summaryrefslogtreecommitdiff
path: root/fs/btrfs/disk-io.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/btrfs/disk-io.c')
-rw-r--r--fs/btrfs/disk-io.c135
1 files changed, 46 insertions, 89 deletions
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 85b207d19aa5..78722aaffecd 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -49,6 +49,7 @@
#include "raid56.h"
#include "sysfs.h"
#include "qgroup.h"
+#include "tree-checker.h"
#ifdef CONFIG_X86
#include <asm/cpufeature.h>
@@ -445,9 +446,9 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
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) {
+ clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
ret = read_extent_buffer_pages(io_tree, eb, start,
WAIT_COMPLETE,
btree_get_extent, mirror_num);
@@ -459,14 +460,6 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
ret = -EIO;
}
- /*
- * This buffer's crc is fine, but its contents are corrupted, so
- * there is no reason to read the other copies, they won't be
- * any less wrong.
- */
- if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags))
- break;
-
num_copies = btrfs_num_copies(root->fs_info,
eb->start, eb->len);
if (num_copies == 1)
@@ -530,72 +523,6 @@ static int check_tree_block_fsid(struct btrfs_fs_info *fs_info,
return ret;
}
-#define CORRUPT(reason, eb, root, slot) \
- btrfs_crit(root->fs_info, "corrupt leaf, %s: block=%llu," \
- "root=%llu, slot=%d", reason, \
- btrfs_header_bytenr(eb), root->objectid, slot)
-
-static noinline int check_leaf(struct btrfs_root *root,
- struct extent_buffer *leaf)
-{
- struct btrfs_key key;
- struct btrfs_key leaf_key;
- u32 nritems = btrfs_header_nritems(leaf);
- int slot;
-
- if (nritems == 0)
- return 0;
-
- /* Check the 0 item */
- if (btrfs_item_offset_nr(leaf, 0) + btrfs_item_size_nr(leaf, 0) !=
- BTRFS_LEAF_DATA_SIZE(root)) {
- CORRUPT("invalid item offset size pair", leaf, root, 0);
- return -EIO;
- }
-
- /*
- * Check to make sure each items keys are in the correct order and their
- * offsets make sense. We only have to loop through nritems-1 because
- * we check the current slot against the next slot, which verifies the
- * next slot's offset+size makes sense and that the current's slot
- * offset is correct.
- */
- for (slot = 0; slot < nritems - 1; slot++) {
- btrfs_item_key_to_cpu(leaf, &leaf_key, slot);
- btrfs_item_key_to_cpu(leaf, &key, slot + 1);
-
- /* Make sure the keys are in the right order */
- if (btrfs_comp_cpu_keys(&leaf_key, &key) >= 0) {
- CORRUPT("bad key order", leaf, root, slot);
- return -EIO;
- }
-
- /*
- * Make sure the offset and ends are right, remember that the
- * item data starts at the end of the leaf and grows towards the
- * front.
- */
- if (btrfs_item_offset_nr(leaf, slot) !=
- btrfs_item_end_nr(leaf, slot + 1)) {
- CORRUPT("slot offset bad", leaf, root, slot);
- return -EIO;
- }
-
- /*
- * Check to make sure that we don't point outside of the leaf,
- * just incase all the items are consistent to eachother, but
- * all point outside of the leaf.
- */
- if (btrfs_item_end_nr(leaf, slot) >
- BTRFS_LEAF_DATA_SIZE(root)) {
- CORRUPT("slot end outside of leaf", leaf, root, slot);
- return -EIO;
- }
- }
-
- return 0;
-}
-
static int btree_readpage_end_io_hook(struct btrfs_io_bio *io_bio,
u64 phy_offset, struct page *page,
u64 start, u64 end, int mirror)
@@ -662,11 +589,14 @@ static int btree_readpage_end_io_hook(struct btrfs_io_bio *io_bio,
* that we don't try and read the other copies of this block, just
* return -EIO.
*/
- if (found_level == 0 && check_leaf(root, eb)) {
+ if (found_level == 0 && btrfs_check_leaf_full(root, eb)) {
set_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
ret = -EIO;
}
+ if (found_level > 0 && btrfs_check_node(root, eb))
+ ret = -EIO;
+
if (!ret)
set_extent_buffer_uptodate(eb);
err:
@@ -923,7 +853,7 @@ static int check_async_write(struct inode *inode, unsigned long bio_flags)
if (bio_flags & EXTENT_BIO_TREE_LOG)
return 0;
#ifdef CONFIG_X86
- if (cpu_has_xmm4_2)
+ if (static_cpu_has(X86_FEATURE_XMM4_2))
return 0;
#endif
return 1;
@@ -1011,8 +941,9 @@ static int btree_writepages(struct address_space *mapping,
fs_info = BTRFS_I(mapping->host)->root->fs_info;
/* this is a bit racy, but that's ok */
- ret = percpu_counter_compare(&fs_info->dirty_metadata_bytes,
- BTRFS_DIRTY_METADATA_THRESH);
+ ret = __percpu_counter_compare(&fs_info->dirty_metadata_bytes,
+ BTRFS_DIRTY_METADATA_THRESH,
+ fs_info->dirty_metadata_batch);
if (ret < 0)
return 0;
}
@@ -1196,7 +1127,7 @@ static struct btrfs_subvolume_writers *btrfs_alloc_subvolume_writers(void)
if (!writers)
return ERR_PTR(-ENOMEM);
- ret = percpu_counter_init(&writers->counter, 0, GFP_KERNEL);
+ ret = percpu_counter_init(&writers->counter, 0, GFP_NOFS);
if (ret < 0) {
kfree(writers);
return ERR_PTR(ret);
@@ -1607,8 +1538,8 @@ fail:
return ret;
}
-static struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
- u64 root_id)
+struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
+ u64 root_id)
{
struct btrfs_root *root;
@@ -3965,7 +3896,13 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
buf->len,
root->fs_info->dirty_metadata_batch);
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
- if (btrfs_header_level(buf) == 0 && check_leaf(root, buf)) {
+ /*
+ * Since btrfs_mark_buffer_dirty() can be called with item pointer set
+ * but item data not updated.
+ * So here we should only check item pointers, not item data.
+ */
+ if (btrfs_header_level(buf) == 0 &&
+ btrfs_check_leaf_relaxed(root, buf)) {
btrfs_print_leaf(root, buf);
ASSERT(0);
}
@@ -3987,8 +3924,9 @@ static void __btrfs_btree_balance_dirty(struct btrfs_root *root,
if (flush_delayed)
btrfs_balance_delayed_items(root);
- ret = percpu_counter_compare(&root->fs_info->dirty_metadata_bytes,
- BTRFS_DIRTY_METADATA_THRESH);
+ ret = __percpu_counter_compare(&root->fs_info->dirty_metadata_bytes,
+ BTRFS_DIRTY_METADATA_THRESH,
+ root->fs_info->dirty_metadata_batch);
if (ret > 0) {
balance_dirty_pages_ratelimited(
root->fs_info->btree_inode->i_mapping);
@@ -4173,6 +4111,14 @@ static void btrfs_destroy_all_ordered_extents(struct btrfs_fs_info *fs_info)
spin_lock(&fs_info->ordered_root_lock);
}
spin_unlock(&fs_info->ordered_root_lock);
+
+ /*
+ * We need this here because if we've been flipped read-only we won't
+ * get sync() from the umount, so we need to make sure any ordered
+ * extents that haven't had their dirty pages IO start writeout yet
+ * actually get run and error out properly.
+ */
+ btrfs_wait_ordered_roots(fs_info, -1);
}
static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
@@ -4331,6 +4277,7 @@ static int btrfs_destroy_marked_extents(struct btrfs_root *root,
static int btrfs_destroy_pinned_extent(struct btrfs_root *root,
struct extent_io_tree *pinned_extents)
{
+ struct btrfs_fs_info *fs_info = root->fs_info;
struct extent_io_tree *unpin;
u64 start;
u64 end;
@@ -4340,21 +4287,31 @@ static int btrfs_destroy_pinned_extent(struct btrfs_root *root,
unpin = pinned_extents;
again:
while (1) {
+ /*
+ * The btrfs_finish_extent_commit() may get the same range as
+ * ours between find_first_extent_bit and clear_extent_dirty.
+ * Hence, hold the unused_bg_unpin_mutex to avoid double unpin
+ * the same extent range.
+ */
+ mutex_lock(&fs_info->unused_bg_unpin_mutex);
ret = find_first_extent_bit(unpin, 0, &start, &end,
EXTENT_DIRTY, NULL);
- if (ret)
+ if (ret) {
+ mutex_unlock(&fs_info->unused_bg_unpin_mutex);
break;
+ }
clear_extent_dirty(unpin, start, end, GFP_NOFS);
btrfs_error_unpin_extent_range(root, start, end);
+ mutex_unlock(&fs_info->unused_bg_unpin_mutex);
cond_resched();
}
if (loop) {
- if (unpin == &root->fs_info->freed_extents[0])
- unpin = &root->fs_info->freed_extents[1];
+ if (unpin == &fs_info->freed_extents[0])
+ unpin = &fs_info->freed_extents[1];
else
- unpin = &root->fs_info->freed_extents[0];
+ unpin = &fs_info->freed_extents[0];
loop = false;
goto again;
}
generated by cgit v1.2.3 (git 2.25.1) at 2025-10-23 12:27:13 +0000