From 1b340d0118da1d7c60c664f17d7c8fce2bb1cd9d Mon Sep 17 00:00:00 2001 From: Stanislav Kinsbursky Date: Fri, 25 Nov 2011 17:13:04 +0300 Subject: NFS: DNS resolver cache per network namespace context introduced This patch implements DNS resolver cache creation and registration for each alive network namespace context. This was done by registering NFS per-net operations, responsible for DNS cache allocation/register and unregister/destructioning instead of initialization and destruction of static "nfs_dns_resolve" cache detail (this one was removed). Pointer to network dns resolver cache is stored in new per-net "nfs_net" structure. This patch also changes nfs_dns_resolve_name() function prototype (and it's calls) by adding network pointer parameter, which is used to get proper DNS resolver cache pointer for do_cache_lookup_wait() call. Note: empty nfs_dns_resolver_init() and nfs_dns_resolver_destroy() functions will be used in next patch in the series. Signed-off-by: Stanislav Kinsbursky Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 33 ++++++++++++++++++++++++++++++--- 1 file changed, 30 insertions(+), 3 deletions(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index f649fba8c384..0335f6e4ff7e 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -51,6 +51,7 @@ #include "fscache.h" #include "dns_resolve.h" #include "pnfs.h" +#include "netns.h" #define NFSDBG_FACILITY NFSDBG_VFS @@ -1552,6 +1553,25 @@ static void nfsiod_stop(void) destroy_workqueue(wq); } +int nfs_net_id; + +static int nfs_net_init(struct net *net) +{ + return nfs_dns_resolver_cache_init(net); +} + +static void nfs_net_exit(struct net *net) +{ + nfs_dns_resolver_cache_destroy(net); +} + +static struct pernet_operations nfs_net_ops = { + .init = nfs_net_init, + .exit = nfs_net_exit, + .id = &nfs_net_id, + .size = sizeof(struct nfs_net), +}; + /* * Initialize NFS */ @@ -1561,9 +1581,13 @@ static int __init init_nfs_fs(void) err = nfs_idmap_init(); if (err < 0) - goto out9; + goto out10; err = nfs_dns_resolver_init(); + if (err < 0) + goto out9; + + err = register_pernet_subsys(&nfs_net_ops); if (err < 0) goto out8; @@ -1625,10 +1649,12 @@ out5: out6: nfs_fscache_unregister(); out7: - nfs_dns_resolver_destroy(); + unregister_pernet_subsys(&nfs_net_ops); out8: - nfs_idmap_quit(); + nfs_dns_resolver_destroy(); out9: + nfs_idmap_quit(); +out10: return err; } @@ -1640,6 +1666,7 @@ static void __exit exit_nfs_fs(void) nfs_destroy_inodecache(); nfs_destroy_nfspagecache(); nfs_fscache_unregister(); + unregister_pernet_subsys(&nfs_net_ops); nfs_dns_resolver_destroy(); nfs_idmap_quit(); #ifdef CONFIG_PROC_FS -- cgit v1.2.3 From 9e2e74dba6ddce94da187369b50a27536147d5df Mon Sep 17 00:00:00 2001 From: Stanislav Kinsbursky Date: Tue, 10 Jan 2012 17:04:24 +0400 Subject: NFS: blocklayout pipe creation per network namespace context introduced This patch implements blocklayout pipe creation and registration per each existent network namespace. This was achived by registering NFS per-net operations, responsible for blocklayout pipe allocation/register and unregister/destruction instead of initialization and destruction of static "bl_device_pipe" pipe (this one was removed). Note, than pointer to network blocklayout pipe is stored in per-net "nfs_net" structure, because allocating of one more per-net structure for blocklayout module looks redundant. This patch also changes dev_remove() function prototype (and all it's callers, where it' requied) by adding network namespace pointer parameter, which is used to discover proper blocklayout pipe for rpc_queue_upcall() call. Signed-off-by: Stanislav Kinsbursky Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 0335f6e4ff7e..577ad5a72a24 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1554,6 +1554,7 @@ static void nfsiod_stop(void) } int nfs_net_id; +EXPORT_SYMBOL_GPL(nfs_net_id); static int nfs_net_init(struct net *net) { -- cgit v1.2.3 From 536e43d12b9517bbbf6114cd1a12be27857a4d7a Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 17 Jan 2012 22:04:26 -0500 Subject: NFS: Optimise away unnecessary setattrs for open(O_TRUNC); Currently, we will correctly optimise away a truncate that doesn't change the file size. However, in the case of open(O_TRUNC), we also want to optimise away the time changes. Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 577ad5a72a24..65486e652943 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -402,7 +402,7 @@ out_no_inode: goto out; } -#define NFS_VALID_ATTRS (ATTR_MODE|ATTR_UID|ATTR_GID|ATTR_SIZE|ATTR_ATIME|ATTR_ATIME_SET|ATTR_MTIME|ATTR_MTIME_SET|ATTR_FILE) +#define NFS_VALID_ATTRS (ATTR_MODE|ATTR_UID|ATTR_GID|ATTR_SIZE|ATTR_ATIME|ATTR_ATIME_SET|ATTR_MTIME|ATTR_MTIME_SET|ATTR_FILE|ATTR_OPEN) int nfs_setattr(struct dentry *dentry, struct iattr *attr) @@ -424,7 +424,7 @@ nfs_setattr(struct dentry *dentry, struct iattr *attr) /* Optimization: if the end result is no change, don't RPC */ attr->ia_valid &= NFS_VALID_ATTRS; - if ((attr->ia_valid & ~ATTR_FILE) == 0) + if ((attr->ia_valid & ~(ATTR_FILE|ATTR_OPEN)) == 0) return 0; /* Write all dirty data */ -- cgit v1.2.3 From ec7652aaf261b7dcb368344369df1e99886c7cd2 Mon Sep 17 00:00:00 2001 From: Stanislav Kinsbursky Date: Tue, 6 Dec 2011 16:42:40 +0300 Subject: SUNRPC: register RPC stats /proc entries in passed network namespace context This patch makes it possible to create NFS program entry ("/proc/net/rpc/nfs") in passed network namespace context instead of hard-coded "init_net". Signed-off-by: Stanislav Kinsbursky Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 65486e652943..d2c760e193f4 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1625,14 +1625,14 @@ static int __init init_nfs_fs(void) goto out0; #ifdef CONFIG_PROC_FS - rpc_proc_register(&nfs_rpcstat); + rpc_proc_register(&init_net, &nfs_rpcstat); #endif if ((err = register_nfs_fs()) != 0) goto out; return 0; out: #ifdef CONFIG_PROC_FS - rpc_proc_unregister("nfs"); + rpc_proc_unregister(&init_net, "nfs"); #endif nfs_destroy_directcache(); out0: @@ -1671,7 +1671,7 @@ static void __exit exit_nfs_fs(void) nfs_dns_resolver_destroy(); nfs_idmap_quit(); #ifdef CONFIG_PROC_FS - rpc_proc_unregister("nfs"); + rpc_proc_unregister(&init_net, "nfs"); #endif nfs_cleanup_cb_ident_idr(); unregister_nfs_fs(); -- cgit v1.2.3 From a030889a01d1bea921e1a7501010b7b891d2abd2 Mon Sep 17 00:00:00 2001 From: Weston Andros Adamson Date: Thu, 26 Jan 2012 13:32:23 -0500 Subject: NFS: start printks w/ NFS: even if __func__ shown This patch addresses printks that have some context to show that they are from fs/nfs/, but for the sake of consistency now start with NFS: Signed-off-by: Weston Andros Adamson Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index d2c760e193f4..028464bcbe0e 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1407,7 +1407,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) /* * Big trouble! The inode has become a different object. */ - printk(KERN_DEBUG "%s: inode %ld mode changed, %07o to %07o\n", + printk(KERN_DEBUG "NFS: %s: inode %ld mode changed, %07o to %07o\n", __func__, inode->i_ino, inode->i_mode, fattr->mode); out_err: /* -- cgit v1.2.3 From 6b13168b36b6a7f603d962c232f1f2f325705832 Mon Sep 17 00:00:00 2001 From: Stanislav Kinsbursky Date: Mon, 23 Jan 2012 17:26:05 +0000 Subject: NFS: make nfs_client_list per net ns This patch splits global list of NFS clients into per-net-ns array of lists. This looks more strict and clearer. BTW, this patch also makes "/proc/fs/nfsfs/servers" entry content depends on /proc mount owner pid namespace. See below for details. NOTE: few words about how was /proc/fs/nfsfs/ entries content show per network namespace done. This is a little bit tricky and not the best is could be. But it's cheap (proper fix for /proc conteinerization is a hard nut to crack). The idea is simple: take proper network namespace from pid namespace child reaper nsproxy of /proc/ mount creator. This actually means, that if there are 2 containers with different net namespace sharing pid namespace, then read of /proc/fs/nfsfs/ entries will always return content, taken from net namespace of pid namespace creator task (and thus second namespace set wil be unvisible). Signed-off-by: Stanislav Kinsbursky Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 028464bcbe0e..0365b84cc2c7 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1558,6 +1558,7 @@ EXPORT_SYMBOL_GPL(nfs_net_id); static int nfs_net_init(struct net *net) { + nfs_clients_init(net); return nfs_dns_resolver_cache_init(net); } -- cgit v1.2.3 From 28cd1b3f262dba56b5e335ba668e342d530f6129 Mon Sep 17 00:00:00 2001 From: Stanislav Kinsbursky Date: Mon, 23 Jan 2012 17:26:22 +0000 Subject: NFS: make cb_ident_idr per net ns This patch makes ID's infrastructure network namespace aware. This was done mainly because of nfs_client_lock, which is desired to be per network namespace, but protects NFS clients ID's. NOTE: NFS client's net pointer have to be set prior to ID initialization, proper assignment was moved. Signed-off-by: Stanislav Kinsbursky Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 0365b84cc2c7..6c662598f885 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1565,6 +1565,7 @@ static int nfs_net_init(struct net *net) static void nfs_net_exit(struct net *net) { nfs_dns_resolver_cache_destroy(net); + nfs_cleanup_cb_ident_idr(net); } static struct pernet_operations nfs_net_ops = { @@ -1674,7 +1675,6 @@ static void __exit exit_nfs_fs(void) #ifdef CONFIG_PROC_FS rpc_proc_unregister(&init_net, "nfs"); #endif - nfs_cleanup_cb_ident_idr(); unregister_nfs_fs(); nfs_fs_proc_exit(); nfsiod_stop(); -- cgit v1.2.3 From 20d27e929fb4790a339a4ddcc9a27f14db06055b Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 1 Mar 2012 17:01:31 -0500 Subject: NFS: Add a client-side function to display NFS file handles For debugging, introduce a simplistic function to print NFS file handles on the system console. The main function is hooked into the dprintk debugging facility, but you can directly call the helper, _nfs_display_fhandle(), if you want to print a handle unconditionally. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 45 insertions(+) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 6c662598f885..99a4f52c14b2 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1045,6 +1045,51 @@ struct nfs_fh *nfs_alloc_fhandle(void) return fh; } +/** + * _nfs_display_fhandle - display an NFS file handle on the console + * + * @fh: file handle to display + * @caption: display caption + * + * For debugging only. + */ +#ifdef RPC_DEBUG +void _nfs_display_fhandle(const struct nfs_fh *fh, const char *caption) +{ + unsigned short i; + + if (fh->size == 0 || fh == NULL) { + printk(KERN_DEFAULT "%s at %p is empty\n", caption, fh); + return; + } + + printk(KERN_DEFAULT "%s at %p is %u bytes:\n", caption, fh, fh->size); + for (i = 0; i < fh->size; i += 16) { + __be32 *pos = (__be32 *)&fh->data[i]; + + switch ((fh->size - i - 1) >> 2) { + case 0: + printk(KERN_DEFAULT " %08x\n", + be32_to_cpup(pos)); + break; + case 1: + printk(KERN_DEFAULT " %08x %08x\n", + be32_to_cpup(pos), be32_to_cpup(pos + 1)); + break; + case 2: + printk(KERN_DEFAULT " %08x %08x %08x\n", + be32_to_cpup(pos), be32_to_cpup(pos + 1), + be32_to_cpup(pos + 2)); + break; + default: + printk(KERN_DEFAULT " %08x %08x %08x %08x\n", + be32_to_cpup(pos), be32_to_cpup(pos + 1), + be32_to_cpup(pos + 2), be32_to_cpup(pos + 3)); + } + } +} +#endif + /** * nfs_inode_attrs_need_update - check if the inode attributes need updating * @inode - pointer to inode -- cgit v1.2.3 From fa68a1ba1de349f0d1fcc54171b95236efe24148 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 6 Mar 2012 10:14:35 -0500 Subject: NFS: Fix a typo in _nfs_display_fhandle The check for 'fh == NULL' needs to come _before_ we dereference fh. Reported-by: Dan Carpenter Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 99a4f52c14b2..ba03b7908149 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1058,7 +1058,7 @@ void _nfs_display_fhandle(const struct nfs_fh *fh, const char *caption) { unsigned short i; - if (fh->size == 0 || fh == NULL) { + if (fh == NULL || fh->size == 0) { printk(KERN_DEFAULT "%s at %p is empty\n", caption, fh); return; } -- cgit v1.2.3 From d8e0539ebdff5ff27fa7eb019715d9dfb5171a3b Mon Sep 17 00:00:00 2001 From: Weston Andros Adamson Date: Tue, 6 Mar 2012 20:46:43 -0500 Subject: NFS: add filehandle crc for debug display Match wireshark's CRC-32 hash for easier debugging Signed-off-by: Weston Andros Adamson Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 23 ++++++++++++++++++++--- 1 file changed, 20 insertions(+), 3 deletions(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index ba03b7908149..0337554f1561 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -39,6 +39,7 @@ #include #include #include +#include #include #include @@ -1045,7 +1046,23 @@ struct nfs_fh *nfs_alloc_fhandle(void) return fh; } -/** +#ifdef RPC_DEBUG +/* + * _nfs_display_fhandle_hash - calculate the crc32 hash for the filehandle + * in the same way that wireshark does + * + * @fh: file handle + * + * For debugging only. + */ +u32 _nfs_display_fhandle_hash(const struct nfs_fh *fh) +{ + /* wireshark uses 32-bit AUTODIN crc and does a bitwise + * not on the result */ + return ~crc32(0xFFFFFFFF, &fh->data[0], fh->size); +} + +/* * _nfs_display_fhandle - display an NFS file handle on the console * * @fh: file handle to display @@ -1053,7 +1070,6 @@ struct nfs_fh *nfs_alloc_fhandle(void) * * For debugging only. */ -#ifdef RPC_DEBUG void _nfs_display_fhandle(const struct nfs_fh *fh, const char *caption) { unsigned short i; @@ -1063,7 +1079,8 @@ void _nfs_display_fhandle(const struct nfs_fh *fh, const char *caption) return; } - printk(KERN_DEFAULT "%s at %p is %u bytes:\n", caption, fh, fh->size); + printk(KERN_DEFAULT "%s at %p is %u bytes, crc: 0x%08x:\n", + caption, fh, fh->size, _nfs_display_fhandle_hash(fh)); for (i = 0; i < fh->size; i += 16) { __be32 *pos = (__be32 *)&fh->data[i]; -- cgit v1.2.3 From 4f1abd226d80ef763c50e3930b369b63dffbb312 Mon Sep 17 00:00:00 2001 From: Weston Andros Adamson Date: Tue, 6 Mar 2012 21:58:20 -0500 Subject: NFS: add fh_crc to debug output Print the filehandle crc in two debug messages Signed-off-by: Weston Andros Adamson Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 0337554f1561..70e25c9c5670 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -390,9 +390,10 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr) unlock_new_inode(inode); } else nfs_refresh_inode(inode, fattr); - dprintk("NFS: nfs_fhget(%s/%Ld ct=%d)\n", + dprintk("NFS: nfs_fhget(%s/%Ld fh_crc=0x%08x ct=%d)\n", inode->i_sb->s_id, (long long)NFS_FILEID(inode), + nfs_display_fhandle_hash(fh), atomic_read(&inode->i_count)); out: @@ -1274,8 +1275,9 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) unsigned long now = jiffies; unsigned long save_cache_validity; - dfprintk(VFS, "NFS: %s(%s/%ld ct=%d info=0x%x)\n", + dfprintk(VFS, "NFS: %s(%s/%ld fh_crc=0x%08x ct=%d info=0x%x)\n", __func__, inode->i_sb->s_id, inode->i_ino, + nfs_display_fhandle_hash(NFS_FH(inode)), atomic_read(&inode->i_count), fattr->valid); if ((fattr->valid & NFS_ATTR_FATTR_FILEID) && nfsi->fileid != fattr->fileid) -- cgit v1.2.3 From d6d6dc7cdfda7c8f49a89a7b7261846f319da6d1 Mon Sep 17 00:00:00 2001 From: Fred Isaman Date: Thu, 8 Mar 2012 17:29:35 -0500 Subject: NFS: remove nfs_inode radix tree The radix tree is only being used to compile lists of reqs needing commit. It is simpler to just put the reqs directly into a list. Signed-off-by: Fred Isaman Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 70e25c9c5670..1a19f8d30c14 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1560,7 +1560,7 @@ static void init_once(void *foo) INIT_LIST_HEAD(&nfsi->open_files); INIT_LIST_HEAD(&nfsi->access_cache_entry_lru); INIT_LIST_HEAD(&nfsi->access_cache_inode_lru); - INIT_RADIX_TREE(&nfsi->nfs_page_tree, GFP_ATOMIC); + INIT_LIST_HEAD(&nfsi->commit_list); nfsi->npages = 0; nfsi->ncommit = 0; atomic_set(&nfsi->silly_count, 1); -- cgit v1.2.3 From e27d359e9b7e446190362cd5c8fe281d02194896 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 18 Mar 2012 14:07:42 -0400 Subject: SUNRPC/NFS: Add Kbuild dependencies for NFS_DEBUG/RPC_DEBUG This allows us to turn on/off the dprintk() debugging interfaces for those distributions that don't ship the 'rpcdebug' utility. It also allows us to add Kbuild dependencies. Specifically, we already know that dprintk() in general relies on CONFIG_SYSCTL. Now it turns out that the NFS dprintks depend on CONFIG_CRC32 after we added support for the filehandle hash. Reported-by: Paul Gortmaker Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 1a19f8d30c14..7bb4d13c1cd5 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1047,7 +1047,7 @@ struct nfs_fh *nfs_alloc_fhandle(void) return fh; } -#ifdef RPC_DEBUG +#ifdef NFS_DEBUG /* * _nfs_display_fhandle_hash - calculate the crc32 hash for the filehandle * in the same way that wireshark does -- cgit v1.2.3 From 9ffc93f203c18a70623f21950f1dd473c9ec48cd Mon Sep 17 00:00:00 2001 From: David Howells Date: Wed, 28 Mar 2012 18:30:03 +0100 Subject: Remove all #inclusions of asm/system.h Remove all #inclusions of asm/system.h preparatory to splitting and killing it. Performed with the following command: perl -p -i -e 's!^#\s*include\s*.*\n!!' `grep -Irl '^#\s*include\s*' *` Signed-off-by: David Howells --- fs/nfs/inode.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs/nfs/inode.c') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index f649fba8c384..c2ce8196912c 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -40,7 +40,6 @@ #include #include -#include #include #include "nfs4_fs.h" -- cgit v1.2.3