->fs_info->pinned_extents); + + /* + memset(cur_trans, 0, sizeof(*cur_trans)); + kmem_cache_free(btrfs_transaction_cachep, cur_trans); + */ +} + +int btrfs_cleanup_transaction(struct btrfs_root *root) { struct btrfs_transaction *t; LIST_HEAD(list); - WARN_ON(1); - mutex_lock(&root->fs_info->transaction_kthread_mutex); spin_lock(&root->fs_info->trans_lock); @@ -3081,17 +3754,23 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) /* FIXME: cleanup wait for commit */ t->in_commit = 1; t->blocked = 1; + smp_mb(); if (waitqueue_active(&root->fs_info->transaction_blocked_wait)) wake_up(&root->fs_info->transaction_blocked_wait); t->blocked = 0; + smp_mb(); if (waitqueue_active(&root->fs_info->transaction_wait)) wake_up(&root->fs_info->transaction_wait); t->commit_done = 1; + smp_mb(); if (waitqueue_active(&t->commit_wait)) wake_up(&t->commit_wait); + btrfs_destroy_delayed_inodes(root); + btrfs_assert_delayed_root_empty(root); + btrfs_destroy_pending_snapshots(t); btrfs_destroy_delalloc_inodes(root); @@ -3123,6 +3802,7 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) static struct extent_io_ops btree_extent_io_ops = { .write_cache_pages_lock_hook = btree_lock_page_hook, .readpage_end_io_hook = btree_readpage_end_io_hook, + .readpage_io_failed_hook = btree_io_failed_hook, .submit_bio_hook = btree_submit_bio_hook, /* note we're sharing with inode.c for the merge bio hook */ .merge_bio_hook = btrfs_merge_bio_hook, diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index bec3ea4bd67f..c5b00a735fef 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -40,19 +40,21 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, u64 parent_transid); int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, u64 parent_transid); +int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize, + int mirror_num, struct extent_buffer **eb); struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize); -int clean_tree_block(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct extent_buffer *buf); -struct btrfs_root *open_ctree(struct super_block *sb, - struct btrfs_fs_devices *fs_devices, - char *options); +void clean_tree_block(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct extent_buffer *buf); +int open_ctree(struct super_block *sb, + struct btrfs_fs_devices *fs_devices, + char *options); int close_ctree(struct btrfs_root *root); int write_ctree_super(struct btrfs_trans_handle *trans, struct btrfs_root *root, int max_mirrors); struct buffer_head *btrfs_read_dev_super(struct block_device *bdev); int btrfs_commit_super(struct btrfs_root *root); -int btrfs_error_commit_super(struct btrfs_root *root); +void btrfs_error_commit_super(struct btrfs_root *root); struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize); struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, @@ -62,9 +64,10 @@ struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info, int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info); void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); -int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root); +void btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root); void btrfs_mark_buffer_dirty(struct extent_buffer *buf); -int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid); +int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid, + int atomic); int btrfs_set_buffer_uptodate(struct extent_buffer *buf); int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid); u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len); @@ -83,8 +86,15 @@ int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info); int btrfs_add_log_tree(struct btrfs_trans_handle *trans, struct btrfs_root *root); -int btree_lock_page_hook(struct page *page); - +int btrfs_cleanup_transaction(struct btrfs_root *root); +void btrfs_cleanup_one_transaction(struct btrfs_transaction *trans, + struct btrfs_root *root); +void btrfs_abort_devices(struct btrfs_root *root); +struct btrfs_root *btrfs_create_tree(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, + u64 objectid); +int btree_lock_page_hook(struct page *page, void *data, + void (*flush_fn)(void *)); #ifdef CONFIG_DEBUG_LOCK_ALLOC void btrfs_init_lockdep(void); diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c index 1b8dc33778f9..614f34a899c2 100644 --- a/fs/btrfs/export.c +++ b/fs/btrfs/export.c @@ -13,15 +13,14 @@ parent_root_objectid) / 4) #define BTRFS_FID_SIZE_CONNECTABLE_ROOT (sizeof(struct btrfs_fid) / 4) -static int btrfs_encode_fh(struct dentry *dentry, u32 *fh, int *max_len, - int connectable) +static int btrfs_encode_fh(struct inode *inode, u32 *fh, int *max_len, + struct inode *parent) { struct btrfs_fid *fid = (struct btrfs_fid *)fh; - struct inode *inode = dentry->d_inode; int len = *max_len; int type; - if (connectable && (len < BTRFS_FID_SIZE_CONNECTABLE)) { + if (parent && (len < BTRFS_FID_SIZE_CONNECTABLE)) { *max_len = BTRFS_FID_SIZE_CONNECTABLE; return 255; } else if (len < BTRFS_FID_SIZE_NON_CONNECTABLE) { @@ -36,19 +35,13 @@ static int btrfs_encode_fh(struct dentry *dentry, u32 *fh, int *max_len, fid->root_objectid = BTRFS_I(inode)->root->objectid; fid->gen = inode->i_generation; - if (connectable && !S_ISDIR(inode->i_mode)) { - struct inode *parent; + if (parent) { u64 parent_root_id; - spin_lock(&dentry->d_lock); - - parent = dentry->d_parent->d_inode; fid->parent_objectid = BTRFS_I(parent)->location.objectid; fid->parent_gen = parent->i_generation; parent_root_id = BTRFS_I(parent)->root->objectid; - spin_unlock(&dentry->d_lock); - if (parent_root_id != fid->root_objectid) { fid->parent_root_objectid = parent_root_id; len = BTRFS_FID_SIZE_CONNECTABLE_ROOT; @@ -67,7 +60,7 @@ static struct dentry *btrfs_get_dentry(struct super_block *sb, u64 objectid, u64 root_objectid, u32 generation, int check_generation) { - struct btrfs_fs_info *fs_info = btrfs_sb(sb)->fs_info; + struct btrfs_fs_info *fs_info = btrfs_sb(sb); struct btrfs_root *root; struct inode *inode; struct btrfs_key key; @@ -193,7 +186,7 @@ static struct dentry *btrfs_get_parent(struct dentry *child) if (ret < 0) goto fail; - BUG_ON(ret == 0); + BUG_ON(ret == 0); /* Key with offset of -1 found */ if (path->slots[0] == 0) { ret = -ENOENT; goto fail; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index f5be06a2462f..ba58024d40d3 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -23,6 +23,7 @@ #include <linux/rcupdate.h> #include <linux/kthread.h> #include <linux/slab.h> +#include <linux/ratelimit.h> #include "compat.h" #include "hash.h" #include "ctree.h" @@ -33,23 +34,41 @@ #include "locking.h" #include "free-space-cache.h" -/* control flags for do_chunk_alloc's force field +#undef SCRAMBLE_DELAYED_REFS + +/* + * control flags for do_chunk_alloc's force field * CHUNK_ALLOC_NO_FORCE means to only allocate a chunk * if we really need one. * - * CHUNK_ALLOC_FORCE means it must try to allocate one - * * CHUNK_ALLOC_LIMITED means to only try and allocate one * if we have very few chunks already allocated. This is * used as part of the clustering code to help make sure * we have a good pool of storage to cluster in, without * filling the FS with empty chunks * + * CHUNK_ALLOC_FORCE means it must try to allocate one + * */ enum { CHUNK_ALLOC_NO_FORCE = 0, - CHUNK_ALLOC_FORCE = 1, - CHUNK_ALLOC_LIMITED = 2, + CHUNK_ALLOC_LIMITED = 1, + CHUNK_ALLOC_FORCE = 2, +}; + +/* + * Control how reservations are dealt with. + * + * RESERVE_FREE - freeing a reservation. + * RESERVE_ALLOC - allocating space and we need to update bytes_may_use for + * ENOSPC accounting + * RESERVE_ALLOC_NO_ACCOUNT - allocating space and we should not update + * bytes_may_use as the ENOSPC accounting is done elsewhere + */ +enum { + RESERVE_FREE = 0, + RESERVE_ALLOC = 1, + RESERVE_ALLOC_NO_ACCOUNT = 2, }; static int update_block_group(struct btrfs_trans_handle *trans, @@ -81,6 +100,8 @@ static int find_next_key(struct btrfs_path *path, int level, struct btrfs_key *key); static void dump_space_info(struct btrfs_space_info *info, u64 bytes, int dump_block_groups); +static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve); static noinline int block_group_cache_done(struct btrfs_block_group_cache *cache) @@ -104,7 +125,6 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache) if (atomic_dec_and_test(&cache->count)) { WARN_ON(cache->pinned > 0); WARN_ON(cache->reserved > 0); - WARN_ON(cache->reserved_pinned > 0); kfree(cache->free_space_ctl); kfree(cache); } @@ -227,7 +247,7 @@ static int exclude_super_stripes(struct btrfs_root *root, cache->bytes_super += stripe_len; ret = add_excluded_extent(root, cache->key.objectid, stripe_len); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { @@ -235,13 +255,13 @@ static int exclude_super_stripes(struct btrfs_root *root, ret = btrfs_rmap_block(&root->fs_info->mapping_tree, cache->key.objectid, bytenr, 0, &logical, &nr, &stripe_len); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ while (nr--) { cache->bytes_super += stripe_len; ret = add_excluded_extent(root, logical[nr], stripe_len); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } kfree(logical); @@ -303,7 +323,7 @@ static u64 add_new_free_space(struct btrfs_block_group_cache *block_group, total_added += size; ret = btrfs_add_free_space(block_group, start, size); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM or logic error */ start = extent_end + 1; } else { break; @@ -314,7 +334,7 @@ static u64 add_new_free_space(struct btrfs_block_group_cache *block_group, size = end - start; total_added += size; ret = btrfs_add_free_space(block_group, start, size); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM or logic error */ } return total_added; @@ -450,13 +470,60 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, struct btrfs_root *root, int load_cache_only) { + DEFINE_WAIT(wait); struct btrfs_fs_info *fs_info = cache->fs_info; struct btrfs_caching_control *caching_ctl; int ret = 0; - smp_mb(); - if (cache->cached != BTRFS_CACHE_NO) + caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); + if (!caching_ctl) + return -ENOMEM; + + INIT_LIST_HEAD(&caching_ctl->list); + mutex_init(&caching_ctl->mutex); + init_waitqueue_head(&caching_ctl->wait); + caching_ctl->block_group = cache; + caching_ctl->progress = cache->key.objectid; + atomic_set(&caching_ctl->count, 1); + caching_ctl->work.func = caching_thread; + + spin_lock(&cache->lock); + /* + * This should be a rare occasion, but this could happen I think in the + * case where one thread starts to load the space cache info, and then + * some other thread starts a transaction commit which tries to do an + * allocation while the other thread is still loading the space cache + * info. The previous loop should have kept us from choosing this block + * group, but if we've moved to the state where we will wait on caching + * block groups we need to first check if we're doing a fast load here, + * so we can wait for it to finish, otherwise we could end up allocating + * from a block group who's cache gets evicted for one reason or + * another. + */ + while (cache->cached == BTRFS_CACHE_FAST) { + struct btrfs_caching_control *ctl; + + ctl = cache->caching_ctl; + atomic_inc(&ctl->count); + prepare_to_wait(&ctl->wait, &wait, TASK_UNINTERRUPTIBLE); + spin_unlock(&cache->lock); + + schedule(); + + finish_wait(&ctl->wait, &wait); + put_caching_control(ctl); + spin_lock(&cache->lock); + } + + if (cache->cached != BTRFS_CACHE_NO) { + spin_unlock(&cache->lock); + kfree(caching_ctl); return 0; + } + WARN_ON(cache->caching_ctl); + cache->caching_ctl = caching_ctl; + cache->cached = BTRFS_CACHE_FAST; + spin_unlock(&cache->lock); /* * We can't do the read from on-disk cache during a commit since we need @@ -464,58 +531,52 @@ static int cache_block_group(struct btrfs_block_group_cache *cache, * allocate blocks for the tree root we can't do the fast caching since * we likely hold important locks. */ - if (trans && (!trans->transaction->in_commit) && - (root && root != root->fs_info->tree_root)) { - spin_lock(&cache->lock); - if (cache->cached != BTRFS_CACHE_NO) { - spin_unlock(&cache->lock); - return 0; - } - cache->cached = BTRFS_CACHE_STARTED; - spin_unlock(&cache->lock); - + if (fs_info->mount_opt & BTRFS_MOUNT_SPACE_CACHE) { ret = load_free_space_cache(fs_info, cache); spin_lock(&cache->lock); if (ret == 1) { + cache->caching_ctl = NULL; cache->cached = BTRFS_CACHE_FINISHED; cache->last_byte_to_unpin = (u64)-1; } else { - cache->cached = BTRFS_CACHE_NO; + if (load_cache_only) { + cache->caching_ctl = NULL; + cache->cached = BTRFS_CACHE_NO; + } else { + cache->cached = BTRFS_CACHE_STARTED; + } } spin_unlock(&cache->lock); + wake_up(&caching_ctl->wait); if (ret == 1) { + put_caching_control(caching_ctl); free_excluded_extents(fs_info->extent_root, cache); return 0; } + } else { + /* + * We are not going to do the fast caching, set cached to the + * appropriate value and wakeup any waiters. + */ + spin_lock(&cache->lock); + if (load_cache_only) { + cache->caching_ctl = NULL; + cache->cached = BTRFS_CACHE_NO; + } else { + cache->cached = BTRFS_CACHE_STARTED; + } + spin_unlock(&cache->lock); + wake_up(&caching_ctl->wait); } - if (load_cache_only) - return 0; - - caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); - BUG_ON(!caching_ctl); - - INIT_LIST_HEAD(&caching_ctl->list); - mutex_init(&caching_ctl->mutex); - init_waitqueue_head(&caching_ctl->wait); - caching_ctl->block_group = cache; - caching_ctl->progress = cache->key.objectid; - /* one for caching kthread, one for caching block group list */ - atomic_set(&caching_ctl->count, 2); - caching_ctl->work.func = caching_thread; - - spin_lock(&cache->lock); - if (cache->cached != BTRFS_CACHE_NO) { - spin_unlock(&cache->lock); - kfree(caching_ctl); + if (load_cache_only) { + put_caching_control(caching_ctl); return 0; } - cache->caching_ctl = caching_ctl; - cache->cached = BTRFS_CACHE_STARTED; - spin_unlock(&cache->lock); down_write(&fs_info->extent_commit_sem); + atomic_inc(&caching_ctl->count); list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups); up_write(&fs_info->extent_commit_sem); @@ -559,8 +620,7 @@ static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info, struct list_head *head = &info->space_info; struct btrfs_space_info *found; - flags &= BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_SYSTEM | - BTRFS_BLOCK_GROUP_METADATA; + flags &= BTRFS_BLOCK_GROUP_TYPE_MASK; rcu_read_lock(); list_for_each_entry_rcu(found, head, list) { @@ -923,7 +983,7 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans, ret = btrfs_next_leaf(root, path); if (ret < 0) return ret; - BUG_ON(ret > 0); + BUG_ON(ret > 0); /* Corruption */ leaf = path->nodes[0]; } btrfs_item_key_to_cpu(leaf, &found_key, @@ -949,9 +1009,9 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans, new_size + extra_size, 1); if (ret < 0) return ret; - BUG_ON(ret); + BUG_ON(ret); /* Corruption */ - ret = btrfs_extend_item(trans, root, path, new_size); + btrfs_extend_item(trans, root, path, new_size); leaf = path->nodes[0]; item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); @@ -1419,7 +1479,11 @@ int lookup_inline_extent_backref(struct btrfs_trans_handle *trans, err = ret; goto out; } - BUG_ON(ret); + if (ret && !insert) { + err = -ENOENT; + goto out; + } + BUG_ON(ret); /* Corruption */ leaf = path->nodes[0]; item_size = btrfs_item_size_nr(leaf, path->slots[0]); @@ -1533,13 +1597,13 @@ out: * helper to add new inline back ref */ static noinline_for_stack -int setup_inline_extent_backref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_extent_inline_ref *iref, - u64 parent, u64 root_objectid, - u64 owner, u64 offset, int refs_to_add, - struct btrfs_delayed_extent_op *extent_op) +void setup_inline_extent_backref(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_extent_inline_ref *iref, + u64 parent, u64 root_objectid, + u64 owner, u64 offset, int refs_to_add, + struct btrfs_delayed_extent_op *extent_op) { struct extent_buffer *leaf; struct btrfs_extent_item *ei; @@ -1549,7 +1613,6 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans, u64 refs; int size; int type; - int ret; leaf = path->nodes[0]; ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); @@ -1558,7 +1621,7 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans, type = extent_ref_type(parent, owner); size = btrfs_extent_inline_ref_size(type); - ret = btrfs_extend_item(trans, root, path, size); + btrfs_extend_item(trans, root, path, size); ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); refs = btrfs_extent_refs(leaf, ei); @@ -1593,7 +1656,6 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans, btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid); } btrfs_mark_buffer_dirty(leaf); - return 0; } static int lookup_extent_backref(struct btrfs_trans_handle *trans, @@ -1628,12 +1690,12 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans, * helper to update/remove inline back ref */ static noinline_for_stack -int update_inline_extent_backref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_extent_inline_ref *iref, - int refs_to_mod, - struct btrfs_delayed_extent_op *extent_op) +void update_inline_extent_backref(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_extent_inline_ref *iref, + int refs_to_mod, + struct btrfs_delayed_extent_op *extent_op) { struct extent_buffer *leaf; struct btrfs_extent_item *ei; @@ -1644,7 +1706,6 @@ int update_inline_extent_backref(struct btrfs_trans_handle *trans, u32 item_size; int size; int type; - int ret; u64 refs; leaf = path->nodes[0]; @@ -1686,10 +1747,9 @@ int update_inline_extent_backref(struct btrfs_trans_handle *trans, memmove_extent_buffer(leaf, ptr, ptr + size, end - ptr - size); item_size -= size; - ret = btrfs_truncate_item(trans, root, path, item_size, 1); + btrfs_truncate_item(trans, root, path, item_size, 1); } btrfs_mark_buffer_dirty(leaf); - return 0; } static noinline_for_stack @@ -1709,13 +1769,13 @@ int insert_inline_extent_backref(struct btrfs_trans_handle *trans, root_objectid, owner, offset, 1); if (ret == 0) { BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID); - ret = update_inline_extent_backref(trans, root, path, iref, - refs_to_add, extent_op); + update_inline_extent_backref(trans, root, path, iref, + refs_to_add, extent_op); } else if (ret == -ENOENT) { - ret = setup_inline_extent_backref(trans, root, path, iref, - parent, root_objectid, - owner, offset, refs_to_add, - extent_op); + setup_inline_extent_backref(trans, root, path, iref, parent, + root_objectid, owner, offset, + refs_to_add, extent_op); + ret = 0; } return ret; } @@ -1745,12 +1805,12 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, struct btrfs_extent_inline_ref *iref, int refs_to_drop, int is_data) { - int ret; + int ret = 0; BUG_ON(!is_data && refs_to_drop != 1); if (iref) { - ret = update_inline_extent_backref(trans, root, path, iref, - -refs_to_drop, NULL); + update_inline_extent_backref(trans, root, path, iref, + -refs_to_drop, NULL); } else if (is_data) { ret = remove_extent_data_ref(trans, root, path, refs_to_drop); } else { @@ -1770,18 +1830,19 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, { int ret; u64 discarded_bytes = 0; - struct btrfs_multi_bio *multi = NULL; + struct btrfs_bio *bbio = NULL; /* Tell the block device(s) that the sectors can be discarded */ ret = btrfs_map_block(&root->fs_info->mapping_tree, REQ_DISCARD, - bytenr, &num_bytes, &multi, 0); + bytenr, &num_bytes, &bbio, 0); + /* Error condition is -ENOMEM */ if (!ret) { - struct btrfs_bio_stripe *stripe = multi->stripes; + struct btrfs_bio_stripe *stripe = bbio->stripes; int i; - for (i = 0; i < multi->num_stripes; i++, stripe++) { + for (i = 0; i < bbio->num_stripes; i++, stripe++) { if (!stripe->dev->can_discard) continue; @@ -1791,7 +1852,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, if (!ret) discarded_bytes += stripe->length; else if (ret != -EOPNOTSUPP) - break; + break; /* Logic errors or -ENOMEM, or -EIO but I don't know how that could happen JDM */ /* * Just in case we get back EOPNOTSUPP for some reason, @@ -1800,7 +1861,7 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, */ ret = 0; } - kfree(multi); + kfree(bbio); } if (actual_bytes) @@ -1810,23 +1871,28 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, return ret; } +/* Can return -ENOMEM */ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 bytenr, u64 num_bytes, u64 parent, - u64 root_objectid, u64 owner, u64 offset) + u64 root_objectid, u64 owner, u64 offset, int for_cow) { int ret; + struct btrfs_fs_info *fs_info = root->fs_info; + BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID && root_objectid == BTRFS_TREE_LOG_OBJECTID); if (owner < BTRFS_FIRST_FREE_OBJECTID) { - ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes, + ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, + num_bytes, parent, root_objectid, (int)owner, - BTRFS_ADD_DELAYED_REF, NULL); + BTRFS_ADD_DELAYED_REF, NULL, for_cow); } else { - ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes, + ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, + num_bytes, parent, root_objectid, owner, offset, - BTRFS_ADD_DELAYED_REF, NULL); + BTRFS_ADD_DELAYED_REF, NULL, for_cow); } return ret; } @@ -1881,7 +1947,8 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, ret = insert_extent_backref(trans, root->fs_info->extent_root, path, bytenr, parent, root_objectid, owner, offset, refs_to_add); - BUG_ON(ret); + if (ret) + btrfs_abort_transaction(trans, root, ret); out: btrfs_free_path(path); return err; @@ -1968,6 +2035,9 @@ static int run_delayed_extent_op(struct btrfs_trans_handle *trans, int ret; int err = 0; + if (trans->aborted) + return 0; + path = btrfs_alloc_path(); if (!path) return -ENOMEM; @@ -2065,7 +2135,11 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans, struct btrfs_delayed_extent_op *extent_op, int insert_reserved) { - int ret; + int ret = 0; + + if (trans->aborted) + return 0; + if (btrfs_delayed_ref_is_head(node)) { struct btrfs_delayed_ref_head *head; /* @@ -2083,11 +2157,10 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans, ret = btrfs_del_csums(trans, root, node->bytenr, node->num_bytes); - BUG_ON(ret); } } mutex_unlock(&head->mutex); - return 0; + return ret; } if (node->type == BTRFS_TREE_BLOCK_REF_KEY || @@ -2134,6 +2207,10 @@ again: return NULL; } +/* + * Returns 0 on success or if called with an already aborted transaction. + * Returns -ENOMEM or -EIO on failure and will abort the transaction. + */ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct list_head *cluster) @@ -2142,6 +2219,7 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_node *ref; struct btrfs_delayed_ref_head *locked_ref = NULL; struct btrfs_delayed_extent_op *extent_op; + struct btrfs_fs_info *fs_info = root->fs_info; int ret; int count = 0; int must_insert_reserved = 0; @@ -2174,6 +2252,38 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans, } /* + * We need to try and merge add/drops of the same ref since we + * can run into issues with relocate dropping the implicit ref + * and then it being added back again before the drop can + * finish. If we merged anything we need to re-loop so we can + * get a good ref. + */ + btrfs_merge_delayed_refs(trans, fs_info, delayed_refs, + locked_ref); + + /* + * locked_ref is the head node, so we have to go one + * node back for any delayed ref updates + */ + ref = select_delayed_ref(locked_ref); + + if (ref && ref->seq && + btrfs_check_delayed_seq(fs_info, delayed_refs, ref->seq)) { + /* + * there are still refs with lower seq numbers in the + * process of being added. Don't run this ref yet. + */ + list_del_init(&locked_ref->cluster); + mutex_unlock(&locked_ref->mutex); + locked_ref = NULL; + delayed_refs->num_heads_ready++; + spin_unlock(&delayed_refs->lock); + cond_resched(); + spin_lock(&delayed_refs->lock); + continue; + } + + /* * record the must insert reserved flag before we * drop the spin lock. */ @@ -2183,11 +2293,6 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans, extent_op = locked_ref->extent_op; locked_ref->extent_op = NULL; - /* - * locked_ref is the head node, so we have to go one - * node back for any delayed ref updates - */ - ref = select_delayed_ref(locked_ref); if (!ref) { /* All delayed refs have been processed, Go ahead * and send the head node to run_one_delayed_ref, @@ -2205,12 +2310,15 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans, ret = run_delayed_extent_op(trans, root, ref, extent_op); - BUG_ON(ret); kfree(extent_op); - cond_resched(); - spin_lock(&delayed_refs->lock); - continue; + if (ret) { + printk(KERN_DEBUG "btrfs: run_delayed_extent_op returned %d\n", ret); + spin_lock(&delayed_refs->lock); + return ret; + } + + goto next; } list_del_init(&locked_ref->cluster); @@ -2220,29 +2328,135 @@ static noinline int run_clustered_refs(struct btrfs_trans_handle *trans, ref->in_tree = 0; rb_erase(&ref->rb_node, &delayed_refs->root); delayed_refs->num_entries--; - + if (locked_ref) { + /* + * when we play the delayed ref, also correct the + * ref_mod on head + */ + switch (ref->action) { + case BTRFS_ADD_DELAYED_REF: + case BTRFS_ADD_DELAYED_EXTENT: + locked_ref->node.ref_mod -= ref->ref_mod; + break; + case BTRFS_DROP_DELAYED_REF: + locked_ref->node.ref_mod += ref->ref_mod; + break; + default: + WARN_ON(1); + } + } spin_unlock(&delayed_refs->lock); ret = run_one_delayed_ref(trans, root, ref, extent_op, must_insert_reserved); - BUG_ON(ret); btrfs_put_delayed_ref(ref); kfree(extent_op); count++; + if (ret) { + printk(KERN_DEBUG "btrfs: run_one_delayed_ref returned %d\n", ret); + spin_lock(&delayed_refs->lock); + return ret; + } + +next: + do_chunk_alloc(trans, fs_info->extent_root, + 2 * 1024 * 1024, + btrfs_get_alloc_profile(root, 0), + CHUNK_ALLOC_NO_FORCE); cond_resched(); spin_lock(&delayed_refs->lock); } return count; } +#ifdef SCRAMBLE_DELAYED_REFS +/* + * Normally delayed refs get processed in ascending bytenr order. This + * correlates in most cases to the order added. To expose dependencies on this + * order, we start to process the tree in the middle instead of the beginning + */ +static u64 find_middle(struct rb_root *root) +{ + struct rb_node *n = root->rb_node; + struct btrfs_delayed_ref_node *entry; + int alt = 1; + u64 middle; + u64 first = 0, last = 0; + + n = rb_first(root); + if (n) { + entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); + first = entry->bytenr; + } + n = rb_last(root); + if (n) { + entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); + last = entry->bytenr; + } + n = root->rb_node; + + while (n) { + entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); + WARN_ON(!entry->in_tree); + + middle = entry->bytenr; + + if (alt) + n = n->rb_left; + else + n = n->rb_right; + + alt = 1 - alt; + } + return middle; +} +#endif + +int btrfs_delayed_refs_qgroup_accounting(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info) +{ + struct qgroup_update *qgroup_update; + int ret = 0; + + if (list_empty(&trans->qgroup_ref_list) != + !trans->delayed_ref_elem.seq) { + /* list without seq or seq without list */ + printk(KERN_ERR "btrfs: qgroup accounting update error, list is%s empty, seq is %llu\n", + list_empty(&trans->qgroup_ref_list) ? "" : " not", + trans->delayed_ref_elem.seq); + BUG(); + } + + if (!trans->delayed_ref_elem.seq) + return 0; + + while (!list_empty(&trans->qgroup_ref_list)) { + qgroup_update = list_first_entry(&trans->qgroup_ref_list, + struct qgroup_update, list); + list_del(&qgroup_update->list); + if (!ret) + ret = btrfs_qgroup_account_ref( + trans, fs_info, qgroup_update->node, + qgroup_update->extent_op); + kfree(qgroup_update); + } + + btrfs_put_tree_mod_seq(fs_info, &trans->delayed_ref_elem); + + return ret; +} + /* * this starts processing the delayed reference count updates and * extent insertions we have queued up so far. count can be * 0, which means to process everything in the tree at the start * of the run (but not newly added entries), or it can be some target * number you'd like to process. + * + * Returns 0 on success or if called with an aborted transaction + * Returns <0 on error and aborts the transaction */ int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count) @@ -2252,16 +2466,34 @@ int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_node *ref; struct list_head cluster; int ret; + u64 delayed_start; int run_all = count == (unsigned long)-1; int run_most = 0; + int loops; + + /* We'll clean this up in btrfs_cleanup_transaction */ + if (trans->aborted) + return 0; if (root == root->fs_info->extent_root) root = root->fs_info->tree_root; + do_chunk_alloc(trans, root->fs_info->extent_root, + 2 * 1024 * 1024, btrfs_get_alloc_profile(root, 0), + CHUNK_ALLOC_NO_FORCE); + + btrfs_delayed_refs_qgroup_accounting(trans, root->fs_info); + delayed_refs = &trans->transaction->delayed_refs; INIT_LIST_HEAD(&cluster); again: + loops = 0; spin_lock(&delayed_refs->lock); + +#ifdef SCRAMBLE_DELAYED_REFS + delayed_refs->run_delayed_start = find_middle(&delayed_refs->root); +#endif + if (count == 0) { count = delayed_refs->num_entries * 2; run_most = 1; @@ -2277,18 +2509,45 @@ again: * of refs to process starting at the first one we are able to * lock */ + delayed_start = delayed_refs->run_delayed_start; ret = btrfs_find_ref_cluster(trans, &cluster, delayed_refs->run_delayed_start); if (ret) break; ret = run_clustered_refs(trans, root, &cluster); - BUG_ON(ret < 0); + if (ret < 0) { + spin_unlock(&delayed_refs->lock); + btrfs_abort_transaction(trans, root, ret); + return ret; + } count -= min_t(unsigned long, ret, count); if (count == 0) break; + + if (delayed_start >= delayed_refs->run_delayed_start) { + if (loops == 0) { + /* + * btrfs_find_ref_cluster looped. let's do one + * more cycle. if we don't run any delayed ref + * during that cycle (because we can't because + * all of them are blocked), bail out. + */ + loops = 1; + } else { + /* + * no runnable refs left, stop trying + */ + BUG_ON(run_all); + break; + } + } + if (ret) { + /* refs were run, let's reset staleness detection */ + loops = 0; + } } if (run_all) { @@ -2326,6 +2585,7 @@ again: } out: spin_unlock(&delayed_refs->lock); + assert_qgroups_uptodate(trans); return 0; } @@ -2346,7 +2606,8 @@ int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans, extent_op->update_key = 0; extent_op->is_data = is_data ? 1 : 0; - ret = btrfs_add_delayed_extent_op(trans, bytenr, num_bytes, extent_op); + ret = btrfs_add_delayed_extent_op(root->fs_info, trans, bytenr, + num_bytes, extent_op); if (ret) kfree(extent_op); return ret; @@ -2404,8 +2665,10 @@ static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, node = rb_prev(node); if (node) { + int seq = ref->seq; + ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node); - if (ref->bytenr == bytenr) + if (ref->bytenr == bytenr && ref->seq == seq) goto out_unlock; } @@ -2442,7 +2705,7 @@ static noinline int check_committed_ref(struct btrfs_trans_handle *trans, ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0); if (ret < 0) goto out; - BUG_ON(ret == 0); + BUG_ON(ret == 0); /* Corruption */ ret = -ENOENT; if (path->slots[0] == 0) @@ -2531,7 +2794,7 @@ out: static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf, - int full_backref, int inc) + int full_backref, int inc, int for_cow) { u64 bytenr; u64 num_bytes; @@ -2544,7 +2807,7 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, int level; int ret = 0; int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *, - u64, u64, u64, u64, u64, u64); + u64, u64, u64, u64, u64, u64, int); ref_root = btrfs_header_owner(buf); nritems = btrfs_header_nritems(buf); @@ -2581,34 +2844,34 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, key.offset -= btrfs_file_extent_offset(buf, fi); ret = process_func(trans, root, bytenr, num_bytes, parent, ref_root, key.objectid, - key.offset); + key.offset, for_cow); if (ret) goto fail; } else { bytenr = btrfs_node_blockptr(buf, i); num_bytes = btrfs_level_size(root, level - 1); ret = process_func(trans, root, bytenr, num_bytes, - parent, ref_root, level - 1, 0); + parent, ref_root, level - 1, 0, + for_cow); if (ret) goto fail; } } return 0; fail: - BUG(); return ret; } int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct extent_buffer *buf, int full_backref) + struct extent_buffer *buf, int full_backref, int for_cow) { - return __btrfs_mod_ref(trans, root, buf, full_backref, 1); + return __btrfs_mod_ref(trans, root, buf, full_backref, 1, for_cow); } int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct extent_buffer *buf, int full_backref) + struct extent_buffer *buf, int full_backref, int for_cow) { - return __btrfs_mod_ref(trans, root, buf, full_backref, 0); + return __btrfs_mod_ref(trans, root, buf, full_backref, 0, for_cow); } static int write_one_cache_group(struct btrfs_trans_handle *trans, @@ -2624,7 +2887,7 @@ static int write_one_cache_group(struct btrfs_trans_handle *trans, ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1); if (ret < 0) goto fail; - BUG_ON(ret); + BUG_ON(ret); /* Corruption */ leaf = path->nodes[0]; bi = btrfs_item_ptr_offset(leaf, path->slots[0]); @@ -2632,8 +2895,10 @@ static int write_one_cache_group(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(leaf); btrfs_release_path(path); fail: - if (ret) + if (ret) { + btrfs_abort_transaction(trans, root, ret); return ret; + } return 0; } @@ -2700,6 +2965,13 @@ again: goto again; } + /* We've already setup this transaction, go ahead and exit */ + if (block_group->cache_generation == trans->transid && + i_size_read(inode)) { + dcs = BTRFS_DC_SETUP; + goto out_put; + } + /* * We want to set the generation to 0, that way if anything goes wrong * from here on out we know not to trust this cache when we load up next @@ -2717,25 +2989,29 @@ again: } spin_lock(&block_group->lock); - if (block_group->cached != BTRFS_CACHE_FINISHED) { - /* We're not cached, don't bother trying to write stuff out */ + if (block_group->cached != BTRFS_CACHE_FINISHED || + !btrfs_test_opt(root, SPACE_CACHE)) { + /* + * don't bother trying to write stuff out _if_ + * a) we're not cached, + * b) we're with nospace_cache mount option. + */ dcs = BTRFS_DC_WRITTEN; spin_unlock(&block_group->lock); goto out_put; } spin_unlock(&block_group->lock); - num_pages = (int)div64_u64(block_group->key.offset, 1024 * 1024 * 1024); + /* + * Try to preallocate enough space based on how big the block group is. + * Keep in mind this has to include any pinned space which could end up + * taking up quite a bit since it's not folded into the other space + * cache. + */ + num_pages = (int)div64_u64(block_group->key.offset, 256 * 1024 * 1024); if (!num_pages) num_pages = 1; - /* - * Just to make absolutely sure we have enough space, we're going to - * preallocate 12 pages worth of space for each block group. In - * practice we ought to use at most 8, but we need extra space so we can - * add our header and have a terminator between the extents and the - * bitmaps. - */ num_pages *= 16; num_pages *= PAGE_CACHE_SIZE; @@ -2749,12 +3025,15 @@ again: if (!ret) dcs = BTRFS_DC_SETUP; btrfs_free_reserved_data_space(inode, num_pages); + out_put: iput(inode); out_free: btrfs_release_path(path); out: spin_lock(&block_group->lock); + if (!ret && dcs == BTRFS_DC_SETUP) + block_group->cache_generation = trans->transid; block_group->disk_cache_state = dcs; spin_unlock(&block_group->lock); @@ -2796,7 +3075,8 @@ again: if (last == 0) { err = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); - BUG_ON(err); + if (err) /* File system offline */ + goto out; } cache = btrfs_lookup_first_block_group(root->fs_info, last); @@ -2823,7 +3103,9 @@ again: last = cache->key.objectid + cache->key.offset; err = write_one_cache_group(trans, root, path, cache); - BUG_ON(err); + if (err) /* File system offline */ + goto out; + btrfs_put_block_group(cache); } @@ -2836,7 +3118,8 @@ again: if (last == 0) { err = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); - BUG_ON(err); + if (err) /* File system offline */ + goto out; } cache = btrfs_lookup_first_block_group(root->fs_info, last); @@ -2861,20 +3144,21 @@ again: continue; } - btrfs_write_out_cache(root, trans, cache, path); + err = btrfs_write_out_cache(root, trans, cache, path); /* * If we didn't have an error then the cache state is still * NEED_WRITE, so we can set it to WRITTEN. */ - if (cache->disk_cache_state == BTRFS_DC_NEED_WRITE) + if (!err && cache->disk_cache_state == BTRFS_DC_NEED_WRITE) cache->disk_cache_state = BTRFS_DC_WRITTEN; last = cache->key.objectid + cache->key.offset; btrfs_put_block_group(cache); } +out: btrfs_free_path(path); - return 0; + return err; } int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr) @@ -2924,9 +3208,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, INIT_LIST_HEAD(&found->block_groups[i]); init_rwsem(&found->groups_sem); spin_lock_init(&found->lock); - found->flags = flags & (BTRFS_BLOCK_GROUP_DATA | - BTRFS_BLOCK_GROUP_SYSTEM | - BTRFS_BLOCK_GROUP_METADATA); + found->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK; found->total_bytes = total_bytes; found->disk_total = total_bytes * factor; found->bytes_used = bytes_used; @@ -2942,25 +3224,59 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, init_waitqueue_head(&found->wait); *space_info = found; list_add_rcu(&found->list, &info->space_info); + if (flags & BTRFS_BLOCK_GROUP_DATA) + info->data_sinfo = found; return 0; } static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags) { - u64 extra_flags = flags & (BTRFS_BLOCK_GROUP_RAID0 | - BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_RAID10 | - BTRFS_BLOCK_GROUP_DUP); - if (extra_flags) { - if (flags & BTRFS_BLOCK_GROUP_DATA) - fs_info->avail_data_alloc_bits |= extra_flags; - if (flags & BTRFS_BLOCK_GROUP_METADATA) - fs_info->avail_metadata_alloc_bits |= extra_flags; - if (flags & BTRFS_BLOCK_GROUP_SYSTEM) - fs_info->avail_system_alloc_bits |= extra_flags; + u64 extra_flags = chunk_to_extended(flags) & + BTRFS_EXTENDED_PROFILE_MASK; + + if (flags & BTRFS_BLOCK_GROUP_DATA) + fs_info->avail_data_alloc_bits |= extra_flags; + if (flags & BTRFS_BLOCK_GROUP_METADATA) + fs_info->avail_metadata_alloc_bits |= extra_flags; + if (flags & BTRFS_BLOCK_GROUP_SYSTEM) + fs_info->avail_system_alloc_bits |= extra_flags; +} + +/* + * returns target flags in extended format or 0 if restripe for this + * chunk_type is not in progress + * + * should be called with either volume_mutex or balance_lock held + */ +static u64 get_restripe_target(struct btrfs_fs_info *fs_info, u64 flags) +{ + struct btrfs_balance_control *bctl = fs_info->balance_ctl; + u64 target = 0; + + if (!bctl) + return 0; + + if (flags & BTRFS_BLOCK_GROUP_DATA && + bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) { + target = BTRFS_BLOCK_GROUP_DATA | bctl->data.target; + } else if (flags & BTRFS_BLOCK_GROUP_SYSTEM && + bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) { + target = BTRFS_BLOCK_GROUP_SYSTEM | bctl->sys.target; + } else if (flags & BTRFS_BLOCK_GROUP_METADATA && + bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) { + target = BTRFS_BLOCK_GROUP_METADATA | bctl->meta.target; } + + return target; } +/* + * @flags: available profiles in extended format (see ctree.h) + * + * Returns reduced profile in chunk format. If profile changing is in + * progress (either running or paused) picks the target profile (if it's + * already available), otherwise falls back to plain reducing. + */ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags) { /* @@ -2970,6 +3286,22 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags) */ u64 num_devices = root->fs_info->fs_devices->rw_devices + root->fs_info->fs_devices->missing_devices; + u64 target; + + /* + * see if restripe for this chunk_type is in progress, if so + * try to reduce to the target profile + */ + spin_lock(&root->fs_info->balance_lock); + target = get_restripe_target(root->fs_info, flags); + if (target) { + /* pick target profile only if it's already available */ + if ((flags & target) & BTRFS_EXTENDED_PROFILE_MASK) { + spin_unlock(&root->fs_info->balance_lock); + return extended_to_chunk(target); + } + } + spin_unlock(&root->fs_info->balance_lock); if (num_devices == 1) flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0); @@ -2990,22 +3322,22 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags) if ((flags & BTRFS_BLOCK_GROUP_RAID0) && ((flags & BTRFS_BLOCK_GROUP_RAID1) | (flags & BTRFS_BLOCK_GROUP_RAID10) | - (flags & BTRFS_BLOCK_GROUP_DUP))) + (flags & BTRFS_BLOCK_GROUP_DUP))) { flags &= ~BTRFS_BLOCK_GROUP_RAID0; - return flags; + } + + return extended_to_chunk(flags); } static u64 get_alloc_profile(struct btrfs_root *root, u64 flags) { if (flags & BTRFS_BLOCK_GROUP_DATA) - flags |= root->fs_info->avail_data_alloc_bits & - root->fs_info->data_alloc_profile; + flags |= root->fs_info->avail_data_alloc_bits; else if (flags & BTRFS_BLOCK_GROUP_SYSTEM) - flags |= root->fs_info->avail_system_alloc_bits & - root->fs_info->system_alloc_profile; + flags |= root->fs_info->avail_system_alloc_bits; else if (flags & BTRFS_BLOCK_GROUP_METADATA) - flags |= root->fs_info->avail_metadata_alloc_bits & - root->fs_info->metadata_alloc_profile; + flags |= root->fs_info->avail_metadata_alloc_bits; + return btrfs_reduce_alloc_profile(root, flags); } @@ -3023,12 +3355,6 @@ u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data) return get_alloc_profile(root, flags); } -void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode) -{ - BTRFS_I(inode)->space_info = __find_space_info(root->fs_info, - BTRFS_BLOCK_GROUP_DATA); -} - /* * This will check the space that the inode allocates from to make sure we have * enough space for bytes. @@ -3037,6 +3363,7 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes) { struct btrfs_space_info *data_sinfo; struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_fs_info *fs_info = root->fs_info; u64 used; int ret = 0, committed = 0, alloc_chunk = 1; @@ -3049,7 +3376,7 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes) committed = 1; } - data_sinfo = BTRFS_I(inode)->space_info; + data_sinfo = fs_info->data_sinfo; if (!data_sinfo) goto alloc; @@ -3090,10 +3417,9 @@ alloc: goto commit_trans; } - if (!data_sinfo) { - btrfs_set_inode_space_info(root, inode); - data_sinfo = BTRFS_I(inode)->space_info; - } + if (!data_sinfo) + data_sinfo = fs_info->data_sinfo; + goto again; } @@ -3122,16 +3448,15 @@ commit_trans: return -ENOSPC; } data_sinfo->bytes_may_use += bytes; - BTRFS_I(inode)->reserved_bytes += bytes; + trace_btrfs_space_reservation(root->fs_info, "space_info", + data_sinfo->flags, bytes, 1); spin_unlock(&data_sinfo->lock); return 0; } /* - * called when we are clearing an delalloc extent from the - * inode's io_tree or there was an error for whatever reason - * after calling btrfs_check_data_free_space + * Called if we need to clear a data reservation for this inode. */ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes) { @@ -3141,10 +3466,11 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes) /* make sure bytes are sectorsize aligned */ bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1); - data_sinfo = BTRFS_I(inode)->space_info; + data_sinfo = root->fs_info->data_sinfo; spin_lock(&data_sinfo->lock); data_sinfo->bytes_may_use -= bytes; - BTRFS_I(inode)->reserved_bytes -= bytes; + trace_btrfs_space_reservation(root->fs_info, "space_info", + data_sinfo->flags, bytes, 0); spin_unlock(&data_sinfo->lock); } @@ -3165,6 +3491,7 @@ static int should_alloc_chunk(struct btrfs_root *root, struct btrfs_space_info *sinfo, u64 alloc_bytes, int force) { + struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly; u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved; u64 thresh; @@ -3173,40 +3500,79 @@ static int should_alloc_chunk(struct btrfs_root *root, return 1; /* + * We need to take into account the global rsv because for all intents + * and purposes it's used space. Don't worry about locking the + * global_rsv, it doesn't change except when the transaction commits. + */ + num_allocated += global_rsv->size; + + /* * in limited mode, we want to have some free space up to * about 1% of the FS size. */ if (force == CHUNK_ALLOC_LIMITED) { - thresh = btrfs_super_total_bytes(&root->fs_info->super_copy); + thresh = btrfs_super_total_bytes(root->fs_info->super_copy); thresh = max_t(u64, 64 * 1024 * 1024, div_factor_fine(thresh, 1)); if (num_bytes - num_allocated < thresh) return 1; } + thresh = btrfs_super_total_bytes(root->fs_info->super_copy); - /* - * we have two similar checks here, one based on percentage - * and once based on a hard number of 256MB. The idea - * is that if we have a good amount of free - * room, don't allocate a chunk. A good mount is - * less than 80% utilized of the chunks we have allocated, - * or more than 256MB free - */ - if (num_allocated + alloc_bytes + 256 * 1024 * 1024 < num_bytes) - return 0; + /* 256MB or 2% of the FS */ + thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 2)); + /* system chunks need a much small threshold */ + if (sinfo->flags & BTRFS_BLOCK_GROUP_SYSTEM) + thresh = 32 * 1024 * 1024; - if (num_allocated + alloc_bytes < div_factor(num_bytes, 8)) + if (num_bytes > thresh && sinfo->bytes_used < div_factor(num_bytes, 8)) return 0; + return 1; +} - thresh = btrfs_super_total_bytes(&root->fs_info->super_copy); +static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type) +{ + u64 num_dev; - /* 256MB or 5% of the FS */ - thresh = max_t(u64, 256 * 1024 * 1024, div_factor_fine(thresh, 5)); + if (type & BTRFS_BLOCK_GROUP_RAID10 || + type & BTRFS_BLOCK_GROUP_RAID0) + num_dev = root->fs_info->fs_devices->rw_devices; + else if (type & BTRFS_BLOCK_GROUP_RAID1) + num_dev = 2; + else + num_dev = 1; /* DUP or single */ - if (num_bytes > thresh && sinfo->bytes_used < div_factor(num_bytes, 3)) - return 0; - return 1; + /* metadata for updaing devices and chunk tree */ + return btrfs_calc_trans_metadata_size(root, num_dev + 1); +} + +static void check_system_chunk(struct btrfs_trans_handle *trans, + struct btrfs_root *root, u64 type) +{ + struct btrfs_space_info *info; + u64 left; + u64 thresh; + + info = __find_space_info(root->fs_info, BTRFS_BLOCK_GROUP_SYSTEM); + spin_lock(&info->lock); + left = info->total_bytes - info->bytes_used - info->bytes_pinned - + info->bytes_reserved - info->bytes_readonly; + spin_unlock(&info->lock); + + thresh = get_system_chunk_thresh(root, type); + if (left < thresh && btrfs_test_opt(root, ENOSPC_DEBUG)) { + printk(KERN_INFO "left=%llu, need=%llu, flags=%llu\n", + left, thresh, type); + dump_space_info(info, 0, 0); + } + + if (left < thresh) { + u64 flags; + + flags = btrfs_get_alloc_profile(root->fs_info->chunk_root, 0); + btrfs_alloc_chunk(trans, root, flags); + } } static int do_chunk_alloc(struct btrfs_trans_handle *trans, @@ -3218,19 +3584,17 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, int wait_for_alloc = 0; int ret = 0; - flags = btrfs_reduce_alloc_profile(extent_root, flags); - space_info = __find_space_info(extent_root->fs_info, flags); if (!space_info) { ret = update_space_info(extent_root->fs_info, flags, 0, 0, &space_info); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } - BUG_ON(!space_info); + BUG_ON(!space_info); /* Logic error */ again: spin_lock(&space_info->lock); - if (space_info->force_alloc) + if (force < space_info->force_alloc) force = space_info->force_alloc; if (space_info->full) { spin_unlock(&space_info->lock); @@ -3281,6 +3645,12 @@ again: force_metadata_allocation(fs_info); } + /* + * Check if we have enough space in SYSTEM chunk because we may need + * to update devices. + */ + check_system_chunk(trans, extent_root, flags); + ret = btrfs_alloc_chunk(trans, extent_root, flags); if (ret < 0 && ret != -ENOSPC) goto out; @@ -3295,114 +3665,201 @@ again: space_info->chunk_alloc = 0; spin_unlock(&space_info->lock); out: - mutex_unlock(&extent_root->fs_info->chunk_mutex); + mutex_unlock(&fs_info->chunk_mutex); return ret; } /* * shrink metadata reservation for delalloc */ -static int shrink_delalloc(struct btrfs_trans_handle *trans, - struct btrfs_root *root, u64 to_reclaim, int sync) +static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, + bool wait_ordered) { struct btrfs_block_rsv *block_rsv; struct btrfs_space_info *space_info; - u64 reserved; + struct btrfs_trans_handle *trans; + u64 delalloc_bytes; u64 max_reclaim; - u64 reclaimed = 0; long time_left; - int nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT; + unsigned long nr_pages = (2 * 1024 * 1024) >> PAGE_CACHE_SHIFT; int loops = 0; - unsigned long progress; + trans = (struct btrfs_trans_handle *)current->journal_info; block_rsv = &root->fs_info->delalloc_block_rsv; space_info = block_rsv->space_info; smp_mb(); - reserved = space_info->bytes_reserved; - progress = space_info->reservation_progress; - - if (reserved == 0) - return 0; - - smp_mb(); - if (root->fs_info->delalloc_bytes == 0) { + delalloc_bytes = root->fs_info->delalloc_bytes; + if (delalloc_bytes == 0) { if (trans) - return 0; + return; btrfs_wait_ordered_extents(root, 0, 0); - return 0; + return; } - max_reclaim = min(reserved, to_reclaim); - - while (loops < 1024) { - /* have the flusher threads jump in and do some IO */ - smp_mb(); - nr_pages = min_t(unsigned long, nr_pages, - root->fs_info->delalloc_bytes >> PAGE_CACHE_SHIFT); - writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages); + while (delalloc_bytes && loops < 3) { + max_reclaim = min(delalloc_bytes, to_reclaim); + nr_pages = max_reclaim >> PAGE_CACHE_SHIFT; + writeback_inodes_sb_nr_if_idle(root->fs_info->sb, nr_pages, + WB_REASON_FS_FREE_SPACE); spin_lock(&space_info->lock); - if (reserved > space_info->bytes_reserved) - reclaimed += reserved - space_info->bytes_reserved; - reserved = space_info->bytes_reserved; + if (space_info->bytes_used + space_info->bytes_reserved + + space_info->bytes_pinned + space_info->bytes_readonly + + space_info->bytes_may_use + orig <= + space_info->total_bytes) { + spin_unlock(&space_info->lock); + break; + } spin_unlock(&space_info->lock); loops++; + if (wait_ordered && !trans) { + btrfs_wait_ordered_extents(root, 0, 0); + } else { + time_left = schedule_timeout_killable(1); + if (time_left) + break; + } + smp_mb(); + delalloc_bytes = root->fs_info->delalloc_bytes; + } +} - if (reserved == 0 || reclaimed >= max_reclaim) - break; - - if (trans && trans->transaction->blocked) - return -EAGAIN; +/** + * maybe_commit_transaction - possibly commit the transaction if its ok to + * @root - the root we're allocating for + * @bytes - the number of bytes we want to reserve + * @force - force the commit + * + * This will check to make sure that committing the transaction will actually + * get us somewhere and then commit the transaction if it does. Otherwise it + * will return -ENOSPC. + */ +static int may_commit_transaction(struct btrfs_root *root, + struct btrfs_space_info *space_info, + u64 bytes, int force) +{ + struct btrfs_block_rsv *delayed_rsv = &root->fs_info->delayed_block_rsv; + struct btrfs_trans_handle *trans; - time_left = schedule_timeout_interruptible(1); + trans = (struct btrfs_trans_handle *)current->journal_info; + if (trans) + return -EAGAIN; - /* We were interrupted, exit */ - if (time_left) - break; + if (force) + goto commit; - /* we've kicked the IO a few times, if anything has been freed, - * exit. There is no sense in looping here for a long time - * when we really need to commit the transaction, or there are - * just too many writers without enough free space - */ + /* See if there is enough pinned space to make this reservation */ + spin_lock(&space_info->lock); + if (space_info->bytes_pinned >= bytes) { + spin_unlock(&space_info->lock); + goto commit; + } + spin_unlock(&space_info->lock); - if (loops > 3) { - smp_mb(); - if (progress != space_info->reservation_progress) - break; - } + /* + * See if there is some space in the delayed insertion reservation for + * this reservation. + */ + if (space_info != delayed_rsv->space_info) + return -ENOSPC; + spin_lock(&space_info->lock); + spin_lock(&delayed_rsv->lock); + if (space_info->bytes_pinned + delayed_rsv->size < bytes) { + spin_unlock(&delayed_rsv->lock); + spin_unlock(&space_info->lock); + return -ENOSPC; } - if (reclaimed >= to_reclaim && !trans) - btrfs_wait_ordered_extents(root, 0, 0); - return reclaimed >= to_reclaim; + spin_unlock(&delayed_rsv->lock); + spin_unlock(&space_info->lock); + +commit: + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) + return -ENOSPC; + + return btrfs_commit_transaction(trans, root); } -/* - * Retries tells us how many times we've called reserve_metadata_bytes. The - * idea is if this is the first call (retries == 0) then we will add to our - * reserved count if we can't make the allocation in order to hold our place - * while we go and try and free up space. That way for retries > 1 we don't try - * and add space, we just check to see if the amount of unused space is >= the - * total space, meaning that our reservation is valid. +enum flush_state { + FLUSH_DELALLOC = 1, + FLUSH_DELALLOC_WAIT = 2, + FLUSH_DELAYED_ITEMS_NR = 3, + FLUSH_DELAYED_ITEMS = 4, + COMMIT_TRANS = 5, +}; + +static int flush_space(struct btrfs_root *root, + struct btrfs_space_info *space_info, u64 num_bytes, + u64 orig_bytes, int state) +{ + struct btrfs_trans_handle *trans; + int nr; + int ret = 0; + + switch (state) { + case FLUSH_DELALLOC: + case FLUSH_DELALLOC_WAIT: + shrink_delalloc(root, num_bytes, orig_bytes, + state == FLUSH_DELALLOC_WAIT); + break; + case FLUSH_DELAYED_ITEMS_NR: + case FLUSH_DELAYED_ITEMS: + if (state == FLUSH_DELAYED_ITEMS_NR) { + u64 bytes = btrfs_calc_trans_metadata_size(root, 1); + + nr = (int)div64_u64(num_bytes, bytes); + if (!nr) + nr = 1; + nr *= 2; + } else { + nr = -1; + } + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + break; + } + ret = btrfs_run_delayed_items_nr(trans, root, nr); + btrfs_end_transaction(trans, root); + break; + case COMMIT_TRANS: + ret = may_commit_transaction(root, space_info, orig_bytes, 0); + break; + default: + ret = -ENOSPC; + break; + } + + return ret; +} +/** + * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space + * @root - the root we're allocating for + * @block_rsv - the block_rsv we're allocating for + * @orig_bytes - the number of bytes we want + * @flush - wether or not we can flush to make our reservation * - * However if we don't intend to retry this reservation, pass -1 as retries so - * that it short circuits this logic. + * This will reserve orgi_bytes number of bytes from the space info associated + * with the block_rsv. If there is not enough space it will make an attempt to + * flush out space to make room. It will do this by flushing delalloc if + * possible or committing the transaction. If flush is 0 then no attempts to + * regain reservations will be made and this will fail if there is not enough + * space already. */ -static int reserve_metadata_bytes(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +static int reserve_metadata_bytes(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 orig_bytes, int flush) { struct btrfs_space_info *space_info = block_rsv->space_info; - u64 unused; + u64 used; u64 num_bytes = orig_bytes; - int retries = 0; + int flush_state = FLUSH_DELALLOC; int ret = 0; - bool committed = false; bool flushing = false; + bool committed = false; again: ret = 0; @@ -3419,11 +3876,10 @@ again: * deadlock since we are waiting for the flusher to finish, but * hold the current transaction open. */ - if (trans) + if (current->journal_info) return -EAGAIN; - ret = wait_event_interruptible(space_info->wait, - !space_info->flush); - /* Must have been interrupted, return */ + ret = wait_event_killable(space_info->wait, !space_info->flush); + /* Must have been killed, return */ if (ret) return -EINTR; @@ -3431,9 +3887,9 @@ again: } ret = -ENOSPC; - unused = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; + used = space_info->bytes_used + space_info->bytes_reserved + + space_info->bytes_pinned + space_info->bytes_readonly + + space_info->bytes_may_use; /* * The idea here is that we've not already over-reserved the block group @@ -3442,10 +3898,11 @@ again: * lets start flushing stuff first and then come back and try to make * our reservation. */ - if (unused <= space_info->total_bytes) { - unused = space_info->total_bytes - unused; - if (unused >= num_bytes) { - space_info->bytes_reserved += orig_bytes; + if (used <= space_info->total_bytes) { + if (used + orig_bytes <= space_info->total_bytes) { + space_info->bytes_may_use += orig_bytes; + trace_btrfs_space_reservation(root->fs_info, + "space_info", space_info->flags, orig_bytes, 1); ret = 0; } else { /* @@ -3461,8 +3918,61 @@ again: * amount plus the amount of bytes that we need for this * reservation. */ - num_bytes = unused - space_info->total_bytes + - (orig_bytes * (retries + 1)); + num_bytes = used - space_info->total_bytes + + (orig_bytes * 2); + } + + if (ret) { + u64 profile = btrfs_get_alloc_profile(root, 0); + u64 avail; + + /* + * If we have a lot of space that's pinned, don't bother doing + * the overcommit dance yet and just commit the transaction. + */ + avail = (space_info->total_bytes - space_info->bytes_used) * 8; + do_div(avail, 10); + if (space_info->bytes_pinned >= avail && flush && !committed) { + space_info->flush = 1; + flushing = true; + spin_unlock(&space_info->lock); + ret = may_commit_transaction(root, space_info, + orig_bytes, 1); + if (ret) + goto out; + committed = true; + goto again; + } + + spin_lock(&root->fs_info->free_chunk_lock); + avail = root->fs_info->free_chunk_space; + + /* + * If we have dup, raid1 or raid10 then only half of the free + * space is actually useable. + */ + if (profile & (BTRFS_BLOCK_GROUP_DUP | + BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10)) + avail >>= 1; + + /* + * If we aren't flushing don't let us overcommit too much, say + * 1/8th of the space. If we can flush, let it overcommit up to + * 1/2 of the space. + */ + if (flush) + avail >>= 3; + else + avail >>= 1; + spin_unlock(&root->fs_info->free_chunk_lock); + + if (used + num_bytes < space_info->total_bytes + avail) { + space_info->bytes_may_use += orig_bytes; + trace_btrfs_space_reservation(root->fs_info, + "space_info", space_info->flags, orig_bytes, 1); + ret = 0; + } } /* @@ -3480,54 +3990,13 @@ again: if (!ret || !flush) goto out; - /* - * We do synchronous shrinking since we don't actually unreserve - * metadata until after the IO is completed. - */ - ret = shrink_delalloc(trans, root, num_bytes, 1); - if (ret < 0) - goto out; - - ret = 0; - - /* - * So if we were overcommitted it's possible that somebody else flushed - * out enough space and we simply didn't have enough space to reclaim, - * so go back around and try again. - */ - if (retries < 2) { - retries++; + ret = flush_space(root, space_info, num_bytes, orig_bytes, + flush_state); + flush_state++; + if (!ret) goto again; - } - - /* - * Not enough space to be reclaimed, don't bother committing the - * transaction. - */ - spin_lock(&space_info->lock); - if (space_info->bytes_pinned < orig_bytes) - ret = -ENOSPC; - spin_unlock(&space_info->lock); - if (ret) - goto out; - - ret = -EAGAIN; - if (trans) - goto out; - - ret = -ENOSPC; - if (committed) - goto out; - - trans = btrfs_join_transaction(root); - if (IS_ERR(trans)) - goto out; - ret = btrfs_commit_transaction(trans, root); - if (!ret) { - trans = NULL; - committed = true; + else if (flush_state <= COMMIT_TRANS) goto again; - } out: if (flushing) { @@ -3539,13 +4008,19 @@ out: return ret; } -static struct btrfs_block_rsv *get_block_rsv(struct btrfs_trans_handle *trans, - struct btrfs_root *root) +static struct btrfs_block_rsv *get_block_rsv( + const struct btrfs_trans_handle *trans, + const struct btrfs_root *root) { - struct btrfs_block_rsv *block_rsv; + struct btrfs_block_rsv *block_rsv = NULL; + if (root->ref_cows) block_rsv = trans->block_rsv; - else + + if (root == root->fs_info->csum_root && trans->adding_csums) + block_rsv = trans->block_rsv; + + if (!block_rsv) block_rsv = root->block_rsv; if (!block_rsv) @@ -3581,7 +4056,8 @@ static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv, spin_unlock(&block_rsv->lock); } -static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, +static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_block_rsv *block_rsv, struct btrfs_block_rsv *dest, u64 num_bytes) { struct btrfs_space_info *space_info = block_rsv->space_info; @@ -3616,7 +4092,9 @@ static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, } if (num_bytes) { spin_lock(&space_info->lock); - space_info->bytes_reserved -= num_bytes; + space_info->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, num_bytes, 0); space_info->reservation_progress++; spin_unlock(&space_info->lock); } @@ -3640,9 +4118,6 @@ void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv) { memset(rsv, 0, sizeof(*rsv)); spin_lock_init(&rsv->lock); - atomic_set(&rsv->usage, 1); - rsv->priority = 6; - INIT_LIST_HEAD(&rsv->list); } struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root) @@ -3663,38 +4138,20 @@ struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root) void btrfs_free_block_rsv(struct btrfs_root *root, struct btrfs_block_rsv *rsv) { - if (rsv && atomic_dec_and_test(&rsv->usage)) { - btrfs_block_rsv_release(root, rsv, (u64)-1); - if (!rsv->durable) - kfree(rsv); - } + btrfs_block_rsv_release(root, rsv, (u64)-1); + kfree(rsv); } -/* - * make the block_rsv struct be able to capture freed space. - * the captured space will re-add to the the block_rsv struct - * after transaction commit - */ -void btrfs_add_durable_block_rsv(struct btrfs_fs_info *fs_info, - struct btrfs_block_rsv *block_rsv) -{ - block_rsv->durable = 1; - mutex_lock(&fs_info->durable_block_rsv_mutex); - list_add_tail(&block_rsv->list, &fs_info->durable_block_rsv_list); - mutex_unlock(&fs_info->durable_block_rsv_mutex); -} - -int btrfs_block_rsv_add(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, - u64 num_bytes) +static inline int __block_rsv_add(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes, int flush) { int ret; if (num_bytes == 0) return 0; - ret = reserve_metadata_bytes(trans, root, block_rsv, num_bytes, 1); + ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); if (!ret) { block_rsv_add_bytes(block_rsv, num_bytes, 1); return 0; @@ -3703,55 +4160,80 @@ int btrfs_block_rsv_add(struct btrfs_trans_handle *trans, return ret; } -int btrfs_block_rsv_check(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, - u64 min_reserved, int min_factor) +int btrfs_block_rsv_add(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes) +{ + return __block_rsv_add(root, block_rsv, num_bytes, 1); +} + +int btrfs_block_rsv_add_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 num_bytes) +{ + return __block_rsv_add(root, block_rsv, num_bytes, 0); +} + +int btrfs_block_rsv_check(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, int min_factor) { u64 num_bytes = 0; - int commit_trans = 0; int ret = -ENOSPC; if (!block_rsv) return 0; spin_lock(&block_rsv->lock); - if (min_factor > 0) - num_bytes = div_factor(block_rsv->size, min_factor); - if (min_reserved > num_bytes) - num_bytes = min_reserved; + num_bytes = div_factor(block_rsv->size, min_factor); + if (block_rsv->reserved >= num_bytes) + ret = 0; + spin_unlock(&block_rsv->lock); - if (block_rsv->reserved >= num_bytes) { + return ret; +} + +static inline int __btrfs_block_rsv_refill(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved, int flush) +{ + u64 num_bytes = 0; + int ret = -ENOSPC; + + if (!block_rsv) + return 0; + + spin_lock(&block_rsv->lock); + num_bytes = min_reserved; + if (block_rsv->reserved >= num_bytes) ret = 0; - } else { + else num_bytes -= block_rsv->reserved; - if (block_rsv->durable && - block_rsv->freed[0] + block_rsv->freed[1] >= num_bytes) - commit_trans = 1; - } spin_unlock(&block_rsv->lock); + if (!ret) return 0; - if (block_rsv->refill_used) { - ret = reserve_metadata_bytes(trans, root, block_rsv, - num_bytes, 0); - if (!ret) { - block_rsv_add_bytes(block_rsv, num_bytes, 0); - return 0; - } - } - - if (commit_trans) { - if (trans) - return -EAGAIN; - trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); - ret = btrfs_commit_transaction(trans, root); + ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); + if (!ret) { + block_rsv_add_bytes(block_rsv, num_bytes, 0); return 0; } - return -ENOSPC; + return ret; +} + +int btrfs_block_rsv_refill(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved) +{ + return __btrfs_block_rsv_refill(root, block_rsv, min_reserved, 1); +} + +int btrfs_block_rsv_refill_noflush(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 min_reserved) +{ + return __btrfs_block_rsv_refill(root, block_rsv, min_reserved, 0); } int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, @@ -3769,7 +4251,8 @@ void btrfs_block_rsv_release(struct btrfs_root *root, if (global_rsv->full || global_rsv == block_rsv || block_rsv->space_info != global_rsv->space_info) global_rsv = NULL; - block_rsv_release_bytes(block_rsv, global_rsv, num_bytes); + block_rsv_release_bytes(root->fs_info, block_rsv, global_rsv, + num_bytes); } /* @@ -3783,7 +4266,7 @@ static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) u64 num_bytes; u64 meta_used; u64 data_used; - int csum_size = btrfs_super_csum_size(&fs_info->super_copy); + int csum_size = btrfs_super_csum_size(fs_info->super_copy); sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); spin_lock(&sinfo->lock); @@ -3815,8 +4298,8 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info) num_bytes = calc_global_metadata_size(fs_info); - spin_lock(&block_rsv->lock); spin_lock(&sinfo->lock); + spin_lock(&block_rsv->lock); block_rsv->size = num_bytes; @@ -3827,19 +4310,23 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info) if (sinfo->total_bytes > num_bytes) { num_bytes = sinfo->total_bytes - num_bytes; block_rsv->reserved += num_bytes; - sinfo->bytes_reserved += num_bytes; + sinfo->bytes_may_use += num_bytes; + trace_btrfs_space_reservation(fs_info, "space_info", + sinfo->flags, num_bytes, 1); } if (block_rsv->reserved >= block_rsv->size) { num_bytes = block_rsv->reserved - block_rsv->size; - sinfo->bytes_reserved -= num_bytes; + sinfo->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(fs_info, "space_info", + sinfo->flags, num_bytes, 0); sinfo->reservation_progress++; block_rsv->reserved = block_rsv->size; block_rsv->full = 1; } - spin_unlock(&sinfo->lock); spin_unlock(&block_rsv->lock); + spin_unlock(&sinfo->lock); } static void init_global_block_rsv(struct btrfs_fs_info *fs_info) @@ -3848,16 +4335,13 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info) space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM); fs_info->chunk_block_rsv.space_info = space_info; - fs_info->chunk_block_rsv.priority = 10; space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); fs_info->global_block_rsv.space_info = space_info; - fs_info->global_block_rsv.priority = 10; - fs_info->global_block_rsv.refill_used = 1; fs_info->delalloc_block_rsv.space_info = space_info; fs_info->trans_block_rsv.space_info = space_info; fs_info->empty_block_rsv.space_info = space_info; - fs_info->empty_block_rsv.priority = 10; + fs_info->delayed_block_rsv.space_info = space_info; fs_info->extent_root->block_rsv = &fs_info->global_block_rsv; fs_info->csum_root->block_rsv = &fs_info->global_block_rsv; @@ -3865,67 +4349,39 @@ static void init_global_block_rsv(struct btrfs_fs_info *fs_info) fs_info->tree_root->block_rsv = &fs_info->global_block_rsv; fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv; - btrfs_add_durable_block_rsv(fs_info, &fs_info->global_block_rsv); - - btrfs_add_durable_block_rsv(fs_info, &fs_info->delalloc_block_rsv); - update_global_block_rsv(fs_info); } static void release_global_block_rsv(struct btrfs_fs_info *fs_info) { - block_rsv_release_bytes(&fs_info->global_block_rsv, NULL, (u64)-1); + block_rsv_release_bytes(fs_info, &fs_info->global_block_rsv, NULL, + (u64)-1); WARN_ON(fs_info->delalloc_block_rsv.size > 0); WARN_ON(fs_info->delalloc_block_rsv.reserved > 0); WARN_ON(fs_info->trans_block_rsv.size > 0); WARN_ON(fs_info->trans_block_rsv.reserved > 0); WARN_ON(fs_info->chunk_block_rsv.size > 0); WARN_ON(fs_info->chunk_block_rsv.reserved > 0); -} - -int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_block_rsv *rsv) -{ - struct btrfs_block_rsv *trans_rsv = &root->fs_info->trans_block_rsv; - u64 num_bytes; - int ret; - - /* - * Truncate should be freeing data, but give us 2 items just in case it - * needs to use some space. We may want to be smarter about this in the - * future. - */ - num_bytes = btrfs_calc_trans_metadata_size(root, 2); - - /* We already have enough bytes, just return */ - if (rsv->reserved >= num_bytes) - return 0; - - num_bytes -= rsv->reserved; - - /* - * You should have reserved enough space before hand to do this, so this - * should not fail. - */ - ret = block_rsv_migrate_bytes(trans_rsv, rsv, num_bytes); - BUG_ON(ret); - - return 0; + WARN_ON(fs_info->delayed_block_rsv.size > 0); + WARN_ON(fs_info->delayed_block_rsv.reserved > 0); } void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans, struct btrfs_root *root) { + if (!trans->block_rsv) + return; + if (!trans->bytes_reserved) return; - BUG_ON(trans->block_rsv != &root->fs_info->trans_block_rsv); - btrfs_block_rsv_release(root, trans->block_rsv, - trans->bytes_reserved); + trace_btrfs_space_reservation(root->fs_info, "transaction", + trans->transid, trans->bytes_reserved, 0); + btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved); trans->bytes_reserved = 0; } +/* Can only return 0 or -ENOSPC */ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, struct inode *inode) { @@ -3939,6 +4395,8 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, * when we are truly done with the orphan item. */ u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); + trace_btrfs_space_reservation(root->fs_info, "orphan", + btrfs_ino(inode), num_bytes, 1); return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); } @@ -3946,6 +4404,8 @@ void btrfs_orphan_release_metadata(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); + trace_btrfs_space_reservation(root->fs_info, "orphan", + btrfs_ino(inode), num_bytes, 0); btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes); } @@ -3964,33 +4424,98 @@ int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans, return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); } +/** + * drop_outstanding_extent - drop an outstanding extent + * @inode: the inode we're dropping the extent for + * + * This is called when we are freeing up an outstanding extent, either called + * after an error or after an extent is written. This will return the number of + * reserved extents that need to be freed. This must be called with + * BTRFS_I(inode)->lock held. + */ static unsigned drop_outstanding_extent(struct inode *inode) { + unsigned drop_inode_space = 0; unsigned dropped_extents = 0; - spin_lock(&BTRFS_I(inode)->lock); BUG_ON(!BTRFS_I(inode)->outstanding_extents); BTRFS_I(inode)->outstanding_extents--; + if (BTRFS_I(inode)->outstanding_extents == 0 && + test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) + drop_inode_space = 1; + /* * If we have more or the same amount of outsanding extents than we have * reserved then we need to leave the reserved extents count alone. */ if (BTRFS_I(inode)->outstanding_extents >= BTRFS_I(inode)->reserved_extents) - goto out; + return drop_inode_space; dropped_extents = BTRFS_I(inode)->reserved_extents - BTRFS_I(inode)->outstanding_extents; BTRFS_I(inode)->reserved_extents -= dropped_extents; -out: - spin_unlock(&BTRFS_I(inode)->lock); - return dropped_extents; + return dropped_extents + drop_inode_space; } -static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes) +/** + * calc_csum_metadata_size - return the amount of metada space that must be + * reserved/free'd for the given bytes. + * @inode: the inode we're manipulating + * @num_bytes: the number of bytes in question + * @reserve: 1 if we are reserving space, 0 if we are freeing space + * + * This adjusts the number of csum_bytes in the inode and then returns the + * correct amount of metadata that must either be reserved or freed. We + * calculate how many checksums we can fit into one leaf and then divide the + * number of bytes that will need to be checksumed by this value to figure out + * how many checksums will be required. If we are adding bytes then the number + * may go up and we will return the number of additional bytes that must be + * reserved. If it is going down we will return the number of bytes that must + * be freed. + * + * This must be called with BTRFS_I(inode)->lock held. + */ +static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes, + int reserve) { - return num_bytes >>= 3; + struct btrfs_root *root = BTRFS_I(inode)->root; + u64 csum_size; + int num_csums_per_leaf; + int num_csums; + int old_csums; + + if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM && + BTRFS_I(inode)->csum_bytes == 0) + return 0; + + old_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); + if (reserve) + BTRFS_I(inode)->csum_bytes += num_bytes; + else + BTRFS_I(inode)->csum_bytes -= num_bytes; + csum_size = BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item); + num_csums_per_leaf = (int)div64_u64(csum_size, + sizeof(struct btrfs_csum_item) + + sizeof(struct btrfs_disk_key)); + num_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); + num_csums = num_csums + num_csums_per_leaf - 1; + num_csums = num_csums / num_csums_per_leaf; + + old_csums = old_csums + num_csums_per_leaf - 1; + old_csums = old_csums / num_csums_per_leaf; + + /* No change, no need to reserve more */ + if (old_csums == num_csums) + return 0; + + if (reserve) + return btrfs_calc_trans_metadata_size(root, + num_csums - old_csums); + + return btrfs_calc_trans_metadata_size(root, old_csums - num_csums); } int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) @@ -3998,45 +4523,115 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv; u64 to_reserve = 0; + u64 csum_bytes; unsigned nr_extents = 0; + int extra_reserve = 0; + int flush = 1; int ret; - if (btrfs_transaction_in_commit(root->fs_info)) + /* Need to be holding the i_mutex here if we aren't free space cache */ + if (btrfs_is_free_space_inode(inode)) + flush = 0; + + if (flush && btrfs_transaction_in_commit(root->fs_info)) schedule_timeout(1); + mutex_lock(&BTRFS_I(inode)->delalloc_mutex); num_bytes = ALIGN(num_bytes, root->sectorsize); spin_lock(&BTRFS_I(inode)->lock); BTRFS_I(inode)->outstanding_extents++; if (BTRFS_I(inode)->outstanding_extents > - BTRFS_I(inode)->reserved_extents) { + BTRFS_I(inode)->reserved_extents) nr_extents = BTRFS_I(inode)->outstanding_extents - BTRFS_I(inode)->reserved_extents; - BTRFS_I(inode)->reserved_extents += nr_extents; - to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); + /* + * Add an item to reserve for updating the inode when we complete the + * delalloc io. + */ + if (!test_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) { + nr_extents++; + extra_reserve = 1; } + + to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); + to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); + csum_bytes = BTRFS_I(inode)->csum_bytes; spin_unlock(&BTRFS_I(inode)->lock); - to_reserve += calc_csum_metadata_size(inode, num_bytes); - ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1); + if (root->fs_info->quota_enabled) { + ret = btrfs_qgroup_reserve(root, num_bytes + + nr_extents * root->leafsize); + if (ret) { + mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); + return ret; + } + } + + ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush); if (ret) { + u64 to_free = 0; unsigned dropped; + + spin_lock(&BTRFS_I(inode)->lock); + dropped = drop_outstanding_extent(inode); /* - * We don't need the return value since our reservation failed, - * we just need to clean up our counter. + * If the inodes csum_bytes is the same as the original + * csum_bytes then we know we haven't raced with any free()ers + * so we can just reduce our inodes csum bytes and carry on. + * Otherwise we have to do the normal free thing to account for + * the case that the free side didn't free up its reserve + * because of this outstanding reservation. */ - dropped = drop_outstanding_extent(inode); - WARN_ON(dropped > 1); + if (BTRFS_I(inode)->csum_bytes == csum_bytes) + calc_csum_metadata_size(inode, num_bytes, 0); + else + to_free = calc_csum_metadata_size(inode, num_bytes, 0); + spin_unlock(&BTRFS_I(inode)->lock); + if (dropped) + to_free += btrfs_calc_trans_metadata_size(root, dropped); + + if (to_free) { + btrfs_block_rsv_release(root, block_rsv, to_free); + trace_btrfs_space_reservation(root->fs_info, + "delalloc", + btrfs_ino(inode), + to_free, 0); + } + mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); return ret; } + spin_lock(&BTRFS_I(inode)->lock); + if (extra_reserve) { + set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags); + nr_extents--; + } + BTRFS_I(inode)->reserved_extents += nr_extents; + spin_unlock(&BTRFS_I(inode)->lock); + mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); + + if (to_reserve) + trace_btrfs_space_reservation(root->fs_info,"delalloc", + btrfs_ino(inode), to_reserve, 1); block_rsv_add_bytes(block_rsv, to_reserve, 1); return 0; } +/** + * btrfs_delalloc_release_metadata - release a metadata reservation for an inode + * @inode: the inode to release the reservation for + * @num_bytes: the number of bytes we're releasing + * + * This will release the metadata reservation for an inode. This can be called + * once we complete IO for a given set of bytes to release their metadata + * reservations. + */ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) { struct btrfs_root *root = BTRFS_I(inode)->root; @@ -4044,16 +4639,40 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) unsigned dropped; num_bytes = ALIGN(num_bytes, root->sectorsize); + spin_lock(&BTRFS_I(inode)->lock); dropped = drop_outstanding_extent(inode); - to_free = calc_csum_metadata_size(inode, num_bytes); + to_free = calc_csum_metadata_size(inode, num_bytes, 0); + spin_unlock(&BTRFS_I(inode)->lock); if (dropped > 0) to_free += btrfs_calc_trans_metadata_size(root, dropped); + trace_btrfs_space_reservation(root->fs_info, "delalloc", + btrfs_ino(inode), to_free, 0); + if (root->fs_info->quota_enabled) { + btrfs_qgroup_free(root, num_bytes + + dropped * root->leafsize); + } + btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv, to_free); } +/** + * btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc + * @inode: inode we're writing to + * @num_bytes: the number of bytes we want to allocate + * + * This will do the following things + * + * o reserve space in the data space info for num_bytes + * o reserve space in the metadata space info based on number of outstanding + * extents and how much csums will be needed + * o add to the inodes ->delalloc_bytes + * o add it to the fs_info's delalloc inodes list. + * + * This will return 0 for success and -ENOSPC if there is no space left. + */ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes) { int ret; @@ -4071,6 +4690,19 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes) return 0; } +/** + * btrfs_delalloc_release_space - release data and metadata space for delalloc + * @inode: inode we're releasing space for + * @num_bytes: the number of bytes we want to free up + * + * This must be matched with a call to btrfs_delalloc_reserve_space. This is + * called in the case that we don't need the metadata AND data reservations + * anymore. So if there is an error or we insert an inline extent. + * + * This function will release the metadata space that was not used and will + * decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes + * list if there are no delalloc bytes left. + */ void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes) { btrfs_delalloc_release_metadata(inode, num_bytes); @@ -4090,18 +4722,18 @@ static int update_block_group(struct btrfs_trans_handle *trans, /* block accounting for super block */ spin_lock(&info->delalloc_lock); - old_val = btrfs_super_bytes_used(&info->super_copy); + old_val = btrfs_super_bytes_used(info->super_copy); if (alloc) old_val += num_bytes; else old_val -= num_bytes; - btrfs_set_super_bytes_used(&info->super_copy, old_val); + btrfs_set_super_bytes_used(info->super_copy, old_val); spin_unlock(&info->delalloc_lock); while (total) { cache = btrfs_lookup_block_group(info, bytenr); if (!cache) - return -1; + return -ENOENT; if (cache->flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10)) @@ -4123,7 +4755,7 @@ static int update_block_group(struct btrfs_trans_handle *trans, spin_lock(&cache->space_info->lock); spin_lock(&cache->lock); - if (btrfs_super_cache_generation(&info->super_copy) != 0 && + if (btrfs_test_opt(root, SPACE_CACHE) && cache->disk_cache_state < BTRFS_DC_CLEAR) cache->disk_cache_state = BTRFS_DC_CLEAR; @@ -4135,7 +4767,6 @@ static int update_block_group(struct btrfs_trans_handle *trans, btrfs_set_block_group_used(&cache->item, old_val); cache->reserved -= num_bytes; cache->space_info->bytes_reserved -= num_bytes; - cache->space_info->reservation_progress++; cache->space_info->bytes_used += num_bytes; cache->space_info->disk_used += num_bytes * factor; spin_unlock(&cache->lock); @@ -4187,7 +4818,6 @@ static int pin_down_extent(struct btrfs_root *root, if (reserved) { cache->reserved -= num_bytes; cache->space_info->bytes_reserved -= num_bytes; - cache->space_info->reservation_progress++; } spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); @@ -4206,7 +4836,7 @@ int btrfs_pin_extent(struct btrfs_root *root, struct btrfs_block_group_cache *cache; cache = btrfs_lookup_block_group(root->fs_info, bytenr); - BUG_ON(!cache); + BUG_ON(!cache); /* Logic error */ pin_down_extent(root, cache, bytenr, num_bytes, reserved); @@ -4215,49 +4845,89 @@ int btrfs_pin_extent(struct btrfs_root *root, } /* - * update size of reserved extents. this function may return -EAGAIN - * if 'reserve' is true or 'sinfo' is false. + * this function must be called within transaction + */ +int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + u64 bytenr, u64 num_bytes) +{ + struct btrfs_block_group_cache *cache; + + cache = btrfs_lookup_block_group(root->fs_info, bytenr); + BUG_ON(!cache); /* Logic error */ + + /* + * pull in the free space cache (if any) so that our pin + * removes the free space from the cache. We have load_only set + * to one because the slow code to read in the free extents does check + * the pinned extents. + */ + cache_block_group(cache, trans, root, 1); + + pin_down_extent(root, cache, bytenr, num_bytes, 0); + + /* remove us from the free space cache (if we're there at all) */ + btrfs_remove_free_space(cache, bytenr, num_bytes); + btrfs_put_block_group(cache); + return 0; +} + +/** + * btrfs_update_reserved_bytes - update the block_group and space info counters + * @cache: The cache we are manipulating + * @num_bytes: The number of bytes in question + * @reserve: One of the reservation enums + * + * This is called by the allocator when it reserves space, or by somebody who is + * freeing space that was never actually used on disk. For example if you + * reserve some space for a new leaf in transaction A and before transaction A + * commits you free that leaf, you call this with reserve set to 0 in order to + * clear the reservation. + * + * Metadata reservations should be called with RESERVE_ALLOC so we do the proper + * ENOSPC accounting. For data we handle the reservation through clearing the + * delalloc bits in the io_tree. We have to do this since we could end up + * allocating less disk space for the amount of data we have reserved in the + * case of compression. + * + * If this is a reservation and the block group has become read only we cannot + * make the reservation and return -EAGAIN, otherwise this function always + * succeeds. */ -int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, - u64 num_bytes, int reserve, int sinfo) +static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, + u64 num_bytes, int reserve) { + struct btrfs_space_info *space_info = cache->space_info; int ret = 0; - if (sinfo) { - struct btrfs_space_info *space_info = cache->space_info; - spin_lock(&space_info->lock); - spin_lock(&cache->lock); - if (reserve) { - if (cache->ro) { - ret = -EAGAIN; - } else { - cache->reserved += num_bytes; - space_info->bytes_reserved += num_bytes; - } - } else { - if (cache->ro) - space_info->bytes_readonly += num_bytes; - cache->reserved -= num_bytes; - space_info->bytes_reserved -= num_bytes; - space_info->reservation_progress++; - } - spin_unlock(&cache->lock); - spin_unlock(&space_info->lock); - } else { - spin_lock(&cache->lock); + + spin_lock(&space_info->lock); + spin_lock(&cache->lock); + if (reserve != RESERVE_FREE) { if (cache->ro) { ret = -EAGAIN; } else { - if (reserve) - cache->reserved += num_bytes; - else - cache->reserved -= num_bytes; + cache->reserved += num_bytes; + space_info->bytes_reserved += num_bytes; + if (reserve == RESERVE_ALLOC) { + trace_btrfs_space_reservation(cache->fs_info, + "space_info", space_info->flags, + num_bytes, 0); + space_info->bytes_may_use -= num_bytes; + } } - spin_unlock(&cache->lock); + } else { + if (cache->ro) + space_info->bytes_readonly += num_bytes; + cache->reserved -= num_bytes; + space_info->bytes_reserved -= num_bytes; + space_info->reservation_progress++; } + spin_unlock(&cache->lock); + spin_unlock(&space_info->lock); return ret; } -int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, +void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; @@ -4287,7 +4957,6 @@ int btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, up_write(&fs_info->extent_commit_sem); update_global_block_rsv(fs_info); - return 0; } static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) @@ -4302,7 +4971,7 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) if (cache) btrfs_put_block_group(cache); cache = btrfs_lookup_block_group(fs_info, start); - BUG_ON(!cache); + BUG_ON(!cache); /* Logic error */ } len = cache->key.objectid + cache->key.offset - start; @@ -4319,13 +4988,8 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) spin_lock(&cache->lock); cache->pinned -= len; cache->space_info->bytes_pinned -= len; - if (cache->ro) { + if (cache->ro) cache->space_info->bytes_readonly += len; - } else if (cache->reserved_pinned > 0) { - len = min(len, cache->reserved_pinned); - cache->reserved_pinned -= len; - cache->space_info->bytes_reserved += len; - } spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); } @@ -4340,13 +5004,13 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, { struct btrfs_fs_info *fs_info = root->fs_info; struct extent_io_tree *unpin; - struct btrfs_block_rsv *block_rsv; - struct btrfs_block_rsv *next_rsv; u64 start; u64 end; - int idx; int ret; + if (trans->aborted) + return 0; + if (fs_info->pinned_extents == &fs_info->freed_extents[0]) unpin = &fs_info->freed_extents[1]; else @@ -4367,30 +5031,6 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, cond_resched(); } - mutex_lock(&fs_info->durable_block_rsv_mutex); - list_for_each_entry_safe(block_rsv, next_rsv, - &fs_info->durable_block_rsv_list, list) { - - idx = trans->transid & 0x1; - if (block_rsv->freed[idx] > 0) { - block_rsv_add_bytes(block_rsv, - block_rsv->freed[idx], 0); - block_rsv->freed[idx] = 0; - } - if (atomic_read(&block_rsv->usage) == 0) { - btrfs_block_rsv_release(root, block_rsv, (u64)-1); - - if (block_rsv->freed[0] == 0 && - block_rsv->freed[1] == 0) { - list_del_init(&block_rsv->list); - kfree(block_rsv); - } - } else { - btrfs_block_rsv_release(root, block_rsv, 0); - } - } - mutex_unlock(&fs_info->durable_block_rsv_mutex); - return 0; } @@ -4456,7 +5096,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = remove_extent_backref(trans, extent_root, path, NULL, refs_to_drop, is_data); - BUG_ON(ret); + if (ret) + goto abort; btrfs_release_path(path); path->leave_spinning = 1; @@ -4474,10 +5115,11 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, btrfs_print_leaf(extent_root, path->nodes[0]); } - BUG_ON(ret); + if (ret < 0) + goto abort; extent_slot = path->slots[0]; } - } else { + } else if (ret == -ENOENT) { btrfs_print_leaf(extent_root, path->nodes[0]); WARN_ON(1); printk(KERN_ERR "btrfs unable to find ref byte nr %llu " @@ -4487,6 +5129,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, (unsigned long long)root_objectid, (unsigned long long)owner_objectid, (unsigned long long)owner_offset); + } else { + goto abort; } leaf = path->nodes[0]; @@ -4496,7 +5140,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, BUG_ON(found_extent || extent_slot != path->slots[0]); ret = convert_extent_item_v0(trans, extent_root, path, owner_objectid, 0); - BUG_ON(ret < 0); + if (ret < 0) + goto abort; btrfs_release_path(path); path->leave_spinning = 1; @@ -4513,7 +5158,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, (unsigned long long)bytenr); btrfs_print_leaf(extent_root, path->nodes[0]); } - BUG_ON(ret); + if (ret < 0) + goto abort; extent_slot = path->slots[0]; leaf = path->nodes[0]; item_size = btrfs_item_size_nr(leaf, extent_slot); @@ -4550,7 +5196,8 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = remove_extent_backref(trans, extent_root, path, iref, refs_to_drop, is_data); - BUG_ON(ret); + if (ret) + goto abort; } } else { if (found_extent) { @@ -4567,23 +5214,27 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = btrfs_del_items(trans, extent_root, path, path->slots[0], num_to_del); - BUG_ON(ret); + if (ret) + goto abort; btrfs_release_path(path); if (is_data) { ret = btrfs_del_csums(trans, root, bytenr, num_bytes); - BUG_ON(ret); - } else { - invalidate_mapping_pages(info->btree_inode->i_mapping, - bytenr >> PAGE_CACHE_SHIFT, - (bytenr + num_bytes - 1) >> PAGE_CACHE_SHIFT); + if (ret) + goto abort; } ret = update_block_group(trans, root, bytenr, num_bytes, 0); - BUG_ON(ret); + if (ret) + goto abort; } +out: btrfs_free_path(path); return ret; + +abort: + btrfs_abort_transaction(trans, extent_root, ret); + goto out; } /* @@ -4668,79 +5319,39 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans, struct extent_buffer *buf, u64 parent, int last_ref) { - struct btrfs_block_rsv *block_rsv; struct btrfs_block_group_cache *cache = NULL; int ret; if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { - ret = btrfs_add_delayed_tree_ref(trans, buf->start, buf->len, - parent, root->root_key.objectid, - btrfs_header_level(buf), - BTRFS_DROP_DELAYED_REF, NULL); - BUG_ON(ret); + ret = btrfs_add_delayed_tree_ref(root->fs_info, trans, + buf->start, buf->len, + parent, root->root_key.objectid, + btrfs_header_level(buf), + BTRFS_DROP_DELAYED_REF, NULL, 0); + BUG_ON(ret); /* -ENOMEM */ } if (!last_ref) return; - block_rsv = get_block_rsv(trans, root); cache = btrfs_lookup_block_group(root->fs_info, buf->start); - if (block_rsv->space_info != cache->space_info) - goto out; if (btrfs_header_generation(buf) == trans->transid) { if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { ret = check_ref_cleanup(trans, root, buf->start); if (!ret) - goto pin; + goto out; } if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) { pin_down_extent(root, cache, buf->start, buf->len, 1); - goto pin; + goto out; } WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)); btrfs_add_free_space(cache, buf->start, buf->len); - ret = btrfs_update_reserved_bytes(cache, buf->len, 0, 0); - if (ret == -EAGAIN) { - /* block group became read-only */ - btrfs_update_reserved_bytes(cache, buf->len, 0, 1); - goto out; - } - - ret = 1; - spin_lock(&block_rsv->lock); - if (block_rsv->reserved < block_rsv->size) { - block_rsv->reserved += buf->len; - ret = 0; - } - spin_unlock(&block_rsv->lock); - - if (ret) { - spin_lock(&cache->space_info->lock); - cache->space_info->bytes_reserved -= buf->len; - cache->space_info->reservation_progress++; - spin_unlock(&cache->space_info->lock); - } - goto out; - } -pin: - if (block_rsv->durable && !cache->ro) { - ret = 0; - spin_lock(&cache->lock); - if (!cache->ro) { - cache->reserved_pinned += buf->len; - ret = 1; - } - spin_unlock(&cache->lock); - - if (ret) { - spin_lock(&block_rsv->lock); - block_rsv->freed[trans->transid & 0x1] += buf->len; - spin_unlock(&block_rsv->lock); - } + btrfs_update_reserved_bytes(cache, buf->len, RESERVE_FREE); } out: /* @@ -4751,12 +5362,13 @@ out: btrfs_put_block_group(cache); } -int btrfs_free_extent(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - u64 bytenr, u64 num_bytes, u64 parent, - u64 root_objectid, u64 owner, u64 offset) +/* Can return -ENOMEM */ +int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, + u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, + u64 owner, u64 offset, int for_cow) { int ret; + struct btrfs_fs_info *fs_info = root->fs_info; /* * tree log blocks never actually go into the extent allocation @@ -4768,15 +5380,16 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, btrfs_pin_extent(root, bytenr, num_bytes, 1); ret = 0; } else if (owner < BTRFS_FIRST_FREE_OBJECTID) { - ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes, + ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, + num_bytes, parent, root_objectid, (int)owner, - BTRFS_DROP_DELAYED_REF, NULL); - BUG_ON(ret); + BTRFS_DROP_DELAYED_REF, NULL, for_cow); } else { - ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes, - parent, root_objectid, owner, - offset, BTRFS_DROP_DELAYED_REF, NULL); - BUG_ON(ret); + ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, + num_bytes, + parent, root_objectid, owner, + offset, BTRFS_DROP_DELAYED_REF, + NULL, for_cow); } return ret; } @@ -4833,28 +5446,34 @@ wait_block_group_cache_done(struct btrfs_block_group_cache *cache) return 0; } -static int get_block_group_index(struct btrfs_block_group_cache *cache) +static int __get_block_group_index(u64 flags) { int index; - if (cache->flags & BTRFS_BLOCK_GROUP_RAID10) + + if (flags & BTRFS_BLOCK_GROUP_RAID10) index = 0; - else if (cache->flags & BTRFS_BLOCK_GROUP_RAID1) + else if (flags & BTRFS_BLOCK_GROUP_RAID1) index = 1; - else if (cache->flags & BTRFS_BLOCK_GROUP_DUP) + else if (flags & BTRFS_BLOCK_GROUP_DUP) index = 2; - else if (cache->flags & BTRFS_BLOCK_GROUP_RAID0) + else if (flags & BTRFS_BLOCK_GROUP_RAID0) index = 3; else index = 4; + return index; } +static int get_block_group_index(struct btrfs_block_group_cache *cache) +{ + return __get_block_group_index(cache->flags); +} + enum btrfs_loop_type { - LOOP_FIND_IDEAL = 0, - LOOP_CACHING_NOWAIT = 1, - LOOP_CACHING_WAIT = 2, - LOOP_ALLOC_CHUNK = 3, - LOOP_NO_EMPTY_SIZE = 4, + LOOP_CACHING_NOWAIT = 0, + LOOP_CACHING_WAIT = 1, + LOOP_ALLOC_CHUNK = 2, + LOOP_NO_EMPTY_SIZE = 3, }; /* @@ -4868,7 +5487,6 @@ enum btrfs_loop_type { static noinline int find_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *orig_root, u64 num_bytes, u64 empty_size, - u64 search_start, u64 search_end, u64 hint_byte, struct btrfs_key *ins, u64 data) { @@ -4876,25 +5494,29 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root = orig_root->fs_info->extent_root; struct btrfs_free_cluster *last_ptr = NULL; struct btrfs_block_group_cache *block_group = NULL; + struct btrfs_block_group_cache *used_block_group; + u64 search_start = 0; int empty_cluster = 2 * 1024 * 1024; int allowed_chunk_alloc = 0; int done_chunk_alloc = 0; struct btrfs_space_info *space_info; - int last_ptr_loop = 0; int loop = 0; int index = 0; + int alloc_type = (data & BTRFS_BLOCK_GROUP_DATA) ? + RESERVE_ALLOC_NO_ACCOUNT : RESERVE_ALLOC; bool found_uncached_bg = false; bool failed_cluster_refill = false; bool failed_alloc = false; bool use_cluster = true; - u64 ideal_cache_percent = 0; - u64 ideal_cache_offset = 0; + bool have_caching_bg = false; WARN_ON(num_bytes < root->sectorsize); btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY); ins->objectid = 0; ins->offset = 0; + trace_find_free_extent(orig_root, num_bytes, empty_size, data); + space_info = __find_space_info(root->fs_info, data); if (!space_info) { printk(KERN_ERR "No space info for %llu\n", data); @@ -4936,9 +5558,9 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, empty_cluster = 0; if (search_start == hint_byte) { -ideal_cache: block_group = btrfs_lookup_block_group(root->fs_info, search_start); + used_block_group = block_group; /* * we don't want to use the block group if it doesn't match our * allocation bits, or if its not cached. @@ -4947,8 +5569,7 @@ ideal_cache: * picked out then we don't care that the block group is cached. */ if (block_group && block_group_bits(block_group, data) && - (block_group->cached != BTRFS_CACHE_NO || - search_start == ideal_cache_offset)) { + block_group->cached != BTRFS_CACHE_NO) { down_read(&space_info->groups_sem); if (list_empty(&block_group->list) || block_group->ro) { @@ -4969,12 +5590,14 @@ ideal_cache: } } search: + have_caching_bg = false; down_read(&space_info->groups_sem); list_for_each_entry(block_group, &space_info->block_groups[index], list) { u64 offset; int cached; + used_block_group = block_group; btrfs_get_block_group(block_group); search_start = block_group->key.objectid; @@ -4998,132 +5621,93 @@ search: } have_block_group: - if (unlikely(block_group->cached == BTRFS_CACHE_NO)) { - u64 free_percent; - - ret = cache_block_group(block_group, trans, - orig_root, 1); - if (block_group->cached == BTRFS_CACHE_FINISHED) - goto have_block_group; - - free_percent = btrfs_block_group_used(&block_group->item); - free_percent *= 100; - free_percent = div64_u64(free_percent, - block_group->key.offset); - free_percent = 100 - free_percent; - if (free_percent > ideal_cache_percent && - likely(!block_group->ro)) { - ideal_cache_offset = block_group->key.objectid; - ideal_cache_percent = free_percent; - } - - /* - * The caching workers are limited to 2 threads, so we - * can queue as much work as we care to. - */ - if (loop > LOOP_FIND_IDEAL) { - ret = cache_block_group(block_group, trans, - orig_root, 0); - BUG_ON(ret); - } - found_uncached_bg = true; - - /* - * If loop is set for cached only, try the next block - * group. - */ - if (loop == LOOP_FIND_IDEAL) - goto loop; - } - cached = block_group_cache_done(block_group); - if (unlikely(!cached)) + if (unlikely(!cached)) { found_uncached_bg = true; + ret = cache_block_group(block_group, trans, + orig_root, 0); + BUG_ON(ret < 0); + ret = 0; + } if (unlikely(block_group->ro)) goto loop; - spin_lock(&block_group->free_space_ctl->tree_lock); - if (cached && - block_group->free_space_ctl->free_space < - num_bytes + empty_size) { - spin_unlock(&block_group->free_space_ctl->tree_lock); - goto loop; - } - spin_unlock(&block_group->free_space_ctl->tree_lock); - /* - * Ok we want to try and use the cluster allocator, so lets look - * there, unless we are on LOOP_NO_EMPTY_SIZE, since we will - * have tried the cluster allocator plenty of times at this - * point and not have found anything, so we are likely way too - * fragmented for the clustering stuff to find anything, so lets - * just skip it and let the allocator find whatever block it can - * find + * Ok we want to try and use the cluster allocator, so + * lets look there */ - if (last_ptr && loop < LOOP_NO_EMPTY_SIZE) { + if (last_ptr) { /* * the refill lock keeps out other * people trying to start a new cluster */ spin_lock(&last_ptr->refill_lock); - if (last_ptr->block_group && - (last_ptr->block_group->ro || - !block_group_bits(last_ptr->block_group, data))) { - offset = 0; + used_block_group = last_ptr->block_group; + if (used_block_group != block_group && + (!used_block_group || + used_block_group->ro || + !block_group_bits(used_block_group, data))) { + used_block_group = block_group; goto refill_cluster; } - offset = btrfs_alloc_from_cluster(block_group, last_ptr, - num_bytes, search_start); + if (used_block_group != block_group) + btrfs_get_block_group(used_block_group); + + offset = btrfs_alloc_from_cluster(used_block_group, + last_ptr, num_bytes, used_block_group->key.objectid); if (offset) { /* we have a block, we're done */ spin_unlock(&last_ptr->refill_lock); + trace_btrfs_reserve_extent_cluster(root, + block_group, search_start, num_bytes); goto checks; } - spin_lock(&last_ptr->lock); - /* - * whoops, this cluster doesn't actually point to - * this block group. Get a ref on the block - * group is does point to and try again - */ - if (!last_ptr_loop && last_ptr->block_group && - last_ptr->block_group != block_group && - index <= - get_block_group_index(last_ptr->block_group)) { - - btrfs_put_block_group(block_group); - block_group = last_ptr->block_group; - btrfs_get_block_group(block_group); - spin_unlock(&last_ptr->lock); - spin_unlock(&last_ptr->refill_lock); - - last_ptr_loop = 1; - search_start = block_group->key.objectid; - /* - * we know this block group is properly - * in the list because - * btrfs_remove_block_group, drops the - * cluster before it removes the block - * group from the list - */ - goto have_block_group; + WARN_ON(last_ptr->block_group != used_block_group); + if (used_block_group != block_group) { + btrfs_put_block_group(used_block_group); + used_block_group = block_group; } - spin_unlock(&last_ptr->lock); refill_cluster: + BUG_ON(used_block_group != block_group); + /* If we are on LOOP_NO_EMPTY_SIZE, we can't + * set up a new clusters, so lets just skip it + * and let the allocator find whatever block + * it can find. If we reach this point, we + * will have tried the cluster allocator + * plenty of times and not have found + * anything, so we are likely way too + * fragmented for the clustering stuff to find + * anything. + * + * However, if the cluster is taken from the + * current block group, release the cluster + * first, so that we stand a better chance of + * succeeding in the unclustered + * allocation. */ + if (loop >= LOOP_NO_EMPTY_SIZE && + last_ptr->block_group != block_group) { + spin_unlock(&last_ptr->refill_lock); + goto unclustered_alloc; + } + /* * this cluster didn't work out, free it and * start over */ btrfs_return_cluster_to_free_space(NULL, last_ptr); - last_ptr_loop = 0; + if (loop >= LOOP_NO_EMPTY_SIZE) { + spin_unlock(&last_ptr->refill_lock); + goto unclustered_alloc; + } /* allocate a cluster in this block group */ ret = btrfs_find_space_cluster(trans, root, block_group, last_ptr, - offset, num_bytes, + search_start, num_bytes, empty_cluster + empty_size); if (ret == 0) { /* @@ -5136,6 +5720,9 @@ refill_cluster: if (offset) { /* we found one, proceed */ spin_unlock(&last_ptr->refill_lock); + trace_btrfs_reserve_extent_cluster(root, + block_group, search_start, + num_bytes); goto checks; } } else if (!cached && loop > LOOP_CACHING_NOWAIT @@ -5159,6 +5746,16 @@ refill_cluster: goto loop; } +unclustered_alloc: + spin_lock(&block_group->free_space_ctl->tree_lock); + if (cached && + block_group->free_space_ctl->free_space < + num_bytes + empty_cluster + empty_size) { + spin_unlock(&block_group->free_space_ctl->tree_lock); + goto loop; + } + spin_unlock(&block_group->free_space_ctl->tree_lock); + offset = btrfs_find_space_for_alloc(block_group, search_start, num_bytes, empty_size); /* @@ -5177,35 +5774,29 @@ refill_cluster: failed_alloc = true; goto have_block_group; } else if (!offset) { + if (!cached) + have_caching_bg = true; goto loop; } checks: search_start = stripe_align(root, offset); - /* move on to the next group */ - if (search_start + num_bytes >= search_end) { - btrfs_add_free_space(block_group, offset, num_bytes); - goto loop; - } /* move on to the next group */ if (search_start + num_bytes > - block_group->key.objectid + block_group->key.offset) { - btrfs_add_free_space(block_group, offset, num_bytes); + used_block_group->key.objectid + used_block_group->key.offset) { + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } - ins->objectid = search_start; - ins->offset = num_bytes; - if (offset < search_start) - btrfs_add_free_space(block_group, offset, + btrfs_add_free_space(used_block_group, offset, search_start - offset); BUG_ON(offset > search_start); - ret = btrfs_update_reserved_bytes(block_group, num_bytes, 1, - (data & BTRFS_BLOCK_GROUP_DATA)); + ret = btrfs_update_reserved_bytes(used_block_group, num_bytes, + alloc_type); if (ret == -EAGAIN) { - btrfs_add_free_space(block_group, offset, num_bytes); + btrfs_add_free_space(used_block_group, offset, num_bytes); goto loop; } @@ -5213,26 +5804,33 @@ checks: ins->objectid = search_start; ins->offset = num_bytes; + trace_btrfs_reserve_extent(orig_root, block_group, + search_start, num_bytes); if (offset < search_start) - btrfs_add_free_space(block_group, offset, + btrfs_add_free_space(used_block_group, offset, search_start - offset); BUG_ON(offset > search_start); + if (used_block_group != block_group) + btrfs_put_block_group(used_block_group); btrfs_put_block_group(block_group); break; loop: failed_cluster_refill = false; failed_alloc = false; BUG_ON(index != get_block_group_index(block_group)); + if (used_block_group != block_group) + btrfs_put_block_group(used_block_group); btrfs_put_block_group(block_group); } up_read(&space_info->groups_sem); + if (!ins->objectid && loop >= LOOP_CACHING_WAIT && have_caching_bg) + goto search; + if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES) goto search; - /* LOOP_FIND_IDEAL, only search caching/cached bg's, and don't wait for - * for them to make caching progress. Also - * determine the best possible bg to cache + /* * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking * caching kthreads as we move along * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching @@ -5242,50 +5840,21 @@ loop: */ if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) { index = 0; - if (loop == LOOP_FIND_IDEAL && found_uncached_bg) { - found_uncached_bg = false; - loop++; - if (!ideal_cache_percent) - goto search; - - /* - * 1 of the following 2 things have happened so far - * - * 1) We found an ideal block group for caching that - * is mostly full and will cache quickly, so we might - * as well wait for it. - * - * 2) We searched for cached only and we didn't find - * anything, and we didn't start any caching kthreads - * either, so chances are we will loop through and - * start a couple caching kthreads, and then come back - * around and just wait for them. This will be slower - * because we will have 2 caching kthreads reading at - * the same time when we could have just started one - * and waited for it to get far enough to give us an - * allocation, so go ahead and go to the wait caching - * loop. - */ - loop = LOOP_CACHING_WAIT; - search_start = ideal_cache_offset; - ideal_cache_percent = 0; - goto ideal_cache; - } else if (loop == LOOP_FIND_IDEAL) { - /* - * Didn't find a uncached bg, wait on anything we find - * next. - */ - loop = LOOP_CACHING_WAIT; - goto search; - } - loop++; - if (loop == LOOP_ALLOC_CHUNK) { if (allowed_chunk_alloc) { ret = do_chunk_alloc(trans, root, num_bytes + 2 * 1024 * 1024, data, CHUNK_ALLOC_LIMITED); + /* + * Do not bail out on ENOSPC since we + * can do more things. + */ + if (ret < 0 && ret != -ENOSPC) { + btrfs_abort_transaction(trans, + root, ret); + goto out; + } allowed_chunk_alloc = 0; if (ret == 1) done_chunk_alloc = 1; @@ -5314,6 +5883,7 @@ loop: } else if (ins->objectid) { ret = 0; } +out: return ret; } @@ -5325,7 +5895,8 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes, int index = 0; spin_lock(&info->lock); - printk(KERN_INFO "space_info has %llu free, is %sfull\n", + printk(KERN_INFO "space_info %llu has %llu free, is %sfull\n", + (unsigned long long)info->flags, (unsigned long long)(info->total_bytes - info->bytes_used - info->bytes_pinned - info->bytes_reserved - info->bytes_readonly), @@ -5347,13 +5918,13 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes, again: list_for_each_entry(cache, &info->block_groups[index], list) { spin_lock(&cache->lock); - printk(KERN_INFO "block group %llu has %llu bytes, %llu used " - "%llu pinned %llu reserved\n", + printk(KERN_INFO "block group %llu has %llu bytes, %llu used %llu pinned %llu reserved %s\n", (unsigned long long)cache->key.objectid, (unsigned long long)cache->key.offset, (unsigned long long)btrfs_block_group_used(&cache->item), (unsigned long long)cache->pinned, - (unsigned long long)cache->reserved); + (unsigned long long)cache->reserved, + cache->ro ? "[readonly]" : ""); btrfs_dump_free_space(cache, bytes); spin_unlock(&cache->lock); } @@ -5366,11 +5937,10 @@ int btrfs_reserve_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 num_bytes, u64 min_alloc_size, u64 empty_size, u64 hint_byte, - u64 search_end, struct btrfs_key *ins, - u64 data) + struct btrfs_key *ins, u64 data) { + bool final_tried = false; int ret; - u64 search_start = 0; data = btrfs_get_alloc_profile(root, data); again: @@ -5378,32 +5948,44 @@ again: * the only place that sets empty_size is btrfs_realloc_node, which * is not called recursively on allocations */ - if (empty_size || root->ref_cows) + if (empty_size || root->ref_cows) { ret = do_chunk_alloc(trans, root->fs_info->extent_root, num_bytes + 2 * 1024 * 1024, data, CHUNK_ALLOC_NO_FORCE); + if (ret < 0 && ret != -ENOSPC) { + btrfs_abort_transaction(trans, root, ret); + return ret; + } + } WARN_ON(num_bytes < root->sectorsize); ret = find_free_extent(trans, root, num_bytes, empty_size, - search_start, search_end, hint_byte, - ins, data); - - if (ret == -ENOSPC && num_bytes > min_alloc_size) { - num_bytes = num_bytes >> 1; - num_bytes = num_bytes & ~(root->sectorsize - 1); - num_bytes = max(num_bytes, min_alloc_size); - do_chunk_alloc(trans, root->fs_info->extent_root, - num_bytes, data, CHUNK_ALLOC_FORCE); - goto again; - } - if (ret == -ENOSPC && btrfs_test_opt(root, ENOSPC_DEBUG)) { - struct btrfs_space_info *sinfo; + hint_byte, ins, data); - sinfo = __find_space_info(root->fs_info, data); - printk(KERN_ERR "btrfs allocation failed flags %llu, " - "wanted %llu\n", (unsigned long long)data, - (unsigned long long)num_bytes); - dump_space_info(sinfo, num_bytes, 1); + if (ret == -ENOSPC) { + if (!final_tried) { + num_bytes = num_bytes >> 1; + num_bytes = num_bytes & ~(root->sectorsize - 1); + num_bytes = max(num_bytes, min_alloc_size); + ret = do_chunk_alloc(trans, root->fs_info->extent_root, + num_bytes, data, CHUNK_ALLOC_FORCE); + if (ret < 0 && ret != -ENOSPC) { + btrfs_abort_transaction(trans, root, ret); + return ret; + } + if (num_bytes == min_alloc_size) + final_tried = true; + goto again; + } else if (btrfs_test_opt(root, ENOSPC_DEBUG)) { + struct btrfs_space_info *sinfo; + + sinfo = __find_space_info(root->fs_info, data); + printk(KERN_ERR "btrfs allocation failed flags %llu, " + "wanted %llu\n", (unsigned long long)data, + (unsigned long long)num_bytes); + if (sinfo) + dump_space_info(sinfo, num_bytes, 1); + } } trace_btrfs_reserved_extent_alloc(root, ins->objectid, ins->offset); @@ -5411,7 +5993,8 @@ again: return ret; } -int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) +static int __btrfs_free_reserved_extent(struct btrfs_root *root, + u64 start, u64 len, int pin) { struct btrfs_block_group_cache *cache; int ret = 0; @@ -5426,8 +6009,12 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) if (btrfs_test_opt(root, DISCARD)) ret = btrfs_discard_extent(root, start, len, NULL); - btrfs_add_free_space(cache, start, len); - btrfs_update_reserved_bytes(cache, len, 0, 1); + if (pin) + pin_down_extent(root, cache, start, len, 1); + else { + btrfs_add_free_space(cache, start, len); + btrfs_update_reserved_bytes(cache, len, RESERVE_FREE); + } btrfs_put_block_group(cache); trace_btrfs_reserved_extent_free(root, start, len); @@ -5435,6 +6022,18 @@ int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len) return ret; } +int btrfs_free_reserved_extent(struct btrfs_root *root, + u64 start, u64 len) +{ + return __btrfs_free_reserved_extent(root, start, len, 0); +} + +int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root, + u64 start, u64 len) +{ + return __btrfs_free_reserved_extent(root, start, len, 1); +} + static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 parent, u64 root_objectid, @@ -5464,7 +6063,10 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, path->leave_spinning = 1; ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, ins, size); - BUG_ON(ret); + if (ret) { + btrfs_free_path(path); + return ret; + } leaf = path->nodes[0]; extent_item = btrfs_item_ptr(leaf, path->slots[0], @@ -5494,7 +6096,7 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, btrfs_free_path(path); ret = update_block_group(trans, root, ins->objectid, ins->offset, 1); - if (ret) { + if (ret) { /* -ENOENT, logic error */ printk(KERN_ERR "btrfs update block group failed for %llu " "%llu\n", (unsigned long long)ins->objectid, (unsigned long long)ins->offset); @@ -5525,7 +6127,10 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans, path->leave_spinning = 1; ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, ins, size); - BUG_ON(ret); + if (ret) { + btrfs_free_path(path); + return ret; + } leaf = path->nodes[0]; extent_item = btrfs_item_ptr(leaf, path->slots[0], @@ -5555,7 +6160,7 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans, btrfs_free_path(path); ret = update_block_group(trans, root, ins->objectid, ins->offset, 1); - if (ret) { + if (ret) { /* -ENOENT, logic error */ printk(KERN_ERR "btrfs update block group failed for %llu " "%llu\n", (unsigned long long)ins->objectid, (unsigned long long)ins->offset); @@ -5573,9 +6178,10 @@ int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans, BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID); - ret = btrfs_add_delayed_data_ref(trans, ins->objectid, ins->offset, - 0, root_objectid, owner, offset, - BTRFS_ADD_DELAYED_EXTENT, NULL); + ret = btrfs_add_delayed_data_ref(root->fs_info, trans, ins->objectid, + ins->offset, 0, + root_objectid, owner, offset, + BTRFS_ADD_DELAYED_EXTENT, NULL, 0); return ret; } @@ -5602,36 +6208,37 @@ int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, if (!caching_ctl) { BUG_ON(!block_group_cache_done(block_group)); ret = btrfs_remove_free_space(block_group, start, num_bytes); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } else { mutex_lock(&caching_ctl->mutex); if (start >= caching_ctl->progress) { ret = add_excluded_extent(root, start, num_bytes); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } else if (start + num_bytes <= caching_ctl->progress) { ret = btrfs_remove_free_space(block_group, start, num_bytes); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } else { num_bytes = caching_ctl->progress - start; ret = btrfs_remove_free_space(block_group, start, num_bytes); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ start = caching_ctl->progress; num_bytes = ins->objectid + ins->offset - caching_ctl->progress; ret = add_excluded_extent(root, start, num_bytes); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } mutex_unlock(&caching_ctl->mutex); put_caching_control(caching_ctl); } - ret = btrfs_update_reserved_bytes(block_group, ins->offset, 1, 1); - BUG_ON(ret); + ret = btrfs_update_reserved_bytes(block_group, ins->offset, + RESERVE_ALLOC_NO_ACCOUNT); + BUG_ON(ret); /* logic error */ btrfs_put_block_group(block_group); ret = alloc_reserved_file_extent(trans, root, 0, root_objectid, 0, owner, offset, ins, 1); @@ -5652,6 +6259,7 @@ struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans, btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level); btrfs_tree_lock(buf); clean_tree_block(trans, root, buf); + clear_bit(EXTENT_BUFFER_STALE, &buf->bflags); btrfs_set_lock_blocking(buf); btrfs_set_buffer_uptodate(buf); @@ -5687,8 +6295,7 @@ use_block_rsv(struct btrfs_trans_handle *trans, block_rsv = get_block_rsv(trans, root); if (block_rsv->size == 0) { - ret = reserve_metadata_bytes(trans, root, block_rsv, - blocksize, 0); + ret = reserve_metadata_bytes(root, block_rsv, blocksize, 0); /* * If we couldn't reserve metadata bytes try and use some from * the global reserve. @@ -5708,13 +6315,15 @@ use_block_rsv(struct btrfs_trans_handle *trans, if (!ret) return block_rsv; if (ret) { - WARN_ON(1); - ret = reserve_metadata_bytes(trans, root, block_rsv, blocksize, - 0); + static DEFINE_RATELIMIT_STATE(_rs, + DEFAULT_RATELIMIT_INTERVAL, + /*DEFAULT_RATELIMIT_BURST*/ 2); + if (__ratelimit(&_rs)) { + printk(KERN_DEBUG "btrfs: block rsv returned %d\n", ret); + WARN_ON(1); + } + ret = reserve_metadata_bytes(root, block_rsv, blocksize, 0); if (!ret) { - spin_lock(&block_rsv->lock); - block_rsv->size += blocksize; - spin_unlock(&block_rsv->lock); return block_rsv; } else if (ret && block_rsv != global_rsv) { ret = block_rsv_use_bytes(global_rsv, blocksize); @@ -5726,10 +6335,11 @@ use_block_rsv(struct btrfs_trans_handle *trans, return ERR_PTR(-ENOSPC); } -static void unuse_block_rsv(struct btrfs_block_rsv *block_rsv, u32 blocksize) +static void unuse_block_rsv(struct btrfs_fs_info *fs_info, + struct btrfs_block_rsv *block_rsv, u32 blocksize) { block_rsv_add_bytes(block_rsv, blocksize, 0); - block_rsv_release_bytes(block_rsv, NULL, 0); + block_rsv_release_bytes(fs_info, block_rsv, NULL, 0); } /* @@ -5757,15 +6367,15 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans, return ERR_CAST(block_rsv); ret = btrfs_reserve_extent(trans, root, blocksize, blocksize, - empty_size, hint, (u64)-1, &ins, 0); + empty_size, hint, &ins, 0); if (ret) { - unuse_block_rsv(block_rsv, blocksize); + unuse_block_rsv(root->fs_info, block_rsv, blocksize); return ERR_PTR(ret); } buf = btrfs_init_new_buffer(trans, root, ins.objectid, blocksize, level); - BUG_ON(IS_ERR(buf)); + BUG_ON(IS_ERR(buf)); /* -ENOMEM */ if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) { if (parent == 0) @@ -5777,7 +6387,7 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans, if (root_objectid != BTRFS_TREE_LOG_OBJECTID) { struct btrfs_delayed_extent_op *extent_op; extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS); - BUG_ON(!extent_op); + BUG_ON(!extent_op); /* -ENOMEM */ if (key) memcpy(&extent_op->key, key, sizeof(extent_op->key)); else @@ -5787,11 +6397,12 @@ struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans, extent_op->update_flags = 1; extent_op->is_data = 0; - ret = btrfs_add_delayed_tree_ref(trans, ins.objectid, + ret = btrfs_add_delayed_tree_ref(root->fs_info, trans, + ins.objectid, ins.offset, parent, root_objectid, level, BTRFS_ADD_DELAYED_EXTENT, - extent_op); - BUG_ON(ret); + extent_op, 0); + BUG_ON(ret); /* -ENOMEM */ } return buf; } @@ -5807,6 +6418,7 @@ struct walk_control { int keep_locks; int reada_slot; int reada_count; + int for_reloc; }; #define DROP_REFERENCE 1 @@ -5860,7 +6472,9 @@ static noinline void reada_walk_down(struct btrfs_trans_handle *trans, /* We don't lock the tree block, it's OK to be racy here */ ret = btrfs_lookup_extent_info(trans, root, bytenr, blocksize, &refs, &flags); - BUG_ON(ret); + /* We don't care about errors in readahead. */ + if (ret < 0) + continue; BUG_ON(refs == 0); if (wc->stage == DROP_REFERENCE) { @@ -5927,7 +6541,9 @@ static noinline int walk_down_proc(struct btrfs_trans_handle *trans, eb->start, eb->len, &wc->refs[level], &wc->flags[level]); - BUG_ON(ret); + BUG_ON(ret == -ENOMEM); + if (ret) + return ret; BUG_ON(wc->refs[level] == 0); } @@ -5945,13 +6561,13 @@ static noinline int walk_down_proc(struct btrfs_trans_handle *trans, /* wc->stage == UPDATE_BACKREF */ if (!(wc->flags[level] & flag)) { BUG_ON(!path->locks[level]); - ret = btrfs_inc_ref(trans, root, eb, 1); - BUG_ON(ret); - ret = btrfs_dec_ref(trans, root, eb, 0); - BUG_ON(ret); + ret = btrfs_inc_ref(trans, root, eb, 1, wc->for_reloc); + BUG_ON(ret); /* -ENOMEM */ + ret = btrfs_dec_ref(trans, root, eb, 0, wc->for_reloc); + BUG_ON(ret); /* -ENOMEM */ ret = btrfs_set_disk_extent_flags(trans, root, eb->start, eb->len, flag, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ wc->flags[level] |= flag; } @@ -6023,7 +6639,11 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans, ret = btrfs_lookup_extent_info(trans, root, bytenr, blocksize, &wc->refs[level - 1], &wc->flags[level - 1]); - BUG_ON(ret); + if (ret < 0) { + btrfs_tree_unlock(next); + return ret; + } + BUG_ON(wc->refs[level - 1] == 0); *lookup_info = 0; @@ -6052,7 +6672,7 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans, goto skip; } - if (!btrfs_buffer_uptodate(next, generation)) { + if (!btrfs_buffer_uptodate(next, generation, 0)) { btrfs_tree_unlock(next); free_extent_buffer(next); next = NULL; @@ -6091,8 +6711,8 @@ skip: } ret = btrfs_free_extent(trans, root, bytenr, blocksize, parent, - root->root_key.objectid, level - 1, 0); - BUG_ON(ret); + root->root_key.objectid, level - 1, 0, 0); + BUG_ON(ret); /* -ENOMEM */ } btrfs_tree_unlock(next); free_extent_buffer(next); @@ -6150,7 +6770,10 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans, eb->start, eb->len, &wc->refs[level], &wc->flags[level]); - BUG_ON(ret); + if (ret < 0) { + btrfs_tree_unlock_rw(eb, path->locks[level]); + return ret; + } BUG_ON(wc->refs[level] == 0); if (wc->refs[level] == 1) { btrfs_tree_unlock_rw(eb, path->locks[level]); @@ -6165,10 +6788,12 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans, if (wc->refs[level] == 1) { if (level == 0) { if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) - ret = btrfs_dec_ref(trans, root, eb, 1); + ret = btrfs_dec_ref(trans, root, eb, 1, + wc->for_reloc); else - ret = btrfs_dec_ref(trans, root, eb, 0); - BUG_ON(ret); + ret = btrfs_dec_ref(trans, root, eb, 0, + wc->for_reloc); + BUG_ON(ret); /* -ENOMEM */ } /* make block locked assertion in clean_tree_block happy */ if (!path->locks[level] && @@ -6277,8 +6902,9 @@ static noinline int walk_up_tree(struct btrfs_trans_handle *trans, * also make sure backrefs for the shared block and all lower level * blocks are properly updated. */ -void btrfs_drop_snapshot(struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, int update_ref) +int btrfs_drop_snapshot(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, int update_ref, + int for_reloc) { struct btrfs_path *path; struct btrfs_trans_handle *trans; @@ -6304,7 +6930,10 @@ void btrfs_drop_snapshot(struct btrfs_root *root, } trans = btrfs_start_transaction(tree_root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + err = PTR_ERR(trans); + goto out_free; + } if (block_rsv) trans->block_rsv = block_rsv; @@ -6329,7 +6958,7 @@ void btrfs_drop_snapshot(struct btrfs_root *root, path->lowest_level = 0; if (ret < 0) { err = ret; - goto out_free; + goto out_end_trans; } WARN_ON(ret > 0); @@ -6349,7 +6978,10 @@ void btrfs_drop_snapshot(struct btrfs_root *root, path->nodes[level]->len, &wc->refs[level], &wc->flags[level]); - BUG_ON(ret); + if (ret < 0) { + err = ret; + goto out_end_trans; + } BUG_ON(wc->refs[level] == 0); if (level == root_item->drop_level) @@ -6366,6 +6998,7 @@ void btrfs_drop_snapshot(struct btrfs_root *root, wc->stage = DROP_REFERENCE; wc->update_ref = update_ref; wc->keep_locks = 0; + wc->for_reloc = for_reloc; wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root); while (1) { @@ -6399,26 +7032,40 @@ void btrfs_drop_snapshot(struct btrfs_root *root, ret = btrfs_update_root(trans, tree_root, &root->root_key, root_item); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, tree_root, ret); + err = ret; + goto out_end_trans; + } btrfs_end_transaction_throttle(trans, tree_root); trans = btrfs_start_transaction(tree_root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + err = PTR_ERR(trans); + goto out_free; + } if (block_rsv) trans->block_rsv = block_rsv; } } btrfs_release_path(path); - BUG_ON(err); + if (err) + goto out_end_trans; ret = btrfs_del_root(trans, tree_root, &root->root_key); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, tree_root, ret); + goto out_end_trans; + } if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) { ret = btrfs_find_last_root(tree_root, root->root_key.objectid, NULL, NULL); - BUG_ON(ret < 0); - if (ret > 0) { + if (ret < 0) { + btrfs_abort_transaction(trans, tree_root, ret); + err = ret; + goto out_end_trans; + } else if (ret > 0) { /* if we fail to delete the orphan item this time * around, it'll get picked up the next time. * @@ -6436,20 +7083,22 @@ void btrfs_drop_snapshot(struct btrfs_root *root, free_extent_buffer(root->commit_root); kfree(root); } -out_free: +out_end_trans: btrfs_end_transaction_throttle(trans, tree_root); +out_free: kfree(wc); btrfs_free_path(path); out: if (err) btrfs_std_error(root->fs_info, err); - return; + return err; } /* * drop subtree rooted at tree block 'node'. * * NOTE: this function will unlock and release tree block 'node' + * only used by relocation code */ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -6494,6 +7143,7 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, wc->stage = DROP_REFERENCE; wc->update_ref = 0; wc->keep_locks = 1; + wc->for_reloc = 1; wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root); while (1) { @@ -6518,8 +7168,15 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) { u64 num_devices; - u64 stripped = BTRFS_BLOCK_GROUP_RAID0 | - BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10; + u64 stripped; + + /* + * if restripe for this chunk_type is on pick target profile and + * return, otherwise do the usual balance + */ + stripped = get_restripe_target(root->fs_info, flags); + if (stripped) + return extended_to_chunk(stripped); /* * we add in the count of missing devices because we want @@ -6529,6 +7186,9 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) num_devices = root->fs_info->fs_devices->rw_devices + root->fs_info->fs_devices->missing_devices; + stripped = BTRFS_BLOCK_GROUP_RAID0 | + BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10; + if (num_devices == 1) { stripped |= BTRFS_BLOCK_GROUP_DUP; stripped = flags & ~stripped; @@ -6541,7 +7201,6 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) if (flags & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10)) return stripped | BTRFS_BLOCK_GROUP_DUP; - return flags; } else { /* they already had raid on here, just return */ if (flags & stripped) @@ -6554,9 +7213,9 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) if (flags & BTRFS_BLOCK_GROUP_DUP) return stripped | BTRFS_BLOCK_GROUP_RAID1; - /* turn single device chunks into raid0 */ - return stripped | BTRFS_BLOCK_GROUP_RAID0; + /* this is drive concat, leave it alone */ } + return flags; } @@ -6592,12 +7251,9 @@ static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force) cache->bytes_super - btrfs_block_group_used(&cache->item); if (sinfo->bytes_used + sinfo->bytes_reserved + sinfo->bytes_pinned + - sinfo->bytes_may_use + sinfo->bytes_readonly + - cache->reserved_pinned + num_bytes + min_allocable_bytes <= - sinfo->total_bytes) { + sinfo->bytes_may_use + sinfo->bytes_readonly + num_bytes + + min_allocable_bytes <= sinfo->total_bytes) { sinfo->bytes_readonly += num_bytes; - sinfo->bytes_reserved += cache->reserved_pinned; - cache->reserved_pinned = 0; cache->ro = 1; ret = 0; } @@ -6618,12 +7274,16 @@ int btrfs_set_block_group_ro(struct btrfs_root *root, BUG_ON(cache->ro); trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); alloc_flags = update_block_group_flags(root, cache->flags); - if (alloc_flags != cache->flags) - do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, - CHUNK_ALLOC_FORCE); + if (alloc_flags != cache->flags) { + ret = do_chunk_alloc(trans, root, 2 * 1024 * 1024, alloc_flags, + CHUNK_ALLOC_FORCE); + if (ret < 0) + goto out; + } ret = set_block_group_ro(cache, 0); if (!ret) @@ -6703,7 +7363,7 @@ u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo) return free_bytes; } -int btrfs_set_block_group_rw(struct btrfs_root *root, +void btrfs_set_block_group_rw(struct btrfs_root *root, struct btrfs_block_group_cache *cache) { struct btrfs_space_info *sinfo = cache->space_info; @@ -6719,7 +7379,6 @@ int btrfs_set_block_group_rw(struct btrfs_root *root, cache->ro = 0; spin_unlock(&cache->lock); spin_unlock(&sinfo->lock); - return 0; } /* @@ -6737,6 +7396,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) u64 min_free; u64 dev_min = 1; u64 dev_nr = 0; + u64 target; int index; int full = 0; int ret = 0; @@ -6777,13 +7437,11 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) /* * ok we don't have enough space, but maybe we have free space on our * devices to allocate new chunks for relocation, so loop through our - * alloc devices and guess if we have enough space. However, if we - * were marked as full, then we know there aren't enough chunks, and we - * can just return. + * alloc devices and guess if we have enough space. if this block + * group is going to be restriped, run checks against the target + * profile instead of the current one. */ ret = -1; - if (full) - goto out; /* * index: @@ -6793,7 +7451,20 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) * 3: raid0 * 4: single */ - index = get_block_group_index(block_group); + target = get_restripe_target(root->fs_info, block_group->flags); + if (target) { + index = __get_block_group_index(extended_to_chunk(target)); + } else { + /* + * this is just a balance, so if we were marked as full + * we know there is no space for a new chunk + */ + if (full) + goto out; + + index = get_block_group_index(block_group); + } + if (index == 0) { dev_min = 4; /* Divide by 2 */ @@ -6817,7 +7488,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) * space to fit our block group in. */ if (device->total_bytes > device->bytes_used + min_free) { - ret = find_free_dev_extent(NULL, device, min_free, + ret = find_free_dev_extent(device, min_free, &dev_offset, NULL); if (!ret) dev_nr++; @@ -6964,7 +7635,8 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info) struct btrfs_space_info, list); if (space_info->bytes_pinned > 0 || - space_info->bytes_reserved > 0) { + space_info->bytes_reserved > 0 || + space_info->bytes_may_use > 0) { WARN_ON(1); dump_space_info(space_info, 0, 0); } @@ -7006,14 +7678,12 @@ int btrfs_read_block_groups(struct btrfs_root *root) return -ENOMEM; path->reada = 1; - cache_gen = btrfs_super_cache_generation(&root->fs_info->super_copy); - if (cache_gen != 0 && - btrfs_super_generation(&root->fs_info->super_copy) != cache_gen) + cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy); + if (btrfs_test_opt(root, SPACE_CACHE) && + btrfs_super_generation(root->fs_info->super_copy) != cache_gen) need_clear = 1; if (btrfs_test_opt(root, CLEAR_CACHE)) need_clear = 1; - if (!btrfs_test_opt(root, SPACE_CACHE) && cache_gen) - printk(KERN_INFO "btrfs: disk space caching is enabled\n"); while (1) { ret = find_first_block_group(root, path, &key); @@ -7042,8 +7712,21 @@ int btrfs_read_block_groups(struct btrfs_root *root) INIT_LIST_HEAD(&cache->list); INIT_LIST_HEAD(&cache->cluster_list); - if (need_clear) + if (need_clear) { + /* + * When we mount with old space cache, we need to + * set BTRFS_DC_CLEAR and set dirty flag. + * + * a) Setting 'BTRFS_DC_CLEAR' makes sure that we + * truncate the old free space cache inode and + * setup a new one. + * b) Setting 'dirty flag' makes sure that we flush + * the new space cache info onto disk. + */ cache->disk_cache_state = BTRFS_DC_CLEAR; + if (btrfs_test_opt(root, SPACE_CACHE)) + cache->dirty = 1; + } read_extent_buffer(leaf, &cache->item, btrfs_item_ptr_offset(leaf, path->slots[0]), @@ -7088,7 +7771,7 @@ int btrfs_read_block_groups(struct btrfs_root *root) ret = update_space_info(info, cache->flags, found_key.offset, btrfs_block_group_used(&cache->item), &space_info); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ cache->space_info = space_info; spin_lock(&cache->space_info->lock); cache->space_info->bytes_readonly += cache->bytes_super; @@ -7097,7 +7780,7 @@ int btrfs_read_block_groups(struct btrfs_root *root) __link_block_group(space_info, cache); ret = btrfs_add_block_group_cache(root->fs_info, cache); - BUG_ON(ret); + BUG_ON(ret); /* Logic error */ set_avail_alloc_bits(root->fs_info, cache->flags); if (btrfs_chunk_readonly(root, cache->key.objectid)) @@ -7179,7 +7862,8 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, ret = update_space_info(root->fs_info, cache->flags, size, bytes_used, &cache->space_info); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ + update_global_block_rsv(root->fs_info); spin_lock(&cache->space_info->lock); cache->space_info->bytes_readonly += cache->bytes_super; @@ -7188,17 +7872,33 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, __link_block_group(cache->space_info, cache); ret = btrfs_add_block_group_cache(root->fs_info, cache); - BUG_ON(ret); + BUG_ON(ret); /* Logic error */ ret = btrfs_insert_item(trans, extent_root, &cache->key, &cache->item, sizeof(cache->item)); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, extent_root, ret); + return ret; + } set_avail_alloc_bits(extent_root->fs_info, type); return 0; } +static void clear_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags) +{ + u64 extra_flags = chunk_to_extended(flags) & + BTRFS_EXTENDED_PROFILE_MASK; + + if (flags & BTRFS_BLOCK_GROUP_DATA) + fs_info->avail_data_alloc_bits &= ~extra_flags; + if (flags & BTRFS_BLOCK_GROUP_METADATA) + fs_info->avail_metadata_alloc_bits &= ~extra_flags; + if (flags & BTRFS_BLOCK_GROUP_SYSTEM) + fs_info->avail_system_alloc_bits &= ~extra_flags; +} + int btrfs_remove_block_group(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 group_start) { @@ -7209,6 +7909,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, struct btrfs_key key; struct inode *inode; int ret; + int index; int factor; root = root->fs_info->extent_root; @@ -7224,6 +7925,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, free_excluded_extents(root, block_group); memcpy(&key, &block_group->key, sizeof(key)); + index = get_block_group_index(block_group); if (block_group->flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10)) @@ -7252,10 +7954,13 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, goto out; } - inode = lookup_free_space_inode(root, block_group, path); + inode = lookup_free_space_inode(tree_root, block_group, path); if (!IS_ERR(inode)) { ret = btrfs_orphan_add(trans, inode); - BUG_ON(ret); + if (ret) { + btrfs_add_delayed_iput(inode); + goto out; + } clear_nlink(inode); /* One for the block groups ref */ spin_lock(&block_group->lock); @@ -7268,7 +7973,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, spin_unlock(&block_group->lock); } /* One for our lookup ref */ - iput(inode); + btrfs_add_delayed_iput(inode); } key.objectid = BTRFS_FREE_SPACE_OBJECTID; @@ -7298,6 +8003,8 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, * are still on the list after taking the semaphore */ list_del_init(&block_group->list); + if (list_empty(&block_group->space_info->block_groups[index])) + clear_avail_alloc_bits(root->fs_info, block_group->flags); up_write(&block_group->space_info->groups_sem); if (block_group->cached == BTRFS_CACHE_STARTED) @@ -7339,7 +8046,7 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info) int mixed = 0; int ret; - disk_super = &fs_info->super_copy; + disk_super = fs_info->super_copy; if (!btrfs_super_root(disk_super)) return 1; @@ -7387,9 +8094,16 @@ int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range) u64 start; u64 end; u64 trimmed = 0; + u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy); int ret = 0; - cache = btrfs_lookup_block_group(fs_info, range->start); + /* + * try to trim all FS space, our block group may start from non-zero. + */ + if (range->len == total_bytes) + cache = btrfs_lookup_first_block_group(fs_info, range->start); + else + cache = btrfs_lookup_block_group(fs_info, range->start); while (cache) { if (cache->key.objectid >= (range->start + range->len)) { diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index d418164a35f1..4c878476bb91 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -17,6 +17,10 @@ #include "compat.h" #include "ctree.h" #include "btrfs_inode.h" +#include "volumes.h" +#include "check-integrity.h" +#include "locking.h" +#include "rcu-string.h" static struct kmem_cache *extent_state_cache; static struct kmem_cache *extent_buffer_cache; @@ -51,6 +55,13 @@ struct extent_page_data { unsigned int sync_io:1; }; +static noinline void flush_write_bio(void *data); +static inline struct btrfs_fs_info * +tree_fs_info(struct extent_io_tree *tree) +{ + return btrfs_sb(tree->mapping->host->i_sb); +} + int __init extent_io_init(void) { extent_state_cache = kmem_cache_create("extent_state", @@ -134,6 +145,7 @@ static struct extent_state *alloc_extent_state(gfp_t mask) #endif atomic_set(&state->refs, 1); init_waitqueue_head(&state->wq); + trace_alloc_extent_state(state, mask, _RET_IP_); return state; } @@ -151,6 +163,7 @@ void free_extent_state(struct extent_state *state) list_del(&state->leak_list); spin_unlock_irqrestore(&leak_lock, flags); #endif + trace_free_extent_state(state, _RET_IP_); kmem_cache_free(extent_state_cache, state); } } @@ -174,7 +187,6 @@ static struct rb_node *tree_insert(struct rb_root *root, u64 offset, return parent; } - entry = rb_entry(node, struct tree_entry, rb_node); rb_link_node(node, parent, p); rb_insert_color(node, root); return NULL; @@ -390,20 +402,28 @@ static int split_state(struct extent_io_tree *tree, struct extent_state *orig, return 0; } +static struct extent_state *next_state(struct extent_state *state) +{ + struct rb_node *next = rb_next(&state->rb_node); + if (next) + return rb_entry(next, struct extent_state, rb_node); + else + return NULL; +} + /* * utility function to clear some bits in an extent state struct. - * it will optionally wake up any one waiting on this state (wake == 1), or - * forcibly remove the state from the tree (delete == 1). + * it will optionally wake up any one waiting on this state (wake == 1). * * If no bits are set on the state struct after clearing things, the * struct is freed and removed from the tree */ -static int clear_state_bit(struct extent_io_tree *tree, - struct extent_state *state, - int *bits, int wake) +static struct extent_state *clear_state_bit(struct extent_io_tree *tree, + struct extent_state *state, + int *bits, int wake) { + struct extent_state *next; int bits_to_clear = *bits & ~EXTENT_CTLBITS; - int ret = state->state & bits_to_clear; if ((bits_to_clear & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) { u64 range = state->end - state->start + 1; @@ -415,6 +435,7 @@ static int clear_state_bit(struct extent_io_tree *tree, if (wake) wake_up(&state->wq); if (state->state == 0) { + next = next_state(state); if (state->tree) { rb_erase(&state->rb_node, &tree->state); state->tree = NULL; @@ -424,8 +445,9 @@ static int clear_state_bit(struct extent_io_tree *tree, } } else { merge_state(tree, state); + next = next_state(state); } - return ret; + return next; } static struct extent_state * @@ -437,6 +459,13 @@ alloc_extent_state_atomic(struct extent_state *prealloc) return prealloc; } +void extent_io_tree_panic(struct extent_io_tree *tree, int err) +{ + btrfs_panic(tree_fs_info(tree), err, "Locking error: " + "Extent tree was modified by another " + "thread while locked."); +} + /* * clear some bits on a range in the tree. This may require splitting * or inserting elements in the tree, so the gfp mask is used to @@ -447,8 +476,7 @@ alloc_extent_state_atomic(struct extent_state *prealloc) * * the range [start, end] is inclusive. * - * This takes the tree lock, and returns < 0 on error, > 0 if any of the - * bits were already set, or zero if none of the bits were already set. + * This takes the tree lock, and returns 0 on success and < 0 on error. */ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits, int wake, int delete, @@ -458,11 +486,9 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state *state; struct extent_state *cached; struct extent_state *prealloc = NULL; - struct rb_node *next_node; struct rb_node *node; u64 last_end; int err; - int set = 0; int clear = 0; if (delete) @@ -511,6 +537,12 @@ hit_next: WARN_ON(state->end < start); last_end = state->end; + /* the state doesn't have the wanted bits, go ahead */ + if (!(state->state & bits)) { + state = next_state(state); + goto next; + } + /* * | ---- desired range ---- | * | state | or @@ -531,15 +563,15 @@ hit_next: prealloc = alloc_extent_state_atomic(prealloc); BUG_ON(!prealloc); err = split_state(tree, state, prealloc, start); - BUG_ON(err == -EEXIST); + if (err) + extent_io_tree_panic(tree, err); + prealloc = NULL; if (err) goto out; if (state->end <= end) { - set |= clear_state_bit(tree, state, &bits, wake); - if (last_end == (u64)-1) - goto out; - start = last_end + 1; + state = clear_state_bit(tree, state, &bits, wake); + goto next; } goto search_again; } @@ -553,31 +585,25 @@ hit_next: prealloc = alloc_extent_state_atomic(prealloc); BUG_ON(!prealloc); err = split_state(tree, state, prealloc, end + 1); - BUG_ON(err == -EEXIST); + if (err) + extent_io_tree_panic(tree, err); + if (wake) wake_up(&state->wq); - set |= clear_state_bit(tree, prealloc, &bits, wake); + clear_state_bit(tree, prealloc, &bits, wake); prealloc = NULL; goto out; } - if (state->end < end && prealloc && !need_resched()) - next_node = rb_next(&state->rb_node); - else - next_node = NULL; - - set |= clear_state_bit(tree, state, &bits, wake); + state = clear_state_bit(tree, state, &bits, wake); +next: if (last_end == (u64)-1) goto out; start = last_end + 1; - if (start <= end && next_node) { - state = rb_entry(next_node, struct extent_state, - rb_node); - if (state->start == start) - goto hit_next; - } + if (start <= end && state && !need_resched()) + goto hit_next; goto search_again; out: @@ -585,7 +611,7 @@ out: if (prealloc) free_extent_state(prealloc); - return set; + return 0; search_again: if (start > end) @@ -596,8 +622,8 @@ search_again: goto again; } -static int wait_on_state(struct extent_io_tree *tree, - struct extent_state *state) +static void wait_on_state(struct extent_io_tree *tree, + struct extent_state *state) __releases(tree->lock) __acquires(tree->lock) { @@ -607,7 +633,6 @@ static int wait_on_state(struct extent_io_tree *tree, schedule(); spin_lock(&tree->lock); finish_wait(&state->wq, &wait); - return 0; } /* @@ -615,7 +640,7 @@ static int wait_on_state(struct extent_io_tree *tree, * The range [start, end] is inclusive. * The tree lock is taken by this function */ -int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits) +void wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits) { struct extent_state *state; struct rb_node *node; @@ -652,7 +677,6 @@ again: } out: spin_unlock(&tree->lock); - return 0; } static void set_state_bits(struct extent_io_tree *tree, @@ -700,9 +724,10 @@ static void uncache_state(struct extent_state **cached_ptr) * [start, end] is inclusive This takes the tree lock. */ -int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, - int bits, int exclusive_bits, u64 *failed_start, - struct extent_state **cached_state, gfp_t mask) +static int __must_check +__set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, + int bits, int exclusive_bits, u64 *failed_start, + struct extent_state **cached_state, gfp_t mask) { struct extent_state *state; struct extent_state *prealloc = NULL; @@ -736,8 +761,10 @@ again: prealloc = alloc_extent_state_atomic(prealloc); BUG_ON(!prealloc); err = insert_state(tree, prealloc, start, end, &bits); + if (err) + extent_io_tree_panic(tree, err); + prealloc = NULL; - BUG_ON(err == -EEXIST); goto out; } state = rb_entry(node, struct extent_state, rb_node); @@ -752,7 +779,6 @@ hit_next: * Just lock what we found and keep going */ if (state->start == start && state->end <= end) { - struct rb_node *next_node; if (state->state & exclusive_bits) { *failed_start = state->start; err = -EEXIST; @@ -760,20 +786,15 @@ hit_next: } set_state_bits(tree, state, &bits); - cache_state(state, cached_state); merge_state(tree, state); if (last_end == (u64)-1) goto out; - start = last_end + 1; - next_node = rb_next(&state->rb_node); - if (next_node && start < end && prealloc && !need_resched()) { - state = rb_entry(next_node, struct extent_state, - rb_node); - if (state->start == start) - goto hit_next; - } + state = next_state(state); + if (start < end && state && state->start == start && + !need_resched()) + goto hit_next; goto search_again; } @@ -803,7 +824,9 @@ hit_next: prealloc = alloc_extent_state_atomic(prealloc); BUG_ON(!prealloc); err = split_state(tree, state, prealloc, start); - BUG_ON(err == -EEXIST); + if (err) + extent_io_tree_panic(tree, err); + prealloc = NULL; if (err) goto out; @@ -814,6 +837,10 @@ hit_next: if (last_end == (u64)-1) goto out; start = last_end + 1; + state = next_state(state); + if (start < end && state && state->start == start && + !need_resched()) + goto hit_next; } goto search_again; } @@ -840,12 +867,9 @@ hit_next: */ err = insert_state(tree, prealloc, start, this_end, &bits); - BUG_ON(err == -EEXIST); - if (err) { - free_extent_state(prealloc); - prealloc = NULL; - goto out; - } + if (err) + extent_io_tree_panic(tree, err); + cache_state(prealloc, cached_state); prealloc = NULL; start = this_end + 1; @@ -867,7 +891,8 @@ hit_next: prealloc = alloc_extent_state_atomic(prealloc); BUG_ON(!prealloc); err = split_state(tree, state, prealloc, end + 1); - BUG_ON(err == -EEXIST); + if (err) + extent_io_tree_panic(tree, err); set_state_bits(tree, prealloc, &bits); cache_state(prealloc, cached_state); @@ -894,18 +919,214 @@ search_again: goto again; } +int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits, + u64 *failed_start, struct extent_state **cached_state, + gfp_t mask) +{ + return __set_extent_bit(tree, start, end, bits, 0, failed_start, + cached_state, mask); +} + + +/** + * convert_extent_bit - convert all bits in a given range from one bit to + * another + * @tree: the io tree to search + * @start: the start offset in bytes + * @end: the end offset in bytes (inclusive) + * @bits: the bits to set in this range + * @clear_bits: the bits to clear in this range + * @mask: the allocation mask + * + * This will go through and set bits for the given range. If any states exist + * already in this range they are set with the given bit and cleared of the + * clear_bits. This is only meant to be used by things that are mergeable, ie + * converting from say DELALLOC to DIRTY. This is not meant to be used with + * boundary bits like LOCK. + */ +int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, + int bits, int clear_bits, gfp_t mask) +{ + struct extent_state *state; + struct extent_state *prealloc = NULL; + struct rb_node *node; + int err = 0; + u64 last_start; + u64 last_end; + +again: + if (!prealloc && (mask & __GFP_WAIT)) { + prealloc = alloc_extent_state(mask); + if (!prealloc) + return -ENOMEM; + } + + spin_lock(&tree->lock); + /* + * this search will find all the extents that end after + * our range starts. + */ + node = tree_search(tree, start); + if (!node) { + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + err = insert_state(tree, prealloc, start, end, &bits); + prealloc = NULL; + if (err) + extent_io_tree_panic(tree, err); + goto out; + } + state = rb_entry(node, struct extent_state, rb_node); +hit_next: + last_start = state->start; + last_end = state->end; + + /* + * | ---- desired range ---- | + * | state | + * + * Just lock what we found and keep going + */ + if (state->start == start && state->end <= end) { + set_state_bits(tree, state, &bits); + state = clear_state_bit(tree, state, &clear_bits, 0); + if (last_end == (u64)-1) + goto out; + start = last_end + 1; + if (start < end && state && state->start == start && + !need_resched()) + goto hit_next; + goto search_again; + } + + /* + * | ---- desired range ---- | + * | state | + * or + * | ------------- state -------------- | + * + * We need to split the extent we found, and may flip bits on + * second half. + * + * If the extent we found extends past our + * range, we just split and search again. It'll get split + * again the next time though. + * + * If the extent we found is inside our range, we set the + * desired bit on it. + */ + if (state->start < start) { + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + err = split_state(tree, state, prealloc, start); + if (err) + extent_io_tree_panic(tree, err); + prealloc = NULL; + if (err) + goto out; + if (state->end <= end) { + set_state_bits(tree, state, &bits); + state = clear_state_bit(tree, state, &clear_bits, 0); + if (last_end == (u64)-1) + goto out; + start = last_end + 1; + if (start < end && state && state->start == start && + !need_resched()) + goto hit_next; + } + goto search_again; + } + /* + * | ---- desired range ---- | + * | state | or | state | + * + * There's a hole, we need to insert something in it and + * ignore the extent we found. + */ + if (state->start > start) { + u64 this_end; + if (end < last_start) + this_end = end; + else + this_end = last_start - 1; + + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + + /* + * Avoid to free 'prealloc' if it can be merged with + * the later extent. + */ + err = insert_state(tree, prealloc, start, this_end, + &bits); + if (err) + extent_io_tree_panic(tree, err); + prealloc = NULL; + start = this_end + 1; + goto search_again; + } + /* + * | ---- desired range ---- | + * | state | + * We need to split the extent, and set the bit + * on the first half + */ + if (state->start <= end && state->end > end) { + prealloc = alloc_extent_state_atomic(prealloc); + if (!prealloc) { + err = -ENOMEM; + goto out; + } + + err = split_state(tree, state, prealloc, end + 1); + if (err) + extent_io_tree_panic(tree, err); + + set_state_bits(tree, prealloc, &bits); + clear_state_bit(tree, prealloc, &clear_bits, 0); + prealloc = NULL; + goto out; + } + + goto search_again; + +out: + spin_unlock(&tree->lock); + if (prealloc) + free_extent_state(prealloc); + + return err; + +search_again: + if (start > end) + goto out; + spin_unlock(&tree->lock); + if (mask & __GFP_WAIT) + cond_resched(); + goto again; +} + /* wrappers around set/clear extent bit */ int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask) { - return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL, + return set_extent_bit(tree, start, end, EXTENT_DIRTY, NULL, NULL, mask); } int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, int bits, gfp_t mask) { - return set_extent_bit(tree, start, end, bits, 0, NULL, + return set_extent_bit(tree, start, end, bits, NULL, NULL, mask); } @@ -919,8 +1140,8 @@ int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask) { return set_extent_bit(tree, start, end, - EXTENT_DELALLOC | EXTENT_DIRTY | EXTENT_UPTODATE, - 0, NULL, cached_state, mask); + EXTENT_DELALLOC | EXTENT_UPTODATE, + NULL, cached_state, mask); } int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, @@ -934,7 +1155,7 @@ int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask) { - return set_extent_bit(tree, start, end, EXTENT_NEW, 0, NULL, + return set_extent_bit(tree, start, end, EXTENT_NEW, NULL, NULL, mask); } @@ -942,12 +1163,11 @@ int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask) { return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, - NULL, cached_state, mask); + cached_state, mask); } -static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, - u64 end, struct extent_state **cached_state, - gfp_t mask) +int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, + struct extent_state **cached_state, gfp_t mask) { return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0, cached_state, mask); @@ -958,42 +1178,40 @@ static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, * us if waiting is desired. */ int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - int bits, struct extent_state **cached_state, gfp_t mask) + int bits, struct extent_state **cached_state) { int err; u64 failed_start; while (1) { - err = set_extent_bit(tree, start, end, EXTENT_LOCKED | bits, - EXTENT_LOCKED, &failed_start, - cached_state, mask); - if (err == -EEXIST && (mask & __GFP_WAIT)) { + err = __set_extent_bit(tree, start, end, EXTENT_LOCKED | bits, + EXTENT_LOCKED, &failed_start, + cached_state, GFP_NOFS); + if (err == -EEXIST) { wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED); start = failed_start; - } else { + } else break; - } WARN_ON(start > end); } return err; } -int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask) +int lock_extent(struct extent_io_tree *tree, u64 start, u64 end) { - return lock_extent_bits(tree, start, end, 0, NULL, mask); + return lock_extent_bits(tree, start, end, 0, NULL); } -int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask) +int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end) { int err; u64 failed_start; - err = set_extent_bit(tree, start, end, EXTENT_LOCKED, EXTENT_LOCKED, - &failed_start, NULL, mask); + err = __set_extent_bit(tree, start, end, EXTENT_LOCKED, EXTENT_LOCKED, + &failed_start, NULL, GFP_NOFS); if (err == -EEXIST) { if (failed_start > start) clear_extent_bit(tree, start, failed_start - 1, - EXTENT_LOCKED, 1, 0, NULL, mask); + EXTENT_LOCKED, 1, 0, NULL, GFP_NOFS); return 0; } return 1; @@ -1006,10 +1224,10 @@ int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end, mask); } -int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask) +int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end) { return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, NULL, - mask); + GFP_NOFS); } /* @@ -1023,7 +1241,7 @@ static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end) while (index <= end_index) { page = find_get_page(tree->mapping, index); - BUG_ON(!page); + BUG_ON(!page); /* Pages should be in the extent_io_tree */ set_page_writeback(page); page_cache_release(page); index++; @@ -1067,7 +1285,7 @@ out: * returned if we find something, and *start_ret and *end_ret are * set to reflect the state struct that was found. * - * If nothing was found, 1 is returned, < 0 on error + * If nothing was found, 1 is returned. If found something, return 0. */ int find_first_extent_bit(struct extent_io_tree *tree, u64 start, u64 *start_ret, u64 *end_ret, int bits) @@ -1146,9 +1364,9 @@ out: return found; } -static noinline int __unlock_for_delalloc(struct inode *inode, - struct page *locked_page, - u64 start, u64 end) +static noinline void __unlock_for_delalloc(struct inode *inode, + struct page *locked_page, + u64 start, u64 end) { int ret; struct page *pages[16]; @@ -1158,7 +1376,7 @@ static noinline int __unlock_for_delalloc(struct inode *inode, int i; if (index == locked_page->index && end_index == index) - return 0; + return; while (nr_pages > 0) { ret = find_get_pages_contig(inode->i_mapping, index, @@ -1173,7 +1391,6 @@ static noinline int __unlock_for_delalloc(struct inode *inode, index += ret; cond_resched(); } - return 0; } static noinline int lock_delalloc_pages(struct inode *inode, @@ -1303,11 +1520,10 @@ again: goto out_failed; } } - BUG_ON(ret); + BUG_ON(ret); /* Only valid values are 0 and -EAGAIN */ /* step three, lock the state bits for the whole range */ - lock_extent_bits(tree, delalloc_start, delalloc_end, - 0, &cached_state, GFP_NOFS); + lock_extent_bits(tree, delalloc_start, delalloc_end, 0, &cached_state); /* then test to make sure it is all still delalloc */ ret = test_range_bit(tree, delalloc_start, delalloc_end, @@ -1564,43 +1780,448 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end, * helper function to set a given page up to date if all the * extents in the tree for that page are up to date */ -static int check_page_uptodate(struct extent_io_tree *tree, - struct page *page) +static void check_page_uptodate(struct extent_io_tree *tree, struct page *page) { u64 start = (u64)page->index << PAGE_CACHE_SHIFT; u64 end = start + PAGE_CACHE_SIZE - 1; if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL)) SetPageUptodate(page); - return 0; } /* * helper function to unlock a page if all the extents in the tree * for that page are unlocked */ -static int check_page_locked(struct extent_io_tree *tree, - struct page *page) +static void check_page_locked(struct extent_io_tree *tree, struct page *page) { u64 start = (u64)page->index << PAGE_CACHE_SHIFT; u64 end = start + PAGE_CACHE_SIZE - 1; if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0, NULL)) unlock_page(page); - return 0; } /* * helper function to end page writeback if all the extents * in the tree for that page are done with writeback */ -static int check_page_writeback(struct extent_io_tree *tree, - struct page *page) +static void check_page_writeback(struct extent_io_tree *tree, + struct page *page) { end_page_writeback(page); +} + +/* + * When IO fails, either with EIO or csum verification fails, we + * try other mirrors that might have a good copy of the data. This + * io_failure_record is used to record state as we go through all the + * mirrors. If another mirror has good data, the page is set up to date + * and things continue. If a good mirror can't be found, the original + * bio end_io callback is called to indicate things have failed. + */ +struct io_failure_record { + struct page *page; + u64 start; + u64 len; + u64 logical; + unsigned long bio_flags; + int this_mirror; + int failed_mirror; + int in_validation; +}; + +static int free_io_failure(struct inode *inode, struct io_failure_record *rec, + int did_repair) +{ + int ret; + int err = 0; + struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree; + + set_state_private(failure_tree, rec->start, 0); + ret = clear_extent_bits(failure_tree, rec->start, + rec->start + rec->len - 1, + EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); + if (ret) + err = ret; + + if (did_repair) { + ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start, + rec->start + rec->len - 1, + EXTENT_DAMAGED, GFP_NOFS); + if (ret && !err) + err = ret; + } + + kfree(rec); + return err; +} + +static void repair_io_failure_callback(struct bio *bio, int err) +{ + complete(bio->bi_private); +} + +/* + * this bypasses the standard btrfs submit functions deliberately, as + * the standard behavior is to write all copies in a raid setup. here we only + * want to write the one bad copy. so we do the mapping for ourselves and issue + * submit_bio directly. + * to avoid any synchonization issues, wait for the data after writing, which + * actually prevents the read that triggered the error from finishing. + * currently, there can be no more than two copies of every data bit. thus, + * exactly one rewrite is required. + */ +int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start, + u64 length, u64 logical, struct page *page, + int mirror_num) +{ + struct bio *bio; + struct btrfs_device *dev; + DECLARE_COMPLETION_ONSTACK(compl); + u64 map_length = 0; + u64 sector; + struct btrfs_bio *bbio = NULL; + int ret; + + BUG_ON(!mirror_num); + + bio = bio_alloc(GFP_NOFS, 1); + if (!bio) + return -EIO; + bio->bi_private = &compl; + bio->bi_end_io = repair_io_failure_callback; + bio->bi_size = 0; + map_length = length; + + ret = btrfs_map_block(map_tree, WRITE, logical, + &map_length, &bbio, mirror_num); + if (ret) { + bio_put(bio); + return -EIO; + } + BUG_ON(mirror_num != bbio->mirror_num); + sector = bbio->stripes[mirror_num-1].physical >> 9; + bio->bi_sector = sector; + dev = bbio->stripes[mirror_num-1].dev; + kfree(bbio); + if (!dev || !dev->bdev || !dev->writeable) { + bio_put(bio); + return -EIO; + } + bio->bi_bdev = dev->bdev; + bio_add_page(bio, page, length, start-page_offset(page)); + btrfsic_submit_bio(WRITE_SYNC, bio); + wait_for_completion(&compl); + + if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) { + /* try to remap that extent elsewhere? */ + bio_put(bio); + btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_WRITE_ERRS); + return -EIO; + } + + printk_ratelimited_in_rcu(KERN_INFO "btrfs read error corrected: ino %lu off %llu " + "(dev %s sector %llu)\n", page->mapping->host->i_ino, + start, rcu_str_deref(dev->name), sector); + + bio_put(bio); return 0; } +int repair_eb_io_failure(struct btrfs_root *root, struct extent_buffer *eb, + int mirror_num) +{ + struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree; + u64 start = eb->start; + unsigned long i, num_pages = num_extent_pages(eb->start, eb->len); + int ret = 0; + + for (i = 0; i < num_pages; i++) { + struct page *p = extent_buffer_page(eb, i); + ret = repair_io_failure(map_tree, start, PAGE_CACHE_SIZE, + start, p, mirror_num); + if (ret) + break; + start += PAGE_CACHE_SIZE; + } + + return ret; +} + +/* + * each time an IO finishes, we do a fast check in the IO failure tree + * to see if we need to process or clean up an io_failure_record + */ +static int clean_io_failure(u64 start, struct page *page) +{ + u64 private; + u64 private_failure; + struct io_failure_record *failrec; + struct btrfs_mapping_tree *map_tree; + struct extent_state *state; + int num_copies; + int did_repair = 0; + int ret; + struct inode *inode = page->mapping->host; + + private = 0; + ret = count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private, + (u64)-1, 1, EXTENT_DIRTY, 0); + if (!ret) + return 0; + + ret = get_state_private(&BTRFS_I(inode)->io_failure_tree, start, + &private_failure); + if (ret) + return 0; + + failrec = (struct io_failure_record *)(unsigned long) private_failure; + BUG_ON(!failrec->this_mirror); + + if (failrec->in_validation) { + /* there was no real error, just free the record */ + pr_debug("clean_io_failure: freeing dummy error at %llu\n", + failrec->start); + did_repair = 1; + goto out; + } + + spin_lock(&BTRFS_I(inode)->io_tree.lock); + state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree, + failrec->start, + EXTENT_LOCKED); + spin_unlock(&BTRFS_I(inode)->io_tree.lock); + + if (state && state->start == failrec->start) { + map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree; + num_copies = btrfs_num_copies(map_tree, failrec->logical, + failrec->len); + if (num_copies > 1) { + ret = repair_io_failure(map_tree, start, failrec->len, + failrec->logical, page, + failrec->failed_mirror); + did_repair = !ret; + } + } + +out: + if (!ret) + ret = free_io_failure(inode, failrec, did_repair); + + return ret; +} + +/* + * this is a generic handler for readpage errors (default + * readpage_io_failed_hook). if other copies exist, read those and write back + * good data to the failed position. does not investigate in remapping the + * failed extent elsewhere, hoping the device will be smart enough to do this as + * needed + */ + +static int bio_readpage_error(struct bio *failed_bio, struct page *page, + u64 start, u64 end, int failed_mirror, + struct extent_state *state) +{ + struct io_failure_record *failrec = NULL; + u64 private; + struct extent_map *em; + struct inode *inode = page->mapping->host; + struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree; + struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree; + struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; + struct bio *bio; + int num_copies; + int ret; + int read_mode; + u64 logical; + + BUG_ON(failed_bio->bi_rw & REQ_WRITE); + + ret = get_state_private(failure_tree, start, &private); + if (ret) { + failrec = kzalloc(sizeof(*failrec), GFP_NOFS); + if (!failrec) + return -ENOMEM; + failrec->start = start; + failrec->len = end - start + 1; + failrec->this_mirror = 0; + failrec->bio_flags = 0; + failrec->in_validation = 0; + + read_lock(&em_tree->lock); + em = lookup_extent_mapping(em_tree, start, failrec->len); + if (!em) { + read_unlock(&em_tree->lock); + kfree(failrec); + return -EIO; + } + + if (em->start > start || em->start + em->len < start) { + free_extent_map(em); + em = NULL; + } + read_unlock(&em_tree->lock); + + if (!em || IS_ERR(em)) { + kfree(failrec); + return -EIO; + } + logical = start - em->start; + logical = em->block_start + logical; + if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) { + logical = em->block_start; + failrec->bio_flags = EXTENT_BIO_COMPRESSED; + extent_set_compress_type(&failrec->bio_flags, + em->compress_type); + } + pr_debug("bio_readpage_error: (new) logical=%llu, start=%llu, " + "len=%llu\n", logical, start, failrec->len); + failrec->logical = logical; + free_extent_map(em); + + /* set the bits in the private failure tree */ + ret = set_extent_bits(failure_tree, start, end, + EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); + if (ret >= 0) + ret = set_state_private(failure_tree, start, + (u64)(unsigned long)failrec); + /* set the bits in the inode's tree */ + if (ret >= 0) + ret = set_extent_bits(tree, start, end, EXTENT_DAMAGED, + GFP_NOFS); + if (ret < 0) { + kfree(failrec); + return ret; + } + } else { + failrec = (struct io_failure_record *)(unsigned long)private; + pr_debug("bio_readpage_error: (found) logical=%llu, " + "start=%llu, len=%llu, validation=%d\n", + failrec->logical, failrec->start, failrec->len, + failrec->in_validation); + /* + * when data can be on disk more than twice, add to failrec here + * (e.g. with a list for failed_mirror) to make + * clean_io_failure() clean all those errors at once. + */ + } + num_copies = btrfs_num_copies( + &BTRFS_I(inode)->root->fs_info->mapping_tree, + failrec->logical, failrec->len); + if (num_copies == 1) { + /* + * we only have a single copy of the data, so don't bother with + * all the retry and error correction code that follows. no + * matter what the error is, it is very likely to persist. + */ + pr_debug("bio_readpage_error: cannot repair, num_copies == 1. " + "state=%p, num_copies=%d, next_mirror %d, " + "failed_mirror %d\n", state, num_copies, + failrec->this_mirror, failed_mirror); + free_io_failure(inode, failrec, 0); + return -EIO; + } + + if (!state) { + spin_lock(&tree->lock); + state = find_first_extent_bit_state(tree, failrec->start, + EXTENT_LOCKED); + if (state && state->start != failrec->start) + state = NULL; + spin_unlock(&tree->lock); + } + + /* + * there are two premises: + * a) deliver good data to the caller + * b) correct the bad sectors on disk + */ + if (failed_bio->bi_vcnt > 1) { + /* + * to fulfill b), we need to know the exact failing sectors, as + * we don't want to rewrite any more than the failed ones. thus, + * we need separate read requests for the failed bio + * + * if the following BUG_ON triggers, our validation request got + * merged. we need separate requests for our algorithm to work. + */ + BUG_ON(failrec->in_validation); + failrec->in_validation = 1; + failrec->this_mirror = failed_mirror; + read_mode = READ_SYNC | REQ_FAILFAST_DEV; + } else { + /* + * we're ready to fulfill a) and b) alongside. get a good copy + * of the failed sector and if we succeed, we have setup + * everything for repair_io_failure to do the rest for us. + */ + if (failrec->in_validation) { + BUG_ON(failrec->this_mirror != failed_mirror); + failrec->in_validation = 0; + failrec->this_mirror = 0; + } + failrec->failed_mirror = failed_mirror; + failrec->this_mirror++; + if (failrec->this_mirror == failed_mirror) + failrec->this_mirror++; + read_mode = READ_SYNC; + } + + if (!state || failrec->this_mirror > num_copies) { + pr_debug("bio_readpage_error: (fail) state=%p, num_copies=%d, " + "next_mirror %d, failed_mirror %d\n", state, + num_copies, failrec->this_mirror, failed_mirror); + free_io_failure(inode, failrec, 0); + return -EIO; + } + + bio = bio_alloc(GFP_NOFS, 1); + if (!bio) { + free_io_failure(inode, failrec, 0); + return -EIO; + } + bio->bi_private = state; + bio->bi_end_io = failed_bio->bi_end_io; + bio->bi_sector = failrec->logical >> 9; + bio->bi_bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev; + bio->bi_size = 0; + + bio_add_page(bio, page, failrec->len, start - page_offset(page)); + + pr_debug("bio_readpage_error: submitting new read[%#x] to " + "this_mirror=%d, num_copies=%d, in_validation=%d\n", read_mode, + failrec->this_mirror, num_copies, failrec->in_validation); + + ret = tree->ops->submit_bio_hook(inode, read_mode, bio, + failrec->this_mirror, + failrec->bio_flags, 0); + return ret; +} + /* lots and lots of room for performance fixes in the end_bio funcs */ +int end_extent_writepage(struct page *page, int err, u64 start, u64 end) +{ + int uptodate = (err == 0); + struct extent_io_tree *tree; + int ret; + + tree = &BTRFS_I(page->mapping->host)->io_tree; + + if (tree->ops && tree->ops->writepage_end_io_hook) { + ret = tree->ops->writepage_end_io_hook(page, start, + end, NULL, uptodate); + if (ret) + uptodate = 0; + } + + if (!uptodate) { + ClearPageUptodate(page); + SetPageError(page); + } + return 0; +} + /* * after a writepage IO is done, we need to: * clear the uptodate bits on error @@ -1612,13 +2233,11 @@ static int check_page_writeback(struct extent_io_tree *tree, */ static void end_bio_extent_writepage(struct bio *bio, int err) { - int uptodate = err == 0; struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1; struct extent_io_tree *tree; u64 start; u64 end; int whole_page; - int ret; do { struct page *page = bvec->bv_page; @@ -1635,28 +2254,9 @@ static void end_bio_extent_writepage(struct bio *bio, int err) if (--bvec >= bio->bi_io_vec) prefetchw(&bvec->bv_page->flags); - if (tree->ops && tree->ops->writepage_end_io_hook) { - ret = tree->ops->writepage_end_io_hook(page, start, - end, NULL, uptodate); - if (ret) - uptodate = 0; - } - if (!uptodate && tree->ops && - tree->ops->writepage_io_failed_hook) { - ret = tree->ops->writepage_io_failed_hook(bio, page, - start, end, NULL); - if (ret == 0) { - uptodate = (err == 0); - continue; - } - } - - if (!uptodate) { - clear_extent_uptodate(tree, start, end, NULL, GFP_NOFS); - ClearPageUptodate(page); - SetPageError(page); - } + if (end_extent_writepage(page, err, start, end)) + continue; if (whole_page) end_page_writeback(page); @@ -1687,6 +2287,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err) u64 start; u64 end; int whole_page; + int mirror; int ret; if (err) @@ -1697,6 +2298,9 @@ static void end_bio_extent_readpage(struct bio *bio, int err) struct extent_state *cached = NULL; struct extent_state *state; + pr_debug("end_bio_extent_readpage: bi_vcnt=%d, idx=%d, err=%d, " + "mirror=%ld\n", bio->bi_vcnt, bio->bi_idx, err, + (long int)bio->bi_bdev); tree = &BTRFS_I(page->mapping->host)->io_tree; start = ((u64)page->index << PAGE_CACHE_SHIFT) + @@ -1722,16 +2326,33 @@ static void end_bio_extent_readpage(struct bio *bio, int err) } spin_unlock(&tree->lock); + mirror = (int)(unsigned long)bio->bi_bdev; if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) { ret = tree->ops->readpage_end_io_hook(page, start, end, - state); + state, mirror); if (ret) uptodate = 0; + else + clean_io_failure(start, page); } - if (!uptodate && tree->ops && - tree->ops->readpage_io_failed_hook) { - ret = tree->ops->readpage_io_failed_hook(bio, page, - start, end, NULL); + + if (!uptodate && tree->ops && tree->ops->readpage_io_failed_hook) { + ret = tree->ops->readpage_io_failed_hook(page, mirror); + if (!ret && !err && + test_bit(BIO_UPTODATE, &bio->bi_flags)) + uptodate = 1; + } else if (!uptodate) { + /* + * The generic bio_readpage_error handles errors the + * following way: If possible, new read requests are + * created and submitted and will end up in + * end_bio_extent_readpage as well (if we're lucky, not + * in the !uptodate case). In that case it returns 0 and + * we just go on with the next page in our bio. If it + * can't handle the error it will return -EIO and we + * remain responsible for that page. + */ + ret = bio_readpage_error(bio, page, start, end, mirror, NULL); if (ret == 0) { uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); @@ -1742,7 +2363,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err) } } - if (uptodate) { + if (uptodate && tree->track_uptodate) { set_extent_uptodate(tree, start, end, &cached, GFP_ATOMIC); } @@ -1791,8 +2412,12 @@ btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs, return bio; } -static int submit_one_bio(int rw, struct bio *bio, int mirror_num, - unsigned long bio_flags) +/* + * Since writes are async, they will only return -ENOMEM. + * Reads can return the full range of I/O error conditions. + */ +static int __must_check submit_one_bio(int rw, struct bio *bio, + int mirror_num, unsigned long bio_flags) { int ret = 0; struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1; @@ -1810,13 +2435,27 @@ static int submit_one_bio(int rw, struct bio *bio, int mirror_num, ret = tree->ops->submit_bio_hook(page->mapping->host, rw, bio, mirror_num, bio_flags, start); else - submit_bio(rw, bio); + btrfsic_submit_bio(rw, bio); + if (bio_flagged(bio, BIO_EOPNOTSUPP)) ret = -EOPNOTSUPP; bio_put(bio); return ret; } +static int merge_bio(struct extent_io_tree *tree, struct page *page, + unsigned long offset, size_t size, struct bio *bio, + unsigned long bio_flags) +{ + int ret = 0; + if (tree->ops && tree->ops->merge_bio_hook) + ret = tree->ops->merge_bio_hook(page, offset, size, bio, + bio_flags); + BUG_ON(ret < 0); + return ret; + +} + static int submit_extent_page(int rw, struct extent_io_tree *tree, struct page *page, sector_t sector, size_t size, unsigned long offset, @@ -1845,12 +2484,12 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, sector; if (prev_bio_flags != bio_flags || !contig || - (tree->ops && tree->ops->merge_bio_hook && - tree->ops->merge_bio_hook(page, offset, page_size, bio, - bio_flags)) || + merge_bio(tree, page, offset, page_size, bio, bio_flags) || bio_add_page(bio, page, page_size, offset) < page_size) { ret = submit_one_bio(rw, bio, mirror_num, prev_bio_flags); + if (ret < 0) + return ret; bio = NULL; } else { return 0; @@ -1877,25 +2516,31 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, return ret; } -void set_page_extent_mapped(struct page *page) +void attach_extent_buffer_page(struct extent_buffer *eb, struct page *page) { if (!PagePrivate(page)) { SetPagePrivate(page); page_cache_get(page); - set_page_private(page, EXTENT_PAGE_PRIVATE); + set_page_private(page, (unsigned long)eb); + } else { + WARN_ON(page->private != (unsigned long)eb); } } -static void set_page_extent_head(struct page *page, unsigned long len) +void set_page_extent_mapped(struct page *page) { - WARN_ON(!PagePrivate(page)); - set_page_private(page, EXTENT_PAGE_PRIVATE_FIRST_PAGE | len << 2); + if (!PagePrivate(page)) { + SetPagePrivate(page); + page_cache_get(page); + set_page_private(page, EXTENT_PAGE_PRIVATE); + } } /* * basic readpage implementation. Locked extent state structs are inserted * into the tree that are removed when the IO is done (by the end_io * handlers) + * XXX JDM: This needs looking at to ensure proper page locking */ static int __extent_read_full_page(struct extent_io_tree *tree, struct page *page, @@ -1935,11 +2580,11 @@ static int __extent_read_full_page(struct extent_io_tree *tree, end = page_end; while (1) { - lock_extent(tree, start, end, GFP_NOFS); + lock_extent(tree, start, end); ordered = btrfs_lookup_ordered_extent(inode, start); if (!ordered) break; - unlock_extent(tree, start, end, GFP_NOFS); + unlock_extent(tree, start, end); btrfs_start_ordered_extent(inode, ordered, 1); btrfs_put_ordered_extent(ordered); } @@ -1950,10 +2595,10 @@ static int __extent_read_full_page(struct extent_io_tree *tree, if (zero_offset) { iosize = PAGE_CACHE_SIZE - zero_offset; - userpage = kmap_atomic(page, KM_USER0); + userpage = kmap_atomic(page); memset(userpage + zero_offset, 0, iosize); flush_dcache_page(page); - kunmap_atomic(userpage, KM_USER0); + kunmap_atomic(userpage); } } while (cur <= end) { @@ -1962,10 +2607,10 @@ static int __extent_read_full_page(struct extent_io_tree *tree, struct extent_state *cached = NULL; iosize = PAGE_CACHE_SIZE - pg_offset; - userpage = kmap_atomic(page, KM_USER0); + userpage = kmap_atomic(page); memset(userpage + pg_offset, 0, iosize); flush_dcache_page(page); - kunmap_atomic(userpage, KM_USER0); + kunmap_atomic(userpage); set_extent_uptodate(tree, cur, cur + iosize - 1, &cached, GFP_NOFS); unlock_extent_cached(tree, cur, cur + iosize - 1, @@ -1976,7 +2621,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, end - cur + 1, 0); if (IS_ERR_OR_NULL(em)) { SetPageError(page); - unlock_extent(tree, cur, end, GFP_NOFS); + unlock_extent(tree, cur, end); break; } extent_offset = cur - em->start; @@ -2011,10 +2656,10 @@ static int __extent_read_full_page(struct extent_io_tree *tree, char *userpage; struct extent_state *cached = NULL; - userpage = kmap_atomic(page, KM_USER0); + userpage = kmap_atomic(page); memset(userpage + pg_offset, 0, iosize); flush_dcache_page(page); - kunmap_atomic(userpage, KM_USER0); + kunmap_atomic(userpage); set_extent_uptodate(tree, cur, cur + iosize - 1, &cached, GFP_NOFS); @@ -2028,7 +2673,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, if (test_range_bit(tree, cur, cur_end, EXTENT_UPTODATE, 1, NULL)) { check_page_uptodate(tree, page); - unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); + unlock_extent(tree, cur, cur + iosize - 1); cur = cur + iosize; pg_offset += iosize; continue; @@ -2038,7 +2683,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, */ if (block_start == EXTENT_MAP_INLINE) { SetPageError(page); - unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); + unlock_extent(tree, cur, cur + iosize - 1); cur = cur + iosize; pg_offset += iosize; continue; @@ -2058,6 +2703,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, end_bio_extent_readpage, mirror_num, *bio_flags, this_bio_flag); + BUG_ON(ret == -ENOMEM); nr++; *bio_flags = this_bio_flag; } @@ -2076,16 +2722,16 @@ out: } int extent_read_full_page(struct extent_io_tree *tree, struct page *page, - get_extent_t *get_extent) + get_extent_t *get_extent, int mirror_num) { struct bio *bio = NULL; unsigned long bio_flags = 0; int ret; - ret = __extent_read_full_page(tree, page, get_extent, &bio, 0, + ret = __extent_read_full_page(tree, page, get_extent, &bio, mirror_num, &bio_flags); if (bio) - ret = submit_one_bio(READ, bio, 0, bio_flags); + ret = submit_one_bio(READ, bio, mirror_num, bio_flags); return ret; } @@ -2136,6 +2782,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, int compressed; int write_flags; unsigned long nr_written = 0; + bool fill_delalloc = true; if (wbc->sync_mode == WB_SYNC_ALL) write_flags = WRITE_SYNC; @@ -2145,6 +2792,9 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, trace___extent_writepage(page, inode, wbc); WARN_ON(!PageLocked(page)); + + ClearPageError(page); + pg_offset = i_size & (PAGE_CACHE_SIZE - 1); if (page->index > end_index || (page->index == end_index && !pg_offset)) { @@ -2156,20 +2806,23 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, if (page->index == end_index) { char *userpage; - userpage = kmap_atomic(page, KM_USER0); + userpage = kmap_atomic(page); memset(userpage + pg_offset, 0, PAGE_CACHE_SIZE - pg_offset); - kunmap_atomic(userpage, KM_USER0); + kunmap_atomic(userpage); flush_dcache_page(page); } pg_offset = 0; set_page_extent_mapped(page); + if (!tree->ops || !tree->ops->fill_delalloc) + fill_delalloc = false; + delalloc_start = start; delalloc_end = 0; page_started = 0; - if (!epd->extent_locked) { + if (!epd->extent_locked && fill_delalloc) { u64 delalloc_to_write = 0; /* * make sure the wbc mapping index is at least updated @@ -2187,9 +2840,16 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, delalloc_start = delalloc_end + 1; continue; } - tree->ops->fill_delalloc(inode, page, delalloc_start, - delalloc_end, &page_started, - &nr_written); + ret = tree->ops->fill_delalloc(inode, page, + delalloc_start, + delalloc_end, + &page_started, + &nr_written); + /* File system has been set read-only */ + if (ret) { + SetPageError(page); + goto done; + } /* * delalloc_end is already one less than the total * length, so we don't subtract one from @@ -2226,8 +2886,12 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, if (tree->ops && tree->ops->writepage_start_hook) { ret = tree->ops->writepage_start_hook(page, start, page_end); - if (ret == -EAGAIN) { - redirty_page_for_writepage(wbc, page); + if (ret) { + /* Fixup worker will requeue */ + if (ret == -EBUSY) + wbc->pages_skipped++; + else + redirty_page_for_writepage(wbc, page); update_nr_written(page, wbc, nr_written); unlock_page(page); ret = 0; @@ -2358,6 +3022,284 @@ done_unlocked: return 0; } +static int eb_wait(void *word) +{ + io_schedule(); + return 0; +} + +static void wait_on_extent_buffer_writeback(struct extent_buffer *eb) +{ + wait_on_bit(&eb->bflags, EXTENT_BUFFER_WRITEBACK, eb_wait, + TASK_UNINTERRUPTIBLE); +} + +static int lock_extent_buffer_for_io(struct extent_buffer *eb, + struct btrfs_fs_info *fs_info, + struct extent_page_data *epd) +{ + unsigned long i, num_pages; + int flush = 0; + int ret = 0; + + if (!btrfs_try_tree_write_lock(eb)) { + flush = 1; + flush_write_bio(epd); + btrfs_tree_lock(eb); + } + + if (test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags)) { + btrfs_tree_unlock(eb); + if (!epd->sync_io) + return 0; + if (!flush) { + flush_write_bio(epd); + flush = 1; + } + while (1) { + wait_on_extent_buffer_writeback(eb); + btrfs_tree_lock(eb); + if (!test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags)) + break; + btrfs_tree_unlock(eb); + } + } + + /* + * We need to do this to prevent races in people who check if the eb is + * under IO since we can end up having no IO bits set for a short period + * of time. + */ + spin_lock(&eb->refs_lock); + if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) { + set_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags); + spin_unlock(&eb->refs_lock); + btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN); + spin_lock(&fs_info->delalloc_lock); + if (fs_info->dirty_metadata_bytes >= eb->len) + fs_info->dirty_metadata_bytes -= eb->len; + else + WARN_ON(1); + spin_unlock(&fs_info->delalloc_lock); + ret = 1; + } else { + spin_unlock(&eb->refs_lock); + } + + btrfs_tree_unlock(eb); + + if (!ret) + return ret; + + num_pages = num_extent_pages(eb->start, eb->len); + for (i = 0; i < num_pages; i++) { + struct page *p = extent_buffer_page(eb, i); + + if (!trylock_page(p)) { + if (!flush) { + flush_write_bio(epd); + flush = 1; + } + lock_page(p); + } + } + + return ret; +} + +static void end_extent_buffer_writeback(struct extent_buffer *eb) +{ + clear_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags); + smp_mb__after_clear_bit(); + wake_up_bit(&eb->bflags, EXTENT_BUFFER_WRITEBACK); +} + +static void end_bio_extent_buffer_writepage(struct bio *bio, int err) +{ + int uptodate = err == 0; + struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1; + struct extent_buffer *eb; + int done; + + do { + struct page *page = bvec->bv_page; + + bvec--; + eb = (struct extent_buffer *)page->private; + BUG_ON(!eb); + done = atomic_dec_and_test(&eb->io_pages); + + if (!uptodate || test_bit(EXTENT_BUFFER_IOERR, &eb->bflags)) { + set_bit(EXTENT_BUFFER_IOERR, &eb->bflags); + ClearPageUptodate(page); + SetPageError(page); + } + + end_page_writeback(page); + + if (!done) + continue; + + end_extent_buffer_writeback(eb); + } while (bvec >= bio->bi_io_vec); + + bio_put(bio); + +} + +static int write_one_eb(struct extent_buffer *eb, + struct btrfs_fs_info *fs_info, + struct writeback_control *wbc, + struct extent_page_data *epd) +{ + struct block_device *bdev = fs_info->fs_devices->latest_bdev; + u64 offset = eb->start; + unsigned long i, num_pages; + int rw = (epd->sync_io ? WRITE_SYNC : WRITE); + int ret = 0; + + clear_bit(EXTENT_BUFFER_IOERR, &eb->bflags); + num_pages = num_extent_pages(eb->start, eb->len); + atomic_set(&eb->io_pages, num_pages); + for (i = 0; i < num_pages; i++) { + struct page *p = extent_buffer_page(eb, i); + + clear_page_dirty_for_io(p); + set_page_writeback(p); + ret = submit_extent_page(rw, eb->tree, p, offset >> 9, + PAGE_CACHE_SIZE, 0, bdev, &epd->bio, + -1, end_bio_extent_buffer_writepage, + 0, 0, 0); + if (ret) { + set_bit(EXTENT_BUFFER_IOERR, &eb->bflags); + SetPageError(p); + if (atomic_sub_and_test(num_pages - i, &eb->io_pages)) + end_extent_buffer_writeback(eb); + ret = -EIO; + break; + } + offset += PAGE_CACHE_SIZE; + update_nr_written(p, wbc, 1); + unlock_page(p); + } + + if (unlikely(ret)) { + for (; i < num_pages; i++) { + struct page *p = extent_buffer_page(eb, i); + unlock_page(p); + } + } + + return ret; +} + +int btree_write_cache_pages(struct address_space *mapping, + struct writeback_control *wbc) +{ + struct extent_io_tree *tree = &BTRFS_I(mapping->host)->io_tree; + struct btrfs_fs_info *fs_info = BTRFS_I(mapping->host)->root->fs_info; + struct extent_buffer *eb, *prev_eb = NULL; + struct extent_page_data epd = { + .bio = NULL, + .tree = tree, + .extent_locked = 0, + .sync_io = wbc->sync_mode == WB_SYNC_ALL, + }; + int ret = 0; + int done = 0; + int nr_to_write_done = 0; + struct pagevec pvec; + int nr_pages; + pgoff_t index; + pgoff_t end; /* Inclusive */ + int scanned = 0; + int tag; + + pagevec_init(&pvec, 0); + if (wbc->range_cyclic) { + index = mapping->writeback_index; /* Start from prev offset */ + end = -1; + } else { + index = wbc->range_start >> PAGE_CACHE_SHIFT; + end = wbc->range_end >> PAGE_CACHE_SHIFT; + scanned = 1; + } + if (wbc->sync_mode == WB_SYNC_ALL) + tag = PAGECACHE_TAG_TOWRITE; + else + tag = PAGECACHE_TAG_DIRTY; +retry: + if (wbc->sync_mode == WB_SYNC_ALL) + tag_pages_for_writeback(mapping, index, end); + while (!done && !nr_to_write_done && (index <= end) && + (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag, + min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) { + unsigned i; + + scanned = 1; + for (i = 0; i < nr_pages; i++) { + struct page *page = pvec.pages[i]; + + if (!PagePrivate(page)) + continue; + + if (!wbc->range_cyclic && page->index > end) { + done = 1; + break; + } + + eb = (struct extent_buffer *)page->private; + if (!eb) { + WARN_ON(1); + continue; + } + + if (eb == prev_eb) + continue; + + if (!atomic_inc_not_zero(&eb->refs)) { + WARN_ON(1); + continue; + } + + prev_eb = eb; + ret = lock_extent_buffer_for_io(eb, fs_info, &epd); + if (!ret) { + free_extent_buffer(eb); + continue; + } + + ret = write_one_eb(eb, fs_info, wbc, &epd); + if (ret) { + done = 1; + free_extent_buffer(eb); + break; + } + free_extent_buffer(eb); + + /* + * the filesystem may choose to bump up nr_to_write. + * We have to make sure to honor the new nr_to_write + * at any time + */ + nr_to_write_done = wbc->nr_to_write <= 0; + } + pagevec_release(&pvec); + cond_resched(); + } + if (!scanned && !done) { + /* + * We hit the last page and there is more work to be done: wrap + * back to the start of the file + */ + scanned = 1; + index = 0; + goto retry; + } + flush_write_bio(&epd); + return ret; +} + /** * write_cache_pages - walk the list of dirty pages of the given address space and write all of them. * @mapping: address space structure to write @@ -2379,6 +3321,7 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, writepage_t writepage, void *data, void (*flush_fn)(void *)) { + struct inode *inode = mapping->host; int ret = 0; int done = 0; int nr_to_write_done = 0; @@ -2389,6 +3332,18 @@ static int extent_write_cache_pages(struct extent_io_tree *tree, int scanned = 0; int tag; + /* + * We have to hold onto the inode so that ordered extents can do their + * work when the IO finishes. The alternative to this is failing to add + * an ordered extent if the igrab() fails there and that is a huge pain + * to deal with, so instead just hold onto the inode throughout the + * writepages operation. If it fails here we are freeing up the inode + * anyway and we'd rather not waste our time writing out stuff that is + * going to be truncated anyway. + */ + if (!igrab(inode)) + return 0; + pagevec_init(&pvec, 0); if (wbc->range_cyclic) { index = mapping->writeback_index; /* Start from prev offset */ @@ -2421,10 +3376,16 @@ retry: * swizzled back from swapper_space to tmpfs file * mapping */ - if (tree->ops && tree->ops->write_cache_pages_lock_hook) - tree->ops->write_cache_pages_lock_hook(page); - else - lock_page(page); + if (tree->ops && + tree->ops->write_cache_pages_lock_hook) { + tree->ops->write_cache_pages_lock_hook(page, + data, flush_fn); + } else { + if (!trylock_page(page)) { + flush_fn(data); + lock_page(page); + } + } if (unlikely(page->mapping != mapping)) { unlock_page(page); @@ -2477,16 +3438,21 @@ retry: index = 0; goto retry; } + btrfs_add_delayed_iput(inode); return ret; } static void flush_epd_write_bio(struct extent_page_data *epd) { if (epd->bio) { + int rw = WRITE; + int ret; + if (epd->sync_io) - submit_one_bio(WRITE_SYNC, epd->bio, 0, 0); - else - submit_one_bio(WRITE, epd->bio, 0, 0); + rw = WRITE_SYNC; + + ret = submit_one_bio(rw, epd->bio, 0, 0); + BUG_ON(ret < 0); /* -ENOMEM */ epd->bio = NULL; } } @@ -2588,22 +3554,41 @@ int extent_readpages(struct extent_io_tree *tree, struct bio *bio = NULL; unsigned page_idx; unsigned long bio_flags = 0; + struct page *pagepool[16]; + struct page *page; + int i = 0; + int nr = 0; for (page_idx = 0; page_idx < nr_pages; page_idx++) { - struct page *page = list_entry(pages->prev, struct page, lru); + page = list_entry(pages->prev, struct page, lru); prefetchw(&page->flags); list_del(&page->lru); - if (!add_to_page_cache_lru(page, mapping, + if (add_to_page_cache_lru(page, mapping, page->index, GFP_NOFS)) { - __extent_read_full_page(tree, page, get_extent, - &bio, 0, &bio_flags); + page_cache_release(page); + continue; } - page_cache_release(page); + + pagepool[nr++] = page; + if (nr < ARRAY_SIZE(pagepool)) + continue; + for (i = 0; i < nr; i++) { + __extent_read_full_page(tree, pagepool[i], get_extent, + &bio, 0, &bio_flags); + page_cache_release(pagepool[i]); + } + nr = 0; + } + for (i = 0; i < nr; i++) { + __extent_read_full_page(tree, pagepool[i], get_extent, + &bio, 0, &bio_flags); + page_cache_release(pagepool[i]); } + BUG_ON(!list_empty(pages)); if (bio) - submit_one_bio(READ, bio, 0, bio_flags); + return submit_one_bio(READ, bio, 0, bio_flags); return 0; } @@ -2624,7 +3609,7 @@ int extent_invalidatepage(struct extent_io_tree *tree, if (start > end) return 0; - lock_extent_bits(tree, start, end, 0, &cached_state, GFP_NOFS); + lock_extent_bits(tree, start, end, 0, &cached_state); wait_on_page_writeback(page); clear_extent_bit(tree, start, end, EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC | @@ -2691,7 +3676,7 @@ int try_release_extent_mapping(struct extent_map_tree *map, len = end - start + 1; write_lock(&map->lock); em = lookup_extent_mapping(map, start, len); - if (IS_ERR_OR_NULL(em)) { + if (!em) { write_unlock(&map->lock); break; } @@ -2790,6 +3775,9 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, return -ENOMEM; path->leave_spinning = 1; + start = ALIGN(start, BTRFS_I(inode)->root->sectorsize); + len = ALIGN(len, BTRFS_I(inode)->root->sectorsize); + /* * lookup the last file extent. We're not using i_size here * because there might be preallocation past i_size @@ -2835,9 +3823,9 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, } lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + len, 0, - &cached_state, GFP_NOFS); + &cached_state); - em = get_extent_skip_holes(inode, off, last_for_get_extent, + em = get_extent_skip_holes(inode, start, last_for_get_extent, get_extent); if (!em) goto out; @@ -2926,37 +3914,31 @@ out: return ret; } -static inline struct page *extent_buffer_page(struct extent_buffer *eb, +inline struct page *extent_buffer_page(struct extent_buffer *eb, unsigned long i) { - struct page *p; - struct address_space *mapping; - - if (i == 0) - return eb->first_page; - i += eb->start >> PAGE_CACHE_SHIFT; - mapping = eb->first_page->mapping; - if (!mapping) - return NULL; - - /* - * extent_buffer_page is only called after pinning the page - * by increasing the reference count. So we know the page must - * be in the radix tree. - */ - rcu_read_lock(); - p = radix_tree_lookup(&mapping->page_tree, i); - rcu_read_unlock(); - - return p; + return eb->pages[i]; } -static inline unsigned long num_extent_pages(u64 start, u64 len) +inline unsigned long num_extent_pages(u64 start, u64 len) { return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) - (start >> PAGE_CACHE_SHIFT); } +static void __free_extent_buffer(struct extent_buffer *eb) +{ +#if LEAK_DEBUG + unsigned long flags; + spin_lock_irqsave(&leak_lock, flags); + list_del(&eb->leak_list); + spin_unlock_irqrestore(&leak_lock, flags); +#endif + if (eb->pages && eb->pages != eb->inline_pages) + kfree(eb->pages); + kmem_cache_free(extent_buffer_cache, eb); +} + static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree, u64 start, unsigned long len, @@ -2972,6 +3954,8 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree, return NULL; eb->start = start; eb->len = len; + eb->tree = tree; + eb->bflags = 0; rwlock_init(&eb->lock); atomic_set(&eb->write_locks, 0); atomic_set(&eb->read_locks, 0); @@ -2979,6 +3963,7 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree, atomic_set(&eb->blocking_writers, 0); atomic_set(&eb->spinning_readers, 0); atomic_set(&eb->spinning_writers, 0); + eb->lock_nested = 0; init_waitqueue_head(&eb->write_lock_wq); init_waitqueue_head(&eb->read_lock_wq); @@ -2987,20 +3972,86 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree, list_add(&eb->leak_list, &buffers); spin_unlock_irqrestore(&leak_lock, flags); #endif + spin_lock_init(&eb->refs_lock); atomic_set(&eb->refs, 1); + atomic_set(&eb->io_pages, 0); + + if (len > MAX_INLINE_EXTENT_BUFFER_SIZE) { + struct page **pages; + int num_pages = (len + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; + pages = kzalloc(num_pages, mask); + if (!pages) { + __free_extent_buffer(eb); + return NULL; + } + eb->pages = pages; + } else { + eb->pages = eb->inline_pages; + } return eb; } -static void __free_extent_buffer(struct extent_buffer *eb) +struct extent_buffer *btrfs_clone_extent_buffer(struct extent_buffer *src) { -#if LEAK_DEBUG - unsigned long flags; - spin_lock_irqsave(&leak_lock, flags); - list_del(&eb->leak_list); - spin_unlock_irqrestore(&leak_lock, flags); -#endif - kmem_cache_free(extent_buffer_cache, eb); + unsigned long i; + struct page *p; + struct extent_buffer *new; + unsigned long num_pages = num_extent_pages(src->start, src->len); + + new = __alloc_extent_buffer(NULL, src->start, src->len, GFP_ATOMIC); + if (new == NULL) + return NULL; + + for (i = 0; i < num_pages; i++) { + p = alloc_page(GFP_ATOMIC); + BUG_ON(!p); + attach_extent_buffer_page(new, p); + WARN_ON(PageDirty(p)); + SetPageUptodate(p); + new->pages[i] = p; + } + + copy_extent_buffer(new, src, 0, 0, src->len); + set_bit(EXTENT_BUFFER_UPTODATE, &new->bflags); + set_bit(EXTENT_BUFFER_DUMMY, &new->bflags); + + return new; +} + +struct extent_buffer *alloc_dummy_extent_buffer(u64 start, unsigned long len) +{ + struct extent_buffer *eb; + unsigned long num_pages = num_extent_pages(0, len); + unsigned long i; + + eb = __alloc_extent_buffer(NULL, start, len, GFP_ATOMIC); + if (!eb) + return NULL; + + for (i = 0; i < num_pages; i++) { + eb->pages[i] = alloc_page(GFP_ATOMIC); + if (!eb->pages[i]) + goto err; + } + set_extent_buffer_uptodate(eb); + btrfs_set_header_nritems(eb, 0); + set_bit(EXTENT_BUFFER_DUMMY, &eb->bflags); + + return eb; +err: + for (i--; i > 0; i--) + __free_page(eb->pages[i]); + __free_extent_buffer(eb); + return NULL; +} + +static int extent_buffer_under_io(struct extent_buffer *eb) +{ + return (atomic_read(&eb->io_pages) || + test_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags) || + test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); } /* @@ -3010,20 +4061,50 @@ static void btrfs_release_extent_buffer_page(struct extent_buffer *eb, unsigned long start_idx) { unsigned long index; + unsigned long num_pages; struct page *page; + int mapped = !test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags); - if (!eb->first_page) - return; + BUG_ON(extent_buffer_under_io(eb)); - index = num_extent_pages(eb->start, eb->len); + num_pages = num_extent_pages(eb->start, eb->len); + index = start_idx + num_pages; if (start_idx >= index) return; do { index--; page = extent_buffer_page(eb, index); - if (page) + if (page && mapped) { + spin_lock(&page->mapping->private_lock); + /* + * We do this since we'll remove the pages after we've + * removed the eb from the radix tree, so we could race + * and have this page now attached to the new eb. So + * only clear page_private if it's still connected to + * this eb. + */ + if (PagePrivate(page) && + page->private == (unsigned long)eb) { + BUG_ON(test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); + BUG_ON(PageDirty(page)); + BUG_ON(PageWriteback(page)); + /* + * We need to make sure we haven't be attached + * to a new eb. + */ + ClearPagePrivate(page); + set_page_private(page, 0); + /* One for the page private */ + page_cache_release(page); + } + spin_unlock(&page->mapping->private_lock); + + } + if (page) { + /* One for when we alloced the page */ page_cache_release(page); + } } while (index != start_idx); } @@ -3036,9 +4117,49 @@ static inline void btrfs_release_extent_buffer(struct extent_buffer *eb) __free_extent_buffer(eb); } +static void check_buffer_tree_ref(struct extent_buffer *eb) +{ + /* the ref bit is tricky. We have to make sure it is set + * if we have the buffer dirty. Otherwise the + * code to free a buffer can end up dropping a dirty + * page + * + * Once the ref bit is set, it won't go away while the + * buffer is dirty or in writeback, and it also won't + * go away while we have the reference count on the + * eb bumped. + * + * We can't just set the ref bit without bumping the + * ref on the eb because free_extent_buffer might + * see the ref bit and try to clear it. If this happens + * free_extent_buffer might end up dropping our original + * ref by mistake and freeing the page before we are able + * to add one more ref. + * + * So bump the ref count first, then set the bit. If someone + * beat us to it, drop the ref we added. + */ + spin_lock(&eb->refs_lock); + if (!test_and_set_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) + atomic_inc(&eb->refs); + spin_unlock(&eb->refs_lock); +} + +static void mark_extent_buffer_accessed(struct extent_buffer *eb) +{ + unsigned long num_pages, i; + + check_buffer_tree_ref(eb); + + num_pages = num_extent_pages(eb->start, eb->len); + for (i = 0; i < num_pages; i++) { + struct page *p = extent_buffer_page(eb, i); + mark_page_accessed(p); + } +} + struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, - u64 start, unsigned long len, - struct page *page0) + u64 start, unsigned long len) { unsigned long num_pages = num_extent_pages(start, len); unsigned long i; @@ -3054,7 +4175,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT); if (eb && atomic_inc_not_zero(&eb->refs)) { rcu_read_unlock(); - mark_page_accessed(eb->first_page); + mark_extent_buffer_accessed(eb); return eb; } rcu_read_unlock(); @@ -3063,32 +4184,44 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, if (!eb) return NULL; - if (page0) { - eb->first_page = page0; - i = 1; - index++; - page_cache_get(page0); - mark_page_accessed(page0); - set_page_extent_mapped(page0); - set_page_extent_head(page0, len); - uptodate = PageUptodate(page0); - } else { - i = 0; - } - for (; i < num_pages; i++, index++) { + for (i = 0; i < num_pages; i++, index++) { p = find_or_create_page(mapping, index, GFP_NOFS); if (!p) { WARN_ON(1); goto free_eb; } - set_page_extent_mapped(p); - mark_page_accessed(p); - if (i == 0) { - eb->first_page = p; - set_page_extent_head(p, len); - } else { - set_page_private(p, EXTENT_PAGE_PRIVATE); + + spin_lock(&mapping->private_lock); + if (PagePrivate(p)) { + /* + * We could have already allocated an eb for this page + * and attached one so lets see if we can get a ref on + * the existing eb, and if we can we know it's good and + * we can just return that one, else we know we can just + * overwrite page->private. + */ + exists = (struct extent_buffer *)p->private; + if (atomic_inc_not_zero(&exists->refs)) { + spin_unlock(&mapping->private_lock); + unlock_page(p); + page_cache_release(p); + mark_extent_buffer_accessed(exists); + goto free_eb; + } + + /* + * Do this so attach doesn't complain and we need to + * drop the ref the old guy had. + */ + ClearPagePrivate(p); + WARN_ON(PageDirty(p)); + page_cache_release(p); } + attach_extent_buffer_page(eb, p); + spin_unlock(&mapping->private_lock); + WARN_ON(PageDirty(p)); + mark_page_accessed(p); + eb->pages[i] = p; if (!PageUptodate(p)) uptodate = 0; @@ -3096,12 +4229,10 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, * see below about how we avoid a nasty race with release page * and why we unlock later */ - if (i != 0) - unlock_page(p); } if (uptodate) set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); - +again: ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); if (ret) goto free_eb; @@ -3111,14 +4242,19 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, if (ret == -EEXIST) { exists = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT); - /* add one reference for the caller */ - atomic_inc(&exists->refs); + if (!atomic_inc_not_zero(&exists->refs)) { + spin_unlock(&tree->buffer_lock); + radix_tree_preload_end(); + exists = NULL; + goto again; + } spin_unlock(&tree->buffer_lock); radix_tree_preload_end(); + mark_extent_buffer_accessed(exists); goto free_eb; } /* add one reference for the tree */ - atomic_inc(&eb->refs); + check_buffer_tree_ref(eb); spin_unlock(&tree->buffer_lock); radix_tree_preload_end(); @@ -3131,18 +4267,22 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, * after the extent buffer is in the radix tree so * it doesn't get lost */ - set_page_extent_mapped(eb->first_page); - set_page_extent_head(eb->first_page, eb->len); - if (!page0) - unlock_page(eb->first_page); + SetPageChecked(eb->pages[0]); + for (i = 1; i < num_pages; i++) { + p = extent_buffer_page(eb, i); + ClearPageChecked(p); + unlock_page(p); + } + unlock_page(eb->pages[0]); return eb; free_eb: - if (eb->first_page && !page0) - unlock_page(eb->first_page); + for (i = 0; i < num_pages; i++) { + if (eb->pages[i]) + unlock_page(eb->pages[i]); + } - if (!atomic_dec_and_test(&eb->refs)) - return exists; + WARN_ON(!atomic_dec_and_test(&eb->refs)); btrfs_release_extent_buffer(eb); return exists; } @@ -3156,7 +4296,7 @@ struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT); if (eb && atomic_inc_not_zero(&eb->refs)) { rcu_read_unlock(); - mark_page_accessed(eb->first_page); + mark_extent_buffer_accessed(eb); return eb; } rcu_read_unlock(); @@ -3164,19 +4304,81 @@ struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, return NULL; } +static inline void btrfs_release_extent_buffer_rcu(struct rcu_head *head) +{ + struct extent_buffer *eb = + container_of(head, struct extent_buffer, rcu_head); + + __free_extent_buffer(eb); +} + +/* Expects to have eb->eb_lock already held */ +static int release_extent_buffer(struct extent_buffer *eb, gfp_t mask) +{ + WARN_ON(atomic_read(&eb->refs) == 0); + if (atomic_dec_and_test(&eb->refs)) { + if (test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags)) { + spin_unlock(&eb->refs_lock); + } else { + struct extent_io_tree *tree = eb->tree; + + spin_unlock(&eb->refs_lock); + + spin_lock(&tree->buffer_lock); + radix_tree_delete(&tree->buffer, + eb->start >> PAGE_CACHE_SHIFT); + spin_unlock(&tree->buffer_lock); + } + + /* Should be safe to release our pages at this point */ + btrfs_release_extent_buffer_page(eb, 0); + + call_rcu(&eb->rcu_head, btrfs_release_extent_buffer_rcu); + return 1; + } + spin_unlock(&eb->refs_lock); + + return 0; +} + void free_extent_buffer(struct extent_buffer *eb) { if (!eb) return; - if (!atomic_dec_and_test(&eb->refs)) + spin_lock(&eb->refs_lock); + if (atomic_read(&eb->refs) == 2 && + test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags)) + atomic_dec(&eb->refs); + + if (atomic_read(&eb->refs) == 2 && + test_bit(EXTENT_BUFFER_STALE, &eb->bflags) && + !extent_buffer_under_io(eb) && + test_and_clear_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) + atomic_dec(&eb->refs); + + /* + * I know this is terrible, but it's temporary until we stop tracking + * the uptodate bits and such for the extent buffers. + */ + release_extent_buffer(eb, GFP_ATOMIC); +} + +void free_extent_buffer_stale(struct extent_buffer *eb) +{ + if (!eb) return; - WARN_ON(1); + spin_lock(&eb->refs_lock); + set_bit(EXTENT_BUFFER_STALE, &eb->bflags); + + if (atomic_read(&eb->refs) == 2 && !extent_buffer_under_io(eb) && + test_and_clear_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) + atomic_dec(&eb->refs); + release_extent_buffer(eb, GFP_NOFS); } -int clear_extent_buffer_dirty(struct extent_io_tree *tree, - struct extent_buffer *eb) +void clear_extent_buffer_dirty(struct extent_buffer *eb) { unsigned long i; unsigned long num_pages; @@ -3192,10 +4394,6 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree, lock_page(page); WARN_ON(!PagePrivate(page)); - set_page_extent_mapped(page); - if (i == 0) - set_page_extent_head(page, eb->len); - clear_page_dirty_for_io(page); spin_lock_irq(&page->mapping->tree_lock); if (!PageDirty(page)) { @@ -3204,26 +4402,32 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree, PAGECACHE_TAG_DIRTY); } spin_unlock_irq(&page->mapping->tree_lock); + ClearPageError(page); unlock_page(page); } - return 0; + WARN_ON(atomic_read(&eb->refs) == 0); } -int set_extent_buffer_dirty(struct extent_io_tree *tree, - struct extent_buffer *eb) +int set_extent_buffer_dirty(struct extent_buffer *eb) { unsigned long i; unsigned long num_pages; int was_dirty = 0; + check_buffer_tree_ref(eb); + was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags); + num_pages = num_extent_pages(eb->start, eb->len); + WARN_ON(atomic_read(&eb->refs) == 0); + WARN_ON(!test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)); + for (i = 0; i < num_pages; i++) - __set_page_dirty_nobuffers(extent_buffer_page(eb, i)); + set_page_dirty(extent_buffer_page(eb, i)); return was_dirty; } -static int __eb_straddles_pages(u64 start, u64 len) +static int range_straddles_pages(u64 start, u64 len) { if (len < PAGE_CACHE_SIZE) return 1; @@ -3234,26 +4438,14 @@ static int __eb_straddles_pages(u64 start, u64 len) return 0; } -static int eb_straddles_pages(struct extent_buffer *eb) -{ - return __eb_straddles_pages(eb->start, eb->len); -} - -int clear_extent_buffer_uptodate(struct extent_io_tree *tree, - struct extent_buffer *eb, - struct extent_state **cached_state) +int clear_extent_buffer_uptodate(struct extent_buffer *eb) { unsigned long i; struct page *page; unsigned long num_pages; - num_pages = num_extent_pages(eb->start, eb->len); clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); - - if (eb_straddles_pages(eb)) { - clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1, - cached_state, GFP_NOFS); - } + num_pages = num_extent_pages(eb->start, eb->len); for (i = 0; i < num_pages; i++) { page = extent_buffer_page(eb, i); if (page) @@ -3262,27 +4454,16 @@ int clear_extent_buffer_uptodate(struct extent_io_tree *tree, return 0; } -int set_extent_buffer_uptodate(struct extent_io_tree *tree, - struct extent_buffer *eb) +int set_extent_buffer_uptodate(struct extent_buffer *eb) { unsigned long i; struct page *page; unsigned long num_pages; + set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); num_pages = num_extent_pages(eb->start, eb->len); - - if (eb_straddles_pages(eb)) { - set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1, - NULL, GFP_NOFS); - } for (i = 0; i < num_pages; i++) { page = extent_buffer_page(eb, i); - if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) || - ((i == num_pages - 1) && - ((eb->start + eb->len) & (PAGE_CACHE_SIZE - 1)))) { - check_page_uptodate(tree, page); - continue; - } SetPageUptodate(page); } return 0; @@ -3297,7 +4478,7 @@ int extent_range_uptodate(struct extent_io_tree *tree, int uptodate; unsigned long index; - if (__eb_straddles_pages(start, end - start + 1)) { + if (range_straddles_pages(start, end - start + 1)) { ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL); if (ret) @@ -3306,6 +4487,8 @@ int extent_range_uptodate(struct extent_io_tree *tree, while (start <= end) { index = start >> PAGE_CACHE_SHIFT; page = find_get_page(tree->mapping, index); + if (!page) + return 1; uptodate = PageUptodate(page); page_cache_release(page); if (!uptodate) { @@ -3317,40 +4500,13 @@ int extent_range_uptodate(struct extent_io_tree *tree, return pg_uptodate; } -int extent_buffer_uptodate(struct extent_io_tree *tree, - struct extent_buffer *eb, - struct extent_state *cached_state) +int extent_buffer_uptodate(struct extent_buffer *eb) { - int ret = 0; - unsigned long num_pages; - unsigned long i; - struct page *page; - int pg_uptodate = 1; - - if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags)) - return 1; - - if (eb_straddles_pages(eb)) { - ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1, - EXTENT_UPTODATE, 1, cached_state); - if (ret) - return ret; - } - - num_pages = num_extent_pages(eb->start, eb->len); - for (i = 0; i < num_pages; i++) { - page = extent_buffer_page(eb, i); - if (!PageUptodate(page)) { - pg_uptodate = 0; - break; - } - } - return pg_uptodate; + return test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); } int read_extent_buffer_pages(struct extent_io_tree *tree, - struct extent_buffer *eb, - u64 start, int wait, + struct extent_buffer *eb, u64 start, int wait, get_extent_t *get_extent, int mirror_num) { unsigned long i; @@ -3360,21 +4516,14 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, int ret = 0; int locked_pages = 0; int all_uptodate = 1; - int inc_all_pages = 0; unsigned long num_pages; + unsigned long num_reads = 0; struct bio *bio = NULL; unsigned long bio_flags = 0; if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags)) return 0; - if (eb_straddles_pages(eb)) { - if (test_range_bit(tree, eb->start, eb->start + eb->len - 1, - EXTENT_UPTODATE, 1, NULL)) { - return 0; - } - } - if (start) { WARN_ON(start < eb->start); start_i = (start >> PAGE_CACHE_SHIFT) - @@ -3386,15 +4535,17 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, num_pages = num_extent_pages(eb->start, eb->len); for (i = start_i; i < num_pages; i++) { page = extent_buffer_page(eb, i); - if (!wait) { + if (wait == WAIT_NONE) { if (!trylock_page(page)) goto unlock_exit; } else { lock_page(page); } locked_pages++; - if (!PageUptodate(page)) + if (!PageUptodate(page)) { + num_reads++; all_uptodate = 0; + } } if (all_uptodate) { if (start_i == 0) @@ -3402,20 +4553,12 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, goto unlock_exit; } + clear_bit(EXTENT_BUFFER_IOERR, &eb->bflags); + eb->read_mirror = 0; + atomic_set(&eb->io_pages, num_reads); for (i = start_i; i < num_pages; i++) { page = extent_buffer_page(eb, i); - - WARN_ON(!PagePrivate(page)); - - set_page_extent_mapped(page); - if (i == 0) - set_page_extent_head(page, eb->len); - - if (inc_all_pages) - page_cache_get(page); if (!PageUptodate(page)) { - if (start_i == 0) - inc_all_pages = 1; ClearPageError(page); err = __extent_read_full_page(tree, page, get_extent, &bio, @@ -3427,10 +4570,13 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, } } - if (bio) - submit_one_bio(READ, bio, mirror_num, bio_flags); + if (bio) { + err = submit_one_bio(READ, bio, mirror_num, bio_flags); + if (err) + return err; + } - if (ret || !wait) + if (ret || wait != WAIT_COMPLETE) return ret; for (i = start_i; i < num_pages; i++) { @@ -3440,8 +4586,6 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, ret = -EIO; } - if (!ret) - set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags); return ret; unlock_exit: @@ -3683,15 +4827,20 @@ static void copy_pages(struct page *dst_page, struct page *src_page, { char *dst_kaddr = page_address(dst_page); char *src_kaddr; + int must_memmove = 0; if (dst_page != src_page) { src_kaddr = page_address(src_page); } else { src_kaddr = dst_kaddr; - BUG_ON(areas_overlap(src_off, dst_off, len)); + if (areas_overlap(src_off, dst_off, len)) + must_memmove = 1; } - memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len); + if (must_memmove) + memmove(dst_kaddr + dst_off, src_kaddr + src_off, len); + else + memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len); } void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, @@ -3761,7 +4910,7 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, "len %lu len %lu\n", dst_offset, len, dst->len); BUG_ON(1); } - if (!areas_overlap(src_offset, dst_offset, len)) { + if (dst_offset < src_offset) { memcpy_extent_buffer(dst, dst_offset, src_offset, len); return; } @@ -3787,47 +4936,47 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, } } -static inline void btrfs_release_extent_buffer_rcu(struct rcu_head *head) +int try_release_extent_buffer(struct page *page, gfp_t mask) { - struct extent_buffer *eb = - container_of(head, struct extent_buffer, rcu_head); - - btrfs_release_extent_buffer(eb); -} - -int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page) -{ - u64 start = page_offset(page); struct extent_buffer *eb; - int ret = 1; - spin_lock(&tree->buffer_lock); - eb = radix_tree_lookup(&tree->buffer, start >> PAGE_CACHE_SHIFT); - if (!eb) { - spin_unlock(&tree->buffer_lock); - return ret; + /* + * We need to make sure noboody is attaching this page to an eb right + * now. + */ + spin_lock(&page->mapping->private_lock); + if (!PagePrivate(page)) { + spin_unlock(&page->mapping->private_lock); + return 1; } - if (test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) { - ret = 0; - goto out; - } + eb = (struct extent_buffer *)page->private; + BUG_ON(!eb); /* - * set @eb->refs to 0 if it is already 1, and then release the @eb. - * Or go back. + * This is a little awful but should be ok, we need to make sure that + * the eb doesn't disappear out from under us while we're looking at + * this page. */ - if (atomic_cmpxchg(&eb->refs, 1, 0) != 1) { - ret = 0; - goto out; + spin_lock(&eb->refs_lock); + if (atomic_read(&eb->refs) != 1 || extent_buffer_under_io(eb)) { + spin_unlock(&eb->refs_lock); + spin_unlock(&page->mapping->private_lock); + return 0; } + spin_unlock(&page->mapping->private_lock); - radix_tree_delete(&tree->buffer, start >> PAGE_CACHE_SHIFT); -out: - spin_unlock(&tree->buffer_lock); + if ((mask & GFP_NOFS) == GFP_NOFS) + mask = GFP_NOFS; - /* at this point we can safely release the extent buffer */ - if (atomic_read(&eb->refs) == 0) - call_rcu(&eb->rcu_head, btrfs_release_extent_buffer_rcu); - return ret; + /* + * If tree ref isn't set then we know the ref on this eb is a real ref, + * so just return, this page will likely be freed soon anyway. + */ + if (!test_and_clear_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) { + spin_unlock(&eb->refs_lock); + return 0; + } + + return release_extent_buffer(eb, mask); } diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 7b2f0c3e7929..25900af5b15d 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -17,6 +17,8 @@ #define EXTENT_NODATASUM (1 << 10) #define EXTENT_DO_ACCOUNTING (1 << 11) #define EXTENT_FIRST_DELALLOC (1 << 12) +#define EXTENT_NEED_WAIT (1 << 13) +#define EXTENT_DAMAGED (1 << 14) #define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK) #define EXTENT_CTLBITS (EXTENT_DO_ACCOUNTING | EXTENT_FIRST_DELALLOC) @@ -32,6 +34,12 @@ #define EXTENT_BUFFER_BLOCKING 1 #define EXTENT_BUFFER_DIRTY 2 #define EXTENT_BUFFER_CORRUPT 3 +#define EXTENT_BUFFER_READAHEAD 4 /* this got triggered by readahead */ +#define EXTENT_BUFFER_TREE_REF 5 +#define EXTENT_BUFFER_STALE 6 +#define EXTENT_BUFFER_WRITEBACK 7 +#define EXTENT_BUFFER_IOERR 8 +#define EXTENT_BUFFER_DUMMY 9 /* these are flags for extent_clear_unlock_delalloc */ #define EXTENT_CLEAR_UNLOCK_PAGE 0x1 @@ -51,6 +59,7 @@ #define EXTENT_PAGE_PRIVATE_FIRST_PAGE 3 struct extent_state; +struct btrfs_root; typedef int (extent_submit_bio_hook_t)(struct inode *inode, int rw, struct bio *bio, int mirror_num, @@ -66,14 +75,9 @@ struct extent_io_ops { size_t size, struct bio *bio, unsigned long bio_flags); int (*readpage_io_hook)(struct page *page, u64 start, u64 end); - int (*readpage_io_failed_hook)(struct bio *bio, struct page *page, - u64 start, u64 end, - struct extent_state *state); - int (*writepage_io_failed_hook)(struct bio *bio, struct page *page, - u64 start, u64 end, - struct extent_state *state); + int (*readpage_io_failed_hook)(struct page *page, int failed_mirror); int (*readpage_end_io_hook)(struct page *page, u64 start, u64 end, - struct extent_state *state); + struct extent_state *state, int mirror); int (*writepage_end_io_hook)(struct page *page, u64 start, u64 end, struct extent_state *state, int uptodate); void (*set_bit_hook)(struct inode *inode, struct extent_state *state, @@ -85,7 +89,8 @@ struct extent_io_ops { struct extent_state *other); void (*split_extent_hook)(struct inode *inode, struct extent_state *orig, u64 split); - int (*write_cache_pages_lock_hook)(struct page *page); + int (*write_cache_pages_lock_hook)(struct page *page, void *data, + void (*flush_fn)(void *)); }; struct extent_io_tree { @@ -93,6 +98,7 @@ struct extent_io_tree { struct radix_tree_root buffer; struct address_space *mapping; u64 dirty_bytes; + int track_uptodate; spinlock_t lock; spinlock_t buffer_lock; struct extent_io_ops *ops; @@ -115,16 +121,22 @@ struct extent_state { struct list_head leak_list; }; +#define INLINE_EXTENT_BUFFER_PAGES 16 +#define MAX_INLINE_EXTENT_BUFFER_SIZE (INLINE_EXTENT_BUFFER_PAGES * PAGE_CACHE_SIZE) struct extent_buffer { u64 start; unsigned long len; unsigned long map_start; unsigned long map_len; - struct page *first_page; unsigned long bflags; + struct extent_io_tree *tree; + spinlock_t refs_lock; + atomic_t refs; + atomic_t io_pages; + int read_mirror; struct list_head leak_list; struct rcu_head rcu_head; - atomic_t refs; + pid_t lock_owner; /* count of read lock holders on the extent buffer */ atomic_t write_locks; @@ -133,6 +145,7 @@ struct extent_buffer { atomic_t blocking_readers; atomic_t spinning_readers; atomic_t spinning_writers; + int lock_nested; /* protects write locks */ rwlock_t lock; @@ -146,6 +159,9 @@ struct extent_buffer { * to unlock */ wait_queue_head_t read_lock_wq; + wait_queue_head_t lock_wq; + struct page *inline_pages[INLINE_EXTENT_BUFFER_PAGES]; + struct page **pages; }; static inline void extent_set_compress_type(unsigned long *bio_flags, @@ -172,20 +188,19 @@ void extent_io_tree_init(struct extent_io_tree *tree, int try_release_extent_mapping(struct extent_map_tree *map, struct extent_io_tree *tree, struct page *page, gfp_t mask); -int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page); +int try_release_extent_buffer(struct page *page, gfp_t mask); int try_release_extent_state(struct extent_map_tree *map, struct extent_io_tree *tree, struct page *page, gfp_t mask); -int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); +int lock_extent(struct extent_io_tree *tree, u64 start, u64 end); int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, - int bits, struct extent_state **cached, gfp_t mask); -int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); + int bits, struct extent_state **cached); +int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end); int unlock_extent_cached(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached, gfp_t mask); -int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask); +int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end); int extent_read_full_page(struct extent_io_tree *tree, struct page *page, - get_extent_t *get_extent); + get_extent_t *get_extent, int mirror_num); int __init extent_io_init(void); void extent_io_exit(void); @@ -204,16 +219,20 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end, int bits, gfp_t mask); int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, - int bits, int exclusive_bits, u64 *failed_start, + int bits, u64 *failed_start, struct extent_state **cached_state, gfp_t mask); int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask); +int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, + struct extent_state **cached_state, gfp_t mask); int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); +int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, + int bits, int clear_bits, gfp_t mask); int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask); int find_first_extent_bit(struct extent_io_tree *tree, u64 start, @@ -232,6 +251,8 @@ int extent_writepages(struct extent_io_tree *tree, struct address_space *mapping, get_extent_t *get_extent, struct writeback_control *wbc); +int btree_write_cache_pages(struct address_space *mapping, + struct writeback_control *wbc); int extent_readpages(struct extent_io_tree *tree, struct address_space *mapping, struct list_head *pages, unsigned nr_pages, @@ -243,14 +264,21 @@ int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private); void set_page_extent_mapped(struct page *page); struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, - u64 start, unsigned long len, - struct page *page0); + u64 start, unsigned long len); +struct extent_buffer *alloc_dummy_extent_buffer(u64 start, unsigned long len); +struct extent_buffer *btrfs_clone_extent_buffer(struct extent_buffer *src); struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, u64 start, unsigned long len); void free_extent_buffer(struct extent_buffer *eb); +void free_extent_buffer_stale(struct extent_buffer *eb); +#define WAIT_NONE 0 +#define WAIT_COMPLETE 1 +#define WAIT_PAGE_LOCK 2 int read_extent_buffer_pages(struct extent_io_tree *tree, struct extent_buffer *eb, u64 start, int wait, get_extent_t *get_extent, int mirror_num); +unsigned long num_extent_pages(u64 start, u64 len); +struct page *extent_buffer_page(struct extent_buffer *eb, unsigned long i); static inline void extent_buffer_get(struct extent_buffer *eb) { @@ -274,19 +302,12 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, unsigned long src_offset, unsigned long len); void memset_extent_buffer(struct extent_buffer *eb, char c, unsigned long start, unsigned long len); -int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits); -int clear_extent_buffer_dirty(struct extent_io_tree *tree, - struct extent_buffer *eb); -int set_extent_buffer_dirty(struct extent_io_tree *tree, - struct extent_buffer *eb); -int set_extent_buffer_uptodate(struct extent_io_tree *tree, - struct extent_buffer *eb); -int clear_extent_buffer_uptodate(struct extent_io_tree *tree, - struct extent_buffer *eb, - struct extent_state **cached_state); -int extent_buffer_uptodate(struct extent_io_tree *tree, - struct extent_buffer *eb, - struct extent_state *cached_state); +void wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits); +void clear_extent_buffer_dirty(struct extent_buffer *eb); +int set_extent_buffer_dirty(struct extent_buffer *eb); +int set_extent_buffer_uptodate(struct extent_buffer *eb); +int clear_extent_buffer_uptodate(struct extent_buffer *eb); +int extent_buffer_uptodate(struct extent_buffer *eb); int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset, unsigned long min_len, char **map, unsigned long *map_start, @@ -300,4 +321,13 @@ int extent_clear_unlock_delalloc(struct inode *inode, struct bio * btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs, gfp_t gfp_flags); + +struct btrfs_mapping_tree; + +int repair_io_failure(struct btrfs_mapping_tree *map_tree, u64 start, + u64 length, u64 logical, struct page *page, + int mirror_num); +int end_extent_writepage(struct page *page, int err, u64 start, u64 end); +int repair_eb_io_failure(struct btrfs_root *root, struct extent_buffer *eb, + int mirror_num); #endif diff --git a/fs/btrfs/extent_map.h b/fs/btrfs/extent_map.h index 33a7890b1f40..1195f09761fe 100644 --- a/fs/btrfs/extent_map.h +++ b/fs/btrfs/extent_map.h @@ -26,8 +26,8 @@ struct extent_map { unsigned long flags; struct block_device *bdev; atomic_t refs; - unsigned int in_tree:1; - unsigned int compress_type:4; + unsigned int in_tree; + unsigned int compress_type; }; struct extent_map_tree { diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index a1cb7821becd..857d93cd01dc 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -25,10 +25,12 @@ #include "transaction.h" #include "print-tree.h" -#define MAX_CSUM_ITEMS(r, size) ((((BTRFS_LEAF_DATA_SIZE(r) - \ +#define __MAX_CSUM_ITEMS(r, size) ((((BTRFS_LEAF_DATA_SIZE(r) - \ sizeof(struct btrfs_item) * 2) / \ size) - 1)) +#define MAX_CSUM_ITEMS(r, size) (min(__MAX_CSUM_ITEMS(r, size), PAGE_CACHE_SIZE)) + #define MAX_ORDERED_SUM_BYTES(r) ((PAGE_SIZE - \ sizeof(struct btrfs_ordered_sum)) / \ sizeof(struct btrfs_sector_sum) * \ @@ -59,7 +61,7 @@ int btrfs_insert_file_extent(struct btrfs_trans_handle *trans, sizeof(*item)); if (ret < 0) goto out; - BUG_ON(ret); + BUG_ON(ret); /* Can't happen */ leaf = path->nodes[0]; item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item); @@ -91,8 +93,7 @@ struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans, struct btrfs_csum_item *item; struct extent_buffer *leaf; u64 csum_offset = 0; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); int csums_in_item; file_key.objectid = BTRFS_EXTENT_CSUM_OBJECTID; @@ -162,8 +163,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, u64 item_last_offset = 0; u64 disk_bytenr; u32 diff; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); int ret; struct btrfs_path *path; struct btrfs_csum_item *item = NULL; @@ -183,7 +183,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, * read from the commit root and sidestep a nasty deadlock * between reading the free space cache and updating the csum tree. */ - if (btrfs_is_free_space_inode(root, inode)) { + if (btrfs_is_free_space_inode(inode)) { path->search_commit_root = 1; path->skip_locking = 1; } @@ -272,9 +272,9 @@ int btrfs_lookup_bio_sums(struct btrfs_root *root, struct inode *inode, } int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode, - struct bio *bio, u64 offset, u32 *dst) + struct bio *bio, u64 offset) { - return __btrfs_lookup_bio_sums(root, inode, bio, offset, dst, 1); + return __btrfs_lookup_bio_sums(root, inode, bio, offset, NULL, 1); } int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, @@ -286,11 +286,12 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, struct btrfs_ordered_sum *sums; struct btrfs_sector_sum *sector_sum; struct btrfs_csum_item *item; + LIST_HEAD(tmplist); unsigned long offset; int ret; size_t size; u64 csum_end; - u16 csum_size = btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); path = btrfs_alloc_path(); if (!path) @@ -360,7 +361,10 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, MAX_ORDERED_SUM_BYTES(root)); sums = kzalloc(btrfs_ordered_sum_size(root, size), GFP_NOFS); - BUG_ON(!sums); + if (!sums) { + ret = -ENOMEM; + goto fail; + } sector_sum = sums->sums; sums->bytenr = start; @@ -382,12 +386,19 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, offset += csum_size; sector_sum++; } - list_add_tail(&sums->list, list); + list_add_tail(&sums->list, &tmplist); } path->slots[0]++; } ret = 0; fail: + while (ret < 0 && !list_empty(&tmplist)) { + sums = list_entry(&tmplist, struct btrfs_ordered_sum, list); + list_del(&sums->list); + kfree(sums); + } + list_splice_tail(&tmplist, list); + btrfs_free_path(path); return ret; } @@ -422,7 +433,7 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode, offset = page_offset(bvec->bv_page) + bvec->bv_offset; ordered = btrfs_lookup_ordered_extent(inode, offset); - BUG_ON(!ordered); + BUG_ON(!ordered); /* Logic error */ sums->bytenr = ordered->start; while (bio_index < bio->bi_vcnt) { @@ -441,21 +452,21 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode, sums = kzalloc(btrfs_ordered_sum_size(root, bytes_left), GFP_NOFS); - BUG_ON(!sums); + BUG_ON(!sums); /* -ENOMEM */ sector_sum = sums->sums; sums->len = bytes_left; ordered = btrfs_lookup_ordered_extent(inode, offset); - BUG_ON(!ordered); + BUG_ON(!ordered); /* Logic error */ sums->bytenr = ordered->start; } - data = kmap_atomic(bvec->bv_page, KM_USER0); + data = kmap_atomic(bvec->bv_page); sector_sum->sum = ~(u32)0; sector_sum->sum = btrfs_csum_data(root, data + bvec->bv_offset, sector_sum->sum, bvec->bv_len); - kunmap_atomic(data, KM_USER0); + kunmap_atomic(data); btrfs_csum_final(sector_sum->sum, (char *)§or_sum->sum); sector_sum->bytenr = disk_bytenr; @@ -485,19 +496,17 @@ int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode, * This calls btrfs_truncate_item with the correct args based on the * overlap, and fixes up the key as required. */ -static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *key, - u64 bytenr, u64 len) +static noinline void truncate_one_csum(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_key *key, + u64 bytenr, u64 len) { struct extent_buffer *leaf; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); u64 csum_end; u64 end_byte = bytenr + len; u32 blocksize_bits = root->fs_info->sb->s_blocksize_bits; - int ret; leaf = path->nodes[0]; csum_end = btrfs_item_size_nr(leaf, path->slots[0]) / csum_size; @@ -513,7 +522,7 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, */ u32 new_size = (bytenr - key->offset) >> blocksize_bits; new_size *= csum_size; - ret = btrfs_truncate_item(trans, root, path, new_size, 1); + btrfs_truncate_item(trans, root, path, new_size, 1); } else if (key->offset >= bytenr && csum_end > end_byte && end_byte > key->offset) { /* @@ -525,15 +534,13 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, u32 new_size = (csum_end - end_byte) >> blocksize_bits; new_size *= csum_size; - ret = btrfs_truncate_item(trans, root, path, new_size, 0); + btrfs_truncate_item(trans, root, path, new_size, 0); key->offset = end_byte; - ret = btrfs_set_item_key_safe(trans, root, path, key); - BUG_ON(ret); + btrfs_set_item_key_safe(trans, root, path, key); } else { BUG(); } - return 0; } /* @@ -549,8 +556,7 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, u64 csum_end; struct extent_buffer *leaf; int ret; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); int blocksize_bits = root->fs_info->sb->s_blocksize_bits; root = root->fs_info->csum_root; @@ -639,13 +645,14 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, * item changed size or key */ ret = btrfs_split_item(trans, root, path, &key, offset); - BUG_ON(ret && ret != -EAGAIN); + if (ret && ret != -EAGAIN) { + btrfs_abort_transaction(trans, root, ret); + goto out; + } key.offset = end_byte - 1; } else { - ret = truncate_one_csum(trans, root, path, - &key, bytenr, len); - BUG_ON(ret); + truncate_one_csum(trans, root, path, &key, bytenr, len); if (key.offset < bytenr) break; } @@ -676,14 +683,14 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans, struct btrfs_sector_sum *sector_sum; u32 nritems; u32 ins_size; - u16 csum_size = - btrfs_super_csum_size(&root->fs_info->super_copy); + u16 csum_size = btrfs_super_csum_size(root->fs_info->super_copy); path = btrfs_alloc_path(); if (!path) return -ENOMEM; sector_sum = sums->sums; + trans->adding_csums = 1; again: next_offset = (u64)-1; found_next = 0; @@ -777,7 +784,7 @@ again: if (diff != csum_size) goto insert; - ret = btrfs_extend_item(trans, root, path, diff); + btrfs_extend_item(trans, root, path, diff); goto csum; } @@ -847,6 +854,7 @@ next_sector: goto again; } out: + trans->adding_csums = 0; btrfs_free_path(path); return ret; diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index e4e57d59edb7..5caf285c6e4d 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -65,6 +65,21 @@ struct inode_defrag { int cycled; }; +static int __compare_inode_defrag(struct inode_defrag *defrag1, + struct inode_defrag *defrag2) +{ + if (defrag1->root > defrag2->root) + return 1; + else if (defrag1->root < defrag2->root) + return -1; + else if (defrag1->ino > defrag2->ino) + return 1; + else if (defrag1->ino < defrag2->ino) + return -1; + else + return 0; +} + /* pop a record for an inode into the defrag tree. The lock * must be held already * @@ -81,15 +96,17 @@ static void __btrfs_add_inode_defrag(struct inode *inode, struct inode_defrag *entry; struct rb_node **p; struct rb_node *parent = NULL; + int ret; p = &root->fs_info->defrag_inodes.rb_node; while (*p) { parent = *p; entry = rb_entry(parent, struct inode_defrag, rb_node); - if (defrag->ino < entry->ino) + ret = __compare_inode_defrag(defrag, entry); + if (ret < 0) p = &parent->rb_left; - else if (defrag->ino > entry->ino) + else if (ret > 0) p = &parent->rb_right; else { /* if we're reinserting an entry for @@ -103,7 +120,7 @@ static void __btrfs_add_inode_defrag(struct inode *inode, goto exists; } } - BTRFS_I(inode)->in_defrag = 1; + set_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags); rb_link_node(&defrag->rb_node, parent, p); rb_insert_color(&defrag->rb_node, &root->fs_info->defrag_inodes); return; @@ -131,7 +148,7 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, if (btrfs_fs_closing(root->fs_info)) return 0; - if (BTRFS_I(inode)->in_defrag) + if (test_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags)) return 0; if (trans) @@ -148,7 +165,7 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, defrag->root = root->root_key.objectid; spin_lock(&root->fs_info->defrag_inodes_lock); - if (!BTRFS_I(inode)->in_defrag) + if (!test_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags)) __btrfs_add_inode_defrag(inode, defrag); else kfree(defrag); @@ -159,28 +176,35 @@ int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, /* * must be called with the defrag_inodes lock held */ -struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, u64 ino, +struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, + u64 root, u64 ino, struct rb_node **next) { struct inode_defrag *entry = NULL; + struct inode_defrag tmp; struct rb_node *p; struct rb_node *parent = NULL; + int ret; + + tmp.ino = ino; + tmp.root = root; p = info->defrag_inodes.rb_node; while (p) { parent = p; entry = rb_entry(parent, struct inode_defrag, rb_node); - if (ino < entry->ino) + ret = __compare_inode_defrag(&tmp, entry); + if (ret < 0) p = parent->rb_left; - else if (ino > entry->ino) + else if (ret > 0) p = parent->rb_right; else return entry; } if (next) { - while (parent && ino > entry->ino) { + while (parent && __compare_inode_defrag(&tmp, entry) > 0) { parent = rb_next(parent); entry = rb_entry(parent, struct inode_defrag, rb_node); } @@ -202,6 +226,7 @@ int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) struct btrfs_key key; struct btrfs_ioctl_defrag_range_args range; u64 first_ino = 0; + u64 root_objectid = 0; int num_defrag; int defrag_batch = 1024; @@ -214,11 +239,14 @@ int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) n = NULL; /* find an inode to defrag */ - defrag = btrfs_find_defrag_inode(fs_info, first_ino, &n); + defrag = btrfs_find_defrag_inode(fs_info, root_objectid, + first_ino, &n); if (!defrag) { - if (n) - defrag = rb_entry(n, struct inode_defrag, rb_node); - else if (first_ino) { + if (n) { + defrag = rb_entry(n, struct inode_defrag, + rb_node); + } else if (root_objectid || first_ino) { + root_objectid = 0; first_ino = 0; continue; } else { @@ -228,6 +256,7 @@ int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) /* remove it from the rbtree */ first_ino = defrag->ino + 1; + root_objectid = defrag->root; rb_erase(&defrag->rb_node, &fs_info->defrag_inodes); if (btrfs_fs_closing(fs_info)) @@ -252,7 +281,7 @@ int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) goto next; /* do a chunk of defrag */ - BTRFS_I(inode)->in_defrag = 0; + clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags); range.start = defrag->last_offset; num_defrag = btrfs_defrag_file(inode, NULL, &range, defrag->transid, defrag_batch); @@ -452,7 +481,7 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, split = alloc_extent_map(); if (!split2) split2 = alloc_extent_map(); - BUG_ON(!split || !split2); + BUG_ON(!split || !split2); /* -ENOMEM */ write_lock(&em_tree->lock); em = lookup_extent_mapping(em_tree, start, len); @@ -494,7 +523,7 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, split->flags = flags; split->compress_type = em->compress_type; ret = add_extent_mapping(em_tree, split); - BUG_ON(ret); + BUG_ON(ret); /* Logic error */ free_extent_map(split); split = split2; split2 = NULL; @@ -520,7 +549,7 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, } ret = add_extent_mapping(em_tree, split); - BUG_ON(ret); + BUG_ON(ret); /* Logic error */ free_extent_map(split); split = NULL; } @@ -567,6 +596,7 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, int extent_type; int recow; int ret; + int modify_tree = -1; if (drop_cache) btrfs_drop_extent_cache(inode, start, end - 1, 0); @@ -575,10 +605,13 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, if (!path) return -ENOMEM; + if (start >= BTRFS_I(inode)->disk_i_size) + modify_tree = 0; + while (1) { recow = 0; ret = btrfs_lookup_file_extent(trans, root, path, ino, - search_start, -1); + search_start, modify_tree); if (ret < 0) break; if (ret > 0 && path->slots[0] > 0 && search_start == start) { @@ -634,7 +667,8 @@ next_slot: } search_start = max(key.offset, start); - if (recow) { + if (recow || !modify_tree) { + modify_tree = -1; btrfs_release_path(path); continue; } @@ -678,8 +712,8 @@ next_slot: disk_bytenr, num_bytes, 0, root->root_key.objectid, new_key.objectid, - start - extent_offset); - BUG_ON(ret); + start - extent_offset, 0); + BUG_ON(ret); /* -ENOMEM */ *hint_byte = disk_bytenr; } key.offset = start; @@ -753,8 +787,8 @@ next_slot: disk_bytenr, num_bytes, 0, root->root_key.objectid, key.objectid, key.offset - - extent_offset); - BUG_ON(ret); + extent_offset, 0); + BUG_ON(ret); /* -ENOMEM */ inode_sub_bytes(inode, extent_end - key.offset); *hint_byte = disk_bytenr; @@ -770,7 +804,10 @@ next_slot: ret = btrfs_del_items(trans, root, path, del_slot, del_nr); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out; + } del_nr = 0; del_slot = 0; @@ -782,11 +819,13 @@ next_slot: BUG_ON(1); } - if (del_nr > 0) { + if (!ret && del_nr > 0) { ret = btrfs_del_items(trans, root, path, del_slot, del_nr); - BUG_ON(ret); + if (ret) + btrfs_abort_transaction(trans, root, ret); } +out: btrfs_free_path(path); return ret; } @@ -944,7 +983,10 @@ again: btrfs_release_path(path); goto again; } - BUG_ON(ret < 0); + if (ret < 0) { + btrfs_abort_transaction(trans, root, ret); + goto out; + } leaf = path->nodes[0]; fi = btrfs_item_ptr(leaf, path->slots[0] - 1, @@ -962,8 +1004,8 @@ again: ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - ino, orig_offset); - BUG_ON(ret); + ino, orig_offset, 0); + BUG_ON(ret); /* -ENOMEM */ if (split == start) { key.offset = start; @@ -989,8 +1031,8 @@ again: del_nr++; ret = btrfs_free_extent(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - ino, orig_offset); - BUG_ON(ret); + ino, orig_offset, 0); + BUG_ON(ret); /* -ENOMEM */ } other_start = 0; other_end = start; @@ -1006,8 +1048,8 @@ again: del_nr++; ret = btrfs_free_extent(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - ino, orig_offset); - BUG_ON(ret); + ino, orig_offset, 0); + BUG_ON(ret); /* -ENOMEM */ } if (del_nr == 0) { fi = btrfs_item_ptr(leaf, path->slots[0], @@ -1025,7 +1067,10 @@ again: btrfs_mark_buffer_dirty(leaf); ret = btrfs_del_items(trans, root, path, del_slot, del_nr); - BUG_ON(ret); + if (ret < 0) { + btrfs_abort_transaction(trans, root, ret); + goto out; + } } out: btrfs_free_path(path); @@ -1069,6 +1114,7 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, int i; unsigned long index = pos >> PAGE_CACHE_SHIFT; struct inode *inode = fdentry(file)->d_inode; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); int err = 0; int faili = 0; u64 start_pos; @@ -1080,7 +1126,7 @@ static noinline int prepare_pages(struct btrfs_root *root, struct file *file, again: for (i = 0; i < num_pages; i++) { pages[i] = find_or_create_page(inode->i_mapping, index + i, - GFP_NOFS); + mask | __GFP_WRITE); if (!pages[i]) { faili = i - 1; err = -ENOMEM; @@ -1104,8 +1150,7 @@ again: if (start_pos < inode->i_size) { struct btrfs_ordered_extent *ordered; lock_extent_bits(&BTRFS_I(inode)->io_tree, - start_pos, last_pos - 1, 0, &cached_state, - GFP_NOFS); + start_pos, last_pos - 1, 0, &cached_state); ordered = btrfs_lookup_first_ordered_extent(inode, last_pos - 1); if (ordered && @@ -1135,7 +1180,8 @@ again: GFP_NOFS); } for (i = 0; i < num_pages; i++) { - clear_page_dirty_for_io(pages[i]); + if (clear_page_dirty_for_io(pages[i])) + account_page_redirty(pages[i]); set_page_extent_mapped(pages[i]); WARN_ON(!PageLocked(pages[i])); } @@ -1166,6 +1212,8 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, nrptrs = min((iov_iter_count(i) + PAGE_CACHE_SIZE - 1) / PAGE_CACHE_SIZE, PAGE_CACHE_SIZE / (sizeof(struct page *))); + nrptrs = min(nrptrs, current->nr_dirtied_pause - current->nr_dirtied); + nrptrs = max(nrptrs, 8); pages = kmalloc(nrptrs * sizeof(struct page *), GFP_KERNEL); if (!pages) return -ENOMEM; @@ -1270,7 +1318,6 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, dirty_pages); if (dirty_pages < (root->leafsize >> PAGE_CACHE_SHIFT) + 1) btrfs_btree_balance_dirty(root, 1); - btrfs_throttle(root); pos += copied; num_written += copied; @@ -1287,7 +1334,6 @@ static ssize_t __btrfs_direct_write(struct kiocb *iocb, loff_t *ppos, size_t count, size_t ocount) { struct file *file = iocb->ki_filp; - struct inode *inode = fdentry(file)->d_inode; struct iov_iter i; ssize_t written; ssize_t written_buffered; @@ -1297,18 +1343,6 @@ static ssize_t __btrfs_direct_write(struct kiocb *iocb, written = generic_file_direct_write(iocb, iov, &nr_segs, pos, ppos, count, ocount); - /* - * the generic O_DIRECT will update in-memory i_size after the - * DIOs are done. But our endio handlers that update the on - * disk i_size never update past the in memory i_size. So we - * need one more update here to catch any additions to the - * file - */ - if (inode->i_size != BTRFS_I(inode)->disk_i_size) { - btrfs_ordered_update_i_size(inode, inode->i_size, NULL); - mark_inode_dirty(inode); - } - if (written < 0 || written == count) return written; @@ -1345,7 +1379,7 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, ssize_t err = 0; size_t count, ocount; - vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE); + sb_start_write(inode->i_sb); mutex_lock(&inode->i_mutex); @@ -1386,8 +1420,11 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, goto out; } - file_update_time(file); - BTRFS_I(inode)->sequence++; + err = file_update_time(file); + if (err) { + mutex_unlock(&inode->i_mutex); + goto out; + } start_pos = round_down(pos, root->sectorsize); if (start_pos > i_size_read(inode)) { @@ -1432,6 +1469,7 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb, num_written = err; } out: + sb_end_write(inode->i_sb); current->backing_dev_info = NULL; return num_written ? num_written : err; } @@ -1444,8 +1482,8 @@ int btrfs_release_file(struct inode *inode, struct file *filp) * flush down new bytes that may have been written if the * application were using truncate to replace a file in place. */ - if (BTRFS_I(inode)->ordered_data_close) { - BTRFS_I(inode)->ordered_data_close = 0; + if (test_and_clear_bit(BTRFS_INODE_ORDERED_DATA_CLOSE, + &BTRFS_I(inode)->runtime_flags)) { btrfs_add_ordered_operation(NULL, BTRFS_I(inode)->root, inode); if (inode->i_size > BTRFS_ORDERED_OPERATIONS_FLUSH_LIMIT) filemap_flush(inode->i_mapping); @@ -1476,14 +1514,15 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) trace_btrfs_sync_file(file, datasync); - ret = filemap_write_and_wait_range(inode->i_mapping, start, end); - if (ret) - return ret; mutex_lock(&inode->i_mutex); - /* we wait first, since the writeback may change the inode */ + /* + * we wait first, since the writeback may change the inode, also wait + * ordered range does a filemape_write_and_wait_range which is why we + * don't do it above like other file systems. + */ root->log_batch++; - btrfs_wait_ordered_range(inode, 0, (u64)-1); + btrfs_wait_ordered_range(inode, start, end); root->log_batch++; /* @@ -1501,7 +1540,8 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) * syncing */ smp_mb(); - if (BTRFS_I(inode)->last_trans <= + if (btrfs_inode_in_log(inode, root->fs_info->generation) || + BTRFS_I(inode)->last_trans <= root->fs_info->last_trans_committed) { BTRFS_I(inode)->last_trans = 0; mutex_unlock(&inode->i_mutex); @@ -1598,6 +1638,14 @@ static long btrfs_fallocate(struct file *file, int mode, return -EOPNOTSUPP; /* + * Make sure we have enough space before we do the + * allocation. + */ + ret = btrfs_check_data_free_space(inode, len); + if (ret) + return ret; + + /* * wait for ordered IO before we have any locks. We'll loop again * below with the locks held. */ @@ -1615,10 +1663,6 @@ static long btrfs_fallocate(struct file *file, int mode, goto out; } - ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start); - if (ret) - goto out; - locked_end = alloc_end - 1; while (1) { struct btrfs_ordered_extent *ordered; @@ -1627,7 +1671,7 @@ static long btrfs_fallocate(struct file *file, int mode, * transaction */ lock_extent_bits(&BTRFS_I(inode)->io_tree, alloc_start, - locked_end, 0, &cached_state, GFP_NOFS); + locked_end, 0, &cached_state); ordered = btrfs_lookup_first_ordered_extent(inode, alloc_end - 1); if (ordered && @@ -1656,7 +1700,13 @@ static long btrfs_fallocate(struct file *file, int mode, em = btrfs_get_extent(inode, NULL, 0, cur_offset, alloc_end - cur_offset, 0); - BUG_ON(IS_ERR_OR_NULL(em)); + if (IS_ERR_OR_NULL(em)) { + if (!em) + ret = -ENOMEM; + else + ret = PTR_ERR(em); + break; + } last_byte = min(extent_map_end(em), alloc_end); actual_end = min_t(u64, extent_map_end(em), offset + len); last_byte = (last_byte + mask) & ~mask; @@ -1669,6 +1719,7 @@ static long btrfs_fallocate(struct file *file, int mode, 1 << inode->i_blkbits, offset + len, &alloc_hint); + if (ret < 0) { free_extent_map(em); break; @@ -1694,10 +1745,10 @@ static long btrfs_fallocate(struct file *file, int mode, } unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end, &cached_state, GFP_NOFS); - - btrfs_free_reserved_data_space(inode, alloc_end - alloc_start); out: mutex_unlock(&inode->i_mutex); + /* Let go of our reservation. */ + btrfs_free_reserved_data_space(inode, len); return ret; } @@ -1725,7 +1776,7 @@ static int find_desired_extent(struct inode *inode, loff_t *offset, int origin) return -ENXIO; lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend, 0, - &cached_state, GFP_NOFS); + &cached_state); /* * Delalloc is such a pain. If we have a hole and we have pending @@ -1744,7 +1795,7 @@ static int find_desired_extent(struct inode *inode, loff_t *offset, int origin) start - root->sectorsize, root->sectorsize, 0); if (IS_ERR(em)) { - ret = -ENXIO; + ret = PTR_ERR(em); goto out; } last_end = em->start + em->len; @@ -1756,7 +1807,7 @@ static int find_desired_extent(struct inode *inode, loff_t *offset, int origin) while (1) { em = btrfs_get_extent_fiemap(inode, NULL, 0, start, len, 0); if (IS_ERR(em)) { - ret = -ENXIO; + ret = PTR_ERR(em); break; } @@ -1821,7 +1872,7 @@ static loff_t btrfs_file_llseek(struct file *file, loff_t offset, int origin) switch (origin) { case SEEK_END: case SEEK_CUR: - offset = generic_file_llseek_unlocked(file, offset, origin); + offset = generic_file_llseek(file, offset, origin); goto out; case SEEK_DATA: case SEEK_HOLE: diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 41ac927401d0..6b10acfc2f5c 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -20,6 +20,7 @@ #include <linux/sched.h> #include <linux/slab.h> #include <linux/math64.h> +#include <linux/ratelimit.h> #include "ctree.h" #include "free-space-cache.h" #include "transaction.h" @@ -32,6 +33,8 @@ static int link_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info); +static void unlink_free_space(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info); static struct inode *__lookup_free_space_inode(struct btrfs_root *root, struct btrfs_path *path, @@ -74,7 +77,8 @@ static struct inode *__lookup_free_space_inode(struct btrfs_root *root, return ERR_PTR(-ENOENT); } - inode->i_mapping->flags &= ~__GFP_FS; + mapping_set_gfp_mask(inode->i_mapping, + mapping_gfp_mask(inode->i_mapping) & ~__GFP_FS); return inode; } @@ -84,6 +88,7 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, *block_group, struct btrfs_path *path) { struct inode *inode = NULL; + u32 flags = BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW; spin_lock(&block_group->lock); if (block_group->inode) @@ -98,13 +103,14 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, return inode; spin_lock(&block_group->lock); - if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) { + if (!((BTRFS_I(inode)->flags & flags) == flags)) { printk(KERN_INFO "Old style space inode found, converting.\n"); - BTRFS_I(inode)->flags &= ~BTRFS_INODE_NODATASUM; + BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM | + BTRFS_INODE_NODATACOW; block_group->disk_cache_state = BTRFS_DC_CLEAR; } - if (!btrfs_fs_closing(root->fs_info)) { + if (!block_group->iref) { block_group->inode = igrab(inode); block_group->iref = 1; } @@ -122,12 +128,17 @@ int __create_free_space_inode(struct btrfs_root *root, struct btrfs_free_space_header *header; struct btrfs_inode_item *inode_item; struct extent_buffer *leaf; + u64 flags = BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC; int ret; ret = btrfs_insert_empty_inode(trans, root, path, ino); if (ret) return ret; + /* We inline crc's for the free disk space cache */ + if (ino != BTRFS_FREE_INO_OBJECTID) + flags |= BTRFS_INODE_NODATASUM | BTRFS_INODE_NODATACOW; + leaf = path->nodes[0]; inode_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item); @@ -140,8 +151,7 @@ int __create_free_space_inode(struct btrfs_root *root, btrfs_set_inode_uid(leaf, inode_item, 0); btrfs_set_inode_gid(leaf, inode_item, 0); btrfs_set_inode_mode(leaf, inode_item, S_IFREG | 0600); - btrfs_set_inode_flags(leaf, inode_item, BTRFS_INODE_NOCOMPRESS | - BTRFS_INODE_PREALLOC); + btrfs_set_inode_flags(leaf, inode_item, flags); btrfs_set_inode_nlink(leaf, inode_item, 1); btrfs_set_inode_transid(leaf, inode_item, trans->transid); btrfs_set_inode_block_group(leaf, inode_item, offset); @@ -191,16 +201,24 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, struct inode *inode) { struct btrfs_block_rsv *rsv; + u64 needed_bytes; loff_t oldsize; int ret = 0; rsv = trans->block_rsv; - trans->block_rsv = root->orphan_block_rsv; - ret = btrfs_block_rsv_check(trans, root, - root->orphan_block_rsv, - 0, 5); - if (ret) - return ret; + trans->block_rsv = &root->fs_info->global_block_rsv; + + /* 1 for slack space, 1 for updating the inode */ + needed_bytes = btrfs_calc_trunc_metadata_size(root, 1) + + btrfs_calc_trans_metadata_size(root, 1); + + spin_lock(&trans->block_rsv->lock); + if (trans->block_rsv->reserved < needed_bytes) { + spin_unlock(&trans->block_rsv->lock); + trans->block_rsv = rsv; + return -ENOSPC; + } + spin_unlock(&trans->block_rsv->lock); oldsize = i_size_read(inode); btrfs_i_size_write(inode, 0); @@ -213,13 +231,17 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, ret = btrfs_truncate_inode_items(trans, root, inode, 0, BTRFS_EXTENT_DATA_KEY); - trans->block_rsv = rsv; if (ret) { - WARN_ON(1); + trans->block_rsv = rsv; + btrfs_abort_transaction(trans, root, ret); return ret; } ret = btrfs_update_inode(trans, root, inode); + if (ret) + btrfs_abort_transaction(trans, root, ret); + trans->block_rsv = rsv; + return ret; } @@ -242,26 +264,388 @@ static int readahead_cache(struct inode *inode) return 0; } +struct io_ctl { + void *cur, *orig; + struct page *page; + struct page **pages; + struct btrfs_root *root; + unsigned long size; + int index; + int num_pages; + unsigned check_crcs:1; +}; + +static int io_ctl_init(struct io_ctl *io_ctl, struct inode *inode, + struct btrfs_root *root) +{ + memset(io_ctl, 0, sizeof(struct io_ctl)); + io_ctl->num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> + PAGE_CACHE_SHIFT; + io_ctl->pages = kzalloc(sizeof(struct page *) * io_ctl->num_pages, + GFP_NOFS); + if (!io_ctl->pages) + return -ENOMEM; + io_ctl->root = root; + if (btrfs_ino(inode) != BTRFS_FREE_INO_OBJECTID) + io_ctl->check_crcs = 1; + return 0; +} + +static void io_ctl_free(struct io_ctl *io_ctl) +{ + kfree(io_ctl->pages); +} + +static void io_ctl_unmap_page(struct io_ctl *io_ctl) +{ + if (io_ctl->cur) { + kunmap(io_ctl->page); + io_ctl->cur = NULL; + io_ctl->orig = NULL; + } +} + +static void io_ctl_map_page(struct io_ctl *io_ctl, int clear) +{ + WARN_ON(io_ctl->cur); + BUG_ON(io_ctl->index >= io_ctl->num_pages); + io_ctl->page = io_ctl->pages[io_ctl->index++]; + io_ctl->cur = kmap(io_ctl->page); + io_ctl->orig = io_ctl->cur; + io_ctl->size = PAGE_CACHE_SIZE; + if (clear) + memset(io_ctl->cur, 0, PAGE_CACHE_SIZE); +} + +static void io_ctl_drop_pages(struct io_ctl *io_ctl) +{ + int i; + + io_ctl_unmap_page(io_ctl); + + for (i = 0; i < io_ctl->num_pages; i++) { + if (io_ctl->pages[i]) { + ClearPageChecked(io_ctl->pages[i]); + unlock_page(io_ctl->pages[i]); + page_cache_release(io_ctl->pages[i]); + } + } +} + +static int io_ctl_prepare_pages(struct io_ctl *io_ctl, struct inode *inode, + int uptodate) +{ + struct page *page; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); + int i; + + for (i = 0; i < io_ctl->num_pages; i++) { + page = find_or_create_page(inode->i_mapping, i, mask); + if (!page) { + io_ctl_drop_pages(io_ctl); + return -ENOMEM; + } + io_ctl->pages[i] = page; + if (uptodate && !PageUptodate(page)) { + btrfs_readpage(NULL, page); + lock_page(page); + if (!PageUptodate(page)) { + printk(KERN_ERR "btrfs: error reading free " + "space cache\n"); + io_ctl_drop_pages(io_ctl); + return -EIO; + } + } + } + + for (i = 0; i < io_ctl->num_pages; i++) { + clear_page_dirty_for_io(io_ctl->pages[i]); + set_page_extent_mapped(io_ctl->pages[i]); + } + + return 0; +} + +static void io_ctl_set_generation(struct io_ctl *io_ctl, u64 generation) +{ + __le64 *val; + + io_ctl_map_page(io_ctl, 1); + + /* + * Skip the csum areas. If we don't check crcs then we just have a + * 64bit chunk at the front of the first page. + */ + if (io_ctl->check_crcs) { + io_ctl->cur += (sizeof(u32) * io_ctl->num_pages); + io_ctl->size -= sizeof(u64) + (sizeof(u32) * io_ctl->num_pages); + } else { + io_ctl->cur += sizeof(u64); + io_ctl->size -= sizeof(u64) * 2; + } + + val = io_ctl->cur; + *val = cpu_to_le64(generation); + io_ctl->cur += sizeof(u64); +} + +static int io_ctl_check_generation(struct io_ctl *io_ctl, u64 generation) +{ + __le64 *gen; + + /* + * Skip the crc area. If we don't check crcs then we just have a 64bit + * chunk at the front of the first page. + */ + if (io_ctl->check_crcs) { + io_ctl->cur += sizeof(u32) * io_ctl->num_pages; + io_ctl->size -= sizeof(u64) + + (sizeof(u32) * io_ctl->num_pages); + } else { + io_ctl->cur += sizeof(u64); + io_ctl->size -= sizeof(u64) * 2; + } + + gen = io_ctl->cur; + if (le64_to_cpu(*gen) != generation) { + printk_ratelimited(KERN_ERR "btrfs: space cache generation " + "(%Lu) does not match inode (%Lu)\n", *gen, + generation); + io_ctl_unmap_page(io_ctl); + return -EIO; + } + io_ctl->cur += sizeof(u64); + return 0; +} + +static void io_ctl_set_crc(struct io_ctl *io_ctl, int index) +{ + u32 *tmp; + u32 crc = ~(u32)0; + unsigned offset = 0; + + if (!io_ctl->check_crcs) { + io_ctl_unmap_page(io_ctl); + return; + } + + if (index == 0) + offset = sizeof(u32) * io_ctl->num_pages; + + crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc, + PAGE_CACHE_SIZE - offset); + btrfs_csum_final(crc, (char *)&crc); + io_ctl_unmap_page(io_ctl); + tmp = kmap(io_ctl->pages[0]); + tmp += index; + *tmp = crc; + kunmap(io_ctl->pages[0]); +} + +static int io_ctl_check_crc(struct io_ctl *io_ctl, int index) +{ + u32 *tmp, val; + u32 crc = ~(u32)0; + unsigned offset = 0; + + if (!io_ctl->check_crcs) { + io_ctl_map_page(io_ctl, 0); + return 0; + } + + if (index == 0) + offset = sizeof(u32) * io_ctl->num_pages; + + tmp = kmap(io_ctl->pages[0]); + tmp += index; + val = *tmp; + kunmap(io_ctl->pages[0]); + + io_ctl_map_page(io_ctl, 0); + crc = btrfs_csum_data(io_ctl->root, io_ctl->orig + offset, crc, + PAGE_CACHE_SIZE - offset); + btrfs_csum_final(crc, (char *)&crc); + if (val != crc) { + printk_ratelimited(KERN_ERR "btrfs: csum mismatch on free " + "space cache\n"); + io_ctl_unmap_page(io_ctl); + return -EIO; + } + + return 0; +} + +static int io_ctl_add_entry(struct io_ctl *io_ctl, u64 offset, u64 bytes, + void *bitmap) +{ + struct btrfs_free_space_entry *entry; + + if (!io_ctl->cur) + return -ENOSPC; + + entry = io_ctl->cur; + entry->offset = cpu_to_le64(offset); + entry->bytes = cpu_to_le64(bytes); + entry->type = (bitmap) ? BTRFS_FREE_SPACE_BITMAP : + BTRFS_FREE_SPACE_EXTENT; + io_ctl->cur += sizeof(struct btrfs_free_space_entry); + io_ctl->size -= sizeof(struct btrfs_free_space_entry); + + if (io_ctl->size >= sizeof(struct btrfs_free_space_entry)) + return 0; + + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + + /* No more pages to map */ + if (io_ctl->index >= io_ctl->num_pages) + return 0; + + /* map the next page */ + io_ctl_map_page(io_ctl, 1); + return 0; +} + +static int io_ctl_add_bitmap(struct io_ctl *io_ctl, void *bitmap) +{ + if (!io_ctl->cur) + return -ENOSPC; + + /* + * If we aren't at the start of the current page, unmap this one and + * map the next one if there is any left. + */ + if (io_ctl->cur != io_ctl->orig) { + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + if (io_ctl->index >= io_ctl->num_pages) + return -ENOSPC; + io_ctl_map_page(io_ctl, 0); + } + + memcpy(io_ctl->cur, bitmap, PAGE_CACHE_SIZE); + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + if (io_ctl->index < io_ctl->num_pages) + io_ctl_map_page(io_ctl, 0); + return 0; +} + +static void io_ctl_zero_remaining_pages(struct io_ctl *io_ctl) +{ + /* + * If we're not on the boundary we know we've modified the page and we + * need to crc the page. + */ + if (io_ctl->cur != io_ctl->orig) + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + else + io_ctl_unmap_page(io_ctl); + + while (io_ctl->index < io_ctl->num_pages) { + io_ctl_map_page(io_ctl, 1); + io_ctl_set_crc(io_ctl, io_ctl->index - 1); + } +} + +static int io_ctl_read_entry(struct io_ctl *io_ctl, + struct btrfs_free_space *entry, u8 *type) +{ + struct btrfs_free_space_entry *e; + int ret; + + if (!io_ctl->cur) { + ret = io_ctl_check_crc(io_ctl, io_ctl->index); + if (ret) + return ret; + } + + e = io_ctl->cur; + entry->offset = le64_to_cpu(e->offset); + entry->bytes = le64_to_cpu(e->bytes); + *type = e->type; + io_ctl->cur += sizeof(struct btrfs_free_space_entry); + io_ctl->size -= sizeof(struct btrfs_free_space_entry); + + if (io_ctl->size >= sizeof(struct btrfs_free_space_entry)) + return 0; + + io_ctl_unmap_page(io_ctl); + + return 0; +} + +static int io_ctl_read_bitmap(struct io_ctl *io_ctl, + struct btrfs_free_space *entry) +{ + int ret; + + ret = io_ctl_check_crc(io_ctl, io_ctl->index); + if (ret) + return ret; + + memcpy(entry->bitmap, io_ctl->cur, PAGE_CACHE_SIZE); + io_ctl_unmap_page(io_ctl); + + return 0; +} + +/* + * Since we attach pinned extents after the fact we can have contiguous sections + * of free space that are split up in entries. This poses a problem with the + * tree logging stuff since it could have allocated across what appears to be 2 + * entries since we would have merged the entries when adding the pinned extents + * back to the free space cache. So run through the space cache that we just + * loaded and merge contiguous entries. This will make the log replay stuff not + * blow up and it will make for nicer allocator behavior. + */ +static void merge_space_tree(struct btrfs_free_space_ctl *ctl) +{ + struct btrfs_free_space *e, *prev = NULL; + struct rb_node *n; + +again: + spin_lock(&ctl->tree_lock); + for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) { + e = rb_entry(n, struct btrfs_free_space, offset_index); + if (!prev) + goto next; + if (e->bitmap || prev->bitmap) + goto next; + if (prev->offset + prev->bytes == e->offset) { + unlink_free_space(ctl, prev); + unlink_free_space(ctl, e); + prev->bytes += e->bytes; + kmem_cache_free(btrfs_free_space_cachep, e); + link_free_space(ctl, prev); + prev = NULL; + spin_unlock(&ctl->tree_lock); + goto again; + } +next: + prev = e; + } + spin_unlock(&ctl->tree_lock); +} + int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, struct btrfs_free_space_ctl *ctl, struct btrfs_path *path, u64 offset) { struct btrfs_free_space_header *header; struct extent_buffer *leaf; - struct page *page; + struct io_ctl io_ctl; struct btrfs_key key; + struct btrfs_free_space *e, *n; struct list_head bitmaps; u64 num_entries; u64 num_bitmaps; u64 generation; - pgoff_t index = 0; + u8 type; int ret = 0; INIT_LIST_HEAD(&bitmaps); /* Nothing in the space cache, goodbye */ if (!i_size_read(inode)) - goto out; + return 0; key.objectid = BTRFS_FREE_SPACE_OBJECTID; key.offset = offset; @@ -269,11 +653,10 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) - goto out; + return 0; else if (ret > 0) { btrfs_release_path(path); - ret = 0; - goto out; + return 0; } ret = -1; @@ -291,169 +674,106 @@ int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, " not match free space cache generation (%llu)\n", (unsigned long long)BTRFS_I(inode)->generation, (unsigned long long)generation); - goto out; + return 0; } if (!num_entries) - goto out; + return 0; + + ret = io_ctl_init(&io_ctl, inode, root); + if (ret) + return ret; ret = readahead_cache(inode); if (ret) goto out; - while (1) { - struct btrfs_free_space_entry *entry; - struct btrfs_free_space *e; - void *addr; - unsigned long offset = 0; - int need_loop = 0; + ret = io_ctl_prepare_pages(&io_ctl, inode, 1); + if (ret) + goto out; - if (!num_entries && !num_bitmaps) - break; + ret = io_ctl_check_crc(&io_ctl, 0); + if (ret) + goto free_cache; + + ret = io_ctl_check_generation(&io_ctl, generation); + if (ret) + goto free_cache; - page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); - if (!page) + while (num_entries) { + e = kmem_cache_zalloc(btrfs_free_space_cachep, + GFP_NOFS); + if (!e) goto free_cache; - if (!PageUptodate(page)) { - btrfs_readpage(NULL, page); - lock_page(page); - if (!PageUptodate(page)) { - unlock_page(page); - page_cache_release(page); - printk(KERN_ERR "btrfs: error reading free " - "space cache\n"); - goto free_cache; - } + ret = io_ctl_read_entry(&io_ctl, e, &type); + if (ret) { + kmem_cache_free(btrfs_free_space_cachep, e); + goto free_cache; } - addr = kmap(page); - if (index == 0) { - u64 *gen; + if (!e->bytes) { + kmem_cache_free(btrfs_free_space_cachep, e); + goto free_cache; + } - /* - * We put a bogus crc in the front of the first page in - * case old kernels try to mount a fs with the new - * format to make sure they discard the cache. - */ - addr += sizeof(u64); - offset += sizeof(u64); - - gen = addr; - if (*gen != BTRFS_I(inode)->generation) { - printk(KERN_ERR "btrfs: space cache generation" - " (%llu) does not match inode (%llu)\n", - (unsigned long long)*gen, - (unsigned long long) - BTRFS_I(inode)->generation); - kunmap(page); - unlock_page(page); - page_cache_release(page); + if (type == BTRFS_FREE_SPACE_EXTENT) { + spin_lock(&ctl->tree_lock); + ret = link_free_space(ctl, e); + spin_unlock(&ctl->tree_lock); + if (ret) { + printk(KERN_ERR "Duplicate entries in " + "free space cache, dumping\n"); + kmem_cache_free(btrfs_free_space_cachep, e); goto free_cache; } - addr += sizeof(u64); - offset += sizeof(u64); - } - entry = addr; - - while (1) { - if (!num_entries) - break; - - need_loop = 1; - e = kmem_cache_zalloc(btrfs_free_space_cachep, - GFP_NOFS); - if (!e) { - kunmap(page); - unlock_page(page); - page_cache_release(page); + } else { + BUG_ON(!num_bitmaps); + num_bitmaps--; + e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); + if (!e->bitmap) { + kmem_cache_free( + btrfs_free_space_cachep, e); goto free_cache; } - - e->offset = le64_to_cpu(entry->offset); - e->bytes = le64_to_cpu(entry->bytes); - if (!e->bytes) { - kunmap(page); + spin_lock(&ctl->tree_lock); + ret = link_free_space(ctl, e); + ctl->total_bitmaps++; + ctl->op->recalc_thresholds(ctl); + spin_unlock(&ctl->tree_lock); + if (ret) { + printk(KERN_ERR "Duplicate entries in " + "free space cache, dumping\n"); kmem_cache_free(btrfs_free_space_cachep, e); - unlock_page(page); - page_cache_release(page); goto free_cache; } - - if (entry->type == BTRFS_FREE_SPACE_EXTENT) { - spin_lock(&ctl->tree_lock); - ret = link_free_space(ctl, e); - spin_unlock(&ctl->tree_lock); - if (ret) { - printk(KERN_ERR "Duplicate entries in " - "free space cache, dumping\n"); - kunmap(page); - unlock_page(page); - page_cache_release(page); - goto free_cache; - } - } else { - e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); - if (!e->bitmap) { - kunmap(page); - kmem_cache_free( - btrfs_free_space_cachep, e); - unlock_page(page); - page_cache_release(page); - goto free_cache; - } - spin_lock(&ctl->tree_lock); - ret = link_free_space(ctl, e); - ctl->total_bitmaps++; - ctl->op->recalc_thresholds(ctl); - spin_unlock(&ctl->tree_lock); - if (ret) { - printk(KERN_ERR "Duplicate entries in " - "free space cache, dumping\n"); - kunmap(page); - unlock_page(page); - page_cache_release(page); - goto free_cache; - } - list_add_tail(&e->list, &bitmaps); - } - - num_entries--; - offset += sizeof(struct btrfs_free_space_entry); - if (offset + sizeof(struct btrfs_free_space_entry) >= - PAGE_CACHE_SIZE) - break; - entry++; + list_add_tail(&e->list, &bitmaps); } - /* - * We read an entry out of this page, we need to move on to the - * next page. - */ - if (need_loop) { - kunmap(page); - goto next; - } + num_entries--; + } - /* - * We add the bitmaps at the end of the entries in order that - * the bitmap entries are added to the cache. - */ - e = list_entry(bitmaps.next, struct btrfs_free_space, list); + io_ctl_unmap_page(&io_ctl); + + /* + * We add the bitmaps at the end of the entries in order that + * the bitmap entries are added to the cache. + */ + list_for_each_entry_safe(e, n, &bitmaps, list) { list_del_init(&e->list); - memcpy(e->bitmap, addr, PAGE_CACHE_SIZE); - kunmap(page); - num_bitmaps--; -next: - unlock_page(page); - page_cache_release(page); - index++; + ret = io_ctl_read_bitmap(&io_ctl, e); + if (ret) + goto free_cache; } + io_ctl_drop_pages(&io_ctl); + merge_space_tree(ctl); ret = 1; out: + io_ctl_free(&io_ctl); return ret; free_cache: + io_ctl_drop_pages(&io_ctl); __btrfs_remove_free_space_cache(ctl); goto out; } @@ -465,18 +785,11 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root = fs_info->tree_root; struct inode *inode; struct btrfs_path *path; - int ret; + int ret = 0; bool matched; u64 used = btrfs_block_group_used(&block_group->item); /* - * If we're unmounting then just return, since this does a search on the - * normal root and not the commit root and we could deadlock. - */ - if (btrfs_fs_closing(fs_info)) - return 0; - - /* * If this block group has been marked to be cleared for one reason or * another then we can't trust the on disk cache, so just return. */ @@ -490,6 +803,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, path = btrfs_alloc_path(); if (!path) return 0; + path->search_commit_root = 1; + path->skip_locking = 1; inode = lookup_free_space_inode(root, block_group, path); if (IS_ERR(inode)) { @@ -497,6 +812,15 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, return 0; } + /* We may have converted the inode and made the cache invalid. */ + spin_lock(&block_group->lock); + if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { + spin_unlock(&block_group->lock); + btrfs_free_path(path); + goto out; + } + spin_unlock(&block_group->lock); + ret = __load_free_space_cache(fs_info->tree_root, inode, ctl, path, block_group->key.objectid); btrfs_free_path(path); @@ -530,6 +854,19 @@ out: return ret; } +/** + * __btrfs_write_out_cache - write out cached info to an inode + * @root - the root the inode belongs to + * @ctl - the free space cache we are going to write out + * @block_group - the block_group for this cache if it belongs to a block_group + * @trans - the trans handle + * @path - the path to use + * @offset - the offset for the key we'll insert + * + * This function writes out a free space cache struct to disk for quick recovery + * on mount. This will return 0 if it was successfull in writing the cache out, + * and -1 if it was not. + */ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, struct btrfs_free_space_ctl *ctl, struct btrfs_block_group_cache *block_group, @@ -540,41 +877,25 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, struct extent_buffer *leaf; struct rb_node *node; struct list_head *pos, *n; - struct page **pages; - struct page *page; struct extent_state *cached_state = NULL; struct btrfs_free_cluster *cluster = NULL; struct extent_io_tree *unpin = NULL; + struct io_ctl io_ctl; struct list_head bitmap_list; struct btrfs_key key; - u64 start, end, len; - u64 bytes = 0; - u32 crc = ~(u32)0; - int index = 0, num_pages = 0; + u64 start, extent_start, extent_end, len; int entries = 0; int bitmaps = 0; - int ret = -1; - bool next_page = false; - bool out_of_space = false; + int ret; + int err = -1; INIT_LIST_HEAD(&bitmap_list); - node = rb_first(&ctl->free_space_offset); - if (!node) - return 0; - if (!i_size_read(inode)) return -1; - num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> - PAGE_CACHE_SHIFT; - - filemap_write_and_wait(inode->i_mapping); - btrfs_wait_ordered_range(inode, inode->i_size & - ~(root->sectorsize - 1), (u64)-1); - - pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS); - if (!pages) + ret = io_ctl_init(&io_ctl, inode, root); + if (ret) return -1; /* Get the cluster for this block_group if it exists */ @@ -583,229 +904,127 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, struct btrfs_free_cluster, block_group_list); - /* - * We shouldn't have switched the pinned extents yet so this is the - * right one - */ - unpin = root->fs_info->pinned_extents; + /* Lock all pages first so we can lock the extent safely. */ + io_ctl_prepare_pages(&io_ctl, inode, 0); - /* - * Lock all pages first so we can lock the extent safely. - * - * NOTE: Because we hold the ref the entire time we're going to write to - * the page find_get_page should never fail, so we don't do a check - * after find_get_page at this point. Just putting this here so people - * know and don't freak out. - */ - while (index < num_pages) { - page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); - if (!page) { - int i; + lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, + 0, &cached_state); - for (i = 0; i < num_pages; i++) { - unlock_page(pages[i]); - page_cache_release(pages[i]); - } - goto out; - } - pages[index] = page; - index++; + node = rb_first(&ctl->free_space_offset); + if (!node && cluster) { + node = rb_first(&cluster->root); + cluster = NULL; } - index = 0; - lock_extent_bits(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, - 0, &cached_state, GFP_NOFS); + /* Make sure we can fit our crcs into the first page */ + if (io_ctl.check_crcs && + (io_ctl.num_pages * sizeof(u32)) >= PAGE_CACHE_SIZE) { + WARN_ON(1); + goto out_nospc; + } - /* - * When searching for pinned extents, we need to start at our start - * offset. - */ - if (block_group) - start = block_group->key.objectid; + io_ctl_set_generation(&io_ctl, trans->transid); /* Write out the extent entries */ - do { - struct btrfs_free_space_entry *entry; - void *addr, *orig; - unsigned long offset = 0; - - next_page = false; - - if (index >= num_pages) { - out_of_space = true; - break; - } - - page = pages[index]; + while (node) { + struct btrfs_free_space *e; - orig = addr = kmap(page); - if (index == 0) { - u64 *gen; + e = rb_entry(node, struct btrfs_free_space, offset_index); + entries++; - /* - * We're going to put in a bogus crc for this page to - * make sure that old kernels who aren't aware of this - * format will be sure to discard the cache. - */ - addr += sizeof(u64); - offset += sizeof(u64); + ret = io_ctl_add_entry(&io_ctl, e->offset, e->bytes, + e->bitmap); + if (ret) + goto out_nospc; - gen = addr; - *gen = trans->transid; - addr += sizeof(u64); - offset += sizeof(u64); + if (e->bitmap) { + list_add_tail(&e->list, &bitmap_list); + bitmaps++; } - entry = addr; - - memset(addr, 0, PAGE_CACHE_SIZE - offset); - while (node && !next_page) { - struct btrfs_free_space *e; - - e = rb_entry(node, struct btrfs_free_space, offset_index); - entries++; - - entry->offset = cpu_to_le64(e->offset); - entry->bytes = cpu_to_le64(e->bytes); - if (e->bitmap) { - entry->type = BTRFS_FREE_SPACE_BITMAP; - list_add_tail(&e->list, &bitmap_list); - bitmaps++; - } else { - entry->type = BTRFS_FREE_SPACE_EXTENT; - } - node = rb_next(node); - if (!node && cluster) { - node = rb_first(&cluster->root); - cluster = NULL; - } - offset += sizeof(struct btrfs_free_space_entry); - if (offset + sizeof(struct btrfs_free_space_entry) >= - PAGE_CACHE_SIZE) - next_page = true; - entry++; + node = rb_next(node); + if (!node && cluster) { + node = rb_first(&cluster->root); + cluster = NULL; } + } - /* - * We want to add any pinned extents to our free space cache - * so we don't leak the space - */ - while (block_group && !next_page && - (start < block_group->key.objectid + - block_group->key.offset)) { - ret = find_first_extent_bit(unpin, start, &start, &end, - EXTENT_DIRTY); - if (ret) { - ret = 0; - break; - } + /* + * We want to add any pinned extents to our free space cache + * so we don't leak the space + */ - /* This pinned extent is out of our range */ - if (start >= block_group->key.objectid + - block_group->key.offset) - break; + /* + * We shouldn't have switched the pinned extents yet so this is the + * right one + */ + unpin = root->fs_info->pinned_extents; - len = block_group->key.objectid + - block_group->key.offset - start; - len = min(len, end + 1 - start); - - entries++; - entry->offset = cpu_to_le64(start); - entry->bytes = cpu_to_le64(len); - entry->type = BTRFS_FREE_SPACE_EXTENT; - - start = end + 1; - offset += sizeof(struct btrfs_free_space_entry); - if (offset + sizeof(struct btrfs_free_space_entry) >= - PAGE_CACHE_SIZE) - next_page = true; - entry++; - } + if (block_group) + start = block_group->key.objectid; - /* Generate bogus crc value */ - if (index == 0) { - u32 *tmp; - crc = btrfs_csum_data(root, orig + sizeof(u64), crc, - PAGE_CACHE_SIZE - sizeof(u64)); - btrfs_csum_final(crc, (char *)&crc); - crc++; - tmp = orig; - *tmp = crc; + while (block_group && (start < block_group->key.objectid + + block_group->key.offset)) { + ret = find_first_extent_bit(unpin, start, + &extent_start, &extent_end, + EXTENT_DIRTY); + if (ret) { + ret = 0; + break; } - kunmap(page); + /* This pinned extent is out of our range */ + if (extent_start >= block_group->key.objectid + + block_group->key.offset) + break; + + extent_start = max(extent_start, start); + extent_end = min(block_group->key.objectid + + block_group->key.offset, extent_end + 1); + len = extent_end - extent_start; - bytes += PAGE_CACHE_SIZE; + entries++; + ret = io_ctl_add_entry(&io_ctl, extent_start, len, NULL); + if (ret) + goto out_nospc; - index++; - } while (node || next_page); + start = extent_end; + } /* Write out the bitmaps */ list_for_each_safe(pos, n, &bitmap_list) { - void *addr; struct btrfs_free_space *entry = list_entry(pos, struct btrfs_free_space, list); - if (index >= num_pages) { - out_of_space = true; - break; - } - page = pages[index]; - - addr = kmap(page); - memcpy(addr, entry->bitmap, PAGE_CACHE_SIZE); - kunmap(page); - bytes += PAGE_CACHE_SIZE; - + ret = io_ctl_add_bitmap(&io_ctl, entry->bitmap); + if (ret) + goto out_nospc; list_del_init(&entry->list); - index++; - } - - if (out_of_space) { - btrfs_drop_pages(pages, num_pages); - unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, - i_size_read(inode) - 1, &cached_state, - GFP_NOFS); - ret = 0; - goto out; } /* Zero out the rest of the pages just to make sure */ - while (index < num_pages) { - void *addr; - - page = pages[index]; - addr = kmap(page); - memset(addr, 0, PAGE_CACHE_SIZE); - kunmap(page); - bytes += PAGE_CACHE_SIZE; - index++; - } + io_ctl_zero_remaining_pages(&io_ctl); - ret = btrfs_dirty_pages(root, inode, pages, num_pages, 0, - bytes, &cached_state); - btrfs_drop_pages(pages, num_pages); + ret = btrfs_dirty_pages(root, inode, io_ctl.pages, io_ctl.num_pages, + 0, i_size_read(inode), &cached_state); + io_ctl_drop_pages(&io_ctl); unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, i_size_read(inode) - 1, &cached_state, GFP_NOFS); - if (ret) { - ret = 0; + if (ret) goto out; - } - BTRFS_I(inode)->generation = trans->transid; - filemap_write_and_wait(inode->i_mapping); + btrfs_wait_ordered_range(inode, 0, (u64)-1); key.objectid = BTRFS_FREE_SPACE_OBJECTID; key.offset = offset; key.type = 0; - ret = btrfs_search_slot(trans, root, &key, path, 1, 1); + ret = btrfs_search_slot(trans, root, &key, path, 0, 1); if (ret < 0) { - ret = -1; - clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, - EXTENT_DIRTY | EXTENT_DELALLOC | - EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS); + clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, inode->i_size - 1, + EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, NULL, + GFP_NOFS); goto out; } leaf = path->nodes[0]; @@ -816,15 +1035,16 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID || found_key.offset != offset) { - ret = -1; - clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, - EXTENT_DIRTY | EXTENT_DELALLOC | - EXTENT_DO_ACCOUNTING, 0, 0, NULL, - GFP_NOFS); + clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, + inode->i_size - 1, + EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, + NULL, GFP_NOFS); btrfs_release_path(path); goto out; } } + + BTRFS_I(inode)->generation = trans->transid; header = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_free_space_header); btrfs_set_free_space_entries(leaf, header, entries); @@ -833,16 +1053,26 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, btrfs_mark_buffer_dirty(leaf); btrfs_release_path(path); - ret = 1; - + err = 0; out: - kfree(pages); - if (ret != 1) { - invalidate_inode_pages2_range(inode->i_mapping, 0, index); + io_ctl_free(&io_ctl); + if (err) { + invalidate_inode_pages2(inode->i_mapping); BTRFS_I(inode)->generation = 0; } btrfs_update_inode(trans, root, inode); - return ret; + return err; + +out_nospc: + list_for_each_safe(pos, n, &bitmap_list) { + struct btrfs_free_space *entry = + list_entry(pos, struct btrfs_free_space, list); + list_del_init(&entry->list); + } + io_ctl_drop_pages(&io_ctl); + unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0, + i_size_read(inode) - 1, &cached_state, GFP_NOFS); + goto out; } int btrfs_write_out_cache(struct btrfs_root *root, @@ -869,14 +1099,15 @@ int btrfs_write_out_cache(struct btrfs_root *root, ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans, path, block_group->key.objectid); - if (ret < 0) { + if (ret) { spin_lock(&block_group->lock); block_group->disk_cache_state = BTRFS_DC_ERROR; spin_unlock(&block_group->lock); ret = 0; - - printk(KERN_ERR "btrfs: failed to write free space cace " +#ifdef DEBUG + printk(KERN_ERR "btrfs: failed to write free space cache " "for block group %llu\n", block_group->key.objectid); +#endif } iput(inode); @@ -1283,6 +1514,7 @@ static void add_new_bitmap(struct btrfs_free_space_ctl *ctl, { info->offset = offset_to_bitmap(ctl, offset); info->bytes = 0; + INIT_LIST_HEAD(&info->list); link_free_space(ctl, info); ctl->total_bitmaps++; @@ -1311,29 +1543,26 @@ again: end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1; /* - * XXX - this can go away after a few releases. - * - * since the only user of btrfs_remove_free_space is the tree logging - * stuff, and the only way to test that is under crash conditions, we - * want to have this debug stuff here just in case somethings not - * working. Search the bitmap for the space we are trying to use to - * make sure its actually there. If its not there then we need to stop - * because something has gone wrong. + * We need to search for bits in this bitmap. We could only cover some + * of the extent in this bitmap thanks to how we add space, so we need + * to search for as much as it as we can and clear that amount, and then + * go searching for the next bit. */ search_start = *offset; - search_bytes = *bytes; + search_bytes = ctl->unit; search_bytes = min(search_bytes, end - search_start + 1); ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes); BUG_ON(ret < 0 || search_start != *offset); - if (*offset > bitmap_info->offset && *offset + *bytes > end) { - bitmap_clear_bits(ctl, bitmap_info, *offset, end - *offset + 1); - *bytes -= end - *offset + 1; - *offset = end + 1; - } else if (*offset >= bitmap_info->offset && *offset + *bytes <= end) { - bitmap_clear_bits(ctl, bitmap_info, *offset, *bytes); - *bytes = 0; - } + /* We may have found more bits than what we need */ + search_bytes = min(search_bytes, *bytes); + + /* Cannot clear past the end of the bitmap */ + search_bytes = min(search_bytes, end - search_start + 1); + + bitmap_clear_bits(ctl, bitmap_info, search_start, search_bytes); + *offset += search_bytes; + *bytes -= search_bytes; if (*bytes) { struct rb_node *next = rb_next(&bitmap_info->offset_index); @@ -1364,7 +1593,7 @@ again: * everything over again. */ search_start = *offset; - search_bytes = *bytes; + search_bytes = ctl->unit; ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes); if (ret < 0 || search_start != *offset) @@ -1647,12 +1876,14 @@ int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group, { struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *info; - struct btrfs_free_space *next_info = NULL; int ret = 0; spin_lock(&ctl->tree_lock); again: + if (!bytes) + goto out_lock; + info = tree_search_offset(ctl, offset, 0, 0); if (!info) { /* @@ -1662,97 +1893,65 @@ again: info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 1, 0); if (!info) { - WARN_ON(1); - goto out_lock; - } - } - - if (info->bytes < bytes && rb_next(&info->offset_index)) { - u64 end; - next_info = rb_entry(rb_next(&info->offset_index), - struct btrfs_free_space, - offset_index); - - if (next_info->bitmap) - end = next_info->offset + - BITS_PER_BITMAP * ctl->unit - 1; - else - end = next_info->offset + next_info->bytes; - - if (next_info->bytes < bytes || - next_info->offset > offset || offset > end) { - printk(KERN_CRIT "Found free space at %llu, size %llu," - " trying to use %llu\n", - (unsigned long long)info->offset, - (unsigned long long)info->bytes, - (unsigned long long)bytes); - WARN_ON(1); - ret = -EINVAL; + /* the tree logging code might be calling us before we + * have fully loaded the free space rbtree for this + * block group. So it is possible the entry won't + * be in the rbtree yet at all. The caching code + * will make sure not to put it in the rbtree if + * the logging code has pinned it. + */ goto out_lock; } - - info = next_info; - } - - if (info->bytes == bytes) { - unlink_free_space(ctl, info); - if (info->bitmap) { - kfree(info->bitmap); - ctl->total_bitmaps--; - } - kmem_cache_free(btrfs_free_space_cachep, info); - goto out_lock; } - if (!info->bitmap && info->offset == offset) { + if (!info->bitmap) { unlink_free_space(ctl, info); - info->offset += bytes; - info->bytes -= bytes; - link_free_space(ctl, info); - goto out_lock; - } + if (offset == info->offset) { + u64 to_free = min(bytes, info->bytes); + + info->bytes -= to_free; + info->offset += to_free; + if (info->bytes) { + ret = link_free_space(ctl, info); + WARN_ON(ret); + } else { + kmem_cache_free(btrfs_free_space_cachep, info); + } - if (!info->bitmap && info->offset <= offset && - info->offset + info->bytes >= offset + bytes) { - u64 old_start = info->offset; - /* - * we're freeing space in the middle of the info, - * this can happen during tree log replay - * - * first unlink the old info and then - * insert it again after the hole we're creating - */ - unlink_free_space(ctl, info); - if (offset + bytes < info->offset + info->bytes) { - u64 old_end = info->offset + info->bytes; + offset += to_free; + bytes -= to_free; + goto again; + } else { + u64 old_end = info->bytes + info->offset; - info->offset = offset + bytes; - info->bytes = old_end - info->offset; + info->bytes = offset - info->offset; ret = link_free_space(ctl, info); WARN_ON(ret); if (ret) goto out_lock; - } else { - /* the hole we're creating ends at the end - * of the info struct, just free the info - */ - kmem_cache_free(btrfs_free_space_cachep, info); - } - spin_unlock(&ctl->tree_lock); - /* step two, insert a new info struct to cover - * anything before the hole - */ - ret = btrfs_add_free_space(block_group, old_start, - offset - old_start); - WARN_ON(ret); - goto out; + /* Not enough bytes in this entry to satisfy us */ + if (old_end < offset + bytes) { + bytes -= old_end - offset; + offset = old_end; + goto again; + } else if (old_end == offset + bytes) { + /* all done */ + goto out_lock; + } + spin_unlock(&ctl->tree_lock); + + ret = btrfs_add_free_space(block_group, offset + bytes, + old_end - (offset + bytes)); + WARN_ON(ret); + goto out; + } } ret = remove_from_bitmap(ctl, info, &offset, &bytes); if (ret == -EAGAIN) goto again; - BUG_ON(ret); + BUG_ON(ret); /* logic error */ out_lock: spin_unlock(&ctl->tree_lock); out: @@ -1769,7 +1968,7 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) { info = rb_entry(n, struct btrfs_free_space, offset_index); - if (info->bytes >= bytes) + if (info->bytes >= bytes && !block_group->ro) count++; printk(KERN_CRIT "entry offset %llu, bytes %llu, bitmap %s\n", (unsigned long long)info->offset, @@ -2040,7 +2239,7 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, if (entry->bitmap) { ret = btrfs_alloc_from_bitmap(block_group, cluster, entry, bytes, - min_start); + cluster->window_start); if (ret == 0) { node = rb_next(&entry->offset_index); if (!node) @@ -2049,6 +2248,7 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, offset_index); continue; } + cluster->window_start += bytes; } else { ret = entry->offset; @@ -2087,23 +2287,23 @@ out: static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, struct btrfs_free_space *entry, struct btrfs_free_cluster *cluster, - u64 offset, u64 bytes, u64 min_bytes) + u64 offset, u64 bytes, + u64 cont1_bytes, u64 min_bytes) { struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; unsigned long next_zero; unsigned long i; - unsigned long search_bits; - unsigned long total_bits; + unsigned long want_bits; + unsigned long min_bits; unsigned long found_bits; unsigned long start = 0; unsigned long total_found = 0; int ret; - bool found = false; i = offset_to_bit(entry->offset, block_group->sectorsize, max_t(u64, offset, entry->offset)); - search_bits = bytes_to_bits(bytes, block_group->sectorsize); - total_bits = bytes_to_bits(min_bytes, block_group->sectorsize); + want_bits = bytes_to_bits(bytes, block_group->sectorsize); + min_bits = bytes_to_bits(min_bytes, block_group->sectorsize); again: found_bits = 0; @@ -2112,7 +2312,7 @@ again: i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, i + 1)) { next_zero = find_next_zero_bit(entry->bitmap, BITS_PER_BITMAP, i); - if (next_zero - i >= search_bits) { + if (next_zero - i >= min_bits) { found_bits = next_zero - i; break; } @@ -2122,9 +2322,9 @@ again: if (!found_bits) return -ENOSPC; - if (!found) { + if (!total_found) { start = i; - found = true; + cluster->max_size = 0; } total_found += found_bits; @@ -2132,13 +2332,8 @@ again: if (cluster->max_size < found_bits * block_group->sectorsize) cluster->max_size = found_bits * block_group->sectorsize; - if (total_found < total_bits) { - i = find_next_bit(entry->bitmap, BITS_PER_BITMAP, next_zero); - if (i - start > total_bits * 2) { - total_found = 0; - cluster->max_size = 0; - found = false; - } + if (total_found < want_bits || cluster->max_size < cont1_bytes) { + i = next_zero + 1; goto again; } @@ -2147,30 +2342,33 @@ again: rb_erase(&entry->offset_index, &ctl->free_space_offset); ret = tree_insert_offset(&cluster->root, entry->offset, &entry->offset_index, 1); - BUG_ON(ret); + BUG_ON(ret); /* -EEXIST; Logic error */ + trace_btrfs_setup_cluster(block_group, cluster, + total_found * block_group->sectorsize, 1); return 0; } /* * This searches the block group for just extents to fill the cluster with. + * Try to find a cluster with at least bytes total bytes, at least one + * extent of cont1_bytes, and other clusters of at least min_bytes. */ static noinline int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, struct list_head *bitmaps, u64 offset, u64 bytes, - u64 min_bytes) + u64 cont1_bytes, u64 min_bytes) { struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *first = NULL; struct btrfs_free_space *entry = NULL; - struct btrfs_free_space *prev = NULL; struct btrfs_free_space *last; struct rb_node *node; u64 window_start; u64 window_free; u64 max_extent; - u64 max_gap = 128 * 1024; + u64 total_size = 0; entry = tree_search_offset(ctl, offset, 0, 1); if (!entry) @@ -2180,8 +2378,8 @@ setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, * We don't want bitmaps, so just move along until we find a normal * extent entry. */ - while (entry->bitmap) { - if (list_empty(&entry->list)) + while (entry->bitmap || entry->bytes < min_bytes) { + if (entry->bitmap && list_empty(&entry->list)) list_add_tail(&entry->list, bitmaps); node = rb_next(&entry->offset_index); if (!node) @@ -2194,12 +2392,9 @@ setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, max_extent = entry->bytes; first = entry; last = entry; - prev = entry; - while (window_free <= min_bytes) { - node = rb_next(&entry->offset_index); - if (!node) - return -ENOSPC; + for (node = rb_next(&entry->offset_index); node; + node = rb_next(&entry->offset_index)) { entry = rb_entry(node, struct btrfs_free_space, offset_index); if (entry->bitmap) { @@ -2208,26 +2403,18 @@ setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, continue; } - /* - * we haven't filled the empty size and the window is - * very large. reset and try again - */ - if (entry->offset - (prev->offset + prev->bytes) > max_gap || - entry->offset - window_start > (min_bytes * 2)) { - first = entry; - window_start = entry->offset; - window_free = entry->bytes; - last = entry; + if (entry->bytes < min_bytes) + continue; + + last = entry; + window_free += entry->bytes; + if (entry->bytes > max_extent) max_extent = entry->bytes; - } else { - last = entry; - window_free += entry->bytes; - if (entry->bytes > max_extent) - max_extent = entry->bytes; - } - prev = entry; } + if (window_free < bytes || max_extent < cont1_bytes) + return -ENOSPC; + cluster->window_start = first->offset; node = &first->offset_index; @@ -2241,17 +2428,18 @@ setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, entry = rb_entry(node, struct btrfs_free_space, offset_index); node = rb_next(&entry->offset_index); - if (entry->bitmap) + if (entry->bitmap || entry->bytes < min_bytes) continue; rb_erase(&entry->offset_index, &ctl->free_space_offset); ret = tree_insert_offset(&cluster->root, entry->offset, &entry->offset_index, 0); - BUG_ON(ret); + total_size += entry->bytes; + BUG_ON(ret); /* -EEXIST; Logic error */ } while (node && entry != last); cluster->max_size = max_extent; - + trace_btrfs_setup_cluster(block_group, cluster, total_size, 0); return 0; } @@ -2263,67 +2451,46 @@ static noinline int setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, struct list_head *bitmaps, u64 offset, u64 bytes, - u64 min_bytes) + u64 cont1_bytes, u64 min_bytes) { struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry; - struct rb_node *node; int ret = -ENOSPC; + u64 bitmap_offset = offset_to_bitmap(ctl, offset); if (ctl->total_bitmaps == 0) return -ENOSPC; /* - * First check our cached list of bitmaps and see if there is an entry - * here that will work. + * The bitmap that covers offset won't be in the list unless offset + * is just its start offset. */ + entry = list_first_entry(bitmaps, struct btrfs_free_space, list); + if (entry->offset != bitmap_offset) { + entry = tree_search_offset(ctl, bitmap_offset, 1, 0); + if (entry && list_empty(&entry->list)) + list_add(&entry->list, bitmaps); + } + list_for_each_entry(entry, bitmaps, list) { - if (entry->bytes < min_bytes) + if (entry->bytes < bytes) continue; ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset, - bytes, min_bytes); + bytes, cont1_bytes, min_bytes); if (!ret) return 0; } /* - * If we do have entries on our list and we are here then we didn't find - * anything, so go ahead and get the next entry after the last entry in - * this list and start the search from there. + * The bitmaps list has all the bitmaps that record free space + * starting after offset, so no more search is required. */ - if (!list_empty(bitmaps)) { - entry = list_entry(bitmaps->prev, struct btrfs_free_space, - list); - node = rb_next(&entry->offset_index); - if (!node) - return -ENOSPC; - entry = rb_entry(node, struct btrfs_free_space, offset_index); - goto search; - } - - entry = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 0, 1); - if (!entry) - return -ENOSPC; - -search: - node = &entry->offset_index; - do { - entry = rb_entry(node, struct btrfs_free_space, offset_index); - node = rb_next(&entry->offset_index); - if (!entry->bitmap) - continue; - if (entry->bytes < min_bytes) - continue; - ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset, - bytes, min_bytes); - } while (ret && node); - - return ret; + return -ENOSPC; } /* * here we try to find a cluster of blocks in a block group. The goal - * is to find at least bytes free and up to empty_size + bytes free. + * is to find at least bytes+empty_size. * We might not find them all in one contiguous area. * * returns zero and sets up cluster if things worked out, otherwise @@ -2336,26 +2503,27 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, u64 offset, u64 bytes, u64 empty_size) { struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; - struct list_head bitmaps; struct btrfs_free_space *entry, *tmp; + LIST_HEAD(bitmaps); u64 min_bytes; + u64 cont1_bytes; int ret; - /* for metadata, allow allocates with more holes */ + /* + * Choose the minimum extent size we'll require for this + * cluster. For SSD_SPREAD, don't allow any fragmentation. + * For metadata, allow allocates with smaller extents. For + * data, keep it dense. + */ if (btrfs_test_opt(root, SSD_SPREAD)) { - min_bytes = bytes + empty_size; + cont1_bytes = min_bytes = bytes + empty_size; } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) { - /* - * we want to do larger allocations when we are - * flushing out the delayed refs, it helps prevent - * making more work as we go along. - */ - if (trans->transaction->delayed_refs.flushing) - min_bytes = max(bytes, (bytes + empty_size) >> 1); - else - min_bytes = max(bytes, (bytes + empty_size) >> 4); - } else - min_bytes = max(bytes, (bytes + empty_size) >> 2); + cont1_bytes = bytes; + min_bytes = block_group->sectorsize; + } else { + cont1_bytes = max(bytes, (bytes + empty_size) >> 2); + min_bytes = block_group->sectorsize; + } spin_lock(&ctl->tree_lock); @@ -2363,7 +2531,7 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, * If we know we don't have enough space to make a cluster don't even * bother doing all the work to try and find one. */ - if (ctl->free_space < min_bytes) { + if (ctl->free_space < bytes) { spin_unlock(&ctl->tree_lock); return -ENOSPC; } @@ -2376,12 +2544,17 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, goto out; } + trace_btrfs_find_cluster(block_group, offset, bytes, empty_size, + min_bytes); + INIT_LIST_HEAD(&bitmaps); ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset, - bytes, min_bytes); + bytes + empty_size, + cont1_bytes, min_bytes); if (ret) ret = setup_cluster_bitmap(block_group, cluster, &bitmaps, - offset, bytes, min_bytes); + offset, bytes + empty_size, + cont1_bytes, min_bytes); /* Clear our temporary list */ list_for_each_entry_safe(entry, tmp, &bitmaps, list) @@ -2392,6 +2565,8 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, list_add_tail(&cluster->block_group_list, &block_group->cluster_list); cluster->block_group = block_group; + } else { + trace_btrfs_failed_cluster_setup(block_group); } out: spin_unlock(&cluster->lock); @@ -2413,17 +2588,57 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster) cluster->block_group = NULL; } -int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, - u64 *trimmed, u64 start, u64 end, u64 minlen) +static int do_trimming(struct btrfs_block_group_cache *block_group, + u64 *total_trimmed, u64 start, u64 bytes, + u64 reserved_start, u64 reserved_bytes) { - struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; - struct btrfs_free_space *entry = NULL; + struct btrfs_space_info *space_info = block_group->space_info; struct btrfs_fs_info *fs_info = block_group->fs_info; - u64 bytes = 0; - u64 actually_trimmed; - int ret = 0; + int ret; + int update = 0; + u64 trimmed = 0; - *trimmed = 0; + spin_lock(&space_info->lock); + spin_lock(&block_group->lock); + if (!block_group->ro) { + block_group->reserved += reserved_bytes; + space_info->bytes_reserved += reserved_bytes; + update = 1; + } + spin_unlock(&block_group->lock); + spin_unlock(&space_info->lock); + + ret = btrfs_error_discard_extent(fs_info->extent_root, + start, bytes, &trimmed); + if (!ret) + *total_trimmed += trimmed; + + btrfs_add_free_space(block_group, reserved_start, reserved_bytes); + + if (update) { + spin_lock(&space_info->lock); + spin_lock(&block_group->lock); + if (block_group->ro) + space_info->bytes_readonly += reserved_bytes; + block_group->reserved -= reserved_bytes; + space_info->bytes_reserved -= reserved_bytes; + spin_unlock(&space_info->lock); + spin_unlock(&block_group->lock); + } + + return ret; +} + +static int trim_no_bitmap(struct btrfs_block_group_cache *block_group, + u64 *total_trimmed, u64 start, u64 end, u64 minlen) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct btrfs_free_space *entry; + struct rb_node *node; + int ret = 0; + u64 extent_start; + u64 extent_bytes; + u64 bytes; while (start < end) { spin_lock(&ctl->tree_lock); @@ -2434,64 +2649,118 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, } entry = tree_search_offset(ctl, start, 0, 1); - if (!entry) - entry = tree_search_offset(ctl, - offset_to_bitmap(ctl, start), - 1, 1); - - if (!entry || entry->offset >= end) { + if (!entry) { spin_unlock(&ctl->tree_lock); break; } - if (entry->bitmap) { - ret = search_bitmap(ctl, entry, &start, &bytes); - if (!ret) { - if (start >= end) { - spin_unlock(&ctl->tree_lock); - break; - } - bytes = min(bytes, end - start); - bitmap_clear_bits(ctl, entry, start, bytes); - if (entry->bytes == 0) - free_bitmap(ctl, entry); - } else { - start = entry->offset + BITS_PER_BITMAP * - block_group->sectorsize; + /* skip bitmaps */ + while (entry->bitmap) { + node = rb_next(&entry->offset_index); + if (!node) { spin_unlock(&ctl->tree_lock); - ret = 0; - continue; + goto out; } - } else { - start = entry->offset; - bytes = min(entry->bytes, end - start); - unlink_free_space(ctl, entry); - kmem_cache_free(btrfs_free_space_cachep, entry); + entry = rb_entry(node, struct btrfs_free_space, + offset_index); + } + + if (entry->offset >= end) { + spin_unlock(&ctl->tree_lock); + break; } + extent_start = entry->offset; + extent_bytes = entry->bytes; + start = max(start, extent_start); + bytes = min(extent_start + extent_bytes, end) - start; + if (bytes < minlen) { + spin_unlock(&ctl->tree_lock); + goto next; + } + + unlink_free_space(ctl, entry); + kmem_cache_free(btrfs_free_space_cachep, entry); + spin_unlock(&ctl->tree_lock); - if (bytes >= minlen) { - int update_ret; - update_ret = btrfs_update_reserved_bytes(block_group, - bytes, 1, 1); + ret = do_trimming(block_group, total_trimmed, start, bytes, + extent_start, extent_bytes); + if (ret) + break; +next: + start += bytes; + + if (fatal_signal_pending(current)) { + ret = -ERESTARTSYS; + break; + } + + cond_resched(); + } +out: + return ret; +} + +static int trim_bitmaps(struct btrfs_block_group_cache *block_group, + u64 *total_trimmed, u64 start, u64 end, u64 minlen) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct btrfs_free_space *entry; + int ret = 0; + int ret2; + u64 bytes; + u64 offset = offset_to_bitmap(ctl, start); - ret = btrfs_error_discard_extent(fs_info->extent_root, - start, - bytes, - &actually_trimmed); + while (offset < end) { + bool next_bitmap = false; - btrfs_add_free_space(block_group, start, bytes); - if (!update_ret) - btrfs_update_reserved_bytes(block_group, - bytes, 0, 1); + spin_lock(&ctl->tree_lock); - if (ret) - break; - *trimmed += actually_trimmed; + if (ctl->free_space < minlen) { + spin_unlock(&ctl->tree_lock); + break; + } + + entry = tree_search_offset(ctl, offset, 1, 0); + if (!entry) { + spin_unlock(&ctl->tree_lock); + next_bitmap = true; + goto next; + } + + bytes = minlen; + ret2 = search_bitmap(ctl, entry, &start, &bytes); + if (ret2 || start >= end) { + spin_unlock(&ctl->tree_lock); + next_bitmap = true; + goto next; + } + + bytes = min(bytes, end - start); + if (bytes < minlen) { + spin_unlock(&ctl->tree_lock); + goto next; + } + + bitmap_clear_bits(ctl, entry, start, bytes); + if (entry->bytes == 0) + free_bitmap(ctl, entry); + + spin_unlock(&ctl->tree_lock); + + ret = do_trimming(block_group, total_trimmed, start, bytes, + start, bytes); + if (ret) + break; +next: + if (next_bitmap) { + offset += BITS_PER_BITMAP * ctl->unit; + } else { + start += bytes; + if (start >= offset + BITS_PER_BITMAP * ctl->unit) + offset += BITS_PER_BITMAP * ctl->unit; } - start += bytes; - bytes = 0; if (fatal_signal_pending(current)) { ret = -ERESTARTSYS; @@ -2504,6 +2773,22 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, return ret; } +int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, + u64 *trimmed, u64 start, u64 end, u64 minlen) +{ + int ret; + + *trimmed = 0; + + ret = trim_no_bitmap(block_group, trimmed, start, end, minlen); + if (ret) + return ret; + + ret = trim_bitmaps(block_group, trimmed, start, end, minlen); + + return ret; +} + /* * Find the left-most item in the cache tree, and then return the * smallest inode number in the item. @@ -2541,6 +2826,7 @@ u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root) int ret; ret = search_bitmap(ctl, entry, &offset, &count); + /* Logic error; Should be empty if it can't find anything */ BUG_ON(ret); ino = offset; @@ -2643,9 +2929,13 @@ int btrfs_write_out_ino_cache(struct btrfs_root *root, return 0; ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0); - if (ret < 0) + if (ret) { + btrfs_delalloc_release_metadata(inode, inode->i_size); +#ifdef DEBUG printk(KERN_ERR "btrfs: failed to write free ino cache " "for root %llu\n", root->root_key.objectid); +#endif + } iput(inode); return ret; diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c index baa74f3db691..a13cf1a96c73 100644 --- a/fs/btrfs/inode-item.c +++ b/fs/btrfs/inode-item.c @@ -19,6 +19,7 @@ #include "ctree.h" #include "disk-io.h" #include "transaction.h" +#include "print-tree.h" static int find_name_in_backref(struct btrfs_path *path, const char *name, int name_len, struct btrfs_inode_ref **ref_ret) @@ -128,13 +129,14 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans, item_start = btrfs_item_ptr_offset(leaf, path->slots[0]); memmove_extent_buffer(leaf, ptr, ptr + sub_item_len, item_size - (ptr + sub_item_len - item_start)); - ret = btrfs_truncate_item(trans, root, path, + btrfs_truncate_item(trans, root, path, item_size - sub_item_len, 1); out: btrfs_free_path(path); return ret; } +/* Will return 0, -ENOMEM, -EMLINK, or -EEXIST or anything from the CoW path */ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, int name_len, @@ -165,7 +167,7 @@ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans, goto out; old_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]); - ret = btrfs_extend_item(trans, root, path, ins_len); + btrfs_extend_item(trans, root, path, ins_len); ref = btrfs_item_ptr(path->nodes[0], path->slots[0], struct btrfs_inode_ref); ref = (struct btrfs_inode_ref *)((unsigned long)ref + old_size); diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c index b4087e0fa871..b1a1c929ba80 100644 --- a/fs/btrfs/inode-map.c +++ b/fs/btrfs/inode-map.c @@ -178,7 +178,7 @@ static void start_caching(struct btrfs_root *root) tsk = kthread_run(caching_kthread, root, "btrfs-ino-cache-%llu\n", root->root_key.objectid); - BUG_ON(IS_ERR(tsk)); + BUG_ON(IS_ERR(tsk)); /* -ENOMEM */ } int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid) @@ -271,7 +271,7 @@ void btrfs_unpin_free_ino(struct btrfs_root *root) break; info = rb_entry(n, struct btrfs_free_space, offset_index); - BUG_ON(info->bitmap); + BUG_ON(info->bitmap); /* Logic error */ if (info->offset > root->cache_progress) goto free; @@ -398,6 +398,8 @@ int btrfs_save_ino_cache(struct btrfs_root *root, struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; struct btrfs_path *path; struct inode *inode; + struct btrfs_block_rsv *rsv; + u64 num_bytes; u64 alloc_hint = 0; int ret; int prealloc; @@ -421,31 +423,53 @@ int btrfs_save_ino_cache(struct btrfs_root *root, if (!path) return -ENOMEM; + rsv = trans->block_rsv; + trans->block_rsv = &root->fs_info->trans_block_rsv; + + num_bytes = trans->bytes_reserved; + /* + * 1 item for inode item insertion if need + * 3 items for inode item update (in the worst case) + * 1 item for free space object + * 3 items for pre-allocation + */ + trans->bytes_reserved = btrfs_calc_trans_metadata_size(root, 8); + ret = btrfs_block_rsv_add_noflush(root, trans->block_rsv, + trans->bytes_reserved); + if (ret) + goto out; + trace_btrfs_space_reservation(root->fs_info, "ino_cache", + trans->transid, trans->bytes_reserved, 1); again: inode = lookup_free_ino_inode(root, path); - if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { + if (IS_ERR(inode) && (PTR_ERR(inode) != -ENOENT || retry)) { ret = PTR_ERR(inode); - goto out; + goto out_release; } if (IS_ERR(inode)) { - BUG_ON(retry); + BUG_ON(retry); /* Logic error */ retry = true; ret = create_free_ino_inode(root, trans, path); if (ret) - goto out; + goto out_release; goto again; } BTRFS_I(inode)->generation = 0; ret = btrfs_update_inode(trans, root, inode); - WARN_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_put; + } if (i_size_read(inode) > 0) { ret = btrfs_truncate_free_space_cache(root, trans, path, inode); - if (ret) + if (ret) { + btrfs_abort_transaction(trans, root, ret); goto out_put; + } } spin_lock(&root->cache_lock); @@ -465,21 +489,28 @@ again: /* Just to make sure we have enough space */ prealloc += 8 * PAGE_CACHE_SIZE; - ret = btrfs_check_data_free_space(inode, prealloc); + ret = btrfs_delalloc_reserve_space(inode, prealloc); if (ret) goto out_put; ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc, prealloc, prealloc, &alloc_hint); - if (ret) + if (ret) { + btrfs_delalloc_release_space(inode, prealloc); goto out_put; + } btrfs_free_reserved_data_space(inode, prealloc); + ret = btrfs_write_out_ino_cache(root, trans, path); out_put: iput(inode); +out_release: + trace_btrfs_space_reservation(root->fs_info, "ino_cache", + trans->transid, trans->bytes_reserved, 0); + btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved); out: - if (ret == 0) - ret = btrfs_write_out_ino_cache(root, trans, path); + trans->block_rsv = rsv; + trans->bytes_reserved = num_bytes; btrfs_free_path(path); return ret; @@ -504,7 +535,7 @@ static int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid) ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0); if (ret < 0) goto error; - BUG_ON(ret == 0); + BUG_ON(ret == 0); /* Corruption */ if (path->slots[0] > 0) { slot = path->slots[0] - 1; l = path->nodes[0]; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index b2d004ad66a0..ec154f954646 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -38,6 +38,7 @@ #include <linux/falloc.h> #include <linux/slab.h> #include <linux/ratelimit.h> +#include <linux/mount.h> #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -45,10 +46,10 @@ #include "btrfs_inode.h" #include "ioctl.h" #include "print-tree.h" -#include "volumes.h" #include "ordered-data.h" #include "xattr.h" #include "tree-log.h" +#include "volumes.h" #include "compression.h" #include "locking.h" #include "free-space-cache.h" @@ -88,11 +89,13 @@ static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { static int btrfs_setsize(struct inode *inode, loff_t newsize); static int btrfs_truncate(struct inode *inode); -static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end); +static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent); static noinline int cow_file_range(struct inode *inode, struct page *locked_page, u64 start, u64 end, int *page_started, unsigned long *nr_written, int unlock); +static noinline int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode); static int btrfs_init_inode_security(struct btrfs_trans_handle *trans, struct inode *inode, struct inode *dir, @@ -147,7 +150,6 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, inode_add_bytes(inode, size); ret = btrfs_insert_empty_item(trans, root, path, &key, datasize); - BUG_ON(ret); if (ret) { err = ret; goto fail; @@ -170,9 +172,9 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, cur_size = min_t(unsigned long, compressed_size, PAGE_CACHE_SIZE); - kaddr = kmap_atomic(cpage, KM_USER0); + kaddr = kmap_atomic(cpage); write_extent_buffer(leaf, kaddr, ptr, cur_size); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); i++; ptr += cur_size; @@ -184,10 +186,10 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, page = find_get_page(inode->i_mapping, start >> PAGE_CACHE_SHIFT); btrfs_set_file_extent_compression(leaf, ei, 0); - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); offset = start & (PAGE_CACHE_SIZE - 1); write_extent_buffer(leaf, kaddr + offset, ptr, size); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); page_cache_release(page); } btrfs_mark_buffer_dirty(leaf); @@ -203,9 +205,9 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, * could end up racing with unlink. */ BTRFS_I(inode)->disk_i_size = inode->i_size; - btrfs_update_inode(trans, root, inode); + ret = btrfs_update_inode(trans, root, inode); - return 0; + return ret; fail: btrfs_free_path(path); return err; @@ -247,14 +249,21 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans, ret = btrfs_drop_extents(trans, inode, start, aligned_end, &hint_byte, 1); - BUG_ON(ret); + if (ret) + return ret; if (isize > actual_end) inline_len = min_t(u64, isize, actual_end); ret = insert_inline_extent(trans, root, inode, start, inline_len, compressed_size, compress_type, compressed_pages); - BUG_ON(ret); + if (ret && ret != -ENOSPC) { + btrfs_abort_transaction(trans, root, ret); + return ret; + } else if (ret == -ENOSPC) { + return 1; + } + btrfs_delalloc_release_metadata(inode, end + 1 - start); btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0); return 0; @@ -290,7 +299,7 @@ static noinline int add_async_extent(struct async_cow *cow, struct async_extent *async_extent; async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS); - BUG_ON(!async_extent); + BUG_ON(!async_extent); /* -ENOMEM */ async_extent->start = start; async_extent->ram_size = ram_size; async_extent->compressed_size = compressed_size; @@ -315,7 +324,8 @@ static noinline int add_async_extent(struct async_cow *cow, * If this code finds it can't get good compression, it puts an * entry onto the work queue to write the uncompressed bytes. This * makes sure that both compressed inodes and uncompressed inodes - * are written in the same order that pdflush sent them down. + * are written in the same order that the flusher thread sent them + * down. */ static noinline int compress_file_range(struct inode *inode, struct page *locked_page, @@ -341,8 +351,9 @@ static noinline int compress_file_range(struct inode *inode, int will_compress; int compress_type = root->fs_info->compress_type; - /* if this is a small write inside eof, kick off a defragbot */ - if (end <= BTRFS_I(inode)->disk_i_size && (end - start + 1) < 16 * 1024) + /* if this is a small write inside eof, kick off a defrag */ + if ((end - start + 1) < 16 * 1024 && + (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) btrfs_add_inode_defrag(NULL, inode); actual_end = min_t(u64, isize, end + 1); @@ -393,7 +404,10 @@ again: (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) { WARN_ON(pages); pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); - BUG_ON(!pages); + if (!pages) { + /* just bail out to the uncompressed code */ + goto cont; + } if (BTRFS_I(inode)->force_compress) compress_type = BTRFS_I(inode)->force_compress; @@ -416,17 +430,22 @@ again: * sending it down to disk */ if (offset) { - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); memset(kaddr + offset, 0, PAGE_CACHE_SIZE - offset); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); } will_compress = 1; } } +cont: if (start == 0) { trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto cleanup_and_out; + } trans->block_rsv = &root->fs_info->delalloc_block_rsv; /* lets try to make an inline extent */ @@ -443,11 +462,11 @@ again: total_compressed, compress_type, pages); } - if (ret == 0) { + if (ret <= 0) { /* - * inline extent creation worked, we don't need - * to create any more async work items. Unlock - * and free up our temp pages. + * inline extent creation worked or returned error, + * we don't need to create any more async work items. + * Unlock and free up our temp pages. */ extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree, @@ -540,7 +559,7 @@ cleanup_and_bail_uncompressed: } out: - return 0; + return ret; free_pages_out: for (i = 0; i < nr_pages_ret; i++) { @@ -550,6 +569,20 @@ free_pages_out: kfree(pages); goto out; + +cleanup_and_out: + extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree, + start, end, NULL, + EXTENT_CLEAR_UNLOCK_PAGE | + EXTENT_CLEAR_DIRTY | + EXTENT_CLEAR_DELALLOC | + EXTENT_SET_WRITEBACK | + EXTENT_END_WRITEBACK); + if (!trans || IS_ERR(trans)) + btrfs_error(root->fs_info, ret, "Failed to join transaction"); + else + btrfs_abort_transaction(trans, root, ret); + goto free_pages_out; } /* @@ -590,7 +623,7 @@ retry: lock_extent(io_tree, async_extent->start, async_extent->start + - async_extent->ram_size - 1, GFP_NOFS); + async_extent->ram_size - 1); /* allocate blocks */ ret = cow_file_range(inode, async_cow->locked_page, @@ -599,6 +632,8 @@ retry: async_extent->ram_size - 1, &page_started, &nr_written, 0); + /* JDM XXX */ + /* * if page_started, cow_file_range inserted an * inline extent and took care of all the unlocking @@ -618,18 +653,21 @@ retry: } lock_extent(io_tree, async_extent->start, - async_extent->start + async_extent->ram_size - 1, - GFP_NOFS); + async_extent->start + async_extent->ram_size - 1); trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); - trans->block_rsv = &root->fs_info->delalloc_block_rsv; - ret = btrfs_reserve_extent(trans, root, + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + } else { + trans->block_rsv = &root->fs_info->delalloc_block_rsv; + ret = btrfs_reserve_extent(trans, root, async_extent->compressed_size, async_extent->compressed_size, - 0, alloc_hint, - (u64)-1, &ins, 1); - btrfs_end_transaction(trans, root); + 0, alloc_hint, &ins, 1); + if (ret) + btrfs_abort_transaction(trans, root, ret); + btrfs_end_transaction(trans, root); + } if (ret) { int i; @@ -642,8 +680,10 @@ retry: async_extent->pages = NULL; unlock_extent(io_tree, async_extent->start, async_extent->start + - async_extent->ram_size - 1, GFP_NOFS); - goto retry; + async_extent->ram_size - 1); + if (ret == -ENOSPC) + goto retry; + goto out_free; /* JDM: Requeue? */ } /* @@ -655,7 +695,7 @@ retry: async_extent->ram_size - 1, 0); em = alloc_extent_map(); - BUG_ON(!em); + BUG_ON(!em); /* -ENOMEM */ em->start = async_extent->start; em->len = async_extent->ram_size; em->orig_start = em->start; @@ -687,7 +727,7 @@ retry: ins.offset, BTRFS_ORDERED_COMPRESSED, async_extent->compress_type); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ /* * clear dirty, set writeback and unlock the pages. @@ -709,13 +749,17 @@ retry: ins.offset, async_extent->pages, async_extent->nr_pages); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ alloc_hint = ins.objectid + ins.offset; kfree(async_extent); cond_resched(); } - - return 0; + ret = 0; +out: + return ret; +out_free: + kfree(async_extent); + goto out; } static u64 get_extent_allocation_hint(struct inode *inode, u64 start, @@ -782,9 +826,20 @@ static noinline int cow_file_range(struct inode *inode, struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; int ret = 0; - BUG_ON(btrfs_is_free_space_inode(root, inode)); + BUG_ON(btrfs_is_free_space_inode(inode)); trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + extent_clear_unlock_delalloc(inode, + &BTRFS_I(inode)->io_tree, + start, end, locked_page, + EXTENT_CLEAR_UNLOCK_PAGE | + EXTENT_CLEAR_UNLOCK | + EXTENT_CLEAR_DELALLOC | + EXTENT_CLEAR_DIRTY | + EXTENT_SET_WRITEBACK | + EXTENT_END_WRITEBACK); + return PTR_ERR(trans); + } trans->block_rsv = &root->fs_info->delalloc_block_rsv; num_bytes = (end - start + blocksize) & ~(blocksize - 1); @@ -793,7 +848,8 @@ static noinline int cow_file_range(struct inode *inode, ret = 0; /* if this is a small write inside eof, kick off defrag */ - if (end <= BTRFS_I(inode)->disk_i_size && num_bytes < 64 * 1024) + if (num_bytes < 64 * 1024 && + (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) btrfs_add_inode_defrag(trans, inode); if (start == 0) { @@ -814,13 +870,15 @@ static noinline int cow_file_range(struct inode *inode, *nr_written = *nr_written + (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE; *page_started = 1; - ret = 0; goto out; + } else if (ret < 0) { + btrfs_abort_transaction(trans, root, ret); + goto out_unlock; } } BUG_ON(disk_num_bytes > - btrfs_super_total_bytes(&root->fs_info->super_copy)); + btrfs_super_total_bytes(root->fs_info->super_copy)); alloc_hint = get_extent_allocation_hint(inode, start, num_bytes); btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0); @@ -831,11 +889,14 @@ static noinline int cow_file_range(struct inode *inode, cur_alloc_size = disk_num_bytes; ret = btrfs_reserve_extent(trans, root, cur_alloc_size, root->sectorsize, 0, alloc_hint, - (u64)-1, &ins, 1); - BUG_ON(ret); + &ins, 1); + if (ret < 0) { + btrfs_abort_transaction(trans, root, ret); + goto out_unlock; + } em = alloc_extent_map(); - BUG_ON(!em); + BUG_ON(!em); /* -ENOMEM */ em->start = start; em->orig_start = em->start; ram_size = ins.offset; @@ -861,13 +922,16 @@ static noinline int cow_file_range(struct inode *inode, cur_alloc_size = ins.offset; ret = btrfs_add_ordered_extent(inode, start, ins.objectid, ram_size, cur_alloc_size, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID) { ret = btrfs_reloc_clone_csums(inode, start, cur_alloc_size); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_unlock; + } } if (disk_num_bytes < cur_alloc_size) @@ -892,11 +956,23 @@ static noinline int cow_file_range(struct inode *inode, alloc_hint = ins.objectid + ins.offset; start += cur_alloc_size; } -out: ret = 0; +out: btrfs_end_transaction(trans, root); return ret; +out_unlock: + extent_clear_unlock_delalloc(inode, + &BTRFS_I(inode)->io_tree, + start, end, locked_page, + EXTENT_CLEAR_UNLOCK_PAGE | + EXTENT_CLEAR_UNLOCK | + EXTENT_CLEAR_DELALLOC | + EXTENT_CLEAR_DIRTY | + EXTENT_SET_WRITEBACK | + EXTENT_END_WRITEBACK); + + goto out; } /* @@ -911,8 +987,10 @@ static noinline void async_cow_start(struct btrfs_work *work) compress_file_range(async_cow->inode, async_cow->locked_page, async_cow->start, async_cow->end, async_cow, &num_added); - if (num_added == 0) + if (num_added == 0) { + btrfs_add_delayed_iput(async_cow->inode); async_cow->inode = NULL; + } } /* @@ -930,10 +1008,8 @@ static noinline void async_cow_submit(struct btrfs_work *work) nr_pages = (async_cow->end - async_cow->start + PAGE_CACHE_SIZE) >> PAGE_CACHE_SHIFT; - atomic_sub(nr_pages, &root->fs_info->async_delalloc_pages); - - if (atomic_read(&root->fs_info->async_delalloc_pages) < - 5 * 1042 * 1024 && + if (atomic_sub_return(nr_pages, &root->fs_info->async_delalloc_pages) < + 5 * 1024 * 1024 && waitqueue_active(&root->fs_info->async_submit_wait)) wake_up(&root->fs_info->async_submit_wait); @@ -945,6 +1021,8 @@ static noinline void async_cow_free(struct btrfs_work *work) { struct async_cow *async_cow; async_cow = container_of(work, struct async_cow, work); + if (async_cow->inode) + btrfs_add_delayed_iput(async_cow->inode); kfree(async_cow); } @@ -956,14 +1034,14 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page, struct btrfs_root *root = BTRFS_I(inode)->root; unsigned long nr_pages; u64 cur_end; - int limit = 10 * 1024 * 1042; + int limit = 10 * 1024 * 1024; clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED, 1, 0, NULL, GFP_NOFS); while (start < end) { async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS); - BUG_ON(!async_cow); - async_cow->inode = inode; + BUG_ON(!async_cow); /* -ENOMEM */ + async_cow->inode = igrab(inode); async_cow->root = root; async_cow->locked_page = locked_page; async_cow->start = start; @@ -1053,7 +1131,7 @@ static noinline int run_delalloc_nocow(struct inode *inode, u64 disk_bytenr; u64 num_bytes; int extent_type; - int ret; + int ret, err; int type; int nocow; int check_prev = 1; @@ -1061,17 +1139,40 @@ static noinline int run_delalloc_nocow(struct inode *inode, u64 ino = btrfs_ino(inode); path = btrfs_alloc_path(); - if (!path) + if (!path) { + extent_clear_unlock_delalloc(inode, + &BTRFS_I(inode)->io_tree, + start, end, locked_page, + EXTENT_CLEAR_UNLOCK_PAGE | + EXTENT_CLEAR_UNLOCK | + EXTENT_CLEAR_DELALLOC | + EXTENT_CLEAR_DIRTY | + EXTENT_SET_WRITEBACK | + EXTENT_END_WRITEBACK); return -ENOMEM; + } - nolock = btrfs_is_free_space_inode(root, inode); + nolock = btrfs_is_free_space_inode(inode); if (nolock) trans = btrfs_join_transaction_nolock(root); else trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + extent_clear_unlock_delalloc(inode, + &BTRFS_I(inode)->io_tree, + start, end, locked_page, + EXTENT_CLEAR_UNLOCK_PAGE | + EXTENT_CLEAR_UNLOCK | + EXTENT_CLEAR_DELALLOC | + EXTENT_CLEAR_DIRTY | + EXTENT_SET_WRITEBACK | + EXTENT_END_WRITEBACK); + btrfs_free_path(path); + return PTR_ERR(trans); + } + trans->block_rsv = &root->fs_info->delalloc_block_rsv; cow_start = (u64)-1; @@ -1079,7 +1180,10 @@ static noinline int run_delalloc_nocow(struct inode *inode, while (1) { ret = btrfs_lookup_file_extent(trans, root, path, ino, cur_offset, 0); - BUG_ON(ret < 0); + if (ret < 0) { + btrfs_abort_transaction(trans, root, ret); + goto error; + } if (ret > 0 && path->slots[0] > 0 && check_prev) { leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, @@ -1093,8 +1197,10 @@ next_slot: leaf = path->nodes[0]; if (path->slots[0] >= btrfs_header_nritems(leaf)) { ret = btrfs_next_leaf(root, path); - if (ret < 0) - BUG_ON(1); + if (ret < 0) { + btrfs_abort_transaction(trans, root, ret); + goto error; + } if (ret > 0) break; leaf = path->nodes[0]; @@ -1182,7 +1288,10 @@ out_check: ret = cow_file_range(inode, locked_page, cow_start, found_key.offset - 1, page_started, nr_written, 1); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto error; + } cow_start = (u64)-1; } @@ -1191,7 +1300,7 @@ out_check: struct extent_map_tree *em_tree; em_tree = &BTRFS_I(inode)->extent_tree; em = alloc_extent_map(); - BUG_ON(!em); + BUG_ON(!em); /* -ENOMEM */ em->start = cur_offset; em->orig_start = em->start; em->len = num_bytes; @@ -1217,13 +1326,16 @@ out_check: ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr, num_bytes, num_bytes, type); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID) { ret = btrfs_reloc_clone_csums(inode, cur_offset, num_bytes); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto error; + } } extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree, @@ -1237,23 +1349,42 @@ out_check: } btrfs_release_path(path); - if (cur_offset <= end && cow_start == (u64)-1) + if (cur_offset <= end && cow_start == (u64)-1) { cow_start = cur_offset; + cur_offset = end; + } + if (cow_start != (u64)-1) { ret = cow_file_range(inode, locked_page, cow_start, end, page_started, nr_written, 1); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto error; + } } +error: if (nolock) { - ret = btrfs_end_transaction_nolock(trans, root); - BUG_ON(ret); + err = btrfs_end_transaction_nolock(trans, root); } else { - ret = btrfs_end_transaction(trans, root); - BUG_ON(ret); + err = btrfs_end_transaction(trans, root); } + if (!ret) + ret = err; + + if (ret && cur_offset < end) + extent_clear_unlock_delalloc(inode, + &BTRFS_I(inode)->io_tree, + cur_offset, end, locked_page, + EXTENT_CLEAR_UNLOCK_PAGE | + EXTENT_CLEAR_UNLOCK | + EXTENT_CLEAR_DELALLOC | + EXTENT_CLEAR_DIRTY | + EXTENT_SET_WRITEBACK | + EXTENT_END_WRITEBACK); + btrfs_free_path(path); - return 0; + return ret; } /* @@ -1266,20 +1397,23 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page, int ret; struct btrfs_root *root = BTRFS_I(inode)->root; - if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) + if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) { ret = run_delalloc_nocow(inode, locked_page, start, end, page_started, 1, nr_written); - else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC) + } else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC) { ret = run_delalloc_nocow(inode, locked_page, start, end, page_started, 0, nr_written); - else if (!btrfs_test_opt(root, COMPRESS) && - !(BTRFS_I(inode)->force_compress) && - !(BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS)) + } else if (!btrfs_test_opt(root, COMPRESS) && + !(BTRFS_I(inode)->force_compress) && + !(BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS)) { ret = cow_file_range(inode, locked_page, start, end, page_started, nr_written, 1); - else + } else { + set_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, + &BTRFS_I(inode)->runtime_flags); ret = cow_file_range_async(inode, locked_page, start, end, page_started, nr_written); + } return ret; } @@ -1331,7 +1465,7 @@ static void btrfs_set_bit_hook(struct inode *inode, if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - bool do_list = !btrfs_is_free_space_inode(root, inode); + bool do_list = !btrfs_is_free_space_inode(inode); if (*bits & EXTENT_FIRST_DELALLOC) { *bits &= ~EXTENT_FIRST_DELALLOC; @@ -1366,7 +1500,7 @@ static void btrfs_clear_bit_hook(struct inode *inode, if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - bool do_list = !btrfs_is_free_space_inode(root, inode); + bool do_list = !btrfs_is_free_space_inode(inode); if (*bits & EXTENT_FIRST_DELALLOC) { *bits &= ~EXTENT_FIRST_DELALLOC; @@ -1418,10 +1552,11 @@ int btrfs_merge_bio_hook(struct page *page, unsigned long offset, map_length = length; ret = btrfs_map_block(map_tree, READ, logical, &map_length, NULL, 0); - + /* Will always return 0 or 1 with map_multi == NULL */ + BUG_ON(ret < 0); if (map_length < length + size) return 1; - return ret; + return 0; } /* @@ -1441,7 +1576,7 @@ static int __btrfs_submit_bio_start(struct inode *inode, int rw, int ret = 0; ret = btrfs_csum_one_bio(root, inode, bio, 0, 0); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ return 0; } @@ -1472,16 +1607,18 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, struct btrfs_root *root = BTRFS_I(inode)->root; int ret = 0; int skip_sum; + int metadata = 0; skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; - if (btrfs_is_free_space_inode(root, inode)) - ret = btrfs_bio_wq_end_io(root->fs_info, bio, 2); - else - ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); - BUG_ON(ret); + if (btrfs_is_free_space_inode(inode)) + metadata = 2; if (!(rw & REQ_WRITE)) { + ret = btrfs_bio_wq_end_io(root->fs_info, bio, metadata); + if (ret) + return ret; + if (bio_flags & EXTENT_BIO_COMPRESSED) { return btrfs_submit_compressed_read(inode, bio, mirror_num, bio_flags); @@ -1548,6 +1685,7 @@ static void btrfs_writepage_fixup_worker(struct btrfs_work *work) struct inode *inode; u64 page_start; u64 page_end; + int ret; fixup = container_of(work, struct btrfs_writepage_fixup, work); page = fixup->page; @@ -1563,7 +1701,7 @@ again: page_end = page_offset(page) + PAGE_CACHE_SIZE - 1; lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end, 0, - &cached_state, GFP_NOFS); + &cached_state); /* already ordered? We're done */ if (PagePrivate2(page)) @@ -1575,12 +1713,21 @@ again: page_end, &cached_state, GFP_NOFS); unlock_page(page); btrfs_start_ordered_extent(inode, ordered, 1); + btrfs_put_ordered_extent(ordered); goto again; } - BUG(); + ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); + if (ret) { + mapping_set_error(page->mapping, ret); + end_extent_writepage(page, ret, page_start, page_end); + ClearPageChecked(page); + goto out; + } + btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state); ClearPageChecked(page); + set_page_dirty(page); out: unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end, &cached_state, GFP_NOFS); @@ -1623,7 +1770,7 @@ static int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end) fixup->work.func = btrfs_writepage_fixup_worker; fixup->page = page; btrfs_queue_worker(&root->fs_info->fixup_workers, &fixup->work); - return -EAGAIN; + return -EBUSY; } static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, @@ -1658,13 +1805,15 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, */ ret = btrfs_drop_extents(trans, inode, file_pos, file_pos + num_bytes, &hint, 0); - BUG_ON(ret); + if (ret) + goto out; ins.objectid = btrfs_ino(inode); ins.offset = file_pos; ins.type = BTRFS_EXTENT_DATA_KEY; ret = btrfs_insert_empty_item(trans, root, path, &ins, sizeof(*fi)); - BUG_ON(ret); + if (ret) + goto out; leaf = path->nodes[0]; fi = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item); @@ -1692,10 +1841,10 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, ret = btrfs_alloc_reserved_file_extent(trans, root, root->root_key.objectid, btrfs_ino(inode), file_pos, &ins); - BUG_ON(ret); +out: btrfs_free_path(path); - return 0; + return ret; } /* @@ -1708,50 +1857,58 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, * an ordered extent if the range of bytes in the file it covers are * fully written. */ -static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) +static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent) { + struct inode *inode = ordered_extent->inode; struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans = NULL; - struct btrfs_ordered_extent *ordered_extent = NULL; struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; struct extent_state *cached_state = NULL; int compress_type = 0; int ret; bool nolock; - ret = btrfs_dec_test_ordered_pending(inode, &ordered_extent, start, - end - start + 1); - if (!ret) - return 0; - BUG_ON(!ordered_extent); + nolock = btrfs_is_free_space_inode(inode); - nolock = btrfs_is_free_space_inode(root, inode); + if (test_bit(BTRFS_ORDERED_IOERR, &ordered_extent->flags)) { + ret = -EIO; + goto out; + } if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) { - BUG_ON(!list_empty(&ordered_extent->list)); + BUG_ON(!list_empty(&ordered_extent->list)); /* Logic error */ ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent); if (!ret) { if (nolock) trans = btrfs_join_transaction_nolock(root); else trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto out; + } trans->block_rsv = &root->fs_info->delalloc_block_rsv; - ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + ret = btrfs_update_inode_fallback(trans, root, inode); + if (ret) /* -ENOMEM or corruption */ + btrfs_abort_transaction(trans, root, ret); } goto out; } lock_extent_bits(io_tree, ordered_extent->file_offset, ordered_extent->file_offset + ordered_extent->len - 1, - 0, &cached_state, GFP_NOFS); + 0, &cached_state); if (nolock) trans = btrfs_join_transaction_nolock(root); else trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto out_unlock; + } trans->block_rsv = &root->fs_info->delalloc_block_rsv; if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags)) @@ -1762,7 +1919,6 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) ordered_extent->file_offset, ordered_extent->file_offset + ordered_extent->len); - BUG_ON(ret); } else { BUG_ON(root == root->fs_info->tree_root); ret = insert_reserved_file_extent(trans, inode, @@ -1776,199 +1932,99 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) unpin_extent_cache(&BTRFS_I(inode)->extent_tree, ordered_extent->file_offset, ordered_extent->len); - BUG_ON(ret); } - unlock_extent_cached(io_tree, ordered_extent->file_offset, - ordered_extent->file_offset + - ordered_extent->len - 1, &cached_state, GFP_NOFS); + + if (ret < 0) { + btrfs_abort_transaction(trans, root, ret); + goto out_unlock; + } add_pending_csums(trans, inode, ordered_extent->file_offset, &ordered_extent->list); ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent); if (!ret || !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) { - ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + ret = btrfs_update_inode_fallback(trans, root, inode); + if (ret) { /* -ENOMEM or corruption */ + btrfs_abort_transaction(trans, root, ret); + goto out_unlock; + } } ret = 0; +out_unlock: + unlock_extent_cached(io_tree, ordered_extent->file_offset, + ordered_extent->file_offset + + ordered_extent->len - 1, &cached_state, GFP_NOFS); out: - if (nolock) { - if (trans) - btrfs_end_transaction_nolock(trans, root); - } else { + if (root != root->fs_info->tree_root) btrfs_delalloc_release_metadata(inode, ordered_extent->len); - if (trans) + if (trans) { + if (nolock) + btrfs_end_transaction_nolock(trans, root); + else btrfs_end_transaction(trans, root); } + if (ret) + clear_extent_uptodate(io_tree, ordered_extent->file_offset, + ordered_extent->file_offset + + ordered_extent->len - 1, NULL, GFP_NOFS); + + /* + * This needs to be dont to make sure anybody waiting knows we are done + * upating everything for this ordered extent. + */ + btrfs_remove_ordered_extent(inode, ordered_extent); + /* once for us */ btrfs_put_ordered_extent(ordered_extent); /* once for the tree */ btrfs_put_ordered_extent(ordered_extent); - return 0; + return ret; } -static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, - struct extent_state *state, int uptodate) +static void finish_ordered_fn(struct btrfs_work *work) { - trace_btrfs_writepage_end_io_hook(page, start, end, uptodate); - - ClearPagePrivate2(page); - return btrfs_finish_ordered_io(page->mapping->host, start, end); + struct btrfs_ordered_extent *ordered_extent; + ordered_extent = container_of(work, struct btrfs_ordered_extent, work); + btrfs_finish_ordered_io(ordered_extent); } -/* - * When IO fails, either with EIO or csum verification fails, we - * try other mirrors that might have a good copy of the data. This - * io_failure_record is used to record state as we go through all the - * mirrors. If another mirror has good data, the page is set up to date - * and things continue. If a good mirror can't be found, the original - * bio end_io callback is called to indicate things have failed. - */ -struct io_failure_record { - struct page *page; - u64 start; - u64 len; - u64 logical; - unsigned long bio_flags; - int last_mirror; -}; - -static int btrfs_io_failed_hook(struct bio *failed_bio, - struct page *page, u64 start, u64 end, - struct extent_state *state) +static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, + struct extent_state *state, int uptodate) { - struct io_failure_record *failrec = NULL; - u64 private; - struct extent_map *em; struct inode *inode = page->mapping->host; - struct extent_io_tree *failure_tree = &BTRFS_I(inode)->io_failure_tree; - struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; - struct bio *bio; - int num_copies; - int ret; - int rw; - u64 logical; - - ret = get_state_private(failure_tree, start, &private); - if (ret) { - failrec = kmalloc(sizeof(*failrec), GFP_NOFS); - if (!failrec) - return -ENOMEM; - failrec->start = start; - failrec->len = end - start + 1; - failrec->last_mirror = 0; - failrec->bio_flags = 0; - - read_lock(&em_tree->lock); - em = lookup_extent_mapping(em_tree, start, failrec->len); - if (em->start > start || em->start + em->len < start) { - free_extent_map(em); - em = NULL; - } - read_unlock(&em_tree->lock); + struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_ordered_extent *ordered_extent = NULL; + struct btrfs_workers *workers; - if (IS_ERR_OR_NULL(em)) { - kfree(failrec); - return -EIO; - } - logical = start - em->start; - logical = em->block_start + logical; - if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) { - logical = em->block_start; - failrec->bio_flags = EXTENT_BIO_COMPRESSED; - extent_set_compress_type(&failrec->bio_flags, - em->compress_type); - } - failrec->logical = logical; - free_extent_map(em); - set_extent_bits(failure_tree, start, end, EXTENT_LOCKED | - EXTENT_DIRTY, GFP_NOFS); - set_state_private(failure_tree, start, - (u64)(unsigned long)failrec); - } else { - failrec = (struct io_failure_record *)(unsigned long)private; - } - num_copies = btrfs_num_copies( - &BTRFS_I(inode)->root->fs_info->mapping_tree, - failrec->logical, failrec->len); - failrec->last_mirror++; - if (!state) { - spin_lock(&BTRFS_I(inode)->io_tree.lock); - state = find_first_extent_bit_state(&BTRFS_I(inode)->io_tree, - failrec->start, - EXTENT_LOCKED); - if (state && state->start != failrec->start) - state = NULL; - spin_unlock(&BTRFS_I(inode)->io_tree.lock); - } - if (!state || failrec->last_mirror > num_copies) { - set_state_private(failure_tree, failrec->start, 0); - clear_extent_bits(failure_tree, failrec->start, - failrec->start + failrec->len - 1, - EXTENT_LOCKED | EXTENT_DIRTY, GFP_NOFS); - kfree(failrec); - return -EIO; - } - bio = bio_alloc(GFP_NOFS, 1); - bio->bi_private = state; - bio->bi_end_io = failed_bio->bi_end_io; - bio->bi_sector = failrec->logical >> 9; - bio->bi_bdev = failed_bio->bi_bdev; - bio->bi_size = 0; + trace_btrfs_writepage_end_io_hook(page, start, end, uptodate); - bio_add_page(bio, page, failrec->len, start - page_offset(page)); - if (failed_bio->bi_rw & REQ_WRITE) - rw = WRITE; - else - rw = READ; + ClearPagePrivate2(page); + if (!btrfs_dec_test_ordered_pending(inode, &ordered_extent, start, + end - start + 1, uptodate)) + return 0; - ret = BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, - failrec->last_mirror, - failrec->bio_flags, 0); - return ret; -} + ordered_extent->work.func = finish_ordered_fn; + ordered_extent->work.flags = 0; -/* - * each time an IO finishes, we do a fast check in the IO failure tree - * to see if we need to process or clean up an io_failure_record - */ -static int btrfs_clean_io_failures(struct inode *inode, u64 start) -{ - u64 private; - u64 private_failure; - struct io_failure_record *failure; - int ret; + if (btrfs_is_free_space_inode(inode)) + workers = &root->fs_info->endio_freespace_worker; + else + workers = &root->fs_info->endio_write_workers; + btrfs_queue_worker(workers, &ordered_extent->work); - private = 0; - if (count_range_bits(&BTRFS_I(inode)->io_failure_tree, &private, - (u64)-1, 1, EXTENT_DIRTY, 0)) { - ret = get_state_private(&BTRFS_I(inode)->io_failure_tree, - start, &private_failure); - if (ret == 0) { - failure = (struct io_failure_record *)(unsigned long) - private_failure; - set_state_private(&BTRFS_I(inode)->io_failure_tree, - failure->start, 0); - clear_extent_bits(&BTRFS_I(inode)->io_failure_tree, - failure->start, - failure->start + failure->len - 1, - EXTENT_DIRTY | EXTENT_LOCKED, - GFP_NOFS); - kfree(failure); - } - } return 0; } /* * when reads are done, we need to check csums to verify the data is correct - * if there's a match, we allow the bio to finish. If not, we go through - * the io_failure_record routines to find good copies + * if there's a match, we allow the bio to finish. If not, the code in + * extent_io.c will try to find good copies for us. */ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end, - struct extent_state *state) + struct extent_state *state, int mirror) { size_t offset = start - ((u64)page->index << PAGE_CACHE_SHIFT); struct inode *inode = page->mapping->host; @@ -2000,7 +2056,7 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end, } else { ret = get_state_private(io_tree, start, &private); } - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); if (ret) goto zeroit; @@ -2009,12 +2065,8 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end, if (csum != private) goto zeroit; - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); good: - /* if the io failure tree for this inode is non-empty, - * check to see if we've recovered from a failed IO - */ - btrfs_clean_io_failures(inode, start); return 0; zeroit: @@ -2025,7 +2077,7 @@ zeroit: (unsigned long long)private); memset(kaddr + offset, 1, end - start + 1); flush_dcache_page(page); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); if (private == 0) return 0; return -EIO; @@ -2036,6 +2088,8 @@ struct delayed_iput { struct inode *inode; }; +/* JDM: If this is fs-wide, why can't we add a pointer to + * btrfs_inode instead and avoid the allocation? */ void btrfs_add_delayed_iput(struct inode *inode) { struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info; @@ -2079,108 +2133,41 @@ void btrfs_run_delayed_iputs(struct btrfs_root *root) up_read(&root->fs_info->cleanup_work_sem); } -/* - * calculate extra metadata reservation when snapshotting a subvolume - * contains orphan files. - */ -void btrfs_orphan_pre_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending, - u64 *bytes_to_reserve) -{ - struct btrfs_root *root; - struct btrfs_block_rsv *block_rsv; - u64 num_bytes; - int index; - - root = pending->root; - if (!root->orphan_block_rsv || list_empty(&root->orphan_list)) - return; - - block_rsv = root->orphan_block_rsv; - - /* orphan block reservation for the snapshot */ - num_bytes = block_rsv->size; - - /* - * after the snapshot is created, COWing tree blocks may use more - * space than it frees. So we should make sure there is enough - * reserved space. - */ - index = trans->transid & 0x1; - if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) { - num_bytes += block_rsv->size - - (block_rsv->reserved + block_rsv->freed[index]); - } - - *bytes_to_reserve += num_bytes; -} - -void btrfs_orphan_post_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_pending_snapshot *pending) -{ - struct btrfs_root *root = pending->root; - struct btrfs_root *snap = pending->snap; - struct btrfs_block_rsv *block_rsv; - u64 num_bytes; - int index; - int ret; - - if (!root->orphan_block_rsv || list_empty(&root->orphan_list)) - return; - - /* refill source subvolume's orphan block reservation */ - block_rsv = root->orphan_block_rsv; - index = trans->transid & 0x1; - if (block_rsv->reserved + block_rsv->freed[index] < block_rsv->size) { - num_bytes = block_rsv->size - - (block_rsv->reserved + block_rsv->freed[index]); - ret = btrfs_block_rsv_migrate(&pending->block_rsv, - root->orphan_block_rsv, - num_bytes); - BUG_ON(ret); - } - - /* setup orphan block reservation for the snapshot */ - block_rsv = btrfs_alloc_block_rsv(snap); - BUG_ON(!block_rsv); - - btrfs_add_durable_block_rsv(root->fs_info, block_rsv); - snap->orphan_block_rsv = block_rsv; - - num_bytes = root->orphan_block_rsv->size; - ret = btrfs_block_rsv_migrate(&pending->block_rsv, - block_rsv, num_bytes); - BUG_ON(ret); - -#if 0 - /* insert orphan item for the snapshot */ - WARN_ON(!root->orphan_item_inserted); - ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root, - snap->root_key.objectid); - BUG_ON(ret); - snap->orphan_item_inserted = 1; -#endif -} - enum btrfs_orphan_cleanup_state { ORPHAN_CLEANUP_STARTED = 1, ORPHAN_CLEANUP_DONE = 2, }; /* - * This is called in transaction commmit time. If there are no orphan + * This is called in transaction commit time. If there are no orphan * files in the subvolume, it removes orphan item and frees block_rsv * structure. */ void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, struct btrfs_root *root) { + struct btrfs_block_rsv *block_rsv; int ret; - if (!list_empty(&root->orphan_list) || + if (atomic_read(&root->orphan_inodes) || root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) return; + spin_lock(&root->orphan_lock); + if (atomic_read(&root->orphan_inodes)) { + spin_unlock(&root->orphan_lock); + return; + } + + if (root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) { + spin_unlock(&root->orphan_lock); + return; + } + + block_rsv = root->orphan_block_rsv; + root->orphan_block_rsv = NULL; + spin_unlock(&root->orphan_lock); + if (root->orphan_item_inserted && btrfs_root_refs(&root->root_item) > 0) { ret = btrfs_del_orphan_item(trans, root->fs_info->tree_root, @@ -2189,10 +2176,9 @@ void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, root->orphan_item_inserted = 0; } - if (root->orphan_block_rsv) { - WARN_ON(root->orphan_block_rsv->size > 0); - btrfs_free_block_rsv(root, root->orphan_block_rsv); - root->orphan_block_rsv = NULL; + if (block_rsv) { + WARN_ON(block_rsv->size > 0); + btrfs_free_block_rsv(root, block_rsv); } } @@ -2225,8 +2211,8 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) block_rsv = NULL; } - if (list_empty(&BTRFS_I(inode)->i_orphan)) { - list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list); + if (!test_and_set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags)) { #if 0 /* * For proper ENOSPC handling, we should do orphan @@ -2239,34 +2225,40 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) insert = 1; #endif insert = 1; + atomic_dec(&root->orphan_inodes); } - if (!BTRFS_I(inode)->orphan_meta_reserved) { - BTRFS_I(inode)->orphan_meta_reserved = 1; + if (!test_and_set_bit(BTRFS_INODE_ORPHAN_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) reserve = 1; - } spin_unlock(&root->orphan_lock); - if (block_rsv) - btrfs_add_durable_block_rsv(root->fs_info, block_rsv); - /* grab metadata reservation from transaction handle */ if (reserve) { ret = btrfs_orphan_reserve_metadata(trans, inode); - BUG_ON(ret); + BUG_ON(ret); /* -ENOSPC in reservation; Logic error? JDM */ } /* insert an orphan item to track this unlinked/truncated file */ if (insert >= 1) { ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); - BUG_ON(ret); + if (ret && ret != -EEXIST) { + clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags); + btrfs_abort_transaction(trans, root, ret); + return ret; + } + ret = 0; } /* insert an orphan item to track subvolume contains orphan files */ if (insert >= 2) { ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root, root->root_key.objectid); - BUG_ON(ret); + if (ret && ret != -EEXIST) { + btrfs_abort_transaction(trans, root, ret); + return ret; + } } return 0; } @@ -2283,24 +2275,24 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode) int ret = 0; spin_lock(&root->orphan_lock); - if (!list_empty(&BTRFS_I(inode)->i_orphan)) { - list_del_init(&BTRFS_I(inode)->i_orphan); + if (test_and_clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags)) delete_item = 1; - } - if (BTRFS_I(inode)->orphan_meta_reserved) { - BTRFS_I(inode)->orphan_meta_reserved = 0; + if (test_and_clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) release_rsv = 1; - } spin_unlock(&root->orphan_lock); if (trans && delete_item) { ret = btrfs_del_orphan_item(trans, root, btrfs_ino(inode)); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM or corruption (JDM: Recheck) */ } - if (release_rsv) + if (release_rsv) { btrfs_orphan_release_metadata(inode); + atomic_dec(&root->orphan_inodes); + } return 0; } @@ -2316,6 +2308,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) struct btrfs_key key, found_key; struct btrfs_trans_handle *trans; struct inode *inode; + u64 last_objectid = 0; int ret = 0, nr_unlink = 0, nr_truncate = 0; if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) @@ -2367,41 +2360,82 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) * crossing root thing. we store the inode number in the * offset of the orphan item. */ + + if (found_key.offset == last_objectid) { + printk(KERN_ERR "btrfs: Error removing orphan entry, " + "stopping orphan cleanup\n"); + ret = -EINVAL; + goto out; + } + + last_objectid = found_key.offset; + found_key.objectid = found_key.offset; found_key.type = BTRFS_INODE_ITEM_KEY; found_key.offset = 0; inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL); - if (IS_ERR(inode)) { - ret = PTR_ERR(inode); + ret = PTR_RET(inode); + if (ret && ret != -ESTALE) goto out; - } - /* - * add this inode to the orphan list so btrfs_orphan_del does - * the proper thing when we hit it - */ - spin_lock(&root->orphan_lock); - list_add(&BTRFS_I(inode)->i_orphan, &root->orphan_list); - spin_unlock(&root->orphan_lock); + if (ret == -ESTALE && root == root->fs_info->tree_root) { + struct btrfs_root *dead_root; + struct btrfs_fs_info *fs_info = root->fs_info; + int is_dead_root = 0; + /* + * this is an orphan in the tree root. Currently these + * could come from 2 sources: + * a) a snapshot deletion in progress + * b) a free space cache inode + * We need to distinguish those two, as the snapshot + * orphan must not get deleted. + * find_dead_roots already ran before us, so if this + * is a snapshot deletion, we should find the root + * in the dead_roots list + */ + spin_lock(&fs_info->trans_lock); + list_for_each_entry(dead_root, &fs_info->dead_roots, + root_list) { + if (dead_root->root_key.objectid == + found_key.objectid) { + is_dead_root = 1; + break; + } + } + spin_unlock(&fs_info->trans_lock); + if (is_dead_root) { + /* prevent this orphan from being found again */ + key.offset = found_key.objectid - 1; + continue; + } + } /* - * if this is a bad inode, means we actually succeeded in - * removing the inode, but not the orphan record, which means - * we need to manually delete the orphan since iput will just - * do a destroy_inode + * Inode is already gone but the orphan item is still there, + * kill the orphan item. */ - if (is_bad_inode(inode)) { - trans = btrfs_start_transaction(root, 0); + if (ret == -ESTALE) { + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { ret = PTR_ERR(trans); goto out; } - btrfs_orphan_del(trans, inode); + printk(KERN_ERR "auto deleting %Lu\n", + found_key.objectid); + ret = btrfs_del_orphan_item(trans, root, + found_key.objectid); + BUG_ON(ret); /* -ENOMEM or corruption (JDM: Recheck) */ btrfs_end_transaction(trans, root); - iput(inode); continue; } + /* + * add this inode to the orphan list so btrfs_orphan_del does + * the proper thing when we hit it + */ + set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags); + /* if we have links, this was a truncate, lets do that */ if (inode->i_nlink) { if (!S_ISREG(inode->i_mode)) { @@ -2420,6 +2454,9 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) if (ret) goto out; } + /* release the path since we're done with it */ + btrfs_release_path(path); + root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; if (root->orphan_block_rsv) @@ -2534,7 +2571,7 @@ static void btrfs_read_locked_inode(struct inode *inode) inode_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item); inode->i_mode = btrfs_inode_mode(leaf, inode_item); - inode->i_nlink = btrfs_inode_nlink(leaf, inode_item); + set_nlink(inode, btrfs_inode_nlink(leaf, inode_item)); inode->i_uid = btrfs_inode_uid(leaf, inode_item); inode->i_gid = btrfs_inode_gid(leaf, inode_item); btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item)); @@ -2553,7 +2590,7 @@ static void btrfs_read_locked_inode(struct inode *inode) inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item)); BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item); - BTRFS_I(inode)->sequence = btrfs_inode_sequence(leaf, inode_item); + inode->i_version = btrfs_inode_sequence(leaf, inode_item); inode->i_generation = BTRFS_I(inode)->generation; inode->i_rdev = 0; rdev = btrfs_inode_rdev(leaf, inode_item); @@ -2637,7 +2674,7 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, btrfs_set_inode_nbytes(leaf, item, inode_get_bytes(inode)); btrfs_set_inode_generation(leaf, item, BTRFS_I(inode)->generation); - btrfs_set_inode_sequence(leaf, item, BTRFS_I(inode)->sequence); + btrfs_set_inode_sequence(leaf, item, inode->i_version); btrfs_set_inode_transid(leaf, item, trans->transid); btrfs_set_inode_rdev(leaf, item, inode->i_rdev); btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags); @@ -2647,7 +2684,7 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, /* * copy everything in the in-memory inode into the btree. */ -noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, +static noinline int btrfs_update_inode_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode) { struct btrfs_inode_item *inode_item; @@ -2655,21 +2692,6 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, struct extent_buffer *leaf; int ret; - /* - * If the inode is a free space inode, we can deadlock during commit - * if we put it into the delayed code. - * - * The data relocation inode should also be directly updated - * without delay - */ - if (!btrfs_is_free_space_inode(root, inode) - && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID) { - ret = btrfs_delayed_update_inode(trans, root, inode); - if (!ret) - btrfs_set_inode_last_trans(trans, inode); - return ret; - } - path = btrfs_alloc_path(); if (!path) return -ENOMEM; @@ -2698,6 +2720,45 @@ failed: } /* + * copy everything in the in-memory inode into the btree. + */ +noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode) +{ + int ret; + + /* + * If the inode is a free space inode, we can deadlock during commit + * if we put it into the delayed code. + * + * The data relocation inode should also be directly updated + * without delay + */ + if (!btrfs_is_free_space_inode(inode) + && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID) { + btrfs_update_root_times(trans, root); + + ret = btrfs_delayed_update_inode(trans, root, inode); + if (!ret) + btrfs_set_inode_last_trans(trans, inode); + return ret; + } + + return btrfs_update_inode_item(trans, root, inode); +} + +static noinline int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode) +{ + int ret; + + ret = btrfs_update_inode(trans, root, inode); + if (ret == -ENOSPC) + return btrfs_update_inode_item(trans, root, inode); + return ret; +} + +/* * unlink helper that gets used here in inode.c and in the tree logging * recovery code. It remove a link in a directory with a given name, and * also drops the back refs in the inode to the directory @@ -2746,16 +2807,22 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, printk(KERN_INFO "btrfs failed to delete reference to %.*s, " "inode %llu parent %llu\n", name_len, name, (unsigned long long)ino, (unsigned long long)dir_ino); + btrfs_abort_transaction(trans, root, ret); goto err; } ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); - if (ret) + if (ret) { + btrfs_abort_transaction(trans, root, ret); goto err; + } ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len, inode, dir_ino); - BUG_ON(ret != 0 && ret != -ENOENT); + if (ret != 0 && ret != -ENOENT) { + btrfs_abort_transaction(trans, root, ret); + goto err; + } ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len, dir, index); @@ -2767,8 +2834,10 @@ err: goto out; btrfs_i_size_write(dir, dir->i_size - name_len * 2); + inode_inc_iversion(inode); + inode_inc_iversion(dir); inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME; - btrfs_update_inode(trans, root, dir); + ret = btrfs_update_inode(trans, root, dir); out: return ret; } @@ -2835,7 +2904,16 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, u64 ino = btrfs_ino(inode); u64 dir_ino = btrfs_ino(dir); - trans = btrfs_start_transaction(root, 10); + /* + * 1 for the possible orphan item + * 1 for the dir item + * 1 for the dir index + * 1 for the inode ref + * 1 for the inode ref in the tree log + * 2 for the dir entries in the log + * 1 for the inode + */ + trans = btrfs_start_transaction(root, 8); if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC) return trans; @@ -2858,7 +2936,8 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, return ERR_PTR(-ENOMEM); } - trans = btrfs_start_transaction(root, 0); + /* 1 for the orphan item */ + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { btrfs_free_path(path); root->fs_info->enospc_unlink = 0; @@ -2903,7 +2982,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, err = ret; goto out; } - BUG_ON(ret == 0); + BUG_ON(ret == 0); /* Corruption */ if (check_path_shared(root, path)) goto out; btrfs_release_path(path); @@ -2936,7 +3015,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, err = PTR_ERR(ref); goto out; } - BUG_ON(!ref); + BUG_ON(!ref); /* Logic error */ if (check_path_shared(root, path)) goto out; index = btrfs_inode_ref_index(path->nodes[0], ref); @@ -2963,6 +3042,12 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, err = 0; out: btrfs_free_path(path); + /* Migrate the orphan reservation over */ + if (!err) + err = btrfs_block_rsv_migrate(trans->block_rsv, + &root->fs_info->global_block_rsv, + trans->bytes_reserved); + if (err) { btrfs_end_transaction(trans, root); root->fs_info->enospc_unlink = 0; @@ -2977,10 +3062,13 @@ static void __unlink_end_trans(struct btrfs_trans_handle *trans, struct btrfs_root *root) { if (trans->block_rsv == &root->fs_info->global_block_rsv) { + btrfs_block_rsv_release(root, trans->block_rsv, + trans->bytes_reserved); + trans->block_rsv = &root->fs_info->trans_block_rsv; BUG_ON(!root->fs_info->enospc_unlink); root->fs_info->enospc_unlink = 0; } - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); } static int btrfs_unlink(struct inode *dir, struct dentry *dentry) @@ -3034,23 +3122,42 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, di = btrfs_lookup_dir_item(trans, root, path, dir_ino, name, name_len, -1); - BUG_ON(IS_ERR_OR_NULL(di)); + if (IS_ERR_OR_NULL(di)) { + if (!di) + ret = -ENOENT; + else + ret = PTR_ERR(di); + goto out; + } leaf = path->nodes[0]; btrfs_dir_item_key_to_cpu(leaf, di, &key); WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); ret = btrfs_delete_one_dir_name(trans, root, path, di); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out; + } btrfs_release_path(path); ret = btrfs_del_root_ref(trans, root->fs_info->tree_root, objectid, root->root_key.objectid, dir_ino, &index, name, name_len); if (ret < 0) { - BUG_ON(ret != -ENOENT); + if (ret != -ENOENT) { + btrfs_abort_transaction(trans, root, ret); + goto out; + } di = btrfs_search_dir_index_item(root, path, dir_ino, name, name_len); - BUG_ON(IS_ERR_OR_NULL(di)); + if (IS_ERR_OR_NULL(di)) { + if (!di) + ret = -ENOENT; + else + ret = PTR_ERR(di); + btrfs_abort_transaction(trans, root, ret); + goto out; + } leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); @@ -3060,15 +3167,20 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, btrfs_release_path(path); ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out; + } btrfs_i_size_write(dir, dir->i_size - name_len * 2); + inode_inc_iversion(dir); dir->i_mtime = dir->i_ctime = CURRENT_TIME; - ret = btrfs_update_inode(trans, root, dir); - BUG_ON(ret); - + ret = btrfs_update_inode_fallback(trans, root, dir); + if (ret) + btrfs_abort_transaction(trans, root, ret); +out: btrfs_free_path(path); - return 0; + return ret; } static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) @@ -3144,7 +3256,6 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, int pending_del_nr = 0; int pending_del_slot = 0; int extent_type = -1; - int encoding; int ret; int err = 0; u64 ino = btrfs_ino(inode); @@ -3194,7 +3305,6 @@ search_again: leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); found_type = btrfs_key_type(&found_key); - encoding = 0; if (found_key.objectid != ino) break; @@ -3207,10 +3317,6 @@ search_again: fi = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item); extent_type = btrfs_file_extent_type(leaf, fi); - encoding = btrfs_file_extent_compression(leaf, fi); - encoding |= btrfs_file_extent_encryption(leaf, fi); - encoding |= btrfs_file_extent_other_encoding(leaf, fi); - if (extent_type != BTRFS_FILE_EXTENT_INLINE) { item_end += btrfs_file_extent_num_bytes(leaf, fi); @@ -3238,7 +3344,7 @@ search_again: if (extent_type != BTRFS_FILE_EXTENT_INLINE) { u64 num_dec; extent_start = btrfs_file_extent_disk_bytenr(leaf, fi); - if (!del_item && !encoding) { + if (!del_item) { u64 orig_num_bytes = btrfs_file_extent_num_bytes(leaf, fi); extent_num_bytes = new_size - @@ -3284,8 +3390,8 @@ search_again: } size = btrfs_file_extent_calc_inline_size(size); - ret = btrfs_truncate_item(trans, root, path, - size, 1); + btrfs_truncate_item(trans, root, path, + size, 1); } else if (root->ref_cows) { inode_sub_bytes(inode, item_end + 1 - found_key.offset); @@ -3314,7 +3420,7 @@ delete: ret = btrfs_free_extent(trans, root, extent_start, extent_num_bytes, 0, btrfs_header_owner(leaf), - ino, extent_offset); + ino, extent_offset, 0); BUG_ON(ret); } @@ -3333,7 +3439,11 @@ delete: ret = btrfs_del_items(trans, root, path, pending_del_slot, pending_del_nr); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, + root, ret); + goto error; + } pending_del_nr = 0; } btrfs_release_path(path); @@ -3346,8 +3456,10 @@ out: if (pending_del_nr) { ret = btrfs_del_items(trans, root, path, pending_del_slot, pending_del_nr); - BUG_ON(ret); + if (ret) + btrfs_abort_transaction(trans, root, ret); } +error: btrfs_free_path(path); return err; } @@ -3368,6 +3480,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from) pgoff_t index = from >> PAGE_CACHE_SHIFT; unsigned offset = from & (PAGE_CACHE_SIZE-1); struct page *page; + gfp_t mask = btrfs_alloc_write_mask(mapping); int ret = 0; u64 page_start; u64 page_end; @@ -3380,7 +3493,7 @@ static int btrfs_truncate_page(struct address_space *mapping, loff_t from) ret = -ENOMEM; again: - page = find_or_create_page(mapping, index, GFP_NOFS); + page = find_or_create_page(mapping, index, mask); if (!page) { btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); goto out; @@ -3404,8 +3517,7 @@ again: } wait_on_page_writeback(page); - lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state, - GFP_NOFS); + lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state); set_page_extent_mapped(page); ordered = btrfs_lookup_ordered_extent(inode, page_start); @@ -3481,7 +3593,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) btrfs_wait_ordered_range(inode, hole_start, block_end - hole_start); lock_extent_bits(io_tree, hole_start, block_end - 1, 0, - &cached_state, GFP_NOFS); + &cached_state); ordered = btrfs_lookup_ordered_extent(inode, hole_start); if (!ordered) break; @@ -3494,14 +3606,17 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) while (1) { em = btrfs_get_extent(inode, NULL, 0, cur_offset, block_end - cur_offset, 0); - BUG_ON(IS_ERR_OR_NULL(em)); + if (IS_ERR(em)) { + err = PTR_ERR(em); + break; + } last_byte = min(extent_map_end(em), block_end); last_byte = (last_byte + mask) & ~mask; if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) { u64 hint_byte = 0; hole_size = last_byte - cur_offset; - trans = btrfs_start_transaction(root, 2); + trans = btrfs_start_transaction(root, 3); if (IS_ERR(trans)) { err = PTR_ERR(trans); break; @@ -3511,6 +3626,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) cur_offset + hole_size, &hint_byte, 1); if (err) { + btrfs_abort_transaction(trans, root, err); btrfs_end_transaction(trans, root); break; } @@ -3520,6 +3636,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) 0, hole_size, 0, hole_size, 0, 0, 0); if (err) { + btrfs_abort_transaction(trans, root, err); btrfs_end_transaction(trans, root); break; } @@ -3527,6 +3644,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) btrfs_drop_extent_cache(inode, hole_start, last_byte - 1, 0); + btrfs_update_inode(trans, root, inode); btrfs_end_transaction(trans, root); } free_extent_map(em); @@ -3544,6 +3662,8 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) static int btrfs_setsize(struct inode *inode, loff_t newsize) { + struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_trans_handle *trans; loff_t oldsize = i_size_read(inode); int ret; @@ -3551,16 +3671,19 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return 0; if (newsize > oldsize) { - i_size_write(inode, newsize); - btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); truncate_pagecache(inode, oldsize, newsize); ret = btrfs_cont_expand(inode, oldsize, newsize); - if (ret) { - btrfs_setsize(inode, oldsize); + if (ret) return ret; - } - mark_inode_dirty(inode); + trans = btrfs_start_transaction(root, 1); + if (IS_ERR(trans)) + return PTR_ERR(trans); + + i_size_write(inode, newsize); + btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); + ret = btrfs_update_inode(trans, root, inode); + btrfs_end_transaction(trans, root); } else { /* @@ -3569,7 +3692,8 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) * any new writes get down to disk quickly. */ if (newsize == 0) - BTRFS_I(inode)->ordered_data_close = 1; + set_bit(BTRFS_INODE_ORDERED_DATA_CLOSE, + &BTRFS_I(inode)->runtime_flags); /* we don't support swapfiles, so vmtruncate shouldn't fail */ truncate_setsize(inode, newsize); @@ -3600,9 +3724,10 @@ static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) if (attr->ia_valid) { setattr_copy(inode, attr); - mark_inode_dirty(inode); + inode_inc_iversion(inode); + err = btrfs_dirty_inode(inode); - if (attr->ia_valid & ATTR_MODE) + if (!err && attr->ia_valid & ATTR_MODE) err = btrfs_acl_chmod(inode); } @@ -3613,6 +3738,8 @@ void btrfs_evict_inode(struct inode *inode) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_block_rsv *rsv, *global_rsv; + u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); unsigned long nr; int ret; @@ -3620,7 +3747,7 @@ void btrfs_evict_inode(struct inode *inode) truncate_inode_pages(&inode->i_data, 0); if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 || - btrfs_is_free_space_inode(root, inode))) + btrfs_is_free_space_inode(inode))) goto no_delete; if (is_bad_inode(inode)) { @@ -3631,7 +3758,8 @@ void btrfs_evict_inode(struct inode *inode) btrfs_wait_ordered_range(inode, 0, (u64)-1); if (root->fs_info->log_root_recovering) { - BUG_ON(!list_empty(&BTRFS_I(inode)->i_orphan)); + BUG_ON(test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags)); goto no_delete; } @@ -3640,22 +3768,55 @@ void btrfs_evict_inode(struct inode *inode) goto no_delete; } + rsv = btrfs_alloc_block_rsv(root); + if (!rsv) { + btrfs_orphan_del(NULL, inode); + goto no_delete; + } + rsv->size = min_size; + global_rsv = &root->fs_info->global_block_rsv; + btrfs_i_size_write(inode, 0); + /* + * This is a bit simpler than btrfs_truncate since + * + * 1) We've already reserved our space for our orphan item in the + * unlink. + * 2) We're going to delete the inode item, so we don't need to update + * it at all. + * + * So we just need to reserve some slack space in case we add bytes when + * doing the truncate. + */ while (1) { - trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); - trans->block_rsv = root->orphan_block_rsv; + ret = btrfs_block_rsv_refill_noflush(root, rsv, min_size); + + /* + * Try and steal from the global reserve since we will + * likely not use this space anyway, we want to try as + * hard as possible to get this to work. + */ + if (ret) + ret = btrfs_block_rsv_migrate(global_rsv, rsv, min_size); - ret = btrfs_block_rsv_check(trans, root, - root->orphan_block_rsv, 0, 5); if (ret) { - BUG_ON(ret != -EAGAIN); - ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); - continue; + printk(KERN_WARNING "Could not get space for a " + "delete, will truncate on mount %d\n", ret); + btrfs_orphan_del(NULL, inode); + btrfs_free_block_rsv(root, rsv); + goto no_delete; } + trans = btrfs_start_transaction(root, 0); + if (IS_ERR(trans)) { + btrfs_orphan_del(NULL, inode); + btrfs_free_block_rsv(root, rsv); + goto no_delete; + } + + trans->block_rsv = rsv; + ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0); if (ret != -EAGAIN) break; @@ -3664,14 +3825,17 @@ void btrfs_evict_inode(struct inode *inode) btrfs_end_transaction(trans, root); trans = NULL; btrfs_btree_balance_dirty(root, nr); - } + btrfs_free_block_rsv(root, rsv); + if (ret == 0) { + trans->block_rsv = root->orphan_block_rsv; ret = btrfs_orphan_del(trans, inode); BUG_ON(ret); } + trans->block_rsv = &root->fs_info->trans_block_rsv; if (!(root == root->fs_info->tree_root || root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) btrfs_return_ino(root, btrfs_ino(inode)); @@ -3680,7 +3844,7 @@ void btrfs_evict_inode(struct inode *inode) btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); no_delete: - end_writeback(inode); + clear_inode(inode); return; } @@ -3855,7 +4019,7 @@ static void inode_tree_del(struct inode *inode) } } -int btrfs_invalidate_inodes(struct btrfs_root *root) +void btrfs_invalidate_inodes(struct btrfs_root *root) { struct rb_node *node; struct rb_node *prev; @@ -3915,7 +4079,6 @@ again: node = rb_next(node); } spin_unlock(&root->inode_lock); - return 0; } static int btrfs_init_locked_inode(struct inode *inode, void *p) @@ -3923,7 +4086,6 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p) struct btrfs_iget_args *args = p; inode->i_ino = args->ino; BTRFS_I(inode)->root = args->root; - btrfs_set_inode_space_info(args->root, inode); return 0; } @@ -3991,10 +4153,10 @@ static struct inode *new_simple_dir(struct super_block *s, BTRFS_I(inode)->root = root; memcpy(&BTRFS_I(inode)->location, key, sizeof(*key)); - BTRFS_I(inode)->dummy_inode = 1; + set_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags); inode->i_ino = BTRFS_EMPTY_SUBVOL_DIR_OBJECTID; - inode->i_op = &simple_dir_inode_operations; + inode->i_op = &btrfs_dir_ro_inode_operations; inode->i_fop = &simple_dir_operations; inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; @@ -4065,14 +4227,18 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry) static int btrfs_dentry_delete(const struct dentry *dentry) { struct btrfs_root *root; + struct inode *inode = dentry->d_inode; - if (!dentry->d_inode && !IS_ROOT(dentry)) - dentry = dentry->d_parent; + if (!inode && !IS_ROOT(dentry)) + inode = dentry->d_parent->d_inode; - if (dentry->d_inode) { - root = BTRFS_I(dentry->d_inode)->root; + if (inode) { + root = BTRFS_I(inode)->root; if (btrfs_root_refs(&root->root_item) == 0) return 1; + + if (btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) + return 1; } return 0; } @@ -4084,7 +4250,7 @@ static void btrfs_dentry_release(struct dentry *dentry) } static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry, - struct nameidata *nd) + unsigned int flags) { struct dentry *ret; @@ -4113,7 +4279,6 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, struct btrfs_path *path; struct list_head ins_list; struct list_head del_list; - struct qstr q; int ret; struct extent_buffer *leaf; int slot; @@ -4204,7 +4369,6 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, while (di_cur < di_total) { struct btrfs_key location; - struct dentry *tmp; if (verify_dir_item(root, leaf, di)) break; @@ -4225,35 +4389,15 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)]; btrfs_dir_item_key_to_cpu(leaf, di, &location); - q.name = name_ptr; - q.len = name_len; - q.hash = full_name_hash(q.name, q.len); - tmp = d_lookup(filp->f_dentry, &q); - if (!tmp) { - struct btrfs_key *newkey; - - newkey = kzalloc(sizeof(struct btrfs_key), - GFP_NOFS); - if (!newkey) - goto no_dentry; - tmp = d_alloc(filp->f_dentry, &q); - if (!tmp) { - kfree(newkey); - dput(tmp); - goto no_dentry; - } - memcpy(newkey, &location, - sizeof(struct btrfs_key)); - tmp->d_fsdata = newkey; - tmp->d_flags |= DCACHE_NEED_LOOKUP; - d_rehash(tmp); - dput(tmp); - } else { - dput(tmp); - } -no_dentry: + /* is this a reference to our own snapshot? If so - * skip it + * skip it. + * + * In contrast to old kernels, we insert the snapshot's + * dir item and dir index after it has been created, so + * we won't find a reference to our own snapshot. We + * still keep the following code for backward + * compatibility. */ if (location.type == BTRFS_ROOT_ITEM_KEY && location.objectid == root->root_key.objectid) { @@ -4313,10 +4457,10 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) int ret = 0; bool nolock = false; - if (BTRFS_I(inode)->dummy_inode) + if (test_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags)) return 0; - if (btrfs_fs_closing(root->fs_info) && btrfs_is_free_space_inode(root, inode)) + if (btrfs_fs_closing(root->fs_info) && btrfs_is_free_space_inode(inode)) nolock = true; if (wbc->sync_mode == WB_SYNC_ALL) { @@ -4340,42 +4484,57 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) * FIXME, needs more benchmarking...there are no reasons other than performance * to keep or drop this code. */ -void btrfs_dirty_inode(struct inode *inode, int flags) +int btrfs_dirty_inode(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; int ret; - if (BTRFS_I(inode)->dummy_inode) - return; + if (test_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags)) + return 0; trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); ret = btrfs_update_inode(trans, root, inode); if (ret && ret == -ENOSPC) { /* whoops, lets try again with the full transaction */ btrfs_end_transaction(trans, root); trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) { - printk_ratelimited(KERN_ERR "btrfs: fail to " - "dirty inode %llu error %ld\n", - (unsigned long long)btrfs_ino(inode), - PTR_ERR(trans)); - return; - } + if (IS_ERR(trans)) + return PTR_ERR(trans); ret = btrfs_update_inode(trans, root, inode); - if (ret) { - printk_ratelimited(KERN_ERR "btrfs: fail to " - "dirty inode %llu error %d\n", - (unsigned long long)btrfs_ino(inode), - ret); - } } btrfs_end_transaction(trans, root); if (BTRFS_I(inode)->delayed_node) btrfs_balance_delayed_items(root); + + return ret; +} + +/* + * This is a copy of file_update_time. We need this so we can return error on + * ENOSPC for updating the inode in the case of file write and mmap writes. + */ +static int btrfs_update_time(struct inode *inode, struct timespec *now, + int flags) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + + if (btrfs_root_readonly(root)) + return -EROFS; + + if (flags & S_VERSION) + inode_inc_iversion(inode); + if (flags & S_CTIME) + inode->i_ctime = *now; + if (flags & S_MTIME) + inode->i_mtime = *now; + if (flags & S_ATIME) + inode->i_atime = *now; + return btrfs_dirty_inode(inode); } /* @@ -4462,8 +4621,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *dir, const char *name, int name_len, - u64 ref_objectid, u64 objectid, int mode, - u64 *index) + u64 ref_objectid, u64 objectid, + umode_t mode, u64 *index) { struct inode *inode; struct btrfs_inode_item *inode_item; @@ -4511,7 +4670,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, BTRFS_I(inode)->root = root; BTRFS_I(inode)->generation = trans->transid; inode->i_generation = BTRFS_I(inode)->generation; - btrfs_set_inode_space_info(root, inode); if (S_ISDIR(mode)) owner = 0; @@ -4539,6 +4697,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0], struct btrfs_inode_item); + memset_extent_buffer(path->nodes[0], 0, (unsigned long)inode_item, + sizeof(*inode_item)); fill_inode_item(trans, path->nodes[0], inode_item, inode); ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1, @@ -4572,6 +4732,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, trace_btrfs_inode_new(inode); btrfs_set_inode_last_trans(trans, inode); + btrfs_update_root_times(trans, root); + return inode; fail: if (dir) @@ -4619,16 +4781,43 @@ int btrfs_add_link(struct btrfs_trans_handle *trans, parent_ino, index); } - if (ret == 0) { - ret = btrfs_insert_dir_item(trans, root, name, name_len, - parent_inode, &key, - btrfs_inode_type(inode), index); - BUG_ON(ret); + /* Nothing to clean up yet */ + if (ret) + return ret; - btrfs_i_size_write(parent_inode, parent_inode->i_size + - name_len * 2); - parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME; - ret = btrfs_update_inode(trans, root, parent_inode); + ret = btrfs_insert_dir_item(trans, root, name, name_len, + parent_inode, &key, + btrfs_inode_type(inode), index); + if (ret == -EEXIST) + goto fail_dir_item; + else if (ret) { + btrfs_abort_transaction(trans, root, ret); + return ret; + } + + btrfs_i_size_write(parent_inode, parent_inode->i_size + + name_len * 2); + inode_inc_iversion(parent_inode); + parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME; + ret = btrfs_update_inode(trans, root, parent_inode); + if (ret) + btrfs_abort_transaction(trans, root, ret); + return ret; + +fail_dir_item: + if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { + u64 local_index; + int err; + err = btrfs_del_root_ref(trans, root->fs_info->tree_root, + key.objectid, root->root_key.objectid, + parent_ino, &local_index, name, name_len); + + } else if (add_backref) { + u64 local_index; + int err; + + err = btrfs_del_inode_ref(trans, root, name, name_len, + ino, parent_ino, &local_index); } return ret; } @@ -4640,17 +4829,13 @@ static int btrfs_add_nondir(struct btrfs_trans_handle *trans, int err = btrfs_add_link(trans, dir, inode, dentry->d_name.name, dentry->d_name.len, backref, index); - if (!err) { - d_instantiate(dentry, inode); - return 0; - } if (err > 0) err = -EEXIST; return err; } static int btrfs_mknod(struct inode *dir, struct dentry *dentry, - int mode, dev_t rdev) + umode_t mode, dev_t rdev) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(dir)->root; @@ -4691,17 +4876,25 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + + inode->i_op = &btrfs_special_inode_operations; err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { - inode->i_op = &btrfs_special_inode_operations; init_special_inode(inode, inode->i_mode, rdev); btrfs_update_inode(trans, root, inode); + d_instantiate(dentry, inode); } out_unlock: nr = trans->blocks_used; - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); if (drop_inode) { inode_dec_link_count(inode); @@ -4711,7 +4904,7 @@ out_unlock: } static int btrfs_create(struct inode *dir, struct dentry *dentry, - int mode, struct nameidata *nd) + umode_t mode, bool excl) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(dir)->root; @@ -4749,19 +4942,27 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + inode->i_fop = &btrfs_file_operations; + inode->i_op = &btrfs_file_inode_operations; + err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { inode->i_mapping->a_ops = &btrfs_aops; inode->i_mapping->backing_dev_info = &root->fs_info->bdi; - inode->i_fop = &btrfs_file_operations; - inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; + d_instantiate(dentry, inode); } out_unlock: nr = trans->blocks_used; - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); if (drop_inode) { inode_dec_link_count(inode); iput(inode); @@ -4804,6 +5005,7 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, } btrfs_inc_nlink(inode); + inode_inc_iversion(inode); inode->i_ctime = CURRENT_TIME; ihold(inode); @@ -4814,12 +5016,14 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, } else { struct dentry *parent = dentry->d_parent; err = btrfs_update_inode(trans, root, inode); - BUG_ON(err); + if (err) + goto fail; + d_instantiate(dentry, inode); btrfs_log_new_name(trans, inode, NULL, parent); } nr = trans->blocks_used; - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); fail: if (drop_inode) { inode_dec_link_count(inode); @@ -4829,7 +5033,7 @@ fail: return err; } -static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) +static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) { struct inode *inode = NULL; struct btrfs_trans_handle *trans; @@ -4885,7 +5089,7 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) out_fail: nr = trans->blocks_used; - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); if (drop_on_err) iput(inode); btrfs_btree_balance_dirty(root, nr); @@ -4944,12 +5148,12 @@ static noinline int uncompress_inline(struct btrfs_path *path, ret = btrfs_decompress(compress_type, tmp, page, extent_offset, inline_size, max_size); if (ret) { - char *kaddr = kmap_atomic(page, KM_USER0); + char *kaddr = kmap_atomic(page); unsigned long copy_size = min_t(u64, PAGE_CACHE_SIZE - pg_offset, max_size - extent_offset); memset(kaddr + pg_offset, 0, copy_size); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); } kfree(tmp); return 0; @@ -5144,7 +5348,7 @@ again: ret = uncompress_inline(path, inode, page, pg_offset, extent_offset, item); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ } else { map = kmap(page); read_extent_buffer(leaf, map + pg_offset, ptr, @@ -5158,7 +5362,7 @@ again: } flush_dcache_page(page); } else if (create && PageUptodate(page)) { - WARN_ON(1); + BUG(); if (!trans) { kunmap(page); free_extent_map(em); @@ -5259,6 +5463,7 @@ out: free_extent_map(em); return ERR_PTR(err); } + BUG_ON(!em); /* Error is always set */ return em; } @@ -5421,7 +5626,7 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, alloc_hint = get_extent_allocation_hint(inode, start, len); ret = btrfs_reserve_extent(trans, root, len, root->sectorsize, 0, - alloc_hint, (u64)-1, &ins, 1); + alloc_hint, &ins, 1); if (ret) { em = ERR_PTR(ret); goto out; @@ -5570,18 +5775,112 @@ out: return ret; } +static int lock_extent_direct(struct inode *inode, u64 lockstart, u64 lockend, + struct extent_state **cached_state, int writing) +{ + struct btrfs_ordered_extent *ordered; + int ret = 0; + + while (1) { + lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend, + 0, cached_state); + /* + * We're concerned with the entire range that we're going to be + * doing DIO to, so we need to make sure theres no ordered + * extents in this range. + */ + ordered = btrfs_lookup_ordered_range(inode, lockstart, + lockend - lockstart + 1); + + /* + * We need to make sure there are no buffered pages in this + * range either, we could have raced between the invalidate in + * generic_file_direct_write and locking the extent. The + * invalidate needs to happen so that reads after a write do not + * get stale data. + */ + if (!ordered && (!writing || + !test_range_bit(&BTRFS_I(inode)->io_tree, + lockstart, lockend, EXTENT_UPTODATE, 0, + *cached_state))) + break; + + unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart, lockend, + cached_state, GFP_NOFS); + + if (ordered) { + btrfs_start_ordered_extent(inode, ordered, 1); + btrfs_put_ordered_extent(ordered); + } else { + /* Screw you mmap */ + ret = filemap_write_and_wait_range(inode->i_mapping, + lockstart, + lockend); + if (ret) + break; + + /* + * If we found a page that couldn't be invalidated just + * fall back to buffered. + */ + ret = invalidate_inode_pages2_range(inode->i_mapping, + lockstart >> PAGE_CACHE_SHIFT, + lockend >> PAGE_CACHE_SHIFT); + if (ret) + break; + } + + cond_resched(); + } + + return ret; +} + static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create) { struct extent_map *em; struct btrfs_root *root = BTRFS_I(inode)->root; + struct extent_state *cached_state = NULL; u64 start = iblock << inode->i_blkbits; + u64 lockstart, lockend; u64 len = bh_result->b_size; struct btrfs_trans_handle *trans; + int unlock_bits = EXTENT_LOCKED; + int ret; + + if (create) { + ret = btrfs_delalloc_reserve_space(inode, len); + if (ret) + return ret; + unlock_bits |= EXTENT_DELALLOC | EXTENT_DIRTY; + } else { + len = min_t(u64, len, root->sectorsize); + } + + lockstart = start; + lockend = start + len - 1; + + /* + * If this errors out it's because we couldn't invalidate pagecache for + * this range and we need to fallback to buffered. + */ + if (lock_extent_direct(inode, lockstart, lockend, &cached_state, create)) + return -ENOTBLK; + + if (create) { + ret = set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, + lockend, EXTENT_DELALLOC, NULL, + &cached_state, GFP_NOFS); + if (ret) + goto unlock_err; + } em = btrfs_get_extent(inode, NULL, 0, start, len, 0); - if (IS_ERR(em)) - return PTR_ERR(em); + if (IS_ERR(em)) { + ret = PTR_ERR(em); + goto unlock_err; + } /* * Ok for INLINE and COMPRESSED extents we need to fallback on buffered @@ -5600,17 +5899,16 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) || em->block_start == EXTENT_MAP_INLINE) { free_extent_map(em); - return -ENOTBLK; + ret = -ENOTBLK; + goto unlock_err; } /* Just a good old fashioned hole, return */ if (!create && (em->block_start == EXTENT_MAP_HOLE || test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) { free_extent_map(em); - /* DIO will do one hole at a time, so just unlock a sector */ - unlock_extent(&BTRFS_I(inode)->io_tree, start, - start + root->sectorsize - 1, GFP_NOFS); - return 0; + ret = 0; + goto unlock_err; } /* @@ -5623,8 +5921,9 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, * */ if (!create) { - len = em->len - (start - em->start); - goto map; + len = min(len, em->len - (start - em->start)); + lockstart = start + len; + goto unlock; } if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) || @@ -5656,7 +5955,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, btrfs_end_transaction(trans, root); if (ret) { free_extent_map(em); - return ret; + goto unlock_err; } goto unlock; } @@ -5669,25 +5968,67 @@ must_cow: */ len = bh_result->b_size; em = btrfs_new_extent_direct(inode, em, start, len); - if (IS_ERR(em)) - return PTR_ERR(em); + if (IS_ERR(em)) { + ret = PTR_ERR(em); + goto unlock_err; + } len = min(len, em->len - (start - em->start)); unlock: - clear_extent_bit(&BTRFS_I(inode)->io_tree, start, start + len - 1, - EXTENT_LOCKED | EXTENT_DELALLOC | EXTENT_DIRTY, 1, - 0, NULL, GFP_NOFS); -map: bh_result->b_blocknr = (em->block_start + (start - em->start)) >> inode->i_blkbits; bh_result->b_size = len; bh_result->b_bdev = em->bdev; set_buffer_mapped(bh_result); - if (create && !test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) - set_buffer_new(bh_result); + if (create) { + if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) + set_buffer_new(bh_result); + + /* + * Need to update the i_size under the extent lock so buffered + * readers will get the updated i_size when we unlock. + */ + if (start + len > i_size_read(inode)) + i_size_write(inode, start + len); + } + + /* + * In the case of write we need to clear and unlock the entire range, + * in the case of read we need to unlock only the end area that we + * aren't using if there is any left over space. + */ + if (lockstart < lockend) { + if (create && len < lockend - lockstart) { + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, + lockstart + len - 1, unlock_bits, 1, 0, + &cached_state, GFP_NOFS); + /* + * Beside unlock, we also need to cleanup reserved space + * for the left range by attaching EXTENT_DO_ACCOUNTING. + */ + clear_extent_bit(&BTRFS_I(inode)->io_tree, + lockstart + len, lockend, + unlock_bits | EXTENT_DO_ACCOUNTING, + 1, 0, NULL, GFP_NOFS); + } else { + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, + lockend, unlock_bits, 1, 0, + &cached_state, GFP_NOFS); + } + } else { + free_extent_state(cached_state); + } free_extent_map(em); return 0; + +unlock_err: + if (create) + unlock_bits |= EXTENT_DO_ACCOUNTING; + + clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, + unlock_bits, 1, 0, &cached_state, GFP_NOFS); + return ret; } struct btrfs_dio_private { @@ -5695,7 +6036,6 @@ struct btrfs_dio_private { u64 logical_offset; u64 disk_bytenr; u64 bytes; - u32 *csums; void *private; /* number of bios pending for this dio */ @@ -5715,7 +6055,6 @@ static void btrfs_endio_direct_read(struct bio *bio, int err) struct inode *inode = dip->inode; struct btrfs_root *root = BTRFS_I(inode)->root; u64 start; - u32 *private = dip->csums; start = dip->logical_offset; do { @@ -5723,37 +6062,40 @@ static void btrfs_endio_direct_read(struct bio *bio, int err) struct page *page = bvec->bv_page; char *kaddr; u32 csum = ~(u32)0; + u64 private = ~(u32)0; unsigned long flags; + if (get_state_private(&BTRFS_I(inode)->io_tree, + start, &private)) + goto failed; local_irq_save(flags); - kaddr = kmap_atomic(page, KM_IRQ0); + kaddr = kmap_atomic(page); csum = btrfs_csum_data(root, kaddr + bvec->bv_offset, csum, bvec->bv_len); btrfs_csum_final(csum, (char *)&csum); - kunmap_atomic(kaddr, KM_IRQ0); + kunmap_atomic(kaddr); local_irq_restore(flags); flush_dcache_page(bvec->bv_page); - if (csum != *private) { + if (csum != private) { +failed: printk(KERN_ERR "btrfs csum failed ino %llu off" " %llu csum %u private %u\n", (unsigned long long)btrfs_ino(inode), (unsigned long long)start, - csum, *private); + csum, (unsigned)private); err = -EIO; } } start += bvec->bv_len; - private++; bvec++; } while (bvec <= bvec_end); unlock_extent(&BTRFS_I(inode)->io_tree, dip->logical_offset, - dip->logical_offset + dip->bytes - 1, GFP_NOFS); + dip->logical_offset + dip->bytes - 1); bio->bi_private = dip->private; - kfree(dip->csums); kfree(dip); /* If we had a csum failure make sure to clear the uptodate flag */ @@ -5767,9 +6109,7 @@ static void btrfs_endio_direct_write(struct bio *bio, int err) struct btrfs_dio_private *dip = bio->bi_private; struct inode *inode = dip->inode; struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_trans_handle *trans; struct btrfs_ordered_extent *ordered = NULL; - struct extent_state *cached_state = NULL; u64 ordered_offset = dip->logical_offset; u64 ordered_bytes = dip->bytes; int ret; @@ -5779,74 +6119,14 @@ static void btrfs_endio_direct_write(struct bio *bio, int err) again: ret = btrfs_dec_test_first_ordered_pending(inode, &ordered, &ordered_offset, - ordered_bytes); + ordered_bytes, !err); if (!ret) goto out_test; - BUG_ON(!ordered); - - trans = btrfs_join_transaction(root); - if (IS_ERR(trans)) { - err = -ENOMEM; - goto out; - } - trans->block_rsv = &root->fs_info->delalloc_block_rsv; - - if (test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags)) { - ret = btrfs_ordered_update_i_size(inode, 0, ordered); - if (!ret) - ret = btrfs_update_inode(trans, root, inode); - err = ret; - goto out; - } - - lock_extent_bits(&BTRFS_I(inode)->io_tree, ordered->file_offset, - ordered->file_offset + ordered->len - 1, 0, - &cached_state, GFP_NOFS); - - if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags)) { - ret = btrfs_mark_extent_written(trans, inode, - ordered->file_offset, - ordered->file_offset + - ordered->len); - if (ret) { - err = ret; - goto out_unlock; - } - } else { - ret = insert_reserved_file_extent(trans, inode, - ordered->file_offset, - ordered->start, - ordered->disk_len, - ordered->len, - ordered->len, - 0, 0, 0, - BTRFS_FILE_EXTENT_REG); - unpin_extent_cache(&BTRFS_I(inode)->extent_tree, - ordered->file_offset, ordered->len); - if (ret) { - err = ret; - WARN_ON(1); - goto out_unlock; - } - } - - add_pending_csums(trans, inode, ordered->file_offset, &ordered->list); - ret = btrfs_ordered_update_i_size(inode, 0, ordered); - if (!ret || !test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags)) - btrfs_update_inode(trans, root, inode); - ret = 0; -out_unlock: - unlock_extent_cached(&BTRFS_I(inode)->io_tree, ordered->file_offset, - ordered->file_offset + ordered->len - 1, - &cached_state, GFP_NOFS); -out: - btrfs_delalloc_release_metadata(inode, ordered->len); - btrfs_end_transaction(trans, root); - ordered_offset = ordered->file_offset + ordered->len; - btrfs_put_ordered_extent(ordered); - btrfs_put_ordered_extent(ordered); - + ordered->work.func = finish_ordered_fn; + ordered->work.flags = 0; + btrfs_queue_worker(&root->fs_info->endio_write_workers, + &ordered->work); out_test: /* * our bio might span multiple ordered extents. If we haven't @@ -5855,12 +6135,12 @@ out_test: if (ordered_offset < dip->logical_offset + dip->bytes) { ordered_bytes = dip->logical_offset + dip->bytes - ordered_offset; + ordered = NULL; goto again; } out_done: bio->bi_private = dip->private; - kfree(dip->csums); kfree(dip); /* If we had an error make sure to clear the uptodate flag */ @@ -5876,7 +6156,7 @@ static int __btrfs_submit_bio_start_direct_io(struct inode *inode, int rw, int ret; struct btrfs_root *root = BTRFS_I(inode)->root; ret = btrfs_csum_one_bio(root, inode, bio, offset, 1); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ return 0; } @@ -5921,16 +6201,19 @@ static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev, static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, int rw, u64 file_offset, int skip_sum, - u32 *csums, int async_submit) + int async_submit) { int write = rw & REQ_WRITE; struct btrfs_root *root = BTRFS_I(inode)->root; int ret; bio_get(bio); - ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); - if (ret) - goto err; + + if (!write) { + ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); + if (ret) + goto err; + } if (skip_sum) goto map; @@ -5951,8 +6234,7 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, if (ret) goto err; } else if (!skip_sum) { - ret = btrfs_lookup_bio_sums_dio(root, inode, bio, - file_offset, csums); + ret = btrfs_lookup_bio_sums_dio(root, inode, bio, file_offset); if (ret) goto err; } @@ -5978,10 +6260,8 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, u64 submit_len = 0; u64 map_length; int nr_pages = 0; - u32 *csums = dip->csums; int ret = 0; int async_submit = 0; - int write = rw & REQ_WRITE; map_length = orig_bio->bi_size; ret = btrfs_map_block(map_tree, READ, start_sector << 9, @@ -6017,16 +6297,13 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, atomic_inc(&dip->pending_bios); ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, - csums, async_submit); + async_submit); if (ret) { bio_put(bio); atomic_dec(&dip->pending_bios); goto out_err; } - /* Write's use the ordered csums */ - if (!write && !skip_sum) - csums = csums + nr_pages; start_sector += submit_len >> 9; file_offset += submit_len; @@ -6056,7 +6333,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, submit: ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, - csums, async_submit); + async_submit); if (!ret) return 0; @@ -6092,17 +6369,6 @@ static void btrfs_submit_direct(int rw, struct bio *bio, struct inode *inode, ret = -ENOMEM; goto free_ordered; } - dip->csums = NULL; - - /* Write's use the ordered csum stuff, so we don't need dip->csums */ - if (!write && !skip_sum) { - dip->csums = kmalloc(sizeof(u32) * bio->bi_vcnt, GFP_NOFS); - if (!dip->csums) { - kfree(dip); - ret = -ENOMEM; - goto free_ordered; - } - } dip->private = bio->bi_private; dip->inode = inode; @@ -6187,96 +6453,22 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io out: return retval; } + static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, loff_t offset, unsigned long nr_segs) { struct file *file = iocb->ki_filp; struct inode *inode = file->f_mapping->host; - struct btrfs_ordered_extent *ordered; - struct extent_state *cached_state = NULL; - u64 lockstart, lockend; - ssize_t ret; - int writing = rw & WRITE; - int write_bits = 0; - size_t count = iov_length(iov, nr_segs); if (check_direct_IO(BTRFS_I(inode)->root, rw, iocb, iov, - offset, nr_segs)) { + offset, nr_segs)) return 0; - } - - lockstart = offset; - lockend = offset + count - 1; - - if (writing) { - ret = btrfs_delalloc_reserve_space(inode, count); - if (ret) - goto out; - } - - while (1) { - lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend, - 0, &cached_state, GFP_NOFS); - /* - * We're concerned with the entire range that we're going to be - * doing DIO to, so we need to make sure theres no ordered - * extents in this range. - */ - ordered = btrfs_lookup_ordered_range(inode, lockstart, - lockend - lockstart + 1); - if (!ordered) - break; - unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart, lockend, - &cached_state, GFP_NOFS); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - cond_resched(); - } - /* - * we don't use btrfs_set_extent_delalloc because we don't want - * the dirty or uptodate bits - */ - if (writing) { - write_bits = EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING; - ret = set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, - EXTENT_DELALLOC, 0, NULL, &cached_state, - GFP_NOFS); - if (ret) { - clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, - lockend, EXTENT_LOCKED | write_bits, - 1, 0, &cached_state, GFP_NOFS); - goto out; - } - } - - free_extent_state(cached_state); - cached_state = NULL; - - ret = __blockdev_direct_IO(rw, iocb, inode, + return __blockdev_direct_IO(rw, iocb, inode, BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev, iov, offset, nr_segs, btrfs_get_blocks_direct, NULL, btrfs_submit_direct, 0); - - if (ret < 0 && ret != -EIOCBQUEUED) { - clear_extent_bit(&BTRFS_I(inode)->io_tree, offset, - offset + iov_length(iov, nr_segs) - 1, - EXTENT_LOCKED | write_bits, 1, 0, - &cached_state, GFP_NOFS); - } else if (ret >= 0 && ret < iov_length(iov, nr_segs)) { - /* - * We're falling back to buffered, unlock the section we didn't - * do IO on. - */ - clear_extent_bit(&BTRFS_I(inode)->io_tree, offset + ret, - offset + iov_length(iov, nr_segs) - 1, - EXTENT_LOCKED | write_bits, 1, 0, - &cached_state, GFP_NOFS); - } -out: - free_extent_state(cached_state); - return ret; } static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, @@ -6289,7 +6481,7 @@ int btrfs_readpage(struct file *file, struct page *page) { struct extent_io_tree *tree; tree = &BTRFS_I(page->mapping->host)->io_tree; - return extent_read_full_page(tree, page, btrfs_get_extent); + return extent_read_full_page(tree, page, btrfs_get_extent, 0); } static int btrfs_writepage(struct page *page, struct writeback_control *wbc) @@ -6350,13 +6542,13 @@ static int btrfs_releasepage(struct page *page, gfp_t gfp_flags) static void btrfs_invalidatepage(struct page *page, unsigned long offset) { + struct inode *inode = page->mapping->host; struct extent_io_tree *tree; struct btrfs_ordered_extent *ordered; struct extent_state *cached_state = NULL; u64 page_start = page_offset(page); u64 page_end = page_start + PAGE_CACHE_SIZE - 1; - /* * we have the page locked, so new writeback can't start, * and the dirty bit won't be cleared while we are here. @@ -6366,14 +6558,13 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset) */ wait_on_page_writeback(page); - tree = &BTRFS_I(page->mapping->host)->io_tree; + tree = &BTRFS_I(inode)->io_tree; if (offset) { btrfs_releasepage(page, GFP_NOFS); return; } - lock_extent_bits(tree, page_start, page_end, 0, &cached_state, - GFP_NOFS); - ordered = btrfs_lookup_ordered_extent(page->mapping->host, + lock_extent_bits(tree, page_start, page_end, 0, &cached_state); + ordered = btrfs_lookup_ordered_extent(inode, page_offset(page)); if (ordered) { /* @@ -6388,14 +6579,14 @@ static void btrfs_invalidatepage(struct page *page, unsigned long offset) * whoever cleared the private bit is responsible * for the finish_ordered_io */ - if (TestClearPagePrivate2(page)) { - btrfs_finish_ordered_io(page->mapping->host, - page_start, page_end); + if (TestClearPagePrivate2(page) && + btrfs_dec_test_ordered_pending(inode, &ordered, page_start, + PAGE_CACHE_SIZE, 1)) { + btrfs_finish_ordered_io(ordered); } btrfs_put_ordered_extent(ordered); cached_state = NULL; - lock_extent_bits(tree, page_start, page_end, 0, &cached_state, - GFP_NOFS); + lock_extent_bits(tree, page_start, page_end, 0, &cached_state); } clear_extent_bit(tree, page_start, page_end, EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC | @@ -6437,16 +6628,24 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) unsigned long zero_start; loff_t size; int ret; + int reserved = 0; u64 page_start; u64 page_end; + sb_start_pagefault(inode->i_sb); ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); + if (!ret) { + ret = file_update_time(vma->vm_file); + reserved = 1; + } if (ret) { if (ret == -ENOMEM) ret = VM_FAULT_OOM; else /* -ENOSPC, -EIO, etc */ ret = VM_FAULT_SIGBUS; - goto out; + if (reserved) + goto out; + goto out_noreserve; } ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */ @@ -6463,8 +6662,7 @@ again: } wait_on_page_writeback(page); - lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state, - GFP_NOFS); + lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state); set_page_extent_mapped(page); /* @@ -6524,11 +6722,15 @@ again: unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS); out_unlock: - if (!ret) + if (!ret) { + sb_end_pagefault(inode->i_sb); return VM_FAULT_LOCKED; + } unlock_page(page); - btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); out: + btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); +out_noreserve: + sb_end_pagefault(inode->i_sb); return ret; } @@ -6541,6 +6743,7 @@ static int btrfs_truncate(struct inode *inode) struct btrfs_trans_handle *trans; unsigned long nr; u64 mask = root->sectorsize - 1; + u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); ret = btrfs_truncate_page(inode->i_mapping, inode->i_size); if (ret) @@ -6588,19 +6791,23 @@ static int btrfs_truncate(struct inode *inode) rsv = btrfs_alloc_block_rsv(root); if (!rsv) return -ENOMEM; - btrfs_add_durable_block_rsv(root->fs_info, rsv); + rsv->size = min_size; + /* + * 1 for the truncate slack space + * 1 for the orphan item we're going to add + * 1 for the orphan item deletion + * 1 for updating the inode. + */ trans = btrfs_start_transaction(root, 4); if (IS_ERR(trans)) { err = PTR_ERR(trans); goto out; } - /* - * Reserve space for the truncate process. Truncate should be adding - * space, but if there are snapshots it may end up using space. - */ - ret = btrfs_truncate_reserve_metadata(trans, root, rsv); + /* Migrate the slack space for the truncate to our reserve */ + ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, + min_size); BUG_ON(ret); ret = btrfs_orphan_add(trans, inode); @@ -6609,21 +6816,6 @@ static int btrfs_truncate(struct inode *inode) goto out; } - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); - - /* - * Ok so we've already migrated our bytes over for the truncate, so here - * just reserve the one slot we need for updating the inode. - */ - trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) { - err = PTR_ERR(trans); - goto out; - } - trans->block_rsv = rsv; - /* * setattr is responsible for setting the ordered_data_close flag, * but that is only tested during the last file release. That @@ -6641,24 +6833,36 @@ static int btrfs_truncate(struct inode *inode) * using truncate to replace the contents of the file will * end up with a zero length file after a crash. */ - if (inode->i_size == 0 && BTRFS_I(inode)->ordered_data_close) + if (inode->i_size == 0 && test_bit(BTRFS_INODE_ORDERED_DATA_CLOSE, + &BTRFS_I(inode)->runtime_flags)) btrfs_add_ordered_operation(trans, root, inode); while (1) { + ret = btrfs_block_rsv_refill(root, rsv, min_size); + if (ret) { + /* + * This can only happen with the original transaction we + * started above, every other time we shouldn't have a + * transaction started yet. + */ + if (ret == -EAGAIN) + goto end_trans; + err = ret; + break; + } + if (!trans) { - trans = btrfs_start_transaction(root, 3); + /* Just need the 1 for updating the inode */ + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { - err = PTR_ERR(trans); - goto out; + ret = err = PTR_ERR(trans); + trans = NULL; + break; } - - ret = btrfs_truncate_reserve_metadata(trans, root, - rsv); - BUG_ON(ret); - - trans->block_rsv = rsv; } + trans->block_rsv = rsv; + ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, BTRFS_EXTENT_DATA_KEY); @@ -6673,7 +6877,7 @@ static int btrfs_truncate(struct inode *inode) err = ret; break; } - +end_trans: nr = trans->blocks_used; btrfs_end_transaction(trans, root); trans = NULL; @@ -6693,14 +6897,16 @@ static int btrfs_truncate(struct inode *inode) ret = btrfs_orphan_del(NULL, inode); } - trans->block_rsv = &root->fs_info->trans_block_rsv; - ret = btrfs_update_inode(trans, root, inode); - if (ret && !err) - err = ret; + if (trans) { + trans->block_rsv = &root->fs_info->trans_block_rsv; + ret = btrfs_update_inode(trans, root, inode); + if (ret && !err) + err = ret; - nr = trans->blocks_used; - ret = btrfs_end_transaction_throttle(trans, root); - btrfs_btree_balance_dirty(root, nr); + nr = trans->blocks_used; + ret = btrfs_end_transaction(trans, root); + btrfs_btree_balance_dirty(root, nr); + } out: btrfs_free_block_rsv(root, rsv); @@ -6721,21 +6927,22 @@ int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, int err; u64 index = 0; - inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, new_dirid, - new_dirid, S_IFDIR | 0700, &index); + inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, + new_dirid, new_dirid, + S_IFDIR | (~current_umask() & S_IRWXUGO), + &index); if (IS_ERR(inode)) return PTR_ERR(inode); inode->i_op = &btrfs_dir_inode_operations; inode->i_fop = &btrfs_dir_file_operations; - inode->i_nlink = 1; + set_nlink(inode, 1); btrfs_i_size_write(inode, 0); err = btrfs_update_inode(trans, new_root, inode); - BUG_ON(err); iput(inode); - return 0; + return err; } struct inode *btrfs_alloc_inode(struct super_block *sb) @@ -6748,16 +6955,14 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) return NULL; ei->root = NULL; - ei->space_info = NULL; ei->generation = 0; - ei->sequence = 0; ei->last_trans = 0; ei->last_sub_trans = 0; ei->logged_trans = 0; ei->delalloc_bytes = 0; - ei->reserved_bytes = 0; ei->disk_i_size = 0; ei->flags = 0; + ei->csum_bytes = 0; ei->index_cnt = (u64)-1; ei->last_unlink_trans = 0; @@ -6765,10 +6970,7 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) ei->outstanding_extents = 0; ei->reserved_extents = 0; - ei->ordered_data_close = 0; - ei->orphan_meta_reserved = 0; - ei->dummy_inode = 0; - ei->in_defrag = 0; + ei->runtime_flags = 0; ei->force_compress = BTRFS_COMPRESS_NONE; ei->delayed_node = NULL; @@ -6777,9 +6979,11 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) extent_map_tree_init(&ei->extent_tree); extent_io_tree_init(&ei->io_tree, &inode->i_data); extent_io_tree_init(&ei->io_failure_tree, &inode->i_data); + ei->io_tree.track_uptodate = 1; + ei->io_failure_tree.track_uptodate = 1; mutex_init(&ei->log_mutex); + mutex_init(&ei->delalloc_mutex); btrfs_ordered_inode_tree_init(&ei->ordered_tree); - INIT_LIST_HEAD(&ei->i_orphan); INIT_LIST_HEAD(&ei->delalloc_inodes); INIT_LIST_HEAD(&ei->ordered_operations); RB_CLEAR_NODE(&ei->rb_node); @@ -6790,7 +6994,6 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) static void btrfs_i_callback(struct rcu_head *head) { struct inode *inode = container_of(head, struct inode, i_rcu); - INIT_LIST_HEAD(&inode->i_dentry); kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode)); } @@ -6799,10 +7002,12 @@ void btrfs_destroy_inode(struct inode *inode) struct btrfs_ordered_extent *ordered; struct btrfs_root *root = BTRFS_I(inode)->root; - WARN_ON(!list_empty(&inode->i_dentry)); + WARN_ON(!hlist_empty(&inode->i_dentry)); WARN_ON(inode->i_data.nrpages); WARN_ON(BTRFS_I(inode)->outstanding_extents); WARN_ON(BTRFS_I(inode)->reserved_extents); + WARN_ON(BTRFS_I(inode)->delalloc_bytes); + WARN_ON(BTRFS_I(inode)->csum_bytes); /* * This can happen where we create an inode, but somebody else also @@ -6823,13 +7028,12 @@ void btrfs_destroy_inode(struct inode *inode) spin_unlock(&root->fs_info->ordered_extent_lock); } - spin_lock(&root->orphan_lock); - if (!list_empty(&BTRFS_I(inode)->i_orphan)) { + if (test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, + &BTRFS_I(inode)->runtime_flags)) { printk(KERN_INFO "BTRFS: inode %llu still on the orphan list\n", (unsigned long long)btrfs_ino(inode)); - list_del_init(&BTRFS_I(inode)->i_orphan); + atomic_dec(&root->orphan_inodes); } - spin_unlock(&root->orphan_lock); while (1) { ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1); @@ -6857,7 +7061,7 @@ int btrfs_drop_inode(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; if (btrfs_root_refs(&root->root_item) == 0 && - !btrfs_is_free_space_inode(root, inode)) + !btrfs_is_free_space_inode(inode)) return 1; else return generic_drop_inode(inode); @@ -6926,11 +7130,13 @@ static int btrfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) { struct inode *inode = dentry->d_inode; + u32 blocksize = inode->i_sb->s_blocksize; + generic_fillattr(inode, stat); stat->dev = BTRFS_I(inode)->root->anon_dev; stat->blksize = PAGE_CACHE_SIZE; - stat->blocks = (inode_get_bytes(inode) + - BTRFS_I(inode)->delalloc_bytes) >> 9; + stat->blocks = (ALIGN(inode_get_bytes(inode), blocksize) + + ALIGN(BTRFS_I(inode)->delalloc_bytes, blocksize)) >> 9; return 0; } @@ -6948,10 +7154,13 @@ static void fixup_inode_flags(struct inode *dir, struct inode *inode) else b_inode->flags &= ~BTRFS_INODE_NODATACOW; - if (b_dir->flags & BTRFS_INODE_COMPRESS) + if (b_dir->flags & BTRFS_INODE_COMPRESS) { b_inode->flags |= BTRFS_INODE_COMPRESS; - else - b_inode->flags &= ~BTRFS_INODE_COMPRESS; + b_inode->flags &= ~BTRFS_INODE_NOCOMPRESS; + } else { + b_inode->flags &= ~(BTRFS_INODE_COMPRESS | + BTRFS_INODE_NOCOMPRESS); + } } static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, @@ -7042,6 +7251,9 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (new_inode && new_inode->i_size && S_ISREG(old_inode->i_mode)) btrfs_add_ordered_operation(trans, root, old_inode); + inode_inc_iversion(old_dir); + inode_inc_iversion(new_dir); + inode_inc_iversion(old_inode); old_dir->i_ctime = old_dir->i_mtime = ctime; new_dir->i_ctime = new_dir->i_mtime = ctime; old_inode->i_ctime = ctime; @@ -7062,9 +7274,13 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (!ret) ret = btrfs_update_inode(trans, root, old_inode); } - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } if (new_inode) { + inode_inc_iversion(new_inode); new_inode->i_ctime = CURRENT_TIME; if (unlikely(btrfs_ino(new_inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { @@ -7080,11 +7296,14 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, new_dentry->d_name.name, new_dentry->d_name.len); } - BUG_ON(ret); - if (new_inode->i_nlink == 0) { + if (!ret && new_inode->i_nlink == 0) { ret = btrfs_orphan_add(trans, new_dentry->d_inode); BUG_ON(ret); } + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } } fixup_inode_flags(new_dir, old_inode); @@ -7092,7 +7311,10 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, ret = btrfs_add_link(trans, new_dir, old_inode, new_dentry->d_name.name, new_dentry->d_name.len, 0, index); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto out_fail; + } if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { struct dentry *parent = new_dentry->d_parent; @@ -7100,7 +7322,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, btrfs_end_log_trans(root); } out_fail: - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); out_notrans: if (old_ino == BTRFS_FIRST_FREE_OBJECTID) up_read(&root->fs_info->subvol_sem); @@ -7206,14 +7428,21 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, goto out_unlock; } + /* + * If the active LSM wants to access the inode during + * d_instantiate it needs these. Smack checks to see + * if the filesystem supports xattrs by looking at the + * ops vector. + */ + inode->i_fop = &btrfs_file_operations; + inode->i_op = &btrfs_file_inode_operations; + err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; else { inode->i_mapping->a_ops = &btrfs_aops; inode->i_mapping->backing_dev_info = &root->fs_info->bdi; - inode->i_fop = &btrfs_file_operations; - inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } if (drop_inode) @@ -7262,8 +7491,10 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, drop_inode = 1; out_unlock: + if (!err) + d_instantiate(dentry, inode); nr = trans->blocks_used; - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); if (drop_inode) { inode_dec_link_count(inode); iput(inode); @@ -7296,7 +7527,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode, } ret = btrfs_reserve_extent(trans, root, num_bytes, min_size, - 0, *alloc_hint, (u64)-1, &ins, 1); + 0, *alloc_hint, &ins, 1); if (ret) { if (own_trans) btrfs_end_transaction(trans, root); @@ -7308,7 +7539,12 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode, ins.offset, ins.offset, ins.offset, 0, 0, 0, BTRFS_FILE_EXTENT_PREALLOC); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + if (own_trans) + btrfs_end_transaction(trans, root); + break; + } btrfs_drop_extent_cache(inode, cur_offset, cur_offset + ins.offset -1, 0); @@ -7316,6 +7552,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode, cur_offset += ins.offset; *alloc_hint = ins.objectid + ins.offset; + inode_inc_iversion(inode); inode->i_ctime = CURRENT_TIME; BTRFS_I(inode)->flags |= BTRFS_INODE_PREALLOC; if (!(mode & FALLOC_FL_KEEP_SIZE) && @@ -7330,7 +7567,13 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode, } ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + + if (ret) { + btrfs_abort_transaction(trans, root, ret); + if (own_trans) + btrfs_end_transaction(trans, root); + break; + } if (own_trans) btrfs_end_transaction(trans, root); @@ -7420,7 +7663,6 @@ static struct extent_io_ops btrfs_extent_io_ops = { .readpage_end_io_hook = btrfs_readpage_end_io_hook, .writepage_end_io_hook = btrfs_writepage_end_io_hook, .writepage_start_hook = btrfs_writepage_start_hook, - .readpage_io_failed_hook = btrfs_io_failed_hook, .set_bit_hook = btrfs_set_bit_hook, .clear_bit_hook = btrfs_clear_bit_hook, .merge_extent_hook = btrfs_merge_extent_hook, @@ -7468,6 +7710,7 @@ static const struct inode_operations btrfs_file_inode_operations = { .permission = btrfs_permission, .fiemap = btrfs_fiemap, .get_acl = btrfs_get_acl, + .update_time = btrfs_update_time, }; static const struct inode_operations btrfs_special_inode_operations = { .getattr = btrfs_getattr, @@ -7478,18 +7721,21 @@ static const struct inode_operations btrfs_special_inode_operations = { .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, .get_acl = btrfs_get_acl, + .update_time = btrfs_update_time, }; static const struct inode_operations btrfs_symlink_inode_operations = { .readlink = generic_readlink, .follow_link = page_follow_link_light, .put_link = page_put_link, .getattr = btrfs_getattr, + .setattr = btrfs_setattr, .permission = btrfs_permission, .setxattr = btrfs_setxattr, .getxattr = btrfs_getxattr, .listxattr = btrfs_listxattr, .removexattr = btrfs_removexattr, .get_acl = btrfs_get_acl, + .update_time = btrfs_update_time, }; const struct dentry_operations btrfs_dentry_operations = { diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index dae5dfe41ba5..9df50fa8a078 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -41,6 +41,7 @@ #include <linux/vmalloc.h> #include <linux/slab.h> #include <linux/blkdev.h> +#include <linux/uuid.h> #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -51,6 +52,9 @@ #include "volumes.h" #include "locking.h" #include "inode-map.h" +#include "backref.h" +#include "rcu-string.h" +#include "send.h" /* Mask out flags that are inappropriate for the given type of inode. */ static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags) @@ -117,7 +121,7 @@ void btrfs_update_iflags(struct inode *inode) /* * Inherit flags from the parent inode. * - * Unlike extN we don't have any flags we don't want to inherit currently. + * Currently only the compression flags and the cow flags are inherited. */ void btrfs_inherit_iflags(struct inode *inode, struct inode *dir) { @@ -128,12 +132,17 @@ void btrfs_inherit_iflags(struct inode *inode, struct inode *dir) flags = BTRFS_I(dir)->flags; - if (S_ISREG(inode->i_mode)) - flags &= ~BTRFS_INODE_DIRSYNC; - else if (!S_ISDIR(inode->i_mode)) - flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME); + if (flags & BTRFS_INODE_NOCOMPRESS) { + BTRFS_I(inode)->flags &= ~BTRFS_INODE_COMPRESS; + BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS; + } else if (flags & BTRFS_INODE_COMPRESS) { + BTRFS_I(inode)->flags &= ~BTRFS_INODE_NOCOMPRESS; + BTRFS_I(inode)->flags |= BTRFS_INODE_COMPRESS; + } + + if (flags & BTRFS_INODE_NODATACOW) + BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW; - BTRFS_I(inode)->flags = flags; btrfs_update_iflags(inode); } @@ -170,6 +179,8 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) struct btrfs_trans_handle *trans; unsigned int flags, oldflags; int ret; + u64 ip_oldflags; + unsigned int i_oldflags; if (btrfs_root_readonly(root)) return -EROFS; @@ -184,8 +195,15 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) if (!inode_owner_or_capable(inode)) return -EACCES; + ret = mnt_want_write_file(file); + if (ret) + return ret; + mutex_lock(&inode->i_mutex); + ip_oldflags = ip->flags; + i_oldflags = inode->i_flags; + flags = btrfs_mask_flags(inode->i_mode, flags); oldflags = btrfs_flags_to_ioctl(ip->flags); if ((flags ^ oldflags) & (FS_APPEND_FL | FS_IMMUTABLE_FL)) { @@ -195,10 +213,6 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) } } - ret = mnt_want_write(file->f_path.mnt); - if (ret) - goto out_unlock; - if (flags & FS_SYNC_FL) ip->flags |= BTRFS_INODE_SYNC; else @@ -243,21 +257,27 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) ip->flags &= ~(BTRFS_INODE_COMPRESS | BTRFS_INODE_NOCOMPRESS); } - trans = btrfs_join_transaction(root); - BUG_ON(IS_ERR(trans)); - - ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + trans = btrfs_start_transaction(root, 1); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out_drop; + } btrfs_update_iflags(inode); + inode_inc_iversion(inode); inode->i_ctime = CURRENT_TIME; - btrfs_end_transaction(trans, root); + ret = btrfs_update_inode(trans, root, inode); - mnt_drop_write(file->f_path.mnt); + btrfs_end_transaction(trans, root); + out_drop: + if (ret) { + ip->flags = ip_oldflags; + inode->i_flags = i_oldflags; + } - ret = 0; out_unlock: mutex_unlock(&inode->i_mutex); + mnt_drop_write_file(file); return ret; } @@ -270,13 +290,13 @@ static int btrfs_ioctl_getversion(struct file *file, int __user *arg) static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) { - struct btrfs_root *root = fdentry(file)->d_sb->s_fs_info; - struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_fs_info *fs_info = btrfs_sb(fdentry(file)->d_sb); struct btrfs_device *device; struct request_queue *q; struct fstrim_range range; u64 minlen = ULLONG_MAX; u64 num_devices = 0; + u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy); int ret; if (!capable(CAP_SYS_ADMIN)) @@ -295,14 +315,17 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) } } rcu_read_unlock(); + if (!num_devices) return -EOPNOTSUPP; - if (copy_from_user(&range, arg, sizeof(range))) return -EFAULT; + if (range.start > total_bytes) + return -EINVAL; + range.len = min(range.len, total_bytes - range.start); range.minlen = max(range.minlen, minlen); - ret = btrfs_trim_fs(root, &range); + ret = btrfs_trim_fs(fs_info->tree_root, &range); if (ret < 0) return ret; @@ -315,7 +338,8 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) static noinline int create_subvol(struct btrfs_root *root, struct dentry *dentry, char *name, int namelen, - u64 *async_transid) + u64 *async_transid, + struct btrfs_qgroup_inherit **inherit) { struct btrfs_trans_handle *trans; struct btrfs_key key; @@ -325,11 +349,13 @@ static noinline int create_subvol(struct btrfs_root *root, struct btrfs_root *new_root; struct dentry *parent = dentry->d_parent; struct inode *dir; + struct timespec cur_time = CURRENT_TIME; int ret; int err; u64 objectid; u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID; u64 index = 0; + uuid_le new_uuid; ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid); if (ret) @@ -347,6 +373,11 @@ static noinline int create_subvol(struct btrfs_root *root, if (IS_ERR(trans)) return PTR_ERR(trans); + ret = btrfs_qgroup_inherit(trans, root->fs_info, 0, objectid, + inherit ? *inherit : NULL); + if (ret) + goto fail; + leaf = btrfs_alloc_free_block(trans, root, root->leafsize, 0, objectid, NULL, 0, 0, 0); if (IS_ERR(leaf)) { @@ -368,8 +399,9 @@ static noinline int create_subvol(struct btrfs_root *root, BTRFS_UUID_SIZE); btrfs_mark_buffer_dirty(leaf); + memset(&root_item, 0, sizeof(root_item)); + inode_item = &root_item.inode; - memset(inode_item, 0, sizeof(*inode_item)); inode_item->generation = cpu_to_le64(1); inode_item->size = cpu_to_le64(3); inode_item->nlink = cpu_to_le32(1); @@ -387,8 +419,15 @@ static noinline int create_subvol(struct btrfs_root *root, btrfs_set_root_used(&root_item, leaf->len); btrfs_set_root_last_snapshot(&root_item, 0); - memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress)); - root_item.drop_level = 0; + btrfs_set_root_generation_v2(&root_item, + btrfs_root_generation(&root_item)); + uuid_le_gen(&new_uuid); + memcpy(root_item.uuid, new_uuid.b, BTRFS_UUID_SIZE); + root_item.otime.sec = cpu_to_le64(cur_time.tv_sec); + root_item.otime.nsec = cpu_to_le32(cur_time.tv_nsec); + root_item.ctime = root_item.otime; + btrfs_set_root_ctransid(&root_item, trans->transid); + btrfs_set_root_otransid(&root_item, trans->transid); btrfs_tree_unlock(leaf); free_extent_buffer(leaf); @@ -406,22 +445,37 @@ static noinline int create_subvol(struct btrfs_root *root, key.offset = (u64)-1; new_root = btrfs_read_fs_root_no_name(root->fs_info, &key); - BUG_ON(IS_ERR(new_root)); + if (IS_ERR(new_root)) { + btrfs_abort_transaction(trans, root, PTR_ERR(new_root)); + ret = PTR_ERR(new_root); + goto fail; + } btrfs_record_root_in_trans(trans, new_root); ret = btrfs_create_subvol_root(trans, new_root, new_dirid); + if (ret) { + /* We potentially lose an unused inode item here */ + btrfs_abort_transaction(trans, root, ret); + goto fail; + } + /* * insert the directory item */ ret = btrfs_set_inode_index(dir, &index); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + goto fail; + } ret = btrfs_insert_dir_item(trans, root, name, namelen, dir, &key, BTRFS_FT_DIR, index); - if (ret) + if (ret) { + btrfs_abort_transaction(trans, root, ret); goto fail; + } btrfs_i_size_write(dir, dir->i_size + namelen * 2); ret = btrfs_update_inode(trans, root, dir); @@ -448,7 +502,7 @@ fail: static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, char *name, int namelen, u64 *async_transid, - bool readonly) + bool readonly, struct btrfs_qgroup_inherit **inherit) { struct inode *inode; struct btrfs_pending_snapshot *pending_snapshot; @@ -466,6 +520,10 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, pending_snapshot->dentry = dentry; pending_snapshot->root = root; pending_snapshot->readonly = readonly; + if (inherit) { + pending_snapshot->inherit = *inherit; + *inherit = NULL; /* take responsibility to free it */ + } trans = btrfs_start_transaction(root->fs_info->extent_root, 5); if (IS_ERR(trans)) { @@ -599,7 +657,8 @@ static inline int btrfs_may_create(struct inode *dir, struct dentry *child) static noinline int btrfs_mksubvol(struct path *parent, char *name, int namelen, struct btrfs_root *snap_src, - u64 *async_transid, bool readonly) + u64 *async_transid, bool readonly, + struct btrfs_qgroup_inherit **inherit) { struct inode *dir = parent->dentry->d_inode; struct dentry *dentry; @@ -616,13 +675,9 @@ static noinline int btrfs_mksubvol(struct path *parent, if (dentry->d_inode) goto out_dput; - error = mnt_want_write(parent->mnt); - if (error) - goto out_dput; - error = btrfs_may_create(dir, dentry); if (error) - goto out_drop_write; + goto out_dput; down_read(&BTRFS_I(dir)->root->fs_info->subvol_sem); @@ -630,18 +685,16 @@ static noinline int btrfs_mksubvol(struct path *parent, goto out_up_read; if (snap_src) { - error = create_snapshot(snap_src, dentry, - name, namelen, async_transid, readonly); + error = create_snapshot(snap_src, dentry, name, namelen, + async_transid, readonly, inherit); } else { error = create_subvol(BTRFS_I(dir)->root, dentry, - name, namelen, async_transid); + name, namelen, async_transid, inherit); } if (!error) fsnotify_mkdir(dir, dentry); out_up_read: up_read(&BTRFS_I(dir)->root->fs_info->subvol_sem); -out_drop_write: - mnt_drop_write(parent->mnt); out_dput: dput(dentry); out_unlock: @@ -750,23 +803,12 @@ none: return -ENOENT; } -static int should_defrag_range(struct inode *inode, u64 start, u64 len, - int thresh, u64 *last_len, u64 *skip, - u64 *defrag_end) +static struct extent_map *defrag_lookup_extent(struct inode *inode, u64 start) { - struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; - struct extent_map *em = NULL; struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; - int ret = 1; - - /* - * make sure that once we start defragging and extent, we keep on - * defragging it - */ - if (start < *defrag_end) - return 1; - - *skip = 0; + struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; + struct extent_map *em; + u64 len = PAGE_CACHE_SIZE; /* * hopefully we have this extent in the tree already, try without @@ -778,24 +820,71 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, if (!em) { /* get the big lock and read metadata off disk */ - lock_extent(io_tree, start, start + len - 1, GFP_NOFS); + lock_extent(io_tree, start, start + len - 1); em = btrfs_get_extent(inode, NULL, 0, start, len, 0); - unlock_extent(io_tree, start, start + len - 1, GFP_NOFS); + unlock_extent(io_tree, start, start + len - 1); if (IS_ERR(em)) - return 0; + return NULL; } + return em; +} + +static bool defrag_check_next_extent(struct inode *inode, struct extent_map *em) +{ + struct extent_map *next; + bool ret = true; + + /* this is the last extent */ + if (em->start + em->len >= i_size_read(inode)) + return false; + + next = defrag_lookup_extent(inode, em->start + em->len); + if (!next || next->block_start >= EXTENT_MAP_LAST_BYTE) + ret = false; + + free_extent_map(next); + return ret; +} + +static int should_defrag_range(struct inode *inode, u64 start, int thresh, + u64 *last_len, u64 *skip, u64 *defrag_end, + int compress) +{ + struct extent_map *em; + int ret = 1; + bool next_mergeable = true; + + /* + * make sure that once we start defragging an extent, we keep on + * defragging it + */ + if (start < *defrag_end) + return 1; + + *skip = 0; + + em = defrag_lookup_extent(inode, start); + if (!em) + return 0; + /* this will cover holes, and inline extents */ - if (em->block_start >= EXTENT_MAP_LAST_BYTE) + if (em->block_start >= EXTENT_MAP_LAST_BYTE) { ret = 0; + goto out; + } + + next_mergeable = defrag_check_next_extent(inode, em); /* - * we hit a real extent, if it is big don't bother defragging it again + * we hit a real extent, if it is big or the next extent is not a + * real extent, don't bother defragging it */ - if ((*last_len == 0 || *last_len >= thresh) && em->len >= thresh) + if (!compress && (*last_len == 0 || *last_len >= thresh) && + (em->len >= thresh || !next_mergeable)) ret = 0; - +out: /* * last_len ends up being a counter of how many bytes we've defragged. * every time we choose not to defrag an extent, we reset *last_len @@ -805,7 +894,6 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, * extent will force at least part of that big extent to be defragged. */ if (ret) { - *last_len += len; *defrag_end = extent_map_end(em); } else { *last_len = 0; @@ -838,32 +926,62 @@ static int cluster_pages_for_defrag(struct inode *inode, u64 isize = i_size_read(inode); u64 page_start; u64 page_end; + u64 page_cnt; int ret; int i; int i_done; struct btrfs_ordered_extent *ordered; struct extent_state *cached_state = NULL; + struct extent_io_tree *tree; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); - if (isize == 0) - return 0; file_end = (isize - 1) >> PAGE_CACHE_SHIFT; + if (!isize || start_index > file_end) + return 0; + + page_cnt = min_t(u64, (u64)num_pages, (u64)file_end - start_index + 1); ret = btrfs_delalloc_reserve_space(inode, - num_pages << PAGE_CACHE_SHIFT); + page_cnt << PAGE_CACHE_SHIFT); if (ret) return ret; -again: - ret = 0; i_done = 0; + tree = &BTRFS_I(inode)->io_tree; /* step one, lock all the pages */ - for (i = 0; i < num_pages; i++) { + for (i = 0; i < page_cnt; i++) { struct page *page; +again: page = find_or_create_page(inode->i_mapping, - start_index + i, GFP_NOFS); + start_index + i, mask); if (!page) break; + page_start = page_offset(page); + page_end = page_start + PAGE_CACHE_SIZE - 1; + while (1) { + lock_extent(tree, page_start, page_end); + ordered = btrfs_lookup_ordered_extent(inode, + page_start); + unlock_extent(tree, page_start, page_end); + if (!ordered) + break; + + unlock_page(page); + btrfs_start_ordered_extent(inode, ordered, 1); + btrfs_put_ordered_extent(ordered); + lock_page(page); + /* + * we unlocked the page above, so we need check if + * it was released or not. + */ + if (page->mapping != inode->i_mapping) { + unlock_page(page); + page_cache_release(page); + goto again; + } + } + if (!PageUptodate(page)) { btrfs_readpage(NULL, page); lock_page(page); @@ -874,15 +992,13 @@ again: break; } } - isize = i_size_read(inode); - file_end = (isize - 1) >> PAGE_CACHE_SHIFT; - if (!isize || page->index > file_end || - page->mapping != inode->i_mapping) { - /* whoops, we blew past eof, skip this page */ + + if (page->mapping != inode->i_mapping) { unlock_page(page); page_cache_release(page); - break; + goto again; } + pages[i] = page; i_done++; } @@ -903,38 +1019,18 @@ again: page_end = page_offset(pages[i_done - 1]) + PAGE_CACHE_SIZE; lock_extent_bits(&BTRFS_I(inode)->io_tree, - page_start, page_end - 1, 0, &cached_state, - GFP_NOFS); - ordered = btrfs_lookup_first_ordered_extent(inode, page_end - 1); - if (ordered && - ordered->file_offset + ordered->len > page_start && - ordered->file_offset < page_end) { - btrfs_put_ordered_extent(ordered); - unlock_extent_cached(&BTRFS_I(inode)->io_tree, - page_start, page_end - 1, - &cached_state, GFP_NOFS); - for (i = 0; i < i_done; i++) { - unlock_page(pages[i]); - page_cache_release(pages[i]); - } - btrfs_wait_ordered_range(inode, page_start, - page_end - page_start); - goto again; - } - if (ordered) - btrfs_put_ordered_extent(ordered); - + page_start, page_end - 1, 0, &cached_state); clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end - 1, EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING, 0, 0, &cached_state, GFP_NOFS); - if (i_done != num_pages) { + if (i_done != page_cnt) { spin_lock(&BTRFS_I(inode)->lock); BTRFS_I(inode)->outstanding_extents++; spin_unlock(&BTRFS_I(inode)->lock); btrfs_delalloc_release_space(inode, - (num_pages - i_done) << PAGE_CACHE_SHIFT); + (page_cnt - i_done) << PAGE_CACHE_SHIFT); } @@ -959,7 +1055,7 @@ out: unlock_page(pages[i]); page_cache_release(pages[i]); } - btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT); + btrfs_delalloc_release_space(inode, page_cnt << PAGE_CACHE_SHIFT); return ret; } @@ -969,21 +1065,21 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, u64 newer_than, unsigned long max_to_defrag) { struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_super_block *disk_super; struct file_ra_state *ra = NULL; unsigned long last_index; - u64 features; + u64 isize = i_size_read(inode); u64 last_len = 0; u64 skip = 0; u64 defrag_end = 0; u64 newer_off = range->start; - int newer_left = 0; unsigned long i; + unsigned long ra_index = 0; int ret; int defrag_count = 0; int compress_type = BTRFS_COMPRESS_ZLIB; int extent_thresh = range->extent_thresh; - int newer_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT; + int max_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT; + int cluster = max_cluster; u64 new_align = ~((u64)128 * 1024 - 1); struct page **pages = NULL; @@ -997,7 +1093,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, compress_type = range->compress_type; } - if (inode->i_size == 0) + if (isize == 0) return 0; /* @@ -1013,7 +1109,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, ra = &file->f_ra; } - pages = kmalloc(sizeof(struct page *) * newer_cluster, + pages = kmalloc(sizeof(struct page *) * max_cluster, GFP_NOFS); if (!pages) { ret = -ENOMEM; @@ -1022,10 +1118,10 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, /* find the last page to defrag */ if (range->start + range->len > range->start) { - last_index = min_t(u64, inode->i_size - 1, + last_index = min_t(u64, isize - 1, range->start + range->len - 1) >> PAGE_CACHE_SHIFT; } else { - last_index = (inode->i_size - 1) >> PAGE_CACHE_SHIFT; + last_index = (isize - 1) >> PAGE_CACHE_SHIFT; } if (newer_than) { @@ -1038,14 +1134,13 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, * the extents in the file evenly spaced */ i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; - newer_left = newer_cluster; } else goto out_ra; } else { i = range->start >> PAGE_CACHE_SHIFT; } if (!max_to_defrag) - max_to_defrag = last_index - 1; + max_to_defrag = last_index + 1; /* * make writeback starts from i, so the defrag range can be @@ -1064,12 +1159,10 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, if (!(inode->i_sb->s_flags & MS_ACTIVE)) break; - if (!newer_than && - !should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT, - PAGE_CACHE_SIZE, - extent_thresh, - &last_len, &skip, - &defrag_end)) { + if (!should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT, + extent_thresh, &last_len, &skip, + &defrag_end, range->flags & + BTRFS_DEFRAG_RANGE_COMPRESS)) { unsigned long next; /* * the should_defrag function tells us how much to skip @@ -1079,23 +1172,43 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, i = max(i + 1, next); continue; } + + if (!newer_than) { + cluster = (PAGE_CACHE_ALIGN(defrag_end) >> + PAGE_CACHE_SHIFT) - i; + cluster = min(cluster, max_cluster); + } else { + cluster = max_cluster; + } + if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) BTRFS_I(inode)->force_compress = compress_type; - btrfs_force_ra(inode->i_mapping, ra, file, i, newer_cluster); + if (i + cluster > ra_index) { + ra_index = max(i, ra_index); + btrfs_force_ra(inode->i_mapping, ra, file, ra_index, + cluster); + ra_index += max_cluster; + } - ret = cluster_pages_for_defrag(inode, pages, i, newer_cluster); - if (ret < 0) + mutex_lock(&inode->i_mutex); + ret = cluster_pages_for_defrag(inode, pages, i, cluster); + if (ret < 0) { + mutex_unlock(&inode->i_mutex); goto out_ra; + } defrag_count += ret; balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret); - i += ret; + mutex_unlock(&inode->i_mutex); if (newer_than) { if (newer_off == (u64)-1) break; + if (ret > 0) + i += ret; + newer_off = max(newer_off + 1, (u64)i << PAGE_CACHE_SHIFT); @@ -1105,12 +1218,17 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, if (!ret) { range->start = newer_off; i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; - newer_left = newer_cluster; } else { break; } } else { - i++; + if (ret > 0) { + i += ret; + last_len += ret << PAGE_CACHE_SHIFT; + } else { + i++; + last_len = 0; + } } } @@ -1136,16 +1254,11 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, mutex_unlock(&inode->i_mutex); } - disk_super = &root->fs_info->super_copy; - features = btrfs_super_incompat_flags(disk_super); if (range->compress_type == BTRFS_COMPRESS_LZO) { - features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO; - btrfs_set_super_incompat_flags(disk_super, features); + btrfs_set_fs_incompat(root->fs_info, COMPRESS_LZO); } - if (!file) - kfree(ra); - return defrag_count; + ret = defrag_count; out_ra: if (!file) @@ -1174,13 +1287,21 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, if (!capable(CAP_SYS_ADMIN)) return -EPERM; + mutex_lock(&root->fs_info->volume_mutex); + if (root->fs_info->balance_ctl) { + printk(KERN_INFO "btrfs: balance in progress\n"); + ret = -EINVAL; + goto out; + } + vol_args = memdup_user(arg, sizeof(*vol_args)); - if (IS_ERR(vol_args)) - return PTR_ERR(vol_args); + if (IS_ERR(vol_args)) { + ret = PTR_ERR(vol_args); + goto out; + } vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; - mutex_lock(&root->fs_info->volume_mutex); sizestr = vol_args->name; devstr = strchr(sizestr, ':'); if (devstr) { @@ -1189,16 +1310,24 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, *devstr = '\0'; devstr = vol_args->name; devid = simple_strtoull(devstr, &end, 10); - printk(KERN_INFO "resizing devid %llu\n", + printk(KERN_INFO "btrfs: resizing devid %llu\n", (unsigned long long)devid); } device = btrfs_find_device(root, devid, NULL, NULL); if (!device) { - printk(KERN_INFO "resizer unable to find device %llu\n", + printk(KERN_INFO "btrfs: resizer unable to find device %llu\n", (unsigned long long)devid); ret = -EINVAL; - goto out_unlock; + goto out_free; } + if (device->fs_devices && device->fs_devices->seeding) { + printk(KERN_INFO "btrfs: resizer unable to apply on " + "seeding device %llu\n", + (unsigned long long)devid); + ret = -EINVAL; + goto out_free; + } + if (!strcmp(sizestr, "max")) new_size = device->bdev->bd_inode->i_size; else { @@ -1212,7 +1341,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, new_size = memparse(sizestr, NULL); if (new_size == 0) { ret = -EINVAL; - goto out_unlock; + goto out_free; } } @@ -1221,7 +1350,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, if (mod < 0) { if (new_size > old_size) { ret = -EINVAL; - goto out_unlock; + goto out_free; } new_size = old_size - new_size; } else if (mod > 0) { @@ -1230,67 +1359,73 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, if (new_size < 256 * 1024 * 1024) { ret = -EINVAL; - goto out_unlock; + goto out_free; } if (new_size > device->bdev->bd_inode->i_size) { ret = -EFBIG; - goto out_unlock; + goto out_free; } do_div(new_size, root->sectorsize); new_size *= root->sectorsize; - printk(KERN_INFO "new size for %s is %llu\n", - device->name, (unsigned long long)new_size); + printk_in_rcu(KERN_INFO "btrfs: new size for %s is %llu\n", + rcu_str_deref(device->name), + (unsigned long long)new_size); if (new_size > old_size) { trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) { ret = PTR_ERR(trans); - goto out_unlock; + goto out_free; } ret = btrfs_grow_device(trans, device, new_size); btrfs_commit_transaction(trans, root); - } else { + } else if (new_size < old_size) { ret = btrfs_shrink_device(device, new_size); } -out_unlock: - mutex_unlock(&root->fs_info->volume_mutex); +out_free: kfree(vol_args); +out: + mutex_unlock(&root->fs_info->volume_mutex); return ret; } static noinline int btrfs_ioctl_snap_create_transid(struct file *file, - char *name, - unsigned long fd, - int subvol, - u64 *transid, - bool readonly) + char *name, unsigned long fd, int subvol, + u64 *transid, bool readonly, + struct btrfs_qgroup_inherit **inherit) { - struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root; struct file *src_file; int namelen; int ret = 0; - if (root->fs_info->sb->s_flags & MS_RDONLY) - return -EROFS; + ret = mnt_want_write_file(file); + if (ret) + goto out; namelen = strlen(name); if (strchr(name, '/')) { ret = -EINVAL; - goto out; + goto out_drop_write; + } + + if (name[0] == '.' && + (namelen == 1 || (name[1] == '.' && namelen == 2))) { + ret = -EEXIST; + goto out_drop_write; } if (subvol) { ret = btrfs_mksubvol(&file->f_path, name, namelen, - NULL, transid, readonly); + NULL, transid, readonly, inherit); } else { struct inode *src_inode; src_file = fget(fd); if (!src_file) { ret = -EINVAL; - goto out; + goto out_drop_write; } src_inode = src_file->f_path.dentry->d_inode; @@ -1299,13 +1434,15 @@ static noinline int btrfs_ioctl_snap_create_transid(struct file *file, "another FS\n"); ret = -EINVAL; fput(src_file); - goto out; + goto out_drop_write; } ret = btrfs_mksubvol(&file->f_path, name, namelen, BTRFS_I(src_inode)->root, - transid, readonly); + transid, readonly, inherit); fput(src_file); } +out_drop_write: + mnt_drop_write_file(file); out: return ret; } @@ -1323,7 +1460,7 @@ static noinline int btrfs_ioctl_snap_create(struct file *file, ret = btrfs_ioctl_snap_create_transid(file, vol_args->name, vol_args->fd, subvol, - NULL, false); + NULL, false, NULL); kfree(vol_args); return ret; @@ -1337,6 +1474,7 @@ static noinline int btrfs_ioctl_snap_create_v2(struct file *file, u64 transid = 0; u64 *ptr = NULL; bool readonly = false; + struct btrfs_qgroup_inherit *inherit = NULL; vol_args = memdup_user(arg, sizeof(*vol_args)); if (IS_ERR(vol_args)) @@ -1344,7 +1482,8 @@ static noinline int btrfs_ioctl_snap_create_v2(struct file *file, vol_args->name[BTRFS_SUBVOL_NAME_MAX] = '\0'; if (vol_args->flags & - ~(BTRFS_SUBVOL_CREATE_ASYNC | BTRFS_SUBVOL_RDONLY)) { + ~(BTRFS_SUBVOL_CREATE_ASYNC | BTRFS_SUBVOL_RDONLY | + BTRFS_SUBVOL_QGROUP_INHERIT)) { ret = -EOPNOTSUPP; goto out; } @@ -1353,10 +1492,21 @@ static noinline int btrfs_ioctl_snap_create_v2(struct file *file, ptr = &transid; if (vol_args->flags & BTRFS_SUBVOL_RDONLY) readonly = true; + if (vol_args->flags & BTRFS_SUBVOL_QGROUP_INHERIT) { + if (vol_args->size > PAGE_CACHE_SIZE) { + ret = -EINVAL; + goto out; + } + inherit = memdup_user(vol_args->qgroup_inherit, vol_args->size); + if (IS_ERR(inherit)) { + ret = PTR_ERR(inherit); + goto out; + } + } ret = btrfs_ioctl_snap_create_transid(file, vol_args->name, - vol_args->fd, subvol, - ptr, readonly); + vol_args->fd, subvol, ptr, + readonly, &inherit); if (ret == 0 && ptr && copy_to_user(arg + @@ -1365,6 +1515,7 @@ static noinline int btrfs_ioctl_snap_create_v2(struct file *file, ret = -EFAULT; out: kfree(vol_args); + kfree(inherit); return ret; } @@ -1400,29 +1551,40 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file, u64 flags; int ret = 0; - if (root->fs_info->sb->s_flags & MS_RDONLY) - return -EROFS; + ret = mnt_want_write_file(file); + if (ret) + goto out; - if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) - return -EINVAL; + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) { + ret = -EINVAL; + goto out_drop_write; + } - if (copy_from_user(&flags, arg, sizeof(flags))) - return -EFAULT; + if (copy_from_user(&flags, arg, sizeof(flags))) { + ret = -EFAULT; + goto out_drop_write; + } - if (flags & BTRFS_SUBVOL_CREATE_ASYNC) - return -EINVAL; + if (flags & BTRFS_SUBVOL_CREATE_ASYNC) { + ret = -EINVAL; + goto out_drop_write; + } - if (flags & ~BTRFS_SUBVOL_RDONLY) - return -EOPNOTSUPP; + if (flags & ~BTRFS_SUBVOL_RDONLY) { + ret = -EOPNOTSUPP; + goto out_drop_write; + } - if (!inode_owner_or_capable(inode)) - return -EACCES; + if (!inode_owner_or_capable(inode)) { + ret = -EACCES; + goto out_drop_write; + } down_write(&root->fs_info->subvol_sem); /* nothing to do */ if (!!(flags & BTRFS_SUBVOL_RDONLY) == btrfs_root_readonly(root)) - goto out; + goto out_drop_sem; root_flags = btrfs_root_flags(&root->root_item); if (flags & BTRFS_SUBVOL_RDONLY) @@ -1445,8 +1607,11 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file, out_reset: if (ret) btrfs_set_root_flags(&root->root_item, root_flags); -out: +out_drop_sem: up_write(&root->fs_info->subvol_sem); +out_drop_write: + mnt_drop_write_file(file); +out: return ret; } @@ -1826,7 +1991,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, goto out; } - err = mnt_want_write(file->f_path.mnt); + err = mnt_want_write_file(file); if (err) goto out; @@ -1910,7 +2075,11 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, dest->root_key.objectid, dentry->d_name.name, dentry->d_name.len); - BUG_ON(ret); + if (ret) { + err = ret; + btrfs_abort_transaction(trans, root, ret); + goto out_end_trans; + } btrfs_record_root_in_trans(trans, dest); @@ -1923,11 +2092,16 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root, dest->root_key.objectid); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + err = ret; + goto out_end_trans; + } } - +out_end_trans: ret = btrfs_end_transaction(trans, root); - BUG_ON(ret); + if (ret && !err) + err = ret; inode->i_flags |= S_DEAD; out_up_write: up_write(&root->fs_info->subvol_sem); @@ -1942,7 +2116,7 @@ out_dput: dput(dentry); out_unlock_dir: mutex_unlock(&dir->i_mutex); - mnt_drop_write(file->f_path.mnt); + mnt_drop_write_file(file); out: kfree(vol_args); return err; @@ -1958,7 +2132,7 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp) if (btrfs_root_readonly(root)) return -EROFS; - ret = mnt_want_write(file->f_path.mnt); + ret = mnt_want_write_file(file); if (ret) return ret; @@ -2011,7 +2185,7 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp) ret = -EINVAL; } out: - mnt_drop_write(file->f_path.mnt); + mnt_drop_write_file(file); return ret; } @@ -2023,14 +2197,25 @@ static long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg) if (!capable(CAP_SYS_ADMIN)) return -EPERM; + mutex_lock(&root->fs_info->volume_mutex); + if (root->fs_info->balance_ctl) { + printk(KERN_INFO "btrfs: balance in progress\n"); + ret = -EINVAL; + goto out; + } + vol_args = memdup_user(arg, sizeof(*vol_args)); - if (IS_ERR(vol_args)) - return PTR_ERR(vol_args); + if (IS_ERR(vol_args)) { + ret = PTR_ERR(vol_args); + goto out; + } vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; ret = btrfs_init_new_device(root, vol_args->name); kfree(vol_args); +out: + mutex_unlock(&root->fs_info->volume_mutex); return ret; } @@ -2045,14 +2230,25 @@ static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg) if (root->fs_info->sb->s_flags & MS_RDONLY) return -EROFS; + mutex_lock(&root->fs_info->volume_mutex); + if (root->fs_info->balance_ctl) { + printk(KERN_INFO "btrfs: balance in progress\n"); + ret = -EINVAL; + goto out; + } + vol_args = memdup_user(arg, sizeof(*vol_args)); - if (IS_ERR(vol_args)) - return PTR_ERR(vol_args); + if (IS_ERR(vol_args)) { + ret = PTR_ERR(vol_args); + goto out; + } vol_args->name[BTRFS_PATH_NAME_MAX] = '\0'; ret = btrfs_rm_device(root, vol_args->name); kfree(vol_args); +out: + mutex_unlock(&root->fs_info->volume_mutex); return ret; } @@ -2120,7 +2316,17 @@ static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg) di_args->bytes_used = dev->bytes_used; di_args->total_bytes = dev->total_bytes; memcpy(di_args->uuid, dev->uuid, sizeof(di_args->uuid)); - strncpy(di_args->path, dev->name, sizeof(di_args->path)); + if (dev->name) { + struct rcu_string *name; + + rcu_read_lock(); + name = rcu_dereference(dev->name); + strncpy(di_args->path, name->str, sizeof(di_args->path)); + rcu_read_unlock(); + di_args->path[sizeof(di_args->path) - 1] = 0; + } else { + di_args->path[0] = '\0'; + } out: if (ret == 0 && copy_to_user(arg, di_args, sizeof(*di_args))) @@ -2166,7 +2372,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, if (btrfs_root_readonly(root)) return -EROFS; - ret = mnt_want_write(file->f_path.mnt); + ret = mnt_want_write_file(file); if (ret) return ret; @@ -2176,6 +2382,10 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, goto out_drop_write; } + ret = -EXDEV; + if (src_file->f_path.mnt != file->f_path.mnt) + goto out_fput; + src = src_file->f_dentry->d_inode; ret = -EINVAL; @@ -2196,7 +2406,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, goto out_fput; ret = -EXDEV; - if (src->i_sb != inode->i_sb || BTRFS_I(src)->root != root) + if (src->i_sb != inode->i_sb) goto out_fput; ret = -ENOMEM; @@ -2248,13 +2458,13 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, another, and lock file content */ while (1) { struct btrfs_ordered_extent *ordered; - lock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); + lock_extent(&BTRFS_I(src)->io_tree, off, off+len); ordered = btrfs_lookup_first_ordered_extent(src, off+len); if (!ordered && !test_range_bit(&BTRFS_I(src)->io_tree, off, off+len, EXTENT_DELALLOC, 0, NULL)) break; - unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); + unlock_extent(&BTRFS_I(src)->io_tree, off, off+len); if (ordered) btrfs_put_ordered_extent(ordered); btrfs_wait_ordered_range(src, off, len); @@ -2270,13 +2480,14 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, * note the key will change type as we walk through the * tree. */ - ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + ret = btrfs_search_slot(NULL, BTRFS_I(src)->root, &key, path, + 0, 0); if (ret < 0) goto out; nritems = btrfs_header_nritems(path->nodes[0]); if (path->slots[0] >= nritems) { - ret = btrfs_next_leaf(root, path); + ret = btrfs_next_leaf(BTRFS_I(src)->root, path); if (ret < 0) goto out; if (ret > 0) @@ -2369,11 +2580,21 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, new_key.offset, new_key.offset + datal, &hint_byte, 1); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, + ret); + btrfs_end_transaction(trans, root); + goto out; + } ret = btrfs_insert_empty_item(trans, root, path, &new_key, size); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, + ret); + btrfs_end_transaction(trans, root); + goto out; + } leaf = path->nodes[0]; slot = path->slots[0]; @@ -2398,8 +2619,17 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, disko, diskl, 0, root->root_key.objectid, btrfs_ino(inode), - new_key.offset - datao); - BUG_ON(ret); + new_key.offset - datao, + 0); + if (ret) { + btrfs_abort_transaction(trans, + root, + ret); + btrfs_end_transaction(trans, + root); + goto out; + + } } } else if (type == BTRFS_FILE_EXTENT_INLINE) { u64 skip = 0; @@ -2424,11 +2654,21 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, new_key.offset, new_key.offset + datal, &hint_byte, 1); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, + ret); + btrfs_end_transaction(trans, root); + goto out; + } ret = btrfs_insert_empty_item(trans, root, path, &new_key, size); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, + ret); + btrfs_end_transaction(trans, root); + goto out; + } if (skip) { u32 start = @@ -2448,6 +2688,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, btrfs_mark_buffer_dirty(leaf); btrfs_release_path(path); + inode_inc_iversion(inode); inode->i_mtime = inode->i_ctime = CURRENT_TIME; /* @@ -2462,8 +2703,12 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, btrfs_i_size_write(inode, endoff); ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); - btrfs_end_transaction(trans, root); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + btrfs_end_transaction(trans, root); + goto out; + } + ret = btrfs_end_transaction(trans, root); } next: btrfs_release_path(path); @@ -2472,7 +2717,7 @@ next: ret = 0; out: btrfs_release_path(path); - unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); + unlock_extent(&BTRFS_I(src)->io_tree, off, off+len); out_unlock: mutex_unlock(&src->i_mutex); mutex_unlock(&inode->i_mutex); @@ -2481,7 +2726,7 @@ out_unlock: out_fput: fput(src_file); out_drop_write: - mnt_drop_write(file->f_path.mnt); + mnt_drop_write_file(file); return ret; } @@ -2520,7 +2765,7 @@ static long btrfs_ioctl_trans_start(struct file *file) if (btrfs_root_readonly(root)) goto out; - ret = mnt_want_write(file->f_path.mnt); + ret = mnt_want_write_file(file); if (ret) goto out; @@ -2536,7 +2781,7 @@ static long btrfs_ioctl_trans_start(struct file *file) out_drop: atomic_dec(&root->fs_info->open_ioctl_trans); - mnt_drop_write(file->f_path.mnt); + mnt_drop_write_file(file); out: return ret; } @@ -2551,8 +2796,6 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp) struct btrfs_path *path; struct btrfs_key location; struct btrfs_disk_key disk_key; - struct btrfs_super_block *disk_super; - u64 features; u64 objectid = 0; u64 dir_id; @@ -2587,7 +2830,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp) return PTR_ERR(trans); } - dir_id = btrfs_super_root_dir(&root->fs_info->super_copy); + dir_id = btrfs_super_root_dir(root->fs_info->super_copy); di = btrfs_lookup_dir_item(trans, root->fs_info->tree_root, path, dir_id, "default", 7, 1); if (IS_ERR_OR_NULL(di)) { @@ -2603,12 +2846,7 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp) btrfs_mark_buffer_dirty(path->nodes[0]); btrfs_free_path(path); - disk_super = &root->fs_info->super_copy; - features = btrfs_super_incompat_flags(disk_super); - if (!(features & BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL)) { - features |= BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL; - btrfs_set_super_incompat_flags(disk_super, features); - } + btrfs_set_fs_incompat(root->fs_info, DEFAULT_SUBVOL); btrfs_end_transaction(trans, root); return 0; @@ -2736,7 +2974,7 @@ long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg) up_read(&info->groups_sem); } - user_dest = (struct btrfs_ioctl_space_info *) + user_dest = (struct btrfs_ioctl_space_info __user *) (arg + sizeof(struct btrfs_ioctl_space_args)); if (copy_to_user(user_dest, dest_orig, alloc_size)) @@ -2771,7 +3009,7 @@ long btrfs_ioctl_trans_end(struct file *file) atomic_dec(&root->fs_info->open_ioctl_trans); - mnt_drop_write(file->f_path.mnt); + mnt_drop_write_file(file); return 0; } @@ -2864,6 +3102,591 @@ static long btrfs_ioctl_scrub_progress(struct btrfs_root *root, return ret; } +static long btrfs_ioctl_get_dev_stats(struct btrfs_root *root, + void __user *arg) +{ + struct btrfs_ioctl_get_dev_stats *sa; + int ret; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + if ((sa->flags & BTRFS_DEV_STATS_RESET) && !capable(CAP_SYS_ADMIN)) { + kfree(sa); + return -EPERM; + } + + ret = btrfs_get_dev_stats(root, sa); + + if (copy_to_user(arg, sa, sizeof(*sa))) + ret = -EFAULT; + + kfree(sa); + return ret; +} + +static long btrfs_ioctl_ino_to_path(struct btrfs_root *root, void __user *arg) +{ + int ret = 0; + int i; + u64 rel_ptr; + int size; + struct btrfs_ioctl_ino_path_args *ipa = NULL; + struct inode_fs_paths *ipath = NULL; + struct btrfs_path *path; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + path = btrfs_alloc_path(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + ipa = memdup_user(arg, sizeof(*ipa)); + if (IS_ERR(ipa)) { + ret = PTR_ERR(ipa); + ipa = NULL; + goto out; + } + + size = min_t(u32, ipa->size, 4096); + ipath = init_ipath(size, root, path); + if (IS_ERR(ipath)) { + ret = PTR_ERR(ipath); + ipath = NULL; + goto out; + } + + ret = paths_from_inode(ipa->inum, ipath); + if (ret < 0) + goto out; + + for (i = 0; i < ipath->fspath->elem_cnt; ++i) { + rel_ptr = ipath->fspath->val[i] - + (u64)(unsigned long)ipath->fspath->val; + ipath->fspath->val[i] = rel_ptr; + } + + ret = copy_to_user((void *)(unsigned long)ipa->fspath, + (void *)(unsigned long)ipath->fspath, size); + if (ret) { + ret = -EFAULT; + goto out; + } + +out: + btrfs_free_path(path); + free_ipath(ipath); + kfree(ipa); + + return ret; +} + +static int build_ino_list(u64 inum, u64 offset, u64 root, void *ctx) +{ + struct btrfs_data_container *inodes = ctx; + const size_t c = 3 * sizeof(u64); + + if (inodes->bytes_left >= c) { + inodes->bytes_left -= c; + inodes->val[inodes->elem_cnt] = inum; + inodes->val[inodes->elem_cnt + 1] = offset; + inodes->val[inodes->elem_cnt + 2] = root; + inodes->elem_cnt += 3; + } else { + inodes->bytes_missing += c - inodes->bytes_left; + inodes->bytes_left = 0; + inodes->elem_missed += 3; + } + + return 0; +} + +static long btrfs_ioctl_logical_to_ino(struct btrfs_root *root, + void __user *arg) +{ + int ret = 0; + int size; + u64 extent_item_pos; + struct btrfs_ioctl_logical_ino_args *loi; + struct btrfs_data_container *inodes = NULL; + struct btrfs_path *path = NULL; + struct btrfs_key key; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + loi = memdup_user(arg, sizeof(*loi)); + if (IS_ERR(loi)) { + ret = PTR_ERR(loi); + loi = NULL; + goto out; + } + + path = btrfs_alloc_path(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + size = min_t(u32, loi->size, 4096); + inodes = init_data_container(size); + if (IS_ERR(inodes)) { + ret = PTR_ERR(inodes); + inodes = NULL; + goto out; + } + + ret = extent_from_logical(root->fs_info, loi->logical, path, &key); + btrfs_release_path(path); + + if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK) + ret = -ENOENT; + if (ret < 0) + goto out; + + extent_item_pos = loi->logical - key.objectid; + ret = iterate_extent_inodes(root->fs_info, key.objectid, + extent_item_pos, 0, build_ino_list, + inodes); + + if (ret < 0) + goto out; + + ret = copy_to_user((void *)(unsigned long)loi->inodes, + (void *)(unsigned long)inodes, size); + if (ret) + ret = -EFAULT; + +out: + btrfs_free_path(path); + kfree(inodes); + kfree(loi); + + return ret; +} + +void update_ioctl_balance_args(struct btrfs_fs_info *fs_info, int lock, + struct btrfs_ioctl_balance_args *bargs) +{ + struct btrfs_balance_control *bctl = fs_info->balance_ctl; + + bargs->flags = bctl->flags; + + if (atomic_read(&fs_info->balance_running)) + bargs->state |= BTRFS_BALANCE_STATE_RUNNING; + if (atomic_read(&fs_info->balance_pause_req)) + bargs->state |= BTRFS_BALANCE_STATE_PAUSE_REQ; + if (atomic_read(&fs_info->balance_cancel_req)) + bargs->state |= BTRFS_BALANCE_STATE_CANCEL_REQ; + + memcpy(&bargs->data, &bctl->data, sizeof(bargs->data)); + memcpy(&bargs->meta, &bctl->meta, sizeof(bargs->meta)); + memcpy(&bargs->sys, &bctl->sys, sizeof(bargs->sys)); + + if (lock) { + spin_lock(&fs_info->balance_lock); + memcpy(&bargs->stat, &bctl->stat, sizeof(bargs->stat)); + spin_unlock(&fs_info->balance_lock); + } else { + memcpy(&bargs->stat, &bctl->stat, sizeof(bargs->stat)); + } +} + +static long btrfs_ioctl_balance(struct file *file, void __user *arg) +{ + struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root; + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_ioctl_balance_args *bargs; + struct btrfs_balance_control *bctl; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + ret = mnt_want_write_file(file); + if (ret) + return ret; + + mutex_lock(&fs_info->volume_mutex); + mutex_lock(&fs_info->balance_mutex); + + if (arg) { + bargs = memdup_user(arg, sizeof(*bargs)); + if (IS_ERR(bargs)) { + ret = PTR_ERR(bargs); + goto out; + } + + if (bargs->flags & BTRFS_BALANCE_RESUME) { + if (!fs_info->balance_ctl) { + ret = -ENOTCONN; + goto out_bargs; + } + + bctl = fs_info->balance_ctl; + spin_lock(&fs_info->balance_lock); + bctl->flags |= BTRFS_BALANCE_RESUME; + spin_unlock(&fs_info->balance_lock); + + goto do_balance; + } + } else { + bargs = NULL; + } + + if (fs_info->balance_ctl) { + ret = -EINPROGRESS; + goto out_bargs; + } + + bctl = kzalloc(sizeof(*bctl), GFP_NOFS); + if (!bctl) { + ret = -ENOMEM; + goto out_bargs; + } + + bctl->fs_info = fs_info; + if (arg) { + memcpy(&bctl->data, &bargs->data, sizeof(bctl->data)); + memcpy(&bctl->meta, &bargs->meta, sizeof(bctl->meta)); + memcpy(&bctl->sys, &bargs->sys, sizeof(bctl->sys)); + + bctl->flags = bargs->flags; + } else { + /* balance everything - no filters */ + bctl->flags |= BTRFS_BALANCE_TYPE_MASK; + } + +do_balance: + ret = btrfs_balance(bctl, bargs); + /* + * bctl is freed in __cancel_balance or in free_fs_info if + * restriper was paused all the way until unmount + */ + if (arg) { + if (copy_to_user(arg, bargs, sizeof(*bargs))) + ret = -EFAULT; + } + +out_bargs: + kfree(bargs); +out: + mutex_unlock(&fs_info->balance_mutex); + mutex_unlock(&fs_info->volume_mutex); + mnt_drop_write_file(file); + return ret; +} + +static long btrfs_ioctl_balance_ctl(struct btrfs_root *root, int cmd) +{ + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + switch (cmd) { + case BTRFS_BALANCE_CTL_PAUSE: + return btrfs_pause_balance(root->fs_info); + case BTRFS_BALANCE_CTL_CANCEL: + return btrfs_cancel_balance(root->fs_info); + } + + return -EINVAL; +} + +static long btrfs_ioctl_balance_progress(struct btrfs_root *root, + void __user *arg) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_ioctl_balance_args *bargs; + int ret = 0; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + mutex_lock(&fs_info->balance_mutex); + if (!fs_info->balance_ctl) { + ret = -ENOTCONN; + goto out; + } + + bargs = kzalloc(sizeof(*bargs), GFP_NOFS); + if (!bargs) { + ret = -ENOMEM; + goto out; + } + + update_ioctl_balance_args(fs_info, 1, bargs); + + if (copy_to_user(arg, bargs, sizeof(*bargs))) + ret = -EFAULT; + + kfree(bargs); +out: + mutex_unlock(&fs_info->balance_mutex); + return ret; +} + +static long btrfs_ioctl_quota_ctl(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_quota_ctl_args *sa; + struct btrfs_trans_handle *trans = NULL; + int ret; + int err; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + if (sa->cmd != BTRFS_QUOTA_CTL_RESCAN) { + trans = btrfs_start_transaction(root, 2); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out; + } + } + + switch (sa->cmd) { + case BTRFS_QUOTA_CTL_ENABLE: + ret = btrfs_quota_enable(trans, root->fs_info); + break; + case BTRFS_QUOTA_CTL_DISABLE: + ret = btrfs_quota_disable(trans, root->fs_info); + break; + case BTRFS_QUOTA_CTL_RESCAN: + ret = btrfs_quota_rescan(root->fs_info); + break; + default: + ret = -EINVAL; + break; + } + + if (copy_to_user(arg, sa, sizeof(*sa))) + ret = -EFAULT; + + if (trans) { + err = btrfs_commit_transaction(trans, root); + if (err && !ret) + ret = err; + } + +out: + kfree(sa); + return ret; +} + +static long btrfs_ioctl_qgroup_assign(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_qgroup_assign_args *sa; + struct btrfs_trans_handle *trans; + int ret; + int err; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out; + } + + /* FIXME: check if the IDs really exist */ + if (sa->assign) { + ret = btrfs_add_qgroup_relation(trans, root->fs_info, + sa->src, sa->dst); + } else { + ret = btrfs_del_qgroup_relation(trans, root->fs_info, + sa->src, sa->dst); + } + + err = btrfs_end_transaction(trans, root); + if (err && !ret) + ret = err; + +out: + kfree(sa); + return ret; +} + +static long btrfs_ioctl_qgroup_create(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_qgroup_create_args *sa; + struct btrfs_trans_handle *trans; + int ret; + int err; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out; + } + + /* FIXME: check if the IDs really exist */ + if (sa->create) { + ret = btrfs_create_qgroup(trans, root->fs_info, sa->qgroupid, + NULL); + } else { + ret = btrfs_remove_qgroup(trans, root->fs_info, sa->qgroupid); + } + + err = btrfs_end_transaction(trans, root); + if (err && !ret) + ret = err; + +out: + kfree(sa); + return ret; +} + +static long btrfs_ioctl_qgroup_limit(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_qgroup_limit_args *sa; + struct btrfs_trans_handle *trans; + int ret; + int err; + u64 qgroupid; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (root->fs_info->sb->s_flags & MS_RDONLY) + return -EROFS; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out; + } + + qgroupid = sa->qgroupid; + if (!qgroupid) { + /* take the current subvol as qgroup */ + qgroupid = root->root_key.objectid; + } + + /* FIXME: check if the IDs really exist */ + ret = btrfs_limit_qgroup(trans, root->fs_info, qgroupid, &sa->lim); + + err = btrfs_end_transaction(trans, root); + if (err && !ret) + ret = err; + +out: + kfree(sa); + return ret; +} + +static long btrfs_ioctl_set_received_subvol(struct file *file, + void __user *arg) +{ + struct btrfs_ioctl_received_subvol_args *sa = NULL; + struct inode *inode = fdentry(file)->d_inode; + struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_root_item *root_item = &root->root_item; + struct btrfs_trans_handle *trans; + struct timespec ct = CURRENT_TIME; + int ret = 0; + + ret = mnt_want_write_file(file); + if (ret < 0) + return ret; + + down_write(&root->fs_info->subvol_sem); + + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) { + ret = -EINVAL; + goto out; + } + + if (btrfs_root_readonly(root)) { + ret = -EROFS; + goto out; + } + + if (!inode_owner_or_capable(inode)) { + ret = -EACCES; + goto out; + } + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) { + ret = PTR_ERR(sa); + sa = NULL; + goto out; + } + + trans = btrfs_start_transaction(root, 1); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto out; + } + + sa->rtransid = trans->transid; + sa->rtime.sec = ct.tv_sec; + sa->rtime.nsec = ct.tv_nsec; + + memcpy(root_item->received_uuid, sa->uuid, BTRFS_UUID_SIZE); + btrfs_set_root_stransid(root_item, sa->stransid); + btrfs_set_root_rtransid(root_item, sa->rtransid); + root_item->stime.sec = cpu_to_le64(sa->stime.sec); + root_item->stime.nsec = cpu_to_le32(sa->stime.nsec); + root_item->rtime.sec = cpu_to_le64(sa->rtime.sec); + root_item->rtime.nsec = cpu_to_le32(sa->rtime.nsec); + + ret = btrfs_update_root(trans, root->fs_info->tree_root, + &root->root_key, &root->root_item); + if (ret < 0) { + btrfs_end_transaction(trans, root); + trans = NULL; + goto out; + } else { + ret = btrfs_commit_transaction(trans, root); + if (ret < 0) + goto out; + } + + ret = copy_to_user(arg, sa, sizeof(*sa)); + if (ret) + ret = -EFAULT; + +out: + kfree(sa); + up_write(&root->fs_info->subvol_sem); + mnt_drop_write_file(file); + return ret; +} + long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { @@ -2885,6 +3708,8 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_snap_create_v2(file, argp, 0); case BTRFS_IOC_SUBVOL_CREATE: return btrfs_ioctl_snap_create(file, argp, 1); + case BTRFS_IOC_SUBVOL_CREATE_V2: + return btrfs_ioctl_snap_create_v2(file, argp, 1); case BTRFS_IOC_SNAP_DESTROY: return btrfs_ioctl_snap_destroy(file, argp); case BTRFS_IOC_SUBVOL_GETFLAGS: @@ -2908,7 +3733,7 @@ long btrfs_ioctl(struct file *file, unsigned int case BTRFS_IOC_DEV_INFO: return btrfs_ioctl_dev_info(root, argp); case BTRFS_IOC_BALANCE: - return btrfs_balance(root->fs_info->dev_root); + return btrfs_ioctl_balance(file, NULL); case BTRFS_IOC_CLONE: return btrfs_ioctl_clone(file, arg, 0, 0, 0); case BTRFS_IOC_CLONE_RANGE: @@ -2921,6 +3746,10 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_tree_search(file, argp); case BTRFS_IOC_INO_LOOKUP: return btrfs_ioctl_ino_lookup(file, argp); + case BTRFS_IOC_INO_PATHS: + return btrfs_ioctl_ino_to_path(root, argp); + case BTRFS_IOC_LOGICAL_INO: + return btrfs_ioctl_logical_to_ino(root, argp); case BTRFS_IOC_SPACE_INFO: return btrfs_ioctl_space_info(root, argp); case BTRFS_IOC_SYNC: @@ -2936,6 +3765,26 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_scrub_cancel(root, argp); case BTRFS_IOC_SCRUB_PROGRESS: return btrfs_ioctl_scrub_progress(root, argp); + case BTRFS_IOC_BALANCE_V2: + return btrfs_ioctl_balance(file, argp); + case BTRFS_IOC_BALANCE_CTL: + return btrfs_ioctl_balance_ctl(root, arg); + case BTRFS_IOC_BALANCE_PROGRESS: + return btrfs_ioctl_balance_progress(root, argp); + case BTRFS_IOC_SET_RECEIVED_SUBVOL: + return btrfs_ioctl_set_received_subvol(file, argp); + case BTRFS_IOC_SEND: + return btrfs_ioctl_send(file, argp); + case BTRFS_IOC_GET_DEV_STATS: + return btrfs_ioctl_get_dev_stats(root, argp); + case BTRFS_IOC_QUOTA_CTL: + return btrfs_ioctl_quota_ctl(root, argp); + case BTRFS_IOC_QGROUP_ASSIGN: + return btrfs_ioctl_qgroup_assign(root, argp); + case BTRFS_IOC_QGROUP_CREATE: + return btrfs_ioctl_qgroup_create(root, argp); + case BTRFS_IOC_QGROUP_LIMIT: + return btrfs_ioctl_qgroup_limit(root, argp); } return -ENOTTY; diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h index ad1ea789fcb4..731e2875ab93 100644 --- a/fs/btrfs/ioctl.h +++ b/fs/btrfs/ioctl.h @@ -32,15 +32,46 @@ struct btrfs_ioctl_vol_args { #define BTRFS_SUBVOL_CREATE_ASYNC (1ULL << 0) #define BTRFS_SUBVOL_RDONLY (1ULL << 1) +#define BTRFS_SUBVOL_QGROUP_INHERIT (1ULL << 2) #define BTRFS_FSID_SIZE 16 #define BTRFS_UUID_SIZE 16 +#define BTRFS_QGROUP_INHERIT_SET_LIMITS (1ULL << 0) + +struct btrfs_qgroup_limit { + __u64 flags; + __u64 max_rfer; + __u64 max_excl; + __u64 rsv_rfer; + __u64 rsv_excl; +}; + +struct btrfs_qgroup_inherit { + __u64 flags; + __u64 num_qgroups; + __u64 num_ref_copies; + __u64 num_excl_copies; + struct btrfs_qgroup_limit lim; + __u64 qgroups[0]; +}; + +struct btrfs_ioctl_qgroup_limit_args { + __u64 qgroupid; + struct btrfs_qgroup_limit lim; +}; + #define BTRFS_SUBVOL_NAME_MAX 4039 struct btrfs_ioctl_vol_args_v2 { __s64 fd; __u64 transid; __u64 flags; - __u64 unused[4]; + union { + struct { + __u64 size; + struct btrfs_qgroup_inherit __user *qgroup_inherit; + }; + __u64 unused[4]; + }; char name[BTRFS_SUBVOL_NAME_MAX + 1]; }; @@ -109,6 +140,55 @@ struct btrfs_ioctl_fs_info_args { __u64 reserved[124]; /* pad to 1k */ }; +/* balance control ioctl modes */ +#define BTRFS_BALANCE_CTL_PAUSE 1 +#define BTRFS_BALANCE_CTL_CANCEL 2 + +/* + * this is packed, because it should be exactly the same as its disk + * byte order counterpart (struct btrfs_disk_balance_args) + */ +struct btrfs_balance_args { + __u64 profiles; + __u64 usage; + __u64 devid; + __u64 pstart; + __u64 pend; + __u64 vstart; + __u64 vend; + + __u64 target; + + __u64 flags; + + __u64 unused[8]; +} __attribute__ ((__packed__)); + +/* report balance progress to userspace */ +struct btrfs_balance_progress { + __u64 expected; /* estimated # of chunks that will be + * relocated to fulfill the request */ + __u64 considered; /* # of chunks we have considered so far */ + __u64 completed; /* # of chunks relocated so far */ +}; + +#define BTRFS_BALANCE_STATE_RUNNING (1ULL << 0) +#define BTRFS_BALANCE_STATE_PAUSE_REQ (1ULL << 1) +#define BTRFS_BALANCE_STATE_CANCEL_REQ (1ULL << 2) + +struct btrfs_ioctl_balance_args { + __u64 flags; /* in/out */ + __u64 state; /* out */ + + struct btrfs_balance_args data; /* in/out */ + struct btrfs_balance_args meta; /* in/out */ + struct btrfs_balance_args sys; /* in/out */ + + struct btrfs_balance_progress stat; /* out */ + + __u64 unused[72]; /* pad to 1k */ +}; + #define BTRFS_INO_LOOKUP_PATH_MAX 4080 struct btrfs_ioctl_ino_lookup_args { __u64 treeid; @@ -193,6 +273,105 @@ struct btrfs_ioctl_space_args { struct btrfs_ioctl_space_info spaces[0]; }; +struct btrfs_data_container { + __u32 bytes_left; /* out -- bytes not needed to deliver output */ + __u32 bytes_missing; /* out -- additional bytes needed for result */ + __u32 elem_cnt; /* out */ + __u32 elem_missed; /* out */ + __u64 val[0]; /* out */ +}; + +struct btrfs_ioctl_ino_path_args { + __u64 inum; /* in */ + __u64 size; /* in */ + __u64 reserved[4]; + /* struct btrfs_data_container *fspath; out */ + __u64 fspath; /* out */ +}; + +struct btrfs_ioctl_logical_ino_args { + __u64 logical; /* in */ + __u64 size; /* in */ + __u64 reserved[4]; + /* struct btrfs_data_container *inodes; out */ + __u64 inodes; +}; + +enum btrfs_dev_stat_values { + /* disk I/O failure stats */ + BTRFS_DEV_STAT_WRITE_ERRS, /* EIO or EREMOTEIO from lower layers */ + BTRFS_DEV_STAT_READ_ERRS, /* EIO or EREMOTEIO from lower layers */ + BTRFS_DEV_STAT_FLUSH_ERRS, /* EIO or EREMOTEIO from lower layers */ + + /* stats for indirect indications for I/O failures */ + BTRFS_DEV_STAT_CORRUPTION_ERRS, /* checksum error, bytenr error or + * contents is illegal: this is an + * indication that the block was damaged + * during read or write, or written to + * wrong location or read from wrong + * location */ + BTRFS_DEV_STAT_GENERATION_ERRS, /* an indication that blocks have not + * been written */ + + BTRFS_DEV_STAT_VALUES_MAX +}; + +/* Reset statistics after reading; needs SYS_ADMIN capability */ +#define BTRFS_DEV_STATS_RESET (1ULL << 0) + +struct btrfs_ioctl_get_dev_stats { + __u64 devid; /* in */ + __u64 nr_items; /* in/out */ + __u64 flags; /* in/out */ + + /* out values: */ + __u64 values[BTRFS_DEV_STAT_VALUES_MAX]; + + __u64 unused[128 - 2 - BTRFS_DEV_STAT_VALUES_MAX]; /* pad to 1k */ +}; + +#define BTRFS_QUOTA_CTL_ENABLE 1 +#define BTRFS_QUOTA_CTL_DISABLE 2 +#define BTRFS_QUOTA_CTL_RESCAN 3 +struct btrfs_ioctl_quota_ctl_args { + __u64 cmd; + __u64 status; +}; + +struct btrfs_ioctl_qgroup_assign_args { + __u64 assign; + __u64 src; + __u64 dst; +}; + +struct btrfs_ioctl_qgroup_create_args { + __u64 create; + __u64 qgroupid; +}; +struct btrfs_ioctl_timespec { + __u64 sec; + __u32 nsec; +}; + +struct btrfs_ioctl_received_subvol_args { + char uuid[BTRFS_UUID_SIZE]; /* in */ + __u64 stransid; /* in */ + __u64 rtransid; /* out */ + struct btrfs_ioctl_timespec stime; /* in */ + struct btrfs_ioctl_timespec rtime; /* out */ + __u64 flags; /* in */ + __u64 reserved[16]; /* in */ +}; + +struct btrfs_ioctl_send_args { + __s64 send_fd; /* in */ + __u64 clone_sources_count; /* in */ + __u64 __user *clone_sources; /* in */ + __u64 parent_root; /* in */ + __u64 flags; /* in */ + __u64 reserved[4]; /* in */ +}; + #define BTRFS_IOC_SNAP_CREATE _IOW(BTRFS_IOCTL_MAGIC, 1, \ struct btrfs_ioctl_vol_args) #define BTRFS_IOC_DEFRAG _IOW(BTRFS_IOCTL_MAGIC, 2, \ @@ -237,7 +416,9 @@ struct btrfs_ioctl_space_args { #define BTRFS_IOC_WAIT_SYNC _IOW(BTRFS_IOCTL_MAGIC, 22, __u64) #define BTRFS_IOC_SNAP_CREATE_V2 _IOW(BTRFS_IOCTL_MAGIC, 23, \ struct btrfs_ioctl_vol_args_v2) -#define BTRFS_IOC_SUBVOL_GETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 25, __u64) +#define BTRFS_IOC_SUBVOL_CREATE_V2 _IOW(BTRFS_IOCTL_MAGIC, 24, \ + struct btrfs_ioctl_vol_args_v2) +#define BTRFS_IOC_SUBVOL_GETFLAGS _IOR(BTRFS_IOCTL_MAGIC, 25, __u64) #define BTRFS_IOC_SUBVOL_SETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 26, __u64) #define BTRFS_IOC_SCRUB _IOWR(BTRFS_IOCTL_MAGIC, 27, \ struct btrfs_ioctl_scrub_args) @@ -248,4 +429,28 @@ struct btrfs_ioctl_space_args { struct btrfs_ioctl_dev_info_args) #define BTRFS_IOC_FS_INFO _IOR(BTRFS_IOCTL_MAGIC, 31, \ struct btrfs_ioctl_fs_info_args) +#define BTRFS_IOC_BALANCE_V2 _IOWR(BTRFS_IOCTL_MAGIC, 32, \ + struct btrfs_ioctl_balance_args) +#define BTRFS_IOC_BALANCE_CTL _IOW(BTRFS_IOCTL_MAGIC, 33, int) +#define BTRFS_IOC_BALANCE_PROGRESS _IOR(BTRFS_IOCTL_MAGIC, 34, \ + struct btrfs_ioctl_balance_args) +#define BTRFS_IOC_INO_PATHS _IOWR(BTRFS_IOCTL_MAGIC, 35, \ + struct btrfs_ioctl_ino_path_args) +#define BTRFS_IOC_LOGICAL_INO _IOWR(BTRFS_IOCTL_MAGIC, 36, \ + struct btrfs_ioctl_ino_path_args) +#define BTRFS_IOC_SET_RECEIVED_SUBVOL _IOWR(BTRFS_IOCTL_MAGIC, 37, \ + struct btrfs_ioctl_received_subvol_args) +#define BTRFS_IOC_SEND _IOW(BTRFS_IOCTL_MAGIC, 38, struct btrfs_ioctl_send_args) +#define BTRFS_IOC_DEVICES_READY _IOR(BTRFS_IOCTL_MAGIC, 39, \ + struct btrfs_ioctl_vol_args) +#define BTRFS_IOC_QUOTA_CTL _IOWR(BTRFS_IOCTL_MAGIC, 40, \ + struct btrfs_ioctl_quota_ctl_args) +#define BTRFS_IOC_QGROUP_ASSIGN _IOW(BTRFS_IOCTL_MAGIC, 41, \ + struct btrfs_ioctl_qgroup_assign_args) +#define BTRFS_IOC_QGROUP_CREATE _IOW(BTRFS_IOCTL_MAGIC, 42, \ + struct btrfs_ioctl_qgroup_create_args) +#define BTRFS_IOC_QGROUP_LIMIT _IOR(BTRFS_IOCTL_MAGIC, 43, \ + struct btrfs_ioctl_qgroup_limit_args) +#define BTRFS_IOC_GET_DEV_STATS _IOWR(BTRFS_IOCTL_MAGIC, 52, \ + struct btrfs_ioctl_get_dev_stats) #endif diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c index d77b67c4b275..2a1762c66041 100644 --- a/fs/btrfs/locking.c +++ b/fs/btrfs/locking.c @@ -33,6 +33,14 @@ void btrfs_assert_tree_read_locked(struct extent_buffer *eb); */ void btrfs_set_lock_blocking_rw(struct extent_buffer *eb, int rw) { + if (eb->lock_nested) { + read_lock(&eb->lock); + if (eb->lock_nested && current->pid == eb->lock_owner) { + read_unlock(&eb->lock); + return; + } + read_unlock(&eb->lock); + } if (rw == BTRFS_WRITE_LOCK) { if (atomic_read(&eb->blocking_writers) == 0) { WARN_ON(atomic_read(&eb->spinning_writers) != 1); @@ -57,18 +65,28 @@ void btrfs_set_lock_blocking_rw(struct extent_buffer *eb, int rw) */ void btrfs_clear_lock_blocking_rw(struct extent_buffer *eb, int rw) { + if (eb->lock_nested) { + read_lock(&eb->lock); + if (eb->lock_nested && current->pid == eb->lock_owner) { + read_unlock(&eb->lock); + return; + } + read_unlock(&eb->lock); + } if (rw == BTRFS_WRITE_LOCK_BLOCKING) { BUG_ON(atomic_read(&eb->blocking_writers) != 1); write_lock(&eb->lock); WARN_ON(atomic_read(&eb->spinning_writers)); atomic_inc(&eb->spinning_writers); - if (atomic_dec_and_test(&eb->blocking_writers)) + if (atomic_dec_and_test(&eb->blocking_writers) && + waitqueue_active(&eb->write_lock_wq)) wake_up(&eb->write_lock_wq); } else if (rw == BTRFS_READ_LOCK_BLOCKING) { BUG_ON(atomic_read(&eb->blocking_readers) == 0); read_lock(&eb->lock); atomic_inc(&eb->spinning_readers); - if (atomic_dec_and_test(&eb->blocking_readers)) + if (atomic_dec_and_test(&eb->blocking_readers) && + waitqueue_active(&eb->read_lock_wq)) wake_up(&eb->read_lock_wq); } return; @@ -81,12 +99,25 @@ void btrfs_clear_lock_blocking_rw(struct extent_buffer *eb, int rw) void btrfs_tree_read_lock(struct extent_buffer *eb) { again: + read_lock(&eb->lock); + if (atomic_read(&eb->blocking_writers) && + current->pid == eb->lock_owner) { + /* + * This extent is already write-locked by our thread. We allow + * an additional read lock to be added because it's for the same + * thread. btrfs_find_all_roots() depends on this as it may be + * called on a partly (write-)locked tree. + */ + BUG_ON(eb->lock_nested); + eb->lock_nested = 1; + read_unlock(&eb->lock); + return; + } + read_unlock(&eb->lock); wait_event(eb->write_lock_wq, atomic_read(&eb->blocking_writers) == 0); read_lock(&eb->lock); if (atomic_read(&eb->blocking_writers)) { read_unlock(&eb->lock); - wait_event(eb->write_lock_wq, - atomic_read(&eb->blocking_writers) == 0); goto again; } atomic_inc(&eb->read_locks); @@ -129,6 +160,7 @@ int btrfs_try_tree_write_lock(struct extent_buffer *eb) } atomic_inc(&eb->write_locks); atomic_inc(&eb->spinning_writers); + eb->lock_owner = current->pid; return 1; } @@ -137,6 +169,15 @@ int btrfs_try_tree_write_lock(struct extent_buffer *eb) */ void btrfs_tree_read_unlock(struct extent_buffer *eb) { + if (eb->lock_nested) { + read_lock(&eb->lock); + if (eb->lock_nested && current->pid == eb->lock_owner) { + eb->lock_nested = 0; + read_unlock(&eb->lock); + return; + } + read_unlock(&eb->lock); + } btrfs_assert_tree_read_locked(eb); WARN_ON(atomic_read(&eb->spinning_readers) == 0); atomic_dec(&eb->spinning_readers); @@ -149,9 +190,19 @@ void btrfs_tree_read_unlock(struct extent_buffer *eb) */ void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb) { + if (eb->lock_nested) { + read_lock(&eb->lock); + if (eb->lock_nested && current->pid == eb->lock_owner) { + eb->lock_nested = 0; + read_unlock(&eb->lock); + return; + } + read_unlock(&eb->lock); + } btrfs_assert_tree_read_locked(eb); WARN_ON(atomic_read(&eb->blocking_readers) == 0); - if (atomic_dec_and_test(&eb->blocking_readers)) + if (atomic_dec_and_test(&eb->blocking_readers) && + waitqueue_active(&eb->read_lock_wq)) wake_up(&eb->read_lock_wq); atomic_dec(&eb->read_locks); } @@ -160,7 +211,7 @@ void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb) * take a spinning write lock. This will wait for both * blocking readers or writers */ -int btrfs_tree_lock(struct extent_buffer *eb) +void btrfs_tree_lock(struct extent_buffer *eb) { again: wait_event(eb->read_lock_wq, atomic_read(&eb->blocking_readers) == 0); @@ -181,13 +232,13 @@ again: WARN_ON(atomic_read(&eb->spinning_writers)); atomic_inc(&eb->spinning_writers); atomic_inc(&eb->write_locks); - return 0; + eb->lock_owner = current->pid; } /* * drop a spinning or a blocking write lock. */ -int btrfs_tree_unlock(struct extent_buffer *eb) +void btrfs_tree_unlock(struct extent_buffer *eb) { int blockers = atomic_read(&eb->blocking_writers); @@ -199,14 +250,14 @@ int btrfs_tree_unlock(struct extent_buffer *eb) if (blockers) { WARN_ON(atomic_read(&eb->spinning_writers)); atomic_dec(&eb->blocking_writers); - smp_wmb(); - wake_up(&eb->write_lock_wq); + smp_mb(); + if (waitqueue_active(&eb->write_lock_wq)) + wake_up(&eb->write_lock_wq); } else { WARN_ON(atomic_read(&eb->spinning_writers) != 1); atomic_dec(&eb->spinning_writers); write_unlock(&eb->lock); } - return 0; } void btrfs_assert_tree_locked(struct extent_buffer *eb) diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h index 17247ddb81a0..ca52681e5f40 100644 --- a/fs/btrfs/locking.h +++ b/fs/btrfs/locking.h @@ -24,8 +24,8 @@ #define BTRFS_WRITE_LOCK_BLOCKING 3 #define BTRFS_READ_LOCK_BLOCKING 4 -int btrfs_tree_lock(struct extent_buffer *eb); -int btrfs_tree_unlock(struct extent_buffer *eb); +void btrfs_tree_lock(struct extent_buffer *eb); +void btrfs_tree_unlock(struct extent_buffer *eb); int btrfs_try_spin_lock(struct extent_buffer *eb); void btrfs_tree_read_lock(struct extent_buffer *eb); diff --git a/fs/btrfs/lzo.c b/fs/btrfs/lzo.c index a178f5ebea78..743b86fa4fcb 100644 --- a/fs/btrfs/lzo.c +++ b/fs/btrfs/lzo.c @@ -411,9 +411,9 @@ static int lzo_decompress(struct list_head *ws, unsigned char *data_in, bytes = min_t(unsigned long, destlen, out_len - start_byte); - kaddr = kmap_atomic(dest_page, KM_USER0); + kaddr = kmap_atomic(dest_page); memcpy(kaddr, workspace->buf + start_byte, bytes); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); out: return ret; } diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index a1c940425307..051c7fe551dd 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -59,6 +59,14 @@ static struct rb_node *tree_insert(struct rb_root *root, u64 file_offset, return NULL; } +static void ordered_data_tree_panic(struct inode *inode, int errno, + u64 offset) +{ + struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb); + btrfs_panic(fs_info, errno, "Inconsistency in ordered tree at offset " + "%llu\n", (unsigned long long)offset); +} + /* * look for a given offset in the tree, and if it can't be found return the * first lesser offset @@ -188,7 +196,7 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset, entry->len = len; entry->disk_len = disk_len; entry->bytes_left = len; - entry->inode = inode; + entry->inode = igrab(inode); entry->compress_type = compress_type; if (type != BTRFS_ORDERED_IO_DONE && type != BTRFS_ORDERED_COMPLETE) set_bit(type, &entry->flags); @@ -204,18 +212,18 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset, trace_btrfs_ordered_extent_add(inode, entry); - spin_lock(&tree->lock); + spin_lock_irq(&tree->lock); node = tree_insert(&tree->tree, file_offset, &entry->rb_node); - BUG_ON(node); - spin_unlock(&tree->lock); + if (node) + ordered_data_tree_panic(inode, -EEXIST, file_offset); + spin_unlock_irq(&tree->lock); spin_lock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock); list_add_tail(&entry->root_extent_list, &BTRFS_I(inode)->root->fs_info->ordered_extents); spin_unlock(&BTRFS_I(inode)->root->fs_info->ordered_extent_lock); - BUG_ON(node); return 0; } @@ -249,17 +257,16 @@ int btrfs_add_ordered_extent_compress(struct inode *inode, u64 file_offset, * when an ordered extent is finished. If the list covers more than one * ordered extent, it is split across multiples. */ -int btrfs_add_ordered_sum(struct inode *inode, - struct btrfs_ordered_extent *entry, - struct btrfs_ordered_sum *sum) +void btrfs_add_ordered_sum(struct inode *inode, + struct btrfs_ordered_extent *entry, + struct btrfs_ordered_sum *sum) { struct btrfs_ordered_inode_tree *tree; tree = &BTRFS_I(inode)->ordered_tree; - spin_lock(&tree->lock); + spin_lock_irq(&tree->lock); list_add_tail(&sum->list, &entry->list); - spin_unlock(&tree->lock); - return 0; + spin_unlock_irq(&tree->lock); } /* @@ -276,18 +283,19 @@ int btrfs_add_ordered_sum(struct inode *inode, */ int btrfs_dec_test_first_ordered_pending(struct inode *inode, struct btrfs_ordered_extent **cached, - u64 *file_offset, u64 io_size) + u64 *file_offset, u64 io_size, int uptodate) { struct btrfs_ordered_inode_tree *tree; struct rb_node *node; struct btrfs_ordered_extent *entry = NULL; int ret; + unsigned long flags; u64 dec_end; u64 dec_start; u64 to_dec; tree = &BTRFS_I(inode)->ordered_tree; - spin_lock(&tree->lock); + spin_lock_irqsave(&tree->lock, flags); node = tree_search(tree, *file_offset); if (!node) { ret = 1; @@ -316,6 +324,9 @@ int btrfs_dec_test_first_ordered_pending(struct inode *inode, (unsigned long long)to_dec); } entry->bytes_left -= to_dec; + if (!uptodate) + set_bit(BTRFS_ORDERED_IOERR, &entry->flags); + if (entry->bytes_left == 0) ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags); else @@ -325,7 +336,7 @@ out: *cached = entry; atomic_inc(&entry->refs); } - spin_unlock(&tree->lock); + spin_unlock_irqrestore(&tree->lock, flags); return ret == 0; } @@ -340,15 +351,21 @@ out: */ int btrfs_dec_test_ordered_pending(struct inode *inode, struct btrfs_ordered_extent **cached, - u64 file_offset, u64 io_size) + u64 file_offset, u64 io_size, int uptodate) { struct btrfs_ordered_inode_tree *tree; struct rb_node *node; struct btrfs_ordered_extent *entry = NULL; + unsigned long flags; int ret; tree = &BTRFS_I(inode)->ordered_tree; - spin_lock(&tree->lock); + spin_lock_irqsave(&tree->lock, flags); + if (cached && *cached) { + entry = *cached; + goto have_entry; + } + node = tree_search(tree, file_offset); if (!node) { ret = 1; @@ -356,6 +373,7 @@ int btrfs_dec_test_ordered_pending(struct inode *inode, } entry = rb_entry(node, struct btrfs_ordered_extent, rb_node); +have_entry: if (!offset_in_entry(entry, file_offset)) { ret = 1; goto out; @@ -367,6 +385,9 @@ int btrfs_dec_test_ordered_pending(struct inode *inode, (unsigned long long)io_size); } entry->bytes_left -= io_size; + if (!uptodate) + set_bit(BTRFS_ORDERED_IOERR, &entry->flags); + if (entry->bytes_left == 0) ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags); else @@ -376,7 +397,7 @@ out: *cached = entry; atomic_inc(&entry->refs); } - spin_unlock(&tree->lock); + spin_unlock_irqrestore(&tree->lock, flags); return ret == 0; } @@ -384,7 +405,7 @@ out: * used to drop a reference on an ordered extent. This will free * the extent if the last reference is dropped */ -int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry) +void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry) { struct list_head *cur; struct btrfs_ordered_sum *sum; @@ -392,6 +413,8 @@ int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry) trace_btrfs_ordered_extent_put(entry->inode, entry); if (atomic_dec_and_test(&entry->refs)) { + if (entry->inode) + btrfs_add_delayed_iput(entry->inode); while (!list_empty(&entry->list)) { cur = entry->list.next; sum = list_entry(cur, struct btrfs_ordered_sum, list); @@ -400,26 +423,26 @@ int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry) } kfree(entry); } - return 0; } /* * remove an ordered extent from the tree. No references are dropped - * and you must wake_up entry->wait. You must hold the tree lock - * while you call this function. + * and waiters are woken up. */ -static int __btrfs_remove_ordered_extent(struct inode *inode, - struct btrfs_ordered_extent *entry) +void btrfs_remove_ordered_extent(struct inode *inode, + struct btrfs_ordered_extent *entry) { struct btrfs_ordered_inode_tree *tree; struct btrfs_root *root = BTRFS_I(inode)->root; struct rb_node *node; tree = &BTRFS_I(inode)->ordered_tree; + spin_lock_irq(&tree->lock); node = &entry->rb_node; rb_erase(node, &tree->tree); tree->last = NULL; set_bit(BTRFS_ORDERED_COMPLETE, &entry->flags); + spin_unlock_irq(&tree->lock); spin_lock(&root->fs_info->ordered_extent_lock); list_del_init(&entry->root_extent_list); @@ -436,35 +459,15 @@ static int __btrfs_remove_ordered_extent(struct inode *inode, list_del_init(&BTRFS_I(inode)->ordered_operations); } spin_unlock(&root->fs_info->ordered_extent_lock); - - return 0; -} - -/* - * remove an ordered extent from the tree. No references are dropped - * but any waiters are woken. - */ -int btrfs_remove_ordered_extent(struct inode *inode, - struct btrfs_ordered_extent *entry) -{ - struct btrfs_ordered_inode_tree *tree; - int ret; - - tree = &BTRFS_I(inode)->ordered_tree; - spin_lock(&tree->lock); - ret = __btrfs_remove_ordered_extent(inode, entry); - spin_unlock(&tree->lock); wake_up(&entry->wait); - - return ret; } /* * wait for all the ordered extents in a root. This is done when balancing * space between drives. */ -int btrfs_wait_ordered_extents(struct btrfs_root *root, - int nocow_only, int delay_iput) +void btrfs_wait_ordered_extents(struct btrfs_root *root, + int nocow_only, int delay_iput) { struct list_head splice; struct list_head *cur; @@ -512,7 +515,6 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root, spin_lock(&root->fs_info->ordered_extent_lock); } spin_unlock(&root->fs_info->ordered_extent_lock); - return 0; } /* @@ -525,7 +527,7 @@ int btrfs_wait_ordered_extents(struct btrfs_root *root, * extra check to make sure the ordered operation list really is empty * before we return */ -int btrfs_run_ordered_operations(struct btrfs_root *root, int wait) +void btrfs_run_ordered_operations(struct btrfs_root *root, int wait) { struct btrfs_inode *btrfs_inode; struct inode *inode; @@ -573,8 +575,6 @@ again: spin_unlock(&root->fs_info->ordered_extent_lock); mutex_unlock(&root->fs_info->ordered_operations_mutex); - - return 0; } /* @@ -596,7 +596,7 @@ void btrfs_start_ordered_extent(struct inode *inode, /* * pages in the range can be dirty, clean or writeback. We * start IO on any dirty ones so the wait doesn't stall waiting - * for pdflush to find them + * for the flusher thread to find them */ if (!test_bit(BTRFS_ORDERED_DIRECT, &entry->flags)) filemap_fdatawrite_range(inode->i_mapping, start, end); @@ -609,7 +609,7 @@ void btrfs_start_ordered_extent(struct inode *inode, /* * Used to wait on ordered extents across a large range of bytes. */ -int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len) +void btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len) { u64 end; u64 orig_end; @@ -623,17 +623,29 @@ int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len) if (orig_end > INT_LIMIT(loff_t)) orig_end = INT_LIMIT(loff_t); } -again: + /* start IO across the range first to instantiate any delalloc * extents */ filemap_fdatawrite_range(inode->i_mapping, start, orig_end); - /* The compression code will leave pages locked but return from - * writepage without setting the page writeback. Starting again - * with WB_SYNC_ALL will end up waiting for the IO to actually start. + /* + * So with compression we will find and lock a dirty page and clear the + * first one as dirty, setup an async extent, and immediately return + * with the entire range locked but with nobody actually marked with + * writeback. So we can't just filemap_write_and_wait_range() and + * expect it to work since it will just kick off a thread to do the + * actual work. So we need to call filemap_fdatawrite_range _again_ + * since it will wait on the page lock, which won't be unlocked until + * after the pages have been marked as writeback and so we're good to go + * from there. We have to do this otherwise we'll miss the ordered + * extents and that results in badness. Please Josef, do not think you + * know better and pull this out at some point in the future, it is + * right and you are wrong. */ - filemap_fdatawrite_range(inode->i_mapping, start, orig_end); + if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, + &BTRFS_I(inode)->runtime_flags)) + filemap_fdatawrite_range(inode->i_mapping, start, orig_end); filemap_fdatawait_range(inode->i_mapping, start, orig_end); @@ -659,12 +671,6 @@ again: break; end--; } - if (found || test_range_bit(&BTRFS_I(inode)->io_tree, start, orig_end, - EXTENT_DELALLOC, 0, NULL)) { - schedule_timeout(1); - goto again; - } - return 0; } /* @@ -679,7 +685,7 @@ struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode, struct btrfs_ordered_extent *entry = NULL; tree = &BTRFS_I(inode)->ordered_tree; - spin_lock(&tree->lock); + spin_lock_irq(&tree->lock); node = tree_search(tree, file_offset); if (!node) goto out; @@ -690,7 +696,7 @@ struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode, if (entry) atomic_inc(&entry->refs); out: - spin_unlock(&tree->lock); + spin_unlock_irq(&tree->lock); return entry; } @@ -706,7 +712,7 @@ struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode, struct btrfs_ordered_extent *entry = NULL; tree = &BTRFS_I(inode)->ordered_tree; - spin_lock(&tree->lock); + spin_lock_irq(&tree->lock); node = tree_search(tree, file_offset); if (!node) { node = tree_search(tree, file_offset + len); @@ -731,7 +737,7 @@ struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode, out: if (entry) atomic_inc(&entry->refs); - spin_unlock(&tree->lock); + spin_unlock_irq(&tree->lock); return entry; } @@ -747,7 +753,7 @@ btrfs_lookup_first_ordered_extent(struct inode *inode, u64 file_offset) struct btrfs_ordered_extent *entry = NULL; tree = &BTRFS_I(inode)->ordered_tree; - spin_lock(&tree->lock); + spin_lock_irq(&tree->lock); node = tree_search(tree, file_offset); if (!node) goto out; @@ -755,7 +761,7 @@ btrfs_lookup_first_ordered_extent(struct inode *inode, u64 file_offset) entry = rb_entry(node, struct btrfs_ordered_extent, rb_node); atomic_inc(&entry->refs); out: - spin_unlock(&tree->lock); + spin_unlock_irq(&tree->lock); return entry; } @@ -767,7 +773,6 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, struct btrfs_ordered_extent *ordered) { struct btrfs_ordered_inode_tree *tree = &BTRFS_I(inode)->ordered_tree; - struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; u64 disk_i_size; u64 new_i_size; u64 i_size_test; @@ -782,7 +787,7 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, else offset = ALIGN(offset, BTRFS_I(inode)->root->sectorsize); - spin_lock(&tree->lock); + spin_lock_irq(&tree->lock); disk_i_size = BTRFS_I(inode)->disk_i_size; /* truncate file */ @@ -801,14 +806,6 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, } /* - * we can't update the disk_isize if there are delalloc bytes - * between disk_i_size and this ordered extent - */ - if (test_range_bit(io_tree, disk_i_size, offset - 1, - EXTENT_DELALLOC, 0, NULL)) { - goto out; - } - /* * walk backward from this ordered extent to disk_i_size. * if we find an ordered extent then we can't update disk i_size * yet @@ -828,15 +825,18 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, } node = prev; } - while (node) { + for (; node; node = rb_prev(node)) { test = rb_entry(node, struct btrfs_ordered_extent, rb_node); + + /* We treat this entry as if it doesnt exist */ + if (test_bit(BTRFS_ORDERED_UPDATED_ISIZE, &test->flags)) + continue; if (test->file_offset + test->len <= disk_i_size) break; if (test->file_offset >= i_size) break; if (test->file_offset >= disk_i_size) goto out; - node = rb_prev(node); } new_i_size = min_t(u64, offset, i_size); @@ -854,43 +854,49 @@ int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, else node = rb_first(&tree->tree); } - i_size_test = 0; - if (node) { - /* - * do we have an area where IO might have finished - * between our ordered extent and the next one. - */ + + /* + * We are looking for an area between our current extent and the next + * ordered extent to update the i_size to. There are 3 cases here + * + * 1) We don't actually have anything and we can update to i_size. + * 2) We have stuff but they already did their i_size update so again we + * can just update to i_size. + * 3) We have an outstanding ordered extent so the most we can update + * our disk_i_size to is the start of the next offset. + */ + i_size_test = i_size; + for (; node; node = rb_next(node)) { test = rb_entry(node, struct btrfs_ordered_extent, rb_node); - if (test->file_offset > offset) + + if (test_bit(BTRFS_ORDERED_UPDATED_ISIZE, &test->flags)) + continue; + if (test->file_offset > offset) { i_size_test = test->file_offset; - } else { - i_size_test = i_size; + break; + } } /* * i_size_test is the end of a region after this ordered - * extent where there are no ordered extents. As long as there - * are no delalloc bytes in this area, it is safe to update - * disk_i_size to the end of the region. + * extent where there are no ordered extents, we can safely set + * disk_i_size to this. */ - if (i_size_test > offset && - !test_range_bit(io_tree, offset, i_size_test - 1, - EXTENT_DELALLOC, 0, NULL)) { + if (i_size_test > offset) new_i_size = min_t(u64, i_size_test, i_size); - } BTRFS_I(inode)->disk_i_size = new_i_size; ret = 0; out: /* - * we need to remove the ordered extent with the tree lock held - * so that other people calling this function don't find our fully - * processed ordered entry and skip updating the i_size + * We need to do this because we can't remove ordered extents until + * after the i_disk_size has been updated and then the inode has been + * updated to reflect the change, so we need to tell anybody who finds + * this ordered extent that we've already done all the real work, we + * just haven't completed all the other work. */ if (ordered) - __btrfs_remove_ordered_extent(inode, ordered); - spin_unlock(&tree->lock); - if (ordered) - wake_up(&ordered->wait); + set_bit(BTRFS_ORDERED_UPDATED_ISIZE, &ordered->flags); + spin_unlock_irq(&tree->lock); return ret; } @@ -915,7 +921,7 @@ int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, if (!ordered) return 1; - spin_lock(&tree->lock); + spin_lock_irq(&tree->lock); list_for_each_entry_reverse(ordered_sum, &ordered->list, list) { if (disk_bytenr >= ordered_sum->bytenr) { num_sectors = ordered_sum->len / sectorsize; @@ -930,7 +936,7 @@ int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, } } out: - spin_unlock(&tree->lock); + spin_unlock_irq(&tree->lock); btrfs_put_ordered_extent(ordered); return ret; } @@ -948,9 +954,8 @@ out: * If trans is not null, we'll do a friendly check for a transaction that * is already flushing things and force the IO down ourselves. */ -int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct inode *inode) +void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode) { u64 last_mod; @@ -961,7 +966,7 @@ int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans, * commit, we can safely return without doing anything */ if (last_mod < root->fs_info->last_trans_committed) - return 0; + return; /* * the transaction is already committing. Just start the IO and @@ -969,7 +974,7 @@ int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans, */ if (trans && root->fs_info->running_transaction->blocked) { btrfs_wait_ordered_range(inode, 0, (u64)-1); - return 0; + return; } spin_lock(&root->fs_info->ordered_extent_lock); @@ -978,6 +983,4 @@ int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans, &root->fs_info->ordered_operations); } spin_unlock(&root->fs_info->ordered_extent_lock); - - return 0; } diff --git a/fs/btrfs/ordered-data.h b/fs/btrfs/ordered-data.h index ff1f69aa1883..e03c560d2997 100644 --- a/fs/btrfs/ordered-data.h +++ b/fs/btrfs/ordered-data.h @@ -74,6 +74,12 @@ struct btrfs_ordered_sum { #define BTRFS_ORDERED_DIRECT 5 /* set when we're doing DIO with this extent */ +#define BTRFS_ORDERED_IOERR 6 /* We had an io error when writing this out */ + +#define BTRFS_ORDERED_UPDATED_ISIZE 7 /* indicates wether this ordered extent + * has done its due diligence in updating + * the isize. */ + struct btrfs_ordered_extent { /* logical offset in the file */ u64 file_offset; @@ -113,6 +119,8 @@ struct btrfs_ordered_extent { /* a per root list of all the pending ordered extents */ struct list_head root_extent_list; + + struct btrfs_work work; }; @@ -138,15 +146,16 @@ btrfs_ordered_inode_tree_init(struct btrfs_ordered_inode_tree *t) t->last = NULL; } -int btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry); -int btrfs_remove_ordered_extent(struct inode *inode, +void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry); +void btrfs_remove_ordered_extent(struct inode *inode, struct btrfs_ordered_extent *entry); int btrfs_dec_test_ordered_pending(struct inode *inode, struct btrfs_ordered_extent **cached, - u64 file_offset, u64 io_size); + u64 file_offset, u64 io_size, int uptodate); int btrfs_dec_test_first_ordered_pending(struct inode *inode, struct btrfs_ordered_extent **cached, - u64 *file_offset, u64 io_size); + u64 *file_offset, u64 io_size, + int uptodate); int btrfs_add_ordered_extent(struct inode *inode, u64 file_offset, u64 start, u64 len, u64 disk_len, int type); int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset, @@ -154,14 +163,14 @@ int btrfs_add_ordered_extent_dio(struct inode *inode, u64 file_offset, int btrfs_add_ordered_extent_compress(struct inode *inode, u64 file_offset, u64 start, u64 len, u64 disk_len, int type, int compress_type); -int btrfs_add_ordered_sum(struct inode *inode, - struct btrfs_ordered_extent *entry, - struct btrfs_ordered_sum *sum); +void btrfs_add_ordered_sum(struct inode *inode, + struct btrfs_ordered_extent *entry, + struct btrfs_ordered_sum *sum); struct btrfs_ordered_extent *btrfs_lookup_ordered_extent(struct inode *inode, u64 file_offset); void btrfs_start_ordered_extent(struct inode *inode, struct btrfs_ordered_extent *entry, int wait); -int btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len); +void btrfs_wait_ordered_range(struct inode *inode, u64 start, u64 len); struct btrfs_ordered_extent * btrfs_lookup_first_ordered_extent(struct inode * inode, u64 file_offset); struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode, @@ -170,10 +179,10 @@ struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode, int btrfs_ordered_update_i_size(struct inode *inode, u64 offset, struct btrfs_ordered_extent *ordered); int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, u32 *sum); -int btrfs_run_ordered_operations(struct btrfs_root *root, int wait); -int btrfs_add_ordered_operation(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct inode *inode); -int btrfs_wait_ordered_extents(struct btrfs_root *root, - int nocow_only, int delay_iput); +void btrfs_run_ordered_operations(struct btrfs_root *root, int wait); +void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *inode); +void btrfs_wait_ordered_extents(struct btrfs_root *root, + int nocow_only, int delay_iput); #endif diff --git a/fs/btrfs/orphan.c b/fs/btrfs/orphan.c index f8be250963a0..24cad1695af7 100644 --- a/fs/btrfs/orphan.c +++ b/fs/btrfs/orphan.c @@ -58,7 +58,7 @@ int btrfs_del_orphan_item(struct btrfs_trans_handle *trans, ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret < 0) goto out; - if (ret) { + if (ret) { /* JDM: Really? */ ret = -ENOENT; goto out; } diff --git a/fs/btrfs/print-tree.c b/fs/btrfs/print-tree.c index fb2605d998e9..5e23684887eb 100644 --- a/fs/btrfs/print-tree.c +++ b/fs/btrfs/print-tree.c @@ -158,8 +158,7 @@ static void print_extent_ref_v0(struct extent_buffer *eb, int slot) void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l) { int i; - u32 type; - u32 nr = btrfs_header_nritems(l); + u32 type, nr; struct btrfs_item *item; struct btrfs_root_item *ri; struct btrfs_dir_item *di; @@ -172,6 +171,11 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l) struct btrfs_key key; struct btrfs_key found_key; + if (!l) + return; + + nr = btrfs_header_nritems(l); + printk(KERN_INFO "leaf %llu total ptrs %d free space %d\n", (unsigned long long)btrfs_header_bytenr(l), nr, btrfs_leaf_free_space(root, l)); @@ -290,6 +294,9 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l) btrfs_dev_extent_chunk_offset(l, dev_extent), (unsigned long long) btrfs_dev_extent_length(l, dev_extent)); + case BTRFS_DEV_STATS_KEY: + printk(KERN_INFO "\t\tdevice stats\n"); + break; }; } } diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c new file mode 100644 index 000000000000..b65015581744 --- /dev/null +++ b/fs/btrfs/qgroup.c @@ -0,0 +1,1573 @@ +/* + * Copyright (C) 2011 STRATO. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/sched.h> +#include <linux/pagemap.h> +#include <linux/writeback.h> +#include <linux/blkdev.h> +#include <linux/rbtree.h> +#include <linux/slab.h> +#include <linux/workqueue.h> + +#include "ctree.h" +#include "transaction.h" +#include "disk-io.h" +#include "locking.h" +#include "ulist.h" +#include "ioctl.h" +#include "backref.h" + +/* TODO XXX FIXME + * - subvol delete -> delete when ref goes to 0? delete limits also? + * - reorganize keys + * - compressed + * - sync + * - rescan + * - copy also limits on subvol creation + * - limit + * - caches fuer ulists + * - performance benchmarks + * - check all ioctl parameters + */ + +/* + * one struct for each qgroup, organized in fs_info->qgroup_tree. + */ +struct btrfs_qgroup { + u64 qgroupid; + + /* + * state + */ + u64 rfer; /* referenced */ + u64 rfer_cmpr; /* referenced compressed */ + u64 excl; /* exclusive */ + u64 excl_cmpr; /* exclusive compressed */ + + /* + * limits + */ + u64 lim_flags; /* which limits are set */ + u64 max_rfer; + u64 max_excl; + u64 rsv_rfer; + u64 rsv_excl; + + /* + * reservation tracking + */ + u64 reserved; + + /* + * lists + */ + struct list_head groups; /* groups this group is member of */ + struct list_head members; /* groups that are members of this group */ + struct list_head dirty; /* dirty groups */ + struct rb_node node; /* tree of qgroups */ + + /* + * temp variables for accounting operations + */ + u64 tag; + u64 refcnt; +}; + +/* + * glue structure to represent the relations between qgroups. + */ +struct btrfs_qgroup_list { + struct list_head next_group; + struct list_head next_member; + struct btrfs_qgroup *group; + struct btrfs_qgroup *member; +}; + +/* must be called with qgroup_lock held */ +static struct btrfs_qgroup *find_qgroup_rb(struct btrfs_fs_info *fs_info, + u64 qgroupid) +{ + struct rb_node *n = fs_info->qgroup_tree.rb_node; + struct btrfs_qgroup *qgroup; + + while (n) { + qgroup = rb_entry(n, struct btrfs_qgroup, node); + if (qgroup->qgroupid < qgroupid) + n = n->rb_left; + else if (qgroup->qgroupid > qgroupid) + n = n->rb_right; + else + return qgroup; + } + return NULL; +} + +/* must be called with qgroup_lock held */ +static struct btrfs_qgroup *add_qgroup_rb(struct btrfs_fs_info *fs_info, + u64 qgroupid) +{ + struct rb_node **p = &fs_info->qgroup_tree.rb_node; + struct rb_node *parent = NULL; + struct btrfs_qgroup *qgroup; + + while (*p) { + parent = *p; + qgroup = rb_entry(parent, struct btrfs_qgroup, node); + + if (qgroup->qgroupid < qgroupid) + p = &(*p)->rb_left; + else if (qgroup->qgroupid > qgroupid) + p = &(*p)->rb_right; + else + return qgroup; + } + + qgroup = kzalloc(sizeof(*qgroup), GFP_ATOMIC); + if (!qgroup) + return ERR_PTR(-ENOMEM); + + qgroup->qgroupid = qgroupid; + INIT_LIST_HEAD(&qgroup->groups); + INIT_LIST_HEAD(&qgroup->members); + INIT_LIST_HEAD(&qgroup->dirty); + + rb_link_node(&qgroup->node, parent, p); + rb_insert_color(&qgroup->node, &fs_info->qgroup_tree); + + return qgroup; +} + +/* must be called with qgroup_lock held */ +static int del_qgroup_rb(struct btrfs_fs_info *fs_info, u64 qgroupid) +{ + struct btrfs_qgroup *qgroup = find_qgroup_rb(fs_info, qgroupid); + struct btrfs_qgroup_list *list; + + if (!qgroup) + return -ENOENT; + + rb_erase(&qgroup->node, &fs_info->qgroup_tree); + list_del(&qgroup->dirty); + + while (!list_empty(&qgroup->groups)) { + list = list_first_entry(&qgroup->groups, + struct btrfs_qgroup_list, next_group); + list_del(&list->next_group); + list_del(&list->next_member); + kfree(list); + } + + while (!list_empty(&qgroup->members)) { + list = list_first_entry(&qgroup->members, + struct btrfs_qgroup_list, next_member); + list_del(&list->next_group); + list_del(&list->next_member); + kfree(list); + } + kfree(qgroup); + + return 0; +} + +/* must be called with qgroup_lock held */ +static int add_relation_rb(struct btrfs_fs_info *fs_info, + u64 memberid, u64 parentid) +{ + struct btrfs_qgroup *member; + struct btrfs_qgroup *parent; + struct btrfs_qgroup_list *list; + + member = find_qgroup_rb(fs_info, memberid); + parent = find_qgroup_rb(fs_info, parentid); + if (!member || !parent) + return -ENOENT; + + list = kzalloc(sizeof(*list), GFP_ATOMIC); + if (!list) + return -ENOMEM; + + list->group = parent; + list->member = member; + list_add_tail(&list->next_group, &member->groups); + list_add_tail(&list->next_member, &parent->members); + + return 0; +} + +/* must be called with qgroup_lock held */ +static int del_relation_rb(struct btrfs_fs_info *fs_info, + u64 memberid, u64 parentid) +{ + struct btrfs_qgroup *member; + struct btrfs_qgroup *parent; + struct btrfs_qgroup_list *list; + + member = find_qgroup_rb(fs_info, memberid); + parent = find_qgroup_rb(fs_info, parentid); + if (!member || !parent) + return -ENOENT; + + list_for_each_entry(list, &member->groups, next_group) { + if (list->group == parent) { + list_del(&list->next_group); + list_del(&list->next_member); + kfree(list); + return 0; + } + } + return -ENOENT; +} + +/* + * The full config is read in one go, only called from open_ctree() + * It doesn't use any locking, as at this point we're still single-threaded + */ +int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info) +{ + struct btrfs_key key; + struct btrfs_key found_key; + struct btrfs_root *quota_root = fs_info->quota_root; + struct btrfs_path *path = NULL; + struct extent_buffer *l; + int slot; + int ret = 0; + u64 flags = 0; + + if (!fs_info->quota_enabled) + return 0; + + path = btrfs_alloc_path(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + /* default this to quota off, in case no status key is found */ + fs_info->qgroup_flags = 0; + + /* + * pass 1: read status, all qgroup infos and limits + */ + key.objectid = 0; + key.type = 0; + key.offset = 0; + ret = btrfs_search_slot_for_read(quota_root, &key, path, 1, 1); + if (ret) + goto out; + + while (1) { + struct btrfs_qgroup *qgroup; + + slot = path->slots[0]; + l = path->nodes[0]; + btrfs_item_key_to_cpu(l, &found_key, slot); + + if (found_key.type == BTRFS_QGROUP_STATUS_KEY) { + struct btrfs_qgroup_status_item *ptr; + + ptr = btrfs_item_ptr(l, slot, + struct btrfs_qgroup_status_item); + + if (btrfs_qgroup_status_version(l, ptr) != + BTRFS_QGROUP_STATUS_VERSION) { + printk(KERN_ERR + "btrfs: old qgroup version, quota disabled\n"); + goto out; + } + if (btrfs_qgroup_status_generation(l, ptr) != + fs_info->generation) { + flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT; + printk(KERN_ERR + "btrfs: qgroup generation mismatch, " + "marked as inconsistent\n"); + } + fs_info->qgroup_flags = btrfs_qgroup_status_flags(l, + ptr); + /* FIXME read scan element */ + goto next1; + } + + if (found_key.type != BTRFS_QGROUP_INFO_KEY && + found_key.type != BTRFS_QGROUP_LIMIT_KEY) + goto next1; + + qgroup = find_qgroup_rb(fs_info, found_key.offset); + if ((qgroup && found_key.type == BTRFS_QGROUP_INFO_KEY) || + (!qgroup && found_key.type == BTRFS_QGROUP_LIMIT_KEY)) { + printk(KERN_ERR "btrfs: inconsitent qgroup config\n"); + flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT; + } + if (!qgroup) { + qgroup = add_qgroup_rb(fs_info, found_key.offset); + if (IS_ERR(qgroup)) { + ret = PTR_ERR(qgroup); + goto out; + } + } + switch (found_key.type) { + case BTRFS_QGROUP_INFO_KEY: { + struct btrfs_qgroup_info_item *ptr; + + ptr = btrfs_item_ptr(l, slot, + struct btrfs_qgroup_info_item); + qgroup->rfer = btrfs_qgroup_info_rfer(l, ptr); + qgroup->rfer_cmpr = btrfs_qgroup_info_rfer_cmpr(l, ptr); + qgroup->excl = btrfs_qgroup_info_excl(l, ptr); + qgroup->excl_cmpr = btrfs_qgroup_info_excl_cmpr(l, ptr); + /* generation currently unused */ + break; + } + case BTRFS_QGROUP_LIMIT_KEY: { + struct btrfs_qgroup_limit_item *ptr; + + ptr = btrfs_item_ptr(l, slot, + struct btrfs_qgroup_limit_item); + qgroup->lim_flags = btrfs_qgroup_limit_flags(l, ptr); + qgroup->max_rfer = btrfs_qgroup_limit_max_rfer(l, ptr); + qgroup->max_excl = btrfs_qgroup_limit_max_excl(l, ptr); + qgroup->rsv_rfer = btrfs_qgroup_limit_rsv_rfer(l, ptr); + qgroup->rsv_excl = btrfs_qgroup_limit_rsv_excl(l, ptr); + break; + } + } +next1: + ret = btrfs_next_item(quota_root, path); + if (ret < 0) + goto out; + if (ret) + break; + } + btrfs_release_path(path); + + /* + * pass 2: read all qgroup relations + */ + key.objectid = 0; + key.type = BTRFS_QGROUP_RELATION_KEY; + key.offset = 0; + ret = btrfs_search_slot_for_read(quota_root, &key, path, 1, 0); + if (ret) + goto out; + while (1) { + slot = path->slots[0]; + l = path->nodes[0]; + btrfs_item_key_to_cpu(l, &found_key, slot); + + if (found_key.type != BTRFS_QGROUP_RELATION_KEY) + goto next2; + + if (found_key.objectid > found_key.offset) { + /* parent <- member, not needed to build config */ + /* FIXME should we omit the key completely? */ + goto next2; + } + + ret = add_relation_rb(fs_info, found_key.objectid, + found_key.offset); + if (ret) + goto out; +next2: + ret = btrfs_next_item(quota_root, path); + if (ret < 0) + goto out; + if (ret) + break; + } +out: + fs_info->qgroup_flags |= flags; + if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON)) { + fs_info->quota_enabled = 0; + fs_info->pending_quota_state = 0; + } + btrfs_free_path(path); + + return ret < 0 ? ret : 0; +} + +/* + * This is only called from close_ctree() or open_ctree(), both in single- + * treaded paths. Clean up the in-memory structures. No locking needed. + */ +void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info) +{ + struct rb_node *n; + struct btrfs_qgroup *qgroup; + struct btrfs_qgroup_list *list; + + while ((n = rb_first(&fs_info->qgroup_tree))) { + qgroup = rb_entry(n, struct btrfs_qgroup, node); + rb_erase(n, &fs_info->qgroup_tree); + + WARN_ON(!list_empty(&qgroup->dirty)); + + while (!list_empty(&qgroup->groups)) { + list = list_first_entry(&qgroup->groups, + struct btrfs_qgroup_list, + next_group); + list_del(&list->next_group); + list_del(&list->next_member); + kfree(list); + } + + while (!list_empty(&qgroup->members)) { + list = list_first_entry(&qgroup->members, + struct btrfs_qgroup_list, + next_member); + list_del(&list->next_group); + list_del(&list->next_member); + kfree(list); + } + kfree(qgroup); + } +} + +static int add_qgroup_relation_item(struct btrfs_trans_handle *trans, + struct btrfs_root *quota_root, + u64 src, u64 dst) +{ + int ret; + struct btrfs_path *path; + struct btrfs_key key; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + key.objectid = src; + key.type = BTRFS_QGROUP_RELATION_KEY; + key.offset = dst; + + ret = btrfs_insert_empty_item(trans, quota_root, path, &key, 0); + + btrfs_mark_buffer_dirty(path->nodes[0]); + + btrfs_free_path(path); + return ret; +} + +static int del_qgroup_relation_item(struct btrfs_trans_handle *trans, + struct btrfs_root *quota_root, + u64 src, u64 dst) +{ + int ret; + struct btrfs_path *path; + struct btrfs_key key; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + key.objectid = src; + key.type = BTRFS_QGROUP_RELATION_KEY; + key.offset = dst; + + ret = btrfs_search_slot(trans, quota_root, &key, path, -1, 1); + if (ret < 0) + goto out; + + if (ret > 0) { + ret = -ENOENT; + goto out; + } + + ret = btrfs_del_item(trans, quota_root, path); +out: + btrfs_free_path(path); + return ret; +} + +static int add_qgroup_item(struct btrfs_trans_handle *trans, + struct btrfs_root *quota_root, u64 qgroupid) +{ + int ret; + struct btrfs_path *path; + struct btrfs_qgroup_info_item *qgroup_info; + struct btrfs_qgroup_limit_item *qgroup_limit; + struct extent_buffer *leaf; + struct btrfs_key key; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + key.objectid = 0; + key.type = BTRFS_QGROUP_INFO_KEY; + key.offset = qgroupid; + + ret = btrfs_insert_empty_item(trans, quota_root, path, &key, + sizeof(*qgroup_info)); + if (ret) + goto out; + + leaf = path->nodes[0]; + qgroup_info = btrfs_item_ptr(leaf, path->slots[0], + struct btrfs_qgroup_info_item); + btrfs_set_qgroup_info_generation(leaf, qgroup_info, trans->transid); + btrfs_set_qgroup_info_rfer(leaf, qgroup_info, 0); + btrfs_set_qgroup_info_rfer_cmpr(leaf, qgroup_info, 0); + btrfs_set_qgroup_info_excl(leaf, qgroup_info, 0); + btrfs_set_qgroup_info_excl_cmpr(leaf, qgroup_info, 0); + + btrfs_mark_buffer_dirty(leaf); + + btrfs_release_path(path); + + key.type = BTRFS_QGROUP_LIMIT_KEY; + ret = btrfs_insert_empty_item(trans, quota_root, path, &key, + sizeof(*qgroup_limit)); + if (ret) + goto out; + + leaf = path->nodes[0]; + qgroup_limit = btrfs_item_ptr(leaf, path->slots[0], + struct btrfs_qgroup_limit_item); + btrfs_set_qgroup_limit_flags(leaf, qgroup_limit, 0); + btrfs_set_qgroup_limit_max_rfer(leaf, qgroup_limit, 0); + btrfs_set_qgroup_limit_max_excl(leaf, qgroup_limit, 0); + btrfs_set_qgroup_limit_rsv_rfer(leaf, qgroup_limit, 0); + btrfs_set_qgroup_limit_rsv_excl(leaf, qgroup_limit, 0); + + btrfs_mark_buffer_dirty(leaf); + + ret = 0; +out: + btrfs_free_path(path); + return ret; +} + +static int del_qgroup_item(struct btrfs_trans_handle *trans, + struct btrfs_root *quota_root, u64 qgroupid) +{ + int ret; + struct btrfs_path *path; + struct btrfs_key key; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + key.objectid = 0; + key.type = BTRFS_QGROUP_INFO_KEY; + key.offset = qgroupid; + ret = btrfs_search_slot(trans, quota_root, &key, path, -1, 1); + if (ret < 0) + goto out; + + if (ret > 0) { + ret = -ENOENT; + goto out; + } + + ret = btrfs_del_item(trans, quota_root, path); + if (ret) + goto out; + + btrfs_release_path(path); + + key.type = BTRFS_QGROUP_LIMIT_KEY; + ret = btrfs_search_slot(trans, quota_root, &key, path, -1, 1); + if (ret < 0) + goto out; + + if (ret > 0) { + ret = -ENOENT; + goto out; + } + + ret = btrfs_del_item(trans, quota_root, path); + +out: + btrfs_free_path(path); + return ret; +} + +static int update_qgroup_limit_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, u64 qgroupid, + u64 flags, u64 max_rfer, u64 max_excl, + u64 rsv_rfer, u64 rsv_excl) +{ + struct btrfs_path *path; + struct btrfs_key key; + struct extent_buffer *l; + struct btrfs_qgroup_limit_item *qgroup_limit; + int ret; + int slot; + + key.objectid = 0; + key.type = BTRFS_QGROUP_LIMIT_KEY; + key.offset = qgroupid; + + path = btrfs_alloc_path(); + BUG_ON(!path); + ret = btrfs_search_slot(trans, root, &key, path, 0, 1); + if (ret > 0) + ret = -ENOENT; + + if (ret) + goto out; + + l = path->nodes[0]; + slot = path->slots[0]; + qgroup_limit = btrfs_item_ptr(l, path->slots[0], + struct btrfs_qgroup_limit_item); + btrfs_set_qgroup_limit_flags(l, qgroup_limit, flags); + btrfs_set_qgroup_limit_max_rfer(l, qgroup_limit, max_rfer); + btrfs_set_qgroup_limit_max_excl(l, qgroup_limit, max_excl); + btrfs_set_qgroup_limit_rsv_rfer(l, qgroup_limit, rsv_rfer); + btrfs_set_qgroup_limit_rsv_excl(l, qgroup_limit, rsv_excl); + + btrfs_mark_buffer_dirty(l); + +out: + btrfs_free_path(path); + return ret; +} + +static int update_qgroup_info_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_qgroup *qgroup) +{ + struct btrfs_path *path; + struct btrfs_key key; + struct extent_buffer *l; + struct btrfs_qgroup_info_item *qgroup_info; + int ret; + int slot; + + key.objectid = 0; + key.type = BTRFS_QGROUP_INFO_KEY; + key.offset = qgroup->qgroupid; + + path = btrfs_alloc_path(); + BUG_ON(!path); + ret = btrfs_search_slot(trans, root, &key, path, 0, 1); + if (ret > 0) + ret = -ENOENT; + + if (ret) + goto out; + + l = path->nodes[0]; + slot = path->slots[0]; + qgroup_info = btrfs_item_ptr(l, path->slots[0], + struct btrfs_qgroup_info_item); + btrfs_set_qgroup_info_generation(l, qgroup_info, trans->transid); + btrfs_set_qgroup_info_rfer(l, qgroup_info, qgroup->rfer); + btrfs_set_qgroup_info_rfer_cmpr(l, qgroup_info, qgroup->rfer_cmpr); + btrfs_set_qgroup_info_excl(l, qgroup_info, qgroup->excl); + btrfs_set_qgroup_info_excl_cmpr(l, qgroup_info, qgroup->excl_cmpr); + + btrfs_mark_buffer_dirty(l); + +out: + btrfs_free_path(path); + return ret; +} + +static int update_qgroup_status_item(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, + struct btrfs_root *root) +{ + struct btrfs_path *path; + struct btrfs_key key; + struct extent_buffer *l; + struct btrfs_qgroup_status_item *ptr; + int ret; + int slot; + + key.objectid = 0; + key.type = BTRFS_QGROUP_STATUS_KEY; + key.offset = 0; + + path = btrfs_alloc_path(); + BUG_ON(!path); + ret = btrfs_search_slot(trans, root, &key, path, 0, 1); + if (ret > 0) + ret = -ENOENT; + + if (ret) + goto out; + + l = path->nodes[0]; + slot = path->slots[0]; + ptr = btrfs_item_ptr(l, slot, struct btrfs_qgroup_status_item); + btrfs_set_qgroup_status_flags(l, ptr, fs_info->qgroup_flags); + btrfs_set_qgroup_status_generation(l, ptr, trans->transid); + /* XXX scan */ + + btrfs_mark_buffer_dirty(l); + +out: + btrfs_free_path(path); + return ret; +} + +/* + * called with qgroup_lock held + */ +static int btrfs_clean_quota_tree(struct btrfs_trans_handle *trans, + struct btrfs_root *root) +{ + struct btrfs_path *path; + struct btrfs_key key; + int ret; + + if (!root) + return -EINVAL; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + while (1) { + key.objectid = 0; + key.offset = 0; + key.type = 0; + + path->leave_spinning = 1; + ret = btrfs_search_slot(trans, root, &key, path, -1, 1); + if (ret > 0) { + if (path->slots[0] == 0) + break; + path->slots[0]--; + } else if (ret < 0) { + break; + } + + ret = btrfs_del_item(trans, root, path); + if (ret) + goto out; + btrfs_release_path(path); + } + ret = 0; +out: + root->fs_info->pending_quota_state = 0; + btrfs_free_path(path); + return ret; +} + +int btrfs_quota_enable(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info) +{ + struct btrfs_root *quota_root; + struct btrfs_path *path = NULL; + struct btrfs_qgroup_status_item *ptr; + struct extent_buffer *leaf; + struct btrfs_key key; + int ret = 0; + + spin_lock(&fs_info->qgroup_lock); + if (fs_info->quota_root) { + fs_info->pending_quota_state = 1; + spin_unlock(&fs_info->qgroup_lock); + goto out; + } + spin_unlock(&fs_info->qgroup_lock); + + /* + * initially create the quota tree + */ + quota_root = btrfs_create_tree(trans, fs_info, + BTRFS_QUOTA_TREE_OBJECTID); + if (IS_ERR(quota_root)) { + ret = PTR_ERR(quota_root); + goto out; + } + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + key.objectid = 0; + key.type = BTRFS_QGROUP_STATUS_KEY; + key.offset = 0; + + ret = btrfs_insert_empty_item(trans, quota_root, path, &key, + sizeof(*ptr)); + if (ret) + goto out; + + leaf = path->nodes[0]; + ptr = btrfs_item_ptr(leaf, path->slots[0], + struct btrfs_qgroup_status_item); + btrfs_set_qgroup_status_generation(leaf, ptr, trans->transid); + btrfs_set_qgroup_status_version(leaf, ptr, BTRFS_QGROUP_STATUS_VERSION); + fs_info->qgroup_flags = BTRFS_QGROUP_STATUS_FLAG_ON | + BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT; + btrfs_set_qgroup_status_flags(leaf, ptr, fs_info->qgroup_flags); + btrfs_set_qgroup_status_scan(leaf, ptr, 0); + + btrfs_mark_buffer_dirty(leaf); + + spin_lock(&fs_info->qgroup_lock); + fs_info->quota_root = quota_root; + fs_info->pending_quota_state = 1; + spin_unlock(&fs_info->qgroup_lock); +out: + btrfs_free_path(path); + return ret; +} + +int btrfs_quota_disable(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info) +{ + struct btrfs_root *tree_root = fs_info->tree_root; + struct btrfs_root *quota_root; + int ret = 0; + + spin_lock(&fs_info->qgroup_lock); + fs_info->quota_enabled = 0; + fs_info->pending_quota_state = 0; + quota_root = fs_info->quota_root; + fs_info->quota_root = NULL; + btrfs_free_qgroup_config(fs_info); + spin_unlock(&fs_info->qgroup_lock); + + if (!quota_root) + return -EINVAL; + + ret = btrfs_clean_quota_tree(trans, quota_root); + if (ret) + goto out; + + ret = btrfs_del_root(trans, tree_root, "a_root->root_key); + if (ret) + goto out; + + list_del("a_root->dirty_list); + + btrfs_tree_lock(quota_root->node); + clean_tree_block(trans, tree_root, quota_root->node); + btrfs_tree_unlock(quota_root->node); + btrfs_free_tree_block(trans, quota_root, quota_root->node, 0, 1); + + free_extent_buffer(quota_root->node); + free_extent_buffer(quota_root->commit_root); + kfree(quota_root); +out: + return ret; +} + +int btrfs_quota_rescan(struct btrfs_fs_info *fs_info) +{ + /* FIXME */ + return 0; +} + +int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 src, u64 dst) +{ + struct btrfs_root *quota_root; + int ret = 0; + + quota_root = fs_info->quota_root; + if (!quota_root) + return -EINVAL; + + ret = add_qgroup_relation_item(trans, quota_root, src, dst); + if (ret) + return ret; + + ret = add_qgroup_relation_item(trans, quota_root, dst, src); + if (ret) { + del_qgroup_relation_item(trans, quota_root, src, dst); + return ret; + } + + spin_lock(&fs_info->qgroup_lock); + ret = add_relation_rb(quota_root->fs_info, src, dst); + spin_unlock(&fs_info->qgroup_lock); + + return ret; +} + +int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 src, u64 dst) +{ + struct btrfs_root *quota_root; + int ret = 0; + int err; + + quota_root = fs_info->quota_root; + if (!quota_root) + return -EINVAL; + + ret = del_qgroup_relation_item(trans, quota_root, src, dst); + err = del_qgroup_relation_item(trans, quota_root, dst, src); + if (err && !ret) + ret = err; + + spin_lock(&fs_info->qgroup_lock); + del_relation_rb(fs_info, src, dst); + + spin_unlock(&fs_info->qgroup_lock); + + return ret; +} + +int btrfs_create_qgroup(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 qgroupid, char *name) +{ + struct btrfs_root *quota_root; + struct btrfs_qgroup *qgroup; + int ret = 0; + + quota_root = fs_info->quota_root; + if (!quota_root) + return -EINVAL; + + ret = add_qgroup_item(trans, quota_root, qgroupid); + + spin_lock(&fs_info->qgroup_lock); + qgroup = add_qgroup_rb(fs_info, qgroupid); + spin_unlock(&fs_info->qgroup_lock); + + if (IS_ERR(qgroup)) + ret = PTR_ERR(qgroup); + + return ret; +} + +int btrfs_remove_qgroup(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 qgroupid) +{ + struct btrfs_root *quota_root; + int ret = 0; + + quota_root = fs_info->quota_root; + if (!quota_root) + return -EINVAL; + + ret = del_qgroup_item(trans, quota_root, qgroupid); + + spin_lock(&fs_info->qgroup_lock); + del_qgroup_rb(quota_root->fs_info, qgroupid); + + spin_unlock(&fs_info->qgroup_lock); + + return ret; +} + +int btrfs_limit_qgroup(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 qgroupid, + struct btrfs_qgroup_limit *limit) +{ + struct btrfs_root *quota_root = fs_info->quota_root; + struct btrfs_qgroup *qgroup; + int ret = 0; + + if (!quota_root) + return -EINVAL; + + ret = update_qgroup_limit_item(trans, quota_root, qgroupid, + limit->flags, limit->max_rfer, + limit->max_excl, limit->rsv_rfer, + limit->rsv_excl); + if (ret) { + fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT; + printk(KERN_INFO "unable to update quota limit for %llu\n", + (unsigned long long)qgroupid); + } + + spin_lock(&fs_info->qgroup_lock); + + qgroup = find_qgroup_rb(fs_info, qgroupid); + if (!qgroup) { + ret = -ENOENT; + goto unlock; + } + qgroup->lim_flags = limit->flags; + qgroup->max_rfer = limit->max_rfer; + qgroup->max_excl = limit->max_excl; + qgroup->rsv_rfer = limit->rsv_rfer; + qgroup->rsv_excl = limit->rsv_excl; + +unlock: + spin_unlock(&fs_info->qgroup_lock); + + return ret; +} + +static void qgroup_dirty(struct btrfs_fs_info *fs_info, + struct btrfs_qgroup *qgroup) +{ + if (list_empty(&qgroup->dirty)) + list_add(&qgroup->dirty, &fs_info->dirty_qgroups); +} + +/* + * btrfs_qgroup_record_ref is called when the ref is added or deleted. it puts + * the modification into a list that's later used by btrfs_end_transaction to + * pass the recorded modifications on to btrfs_qgroup_account_ref. + */ +int btrfs_qgroup_record_ref(struct btrfs_trans_handle *trans, + struct btrfs_delayed_ref_node *node, + struct btrfs_delayed_extent_op *extent_op) +{ + struct qgroup_update *u; + + BUG_ON(!trans->delayed_ref_elem.seq); + u = kmalloc(sizeof(*u), GFP_NOFS); + if (!u) + return -ENOMEM; + + u->node = node; + u->extent_op = extent_op; + list_add_tail(&u->list, &trans->qgroup_ref_list); + + return 0; +} + +/* + * btrfs_qgroup_account_ref is called for every ref that is added to or deleted + * from the fs. First, all roots referencing the extent are searched, and + * then the space is accounted accordingly to the different roots. The + * accounting algorithm works in 3 steps documented inline. + */ +int btrfs_qgroup_account_ref(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *node, + struct btrfs_delayed_extent_op *extent_op) +{ + struct btrfs_key ins; + struct btrfs_root *quota_root; + u64 ref_root; + struct btrfs_qgroup *qgroup; + struct ulist_node *unode; + struct ulist *roots = NULL; + struct ulist *tmp = NULL; + struct ulist_iterator uiter; + u64 seq; + int ret = 0; + int sgn; + + if (!fs_info->quota_enabled) + return 0; + + BUG_ON(!fs_info->quota_root); + + ins.objectid = node->bytenr; + ins.offset = node->num_bytes; + ins.type = BTRFS_EXTENT_ITEM_KEY; + + if (node->type == BTRFS_TREE_BLOCK_REF_KEY || + node->type == BTRFS_SHARED_BLOCK_REF_KEY) { + struct btrfs_delayed_tree_ref *ref; + ref = btrfs_delayed_node_to_tree_ref(node); + ref_root = ref->root; + } else if (node->type == BTRFS_EXTENT_DATA_REF_KEY || + node->type == BTRFS_SHARED_DATA_REF_KEY) { + struct btrfs_delayed_data_ref *ref; + ref = btrfs_delayed_node_to_data_ref(node); + ref_root = ref->root; + } else { + BUG(); + } + + if (!is_fstree(ref_root)) { + /* + * non-fs-trees are not being accounted + */ + return 0; + } + + switch (node->action) { + case BTRFS_ADD_DELAYED_REF: + case BTRFS_ADD_DELAYED_EXTENT: + sgn = 1; + break; + case BTRFS_DROP_DELAYED_REF: + sgn = -1; + break; + case BTRFS_UPDATE_DELAYED_HEAD: + return 0; + default: + BUG(); + } + + /* + * the delayed ref sequence number we pass depends on the direction of + * the operation. for add operations, we pass (node->seq - 1) to skip + * the delayed ref's current sequence number, because we need the state + * of the tree before the add operation. for delete operations, we pass + * (node->seq) to include the delayed ref's current sequence number, + * because we need the state of the tree after the delete operation. + */ + ret = btrfs_find_all_roots(trans, fs_info, node->bytenr, + sgn > 0 ? node->seq - 1 : node->seq, &roots); + if (ret < 0) + goto out; + + spin_lock(&fs_info->qgroup_lock); + quota_root = fs_info->quota_root; + if (!quota_root) + goto unlock; + + qgroup = find_qgroup_rb(fs_info, ref_root); + if (!qgroup) + goto unlock; + + /* + * step 1: for each old ref, visit all nodes once and inc refcnt + */ + tmp = ulist_alloc(GFP_ATOMIC); + if (!tmp) { + ret = -ENOMEM; + goto unlock; + } + seq = fs_info->qgroup_seq; + fs_info->qgroup_seq += roots->nnodes + 1; /* max refcnt */ + + ULIST_ITER_INIT(&uiter); + while ((unode = ulist_next(roots, &uiter))) { + struct ulist_node *tmp_unode; + struct ulist_iterator tmp_uiter; + struct btrfs_qgroup *qg; + + qg = find_qgroup_rb(fs_info, unode->val); + if (!qg) + continue; + + ulist_reinit(tmp); + /* XXX id not needed */ + ulist_add(tmp, qg->qgroupid, (unsigned long)qg, GFP_ATOMIC); + ULIST_ITER_INIT(&tmp_uiter); + while ((tmp_unode = ulist_next(tmp, &tmp_uiter))) { + struct btrfs_qgroup_list *glist; + + qg = (struct btrfs_qgroup *)tmp_unode->aux; + if (qg->refcnt < seq) + qg->refcnt = seq + 1; + else + ++qg->refcnt; + + list_for_each_entry(glist, &qg->groups, next_group) { + ulist_add(tmp, glist->group->qgroupid, + (unsigned long)glist->group, + GFP_ATOMIC); + } + } + } + + /* + * step 2: walk from the new root + */ + ulist_reinit(tmp); + ulist_add(tmp, qgroup->qgroupid, (unsigned long)qgroup, GFP_ATOMIC); + ULIST_ITER_INIT(&uiter); + while ((unode = ulist_next(tmp, &uiter))) { + struct btrfs_qgroup *qg; + struct btrfs_qgroup_list *glist; + + qg = (struct btrfs_qgroup *)unode->aux; + if (qg->refcnt < seq) { + /* not visited by step 1 */ + qg->rfer += sgn * node->num_bytes; + qg->rfer_cmpr += sgn * node->num_bytes; + if (roots->nnodes == 0) { + qg->excl += sgn * node->num_bytes; + qg->excl_cmpr += sgn * node->num_bytes; + } + qgroup_dirty(fs_info, qg); + } + WARN_ON(qg->tag >= seq); + qg->tag = seq; + + list_for_each_entry(glist, &qg->groups, next_group) { + ulist_add(tmp, glist->group->qgroupid, + (unsigned long)glist->group, GFP_ATOMIC); + } + } + + /* + * step 3: walk again from old refs + */ + ULIST_ITER_INIT(&uiter); + while ((unode = ulist_next(roots, &uiter))) { + struct btrfs_qgroup *qg; + struct ulist_node *tmp_unode; + struct ulist_iterator tmp_uiter; + + qg = find_qgroup_rb(fs_info, unode->val); + if (!qg) + continue; + + ulist_reinit(tmp); + ulist_add(tmp, qg->qgroupid, (unsigned long)qg, GFP_ATOMIC); + ULIST_ITER_INIT(&tmp_uiter); + while ((tmp_unode = ulist_next(tmp, &tmp_uiter))) { + struct btrfs_qgroup_list *glist; + + qg = (struct btrfs_qgroup *)tmp_unode->aux; + if (qg->tag == seq) + continue; + + if (qg->refcnt - seq == roots->nnodes) { + qg->excl -= sgn * node->num_bytes; + qg->excl_cmpr -= sgn * node->num_bytes; + qgroup_dirty(fs_info, qg); + } + + list_for_each_entry(glist, &qg->groups, next_group) { + ulist_add(tmp, glist->group->qgroupid, + (unsigned long)glist->group, + GFP_ATOMIC); + } + } + } + ret = 0; +unlock: + spin_unlock(&fs_info->qgroup_lock); +out: + ulist_free(roots); + ulist_free(tmp); + + return ret; +} + +/* + * called from commit_transaction. Writes all changed qgroups to disk. + */ +int btrfs_run_qgroups(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info) +{ + struct btrfs_root *quota_root = fs_info->quota_root; + int ret = 0; + + if (!quota_root) + goto out; + + fs_info->quota_enabled = fs_info->pending_quota_state; + + spin_lock(&fs_info->qgroup_lock); + while (!list_empty(&fs_info->dirty_qgroups)) { + struct btrfs_qgroup *qgroup; + qgroup = list_first_entry(&fs_info->dirty_qgroups, + struct btrfs_qgroup, dirty); + list_del_init(&qgroup->dirty); + spin_unlock(&fs_info->qgroup_lock); + ret = update_qgroup_info_item(trans, quota_root, qgroup); + if (ret) + fs_info->qgroup_flags |= + BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT; + spin_lock(&fs_info->qgroup_lock); + } + if (fs_info->quota_enabled) + fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_ON; + else + fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_ON; + spin_unlock(&fs_info->qgroup_lock); + + ret = update_qgroup_status_item(trans, fs_info, quota_root); + if (ret) + fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT; + +out: + + return ret; +} + +/* + * copy the acounting information between qgroups. This is necessary when a + * snapshot or a subvolume is created + */ +int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info, u64 srcid, u64 objectid, + struct btrfs_qgroup_inherit *inherit) +{ + int ret = 0; + int i; + u64 *i_qgroups; + struct btrfs_root *quota_root = fs_info->quota_root; + struct btrfs_qgroup *srcgroup; + struct btrfs_qgroup *dstgroup; + u32 level_size = 0; + + if (!fs_info->quota_enabled) + return 0; + + if (!quota_root) + return -EINVAL; + + /* + * create a tracking group for the subvol itself + */ + ret = add_qgroup_item(trans, quota_root, objectid); + if (ret) + goto out; + + if (inherit && inherit->flags & BTRFS_QGROUP_INHERIT_SET_LIMITS) { + ret = update_qgroup_limit_item(trans, quota_root, objectid, + inherit->lim.flags, + inherit->lim.max_rfer, + inherit->lim.max_excl, + inherit->lim.rsv_rfer, + inherit->lim.rsv_excl); + if (ret) + goto out; + } + + if (srcid) { + struct btrfs_root *srcroot; + struct btrfs_key srckey; + int srcroot_level; + + srckey.objectid = srcid; + srckey.type = BTRFS_ROOT_ITEM_KEY; + srckey.offset = (u64)-1; + srcroot = btrfs_read_fs_root_no_name(fs_info, &srckey); + if (IS_ERR(srcroot)) { + ret = PTR_ERR(srcroot); + goto out; + } + + rcu_read_lock(); + srcroot_level = btrfs_header_level(srcroot->node); + level_size = btrfs_level_size(srcroot, srcroot_level); + rcu_read_unlock(); + } + + /* + * add qgroup to all inherited groups + */ + if (inherit) { + i_qgroups = (u64 *)(inherit + 1); + for (i = 0; i < inherit->num_qgroups; ++i) { + ret = add_qgroup_relation_item(trans, quota_root, + objectid, *i_qgroups); + if (ret) + goto out; + ret = add_qgroup_relation_item(trans, quota_root, + *i_qgroups, objectid); + if (ret) + goto out; + ++i_qgroups; + } + } + + + spin_lock(&fs_info->qgroup_lock); + + dstgroup = add_qgroup_rb(fs_info, objectid); + if (IS_ERR(dstgroup)) { + ret = PTR_ERR(dstgroup); + goto unlock; + } + + if (srcid) { + srcgroup = find_qgroup_rb(fs_info, srcid); + if (!srcgroup) + goto unlock; + dstgroup->rfer = srcgroup->rfer - level_size; + dstgroup->rfer_cmpr = srcgroup->rfer_cmpr - level_size; + srcgroup->excl = level_size; + srcgroup->excl_cmpr = level_size; + qgroup_dirty(fs_info, dstgroup); + qgroup_dirty(fs_info, srcgroup); + } + + if (!inherit) + goto unlock; + + i_qgroups = (u64 *)(inherit + 1); + for (i = 0; i < inherit->num_qgroups; ++i) { + ret = add_relation_rb(quota_root->fs_info, objectid, + *i_qgroups); + if (ret) + goto unlock; + ++i_qgroups; + } + + for (i = 0; i < inherit->num_ref_copies; ++i) { + struct btrfs_qgroup *src; + struct btrfs_qgroup *dst; + + src = find_qgroup_rb(fs_info, i_qgroups[0]); + dst = find_qgroup_rb(fs_info, i_qgroups[1]); + + if (!src || !dst) { + ret = -EINVAL; + goto unlock; + } + + dst->rfer = src->rfer - level_size; + dst->rfer_cmpr = src->rfer_cmpr - level_size; + i_qgroups += 2; + } + for (i = 0; i < inherit->num_excl_copies; ++i) { + struct btrfs_qgroup *src; + struct btrfs_qgroup *dst; + + src = find_qgroup_rb(fs_info, i_qgroups[0]); + dst = find_qgroup_rb(fs_info, i_qgroups[1]); + + if (!src || !dst) { + ret = -EINVAL; + goto unlock; + } + + dst->excl = src->excl + level_size; + dst->excl_cmpr = src->excl_cmpr + level_size; + i_qgroups += 2; + } + +unlock: + spin_unlock(&fs_info->qgroup_lock); +out: + return ret; +} + +/* + * reserve some space for a qgroup and all its parents. The reservation takes + * place with start_transaction or dealloc_reserve, similar to ENOSPC + * accounting. If not enough space is available, EDQUOT is returned. + * We assume that the requested space is new for all qgroups. + */ +int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes) +{ + struct btrfs_root *quota_root; + struct btrfs_qgroup *qgroup; + struct btrfs_fs_info *fs_info = root->fs_info; + u64 ref_root = root->root_key.objectid; + int ret = 0; + struct ulist *ulist = NULL; + struct ulist_node *unode; + struct ulist_iterator uiter; + + if (!is_fstree(ref_root)) + return 0; + + if (num_bytes == 0) + return 0; + + spin_lock(&fs_info->qgroup_lock); + quota_root = fs_info->quota_root; + if (!quota_root) + goto out; + + qgroup = find_qgroup_rb(fs_info, ref_root); + if (!qgroup) + goto out; + + /* + * in a first step, we check all affected qgroups if any limits would + * be exceeded + */ + ulist = ulist_alloc(GFP_ATOMIC); + ulist_add(ulist, qgroup->qgroupid, (unsigned long)qgroup, GFP_ATOMIC); + ULIST_ITER_INIT(&uiter); + while ((unode = ulist_next(ulist, &uiter))) { + struct btrfs_qgroup *qg; + struct btrfs_qgroup_list *glist; + + qg = (struct btrfs_qgroup *)unode->aux; + + if ((qg->lim_flags & BTRFS_QGROUP_LIMIT_MAX_RFER) && + qg->reserved + qg->rfer + num_bytes > + qg->max_rfer) + ret = -EDQUOT; + + if ((qg->lim_flags & BTRFS_QGROUP_LIMIT_MAX_EXCL) && + qg->reserved + qg->excl + num_bytes > + qg->max_excl) + ret = -EDQUOT; + + list_for_each_entry(glist, &qg->groups, next_group) { + ulist_add(ulist, glist->group->qgroupid, + (unsigned long)glist->group, GFP_ATOMIC); + } + } + if (ret) + goto out; + + /* + * no limits exceeded, now record the reservation into all qgroups + */ + ULIST_ITER_INIT(&uiter); + while ((unode = ulist_next(ulist, &uiter))) { + struct btrfs_qgroup *qg; + + qg = (struct btrfs_qgroup *)unode->aux; + + qg->reserved += num_bytes; + } + +out: + spin_unlock(&fs_info->qgroup_lock); + ulist_free(ulist); + + return ret; +} + +void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes) +{ + struct btrfs_root *quota_root; + struct btrfs_qgroup *qgroup; + struct btrfs_fs_info *fs_info = root->fs_info; + struct ulist *ulist = NULL; + struct ulist_node *unode; + struct ulist_iterator uiter; + u64 ref_root = root->root_key.objectid; + + if (!is_fstree(ref_root)) + return; + + if (num_bytes == 0) + return; + + spin_lock(&fs_info->qgroup_lock); + + quota_root = fs_info->quota_root; + if (!quota_root) + goto out; + + qgroup = find_qgroup_rb(fs_info, ref_root); + if (!qgroup) + goto out; + + ulist = ulist_alloc(GFP_ATOMIC); + ulist_add(ulist, qgroup->qgroupid, (unsigned long)qgroup, GFP_ATOMIC); + ULIST_ITER_INIT(&uiter); + while ((unode = ulist_next(ulist, &uiter))) { + struct btrfs_qgroup *qg; + struct btrfs_qgroup_list *glist; + + qg = (struct btrfs_qgroup *)unode->aux; + + qg->reserved -= num_bytes; + + list_for_each_entry(glist, &qg->groups, next_group) { + ulist_add(ulist, glist->group->qgroupid, + (unsigned long)glist->group, GFP_ATOMIC); + } + } + +out: + spin_unlock(&fs_info->qgroup_lock); + ulist_free(ulist); +} + +void assert_qgroups_uptodate(struct btrfs_trans_handle *trans) +{ + if (list_empty(&trans->qgroup_ref_list) && !trans->delayed_ref_elem.seq) + return; + printk(KERN_ERR "btrfs: qgroups not uptodate in trans handle %p: list is%s empty, seq is %llu\n", + trans, list_empty(&trans->qgroup_ref_list) ? "" : " not", + trans->delayed_ref_elem.seq); + BUG(); +} diff --git a/fs/btrfs/rcu-string.h b/fs/btrfs/rcu-string.h new file mode 100644 index 000000000000..9e111e4576d4 --- /dev/null +++ b/fs/btrfs/rcu-string.h @@ -0,0 +1,56 @@ +/* + * Copyright (C) 2012 Red Hat. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +struct rcu_string { + struct rcu_head rcu; + char str[0]; +}; + +static inline struct rcu_string *rcu_string_strdup(const char *src, gfp_t mask) +{ + size_t len = strlen(src) + 1; + struct rcu_string *ret = kzalloc(sizeof(struct rcu_string) + + (len * sizeof(char)), mask); + if (!ret) + return ret; + strncpy(ret->str, src, len); + return ret; +} + +static inline void rcu_string_free(struct rcu_string *str) +{ + if (str) + kfree_rcu(str, rcu); +} + +#define printk_in_rcu(fmt, ...) do { \ + rcu_read_lock(); \ + printk(fmt, __VA_ARGS__); \ + rcu_read_unlock(); \ +} while (0) + +#define printk_ratelimited_in_rcu(fmt, ...) do { \ + rcu_read_lock(); \ + printk_ratelimited(fmt, __VA_ARGS__); \ + rcu_read_unlock(); \ +} while (0) + +#define rcu_str_deref(rcu_str) ({ \ + struct rcu_string *__str = rcu_dereference(rcu_str); \ + __str->str; \ +}) diff --git a/fs/btrfs/reada.c b/fs/btrfs/reada.c new file mode 100644 index 000000000000..48a4882d8ad5 --- /dev/null +++ b/fs/btrfs/reada.c @@ -0,0 +1,966 @@ +/* + * Copyright (C) 2011 STRATO. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/sched.h> +#include <linux/pagemap.h> +#include <linux/writeback.h> +#include <linux/blkdev.h> +#include <linux/rbtree.h> +#include <linux/slab.h> +#include <linux/workqueue.h> +#include "ctree.h" +#include "volumes.h" +#include "disk-io.h" +#include "transaction.h" + +#undef DEBUG + +/* + * This is the implementation for the generic read ahead framework. + * + * To trigger a readahead, btrfs_reada_add must be called. It will start + * a read ahead for the given range [start, end) on tree root. The returned + * handle can either be used to wait on the readahead to finish + * (btrfs_reada_wait), or to send it to the background (btrfs_reada_detach). + * + * The read ahead works as follows: + * On btrfs_reada_add, the root of the tree is inserted into a radix_tree. + * reada_start_machine will then search for extents to prefetch and trigger + * some reads. When a read finishes for a node, all contained node/leaf + * pointers that lie in the given range will also be enqueued. The reads will + * be triggered in sequential order, thus giving a big win over a naive + * enumeration. It will also make use of multi-device layouts. Each disk + * will have its on read pointer and all disks will by utilized in parallel. + * Also will no two disks read both sides of a mirror simultaneously, as this + * would waste seeking capacity. Instead both disks will read different parts + * of the filesystem. + * Any number of readaheads can be started in parallel. The read order will be + * determined globally, i.e. 2 parallel readaheads will normally finish faster + * than the 2 started one after another. + */ + +#define MAX_IN_FLIGHT 6 + +struct reada_extctl { + struct list_head list; + struct reada_control *rc; + u64 generation; +}; + +struct reada_extent { + u64 logical; + struct btrfs_key top; + u32 blocksize; + int err; + struct list_head extctl; + struct kref refcnt; + spinlock_t lock; + struct reada_zone *zones[BTRFS_MAX_MIRRORS]; + int nzones; + struct btrfs_device *scheduled_for; +}; + +struct reada_zone { + u64 start; + u64 end; + u64 elems; + struct list_head list; + spinlock_t lock; + int locked; + struct btrfs_device *device; + struct btrfs_device *devs[BTRFS_MAX_MIRRORS]; /* full list, incl + * self */ + int ndevs; + struct kref refcnt; +}; + +struct reada_machine_work { + struct btrfs_work work; + struct btrfs_fs_info *fs_info; +}; + +static void reada_extent_put(struct btrfs_fs_info *, struct reada_extent *); +static void reada_control_release(struct kref *kref); +static void reada_zone_release(struct kref *kref); +static void reada_start_machine(struct btrfs_fs_info *fs_info); +static void __reada_start_machine(struct btrfs_fs_info *fs_info); + +static int reada_add_block(struct reada_control *rc, u64 logical, + struct btrfs_key *top, int level, u64 generation); + +/* recurses */ +/* in case of err, eb might be NULL */ +static int __readahead_hook(struct btrfs_root *root, struct extent_buffer *eb, + u64 start, int err) +{ + int level = 0; + int nritems; + int i; + u64 bytenr; + u64 generation; + struct reada_extent *re; + struct btrfs_fs_info *fs_info = root->fs_info; + struct list_head list; + unsigned long index = start >> PAGE_CACHE_SHIFT; + struct btrfs_device *for_dev; + + if (eb) + level = btrfs_header_level(eb); + + /* find extent */ + spin_lock(&fs_info->reada_lock); + re = radix_tree_lookup(&fs_info->reada_tree, index); + if (re) + kref_get(&re->refcnt); + spin_unlock(&fs_info->reada_lock); + + if (!re) + return -1; + + spin_lock(&re->lock); + /* + * just take the full list from the extent. afterwards we + * don't need the lock anymore + */ + list_replace_init(&re->extctl, &list); + for_dev = re->scheduled_for; + re->scheduled_for = NULL; + spin_unlock(&re->lock); + + if (err == 0) { + nritems = level ? btrfs_header_nritems(eb) : 0; + generation = btrfs_header_generation(eb); + /* + * FIXME: currently we just set nritems to 0 if this is a leaf, + * effectively ignoring the content. In a next step we could + * trigger more readahead depending from the content, e.g. + * fetch the checksums for the extents in the leaf. + */ + } else { + /* + * this is the error case, the extent buffer has not been + * read correctly. We won't access anything from it and + * just cleanup our data structures. Effectively this will + * cut the branch below this node from read ahead. + */ + nritems = 0; + generation = 0; + } + + for (i = 0; i < nritems; i++) { + struct reada_extctl *rec; + u64 n_gen; + struct btrfs_key key; + struct btrfs_key next_key; + + btrfs_node_key_to_cpu(eb, &key, i); + if (i + 1 < nritems) + btrfs_node_key_to_cpu(eb, &next_key, i + 1); + else + next_key = re->top; + bytenr = btrfs_node_blockptr(eb, i); + n_gen = btrfs_node_ptr_generation(eb, i); + + list_for_each_entry(rec, &list, list) { + struct reada_control *rc = rec->rc; + + /* + * if the generation doesn't match, just ignore this + * extctl. This will probably cut off a branch from + * prefetch. Alternatively one could start a new (sub-) + * prefetch for this branch, starting again from root. + * FIXME: move the generation check out of this loop + */ +#ifdef DEBUG + if (rec->generation != generation) { + printk(KERN_DEBUG "generation mismatch for " + "(%llu,%d,%llu) %llu != %llu\n", + key.objectid, key.type, key.offset, + rec->generation, generation); + } +#endif + if (rec->generation == generation && + btrfs_comp_cpu_keys(&key, &rc->key_end) < 0 && + btrfs_comp_cpu_keys(&next_key, &rc->key_start) > 0) + reada_add_block(rc, bytenr, &next_key, + level - 1, n_gen); + } + } + /* + * free extctl records + */ + while (!list_empty(&list)) { + struct reada_control *rc; + struct reada_extctl *rec; + + rec = list_first_entry(&list, struct reada_extctl, list); + list_del(&rec->list); + rc = rec->rc; + kfree(rec); + + kref_get(&rc->refcnt); + if (atomic_dec_and_test(&rc->elems)) { + kref_put(&rc->refcnt, reada_control_release); + wake_up(&rc->wait); + } + kref_put(&rc->refcnt, reada_control_release); + + reada_extent_put(fs_info, re); /* one ref for each entry */ + } + reada_extent_put(fs_info, re); /* our ref */ + if (for_dev) + atomic_dec(&for_dev->reada_in_flight); + + return 0; +} + +/* + * start is passed separately in case eb in NULL, which may be the case with + * failed I/O + */ +int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb, + u64 start, int err) +{ + int ret; + + ret = __readahead_hook(root, eb, start, err); + + reada_start_machine(root->fs_info); + + return ret; +} + +static struct reada_zone *reada_find_zone(struct btrfs_fs_info *fs_info, + struct btrfs_device *dev, u64 logical, + struct btrfs_bio *bbio) +{ + int ret; + struct reada_zone *zone; + struct btrfs_block_group_cache *cache = NULL; + u64 start; + u64 end; + int i; + + zone = NULL; + spin_lock(&fs_info->reada_lock); + ret = radix_tree_gang_lookup(&dev->reada_zones, (void **)&zone, + logical >> PAGE_CACHE_SHIFT, 1); + if (ret == 1) + kref_get(&zone->refcnt); + spin_unlock(&fs_info->reada_lock); + + if (ret == 1) { + if (logical >= zone->start && logical < zone->end) + return zone; + spin_lock(&fs_info->reada_lock); + kref_put(&zone->refcnt, reada_zone_release); + spin_unlock(&fs_info->reada_lock); + } + + cache = btrfs_lookup_block_group(fs_info, logical); + if (!cache) + return NULL; + + start = cache->key.objectid; + end = start + cache->key.offset - 1; + btrfs_put_block_group(cache); + + zone = kzalloc(sizeof(*zone), GFP_NOFS); + if (!zone) + return NULL; + + zone->start = start; + zone->end = end; + INIT_LIST_HEAD(&zone->list); + spin_lock_init(&zone->lock); + zone->locked = 0; + kref_init(&zone->refcnt); + zone->elems = 0; + zone->device = dev; /* our device always sits at index 0 */ + for (i = 0; i < bbio->num_stripes; ++i) { + /* bounds have already been checked */ + zone->devs[i] = bbio->stripes[i].dev; + } + zone->ndevs = bbio->num_stripes; + + spin_lock(&fs_info->reada_lock); + ret = radix_tree_insert(&dev->reada_zones, + (unsigned long)(zone->end >> PAGE_CACHE_SHIFT), + zone); + + if (ret == -EEXIST) { + kfree(zone); + ret = radix_tree_gang_lookup(&dev->reada_zones, (void **)&zone, + logical >> PAGE_CACHE_SHIFT, 1); + if (ret == 1) + kref_get(&zone->refcnt); + } + spin_unlock(&fs_info->reada_lock); + + return zone; +} + +static struct reada_extent *reada_find_extent(struct btrfs_root *root, + u64 logical, + struct btrfs_key *top, int level) +{ + int ret; + struct reada_extent *re = NULL; + struct reada_extent *re_exist = NULL; + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; + struct btrfs_bio *bbio = NULL; + struct btrfs_device *dev; + struct btrfs_device *prev_dev; + u32 blocksize; + u64 length; + int nzones = 0; + int i; + unsigned long index = logical >> PAGE_CACHE_SHIFT; + + spin_lock(&fs_info->reada_lock); + re = radix_tree_lookup(&fs_info->reada_tree, index); + if (re) + kref_get(&re->refcnt); + spin_unlock(&fs_info->reada_lock); + + if (re) + return re; + + re = kzalloc(sizeof(*re), GFP_NOFS); + if (!re) + return NULL; + + blocksize = btrfs_level_size(root, level); + re->logical = logical; + re->blocksize = blocksize; + re->top = *top; + INIT_LIST_HEAD(&re->extctl); + spin_lock_init(&re->lock); + kref_init(&re->refcnt); + + /* + * map block + */ + length = blocksize; + ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length, &bbio, 0); + if (ret || !bbio || length < blocksize) + goto error; + + if (bbio->num_stripes > BTRFS_MAX_MIRRORS) { + printk(KERN_ERR "btrfs readahead: more than %d copies not " + "supported", BTRFS_MAX_MIRRORS); + goto error; + } + + for (nzones = 0; nzones < bbio->num_stripes; ++nzones) { + struct reada_zone *zone; + + dev = bbio->stripes[nzones].dev; + zone = reada_find_zone(fs_info, dev, logical, bbio); + if (!zone) + break; + + re->zones[nzones] = zone; + spin_lock(&zone->lock); + if (!zone->elems) + kref_get(&zone->refcnt); + ++zone->elems; + spin_unlock(&zone->lock); + spin_lock(&fs_info->reada_lock); + kref_put(&zone->refcnt, reada_zone_release); + spin_unlock(&fs_info->reada_lock); + } + re->nzones = nzones; + if (nzones == 0) { + /* not a single zone found, error and out */ + goto error; + } + + /* insert extent in reada_tree + all per-device trees, all or nothing */ + spin_lock(&fs_info->reada_lock); + ret = radix_tree_insert(&fs_info->reada_tree, index, re); + if (ret == -EEXIST) { + re_exist = radix_tree_lookup(&fs_info->reada_tree, index); + BUG_ON(!re_exist); + kref_get(&re_exist->refcnt); + spin_unlock(&fs_info->reada_lock); + goto error; + } + if (ret) { + spin_unlock(&fs_info->reada_lock); + goto error; + } + prev_dev = NULL; + for (i = 0; i < nzones; ++i) { + dev = bbio->stripes[i].dev; + if (dev == prev_dev) { + /* + * in case of DUP, just add the first zone. As both + * are on the same device, there's nothing to gain + * from adding both. + * Also, it wouldn't work, as the tree is per device + * and adding would fail with EEXIST + */ + continue; + } + prev_dev = dev; + ret = radix_tree_insert(&dev->reada_extents, index, re); + if (ret) { + while (--i >= 0) { + dev = bbio->stripes[i].dev; + BUG_ON(dev == NULL); + radix_tree_delete(&dev->reada_extents, index); + } + BUG_ON(fs_info == NULL); + radix_tree_delete(&fs_info->reada_tree, index); + spin_unlock(&fs_info->reada_lock); + goto error; + } + } + spin_unlock(&fs_info->reada_lock); + + kfree(bbio); + return re; + +error: + while (nzones) { + struct reada_zone *zone; + + --nzones; + zone = re->zones[nzones]; + kref_get(&zone->refcnt); + spin_lock(&zone->lock); + --zone->elems; + if (zone->elems == 0) { + /* + * no fs_info->reada_lock needed, as this can't be + * the last ref + */ + kref_put(&zone->refcnt, reada_zone_release); + } + spin_unlock(&zone->lock); + + spin_lock(&fs_info->reada_lock); + kref_put(&zone->refcnt, reada_zone_release); + spin_unlock(&fs_info->reada_lock); + } + kfree(bbio); + kfree(re); + return re_exist; +} + +static void reada_kref_dummy(struct kref *kr) +{ +} + +static void reada_extent_put(struct btrfs_fs_info *fs_info, + struct reada_extent *re) +{ + int i; + unsigned long index = re->logical >> PAGE_CACHE_SHIFT; + + spin_lock(&fs_info->reada_lock); + if (!kref_put(&re->refcnt, reada_kref_dummy)) { + spin_unlock(&fs_info->reada_lock); + return; + } + + radix_tree_delete(&fs_info->reada_tree, index); + for (i = 0; i < re->nzones; ++i) { + struct reada_zone *zone = re->zones[i]; + + radix_tree_delete(&zone->device->reada_extents, index); + } + + spin_unlock(&fs_info->reada_lock); + + for (i = 0; i < re->nzones; ++i) { + struct reada_zone *zone = re->zones[i]; + + kref_get(&zone->refcnt); + spin_lock(&zone->lock); + --zone->elems; + if (zone->elems == 0) { + /* no fs_info->reada_lock needed, as this can't be + * the last ref */ + kref_put(&zone->refcnt, reada_zone_release); + } + spin_unlock(&zone->lock); + + spin_lock(&fs_info->reada_lock); + kref_put(&zone->refcnt, reada_zone_release); + spin_unlock(&fs_info->reada_lock); + } + if (re->scheduled_for) + atomic_dec(&re->scheduled_for->reada_in_flight); + + kfree(re); +} + +static void reada_zone_release(struct kref *kref) +{ + struct reada_zone *zone = container_of(kref, struct reada_zone, refcnt); + + radix_tree_delete(&zone->device->reada_zones, + zone->end >> PAGE_CACHE_SHIFT); + + kfree(zone); +} + +static void reada_control_release(struct kref *kref) +{ + struct reada_control *rc = container_of(kref, struct reada_control, + refcnt); + + kfree(rc); +} + +static int reada_add_block(struct reada_control *rc, u64 logical, + struct btrfs_key *top, int level, u64 generation) +{ + struct btrfs_root *root = rc->root; + struct reada_extent *re; + struct reada_extctl *rec; + + re = reada_find_extent(root, logical, top, level); /* takes one ref */ + if (!re) + return -1; + + rec = kzalloc(sizeof(*rec), GFP_NOFS); + if (!rec) { + reada_extent_put(root->fs_info, re); + return -1; + } + + rec->rc = rc; + rec->generation = generation; + atomic_inc(&rc->elems); + + spin_lock(&re->lock); + list_add_tail(&rec->list, &re->extctl); + spin_unlock(&re->lock); + + /* leave the ref on the extent */ + + return 0; +} + +/* + * called with fs_info->reada_lock held + */ +static void reada_peer_zones_set_lock(struct reada_zone *zone, int lock) +{ + int i; + unsigned long index = zone->end >> PAGE_CACHE_SHIFT; + + for (i = 0; i < zone->ndevs; ++i) { + struct reada_zone *peer; + peer = radix_tree_lookup(&zone->devs[i]->reada_zones, index); + if (peer && peer->device != zone->device) + peer->locked = lock; + } +} + +/* + * called with fs_info->reada_lock held + */ +static int reada_pick_zone(struct btrfs_device *dev) +{ + struct reada_zone *top_zone = NULL; + struct reada_zone *top_locked_zone = NULL; + u64 top_elems = 0; + u64 top_locked_elems = 0; + unsigned long index = 0; + int ret; + + if (dev->reada_curr_zone) { + reada_peer_zones_set_lock(dev->reada_curr_zone, 0); + kref_put(&dev->reada_curr_zone->refcnt, reada_zone_release); + dev->reada_curr_zone = NULL; + } + /* pick the zone with the most elements */ + while (1) { + struct reada_zone *zone; + + ret = radix_tree_gang_lookup(&dev->reada_zones, + (void **)&zone, index, 1); + if (ret == 0) + break; + index = (zone->end >> PAGE_CACHE_SHIFT) + 1; + if (zone->locked) { + if (zone->elems > top_locked_elems) { + top_locked_elems = zone->elems; + top_locked_zone = zone; + } + } else { + if (zone->elems > top_elems) { + top_elems = zone->elems; + top_zone = zone; + } + } + } + if (top_zone) + dev->reada_curr_zone = top_zone; + else if (top_locked_zone) + dev->reada_curr_zone = top_locked_zone; + else + return 0; + + dev->reada_next = dev->reada_curr_zone->start; + kref_get(&dev->reada_curr_zone->refcnt); + reada_peer_zones_set_lock(dev->reada_curr_zone, 1); + + return 1; +} + +static int reada_start_machine_dev(struct btrfs_fs_info *fs_info, + struct btrfs_device *dev) +{ + struct reada_extent *re = NULL; + int mirror_num = 0; + struct extent_buffer *eb = NULL; + u64 logical; + u32 blocksize; + int ret; + int i; + int need_kick = 0; + + spin_lock(&fs_info->reada_lock); + if (dev->reada_curr_zone == NULL) { + ret = reada_pick_zone(dev); + if (!ret) { + spin_unlock(&fs_info->reada_lock); + return 0; + } + } + /* + * FIXME currently we issue the reads one extent at a time. If we have + * a contiguous block of extents, we could also coagulate them or use + * plugging to speed things up + */ + ret = radix_tree_gang_lookup(&dev->reada_extents, (void **)&re, + dev->reada_next >> PAGE_CACHE_SHIFT, 1); + if (ret == 0 || re->logical >= dev->reada_curr_zone->end) { + ret = reada_pick_zone(dev); + if (!ret) { + spin_unlock(&fs_info->reada_lock); + return 0; + } + re = NULL; + ret = radix_tree_gang_lookup(&dev->reada_extents, (void **)&re, + dev->reada_next >> PAGE_CACHE_SHIFT, 1); + } + if (ret == 0) { + spin_unlock(&fs_info->reada_lock); + return 0; + } + dev->reada_next = re->logical + re->blocksize; + kref_get(&re->refcnt); + + spin_unlock(&fs_info->reada_lock); + + /* + * find mirror num + */ + for (i = 0; i < re->nzones; ++i) { + if (re->zones[i]->device == dev) { + mirror_num = i + 1; + break; + } + } + logical = re->logical; + blocksize = re->blocksize; + + spin_lock(&re->lock); + if (re->scheduled_for == NULL) { + re->scheduled_for = dev; + need_kick = 1; + } + spin_unlock(&re->lock); + + reada_extent_put(fs_info, re); + + if (!need_kick) + return 0; + + atomic_inc(&dev->reada_in_flight); + ret = reada_tree_block_flagged(fs_info->extent_root, logical, blocksize, + mirror_num, &eb); + if (ret) + __readahead_hook(fs_info->extent_root, NULL, logical, ret); + else if (eb) + __readahead_hook(fs_info->extent_root, eb, eb->start, ret); + + if (eb) + free_extent_buffer(eb); + + return 1; + +} + +static void reada_start_machine_worker(struct btrfs_work *work) +{ + struct reada_machine_work *rmw; + struct btrfs_fs_info *fs_info; + int old_ioprio; + + rmw = container_of(work, struct reada_machine_work, work); + fs_info = rmw->fs_info; + + kfree(rmw); + + old_ioprio = IOPRIO_PRIO_VALUE(task_nice_ioclass(current), + task_nice_ioprio(current)); + set_task_ioprio(current, BTRFS_IOPRIO_READA); + __reada_start_machine(fs_info); + set_task_ioprio(current, old_ioprio); +} + +static void __reada_start_machine(struct btrfs_fs_info *fs_info) +{ + struct btrfs_device *device; + struct btrfs_fs_devices *fs_devices = fs_info->fs_devices; + u64 enqueued; + u64 total = 0; + int i; + + do { + enqueued = 0; + list_for_each_entry(device, &fs_devices->devices, dev_list) { + if (atomic_read(&device->reada_in_flight) < + MAX_IN_FLIGHT) + enqueued += reada_start_machine_dev(fs_info, + device); + } + total += enqueued; + } while (enqueued && total < 10000); + + if (enqueued == 0) + return; + + /* + * If everything is already in the cache, this is effectively single + * threaded. To a) not hold the caller for too long and b) to utilize + * more cores, we broke the loop above after 10000 iterations and now + * enqueue to workers to finish it. This will distribute the load to + * the cores. + */ + for (i = 0; i < 2; ++i) + reada_start_machine(fs_info); +} + +static void reada_start_machine(struct btrfs_fs_info *fs_info) +{ + struct reada_machine_work *rmw; + + rmw = kzalloc(sizeof(*rmw), GFP_NOFS); + if (!rmw) { + /* FIXME we cannot handle this properly right now */ + BUG(); + } + rmw->work.func = reada_start_machine_worker; + rmw->fs_info = fs_info; + + btrfs_queue_worker(&fs_info->readahead_workers, &rmw->work); +} + +#ifdef DEBUG +static void dump_devs(struct btrfs_fs_info *fs_info, int all) +{ + struct btrfs_device *device; + struct btrfs_fs_devices *fs_devices = fs_info->fs_devices; + unsigned long index; + int ret; + int i; + int j; + int cnt; + + spin_lock(&fs_info->reada_lock); + list_for_each_entry(device, &fs_devices->devices, dev_list) { + printk(KERN_DEBUG "dev %lld has %d in flight\n", device->devid, + atomic_read(&device->reada_in_flight)); + index = 0; + while (1) { + struct reada_zone *zone; + ret = radix_tree_gang_lookup(&device->reada_zones, + (void **)&zone, index, 1); + if (ret == 0) + break; + printk(KERN_DEBUG " zone %llu-%llu elems %llu locked " + "%d devs", zone->start, zone->end, zone->elems, + zone->locked); + for (j = 0; j < zone->ndevs; ++j) { + printk(KERN_CONT " %lld", + zone->devs[j]->devid); + } + if (device->reada_curr_zone == zone) + printk(KERN_CONT " curr off %llu", + device->reada_next - zone->start); + printk(KERN_CONT "\n"); + index = (zone->end >> PAGE_CACHE_SHIFT) + 1; + } + cnt = 0; + index = 0; + while (all) { + struct reada_extent *re = NULL; + + ret = radix_tree_gang_lookup(&device->reada_extents, + (void **)&re, index, 1); + if (ret == 0) + break; + printk(KERN_DEBUG + " re: logical %llu size %u empty %d for %lld", + re->logical, re->blocksize, + list_empty(&re->extctl), re->scheduled_for ? + re->scheduled_for->devid : -1); + + for (i = 0; i < re->nzones; ++i) { + printk(KERN_CONT " zone %llu-%llu devs", + re->zones[i]->start, + re->zones[i]->end); + for (j = 0; j < re->zones[i]->ndevs; ++j) { + printk(KERN_CONT " %lld", + re->zones[i]->devs[j]->devid); + } + } + printk(KERN_CONT "\n"); + index = (re->logical >> PAGE_CACHE_SHIFT) + 1; + if (++cnt > 15) + break; + } + } + + index = 0; + cnt = 0; + while (all) { + struct reada_extent *re = NULL; + + ret = radix_tree_gang_lookup(&fs_info->reada_tree, (void **)&re, + index, 1); + if (ret == 0) + break; + if (!re->scheduled_for) { + index = (re->logical >> PAGE_CACHE_SHIFT) + 1; + continue; + } + printk(KERN_DEBUG + "re: logical %llu size %u list empty %d for %lld", + re->logical, re->blocksize, list_empty(&re->extctl), + re->scheduled_for ? re->scheduled_for->devid : -1); + for (i = 0; i < re->nzones; ++i) { + printk(KERN_CONT " zone %llu-%llu devs", + re->zones[i]->start, + re->zones[i]->end); + for (i = 0; i < re->nzones; ++i) { + printk(KERN_CONT " zone %llu-%llu devs", + re->zones[i]->start, + re->zones[i]->end); + for (j = 0; j < re->zones[i]->ndevs; ++j) { + printk(KERN_CONT " %lld", + re->zones[i]->devs[j]->devid); + } + } + } + printk(KERN_CONT "\n"); + index = (re->logical >> PAGE_CACHE_SHIFT) + 1; + } + spin_unlock(&fs_info->reada_lock); +} +#endif + +/* + * interface + */ +struct reada_control *btrfs_reada_add(struct btrfs_root *root, + struct btrfs_key *key_start, struct btrfs_key *key_end) +{ + struct reada_control *rc; + u64 start; + u64 generation; + int level; + struct extent_buffer *node; + static struct btrfs_key max_key = { + .objectid = (u64)-1, + .type = (u8)-1, + .offset = (u64)-1 + }; + + rc = kzalloc(sizeof(*rc), GFP_NOFS); + if (!rc) + return ERR_PTR(-ENOMEM); + + rc->root = root; + rc->key_start = *key_start; + rc->key_end = *key_end; + atomic_set(&rc->elems, 0); + init_waitqueue_head(&rc->wait); + kref_init(&rc->refcnt); + kref_get(&rc->refcnt); /* one ref for having elements */ + + node = btrfs_root_node(root); + start = node->start; + level = btrfs_header_level(node); + generation = btrfs_header_generation(node); + free_extent_buffer(node); + + reada_add_block(rc, start, &max_key, level, generation); + + reada_start_machine(root->fs_info); + + return rc; +} + +#ifdef DEBUG +int btrfs_reada_wait(void *handle) +{ + struct reada_control *rc = handle; + + while (atomic_read(&rc->elems)) { + wait_event_timeout(rc->wait, atomic_read(&rc->elems) == 0, + 5 * HZ); + dump_devs(rc->root->fs_info, rc->elems < 10 ? 1 : 0); + } + + dump_devs(rc->root->fs_info, rc->elems < 10 ? 1 : 0); + + kref_put(&rc->refcnt, reada_control_release); + + return 0; +} +#else +int btrfs_reada_wait(void *handle) +{ + struct reada_control *rc = handle; + + while (atomic_read(&rc->elems)) { + wait_event(rc->wait, atomic_read(&rc->elems) == 0); + } + + kref_put(&rc->refcnt, reada_control_release); + + return 0; +} +#endif + +void btrfs_reada_detach(void *handle) +{ + struct reada_control *rc = handle; + + kref_put(&rc->refcnt, reada_control_release); +} diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 59bb1764273d..4da08652004d 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -326,6 +326,19 @@ static struct rb_node *tree_search(struct rb_root *root, u64 bytenr) return NULL; } +void backref_tree_panic(struct rb_node *rb_node, int errno, + u64 bytenr) +{ + + struct btrfs_fs_info *fs_info = NULL; + struct backref_node *bnode = rb_entry(rb_node, struct backref_node, + rb_node); + if (bnode->root) + fs_info = bnode->root->fs_info; + btrfs_panic(fs_info, errno, "Inconsistency in backref cache " + "found at offset %llu\n", (unsigned long long)bytenr); +} + /* * walk up backref nodes until reach node presents tree root */ @@ -452,7 +465,8 @@ static void update_backref_node(struct backref_cache *cache, rb_erase(&node->rb_node, &cache->rb_root); node->bytenr = bytenr; rb_node = tree_insert(&cache->rb_root, node->bytenr, &node->rb_node); - BUG_ON(rb_node); + if (rb_node) + backref_tree_panic(rb_node, -EEXIST, bytenr); } /* @@ -999,7 +1013,8 @@ next: if (!cowonly) { rb_node = tree_insert(&cache->rb_root, node->bytenr, &node->rb_node); - BUG_ON(rb_node); + if (rb_node) + backref_tree_panic(rb_node, -EEXIST, node->bytenr); list_add_tail(&node->lower, &cache->leaves); } @@ -1034,7 +1049,9 @@ next: if (!cowonly) { rb_node = tree_insert(&cache->rb_root, upper->bytenr, &upper->rb_node); - BUG_ON(rb_node); + if (rb_node) + backref_tree_panic(rb_node, -EEXIST, + upper->bytenr); } list_add_tail(&edge->list[UPPER], &upper->lower); @@ -1174,11 +1191,14 @@ static int clone_backref_node(struct btrfs_trans_handle *trans, list_add_tail(&new_edge->list[UPPER], &new_node->lower); } + } else { + list_add_tail(&new_node->lower, &cache->leaves); } rb_node = tree_insert(&cache->rb_root, new_node->bytenr, &new_node->rb_node); - BUG_ON(rb_node); + if (rb_node) + backref_tree_panic(rb_node, -EEXIST, new_node->bytenr); if (!new_node->lowest) { list_for_each_entry(new_edge, &new_node->lower, list[UPPER]) { @@ -1201,14 +1221,15 @@ fail: /* * helper to add 'address of tree root -> reloc tree' mapping */ -static int __add_reloc_root(struct btrfs_root *root) +static int __must_check __add_reloc_root(struct btrfs_root *root) { struct rb_node *rb_node; struct mapping_node *node; struct reloc_control *rc = root->fs_info->reloc_ctl; node = kmalloc(sizeof(*node), GFP_NOFS); - BUG_ON(!node); + if (!node) + return -ENOMEM; node->bytenr = root->node->start; node->data = root; @@ -1217,7 +1238,13 @@ static int __add_reloc_root(struct btrfs_root *root) rb_node = tree_insert(&rc->reloc_root_tree.rb_root, node->bytenr, &node->rb_node); spin_unlock(&rc->reloc_root_tree.lock); - BUG_ON(rb_node); + if (rb_node) { + btrfs_panic(root->fs_info, -EEXIST, "Duplicate root found " + "for start=%llu while inserting into relocation " + "tree\n", node->bytenr); + kfree(node); + return -EEXIST; + } list_add_tail(&root->root_list, &rc->reloc_roots); return 0; @@ -1250,9 +1277,12 @@ static int __update_reloc_root(struct btrfs_root *root, int del) rb_node = tree_insert(&rc->reloc_root_tree.rb_root, node->bytenr, &node->rb_node); spin_unlock(&rc->reloc_root_tree.lock); - BUG_ON(rb_node); + if (rb_node) + backref_tree_panic(rb_node, -EEXIST, node->bytenr); } else { + spin_lock(&root->fs_info->trans_lock); list_del_init(&root->root_list); + spin_unlock(&root->fs_info->trans_lock); kfree(node); } return 0; @@ -1332,6 +1362,7 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans, struct btrfs_root *reloc_root; struct reloc_control *rc = root->fs_info->reloc_ctl; int clear_rsv = 0; + int ret; if (root->reloc_root) { reloc_root = root->reloc_root; @@ -1351,7 +1382,8 @@ int btrfs_init_reloc_root(struct btrfs_trans_handle *trans, if (clear_rsv) trans->block_rsv = NULL; - __add_reloc_root(reloc_root); + ret = __add_reloc_root(reloc_root); + BUG_ON(ret < 0); root->reloc_root = reloc_root; return 0; } @@ -1575,15 +1607,14 @@ int replace_file_extents(struct btrfs_trans_handle *trans, WARN_ON(!IS_ALIGNED(end, root->sectorsize)); end--; ret = try_lock_extent(&BTRFS_I(inode)->io_tree, - key.offset, end, - GFP_NOFS); + key.offset, end); if (!ret) continue; btrfs_drop_extent_cache(inode, key.offset, end, 1); unlock_extent(&BTRFS_I(inode)->io_tree, - key.offset, end, GFP_NOFS); + key.offset, end); } } @@ -1602,12 +1633,12 @@ int replace_file_extents(struct btrfs_trans_handle *trans, ret = btrfs_inc_extent_ref(trans, root, new_bytenr, num_bytes, parent, btrfs_header_owner(leaf), - key.objectid, key.offset); + key.objectid, key.offset, 1); BUG_ON(ret); ret = btrfs_free_extent(trans, root, bytenr, num_bytes, parent, btrfs_header_owner(leaf), - key.objectid, key.offset); + key.objectid, key.offset, 1); BUG_ON(ret); } if (dirty) @@ -1776,21 +1807,23 @@ again: ret = btrfs_inc_extent_ref(trans, src, old_bytenr, blocksize, path->nodes[level]->start, - src->root_key.objectid, level - 1, 0); + src->root_key.objectid, level - 1, 0, + 1); BUG_ON(ret); ret = btrfs_inc_extent_ref(trans, dest, new_bytenr, blocksize, 0, dest->root_key.objectid, level - 1, - 0); + 0, 1); BUG_ON(ret); ret = btrfs_free_extent(trans, src, new_bytenr, blocksize, path->nodes[level]->start, - src->root_key.objectid, level - 1, 0); + src->root_key.objectid, level - 1, 0, + 1); BUG_ON(ret); ret = btrfs_free_extent(trans, dest, old_bytenr, blocksize, 0, dest->root_key.objectid, level - 1, - 0); + 0, 1); BUG_ON(ret); btrfs_unlock_up_safe(path, 0); @@ -1952,9 +1985,9 @@ static int invalidate_extent_cache(struct btrfs_root *root, } /* the lock_extent waits for readpage to complete */ - lock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS); + lock_extent(&BTRFS_I(inode)->io_tree, start, end); btrfs_drop_extent_cache(inode, start, end, 1); - unlock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS); + unlock_extent(&BTRFS_I(inode)->io_tree, start, end); } return 0; } @@ -2041,8 +2074,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc, BUG_ON(IS_ERR(trans)); trans->block_rsv = rc->block_rsv; - ret = btrfs_block_rsv_check(trans, root, rc->block_rsv, - min_reserved, 0); + ret = btrfs_block_rsv_refill(root, rc->block_rsv, min_reserved); if (ret) { BUG_ON(ret != -EAGAIN); ret = btrfs_commit_transaction(trans, root); @@ -2152,8 +2184,7 @@ int prepare_to_merge(struct reloc_control *rc, int err) again: if (!err) { num_bytes = rc->merging_rsv_size; - ret = btrfs_block_rsv_add(NULL, root, rc->block_rsv, - num_bytes); + ret = btrfs_block_rsv_add(root, rc->block_rsv, num_bytes); if (ret) err = ret; } @@ -2244,7 +2275,8 @@ again: } else { list_del_init(&reloc_root->root_list); } - btrfs_drop_snapshot(reloc_root, rc->block_rsv, 0); + ret = btrfs_drop_snapshot(reloc_root, rc->block_rsv, 0, 1); + BUG_ON(ret < 0); } if (found) { @@ -2427,7 +2459,7 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans, num_bytes = calcu_metadata_size(rc, node, 1) * 2; trans->block_rsv = rc->block_rsv; - ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes); + ret = btrfs_block_rsv_add(root, rc->block_rsv, num_bytes); if (ret) { if (ret == -EAGAIN) rc->commit_transaction = 1; @@ -2558,7 +2590,7 @@ static int do_relocation(struct btrfs_trans_handle *trans, node->eb->start, blocksize, upper->eb->start, btrfs_header_owner(upper->eb), - node->level, 0); + node->level, 0, 1); BUG_ON(ret); ret = btrfs_drop_subtree(trans, root, eb, upper->eb); @@ -2860,12 +2892,12 @@ int prealloc_file_extent_cluster(struct inode *inode, else end = cluster->end - offset; - lock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS); + lock_extent(&BTRFS_I(inode)->io_tree, start, end); num_bytes = end + 1 - start; ret = btrfs_prealloc_file_range(inode, 0, start, num_bytes, num_bytes, end + 1, &alloc_hint); - unlock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS); + unlock_extent(&BTRFS_I(inode)->io_tree, start, end); if (ret) break; nr++; @@ -2897,7 +2929,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end, em->bdev = root->fs_info->fs_devices->latest_bdev; set_bit(EXTENT_FLAG_PINNED, &em->flags); - lock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS); + lock_extent(&BTRFS_I(inode)->io_tree, start, end); while (1) { write_lock(&em_tree->lock); ret = add_extent_mapping(em_tree, em); @@ -2908,7 +2940,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end, } btrfs_drop_extent_cache(inode, start, end, 0); } - unlock_extent(&BTRFS_I(inode)->io_tree, start, end, GFP_NOFS); + unlock_extent(&BTRFS_I(inode)->io_tree, start, end); return ret; } @@ -2922,6 +2954,7 @@ static int relocate_file_extent_cluster(struct inode *inode, unsigned long last_index; struct page *page; struct file_ra_state *ra; + gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping); int nr = 0; int ret = 0; @@ -2956,7 +2989,7 @@ static int relocate_file_extent_cluster(struct inode *inode, ra, NULL, index, last_index + 1 - index); page = find_or_create_page(inode->i_mapping, index, - GFP_NOFS); + mask); if (!page) { btrfs_delalloc_release_metadata(inode, PAGE_CACHE_SIZE); @@ -2987,8 +3020,7 @@ static int relocate_file_extent_cluster(struct inode *inode, page_start = (u64)page->index << PAGE_CACHE_SHIFT; page_end = page_start + PAGE_CACHE_SIZE - 1; - lock_extent(&BTRFS_I(inode)->io_tree, - page_start, page_end, GFP_NOFS); + lock_extent(&BTRFS_I(inode)->io_tree, page_start, page_end); set_page_extent_mapped(page); @@ -3004,7 +3036,7 @@ static int relocate_file_extent_cluster(struct inode *inode, set_page_dirty(page); unlock_extent(&BTRFS_I(inode)->io_tree, - page_start, page_end, GFP_NOFS); + page_start, page_end); unlock_page(page); page_cache_release(page); @@ -3151,7 +3183,8 @@ static int add_tree_block(struct reloc_control *rc, block->key_ready = 0; rb_node = tree_insert(blocks, block->bytenr, &block->rb_node); - BUG_ON(rb_node); + if (rb_node) + backref_tree_panic(rb_node, -EEXIST, block->bytenr); return 0; } @@ -3323,8 +3356,11 @@ static int find_data_references(struct reloc_control *rc, } key.objectid = ref_objectid; - key.offset = ref_offset; key.type = BTRFS_EXTENT_DATA_KEY; + if (ref_offset > ((u64)-1 << 32)) + key.offset = 0; + else + key.offset = ref_offset; path->search_commit_root = 1; path->skip_locking = 1; @@ -3420,7 +3456,9 @@ static int find_data_references(struct reloc_control *rc, block->key_ready = 1; rb_node = tree_insert(blocks, block->bytenr, &block->rb_node); - BUG_ON(rb_node); + if (rb_node) + backref_tree_panic(rb_node, -EEXIST, + block->bytenr); } if (counted) added = 1; @@ -3645,14 +3683,11 @@ int prepare_to_relocate(struct reloc_control *rc) * btrfs_init_reloc_root will use them when there * is no reservation in transaction handle. */ - ret = btrfs_block_rsv_add(NULL, rc->extent_root, rc->block_rsv, + ret = btrfs_block_rsv_add(rc->extent_root, rc->block_rsv, rc->extent_root->nodesize * 256); if (ret) return ret; - rc->block_rsv->refill_used = 1; - btrfs_add_durable_block_rsv(rc->extent_root->fs_info, rc->block_rsv); - memset(&rc->cluster, 0, sizeof(rc->cluster)); rc->search_start = rc->block_group->key.objectid; rc->extents_found = 0; @@ -3777,10 +3812,9 @@ restart: } } - ret = btrfs_block_rsv_check(trans, rc->extent_root, - rc->block_rsv, 0, 5); + ret = btrfs_block_rsv_check(rc->extent_root, rc->block_rsv, 5); if (ret < 0) { - if (ret != -EAGAIN) { + if (ret != -ENOSPC) { err = ret; WARN_ON(1); break; @@ -4071,10 +4105,11 @@ out: static noinline_for_stack int mark_garbage_root(struct btrfs_root *root) { struct btrfs_trans_handle *trans; - int ret; + int ret, err; trans = btrfs_start_transaction(root->fs_info->tree_root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); memset(&root->root_item.drop_progress, 0, sizeof(root->root_item.drop_progress)); @@ -4082,11 +4117,11 @@ static noinline_for_stack int mark_garbage_root(struct btrfs_root *root) btrfs_set_root_refs(&root->root_item, 0); ret = btrfs_update_root(trans, root->fs_info->tree_root, &root->root_key, &root->root_item); - BUG_ON(ret); - ret = btrfs_end_transaction(trans, root->fs_info->tree_root); - BUG_ON(ret); - return 0; + err = btrfs_end_transaction(trans, root->fs_info->tree_root); + if (err) + return err; + return ret; } /* @@ -4154,7 +4189,11 @@ int btrfs_recover_relocation(struct btrfs_root *root) err = ret; goto out; } - mark_garbage_root(reloc_root); + ret = mark_garbage_root(reloc_root); + if (ret < 0) { + err = ret; + goto out; + } } } @@ -4200,13 +4239,19 @@ int btrfs_recover_relocation(struct btrfs_root *root) fs_root = read_fs_root(root->fs_info, reloc_root->root_key.offset); - BUG_ON(IS_ERR(fs_root)); + if (IS_ERR(fs_root)) { + err = PTR_ERR(fs_root); + goto out_free; + } - __add_reloc_root(reloc_root); + err = __add_reloc_root(reloc_root); + BUG_ON(err < 0); /* -ENOMEM or logic error */ fs_root->reloc_root = reloc_root; } - btrfs_commit_transaction(trans, rc->extent_root); + err = btrfs_commit_transaction(trans, rc->extent_root); + if (err) + goto out_free; merge_reloc_roots(rc); @@ -4216,7 +4261,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) if (IS_ERR(trans)) err = PTR_ERR(trans); else - btrfs_commit_transaction(trans, rc->extent_root); + err = btrfs_commit_transaction(trans, rc->extent_root); out_free: kfree(rc); out: @@ -4265,6 +4310,8 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len) disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt; ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr, disk_bytenr + len - 1, &list, 0); + if (ret) + goto out; while (!list_empty(&list)) { sums = list_entry(list.next, struct btrfs_ordered_sum, list); @@ -4282,6 +4329,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len) btrfs_add_ordered_sum(inode, ordered, sums); } +out: btrfs_put_ordered_extent(ordered); return ret; } @@ -4378,7 +4426,7 @@ void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans, * called after snapshot is created. migrate block reservation * and create reloc root for the newly created snapshot */ -void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, +int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending) { struct btrfs_root *root = pending->root; @@ -4388,7 +4436,7 @@ void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, int ret; if (!root->reloc_root) - return; + return 0; rc = root->fs_info->reloc_ctl; rc->merging_rsv_size += rc->nodes_relocated; @@ -4397,18 +4445,21 @@ void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, ret = btrfs_block_rsv_migrate(&pending->block_rsv, rc->block_rsv, rc->nodes_relocated); - BUG_ON(ret); + if (ret) + return ret; } new_root = pending->snap; reloc_root = create_reloc_root(trans, root->reloc_root, new_root->root_key.objectid); + if (IS_ERR(reloc_root)) + return PTR_ERR(reloc_root); - __add_reloc_root(reloc_root); + ret = __add_reloc_root(reloc_root); + BUG_ON(ret < 0); new_root->reloc_root = reloc_root; - if (rc->create_reloc_tree) { + if (rc->create_reloc_tree) ret = clone_backref_node(trans, rc, root, reloc_root); - BUG_ON(ret); - } + return ret; } diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index f4099904565a..10d8e4d88071 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -16,12 +16,55 @@ * Boston, MA 021110-1307, USA. */ +#include <linux/uuid.h> #include "ctree.h" #include "transaction.h" #include "disk-io.h" #include "print-tree.h" /* + * Read a root item from the tree. In case we detect a root item smaller then + * sizeof(root_item), we know it's an old version of the root structure and + * initialize all new fields to zero. The same happens if we detect mismatching + * generation numbers as then we know the root was once mounted with an older + * kernel that was not aware of the root item structure change. + */ +void btrfs_read_root_item(struct btrfs_root *root, + struct extent_buffer *eb, int slot, + struct btrfs_root_item *item) +{ + uuid_le uuid; + int len; + int need_reset = 0; + + len = btrfs_item_size_nr(eb, slot); + read_extent_buffer(eb, item, btrfs_item_ptr_offset(eb, slot), + min_t(int, len, (int)sizeof(*item))); + if (len < sizeof(*item)) + need_reset = 1; + if (!need_reset && btrfs_root_generation(item) + != btrfs_root_generation_v2(item)) { + if (btrfs_root_generation_v2(item) != 0) { + printk(KERN_WARNING "btrfs: mismatching " + "generation and generation_v2 " + "found in root item. This root " + "was probably mounted with an " + "older kernel. Resetting all " + "new fields.\n"); + } + need_reset = 1; + } + if (need_reset) { + memset(&item->generation_v2, 0, + sizeof(*item) - offsetof(struct btrfs_root_item, + generation_v2)); + + uuid_le_gen(&uuid); + memcpy(item->uuid, uuid.b, BTRFS_UUID_SIZE); + } +} + +/* * lookup the root with the highest offset for a given objectid. The key we do * find is copied into 'key'. If we find something return 0, otherwise 1, < 0 * on error. @@ -61,10 +104,10 @@ int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, goto out; } if (item) - read_extent_buffer(l, item, btrfs_item_ptr_offset(l, slot), - sizeof(*item)); + btrfs_read_root_item(root, l, slot, item); if (key) memcpy(key, &found_key, sizeof(found_key)); + ret = 0; out: btrfs_free_path(path); @@ -91,12 +134,15 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root int ret; int slot; unsigned long ptr; + int old_len; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; + ret = btrfs_search_slot(trans, root, key, path, 0, 1); if (ret < 0) - goto out; + goto out_abort; if (ret != 0) { btrfs_print_leaf(root, path->nodes[0]); @@ -109,20 +155,57 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root l = path->nodes[0]; slot = path->slots[0]; ptr = btrfs_item_ptr_offset(l, slot); + old_len = btrfs_item_size_nr(l, slot); + + /* + * If this is the first time we update the root item which originated + * from an older kernel, we need to enlarge the item size to make room + * for the added fields. + */ + if (old_len < sizeof(*item)) { + btrfs_release_path(path); + ret = btrfs_search_slot(trans, root, key, path, + -1, 1); + if (ret < 0) + goto out_abort; + ret = btrfs_del_item(trans, root, path); + if (ret < 0) + goto out_abort; + btrfs_release_path(path); + ret = btrfs_insert_empty_item(trans, root, path, + key, sizeof(*item)); + if (ret < 0) + goto out_abort; + l = path->nodes[0]; + slot = path->slots[0]; + ptr = btrfs_item_ptr_offset(l, slot); + } + + /* + * Update generation_v2 so at the next mount we know the new root + * fields are valid. + */ + btrfs_set_root_generation_v2(item, btrfs_root_generation(item)); + write_extent_buffer(l, item, ptr, sizeof(*item)); btrfs_mark_buffer_dirty(path->nodes[0]); out: btrfs_free_path(path); return ret; + +out_abort: + btrfs_abort_transaction(trans, root, ret); + goto out; } -int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root - *root, struct btrfs_key *key, struct btrfs_root_item - *item) +int btrfs_insert_root(struct btrfs_trans_handle *trans, struct btrfs_root *root, + struct btrfs_key *key, struct btrfs_root_item *item) { - int ret; - ret = btrfs_insert_item(trans, root, key, item, sizeof(*item)); - return ret; + /* + * Make sure generation v1 and v2 match. See update_root for details. + */ + btrfs_set_root_generation_v2(item, btrfs_root_generation(item)); + return btrfs_insert_item(trans, root, key, item, sizeof(*item)); } /* @@ -384,6 +467,8 @@ int btrfs_find_root_ref(struct btrfs_root *tree_root, * * For a back ref the root_id is the id of the subvol or snapshot and * ref_id is the id of the tree referencing it. + * + * Will return 0, -ENOMEM, or anything from the CoW path */ int btrfs_add_root_ref(struct btrfs_trans_handle *trans, struct btrfs_root *tree_root, @@ -407,7 +492,11 @@ int btrfs_add_root_ref(struct btrfs_trans_handle *trans, again: ret = btrfs_insert_empty_item(trans, tree_root, path, &key, sizeof(*ref) + name_len); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, tree_root, ret); + btrfs_free_path(path); + return ret; + } leaf = path->nodes[0]; ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); @@ -447,3 +536,16 @@ void btrfs_check_and_init_root_item(struct btrfs_root_item *root_item) root_item->byte_limit = 0; } } + +void btrfs_update_root_times(struct btrfs_trans_handle *trans, + struct btrfs_root *root) +{ + struct btrfs_root_item *item = &root->root_item; + struct timespec ct = CURRENT_TIME; + + spin_lock(&root->root_times_lock); + item->ctransid = cpu_to_le64(trans->transid); + item->ctime.sec = cpu_to_le64(ct.tv_sec); + item->ctime.nsec = cpu_to_le32(ct.tv_nsec); + spin_unlock(&root->root_times_lock); +} diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index a8d03d5efb5d..b223620cd5a6 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -17,10 +17,16 @@ */ #include <linux/blkdev.h> +#include <linux/ratelimit.h> #include "ctree.h" #include "volumes.h" #include "disk-io.h" #include "ordered-data.h" +#include "transaction.h" +#include "backref.h" +#include "extent_io.h" +#include "check-integrity.h" +#include "rcu-string.h" /* * This is only the first step towards a full-features scrub. It reads all @@ -29,42 +35,32 @@ * any can be found. * * Future enhancements: - * - To enhance the performance, better read-ahead strategies for the - * extent-tree can be employed. * - In case an unrepairable extent is encountered, track which files are * affected and report them - * - In case of a read error on files with nodatasum, map the file and read - * the extent to trigger a writeback of the good copy * - track and record media errors, throw out bad devices * - add a mode to also read unallocated space - * - make the prefetch cancellable */ -struct scrub_bio; -struct scrub_page; +struct scrub_block; struct scrub_dev; -static void scrub_bio_end_io(struct bio *bio, int err); -static void scrub_checksum(struct btrfs_work *work); -static int scrub_checksum_data(struct scrub_dev *sdev, - struct scrub_page *spag, void *buffer); -static int scrub_checksum_tree_block(struct scrub_dev *sdev, - struct scrub_page *spag, u64 logical, - void *buffer); -static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer); -static int scrub_fixup_check(struct scrub_bio *sbio, int ix); -static void scrub_fixup_end_io(struct bio *bio, int err); -static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector, - struct page *page); -static void scrub_fixup(struct scrub_bio *sbio, int ix); #define SCRUB_PAGES_PER_BIO 16 /* 64k per bio */ #define SCRUB_BIOS_PER_DEV 16 /* 1 MB per device in flight */ +#define SCRUB_MAX_PAGES_PER_BLOCK 16 /* 64k per node/leaf/sector */ struct scrub_page { + struct scrub_block *sblock; + struct page *page; + struct btrfs_device *dev; u64 flags; /* extent flags */ u64 generation; - u64 mirror_num; - int have_csum; + u64 logical; + u64 physical; + struct { + unsigned int mirror_num:8; + unsigned int have_csum:1; + unsigned int io_error:1; + }; u8 csum[BTRFS_CSUM_SIZE]; }; @@ -75,24 +71,43 @@ struct scrub_bio { int err; u64 logical; u64 physical; - struct scrub_page spag[SCRUB_PAGES_PER_BIO]; - u64 count; + struct scrub_page *pagev[SCRUB_PAGES_PER_BIO]; + int page_count; int next_free; struct btrfs_work work; }; +struct scrub_block { + struct scrub_page pagev[SCRUB_MAX_PAGES_PER_BLOCK]; + int page_count; + atomic_t outstanding_pages; + atomic_t ref_count; /* free mem on transition to zero */ + struct scrub_dev *sdev; + struct { + unsigned int header_error:1; + unsigned int checksum_error:1; + unsigned int no_io_error_seen:1; + unsigned int generation_error:1; /* also sets header_error */ + }; +}; + struct scrub_dev { struct scrub_bio *bios[SCRUB_BIOS_PER_DEV]; struct btrfs_device *dev; int first_free; int curr; atomic_t in_flight; + atomic_t fixup_cnt; spinlock_t list_lock; wait_queue_head_t list_wait; u16 csum_size; struct list_head csum_list; atomic_t cancel_req; int readonly; + int pages_per_bio; /* <= SCRUB_PAGES_PER_BIO */ + u32 sectorsize; + u32 nodesize; + u32 leafsize; /* * statistics */ @@ -100,6 +115,64 @@ struct scrub_dev { spinlock_t stat_lock; }; +struct scrub_fixup_nodatasum { + struct scrub_dev *sdev; + u64 logical; + struct btrfs_root *root; + struct btrfs_work work; + int mirror_num; +}; + +struct scrub_warning { + struct btrfs_path *path; + u64 extent_item_size; + char *scratch_buf; + char *msg_buf; + const char *errstr; + sector_t sector; + u64 logical; + struct btrfs_device *dev; + int msg_bufsize; + int scratch_bufsize; +}; + + +static int scrub_handle_errored_block(struct scrub_block *sblock_to_check); +static int scrub_setup_recheck_block(struct scrub_dev *sdev, + struct btrfs_mapping_tree *map_tree, + u64 length, u64 logical, + struct scrub_block *sblock); +static int scrub_recheck_block(struct btrfs_fs_info *fs_info, + struct scrub_block *sblock, int is_metadata, + int have_csum, u8 *csum, u64 generation, + u16 csum_size); +static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info, + struct scrub_block *sblock, + int is_metadata, int have_csum, + const u8 *csum, u64 generation, + u16 csum_size); +static void scrub_complete_bio_end_io(struct bio *bio, int err); +static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad, + struct scrub_block *sblock_good, + int force_write); +static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad, + struct scrub_block *sblock_good, + int page_num, int force_write); +static int scrub_checksum_data(struct scrub_block *sblock); +static int scrub_checksum_tree_block(struct scrub_block *sblock); +static int scrub_checksum_super(struct scrub_block *sblock); +static void scrub_block_get(struct scrub_block *sblock); +static void scrub_block_put(struct scrub_block *sblock); +static int scrub_add_page_to_bio(struct scrub_dev *sdev, + struct scrub_page *spage); +static int scrub_pages(struct scrub_dev *sdev, u64 logical, u64 len, + u64 physical, u64 flags, u64 gen, int mirror_num, + u8 *csum, int force); +static void scrub_bio_end_io(struct bio *bio, int err); +static void scrub_bio_end_io_worker(struct btrfs_work *work); +static void scrub_block_complete(struct scrub_block *sblock); + + static void scrub_free_csums(struct scrub_dev *sdev) { while (!list_empty(&sdev->csum_list)) { @@ -111,23 +184,6 @@ static void scrub_free_csums(struct scrub_dev *sdev) } } -static void scrub_free_bio(struct bio *bio) -{ - int i; - struct page *last_page = NULL; - - if (!bio) - return; - - for (i = 0; i < bio->bi_vcnt; ++i) { - if (bio->bi_io_vec[i].bv_page == last_page) - continue; - last_page = bio->bi_io_vec[i].bv_page; - __free_page(last_page); - } - bio_put(bio); -} - static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev) { int i; @@ -135,13 +191,23 @@ static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev) if (!sdev) return; + /* this can happen when scrub is cancelled */ + if (sdev->curr != -1) { + struct scrub_bio *sbio = sdev->bios[sdev->curr]; + + for (i = 0; i < sbio->page_count; i++) { + BUG_ON(!sbio->pagev[i]); + BUG_ON(!sbio->pagev[i]->page); + scrub_block_put(sbio->pagev[i]->sblock); + } + bio_put(sbio->bio); + } + for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) { struct scrub_bio *sbio = sdev->bios[i]; if (!sbio) break; - - scrub_free_bio(sbio->bio); kfree(sbio); } @@ -155,11 +221,16 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev) struct scrub_dev *sdev; int i; struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; + int pages_per_bio; + pages_per_bio = min_t(int, SCRUB_PAGES_PER_BIO, + bio_get_nr_vecs(dev->bdev)); sdev = kzalloc(sizeof(*sdev), GFP_NOFS); if (!sdev) goto nomem; sdev->dev = dev; + sdev->pages_per_bio = pages_per_bio; + sdev->curr = -1; for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) { struct scrub_bio *sbio; @@ -170,19 +241,22 @@ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev) sbio->index = i; sbio->sdev = sdev; - sbio->count = 0; - sbio->work.func = scrub_checksum; + sbio->page_count = 0; + sbio->work.func = scrub_bio_end_io_worker; if (i != SCRUB_BIOS_PER_DEV-1) sdev->bios[i]->next_free = i + 1; - else + else sdev->bios[i]->next_free = -1; } sdev->first_free = 0; - sdev->curr = -1; + sdev->nodesize = dev->dev_root->nodesize; + sdev->leafsize = dev->dev_root->leafsize; + sdev->sectorsize = dev->dev_root->sectorsize; atomic_set(&sdev->in_flight, 0); + atomic_set(&sdev->fixup_cnt, 0); atomic_set(&sdev->cancel_req, 0); - sdev->csum_size = btrfs_super_csum_size(&fs_info->super_copy); + sdev->csum_size = btrfs_super_csum_size(fs_info->super_copy); INIT_LIST_HEAD(&sdev->csum_list); spin_lock_init(&sdev->list_lock); @@ -195,295 +269,1070 @@ nomem: return ERR_PTR(-ENOMEM); } -/* - * scrub_recheck_error gets called when either verification of the page - * failed or the bio failed to read, e.g. with EIO. In the latter case, - * recheck_error gets called for every page in the bio, even though only - * one may be bad - */ -static void scrub_recheck_error(struct scrub_bio *sbio, int ix) +static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root, void *ctx) { - if (sbio->err) { - if (scrub_fixup_io(READ, sbio->sdev->dev->bdev, - (sbio->physical + ix * PAGE_SIZE) >> 9, - sbio->bio->bi_io_vec[ix].bv_page) == 0) { - if (scrub_fixup_check(sbio, ix) == 0) - return; - } + u64 isize; + u32 nlink; + int ret; + int i; + struct extent_buffer *eb; + struct btrfs_inode_item *inode_item; + struct scrub_warning *swarn = ctx; + struct btrfs_fs_info *fs_info = swarn->dev->dev_root->fs_info; + struct inode_fs_paths *ipath = NULL; + struct btrfs_root *local_root; + struct btrfs_key root_key; + + root_key.objectid = root; + root_key.type = BTRFS_ROOT_ITEM_KEY; + root_key.offset = (u64)-1; + local_root = btrfs_read_fs_root_no_name(fs_info, &root_key); + if (IS_ERR(local_root)) { + ret = PTR_ERR(local_root); + goto err; + } + + ret = inode_item_info(inum, 0, local_root, swarn->path); + if (ret) { + btrfs_release_path(swarn->path); + goto err; + } + + eb = swarn->path->nodes[0]; + inode_item = btrfs_item_ptr(eb, swarn->path->slots[0], + struct btrfs_inode_item); + isize = btrfs_inode_size(eb, inode_item); + nlink = btrfs_inode_nlink(eb, inode_item); + btrfs_release_path(swarn->path); + + ipath = init_ipath(4096, local_root, swarn->path); + if (IS_ERR(ipath)) { + ret = PTR_ERR(ipath); + ipath = NULL; + goto err; } + ret = paths_from_inode(inum, ipath); + + if (ret < 0) + goto err; - scrub_fixup(sbio, ix); + /* + * we deliberately ignore the bit ipath might have been too small to + * hold all of the paths here + */ + for (i = 0; i < ipath->fspath->elem_cnt; ++i) + printk_in_rcu(KERN_WARNING "btrfs: %s at logical %llu on dev " + "%s, sector %llu, root %llu, inode %llu, offset %llu, " + "length %llu, links %u (path: %s)\n", swarn->errstr, + swarn->logical, rcu_str_deref(swarn->dev->name), + (unsigned long long)swarn->sector, root, inum, offset, + min(isize - offset, (u64)PAGE_SIZE), nlink, + (char *)(unsigned long)ipath->fspath->val[i]); + + free_ipath(ipath); + return 0; + +err: + printk_in_rcu(KERN_WARNING "btrfs: %s at logical %llu on dev " + "%s, sector %llu, root %llu, inode %llu, offset %llu: path " + "resolving failed with ret=%d\n", swarn->errstr, + swarn->logical, rcu_str_deref(swarn->dev->name), + (unsigned long long)swarn->sector, root, inum, offset, ret); + + free_ipath(ipath); + return 0; } -static int scrub_fixup_check(struct scrub_bio *sbio, int ix) +static void scrub_print_warning(const char *errstr, struct scrub_block *sblock) { - int ret = 1; - struct page *page; - void *buffer; - u64 flags = sbio->spag[ix].flags; + struct btrfs_device *dev = sblock->sdev->dev; + struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; + struct btrfs_path *path; + struct btrfs_key found_key; + struct extent_buffer *eb; + struct btrfs_extent_item *ei; + struct scrub_warning swarn; + u32 item_size; + int ret; + u64 ref_root; + u8 ref_level; + unsigned long ptr = 0; + const int bufsize = 4096; + u64 extent_item_pos; - page = sbio->bio->bi_io_vec[ix].bv_page; - buffer = kmap_atomic(page, KM_USER0); - if (flags & BTRFS_EXTENT_FLAG_DATA) { - ret = scrub_checksum_data(sbio->sdev, - sbio->spag + ix, buffer); - } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { - ret = scrub_checksum_tree_block(sbio->sdev, - sbio->spag + ix, - sbio->logical + ix * PAGE_SIZE, - buffer); + path = btrfs_alloc_path(); + + swarn.scratch_buf = kmalloc(bufsize, GFP_NOFS); + swarn.msg_buf = kmalloc(bufsize, GFP_NOFS); + BUG_ON(sblock->page_count < 1); + swarn.sector = (sblock->pagev[0].physical) >> 9; + swarn.logical = sblock->pagev[0].logical; + swarn.errstr = errstr; + swarn.dev = dev; + swarn.msg_bufsize = bufsize; + swarn.scratch_bufsize = bufsize; + + if (!path || !swarn.scratch_buf || !swarn.msg_buf) + goto out; + + ret = extent_from_logical(fs_info, swarn.logical, path, &found_key); + if (ret < 0) + goto out; + + extent_item_pos = swarn.logical - found_key.objectid; + swarn.extent_item_size = found_key.offset; + + eb = path->nodes[0]; + ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item); + item_size = btrfs_item_size_nr(eb, path->slots[0]); + btrfs_release_path(path); + + if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + do { + ret = tree_backref_for_extent(&ptr, eb, ei, item_size, + &ref_root, &ref_level); + printk_in_rcu(KERN_WARNING + "btrfs: %s at logical %llu on dev %s, " + "sector %llu: metadata %s (level %d) in tree " + "%llu\n", errstr, swarn.logical, + rcu_str_deref(dev->name), + (unsigned long long)swarn.sector, + ref_level ? "node" : "leaf", + ret < 0 ? -1 : ref_level, + ret < 0 ? -1 : ref_root); + } while (ret != 1); } else { - WARN_ON(1); + swarn.path = path; + iterate_extent_inodes(fs_info, found_key.objectid, + extent_item_pos, 1, + scrub_print_warning_inode, &swarn); } - kunmap_atomic(buffer, KM_USER0); - return ret; +out: + btrfs_free_path(path); + kfree(swarn.scratch_buf); + kfree(swarn.msg_buf); } -static void scrub_fixup_end_io(struct bio *bio, int err) +static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *ctx) { - complete((struct completion *)bio->bi_private); + struct page *page = NULL; + unsigned long index; + struct scrub_fixup_nodatasum *fixup = ctx; + int ret; + int corrected = 0; + struct btrfs_key key; + struct inode *inode = NULL; + u64 end = offset + PAGE_SIZE - 1; + struct btrfs_root *local_root; + + key.objectid = root; + key.type = BTRFS_ROOT_ITEM_KEY; + key.offset = (u64)-1; + local_root = btrfs_read_fs_root_no_name(fixup->root->fs_info, &key); + if (IS_ERR(local_root)) + return PTR_ERR(local_root); + + key.type = BTRFS_INODE_ITEM_KEY; + key.objectid = inum; + key.offset = 0; + inode = btrfs_iget(fixup->root->fs_info->sb, &key, local_root, NULL); + if (IS_ERR(inode)) + return PTR_ERR(inode); + + index = offset >> PAGE_CACHE_SHIFT; + + page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); + if (!page) { + ret = -ENOMEM; + goto out; + } + + if (PageUptodate(page)) { + struct btrfs_mapping_tree *map_tree; + if (PageDirty(page)) { + /* + * we need to write the data to the defect sector. the + * data that was in that sector is not in memory, + * because the page was modified. we must not write the + * modified page to that sector. + * + * TODO: what could be done here: wait for the delalloc + * runner to write out that page (might involve + * COW) and see whether the sector is still + * referenced afterwards. + * + * For the meantime, we'll treat this error + * incorrectable, although there is a chance that a + * later scrub will find the bad sector again and that + * there's no dirty page in memory, then. + */ + ret = -EIO; + goto out; + } + map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree; + ret = repair_io_failure(map_tree, offset, PAGE_SIZE, + fixup->logical, page, + fixup->mirror_num); + unlock_page(page); + corrected = !ret; + } else { + /* + * we need to get good data first. the general readpage path + * will call repair_io_failure for us, we just have to make + * sure we read the bad mirror. + */ + ret = set_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, + EXTENT_DAMAGED, GFP_NOFS); + if (ret) { + /* set_extent_bits should give proper error */ + WARN_ON(ret > 0); + if (ret > 0) + ret = -EFAULT; + goto out; + } + + ret = extent_read_full_page(&BTRFS_I(inode)->io_tree, page, + btrfs_get_extent, + fixup->mirror_num); + wait_on_page_locked(page); + + corrected = !test_range_bit(&BTRFS_I(inode)->io_tree, offset, + end, EXTENT_DAMAGED, 0, NULL); + if (!corrected) + clear_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, + EXTENT_DAMAGED, GFP_NOFS); + } + +out: + if (page) + put_page(page); + if (inode) + iput(inode); + + if (ret < 0) + return ret; + + if (ret == 0 && corrected) { + /* + * we only need to call readpage for one of the inodes belonging + * to this extent. so make iterate_extent_inodes stop + */ + return 1; + } + + return -EIO; } -static void scrub_fixup(struct scrub_bio *sbio, int ix) +static void scrub_fixup_nodatasum(struct btrfs_work *work) { - struct scrub_dev *sdev = sbio->sdev; - struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; - struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; - struct btrfs_multi_bio *multi = NULL; - u64 logical = sbio->logical + ix * PAGE_SIZE; + int ret; + struct scrub_fixup_nodatasum *fixup; + struct scrub_dev *sdev; + struct btrfs_trans_handle *trans = NULL; + struct btrfs_fs_info *fs_info; + struct btrfs_path *path; + int uncorrectable = 0; + + fixup = container_of(work, struct scrub_fixup_nodatasum, work); + sdev = fixup->sdev; + fs_info = fixup->root->fs_info; + + path = btrfs_alloc_path(); + if (!path) { + spin_lock(&sdev->stat_lock); + ++sdev->stat.malloc_errors; + spin_unlock(&sdev->stat_lock); + uncorrectable = 1; + goto out; + } + + trans = btrfs_join_transaction(fixup->root); + if (IS_ERR(trans)) { + uncorrectable = 1; + goto out; + } + + /* + * the idea is to trigger a regular read through the standard path. we + * read a page from the (failed) logical address by specifying the + * corresponding copynum of the failed sector. thus, that readpage is + * expected to fail. + * that is the point where on-the-fly error correction will kick in + * (once it's finished) and rewrite the failed sector if a good copy + * can be found. + */ + ret = iterate_inodes_from_logical(fixup->logical, fixup->root->fs_info, + path, scrub_fixup_readpage, + fixup); + if (ret < 0) { + uncorrectable = 1; + goto out; + } + WARN_ON(ret != 1); + + spin_lock(&sdev->stat_lock); + ++sdev->stat.corrected_errors; + spin_unlock(&sdev->stat_lock); + +out: + if (trans && !IS_ERR(trans)) + btrfs_end_transaction(trans, fixup->root); + if (uncorrectable) { + spin_lock(&sdev->stat_lock); + ++sdev->stat.uncorrectable_errors; + spin_unlock(&sdev->stat_lock); + + printk_ratelimited_in_rcu(KERN_ERR + "btrfs: unable to fixup (nodatasum) error at logical %llu on dev %s\n", + (unsigned long long)fixup->logical, + rcu_str_deref(sdev->dev->name)); + } + + btrfs_free_path(path); + kfree(fixup); + + /* see caller why we're pretending to be paused in the scrub counters */ + mutex_lock(&fs_info->scrub_lock); + atomic_dec(&fs_info->scrubs_running); + atomic_dec(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + atomic_dec(&sdev->fixup_cnt); + wake_up(&fs_info->scrub_pause_wait); + wake_up(&sdev->list_wait); +} + +/* + * scrub_handle_errored_block gets called when either verification of the + * pages failed or the bio failed to read, e.g. with EIO. In the latter + * case, this function handles all pages in the bio, even though only one + * may be bad. + * The goal of this function is to repair the errored block by using the + * contents of one of the mirrors. + */ +static int scrub_handle_errored_block(struct scrub_block *sblock_to_check) +{ + struct scrub_dev *sdev = sblock_to_check->sdev; + struct btrfs_fs_info *fs_info; u64 length; - int i; + u64 logical; + u64 generation; + unsigned int failed_mirror_index; + unsigned int is_metadata; + unsigned int have_csum; + u8 *csum; + struct scrub_block *sblocks_for_recheck; /* holds one for each mirror */ + struct scrub_block *sblock_bad; int ret; - DECLARE_COMPLETION_ONSTACK(complete); + int mirror_index; + int page_num; + int success; + static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, + DEFAULT_RATELIMIT_BURST); + + BUG_ON(sblock_to_check->page_count < 1); + fs_info = sdev->dev->dev_root->fs_info; + length = sblock_to_check->page_count * PAGE_SIZE; + logical = sblock_to_check->pagev[0].logical; + generation = sblock_to_check->pagev[0].generation; + BUG_ON(sblock_to_check->pagev[0].mirror_num < 1); + failed_mirror_index = sblock_to_check->pagev[0].mirror_num - 1; + is_metadata = !(sblock_to_check->pagev[0].flags & + BTRFS_EXTENT_FLAG_DATA); + have_csum = sblock_to_check->pagev[0].have_csum; + csum = sblock_to_check->pagev[0].csum; + + /* + * read all mirrors one after the other. This includes to + * re-read the extent or metadata block that failed (that was + * the cause that this fixup code is called) another time, + * page by page this time in order to know which pages + * caused I/O errors and which ones are good (for all mirrors). + * It is the goal to handle the situation when more than one + * mirror contains I/O errors, but the errors do not + * overlap, i.e. the data can be repaired by selecting the + * pages from those mirrors without I/O error on the + * particular pages. One example (with blocks >= 2 * PAGE_SIZE) + * would be that mirror #1 has an I/O error on the first page, + * the second page is good, and mirror #2 has an I/O error on + * the second page, but the first page is good. + * Then the first page of the first mirror can be repaired by + * taking the first page of the second mirror, and the + * second page of the second mirror can be repaired by + * copying the contents of the 2nd page of the 1st mirror. + * One more note: if the pages of one mirror contain I/O + * errors, the checksum cannot be verified. In order to get + * the best data for repairing, the first attempt is to find + * a mirror without I/O errors and with a validated checksum. + * Only if this is not possible, the pages are picked from + * mirrors with I/O errors without considering the checksum. + * If the latter is the case, at the end, the checksum of the + * repaired area is verified in order to correctly maintain + * the statistics. + */ + + sblocks_for_recheck = kzalloc(BTRFS_MAX_MIRRORS * + sizeof(*sblocks_for_recheck), + GFP_NOFS); + if (!sblocks_for_recheck) { + spin_lock(&sdev->stat_lock); + sdev->stat.malloc_errors++; + sdev->stat.read_errors++; + sdev->stat.uncorrectable_errors++; + spin_unlock(&sdev->stat_lock); + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_READ_ERRS); + goto out; + } + + /* setup the context, map the logical blocks and alloc the pages */ + ret = scrub_setup_recheck_block(sdev, &fs_info->mapping_tree, length, + logical, sblocks_for_recheck); + if (ret) { + spin_lock(&sdev->stat_lock); + sdev->stat.read_errors++; + sdev->stat.uncorrectable_errors++; + spin_unlock(&sdev->stat_lock); + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_READ_ERRS); + goto out; + } + BUG_ON(failed_mirror_index >= BTRFS_MAX_MIRRORS); + sblock_bad = sblocks_for_recheck + failed_mirror_index; + + /* build and submit the bios for the failed mirror, check checksums */ + ret = scrub_recheck_block(fs_info, sblock_bad, is_metadata, have_csum, + csum, generation, sdev->csum_size); + if (ret) { + spin_lock(&sdev->stat_lock); + sdev->stat.read_errors++; + sdev->stat.uncorrectable_errors++; + spin_unlock(&sdev->stat_lock); + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_READ_ERRS); + goto out; + } - if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) && - (sbio->spag[ix].have_csum == 0)) { + if (!sblock_bad->header_error && !sblock_bad->checksum_error && + sblock_bad->no_io_error_seen) { /* - * nodatasum, don't try to fix anything - * FIXME: we can do better, open the inode and trigger a - * writeback + * the error disappeared after reading page by page, or + * the area was part of a huge bio and other parts of the + * bio caused I/O errors, or the block layer merged several + * read requests into one and the error is caused by a + * different bio (usually one of the two latter cases is + * the cause) */ - goto uncorrectable; + spin_lock(&sdev->stat_lock); + sdev->stat.unverified_errors++; + spin_unlock(&sdev->stat_lock); + + goto out; } - length = PAGE_SIZE; - ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length, - &multi, 0); - if (ret || !multi || length < PAGE_SIZE) { - printk(KERN_ERR - "scrub_fixup: btrfs_map_block failed us for %llu\n", - (unsigned long long)logical); - WARN_ON(1); - return; + if (!sblock_bad->no_io_error_seen) { + spin_lock(&sdev->stat_lock); + sdev->stat.read_errors++; + spin_unlock(&sdev->stat_lock); + if (__ratelimit(&_rs)) + scrub_print_warning("i/o error", sblock_to_check); + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_READ_ERRS); + } else if (sblock_bad->checksum_error) { + spin_lock(&sdev->stat_lock); + sdev->stat.csum_errors++; + spin_unlock(&sdev->stat_lock); + if (__ratelimit(&_rs)) + scrub_print_warning("checksum error", sblock_to_check); + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_CORRUPTION_ERRS); + } else if (sblock_bad->header_error) { + spin_lock(&sdev->stat_lock); + sdev->stat.verify_errors++; + spin_unlock(&sdev->stat_lock); + if (__ratelimit(&_rs)) + scrub_print_warning("checksum/header error", + sblock_to_check); + if (sblock_bad->generation_error) + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_GENERATION_ERRS); + else + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_CORRUPTION_ERRS); } - if (multi->num_stripes == 1) - /* there aren't any replicas */ - goto uncorrectable; + if (sdev->readonly) + goto did_not_correct_error; + + if (!is_metadata && !have_csum) { + struct scrub_fixup_nodatasum *fixup_nodatasum; + + /* + * !is_metadata and !have_csum, this means that the data + * might not be COW'ed, that it might be modified + * concurrently. The general strategy to work on the + * commit root does not help in the case when COW is not + * used. + */ + fixup_nodatasum = kzalloc(sizeof(*fixup_nodatasum), GFP_NOFS); + if (!fixup_nodatasum) + goto did_not_correct_error; + fixup_nodatasum->sdev = sdev; + fixup_nodatasum->logical = logical; + fixup_nodatasum->root = fs_info->extent_root; + fixup_nodatasum->mirror_num = failed_mirror_index + 1; + /* + * increment scrubs_running to prevent cancel requests from + * completing as long as a fixup worker is running. we must also + * increment scrubs_paused to prevent deadlocking on pause + * requests used for transactions commits (as the worker uses a + * transaction context). it is safe to regard the fixup worker + * as paused for all matters practical. effectively, we only + * avoid cancellation requests from completing. + */ + mutex_lock(&fs_info->scrub_lock); + atomic_inc(&fs_info->scrubs_running); + atomic_inc(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + atomic_inc(&sdev->fixup_cnt); + fixup_nodatasum->work.func = scrub_fixup_nodatasum; + btrfs_queue_worker(&fs_info->scrub_workers, + &fixup_nodatasum->work); + goto out; + } /* - * first find a good copy + * now build and submit the bios for the other mirrors, check + * checksums */ - for (i = 0; i < multi->num_stripes; ++i) { - if (i == sbio->spag[ix].mirror_num) + for (mirror_index = 0; + mirror_index < BTRFS_MAX_MIRRORS && + sblocks_for_recheck[mirror_index].page_count > 0; + mirror_index++) { + if (mirror_index == failed_mirror_index) continue; - if (scrub_fixup_io(READ, multi->stripes[i].dev->bdev, - multi->stripes[i].physical >> 9, - sbio->bio->bi_io_vec[ix].bv_page)) { - /* I/O-error, this is not a good copy */ + /* build and submit the bios, check checksums */ + ret = scrub_recheck_block(fs_info, + sblocks_for_recheck + mirror_index, + is_metadata, have_csum, csum, + generation, sdev->csum_size); + if (ret) + goto did_not_correct_error; + } + + /* + * first try to pick the mirror which is completely without I/O + * errors and also does not have a checksum error. + * If one is found, and if a checksum is present, the full block + * that is known to contain an error is rewritten. Afterwards + * the block is known to be corrected. + * If a mirror is found which is completely correct, and no + * checksum is present, only those pages are rewritten that had + * an I/O error in the block to be repaired, since it cannot be + * determined, which copy of the other pages is better (and it + * could happen otherwise that a correct page would be + * overwritten by a bad one). + */ + for (mirror_index = 0; + mirror_index < BTRFS_MAX_MIRRORS && + sblocks_for_recheck[mirror_index].page_count > 0; + mirror_index++) { + struct scrub_block *sblock_other = sblocks_for_recheck + + mirror_index; + + if (!sblock_other->header_error && + !sblock_other->checksum_error && + sblock_other->no_io_error_seen) { + int force_write = is_metadata || have_csum; + + ret = scrub_repair_block_from_good_copy(sblock_bad, + sblock_other, + force_write); + if (0 == ret) + goto corrected_error; + } + } + + /* + * in case of I/O errors in the area that is supposed to be + * repaired, continue by picking good copies of those pages. + * Select the good pages from mirrors to rewrite bad pages from + * the area to fix. Afterwards verify the checksum of the block + * that is supposed to be repaired. This verification step is + * only done for the purpose of statistic counting and for the + * final scrub report, whether errors remain. + * A perfect algorithm could make use of the checksum and try + * all possible combinations of pages from the different mirrors + * until the checksum verification succeeds. For example, when + * the 2nd page of mirror #1 faces I/O errors, and the 2nd page + * of mirror #2 is readable but the final checksum test fails, + * then the 2nd page of mirror #3 could be tried, whether now + * the final checksum succeedes. But this would be a rare + * exception and is therefore not implemented. At least it is + * avoided that the good copy is overwritten. + * A more useful improvement would be to pick the sectors + * without I/O error based on sector sizes (512 bytes on legacy + * disks) instead of on PAGE_SIZE. Then maybe 512 byte of one + * mirror could be repaired by taking 512 byte of a different + * mirror, even if other 512 byte sectors in the same PAGE_SIZE + * area are unreadable. + */ + + /* can only fix I/O errors from here on */ + if (sblock_bad->no_io_error_seen) + goto did_not_correct_error; + + success = 1; + for (page_num = 0; page_num < sblock_bad->page_count; page_num++) { + struct scrub_page *page_bad = sblock_bad->pagev + page_num; + + if (!page_bad->io_error) continue; + + for (mirror_index = 0; + mirror_index < BTRFS_MAX_MIRRORS && + sblocks_for_recheck[mirror_index].page_count > 0; + mirror_index++) { + struct scrub_block *sblock_other = sblocks_for_recheck + + mirror_index; + struct scrub_page *page_other = sblock_other->pagev + + page_num; + + if (!page_other->io_error) { + ret = scrub_repair_page_from_good_copy( + sblock_bad, sblock_other, page_num, 0); + if (0 == ret) { + page_bad->io_error = 0; + break; /* succeeded for this page */ + } + } } - if (scrub_fixup_check(sbio, ix) == 0) - break; + if (page_bad->io_error) { + /* did not find a mirror to copy the page from */ + success = 0; + } } - if (i == multi->num_stripes) - goto uncorrectable; - if (!sdev->readonly) { - /* - * bi_io_vec[ix].bv_page now contains good data, write it back - */ - if (scrub_fixup_io(WRITE, sdev->dev->bdev, - (sbio->physical + ix * PAGE_SIZE) >> 9, - sbio->bio->bi_io_vec[ix].bv_page)) { - /* I/O-error, writeback failed, give up */ - goto uncorrectable; + if (success) { + if (is_metadata || have_csum) { + /* + * need to verify the checksum now that all + * sectors on disk are repaired (the write + * request for data to be repaired is on its way). + * Just be lazy and use scrub_recheck_block() + * which re-reads the data before the checksum + * is verified, but most likely the data comes out + * of the page cache. + */ + ret = scrub_recheck_block(fs_info, sblock_bad, + is_metadata, have_csum, csum, + generation, sdev->csum_size); + if (!ret && !sblock_bad->header_error && + !sblock_bad->checksum_error && + sblock_bad->no_io_error_seen) + goto corrected_error; + else + goto did_not_correct_error; + } else { +corrected_error: + spin_lock(&sdev->stat_lock); + sdev->stat.corrected_errors++; + spin_unlock(&sdev->stat_lock); + printk_ratelimited_in_rcu(KERN_ERR + "btrfs: fixed up error at logical %llu on dev %s\n", + (unsigned long long)logical, + rcu_str_deref(sdev->dev->name)); } + } else { +did_not_correct_error: + spin_lock(&sdev->stat_lock); + sdev->stat.uncorrectable_errors++; + spin_unlock(&sdev->stat_lock); + printk_ratelimited_in_rcu(KERN_ERR + "btrfs: unable to fixup (regular) error at logical %llu on dev %s\n", + (unsigned long long)logical, + rcu_str_deref(sdev->dev->name)); } - kfree(multi); - spin_lock(&sdev->stat_lock); - ++sdev->stat.corrected_errors; - spin_unlock(&sdev->stat_lock); - - if (printk_ratelimit()) - printk(KERN_ERR "btrfs: fixed up at %llu\n", - (unsigned long long)logical); - return; - -uncorrectable: - kfree(multi); - spin_lock(&sdev->stat_lock); - ++sdev->stat.uncorrectable_errors; - spin_unlock(&sdev->stat_lock); +out: + if (sblocks_for_recheck) { + for (mirror_index = 0; mirror_index < BTRFS_MAX_MIRRORS; + mirror_index++) { + struct scrub_block *sblock = sblocks_for_recheck + + mirror_index; + int page_index; + + for (page_index = 0; page_index < SCRUB_PAGES_PER_BIO; + page_index++) + if (sblock->pagev[page_index].page) + __free_page( + sblock->pagev[page_index].page); + } + kfree(sblocks_for_recheck); + } - if (printk_ratelimit()) - printk(KERN_ERR "btrfs: unable to fixup at %llu\n", - (unsigned long long)logical); + return 0; } -static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector, - struct page *page) +static int scrub_setup_recheck_block(struct scrub_dev *sdev, + struct btrfs_mapping_tree *map_tree, + u64 length, u64 logical, + struct scrub_block *sblocks_for_recheck) { - struct bio *bio = NULL; + int page_index; + int mirror_index; int ret; - DECLARE_COMPLETION_ONSTACK(complete); - bio = bio_alloc(GFP_NOFS, 1); - bio->bi_bdev = bdev; - bio->bi_sector = sector; - bio_add_page(bio, page, PAGE_SIZE, 0); - bio->bi_end_io = scrub_fixup_end_io; - bio->bi_private = &complete; - submit_bio(rw, bio); + /* + * note: the three members sdev, ref_count and outstanding_pages + * are not used (and not set) in the blocks that are used for + * the recheck procedure + */ - /* this will also unplug the queue */ - wait_for_completion(&complete); + page_index = 0; + while (length > 0) { + u64 sublen = min_t(u64, length, PAGE_SIZE); + u64 mapped_length = sublen; + struct btrfs_bio *bbio = NULL; - ret = !test_bit(BIO_UPTODATE, &bio->bi_flags); - bio_put(bio); - return ret; + /* + * with a length of PAGE_SIZE, each returned stripe + * represents one mirror + */ + ret = btrfs_map_block(map_tree, WRITE, logical, &mapped_length, + &bbio, 0); + if (ret || !bbio || mapped_length < sublen) { + kfree(bbio); + return -EIO; + } + + BUG_ON(page_index >= SCRUB_PAGES_PER_BIO); + for (mirror_index = 0; mirror_index < (int)bbio->num_stripes; + mirror_index++) { + struct scrub_block *sblock; + struct scrub_page *page; + + if (mirror_index >= BTRFS_MAX_MIRRORS) + continue; + + sblock = sblocks_for_recheck + mirror_index; + page = sblock->pagev + page_index; + page->logical = logical; + page->physical = bbio->stripes[mirror_index].physical; + /* for missing devices, dev->bdev is NULL */ + page->dev = bbio->stripes[mirror_index].dev; + page->mirror_num = mirror_index + 1; + page->page = alloc_page(GFP_NOFS); + if (!page->page) { + spin_lock(&sdev->stat_lock); + sdev->stat.malloc_errors++; + spin_unlock(&sdev->stat_lock); + return -ENOMEM; + } + sblock->page_count++; + } + kfree(bbio); + length -= sublen; + logical += sublen; + page_index++; + } + + return 0; } -static void scrub_bio_end_io(struct bio *bio, int err) +/* + * this function will check the on disk data for checksum errors, header + * errors and read I/O errors. If any I/O errors happen, the exact pages + * which are errored are marked as being bad. The goal is to enable scrub + * to take those pages that are not errored from all the mirrors so that + * the pages that are errored in the just handled mirror can be repaired. + */ +static int scrub_recheck_block(struct btrfs_fs_info *fs_info, + struct scrub_block *sblock, int is_metadata, + int have_csum, u8 *csum, u64 generation, + u16 csum_size) { - struct scrub_bio *sbio = bio->bi_private; - struct scrub_dev *sdev = sbio->sdev; - struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + int page_num; - sbio->err = err; - sbio->bio = bio; + sblock->no_io_error_seen = 1; + sblock->header_error = 0; + sblock->checksum_error = 0; - btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work); + for (page_num = 0; page_num < sblock->page_count; page_num++) { + struct bio *bio; + int ret; + struct scrub_page *page = sblock->pagev + page_num; + DECLARE_COMPLETION_ONSTACK(complete); + + if (page->dev->bdev == NULL) { + page->io_error = 1; + sblock->no_io_error_seen = 0; + continue; + } + + BUG_ON(!page->page); + bio = bio_alloc(GFP_NOFS, 1); + if (!bio) + return -EIO; + bio->bi_bdev = page->dev->bdev; + bio->bi_sector = page->physical >> 9; + bio->bi_end_io = scrub_complete_bio_end_io; + bio->bi_private = &complete; + + ret = bio_add_page(bio, page->page, PAGE_SIZE, 0); + if (PAGE_SIZE != ret) { + bio_put(bio); + return -EIO; + } + btrfsic_submit_bio(READ, bio); + + /* this will also unplug the queue */ + wait_for_completion(&complete); + + page->io_error = !test_bit(BIO_UPTODATE, &bio->bi_flags); + if (!test_bit(BIO_UPTODATE, &bio->bi_flags)) + sblock->no_io_error_seen = 0; + bio_put(bio); + } + + if (sblock->no_io_error_seen) + scrub_recheck_block_checksum(fs_info, sblock, is_metadata, + have_csum, csum, generation, + csum_size); + + return 0; } -static void scrub_checksum(struct btrfs_work *work) +static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info, + struct scrub_block *sblock, + int is_metadata, int have_csum, + const u8 *csum, u64 generation, + u16 csum_size) { - struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); - struct scrub_dev *sdev = sbio->sdev; - struct page *page; - void *buffer; - int i; - u64 flags; - u64 logical; - int ret; + int page_num; + u8 calculated_csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + struct btrfs_root *root = fs_info->extent_root; + void *mapped_buffer; + + BUG_ON(!sblock->pagev[0].page); + if (is_metadata) { + struct btrfs_header *h; + + mapped_buffer = kmap_atomic(sblock->pagev[0].page); + h = (struct btrfs_header *)mapped_buffer; + + if (sblock->pagev[0].logical != le64_to_cpu(h->bytenr) || + memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE) || + memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid, + BTRFS_UUID_SIZE)) { + sblock->header_error = 1; + } else if (generation != le64_to_cpu(h->generation)) { + sblock->header_error = 1; + sblock->generation_error = 1; + } + csum = h->csum; + } else { + if (!have_csum) + return; - if (sbio->err) { - for (i = 0; i < sbio->count; ++i) - scrub_recheck_error(sbio, i); + mapped_buffer = kmap_atomic(sblock->pagev[0].page); + } - sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1); - sbio->bio->bi_flags |= 1 << BIO_UPTODATE; - sbio->bio->bi_phys_segments = 0; - sbio->bio->bi_idx = 0; + for (page_num = 0;;) { + if (page_num == 0 && is_metadata) + crc = btrfs_csum_data(root, + ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE, + crc, PAGE_SIZE - BTRFS_CSUM_SIZE); + else + crc = btrfs_csum_data(root, mapped_buffer, crc, + PAGE_SIZE); + + kunmap_atomic(mapped_buffer); + page_num++; + if (page_num >= sblock->page_count) + break; + BUG_ON(!sblock->pagev[page_num].page); - for (i = 0; i < sbio->count; i++) { - struct bio_vec *bi; - bi = &sbio->bio->bi_io_vec[i]; - bi->bv_offset = 0; - bi->bv_len = PAGE_SIZE; - } + mapped_buffer = kmap_atomic(sblock->pagev[page_num].page); + } - spin_lock(&sdev->stat_lock); - ++sdev->stat.read_errors; - spin_unlock(&sdev->stat_lock); - goto out; + btrfs_csum_final(crc, calculated_csum); + if (memcmp(calculated_csum, csum, csum_size)) + sblock->checksum_error = 1; +} + +static void scrub_complete_bio_end_io(struct bio *bio, int err) +{ + complete((struct completion *)bio->bi_private); +} + +static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad, + struct scrub_block *sblock_good, + int force_write) +{ + int page_num; + int ret = 0; + + for (page_num = 0; page_num < sblock_bad->page_count; page_num++) { + int ret_sub; + + ret_sub = scrub_repair_page_from_good_copy(sblock_bad, + sblock_good, + page_num, + force_write); + if (ret_sub) + ret = ret_sub; } - for (i = 0; i < sbio->count; ++i) { - page = sbio->bio->bi_io_vec[i].bv_page; - buffer = kmap_atomic(page, KM_USER0); - flags = sbio->spag[i].flags; - logical = sbio->logical + i * PAGE_SIZE; - ret = 0; - if (flags & BTRFS_EXTENT_FLAG_DATA) { - ret = scrub_checksum_data(sdev, sbio->spag + i, buffer); - } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { - ret = scrub_checksum_tree_block(sdev, sbio->spag + i, - logical, buffer); - } else if (flags & BTRFS_EXTENT_FLAG_SUPER) { - BUG_ON(i); - (void)scrub_checksum_super(sbio, buffer); - } else { - WARN_ON(1); + + return ret; +} + +static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad, + struct scrub_block *sblock_good, + int page_num, int force_write) +{ + struct scrub_page *page_bad = sblock_bad->pagev + page_num; + struct scrub_page *page_good = sblock_good->pagev + page_num; + + BUG_ON(sblock_bad->pagev[page_num].page == NULL); + BUG_ON(sblock_good->pagev[page_num].page == NULL); + if (force_write || sblock_bad->header_error || + sblock_bad->checksum_error || page_bad->io_error) { + struct bio *bio; + int ret; + DECLARE_COMPLETION_ONSTACK(complete); + + bio = bio_alloc(GFP_NOFS, 1); + if (!bio) + return -EIO; + bio->bi_bdev = page_bad->dev->bdev; + bio->bi_sector = page_bad->physical >> 9; + bio->bi_end_io = scrub_complete_bio_end_io; + bio->bi_private = &complete; + + ret = bio_add_page(bio, page_good->page, PAGE_SIZE, 0); + if (PAGE_SIZE != ret) { + bio_put(bio); + return -EIO; } - kunmap_atomic(buffer, KM_USER0); - if (ret) - scrub_recheck_error(sbio, i); + btrfsic_submit_bio(WRITE, bio); + + /* this will also unplug the queue */ + wait_for_completion(&complete); + if (!bio_flagged(bio, BIO_UPTODATE)) { + btrfs_dev_stat_inc_and_print(page_bad->dev, + BTRFS_DEV_STAT_WRITE_ERRS); + bio_put(bio); + return -EIO; + } + bio_put(bio); } -out: - scrub_free_bio(sbio->bio); - sbio->bio = NULL; - spin_lock(&sdev->list_lock); - sbio->next_free = sdev->first_free; - sdev->first_free = sbio->index; - spin_unlock(&sdev->list_lock); - atomic_dec(&sdev->in_flight); - wake_up(&sdev->list_wait); + return 0; +} + +static void scrub_checksum(struct scrub_block *sblock) +{ + u64 flags; + int ret; + + BUG_ON(sblock->page_count < 1); + flags = sblock->pagev[0].flags; + ret = 0; + if (flags & BTRFS_EXTENT_FLAG_DATA) + ret = scrub_checksum_data(sblock); + else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) + ret = scrub_checksum_tree_block(sblock); + else if (flags & BTRFS_EXTENT_FLAG_SUPER) + (void)scrub_checksum_super(sblock); + else + WARN_ON(1); + if (ret) + scrub_handle_errored_block(sblock); } -static int scrub_checksum_data(struct scrub_dev *sdev, - struct scrub_page *spag, void *buffer) +static int scrub_checksum_data(struct scrub_block *sblock) { + struct scrub_dev *sdev = sblock->sdev; u8 csum[BTRFS_CSUM_SIZE]; + u8 *on_disk_csum; + struct page *page; + void *buffer; u32 crc = ~(u32)0; int fail = 0; struct btrfs_root *root = sdev->dev->dev_root; + u64 len; + int index; - if (!spag->have_csum) + BUG_ON(sblock->page_count < 1); + if (!sblock->pagev[0].have_csum) return 0; - crc = btrfs_csum_data(root, buffer, crc, PAGE_SIZE); + on_disk_csum = sblock->pagev[0].csum; + page = sblock->pagev[0].page; + buffer = kmap_atomic(page); + + len = sdev->sectorsize; + index = 0; + for (;;) { + u64 l = min_t(u64, len, PAGE_SIZE); + + crc = btrfs_csum_data(root, buffer, crc, l); + kunmap_atomic(buffer); + len -= l; + if (len == 0) + break; + index++; + BUG_ON(index >= sblock->page_count); + BUG_ON(!sblock->pagev[index].page); + page = sblock->pagev[index].page; + buffer = kmap_atomic(page); + } + btrfs_csum_final(crc, csum); - if (memcmp(csum, spag->csum, sdev->csum_size)) + if (memcmp(csum, on_disk_csum, sdev->csum_size)) fail = 1; - spin_lock(&sdev->stat_lock); - ++sdev->stat.data_extents_scrubbed; - sdev->stat.data_bytes_scrubbed += PAGE_SIZE; - if (fail) - ++sdev->stat.csum_errors; - spin_unlock(&sdev->stat_lock); - return fail; } -static int scrub_checksum_tree_block(struct scrub_dev *sdev, - struct scrub_page *spag, u64 logical, - void *buffer) +static int scrub_checksum_tree_block(struct scrub_block *sblock) { + struct scrub_dev *sdev = sblock->sdev; struct btrfs_header *h; struct btrfs_root *root = sdev->dev->dev_root; struct btrfs_fs_info *fs_info = root->fs_info; - u8 csum[BTRFS_CSUM_SIZE]; + u8 calculated_csum[BTRFS_CSUM_SIZE]; + u8 on_disk_csum[BTRFS_CSUM_SIZE]; + struct page *page; + void *mapped_buffer; + u64 mapped_size; + void *p; u32 crc = ~(u32)0; int fail = 0; int crc_fail = 0; + u64 len; + int index; + + BUG_ON(sblock->page_count < 1); + page = sblock->pagev[0].page; + mapped_buffer = kmap_atomic(page); + h = (struct btrfs_header *)mapped_buffer; + memcpy(on_disk_csum, h->csum, sdev->csum_size); /* * we don't use the getter functions here, as we * a) don't have an extent buffer and * b) the page is already kmapped */ - h = (struct btrfs_header *)buffer; - if (logical != le64_to_cpu(h->bytenr)) + if (sblock->pagev[0].logical != le64_to_cpu(h->bytenr)) ++fail; - if (spag->generation != le64_to_cpu(h->generation)) + if (sblock->pagev[0].generation != le64_to_cpu(h->generation)) ++fail; if (memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE)) @@ -493,54 +1342,94 @@ static int scrub_checksum_tree_block(struct scrub_dev *sdev, BTRFS_UUID_SIZE)) ++fail; - crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc, - PAGE_SIZE - BTRFS_CSUM_SIZE); - btrfs_csum_final(crc, csum); - if (memcmp(csum, h->csum, sdev->csum_size)) - ++crc_fail; + BUG_ON(sdev->nodesize != sdev->leafsize); + len = sdev->nodesize - BTRFS_CSUM_SIZE; + mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE; + p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE; + index = 0; + for (;;) { + u64 l = min_t(u64, len, mapped_size); - spin_lock(&sdev->stat_lock); - ++sdev->stat.tree_extents_scrubbed; - sdev->stat.tree_bytes_scrubbed += PAGE_SIZE; - if (crc_fail) - ++sdev->stat.csum_errors; - if (fail) - ++sdev->stat.verify_errors; - spin_unlock(&sdev->stat_lock); + crc = btrfs_csum_data(root, p, crc, l); + kunmap_atomic(mapped_buffer); + len -= l; + if (len == 0) + break; + index++; + BUG_ON(index >= sblock->page_count); + BUG_ON(!sblock->pagev[index].page); + page = sblock->pagev[index].page; + mapped_buffer = kmap_atomic(page); + mapped_size = PAGE_SIZE; + p = mapped_buffer; + } + + btrfs_csum_final(crc, calculated_csum); + if (memcmp(calculated_csum, on_disk_csum, sdev->csum_size)) + ++crc_fail; return fail || crc_fail; } -static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer) +static int scrub_checksum_super(struct scrub_block *sblock) { struct btrfs_super_block *s; - u64 logical; - struct scrub_dev *sdev = sbio->sdev; + struct scrub_dev *sdev = sblock->sdev; struct btrfs_root *root = sdev->dev->dev_root; struct btrfs_fs_info *fs_info = root->fs_info; - u8 csum[BTRFS_CSUM_SIZE]; + u8 calculated_csum[BTRFS_CSUM_SIZE]; + u8 on_disk_csum[BTRFS_CSUM_SIZE]; + struct page *page; + void *mapped_buffer; + u64 mapped_size; + void *p; u32 crc = ~(u32)0; - int fail = 0; + int fail_gen = 0; + int fail_cor = 0; + u64 len; + int index; - s = (struct btrfs_super_block *)buffer; - logical = sbio->logical; + BUG_ON(sblock->page_count < 1); + page = sblock->pagev[0].page; + mapped_buffer = kmap_atomic(page); + s = (struct btrfs_super_block *)mapped_buffer; + memcpy(on_disk_csum, s->csum, sdev->csum_size); - if (logical != le64_to_cpu(s->bytenr)) - ++fail; + if (sblock->pagev[0].logical != le64_to_cpu(s->bytenr)) + ++fail_cor; - if (sbio->spag[0].generation != le64_to_cpu(s->generation)) - ++fail; + if (sblock->pagev[0].generation != le64_to_cpu(s->generation)) + ++fail_gen; if (memcmp(s->fsid, fs_info->fsid, BTRFS_UUID_SIZE)) - ++fail; + ++fail_cor; - crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc, - PAGE_SIZE - BTRFS_CSUM_SIZE); - btrfs_csum_final(crc, csum); - if (memcmp(csum, s->csum, sbio->sdev->csum_size)) - ++fail; + len = BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE; + mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE; + p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE; + index = 0; + for (;;) { + u64 l = min_t(u64, len, mapped_size); + + crc = btrfs_csum_data(root, p, crc, l); + kunmap_atomic(mapped_buffer); + len -= l; + if (len == 0) + break; + index++; + BUG_ON(index >= sblock->page_count); + BUG_ON(!sblock->pagev[index].page); + page = sblock->pagev[index].page; + mapped_buffer = kmap_atomic(page); + mapped_size = PAGE_SIZE; + p = mapped_buffer; + } - if (fail) { + btrfs_csum_final(crc, calculated_csum); + if (memcmp(calculated_csum, on_disk_csum, sdev->csum_size)) + ++fail_cor; + + if (fail_cor + fail_gen) { /* * if we find an error in a super block, we just report it. * They will get written with the next transaction commit @@ -549,65 +1438,54 @@ static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer) spin_lock(&sdev->stat_lock); ++sdev->stat.super_errors; spin_unlock(&sdev->stat_lock); + if (fail_cor) + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_CORRUPTION_ERRS); + else + btrfs_dev_stat_inc_and_print(sdev->dev, + BTRFS_DEV_STAT_GENERATION_ERRS); } - return fail; + return fail_cor + fail_gen; } -static int scrub_submit(struct scrub_dev *sdev) +static void scrub_block_get(struct scrub_block *sblock) { - struct scrub_bio *sbio; - struct bio *bio; - int i; - - if (sdev->curr == -1) - return 0; - - sbio = sdev->bios[sdev->curr]; - - bio = bio_alloc(GFP_NOFS, sbio->count); - if (!bio) - goto nomem; + atomic_inc(&sblock->ref_count); +} - bio->bi_private = sbio; - bio->bi_end_io = scrub_bio_end_io; - bio->bi_bdev = sdev->dev->bdev; - bio->bi_sector = sbio->physical >> 9; +static void scrub_block_put(struct scrub_block *sblock) +{ + if (atomic_dec_and_test(&sblock->ref_count)) { + int i; - for (i = 0; i < sbio->count; ++i) { - struct page *page; - int ret; + for (i = 0; i < sblock->page_count; i++) + if (sblock->pagev[i].page) + __free_page(sblock->pagev[i].page); + kfree(sblock); + } +} - page = alloc_page(GFP_NOFS); - if (!page) - goto nomem; +static void scrub_submit(struct scrub_dev *sdev) +{ + struct scrub_bio *sbio; - ret = bio_add_page(bio, page, PAGE_SIZE, 0); - if (!ret) { - __free_page(page); - goto nomem; - } - } + if (sdev->curr == -1) + return; - sbio->err = 0; + sbio = sdev->bios[sdev->curr]; sdev->curr = -1; atomic_inc(&sdev->in_flight); - submit_bio(READ, bio); - - return 0; - -nomem: - scrub_free_bio(bio); - - return -ENOMEM; + btrfsic_submit_bio(READ, sbio->bio); } -static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len, - u64 physical, u64 flags, u64 gen, u64 mirror_num, - u8 *csum, int force) +static int scrub_add_page_to_bio(struct scrub_dev *sdev, + struct scrub_page *spage) { + struct scrub_block *sblock = spage->sblock; struct scrub_bio *sbio; + int ret; again: /* @@ -619,7 +1497,7 @@ again: if (sdev->curr != -1) { sdev->first_free = sdev->bios[sdev->curr]->next_free; sdev->bios[sdev->curr]->next_free = -1; - sdev->bios[sdev->curr]->count = 0; + sdev->bios[sdev->curr]->page_count = 0; spin_unlock(&sdev->list_lock); } else { spin_unlock(&sdev->list_lock); @@ -627,38 +1505,200 @@ again: } } sbio = sdev->bios[sdev->curr]; - if (sbio->count == 0) { - sbio->physical = physical; - sbio->logical = logical; - } else if (sbio->physical + sbio->count * PAGE_SIZE != physical || - sbio->logical + sbio->count * PAGE_SIZE != logical) { - int ret; + if (sbio->page_count == 0) { + struct bio *bio; + + sbio->physical = spage->physical; + sbio->logical = spage->logical; + bio = sbio->bio; + if (!bio) { + bio = bio_alloc(GFP_NOFS, sdev->pages_per_bio); + if (!bio) + return -ENOMEM; + sbio->bio = bio; + } - ret = scrub_submit(sdev); - if (ret) - return ret; + bio->bi_private = sbio; + bio->bi_end_io = scrub_bio_end_io; + bio->bi_bdev = sdev->dev->bdev; + bio->bi_sector = spage->physical >> 9; + sbio->err = 0; + } else if (sbio->physical + sbio->page_count * PAGE_SIZE != + spage->physical || + sbio->logical + sbio->page_count * PAGE_SIZE != + spage->logical) { + scrub_submit(sdev); goto again; } - sbio->spag[sbio->count].flags = flags; - sbio->spag[sbio->count].generation = gen; - sbio->spag[sbio->count].have_csum = 0; - sbio->spag[sbio->count].mirror_num = mirror_num; - if (csum) { - sbio->spag[sbio->count].have_csum = 1; - memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size); + + sbio->pagev[sbio->page_count] = spage; + ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0); + if (ret != PAGE_SIZE) { + if (sbio->page_count < 1) { + bio_put(sbio->bio); + sbio->bio = NULL; + return -EIO; + } + scrub_submit(sdev); + goto again; + } + + scrub_block_get(sblock); /* one for the added page */ + atomic_inc(&sblock->outstanding_pages); + sbio->page_count++; + if (sbio->page_count == sdev->pages_per_bio) + scrub_submit(sdev); + + return 0; +} + +static int scrub_pages(struct scrub_dev *sdev, u64 logical, u64 len, + u64 physical, u64 flags, u64 gen, int mirror_num, + u8 *csum, int force) +{ + struct scrub_block *sblock; + int index; + + sblock = kzalloc(sizeof(*sblock), GFP_NOFS); + if (!sblock) { + spin_lock(&sdev->stat_lock); + sdev->stat.malloc_errors++; + spin_unlock(&sdev->stat_lock); + return -ENOMEM; } - ++sbio->count; - if (sbio->count == SCRUB_PAGES_PER_BIO || force) { + + /* one ref inside this function, plus one for each page later on */ + atomic_set(&sblock->ref_count, 1); + sblock->sdev = sdev; + sblock->no_io_error_seen = 1; + + for (index = 0; len > 0; index++) { + struct scrub_page *spage = sblock->pagev + index; + u64 l = min_t(u64, len, PAGE_SIZE); + + BUG_ON(index >= SCRUB_MAX_PAGES_PER_BLOCK); + spage->page = alloc_page(GFP_NOFS); + if (!spage->page) { + spin_lock(&sdev->stat_lock); + sdev->stat.malloc_errors++; + spin_unlock(&sdev->stat_lock); + while (index > 0) { + index--; + __free_page(sblock->pagev[index].page); + } + kfree(sblock); + return -ENOMEM; + } + spage->sblock = sblock; + spage->dev = sdev->dev; + spage->flags = flags; + spage->generation = gen; + spage->logical = logical; + spage->physical = physical; + spage->mirror_num = mirror_num; + if (csum) { + spage->have_csum = 1; + memcpy(spage->csum, csum, sdev->csum_size); + } else { + spage->have_csum = 0; + } + sblock->page_count++; + len -= l; + logical += l; + physical += l; + } + + BUG_ON(sblock->page_count == 0); + for (index = 0; index < sblock->page_count; index++) { + struct scrub_page *spage = sblock->pagev + index; int ret; - ret = scrub_submit(sdev); - if (ret) + ret = scrub_add_page_to_bio(sdev, spage); + if (ret) { + scrub_block_put(sblock); return ret; + } } + if (force) + scrub_submit(sdev); + + /* last one frees, either here or in bio completion for last page */ + scrub_block_put(sblock); return 0; } +static void scrub_bio_end_io(struct bio *bio, int err) +{ + struct scrub_bio *sbio = bio->bi_private; + struct scrub_dev *sdev = sbio->sdev; + struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + + sbio->err = err; + sbio->bio = bio; + + btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work); +} + +static void scrub_bio_end_io_worker(struct btrfs_work *work) +{ + struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); + struct scrub_dev *sdev = sbio->sdev; + int i; + + BUG_ON(sbio->page_count > SCRUB_PAGES_PER_BIO); + if (sbio->err) { + for (i = 0; i < sbio->page_count; i++) { + struct scrub_page *spage = sbio->pagev[i]; + + spage->io_error = 1; + spage->sblock->no_io_error_seen = 0; + } + } + + /* now complete the scrub_block items that have all pages completed */ + for (i = 0; i < sbio->page_count; i++) { + struct scrub_page *spage = sbio->pagev[i]; + struct scrub_block *sblock = spage->sblock; + + if (atomic_dec_and_test(&sblock->outstanding_pages)) + scrub_block_complete(sblock); + scrub_block_put(sblock); + } + + if (sbio->err) { + /* what is this good for??? */ + sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1); + sbio->bio->bi_flags |= 1 << BIO_UPTODATE; + sbio->bio->bi_phys_segments = 0; + sbio->bio->bi_idx = 0; + + for (i = 0; i < sbio->page_count; i++) { + struct bio_vec *bi; + bi = &sbio->bio->bi_io_vec[i]; + bi->bv_offset = 0; + bi->bv_len = PAGE_SIZE; + } + } + + bio_put(sbio->bio); + sbio->bio = NULL; + spin_lock(&sdev->list_lock); + sbio->next_free = sdev->first_free; + sdev->first_free = sbio->index; + spin_unlock(&sdev->list_lock); + atomic_dec(&sdev->in_flight); + wake_up(&sdev->list_wait); +} + +static void scrub_block_complete(struct scrub_block *sblock) +{ + if (!sblock->no_io_error_seen) + scrub_handle_errored_block(sblock); + else + scrub_checksum(sblock); +} + static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len, u8 *csum) { @@ -666,7 +1706,6 @@ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len, int ret = 0; unsigned long i; unsigned long num_sectors; - u32 sectorsize = sdev->dev->dev_root->sectorsize; while (!list_empty(&sdev->csum_list)) { sum = list_first_entry(&sdev->csum_list, @@ -684,7 +1723,7 @@ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len, if (!sum) return 0; - num_sectors = sum->len / sectorsize; + num_sectors = sum->len / sdev->sectorsize; for (i = 0; i < num_sectors; ++i) { if (sum->sums[i].bytenr == logical) { memcpy(csum, &sum->sums[i].sum, sdev->csum_size); @@ -701,13 +1740,32 @@ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len, /* scrub extent tries to collect up to 64 kB for each bio */ static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len, - u64 physical, u64 flags, u64 gen, u64 mirror_num) + u64 physical, u64 flags, u64 gen, int mirror_num) { int ret; u8 csum[BTRFS_CSUM_SIZE]; + u32 blocksize; + + if (flags & BTRFS_EXTENT_FLAG_DATA) { + blocksize = sdev->sectorsize; + spin_lock(&sdev->stat_lock); + sdev->stat.data_extents_scrubbed++; + sdev->stat.data_bytes_scrubbed += len; + spin_unlock(&sdev->stat_lock); + } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + BUG_ON(sdev->nodesize != sdev->leafsize); + blocksize = sdev->nodesize; + spin_lock(&sdev->stat_lock); + sdev->stat.tree_extents_scrubbed++; + sdev->stat.tree_bytes_scrubbed += len; + spin_unlock(&sdev->stat_lock); + } else { + blocksize = sdev->sectorsize; + BUG_ON(1); + } while (len) { - u64 l = min_t(u64, len, PAGE_SIZE); + u64 l = min_t(u64, len, blocksize); int have_csum = 0; if (flags & BTRFS_EXTENT_FLAG_DATA) { @@ -716,8 +1774,8 @@ static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len, if (have_csum == 0) ++sdev->stat.no_csum; } - ret = scrub_page(sdev, logical, l, physical, flags, gen, - mirror_num, have_csum ? csum : NULL, 0); + ret = scrub_pages(sdev, logical, l, physical, flags, gen, + mirror_num, have_csum ? csum : NULL, 0); if (ret) return ret; len -= l; @@ -741,13 +1799,16 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev, int slot; int i; u64 nstripes; - int start_stripe; struct extent_buffer *l; struct btrfs_key key; u64 physical; u64 logical; u64 generation; - u64 mirror_num; + int mirror_num; + struct reada_control *reada1; + struct reada_control *reada2; + struct btrfs_key key_start; + struct btrfs_key key_end; u64 increment = map->stripe_len; u64 offset; @@ -758,102 +1819,93 @@ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev, if (map->type & BTRFS_BLOCK_GROUP_RAID0) { offset = map->stripe_len * num; increment = map->stripe_len * map->num_stripes; - mirror_num = 0; + mirror_num = 1; } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { int factor = map->num_stripes / map->sub_stripes; offset = map->stripe_len * (num / map->sub_stripes); increment = map->stripe_len * factor; - mirror_num = num % map->sub_stripes; + mirror_num = num % map->sub_stripes + 1; } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { increment = map->stripe_len; - mirror_num = num % map->num_stripes; + mirror_num = num % map->num_stripes + 1; } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { increment = map->stripe_len; - mirror_num = num % map->num_stripes; + mirror_num = num % map->num_stripes + 1; } else { increment = map->stripe_len; - mirror_num = 0; + mirror_num = 1; } path = btrfs_alloc_path(); if (!path) return -ENOMEM; - path->reada = 2; + /* + * work on commit root. The related disk blocks are static as + * long as COW is applied. This means, it is save to rewrite + * them to repair disk errors without any race conditions + */ path->search_commit_root = 1; path->skip_locking = 1; /* - * find all extents for each stripe and just read them to get - * them into the page cache - * FIXME: we can do better. build a more intelligent prefetching + * trigger the readahead for extent tree csum tree and wait for + * completion. During readahead, the scrub is officially paused + * to not hold off transaction commits */ logical = base + offset; - physical = map->stripes[num].physical; - ret = 0; - for (i = 0; i < nstripes; ++i) { - key.objectid = logical; - key.type = BTRFS_EXTENT_ITEM_KEY; - key.offset = (u64)0; - ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); - if (ret < 0) - goto out_noplug; - - /* - * we might miss half an extent here, but that doesn't matter, - * as it's only the prefetch - */ - while (1) { - l = path->nodes[0]; - slot = path->slots[0]; - if (slot >= btrfs_header_nritems(l)) { - ret = btrfs_next_leaf(root, path); - if (ret == 0) - continue; - if (ret < 0) - goto out_noplug; - - break; - } - btrfs_item_key_to_cpu(l, &key, slot); + wait_event(sdev->list_wait, + atomic_read(&sdev->in_flight) == 0); + atomic_inc(&fs_info->scrubs_paused); + wake_up(&fs_info->scrub_pause_wait); - if (key.objectid >= logical + map->stripe_len) - break; + /* FIXME it might be better to start readahead at commit root */ + key_start.objectid = logical; + key_start.type = BTRFS_EXTENT_ITEM_KEY; + key_start.offset = (u64)0; + key_end.objectid = base + offset + nstripes * increment; + key_end.type = BTRFS_EXTENT_ITEM_KEY; + key_end.offset = (u64)0; + reada1 = btrfs_reada_add(root, &key_start, &key_end); + + key_start.objectid = BTRFS_EXTENT_CSUM_OBJECTID; + key_start.type = BTRFS_EXTENT_CSUM_KEY; + key_start.offset = logical; + key_end.objectid = BTRFS_EXTENT_CSUM_OBJECTID; + key_end.type = BTRFS_EXTENT_CSUM_KEY; + key_end.offset = base + offset + nstripes * increment; + reada2 = btrfs_reada_add(csum_root, &key_start, &key_end); + + if (!IS_ERR(reada1)) + btrfs_reada_wait(reada1); + if (!IS_ERR(reada2)) + btrfs_reada_wait(reada2); - path->slots[0]++; - } - btrfs_release_path(path); - logical += increment; - physical += map->stripe_len; - cond_resched(); + mutex_lock(&fs_info->scrub_lock); + while (atomic_read(&fs_info->scrub_pause_req)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrub_pause_req) == 0); + mutex_lock(&fs_info->scrub_lock); } + atomic_dec(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + wake_up(&fs_info->scrub_pause_wait); /* * collect all data csums for the stripe to avoid seeking during * the scrub. This might currently (crc32) end up to be about 1MB */ - start_stripe = 0; blk_start_plug(&plug); -again: - logical = base + offset + start_stripe * increment; - for (i = start_stripe; i < nstripes; ++i) { - ret = btrfs_lookup_csums_range(csum_root, logical, - logical + map->stripe_len - 1, - &sdev->csum_list, 1); - if (ret) - goto out; - logical += increment; - cond_resched(); - } /* * now find all extents for each stripe and scrub them */ - logical = base + offset + start_stripe * increment; - physical = map->stripes[num].physical + start_stripe * map->stripe_len; + logical = base + offset; + physical = map->stripes[num].physical; ret = 0; - for (i = start_stripe; i < nstripes; ++i) { + for (i = 0; i < nstripes; ++i) { /* * canceled? */ @@ -882,11 +1934,14 @@ again: atomic_dec(&fs_info->scrubs_paused); mutex_unlock(&fs_info->scrub_lock); wake_up(&fs_info->scrub_pause_wait); - scrub_free_csums(sdev); - start_stripe = i; - goto again; } + ret = btrfs_lookup_csums_range(csum_root, logical, + logical + map->stripe_len - 1, + &sdev->csum_list, 1); + if (ret) + goto out; + key.objectid = logical; key.type = BTRFS_EXTENT_ITEM_KEY; key.offset = (u64)0; @@ -982,13 +2037,13 @@ next: out: blk_finish_plug(&plug); -out_noplug: btrfs_free_path(path); return ret < 0 ? ret : 0; } static noinline_for_stack int scrub_chunk(struct scrub_dev *sdev, - u64 chunk_tree, u64 chunk_objectid, u64 chunk_offset, u64 length) + u64 chunk_tree, u64 chunk_objectid, u64 chunk_offset, u64 length, + u64 dev_offset) { struct btrfs_mapping_tree *map_tree = &sdev->dev->dev_root->fs_info->mapping_tree; @@ -1012,7 +2067,8 @@ static noinline_for_stack int scrub_chunk(struct scrub_dev *sdev, goto out; for (i = 0; i < map->num_stripes; ++i) { - if (map->stripes[i].dev == sdev->dev) { + if (map->stripes[i].dev == sdev->dev && + map->stripes[i].physical == dev_offset) { ret = scrub_stripe(sdev, map, i, chunk_offset, length); if (ret) goto out; @@ -1108,7 +2164,7 @@ int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end) break; } ret = scrub_chunk(sdev, chunk_tree, chunk_objectid, - chunk_offset, length); + chunk_offset, length, found_key.offset); btrfs_put_block_group(cache); if (ret) break; @@ -1135,15 +2191,18 @@ static noinline_for_stack int scrub_supers(struct scrub_dev *sdev) struct btrfs_device *device = sdev->dev; struct btrfs_root *root = device->dev_root; + if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) + return -EIO; + gen = root->fs_info->last_trans_committed; for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { bytenr = btrfs_sb_offset(i); - if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes) + if (bytenr + BTRFS_SUPER_INFO_SIZE > device->total_bytes) break; - ret = scrub_page(sdev, bytenr, PAGE_SIZE, bytenr, - BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1); + ret = scrub_pages(sdev, bytenr, BTRFS_SUPER_INFO_SIZE, bytenr, + BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1); if (ret) return ret; } @@ -1158,18 +2217,22 @@ static noinline_for_stack int scrub_supers(struct scrub_dev *sdev) static noinline_for_stack int scrub_workers_get(struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; + int ret = 0; mutex_lock(&fs_info->scrub_lock); if (fs_info->scrub_workers_refcnt == 0) { btrfs_init_workers(&fs_info->scrub_workers, "scrub", fs_info->thread_pool_size, &fs_info->generic_worker); fs_info->scrub_workers.idle_thresh = 4; - btrfs_start_workers(&fs_info->scrub_workers, 1); + ret = btrfs_start_workers(&fs_info->scrub_workers); + if (ret) + goto out; } ++fs_info->scrub_workers_refcnt; +out: mutex_unlock(&fs_info->scrub_lock); - return 0; + return ret; } static noinline_for_stack void scrub_workers_put(struct btrfs_root *root) @@ -1198,10 +2261,30 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, /* * check some assumptions */ - if (root->sectorsize != PAGE_SIZE || - root->sectorsize != root->leafsize || - root->sectorsize != root->nodesize) { - printk(KERN_ERR "btrfs_scrub: size assumptions fail\n"); + if (root->nodesize != root->leafsize) { + printk(KERN_ERR + "btrfs_scrub: size assumption nodesize == leafsize (%d == %d) fails\n", + root->nodesize, root->leafsize); + return -EINVAL; + } + + if (root->nodesize > BTRFS_STRIPE_LEN) { + /* + * in this case scrub is unable to calculate the checksum + * the way scrub is implemented. Do not handle this + * situation at all because it won't ever happen. + */ + printk(KERN_ERR + "btrfs_scrub: size assumption nodesize <= BTRFS_STRIPE_LEN (%d <= %d) fails\n", + root->nodesize, BTRFS_STRIPE_LEN); + return -EINVAL; + } + + if (root->sectorsize != PAGE_SIZE) { + /* not supported for data w/o checksums */ + printk(KERN_ERR + "btrfs_scrub: size assumption sectorsize != PAGE_SIZE (%d != %lld) fails\n", + root->sectorsize, (unsigned long long)PAGE_SIZE); return -EINVAL; } @@ -1253,10 +2336,11 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, ret = scrub_enumerate_chunks(sdev, start, end); wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0); - atomic_dec(&fs_info->scrubs_running); wake_up(&fs_info->scrub_pause_wait); + wait_event(sdev->list_wait, atomic_read(&sdev->fixup_cnt) == 0); + if (progress) memcpy(progress, &sdev->stat, sizeof(*progress)); @@ -1270,7 +2354,7 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, return ret; } -int btrfs_scrub_pause(struct btrfs_root *root) +void btrfs_scrub_pause(struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; @@ -1285,34 +2369,28 @@ int btrfs_scrub_pause(struct btrfs_root *root) mutex_lock(&fs_info->scrub_lock); } mutex_unlock(&fs_info->scrub_lock); - - return 0; } -int btrfs_scrub_continue(struct btrfs_root *root) +void btrfs_scrub_continue(struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; atomic_dec(&fs_info->scrub_pause_req); wake_up(&fs_info->scrub_pause_wait); - return 0; } -int btrfs_scrub_pause_super(struct btrfs_root *root) +void btrfs_scrub_pause_super(struct btrfs_root *root) { down_write(&root->fs_info->scrub_super_lock); - return 0; } -int btrfs_scrub_continue_super(struct btrfs_root *root) +void btrfs_scrub_continue_super(struct btrfs_root *root) { up_write(&root->fs_info->scrub_super_lock); - return 0; } -int btrfs_scrub_cancel(struct btrfs_root *root) +int __btrfs_scrub_cancel(struct btrfs_fs_info *fs_info) { - struct btrfs_fs_info *fs_info = root->fs_info; mutex_lock(&fs_info->scrub_lock); if (!atomic_read(&fs_info->scrubs_running)) { @@ -1333,6 +2411,11 @@ int btrfs_scrub_cancel(struct btrfs_root *root) return 0; } +int btrfs_scrub_cancel(struct btrfs_root *root) +{ + return __btrfs_scrub_cancel(root->fs_info); +} + int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev) { struct btrfs_fs_info *fs_info = root->fs_info; @@ -1355,6 +2438,7 @@ int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev) return 0; } + int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid) { struct btrfs_fs_info *fs_info = root->fs_info; diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c new file mode 100644 index 000000000000..fb5ffe95f869 --- /dev/null +++ b/fs/btrfs/send.c @@ -0,0 +1,4572 @@ +/* + * Copyright (C) 2012 Alexander Block. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/bsearch.h> +#include <linux/fs.h> +#include <linux/file.h> +#include <linux/sort.h> +#include <linux/mount.h> +#include <linux/xattr.h> +#include <linux/posix_acl_xattr.h> +#include <linux/radix-tree.h> +#include <linux/crc32c.h> +#include <linux/vmalloc.h> + +#include "send.h" +#include "backref.h" +#include "locking.h" +#include "disk-io.h" +#include "btrfs_inode.h" +#include "transaction.h" + +static int g_verbose = 0; + +#define verbose_printk(...) if (g_verbose) printk(__VA_ARGS__) + +/* + * A fs_path is a helper to dynamically build path names with unknown size. + * It reallocates the internal buffer on demand. + * It allows fast adding of path elements on the right side (normal path) and + * fast adding to the left side (reversed path). A reversed path can also be + * unreversed if needed. + */ +struct fs_path { + union { + struct { + char *start; + char *end; + char *prepared; + + char *buf; + int buf_len; + int reversed:1; + int virtual_mem:1; + char inline_buf[]; + }; + char pad[PAGE_SIZE]; + }; +}; +#define FS_PATH_INLINE_SIZE \ + (sizeof(struct fs_path) - offsetof(struct fs_path, inline_buf)) + + +/* reused for each extent */ +struct clone_root { + struct btrfs_root *root; + u64 ino; + u64 offset; + + u64 found_refs; +}; + +#define SEND_CTX_MAX_NAME_CACHE_SIZE 128 +#define SEND_CTX_NAME_CACHE_CLEAN_SIZE (SEND_CTX_MAX_NAME_CACHE_SIZE * 2) + +struct send_ctx { + struct file *send_filp; + loff_t send_off; + char *send_buf; + u32 send_size; + u32 send_max_size; + u64 total_send_size; + u64 cmd_send_size[BTRFS_SEND_C_MAX + 1]; + + struct vfsmount *mnt; + + struct btrfs_root *send_root; + struct btrfs_root *parent_root; + struct clone_root *clone_roots; + int clone_roots_cnt; + + /* current state of the compare_tree call */ + struct btrfs_path *left_path; + struct btrfs_path *right_path; + struct btrfs_key *cmp_key; + + /* + * infos of the currently processed inode. In case of deleted inodes, + * these are the values from the deleted inode. + */ + u64 cur_ino; + u64 cur_inode_gen; + int cur_inode_new; + int cur_inode_new_gen; + int cur_inode_deleted; + int cur_inode_first_ref_orphan; + u64 cur_inode_size; + u64 cur_inode_mode; + + u64 send_progress; + + struct list_head new_refs; + struct list_head deleted_refs; + + struct radix_tree_root name_cache; + struct list_head name_cache_list; + int name_cache_size; + + struct file *cur_inode_filp; + char *read_buf; +}; + +struct name_cache_entry { + struct list_head list; + struct list_head use_list; + u64 ino; + u64 gen; + u64 parent_ino; + u64 parent_gen; + int ret; + int need_later_update; + int name_len; + char name[]; +}; + +static void fs_path_reset(struct fs_path *p) +{ + if (p->reversed) { + p->start = p->buf + p->buf_len - 1; + p->end = p->start; + *p->start = 0; + } else { + p->start = p->buf; + p->end = p->start; + *p->start = 0; + } +} + +static struct fs_path *fs_path_alloc(struct send_ctx *sctx) +{ + struct fs_path *p; + + p = kmalloc(sizeof(*p), GFP_NOFS); + if (!p) + return NULL; + p->reversed = 0; + p->virtual_mem = 0; + p->buf = p->inline_buf; + p->buf_len = FS_PATH_INLINE_SIZE; + fs_path_reset(p); + return p; +} + +static struct fs_path *fs_path_alloc_reversed(struct send_ctx *sctx) +{ + struct fs_path *p; + + p = fs_path_alloc(sctx); + if (!p) + return NULL; + p->reversed = 1; + fs_path_reset(p); + return p; +} + +static void fs_path_free(struct send_ctx *sctx, struct fs_path *p) +{ + if (!p) + return; + if (p->buf != p->inline_buf) { + if (p->virtual_mem) + vfree(p->buf); + else + kfree(p->buf); + } + kfree(p); +} + +static int fs_path_len(struct fs_path *p) +{ + return p->end - p->start; +} + +static int fs_path_ensure_buf(struct fs_path *p, int len) +{ + char *tmp_buf; + int path_len; + int old_buf_len; + + len++; + + if (p->buf_len >= len) + return 0; + + path_len = p->end - p->start; + old_buf_len = p->buf_len; + len = PAGE_ALIGN(len); + + if (p->buf == p->inline_buf) { + tmp_buf = kmalloc(len, GFP_NOFS); + if (!tmp_buf) { + tmp_buf = vmalloc(len); + if (!tmp_buf) + return -ENOMEM; + p->virtual_mem = 1; + } + memcpy(tmp_buf, p->buf, p->buf_len); + p->buf = tmp_buf; + p->buf_len = len; + } else { + if (p->virtual_mem) { + tmp_buf = vmalloc(len); + if (!tmp_buf) + return -ENOMEM; + memcpy(tmp_buf, p->buf, p->buf_len); + vfree(p->buf); + } else { + tmp_buf = krealloc(p->buf, len, GFP_NOFS); + if (!tmp_buf) { + tmp_buf = vmalloc(len); + if (!tmp_buf) + return -ENOMEM; + memcpy(tmp_buf, p->buf, p->buf_len); + kfree(p->buf); + p->virtual_mem = 1; + } + } + p->buf = tmp_buf; + p->buf_len = len; + } + if (p->reversed) { + tmp_buf = p->buf + old_buf_len - path_len - 1; + p->end = p->buf + p->buf_len - 1; + p->start = p->end - path_len; + memmove(p->start, tmp_buf, path_len + 1); + } else { + p->start = p->buf; + p->end = p->start + path_len; + } + return 0; +} + +static int fs_path_prepare_for_add(struct fs_path *p, int name_len) +{ + int ret; + int new_len; + + new_len = p->end - p->start + name_len; + if (p->start != p->end) + new_len++; + ret = fs_path_ensure_buf(p, new_len); + if (ret < 0) + goto out; + + if (p->reversed) { + if (p->start != p->end) + *--p->start = '/'; + p->start -= name_len; + p->prepared = p->start; + } else { + if (p->start != p->end) + *p->end++ = '/'; + p->prepared = p->end; + p->end += name_len; + *p->end = 0; + } + +out: + return ret; +} + +static int fs_path_add(struct fs_path *p, const char *name, int name_len) +{ + int ret; + + ret = fs_path_prepare_for_add(p, name_len); + if (ret < 0) + goto out; + memcpy(p->prepared, name, name_len); + p->prepared = NULL; + +out: + return ret; +} + +static int fs_path_add_path(struct fs_path *p, struct fs_path *p2) +{ + int ret; + + ret = fs_path_prepare_for_add(p, p2->end - p2->start); + if (ret < 0) + goto out; + memcpy(p->prepared, p2->start, p2->end - p2->start); + p->prepared = NULL; + +out: + return ret; +} + +static int fs_path_add_from_extent_buffer(struct fs_path *p, + struct extent_buffer *eb, + unsigned long off, int len) +{ + int ret; + + ret = fs_path_prepare_for_add(p, len); + if (ret < 0) + goto out; + + read_extent_buffer(eb, p->prepared, off, len); + p->prepared = NULL; + +out: + return ret; +} + +static void fs_path_remove(struct fs_path *p) +{ + BUG_ON(p->reversed); + while (p->start != p->end && *p->end != '/') + p->end--; + *p->end = 0; +} + +static int fs_path_copy(struct fs_path *p, struct fs_path *from) +{ + int ret; + + p->reversed = from->reversed; + fs_path_reset(p); + + ret = fs_path_add_path(p, from); + + return ret; +} + + +static void fs_path_unreverse(struct fs_path *p) +{ + char *tmp; + int len; + + if (!p->reversed) + return; + + tmp = p->start; + len = p->end - p->start; + p->start = p->buf; + p->end = p->start + len; + memmove(p->start, tmp, len + 1); + p->reversed = 0; +} + +static struct btrfs_path *alloc_path_for_send(void) +{ + struct btrfs_path *path; + + path = btrfs_alloc_path(); + if (!path) + return NULL; + path->search_commit_root = 1; + path->skip_locking = 1; + return path; +} + +static int write_buf(struct send_ctx *sctx, const void *buf, u32 len) +{ + int ret; + mm_segment_t old_fs; + u32 pos = 0; + + old_fs = get_fs(); + set_fs(KERNEL_DS); + + while (pos < len) { + ret = vfs_write(sctx->send_filp, (char *)buf + pos, len - pos, + &sctx->send_off); + /* TODO handle that correctly */ + /*if (ret == -ERESTARTSYS) { + continue; + }*/ + if (ret < 0) + goto out; + if (ret == 0) { + ret = -EIO; + goto out; + } + pos += ret; + } + + ret = 0; + +out: + set_fs(old_fs); + return ret; +} + +static int tlv_put(struct send_ctx *sctx, u16 attr, const void *data, int len) +{ + struct btrfs_tlv_header *hdr; + int total_len = sizeof(*hdr) + len; + int left = sctx->send_max_size - sctx->send_size; + + if (unlikely(left < total_len)) + return -EOVERFLOW; + + hdr = (struct btrfs_tlv_header *) (sctx->send_buf + sctx->send_size); + hdr->tlv_type = cpu_to_le16(attr); + hdr->tlv_len = cpu_to_le16(len); + memcpy(hdr + 1, data, len); + sctx->send_size += total_len; + + return 0; +} + +#if 0 +static int tlv_put_u8(struct send_ctx *sctx, u16 attr, u8 value) +{ + return tlv_put(sctx, attr, &value, sizeof(value)); +} + +static int tlv_put_u16(struct send_ctx *sctx, u16 attr, u16 value) +{ + __le16 tmp = cpu_to_le16(value); + return tlv_put(sctx, attr, &tmp, sizeof(tmp)); +} + +static int tlv_put_u32(struct send_ctx *sctx, u16 attr, u32 value) +{ + __le32 tmp = cpu_to_le32(value); + return tlv_put(sctx, attr, &tmp, sizeof(tmp)); +} +#endif + +static int tlv_put_u64(struct send_ctx *sctx, u16 attr, u64 value) +{ + __le64 tmp = cpu_to_le64(value); + return tlv_put(sctx, attr, &tmp, sizeof(tmp)); +} + +static int tlv_put_string(struct send_ctx *sctx, u16 attr, + const char *str, int len) +{ + if (len == -1) + len = strlen(str); + return tlv_put(sctx, attr, str, len); +} + +static int tlv_put_uuid(struct send_ctx *sctx, u16 attr, + const u8 *uuid) +{ + return tlv_put(sctx, attr, uuid, BTRFS_UUID_SIZE); +} + +#if 0 +static int tlv_put_timespec(struct send_ctx *sctx, u16 attr, + struct timespec *ts) +{ + struct btrfs_timespec bts; + bts.sec = cpu_to_le64(ts->tv_sec); + bts.nsec = cpu_to_le32(ts->tv_nsec); + return tlv_put(sctx, attr, &bts, sizeof(bts)); +} +#endif + +static int tlv_put_btrfs_timespec(struct send_ctx *sctx, u16 attr, + struct extent_buffer *eb, + struct btrfs_timespec *ts) +{ + struct btrfs_timespec bts; + read_extent_buffer(eb, &bts, (unsigned long)ts, sizeof(bts)); + return tlv_put(sctx, attr, &bts, sizeof(bts)); +} + + +#define TLV_PUT(sctx, attrtype, attrlen, data) \ + do { \ + ret = tlv_put(sctx, attrtype, attrlen, data); \ + if (ret < 0) \ + goto tlv_put_failure; \ + } while (0) + +#define TLV_PUT_INT(sctx, attrtype, bits, value) \ + do { \ + ret = tlv_put_u##bits(sctx, attrtype, value); \ + if (ret < 0) \ + goto tlv_put_failure; \ + } while (0) + +#define TLV_PUT_U8(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 8, data) +#define TLV_PUT_U16(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 16, data) +#define TLV_PUT_U32(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 32, data) +#define TLV_PUT_U64(sctx, attrtype, data) TLV_PUT_INT(sctx, attrtype, 64, data) +#define TLV_PUT_STRING(sctx, attrtype, str, len) \ + do { \ + ret = tlv_put_string(sctx, attrtype, str, len); \ + if (ret < 0) \ + goto tlv_put_failure; \ + } while (0) +#define TLV_PUT_PATH(sctx, attrtype, p) \ + do { \ + ret = tlv_put_string(sctx, attrtype, p->start, \ + p->end - p->start); \ + if (ret < 0) \ + goto tlv_put_failure; \ + } while(0) +#define TLV_PUT_UUID(sctx, attrtype, uuid) \ + do { \ + ret = tlv_put_uuid(sctx, attrtype, uuid); \ + if (ret < 0) \ + goto tlv_put_failure; \ + } while (0) +#define TLV_PUT_TIMESPEC(sctx, attrtype, ts) \ + do { \ + ret = tlv_put_timespec(sctx, attrtype, ts); \ + if (ret < 0) \ + goto tlv_put_failure; \ + } while (0) +#define TLV_PUT_BTRFS_TIMESPEC(sctx, attrtype, eb, ts) \ + do { \ + ret = tlv_put_btrfs_timespec(sctx, attrtype, eb, ts); \ + if (ret < 0) \ + goto tlv_put_failure; \ + } while (0) + +static int send_header(struct send_ctx *sctx) +{ + struct btrfs_stream_header hdr; + + strcpy(hdr.magic, BTRFS_SEND_STREAM_MAGIC); + hdr.version = cpu_to_le32(BTRFS_SEND_STREAM_VERSION); + + return write_buf(sctx, &hdr, sizeof(hdr)); +} + +/* + * For each command/item we want to send to userspace, we call this function. + */ +static int begin_cmd(struct send_ctx *sctx, int cmd) +{ + struct btrfs_cmd_header *hdr; + + if (!sctx->send_buf) { + WARN_ON(1); + return -EINVAL; + } + + BUG_ON(sctx->send_size); + + sctx->send_size += sizeof(*hdr); + hdr = (struct btrfs_cmd_header *)sctx->send_buf; + hdr->cmd = cpu_to_le16(cmd); + + return 0; +} + +static int send_cmd(struct send_ctx *sctx) +{ + int ret; + struct btrfs_cmd_header *hdr; + u32 crc; + + hdr = (struct btrfs_cmd_header *)sctx->send_buf; + hdr->len = cpu_to_le32(sctx->send_size - sizeof(*hdr)); + hdr->crc = 0; + + crc = crc32c(0, (unsigned char *)sctx->send_buf, sctx->send_size); + hdr->crc = cpu_to_le32(crc); + + ret = write_buf(sctx, sctx->send_buf, sctx->send_size); + + sctx->total_send_size += sctx->send_size; + sctx->cmd_send_size[le16_to_cpu(hdr->cmd)] += sctx->send_size; + sctx->send_size = 0; + + return ret; +} + +/* + * Sends a move instruction to user space + */ +static int send_rename(struct send_ctx *sctx, + struct fs_path *from, struct fs_path *to) +{ + int ret; + +verbose_printk("btrfs: send_rename %s -> %s\n", from->start, to->start); + + ret = begin_cmd(sctx, BTRFS_SEND_C_RENAME); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, from); + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_TO, to); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + return ret; +} + +/* + * Sends a link instruction to user space + */ +static int send_link(struct send_ctx *sctx, + struct fs_path *path, struct fs_path *lnk) +{ + int ret; + +verbose_printk("btrfs: send_link %s -> %s\n", path->start, lnk->start); + + ret = begin_cmd(sctx, BTRFS_SEND_C_LINK); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path); + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_LINK, lnk); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + return ret; +} + +/* + * Sends an unlink instruction to user space + */ +static int send_unlink(struct send_ctx *sctx, struct fs_path *path) +{ + int ret; + +verbose_printk("btrfs: send_unlink %s\n", path->start); + + ret = begin_cmd(sctx, BTRFS_SEND_C_UNLINK); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + return ret; +} + +/* + * Sends a rmdir instruction to user space + */ +static int send_rmdir(struct send_ctx *sctx, struct fs_path *path) +{ + int ret; + +verbose_printk("btrfs: send_rmdir %s\n", path->start); + + ret = begin_cmd(sctx, BTRFS_SEND_C_RMDIR); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + return ret; +} + +/* + * Helper function to retrieve some fields from an inode item. + */ +static int get_inode_info(struct btrfs_root *root, + u64 ino, u64 *size, u64 *gen, + u64 *mode, u64 *uid, u64 *gid) +{ + int ret; + struct btrfs_inode_item *ii; + struct btrfs_key key; + struct btrfs_path *path; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + key.objectid = ino; + key.type = BTRFS_INODE_ITEM_KEY; + key.offset = 0; + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + if (ret) { + ret = -ENOENT; + goto out; + } + + ii = btrfs_item_ptr(path->nodes[0], path->slots[0], + struct btrfs_inode_item); + if (size) + *size = btrfs_inode_size(path->nodes[0], ii); + if (gen) + *gen = btrfs_inode_generation(path->nodes[0], ii); + if (mode) + *mode = btrfs_inode_mode(path->nodes[0], ii); + if (uid) + *uid = btrfs_inode_uid(path->nodes[0], ii); + if (gid) + *gid = btrfs_inode_gid(path->nodes[0], ii); + +out: + btrfs_free_path(path); + return ret; +} + +typedef int (*iterate_inode_ref_t)(int num, u64 dir, int index, + struct fs_path *p, + void *ctx); + +/* + * Helper function to iterate the entries in ONE btrfs_inode_ref. + * The iterate callback may return a non zero value to stop iteration. This can + * be a negative value for error codes or 1 to simply stop it. + * + * path must point to the INODE_REF when called. + */ +static int iterate_inode_ref(struct send_ctx *sctx, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *found_key, int resolve, + iterate_inode_ref_t iterate, void *ctx) +{ + struct extent_buffer *eb; + struct btrfs_item *item; + struct btrfs_inode_ref *iref; + struct btrfs_path *tmp_path; + struct fs_path *p; + u32 cur; + u32 len; + u32 total; + int slot; + u32 name_len; + char *start; + int ret = 0; + int num; + int index; + + p = fs_path_alloc_reversed(sctx); + if (!p) + return -ENOMEM; + + tmp_path = alloc_path_for_send(); + if (!tmp_path) { + fs_path_free(sctx, p); + return -ENOMEM; + } + + eb = path->nodes[0]; + slot = path->slots[0]; + item = btrfs_item_nr(eb, slot); + iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref); + cur = 0; + len = 0; + total = btrfs_item_size(eb, item); + + num = 0; + while (cur < total) { + fs_path_reset(p); + + name_len = btrfs_inode_ref_name_len(eb, iref); + index = btrfs_inode_ref_index(eb, iref); + if (resolve) { + start = btrfs_iref_to_path(root, tmp_path, iref, eb, + found_key->offset, p->buf, + p->buf_len); + if (IS_ERR(start)) { + ret = PTR_ERR(start); + goto out; + } + if (start < p->buf) { + /* overflow , try again with larger buffer */ + ret = fs_path_ensure_buf(p, + p->buf_len + p->buf - start); + if (ret < 0) + goto out; + start = btrfs_iref_to_path(root, tmp_path, iref, + eb, found_key->offset, p->buf, + p->buf_len); + if (IS_ERR(start)) { + ret = PTR_ERR(start); + goto out; + } + BUG_ON(start < p->buf); + } + p->start = start; + } else { + ret = fs_path_add_from_extent_buffer(p, eb, + (unsigned long)(iref + 1), name_len); + if (ret < 0) + goto out; + } + + + len = sizeof(*iref) + name_len; + iref = (struct btrfs_inode_ref *)((char *)iref + len); + cur += len; + + ret = iterate(num, found_key->offset, index, p, ctx); + if (ret) + goto out; + + num++; + } + +out: + btrfs_free_path(tmp_path); + fs_path_free(sctx, p); + return ret; +} + +typedef int (*iterate_dir_item_t)(int num, struct btrfs_key *di_key, + const char *name, int name_len, + const char *data, int data_len, + u8 type, void *ctx); + +/* + * Helper function to iterate the entries in ONE btrfs_dir_item. + * The iterate callback may return a non zero value to stop iteration. This can + * be a negative value for error codes or 1 to simply stop it. + * + * path must point to the dir item when called. + */ +static int iterate_dir_item(struct send_ctx *sctx, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *found_key, + iterate_dir_item_t iterate, void *ctx) +{ + int ret = 0; + struct extent_buffer *eb; + struct btrfs_item *item; + struct btrfs_dir_item *di; + struct btrfs_path *tmp_path = NULL; + struct btrfs_key di_key; + char *buf = NULL; + char *buf2 = NULL; + int buf_len; + int buf_virtual = 0; + u32 name_len; + u32 data_len; + u32 cur; + u32 len; + u32 total; + int slot; + int num; + u8 type; + + buf_len = PAGE_SIZE; + buf = kmalloc(buf_len, GFP_NOFS); + if (!buf) { + ret = -ENOMEM; + goto out; + } + + tmp_path = alloc_path_for_send(); + if (!tmp_path) { + ret = -ENOMEM; + goto out; + } + + eb = path->nodes[0]; + slot = path->slots[0]; + item = btrfs_item_nr(eb, slot); + di = btrfs_item_ptr(eb, slot, struct btrfs_dir_item); + cur = 0; + len = 0; + total = btrfs_item_size(eb, item); + + num = 0; + while (cur < total) { + name_len = btrfs_dir_name_len(eb, di); + data_len = btrfs_dir_data_len(eb, di); + type = btrfs_dir_type(eb, di); + btrfs_dir_item_key_to_cpu(eb, di, &di_key); + + if (name_len + data_len > buf_len) { + buf_len = PAGE_ALIGN(name_len + data_len); + if (buf_virtual) { + buf2 = vmalloc(buf_len); + if (!buf2) { + ret = -ENOMEM; + goto out; + } + vfree(buf); + } else { + buf2 = krealloc(buf, buf_len, GFP_NOFS); + if (!buf2) { + buf2 = vmalloc(buf_len); + if (!buf2) { + ret = -ENOMEM; + goto out; + } + kfree(buf); + buf_virtual = 1; + } + } + + buf = buf2; + buf2 = NULL; + } + + read_extent_buffer(eb, buf, (unsigned long)(di + 1), + name_len + data_len); + + len = sizeof(*di) + name_len + data_len; + di = (struct btrfs_dir_item *)((char *)di + len); + cur += len; + + ret = iterate(num, &di_key, buf, name_len, buf + name_len, + data_len, type, ctx); + if (ret < 0) + goto out; + if (ret) { + ret = 0; + goto out; + } + + num++; + } + +out: + btrfs_free_path(tmp_path); + if (buf_virtual) + vfree(buf); + else + kfree(buf); + return ret; +} + +static int __copy_first_ref(int num, u64 dir, int index, + struct fs_path *p, void *ctx) +{ + int ret; + struct fs_path *pt = ctx; + + ret = fs_path_copy(pt, p); + if (ret < 0) + return ret; + + /* we want the first only */ + return 1; +} + +/* + * Retrieve the first path of an inode. If an inode has more then one + * ref/hardlink, this is ignored. + */ +static int get_inode_path(struct send_ctx *sctx, struct btrfs_root *root, + u64 ino, struct fs_path *path) +{ + int ret; + struct btrfs_key key, found_key; + struct btrfs_path *p; + + p = alloc_path_for_send(); + if (!p) + return -ENOMEM; + + fs_path_reset(path); + + key.objectid = ino; + key.type = BTRFS_INODE_REF_KEY; + key.offset = 0; + + ret = btrfs_search_slot_for_read(root, &key, p, 1, 0); + if (ret < 0) + goto out; + if (ret) { + ret = 1; + goto out; + } + btrfs_item_key_to_cpu(p->nodes[0], &found_key, p->slots[0]); + if (found_key.objectid != ino || + found_key.type != BTRFS_INODE_REF_KEY) { + ret = -ENOENT; + goto out; + } + + ret = iterate_inode_ref(sctx, root, p, &found_key, 1, + __copy_first_ref, path); + if (ret < 0) + goto out; + ret = 0; + +out: + btrfs_free_path(p); + return ret; +} + +struct backref_ctx { + struct send_ctx *sctx; + + /* number of total found references */ + u64 found; + + /* + * used for clones found in send_root. clones found behind cur_objectid + * and cur_offset are not considered as allowed clones. + */ + u64 cur_objectid; + u64 cur_offset; + + /* may be truncated in case it's the last extent in a file */ + u64 extent_len; + + /* Just to check for bugs in backref resolving */ + int found_in_send_root; +}; + +static int __clone_root_cmp_bsearch(const void *key, const void *elt) +{ + u64 root = (u64)key; + struct clone_root *cr = (struct clone_root *)elt; + + if (root < cr->root->objectid) + return -1; + if (root > cr->root->objectid) + return 1; + return 0; +} + +static int __clone_root_cmp_sort(const void *e1, const void *e2) +{ + struct clone_root *cr1 = (struct clone_root *)e1; + struct clone_root *cr2 = (struct clone_root *)e2; + + if (cr1->root->objectid < cr2->root->objectid) + return -1; + if (cr1->root->objectid > cr2->root->objectid) + return 1; + return 0; +} + +/* + * Called for every backref that is found for the current extent. + */ +static int __iterate_backrefs(u64 ino, u64 offset, u64 root, void *ctx_) +{ + struct backref_ctx *bctx = ctx_; + struct clone_root *found; + int ret; + u64 i_size; + + /* First check if the root is in the list of accepted clone sources */ + found = bsearch((void *)root, bctx->sctx->clone_roots, + bctx->sctx->clone_roots_cnt, + sizeof(struct clone_root), + __clone_root_cmp_bsearch); + if (!found) + return 0; + + if (found->root == bctx->sctx->send_root && + ino == bctx->cur_objectid && + offset == bctx->cur_offset) { + bctx->found_in_send_root = 1; + } + + /* + * There are inodes that have extents that lie behind it's i_size. Don't + * accept clones from these extents. + */ + ret = get_inode_info(found->root, ino, &i_size, NULL, NULL, NULL, NULL); + if (ret < 0) + return ret; + + if (offset + bctx->extent_len > i_size) + return 0; + + /* + * Make sure we don't consider clones from send_root that are + * behind the current inode/offset. + */ + if (found->root == bctx->sctx->send_root) { + /* + * TODO for the moment we don't accept clones from the inode + * that is currently send. We may change this when + * BTRFS_IOC_CLONE_RANGE supports cloning from and to the same + * file. + */ + if (ino >= bctx->cur_objectid) + return 0; + /*if (ino > ctx->cur_objectid) + return 0; + if (offset + ctx->extent_len > ctx->cur_offset) + return 0;*/ + + bctx->found++; + found->found_refs++; + found->ino = ino; + found->offset = offset; + return 0; + } + + bctx->found++; + found->found_refs++; + if (ino < found->ino) { + found->ino = ino; + found->offset = offset; + } else if (found->ino == ino) { + /* + * same extent found more then once in the same file. + */ + if (found->offset > offset + bctx->extent_len) + found->offset = offset; + } + + return 0; +} + +/* + * path must point to the extent item when called. + */ +static int find_extent_clone(struct send_ctx *sctx, + struct btrfs_path *path, + u64 ino, u64 data_offset, + u64 ino_size, + struct clone_root **found) +{ + int ret; + int extent_type; + u64 logical; + u64 num_bytes; + u64 extent_item_pos; + struct btrfs_file_extent_item *fi; + struct extent_buffer *eb = path->nodes[0]; + struct backref_ctx backref_ctx; + struct clone_root *cur_clone_root; + struct btrfs_key found_key; + struct btrfs_path *tmp_path; + u32 i; + + tmp_path = alloc_path_for_send(); + if (!tmp_path) + return -ENOMEM; + + if (data_offset >= ino_size) { + /* + * There may be extents that lie behind the file's size. + * I at least had this in combination with snapshotting while + * writing large files. + */ + ret = 0; + goto out; + } + + fi = btrfs_item_ptr(eb, path->slots[0], + struct btrfs_file_extent_item); + extent_type = btrfs_file_extent_type(eb, fi); + if (extent_type == BTRFS_FILE_EXTENT_INLINE) { + ret = -ENOENT; + goto out; + } + + num_bytes = btrfs_file_extent_num_bytes(eb, fi); + logical = btrfs_file_extent_disk_bytenr(eb, fi); + if (logical == 0) { + ret = -ENOENT; + goto out; + } + logical += btrfs_file_extent_offset(eb, fi); + + ret = extent_from_logical(sctx->send_root->fs_info, + logical, tmp_path, &found_key); + btrfs_release_path(tmp_path); + + if (ret < 0) + goto out; + if (ret & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + ret = -EIO; + goto out; + } + + /* + * Setup the clone roots. + */ + for (i = 0; i < sctx->clone_roots_cnt; i++) { + cur_clone_root = sctx->clone_roots + i; + cur_clone_root->ino = (u64)-1; + cur_clone_root->offset = 0; + cur_clone_root->found_refs = 0; + } + + backref_ctx.sctx = sctx; + backref_ctx.found = 0; + backref_ctx.cur_objectid = ino; + backref_ctx.cur_offset = data_offset; + backref_ctx.found_in_send_root = 0; + backref_ctx.extent_len = num_bytes; + + /* + * The last extent of a file may be too large due to page alignment. + * We need to adjust extent_len in this case so that the checks in + * __iterate_backrefs work. + */ + if (data_offset + num_bytes >= ino_size) + backref_ctx.extent_len = ino_size - data_offset; + + /* + * Now collect all backrefs. + */ + extent_item_pos = logical - found_key.objectid; + ret = iterate_extent_inodes(sctx->send_root->fs_info, + found_key.objectid, extent_item_pos, 1, + __iterate_backrefs, &backref_ctx); + if (ret < 0) + goto out; + + if (!backref_ctx.found_in_send_root) { + /* found a bug in backref code? */ + ret = -EIO; + printk(KERN_ERR "btrfs: ERROR did not find backref in " + "send_root. inode=%llu, offset=%llu, " + "logical=%llu\n", + ino, data_offset, logical); + goto out; + } + +verbose_printk(KERN_DEBUG "btrfs: find_extent_clone: data_offset=%llu, " + "ino=%llu, " + "num_bytes=%llu, logical=%llu\n", + data_offset, ino, num_bytes, logical); + + if (!backref_ctx.found) + verbose_printk("btrfs: no clones found\n"); + + cur_clone_root = NULL; + for (i = 0; i < sctx->clone_roots_cnt; i++) { + if (sctx->clone_roots[i].found_refs) { + if (!cur_clone_root) + cur_clone_root = sctx->clone_roots + i; + else if (sctx->clone_roots[i].root == sctx->send_root) + /* prefer clones from send_root over others */ + cur_clone_root = sctx->clone_roots + i; + break; + } + + } + + if (cur_clone_root) { + *found = cur_clone_root; + ret = 0; + } else { + ret = -ENOENT; + } + +out: + btrfs_free_path(tmp_path); + return ret; +} + +static int read_symlink(struct send_ctx *sctx, + struct btrfs_root *root, + u64 ino, + struct fs_path *dest) +{ + int ret; + struct btrfs_path *path; + struct btrfs_key key; + struct btrfs_file_extent_item *ei; + u8 type; + u8 compression; + unsigned long off; + int len; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + key.objectid = ino; + key.type = BTRFS_EXTENT_DATA_KEY; + key.offset = 0; + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + BUG_ON(ret); + + ei = btrfs_item_ptr(path->nodes[0], path->slots[0], + struct btrfs_file_extent_item); + type = btrfs_file_extent_type(path->nodes[0], ei); + compression = btrfs_file_extent_compression(path->nodes[0], ei); + BUG_ON(type != BTRFS_FILE_EXTENT_INLINE); + BUG_ON(compression); + + off = btrfs_file_extent_inline_start(ei); + len = btrfs_file_extent_inline_len(path->nodes[0], ei); + + ret = fs_path_add_from_extent_buffer(dest, path->nodes[0], off, len); + if (ret < 0) + goto out; + +out: + btrfs_free_path(path); + return ret; +} + +/* + * Helper function to generate a file name that is unique in the root of + * send_root and parent_root. This is used to generate names for orphan inodes. + */ +static int gen_unique_name(struct send_ctx *sctx, + u64 ino, u64 gen, + struct fs_path *dest) +{ + int ret = 0; + struct btrfs_path *path; + struct btrfs_dir_item *di; + char tmp[64]; + int len; + u64 idx = 0; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + while (1) { + len = snprintf(tmp, sizeof(tmp) - 1, "o%llu-%llu-%llu", + ino, gen, idx); + if (len >= sizeof(tmp)) { + /* should really not happen */ + ret = -EOVERFLOW; + goto out; + } + + di = btrfs_lookup_dir_item(NULL, sctx->send_root, + path, BTRFS_FIRST_FREE_OBJECTID, + tmp, strlen(tmp), 0); + btrfs_release_path(path); + if (IS_ERR(di)) { + ret = PTR_ERR(di); + goto out; + } + if (di) { + /* not unique, try again */ + idx++; + continue; + } + + if (!sctx->parent_root) { + /* unique */ + ret = 0; + break; + } + + di = btrfs_lookup_dir_item(NULL, sctx->parent_root, + path, BTRFS_FIRST_FREE_OBJECTID, + tmp, strlen(tmp), 0); + btrfs_release_path(path); + if (IS_ERR(di)) { + ret = PTR_ERR(di); + goto out; + } + if (di) { + /* not unique, try again */ + idx++; + continue; + } + /* unique */ + break; + } + + ret = fs_path_add(dest, tmp, strlen(tmp)); + +out: + btrfs_free_path(path); + return ret; +} + +enum inode_state { + inode_state_no_change, + inode_state_will_create, + inode_state_did_create, + inode_state_will_delete, + inode_state_did_delete, +}; + +static int get_cur_inode_state(struct send_ctx *sctx, u64 ino, u64 gen) +{ + int ret; + int left_ret; + int right_ret; + u64 left_gen; + u64 right_gen; + + ret = get_inode_info(sctx->send_root, ino, NULL, &left_gen, NULL, NULL, + NULL); + if (ret < 0 && ret != -ENOENT) + goto out; + left_ret = ret; + + if (!sctx->parent_root) { + right_ret = -ENOENT; + } else { + ret = get_inode_info(sctx->parent_root, ino, NULL, &right_gen, + NULL, NULL, NULL); + if (ret < 0 && ret != -ENOENT) + goto out; + right_ret = ret; + } + + if (!left_ret && !right_ret) { + if (left_gen == gen && right_gen == gen) + ret = inode_state_no_change; + else if (left_gen == gen) { + if (ino < sctx->send_progress) + ret = inode_state_did_create; + else + ret = inode_state_will_create; + } else if (right_gen == gen) { + if (ino < sctx->send_progress) + ret = inode_state_did_delete; + else + ret = inode_state_will_delete; + } else { + ret = -ENOENT; + } + } else if (!left_ret) { + if (left_gen == gen) { + if (ino < sctx->send_progress) + ret = inode_state_did_create; + else + ret = inode_state_will_create; + } else { + ret = -ENOENT; + } + } else if (!right_ret) { + if (right_gen == gen) { + if (ino < sctx->send_progress) + ret = inode_state_did_delete; + else + ret = inode_state_will_delete; + } else { + ret = -ENOENT; + } + } else { + ret = -ENOENT; + } + +out: + return ret; +} + +static int is_inode_existent(struct send_ctx *sctx, u64 ino, u64 gen) +{ + int ret; + + ret = get_cur_inode_state(sctx, ino, gen); + if (ret < 0) + goto out; + + if (ret == inode_state_no_change || + ret == inode_state_did_create || + ret == inode_state_will_delete) + ret = 1; + else + ret = 0; + +out: + return ret; +} + +/* + * Helper function to lookup a dir item in a dir. + */ +static int lookup_dir_item_inode(struct btrfs_root *root, + u64 dir, const char *name, int name_len, + u64 *found_inode, + u8 *found_type) +{ + int ret = 0; + struct btrfs_dir_item *di; + struct btrfs_key key; + struct btrfs_path *path; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + di = btrfs_lookup_dir_item(NULL, root, path, + dir, name, name_len, 0); + if (!di) { + ret = -ENOENT; + goto out; + } + if (IS_ERR(di)) { + ret = PTR_ERR(di); + goto out; + } + btrfs_dir_item_key_to_cpu(path->nodes[0], di, &key); + *found_inode = key.objectid; + *found_type = btrfs_dir_type(path->nodes[0], di); + +out: + btrfs_free_path(path); + return ret; +} + +static int get_first_ref(struct send_ctx *sctx, + struct btrfs_root *root, u64 ino, + u64 *dir, u64 *dir_gen, struct fs_path *name) +{ + int ret; + struct btrfs_key key; + struct btrfs_key found_key; + struct btrfs_path *path; + struct btrfs_inode_ref *iref; + int len; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + key.objectid = ino; + key.type = BTRFS_INODE_REF_KEY; + key.offset = 0; + + ret = btrfs_search_slot_for_read(root, &key, path, 1, 0); + if (ret < 0) + goto out; + if (!ret) + btrfs_item_key_to_cpu(path->nodes[0], &found_key, + path->slots[0]); + if (ret || found_key.objectid != key.objectid || + found_key.type != key.type) { + ret = -ENOENT; + goto out; + } + + iref = btrfs_item_ptr(path->nodes[0], path->slots[0], + struct btrfs_inode_ref); + len = btrfs_inode_ref_name_len(path->nodes[0], iref); + ret = fs_path_add_from_extent_buffer(name, path->nodes[0], + (unsigned long)(iref + 1), len); + if (ret < 0) + goto out; + btrfs_release_path(path); + + ret = get_inode_info(root, found_key.offset, NULL, dir_gen, NULL, NULL, + NULL); + if (ret < 0) + goto out; + + *dir = found_key.offset; + +out: + btrfs_free_path(path); + return ret; +} + +static int is_first_ref(struct send_ctx *sctx, + struct btrfs_root *root, + u64 ino, u64 dir, + const char *name, int name_len) +{ + int ret; + struct fs_path *tmp_name; + u64 tmp_dir; + u64 tmp_dir_gen; + + tmp_name = fs_path_alloc(sctx); + if (!tmp_name) + return -ENOMEM; + + ret = get_first_ref(sctx, root, ino, &tmp_dir, &tmp_dir_gen, tmp_name); + if (ret < 0) + goto out; + + if (name_len != fs_path_len(tmp_name)) { + ret = 0; + goto out; + } + + ret = memcmp(tmp_name->start, name, name_len); + if (ret) + ret = 0; + else + ret = 1; + +out: + fs_path_free(sctx, tmp_name); + return ret; +} + +static int will_overwrite_ref(struct send_ctx *sctx, u64 dir, u64 dir_gen, + const char *name, int name_len, + u64 *who_ino, u64 *who_gen) +{ + int ret = 0; + u64 other_inode = 0; + u8 other_type = 0; + + if (!sctx->parent_root) + goto out; + + ret = is_inode_existent(sctx, dir, dir_gen); + if (ret <= 0) + goto out; + + ret = lookup_dir_item_inode(sctx->parent_root, dir, name, name_len, + &other_inode, &other_type); + if (ret < 0 && ret != -ENOENT) + goto out; + if (ret) { + ret = 0; + goto out; + } + + if (other_inode > sctx->send_progress) { + ret = get_inode_info(sctx->parent_root, other_inode, NULL, + who_gen, NULL, NULL, NULL); + if (ret < 0) + goto out; + + ret = 1; + *who_ino = other_inode; + } else { + ret = 0; + } + +out: + return ret; +} + +static int did_overwrite_ref(struct send_ctx *sctx, + u64 dir, u64 dir_gen, + u64 ino, u64 ino_gen, + const char *name, int name_len) +{ + int ret = 0; + u64 gen; + u64 ow_inode; + u8 other_type; + + if (!sctx->parent_root) + goto out; + + ret = is_inode_existent(sctx, dir, dir_gen); + if (ret <= 0) + goto out; + + /* check if the ref was overwritten by another ref */ + ret = lookup_dir_item_inode(sctx->send_root, dir, name, name_len, + &ow_inode, &other_type); + if (ret < 0 && ret != -ENOENT) + goto out; + if (ret) { + /* was never and will never be overwritten */ + ret = 0; + goto out; + } + + ret = get_inode_info(sctx->send_root, ow_inode, NULL, &gen, NULL, NULL, + NULL); + if (ret < 0) + goto out; + + if (ow_inode == ino && gen == ino_gen) { + ret = 0; + goto out; + } + + /* we know that it is or will be overwritten. check this now */ + if (ow_inode < sctx->send_progress) + ret = 1; + else + ret = 0; + +out: + return ret; +} + +static int did_overwrite_first_ref(struct send_ctx *sctx, u64 ino, u64 gen) +{ + int ret = 0; + struct fs_path *name = NULL; + u64 dir; + u64 dir_gen; + + if (!sctx->parent_root) + goto out; + + name = fs_path_alloc(sctx); + if (!name) + return -ENOMEM; + + ret = get_first_ref(sctx, sctx->parent_root, ino, &dir, &dir_gen, name); + if (ret < 0) + goto out; + + ret = did_overwrite_ref(sctx, dir, dir_gen, ino, gen, + name->start, fs_path_len(name)); + if (ret < 0) + goto out; + +out: + fs_path_free(sctx, name); + return ret; +} + +static int name_cache_insert(struct send_ctx *sctx, + struct name_cache_entry *nce) +{ + int ret = 0; + struct name_cache_entry **ncea; + + ncea = radix_tree_lookup(&sctx->name_cache, nce->ino); + if (ncea) { + if (!ncea[0]) + ncea[0] = nce; + else if (!ncea[1]) + ncea[1] = nce; + else + BUG(); + } else { + ncea = kmalloc(sizeof(void *) * 2, GFP_NOFS); + if (!ncea) + return -ENOMEM; + + ncea[0] = nce; + ncea[1] = NULL; + ret = radix_tree_insert(&sctx->name_cache, nce->ino, ncea); + if (ret < 0) + return ret; + } + list_add_tail(&nce->list, &sctx->name_cache_list); + sctx->name_cache_size++; + + return ret; +} + +static void name_cache_delete(struct send_ctx *sctx, + struct name_cache_entry *nce) +{ + struct name_cache_entry **ncea; + + ncea = radix_tree_lookup(&sctx->name_cache, nce->ino); + BUG_ON(!ncea); + + if (ncea[0] == nce) + ncea[0] = NULL; + else if (ncea[1] == nce) + ncea[1] = NULL; + else + BUG(); + + if (!ncea[0] && !ncea[1]) { + radix_tree_delete(&sctx->name_cache, nce->ino); + kfree(ncea); + } + + list_del(&nce->list); + + sctx->name_cache_size--; +} + +static struct name_cache_entry *name_cache_search(struct send_ctx *sctx, + u64 ino, u64 gen) +{ + struct name_cache_entry **ncea; + + ncea = radix_tree_lookup(&sctx->name_cache, ino); + if (!ncea) + return NULL; + + if (ncea[0] && ncea[0]->gen == gen) + return ncea[0]; + else if (ncea[1] && ncea[1]->gen == gen) + return ncea[1]; + return NULL; +} + +static void name_cache_used(struct send_ctx *sctx, struct name_cache_entry *nce) +{ + list_del(&nce->list); + list_add_tail(&nce->list, &sctx->name_cache_list); +} + +static void name_cache_clean_unused(struct send_ctx *sctx) +{ + struct name_cache_entry *nce; + + if (sctx->name_cache_size < SEND_CTX_NAME_CACHE_CLEAN_SIZE) + return; + + while (sctx->name_cache_size > SEND_CTX_MAX_NAME_CACHE_SIZE) { + nce = list_entry(sctx->name_cache_list.next, + struct name_cache_entry, list); + name_cache_delete(sctx, nce); + kfree(nce); + } +} + +static void name_cache_free(struct send_ctx *sctx) +{ + struct name_cache_entry *nce; + struct name_cache_entry *tmp; + + list_for_each_entry_safe(nce, tmp, &sctx->name_cache_list, list) { + name_cache_delete(sctx, nce); + } +} + +static int __get_cur_name_and_parent(struct send_ctx *sctx, + u64 ino, u64 gen, + u64 *parent_ino, + u64 *parent_gen, + struct fs_path *dest) +{ + int ret; + int nce_ret; + struct btrfs_path *path = NULL; + struct name_cache_entry *nce = NULL; + + nce = name_cache_search(sctx, ino, gen); + if (nce) { + if (ino < sctx->send_progress && nce->need_later_update) { + name_cache_delete(sctx, nce); + kfree(nce); + nce = NULL; + } else { + name_cache_used(sctx, nce); + *parent_ino = nce->parent_ino; + *parent_gen = nce->parent_gen; + ret = fs_path_add(dest, nce->name, nce->name_len); + if (ret < 0) + goto out; + ret = nce->ret; + goto out; + } + } + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + ret = is_inode_existent(sctx, ino, gen); + if (ret < 0) + goto out; + + if (!ret) { + ret = gen_unique_name(sctx, ino, gen, dest); + if (ret < 0) + goto out; + ret = 1; + goto out_cache; + } + + if (ino < sctx->send_progress) + ret = get_first_ref(sctx, sctx->send_root, ino, + parent_ino, parent_gen, dest); + else + ret = get_first_ref(sctx, sctx->parent_root, ino, + parent_ino, parent_gen, dest); + if (ret < 0) + goto out; + + ret = did_overwrite_ref(sctx, *parent_ino, *parent_gen, ino, gen, + dest->start, dest->end - dest->start); + if (ret < 0) + goto out; + if (ret) { + fs_path_reset(dest); + ret = gen_unique_name(sctx, ino, gen, dest); + if (ret < 0) + goto out; + ret = 1; + } + +out_cache: + nce = kmalloc(sizeof(*nce) + fs_path_len(dest) + 1, GFP_NOFS); + if (!nce) { + ret = -ENOMEM; + goto out; + } + + nce->ino = ino; + nce->gen = gen; + nce->parent_ino = *parent_ino; + nce->parent_gen = *parent_gen; + nce->name_len = fs_path_len(dest); + nce->ret = ret; + strcpy(nce->name, dest->start); + memset(&nce->use_list, 0, sizeof(nce->use_list)); + + if (ino < sctx->send_progress) + nce->need_later_update = 0; + else + nce->need_later_update = 1; + + nce_ret = name_cache_insert(sctx, nce); + if (nce_ret < 0) + ret = nce_ret; + name_cache_clean_unused(sctx); + +out: + btrfs_free_path(path); + return ret; +} + +/* + * Magic happens here. This function returns the first ref to an inode as it + * would look like while receiving the stream at this point in time. + * We walk the path up to the root. For every inode in between, we check if it + * was already processed/sent. If yes, we continue with the parent as found + * in send_root. If not, we continue with the parent as found in parent_root. + * If we encounter an inode that was deleted at this point in time, we use the + * inodes "orphan" name instead of the real name and stop. Same with new inodes + * that were not created yet and overwritten inodes/refs. + * + * When do we have have orphan inodes: + * 1. When an inode is freshly created and thus no valid refs are available yet + * 2. When a directory lost all it's refs (deleted) but still has dir items + * inside which were not processed yet (pending for move/delete). If anyone + * tried to get the path to the dir items, it would get a path inside that + * orphan directory. + * 3. When an inode is moved around or gets new links, it may overwrite the ref + * of an unprocessed inode. If in that case the first ref would be + * overwritten, the overwritten inode gets "orphanized". Later when we + * process this overwritten inode, it is restored at a new place by moving + * the orphan inode. + * + * sctx->send_progress tells this function at which point in time receiving + * would be. + */ +static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen, + struct fs_path *dest) +{ + int ret = 0; + struct fs_path *name = NULL; + u64 parent_inode = 0; + u64 parent_gen = 0; + int stop = 0; + + name = fs_path_alloc(sctx); + if (!name) { + ret = -ENOMEM; + goto out; + } + + dest->reversed = 1; + fs_path_reset(dest); + + while (!stop && ino != BTRFS_FIRST_FREE_OBJECTID) { + fs_path_reset(name); + + ret = __get_cur_name_and_parent(sctx, ino, gen, + &parent_inode, &parent_gen, name); + if (ret < 0) + goto out; + if (ret) + stop = 1; + + ret = fs_path_add_path(dest, name); + if (ret < 0) + goto out; + + ino = parent_inode; + gen = parent_gen; + } + +out: + fs_path_free(sctx, name); + if (!ret) + fs_path_unreverse(dest); + return ret; +} + +/* + * Called for regular files when sending extents data. Opens a struct file + * to read from the file. + */ +static int open_cur_inode_file(struct send_ctx *sctx) +{ + int ret = 0; + struct btrfs_key key; + struct path path; + struct inode *inode; + struct dentry *dentry; + struct file *filp; + int new = 0; + + if (sctx->cur_inode_filp) + goto out; + + key.objectid = sctx->cur_ino; + key.type = BTRFS_INODE_ITEM_KEY; + key.offset = 0; + + inode = btrfs_iget(sctx->send_root->fs_info->sb, &key, sctx->send_root, + &new); + if (IS_ERR(inode)) { + ret = PTR_ERR(inode); + goto out; + } + + dentry = d_obtain_alias(inode); + inode = NULL; + if (IS_ERR(dentry)) { + ret = PTR_ERR(dentry); + goto out; + } + + path.mnt = sctx->mnt; + path.dentry = dentry; + filp = dentry_open(&path, O_RDONLY | O_LARGEFILE, current_cred()); + dput(dentry); + dentry = NULL; + if (IS_ERR(filp)) { + ret = PTR_ERR(filp); + goto out; + } + sctx->cur_inode_filp = filp; + +out: + /* + * no xxxput required here as every vfs op + * does it by itself on failure + */ + return ret; +} + +/* + * Closes the struct file that was created in open_cur_inode_file + */ +static int close_cur_inode_file(struct send_ctx *sctx) +{ + int ret = 0; + + if (!sctx->cur_inode_filp) + goto out; + + ret = filp_close(sctx->cur_inode_filp, NULL); + sctx->cur_inode_filp = NULL; + +out: + return ret; +} + +/* + * Sends a BTRFS_SEND_C_SUBVOL command/item to userspace + */ +static int send_subvol_begin(struct send_ctx *sctx) +{ + int ret; + struct btrfs_root *send_root = sctx->send_root; + struct btrfs_root *parent_root = sctx->parent_root; + struct btrfs_path *path; + struct btrfs_key key; + struct btrfs_root_ref *ref; + struct extent_buffer *leaf; + char *name = NULL; + int namelen; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + name = kmalloc(BTRFS_PATH_NAME_MAX, GFP_NOFS); + if (!name) { + btrfs_free_path(path); + return -ENOMEM; + } + + key.objectid = send_root->objectid; + key.type = BTRFS_ROOT_BACKREF_KEY; + key.offset = 0; + + ret = btrfs_search_slot_for_read(send_root->fs_info->tree_root, + &key, path, 1, 0); + if (ret < 0) + goto out; + if (ret) { + ret = -ENOENT; + goto out; + } + + leaf = path->nodes[0]; + btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); + if (key.type != BTRFS_ROOT_BACKREF_KEY || + key.objectid != send_root->objectid) { + ret = -ENOENT; + goto out; + } + ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); + namelen = btrfs_root_ref_name_len(leaf, ref); + read_extent_buffer(leaf, name, (unsigned long)(ref + 1), namelen); + btrfs_release_path(path); + + if (ret < 0) + goto out; + + if (parent_root) { + ret = begin_cmd(sctx, BTRFS_SEND_C_SNAPSHOT); + if (ret < 0) + goto out; + } else { + ret = begin_cmd(sctx, BTRFS_SEND_C_SUBVOL); + if (ret < 0) + goto out; + } + + TLV_PUT_STRING(sctx, BTRFS_SEND_A_PATH, name, namelen); + TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID, + sctx->send_root->root_item.uuid); + TLV_PUT_U64(sctx, BTRFS_SEND_A_CTRANSID, + sctx->send_root->root_item.ctransid); + if (parent_root) { + TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID, + sctx->parent_root->root_item.uuid); + TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_CTRANSID, + sctx->parent_root->root_item.ctransid); + } + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + btrfs_free_path(path); + kfree(name); + return ret; +} + +static int send_truncate(struct send_ctx *sctx, u64 ino, u64 gen, u64 size) +{ + int ret = 0; + struct fs_path *p; + +verbose_printk("btrfs: send_truncate %llu size=%llu\n", ino, size); + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ret = begin_cmd(sctx, BTRFS_SEND_C_TRUNCATE); + if (ret < 0) + goto out; + + ret = get_cur_path(sctx, ino, gen, p); + if (ret < 0) + goto out; + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p); + TLV_PUT_U64(sctx, BTRFS_SEND_A_SIZE, size); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + fs_path_free(sctx, p); + return ret; +} + +static int send_chmod(struct send_ctx *sctx, u64 ino, u64 gen, u64 mode) +{ + int ret = 0; + struct fs_path *p; + +verbose_printk("btrfs: send_chmod %llu mode=%llu\n", ino, mode); + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ret = begin_cmd(sctx, BTRFS_SEND_C_CHMOD); + if (ret < 0) + goto out; + + ret = get_cur_path(sctx, ino, gen, p); + if (ret < 0) + goto out; + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p); + TLV_PUT_U64(sctx, BTRFS_SEND_A_MODE, mode & 07777); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + fs_path_free(sctx, p); + return ret; +} + +static int send_chown(struct send_ctx *sctx, u64 ino, u64 gen, u64 uid, u64 gid) +{ + int ret = 0; + struct fs_path *p; + +verbose_printk("btrfs: send_chown %llu uid=%llu, gid=%llu\n", ino, uid, gid); + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ret = begin_cmd(sctx, BTRFS_SEND_C_CHOWN); + if (ret < 0) + goto out; + + ret = get_cur_path(sctx, ino, gen, p); + if (ret < 0) + goto out; + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p); + TLV_PUT_U64(sctx, BTRFS_SEND_A_UID, uid); + TLV_PUT_U64(sctx, BTRFS_SEND_A_GID, gid); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + fs_path_free(sctx, p); + return ret; +} + +static int send_utimes(struct send_ctx *sctx, u64 ino, u64 gen) +{ + int ret = 0; + struct fs_path *p = NULL; + struct btrfs_inode_item *ii; + struct btrfs_path *path = NULL; + struct extent_buffer *eb; + struct btrfs_key key; + int slot; + +verbose_printk("btrfs: send_utimes %llu\n", ino); + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + path = alloc_path_for_send(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + key.objectid = ino; + key.type = BTRFS_INODE_ITEM_KEY; + key.offset = 0; + ret = btrfs_search_slot(NULL, sctx->send_root, &key, path, 0, 0); + if (ret < 0) + goto out; + + eb = path->nodes[0]; + slot = path->slots[0]; + ii = btrfs_item_ptr(eb, slot, struct btrfs_inode_item); + + ret = begin_cmd(sctx, BTRFS_SEND_C_UTIMES); + if (ret < 0) + goto out; + + ret = get_cur_path(sctx, ino, gen, p); + if (ret < 0) + goto out; + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p); + TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_ATIME, eb, + btrfs_inode_atime(ii)); + TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_MTIME, eb, + btrfs_inode_mtime(ii)); + TLV_PUT_BTRFS_TIMESPEC(sctx, BTRFS_SEND_A_CTIME, eb, + btrfs_inode_ctime(ii)); + /* TODO otime? */ + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + fs_path_free(sctx, p); + btrfs_free_path(path); + return ret; +} + +/* + * Sends a BTRFS_SEND_C_MKXXX or SYMLINK command to user space. We don't have + * a valid path yet because we did not process the refs yet. So, the inode + * is created as orphan. + */ +static int send_create_inode(struct send_ctx *sctx, struct btrfs_path *path, + struct btrfs_key *key) +{ + int ret = 0; + struct extent_buffer *eb = path->nodes[0]; + struct btrfs_inode_item *ii; + struct fs_path *p; + int slot = path->slots[0]; + int cmd; + u64 mode; + +verbose_printk("btrfs: send_create_inode %llu\n", sctx->cur_ino); + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ii = btrfs_item_ptr(eb, slot, struct btrfs_inode_item); + mode = btrfs_inode_mode(eb, ii); + + if (S_ISREG(mode)) + cmd = BTRFS_SEND_C_MKFILE; + else if (S_ISDIR(mode)) + cmd = BTRFS_SEND_C_MKDIR; + else if (S_ISLNK(mode)) + cmd = BTRFS_SEND_C_SYMLINK; + else if (S_ISCHR(mode) || S_ISBLK(mode)) + cmd = BTRFS_SEND_C_MKNOD; + else if (S_ISFIFO(mode)) + cmd = BTRFS_SEND_C_MKFIFO; + else if (S_ISSOCK(mode)) + cmd = BTRFS_SEND_C_MKSOCK; + else { + printk(KERN_WARNING "btrfs: unexpected inode type %o", + (int)(mode & S_IFMT)); + ret = -ENOTSUPP; + goto out; + } + + ret = begin_cmd(sctx, cmd); + if (ret < 0) + goto out; + + ret = gen_unique_name(sctx, sctx->cur_ino, sctx->cur_inode_gen, p); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p); + TLV_PUT_U64(sctx, BTRFS_SEND_A_INO, sctx->cur_ino); + + if (S_ISLNK(mode)) { + fs_path_reset(p); + ret = read_symlink(sctx, sctx->send_root, sctx->cur_ino, p); + if (ret < 0) + goto out; + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH_LINK, p); + } else if (S_ISCHR(mode) || S_ISBLK(mode) || + S_ISFIFO(mode) || S_ISSOCK(mode)) { + TLV_PUT_U64(sctx, BTRFS_SEND_A_RDEV, btrfs_inode_rdev(eb, ii)); + } + + ret = send_cmd(sctx); + if (ret < 0) + goto out; + + +tlv_put_failure: +out: + fs_path_free(sctx, p); + return ret; +} + +struct recorded_ref { + struct list_head list; + char *dir_path; + char *name; + struct fs_path *full_path; + u64 dir; + u64 dir_gen; + int dir_path_len; + int name_len; +}; + +/* + * We need to process new refs before deleted refs, but compare_tree gives us + * everything mixed. So we first record all refs and later process them. + * This function is a helper to record one ref. + */ +static int record_ref(struct list_head *head, u64 dir, + u64 dir_gen, struct fs_path *path) +{ + struct recorded_ref *ref; + char *tmp; + + ref = kmalloc(sizeof(*ref), GFP_NOFS); + if (!ref) + return -ENOMEM; + + ref->dir = dir; + ref->dir_gen = dir_gen; + ref->full_path = path; + + tmp = strrchr(ref->full_path->start, '/'); + if (!tmp) { + ref->name_len = ref->full_path->end - ref->full_path->start; + ref->name = ref->full_path->start; + ref->dir_path_len = 0; + ref->dir_path = ref->full_path->start; + } else { + tmp++; + ref->name_len = ref->full_path->end - tmp; + ref->name = tmp; + ref->dir_path = ref->full_path->start; + ref->dir_path_len = ref->full_path->end - + ref->full_path->start - 1 - ref->name_len; + } + + list_add_tail(&ref->list, head); + return 0; +} + +static void __free_recorded_refs(struct send_ctx *sctx, struct list_head *head) +{ + struct recorded_ref *cur; + struct recorded_ref *tmp; + + list_for_each_entry_safe(cur, tmp, head, list) { + fs_path_free(sctx, cur->full_path); + kfree(cur); + } + INIT_LIST_HEAD(head); +} + +static void free_recorded_refs(struct send_ctx *sctx) +{ + __free_recorded_refs(sctx, &sctx->new_refs); + __free_recorded_refs(sctx, &sctx->deleted_refs); +} + +/* + * Renames/moves a file/dir to it's orphan name. Used when the first + * ref of an unprocessed inode gets overwritten and for all non empty + * directories. + */ +static int orphanize_inode(struct send_ctx *sctx, u64 ino, u64 gen, + struct fs_path *path) +{ + int ret; + struct fs_path *orphan; + + orphan = fs_path_alloc(sctx); + if (!orphan) + return -ENOMEM; + + ret = gen_unique_name(sctx, ino, gen, orphan); + if (ret < 0) + goto out; + + ret = send_rename(sctx, path, orphan); + +out: + fs_path_free(sctx, orphan); + return ret; +} + +/* + * Returns 1 if a directory can be removed at this point in time. + * We check this by iterating all dir items and checking if the inode behind + * the dir item was already processed. + */ +static int can_rmdir(struct send_ctx *sctx, u64 dir, u64 send_progress) +{ + int ret = 0; + struct btrfs_root *root = sctx->parent_root; + struct btrfs_path *path; + struct btrfs_key key; + struct btrfs_key found_key; + struct btrfs_key loc; + struct btrfs_dir_item *di; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + key.objectid = dir; + key.type = BTRFS_DIR_INDEX_KEY; + key.offset = 0; + + while (1) { + ret = btrfs_search_slot_for_read(root, &key, path, 1, 0); + if (ret < 0) + goto out; + if (!ret) { + btrfs_item_key_to_cpu(path->nodes[0], &found_key, + path->slots[0]); + } + if (ret || found_key.objectid != key.objectid || + found_key.type != key.type) { + break; + } + + di = btrfs_item_ptr(path->nodes[0], path->slots[0], + struct btrfs_dir_item); + btrfs_dir_item_key_to_cpu(path->nodes[0], di, &loc); + + if (loc.objectid > send_progress) { + ret = 0; + goto out; + } + + btrfs_release_path(path); + key.offset = found_key.offset + 1; + } + + ret = 1; + +out: + btrfs_free_path(path); + return ret; +} + +struct finish_unordered_dir_ctx { + struct send_ctx *sctx; + struct fs_path *cur_path; + struct fs_path *dir_path; + u64 dir_ino; + int need_delete; + int delete_pass; +}; + +int __finish_unordered_dir(int num, struct btrfs_key *di_key, + const char *name, int name_len, + const char *data, int data_len, + u8 type, void *ctx) +{ + int ret = 0; + struct finish_unordered_dir_ctx *fctx = ctx; + struct send_ctx *sctx = fctx->sctx; + u64 di_gen; + u64 di_mode; + int is_orphan = 0; + + if (di_key->objectid >= fctx->dir_ino) + goto out; + + fs_path_reset(fctx->cur_path); + + ret = get_inode_info(sctx->send_root, di_key->objectid, + NULL, &di_gen, &di_mode, NULL, NULL); + if (ret < 0) + goto out; + + ret = is_first_ref(sctx, sctx->send_root, di_key->objectid, + fctx->dir_ino, name, name_len); + if (ret < 0) + goto out; + if (ret) { + is_orphan = 1; + ret = gen_unique_name(sctx, di_key->objectid, di_gen, + fctx->cur_path); + } else { + ret = get_cur_path(sctx, di_key->objectid, di_gen, + fctx->cur_path); + } + if (ret < 0) + goto out; + + ret = fs_path_add(fctx->dir_path, name, name_len); + if (ret < 0) + goto out; + + if (!fctx->delete_pass) { + if (S_ISDIR(di_mode)) { + ret = send_rename(sctx, fctx->cur_path, + fctx->dir_path); + } else { + ret = send_link(sctx, fctx->dir_path, + fctx->cur_path); + if (is_orphan) + fctx->need_delete = 1; + } + } else if (!S_ISDIR(di_mode)) { + ret = send_unlink(sctx, fctx->cur_path); + } else { + ret = 0; + } + + fs_path_remove(fctx->dir_path); + +out: + return ret; +} + +/* + * Go through all dir items and see if we find refs which could not be created + * in the past because the dir did not exist at that time. + */ +static int finish_outoforder_dir(struct send_ctx *sctx, u64 dir, u64 dir_gen) +{ + int ret = 0; + struct btrfs_path *path = NULL; + struct btrfs_key key; + struct btrfs_key found_key; + struct extent_buffer *eb; + struct finish_unordered_dir_ctx fctx; + int slot; + + path = alloc_path_for_send(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + memset(&fctx, 0, sizeof(fctx)); + fctx.sctx = sctx; + fctx.cur_path = fs_path_alloc(sctx); + fctx.dir_path = fs_path_alloc(sctx); + if (!fctx.cur_path || !fctx.dir_path) { + ret = -ENOMEM; + goto out; + } + fctx.dir_ino = dir; + + ret = get_cur_path(sctx, dir, dir_gen, fctx.dir_path); + if (ret < 0) + goto out; + + /* + * We do two passes. The first links in the new refs and the second + * deletes orphans if required. Deletion of orphans is not required for + * directory inodes, as we always have only one ref and use rename + * instead of link for those. + */ + +again: + key.objectid = dir; + key.type = BTRFS_DIR_ITEM_KEY; + key.offset = 0; + while (1) { + ret = btrfs_search_slot_for_read(sctx->send_root, &key, path, + 1, 0); + if (ret < 0) + goto out; + eb = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + + if (found_key.objectid != key.objectid || + found_key.type != key.type) { + btrfs_release_path(path); + break; + } + + ret = iterate_dir_item(sctx, sctx->send_root, path, + &found_key, __finish_unordered_dir, + &fctx); + if (ret < 0) + goto out; + + key.offset = found_key.offset + 1; + btrfs_release_path(path); + } + + if (!fctx.delete_pass && fctx.need_delete) { + fctx.delete_pass = 1; + goto again; + } + +out: + btrfs_free_path(path); + fs_path_free(sctx, fctx.cur_path); + fs_path_free(sctx, fctx.dir_path); + return ret; +} + +/* + * This does all the move/link/unlink/rmdir magic. + */ +static int process_recorded_refs(struct send_ctx *sctx) +{ + int ret = 0; + struct recorded_ref *cur; + struct ulist *check_dirs = NULL; + struct ulist_iterator uit; + struct ulist_node *un; + struct fs_path *valid_path = NULL; + u64 ow_inode = 0; + u64 ow_gen; + int did_overwrite = 0; + int is_orphan = 0; + +verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino); + + valid_path = fs_path_alloc(sctx); + if (!valid_path) { + ret = -ENOMEM; + goto out; + } + + check_dirs = ulist_alloc(GFP_NOFS); + if (!check_dirs) { + ret = -ENOMEM; + goto out; + } + + /* + * First, check if the first ref of the current inode was overwritten + * before. If yes, we know that the current inode was already orphanized + * and thus use the orphan name. If not, we can use get_cur_path to + * get the path of the first ref as it would like while receiving at + * this point in time. + * New inodes are always orphan at the beginning, so force to use the + * orphan name in this case. + * The first ref is stored in valid_path and will be updated if it + * gets moved around. + */ + if (!sctx->cur_inode_new) { + ret = did_overwrite_first_ref(sctx, sctx->cur_ino, + sctx->cur_inode_gen); + if (ret < 0) + goto out; + if (ret) + did_overwrite = 1; + } + if (sctx->cur_inode_new || did_overwrite) { + ret = gen_unique_name(sctx, sctx->cur_ino, + sctx->cur_inode_gen, valid_path); + if (ret < 0) + goto out; + is_orphan = 1; + } else { + ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, + valid_path); + if (ret < 0) + goto out; + } + + list_for_each_entry(cur, &sctx->new_refs, list) { + /* + * Check if this new ref would overwrite the first ref of + * another unprocessed inode. If yes, orphanize the + * overwritten inode. If we find an overwritten ref that is + * not the first ref, simply unlink it. + */ + ret = will_overwrite_ref(sctx, cur->dir, cur->dir_gen, + cur->name, cur->name_len, + &ow_inode, &ow_gen); + if (ret < 0) + goto out; + if (ret) { + ret = is_first_ref(sctx, sctx->parent_root, + ow_inode, cur->dir, cur->name, + cur->name_len); + if (ret < 0) + goto out; + if (ret) { + ret = orphanize_inode(sctx, ow_inode, ow_gen, + cur->full_path); + if (ret < 0) + goto out; + } else { + ret = send_unlink(sctx, cur->full_path); + if (ret < 0) + goto out; + } + } + + /* + * link/move the ref to the new place. If we have an orphan + * inode, move it and update valid_path. If not, link or move + * it depending on the inode mode. + */ + if (is_orphan && !sctx->cur_inode_first_ref_orphan) { + ret = send_rename(sctx, valid_path, cur->full_path); + if (ret < 0) + goto out; + is_orphan = 0; + ret = fs_path_copy(valid_path, cur->full_path); + if (ret < 0) + goto out; + } else { + if (S_ISDIR(sctx->cur_inode_mode)) { + /* + * Dirs can't be linked, so move it. For moved + * dirs, we always have one new and one deleted + * ref. The deleted ref is ignored later. + */ + ret = send_rename(sctx, valid_path, + cur->full_path); + if (ret < 0) + goto out; + ret = fs_path_copy(valid_path, cur->full_path); + if (ret < 0) + goto out; + } else { + ret = send_link(sctx, cur->full_path, + valid_path); + if (ret < 0) + goto out; + } + } + ret = ulist_add(check_dirs, cur->dir, cur->dir_gen, + GFP_NOFS); + if (ret < 0) + goto out; + } + + if (S_ISDIR(sctx->cur_inode_mode) && sctx->cur_inode_deleted) { + /* + * Check if we can already rmdir the directory. If not, + * orphanize it. For every dir item inside that gets deleted + * later, we do this check again and rmdir it then if possible. + * See the use of check_dirs for more details. + */ + ret = can_rmdir(sctx, sctx->cur_ino, sctx->cur_ino); + if (ret < 0) + goto out; + if (ret) { + ret = send_rmdir(sctx, valid_path); + if (ret < 0) + goto out; + } else if (!is_orphan) { + ret = orphanize_inode(sctx, sctx->cur_ino, + sctx->cur_inode_gen, valid_path); + if (ret < 0) + goto out; + is_orphan = 1; + } + + list_for_each_entry(cur, &sctx->deleted_refs, list) { + ret = ulist_add(check_dirs, cur->dir, cur->dir_gen, + GFP_NOFS); + if (ret < 0) + goto out; + } + } else if (!S_ISDIR(sctx->cur_inode_mode)) { + /* + * We have a non dir inode. Go through all deleted refs and + * unlink them if they were not already overwritten by other + * inodes. + */ + list_for_each_entry(cur, &sctx->deleted_refs, list) { + ret = did_overwrite_ref(sctx, cur->dir, cur->dir_gen, + sctx->cur_ino, sctx->cur_inode_gen, + cur->name, cur->name_len); + if (ret < 0) + goto out; + if (!ret) { + /* + * In case the inode was moved to a directory + * that was not created yet (see + * __record_new_ref), we can not unlink the ref + * as it will be needed later when the parent + * directory is created, so that we can move in + * the inode to the new dir. + */ + if (!is_orphan && + sctx->cur_inode_first_ref_orphan) { + ret = orphanize_inode(sctx, + sctx->cur_ino, + sctx->cur_inode_gen, + cur->full_path); + if (ret < 0) + goto out; + ret = gen_unique_name(sctx, + sctx->cur_ino, + sctx->cur_inode_gen, + valid_path); + if (ret < 0) + goto out; + is_orphan = 1; + + } else { + ret = send_unlink(sctx, cur->full_path); + if (ret < 0) + goto out; + } + } + ret = ulist_add(check_dirs, cur->dir, cur->dir_gen, + GFP_NOFS); + if (ret < 0) + goto out; + } + + /* + * If the inode is still orphan, unlink the orphan. This may + * happen when a previous inode did overwrite the first ref + * of this inode and no new refs were added for the current + * inode. + * We can however not delete the orphan in case the inode relies + * in a directory that was not created yet (see + * __record_new_ref) + */ + if (is_orphan && !sctx->cur_inode_first_ref_orphan) { + ret = send_unlink(sctx, valid_path); + if (ret < 0) + goto out; + } + } + + /* + * We did collect all parent dirs where cur_inode was once located. We + * now go through all these dirs and check if they are pending for + * deletion and if it's finally possible to perform the rmdir now. + * We also update the inode stats of the parent dirs here. + */ + ULIST_ITER_INIT(&uit); + while ((un = ulist_next(check_dirs, &uit))) { + if (un->val > sctx->cur_ino) + continue; + + ret = get_cur_inode_state(sctx, un->val, un->aux); + if (ret < 0) + goto out; + + if (ret == inode_state_did_create || + ret == inode_state_no_change) { + /* TODO delayed utimes */ + ret = send_utimes(sctx, un->val, un->aux); + if (ret < 0) + goto out; + } else if (ret == inode_state_did_delete) { + ret = can_rmdir(sctx, un->val, sctx->cur_ino); + if (ret < 0) + goto out; + if (ret) { + ret = get_cur_path(sctx, un->val, un->aux, + valid_path); + if (ret < 0) + goto out; + ret = send_rmdir(sctx, valid_path); + if (ret < 0) + goto out; + } + } + } + + /* + * Current inode is now at it's new position, so we must increase + * send_progress + */ + sctx->send_progress = sctx->cur_ino + 1; + + /* + * We may have a directory here that has pending refs which could not + * be created before (because the dir did not exist before, see + * __record_new_ref). finish_outoforder_dir will link/move the pending + * refs. + */ + if (S_ISDIR(sctx->cur_inode_mode) && sctx->cur_inode_new) { + ret = finish_outoforder_dir(sctx, sctx->cur_ino, + sctx->cur_inode_gen); + if (ret < 0) + goto out; + } + + ret = 0; + +out: + free_recorded_refs(sctx); + ulist_free(check_dirs); + fs_path_free(sctx, valid_path); + return ret; +} + +static int __record_new_ref(int num, u64 dir, int index, + struct fs_path *name, + void *ctx) +{ + int ret = 0; + struct send_ctx *sctx = ctx; + struct fs_path *p; + u64 gen; + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ret = get_inode_info(sctx->send_root, dir, NULL, &gen, NULL, NULL, + NULL); + if (ret < 0) + goto out; + + /* + * The parent may be non-existent at this point in time. This happens + * if the ino of the parent dir is higher then the current ino. In this + * case, we can not process this ref until the parent dir is finally + * created. If we reach the parent dir later, process_recorded_refs + * will go through all dir items and process the refs that could not be + * processed before. In case this is the first ref, we set + * cur_inode_first_ref_orphan to 1 to inform process_recorded_refs to + * keep an orphan of the inode so that it later can be used for + * link/move + */ + ret = is_inode_existent(sctx, dir, gen); + if (ret < 0) + goto out; + if (!ret) { + ret = is_first_ref(sctx, sctx->send_root, sctx->cur_ino, dir, + name->start, fs_path_len(name)); + if (ret < 0) + goto out; + if (ret) + sctx->cur_inode_first_ref_orphan = 1; + ret = 0; + goto out; + } + + ret = get_cur_path(sctx, dir, gen, p); + if (ret < 0) + goto out; + ret = fs_path_add_path(p, name); + if (ret < 0) + goto out; + + ret = record_ref(&sctx->new_refs, dir, gen, p); + +out: + if (ret) + fs_path_free(sctx, p); + return ret; +} + +static int __record_deleted_ref(int num, u64 dir, int index, + struct fs_path *name, + void *ctx) +{ + int ret = 0; + struct send_ctx *sctx = ctx; + struct fs_path *p; + u64 gen; + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ret = get_inode_info(sctx->parent_root, dir, NULL, &gen, NULL, NULL, + NULL); + if (ret < 0) + goto out; + + ret = get_cur_path(sctx, dir, gen, p); + if (ret < 0) + goto out; + ret = fs_path_add_path(p, name); + if (ret < 0) + goto out; + + ret = record_ref(&sctx->deleted_refs, dir, gen, p); + +out: + if (ret) + fs_path_free(sctx, p); + return ret; +} + +static int record_new_ref(struct send_ctx *sctx) +{ + int ret; + + ret = iterate_inode_ref(sctx, sctx->send_root, sctx->left_path, + sctx->cmp_key, 0, __record_new_ref, sctx); + if (ret < 0) + goto out; + ret = 0; + +out: + return ret; +} + +static int record_deleted_ref(struct send_ctx *sctx) +{ + int ret; + + ret = iterate_inode_ref(sctx, sctx->parent_root, sctx->right_path, + sctx->cmp_key, 0, __record_deleted_ref, sctx); + if (ret < 0) + goto out; + ret = 0; + +out: + return ret; +} + +struct find_ref_ctx { + u64 dir; + struct fs_path *name; + int found_idx; +}; + +static int __find_iref(int num, u64 dir, int index, + struct fs_path *name, + void *ctx_) +{ + struct find_ref_ctx *ctx = ctx_; + + if (dir == ctx->dir && fs_path_len(name) == fs_path_len(ctx->name) && + strncmp(name->start, ctx->name->start, fs_path_len(name)) == 0) { + ctx->found_idx = num; + return 1; + } + return 0; +} + +static int find_iref(struct send_ctx *sctx, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_key *key, + u64 dir, struct fs_path *name) +{ + int ret; + struct find_ref_ctx ctx; + + ctx.dir = dir; + ctx.name = name; + ctx.found_idx = -1; + + ret = iterate_inode_ref(sctx, root, path, key, 0, __find_iref, &ctx); + if (ret < 0) + return ret; + + if (ctx.found_idx == -1) + return -ENOENT; + + return ctx.found_idx; +} + +static int __record_changed_new_ref(int num, u64 dir, int index, + struct fs_path *name, + void *ctx) +{ + int ret; + struct send_ctx *sctx = ctx; + + ret = find_iref(sctx, sctx->parent_root, sctx->right_path, + sctx->cmp_key, dir, name); + if (ret == -ENOENT) + ret = __record_new_ref(num, dir, index, name, sctx); + else if (ret > 0) + ret = 0; + + return ret; +} + +static int __record_changed_deleted_ref(int num, u64 dir, int index, + struct fs_path *name, + void *ctx) +{ + int ret; + struct send_ctx *sctx = ctx; + + ret = find_iref(sctx, sctx->send_root, sctx->left_path, sctx->cmp_key, + dir, name); + if (ret == -ENOENT) + ret = __record_deleted_ref(num, dir, index, name, sctx); + else if (ret > 0) + ret = 0; + + return ret; +} + +static int record_changed_ref(struct send_ctx *sctx) +{ + int ret = 0; + + ret = iterate_inode_ref(sctx, sctx->send_root, sctx->left_path, + sctx->cmp_key, 0, __record_changed_new_ref, sctx); + if (ret < 0) + goto out; + ret = iterate_inode_ref(sctx, sctx->parent_root, sctx->right_path, + sctx->cmp_key, 0, __record_changed_deleted_ref, sctx); + if (ret < 0) + goto out; + ret = 0; + +out: + return ret; +} + +/* + * Record and process all refs at once. Needed when an inode changes the + * generation number, which means that it was deleted and recreated. + */ +static int process_all_refs(struct send_ctx *sctx, + enum btrfs_compare_tree_result cmd) +{ + int ret; + struct btrfs_root *root; + struct btrfs_path *path; + struct btrfs_key key; + struct btrfs_key found_key; + struct extent_buffer *eb; + int slot; + iterate_inode_ref_t cb; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + if (cmd == BTRFS_COMPARE_TREE_NEW) { + root = sctx->send_root; + cb = __record_new_ref; + } else if (cmd == BTRFS_COMPARE_TREE_DELETED) { + root = sctx->parent_root; + cb = __record_deleted_ref; + } else { + BUG(); + } + + key.objectid = sctx->cmp_key->objectid; + key.type = BTRFS_INODE_REF_KEY; + key.offset = 0; + while (1) { + ret = btrfs_search_slot_for_read(root, &key, path, 1, 0); + if (ret < 0) { + btrfs_release_path(path); + goto out; + } + if (ret) { + btrfs_release_path(path); + break; + } + + eb = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + + if (found_key.objectid != key.objectid || + found_key.type != key.type) { + btrfs_release_path(path); + break; + } + + ret = iterate_inode_ref(sctx, sctx->parent_root, path, + &found_key, 0, cb, sctx); + btrfs_release_path(path); + if (ret < 0) + goto out; + + key.offset = found_key.offset + 1; + } + + ret = process_recorded_refs(sctx); + +out: + btrfs_free_path(path); + return ret; +} + +static int send_set_xattr(struct send_ctx *sctx, + struct fs_path *path, + const char *name, int name_len, + const char *data, int data_len) +{ + int ret = 0; + + ret = begin_cmd(sctx, BTRFS_SEND_C_SET_XATTR); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path); + TLV_PUT_STRING(sctx, BTRFS_SEND_A_XATTR_NAME, name, name_len); + TLV_PUT(sctx, BTRFS_SEND_A_XATTR_DATA, data, data_len); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + return ret; +} + +static int send_remove_xattr(struct send_ctx *sctx, + struct fs_path *path, + const char *name, int name_len) +{ + int ret = 0; + + ret = begin_cmd(sctx, BTRFS_SEND_C_REMOVE_XATTR); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, path); + TLV_PUT_STRING(sctx, BTRFS_SEND_A_XATTR_NAME, name, name_len); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + return ret; +} + +static int __process_new_xattr(int num, struct btrfs_key *di_key, + const char *name, int name_len, + const char *data, int data_len, + u8 type, void *ctx) +{ + int ret; + struct send_ctx *sctx = ctx; + struct fs_path *p; + posix_acl_xattr_header dummy_acl; + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + /* + * This hack is needed because empty acl's are stored as zero byte + * data in xattrs. Problem with that is, that receiving these zero byte + * acl's will fail later. To fix this, we send a dummy acl list that + * only contains the version number and no entries. + */ + if (!strncmp(name, XATTR_NAME_POSIX_ACL_ACCESS, name_len) || + !strncmp(name, XATTR_NAME_POSIX_ACL_DEFAULT, name_len)) { + if (data_len == 0) { + dummy_acl.a_version = + cpu_to_le32(POSIX_ACL_XATTR_VERSION); + data = (char *)&dummy_acl; + data_len = sizeof(dummy_acl); + } + } + + ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p); + if (ret < 0) + goto out; + + ret = send_set_xattr(sctx, p, name, name_len, data, data_len); + +out: + fs_path_free(sctx, p); + return ret; +} + +static int __process_deleted_xattr(int num, struct btrfs_key *di_key, + const char *name, int name_len, + const char *data, int data_len, + u8 type, void *ctx) +{ + int ret; + struct send_ctx *sctx = ctx; + struct fs_path *p; + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p); + if (ret < 0) + goto out; + + ret = send_remove_xattr(sctx, p, name, name_len); + +out: + fs_path_free(sctx, p); + return ret; +} + +static int process_new_xattr(struct send_ctx *sctx) +{ + int ret = 0; + + ret = iterate_dir_item(sctx, sctx->send_root, sctx->left_path, + sctx->cmp_key, __process_new_xattr, sctx); + + return ret; +} + +static int process_deleted_xattr(struct send_ctx *sctx) +{ + int ret; + + ret = iterate_dir_item(sctx, sctx->parent_root, sctx->right_path, + sctx->cmp_key, __process_deleted_xattr, sctx); + + return ret; +} + +struct find_xattr_ctx { + const char *name; + int name_len; + int found_idx; + char *found_data; + int found_data_len; +}; + +static int __find_xattr(int num, struct btrfs_key *di_key, + const char *name, int name_len, + const char *data, int data_len, + u8 type, void *vctx) +{ + struct find_xattr_ctx *ctx = vctx; + + if (name_len == ctx->name_len && + strncmp(name, ctx->name, name_len) == 0) { + ctx->found_idx = num; + ctx->found_data_len = data_len; + ctx->found_data = kmalloc(data_len, GFP_NOFS); + if (!ctx->found_data) + return -ENOMEM; + memcpy(ctx->found_data, data, data_len); + return 1; + } + return 0; +} + +static int find_xattr(struct send_ctx *sctx, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_key *key, + const char *name, int name_len, + char **data, int *data_len) +{ + int ret; + struct find_xattr_ctx ctx; + + ctx.name = name; + ctx.name_len = name_len; + ctx.found_idx = -1; + ctx.found_data = NULL; + ctx.found_data_len = 0; + + ret = iterate_dir_item(sctx, root, path, key, __find_xattr, &ctx); + if (ret < 0) + return ret; + + if (ctx.found_idx == -1) + return -ENOENT; + if (data) { + *data = ctx.found_data; + *data_len = ctx.found_data_len; + } else { + kfree(ctx.found_data); + } + return ctx.found_idx; +} + + +static int __process_changed_new_xattr(int num, struct btrfs_key *di_key, + const char *name, int name_len, + const char *data, int data_len, + u8 type, void *ctx) +{ + int ret; + struct send_ctx *sctx = ctx; + char *found_data = NULL; + int found_data_len = 0; + struct fs_path *p = NULL; + + ret = find_xattr(sctx, sctx->parent_root, sctx->right_path, + sctx->cmp_key, name, name_len, &found_data, + &found_data_len); + if (ret == -ENOENT) { + ret = __process_new_xattr(num, di_key, name, name_len, data, + data_len, type, ctx); + } else if (ret >= 0) { + if (data_len != found_data_len || + memcmp(data, found_data, data_len)) { + ret = __process_new_xattr(num, di_key, name, name_len, + data, data_len, type, ctx); + } else { + ret = 0; + } + } + + kfree(found_data); + fs_path_free(sctx, p); + return ret; +} + +static int __process_changed_deleted_xattr(int num, struct btrfs_key *di_key, + const char *name, int name_len, + const char *data, int data_len, + u8 type, void *ctx) +{ + int ret; + struct send_ctx *sctx = ctx; + + ret = find_xattr(sctx, sctx->send_root, sctx->left_path, sctx->cmp_key, + name, name_len, NULL, NULL); + if (ret == -ENOENT) + ret = __process_deleted_xattr(num, di_key, name, name_len, data, + data_len, type, ctx); + else if (ret >= 0) + ret = 0; + + return ret; +} + +static int process_changed_xattr(struct send_ctx *sctx) +{ + int ret = 0; + + ret = iterate_dir_item(sctx, sctx->send_root, sctx->left_path, + sctx->cmp_key, __process_changed_new_xattr, sctx); + if (ret < 0) + goto out; + ret = iterate_dir_item(sctx, sctx->parent_root, sctx->right_path, + sctx->cmp_key, __process_changed_deleted_xattr, sctx); + +out: + return ret; +} + +static int process_all_new_xattrs(struct send_ctx *sctx) +{ + int ret; + struct btrfs_root *root; + struct btrfs_path *path; + struct btrfs_key key; + struct btrfs_key found_key; + struct extent_buffer *eb; + int slot; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + root = sctx->send_root; + + key.objectid = sctx->cmp_key->objectid; + key.type = BTRFS_XATTR_ITEM_KEY; + key.offset = 0; + while (1) { + ret = btrfs_search_slot_for_read(root, &key, path, 1, 0); + if (ret < 0) + goto out; + if (ret) { + ret = 0; + goto out; + } + + eb = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + + if (found_key.objectid != key.objectid || + found_key.type != key.type) { + ret = 0; + goto out; + } + + ret = iterate_dir_item(sctx, root, path, &found_key, + __process_new_xattr, sctx); + if (ret < 0) + goto out; + + btrfs_release_path(path); + key.offset = found_key.offset + 1; + } + +out: + btrfs_free_path(path); + return ret; +} + +/* + * Read some bytes from the current inode/file and send a write command to + * user space. + */ +static int send_write(struct send_ctx *sctx, u64 offset, u32 len) +{ + int ret = 0; + struct fs_path *p; + loff_t pos = offset; + int readed = 0; + mm_segment_t old_fs; + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + /* + * vfs normally only accepts user space buffers for security reasons. + * we only read from the file and also only provide the read_buf buffer + * to vfs. As this buffer does not come from a user space call, it's + * ok to temporary allow kernel space buffers. + */ + old_fs = get_fs(); + set_fs(KERNEL_DS); + +verbose_printk("btrfs: send_write offset=%llu, len=%d\n", offset, len); + + ret = open_cur_inode_file(sctx); + if (ret < 0) + goto out; + + ret = vfs_read(sctx->cur_inode_filp, sctx->read_buf, len, &pos); + if (ret < 0) + goto out; + readed = ret; + if (!readed) + goto out; + + ret = begin_cmd(sctx, BTRFS_SEND_C_WRITE); + if (ret < 0) + goto out; + + ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p); + if (ret < 0) + goto out; + + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p); + TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset); + TLV_PUT(sctx, BTRFS_SEND_A_DATA, sctx->read_buf, readed); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + fs_path_free(sctx, p); + set_fs(old_fs); + if (ret < 0) + return ret; + return readed; +} + +/* + * Send a clone command to user space. + */ +static int send_clone(struct send_ctx *sctx, + u64 offset, u32 len, + struct clone_root *clone_root) +{ + int ret = 0; + struct btrfs_root *clone_root2 = clone_root->root; + struct fs_path *p; + u64 gen; + +verbose_printk("btrfs: send_clone offset=%llu, len=%d, clone_root=%llu, " + "clone_inode=%llu, clone_offset=%llu\n", offset, len, + clone_root->root->objectid, clone_root->ino, + clone_root->offset); + + p = fs_path_alloc(sctx); + if (!p) + return -ENOMEM; + + ret = begin_cmd(sctx, BTRFS_SEND_C_CLONE); + if (ret < 0) + goto out; + + ret = get_cur_path(sctx, sctx->cur_ino, sctx->cur_inode_gen, p); + if (ret < 0) + goto out; + + TLV_PUT_U64(sctx, BTRFS_SEND_A_FILE_OFFSET, offset); + TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_LEN, len); + TLV_PUT_PATH(sctx, BTRFS_SEND_A_PATH, p); + + if (clone_root2 == sctx->send_root) { + ret = get_inode_info(sctx->send_root, clone_root->ino, NULL, + &gen, NULL, NULL, NULL); + if (ret < 0) + goto out; + ret = get_cur_path(sctx, clone_root->ino, gen, p); + } else { + ret = get_inode_path(sctx, clone_root2, clone_root->ino, p); + } + if (ret < 0) + goto out; + + TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID, + clone_root2->root_item.uuid); + TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_CTRANSID, + clone_root2->root_item.ctransid); + TLV_PUT_PATH(sctx, BTRFS_SEND_A_CLONE_PATH, p); + TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_OFFSET, + clone_root->offset); + + ret = send_cmd(sctx); + +tlv_put_failure: +out: + fs_path_free(sctx, p); + return ret; +} + +static int send_write_or_clone(struct send_ctx *sctx, + struct btrfs_path *path, + struct btrfs_key *key, + struct clone_root *clone_root) +{ + int ret = 0; + struct btrfs_file_extent_item *ei; + u64 offset = key->offset; + u64 pos = 0; + u64 len; + u32 l; + u8 type; + + ei = btrfs_item_ptr(path->nodes[0], path->slots[0], + struct btrfs_file_extent_item); + type = btrfs_file_extent_type(path->nodes[0], ei); + if (type == BTRFS_FILE_EXTENT_INLINE) + len = btrfs_file_extent_inline_len(path->nodes[0], ei); + else + len = btrfs_file_extent_num_bytes(path->nodes[0], ei); + + if (offset + len > sctx->cur_inode_size) + len = sctx->cur_inode_size - offset; + if (len == 0) { + ret = 0; + goto out; + } + + if (!clone_root) { + while (pos < len) { + l = len - pos; + if (l > BTRFS_SEND_READ_SIZE) + l = BTRFS_SEND_READ_SIZE; + ret = send_write(sctx, pos + offset, l); + if (ret < 0) + goto out; + if (!ret) + break; + pos += ret; + } + ret = 0; + } else { + ret = send_clone(sctx, offset, len, clone_root); + } + +out: + return ret; +} + +static int is_extent_unchanged(struct send_ctx *sctx, + struct btrfs_path *left_path, + struct btrfs_key *ekey) +{ + int ret = 0; + struct btrfs_key key; + struct btrfs_path *path = NULL; + struct extent_buffer *eb; + int slot; + struct btrfs_key found_key; + struct btrfs_file_extent_item *ei; + u64 left_disknr; + u64 right_disknr; + u64 left_offset; + u64 right_offset; + u64 left_offset_fixed; + u64 left_len; + u64 right_len; + u8 left_type; + u8 right_type; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + eb = left_path->nodes[0]; + slot = left_path->slots[0]; + + ei = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); + left_type = btrfs_file_extent_type(eb, ei); + left_disknr = btrfs_file_extent_disk_bytenr(eb, ei); + left_len = btrfs_file_extent_num_bytes(eb, ei); + left_offset = btrfs_file_extent_offset(eb, ei); + + if (left_type != BTRFS_FILE_EXTENT_REG) { + ret = 0; + goto out; + } + + /* + * Following comments will refer to these graphics. L is the left + * extents which we are checking at the moment. 1-8 are the right + * extents that we iterate. + * + * |-----L-----| + * |-1-|-2a-|-3-|-4-|-5-|-6-| + * + * |-----L-----| + * |--1--|-2b-|...(same as above) + * + * Alternative situation. Happens on files where extents got split. + * |-----L-----| + * |-----------7-----------|-6-| + * + * Alternative situation. Happens on files which got larger. + * |-----L-----| + * |-8-| + * Nothing follows after 8. + */ + + key.objectid = ekey->objectid; + key.type = BTRFS_EXTENT_DATA_KEY; + key.offset = ekey->offset; + ret = btrfs_search_slot_for_read(sctx->parent_root, &key, path, 0, 0); + if (ret < 0) + goto out; + if (ret) { + ret = 0; + goto out; + } + + /* + * Handle special case where the right side has no extents at all. + */ + eb = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + if (found_key.objectid != key.objectid || + found_key.type != key.type) { + ret = 0; + goto out; + } + + /* + * We're now on 2a, 2b or 7. + */ + key = found_key; + while (key.offset < ekey->offset + left_len) { + ei = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); + right_type = btrfs_file_extent_type(eb, ei); + right_disknr = btrfs_file_extent_disk_bytenr(eb, ei); + right_len = btrfs_file_extent_num_bytes(eb, ei); + right_offset = btrfs_file_extent_offset(eb, ei); + + if (right_type != BTRFS_FILE_EXTENT_REG) { + ret = 0; + goto out; + } + + /* + * Are we at extent 8? If yes, we know the extent is changed. + * This may only happen on the first iteration. + */ + if (found_key.offset + right_len < ekey->offset) { + ret = 0; + goto out; + } + + left_offset_fixed = left_offset; + if (key.offset < ekey->offset) { + /* Fix the right offset for 2a and 7. */ + right_offset += ekey->offset - key.offset; + } else { + /* Fix the left offset for all behind 2a and 2b */ + left_offset_fixed += key.offset - ekey->offset; + } + + /* + * Check if we have the same extent. + */ + if (left_disknr + left_offset_fixed != + right_disknr + right_offset) { + ret = 0; + goto out; + } + + /* + * Go to the next extent. + */ + ret = btrfs_next_item(sctx->parent_root, path); + if (ret < 0) + goto out; + if (!ret) { + eb = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + } + if (ret || found_key.objectid != key.objectid || + found_key.type != key.type) { + key.offset += right_len; + break; + } else { + if (found_key.offset != key.offset + right_len) { + /* Should really not happen */ + ret = -EIO; + goto out; + } + } + key = found_key; + } + + /* + * We're now behind the left extent (treat as unchanged) or at the end + * of the right side (treat as changed). + */ + if (key.offset >= ekey->offset + left_len) + ret = 1; + else + ret = 0; + + +out: + btrfs_free_path(path); + return ret; +} + +static int process_extent(struct send_ctx *sctx, + struct btrfs_path *path, + struct btrfs_key *key) +{ + int ret = 0; + struct clone_root *found_clone = NULL; + + if (S_ISLNK(sctx->cur_inode_mode)) + return 0; + + if (sctx->parent_root && !sctx->cur_inode_new) { + ret = is_extent_unchanged(sctx, path, key); + if (ret < 0) + goto out; + if (ret) { + ret = 0; + goto out; + } + } + + ret = find_extent_clone(sctx, path, key->objectid, key->offset, + sctx->cur_inode_size, &found_clone); + if (ret != -ENOENT && ret < 0) + goto out; + + ret = send_write_or_clone(sctx, path, key, found_clone); + +out: + return ret; +} + +static int process_all_extents(struct send_ctx *sctx) +{ + int ret; + struct btrfs_root *root; + struct btrfs_path *path; + struct btrfs_key key; + struct btrfs_key found_key; + struct extent_buffer *eb; + int slot; + + root = sctx->send_root; + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + key.objectid = sctx->cmp_key->objectid; + key.type = BTRFS_EXTENT_DATA_KEY; + key.offset = 0; + while (1) { + ret = btrfs_search_slot_for_read(root, &key, path, 1, 0); + if (ret < 0) + goto out; + if (ret) { + ret = 0; + goto out; + } + + eb = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + + if (found_key.objectid != key.objectid || + found_key.type != key.type) { + ret = 0; + goto out; + } + + ret = process_extent(sctx, path, &found_key); + if (ret < 0) + goto out; + + btrfs_release_path(path); + key.offset = found_key.offset + 1; + } + +out: + btrfs_free_path(path); + return ret; +} + +static int process_recorded_refs_if_needed(struct send_ctx *sctx, int at_end) +{ + int ret = 0; + + if (sctx->cur_ino == 0) + goto out; + if (!at_end && sctx->cur_ino == sctx->cmp_key->objectid && + sctx->cmp_key->type <= BTRFS_INODE_REF_KEY) + goto out; + if (list_empty(&sctx->new_refs) && list_empty(&sctx->deleted_refs)) + goto out; + + ret = process_recorded_refs(sctx); + +out: + return ret; +} + +static int finish_inode_if_needed(struct send_ctx *sctx, int at_end) +{ + int ret = 0; + u64 left_mode; + u64 left_uid; + u64 left_gid; + u64 right_mode; + u64 right_uid; + u64 right_gid; + int need_chmod = 0; + int need_chown = 0; + + ret = process_recorded_refs_if_needed(sctx, at_end); + if (ret < 0) + goto out; + + if (sctx->cur_ino == 0 || sctx->cur_inode_deleted) + goto out; + if (!at_end && sctx->cmp_key->objectid == sctx->cur_ino) + goto out; + + ret = get_inode_info(sctx->send_root, sctx->cur_ino, NULL, NULL, + &left_mode, &left_uid, &left_gid); + if (ret < 0) + goto out; + + if (!S_ISLNK(sctx->cur_inode_mode)) { + if (!sctx->parent_root || sctx->cur_inode_new) { + need_chmod = 1; + need_chown = 1; + } else { + ret = get_inode_info(sctx->parent_root, sctx->cur_ino, + NULL, NULL, &right_mode, &right_uid, + &right_gid); + if (ret < 0) + goto out; + + if (left_uid != right_uid || left_gid != right_gid) + need_chown = 1; + if (left_mode != right_mode) + need_chmod = 1; + } + } + + if (S_ISREG(sctx->cur_inode_mode)) { + ret = send_truncate(sctx, sctx->cur_ino, sctx->cur_inode_gen, + sctx->cur_inode_size); + if (ret < 0) + goto out; + } + + if (need_chown) { + ret = send_chown(sctx, sctx->cur_ino, sctx->cur_inode_gen, + left_uid, left_gid); + if (ret < 0) + goto out; + } + if (need_chmod) { + ret = send_chmod(sctx, sctx->cur_ino, sctx->cur_inode_gen, + left_mode); + if (ret < 0) + goto out; + } + + /* + * Need to send that every time, no matter if it actually changed + * between the two trees as we have done changes to the inode before. + */ + ret = send_utimes(sctx, sctx->cur_ino, sctx->cur_inode_gen); + if (ret < 0) + goto out; + +out: + return ret; +} + +static int changed_inode(struct send_ctx *sctx, + enum btrfs_compare_tree_result result) +{ + int ret = 0; + struct btrfs_key *key = sctx->cmp_key; + struct btrfs_inode_item *left_ii = NULL; + struct btrfs_inode_item *right_ii = NULL; + u64 left_gen = 0; + u64 right_gen = 0; + + ret = close_cur_inode_file(sctx); + if (ret < 0) + goto out; + + sctx->cur_ino = key->objectid; + sctx->cur_inode_new_gen = 0; + sctx->cur_inode_first_ref_orphan = 0; + sctx->send_progress = sctx->cur_ino; + + if (result == BTRFS_COMPARE_TREE_NEW || + result == BTRFS_COMPARE_TREE_CHANGED) { + left_ii = btrfs_item_ptr(sctx->left_path->nodes[0], + sctx->left_path->slots[0], + struct btrfs_inode_item); + left_gen = btrfs_inode_generation(sctx->left_path->nodes[0], + left_ii); + } else { + right_ii = btrfs_item_ptr(sctx->right_path->nodes[0], + sctx->right_path->slots[0], + struct btrfs_inode_item); + right_gen = btrfs_inode_generation(sctx->right_path->nodes[0], + right_ii); + } + if (result == BTRFS_COMPARE_TREE_CHANGED) { + right_ii = btrfs_item_ptr(sctx->right_path->nodes[0], + sctx->right_path->slots[0], + struct btrfs_inode_item); + + right_gen = btrfs_inode_generation(sctx->right_path->nodes[0], + right_ii); + if (left_gen != right_gen) + sctx->cur_inode_new_gen = 1; + } + + if (result == BTRFS_COMPARE_TREE_NEW) { + sctx->cur_inode_gen = left_gen; + sctx->cur_inode_new = 1; + sctx->cur_inode_deleted = 0; + sctx->cur_inode_size = btrfs_inode_size( + sctx->left_path->nodes[0], left_ii); + sctx->cur_inode_mode = btrfs_inode_mode( + sctx->left_path->nodes[0], left_ii); + if (sctx->cur_ino != BTRFS_FIRST_FREE_OBJECTID) + ret = send_create_inode(sctx, sctx->left_path, + sctx->cmp_key); + } else if (result == BTRFS_COMPARE_TREE_DELETED) { + sctx->cur_inode_gen = right_gen; + sctx->cur_inode_new = 0; + sctx->cur_inode_deleted = 1; + sctx->cur_inode_size = btrfs_inode_size( + sctx->right_path->nodes[0], right_ii); + sctx->cur_inode_mode = btrfs_inode_mode( + sctx->right_path->nodes[0], right_ii); + } else if (result == BTRFS_COMPARE_TREE_CHANGED) { + if (sctx->cur_inode_new_gen) { + sctx->cur_inode_gen = right_gen; + sctx->cur_inode_new = 0; + sctx->cur_inode_deleted = 1; + sctx->cur_inode_size = btrfs_inode_size( + sctx->right_path->nodes[0], right_ii); + sctx->cur_inode_mode = btrfs_inode_mode( + sctx->right_path->nodes[0], right_ii); + ret = process_all_refs(sctx, + BTRFS_COMPARE_TREE_DELETED); + if (ret < 0) + goto out; + + sctx->cur_inode_gen = left_gen; + sctx->cur_inode_new = 1; + sctx->cur_inode_deleted = 0; + sctx->cur_inode_size = btrfs_inode_size( + sctx->left_path->nodes[0], left_ii); + sctx->cur_inode_mode = btrfs_inode_mode( + sctx->left_path->nodes[0], left_ii); + ret = send_create_inode(sctx, sctx->left_path, + sctx->cmp_key); + if (ret < 0) + goto out; + + ret = process_all_refs(sctx, BTRFS_COMPARE_TREE_NEW); + if (ret < 0) + goto out; + ret = process_all_extents(sctx); + if (ret < 0) + goto out; + ret = process_all_new_xattrs(sctx); + if (ret < 0) + goto out; + } else { + sctx->cur_inode_gen = left_gen; + sctx->cur_inode_new = 0; + sctx->cur_inode_new_gen = 0; + sctx->cur_inode_deleted = 0; + sctx->cur_inode_size = btrfs_inode_size( + sctx->left_path->nodes[0], left_ii); + sctx->cur_inode_mode = btrfs_inode_mode( + sctx->left_path->nodes[0], left_ii); + } + } + +out: + return ret; +} + +static int changed_ref(struct send_ctx *sctx, + enum btrfs_compare_tree_result result) +{ + int ret = 0; + + BUG_ON(sctx->cur_ino != sctx->cmp_key->objectid); + + if (!sctx->cur_inode_new_gen && + sctx->cur_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (result == BTRFS_COMPARE_TREE_NEW) + ret = record_new_ref(sctx); + else if (result == BTRFS_COMPARE_TREE_DELETED) + ret = record_deleted_ref(sctx); + else if (result == BTRFS_COMPARE_TREE_CHANGED) + ret = record_changed_ref(sctx); + } + + return ret; +} + +static int changed_xattr(struct send_ctx *sctx, + enum btrfs_compare_tree_result result) +{ + int ret = 0; + + BUG_ON(sctx->cur_ino != sctx->cmp_key->objectid); + + if (!sctx->cur_inode_new_gen && !sctx->cur_inode_deleted) { + if (result == BTRFS_COMPARE_TREE_NEW) + ret = process_new_xattr(sctx); + else if (result == BTRFS_COMPARE_TREE_DELETED) + ret = process_deleted_xattr(sctx); + else if (result == BTRFS_COMPARE_TREE_CHANGED) + ret = process_changed_xattr(sctx); + } + + return ret; +} + +static int changed_extent(struct send_ctx *sctx, + enum btrfs_compare_tree_result result) +{ + int ret = 0; + + BUG_ON(sctx->cur_ino != sctx->cmp_key->objectid); + + if (!sctx->cur_inode_new_gen && !sctx->cur_inode_deleted) { + if (result != BTRFS_COMPARE_TREE_DELETED) + ret = process_extent(sctx, sctx->left_path, + sctx->cmp_key); + } + + return ret; +} + + +static int changed_cb(struct btrfs_root *left_root, + struct btrfs_root *right_root, + struct btrfs_path *left_path, + struct btrfs_path *right_path, + struct btrfs_key *key, + enum btrfs_compare_tree_result result, + void *ctx) +{ + int ret = 0; + struct send_ctx *sctx = ctx; + + sctx->left_path = left_path; + sctx->right_path = right_path; + sctx->cmp_key = key; + + ret = finish_inode_if_needed(sctx, 0); + if (ret < 0) + goto out; + + if (key->type == BTRFS_INODE_ITEM_KEY) + ret = changed_inode(sctx, result); + else if (key->type == BTRFS_INODE_REF_KEY) + ret = changed_ref(sctx, result); + else if (key->type == BTRFS_XATTR_ITEM_KEY) + ret = changed_xattr(sctx, result); + else if (key->type == BTRFS_EXTENT_DATA_KEY) + ret = changed_extent(sctx, result); + +out: + return ret; +} + +static int full_send_tree(struct send_ctx *sctx) +{ + int ret; + struct btrfs_trans_handle *trans = NULL; + struct btrfs_root *send_root = sctx->send_root; + struct btrfs_key key; + struct btrfs_key found_key; + struct btrfs_path *path; + struct extent_buffer *eb; + int slot; + u64 start_ctransid; + u64 ctransid; + + path = alloc_path_for_send(); + if (!path) + return -ENOMEM; + + spin_lock(&send_root->root_times_lock); + start_ctransid = btrfs_root_ctransid(&send_root->root_item); + spin_unlock(&send_root->root_times_lock); + + key.objectid = BTRFS_FIRST_FREE_OBJECTID; + key.type = BTRFS_INODE_ITEM_KEY; + key.offset = 0; + +join_trans: + /* + * We need to make sure the transaction does not get committed + * while we do anything on commit roots. Join a transaction to prevent + * this. + */ + trans = btrfs_join_transaction(send_root); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + trans = NULL; + goto out; + } + + /* + * Make sure the tree has not changed + */ + spin_lock(&send_root->root_times_lock); + ctransid = btrfs_root_ctransid(&send_root->root_item); + spin_unlock(&send_root->root_times_lock); + + if (ctransid != start_ctransid) { + WARN(1, KERN_WARNING "btrfs: the root that you're trying to " + "send was modified in between. This is " + "probably a bug.\n"); + ret = -EIO; + goto out; + } + + ret = btrfs_search_slot_for_read(send_root, &key, path, 1, 0); + if (ret < 0) + goto out; + if (ret) + goto out_finish; + + while (1) { + /* + * When someone want to commit while we iterate, end the + * joined transaction and rejoin. + */ + if (btrfs_should_end_transaction(trans, send_root)) { + ret = btrfs_end_transaction(trans, send_root); + trans = NULL; + if (ret < 0) + goto out; + btrfs_release_path(path); + goto join_trans; + } + + eb = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + + ret = changed_cb(send_root, NULL, path, NULL, + &found_key, BTRFS_COMPARE_TREE_NEW, sctx); + if (ret < 0) + goto out; + + key.objectid = found_key.objectid; + key.type = found_key.type; + key.offset = found_key.offset + 1; + + ret = btrfs_next_item(send_root, path); + if (ret < 0) + goto out; + if (ret) { + ret = 0; + break; + } + } + +out_finish: + ret = finish_inode_if_needed(sctx, 1); + +out: + btrfs_free_path(path); + if (trans) { + if (!ret) + ret = btrfs_end_transaction(trans, send_root); + else + btrfs_end_transaction(trans, send_root); + } + return ret; +} + +static int send_subvol(struct send_ctx *sctx) +{ + int ret; + + ret = send_header(sctx); + if (ret < 0) + goto out; + + ret = send_subvol_begin(sctx); + if (ret < 0) + goto out; + + if (sctx->parent_root) { + ret = btrfs_compare_trees(sctx->send_root, sctx->parent_root, + changed_cb, sctx); + if (ret < 0) + goto out; + ret = finish_inode_if_needed(sctx, 1); + if (ret < 0) + goto out; + } else { + ret = full_send_tree(sctx); + if (ret < 0) + goto out; + } + +out: + if (!ret) + ret = close_cur_inode_file(sctx); + else + close_cur_inode_file(sctx); + + free_recorded_refs(sctx); + return ret; +} + +long btrfs_ioctl_send(struct file *mnt_file, void __user *arg_) +{ + int ret = 0; + struct btrfs_root *send_root; + struct btrfs_root *clone_root; + struct btrfs_fs_info *fs_info; + struct btrfs_ioctl_send_args *arg = NULL; + struct btrfs_key key; + struct file *filp = NULL; + struct send_ctx *sctx = NULL; + u32 i; + u64 *clone_sources_tmp = NULL; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + send_root = BTRFS_I(fdentry(mnt_file)->d_inode)->root; + fs_info = send_root->fs_info; + + arg = memdup_user(arg_, sizeof(*arg)); + if (IS_ERR(arg)) { + ret = PTR_ERR(arg); + arg = NULL; + goto out; + } + + if (!access_ok(VERIFY_READ, arg->clone_sources, + sizeof(*arg->clone_sources * + arg->clone_sources_count))) { + ret = -EFAULT; + goto out; + } + + sctx = kzalloc(sizeof(struct send_ctx), GFP_NOFS); + if (!sctx) { + ret = -ENOMEM; + goto out; + } + + INIT_LIST_HEAD(&sctx->new_refs); + INIT_LIST_HEAD(&sctx->deleted_refs); + INIT_RADIX_TREE(&sctx->name_cache, GFP_NOFS); + INIT_LIST_HEAD(&sctx->name_cache_list); + + sctx->send_filp = fget(arg->send_fd); + if (IS_ERR(sctx->send_filp)) { + ret = PTR_ERR(sctx->send_filp); + goto out; + } + + sctx->mnt = mnt_file->f_path.mnt; + + sctx->send_root = send_root; + sctx->clone_roots_cnt = arg->clone_sources_count; + + sctx->send_max_size = BTRFS_SEND_BUF_SIZE; + sctx->send_buf = vmalloc(sctx->send_max_size); + if (!sctx->send_buf) { + ret = -ENOMEM; + goto out; + } + + sctx->read_buf = vmalloc(BTRFS_SEND_READ_SIZE); + if (!sctx->read_buf) { + ret = -ENOMEM; + goto out; + } + + sctx->clone_roots = vzalloc(sizeof(struct clone_root) * + (arg->clone_sources_count + 1)); + if (!sctx->clone_roots) { + ret = -ENOMEM; + goto out; + } + + if (arg->clone_sources_count) { + clone_sources_tmp = vmalloc(arg->clone_sources_count * + sizeof(*arg->clone_sources)); + if (!clone_sources_tmp) { + ret = -ENOMEM; + goto out; + } + + ret = copy_from_user(clone_sources_tmp, arg->clone_sources, + arg->clone_sources_count * + sizeof(*arg->clone_sources)); + if (ret) { + ret = -EFAULT; + goto out; + } + + for (i = 0; i < arg->clone_sources_count; i++) { + key.objectid = clone_sources_tmp[i]; + key.type = BTRFS_ROOT_ITEM_KEY; + key.offset = (u64)-1; + clone_root = btrfs_read_fs_root_no_name(fs_info, &key); + if (!clone_root) { + ret = -EINVAL; + goto out; + } + if (IS_ERR(clone_root)) { + ret = PTR_ERR(clone_root); + goto out; + } + sctx->clone_roots[i].root = clone_root; + } + vfree(clone_sources_tmp); + clone_sources_tmp = NULL; + } + + if (arg->parent_root) { + key.objectid = arg->parent_root; + key.type = BTRFS_ROOT_ITEM_KEY; + key.offset = (u64)-1; + sctx->parent_root = btrfs_read_fs_root_no_name(fs_info, &key); + if (!sctx->parent_root) { + ret = -EINVAL; + goto out; + } + } + + /* + * Clones from send_root are allowed, but only if the clone source + * is behind the current send position. This is checked while searching + * for possible clone sources. + */ + sctx->clone_roots[sctx->clone_roots_cnt++].root = sctx->send_root; + + /* We do a bsearch later */ + sort(sctx->clone_roots, sctx->clone_roots_cnt, + sizeof(*sctx->clone_roots), __clone_root_cmp_sort, + NULL); + + ret = send_subvol(sctx); + if (ret < 0) + goto out; + + ret = begin_cmd(sctx, BTRFS_SEND_C_END); + if (ret < 0) + goto out; + ret = send_cmd(sctx); + if (ret < 0) + goto out; + +out: + if (filp) + fput(filp); + kfree(arg); + vfree(clone_sources_tmp); + + if (sctx) { + if (sctx->send_filp) + fput(sctx->send_filp); + + vfree(sctx->clone_roots); + vfree(sctx->send_buf); + vfree(sctx->read_buf); + + name_cache_free(sctx); + + kfree(sctx); + } + + return ret; +} diff --git a/fs/btrfs/send.h b/fs/btrfs/send.h new file mode 100644 index 000000000000..9934e948e57f --- /dev/null +++ b/fs/btrfs/send.h @@ -0,0 +1,133 @@ +/* + * Copyright (C) 2012 Alexander Block. All rights reserved. + * Copyright (C) 2012 STRATO. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include "ctree.h" + +#define BTRFS_SEND_STREAM_MAGIC "btrfs-stream" +#define BTRFS_SEND_STREAM_VERSION 1 + +#define BTRFS_SEND_BUF_SIZE (1024 * 64) +#define BTRFS_SEND_READ_SIZE (1024 * 48) + +enum btrfs_tlv_type { + BTRFS_TLV_U8, + BTRFS_TLV_U16, + BTRFS_TLV_U32, + BTRFS_TLV_U64, + BTRFS_TLV_BINARY, + BTRFS_TLV_STRING, + BTRFS_TLV_UUID, + BTRFS_TLV_TIMESPEC, +}; + +struct btrfs_stream_header { + char magic[sizeof(BTRFS_SEND_STREAM_MAGIC)]; + __le32 version; +} __attribute__ ((__packed__)); + +struct btrfs_cmd_header { + /* len excluding the header */ + __le32 len; + __le16 cmd; + /* crc including the header with zero crc field */ + __le32 crc; +} __attribute__ ((__packed__)); + +struct btrfs_tlv_header { + __le16 tlv_type; + /* len excluding the header */ + __le16 tlv_len; +} __attribute__ ((__packed__)); + +/* commands */ +enum btrfs_send_cmd { + BTRFS_SEND_C_UNSPEC, + + BTRFS_SEND_C_SUBVOL, + BTRFS_SEND_C_SNAPSHOT, + + BTRFS_SEND_C_MKFILE, + BTRFS_SEND_C_MKDIR, + BTRFS_SEND_C_MKNOD, + BTRFS_SEND_C_MKFIFO, + BTRFS_SEND_C_MKSOCK, + BTRFS_SEND_C_SYMLINK, + + BTRFS_SEND_C_RENAME, + BTRFS_SEND_C_LINK, + BTRFS_SEND_C_UNLINK, + BTRFS_SEND_C_RMDIR, + + BTRFS_SEND_C_SET_XATTR, + BTRFS_SEND_C_REMOVE_XATTR, + + BTRFS_SEND_C_WRITE, + BTRFS_SEND_C_CLONE, + + BTRFS_SEND_C_TRUNCATE, + BTRFS_SEND_C_CHMOD, + BTRFS_SEND_C_CHOWN, + BTRFS_SEND_C_UTIMES, + + BTRFS_SEND_C_END, + __BTRFS_SEND_C_MAX, +}; +#define BTRFS_SEND_C_MAX (__BTRFS_SEND_C_MAX - 1) + +/* attributes in send stream */ +enum { + BTRFS_SEND_A_UNSPEC, + + BTRFS_SEND_A_UUID, + BTRFS_SEND_A_CTRANSID, + + BTRFS_SEND_A_INO, + BTRFS_SEND_A_SIZE, + BTRFS_SEND_A_MODE, + BTRFS_SEND_A_UID, + BTRFS_SEND_A_GID, + BTRFS_SEND_A_RDEV, + BTRFS_SEND_A_CTIME, + BTRFS_SEND_A_MTIME, + BTRFS_SEND_A_ATIME, + BTRFS_SEND_A_OTIME, + + BTRFS_SEND_A_XATTR_NAME, + BTRFS_SEND_A_XATTR_DATA, + + BTRFS_SEND_A_PATH, + BTRFS_SEND_A_PATH_TO, + BTRFS_SEND_A_PATH_LINK, + + BTRFS_SEND_A_FILE_OFFSET, + BTRFS_SEND_A_DATA, + + BTRFS_SEND_A_CLONE_UUID, + BTRFS_SEND_A_CLONE_CTRANSID, + BTRFS_SEND_A_CLONE_PATH, + BTRFS_SEND_A_CLONE_OFFSET, + BTRFS_SEND_A_CLONE_LEN, + + __BTRFS_SEND_A_MAX, +}; +#define BTRFS_SEND_A_MAX (__BTRFS_SEND_A_MAX - 1) + +#ifdef __KERNEL__ +long btrfs_ioctl_send(struct file *mnt_file, void __user *arg); +#endif diff --git a/fs/btrfs/struct-funcs.c b/fs/btrfs/struct-funcs.c index bc1f6ad18442..b976597b0721 100644 --- a/fs/btrfs/struct-funcs.c +++ b/fs/btrfs/struct-funcs.c @@ -17,15 +17,27 @@ */ #include <linux/highmem.h> +#include <asm/unaligned.h> -/* this is some deeply nasty code. ctree.h has a different - * definition for this BTRFS_SETGET_FUNCS macro, behind a #ifndef +#include "ctree.h" + +static inline u8 get_unaligned_le8(const void *p) +{ + return *(u8 *)p; +} + +static inline void put_unaligned_le8(u8 val, void *p) +{ + *(u8 *)p = val; +} + +/* + * this is some deeply nasty code. * * The end result is that anyone who #includes ctree.h gets a - * declaration for the btrfs_set_foo functions and btrfs_foo functions - * - * This file declares the macros and then #includes ctree.h, which results - * in cpp creating the function here based on the template below. + * declaration for the btrfs_set_foo functions and btrfs_foo functions, + * which are wappers of btrfs_set_token_#bits functions and + * btrfs_get_token_#bits functions, which are defined in this file. * * These setget functions do all the extent_buffer related mapping * required to efficiently read and write specific fields in the extent @@ -33,64 +45,93 @@ * an unsigned long offset into the extent buffer which has been * cast to a specific type. This gives us all the gcc type checking. * - * The extent buffer api is used to do all the kmapping and page - * spanning work required to get extent buffers in highmem and have - * a metadata blocksize different from the page size. - * - * The macro starts with a simple function prototype declaration so that - * sparse won't complain about it being static. + * The extent buffer api is used to do the page spanning work required to + * have a metadata blocksize different from the page size. */ -#define BTRFS_SETGET_FUNCS(name, type, member, bits) \ -u##bits btrfs_##name(struct extent_buffer *eb, type *s); \ -void btrfs_set_##name(struct extent_buffer *eb, type *s, u##bits val); \ -u##bits btrfs_##name(struct extent_buffer *eb, \ - type *s) \ +#define DEFINE_BTRFS_SETGET_BITS(bits) \ +u##bits btrfs_get_token_##bits(struct extent_buffer *eb, void *ptr, \ + unsigned long off, \ + struct btrfs_map_token *token) \ { \ - unsigned long part_offset = (unsigned long)s; \ - unsigned long offset = part_offset + offsetof(type, member); \ - type *p; \ - int err; \ - char *kaddr; \ - unsigned long map_start; \ - unsigned long map_len; \ - u##bits res; \ - err = map_private_extent_buffer(eb, offset, \ - sizeof(((type *)0)->member), \ - &kaddr, &map_start, &map_len); \ - if (err) { \ - __le##bits leres; \ - read_eb_member(eb, s, type, member, &leres); \ - return le##bits##_to_cpu(leres); \ - } \ - p = (type *)(kaddr + part_offset - map_start); \ - res = le##bits##_to_cpu(p->member); \ - return res; \ + unsigned long part_offset = (unsigned long)ptr; \ + unsigned long offset = part_offset + off; \ + void *p; \ + int err; \ + char *kaddr; \ + unsigned long map_start; \ + unsigned long map_len; \ + int size = sizeof(u##bits); \ + u##bits res; \ + \ + if (token && token->kaddr && token->offset <= offset && \ + token->eb == eb && \ + (token->offset + PAGE_CACHE_SIZE >= offset + size)) { \ + kaddr = token->kaddr; \ + p = kaddr + part_offset - token->offset; \ + res = get_unaligned_le##bits(p + off); \ + return res; \ + } \ + err = map_private_extent_buffer(eb, offset, size, \ + &kaddr, &map_start, &map_len); \ + if (err) { \ + __le##bits leres; \ + \ + read_extent_buffer(eb, &leres, offset, size); \ + return le##bits##_to_cpu(leres); \ + } \ + p = kaddr + part_offset - map_start; \ + res = get_unaligned_le##bits(p + off); \ + if (token) { \ + token->kaddr = kaddr; \ + token->offset = map_start; \ + token->eb = eb; \ + } \ + return res; \ } \ -void btrfs_set_##name(struct extent_buffer *eb, \ - type *s, u##bits val) \ +void btrfs_set_token_##bits(struct extent_buffer *eb, \ + void *ptr, unsigned long off, u##bits val, \ + struct btrfs_map_token *token) \ { \ - unsigned long part_offset = (unsigned long)s; \ - unsigned long offset = part_offset + offsetof(type, member); \ - type *p; \ - int err; \ - char *kaddr; \ - unsigned long map_start; \ - unsigned long map_len; \ - err = map_private_extent_buffer(eb, offset, \ - sizeof(((type *)0)->member), \ - &kaddr, &map_start, &map_len); \ - if (err) { \ - __le##bits val2; \ - val2 = cpu_to_le##bits(val); \ - write_eb_member(eb, s, type, member, &val2); \ - return; \ - } \ - p = (type *)(kaddr + part_offset - map_start); \ - p->member = cpu_to_le##bits(val); \ + unsigned long part_offset = (unsigned long)ptr; \ + unsigned long offset = part_offset + off; \ + void *p; \ + int err; \ + char *kaddr; \ + unsigned long map_start; \ + unsigned long map_len; \ + int size = sizeof(u##bits); \ + \ + if (token && token->kaddr && token->offset <= offset && \ + token->eb == eb && \ + (token->offset + PAGE_CACHE_SIZE >= offset + size)) { \ + kaddr = token->kaddr; \ + p = kaddr + part_offset - token->offset; \ + put_unaligned_le##bits(val, p + off); \ + return; \ + } \ + err = map_private_extent_buffer(eb, offset, size, \ + &kaddr, &map_start, &map_len); \ + if (err) { \ + __le##bits val2; \ + \ + val2 = cpu_to_le##bits(val); \ + write_extent_buffer(eb, &val2, offset, size); \ + return; \ + } \ + p = kaddr + part_offset - map_start; \ + put_unaligned_le##bits(val, p + off); \ + if (token) { \ + token->kaddr = kaddr; \ + token->offset = map_start; \ + token->eb = eb; \ + } \ } -#include "ctree.h" +DEFINE_BTRFS_SETGET_BITS(8) +DEFINE_BTRFS_SETGET_BITS(16) +DEFINE_BTRFS_SETGET_BITS(32) +DEFINE_BTRFS_SETGET_BITS(64) void btrfs_node_key(struct extent_buffer *eb, struct btrfs_disk_key *disk_key, int nr) diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 15634d4648d7..83d6f9f9c220 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -40,6 +40,7 @@ #include <linux/magic.h> #include <linux/slab.h> #include <linux/cleancache.h> +#include <linux/ratelimit.h> #include "compat.h" #include "delayed-inode.h" #include "ctree.h" @@ -53,11 +54,13 @@ #include "version.h" #include "export.h" #include "compression.h" +#include "rcu-string.h" #define CREATE_TRACE_POINTS #include <trace/events/btrfs.h> static const struct super_operations btrfs_super_ops; +static struct file_system_type btrfs_fs_type; static const char *btrfs_decode_error(struct btrfs_fs_info *fs_info, int errno, char nbuf[16]) @@ -74,6 +77,9 @@ static const char *btrfs_decode_error(struct btrfs_fs_info *fs_info, int errno, case -EROFS: errstr = "Readonly filesystem"; break; + case -EEXIST: + errstr = "Object already exists"; + break; default: if (nbuf) { if (snprintf(nbuf, 16, "error %d", -errno) >= 0) @@ -94,10 +100,6 @@ static void __save_error_info(struct btrfs_fs_info *fs_info) fs_info->fs_state = BTRFS_SUPER_FLAG_ERROR; } -/* NOTE: - * We move write_super stuff at umount in order to avoid deadlock - * for umount hold all lock. - */ static void save_error_info(struct btrfs_fs_info *fs_info) { __save_error_info(fs_info); @@ -114,44 +116,183 @@ static void btrfs_handle_error(struct btrfs_fs_info *fs_info) if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { sb->s_flags |= MS_RDONLY; printk(KERN_INFO "btrfs is forced readonly\n"); + __btrfs_scrub_cancel(fs_info); +// WARN_ON(1); } } +#ifdef CONFIG_PRINTK /* * __btrfs_std_error decodes expected errors from the caller and * invokes the approciate error response. */ void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function, - unsigned int line, int errno) + unsigned int line, int errno, const char *fmt, ...) { struct super_block *sb = fs_info->sb; char nbuf[16]; const char *errstr; + va_list args; + va_start(args, fmt); /* * Special case: if the error is EROFS, and we're already * under MS_RDONLY, then it is safe here. */ if (errno == -EROFS && (sb->s_flags & MS_RDONLY)) - return; + return; - errstr = btrfs_decode_error(fs_info, errno, nbuf); - printk(KERN_CRIT "BTRFS error (device %s) in %s:%d: %s\n", - sb->s_id, function, line, errstr); - save_error_info(fs_info); + errstr = btrfs_decode_error(fs_info, errno, nbuf); + if (fmt) { + struct va_format vaf = { + .fmt = fmt, + .va = &args, + }; - btrfs_handle_error(fs_info); + printk(KERN_CRIT "BTRFS error (device %s) in %s:%d: %s (%pV)\n", + sb->s_id, function, line, errstr, &vaf); + } else { + printk(KERN_CRIT "BTRFS error (device %s) in %s:%d: %s\n", + sb->s_id, function, line, errstr); + } + + /* Don't go through full error handling during mount */ + if (sb->s_flags & MS_BORN) { + save_error_info(fs_info); + btrfs_handle_error(fs_info); + } + va_end(args); } -static void btrfs_put_super(struct super_block *sb) +static const char * const logtypes[] = { + "emergency", + "alert", + "critical", + "error", + "warning", + "notice", + "info", + "debug", +}; + +void btrfs_printk(struct btrfs_fs_info *fs_info, const char *fmt, ...) { - struct btrfs_root *root = btrfs_sb(sb); - int ret; + struct super_block *sb = fs_info->sb; + char lvl[4]; + struct va_format vaf; + va_list args; + const char *type = logtypes[4]; + int kern_level; + + va_start(args, fmt); + + kern_level = printk_get_level(fmt); + if (kern_level) { + size_t size = printk_skip_level(fmt) - fmt; + memcpy(lvl, fmt, size); + lvl[size] = '\0'; + fmt += size; + type = logtypes[kern_level - '0']; + } else + *lvl = '\0'; + + vaf.fmt = fmt; + vaf.va = &args; + + printk("%sBTRFS %s (device %s): %pV", lvl, type, sb->s_id, &vaf); + + va_end(args); +} - ret = close_ctree(root); - sb->s_fs_info = NULL; +#else - (void)ret; /* FIXME: need to fix VFS to return error? */ +void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function, + unsigned int line, int errno, const char *fmt, ...) +{ + struct super_block *sb = fs_info->sb; + + /* + * Special case: if the error is EROFS, and we're already + * under MS_RDONLY, then it is safe here. + */ + if (errno == -EROFS && (sb->s_flags & MS_RDONLY)) + return; + + /* Don't go through full error handling during mount */ + if (sb->s_flags & MS_BORN) { + save_error_info(fs_info); + btrfs_handle_error(fs_info); + } +} +#endif + +/* + * We only mark the transaction aborted and then set the file system read-only. + * This will prevent new transactions from starting or trying to join this + * one. + * + * This means that error recovery at the call site is limited to freeing + * any local memory allocations and passing the error code up without + * further cleanup. The transaction should complete as it normally would + * in the call path but will return -EIO. + * + * We'll complete the cleanup in btrfs_end_transaction and + * btrfs_commit_transaction. + */ +void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, + struct btrfs_root *root, const char *function, + unsigned int line, int errno) +{ + WARN_ONCE(1, KERN_DEBUG "btrfs: Transaction aborted"); + trans->aborted = errno; + /* Nothing used. The other threads that have joined this + * transaction may be able to continue. */ + if (!trans->blocks_used) { + btrfs_printk(root->fs_info, "Aborting unused transaction.\n"); + return; + } + trans->transaction->aborted = errno; + __btrfs_std_error(root->fs_info, function, line, errno, NULL); +} +/* + * __btrfs_panic decodes unexpected, fatal errors from the caller, + * issues an alert, and either panics or BUGs, depending on mount options. + */ +void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function, + unsigned int line, int errno, const char *fmt, ...) +{ + char nbuf[16]; + char *s_id = "<unknown>"; + const char *errstr; + struct va_format vaf = { .fmt = fmt }; + va_list args; + + if (fs_info) + s_id = fs_info->sb->s_id; + + va_start(args, fmt); + vaf.va = &args; + + errstr = btrfs_decode_error(fs_info, errno, nbuf); + if (fs_info->mount_opt & BTRFS_MOUNT_PANIC_ON_FATAL_ERROR) + panic(KERN_CRIT "BTRFS panic (device %s) in %s:%d: %pV (%s)\n", + s_id, function, line, &vaf, errstr); + + printk(KERN_CRIT "BTRFS panic (device %s) in %s:%d: %pV (%s)\n", + s_id, function, line, &vaf, errstr); + va_end(args); + /* Caller calls BUG() */ +} + +static void btrfs_put_super(struct super_block *sb) +{ + (void)close_ctree(btrfs_sb(sb)->tree_root); + /* FIXME: need to fix VFS to return error? */ + /* AV: return it _where_? ->put_super() can be triggered by any number + * of async events, up to and including delivery of SIGKILL to the + * last process that kept it busy. Or segfault in the aforementioned + * process... Whom would you report that to? + */ } enum { @@ -161,8 +302,11 @@ enum { Opt_compress_type, Opt_compress_force, Opt_compress_force_type, Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard, Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed, - Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, - Opt_inode_cache, Opt_err, + Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, Opt_inode_cache, + Opt_no_space_cache, Opt_recovery, Opt_skip_balance, + Opt_check_integrity, Opt_check_integrity_including_extent_data, + Opt_check_integrity_print_mask, Opt_fatal_errors, + Opt_err, }; static match_table_t tokens = { @@ -195,25 +339,38 @@ static match_table_t tokens = { {Opt_subvolrootid, "subvolrootid=%d"}, {Opt_defrag, "autodefrag"}, {Opt_inode_cache, "inode_cache"}, + {Opt_no_space_cache, "nospace_cache"}, + {Opt_recovery, "recovery"}, + {Opt_skip_balance, "skip_balance"}, + {Opt_check_integrity, "check_int"}, + {Opt_check_integrity_including_extent_data, "check_int_data"}, + {Opt_check_integrity_print_mask, "check_int_print_mask=%d"}, + {Opt_fatal_errors, "fatal_errors=%s"}, {Opt_err, NULL}, }; /* * Regular mount options parser. Everything that is needed only when * reading in a new superblock is parsed here. + * XXX JDM: This needs to be cleaned up for remount. */ int btrfs_parse_options(struct btrfs_root *root, char *options) { struct btrfs_fs_info *info = root->fs_info; substring_t args[MAX_OPT_ARGS]; - char *p, *num, *orig; + char *p, *num, *orig = NULL; + u64 cache_gen; int intarg; int ret = 0; char *compress_type; bool compress_force = false; + cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy); + if (cache_gen) + btrfs_set_opt(info->mount_opt, SPACE_CACHE); + if (!options) - return 0; + goto out; /* * strsep changes the string, duplicate it because parse_options @@ -264,15 +421,23 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) strcmp(args[0].from, "zlib") == 0) { compress_type = "zlib"; info->compress_type = BTRFS_COMPRESS_ZLIB; + btrfs_set_opt(info->mount_opt, COMPRESS); } else if (strcmp(args[0].from, "lzo") == 0) { compress_type = "lzo"; info->compress_type = BTRFS_COMPRESS_LZO; + btrfs_set_opt(info->mount_opt, COMPRESS); + btrfs_set_fs_incompat(info, COMPRESS_LZO); + } else if (strncmp(args[0].from, "no", 2) == 0) { + compress_type = "no"; + info->compress_type = BTRFS_COMPRESS_NONE; + btrfs_clear_opt(info->mount_opt, COMPRESS); + btrfs_clear_opt(info->mount_opt, FORCE_COMPRESS); + compress_force = false; } else { ret = -EINVAL; goto out; } - btrfs_set_opt(info->mount_opt, COMPRESS); if (compress_force) { btrfs_set_opt(info->mount_opt, FORCE_COMPRESS); pr_info("btrfs: force %s compression\n", @@ -305,11 +470,8 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) case Opt_thread_pool: intarg = 0; match_int(&args[0], &intarg); - if (intarg) { + if (intarg) info->thread_pool_size = intarg; - printk(KERN_INFO "btrfs: thread pool %d\n", - info->thread_pool_size); - } break; case Opt_max_inline: num = match_strdup(&args[0]); @@ -360,9 +522,12 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) btrfs_set_opt(info->mount_opt, DISCARD); break; case Opt_space_cache: - printk(KERN_INFO "btrfs: enabling disk space caching\n"); btrfs_set_opt(info->mount_opt, SPACE_CACHE); break; + case Opt_no_space_cache: + printk(KERN_INFO "btrfs: disabling disk space caching\n"); + btrfs_clear_opt(info->mount_opt, SPACE_CACHE); + break; case Opt_inode_cache: printk(KERN_INFO "btrfs: enabling inode map caching\n"); btrfs_set_opt(info->mount_opt, INODE_MAP_CACHE); @@ -381,6 +546,56 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) printk(KERN_INFO "btrfs: enabling auto defrag"); btrfs_set_opt(info->mount_opt, AUTO_DEFRAG); break; + case Opt_recovery: + printk(KERN_INFO "btrfs: enabling auto recovery"); + btrfs_set_opt(info->mount_opt, RECOVERY); + break; + case Opt_skip_balance: + btrfs_set_opt(info->mount_opt, SKIP_BALANCE); + break; +#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY + case Opt_check_integrity_including_extent_data: + printk(KERN_INFO "btrfs: enabling check integrity" + " including extent data\n"); + btrfs_set_opt(info->mount_opt, + CHECK_INTEGRITY_INCLUDING_EXTENT_DATA); + btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY); + break; + case Opt_check_integrity: + printk(KERN_INFO "btrfs: enabling check integrity\n"); + btrfs_set_opt(info->mount_opt, CHECK_INTEGRITY); + break; + case Opt_check_integrity_print_mask: + intarg = 0; + match_int(&args[0], &intarg); + if (intarg) { + info->check_integrity_print_mask = intarg; + printk(KERN_INFO "btrfs:" + " check_integrity_print_mask 0x%x\n", + info->check_integrity_print_mask); + } + break; +#else + case Opt_check_integrity_including_extent_data: + case Opt_check_integrity: + case Opt_check_integrity_print_mask: + printk(KERN_ERR "btrfs: support for check_integrity*" + " not compiled in!\n"); + ret = -EINVAL; + goto out; +#endif + case Opt_fatal_errors: + if (strcmp(args[0].from, "panic") == 0) + btrfs_set_opt(info->mount_opt, + PANIC_ON_FATAL_ERROR); + else if (strcmp(args[0].from, "bug") == 0) + btrfs_clear_opt(info->mount_opt, + PANIC_ON_FATAL_ERROR); + else { + ret = -EINVAL; + goto out; + } + break; case Opt_err: printk(KERN_INFO "btrfs: unrecognized mount option " "'%s'\n", p); @@ -391,6 +606,8 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) } } out: + if (!ret && btrfs_test_opt(root, SPACE_CACHE)) + printk(KERN_INFO "btrfs: disk space caching is enabled\n"); kfree(orig); return ret; } @@ -406,12 +623,12 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags, u64 *subvol_rootid, struct btrfs_fs_devices **fs_devices) { substring_t args[MAX_OPT_ARGS]; - char *opts, *orig, *p; + char *device_name, *opts, *orig, *p; int error = 0; int intarg; if (!options) - goto out; + return 0; /* * strsep changes the string, duplicate it because parse_options @@ -430,6 +647,7 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags, token = match_token(p, tokens, args); switch (token) { case Opt_subvol: + kfree(*subvol_name); *subvol_name = match_strdup(&args[0]); break; case Opt_subvolid: @@ -457,42 +675,37 @@ static int btrfs_parse_early_options(const char *options, fmode_t flags, } break; case Opt_device: - error = btrfs_scan_one_device(match_strdup(&args[0]), + device_name = match_strdup(&args[0]); + if (!device_name) { + error = -ENOMEM; + goto out; + } + error = btrfs_scan_one_device(device_name, flags, holder, fs_devices); + kfree(device_name); if (error) - goto out_free_opts; + goto out; break; default: break; } } - out_free_opts: +out: kfree(orig); - out: - /* - * If no subvolume name is specified we use the default one. Allocate - * a copy of the string "." here so that code later in the - * mount path doesn't care if it's the default volume or another one. - */ - if (!*subvol_name) { - *subvol_name = kstrdup(".", GFP_KERNEL); - if (!*subvol_name) - return -ENOMEM; - } return error; } static struct dentry *get_default_root(struct super_block *sb, u64 subvol_objectid) { - struct btrfs_root *root = sb->s_fs_info; + struct btrfs_fs_info *fs_info = btrfs_sb(sb); + struct btrfs_root *root = fs_info->tree_root; struct btrfs_root *new_root; struct btrfs_dir_item *di; struct btrfs_path *path; struct btrfs_key location; struct inode *inode; - struct dentry *dentry; u64 dir_id; int new = 0; @@ -517,7 +730,7 @@ static struct dentry *get_default_root(struct super_block *sb, * will mount by default if we haven't been given a specific subvolume * to mount. */ - dir_id = btrfs_super_root_dir(&root->fs_info->super_copy); + dir_id = btrfs_super_root_dir(fs_info->super_copy); di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0); if (IS_ERR(di)) { btrfs_free_path(path); @@ -531,7 +744,7 @@ static struct dentry *get_default_root(struct super_block *sb, */ btrfs_free_path(path); dir_id = BTRFS_FIRST_FREE_OBJECTID; - new_root = root->fs_info->fs_root; + new_root = fs_info->fs_root; goto setup_root; } @@ -539,7 +752,7 @@ static struct dentry *get_default_root(struct super_block *sb, btrfs_free_path(path); find_root: - new_root = btrfs_read_fs_root_no_name(root->fs_info, &location); + new_root = btrfs_read_fs_root_no_name(fs_info, &location); if (IS_ERR(new_root)) return ERR_CAST(new_root); @@ -566,29 +779,7 @@ setup_root: return dget(sb->s_root); } - if (new) { - const struct qstr name = { .name = "/", .len = 1 }; - - /* - * New inode, we need to make the dentry a sibling of s_root so - * everything gets cleaned up properly on unmount. - */ - dentry = d_alloc(sb->s_root, &name); - if (!dentry) { - iput(inode); - return ERR_PTR(-ENOMEM); - } - d_splice_alias(inode, dentry); - } else { - /* - * We found the inode in cache, just find a dentry for it and - * put the reference to the inode we just got. - */ - dentry = d_find_alias(inode); - iput(inode); - } - - return dentry; + return d_obtain_alias(inode); } static int btrfs_fill_super(struct super_block *sb, @@ -596,8 +787,7 @@ static int btrfs_fill_super(struct super_block *sb, void *data, int silent) { struct inode *inode; - struct dentry *root_dentry; - struct btrfs_root *tree_root; + struct btrfs_fs_info *fs_info = btrfs_sb(sb); struct btrfs_key key; int err; @@ -611,69 +801,70 @@ static int btrfs_fill_super(struct super_block *sb, #ifdef CONFIG_BTRFS_FS_POSIX_ACL sb->s_flags |= MS_POSIXACL; #endif - - tree_root = open_ctree(sb, fs_devices, (char *)data); - - if (IS_ERR(tree_root)) { + sb->s_flags |= MS_I_VERSION; + err = open_ctree(sb, fs_devices, (char *)data); + if (err) { printk("btrfs: open_ctree failed\n"); - return PTR_ERR(tree_root); + return err; } - sb->s_fs_info = tree_root; key.objectid = BTRFS_FIRST_FREE_OBJECTID; key.type = BTRFS_INODE_ITEM_KEY; key.offset = 0; - inode = btrfs_iget(sb, &key, tree_root->fs_info->fs_root, NULL); + inode = btrfs_iget(sb, &key, fs_info->fs_root, NULL); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto fail_close; } - root_dentry = d_alloc_root(inode); - if (!root_dentry) { - iput(inode); + sb->s_root = d_make_root(inode); + if (!sb->s_root) { err = -ENOMEM; goto fail_close; } - sb->s_root = root_dentry; - save_mount_options(sb, data); cleancache_init_fs(sb); + sb->s_flags |= MS_ACTIVE; return 0; fail_close: - close_ctree(tree_root); + close_ctree(fs_info->tree_root); return err; } int btrfs_sync_fs(struct super_block *sb, int wait) { struct btrfs_trans_handle *trans; - struct btrfs_root *root = btrfs_sb(sb); - int ret; + struct btrfs_fs_info *fs_info = btrfs_sb(sb); + struct btrfs_root *root = fs_info->tree_root; trace_btrfs_sync_fs(wait); if (!wait) { - filemap_flush(root->fs_info->btree_inode->i_mapping); + filemap_flush(fs_info->btree_inode->i_mapping); return 0; } - btrfs_start_delalloc_inodes(root, 0); btrfs_wait_ordered_extents(root, 0, 0); - trans = btrfs_start_transaction(root, 0); + spin_lock(&fs_info->trans_lock); + if (!fs_info->running_transaction) { + spin_unlock(&fs_info->trans_lock); + return 0; + } + spin_unlock(&fs_info->trans_lock); + + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); - ret = btrfs_commit_transaction(trans, root); - return ret; + return btrfs_commit_transaction(trans, root); } -static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) +static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry) { - struct btrfs_root *root = btrfs_sb(vfs->mnt_sb); - struct btrfs_fs_info *info = root->fs_info; + struct btrfs_fs_info *info = btrfs_sb(dentry->d_sb); + struct btrfs_root *root = info->tree_root; char *compress_type; if (btrfs_test_opt(root, DEGRADED)) @@ -719,6 +910,8 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) seq_puts(seq, ",noacl"); if (btrfs_test_opt(root, SPACE_CACHE)) seq_puts(seq, ",space_cache"); + else + seq_puts(seq, ",nospace_cache"); if (btrfs_test_opt(root, CLEAR_CACHE)) seq_puts(seq, ",clear_cache"); if (btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED)) @@ -729,30 +922,119 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs) seq_puts(seq, ",autodefrag"); if (btrfs_test_opt(root, INODE_MAP_CACHE)) seq_puts(seq, ",inode_cache"); + if (btrfs_test_opt(root, SKIP_BALANCE)) + seq_puts(seq, ",skip_balance"); + if (btrfs_test_opt(root, PANIC_ON_FATAL_ERROR)) + seq_puts(seq, ",fatal_errors=panic"); return 0; } static int btrfs_test_super(struct super_block *s, void *data) { - struct btrfs_root *test_root = data; - struct btrfs_root *root = btrfs_sb(s); + struct btrfs_fs_info *p = data; + struct btrfs_fs_info *fs_info = btrfs_sb(s); - /* - * If this super block is going away, return false as it - * can't match as an existing super block. - */ - if (!atomic_read(&s->s_active)) - return 0; - return root->fs_info->fs_devices == test_root->fs_info->fs_devices; + return fs_info->fs_devices == p->fs_devices; } static int btrfs_set_super(struct super_block *s, void *data) { - s->s_fs_info = data; + int err = set_anon_super(s, data); + if (!err) + s->s_fs_info = data; + return err; +} - return set_anon_super(s, data); +/* + * subvolumes are identified by ino 256 + */ +static inline int is_subvolume_inode(struct inode *inode) +{ + if (inode && inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) + return 1; + return 0; +} + +/* + * This will strip out the subvol=%s argument for an argument string and add + * subvolid=0 to make sure we get the actual tree root for path walking to the + * subvol we want. + */ +static char *setup_root_args(char *args) +{ + unsigned len = strlen(args) + 2 + 1; + char *src, *dst, *buf; + + /* + * We need the same args as before, but with this substitution: + * s!subvol=[^,]+!subvolid=0! + * + * Since the replacement string is up to 2 bytes longer than the + * original, allocate strlen(args) + 2 + 1 bytes. + */ + + src = strstr(args, "subvol="); + /* This shouldn't happen, but just in case.. */ + if (!src) + return NULL; + + buf = dst = kmalloc(len, GFP_NOFS); + if (!buf) + return NULL; + + /* + * If the subvol= arg is not at the start of the string, + * copy whatever precedes it into buf. + */ + if (src != args) { + *src++ = '\0'; + strcpy(buf, args); + dst += strlen(args); + } + + strcpy(dst, "subvolid=0"); + dst += strlen("subvolid=0"); + + /* + * If there is a "," after the original subvol=... string, + * copy that suffix into our buffer. Otherwise, we're done. + */ + src = strchr(src, ','); + if (src) + strcpy(dst, src); + + return buf; } +static struct dentry *mount_subvol(const char *subvol_name, int flags, + const char *device_name, char *data) +{ + struct dentry *root; + struct vfsmount *mnt; + char *newargs; + + newargs = setup_root_args(data); + if (!newargs) + return ERR_PTR(-ENOMEM); + mnt = vfs_kern_mount(&btrfs_fs_type, flags, device_name, + newargs); + kfree(newargs); + if (IS_ERR(mnt)) + return ERR_CAST(mnt); + + root = mount_subtree(mnt, subvol_name); + + if (!IS_ERR(root) && !is_subvolume_inode(root->d_inode)) { + struct super_block *s = root->d_sb; + dput(root); + root = ERR_PTR(-EINVAL); + deactivate_locked_super(s); + printk(KERN_ERR "btrfs: '%s' is not a valid subvolume\n", + subvol_name); + } + + return root; +} /* * Find a superblock for the given device / mount point. @@ -767,7 +1049,6 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, struct super_block *s; struct dentry *root; struct btrfs_fs_devices *fs_devices = NULL; - struct btrfs_root *tree_root = NULL; struct btrfs_fs_info *fs_info = NULL; fmode_t mode = FMODE_READ; char *subvol_name = NULL; @@ -781,21 +1062,20 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, error = btrfs_parse_early_options(data, mode, fs_type, &subvol_name, &subvol_objectid, &subvol_rootid, &fs_devices); - if (error) + if (error) { + kfree(subvol_name); return ERR_PTR(error); + } - error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices); - if (error) - goto error_free_subvol_name; + if (subvol_name) { + root = mount_subvol(subvol_name, flags, device_name, data); + kfree(subvol_name); + return root; + } - error = btrfs_open_devices(fs_devices, mode, fs_type); + error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices); if (error) - goto error_free_subvol_name; - - if (!(flags & MS_RDONLY) && fs_devices->rw_devices == 0) { - error = -EACCES; - goto error_close_devices; - } + return ERR_PTR(error); /* * Setup a dummy root and fs_info for test/set super. This is because @@ -804,108 +1084,117 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, * then open_ctree will properly initialize everything later. */ fs_info = kzalloc(sizeof(struct btrfs_fs_info), GFP_NOFS); - tree_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); - if (!fs_info || !tree_root) { + if (!fs_info) + return ERR_PTR(-ENOMEM); + + fs_info->fs_devices = fs_devices; + + fs_info->super_copy = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS); + fs_info->super_for_commit = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_NOFS); + if (!fs_info->super_copy || !fs_info->super_for_commit) { error = -ENOMEM; + goto error_fs_info; + } + + error = btrfs_open_devices(fs_devices, mode, fs_type); + if (error) + goto error_fs_info; + + if (!(flags & MS_RDONLY) && fs_devices->rw_devices == 0) { + error = -EACCES; goto error_close_devices; } - fs_info->tree_root = tree_root; - fs_info->fs_devices = fs_devices; - tree_root->fs_info = fs_info; bdev = fs_devices->latest_bdev; - s = sget(fs_type, btrfs_test_super, btrfs_set_super, tree_root); - if (IS_ERR(s)) - goto error_s; + s = sget(fs_type, btrfs_test_super, btrfs_set_super, flags | MS_NOSEC, + fs_info); + if (IS_ERR(s)) { + error = PTR_ERR(s); + goto error_close_devices; + } if (s->s_root) { - if ((flags ^ s->s_flags) & MS_RDONLY) { - deactivate_locked_super(s); - error = -EBUSY; - goto error_close_devices; - } - btrfs_close_devices(fs_devices); - kfree(fs_info); - kfree(tree_root); + free_fs_info(fs_info); + if ((flags ^ s->s_flags) & MS_RDONLY) + error = -EBUSY; } else { char b[BDEVNAME_SIZE]; - s->s_flags = flags | MS_NOSEC; strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id)); + btrfs_sb(s)->bdev_holder = fs_type; error = btrfs_fill_super(s, fs_devices, data, flags & MS_SILENT ? 1 : 0); - if (error) { - deactivate_locked_super(s); - goto error_free_subvol_name; - } - - btrfs_sb(s)->fs_info->bdev_holder = fs_type; - s->s_flags |= MS_ACTIVE; } - /* if they gave us a subvolume name bind mount into that */ - if (strcmp(subvol_name, ".")) { - struct dentry *new_root; - - root = get_default_root(s, subvol_rootid); - if (IS_ERR(root)) { - error = PTR_ERR(root); - deactivate_locked_super(s); - goto error_free_subvol_name; - } - - mutex_lock(&root->d_inode->i_mutex); - new_root = lookup_one_len(subvol_name, root, - strlen(subvol_name)); - mutex_unlock(&root->d_inode->i_mutex); - - if (IS_ERR(new_root)) { - dput(root); - deactivate_locked_super(s); - error = PTR_ERR(new_root); - goto error_free_subvol_name; - } - if (!new_root->d_inode) { - dput(root); - dput(new_root); - deactivate_locked_super(s); - error = -ENXIO; - goto error_free_subvol_name; - } - dput(root); - root = new_root; - } else { - root = get_default_root(s, subvol_objectid); - if (IS_ERR(root)) { - error = PTR_ERR(root); - deactivate_locked_super(s); - goto error_free_subvol_name; - } - } + root = !error ? get_default_root(s, subvol_objectid) : ERR_PTR(error); + if (IS_ERR(root)) + deactivate_locked_super(s); - kfree(subvol_name); return root; -error_s: - error = PTR_ERR(s); error_close_devices: btrfs_close_devices(fs_devices); - kfree(fs_info); - kfree(tree_root); -error_free_subvol_name: - kfree(subvol_name); +error_fs_info: + free_fs_info(fs_info); return ERR_PTR(error); } +static void btrfs_set_max_workers(struct btrfs_workers *workers, int new_limit) +{ + spin_lock_irq(&workers->lock); + workers->max_workers = new_limit; + spin_unlock_irq(&workers->lock); +} + +static void btrfs_resize_thread_pool(struct btrfs_fs_info *fs_info, + int new_pool_size, int old_pool_size) +{ + if (new_pool_size == old_pool_size) + return; + + fs_info->thread_pool_size = new_pool_size; + + printk(KERN_INFO "btrfs: resize thread pool %d -> %d\n", + old_pool_size, new_pool_size); + + btrfs_set_max_workers(&fs_info->generic_worker, new_pool_size); + btrfs_set_max_workers(&fs_info->workers, new_pool_size); + btrfs_set_max_workers(&fs_info->delalloc_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->submit_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->caching_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->fixup_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->endio_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->endio_meta_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->endio_meta_write_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->endio_write_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->endio_freespace_worker, new_pool_size); + btrfs_set_max_workers(&fs_info->delayed_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->readahead_workers, new_pool_size); + btrfs_set_max_workers(&fs_info->scrub_workers, new_pool_size); +} + static int btrfs_remount(struct super_block *sb, int *flags, char *data) { - struct btrfs_root *root = btrfs_sb(sb); + struct btrfs_fs_info *fs_info = btrfs_sb(sb); + struct btrfs_root *root = fs_info->tree_root; + unsigned old_flags = sb->s_flags; + unsigned long old_opts = fs_info->mount_opt; + unsigned long old_compress_type = fs_info->compress_type; + u64 old_max_inline = fs_info->max_inline; + u64 old_alloc_start = fs_info->alloc_start; + int old_thread_pool_size = fs_info->thread_pool_size; + unsigned int old_metadata_ratio = fs_info->metadata_ratio; int ret; ret = btrfs_parse_options(root, data); - if (ret) - return -EINVAL; + if (ret) { + ret = -EINVAL; + goto restore; + } + + btrfs_resize_thread_pool(fs_info, + fs_info->thread_pool_size, old_thread_pool_size); if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) return 0; @@ -913,26 +1202,51 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) if (*flags & MS_RDONLY) { sb->s_flags |= MS_RDONLY; - ret = btrfs_commit_super(root); - WARN_ON(ret); + ret = btrfs_commit_super(root); + if (ret) + goto restore; } else { - if (root->fs_info->fs_devices->rw_devices == 0) - return -EACCES; + if (fs_info->fs_devices->rw_devices == 0) { + ret = -EACCES; + goto restore; + } - if (btrfs_super_log_root(&root->fs_info->super_copy) != 0) - return -EINVAL; + if (btrfs_super_log_root(fs_info->super_copy) != 0) { + ret = -EINVAL; + goto restore; + } - ret = btrfs_cleanup_fs_roots(root->fs_info); - WARN_ON(ret); + ret = btrfs_cleanup_fs_roots(fs_info); + if (ret) + goto restore; /* recover relocation */ ret = btrfs_recover_relocation(root); - WARN_ON(ret); + if (ret) + goto restore; + + ret = btrfs_resume_balance_async(fs_info); + if (ret) + goto restore; sb->s_flags &= ~MS_RDONLY; } return 0; + +restore: + /* We've hit an error - don't reset MS_RDONLY */ + if (sb->s_flags & MS_RDONLY) + old_flags |= MS_RDONLY; + sb->s_flags = old_flags; + fs_info->mount_opt = old_opts; + fs_info->compress_type = old_compress_type; + fs_info->max_inline = old_max_inline; + fs_info->alloc_start = old_alloc_start; + btrfs_resize_thread_pool(fs_info, + old_thread_pool_size, fs_info->thread_pool_size); + fs_info->metadata_ratio = old_metadata_ratio; + return ret; } /* Used to sort the devices by max_avail(descending sort) */ @@ -976,11 +1290,11 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) u64 avail_space; u64 used_space; u64 min_stripe_size; - int min_stripes = 1; + int min_stripes = 1, num_stripes = 1; int i = 0, nr_devices; int ret; - nr_devices = fs_info->fs_devices->rw_devices; + nr_devices = fs_info->fs_devices->open_devices; BUG_ON(!nr_devices); devices_info = kmalloc(sizeof(*devices_info) * nr_devices, @@ -990,20 +1304,24 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) /* calc min stripe number for data space alloction */ type = btrfs_get_alloc_profile(root, 1); - if (type & BTRFS_BLOCK_GROUP_RAID0) + if (type & BTRFS_BLOCK_GROUP_RAID0) { min_stripes = 2; - else if (type & BTRFS_BLOCK_GROUP_RAID1) + num_stripes = nr_devices; + } else if (type & BTRFS_BLOCK_GROUP_RAID1) { min_stripes = 2; - else if (type & BTRFS_BLOCK_GROUP_RAID10) + num_stripes = 2; + } else if (type & BTRFS_BLOCK_GROUP_RAID10) { min_stripes = 4; + num_stripes = 4; + } if (type & BTRFS_BLOCK_GROUP_DUP) min_stripe_size = 2 * BTRFS_STRIPE_LEN; else min_stripe_size = BTRFS_STRIPE_LEN; - list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) { - if (!device->in_fs_metadata) + list_for_each_entry(device, &fs_devices->devices, dev_list) { + if (!device->in_fs_metadata || !device->bdev) continue; avail_space = device->total_bytes - device->bytes_used; @@ -1064,13 +1382,16 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) i = nr_devices - 1; avail_space = 0; while (nr_devices >= min_stripes) { + if (num_stripes > nr_devices) + num_stripes = nr_devices; + if (devices_info[i].max_avail >= min_stripe_size) { int j; u64 alloc_size; - avail_space += devices_info[i].max_avail * min_stripes; + avail_space += devices_info[i].max_avail * num_stripes; alloc_size = devices_info[i].max_avail; - for (j = i + 1 - min_stripes; j <= i; j++) + for (j = i + 1 - num_stripes; j <= i; j++) devices_info[j].max_avail -= alloc_size; } i--; @@ -1084,18 +1405,18 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) { - struct btrfs_root *root = btrfs_sb(dentry->d_sb); - struct btrfs_super_block *disk_super = &root->fs_info->super_copy; - struct list_head *head = &root->fs_info->space_info; + struct btrfs_fs_info *fs_info = btrfs_sb(dentry->d_sb); + struct btrfs_super_block *disk_super = fs_info->super_copy; + struct list_head *head = &fs_info->space_info; struct btrfs_space_info *found; u64 total_used = 0; u64 total_free_data = 0; int bits = dentry->d_sb->s_blocksize_bits; - __be32 *fsid = (__be32 *)root->fs_info->fsid; + __be32 *fsid = (__be32 *)fs_info->fsid; int ret; /* holding chunk_muext to avoid allocating new chunks */ - mutex_lock(&root->fs_info->chunk_mutex); + mutex_lock(&fs_info->chunk_mutex); rcu_read_lock(); list_for_each_entry_rcu(found, head, list) { if (found->flags & BTRFS_BLOCK_GROUP_DATA) { @@ -1114,14 +1435,14 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_bsize = dentry->d_sb->s_blocksize; buf->f_type = BTRFS_SUPER_MAGIC; buf->f_bavail = total_free_data; - ret = btrfs_calc_avail_data_space(root, &total_free_data); + ret = btrfs_calc_avail_data_space(fs_info->tree_root, &total_free_data); if (ret) { - mutex_unlock(&root->fs_info->chunk_mutex); + mutex_unlock(&fs_info->chunk_mutex); return ret; } buf->f_bavail += total_free_data; buf->f_bavail = buf->f_bavail >> bits; - mutex_unlock(&root->fs_info->chunk_mutex); + mutex_unlock(&fs_info->chunk_mutex); /* We treat it as constant endianness (it doesn't matter _which_) because we want the fsid to come out the same whether mounted @@ -1135,11 +1456,18 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) return 0; } +static void btrfs_kill_super(struct super_block *sb) +{ + struct btrfs_fs_info *fs_info = btrfs_sb(sb); + kill_anon_super(sb); + free_fs_info(fs_info); +} + static struct file_system_type btrfs_fs_type = { .owner = THIS_MODULE, .name = "btrfs", .mount = btrfs_mount, - .kill_sb = kill_anon_super, + .kill_sb = btrfs_kill_super, .fs_flags = FS_REQUIRES_DEV, }; @@ -1165,6 +1493,13 @@ static long btrfs_control_ioctl(struct file *file, unsigned int cmd, ret = btrfs_scan_one_device(vol->name, FMODE_READ, &btrfs_fs_type, &fs_devices); break; + case BTRFS_IOC_DEVICES_READY: + ret = btrfs_scan_one_device(vol->name, FMODE_READ, + &btrfs_fs_type, &fs_devices); + if (ret) + break; + ret = !(fs_devices->num_devices == fs_devices->total_devices); + break; } kfree(vol); @@ -1173,17 +1508,50 @@ static long btrfs_control_ioctl(struct file *file, unsigned int cmd, static int btrfs_freeze(struct super_block *sb) { - struct btrfs_root *root = btrfs_sb(sb); - mutex_lock(&root->fs_info->transaction_kthread_mutex); - mutex_lock(&root->fs_info->cleaner_mutex); + struct btrfs_fs_info *fs_info = btrfs_sb(sb); + mutex_lock(&fs_info->transaction_kthread_mutex); + mutex_lock(&fs_info->cleaner_mutex); return 0; } static int btrfs_unfreeze(struct super_block *sb) { - struct btrfs_root *root = btrfs_sb(sb); - mutex_unlock(&root->fs_info->cleaner_mutex); - mutex_unlock(&root->fs_info->transaction_kthread_mutex); + struct btrfs_fs_info *fs_info = btrfs_sb(sb); + mutex_unlock(&fs_info->cleaner_mutex); + mutex_unlock(&fs_info->transaction_kthread_mutex); + return 0; +} + +static int btrfs_show_devname(struct seq_file *m, struct dentry *root) +{ + struct btrfs_fs_info *fs_info = btrfs_sb(root->d_sb); + struct btrfs_fs_devices *cur_devices; + struct btrfs_device *dev, *first_dev = NULL; + struct list_head *head; + struct rcu_string *name; + + mutex_lock(&fs_info->fs_devices->device_list_mutex); + cur_devices = fs_info->fs_devices; + while (cur_devices) { + head = &cur_devices->devices; + list_for_each_entry(dev, head, dev_list) { + if (dev->missing) + continue; + if (!first_dev || dev->devid < first_dev->devid) + first_dev = dev; + } + cur_devices = cur_devices->seed; + } + + if (first_dev) { + rcu_read_lock(); + name = rcu_dereference(first_dev->name); + seq_escape(m, name->str, " \t\n\\"); + rcu_read_unlock(); + } else { + WARN_ON(1); + } + mutex_unlock(&fs_info->fs_devices->device_list_mutex); return 0; } @@ -1193,8 +1561,8 @@ static const struct super_operations btrfs_super_ops = { .put_super = btrfs_put_super, .sync_fs = btrfs_sync_fs, .show_options = btrfs_show_options, + .show_devname = btrfs_show_devname, .write_inode = btrfs_write_inode, - .dirty_inode = btrfs_dirty_inode, .alloc_inode = btrfs_alloc_inode, .destroy_inode = btrfs_destroy_inode, .statfs = btrfs_statfs, @@ -1238,9 +1606,7 @@ static int __init init_btrfs_fs(void) if (err) return err; - err = btrfs_init_compress(); - if (err) - goto free_sysfs; + btrfs_init_compress(); err = btrfs_init_cachep(); if (err) @@ -1266,6 +1632,8 @@ static int __init init_btrfs_fs(void) if (err) goto unregister_ioctl; + btrfs_init_lockdep(); + printk(KERN_INFO "%s loaded\n", BTRFS_BUILD_VERSION); return 0; @@ -1281,7 +1649,6 @@ free_cachep: btrfs_destroy_cachep(); free_compress: btrfs_exit_compress(); -free_sysfs: btrfs_exit_sysfs(); return err; } diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index e24b7964a155..27c26004e050 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -22,20 +22,23 @@ #include <linux/writeback.h> #include <linux/pagemap.h> #include <linux/blkdev.h> +#include <linux/uuid.h> #include "ctree.h" #include "disk-io.h" #include "transaction.h" #include "locking.h" #include "tree-log.h" #include "inode-map.h" +#include "volumes.h" #define BTRFS_ROOT_TRANS_TAG 0 -static noinline void put_transaction(struct btrfs_transaction *transaction) +void put_transaction(struct btrfs_transaction *transaction) { WARN_ON(atomic_read(&transaction->use_count) == 0); if (atomic_dec_and_test(&transaction->use_count)) { BUG_ON(!list_empty(&transaction->list)); + WARN_ON(transaction->delayed_refs.root.rb_node); memset(transaction, 0, sizeof(*transaction)); kmem_cache_free(btrfs_transaction_cachep, transaction); } @@ -53,38 +56,56 @@ static noinline void switch_commit_root(struct btrfs_root *root) static noinline int join_transaction(struct btrfs_root *root, int nofail) { struct btrfs_transaction *cur_trans; + struct btrfs_fs_info *fs_info = root->fs_info; - spin_lock(&root->fs_info->trans_lock); - if (root->fs_info->trans_no_join) { + spin_lock(&fs_info->trans_lock); +loop: + /* The file system has been taken offline. No new transactions. */ + if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { + spin_unlock(&fs_info->trans_lock); + return -EROFS; + } + + if (fs_info->trans_no_join) { if (!nofail) { - spin_unlock(&root->fs_info->trans_lock); + spin_unlock(&fs_info->trans_lock); return -EBUSY; } } - cur_trans = root->fs_info->running_transaction; + cur_trans = fs_info->running_transaction; if (cur_trans) { + if (cur_trans->aborted) { + spin_unlock(&fs_info->trans_lock); + return cur_trans->aborted; + } atomic_inc(&cur_trans->use_count); atomic_inc(&cur_trans->num_writers); cur_trans->num_joined++; - spin_unlock(&root->fs_info->trans_lock); + spin_unlock(&fs_info->trans_lock); return 0; } - spin_unlock(&root->fs_info->trans_lock); + spin_unlock(&fs_info->trans_lock); cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS); if (!cur_trans) return -ENOMEM; - spin_lock(&root->fs_info->trans_lock); - if (root->fs_info->running_transaction) { + + spin_lock(&fs_info->trans_lock); + if (fs_info->running_transaction) { + /* + * someone started a transaction after we unlocked. Make sure + * to redo the trans_no_join checks above + */ kmem_cache_free(btrfs_transaction_cachep, cur_trans); - cur_trans = root->fs_info->running_transaction; - atomic_inc(&cur_trans->use_count); - atomic_inc(&cur_trans->num_writers); - cur_trans->num_joined++; - spin_unlock(&root->fs_info->trans_lock); - return 0; + cur_trans = fs_info->running_transaction; + goto loop; + } else if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { + spin_unlock(&fs_info->trans_lock); + kmem_cache_free(btrfs_transaction_cachep, cur_trans); + return -EROFS; } + atomic_set(&cur_trans->num_writers, 1); cur_trans->num_joined = 0; init_waitqueue_head(&cur_trans->writer_wait); @@ -105,17 +126,36 @@ static noinline int join_transaction(struct btrfs_root *root, int nofail) cur_trans->delayed_refs.num_heads = 0; cur_trans->delayed_refs.flushing = 0; cur_trans->delayed_refs.run_delayed_start = 0; + + /* + * although the tree mod log is per file system and not per transaction, + * the log must never go across transaction boundaries. + */ + smp_mb(); + if (!list_empty(&fs_info->tree_mod_seq_list)) { + printk(KERN_ERR "btrfs: tree_mod_seq_list not empty when " + "creating a fresh transaction\n"); + WARN_ON(1); + } + if (!RB_EMPTY_ROOT(&fs_info->tree_mod_log)) { + printk(KERN_ERR "btrfs: tree_mod_log rb tree not empty when " + "creating a fresh transaction\n"); + WARN_ON(1); + } + atomic_set(&fs_info->tree_mod_seq, 0); + spin_lock_init(&cur_trans->commit_lock); spin_lock_init(&cur_trans->delayed_refs.lock); INIT_LIST_HEAD(&cur_trans->pending_snapshots); - list_add_tail(&cur_trans->list, &root->fs_info->trans_list); + list_add_tail(&cur_trans->list, &fs_info->trans_list); extent_io_tree_init(&cur_trans->dirty_pages, - root->fs_info->btree_inode->i_mapping); - root->fs_info->generation++; - cur_trans->transid = root->fs_info->generation; - root->fs_info->running_transaction = cur_trans; - spin_unlock(&root->fs_info->trans_lock); + fs_info->btree_inode->i_mapping); + fs_info->generation++; + cur_trans->transid = fs_info->generation; + fs_info->running_transaction = cur_trans; + cur_trans->aborted = 0; + spin_unlock(&fs_info->trans_lock); return 0; } @@ -256,6 +296,7 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root, struct btrfs_transaction *cur_trans; u64 num_bytes = 0; int ret; + u64 qgroup_reserved = 0; if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) return ERR_PTR(-EROFS); @@ -274,8 +315,16 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root, * the appropriate flushing if need be. */ if (num_items > 0 && root != root->fs_info->chunk_root) { + if (root->fs_info->quota_enabled && + is_fstree(root->root_key.objectid)) { + qgroup_reserved = num_items * root->leafsize; + ret = btrfs_qgroup_reserve(root, qgroup_reserved); + if (ret) + return ERR_PTR(ret); + } + num_bytes = btrfs_calc_trans_metadata_size(root, num_items); - ret = btrfs_block_rsv_add(NULL, root, + ret = btrfs_block_rsv_add(root, &root->fs_info->trans_block_rsv, num_bytes); if (ret) @@ -286,6 +335,8 @@ again: if (!h) return ERR_PTR(-ENOMEM); + sb_start_intwrite(root->fs_info->sb); + if (may_wait_transaction(root, type)) wait_current_trans(root); @@ -296,6 +347,7 @@ again: } while (ret == -EBUSY); if (ret < 0) { + sb_end_intwrite(root->fs_info->sb); kmem_cache_free(btrfs_trans_handle_cachep, h); return ERR_PTR(ret); } @@ -306,10 +358,16 @@ again: h->transaction = cur_trans; h->blocks_used = 0; h->bytes_reserved = 0; + h->root = root; h->delayed_ref_updates = 0; h->use_count = 1; + h->adding_csums = 0; h->block_rsv = NULL; h->orig_rsv = NULL; + h->aborted = 0; + h->qgroup_reserved = qgroup_reserved; + h->delayed_ref_elem.seq = 0; + INIT_LIST_HEAD(&h->qgroup_ref_list); smp_mb(); if (cur_trans->blocked && may_wait_transaction(root, type)) { @@ -318,6 +376,8 @@ again: } if (num_bytes) { + trace_btrfs_space_reservation(root->fs_info, "transaction", + h->transid, num_bytes, 1); h->block_rsv = &root->fs_info->trans_block_rsv; h->bytes_reserved = num_bytes; } @@ -418,8 +478,8 @@ static int should_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root) { int ret; - ret = btrfs_block_rsv_check(trans, root, - &root->fs_info->global_block_rsv, 0, 5); + + ret = btrfs_block_rsv_check(root, &root->fs_info->global_block_rsv, 5); return ret ? 1 : 0; } @@ -428,6 +488,7 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans, { struct btrfs_transaction *cur_trans = trans->transaction; int updates; + int err; smp_mb(); if (cur_trans->blocked || cur_trans->delayed_refs.flushing) @@ -435,8 +496,11 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans, updates = trans->delayed_ref_updates; trans->delayed_ref_updates = 0; - if (updates) - btrfs_run_delayed_refs(trans, root, updates); + if (updates) { + err = btrfs_run_delayed_refs(trans, root, updates); + if (err) /* Error code will also eval true */ + return err; + } return should_end_transaction(trans, root); } @@ -447,33 +511,47 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_transaction *cur_trans = trans->transaction; struct btrfs_fs_info *info = root->fs_info; int count = 0; + int err = 0; if (--trans->use_count) { trans->block_rsv = trans->orig_rsv; return 0; } - while (count < 4) { + /* + * do the qgroup accounting as early as possible + */ + err = btrfs_delayed_refs_qgroup_accounting(trans, info); + + btrfs_trans_release_metadata(trans, root); + trans->block_rsv = NULL; + /* + * the same root has to be passed to start_transaction and + * end_transaction. Subvolume quota depends on this. + */ + WARN_ON(trans->root != root); + + if (trans->qgroup_reserved) { + btrfs_qgroup_free(root, trans->qgroup_reserved); + trans->qgroup_reserved = 0; + } + + while (count < 2) { unsigned long cur = trans->delayed_ref_updates; trans->delayed_ref_updates = 0; if (cur && trans->transaction->delayed_refs.num_heads_ready > 64) { trans->delayed_ref_updates = 0; - - /* - * do a full flush if the transaction is trying - * to close - */ - if (trans->transaction->delayed_refs.flushing) - cur = 0; btrfs_run_delayed_refs(trans, root, cur); } else { break; } count++; } - btrfs_trans_release_metadata(trans, root); + trans->block_rsv = NULL; + + sb_end_intwrite(root->fs_info->sb); if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) && should_end_transaction(trans, root)) { @@ -506,13 +584,19 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, if (current->journal_info == trans) current->journal_info = NULL; - memset(trans, 0, sizeof(*trans)); - kmem_cache_free(btrfs_trans_handle_cachep, trans); if (throttle) btrfs_run_delayed_iputs(root); - return 0; + if (trans->aborted || + root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { + err = -EIO; + } + assert_qgroups_uptodate(trans); + + memset(trans, 0, sizeof(*trans)); + kmem_cache_free(btrfs_trans_handle_cachep, trans); + return err; } int btrfs_end_transaction(struct btrfs_trans_handle *trans, @@ -562,50 +646,21 @@ int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans, int btrfs_write_marked_extents(struct btrfs_root *root, struct extent_io_tree *dirty_pages, int mark) { - int ret; int err = 0; int werr = 0; - struct page *page; - struct inode *btree_inode = root->fs_info->btree_inode; + struct address_space *mapping = root->fs_info->btree_inode->i_mapping; u64 start = 0; u64 end; - unsigned long index; - - while (1) { - ret = find_first_extent_bit(dirty_pages, start, &start, &end, - mark); - if (ret) - break; - while (start <= end) { - cond_resched(); - - index = start >> PAGE_CACHE_SHIFT; - start = (u64)(index + 1) << PAGE_CACHE_SHIFT; - page = find_get_page(btree_inode->i_mapping, index); - if (!page) - continue; - - btree_lock_page_hook(page); - if (!page->mapping) { - unlock_page(page); - page_cache_release(page); - continue; - } - if (PageWriteback(page)) { - if (PageDirty(page)) - wait_on_page_writeback(page); - else { - unlock_page(page); - page_cache_release(page); - continue; - } - } - err = write_one_page(page, 0); - if (err) - werr = err; - page_cache_release(page); - } + while (!find_first_extent_bit(dirty_pages, start, &start, &end, + mark)) { + convert_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT, mark, + GFP_NOFS); + err = filemap_fdatawrite_range(mapping, start, end); + if (err) + werr = err; + cond_resched(); + start = end + 1; } if (err) werr = err; @@ -621,39 +676,20 @@ int btrfs_write_marked_extents(struct btrfs_root *root, int btrfs_wait_marked_extents(struct btrfs_root *root, struct extent_io_tree *dirty_pages, int mark) { - int ret; int err = 0; int werr = 0; - struct page *page; - struct inode *btree_inode = root->fs_info->btree_inode; + struct address_space *mapping = root->fs_info->btree_inode->i_mapping; u64 start = 0; u64 end; - unsigned long index; - while (1) { - ret = find_first_extent_bit(dirty_pages, start, &start, &end, - mark); - if (ret) - break; - - clear_extent_bits(dirty_pages, start, end, mark, GFP_NOFS); - while (start <= end) { - index = start >> PAGE_CACHE_SHIFT; - start = (u64)(index + 1) << PAGE_CACHE_SHIFT; - page = find_get_page(btree_inode->i_mapping, index); - if (!page) - continue; - if (PageDirty(page)) { - btree_lock_page_hook(page); - wait_on_page_writeback(page); - err = write_one_page(page, 0); - if (err) - werr = err; - } - wait_on_page_writeback(page); - page_cache_release(page); - cond_resched(); - } + while (!find_first_extent_bit(dirty_pages, start, &start, &end, + EXTENT_NEED_WAIT)) { + clear_extent_bits(dirty_pages, start, end, EXTENT_NEED_WAIT, GFP_NOFS); + err = filemap_fdatawait_range(mapping, start, end); + if (err) + werr = err; + cond_resched(); + start = end + 1; } if (err) werr = err; @@ -673,7 +709,12 @@ int btrfs_write_and_wait_marked_extents(struct btrfs_root *root, ret = btrfs_write_marked_extents(root, dirty_pages, mark); ret2 = btrfs_wait_marked_extents(root, dirty_pages, mark); - return ret || ret2; + + if (ret) + return ret; + if (ret2) + return ret2; + return 0; } int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans, @@ -720,11 +761,13 @@ static int update_cowonly_root(struct btrfs_trans_handle *trans, ret = btrfs_update_root(trans, tree_root, &root->root_key, &root->root_item); - BUG_ON(ret); + if (ret) + return ret; old_root_used = btrfs_root_used(&root->root_item); ret = btrfs_write_dirty_block_groups(trans, root); - BUG_ON(ret); + if (ret) + return ret; } if (root != root->fs_info->extent_root) @@ -735,6 +778,10 @@ static int update_cowonly_root(struct btrfs_trans_handle *trans, /* * update all the cowonly tree roots on disk + * + * The error handling in this function may not be obvious. Any of the + * failures will cause the file system to go offline. We still need + * to clean up the delayed refs. */ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans, struct btrfs_root *root) @@ -745,13 +792,29 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans, int ret; ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); - BUG_ON(ret); + if (ret) + return ret; eb = btrfs_lock_root_node(fs_info->tree_root); - btrfs_cow_block(trans, fs_info->tree_root, eb, NULL, 0, &eb); + ret = btrfs_cow_block(trans, fs_info->tree_root, eb, NULL, + 0, &eb); btrfs_tree_unlock(eb); free_extent_buffer(eb); + if (ret) + return ret; + + ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); + if (ret) + return ret; + + ret = btrfs_run_dev_stats(trans, root->fs_info); + BUG_ON(ret); + + ret = btrfs_run_qgroups(trans, root->fs_info); + BUG_ON(ret); + + /* run_qgroups might have added some more refs */ ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); BUG_ON(ret); @@ -760,7 +823,9 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans, list_del_init(next); root = list_entry(next, struct btrfs_root, dirty_list); - update_cowonly_root(trans, root); + ret = update_cowonly_root(trans, root); + if (ret) + return ret; } down_write(&fs_info->extent_commit_sem); @@ -816,6 +881,10 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, btrfs_save_ino_cache(root, trans); + /* see comments in should_cow_block() */ + root->force_cow = 0; + smp_wmb(); + if (root->commit_root != root->node) { mutex_lock(&root->fs_commit_mutex); switch_commit_root(root); @@ -890,17 +959,19 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, struct dentry *dentry; struct extent_buffer *tmp; struct extent_buffer *old; + struct timespec cur_time = CURRENT_TIME; int ret; u64 to_reserve = 0; u64 index = 0; u64 objectid; u64 root_flags; + uuid_le new_uuid; rsv = trans->block_rsv; new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS); if (!new_root_item) { - pending->error = -ENOMEM; + ret = pending->error = -ENOMEM; goto fail; } @@ -911,17 +982,24 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, } btrfs_reloc_pre_snapshot(trans, pending, &to_reserve); - btrfs_orphan_pre_snapshot(trans, pending, &to_reserve); if (to_reserve > 0) { - ret = btrfs_block_rsv_add(trans, root, &pending->block_rsv, - to_reserve); + ret = btrfs_block_rsv_add_noflush(root, &pending->block_rsv, + to_reserve); if (ret) { pending->error = ret; goto fail; } } + ret = btrfs_qgroup_inherit(trans, fs_info, root->root_key.objectid, + objectid, pending->inherit); + kfree(pending->inherit); + if (ret) { + pending->error = ret; + goto fail; + } + key.objectid = objectid; key.offset = (u64)-1; key.type = BTRFS_ROOT_ITEM_KEY; @@ -938,17 +1016,25 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, * insert the directory item */ ret = btrfs_set_inode_index(parent_inode, &index); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM */ ret = btrfs_insert_dir_item(trans, parent_root, dentry->d_name.name, dentry->d_name.len, parent_inode, &key, BTRFS_FT_DIR, index); - BUG_ON(ret); + if (ret == -EEXIST) { + pending->error = -EEXIST; + dput(parent); + goto fail; + } else if (ret) { + goto abort_trans_dput; + } btrfs_i_size_write(parent_inode, parent_inode->i_size + dentry->d_name.len * 2); + parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME; ret = btrfs_update_inode(trans, parent_root, parent_inode); - BUG_ON(ret); + if (ret) + goto abort_trans_dput; /* * pull in the delayed directory update @@ -957,7 +1043,10 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, * snapshot */ ret = btrfs_run_delayed_items(trans, root); - BUG_ON(ret); + if (ret) { /* Transaction aborted */ + dput(parent); + goto fail; + } record_root_in_trans(trans, root); btrfs_set_root_last_snapshot(&root->root_item, trans->transid); @@ -971,13 +1060,40 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, root_flags &= ~BTRFS_ROOT_SUBVOL_RDONLY; btrfs_set_root_flags(new_root_item, root_flags); + btrfs_set_root_generation_v2(new_root_item, + trans->transid); + uuid_le_gen(&new_uuid); + memcpy(new_root_item->uuid, new_uuid.b, BTRFS_UUID_SIZE); + memcpy(new_root_item->parent_uuid, root->root_item.uuid, + BTRFS_UUID_SIZE); + new_root_item->otime.sec = cpu_to_le64(cur_time.tv_sec); + new_root_item->otime.nsec = cpu_to_le32(cur_time.tv_nsec); + btrfs_set_root_otransid(new_root_item, trans->transid); + memset(&new_root_item->stime, 0, sizeof(new_root_item->stime)); + memset(&new_root_item->rtime, 0, sizeof(new_root_item->rtime)); + btrfs_set_root_stransid(new_root_item, 0); + btrfs_set_root_rtransid(new_root_item, 0); + old = btrfs_lock_root_node(root); - btrfs_cow_block(trans, root, old, NULL, 0, &old); + ret = btrfs_cow_block(trans, root, old, NULL, 0, &old); + if (ret) { + btrfs_tree_unlock(old); + free_extent_buffer(old); + goto abort_trans_dput; + } + btrfs_set_lock_blocking(old); - btrfs_copy_root(trans, root, old, &tmp, objectid); + ret = btrfs_copy_root(trans, root, old, &tmp, objectid); + /* clean up in any case */ btrfs_tree_unlock(old); free_extent_buffer(old); + if (ret) + goto abort_trans_dput; + + /* see comments in should_cow_block() */ + root->force_cow = 1; + smp_wmb(); btrfs_set_root_node(new_root_item, tmp); /* record when the snapshot was created in key.offset */ @@ -985,7 +1101,8 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, ret = btrfs_insert_root(trans, tree_root, &key, new_root_item); btrfs_tree_unlock(tmp); free_extent_buffer(tmp); - BUG_ON(ret); + if (ret) + goto abort_trans_dput; /* * insert root back/forward references @@ -994,20 +1111,32 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, parent_root->root_key.objectid, btrfs_ino(parent_inode), index, dentry->d_name.name, dentry->d_name.len); - BUG_ON(ret); dput(parent); + if (ret) + goto fail; key.offset = (u64)-1; pending->snap = btrfs_read_fs_root_no_name(root->fs_info, &key); - BUG_ON(IS_ERR(pending->snap)); + if (IS_ERR(pending->snap)) { + ret = PTR_ERR(pending->snap); + goto abort_trans; + } - btrfs_reloc_post_snapshot(trans, pending); - btrfs_orphan_post_snapshot(trans, pending); + ret = btrfs_reloc_post_snapshot(trans, pending); + if (ret) + goto abort_trans; + ret = 0; fail: kfree(new_root_item); trans->block_rsv = rsv; btrfs_block_rsv_release(root, &pending->block_rsv, (u64)-1); - return 0; + return ret; + +abort_trans_dput: + dput(parent); +abort_trans: + btrfs_abort_transaction(trans, root, ret); + goto fail; } /* @@ -1018,12 +1147,9 @@ static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans, { struct btrfs_pending_snapshot *pending; struct list_head *head = &trans->transaction->pending_snapshots; - int ret; - list_for_each_entry(pending, head, list) { - ret = create_pending_snapshot(trans, fs_info, pending); - BUG_ON(ret); - } + list_for_each_entry(pending, head, list) + create_pending_snapshot(trans, fs_info, pending); return 0; } @@ -1032,7 +1158,7 @@ static void update_super_roots(struct btrfs_root *root) struct btrfs_root_item *root_item; struct btrfs_super_block *super; - super = &root->fs_info->super_copy; + super = root->fs_info->super_copy; root_item = &root->fs_info->chunk_root->root_item; super->chunk_root = root_item->bytenr; @@ -1043,7 +1169,7 @@ static void update_super_roots(struct btrfs_root *root) super->root = root_item->bytenr; super->generation = root_item->generation; super->root_level = root_item->level; - if (super->cache_generation != 0 || btrfs_test_opt(root, SPACE_CACHE)) + if (btrfs_test_opt(root, SPACE_CACHE)) super->cache_generation = root_item->generation; } @@ -1147,6 +1273,39 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, return 0; } + +static void cleanup_transaction(struct btrfs_trans_handle *trans, + struct btrfs_root *root, int err) +{ + struct btrfs_transaction *cur_trans = trans->transaction; + + WARN_ON(trans->use_count > 1); + + btrfs_abort_transaction(trans, root, err); + + spin_lock(&root->fs_info->trans_lock); + list_del_init(&cur_trans->list); + if (cur_trans == root->fs_info->running_transaction) { + root->fs_info->running_transaction = NULL; + root->fs_info->trans_no_join = 0; + } + spin_unlock(&root->fs_info->trans_lock); + + btrfs_cleanup_one_transaction(trans->transaction, root); + + put_transaction(cur_trans); + put_transaction(cur_trans); + + trace_btrfs_transaction_commit(root); + + btrfs_scrub_continue(root); + + if (current->journal_info == trans) + current->journal_info = NULL; + + kmem_cache_free(btrfs_trans_handle_cachep, trans); +} + /* * btrfs_transaction state sequence: * in_commit = 0, blocked = 0 (initial) @@ -1158,25 +1317,31 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root) { unsigned long joined = 0; - struct btrfs_transaction *cur_trans; + struct btrfs_transaction *cur_trans = trans->transaction; struct btrfs_transaction *prev_trans = NULL; DEFINE_WAIT(wait); - int ret; + int ret = -EIO; int should_grow = 0; unsigned long now = get_seconds(); int flush_on_commit = btrfs_test_opt(root, FLUSHONCOMMIT); btrfs_run_ordered_operations(root, 0); + if (cur_trans->aborted) + goto cleanup_transaction; + /* make a pass through all the delayed refs we have so far * any runnings procs may add more while we are here */ ret = btrfs_run_delayed_refs(trans, root, 0); - BUG_ON(ret); + if (ret) + goto cleanup_transaction; btrfs_trans_release_metadata(trans, root); + trans->block_rsv = NULL; cur_trans = trans->transaction; + /* * set the flushing flag so procs in this transaction have to * start sending their work down. @@ -1184,19 +1349,20 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, cur_trans->delayed_refs.flushing = 1; ret = btrfs_run_delayed_refs(trans, root, 0); - BUG_ON(ret); + if (ret) + goto cleanup_transaction; spin_lock(&cur_trans->commit_lock); if (cur_trans->in_commit) { spin_unlock(&cur_trans->commit_lock); atomic_inc(&cur_trans->use_count); - btrfs_end_transaction(trans, root); + ret = btrfs_end_transaction(trans, root); wait_for_commit(root, cur_trans); put_transaction(cur_trans); - return 0; + return ret; } trans->transaction->in_commit = 1; @@ -1222,7 +1388,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, spin_unlock(&root->fs_info->trans_lock); } - if (now < cur_trans->start_time || now - cur_trans->start_time < 1) + if (!btrfs_test_opt(root, SSD) && + (now < cur_trans->start_time || now - cur_trans->start_time < 1)) should_grow = 1; do { @@ -1236,12 +1403,19 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, if (flush_on_commit || snap_pending) { btrfs_start_delalloc_inodes(root, 1); - ret = btrfs_wait_ordered_extents(root, 0, 1); - BUG_ON(ret); + btrfs_wait_ordered_extents(root, 0, 1); } ret = btrfs_run_delayed_items(trans, root); - BUG_ON(ret); + if (ret) + goto cleanup_transaction; + + /* + * running the delayed items may have added new refs. account + * them now so that they hinder processing of more delayed refs + * as little as possible. + */ + btrfs_delayed_refs_qgroup_accounting(trans, root->fs_info); /* * rename don't use btrfs_join_transaction, so, once we @@ -1283,13 +1457,22 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, mutex_lock(&root->fs_info->reloc_mutex); ret = btrfs_run_delayed_items(trans, root); - BUG_ON(ret); + if (ret) { + mutex_unlock(&root->fs_info->reloc_mutex); + goto cleanup_transaction; + } ret = create_pending_snapshots(trans, root->fs_info); - BUG_ON(ret); + if (ret) { + mutex_unlock(&root->fs_info->reloc_mutex); + goto cleanup_transaction; + } ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); - BUG_ON(ret); + if (ret) { + mutex_unlock(&root->fs_info->reloc_mutex); + goto cleanup_transaction; + } /* * make sure none of the code above managed to slip in a @@ -1316,7 +1499,11 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, mutex_lock(&root->fs_info->tree_log_mutex); ret = commit_fs_roots(trans, root); - BUG_ON(ret); + if (ret) { + mutex_unlock(&root->fs_info->tree_log_mutex); + mutex_unlock(&root->fs_info->reloc_mutex); + goto cleanup_transaction; + } /* commit_fs_roots gets rid of all the tree log roots, it is now * safe to free the root of tree log roots @@ -1324,7 +1511,11 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, btrfs_free_log_root_tree(trans, root->fs_info); ret = commit_cowonly_roots(trans, root); - BUG_ON(ret); + if (ret) { + mutex_unlock(&root->fs_info->tree_log_mutex); + mutex_unlock(&root->fs_info->reloc_mutex); + goto cleanup_transaction; + } btrfs_prepare_extent_commit(trans, root); @@ -1338,15 +1529,16 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, root->fs_info->chunk_root->node); switch_commit_root(root->fs_info->chunk_root); + assert_qgroups_uptodate(trans); update_super_roots(root); if (!root->fs_info->log_root_recovering) { - btrfs_set_super_log_root(&root->fs_info->super_copy, 0); - btrfs_set_super_log_root_level(&root->fs_info->super_copy, 0); + btrfs_set_super_log_root(root->fs_info->super_copy, 0); + btrfs_set_super_log_root_level(root->fs_info->super_copy, 0); } - memcpy(&root->fs_info->super_for_commit, &root->fs_info->super_copy, - sizeof(root->fs_info->super_copy)); + memcpy(root->fs_info->super_for_commit, root->fs_info->super_copy, + sizeof(*root->fs_info->super_copy)); trans->transaction->blocked = 0; spin_lock(&root->fs_info->trans_lock); @@ -1358,8 +1550,18 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, wake_up(&root->fs_info->transaction_wait); ret = btrfs_write_and_wait_transaction(trans, root); - BUG_ON(ret); - write_ctree_super(trans, root, 0); + if (ret) { + btrfs_error(root->fs_info, ret, + "Error while writing out transaction."); + mutex_unlock(&root->fs_info->tree_log_mutex); + goto cleanup_transaction; + } + + ret = write_ctree_super(trans, root, 0); + if (ret) { + mutex_unlock(&root->fs_info->tree_log_mutex); + goto cleanup_transaction; + } /* * the super is written, we can safely allow the tree-loggers @@ -1382,6 +1584,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, put_transaction(cur_trans); put_transaction(cur_trans); + sb_end_intwrite(root->fs_info->sb); + trace_btrfs_transaction_commit(root); btrfs_scrub_continue(root); @@ -1395,6 +1599,17 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, btrfs_run_delayed_iputs(root); return ret; + +cleanup_transaction: + btrfs_trans_release_metadata(trans, root); + trans->block_rsv = NULL; + btrfs_printk(root->fs_info, "Skipping commit of aborted transaction.\n"); +// WARN_ON(1); + if (current->journal_info == trans) + current->journal_info = NULL; + cleanup_transaction(trans, root, ret); + + return ret; } /* @@ -1410,6 +1625,8 @@ int btrfs_clean_old_snapshots(struct btrfs_root *root) spin_unlock(&fs_info->trans_lock); while (!list_empty(&list)) { + int ret; + root = list_entry(list.next, struct btrfs_root, root_list); list_del(&root->root_list); @@ -1417,9 +1634,10 @@ int btrfs_clean_old_snapshots(struct btrfs_root *root) if (btrfs_header_backref_rev(root->node) < BTRFS_MIXED_BACKREF_REV) - btrfs_drop_snapshot(root, NULL, 0); + ret = btrfs_drop_snapshot(root, NULL, 0, 0); else - btrfs_drop_snapshot(root, NULL, 1); + ret =btrfs_drop_snapshot(root, NULL, 1, 0); + BUG_ON(ret < 0); } return 0; } diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h index 02564e6230ac..e8b8416c688b 100644 --- a/fs/btrfs/transaction.h +++ b/fs/btrfs/transaction.h @@ -20,6 +20,7 @@ #define __BTRFS_TRANSACTION__ #include "btrfs_inode.h" #include "delayed-ref.h" +#include "ctree.h" struct btrfs_transaction { u64 transid; @@ -43,11 +44,13 @@ struct btrfs_transaction { wait_queue_head_t commit_wait; struct list_head pending_snapshots; struct btrfs_delayed_ref_root delayed_refs; + int aborted; }; struct btrfs_trans_handle { u64 transid; u64 bytes_reserved; + u64 qgroup_reserved; unsigned long use_count; unsigned long blocks_reserved; unsigned long blocks_used; @@ -55,12 +58,23 @@ struct btrfs_trans_handle { struct btrfs_transaction *transaction; struct btrfs_block_rsv *block_rsv; struct btrfs_block_rsv *orig_rsv; + int aborted; + int adding_csums; + /* + * this root is only needed to validate that the root passed to + * start_transaction is the same as the one passed to end_transaction. + * Subvolume quota depends on this + */ + struct btrfs_root *root; + struct seq_list delayed_ref_elem; + struct list_head qgroup_ref_list; }; struct btrfs_pending_snapshot { struct dentry *dentry; struct btrfs_root *root; struct btrfs_root *snap; + struct btrfs_qgroup_inherit *inherit; /* block reservation for the operation */ struct btrfs_block_rsv block_rsv; /* extra metadata reseration for relocation */ @@ -114,4 +128,5 @@ int btrfs_wait_marked_extents(struct btrfs_root *root, struct extent_io_tree *dirty_pages, int mark); int btrfs_transaction_blocked(struct btrfs_fs_info *info); int btrfs_transaction_in_commit(struct btrfs_fs_info *info); +void put_transaction(struct btrfs_transaction *transaction); #endif diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 786639fca067..c86670f4f285 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -212,14 +212,13 @@ int btrfs_pin_log_trans(struct btrfs_root *root) * indicate we're done making changes to the log tree * and wake up anyone waiting to do a sync */ -int btrfs_end_log_trans(struct btrfs_root *root) +void btrfs_end_log_trans(struct btrfs_root *root) { if (atomic_dec_and_test(&root->log_writers)) { smp_mb(); if (waitqueue_active(&root->log_writer_wait)) wake_up(&root->log_writer_wait); } - return 0; } @@ -276,10 +275,11 @@ static int process_one_buffer(struct btrfs_root *log, struct walk_control *wc, u64 gen) { if (wc->pin) - btrfs_pin_extent(log->fs_info->extent_root, - eb->start, eb->len, 0); + btrfs_pin_extent_for_log_replay(wc->trans, + log->fs_info->extent_root, + eb->start, eb->len); - if (btrfs_buffer_uptodate(eb, gen)) { + if (btrfs_buffer_uptodate(eb, gen, 0)) { if (wc->write) btrfs_write_tree_block(eb); if (wc->wait) @@ -377,12 +377,11 @@ insert: u32 found_size; found_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]); - if (found_size > item_size) { + if (found_size > item_size) btrfs_truncate_item(trans, root, path, item_size, 1); - } else if (found_size < item_size) { - ret = btrfs_extend_item(trans, root, path, - item_size - found_size); - } + else if (found_size < item_size) + btrfs_extend_item(trans, root, path, + item_size - found_size); } else if (ret) { return ret; } @@ -588,7 +587,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, ret = btrfs_inc_extent_ref(trans, root, ins.objectid, ins.offset, 0, root->root_key.objectid, - key->objectid, offset); + key->objectid, offset, 0); BUG_ON(ret); } else { /* @@ -638,7 +637,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, } inode_set_bytes(inode, saved_nbytes); - btrfs_update_inode(trans, root, inode); + ret = btrfs_update_inode(trans, root, inode); out: if (inode) iput(inode); @@ -691,6 +690,8 @@ static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans, kfree(name); iput(inode); + + btrfs_run_delayed_items(trans, root); return ret; } @@ -896,6 +897,7 @@ again: ret = btrfs_unlink_inode(trans, root, dir, inode, victim_name, victim_name_len); + btrfs_run_delayed_items(trans, root); } kfree(victim_name); ptr = (unsigned long)(victim_ref + 1) + victim_name_len; @@ -1030,7 +1032,7 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, } btrfs_release_path(path); if (nlink != inode->i_nlink) { - inode->i_nlink = nlink; + set_nlink(inode, nlink); btrfs_update_inode(trans, root, inode); } BTRFS_I(inode)->index_cnt = (u64)-1; @@ -1131,7 +1133,7 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans, btrfs_release_path(path); if (ret == 0) { btrfs_inc_nlink(inode); - btrfs_update_inode(trans, root, inode); + ret = btrfs_update_inode(trans, root, inode); } else if (ret == -EEXIST) { ret = 0; } else { @@ -1476,6 +1478,9 @@ again: ret = btrfs_unlink_inode(trans, root, dir, inode, name, name_len); BUG_ON(ret); + + btrfs_run_delayed_items(trans, root); + kfree(name); iput(inode); @@ -1629,7 +1634,9 @@ static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb, int i; int ret; - btrfs_read_buffer(eb, gen); + ret = btrfs_read_buffer(eb, gen); + if (ret) + return ret; level = btrfs_header_level(eb); @@ -1750,7 +1757,11 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans, path->slots[*level]++; if (wc->free) { - btrfs_read_buffer(next, ptr_gen); + ret = btrfs_read_buffer(next, ptr_gen); + if (ret) { + free_extent_buffer(next); + return ret; + } btrfs_tree_lock(next); btrfs_set_lock_blocking(next); @@ -1760,14 +1771,18 @@ static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans, WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID); - ret = btrfs_free_reserved_extent(root, + ret = btrfs_free_and_pin_reserved_extent(root, bytenr, blocksize); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM or logic errors */ } free_extent_buffer(next); continue; } - btrfs_read_buffer(next, ptr_gen); + ret = btrfs_read_buffer(next, ptr_gen); + if (ret) { + free_extent_buffer(next); + return ret; + } WARN_ON(*level <= 0); if (path->nodes[*level-1]) @@ -1828,7 +1843,7 @@ static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans, btrfs_tree_unlock(next); WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID); - ret = btrfs_free_reserved_extent(root, + ret = btrfs_free_and_pin_reserved_extent(root, path->nodes[*level]->start, path->nodes[*level]->len); BUG_ON(ret); @@ -1870,20 +1885,26 @@ static int walk_log_tree(struct btrfs_trans_handle *trans, wret = walk_down_log_tree(trans, log, path, &level, wc); if (wret > 0) break; - if (wret < 0) + if (wret < 0) { ret = wret; + goto out; + } wret = walk_up_log_tree(trans, log, path, &level, wc); if (wret > 0) break; - if (wret < 0) + if (wret < 0) { ret = wret; + goto out; + } } /* was the root node processed? if not, catch it here */ if (path->nodes[orig_level]) { - wc->process_func(log, path->nodes[orig_level], wc, + ret = wc->process_func(log, path->nodes[orig_level], wc, btrfs_header_generation(path->nodes[orig_level])); + if (ret) + goto out; if (wc->free) { struct extent_buffer *next; @@ -1897,12 +1918,13 @@ static int walk_log_tree(struct btrfs_trans_handle *trans, WARN_ON(log->root_key.objectid != BTRFS_TREE_LOG_OBJECTID); - ret = btrfs_free_reserved_extent(log, next->start, + ret = btrfs_free_and_pin_reserved_extent(log, next->start, next->len); - BUG_ON(ret); + BUG_ON(ret); /* -ENOMEM or logic errors */ } } +out: for (i = 0; i <= orig_level; i++) { if (path->nodes[i]) { free_extent_buffer(path->nodes[i]); @@ -1956,16 +1978,18 @@ static int wait_log_commit(struct btrfs_trans_handle *trans, finish_wait(&root->log_commit_wait[index], &wait); mutex_lock(&root->log_mutex); - } while (root->log_transid < transid + 2 && + } while (root->fs_info->last_trans_log_full_commit != + trans->transid && root->log_transid < transid + 2 && atomic_read(&root->log_commit[index])); return 0; } -static int wait_for_writer(struct btrfs_trans_handle *trans, - struct btrfs_root *root) +static void wait_for_writer(struct btrfs_trans_handle *trans, + struct btrfs_root *root) { DEFINE_WAIT(wait); - while (atomic_read(&root->log_writers)) { + while (root->fs_info->last_trans_log_full_commit != + trans->transid && atomic_read(&root->log_writers)) { prepare_to_wait(&root->log_writer_wait, &wait, TASK_UNINTERRUPTIBLE); mutex_unlock(&root->log_mutex); @@ -1975,7 +1999,6 @@ static int wait_for_writer(struct btrfs_trans_handle *trans, mutex_lock(&root->log_mutex); finish_wait(&root->log_writer_wait, &wait); } - return 0; } /* @@ -2013,10 +2036,10 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, /* wait for previous tree log sync to complete */ if (atomic_read(&root->log_commit[(index1 + 1) % 2])) wait_log_commit(trans, root, root->log_transid - 1); - while (1) { unsigned long batch = root->log_batch; - if (root->log_multiple_pids) { + /* when we're on an ssd, just kick the log commit out */ + if (!btrfs_test_opt(root, SSD) && root->log_multiple_pids) { mutex_unlock(&root->log_mutex); schedule_timeout_uninterruptible(1); mutex_lock(&root->log_mutex); @@ -2043,7 +2066,11 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, * wait for them until later. */ ret = btrfs_write_marked_extents(log, &log->dirty_log_pages, mark); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + mutex_unlock(&root->log_mutex); + goto out; + } btrfs_set_root_node(&log->root_item, log->node); @@ -2074,7 +2101,11 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, } if (ret) { - BUG_ON(ret != -ENOSPC); + if (ret != -ENOSPC) { + btrfs_abort_transaction(trans, root, ret); + mutex_unlock(&log_root_tree->log_mutex); + goto out; + } root->fs_info->last_trans_log_full_commit = trans->transid; btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark); mutex_unlock(&log_root_tree->log_mutex); @@ -2114,12 +2145,16 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, ret = btrfs_write_and_wait_marked_extents(log_root_tree, &log_root_tree->dirty_log_pages, EXTENT_DIRTY | EXTENT_NEW); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, root, ret); + mutex_unlock(&log_root_tree->log_mutex); + goto out_wake_log_root; + } btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark); - btrfs_set_super_log_root(&root->fs_info->super_for_commit, + btrfs_set_super_log_root(root->fs_info->super_for_commit, log_root_tree->node->start); - btrfs_set_super_log_root_level(&root->fs_info->super_for_commit, + btrfs_set_super_log_root_level(root->fs_info->super_for_commit, btrfs_header_level(log_root_tree->node)); log_root_tree->log_batch = 0; @@ -2323,7 +2358,9 @@ out_unlock: if (ret == -ENOSPC) { root->fs_info->last_trans_log_full_commit = trans->transid; ret = 0; - } + } else if (ret < 0) + btrfs_abort_transaction(trans, root, ret); + btrfs_end_log_trans(root); return err; @@ -2354,7 +2391,8 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans, if (ret == -ENOSPC) { root->fs_info->last_trans_log_full_commit = trans->transid; ret = 0; - } + } else if (ret < 0 && ret != -ENOENT) + btrfs_abort_transaction(trans, root, ret); btrfs_end_log_trans(root); return ret; @@ -2635,6 +2673,8 @@ static int drop_objectid_items(struct btrfs_trans_handle *trans, btrfs_release_path(path); } btrfs_release_path(path); + if (ret > 0) + ret = 0; return ret; } @@ -3006,21 +3046,6 @@ out: return ret; } -static int inode_in_log(struct btrfs_trans_handle *trans, - struct inode *inode) -{ - struct btrfs_root *root = BTRFS_I(inode)->root; - int ret = 0; - - mutex_lock(&root->log_mutex); - if (BTRFS_I(inode)->logged_trans == trans->transid && - BTRFS_I(inode)->last_sub_trans <= root->last_log_commit) - ret = 1; - mutex_unlock(&root->log_mutex); - return ret; -} - - /* * helper function around btrfs_log_inode to make sure newly created * parent directories also end up in the log. A minimal inode and backref @@ -3061,7 +3086,7 @@ int btrfs_log_inode_parent(struct btrfs_trans_handle *trans, if (ret) goto end_no_trans; - if (inode_in_log(trans, inode)) { + if (btrfs_inode_in_log(inode, trans->transid)) { ret = BTRFS_NO_LOG_SYNC; goto end_no_trans; } @@ -3166,13 +3191,20 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree) fs_info->log_root_recovering = 1; trans = btrfs_start_transaction(fs_info->tree_root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto error; + } wc.trans = trans; wc.pin = 1; ret = walk_log_tree(trans, log_root_tree, &wc); - BUG_ON(ret); + if (ret) { + btrfs_error(fs_info, ret, "Failed to pin buffers while " + "recovering log root tree."); + goto error; + } again: key.objectid = BTRFS_TREE_LOG_OBJECTID; @@ -3181,8 +3213,12 @@ again: while (1) { ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0); - if (ret < 0) - break; + + if (ret < 0) { + btrfs_error(fs_info, ret, + "Couldn't find tree log root."); + goto error; + } if (ret > 0) { if (path->slots[0] == 0) break; @@ -3196,14 +3232,24 @@ again: log = btrfs_read_fs_root_no_radix(log_root_tree, &found_key); - BUG_ON(IS_ERR(log)); + if (IS_ERR(log)) { + ret = PTR_ERR(log); + btrfs_error(fs_info, ret, + "Couldn't read tree log root."); + goto error; + } tmp_key.objectid = found_key.offset; tmp_key.type = BTRFS_ROOT_ITEM_KEY; tmp_key.offset = (u64)-1; wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key); - BUG_ON(IS_ERR_OR_NULL(wc.replay_dest)); + if (IS_ERR(wc.replay_dest)) { + ret = PTR_ERR(wc.replay_dest); + btrfs_error(fs_info, ret, "Couldn't read target root " + "for tree log recovery."); + goto error; + } wc.replay_dest->log_root = log; btrfs_record_root_in_trans(trans, wc.replay_dest); @@ -3251,6 +3297,10 @@ again: kfree(log_root_tree); return 0; + +error: + btrfs_free_path(path); + return ret; } /* diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h index 2270ac58d746..862ac813f6b8 100644 --- a/fs/btrfs/tree-log.h +++ b/fs/btrfs/tree-log.h @@ -38,7 +38,7 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, int name_len, struct inode *inode, u64 dirid); -int btrfs_end_log_trans(struct btrfs_root *root); +void btrfs_end_log_trans(struct btrfs_root *root); int btrfs_pin_log_trans(struct btrfs_root *root); int btrfs_log_inode_parent(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode, diff --git a/fs/btrfs/ulist.c b/fs/btrfs/ulist.c new file mode 100644 index 000000000000..ab942f46b3dd --- /dev/null +++ b/fs/btrfs/ulist.c @@ -0,0 +1,222 @@ +/* + * Copyright (C) 2011 STRATO AG + * written by Arne Jansen <sensille@gmx.net> + * Distributed under the GNU GPL license version 2. + */ + +#include <linux/slab.h> +#include <linux/module.h> +#include "ulist.h" + +/* + * ulist is a generic data structure to hold a collection of unique u64 + * values. The only operations it supports is adding to the list and + * enumerating it. + * It is possible to store an auxiliary value along with the key. + * + * The implementation is preliminary and can probably be sped up + * significantly. A first step would be to store the values in an rbtree + * as soon as ULIST_SIZE is exceeded. + * + * A sample usage for ulists is the enumeration of directed graphs without + * visiting a node twice. The pseudo-code could look like this: + * + * ulist = ulist_alloc(); + * ulist_add(ulist, root); + * ULIST_ITER_INIT(&uiter); + * + * while ((elem = ulist_next(ulist, &uiter)) { + * for (all child nodes n in elem) + * ulist_add(ulist, n); + * do something useful with the node; + * } + * ulist_free(ulist); + * + * This assumes the graph nodes are adressable by u64. This stems from the + * usage for tree enumeration in btrfs, where the logical addresses are + * 64 bit. + * + * It is also useful for tree enumeration which could be done elegantly + * recursively, but is not possible due to kernel stack limitations. The + * loop would be similar to the above. + */ + +/** + * ulist_init - freshly initialize a ulist + * @ulist: the ulist to initialize + * + * Note: don't use this function to init an already used ulist, use + * ulist_reinit instead. + */ +void ulist_init(struct ulist *ulist) +{ + ulist->nnodes = 0; + ulist->nodes = ulist->int_nodes; + ulist->nodes_alloced = ULIST_SIZE; +} +EXPORT_SYMBOL(ulist_init); + +/** + * ulist_fini - free up additionally allocated memory for the ulist + * @ulist: the ulist from which to free the additional memory + * + * This is useful in cases where the base 'struct ulist' has been statically + * allocated. + */ +void ulist_fini(struct ulist *ulist) +{ + /* + * The first ULIST_SIZE elements are stored inline in struct ulist. + * Only if more elements are alocated they need to be freed. + */ + if (ulist->nodes_alloced > ULIST_SIZE) + kfree(ulist->nodes); + ulist->nodes_alloced = 0; /* in case ulist_fini is called twice */ +} +EXPORT_SYMBOL(ulist_fini); + +/** + * ulist_reinit - prepare a ulist for reuse + * @ulist: ulist to be reused + * + * Free up all additional memory allocated for the list elements and reinit + * the ulist. + */ +void ulist_reinit(struct ulist *ulist) +{ + ulist_fini(ulist); + ulist_init(ulist); +} +EXPORT_SYMBOL(ulist_reinit); + +/** + * ulist_alloc - dynamically allocate a ulist + * @gfp_mask: allocation flags to for base allocation + * + * The allocated ulist will be returned in an initialized state. + */ +struct ulist *ulist_alloc(gfp_t gfp_mask) +{ + struct ulist *ulist = kmalloc(sizeof(*ulist), gfp_mask); + + if (!ulist) + return NULL; + + ulist_init(ulist); + + return ulist; +} +EXPORT_SYMBOL(ulist_alloc); + +/** + * ulist_free - free dynamically allocated ulist + * @ulist: ulist to free + * + * It is not necessary to call ulist_fini before. + */ +void ulist_free(struct ulist *ulist) +{ + if (!ulist) + return; + ulist_fini(ulist); + kfree(ulist); +} +EXPORT_SYMBOL(ulist_free); + +/** + * ulist_add - add an element to the ulist + * @ulist: ulist to add the element to + * @val: value to add to ulist + * @aux: auxiliary value to store along with val + * @gfp_mask: flags to use for allocation + * + * Note: locking must be provided by the caller. In case of rwlocks write + * locking is needed + * + * Add an element to a ulist. The @val will only be added if it doesn't + * already exist. If it is added, the auxiliary value @aux is stored along with + * it. In case @val already exists in the ulist, @aux is ignored, even if + * it differs from the already stored value. + * + * ulist_add returns 0 if @val already exists in ulist and 1 if @val has been + * inserted. + * In case of allocation failure -ENOMEM is returned and the ulist stays + * unaltered. + */ +int ulist_add(struct ulist *ulist, u64 val, unsigned long aux, + gfp_t gfp_mask) +{ + return ulist_add_merge(ulist, val, aux, NULL, gfp_mask); +} + +int ulist_add_merge(struct ulist *ulist, u64 val, unsigned long aux, + unsigned long *old_aux, gfp_t gfp_mask) +{ + int i; + + for (i = 0; i < ulist->nnodes; ++i) { + if (ulist->nodes[i].val == val) { + if (old_aux) + *old_aux = ulist->nodes[i].aux; + return 0; + } + } + + if (ulist->nnodes >= ulist->nodes_alloced) { + u64 new_alloced = ulist->nodes_alloced + 128; + struct ulist_node *new_nodes; + void *old = NULL; + + /* + * if nodes_alloced == ULIST_SIZE no memory has been allocated + * yet, so pass NULL to krealloc + */ + if (ulist->nodes_alloced > ULIST_SIZE) + old = ulist->nodes; + + new_nodes = krealloc(old, sizeof(*new_nodes) * new_alloced, + gfp_mask); + if (!new_nodes) + return -ENOMEM; + + if (!old) + memcpy(new_nodes, ulist->int_nodes, + sizeof(ulist->int_nodes)); + + ulist->nodes = new_nodes; + ulist->nodes_alloced = new_alloced; + } + ulist->nodes[ulist->nnodes].val = val; + ulist->nodes[ulist->nnodes].aux = aux; + ++ulist->nnodes; + + return 1; +} +EXPORT_SYMBOL(ulist_add); + +/** + * ulist_next - iterate ulist + * @ulist: ulist to iterate + * @uiter: iterator variable, initialized with ULIST_ITER_INIT(&iterator) + * + * Note: locking must be provided by the caller. In case of rwlocks only read + * locking is needed + * + * This function is used to iterate an ulist. + * It returns the next element from the ulist or %NULL when the + * end is reached. No guarantee is made with respect to the order in which + * the elements are returned. They might neither be returned in order of + * addition nor in ascending order. + * It is allowed to call ulist_add during an enumeration. Newly added items + * are guaranteed to show up in the running enumeration. + */ +struct ulist_node *ulist_next(struct ulist *ulist, struct ulist_iterator *uiter) +{ + if (ulist->nnodes == 0) + return NULL; + if (uiter->i < 0 || uiter->i >= ulist->nnodes) + return NULL; + + return &ulist->nodes[uiter->i++]; +} +EXPORT_SYMBOL(ulist_next); diff --git a/fs/btrfs/ulist.h b/fs/btrfs/ulist.h new file mode 100644 index 000000000000..21bdc8ec8130 --- /dev/null +++ b/fs/btrfs/ulist.h @@ -0,0 +1,77 @@ +/* + * Copyright (C) 2011 STRATO AG + * written by Arne Jansen <sensille@gmx.net> + * Distributed under the GNU GPL license version 2. + * + */ + +#ifndef __ULIST__ +#define __ULIST__ + +/* + * ulist is a generic data structure to hold a collection of unique u64 + * values. The only operations it supports is adding to the list and + * enumerating it. + * It is possible to store an auxiliary value along with the key. + * + * The implementation is preliminary and can probably be sped up + * significantly. A first step would be to store the values in an rbtree + * as soon as ULIST_SIZE is exceeded. + */ + +/* + * number of elements statically allocated inside struct ulist + */ +#define ULIST_SIZE 16 + +struct ulist_iterator { + int i; +}; + +/* + * element of the list + */ +struct ulist_node { + u64 val; /* value to store */ + unsigned long aux; /* auxiliary value saved along with the val */ +}; + +struct ulist { + /* + * number of elements stored in list + */ + unsigned long nnodes; + + /* + * number of nodes we already have room for + */ + unsigned long nodes_alloced; + + /* + * pointer to the array storing the elements. The first ULIST_SIZE + * elements are stored inline. In this case the it points to int_nodes. + * After exceeding ULIST_SIZE, dynamic memory is allocated. + */ + struct ulist_node *nodes; + + /* + * inline storage space for the first ULIST_SIZE entries + */ + struct ulist_node int_nodes[ULIST_SIZE]; +}; + +void ulist_init(struct ulist *ulist); +void ulist_fini(struct ulist *ulist); +void ulist_reinit(struct ulist *ulist); +struct ulist *ulist_alloc(gfp_t gfp_mask); +void ulist_free(struct ulist *ulist); +int ulist_add(struct ulist *ulist, u64 val, unsigned long aux, + gfp_t gfp_mask); +int ulist_add_merge(struct ulist *ulist, u64 val, unsigned long aux, + unsigned long *old_aux, gfp_t gfp_mask); +struct ulist_node *ulist_next(struct ulist *ulist, + struct ulist_iterator *uiter); + +#define ULIST_ITER_INIT(uiter) ((uiter)->i = 0) + +#endif diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index f2a4cc79da61..88b969aeeb71 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -23,6 +23,8 @@ #include <linux/random.h> #include <linux/iocontext.h> #include <linux/capability.h> +#include <linux/ratelimit.h> +#include <linux/kthread.h> #include <asm/div64.h> #include "compat.h" #include "ctree.h" @@ -32,11 +34,15 @@ #include "print-tree.h" #include "volumes.h" #include "async-thread.h" +#include "check-integrity.h" +#include "rcu-string.h" static int init_first_rw_device(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_device *device); static int btrfs_relocate_sys_chunks(struct btrfs_root *root); +static void __btrfs_reset_dev_stats(struct btrfs_device *dev); +static void btrfs_dev_stat_print_on_load(struct btrfs_device *device); static DEFINE_MUTEX(uuid_mutex); static LIST_HEAD(fs_uuids); @@ -59,13 +65,13 @@ static void free_fs_devices(struct btrfs_fs_devices *fs_devices) device = list_entry(fs_devices->devices.next, struct btrfs_device, dev_list); list_del(&device->dev_list); - kfree(device->name); + rcu_string_free(device->name); kfree(device); } kfree(fs_devices); } -int btrfs_cleanup_fs_uuids(void) +void btrfs_cleanup_fs_uuids(void) { struct btrfs_fs_devices *fs_devices; @@ -75,7 +81,6 @@ int btrfs_cleanup_fs_uuids(void) list_del(&fs_devices->list); free_fs_devices(fs_devices); } - return 0; } static noinline struct btrfs_device *__find_device(struct list_head *head, @@ -128,7 +133,7 @@ static void requeue_list(struct btrfs_pending_bios *pending_bios, * the list if the block device is congested. This way, multiple devices * can make progress from a single worker thread. */ -static noinline int run_scheduled_bios(struct btrfs_device *device) +static noinline void run_scheduled_bios(struct btrfs_device *device) { struct bio *pending; struct backing_dev_info *bdi; @@ -222,9 +227,8 @@ loop_lock: cur = pending; pending = pending->bi_next; cur->bi_next = NULL; - atomic_dec(&fs_info->nr_async_bios); - if (atomic_read(&fs_info->nr_async_bios) < limit && + if (atomic_dec_return(&fs_info->nr_async_bios) < limit && waitqueue_active(&fs_info->async_submit_wait)) wake_up(&fs_info->async_submit_wait); @@ -246,7 +250,7 @@ loop_lock: sync_pending = 0; } - submit_bio(cur->bi_rw, cur); + btrfsic_submit_bio(cur->bi_rw, cur); num_run++; batch_run++; if (need_resched()) @@ -295,6 +299,12 @@ loop_lock: btrfs_requeue_work(&device->work); goto done; } + /* unplug every 64 requests just for good measure */ + if (batch_run % 64 == 0) { + blk_finish_plug(&plug); + blk_start_plug(&plug); + sync_pending = 0; + } } cond_resched(); @@ -308,7 +318,6 @@ loop_lock: done: blk_finish_plug(&plug); - return 0; } static void pending_bios_fn(struct btrfs_work *work) @@ -325,8 +334,8 @@ static noinline int device_list_add(const char *path, { struct btrfs_device *device; struct btrfs_fs_devices *fs_devices; + struct rcu_string *name; u64 found_transid = btrfs_super_generation(disk_super); - char *name; fs_devices = find_fsid(disk_super->fsid); if (!fs_devices) { @@ -355,29 +364,40 @@ static noinline int device_list_add(const char *path, return -ENOMEM; } device->devid = devid; + device->dev_stats_valid = 0; device->work.func = pending_bios_fn; memcpy(device->uuid, disk_super->dev_item.uuid, BTRFS_UUID_SIZE); spin_lock_init(&device->io_lock); - device->name = kstrdup(path, GFP_NOFS); - if (!device->name) { + + name = rcu_string_strdup(path, GFP_NOFS); + if (!name) { kfree(device); return -ENOMEM; } + rcu_assign_pointer(device->name, name); INIT_LIST_HEAD(&device->dev_alloc_list); + /* init readahead state */ + spin_lock_init(&device->reada_lock); + device->reada_curr_zone = NULL; + atomic_set(&device->reada_in_flight, 0); + device->reada_next = 0; + INIT_RADIX_TREE(&device->reada_zones, GFP_NOFS & ~__GFP_WAIT); + INIT_RADIX_TREE(&device->reada_extents, GFP_NOFS & ~__GFP_WAIT); + mutex_lock(&fs_devices->device_list_mutex); list_add_rcu(&device->dev_list, &fs_devices->devices); mutex_unlock(&fs_devices->device_list_mutex); device->fs_devices = fs_devices; fs_devices->num_devices++; - } else if (!device->name || strcmp(device->name, path)) { - name = kstrdup(path, GFP_NOFS); + } else if (!device->name || strcmp(device->name->str, path)) { + name = rcu_string_strdup(path, GFP_NOFS); if (!name) return -ENOMEM; - kfree(device->name); - device->name = name; + rcu_string_free(device->name); + rcu_assign_pointer(device->name, name); if (device->missing) { fs_devices->missing_devices--; device->missing = 0; @@ -408,19 +428,27 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig) mutex_init(&fs_devices->device_list_mutex); fs_devices->latest_devid = orig->latest_devid; fs_devices->latest_trans = orig->latest_trans; + fs_devices->total_devices = orig->total_devices; memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid)); /* We have held the volume lock, it is safe to get the devices. */ list_for_each_entry(orig_dev, &orig->devices, dev_list) { + struct rcu_string *name; + device = kzalloc(sizeof(*device), GFP_NOFS); if (!device) goto error; - device->name = kstrdup(orig_dev->name, GFP_NOFS); - if (!device->name) { + /* + * This is ok to do without rcu read locked because we hold the + * uuid mutex so nothing we touch in here is going to disappear. + */ + name = rcu_string_strdup(orig_dev->name->str, GFP_NOFS); + if (!name) { kfree(device); goto error; } + rcu_assign_pointer(device->name, name); device->devid = orig_dev->devid; device->work.func = pending_bios_fn; @@ -439,16 +467,27 @@ error: return ERR_PTR(-ENOMEM); } -int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices) +void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices) { struct btrfs_device *device, *next; + struct block_device *latest_bdev = NULL; + u64 latest_devid = 0; + u64 latest_transid = 0; + mutex_lock(&uuid_mutex); again: /* This is the initialized path, it is safe to release the devices. */ list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) { - if (device->in_fs_metadata) + if (device->in_fs_metadata) { + if (!latest_transid || + device->generation > latest_transid) { + latest_devid = device->devid; + latest_transid = device->generation; + latest_bdev = device->bdev; + } continue; + } if (device->bdev) { blkdev_put(device->bdev, device->mode); @@ -462,7 +501,7 @@ again: } list_del_init(&device->dev_list); fs_devices->num_devices--; - kfree(device->name); + rcu_string_free(device->name); kfree(device); } @@ -471,8 +510,11 @@ again: goto again; } + fs_devices->latest_bdev = latest_bdev; + fs_devices->latest_devid = latest_devid; + fs_devices->latest_trans = latest_transid; + mutex_unlock(&uuid_mutex); - return 0; } static void __free_device(struct work_struct *work) @@ -484,7 +526,7 @@ static void __free_device(struct work_struct *work) if (device->bdev) blkdev_put(device->bdev, device->mode); - kfree(device->name); + rcu_string_free(device->name); kfree(device); } @@ -508,6 +550,7 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) mutex_lock(&fs_devices->device_list_mutex); list_for_each_entry(device, &fs_devices->devices, dev_list) { struct btrfs_device *new_device; + struct rcu_string *name; if (device->bdev) fs_devices->open_devices--; @@ -521,10 +564,15 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) fs_devices->num_can_discard--; new_device = kmalloc(sizeof(*new_device), GFP_NOFS); - BUG_ON(!new_device); + BUG_ON(!new_device); /* -ENOMEM */ memcpy(new_device, device, sizeof(*new_device)); - new_device->name = kstrdup(device->name, GFP_NOFS); - BUG_ON(device->name && !new_device->name); + + /* Safe because we are under uuid_mutex */ + if (device->name) { + name = rcu_string_strdup(device->name->str, GFP_NOFS); + BUG_ON(device->name && !name); /* -ENOMEM */ + rcu_assign_pointer(new_device->name, name); + } new_device->bdev = NULL; new_device->writeable = 0; new_device->in_fs_metadata = 0; @@ -589,18 +637,18 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, if (!device->name) continue; - bdev = blkdev_get_by_path(device->name, flags, holder); + bdev = blkdev_get_by_path(device->name->str, flags, holder); if (IS_ERR(bdev)) { - printk(KERN_INFO "open %s failed\n", device->name); + printk(KERN_INFO "open %s failed\n", device->name->str); goto error; } + filemap_write_and_wait(bdev->bd_inode->i_mapping); + invalidate_bdev(bdev); set_blocksize(bdev, 4096); bh = btrfs_read_dev_super(bdev); - if (!bh) { - ret = -EINVAL; + if (!bh) goto error_close; - } disk_super = (struct btrfs_super_block *)bh->b_data; devid = btrfs_stack_device_id(&disk_super->dev_item); @@ -655,7 +703,7 @@ error: continue; } if (fs_devices->open_devices == 0) { - ret = -EIO; + ret = -EINVAL; goto out; } fs_devices->seeding = seeding; @@ -693,8 +741,7 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, int ret; u64 devid; u64 transid; - - mutex_lock(&uuid_mutex); + u64 total_devices; flags |= FMODE_EXCL; bdev = blkdev_get_by_path(path, flags, holder); @@ -704,6 +751,7 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, goto error; } + mutex_lock(&uuid_mutex); ret = set_blocksize(bdev, 4096); if (ret) goto error_close; @@ -715,6 +763,7 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, disk_super = (struct btrfs_super_block *)bh->b_data; devid = btrfs_stack_device_id(&disk_super->dev_item); transid = btrfs_super_generation(disk_super); + total_devices = btrfs_super_num_devices(disk_super); if (disk_super->label[0]) printk(KERN_INFO "device label %s ", disk_super->label); else @@ -722,12 +771,13 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, printk(KERN_CONT "devid %llu transid %llu %s\n", (unsigned long long)devid, (unsigned long long)transid, path); ret = device_list_add(path, disk_super, devid, fs_devices_ret); - + if (!ret && fs_devices_ret) + (*fs_devices_ret)->total_devices = total_devices; brelse(bh); error_close: + mutex_unlock(&uuid_mutex); blkdev_put(bdev, flags); error: - mutex_unlock(&uuid_mutex); return ret; } @@ -817,7 +867,6 @@ out: /* * find_free_dev_extent - find free space in the specified device - * @trans: transaction handler * @device: the device which we search the free space in * @num_bytes: the size of the free space that we need * @start: store the start of the free space. @@ -836,8 +885,7 @@ out: * But if we don't find suitable free space, it is used to store the size of * the max free space. */ -int find_free_dev_extent(struct btrfs_trans_handle *trans, - struct btrfs_device *device, u64 num_bytes, +int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes, u64 *start, u64 *len) { struct btrfs_key key; @@ -881,7 +929,7 @@ int find_free_dev_extent(struct btrfs_trans_handle *trans, key.offset = search_start; key.type = BTRFS_DEV_EXTENT_KEY; - ret = btrfs_search_slot(trans, root, &key, path, 0, 0); + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) goto out; if (ret > 0) { @@ -993,7 +1041,7 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, key.objectid = device->devid; key.offset = start; key.type = BTRFS_DEV_EXTENT_KEY; - +again: ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret > 0) { ret = btrfs_previous_item(root, path, key.objectid, @@ -1006,17 +1054,30 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, struct btrfs_dev_extent); BUG_ON(found_key.offset > start || found_key.offset + btrfs_dev_extent_length(leaf, extent) < start); + key = found_key; + btrfs_release_path(path); + goto again; } else if (ret == 0) { leaf = path->nodes[0]; extent = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_extent); + } else { + btrfs_error(root->fs_info, ret, "Slot search failed"); + goto out; } - BUG_ON(ret); - if (device->bytes_used > 0) - device->bytes_used -= btrfs_dev_extent_length(leaf, extent); + if (device->bytes_used > 0) { + u64 len = btrfs_dev_extent_length(leaf, extent); + device->bytes_used -= len; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += len; + spin_unlock(&root->fs_info->free_chunk_lock); + } ret = btrfs_del_item(trans, root, path); - + if (ret) { + btrfs_error(root->fs_info, ret, + "Failed to remove dev extent item"); + } out: btrfs_free_path(path); return ret; @@ -1044,7 +1105,8 @@ int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans, key.type = BTRFS_DEV_EXTENT_KEY; ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*extent)); - BUG_ON(ret); + if (ret) + goto out; leaf = path->nodes[0]; extent = btrfs_item_ptr(leaf, path->slots[0], @@ -1059,6 +1121,7 @@ int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans, btrfs_set_dev_extent_length(leaf, extent, num_bytes); btrfs_mark_buffer_dirty(leaf); +out: btrfs_free_path(path); return ret; } @@ -1084,7 +1147,7 @@ static noinline int find_next_chunk(struct btrfs_root *root, if (ret < 0) goto error; - BUG_ON(ret == 0); + BUG_ON(ret == 0); /* Corruption */ ret = btrfs_previous_item(root, path, 0, BTRFS_CHUNK_ITEM_KEY); if (ret) { @@ -1128,7 +1191,7 @@ static noinline int find_next_devid(struct btrfs_root *root, u64 *objectid) if (ret < 0) goto error; - BUG_ON(ret == 0); + BUG_ON(ret == 0); /* Corruption */ ret = btrfs_previous_item(root, path, BTRFS_DEV_ITEMS_OBJECTID, BTRFS_DEV_ITEM_KEY); @@ -1262,7 +1325,6 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) bool clear_super = false; mutex_lock(&uuid_mutex); - mutex_lock(&root->fs_info->volume_mutex); all_avail = root->fs_info->avail_data_alloc_bits | root->fs_info->avail_system_alloc_bits | @@ -1317,6 +1379,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) } set_blocksize(bdev, 4096); + invalidate_bdev(bdev); bh = btrfs_read_dev_super(bdev); if (!bh) { ret = -EINVAL; @@ -1356,6 +1419,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (ret) goto error_undo; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space = device->total_bytes - + device->bytes_used; + spin_unlock(&root->fs_info->free_chunk_lock); + device->in_fs_metadata = 0; btrfs_scrub_cancel_dev(root, device); @@ -1370,6 +1438,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) list_del_rcu(&device->dev_list); device->fs_devices->num_devices--; + device->fs_devices->total_devices--; if (device->missing) root->fs_info->fs_devices->missing_devices--; @@ -1387,8 +1456,8 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) call_rcu(&device->rcu, free_device); mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); - num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1; - btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices); + num_devices = btrfs_super_num_devices(root->fs_info->super_copy) - 1; + btrfs_set_super_num_devices(root->fs_info->super_copy, num_devices); if (cur_devices->open_devices == 0) { struct btrfs_fs_devices *fs_devices; @@ -1427,7 +1496,6 @@ error_close: if (bdev) blkdev_put(bdev, FMODE_READ | FMODE_EXCL); out: - mutex_unlock(&root->fs_info->volume_mutex); mutex_unlock(&uuid_mutex); return ret; error_undo: @@ -1444,13 +1512,12 @@ error_undo: /* * does all the dirty work required for changing file system's UUID. */ -static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans, - struct btrfs_root *root) +static int btrfs_prepare_sprout(struct btrfs_root *root) { struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; struct btrfs_fs_devices *old_devices; struct btrfs_fs_devices *seed_devices; - struct btrfs_super_block *disk_super = &root->fs_info->super_copy; + struct btrfs_super_block *disk_super = root->fs_info->super_copy; struct btrfs_device *device; u64 super_flags; @@ -1489,6 +1556,7 @@ static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans, fs_devices->seeding = 0; fs_devices->num_devices = 0; fs_devices->open_devices = 0; + fs_devices->total_devices = 0; fs_devices->seed = seed_devices; generate_random_uuid(fs_devices->fsid); @@ -1560,7 +1628,7 @@ next_slot: (unsigned long)btrfs_device_fsid(dev_item), BTRFS_UUID_SIZE); device = btrfs_find_device(root, devid, dev_uuid, fs_uuid); - BUG_ON(!device); + BUG_ON(!device); /* Logic error */ if (device->fs_devices->seeding) { btrfs_set_device_generation(leaf, dev_item, @@ -1585,14 +1653,15 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) struct block_device *bdev; struct list_head *devices; struct super_block *sb = root->fs_info->sb; + struct rcu_string *name; u64 total_bytes; int seeding_dev = 0; int ret = 0; if ((sb->s_flags & MS_RDONLY) && !root->fs_info->fs_devices->seeding) - return -EINVAL; + return -EROFS; - bdev = blkdev_get_by_path(device_path, FMODE_EXCL, + bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL, root->fs_info->bdev_holder); if (IS_ERR(bdev)) return PTR_ERR(bdev); @@ -1604,7 +1673,6 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) } filemap_write_and_wait(bdev->bd_inode->i_mapping); - mutex_lock(&root->fs_info->volume_mutex); devices = &root->fs_info->fs_devices->devices; /* @@ -1625,23 +1693,24 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) goto error; } - device->name = kstrdup(device_path, GFP_NOFS); - if (!device->name) { + name = rcu_string_strdup(device_path, GFP_NOFS); + if (!name) { kfree(device); ret = -ENOMEM; goto error; } + rcu_assign_pointer(device->name, name); ret = find_next_devid(root, &device->devid); if (ret) { - kfree(device->name); + rcu_string_free(device->name); kfree(device); goto error; } trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) { - kfree(device->name); + rcu_string_free(device->name); kfree(device); ret = PTR_ERR(trans); goto error; @@ -1670,16 +1739,12 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) if (seeding_dev) { sb->s_flags &= ~MS_RDONLY; - ret = btrfs_prepare_sprout(trans, root); - BUG_ON(ret); + ret = btrfs_prepare_sprout(root); + BUG_ON(ret); /* -ENOMEM */ } device->fs_devices = root->fs_info->fs_devices; - /* - * we don't want write_supers to jump in here with our device - * half setup - */ mutex_lock(&root->fs_info->fs_devices->device_list_mutex); list_add_rcu(&device->dev_list, &root->fs_info->fs_devices->devices); list_add(&device->dev_alloc_list, @@ -1687,29 +1752,38 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) root->fs_info->fs_devices->num_devices++; root->fs_info->fs_devices->open_devices++; root->fs_info->fs_devices->rw_devices++; + root->fs_info->fs_devices->total_devices++; if (device->can_discard) root->fs_info->fs_devices->num_can_discard++; root->fs_info->fs_devices->total_rw_bytes += device->total_bytes; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += device->total_bytes; + spin_unlock(&root->fs_info->free_chunk_lock); + if (!blk_queue_nonrot(bdev_get_queue(bdev))) root->fs_info->fs_devices->rotating = 1; - total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy); - btrfs_set_super_total_bytes(&root->fs_info->super_copy, + total_bytes = btrfs_super_total_bytes(root->fs_info->super_copy); + btrfs_set_super_total_bytes(root->fs_info->super_copy, total_bytes + device->total_bytes); - total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy); - btrfs_set_super_num_devices(&root->fs_info->super_copy, + total_bytes = btrfs_super_num_devices(root->fs_info->super_copy); + btrfs_set_super_num_devices(root->fs_info->super_copy, total_bytes + 1); mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); if (seeding_dev) { ret = init_first_rw_device(trans, root, device); - BUG_ON(ret); + if (ret) + goto error_trans; ret = btrfs_finish_sprout(trans, root); - BUG_ON(ret); + if (ret) + goto error_trans; } else { ret = btrfs_add_device(trans, root, device); + if (ret) + goto error_trans; } /* @@ -1719,25 +1793,38 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) btrfs_clear_space_info_full(root->fs_info); unlock_chunks(root); - btrfs_commit_transaction(trans, root); + ret = btrfs_commit_transaction(trans, root); if (seeding_dev) { mutex_unlock(&uuid_mutex); up_write(&sb->s_umount); + if (ret) /* transaction commit */ + return ret; + ret = btrfs_relocate_sys_chunks(root); - BUG_ON(ret); + if (ret < 0) + btrfs_error(root->fs_info, ret, + "Failed to relocate sys chunks after " + "device initialization. This can be fixed " + "using the \"btrfs balance\" command."); } -out: - mutex_unlock(&root->fs_info->volume_mutex); + return ret; + +error_trans: + unlock_chunks(root); + btrfs_abort_transaction(trans, root, ret); + btrfs_end_transaction(trans, root); + rcu_string_free(device->name); + kfree(device); error: blkdev_put(bdev, FMODE_EXCL); if (seeding_dev) { mutex_unlock(&uuid_mutex); up_write(&sb->s_umount); } - goto out; + return ret; } static noinline int btrfs_update_device(struct btrfs_trans_handle *trans, @@ -1790,7 +1877,7 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 new_size) { struct btrfs_super_block *super_copy = - &device->dev_root->fs_info->super_copy; + device->dev_root->fs_info->super_copy; u64 old_total = btrfs_super_total_bytes(super_copy); u64 diff = new_size - device->total_bytes; @@ -1838,10 +1925,20 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans, key.type = BTRFS_CHUNK_ITEM_KEY; ret = btrfs_search_slot(trans, root, &key, path, -1, 1); - BUG_ON(ret); + if (ret < 0) + goto out; + else if (ret > 0) { /* Logic error or corruption */ + btrfs_error(root->fs_info, -ENOENT, + "Failed lookup while freeing chunk."); + ret = -ENOENT; + goto out; + } ret = btrfs_del_item(trans, root, path); - + if (ret < 0) + btrfs_error(root->fs_info, ret, + "Failed to delete chunk item."); +out: btrfs_free_path(path); return ret; } @@ -1849,7 +1946,7 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans, static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64 chunk_offset) { - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; struct btrfs_disk_key *disk_key; struct btrfs_chunk *chunk; u8 *ptr; @@ -1930,7 +2027,7 @@ static int btrfs_relocate_chunk(struct btrfs_root *root, em = lookup_extent_mapping(em_tree, chunk_offset, 1); read_unlock(&em_tree->lock); - BUG_ON(em->start > chunk_offset || + BUG_ON(!em || em->start > chunk_offset || em->start + em->len < chunk_offset); map = (struct map_lookup *)em->bdev; @@ -2003,7 +2100,7 @@ again: ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0); if (ret < 0) goto error; - BUG_ON(ret == 0); + BUG_ON(ret == 0); /* Corruption */ ret = btrfs_previous_item(chunk_root, path, key.objectid, key.type); @@ -2048,6 +2145,358 @@ error: return ret; } +static int insert_balance_item(struct btrfs_root *root, + struct btrfs_balance_control *bctl) +{ + struct btrfs_trans_handle *trans; + struct btrfs_balance_item *item; + struct btrfs_disk_balance_args disk_bargs; + struct btrfs_path *path; + struct extent_buffer *leaf; + struct btrfs_key key; + int ret, err; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + trans = btrfs_start_transaction(root, 0); + if (IS_ERR(trans)) { + btrfs_free_path(path); + return PTR_ERR(trans); + } + + key.objectid = BTRFS_BALANCE_OBJECTID; + key.type = BTRFS_BALANCE_ITEM_KEY; + key.offset = 0; + + ret = btrfs_insert_empty_item(trans, root, path, &key, + sizeof(*item)); + if (ret) + goto out; + + leaf = path->nodes[0]; + item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_balance_item); + + memset_extent_buffer(leaf, 0, (unsigned long)item, sizeof(*item)); + + btrfs_cpu_balance_args_to_disk(&disk_bargs, &bctl->data); + btrfs_set_balance_data(leaf, item, &disk_bargs); + btrfs_cpu_balance_args_to_disk(&disk_bargs, &bctl->meta); + btrfs_set_balance_meta(leaf, item, &disk_bargs); + btrfs_cpu_balance_args_to_disk(&disk_bargs, &bctl->sys); + btrfs_set_balance_sys(leaf, item, &disk_bargs); + + btrfs_set_balance_flags(leaf, item, bctl->flags); + + btrfs_mark_buffer_dirty(leaf); +out: + btrfs_free_path(path); + err = btrfs_commit_transaction(trans, root); + if (err && !ret) + ret = err; + return ret; +} + +static int del_balance_item(struct btrfs_root *root) +{ + struct btrfs_trans_handle *trans; + struct btrfs_path *path; + struct btrfs_key key; + int ret, err; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + trans = btrfs_start_transaction(root, 0); + if (IS_ERR(trans)) { + btrfs_free_path(path); + return PTR_ERR(trans); + } + + key.objectid = BTRFS_BALANCE_OBJECTID; + key.type = BTRFS_BALANCE_ITEM_KEY; + key.offset = 0; + + ret = btrfs_search_slot(trans, root, &key, path, -1, 1); + if (ret < 0) + goto out; + if (ret > 0) { + ret = -ENOENT; + goto out; + } + + ret = btrfs_del_item(trans, root, path); +out: + btrfs_free_path(path); + err = btrfs_commit_transaction(trans, root); + if (err && !ret) + ret = err; + return ret; +} + +/* + * This is a heuristic used to reduce the number of chunks balanced on + * resume after balance was interrupted. + */ +static void update_balance_args(struct btrfs_balance_control *bctl) +{ + /* + * Turn on soft mode for chunk types that were being converted. + */ + if (bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) + bctl->data.flags |= BTRFS_BALANCE_ARGS_SOFT; + if (bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) + bctl->sys.flags |= BTRFS_BALANCE_ARGS_SOFT; + if (bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) + bctl->meta.flags |= BTRFS_BALANCE_ARGS_SOFT; + + /* + * Turn on usage filter if is not already used. The idea is + * that chunks that we have already balanced should be + * reasonably full. Don't do it for chunks that are being + * converted - that will keep us from relocating unconverted + * (albeit full) chunks. + */ + if (!(bctl->data.flags & BTRFS_BALANCE_ARGS_USAGE) && + !(bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT)) { + bctl->data.flags |= BTRFS_BALANCE_ARGS_USAGE; + bctl->data.usage = 90; + } + if (!(bctl->sys.flags & BTRFS_BALANCE_ARGS_USAGE) && + !(bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT)) { + bctl->sys.flags |= BTRFS_BALANCE_ARGS_USAGE; + bctl->sys.usage = 90; + } + if (!(bctl->meta.flags & BTRFS_BALANCE_ARGS_USAGE) && + !(bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT)) { + bctl->meta.flags |= BTRFS_BALANCE_ARGS_USAGE; + bctl->meta.usage = 90; + } +} + +/* + * Should be called with both balance and volume mutexes held to + * serialize other volume operations (add_dev/rm_dev/resize) with + * restriper. Same goes for unset_balance_control. + */ +static void set_balance_control(struct btrfs_balance_control *bctl) +{ + struct btrfs_fs_info *fs_info = bctl->fs_info; + + BUG_ON(fs_info->balance_ctl); + + spin_lock(&fs_info->balance_lock); + fs_info->balance_ctl = bctl; + spin_unlock(&fs_info->balance_lock); +} + +static void unset_balance_control(struct btrfs_fs_info *fs_info) +{ + struct btrfs_balance_control *bctl = fs_info->balance_ctl; + + BUG_ON(!fs_info->balance_ctl); + + spin_lock(&fs_info->balance_lock); + fs_info->balance_ctl = NULL; + spin_unlock(&fs_info->balance_lock); + + kfree(bctl); +} + +/* + * Balance filters. Return 1 if chunk should be filtered out + * (should not be balanced). + */ +static int chunk_profiles_filter(u64 chunk_type, + struct btrfs_balance_args *bargs) +{ + chunk_type = chunk_to_extended(chunk_type) & + BTRFS_EXTENDED_PROFILE_MASK; + + if (bargs->profiles & chunk_type) + return 0; + + return 1; +} + +static u64 div_factor_fine(u64 num, int factor) +{ + if (factor <= 0) + return 0; + if (factor >= 100) + return num; + + num *= factor; + do_div(num, 100); + return num; +} + +static int chunk_usage_filter(struct btrfs_fs_info *fs_info, u64 chunk_offset, + struct btrfs_balance_args *bargs) +{ + struct btrfs_block_group_cache *cache; + u64 chunk_used, user_thresh; + int ret = 1; + + cache = btrfs_lookup_block_group(fs_info, chunk_offset); + chunk_used = btrfs_block_group_used(&cache->item); + + user_thresh = div_factor_fine(cache->key.offset, bargs->usage); + if (chunk_used < user_thresh) + ret = 0; + + btrfs_put_block_group(cache); + return ret; +} + +static int chunk_devid_filter(struct extent_buffer *leaf, + struct btrfs_chunk *chunk, + struct btrfs_balance_args *bargs) +{ + struct btrfs_stripe *stripe; + int num_stripes = btrfs_chunk_num_stripes(leaf, chunk); + int i; + + for (i = 0; i < num_stripes; i++) { + stripe = btrfs_stripe_nr(chunk, i); + if (btrfs_stripe_devid(leaf, stripe) == bargs->devid) + return 0; + } + + return 1; +} + +/* [pstart, pend) */ +static int chunk_drange_filter(struct extent_buffer *leaf, + struct btrfs_chunk *chunk, + u64 chunk_offset, + struct btrfs_balance_args *bargs) +{ + struct btrfs_stripe *stripe; + int num_stripes = btrfs_chunk_num_stripes(leaf, chunk); + u64 stripe_offset; + u64 stripe_length; + int factor; + int i; + + if (!(bargs->flags & BTRFS_BALANCE_ARGS_DEVID)) + return 0; + + if (btrfs_chunk_type(leaf, chunk) & (BTRFS_BLOCK_GROUP_DUP | + BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10)) + factor = 2; + else + factor = 1; + factor = num_stripes / factor; + + for (i = 0; i < num_stripes; i++) { + stripe = btrfs_stripe_nr(chunk, i); + if (btrfs_stripe_devid(leaf, stripe) != bargs->devid) + continue; + + stripe_offset = btrfs_stripe_offset(leaf, stripe); + stripe_length = btrfs_chunk_length(leaf, chunk); + do_div(stripe_length, factor); + + if (stripe_offset < bargs->pend && + stripe_offset + stripe_length > bargs->pstart) + return 0; + } + + return 1; +} + +/* [vstart, vend) */ +static int chunk_vrange_filter(struct extent_buffer *leaf, + struct btrfs_chunk *chunk, + u64 chunk_offset, + struct btrfs_balance_args *bargs) +{ + if (chunk_offset < bargs->vend && + chunk_offset + btrfs_chunk_length(leaf, chunk) > bargs->vstart) + /* at least part of the chunk is inside this vrange */ + return 0; + + return 1; +} + +static int chunk_soft_convert_filter(u64 chunk_type, + struct btrfs_balance_args *bargs) +{ + if (!(bargs->flags & BTRFS_BALANCE_ARGS_CONVERT)) + return 0; + + chunk_type = chunk_to_extended(chunk_type) & + BTRFS_EXTENDED_PROFILE_MASK; + + if (bargs->target == chunk_type) + return 1; + + return 0; +} + +static int should_balance_chunk(struct btrfs_root *root, + struct extent_buffer *leaf, + struct btrfs_chunk *chunk, u64 chunk_offset) +{ + struct btrfs_balance_control *bctl = root->fs_info->balance_ctl; + struct btrfs_balance_args *bargs = NULL; + u64 chunk_type = btrfs_chunk_type(leaf, chunk); + + /* type filter */ + if (!((chunk_type & BTRFS_BLOCK_GROUP_TYPE_MASK) & + (bctl->flags & BTRFS_BALANCE_TYPE_MASK))) { + return 0; + } + + if (chunk_type & BTRFS_BLOCK_GROUP_DATA) + bargs = &bctl->data; + else if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) + bargs = &bctl->sys; + else if (chunk_type & BTRFS_BLOCK_GROUP_METADATA) + bargs = &bctl->meta; + + /* profiles filter */ + if ((bargs->flags & BTRFS_BALANCE_ARGS_PROFILES) && + chunk_profiles_filter(chunk_type, bargs)) { + return 0; + } + + /* usage filter */ + if ((bargs->flags & BTRFS_BALANCE_ARGS_USAGE) && + chunk_usage_filter(bctl->fs_info, chunk_offset, bargs)) { + return 0; + } + + /* devid filter */ + if ((bargs->flags & BTRFS_BALANCE_ARGS_DEVID) && + chunk_devid_filter(leaf, chunk, bargs)) { + return 0; + } + + /* drange filter, makes sense only with devid filter */ + if ((bargs->flags & BTRFS_BALANCE_ARGS_DRANGE) && + chunk_drange_filter(leaf, chunk, chunk_offset, bargs)) { + return 0; + } + + /* vrange filter */ + if ((bargs->flags & BTRFS_BALANCE_ARGS_VRANGE) && + chunk_vrange_filter(leaf, chunk, chunk_offset, bargs)) { + return 0; + } + + /* soft profile changing mode */ + if ((bargs->flags & BTRFS_BALANCE_ARGS_SOFT) && + chunk_soft_convert_filter(chunk_type, bargs)) { + return 0; + } + + return 1; +} + static u64 div_factor(u64 num, int factor) { if (factor == 10) @@ -2057,29 +2506,28 @@ static u64 div_factor(u64 num, int factor) return num; } -int btrfs_balance(struct btrfs_root *dev_root) +static int __btrfs_balance(struct btrfs_fs_info *fs_info) { - int ret; - struct list_head *devices = &dev_root->fs_info->fs_devices->devices; + struct btrfs_balance_control *bctl = fs_info->balance_ctl; + struct btrfs_root *chunk_root = fs_info->chunk_root; + struct btrfs_root *dev_root = fs_info->dev_root; + struct list_head *devices; struct btrfs_device *device; u64 old_size; u64 size_to_free; + struct btrfs_chunk *chunk; struct btrfs_path *path; struct btrfs_key key; - struct btrfs_root *chunk_root = dev_root->fs_info->chunk_root; - struct btrfs_trans_handle *trans; struct btrfs_key found_key; - - if (dev_root->fs_info->sb->s_flags & MS_RDONLY) - return -EROFS; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - mutex_lock(&dev_root->fs_info->volume_mutex); - dev_root = dev_root->fs_info->dev_root; + struct btrfs_trans_handle *trans; + struct extent_buffer *leaf; + int slot; + int ret; + int enospc_errors = 0; + bool counting = true; /* step one make some room on all the devices */ + devices = &fs_info->fs_devices->devices; list_for_each_entry(device, devices, dev_list) { old_size = device->total_bytes; size_to_free = div_factor(old_size, 1); @@ -2108,11 +2556,23 @@ int btrfs_balance(struct btrfs_root *dev_root) ret = -ENOMEM; goto error; } + + /* zero out stat counters */ + spin_lock(&fs_info->balance_lock); + memset(&bctl->stat, 0, sizeof(bctl->stat)); + spin_unlock(&fs_info->balance_lock); +again: key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID; key.offset = (u64)-1; key.type = BTRFS_CHUNK_ITEM_KEY; while (1) { + if ((!counting && atomic_read(&fs_info->balance_pause_req)) || + atomic_read(&fs_info->balance_cancel_req)) { + ret = -ECANCELED; + goto error; + } + ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0); if (ret < 0) goto error; @@ -2122,15 +2582,19 @@ int btrfs_balance(struct btrfs_root *dev_root) * failed */ if (ret == 0) - break; + BUG(); /* FIXME break ? */ ret = btrfs_previous_item(chunk_root, path, 0, BTRFS_CHUNK_ITEM_KEY); - if (ret) + if (ret) { + ret = 0; break; + } + + leaf = path->nodes[0]; + slot = path->slots[0]; + btrfs_item_key_to_cpu(leaf, &found_key, slot); - btrfs_item_key_to_cpu(path->nodes[0], &found_key, - path->slots[0]); if (found_key.objectid != key.objectid) break; @@ -2138,22 +2602,416 @@ int btrfs_balance(struct btrfs_root *dev_root) if (found_key.offset == 0) break; + chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk); + + if (!counting) { + spin_lock(&fs_info->balance_lock); + bctl->stat.considered++; + spin_unlock(&fs_info->balance_lock); + } + + ret = should_balance_chunk(chunk_root, leaf, chunk, + found_key.offset); btrfs_release_path(path); + if (!ret) + goto loop; + + if (counting) { + spin_lock(&fs_info->balance_lock); + bctl->stat.expected++; + spin_unlock(&fs_info->balance_lock); + goto loop; + } + ret = btrfs_relocate_chunk(chunk_root, chunk_root->root_key.objectid, found_key.objectid, found_key.offset); if (ret && ret != -ENOSPC) goto error; + if (ret == -ENOSPC) { + enospc_errors++; + } else { + spin_lock(&fs_info->balance_lock); + bctl->stat.completed++; + spin_unlock(&fs_info->balance_lock); + } +loop: key.offset = found_key.offset - 1; } - ret = 0; + + if (counting) { + btrfs_release_path(path); + counting = false; + goto again; + } error: btrfs_free_path(path); - mutex_unlock(&dev_root->fs_info->volume_mutex); + if (enospc_errors) { + printk(KERN_INFO "btrfs: %d enospc errors during balance\n", + enospc_errors); + if (!ret) + ret = -ENOSPC; + } + return ret; } +/** + * alloc_profile_is_valid - see if a given profile is valid and reduced + * @flags: profile to validate + * @extended: if true @flags is treated as an extended profile + */ +static int alloc_profile_is_valid(u64 flags, int extended) +{ + u64 mask = (extended ? BTRFS_EXTENDED_PROFILE_MASK : + BTRFS_BLOCK_GROUP_PROFILE_MASK); + + flags &= ~BTRFS_BLOCK_GROUP_TYPE_MASK; + + /* 1) check that all other bits are zeroed */ + if (flags & ~mask) + return 0; + + /* 2) see if profile is reduced */ + if (flags == 0) + return !extended; /* "0" is valid for usual profiles */ + + /* true if exactly one bit set */ + return (flags & (flags - 1)) == 0; +} + +static inline int balance_need_close(struct btrfs_fs_info *fs_info) +{ + /* cancel requested || normal exit path */ + return atomic_read(&fs_info->balance_cancel_req) || + (atomic_read(&fs_info->balance_pause_req) == 0 && + atomic_read(&fs_info->balance_cancel_req) == 0); +} + +static void __cancel_balance(struct btrfs_fs_info *fs_info) +{ + int ret; + + unset_balance_control(fs_info); + ret = del_balance_item(fs_info->tree_root); + BUG_ON(ret); +} + +void update_ioctl_balance_args(struct btrfs_fs_info *fs_info, int lock, + struct btrfs_ioctl_balance_args *bargs); + +/* + * Should be called with both balance and volume mutexes held + */ +int btrfs_balance(struct btrfs_balance_control *bctl, + struct btrfs_ioctl_balance_args *bargs) +{ + struct btrfs_fs_info *fs_info = bctl->fs_info; + u64 allowed; + int mixed = 0; + int ret; + + if (btrfs_fs_closing(fs_info) || + atomic_read(&fs_info->balance_pause_req) || + atomic_read(&fs_info->balance_cancel_req)) { + ret = -EINVAL; + goto out; + } + + allowed = btrfs_super_incompat_flags(fs_info->super_copy); + if (allowed & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) + mixed = 1; + + /* + * In case of mixed groups both data and meta should be picked, + * and identical options should be given for both of them. + */ + allowed = BTRFS_BALANCE_DATA | BTRFS_BALANCE_METADATA; + if (mixed && (bctl->flags & allowed)) { + if (!(bctl->flags & BTRFS_BALANCE_DATA) || + !(bctl->flags & BTRFS_BALANCE_METADATA) || + memcmp(&bctl->data, &bctl->meta, sizeof(bctl->data))) { + printk(KERN_ERR "btrfs: with mixed groups data and " + "metadata balance options must be the same\n"); + ret = -EINVAL; + goto out; + } + } + + allowed = BTRFS_AVAIL_ALLOC_BIT_SINGLE; + if (fs_info->fs_devices->num_devices == 1) + allowed |= BTRFS_BLOCK_GROUP_DUP; + else if (fs_info->fs_devices->num_devices < 4) + allowed |= (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1); + else + allowed |= (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10); + + if ((bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (!alloc_profile_is_valid(bctl->data.target, 1) || + (bctl->data.target & ~allowed))) { + printk(KERN_ERR "btrfs: unable to start balance with target " + "data profile %llu\n", + (unsigned long long)bctl->data.target); + ret = -EINVAL; + goto out; + } + if ((bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (!alloc_profile_is_valid(bctl->meta.target, 1) || + (bctl->meta.target & ~allowed))) { + printk(KERN_ERR "btrfs: unable to start balance with target " + "metadata profile %llu\n", + (unsigned long long)bctl->meta.target); + ret = -EINVAL; + goto out; + } + if ((bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (!alloc_profile_is_valid(bctl->sys.target, 1) || + (bctl->sys.target & ~allowed))) { + printk(KERN_ERR "btrfs: unable to start balance with target " + "system profile %llu\n", + (unsigned long long)bctl->sys.target); + ret = -EINVAL; + goto out; + } + + /* allow dup'ed data chunks only in mixed mode */ + if (!mixed && (bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (bctl->data.target & BTRFS_BLOCK_GROUP_DUP)) { + printk(KERN_ERR "btrfs: dup for data is not allowed\n"); + ret = -EINVAL; + goto out; + } + + /* allow to reduce meta or sys integrity only if force set */ + allowed = BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10; + if (((bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (fs_info->avail_system_alloc_bits & allowed) && + !(bctl->sys.target & allowed)) || + ((bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (fs_info->avail_metadata_alloc_bits & allowed) && + !(bctl->meta.target & allowed))) { + if (bctl->flags & BTRFS_BALANCE_FORCE) { + printk(KERN_INFO "btrfs: force reducing metadata " + "integrity\n"); + } else { + printk(KERN_ERR "btrfs: balance will reduce metadata " + "integrity, use force if you want this\n"); + ret = -EINVAL; + goto out; + } + } + + ret = insert_balance_item(fs_info->tree_root, bctl); + if (ret && ret != -EEXIST) + goto out; + + if (!(bctl->flags & BTRFS_BALANCE_RESUME)) { + BUG_ON(ret == -EEXIST); + set_balance_control(bctl); + } else { + BUG_ON(ret != -EEXIST); + spin_lock(&fs_info->balance_lock); + update_balance_args(bctl); + spin_unlock(&fs_info->balance_lock); + } + + atomic_inc(&fs_info->balance_running); + mutex_unlock(&fs_info->balance_mutex); + + ret = __btrfs_balance(fs_info); + + mutex_lock(&fs_info->balance_mutex); + atomic_dec(&fs_info->balance_running); + + if (bargs) { + memset(bargs, 0, sizeof(*bargs)); + update_ioctl_balance_args(fs_info, 0, bargs); + } + + if ((ret && ret != -ECANCELED && ret != -ENOSPC) || + balance_need_close(fs_info)) { + __cancel_balance(fs_info); + } + + wake_up(&fs_info->balance_wait_q); + + return ret; +out: + if (bctl->flags & BTRFS_BALANCE_RESUME) + __cancel_balance(fs_info); + else + kfree(bctl); + return ret; +} + +static int balance_kthread(void *data) +{ + struct btrfs_fs_info *fs_info = data; + int ret = 0; + + mutex_lock(&fs_info->volume_mutex); + mutex_lock(&fs_info->balance_mutex); + + if (fs_info->balance_ctl) { + printk(KERN_INFO "btrfs: continuing balance\n"); + ret = btrfs_balance(fs_info->balance_ctl, NULL); + } + + mutex_unlock(&fs_info->balance_mutex); + mutex_unlock(&fs_info->volume_mutex); + + return ret; +} + +int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info) +{ + struct task_struct *tsk; + + spin_lock(&fs_info->balance_lock); + if (!fs_info->balance_ctl) { + spin_unlock(&fs_info->balance_lock); + return 0; + } + spin_unlock(&fs_info->balance_lock); + + if (btrfs_test_opt(fs_info->tree_root, SKIP_BALANCE)) { + printk(KERN_INFO "btrfs: force skipping balance\n"); + return 0; + } + + tsk = kthread_run(balance_kthread, fs_info, "btrfs-balance"); + if (IS_ERR(tsk)) + return PTR_ERR(tsk); + + return 0; +} + +int btrfs_recover_balance(struct btrfs_fs_info *fs_info) +{ + struct btrfs_balance_control *bctl; + struct btrfs_balance_item *item; + struct btrfs_disk_balance_args disk_bargs; + struct btrfs_path *path; + struct extent_buffer *leaf; + struct btrfs_key key; + int ret; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + key.objectid = BTRFS_BALANCE_OBJECTID; + key.type = BTRFS_BALANCE_ITEM_KEY; + key.offset = 0; + + ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, path, 0, 0); + if (ret < 0) + goto out; + if (ret > 0) { /* ret = -ENOENT; */ + ret = 0; + goto out; + } + + bctl = kzalloc(sizeof(*bctl), GFP_NOFS); + if (!bctl) { + ret = -ENOMEM; + goto out; + } + + leaf = path->nodes[0]; + item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_balance_item); + + bctl->fs_info = fs_info; + bctl->flags = btrfs_balance_flags(leaf, item); + bctl->flags |= BTRFS_BALANCE_RESUME; + + btrfs_balance_data(leaf, item, &disk_bargs); + btrfs_disk_balance_args_to_cpu(&bctl->data, &disk_bargs); + btrfs_balance_meta(leaf, item, &disk_bargs); + btrfs_disk_balance_args_to_cpu(&bctl->meta, &disk_bargs); + btrfs_balance_sys(leaf, item, &disk_bargs); + btrfs_disk_balance_args_to_cpu(&bctl->sys, &disk_bargs); + + mutex_lock(&fs_info->volume_mutex); + mutex_lock(&fs_info->balance_mutex); + + set_balance_control(bctl); + + mutex_unlock(&fs_info->balance_mutex); + mutex_unlock(&fs_info->volume_mutex); +out: + btrfs_free_path(path); + return ret; +} + +int btrfs_pause_balance(struct btrfs_fs_info *fs_info) +{ + int ret = 0; + + mutex_lock(&fs_info->balance_mutex); + if (!fs_info->balance_ctl) { + mutex_unlock(&fs_info->balance_mutex); + return -ENOTCONN; + } + + if (atomic_read(&fs_info->balance_running)) { + atomic_inc(&fs_info->balance_pause_req); + mutex_unlock(&fs_info->balance_mutex); + + wait_event(fs_info->balance_wait_q, + atomic_read(&fs_info->balance_running) == 0); + + mutex_lock(&fs_info->balance_mutex); + /* we are good with balance_ctl ripped off from under us */ + BUG_ON(atomic_read(&fs_info->balance_running)); + atomic_dec(&fs_info->balance_pause_req); + } else { + ret = -ENOTCONN; + } + + mutex_unlock(&fs_info->balance_mutex); + return ret; +} + +int btrfs_cancel_balance(struct btrfs_fs_info *fs_info) +{ + mutex_lock(&fs_info->balance_mutex); + if (!fs_info->balance_ctl) { + mutex_unlock(&fs_info->balance_mutex); + return -ENOTCONN; + } + + atomic_inc(&fs_info->balance_cancel_req); + /* + * if we are running just wait and return, balance item is + * deleted in btrfs_balance in this case + */ + if (atomic_read(&fs_info->balance_running)) { + mutex_unlock(&fs_info->balance_mutex); + wait_event(fs_info->balance_wait_q, + atomic_read(&fs_info->balance_running) == 0); + mutex_lock(&fs_info->balance_mutex); + } else { + /* __cancel_balance needs volume_mutex */ + mutex_unlock(&fs_info->balance_mutex); + mutex_lock(&fs_info->volume_mutex); + mutex_lock(&fs_info->balance_mutex); + + if (fs_info->balance_ctl) + __cancel_balance(fs_info); + + mutex_unlock(&fs_info->volume_mutex); + } + + BUG_ON(fs_info->balance_ctl || atomic_read(&fs_info->balance_running)); + atomic_dec(&fs_info->balance_cancel_req); + mutex_unlock(&fs_info->balance_mutex); + return 0; +} + /* * shrinking a device means finding all of the device extents past * the new size, and then following the back refs to the chunks. @@ -2175,7 +3033,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size) bool retried = false; struct extent_buffer *l; struct btrfs_key key; - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; u64 old_total = btrfs_super_total_bytes(super_copy); u64 old_size = device->total_bytes; u64 diff = device->total_bytes - new_size; @@ -2192,8 +3050,12 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size) lock_chunks(root); device->total_bytes = new_size; - if (device->writeable) + if (device->writeable) { device->fs_devices->total_rw_bytes -= diff; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space -= diff; + spin_unlock(&root->fs_info->free_chunk_lock); + } unlock_chunks(root); again: @@ -2201,7 +3063,7 @@ again: key.offset = (u64)-1; key.type = BTRFS_DEV_EXTENT_KEY; - while (1) { + do { ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) goto done; @@ -2243,8 +3105,7 @@ again: goto done; if (ret == -ENOSPC) failed++; - key.offset -= 1; - } + } while (key.offset-- > 0); if (failed && !retried) { failed = 0; @@ -2257,6 +3118,9 @@ again: device->total_bytes = old_size; if (device->writeable) device->fs_devices->total_rw_bytes += diff; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += diff; + spin_unlock(&root->fs_info->free_chunk_lock); unlock_chunks(root); goto done; } @@ -2287,12 +3151,11 @@ done: return ret; } -static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans, - struct btrfs_root *root, +static int btrfs_add_system_chunk(struct btrfs_root *root, struct btrfs_key *key, struct btrfs_chunk *chunk, int item_size) { - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; struct btrfs_disk_key disk_key; u32 array_size; u8 *ptr; @@ -2360,11 +3223,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, int i; int j; - if ((type & BTRFS_BLOCK_GROUP_RAID1) && - (type & BTRFS_BLOCK_GROUP_DUP)) { - WARN_ON(1); - type &= ~BTRFS_BLOCK_GROUP_DUP; - } + BUG_ON(!alloc_profile_is_valid(type, 0)); if (list_empty(&fs_devices->alloc_list)) return -ENOSPC; @@ -2405,10 +3264,14 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, max_stripe_size = 1024 * 1024 * 1024; max_chunk_size = 10 * max_stripe_size; } else if (type & BTRFS_BLOCK_GROUP_METADATA) { - max_stripe_size = 256 * 1024 * 1024; + /* for larger filesystems, use larger metadata chunks */ + if (fs_devices->total_rw_bytes > 50ULL * 1024 * 1024 * 1024) + max_stripe_size = 1024 * 1024 * 1024; + else + max_stripe_size = 256 * 1024 * 1024; max_chunk_size = max_stripe_size; } else if (type & BTRFS_BLOCK_GROUP_SYSTEM) { - max_stripe_size = 8 * 1024 * 1024; + max_stripe_size = 32 * 1024 * 1024; max_chunk_size = 2 * max_stripe_size; } else { printk(KERN_ERR "btrfs: invalid chunk type 0x%llx requested\n", @@ -2460,7 +3323,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, if (total_avail == 0) continue; - ret = find_free_dev_extent(trans, device, + ret = find_free_dev_extent(device, max_stripe_size * dev_stripes, &dev_offset, &max_avail); if (ret && ret != -ENOSPC) @@ -2502,12 +3365,14 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, stripe_size = devices_info[ndevs-1].max_avail; num_stripes = ndevs * dev_stripes; - if (stripe_size * num_stripes > max_chunk_size * ncopies) { + if (stripe_size * ndevs > max_chunk_size * ncopies) { stripe_size = max_chunk_size * ncopies; - do_div(stripe_size, num_stripes); + do_div(stripe_size, ndevs); } do_div(stripe_size, dev_stripes); + + /* align to BTRFS_STRIPE_LEN */ do_div(stripe_size, BTRFS_STRIPE_LEN); stripe_size *= BTRFS_STRIPE_LEN; @@ -2556,13 +3421,15 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, write_lock(&em_tree->lock); ret = add_extent_mapping(em_tree, em); write_unlock(&em_tree->lock); - BUG_ON(ret); free_extent_map(em); + if (ret) + goto error; ret = btrfs_make_block_group(trans, extent_root, 0, type, BTRFS_FIRST_CHUNK_TREE_OBJECTID, start, num_bytes); - BUG_ON(ret); + if (ret) + goto error; for (i = 0; i < map->num_stripes; ++i) { struct btrfs_device *device; @@ -2575,7 +3442,10 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, info->chunk_root->root_key.objectid, BTRFS_FIRST_CHUNK_TREE_OBJECTID, start, dev_offset, stripe_size); - BUG_ON(ret); + if (ret) { + btrfs_abort_transaction(trans, extent_root, ret); + goto error; + } } kfree(devices_info); @@ -2611,10 +3481,16 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans, device = map->stripes[index].dev; device->bytes_used += stripe_size; ret = btrfs_update_device(trans, device); - BUG_ON(ret); + if (ret) + goto out_free; index++; } + spin_lock(&extent_root->fs_info->free_chunk_lock); + extent_root->fs_info->free_chunk_space -= (stripe_size * + map->num_stripes); + spin_unlock(&extent_root->fs_info->free_chunk_lock); + index = 0; stripe = &chunk->stripe; while (index < map->num_stripes) { @@ -2643,16 +3519,19 @@ static int __finish_chunk_alloc(struct btrfs_trans_handle *trans, key.offset = chunk_offset; ret = btrfs_insert_item(trans, chunk_root, &key, chunk, item_size); - BUG_ON(ret); - if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) { - ret = btrfs_add_system_chunk(trans, chunk_root, &key, chunk, + if (ret == 0 && map->type & BTRFS_BLOCK_GROUP_SYSTEM) { + /* + * TODO: Cleanup of inserted chunk root in case of + * failure. + */ + ret = btrfs_add_system_chunk(chunk_root, &key, chunk, item_size); - BUG_ON(ret); } +out_free: kfree(chunk); - return 0; + return ret; } /* @@ -2684,7 +3563,8 @@ int btrfs_alloc_chunk(struct btrfs_trans_handle *trans, ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset, chunk_size, stripe_size); - BUG_ON(ret); + if (ret) + return ret; return 0; } @@ -2711,28 +3591,29 @@ static noinline int init_first_rw_device(struct btrfs_trans_handle *trans, return ret; alloc_profile = BTRFS_BLOCK_GROUP_METADATA | - (fs_info->metadata_alloc_profile & - fs_info->avail_metadata_alloc_bits); + fs_info->avail_metadata_alloc_bits; alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile); ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size, &stripe_size, chunk_offset, alloc_profile); - BUG_ON(ret); + if (ret) + return ret; sys_chunk_offset = chunk_offset + chunk_size; alloc_profile = BTRFS_BLOCK_GROUP_SYSTEM | - (fs_info->system_alloc_profile & - fs_info->avail_system_alloc_bits); + fs_info->avail_system_alloc_bits; alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile); ret = __btrfs_alloc_chunk(trans, extent_root, &sys_map, &sys_chunk_size, &sys_stripe_size, sys_chunk_offset, alloc_profile); - BUG_ON(ret); + if (ret) + goto abort; ret = btrfs_add_device(trans, fs_info->chunk_root, device); - BUG_ON(ret); + if (ret) + goto abort; /* * Modifying chunk tree needs allocating new blocks from both @@ -2742,13 +3623,20 @@ static noinline int init_first_rw_device(struct btrfs_trans_handle *trans, */ ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset, chunk_size, stripe_size); - BUG_ON(ret); + if (ret) + goto abort; ret = __finish_chunk_alloc(trans, extent_root, sys_map, sys_chunk_offset, sys_chunk_size, sys_stripe_size); - BUG_ON(ret); + if (ret) + goto abort; + return 0; + +abort: + btrfs_abort_transaction(trans, root, ret); + return ret; } int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset) @@ -2848,7 +3736,7 @@ static int find_live_mirror(struct map_lookup *map, int first, int num, static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 logical, u64 *length, - struct btrfs_multi_bio **multi_ret, + struct btrfs_bio **bbio_ret, int mirror_num) { struct extent_map *em; @@ -2860,25 +3748,12 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 stripe_nr; u64 stripe_nr_orig; u64 stripe_nr_end; - int stripes_allocated = 8; - int stripes_required = 1; int stripe_index; int i; + int ret = 0; int num_stripes; int max_errors = 0; - struct btrfs_multi_bio *multi = NULL; - - if (multi_ret && !(rw & (REQ_WRITE | REQ_DISCARD))) - stripes_allocated = 1; -again: - if (multi_ret) { - multi = kzalloc(btrfs_multi_bio_size(stripes_allocated), - GFP_NOFS); - if (!multi) - return -ENOMEM; - - atomic_set(&multi->error, 0); - } + struct btrfs_bio *bbio = NULL; read_lock(&em_tree->lock); em = lookup_extent_mapping(em_tree, logical, *length); @@ -2898,32 +3773,6 @@ again: if (mirror_num > map->num_stripes) mirror_num = 0; - /* if our multi bio struct is too small, back off and try again */ - if (rw & REQ_WRITE) { - if (map->type & (BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_DUP)) { - stripes_required = map->num_stripes; - max_errors = 1; - } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { - stripes_required = map->sub_stripes; - max_errors = 1; - } - } - if (rw & REQ_DISCARD) { - if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | - BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_DUP | - BTRFS_BLOCK_GROUP_RAID10)) { - stripes_required = map->num_stripes; - } - } - if (multi_ret && (rw & (REQ_WRITE | REQ_DISCARD)) && - stripes_allocated < stripes_required) { - stripes_allocated = map->num_stripes; - free_extent_map(em); - kfree(multi); - goto again; - } stripe_nr = offset; /* * stripe_nr counts the total number of stripes we have to stride @@ -2939,10 +3788,7 @@ again: if (rw & REQ_DISCARD) *length = min_t(u64, em->len - offset, *length); - else if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | - BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_RAID10 | - BTRFS_BLOCK_GROUP_DUP)) { + else if (map->type & BTRFS_BLOCK_GROUP_PROFILE_MASK) { /* we limit the length of each bio to what fits in a stripe */ *length = min_t(u64, em->len - offset, map->stripe_len - stripe_offset); @@ -2950,7 +3796,7 @@ again: *length = em->len - offset; } - if (!multi_ret) + if (!bbio_ret) goto out; num_stripes = 1; @@ -2975,13 +3821,17 @@ again: stripe_index = find_live_mirror(map, 0, map->num_stripes, current->pid % map->num_stripes); + mirror_num = stripe_index + 1; } } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { - if (rw & (REQ_WRITE | REQ_DISCARD)) + if (rw & (REQ_WRITE | REQ_DISCARD)) { num_stripes = map->num_stripes; - else if (mirror_num) + } else if (mirror_num) { stripe_index = mirror_num - 1; + } else { + mirror_num = 1; + } } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { int factor = map->num_stripes / map->sub_stripes; @@ -2998,9 +3848,11 @@ again: else if (mirror_num) stripe_index += mirror_num - 1; else { + int old_stripe_index = stripe_index; stripe_index = find_live_mirror(map, stripe_index, map->sub_stripes, stripe_index + current->pid % map->sub_stripes); + mirror_num = stripe_index - old_stripe_index + 1; } } else { /* @@ -3009,86 +3861,77 @@ again: * stripe_index is the number of our device in the stripe array */ stripe_index = do_div(stripe_nr, map->num_stripes); + mirror_num = stripe_index + 1; } BUG_ON(stripe_index >= map->num_stripes); + bbio = kzalloc(btrfs_bio_size(num_stripes), GFP_NOFS); + if (!bbio) { + ret = -ENOMEM; + goto out; + } + atomic_set(&bbio->error, 0); + if (rw & REQ_DISCARD) { + int factor = 0; + int sub_stripes = 0; + u64 stripes_per_dev = 0; + u32 remaining_stripes = 0; + u32 last_stripe = 0; + + if (map->type & + (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID10)) { + if (map->type & BTRFS_BLOCK_GROUP_RAID0) + sub_stripes = 1; + else + sub_stripes = map->sub_stripes; + + factor = map->num_stripes / sub_stripes; + stripes_per_dev = div_u64_rem(stripe_nr_end - + stripe_nr_orig, + factor, + &remaining_stripes); + div_u64_rem(stripe_nr_end - 1, factor, &last_stripe); + last_stripe *= sub_stripes; + } + for (i = 0; i < num_stripes; i++) { - multi->stripes[i].physical = + bbio->stripes[i].physical = map->stripes[stripe_index].physical + stripe_offset + stripe_nr * map->stripe_len; - multi->stripes[i].dev = map->stripes[stripe_index].dev; + bbio->stripes[i].dev = map->stripes[stripe_index].dev; - if (map->type & BTRFS_BLOCK_GROUP_RAID0) { - u64 stripes; - u32 last_stripe = 0; - int j; + if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | + BTRFS_BLOCK_GROUP_RAID10)) { + bbio->stripes[i].length = stripes_per_dev * + map->stripe_len; - div_u64_rem(stripe_nr_end - 1, - map->num_stripes, - &last_stripe); - - for (j = 0; j < map->num_stripes; j++) { - u32 test; - - div_u64_rem(stripe_nr_end - 1 - j, - map->num_stripes, &test); - if (test == stripe_index) - break; - } - stripes = stripe_nr_end - 1 - j; - do_div(stripes, map->num_stripes); - multi->stripes[i].length = map->stripe_len * - (stripes - stripe_nr + 1); - - if (i == 0) { - multi->stripes[i].length -= - stripe_offset; - stripe_offset = 0; - } - if (stripe_index == last_stripe) - multi->stripes[i].length -= - stripe_end_offset; - } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { - u64 stripes; - int j; - int factor = map->num_stripes / - map->sub_stripes; - u32 last_stripe = 0; - - div_u64_rem(stripe_nr_end - 1, - factor, &last_stripe); - last_stripe *= map->sub_stripes; - - for (j = 0; j < factor; j++) { - u32 test; - - div_u64_rem(stripe_nr_end - 1 - j, - factor, &test); - - if (test == - stripe_index / map->sub_stripes) - break; - } - stripes = stripe_nr_end - 1 - j; - do_div(stripes, factor); - multi->stripes[i].length = map->stripe_len * - (stripes - stripe_nr + 1); - - if (i < map->sub_stripes) { - multi->stripes[i].length -= + if (i / sub_stripes < remaining_stripes) + bbio->stripes[i].length += + map->stripe_len; + + /* + * Special for the first stripe and + * the last stripe: + * + * |-------|...|-------| + * |----------| + * off end_off + */ + if (i < sub_stripes) + bbio->stripes[i].length -= stripe_offset; - if (i == map->sub_stripes - 1) - stripe_offset = 0; - } + if (stripe_index >= last_stripe && stripe_index <= (last_stripe + - map->sub_stripes - 1)) { - multi->stripes[i].length -= + sub_stripes - 1)) + bbio->stripes[i].length -= stripe_end_offset; - } + + if (i == sub_stripes - 1) + stripe_offset = 0; } else - multi->stripes[i].length = *length; + bbio->stripes[i].length = *length; stripe_index++; if (stripe_index == map->num_stripes) { @@ -3099,30 +3942,38 @@ again: } } else { for (i = 0; i < num_stripes; i++) { - multi->stripes[i].physical = + bbio->stripes[i].physical = map->stripes[stripe_index].physical + stripe_offset + stripe_nr * map->stripe_len; - multi->stripes[i].dev = + bbio->stripes[i].dev = map->stripes[stripe_index].dev; stripe_index++; } } - if (multi_ret) { - *multi_ret = multi; - multi->num_stripes = num_stripes; - multi->max_errors = max_errors; + + if (rw & REQ_WRITE) { + if (map->type & (BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID10 | + BTRFS_BLOCK_GROUP_DUP)) { + max_errors = 1; + } } + + *bbio_ret = bbio; + bbio->num_stripes = num_stripes; + bbio->max_errors = max_errors; + bbio->mirror_num = mirror_num; out: free_extent_map(em); - return 0; + return ret; } int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 logical, u64 *length, - struct btrfs_multi_bio **multi_ret, int mirror_num) + struct btrfs_bio **bbio_ret, int mirror_num) { - return __btrfs_map_block(map_tree, rw, logical, length, multi_ret, + return __btrfs_map_block(map_tree, rw, logical, length, bbio_ret, mirror_num); } @@ -3153,7 +4004,7 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, do_div(length, map->num_stripes); buf = kzalloc(sizeof(u64) * map->num_stripes, GFP_NOFS); - BUG_ON(!buf); + BUG_ON(!buf); /* -ENOMEM */ for (i = 0; i < map->num_stripes; i++) { if (devid && map->stripes[i].dev->devid != devid) @@ -3191,30 +4042,79 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, return 0; } -static void end_bio_multi_stripe(struct bio *bio, int err) +static void *merge_stripe_index_into_bio_private(void *bi_private, + unsigned int stripe_index) { - struct btrfs_multi_bio *multi = bio->bi_private; + /* + * with single, dup, RAID0, RAID1 and RAID10, stripe_index is + * at most 1. + * The alternative solution (instead of stealing bits from the + * pointer) would be to allocate an intermediate structure + * that contains the old private pointer plus the stripe_index. + */ + BUG_ON((((uintptr_t)bi_private) & 3) != 0); + BUG_ON(stripe_index > 3); + return (void *)(((uintptr_t)bi_private) | stripe_index); +} + +static struct btrfs_bio *extract_bbio_from_bio_private(void *bi_private) +{ + return (struct btrfs_bio *)(((uintptr_t)bi_private) & ~((uintptr_t)3)); +} + +static unsigned int extract_stripe_index_from_bio_private(void *bi_private) +{ + return (unsigned int)((uintptr_t)bi_private) & 3; +} + +static void btrfs_end_bio(struct bio *bio, int err) +{ + struct btrfs_bio *bbio = extract_bbio_from_bio_private(bio->bi_private); int is_orig_bio = 0; - if (err) - atomic_inc(&multi->error); + if (err) { + atomic_inc(&bbio->error); + if (err == -EIO || err == -EREMOTEIO) { + unsigned int stripe_index = + extract_stripe_index_from_bio_private( + bio->bi_private); + struct btrfs_device *dev; + + BUG_ON(stripe_index >= bbio->num_stripes); + dev = bbio->stripes[stripe_index].dev; + if (dev->bdev) { + if (bio->bi_rw & WRITE) + btrfs_dev_stat_inc(dev, + BTRFS_DEV_STAT_WRITE_ERRS); + else + btrfs_dev_stat_inc(dev, + BTRFS_DEV_STAT_READ_ERRS); + if ((bio->bi_rw & WRITE_FLUSH) == WRITE_FLUSH) + btrfs_dev_stat_inc(dev, + BTRFS_DEV_STAT_FLUSH_ERRS); + btrfs_dev_stat_print_on_error(dev); + } + } + } - if (bio == multi->orig_bio) + if (bio == bbio->orig_bio) is_orig_bio = 1; - if (atomic_dec_and_test(&multi->stripes_pending)) { + if (atomic_dec_and_test(&bbio->stripes_pending)) { if (!is_orig_bio) { bio_put(bio); - bio = multi->orig_bio; + bio = bbio->orig_bio; } - bio->bi_private = multi->private; - bio->bi_end_io = multi->end_io; + bio->bi_private = bbio->private; + bio->bi_end_io = bbio->end_io; + bio->bi_bdev = (struct block_device *) + (unsigned long)bbio->mirror_num; /* only send an error to the higher layers if it is * beyond the tolerance of the multi-bio */ - if (atomic_read(&multi->error) > multi->max_errors) { + if (atomic_read(&bbio->error) > bbio->max_errors) { err = -EIO; - } else if (err) { + } else { /* * this bio is actually up to date, we didn't * go over the max number of errors @@ -3222,7 +4122,7 @@ static void end_bio_multi_stripe(struct bio *bio, int err) set_bit(BIO_UPTODATE, &bio->bi_flags); err = 0; } - kfree(multi); + kfree(bbio); bio_endio(bio, err); } else if (!is_orig_bio) { @@ -3244,7 +4144,7 @@ struct async_sched { * This will add one bio to the pending list for a device and make sure * the work struct is scheduled. */ -static noinline int schedule_bio(struct btrfs_root *root, +static noinline void schedule_bio(struct btrfs_root *root, struct btrfs_device *device, int rw, struct bio *bio) { @@ -3254,9 +4154,9 @@ static noinline int schedule_bio(struct btrfs_root *root, /* don't bother with additional async steps for reads, right now */ if (!(rw & REQ_WRITE)) { bio_get(bio); - submit_bio(rw, bio); + btrfsic_submit_bio(rw, bio); bio_put(bio); - return 0; + return; } /* @@ -3290,7 +4190,6 @@ static noinline int schedule_bio(struct btrfs_root *root, if (should_queue) btrfs_queue_worker(&root->fs_info->submit_workers, &device->work); - return 0; } int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, @@ -3302,20 +4201,21 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, u64 logical = (u64)bio->bi_sector << 9; u64 length = 0; u64 map_length; - struct btrfs_multi_bio *multi = NULL; int ret; int dev_nr = 0; int total_devs = 1; + struct btrfs_bio *bbio = NULL; length = bio->bi_size; map_tree = &root->fs_info->mapping_tree; map_length = length; - ret = btrfs_map_block(map_tree, rw, logical, &map_length, &multi, + ret = btrfs_map_block(map_tree, rw, logical, &map_length, &bbio, mirror_num); - BUG_ON(ret); + if (ret) /* -ENOMEM */ + return ret; - total_devs = multi->num_stripes; + total_devs = bbio->num_stripes; if (map_length < length) { printk(KERN_CRIT "mapping failed logical %llu bio len %llu " "len %llu\n", (unsigned long long)logical, @@ -3323,30 +4223,42 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, (unsigned long long)map_length); BUG(); } - multi->end_io = first_bio->bi_end_io; - multi->private = first_bio->bi_private; - multi->orig_bio = first_bio; - atomic_set(&multi->stripes_pending, multi->num_stripes); + + bbio->orig_bio = first_bio; + bbio->private = first_bio->bi_private; + bbio->end_io = first_bio->bi_end_io; + atomic_set(&bbio->stripes_pending, bbio->num_stripes); while (dev_nr < total_devs) { - if (total_devs > 1) { - if (dev_nr < total_devs - 1) { - bio = bio_clone(first_bio, GFP_NOFS); - BUG_ON(!bio); - } else { - bio = first_bio; - } - bio->bi_private = multi; - bio->bi_end_io = end_bio_multi_stripe; + if (dev_nr < total_devs - 1) { + bio = bio_clone(first_bio, GFP_NOFS); + BUG_ON(!bio); /* -ENOMEM */ + } else { + bio = first_bio; } - bio->bi_sector = multi->stripes[dev_nr].physical >> 9; - dev = multi->stripes[dev_nr].dev; + bio->bi_private = bbio; + bio->bi_private = merge_stripe_index_into_bio_private( + bio->bi_private, (unsigned int)dev_nr); + bio->bi_end_io = btrfs_end_bio; + bio->bi_sector = bbio->stripes[dev_nr].physical >> 9; + dev = bbio->stripes[dev_nr].dev; if (dev && dev->bdev && (rw != WRITE || dev->writeable)) { +#ifdef DEBUG + struct rcu_string *name; + + rcu_read_lock(); + name = rcu_dereference(dev->name); + pr_debug("btrfs_map_bio: rw %d, secor=%llu, dev=%lu " + "(%s id %llu), size=%u\n", rw, + (u64)bio->bi_sector, (u_long)dev->bdev->bd_dev, + name->str, dev->devid, bio->bi_size); + rcu_read_unlock(); +#endif bio->bi_bdev = dev->bdev; if (async_submit) schedule_bio(root, dev, rw, bio); else - submit_bio(rw, bio); + btrfsic_submit_bio(rw, bio); } else { bio->bi_bdev = root->fs_info->fs_devices->latest_bdev; bio->bi_sector = logical >> 9; @@ -3354,8 +4266,6 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, } dev_nr++; } - if (total_devs == 1) - kfree(multi); return 0; } @@ -3485,13 +4395,13 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, write_lock(&map_tree->map_tree.lock); ret = add_extent_mapping(&map_tree->map_tree, em); write_unlock(&map_tree->map_tree.lock); - BUG_ON(ret); + BUG_ON(ret); /* Tree corruption */ free_extent_map(em); return 0; } -static int fill_device_from_item(struct extent_buffer *leaf, +static void fill_device_from_item(struct extent_buffer *leaf, struct btrfs_dev_item *dev_item, struct btrfs_device *device) { @@ -3508,8 +4418,6 @@ static int fill_device_from_item(struct extent_buffer *leaf, ptr = (unsigned long)btrfs_device_uuid(dev_item); read_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE); - - return 0; } static int open_seed_devices(struct btrfs_root *root, u8 *fsid) @@ -3517,7 +4425,7 @@ static int open_seed_devices(struct btrfs_root *root, u8 *fsid) struct btrfs_fs_devices *fs_devices; int ret; - mutex_lock(&uuid_mutex); + BUG_ON(!mutex_is_locked(&uuid_mutex)); fs_devices = root->fs_info->fs_devices->seed; while (fs_devices) { @@ -3542,8 +4450,10 @@ static int open_seed_devices(struct btrfs_root *root, u8 *fsid) ret = __btrfs_open_devices(fs_devices, FMODE_READ, root->fs_info->bdev_holder); - if (ret) + if (ret) { + free_fs_devices(fs_devices); goto out; + } if (!fs_devices->seeding) { __btrfs_close_devices(fs_devices); @@ -3555,7 +4465,6 @@ static int open_seed_devices(struct btrfs_root *root, u8 *fsid) fs_devices->seed = root->fs_info->fs_devices->seed; root->fs_info->fs_devices->seed = fs_devices; out: - mutex_unlock(&uuid_mutex); return ret; } @@ -3616,15 +4525,20 @@ static int read_one_dev(struct btrfs_root *root, fill_device_from_item(leaf, dev_item, device); device->dev_root = root->fs_info->dev_root; device->in_fs_metadata = 1; - if (device->writeable) + if (device->writeable) { device->fs_devices->total_rw_bytes += device->total_bytes; + spin_lock(&root->fs_info->free_chunk_lock); + root->fs_info->free_chunk_space += device->total_bytes - + device->bytes_used; + spin_unlock(&root->fs_info->free_chunk_lock); + } ret = 0; return ret; } int btrfs_read_sys_array(struct btrfs_root *root) { - struct btrfs_super_block *super_copy = &root->fs_info->super_copy; + struct btrfs_super_block *super_copy = root->fs_info->super_copy; struct extent_buffer *sb; struct btrfs_disk_key *disk_key; struct btrfs_chunk *chunk; @@ -3643,6 +4557,20 @@ int btrfs_read_sys_array(struct btrfs_root *root) return -ENOMEM; btrfs_set_buffer_uptodate(sb); btrfs_set_buffer_lockdep_class(root->root_key.objectid, sb, 0); + /* + * The sb extent buffer is artifical and just used to read the system array. + * btrfs_set_buffer_uptodate() call does not properly mark all it's + * pages up-to-date when the page is larger: extent does not cover the + * whole page and consequently check_page_uptodate does not find all + * the page's extents up-to-date (the hole beyond sb), + * write_extent_buffer then triggers a WARN_ON. + * + * Regular short extents go through mark_extent_buffer_dirty/writeback cycle, + * but sb spans only this function. Add an explicit SetPageUptodate call + * to silence the warning eg. on PowerPC 64. + */ + if (PAGE_CACHE_SIZE > BTRFS_SUPER_INFO_SIZE) + SetPageUptodate(sb->pages[0]); write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE); array_size = btrfs_super_sys_array_size(super_copy); @@ -3693,6 +4621,9 @@ int btrfs_read_chunk_tree(struct btrfs_root *root) if (!path) return -ENOMEM; + mutex_lock(&uuid_mutex); + lock_chunks(root); + /* first we search for all of the device items, and then we * read in all of the chunk items. This way we can create chunk * mappings that reference all of the devices that are afound @@ -3743,6 +4674,241 @@ again: } ret = 0; error: + unlock_chunks(root); + mutex_unlock(&uuid_mutex); + + btrfs_free_path(path); + return ret; +} + +static void __btrfs_reset_dev_stats(struct btrfs_device *dev) +{ + int i; + + for (i = 0; i < BTRFS_DEV_STAT_VALUES_MAX; i++) + btrfs_dev_stat_reset(dev, i); +} + +int btrfs_init_dev_stats(struct btrfs_fs_info *fs_info) +{ + struct btrfs_key key; + struct btrfs_key found_key; + struct btrfs_root *dev_root = fs_info->dev_root; + struct btrfs_fs_devices *fs_devices = fs_info->fs_devices; + struct extent_buffer *eb; + int slot; + int ret = 0; + struct btrfs_device *device; + struct btrfs_path *path = NULL; + int i; + + path = btrfs_alloc_path(); + if (!path) { + ret = -ENOMEM; + goto out; + } + + mutex_lock(&fs_devices->device_list_mutex); + list_for_each_entry(device, &fs_devices->devices, dev_list) { + int item_size; + struct btrfs_dev_stats_item *ptr; + + key.objectid = 0; + key.type = BTRFS_DEV_STATS_KEY; + key.offset = device->devid; + ret = btrfs_search_slot(NULL, dev_root, &key, path, 0, 0); + if (ret) { + __btrfs_reset_dev_stats(device); + device->dev_stats_valid = 1; + btrfs_release_path(path); + continue; + } + slot = path->slots[0]; + eb = path->nodes[0]; + btrfs_item_key_to_cpu(eb, &found_key, slot); + item_size = btrfs_item_size_nr(eb, slot); + + ptr = btrfs_item_ptr(eb, slot, + struct btrfs_dev_stats_item); + + for (i = 0; i < BTRFS_DEV_STAT_VALUES_MAX; i++) { + if (item_size >= (1 + i) * sizeof(__le64)) + btrfs_dev_stat_set(device, i, + btrfs_dev_stats_value(eb, ptr, i)); + else + btrfs_dev_stat_reset(device, i); + } + + device->dev_stats_valid = 1; + btrfs_dev_stat_print_on_load(device); + btrfs_release_path(path); + } + mutex_unlock(&fs_devices->device_list_mutex); + +out: btrfs_free_path(path); + return ret < 0 ? ret : 0; +} + +static int update_dev_stat_item(struct btrfs_trans_handle *trans, + struct btrfs_root *dev_root, + struct btrfs_device *device) +{ + struct btrfs_path *path; + struct btrfs_key key; + struct extent_buffer *eb; + struct btrfs_dev_stats_item *ptr; + int ret; + int i; + + key.objectid = 0; + key.type = BTRFS_DEV_STATS_KEY; + key.offset = device->devid; + + path = btrfs_alloc_path(); + BUG_ON(!path); + ret = btrfs_search_slot(trans, dev_root, &key, path, -1, 1); + if (ret < 0) { + printk_in_rcu(KERN_WARNING "btrfs: error %d while searching for dev_stats item for device %s!\n", + ret, rcu_str_deref(device->name)); + goto out; + } + + if (ret == 0 && + btrfs_item_size_nr(path->nodes[0], path->slots[0]) < sizeof(*ptr)) { + /* need to delete old one and insert a new one */ + ret = btrfs_del_item(trans, dev_root, path); + if (ret != 0) { + printk_in_rcu(KERN_WARNING "btrfs: delete too small dev_stats item for device %s failed %d!\n", + rcu_str_deref(device->name), ret); + goto out; + } + ret = 1; + } + + if (ret == 1) { + /* need to insert a new item */ + btrfs_release_path(path); + ret = btrfs_insert_empty_item(trans, dev_root, path, + &key, sizeof(*ptr)); + if (ret < 0) { + printk_in_rcu(KERN_WARNING "btrfs: insert dev_stats item for device %s failed %d!\n", + rcu_str_deref(device->name), ret); + goto out; + } + } + + eb = path->nodes[0]; + ptr = btrfs_item_ptr(eb, path->slots[0], struct btrfs_dev_stats_item); + for (i = 0; i < BTRFS_DEV_STAT_VALUES_MAX; i++) + btrfs_set_dev_stats_value(eb, ptr, i, + btrfs_dev_stat_read(device, i)); + btrfs_mark_buffer_dirty(eb); + +out: + btrfs_free_path(path); + return ret; +} + +/* + * called from commit_transaction. Writes all changed device stats to disk. + */ +int btrfs_run_dev_stats(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info) +{ + struct btrfs_root *dev_root = fs_info->dev_root; + struct btrfs_fs_devices *fs_devices = fs_info->fs_devices; + struct btrfs_device *device; + int ret = 0; + + mutex_lock(&fs_devices->device_list_mutex); + list_for_each_entry(device, &fs_devices->devices, dev_list) { + if (!device->dev_stats_valid || !device->dev_stats_dirty) + continue; + + ret = update_dev_stat_item(trans, dev_root, device); + if (!ret) + device->dev_stats_dirty = 0; + } + mutex_unlock(&fs_devices->device_list_mutex); + return ret; } + +void btrfs_dev_stat_inc_and_print(struct btrfs_device *dev, int index) +{ + btrfs_dev_stat_inc(dev, index); + btrfs_dev_stat_print_on_error(dev); +} + +void btrfs_dev_stat_print_on_error(struct btrfs_device *dev) +{ + if (!dev->dev_stats_valid) + return; + printk_ratelimited_in_rcu(KERN_ERR + "btrfs: bdev %s errs: wr %u, rd %u, flush %u, corrupt %u, gen %u\n", + rcu_str_deref(dev->name), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_WRITE_ERRS), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_READ_ERRS), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_FLUSH_ERRS), + btrfs_dev_stat_read(dev, + BTRFS_DEV_STAT_CORRUPTION_ERRS), + btrfs_dev_stat_read(dev, + BTRFS_DEV_STAT_GENERATION_ERRS)); +} + +static void btrfs_dev_stat_print_on_load(struct btrfs_device *dev) +{ + int i; + + for (i = 0; i < BTRFS_DEV_STAT_VALUES_MAX; i++) + if (btrfs_dev_stat_read(dev, i) != 0) + break; + if (i == BTRFS_DEV_STAT_VALUES_MAX) + return; /* all values == 0, suppress message */ + + printk_in_rcu(KERN_INFO "btrfs: bdev %s errs: wr %u, rd %u, flush %u, corrupt %u, gen %u\n", + rcu_str_deref(dev->name), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_WRITE_ERRS), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_READ_ERRS), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_FLUSH_ERRS), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_CORRUPTION_ERRS), + btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_GENERATION_ERRS)); +} + +int btrfs_get_dev_stats(struct btrfs_root *root, + struct btrfs_ioctl_get_dev_stats *stats) +{ + struct btrfs_device *dev; + struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; + int i; + + mutex_lock(&fs_devices->device_list_mutex); + dev = btrfs_find_device(root, stats->devid, NULL, NULL); + mutex_unlock(&fs_devices->device_list_mutex); + + if (!dev) { + printk(KERN_WARNING + "btrfs: get dev_stats failed, device not found\n"); + return -ENODEV; + } else if (!dev->dev_stats_valid) { + printk(KERN_WARNING + "btrfs: get dev_stats failed, not yet valid\n"); + return -ENODEV; + } else if (stats->flags & BTRFS_DEV_STATS_RESET) { + for (i = 0; i < BTRFS_DEV_STAT_VALUES_MAX; i++) { + if (stats->nr_items > i) + stats->values[i] = + btrfs_dev_stat_read_and_reset(dev, i); + else + btrfs_dev_stat_reset(dev, i); + } + } else { + for (i = 0; i < BTRFS_DEV_STAT_VALUES_MAX; i++) + if (stats->nr_items > i) + stats->values[i] = btrfs_dev_stat_read(dev, i); + } + if (stats->nr_items > BTRFS_DEV_STAT_VALUES_MAX) + stats->nr_items = BTRFS_DEV_STAT_VALUES_MAX; + return 0; +} diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 6d866db4e177..53c06af92e8d 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -22,6 +22,7 @@ #include <linux/bio.h> #include <linux/sort.h> #include "async-thread.h" +#include "ioctl.h" #define BTRFS_STRIPE_LEN (64 * 1024) @@ -57,7 +58,7 @@ struct btrfs_device { /* the mode sent to blkdev_get */ fmode_t mode; - char *name; + struct rcu_string *name; /* the internal btrfs device id */ u64 devid; @@ -92,6 +93,25 @@ struct btrfs_device { struct btrfs_work work; struct rcu_head rcu; struct work_struct rcu_work; + + /* readahead state */ + spinlock_t reada_lock; + atomic_t reada_in_flight; + u64 reada_next; + struct reada_zone *reada_curr_zone; + struct radix_tree_root reada_zones; + struct radix_tree_root reada_extents; + + /* for sending down flush barriers */ + struct bio *flush_bio; + struct completion flush_wait; + int nobarriers; + + /* disk I/O failure stats. For detailed description refer to + * enum btrfs_dev_stat_values in ioctl.h */ + int dev_stats_valid; + int dev_stats_dirty; /* counters need to be written to disk */ + atomic_t dev_stat_values[BTRFS_DEV_STAT_VALUES_MAX]; }; struct btrfs_fs_devices { @@ -106,6 +126,7 @@ struct btrfs_fs_devices { u64 missing_devices; u64 total_rw_bytes; u64 num_can_discard; + u64 total_devices; struct block_device *latest_bdev; /* all of the devices in the FS, protected by a mutex @@ -136,7 +157,10 @@ struct btrfs_bio_stripe { u64 length; /* only used for discard mappings */ }; -struct btrfs_multi_bio { +struct btrfs_bio; +typedef void (btrfs_bio_end_io_t) (struct btrfs_bio *bio, int err); + +struct btrfs_bio { atomic_t stripes_pending; bio_end_io_t *end_io; struct bio *orig_bio; @@ -144,6 +168,7 @@ struct btrfs_multi_bio { atomic_t error; int max_errors; int num_stripes; + int mirror_num; struct btrfs_bio_stripe stripes[]; }; @@ -168,10 +193,55 @@ struct map_lookup { #define map_lookup_size(n) (sizeof(struct map_lookup) + \ (sizeof(struct btrfs_bio_stripe) * (n))) +/* + * Restriper's general type filter + */ +#define BTRFS_BALANCE_DATA (1ULL << 0) +#define BTRFS_BALANCE_SYSTEM (1ULL << 1) +#define BTRFS_BALANCE_METADATA (1ULL << 2) + +#define BTRFS_BALANCE_TYPE_MASK (BTRFS_BALANCE_DATA | \ + BTRFS_BALANCE_SYSTEM | \ + BTRFS_BALANCE_METADATA) + +#define BTRFS_BALANCE_FORCE (1ULL << 3) +#define BTRFS_BALANCE_RESUME (1ULL << 4) + +/* + * Balance filters + */ +#define BTRFS_BALANCE_ARGS_PROFILES (1ULL << 0) +#define BTRFS_BALANCE_ARGS_USAGE (1ULL << 1) +#define BTRFS_BALANCE_ARGS_DEVID (1ULL << 2) +#define BTRFS_BALANCE_ARGS_DRANGE (1ULL << 3) +#define BTRFS_BALANCE_ARGS_VRANGE (1ULL << 4) + +/* + * Profile changing flags. When SOFT is set we won't relocate chunk if + * it already has the target profile (even though it may be + * half-filled). + */ +#define BTRFS_BALANCE_ARGS_CONVERT (1ULL << 8) +#define BTRFS_BALANCE_ARGS_SOFT (1ULL << 9) + +struct btrfs_balance_args; +struct btrfs_balance_progress; +struct btrfs_balance_control { + struct btrfs_fs_info *fs_info; + + struct btrfs_balance_args data; + struct btrfs_balance_args meta; + struct btrfs_balance_args sys; + + u64 flags; + + struct btrfs_balance_progress stat; +}; + int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start, u64 end, u64 *length); -#define btrfs_multi_bio_size(n) (sizeof(struct btrfs_multi_bio) + \ +#define btrfs_bio_size(n) (sizeof(struct btrfs_bio) + \ (sizeof(struct btrfs_bio_stripe) * (n))) int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans, @@ -180,7 +250,7 @@ int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans, u64 chunk_offset, u64 start, u64 num_bytes); int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, u64 logical, u64 *length, - struct btrfs_multi_bio **multi_ret, int mirror_num); + struct btrfs_bio **bbio_ret, int mirror_num); int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, u64 chunk_start, u64 physical, u64 devid, u64 **logical, int *naddrs, int *stripe_len); @@ -197,12 +267,12 @@ int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, struct btrfs_fs_devices **fs_devices_ret); int btrfs_close_devices(struct btrfs_fs_devices *fs_devices); -int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices); +void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices); int btrfs_add_device(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_device *device); int btrfs_rm_device(struct btrfs_root *root, char *device_path); -int btrfs_cleanup_fs_uuids(void); +void btrfs_cleanup_fs_uuids(void); int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len); int btrfs_grow_device(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 new_size); @@ -210,9 +280,56 @@ struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, u8 *uuid, u8 *fsid); int btrfs_shrink_device(struct btrfs_device *device, u64 new_size); int btrfs_init_new_device(struct btrfs_root *root, char *path); -int btrfs_balance(struct btrfs_root *dev_root); +int btrfs_balance(struct btrfs_balance_control *bctl, + struct btrfs_ioctl_balance_args *bargs); +int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info); +int btrfs_recover_balance(struct btrfs_fs_info *fs_info); +int btrfs_pause_balance(struct btrfs_fs_info *fs_info); +int btrfs_cancel_balance(struct btrfs_fs_info *fs_info); int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset); -int find_free_dev_extent(struct btrfs_trans_handle *trans, - struct btrfs_device *device, u64 num_bytes, +int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes, u64 *start, u64 *max_avail); +void btrfs_dev_stat_print_on_error(struct btrfs_device *device); +void btrfs_dev_stat_inc_and_print(struct btrfs_device *dev, int index); +int btrfs_get_dev_stats(struct btrfs_root *root, + struct btrfs_ioctl_get_dev_stats *stats); +int btrfs_init_dev_stats(struct btrfs_fs_info *fs_info); +int btrfs_run_dev_stats(struct btrfs_trans_handle *trans, + struct btrfs_fs_info *fs_info); + +static inline void btrfs_dev_stat_inc(struct btrfs_device *dev, + int index) +{ + atomic_inc(dev->dev_stat_values + index); + dev->dev_stats_dirty = 1; +} + +static inline int btrfs_dev_stat_read(struct btrfs_device *dev, + int index) +{ + return atomic_read(dev->dev_stat_values + index); +} + +static inline int btrfs_dev_stat_read_and_reset(struct btrfs_device *dev, + int index) +{ + int ret; + + ret = atomic_xchg(dev->dev_stat_values + index, 0); + dev->dev_stats_dirty = 1; + return ret; +} + +static inline void btrfs_dev_stat_set(struct btrfs_device *dev, + int index, unsigned long val) +{ + atomic_set(dev->dev_stat_values + index, val); + dev->dev_stats_dirty = 1; +} + +static inline void btrfs_dev_stat_reset(struct btrfs_device *dev, + int index) +{ + btrfs_dev_stat_set(dev, index, 0); +} #endif diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index 69565e5fc6a0..3f4e2d69e83a 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -127,6 +127,17 @@ static int do_setxattr(struct btrfs_trans_handle *trans, again: ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode), name, name_len, value, size); + /* + * If we're setting an xattr to a new value but the new value is say + * exactly BTRFS_MAX_XATTR_SIZE, we could end up with EOVERFLOW getting + * back from split_leaf. This is because it thinks we'll be extending + * the existing item size, but we're asking for enough space to add the + * item itself. So if we get EOVERFLOW just set ret to EEXIST and let + * the rest of the function figure it out. + */ + if (ret == -EOVERFLOW) + ret = -EEXIST; + if (ret == -EEXIST) { if (flags & XATTR_CREATE) goto out; @@ -185,11 +196,12 @@ int __btrfs_setxattr(struct btrfs_trans_handle *trans, if (ret) goto out; + inode_inc_iversion(inode); inode->i_ctime = CURRENT_TIME; ret = btrfs_update_inode(trans, root, inode); BUG_ON(ret); out: - btrfs_end_transaction_throttle(trans, root); + btrfs_end_transaction(trans, root); return ret; } @@ -383,36 +395,36 @@ int btrfs_removexattr(struct dentry *dentry, const char *name) XATTR_REPLACE); } -int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, - struct inode *inode, struct inode *dir, - const struct qstr *qstr) +int btrfs_initxattrs(struct inode *inode, const struct xattr *xattr_array, + void *fs_info) { - int err; - size_t len; - void *value; - char *suffix; + const struct xattr *xattr; + struct btrfs_trans_handle *trans = fs_info; char *name; + int err = 0; - err = security_inode_init_security(inode, dir, qstr, &suffix, &value, - &len); - if (err) { - if (err == -EOPNOTSUPP) - return 0; - return err; - } - - name = kmalloc(XATTR_SECURITY_PREFIX_LEN + strlen(suffix) + 1, - GFP_NOFS); - if (!name) { - err = -ENOMEM; - } else { + for (xattr = xattr_array; xattr->name != NULL; xattr++) { + name = kmalloc(XATTR_SECURITY_PREFIX_LEN + + strlen(xattr->name) + 1, GFP_NOFS); + if (!name) { + err = -ENOMEM; + break; + } strcpy(name, XATTR_SECURITY_PREFIX); - strcpy(name + XATTR_SECURITY_PREFIX_LEN, suffix); - err = __btrfs_setxattr(trans, inode, name, value, len, 0); + strcpy(name + XATTR_SECURITY_PREFIX_LEN, xattr->name); + err = __btrfs_setxattr(trans, inode, name, + xattr->value, xattr->value_len, 0); kfree(name); + if (err < 0) + break; } - - kfree(suffix); - kfree(value); return err; } + +int btrfs_xattr_security_init(struct btrfs_trans_handle *trans, + struct inode *inode, struct inode *dir, + const struct qstr *qstr) +{ + return security_inode_init_security(inode, dir, qstr, + &btrfs_initxattrs, trans); +} diff --git a/fs/btrfs/zlib.c b/fs/btrfs/zlib.c index faccd47c6c46..92c20654cc55 100644 --- a/fs/btrfs/zlib.c +++ b/fs/btrfs/zlib.c @@ -370,9 +370,9 @@ static int zlib_decompress(struct list_head *ws, unsigned char *data_in, PAGE_CACHE_SIZE - buf_offset); bytes = min(bytes, bytes_left); - kaddr = kmap_atomic(dest_page, KM_USER0); + kaddr = kmap_atomic(dest_page); memcpy(kaddr + pg_offset, workspace->buf + buf_offset, bytes); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); pg_offset += bytes; bytes_left -= bytes; |