From 1c3c07e9f6cc50dab2aeb8051325e317d4f6c70e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 25 Jul 2006 11:28:18 -0400 Subject: [PATCH 01/74] NFS: Add a new ACCESS rpc call cache to the linux nfs client The current access cache only allows one entry at a time to be cached for each inode. Add a per-inode red-black tree in order to allow more than one to be cached at a time. Should significantly cut down the time spent in path traversal for shared directories such as ${PATH}, /usr/share, etc. Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 137 +++++++++++++++++++++++++++++++++++------ fs/nfs/inode.c | 13 ++-- include/linux/nfs_fs.h | 5 +- 3 files changed, 126 insertions(+), 29 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index e7ffb4deb3e5..094afded2b11 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1638,35 +1638,134 @@ static int nfs_rename(struct inode *old_dir, struct dentry *old_dentry, return error; } +static void nfs_access_free_entry(struct nfs_access_entry *entry) +{ + put_rpccred(entry->cred); + kfree(entry); +} + +static void __nfs_access_zap_cache(struct inode *inode) +{ + struct nfs_inode *nfsi = NFS_I(inode); + struct rb_root *root_node = &nfsi->access_cache; + struct rb_node *n, *dispose = NULL; + struct nfs_access_entry *entry; + + /* Unhook entries from the cache */ + while ((n = rb_first(root_node)) != NULL) { + entry = rb_entry(n, struct nfs_access_entry, rb_node); + rb_erase(n, root_node); + n->rb_left = dispose; + dispose = n; + } + nfsi->cache_validity &= ~NFS_INO_INVALID_ACCESS; + spin_unlock(&inode->i_lock); + + /* Now kill them all! */ + while (dispose != NULL) { + n = dispose; + dispose = n->rb_left; + nfs_access_free_entry(rb_entry(n, struct nfs_access_entry, rb_node)); + } +} + +void nfs_access_zap_cache(struct inode *inode) +{ + spin_lock(&inode->i_lock); + /* This will release the spinlock */ + __nfs_access_zap_cache(inode); +} + +static struct nfs_access_entry *nfs_access_search_rbtree(struct inode *inode, struct rpc_cred *cred) +{ + struct rb_node *n = NFS_I(inode)->access_cache.rb_node; + struct nfs_access_entry *entry; + + while (n != NULL) { + entry = rb_entry(n, struct nfs_access_entry, rb_node); + + if (cred < entry->cred) + n = n->rb_left; + else if (cred > entry->cred) + n = n->rb_right; + else + return entry; + } + return NULL; +} + int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs_access_entry *res) { struct nfs_inode *nfsi = NFS_I(inode); - struct nfs_access_entry *cache = &nfsi->cache_access; + struct nfs_access_entry *cache; + int err = -ENOENT; - if (cache->cred != cred - || time_after(jiffies, cache->jiffies + NFS_ATTRTIMEO(inode)) - || (nfsi->cache_validity & NFS_INO_INVALID_ACCESS)) - return -ENOENT; - memcpy(res, cache, sizeof(*res)); - return 0; + spin_lock(&inode->i_lock); + if (nfsi->cache_validity & NFS_INO_INVALID_ACCESS) + goto out_zap; + cache = nfs_access_search_rbtree(inode, cred); + if (cache == NULL) + goto out; + if (time_after(jiffies, cache->jiffies + NFS_ATTRTIMEO(inode))) + goto out_stale; + res->jiffies = cache->jiffies; + res->cred = cache->cred; + res->mask = cache->mask; + err = 0; +out: + spin_unlock(&inode->i_lock); + return err; +out_stale: + rb_erase(&cache->rb_node, &nfsi->access_cache); + spin_unlock(&inode->i_lock); + nfs_access_free_entry(cache); + return -ENOENT; +out_zap: + /* This will release the spinlock */ + __nfs_access_zap_cache(inode); + return -ENOENT; +} + +static void nfs_access_add_rbtree(struct inode *inode, struct nfs_access_entry *set) +{ + struct rb_root *root_node = &NFS_I(inode)->access_cache; + struct rb_node **p = &root_node->rb_node; + struct rb_node *parent = NULL; + struct nfs_access_entry *entry; + + spin_lock(&inode->i_lock); + while (*p != NULL) { + parent = *p; + entry = rb_entry(parent, struct nfs_access_entry, rb_node); + + if (set->cred < entry->cred) + p = &parent->rb_left; + else if (set->cred > entry->cred) + p = &parent->rb_right; + else + goto found; + } + rb_link_node(&set->rb_node, parent, p); + rb_insert_color(&set->rb_node, root_node); + spin_unlock(&inode->i_lock); + return; +found: + rb_replace_node(parent, &set->rb_node, root_node); + spin_unlock(&inode->i_lock); + nfs_access_free_entry(entry); } void nfs_access_add_cache(struct inode *inode, struct nfs_access_entry *set) { - struct nfs_inode *nfsi = NFS_I(inode); - struct nfs_access_entry *cache = &nfsi->cache_access; - - if (cache->cred != set->cred) { - if (cache->cred) - put_rpccred(cache->cred); - cache->cred = get_rpccred(set->cred); - } - /* FIXME: replace current access_cache BKL reliance with inode->i_lock */ - spin_lock(&inode->i_lock); - nfsi->cache_validity &= ~NFS_INO_INVALID_ACCESS; - spin_unlock(&inode->i_lock); + struct nfs_access_entry *cache = kmalloc(sizeof(*cache), GFP_KERNEL); + if (cache == NULL) + return; + RB_CLEAR_NODE(&cache->rb_node); cache->jiffies = set->jiffies; + cache->cred = get_rpccred(set->cred); cache->mask = set->mask; + + nfs_access_add_rbtree(inode, cache); } static int nfs_do_access(struct inode *inode, struct rpc_cred *cred, int mask) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index d349fb2245da..b94ab060bb1e 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -76,19 +76,14 @@ int nfs_write_inode(struct inode *inode, int sync) void nfs_clear_inode(struct inode *inode) { - struct nfs_inode *nfsi = NFS_I(inode); - struct rpc_cred *cred; - /* * The following should never happen... */ BUG_ON(nfs_have_writebacks(inode)); - BUG_ON (!list_empty(&nfsi->open_files)); + BUG_ON(!list_empty(&NFS_I(inode)->open_files)); + BUG_ON(atomic_read(&NFS_I(inode)->data_updates) != 0); nfs_zap_acl_cache(inode); - cred = nfsi->cache_access.cred; - if (cred) - put_rpccred(cred); - BUG_ON(atomic_read(&nfsi->data_updates) != 0); + nfs_access_zap_cache(inode); } /** @@ -290,7 +285,7 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr) nfsi->attrtimeo = NFS_MINATTRTIMEO(inode); nfsi->attrtimeo_timestamp = jiffies; memset(nfsi->cookieverf, 0, sizeof(nfsi->cookieverf)); - nfsi->cache_access.cred = NULL; + nfsi->access_cache = RB_ROOT; unlock_new_inode(inode); } else diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 6c2066caeaab..cc013ed2e52e 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -42,6 +42,7 @@ #include #include #include +#include #include #include @@ -69,6 +70,7 @@ * NFSv3/v4 Access mode cache entry */ struct nfs_access_entry { + struct rb_node rb_node; unsigned long jiffies; struct rpc_cred * cred; int mask; @@ -145,7 +147,7 @@ struct nfs_inode { */ atomic_t data_updates; - struct nfs_access_entry cache_access; + struct rb_root access_cache; #ifdef CONFIG_NFS_V3_ACL struct posix_acl *acl_access; struct posix_acl *acl_default; @@ -297,6 +299,7 @@ extern int nfs_getattr(struct vfsmount *, struct dentry *, struct kstat *); extern int nfs_permission(struct inode *, int, struct nameidata *); extern int nfs_access_get_cached(struct inode *, struct rpc_cred *, struct nfs_access_entry *); extern void nfs_access_add_cache(struct inode *, struct nfs_access_entry *); +extern void nfs_access_zap_cache(struct inode *inode); extern int nfs_open(struct inode *, struct file *); extern int nfs_release(struct inode *, struct file *); extern int nfs_attribute_timeout(struct inode *inode); From cfcea3e8c66c2dcde98d5c2693d4bff50b5cac97 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 25 Jul 2006 11:28:18 -0400 Subject: [PATCH 02/74] NFS: Add a global LRU list for the ACCESS cache ...in order to allow the addition of a memory shrinker. Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 35 ++++++++++++++++++++++++++++++++++- fs/nfs/inode.c | 2 ++ include/linux/nfs_fs.h | 4 ++++ 3 files changed, 40 insertions(+), 1 deletion(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 094afded2b11..bf4f5ffda703 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1638,10 +1638,17 @@ static int nfs_rename(struct inode *old_dir, struct dentry *old_dentry, return error; } +static DEFINE_SPINLOCK(nfs_access_lru_lock); +static LIST_HEAD(nfs_access_lru_list); +static atomic_long_t nfs_access_nr_entries; + static void nfs_access_free_entry(struct nfs_access_entry *entry) { put_rpccred(entry->cred); kfree(entry); + smp_mb__before_atomic_dec(); + atomic_long_dec(&nfs_access_nr_entries); + smp_mb__after_atomic_dec(); } static void __nfs_access_zap_cache(struct inode *inode) @@ -1655,6 +1662,7 @@ static void __nfs_access_zap_cache(struct inode *inode) while ((n = rb_first(root_node)) != NULL) { entry = rb_entry(n, struct nfs_access_entry, rb_node); rb_erase(n, root_node); + list_del(&entry->lru); n->rb_left = dispose; dispose = n; } @@ -1671,6 +1679,13 @@ static void __nfs_access_zap_cache(struct inode *inode) void nfs_access_zap_cache(struct inode *inode) { + /* Remove from global LRU init */ + if (test_and_clear_bit(NFS_INO_ACL_LRU_SET, &NFS_FLAGS(inode))) { + spin_lock(&nfs_access_lru_lock); + list_del_init(&NFS_I(inode)->access_cache_inode_lru); + spin_unlock(&nfs_access_lru_lock); + } + spin_lock(&inode->i_lock); /* This will release the spinlock */ __nfs_access_zap_cache(inode); @@ -1711,12 +1726,14 @@ int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs res->jiffies = cache->jiffies; res->cred = cache->cred; res->mask = cache->mask; + list_move_tail(&cache->lru, &nfsi->access_cache_entry_lru); err = 0; out: spin_unlock(&inode->i_lock); return err; out_stale: rb_erase(&cache->rb_node, &nfsi->access_cache); + list_del(&cache->lru); spin_unlock(&inode->i_lock); nfs_access_free_entry(cache); return -ENOENT; @@ -1728,7 +1745,8 @@ int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs static void nfs_access_add_rbtree(struct inode *inode, struct nfs_access_entry *set) { - struct rb_root *root_node = &NFS_I(inode)->access_cache; + struct nfs_inode *nfsi = NFS_I(inode); + struct rb_root *root_node = &nfsi->access_cache; struct rb_node **p = &root_node->rb_node; struct rb_node *parent = NULL; struct nfs_access_entry *entry; @@ -1747,10 +1765,13 @@ static void nfs_access_add_rbtree(struct inode *inode, struct nfs_access_entry * } rb_link_node(&set->rb_node, parent, p); rb_insert_color(&set->rb_node, root_node); + list_add_tail(&set->lru, &nfsi->access_cache_entry_lru); spin_unlock(&inode->i_lock); return; found: rb_replace_node(parent, &set->rb_node, root_node); + list_add_tail(&set->lru, &nfsi->access_cache_entry_lru); + list_del(&entry->lru); spin_unlock(&inode->i_lock); nfs_access_free_entry(entry); } @@ -1766,6 +1787,18 @@ void nfs_access_add_cache(struct inode *inode, struct nfs_access_entry *set) cache->mask = set->mask; nfs_access_add_rbtree(inode, cache); + + /* Update accounting */ + smp_mb__before_atomic_inc(); + atomic_long_inc(&nfs_access_nr_entries); + smp_mb__after_atomic_inc(); + + /* Add inode to global LRU list */ + if (!test_and_set_bit(NFS_INO_ACL_LRU_SET, &NFS_FLAGS(inode))) { + spin_lock(&nfs_access_lru_lock); + list_add_tail(&NFS_I(inode)->access_cache_inode_lru, &nfs_access_lru_list); + spin_unlock(&nfs_access_lru_lock); + } } static int nfs_do_access(struct inode *inode, struct rpc_cred *cred, int mask) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index b94ab060bb1e..6ed018c9aad2 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1104,6 +1104,8 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags) INIT_LIST_HEAD(&nfsi->dirty); INIT_LIST_HEAD(&nfsi->commit); 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); atomic_set(&nfsi->data_updates, 0); nfsi->ndirty = 0; diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index cc013ed2e52e..a36e01cd6321 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -71,6 +71,7 @@ */ struct nfs_access_entry { struct rb_node rb_node; + struct list_head lru; unsigned long jiffies; struct rpc_cred * cred; int mask; @@ -148,6 +149,8 @@ struct nfs_inode { atomic_t data_updates; struct rb_root access_cache; + struct list_head access_cache_entry_lru; + struct list_head access_cache_inode_lru; #ifdef CONFIG_NFS_V3_ACL struct posix_acl *acl_access; struct posix_acl *acl_default; @@ -201,6 +204,7 @@ struct nfs_inode { #define NFS_INO_REVALIDATING (0) /* revalidating attrs */ #define NFS_INO_ADVISE_RDPLUS (1) /* advise readdirplus */ #define NFS_INO_STALE (2) /* possible stale inode */ +#define NFS_INO_ACL_LRU_SET (3) /* Inode is on the LRU list */ static inline struct nfs_inode *NFS_I(struct inode *inode) { From 979df72e6f963b42ee484f2eca049c3344da0ba7 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 25 Jul 2006 11:28:19 -0400 Subject: [PATCH 03/74] NFS: Add an ACCESS cache memory shrinker A pinned inode may in theory end up filling memory with cached ACCESS calls. This patch ensures that the VM may shrink away the cache in these particular cases. The shrinker works by iterating through the list of inodes on the global nfs_access_lru_list, and removing the least recently used access cache entry until it is done (or until the entire cache is empty). Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 44 ++++++++++++++++++++++++++++++++++++++++++++ fs/nfs/internal.h | 3 +++ fs/nfs/super.c | 5 +++++ 3 files changed, 52 insertions(+) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index bf4f5ffda703..067d144d141b 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1651,6 +1651,50 @@ static void nfs_access_free_entry(struct nfs_access_entry *entry) smp_mb__after_atomic_dec(); } +int nfs_access_cache_shrinker(int nr_to_scan, gfp_t gfp_mask) +{ + LIST_HEAD(head); + struct nfs_inode *nfsi; + struct nfs_access_entry *cache; + + spin_lock(&nfs_access_lru_lock); +restart: + list_for_each_entry(nfsi, &nfs_access_lru_list, access_cache_inode_lru) { + struct inode *inode; + + if (nr_to_scan-- == 0) + break; + inode = igrab(&nfsi->vfs_inode); + if (inode == NULL) + continue; + spin_lock(&inode->i_lock); + if (list_empty(&nfsi->access_cache_entry_lru)) + goto remove_lru_entry; + cache = list_entry(nfsi->access_cache_entry_lru.next, + struct nfs_access_entry, lru); + list_move(&cache->lru, &head); + rb_erase(&cache->rb_node, &nfsi->access_cache); + if (!list_empty(&nfsi->access_cache_entry_lru)) + list_move_tail(&nfsi->access_cache_inode_lru, + &nfs_access_lru_list); + else { +remove_lru_entry: + list_del_init(&nfsi->access_cache_inode_lru); + clear_bit(NFS_INO_ACL_LRU_SET, &nfsi->flags); + } + spin_unlock(&inode->i_lock); + iput(inode); + goto restart; + } + spin_unlock(&nfs_access_lru_lock); + while (!list_empty(&head)) { + cache = list_entry(head.next, struct nfs_access_entry, lru); + list_del(&cache->lru); + nfs_access_free_entry(cache); + } + return (atomic_long_read(&nfs_access_nr_entries) / 100) * sysctl_vfs_cache_pressure; +} + static void __nfs_access_zap_cache(struct inode *inode) { struct nfs_inode *nfsi = NFS_I(inode); diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index e4f4e5def0fc..660e9ff5341c 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -66,6 +66,9 @@ extern int nfs4_proc_fs_locations(struct inode *dir, struct dentry *dentry, struct page *page); #endif +/* dir.c */ +extern int nfs_access_cache_shrinker(int nr_to_scan, gfp_t gfp_mask); + /* inode.c */ extern struct inode *nfs_alloc_inode(struct super_block *sb); extern void nfs_destroy_inode(struct inode *); diff --git a/fs/nfs/super.c b/fs/nfs/super.c index e8a9bee74d9d..06c321beacfe 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -221,6 +221,8 @@ module_param_call(idmap_cache_timeout, param_set_idmap_timeout, param_get_int, &nfs_idmap_cache_timeout, 0644); #endif +static struct shrinker *acl_shrinker; + /* * Register the NFS filesystems */ @@ -240,6 +242,7 @@ int __init register_nfs_fs(void) if (ret < 0) goto error_2; #endif + acl_shrinker = set_shrinker(DEFAULT_SEEKS, nfs_access_cache_shrinker); return 0; #ifdef CONFIG_NFS_V4 @@ -257,6 +260,8 @@ int __init register_nfs_fs(void) */ void __exit unregister_nfs_fs(void) { + if (acl_shrinker != NULL) + remove_shrinker(acl_shrinker); #ifdef CONFIG_NFS_V4 unregister_filesystem(&nfs4_fs_type); nfs_unregister_sysctl(); From 770bfad846ab6628444428467b11fa6773ae9ea1 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:07 -0400 Subject: [PATCH 04/74] NFS: Add dentry materialisation op The attached patch adds a new directory cache management function that prepares a disconnected anonymous function to be connected into the dentry tree. The anonymous dentry is transferred the name and parentage from another dentry. The following changes were made in [try #2]: (*) d_materialise_dentry() now switches the parentage of the two nodes around correctly when one or other of them is self-referential. The following changes were made in [try #7]: (*) d_instantiate_unique() has had the interior part split out as function __d_instantiate_unique(). Callers of this latter function must be holding the appropriate locks. (*) _d_rehash() has been added as a wrapper around __d_rehash() to call it with the most obvious hash list (the one from the name). d_rehash() now calls _d_rehash(). (*) d_materialise_dentry() is now __d_materialise_dentry() and is static. (*) d_materialise_unique() added to perform the combination of d_find_alias(), d_materialise_dentry() and d_add_unique() that the NFS client was doing twice, all within a single dcache_lock critical section. This reduces the number of times two different spinlocks were being accessed. The following further changes were made: (*) Add the dentries onto their parents d_subdirs lists. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/dcache.c | 164 +++++++++++++++++++++++++++++++++++++---- include/linux/dcache.h | 1 + 2 files changed, 151 insertions(+), 14 deletions(-) diff --git a/fs/dcache.c b/fs/dcache.c index 1b4a3a34ec57..17b392a2049e 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -828,17 +828,19 @@ void d_instantiate(struct dentry *entry, struct inode * inode) * (or otherwise set) by the caller to indicate that it is now * in use by the dcache. */ -struct dentry *d_instantiate_unique(struct dentry *entry, struct inode *inode) +static struct dentry *__d_instantiate_unique(struct dentry *entry, + struct inode *inode) { struct dentry *alias; int len = entry->d_name.len; const char *name = entry->d_name.name; unsigned int hash = entry->d_name.hash; - BUG_ON(!list_empty(&entry->d_alias)); - spin_lock(&dcache_lock); - if (!inode) - goto do_negative; + if (!inode) { + entry->d_inode = NULL; + return NULL; + } + list_for_each_entry(alias, &inode->i_dentry, d_alias) { struct qstr *qstr = &alias->d_name; @@ -851,19 +853,35 @@ struct dentry *d_instantiate_unique(struct dentry *entry, struct inode *inode) if (memcmp(qstr->name, name, len)) continue; dget_locked(alias); - spin_unlock(&dcache_lock); - BUG_ON(!d_unhashed(alias)); - iput(inode); return alias; } + list_add(&entry->d_alias, &inode->i_dentry); -do_negative: entry->d_inode = inode; fsnotify_d_instantiate(entry, inode); - spin_unlock(&dcache_lock); - security_d_instantiate(entry, inode); return NULL; } + +struct dentry *d_instantiate_unique(struct dentry *entry, struct inode *inode) +{ + struct dentry *result; + + BUG_ON(!list_empty(&entry->d_alias)); + + spin_lock(&dcache_lock); + result = __d_instantiate_unique(entry, inode); + spin_unlock(&dcache_lock); + + if (!result) { + security_d_instantiate(entry, inode); + return NULL; + } + + BUG_ON(!d_unhashed(result)); + iput(inode); + return result; +} + EXPORT_SYMBOL(d_instantiate_unique); /** @@ -1235,6 +1253,11 @@ static void __d_rehash(struct dentry * entry, struct hlist_head *list) hlist_add_head_rcu(&entry->d_hash, list); } +static void _d_rehash(struct dentry * entry) +{ + __d_rehash(entry, d_hash(entry->d_parent, entry->d_name.hash)); +} + /** * d_rehash - add an entry back to the hash * @entry: dentry to add to the hash @@ -1244,11 +1267,9 @@ static void __d_rehash(struct dentry * entry, struct hlist_head *list) void d_rehash(struct dentry * entry) { - struct hlist_head *list = d_hash(entry->d_parent, entry->d_name.hash); - spin_lock(&dcache_lock); spin_lock(&entry->d_lock); - __d_rehash(entry, list); + _d_rehash(entry); spin_unlock(&entry->d_lock); spin_unlock(&dcache_lock); } @@ -1386,6 +1407,120 @@ void d_move(struct dentry * dentry, struct dentry * target) spin_unlock(&dcache_lock); } +/* + * Prepare an anonymous dentry for life in the superblock's dentry tree as a + * named dentry in place of the dentry to be replaced. + */ +static void __d_materialise_dentry(struct dentry *dentry, struct dentry *anon) +{ + struct dentry *dparent, *aparent; + + switch_names(dentry, anon); + do_switch(dentry->d_name.len, anon->d_name.len); + do_switch(dentry->d_name.hash, anon->d_name.hash); + + dparent = dentry->d_parent; + aparent = anon->d_parent; + + dentry->d_parent = (aparent == anon) ? dentry : aparent; + list_del(&dentry->d_u.d_child); + if (!IS_ROOT(dentry)) + list_add(&dentry->d_u.d_child, &dentry->d_parent->d_subdirs); + else + INIT_LIST_HEAD(&dentry->d_u.d_child); + + anon->d_parent = (dparent == dentry) ? anon : dparent; + list_del(&anon->d_u.d_child); + if (!IS_ROOT(anon)) + list_add(&anon->d_u.d_child, &anon->d_parent->d_subdirs); + else + INIT_LIST_HEAD(&anon->d_u.d_child); + + anon->d_flags &= ~DCACHE_DISCONNECTED; +} + +/** + * d_materialise_unique - introduce an inode into the tree + * @dentry: candidate dentry + * @inode: inode to bind to the dentry, to which aliases may be attached + * + * Introduces an dentry into the tree, substituting an extant disconnected + * root directory alias in its place if there is one + */ +struct dentry *d_materialise_unique(struct dentry *dentry, struct inode *inode) +{ + struct dentry *alias, *actual; + + BUG_ON(!d_unhashed(dentry)); + + spin_lock(&dcache_lock); + + if (!inode) { + actual = dentry; + dentry->d_inode = NULL; + goto found_lock; + } + + /* See if a disconnected directory already exists as an anonymous root + * that we should splice into the tree instead */ + if (S_ISDIR(inode->i_mode) && (alias = __d_find_alias(inode, 1))) { + spin_lock(&alias->d_lock); + + /* Is this a mountpoint that we could splice into our tree? */ + if (IS_ROOT(alias)) + goto connect_mountpoint; + + if (alias->d_name.len == dentry->d_name.len && + alias->d_parent == dentry->d_parent && + memcmp(alias->d_name.name, + dentry->d_name.name, + dentry->d_name.len) == 0) + goto replace_with_alias; + + spin_unlock(&alias->d_lock); + + /* Doh! Seem to be aliasing directories for some reason... */ + dput(alias); + } + + /* Add a unique reference */ + actual = __d_instantiate_unique(dentry, inode); + if (!actual) + actual = dentry; + else if (unlikely(!d_unhashed(actual))) + goto shouldnt_be_hashed; + +found_lock: + spin_lock(&actual->d_lock); +found: + _d_rehash(actual); + spin_unlock(&actual->d_lock); + spin_unlock(&dcache_lock); + + if (actual == dentry) { + security_d_instantiate(dentry, inode); + return NULL; + } + + iput(inode); + return actual; + + /* Convert the anonymous/root alias into an ordinary dentry */ +connect_mountpoint: + __d_materialise_dentry(dentry, alias); + + /* Replace the candidate dentry with the alias in the tree */ +replace_with_alias: + __d_drop(alias); + actual = alias; + goto found; + +shouldnt_be_hashed: + spin_unlock(&dcache_lock); + BUG(); + goto shouldnt_be_hashed; +} + /** * d_path - return the path of a dentry * @dentry: dentry to report @@ -1784,6 +1919,7 @@ EXPORT_SYMBOL(d_instantiate); EXPORT_SYMBOL(d_invalidate); EXPORT_SYMBOL(d_lookup); EXPORT_SYMBOL(d_move); +EXPORT_SYMBOL_GPL(d_materialise_unique); EXPORT_SYMBOL(d_path); EXPORT_SYMBOL(d_prune_aliases); EXPORT_SYMBOL(d_rehash); diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 471781ffeab1..44605be59409 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -221,6 +221,7 @@ static inline int dname_external(struct dentry *dentry) */ extern void d_instantiate(struct dentry *, struct inode *); extern struct dentry * d_instantiate_unique(struct dentry *, struct inode *); +extern struct dentry * d_materialise_unique(struct dentry *, struct inode *); extern void d_delete(struct dentry *); /* allocate/de-allocate */ From 7d4e2747a0412583526a162fbbd6edeeafcceb08 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:07 -0400 Subject: [PATCH 05/74] NFS: Fix up split of fs/nfs/inode.c Fix ups for the splitting of the superblock stuff out of fs/nfs/inode.c, including: (*) Move the callback tcpport module param into callback.c. (*) Move the idmap cache timeout module param into idmap.c. (*) Changes to internal.h: (*) namespace-nfs4.c was renamed to nfs4namespace.c. (*) nfs_stat_to_errno() is in nfs2xdr.c, not nfs4xdr.c. (*) nfs4xdr.c is contingent on CONFIG_NFS_V4. (*) nfs4_path() is only uses if CONFIG_NFS_V4 is set. Plus also: (*) The sec_flavours[] table should really be const. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/callback.c | 15 +++++++++++++++ fs/nfs/idmap.c | 14 ++++++++++++++ fs/nfs/internal.h | 12 ++++++------ fs/nfs/super.c | 40 ++++------------------------------------ 4 files changed, 39 insertions(+), 42 deletions(-) diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index fe0a6b8ac149..d6c4bae14bb9 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -36,6 +36,21 @@ static struct svc_program nfs4_callback_program; unsigned int nfs_callback_set_tcpport; unsigned short nfs_callback_tcpport; +static const int nfs_set_port_min = 0; +static const int nfs_set_port_max = 65535; + +static int param_set_port(const char *val, struct kernel_param *kp) +{ + char *endp; + int num = simple_strtol(val, &endp, 0); + if (endp == val || *endp || num < nfs_set_port_min || num > nfs_set_port_max) + return -EINVAL; + *((int *)kp->arg) = num; + return 0; +} + +module_param_call(callback_tcpport, param_set_port, param_get_int, + &nfs_callback_set_tcpport, 0644); /* * This is the callback kernel thread. diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c index 07a5dd57646e..873deb96a6cd 100644 --- a/fs/nfs/idmap.c +++ b/fs/nfs/idmap.c @@ -57,6 +57,20 @@ /* Default cache timeout is 10 minutes */ unsigned int nfs_idmap_cache_timeout = 600 * HZ; +static int param_set_idmap_timeout(const char *val, struct kernel_param *kp) +{ + char *endp; + int num = simple_strtol(val, &endp, 0); + int jif = num * HZ; + if (endp == val || *endp || num < 0 || jif < num) + return -EINVAL; + *((int *)kp->arg) = jif; + return 0; +} + +module_param_call(idmap_cache_timeout, param_set_idmap_timeout, param_get_int, + &nfs_idmap_cache_timeout, 0644); + struct idmap_hashent { unsigned long ih_expires; __u32 ih_id; diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 660e9ff5341c..4802157963f8 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -15,7 +15,7 @@ struct nfs_clone_mount { rpc_authflavor_t authflavor; }; -/* namespace-nfs4.c */ +/* nfs4namespace.c */ #ifdef CONFIG_NFS_V4 extern struct vfsmount *nfs_do_refmount(const struct vfsmount *mnt_parent, struct dentry *dentry); #else @@ -46,6 +46,7 @@ extern void nfs_destroy_directcache(void); #endif /* nfs2xdr.c */ +extern int nfs_stat_to_errno(int); extern struct rpc_procinfo nfs_procedures[]; extern u32 * nfs_decode_dirent(u32 *, struct nfs_entry *, int); @@ -54,8 +55,9 @@ extern struct rpc_procinfo nfs3_procedures[]; extern u32 *nfs3_decode_dirent(u32 *, struct nfs_entry *, int); /* nfs4xdr.c */ -extern int nfs_stat_to_errno(int); +#ifdef CONFIG_NFS_V4 extern u32 *nfs4_decode_dirent(u32 *p, struct nfs_entry *entry, int plus); +#endif /* nfs4proc.c */ #ifdef CONFIG_NFS_V4 @@ -97,15 +99,13 @@ extern char *nfs_path(const char *base, const struct dentry *dentry, /* * Determine the mount path as a string */ +#ifdef CONFIG_NFS_V4 static inline char * nfs4_path(const struct dentry *dentry, char *buffer, ssize_t buflen) { -#ifdef CONFIG_NFS_V4 return nfs_path(NFS_SB(dentry->d_sb)->mnt_path, dentry, buffer, buflen); -#else - return NULL; -#endif } +#endif /* * Determine the device name as a string diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 06c321beacfe..63497345806b 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -187,40 +187,6 @@ static struct super_operations nfs4_sops = { }; #endif -#ifdef CONFIG_NFS_V4 -static const int nfs_set_port_min = 0; -static const int nfs_set_port_max = 65535; - -static int param_set_port(const char *val, struct kernel_param *kp) -{ - char *endp; - int num = simple_strtol(val, &endp, 0); - if (endp == val || *endp || num < nfs_set_port_min || num > nfs_set_port_max) - return -EINVAL; - *((int *)kp->arg) = num; - return 0; -} - -module_param_call(callback_tcpport, param_set_port, param_get_int, - &nfs_callback_set_tcpport, 0644); -#endif - -#ifdef CONFIG_NFS_V4 -static int param_set_idmap_timeout(const char *val, struct kernel_param *kp) -{ - char *endp; - int num = simple_strtol(val, &endp, 0); - int jif = num * HZ; - if (endp == val || *endp || num < 0 || jif < num) - return -EINVAL; - *((int *)kp->arg) = jif; - return 0; -} - -module_param_call(idmap_cache_timeout, param_set_idmap_timeout, param_get_int, - &nfs_idmap_cache_timeout, 0644); -#endif - static struct shrinker *acl_shrinker; /* @@ -328,9 +294,12 @@ static int nfs_statfs(struct dentry *dentry, struct kstatfs *buf) } +/* + * Map the security flavour number to a name + */ static const char *nfs_pseudoflavour_to_name(rpc_authflavor_t flavour) { - static struct { + static const struct { rpc_authflavor_t flavour; const char *str; } sec_flavours[] = { @@ -1368,7 +1337,6 @@ static int nfs4_get_sb(struct file_system_type *fs_type, } s = sget(fs_type, nfs4_compare_super, nfs_set_super, server); - if (IS_ERR(s)) { error = PTR_ERR(s); goto out_free; From 0a8ea4372b2868842986118ca90912f3382e6c5a Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:08 -0400 Subject: [PATCH 06/74] NFS: Disambiguate nfs_stat_to_errno() Rename the NFS4 version of nfs_stat_to_errno() so that it doesn't conflict with the common one used by NFS2 and NFS3. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/nfs4xdr.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 730ec8fb31c6..1dee6ef7e5a9 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -58,7 +58,7 @@ /* Mapping from NFS error code to "errno" error code. */ #define errno_NFSERR_IO EIO -static int nfs_stat_to_errno(int); +static int nfs4_stat_to_errno(int); /* NFSv4 COMPOUND tags are only wanted for debugging purposes */ #ifdef DEBUG @@ -2127,7 +2127,7 @@ static int decode_op_hdr(struct xdr_stream *xdr, enum nfs_opnum4 expected) } READ32(nfserr); if (nfserr != NFS_OK) - return -nfs_stat_to_errno(nfserr); + return -nfs4_stat_to_errno(nfserr); return 0; } @@ -3598,7 +3598,7 @@ static int decode_setclientid(struct xdr_stream *xdr, struct nfs4_client *clp) READ_BUF(len); return -NFSERR_CLID_INUSE; } else - return -nfs_stat_to_errno(nfserr); + return -nfs4_stat_to_errno(nfserr); return 0; } @@ -4256,7 +4256,7 @@ static int nfs4_xdr_dec_fsinfo(struct rpc_rqst *req, uint32_t *p, struct nfs_fsi if (!status) status = decode_fsinfo(&xdr, fsinfo); if (!status) - status = -nfs_stat_to_errno(hdr.status); + status = -nfs4_stat_to_errno(hdr.status); return status; } @@ -4346,7 +4346,7 @@ static int nfs4_xdr_dec_setclientid(struct rpc_rqst *req, uint32_t *p, if (!status) status = decode_setclientid(&xdr, clp); if (!status) - status = -nfs_stat_to_errno(hdr.status); + status = -nfs4_stat_to_errno(hdr.status); return status; } @@ -4368,7 +4368,7 @@ static int nfs4_xdr_dec_setclientid_confirm(struct rpc_rqst *req, uint32_t *p, s if (!status) status = decode_fsinfo(&xdr, fsinfo); if (!status) - status = -nfs_stat_to_errno(hdr.status); + status = -nfs4_stat_to_errno(hdr.status); return status; } @@ -4521,7 +4521,7 @@ static struct { * This one is used jointly by NFSv2 and NFSv3. */ static int -nfs_stat_to_errno(int stat) +nfs4_stat_to_errno(int stat) { int i; for (i = 0; nfs_errtbl[i].stat != -1; i++) { From 5ae1fbce142b67bf59e15fb1af96e88a96abde7b Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:08 -0400 Subject: [PATCH 07/74] NFS: Fix NFS4 callback up/down prototypes Make the nfs_callback_up()/down() prototypes just do nothing if NFS4 is not enabled. Also make the down function void type since we can't really do anything if it fails. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/callback.c | 5 +---- fs/nfs/callback.h | 7 ++++++- 2 files changed, 7 insertions(+), 5 deletions(-) diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index d6c4bae14bb9..b1f7dc415392 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -149,10 +149,8 @@ int nfs_callback_up(void) /* * Kill the server process if it is not already up. */ -int nfs_callback_down(void) +void nfs_callback_down(void) { - int ret = 0; - lock_kernel(); mutex_lock(&nfs_callback_mutex); nfs_callback_info.users--; @@ -164,7 +162,6 @@ int nfs_callback_down(void) } while (wait_for_completion_timeout(&nfs_callback_info.stopped, 5*HZ) == 0); mutex_unlock(&nfs_callback_mutex); unlock_kernel(); - return ret; } static int nfs_callback_authenticate(struct svc_rqst *rqstp) diff --git a/fs/nfs/callback.h b/fs/nfs/callback.h index b252e7fe53a5..5676163d26e8 100644 --- a/fs/nfs/callback.h +++ b/fs/nfs/callback.h @@ -62,8 +62,13 @@ struct cb_recallargs { extern unsigned nfs4_callback_getattr(struct cb_getattrargs *args, struct cb_getattrres *res); extern unsigned nfs4_callback_recall(struct cb_recallargs *args, void *dummy); +#ifdef CONFIG_NFS_V4 extern int nfs_callback_up(void); -extern int nfs_callback_down(void); +extern void nfs_callback_down(void); +#else +#define nfs_callback_up() (0) +#define nfs_callback_down() do {} while(0) +#endif extern unsigned int nfs_callback_set_tcpport; extern unsigned short nfs_callback_tcpport; From adfa6f980bd46974e6b32b22dd0c45e3f52063f4 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:08 -0400 Subject: [PATCH 08/74] NFS: Rename struct nfs4_client to struct nfs_client Rename struct nfs4_client to struct nfs_client so that it can become the basis for a general client record for NFS2 and NFS3 in addition to NFS4. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/callback.c | 2 +- fs/nfs/callback_proc.c | 4 ++-- fs/nfs/delegation.c | 24 +++++++++---------- fs/nfs/delegation.h | 10 ++++---- fs/nfs/idmap.c | 12 +++++----- fs/nfs/nfs4_fs.h | 30 +++++++++++------------ fs/nfs/nfs4proc.c | 32 ++++++++++++------------- fs/nfs/nfs4renewd.c | 8 +++---- fs/nfs/nfs4state.c | 50 +++++++++++++++++++-------------------- fs/nfs/nfs4xdr.c | 18 +++++++------- fs/nfs/super.c | 4 ++-- include/linux/nfs_fs_sb.h | 2 +- include/linux/nfs_idmap.h | 14 +++++------ 13 files changed, 105 insertions(+), 105 deletions(-) diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index b1f7dc415392..1b596b6d9dc2 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -167,7 +167,7 @@ void nfs_callback_down(void) static int nfs_callback_authenticate(struct svc_rqst *rqstp) { struct in_addr *addr = &rqstp->rq_addr.sin_addr; - struct nfs4_client *clp; + struct nfs_client *clp; /* Don't talk to strangers */ clp = nfs4_find_client(addr); diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 7719483ecdfc..55d6e2ec157f 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -15,7 +15,7 @@ unsigned nfs4_callback_getattr(struct cb_getattrargs *args, struct cb_getattrres *res) { - struct nfs4_client *clp; + struct nfs_client *clp; struct nfs_delegation *delegation; struct nfs_inode *nfsi; struct inode *inode; @@ -56,7 +56,7 @@ unsigned nfs4_callback_getattr(struct cb_getattrargs *args, struct cb_getattrres unsigned nfs4_callback_recall(struct cb_recallargs *args, void *dummy) { - struct nfs4_client *clp; + struct nfs_client *clp; struct inode *inode; unsigned res; diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index 9540a316c05e..5a1105c258bd 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -114,7 +114,7 @@ void nfs_inode_reclaim_delegation(struct inode *inode, struct rpc_cred *cred, st */ int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res) { - struct nfs4_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; int status = 0; @@ -176,7 +176,7 @@ static void nfs_msync_inode(struct inode *inode) */ int __nfs_inode_return_delegation(struct inode *inode) { - struct nfs4_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; int res = 0; @@ -208,7 +208,7 @@ int __nfs_inode_return_delegation(struct inode *inode) */ void nfs_return_all_delegations(struct super_block *sb) { - struct nfs4_client *clp = NFS_SB(sb)->nfs4_state; + struct nfs_client *clp = NFS_SB(sb)->nfs4_state; struct nfs_delegation *delegation; struct inode *inode; @@ -232,7 +232,7 @@ void nfs_return_all_delegations(struct super_block *sb) int nfs_do_expire_all_delegations(void *ptr) { - struct nfs4_client *clp = ptr; + struct nfs_client *clp = ptr; struct nfs_delegation *delegation; struct inode *inode; @@ -258,7 +258,7 @@ int nfs_do_expire_all_delegations(void *ptr) module_put_and_exit(0); } -void nfs_expire_all_delegations(struct nfs4_client *clp) +void nfs_expire_all_delegations(struct nfs_client *clp) { struct task_struct *task; @@ -276,7 +276,7 @@ void nfs_expire_all_delegations(struct nfs4_client *clp) /* * Return all delegations following an NFS4ERR_CB_PATH_DOWN error. */ -void nfs_handle_cb_pathdown(struct nfs4_client *clp) +void nfs_handle_cb_pathdown(struct nfs_client *clp) { struct nfs_delegation *delegation; struct inode *inode; @@ -299,7 +299,7 @@ void nfs_handle_cb_pathdown(struct nfs4_client *clp) struct recall_threadargs { struct inode *inode; - struct nfs4_client *clp; + struct nfs_client *clp; const nfs4_stateid *stateid; struct completion started; @@ -310,7 +310,7 @@ static int recall_thread(void *data) { struct recall_threadargs *args = (struct recall_threadargs *)data; struct inode *inode = igrab(args->inode); - struct nfs4_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; @@ -371,7 +371,7 @@ int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *s /* * Retrieve the inode associated with a delegation */ -struct inode *nfs_delegation_find_inode(struct nfs4_client *clp, const struct nfs_fh *fhandle) +struct inode *nfs_delegation_find_inode(struct nfs_client *clp, const struct nfs_fh *fhandle) { struct nfs_delegation *delegation; struct inode *res = NULL; @@ -389,7 +389,7 @@ struct inode *nfs_delegation_find_inode(struct nfs4_client *clp, const struct nf /* * Mark all delegations as needing to be reclaimed */ -void nfs_delegation_mark_reclaim(struct nfs4_client *clp) +void nfs_delegation_mark_reclaim(struct nfs_client *clp) { struct nfs_delegation *delegation; spin_lock(&clp->cl_lock); @@ -401,7 +401,7 @@ void nfs_delegation_mark_reclaim(struct nfs4_client *clp) /* * Reap all unclaimed delegations after reboot recovery is done */ -void nfs_delegation_reap_unclaimed(struct nfs4_client *clp) +void nfs_delegation_reap_unclaimed(struct nfs_client *clp) { struct nfs_delegation *delegation, *n; LIST_HEAD(head); @@ -423,7 +423,7 @@ void nfs_delegation_reap_unclaimed(struct nfs4_client *clp) int nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode) { - struct nfs4_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; int res = 0; diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h index 3858694652fa..2cfd4b24c7fe 100644 --- a/fs/nfs/delegation.h +++ b/fs/nfs/delegation.h @@ -29,13 +29,13 @@ void nfs_inode_reclaim_delegation(struct inode *inode, struct rpc_cred *cred, st int __nfs_inode_return_delegation(struct inode *inode); int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *stateid); -struct inode *nfs_delegation_find_inode(struct nfs4_client *clp, const struct nfs_fh *fhandle); +struct inode *nfs_delegation_find_inode(struct nfs_client *clp, const struct nfs_fh *fhandle); void nfs_return_all_delegations(struct super_block *sb); -void nfs_expire_all_delegations(struct nfs4_client *clp); -void nfs_handle_cb_pathdown(struct nfs4_client *clp); +void nfs_expire_all_delegations(struct nfs_client *clp); +void nfs_handle_cb_pathdown(struct nfs_client *clp); -void nfs_delegation_mark_reclaim(struct nfs4_client *clp); -void nfs_delegation_reap_unclaimed(struct nfs4_client *clp); +void nfs_delegation_mark_reclaim(struct nfs_client *clp); +void nfs_delegation_reap_unclaimed(struct nfs_client *clp); /* NFSv4 delegation-related procedures */ int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid); diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c index 873deb96a6cd..d05148ec9414 100644 --- a/fs/nfs/idmap.c +++ b/fs/nfs/idmap.c @@ -109,7 +109,7 @@ static struct rpc_pipe_ops idmap_upcall_ops = { }; void -nfs_idmap_new(struct nfs4_client *clp) +nfs_idmap_new(struct nfs_client *clp) { struct idmap *idmap; @@ -138,7 +138,7 @@ nfs_idmap_new(struct nfs4_client *clp) } void -nfs_idmap_delete(struct nfs4_client *clp) +nfs_idmap_delete(struct nfs_client *clp) { struct idmap *idmap = clp->cl_idmap; @@ -491,27 +491,27 @@ static unsigned int fnvhash32(const void *buf, size_t buflen) return (hash); } -int nfs_map_name_to_uid(struct nfs4_client *clp, const char *name, size_t namelen, __u32 *uid) +int nfs_map_name_to_uid(struct nfs_client *clp, const char *name, size_t namelen, __u32 *uid) { struct idmap *idmap = clp->cl_idmap; return nfs_idmap_id(idmap, &idmap->idmap_user_hash, name, namelen, uid); } -int nfs_map_group_to_gid(struct nfs4_client *clp, const char *name, size_t namelen, __u32 *uid) +int nfs_map_group_to_gid(struct nfs_client *clp, const char *name, size_t namelen, __u32 *uid) { struct idmap *idmap = clp->cl_idmap; return nfs_idmap_id(idmap, &idmap->idmap_group_hash, name, namelen, uid); } -int nfs_map_uid_to_name(struct nfs4_client *clp, __u32 uid, char *buf) +int nfs_map_uid_to_name(struct nfs_client *clp, __u32 uid, char *buf) { struct idmap *idmap = clp->cl_idmap; return nfs_idmap_name(idmap, &idmap->idmap_user_hash, uid, buf); } -int nfs_map_gid_to_group(struct nfs4_client *clp, __u32 uid, char *buf) +int nfs_map_gid_to_group(struct nfs_client *clp, __u32 uid, char *buf) { struct idmap *idmap = clp->cl_idmap; diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 9a102860df37..4e334cb48498 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -43,9 +43,9 @@ enum nfs4_client_state { }; /* - * The nfs4_client identifies our client state to the server. + * The nfs_client identifies our client state to the server. */ -struct nfs4_client { +struct nfs_client { struct list_head cl_servers; /* Global list of servers */ struct in_addr cl_addr; /* Server identifier */ u64 cl_clientid; /* constant */ @@ -127,7 +127,7 @@ static inline void nfs_confirm_seqid(struct nfs_seqid_counter *seqid, int status struct nfs4_state_owner { spinlock_t so_lock; struct list_head so_list; /* per-clientid list of state_owners */ - struct nfs4_client *so_client; + struct nfs_client *so_client; u32 so_id; /* 32-bit identifier, unique */ atomic_t so_count; @@ -210,10 +210,10 @@ extern ssize_t nfs4_listxattr(struct dentry *, char *, size_t); /* nfs4proc.c */ extern int nfs4_map_errors(int err); -extern int nfs4_proc_setclientid(struct nfs4_client *, u32, unsigned short, struct rpc_cred *); -extern int nfs4_proc_setclientid_confirm(struct nfs4_client *, struct rpc_cred *); -extern int nfs4_proc_async_renew(struct nfs4_client *, struct rpc_cred *); -extern int nfs4_proc_renew(struct nfs4_client *, struct rpc_cred *); +extern int nfs4_proc_setclientid(struct nfs_client *, u32, unsigned short, struct rpc_cred *); +extern int nfs4_proc_setclientid_confirm(struct nfs_client *, struct rpc_cred *); +extern int nfs4_proc_async_renew(struct nfs_client *, struct rpc_cred *); +extern int nfs4_proc_renew(struct nfs_client *, struct rpc_cred *); extern int nfs4_do_close(struct inode *inode, struct nfs4_state *state); extern struct dentry *nfs4_atomic_open(struct inode *, struct dentry *, struct nameidata *); extern int nfs4_open_revalidate(struct inode *, struct dentry *, int, struct nameidata *); @@ -231,19 +231,19 @@ extern const u32 nfs4_fsinfo_bitmap[2]; extern const u32 nfs4_fs_locations_bitmap[2]; /* nfs4renewd.c */ -extern void nfs4_schedule_state_renewal(struct nfs4_client *); +extern void nfs4_schedule_state_renewal(struct nfs_client *); extern void nfs4_renewd_prepare_shutdown(struct nfs_server *); -extern void nfs4_kill_renewd(struct nfs4_client *); +extern void nfs4_kill_renewd(struct nfs_client *); extern void nfs4_renew_state(void *); /* nfs4state.c */ extern void init_nfsv4_state(struct nfs_server *); extern void destroy_nfsv4_state(struct nfs_server *); -extern struct nfs4_client *nfs4_get_client(struct in_addr *); -extern void nfs4_put_client(struct nfs4_client *clp); -extern struct nfs4_client *nfs4_find_client(struct in_addr *); -struct rpc_cred *nfs4_get_renew_cred(struct nfs4_client *clp); -extern u32 nfs4_alloc_lockowner_id(struct nfs4_client *); +extern struct nfs_client *nfs4_get_client(struct in_addr *); +extern void nfs4_put_client(struct nfs_client *clp); +extern struct nfs_client *nfs4_find_client(struct in_addr *); +struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp); +extern u32 nfs4_alloc_lockowner_id(struct nfs_client *); extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struct rpc_cred *); extern void nfs4_put_state_owner(struct nfs4_state_owner *); @@ -252,7 +252,7 @@ extern struct nfs4_state * nfs4_get_open_state(struct inode *, struct nfs4_state extern void nfs4_put_open_state(struct nfs4_state *); extern void nfs4_close_state(struct nfs4_state *, mode_t); extern void nfs4_state_set_mode_locked(struct nfs4_state *, mode_t); -extern void nfs4_schedule_state_recovery(struct nfs4_client *); +extern void nfs4_schedule_state_recovery(struct nfs_client *); extern void nfs4_put_lock_state(struct nfs4_lock_state *lsp); extern int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl); extern void nfs4_copy_stateid(nfs4_stateid *, struct nfs4_state *, fl_owner_t); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index b14145b7b87f..168f3ffb059f 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -64,7 +64,7 @@ static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinf static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *); static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry); static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception); -static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs4_client *clp); +static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp); /* Prevent leaks of NFSv4 errors into userland */ int nfs4_map_errors(int err) @@ -195,7 +195,7 @@ static void nfs4_setup_readdir(u64 cookie, u32 *verifier, struct dentry *dentry, static void renew_lease(const struct nfs_server *server, unsigned long timestamp) { - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; spin_lock(&clp->cl_lock); if (time_before(clp->cl_last_renewal,timestamp)) clp->cl_last_renewal = timestamp; @@ -792,7 +792,7 @@ static int _nfs4_do_access(struct inode *inode, struct rpc_cred *cred, int openf int nfs4_recover_expired_lease(struct nfs_server *server) { - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; if (test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) nfs4_schedule_state_recovery(clp); @@ -867,7 +867,7 @@ static int _nfs4_open_delegated(struct inode *inode, int flags, struct rpc_cred { struct nfs_delegation *delegation; struct nfs_server *server = NFS_SERVER(inode); - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; struct nfs_inode *nfsi = NFS_I(inode); struct nfs4_state_owner *sp = NULL; struct nfs4_state *state = NULL; @@ -953,7 +953,7 @@ static int _nfs4_do_open(struct inode *dir, struct dentry *dentry, int flags, st struct nfs4_state_owner *sp; struct nfs4_state *state = NULL; struct nfs_server *server = NFS_SERVER(dir); - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; struct nfs4_opendata *opendata; int status; @@ -2521,7 +2521,7 @@ static void nfs4_proc_commit_setup(struct nfs_write_data *data, int how) */ static void nfs4_renew_done(struct rpc_task *task, void *data) { - struct nfs4_client *clp = (struct nfs4_client *)task->tk_msg.rpc_argp; + struct nfs_client *clp = (struct nfs_client *)task->tk_msg.rpc_argp; unsigned long timestamp = (unsigned long)data; if (task->tk_status < 0) { @@ -2543,7 +2543,7 @@ static const struct rpc_call_ops nfs4_renew_ops = { .rpc_call_done = nfs4_renew_done, }; -int nfs4_proc_async_renew(struct nfs4_client *clp, struct rpc_cred *cred) +int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred) { struct rpc_message msg = { .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENEW], @@ -2555,7 +2555,7 @@ int nfs4_proc_async_renew(struct nfs4_client *clp, struct rpc_cred *cred) &nfs4_renew_ops, (void *)jiffies); } -int nfs4_proc_renew(struct nfs4_client *clp, struct rpc_cred *cred) +int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred) { struct rpc_message msg = { .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENEW], @@ -2791,7 +2791,7 @@ static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen static int nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server) { - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; if (!clp || task->tk_status >= 0) return 0; @@ -2828,7 +2828,7 @@ static int nfs4_wait_bit_interruptible(void *word) return 0; } -static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs4_client *clp) +static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp) { sigset_t oldset; int res; @@ -2871,7 +2871,7 @@ static int nfs4_delay(struct rpc_clnt *clnt, long *timeout) */ int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception) { - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; int ret = errorcode; exception->retry = 0; @@ -2898,7 +2898,7 @@ int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct return nfs4_map_errors(ret); } -int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short port, struct rpc_cred *cred) +int nfs4_proc_setclientid(struct nfs_client *clp, u32 program, unsigned short port, struct rpc_cred *cred) { nfs4_verifier sc_verifier; struct nfs4_setclientid setclientid = { @@ -2945,7 +2945,7 @@ int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short p return status; } -static int _nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cred *cred) +static int _nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred) { struct nfs_fsinfo fsinfo; struct rpc_message msg = { @@ -2969,7 +2969,7 @@ static int _nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cr return status; } -int nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cred *cred) +int nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred) { long timeout; int err; @@ -3106,7 +3106,7 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock { struct inode *inode = state->inode; struct nfs_server *server = NFS_SERVER(inode); - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; struct nfs_lockt_args arg = { .fh = NFS_FH(inode), .fl = request, @@ -3513,7 +3513,7 @@ static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request) { - struct nfs4_client *clp = state->owner->so_client; + struct nfs_client *clp = state->owner->so_client; unsigned char fl_flags = request->fl_flags; int status; diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c index 5d764d8e6d8a..208764069f61 100644 --- a/fs/nfs/nfs4renewd.c +++ b/fs/nfs/nfs4renewd.c @@ -61,7 +61,7 @@ void nfs4_renew_state(void *data) { - struct nfs4_client *clp = (struct nfs4_client *)data; + struct nfs_client *clp = (struct nfs_client *)data; struct rpc_cred *cred; long lease, timeout; unsigned long last, now; @@ -108,7 +108,7 @@ nfs4_renew_state(void *data) /* Must be called with clp->cl_sem locked for writes */ void -nfs4_schedule_state_renewal(struct nfs4_client *clp) +nfs4_schedule_state_renewal(struct nfs_client *clp) { long timeout; @@ -127,7 +127,7 @@ nfs4_schedule_state_renewal(struct nfs4_client *clp) void nfs4_renewd_prepare_shutdown(struct nfs_server *server) { - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; if (!clp) return; @@ -140,7 +140,7 @@ nfs4_renewd_prepare_shutdown(struct nfs_server *server) /* Must be called with clp->cl_sem locked for writes */ void -nfs4_kill_renewd(struct nfs4_client *clp) +nfs4_kill_renewd(struct nfs_client *clp) { down_read(&clp->cl_sem); if (!list_empty(&clp->cl_superblocks)) { diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 090a36b07a22..c0b6439f1f71 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -83,10 +83,10 @@ destroy_nfsv4_state(struct nfs_server *server) * Since these are allocated/deallocated very rarely, we don't * bother putting them in a slab cache... */ -static struct nfs4_client * +static struct nfs_client * nfs4_alloc_client(struct in_addr *addr) { - struct nfs4_client *clp; + struct nfs_client *clp; if (nfs_callback_up() < 0) return NULL; @@ -111,7 +111,7 @@ nfs4_alloc_client(struct in_addr *addr) } static void -nfs4_free_client(struct nfs4_client *clp) +nfs4_free_client(struct nfs_client *clp) { struct nfs4_state_owner *sp; @@ -130,9 +130,9 @@ nfs4_free_client(struct nfs4_client *clp) nfs_callback_down(); } -static struct nfs4_client *__nfs4_find_client(struct in_addr *addr) +static struct nfs_client *__nfs4_find_client(struct in_addr *addr) { - struct nfs4_client *clp; + struct nfs_client *clp; list_for_each_entry(clp, &nfs4_clientid_list, cl_servers) { if (memcmp(&clp->cl_addr, addr, sizeof(clp->cl_addr)) == 0) { atomic_inc(&clp->cl_count); @@ -142,19 +142,19 @@ static struct nfs4_client *__nfs4_find_client(struct in_addr *addr) return NULL; } -struct nfs4_client *nfs4_find_client(struct in_addr *addr) +struct nfs_client *nfs4_find_client(struct in_addr *addr) { - struct nfs4_client *clp; + struct nfs_client *clp; spin_lock(&state_spinlock); clp = __nfs4_find_client(addr); spin_unlock(&state_spinlock); return clp; } -struct nfs4_client * +struct nfs_client * nfs4_get_client(struct in_addr *addr) { - struct nfs4_client *clp, *new = NULL; + struct nfs_client *clp, *new = NULL; spin_lock(&state_spinlock); for (;;) { @@ -180,7 +180,7 @@ nfs4_get_client(struct in_addr *addr) } void -nfs4_put_client(struct nfs4_client *clp) +nfs4_put_client(struct nfs_client *clp) { if (!atomic_dec_and_lock(&clp->cl_count, &state_spinlock)) return; @@ -192,7 +192,7 @@ nfs4_put_client(struct nfs4_client *clp) nfs4_free_client(clp); } -static int nfs4_init_client(struct nfs4_client *clp, struct rpc_cred *cred) +static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred) { int status = nfs4_proc_setclientid(clp, NFS4_CALLBACK, nfs_callback_tcpport, cred); @@ -204,13 +204,13 @@ static int nfs4_init_client(struct nfs4_client *clp, struct rpc_cred *cred) } u32 -nfs4_alloc_lockowner_id(struct nfs4_client *clp) +nfs4_alloc_lockowner_id(struct nfs_client *clp) { return clp->cl_lockowner_id ++; } static struct nfs4_state_owner * -nfs4_client_grab_unused(struct nfs4_client *clp, struct rpc_cred *cred) +nfs4_client_grab_unused(struct nfs_client *clp, struct rpc_cred *cred) { struct nfs4_state_owner *sp = NULL; @@ -224,7 +224,7 @@ nfs4_client_grab_unused(struct nfs4_client *clp, struct rpc_cred *cred) return sp; } -struct rpc_cred *nfs4_get_renew_cred(struct nfs4_client *clp) +struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp) { struct nfs4_state_owner *sp; struct rpc_cred *cred = NULL; @@ -238,7 +238,7 @@ struct rpc_cred *nfs4_get_renew_cred(struct nfs4_client *clp) return cred; } -struct rpc_cred *nfs4_get_setclientid_cred(struct nfs4_client *clp) +struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp) { struct nfs4_state_owner *sp; @@ -251,7 +251,7 @@ struct rpc_cred *nfs4_get_setclientid_cred(struct nfs4_client *clp) } static struct nfs4_state_owner * -nfs4_find_state_owner(struct nfs4_client *clp, struct rpc_cred *cred) +nfs4_find_state_owner(struct nfs_client *clp, struct rpc_cred *cred) { struct nfs4_state_owner *sp, *res = NULL; @@ -294,7 +294,7 @@ nfs4_alloc_state_owner(void) void nfs4_drop_state_owner(struct nfs4_state_owner *sp) { - struct nfs4_client *clp = sp->so_client; + struct nfs_client *clp = sp->so_client; spin_lock(&clp->cl_lock); list_del_init(&sp->so_list); spin_unlock(&clp->cl_lock); @@ -306,7 +306,7 @@ nfs4_drop_state_owner(struct nfs4_state_owner *sp) */ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct rpc_cred *cred) { - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; struct nfs4_state_owner *sp, *new; get_rpccred(cred); @@ -337,7 +337,7 @@ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct */ void nfs4_put_state_owner(struct nfs4_state_owner *sp) { - struct nfs4_client *clp = sp->so_client; + struct nfs_client *clp = sp->so_client; struct rpc_cred *cred = sp->so_cred; if (!atomic_dec_and_lock(&sp->so_count, &clp->cl_lock)) @@ -540,7 +540,7 @@ __nfs4_find_lock_state(struct nfs4_state *state, fl_owner_t fl_owner) static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, fl_owner_t fl_owner) { struct nfs4_lock_state *lsp; - struct nfs4_client *clp = state->owner->so_client; + struct nfs_client *clp = state->owner->so_client; lsp = kzalloc(sizeof(*lsp), GFP_KERNEL); if (lsp == NULL) @@ -752,7 +752,7 @@ int nfs_wait_on_sequence(struct nfs_seqid *seqid, struct rpc_task *task) static int reclaimer(void *); -static inline void nfs4_clear_recover_bit(struct nfs4_client *clp) +static inline void nfs4_clear_recover_bit(struct nfs_client *clp) { smp_mb__before_clear_bit(); clear_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state); @@ -764,7 +764,7 @@ static inline void nfs4_clear_recover_bit(struct nfs4_client *clp) /* * State recovery routine */ -static void nfs4_recover_state(struct nfs4_client *clp) +static void nfs4_recover_state(struct nfs_client *clp) { struct task_struct *task; @@ -782,7 +782,7 @@ static void nfs4_recover_state(struct nfs4_client *clp) /* * Schedule a state recovery attempt */ -void nfs4_schedule_state_recovery(struct nfs4_client *clp) +void nfs4_schedule_state_recovery(struct nfs_client *clp) { if (!clp) return; @@ -879,7 +879,7 @@ static int nfs4_reclaim_open_state(struct nfs4_state_recovery_ops *ops, struct n return status; } -static void nfs4_state_mark_reclaim(struct nfs4_client *clp) +static void nfs4_state_mark_reclaim(struct nfs_client *clp) { struct nfs4_state_owner *sp; struct nfs4_state *state; @@ -903,7 +903,7 @@ static void nfs4_state_mark_reclaim(struct nfs4_client *clp) static int reclaimer(void *ptr) { - struct nfs4_client *clp = ptr; + struct nfs_client *clp = ptr; struct nfs4_state_owner *sp; struct nfs4_state_recovery_ops *ops; struct rpc_cred *cred; diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 1dee6ef7e5a9..04748ab9ed55 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1160,7 +1160,7 @@ static int encode_rename(struct xdr_stream *xdr, const struct qstr *oldname, con return 0; } -static int encode_renew(struct xdr_stream *xdr, const struct nfs4_client *client_stateid) +static int encode_renew(struct xdr_stream *xdr, const struct nfs_client *client_stateid) { uint32_t *p; @@ -1246,7 +1246,7 @@ static int encode_setclientid(struct xdr_stream *xdr, const struct nfs4_setclien return 0; } -static int encode_setclientid_confirm(struct xdr_stream *xdr, const struct nfs4_client *client_state) +static int encode_setclientid_confirm(struct xdr_stream *xdr, const struct nfs_client *client_state) { uint32_t *p; @@ -1945,7 +1945,7 @@ static int nfs4_xdr_enc_server_caps(struct rpc_rqst *req, uint32_t *p, const str /* * a RENEW request */ -static int nfs4_xdr_enc_renew(struct rpc_rqst *req, uint32_t *p, struct nfs4_client *clp) +static int nfs4_xdr_enc_renew(struct rpc_rqst *req, uint32_t *p, struct nfs_client *clp) { struct xdr_stream xdr; struct compound_hdr hdr = { @@ -1975,7 +1975,7 @@ static int nfs4_xdr_enc_setclientid(struct rpc_rqst *req, uint32_t *p, struct nf /* * a SETCLIENTID_CONFIRM request */ -static int nfs4_xdr_enc_setclientid_confirm(struct rpc_rqst *req, uint32_t *p, struct nfs4_client *clp) +static int nfs4_xdr_enc_setclientid_confirm(struct rpc_rqst *req, uint32_t *p, struct nfs_client *clp) { struct xdr_stream xdr; struct compound_hdr hdr = { @@ -2132,7 +2132,7 @@ static int decode_op_hdr(struct xdr_stream *xdr, enum nfs_opnum4 expected) } /* Dummy routine */ -static int decode_ace(struct xdr_stream *xdr, void *ace, struct nfs4_client *clp) +static int decode_ace(struct xdr_stream *xdr, void *ace, struct nfs_client *clp) { uint32_t *p; unsigned int strlen; @@ -2636,7 +2636,7 @@ static int decode_attr_nlink(struct xdr_stream *xdr, uint32_t *bitmap, uint32_t return 0; } -static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs4_client *clp, int32_t *uid) +static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs_client *clp, int32_t *uid) { uint32_t len, *p; @@ -2660,7 +2660,7 @@ static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap, struct nf return 0; } -static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs4_client *clp, int32_t *gid) +static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap, struct nfs_client *clp, int32_t *gid) { uint32_t len, *p; @@ -3565,7 +3565,7 @@ static int decode_setattr(struct xdr_stream *xdr, struct nfs_setattrres *res) return 0; } -static int decode_setclientid(struct xdr_stream *xdr, struct nfs4_client *clp) +static int decode_setclientid(struct xdr_stream *xdr, struct nfs_client *clp) { uint32_t *p; uint32_t opnum; @@ -4335,7 +4335,7 @@ static int nfs4_xdr_dec_renew(struct rpc_rqst *rqstp, uint32_t *p, void *dummy) * a SETCLIENTID request */ static int nfs4_xdr_dec_setclientid(struct rpc_rqst *req, uint32_t *p, - struct nfs4_client *clp) + struct nfs_client *clp) { struct xdr_stream xdr; struct compound_hdr hdr; diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 63497345806b..d03ede5b1aca 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1099,7 +1099,7 @@ static int nfs_clone_nfs_sb(struct file_system_type *fs_type, static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, struct rpc_timeout *timeparms, int proto, rpc_authflavor_t flavor) { - struct nfs4_client *clp; + struct nfs_client *clp; struct rpc_xprt *xprt = NULL; struct rpc_clnt *clnt = NULL; int err = -EIO; @@ -1416,7 +1416,7 @@ static inline char *nfs4_dup_path(const struct dentry *dentry) static struct super_block *nfs4_clone_sb(struct nfs_server *server, struct nfs_clone_mount *data) { const struct dentry *dentry = data->dentry; - struct nfs4_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs4_state; struct super_block *sb; server->fsid = data->fattr->fsid; diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 6b4a13c79474..4db90df2aed0 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -43,7 +43,7 @@ struct nfs_server { */ char ip_addr[16]; char * mnt_path; - struct nfs4_client * nfs4_state; /* all NFSv4 state starts here */ + struct nfs_client * nfs4_state; /* all NFSv4 state starts here */ struct list_head nfs4_siblings; /* List of other nfs_server structs * that share the same clientid */ diff --git a/include/linux/nfs_idmap.h b/include/linux/nfs_idmap.h index 102e56094296..678fe68982ef 100644 --- a/include/linux/nfs_idmap.h +++ b/include/linux/nfs_idmap.h @@ -62,15 +62,15 @@ struct idmap_msg { #ifdef __KERNEL__ /* Forward declaration to make this header independent of others */ -struct nfs4_client; +struct nfs_client; -void nfs_idmap_new(struct nfs4_client *); -void nfs_idmap_delete(struct nfs4_client *); +void nfs_idmap_new(struct nfs_client *); +void nfs_idmap_delete(struct nfs_client *); -int nfs_map_name_to_uid(struct nfs4_client *, const char *, size_t, __u32 *); -int nfs_map_group_to_gid(struct nfs4_client *, const char *, size_t, __u32 *); -int nfs_map_uid_to_name(struct nfs4_client *, __u32, char *); -int nfs_map_gid_to_group(struct nfs4_client *, __u32, char *); +int nfs_map_name_to_uid(struct nfs_client *, const char *, size_t, __u32 *); +int nfs_map_group_to_gid(struct nfs_client *, const char *, size_t, __u32 *); +int nfs_map_uid_to_name(struct nfs_client *, __u32, char *); +int nfs_map_gid_to_group(struct nfs_client *, __u32, char *); extern unsigned int nfs_idmap_cache_timeout; #endif /* __KERNEL__ */ From 7539bbab8062aadc1db95a22b377146843cfa88f Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:09 -0400 Subject: [PATCH 09/74] NFS: Rename nfs_server::nfs4_state Rename nfs_server::nfs4_state to nfs_client as it will be used to represent the client state for NFS2 and NFS3 also. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/delegation.c | 12 ++++++------ fs/nfs/nfs4proc.c | 26 +++++++++++++------------- fs/nfs/nfs4renewd.c | 2 +- fs/nfs/nfs4state.c | 10 +++++----- fs/nfs/nfs4xdr.c | 10 +++++----- fs/nfs/super.c | 6 +++--- include/linux/nfs_fs_sb.h | 2 +- 7 files changed, 34 insertions(+), 34 deletions(-) diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index 5a1105c258bd..cfe239736ac0 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -52,7 +52,7 @@ static int nfs_delegation_claim_locks(struct nfs_open_context *ctx, struct nfs4_ case -NFS4ERR_EXPIRED: /* kill_proc(fl->fl_pid, SIGLOST, 1); */ case -NFS4ERR_STALE_CLIENTID: - nfs4_schedule_state_recovery(NFS_SERVER(inode)->nfs4_state); + nfs4_schedule_state_recovery(NFS_SERVER(inode)->nfs_client); goto out_err; } } @@ -114,7 +114,7 @@ void nfs_inode_reclaim_delegation(struct inode *inode, struct rpc_cred *cred, st */ int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res) { - struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs_client; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; int status = 0; @@ -176,7 +176,7 @@ static void nfs_msync_inode(struct inode *inode) */ int __nfs_inode_return_delegation(struct inode *inode) { - struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs_client; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; int res = 0; @@ -208,7 +208,7 @@ int __nfs_inode_return_delegation(struct inode *inode) */ void nfs_return_all_delegations(struct super_block *sb) { - struct nfs_client *clp = NFS_SB(sb)->nfs4_state; + struct nfs_client *clp = NFS_SB(sb)->nfs_client; struct nfs_delegation *delegation; struct inode *inode; @@ -310,7 +310,7 @@ static int recall_thread(void *data) { struct recall_threadargs *args = (struct recall_threadargs *)data; struct inode *inode = igrab(args->inode); - struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs_client; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; @@ -423,7 +423,7 @@ void nfs_delegation_reap_unclaimed(struct nfs_client *clp) int nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode) { - struct nfs_client *clp = NFS_SERVER(inode)->nfs4_state; + struct nfs_client *clp = NFS_SERVER(inode)->nfs_client; struct nfs_inode *nfsi = NFS_I(inode); struct nfs_delegation *delegation; int res = 0; diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 168f3ffb059f..b46597fc81e1 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -195,7 +195,7 @@ static void nfs4_setup_readdir(u64 cookie, u32 *verifier, struct dentry *dentry, static void renew_lease(const struct nfs_server *server, unsigned long timestamp) { - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; spin_lock(&clp->cl_lock); if (time_before(clp->cl_last_renewal,timestamp)) clp->cl_last_renewal = timestamp; @@ -252,7 +252,7 @@ static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry, atomic_inc(&sp->so_count); p->o_arg.fh = NFS_FH(dir); p->o_arg.open_flags = flags, - p->o_arg.clientid = server->nfs4_state->cl_clientid; + p->o_arg.clientid = server->nfs_client->cl_clientid; p->o_arg.id = sp->so_id; p->o_arg.name = &dentry->d_name; p->o_arg.server = server; @@ -550,7 +550,7 @@ int nfs4_open_delegation_recall(struct dentry *dentry, struct nfs4_state *state) case -NFS4ERR_STALE_STATEID: case -NFS4ERR_EXPIRED: /* Don't recall a delegation if it was lost */ - nfs4_schedule_state_recovery(server->nfs4_state); + nfs4_schedule_state_recovery(server->nfs_client); return err; } err = nfs4_handle_exception(server, err, &exception); @@ -792,7 +792,7 @@ static int _nfs4_do_access(struct inode *inode, struct rpc_cred *cred, int openf int nfs4_recover_expired_lease(struct nfs_server *server) { - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; if (test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) nfs4_schedule_state_recovery(clp); @@ -867,7 +867,7 @@ static int _nfs4_open_delegated(struct inode *inode, int flags, struct rpc_cred { struct nfs_delegation *delegation; struct nfs_server *server = NFS_SERVER(inode); - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; struct nfs_inode *nfsi = NFS_I(inode); struct nfs4_state_owner *sp = NULL; struct nfs4_state *state = NULL; @@ -953,7 +953,7 @@ static int _nfs4_do_open(struct inode *dir, struct dentry *dentry, int flags, st struct nfs4_state_owner *sp; struct nfs4_state *state = NULL; struct nfs_server *server = NFS_SERVER(dir); - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; struct nfs4_opendata *opendata; int status; @@ -1133,7 +1133,7 @@ static void nfs4_close_done(struct rpc_task *task, void *data) break; case -NFS4ERR_STALE_STATEID: case -NFS4ERR_EXPIRED: - nfs4_schedule_state_recovery(server->nfs4_state); + nfs4_schedule_state_recovery(server->nfs_client); break; default: if (nfs4_async_handle_error(task, server) == -EAGAIN) { @@ -2791,7 +2791,7 @@ static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen static int nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server) { - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; if (!clp || task->tk_status >= 0) return 0; @@ -2871,7 +2871,7 @@ static int nfs4_delay(struct rpc_clnt *clnt, long *timeout) */ int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception) { - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; int ret = errorcode; exception->retry = 0; @@ -3077,7 +3077,7 @@ int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4 switch (err) { case -NFS4ERR_STALE_STATEID: case -NFS4ERR_EXPIRED: - nfs4_schedule_state_recovery(server->nfs4_state); + nfs4_schedule_state_recovery(server->nfs_client); case 0: return 0; } @@ -3106,7 +3106,7 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock { struct inode *inode = state->inode; struct nfs_server *server = NFS_SERVER(inode); - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; struct nfs_lockt_args arg = { .fh = NFS_FH(inode), .fl = request, @@ -3231,7 +3231,7 @@ static void nfs4_locku_done(struct rpc_task *task, void *data) break; case -NFS4ERR_STALE_STATEID: case -NFS4ERR_EXPIRED: - nfs4_schedule_state_recovery(calldata->server->nfs4_state); + nfs4_schedule_state_recovery(calldata->server->nfs_client); break; default: if (nfs4_async_handle_error(task, calldata->server) == -EAGAIN) { @@ -3343,7 +3343,7 @@ static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl, if (p->arg.lock_seqid == NULL) goto out_free; p->arg.lock_stateid = &lsp->ls_stateid; - p->arg.lock_owner.clientid = server->nfs4_state->cl_clientid; + p->arg.lock_owner.clientid = server->nfs_client->cl_clientid; p->arg.lock_owner.id = lsp->ls_id; p->lsp = lsp; atomic_inc(&lsp->ls_count); diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c index 208764069f61..ff947ecb8b81 100644 --- a/fs/nfs/nfs4renewd.c +++ b/fs/nfs/nfs4renewd.c @@ -127,7 +127,7 @@ nfs4_schedule_state_renewal(struct nfs_client *clp) void nfs4_renewd_prepare_shutdown(struct nfs_server *server) { - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; if (!clp) return; diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index c0b6439f1f71..fa51a7d4c022 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -61,7 +61,7 @@ static LIST_HEAD(nfs4_clientid_list); void init_nfsv4_state(struct nfs_server *server) { - server->nfs4_state = NULL; + server->nfs_client = NULL; INIT_LIST_HEAD(&server->nfs4_siblings); } @@ -70,9 +70,9 @@ destroy_nfsv4_state(struct nfs_server *server) { kfree(server->mnt_path); server->mnt_path = NULL; - if (server->nfs4_state) { - nfs4_put_client(server->nfs4_state); - server->nfs4_state = NULL; + if (server->nfs_client) { + nfs4_put_client(server->nfs_client); + server->nfs_client = NULL; } } @@ -306,7 +306,7 @@ nfs4_drop_state_owner(struct nfs4_state_owner *sp) */ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct rpc_cred *cred) { - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; struct nfs4_state_owner *sp, *new; get_rpccred(cred); diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 04748ab9ed55..99926067eca4 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -529,7 +529,7 @@ static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const s if (iap->ia_valid & ATTR_MODE) len += 4; if (iap->ia_valid & ATTR_UID) { - owner_namelen = nfs_map_uid_to_name(server->nfs4_state, iap->ia_uid, owner_name); + owner_namelen = nfs_map_uid_to_name(server->nfs_client, iap->ia_uid, owner_name); if (owner_namelen < 0) { printk(KERN_WARNING "nfs: couldn't resolve uid %d to string\n", iap->ia_uid); @@ -541,7 +541,7 @@ static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const s len += 4 + (XDR_QUADLEN(owner_namelen) << 2); } if (iap->ia_valid & ATTR_GID) { - owner_grouplen = nfs_map_gid_to_group(server->nfs4_state, iap->ia_gid, owner_group); + owner_grouplen = nfs_map_gid_to_group(server->nfs_client, iap->ia_gid, owner_group); if (owner_grouplen < 0) { printk(KERN_WARNING "nfs4: couldn't resolve gid %d to string\n", iap->ia_gid); @@ -3051,9 +3051,9 @@ static int decode_getfattr(struct xdr_stream *xdr, struct nfs_fattr *fattr, cons fattr->mode |= fmode; if ((status = decode_attr_nlink(xdr, bitmap, &fattr->nlink)) != 0) goto xdr_error; - if ((status = decode_attr_owner(xdr, bitmap, server->nfs4_state, &fattr->uid)) != 0) + if ((status = decode_attr_owner(xdr, bitmap, server->nfs_client, &fattr->uid)) != 0) goto xdr_error; - if ((status = decode_attr_group(xdr, bitmap, server->nfs4_state, &fattr->gid)) != 0) + if ((status = decode_attr_group(xdr, bitmap, server->nfs_client, &fattr->gid)) != 0) goto xdr_error; if ((status = decode_attr_rdev(xdr, bitmap, &fattr->rdev)) != 0) goto xdr_error; @@ -3254,7 +3254,7 @@ static int decode_delegation(struct xdr_stream *xdr, struct nfs_openres *res) if (decode_space_limit(xdr, &res->maxsize) < 0) return -EIO; } - return decode_ace(xdr, NULL, res->server->nfs4_state); + return decode_ace(xdr, NULL, res->server->nfs_client); } static int decode_open(struct xdr_stream *xdr, struct nfs_openres *res) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index d03ede5b1aca..ab4c78ee840c 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1141,7 +1141,7 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks); clnt = rpc_clone_client(clp->cl_rpcclient); if (!IS_ERR(clnt)) - server->nfs4_state = clp; + server->nfs_client = clp; up_write(&clp->cl_sem); clp = NULL; @@ -1151,7 +1151,7 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, return clnt; } - if (server->nfs4_state->cl_idmap == NULL) { + if (server->nfs_client->cl_idmap == NULL) { dprintk("%s: failed to create idmapper.\n", __FUNCTION__); return ERR_PTR(-ENOMEM); } @@ -1416,7 +1416,7 @@ static inline char *nfs4_dup_path(const struct dentry *dentry) static struct super_block *nfs4_clone_sb(struct nfs_server *server, struct nfs_clone_mount *data) { const struct dentry *dentry = data->dentry; - struct nfs_client *clp = server->nfs4_state; + struct nfs_client *clp = server->nfs_client; struct super_block *sb; server->fsid = data->fattr->fsid; diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 4db90df2aed0..fc20d6b934fb 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -43,7 +43,7 @@ struct nfs_server { */ char ip_addr[16]; char * mnt_path; - struct nfs_client * nfs4_state; /* all NFSv4 state starts here */ + struct nfs_client * nfs_client; /* all NFSv4 state starts here */ struct list_head nfs4_siblings; /* List of other nfs_server structs * that share the same clientid */ From b7162792b5c0e0f6e91b8997f8e6bbc76ec5420a Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:09 -0400 Subject: [PATCH 10/74] NFS: Return an error when starting the idmapping pipe Return an error when starting the idmapping pipe so that we can detect it failing. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/idmap.c | 12 ++++++++---- fs/nfs/super.c | 3 ++- include/linux/nfs_idmap.h | 2 +- 3 files changed, 11 insertions(+), 6 deletions(-) diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c index d05148ec9414..231c20ffc0ff 100644 --- a/fs/nfs/idmap.c +++ b/fs/nfs/idmap.c @@ -108,15 +108,17 @@ static struct rpc_pipe_ops idmap_upcall_ops = { .destroy_msg = idmap_pipe_destroy_msg, }; -void +int nfs_idmap_new(struct nfs_client *clp) { struct idmap *idmap; + int error; if (clp->cl_idmap != NULL) - return; + return 0; + if ((idmap = kzalloc(sizeof(*idmap), GFP_KERNEL)) == NULL) - return; + return -ENOMEM; snprintf(idmap->idmap_path, sizeof(idmap->idmap_path), "%s/idmap", clp->cl_rpcclient->cl_pathname); @@ -124,8 +126,9 @@ nfs_idmap_new(struct nfs_client *clp) idmap->idmap_dentry = rpc_mkpipe(idmap->idmap_path, idmap, &idmap_upcall_ops, 0); if (IS_ERR(idmap->idmap_dentry)) { + error = PTR_ERR(idmap->idmap_dentry); kfree(idmap); - return; + return error; } mutex_init(&idmap->idmap_lock); @@ -135,6 +138,7 @@ nfs_idmap_new(struct nfs_client *clp) idmap->idmap_group_hash.h_type = IDMAP_TYPE_GROUP; clp->cl_idmap = idmap; + return 0; } void diff --git a/fs/nfs/super.c b/fs/nfs/super.c index ab4c78ee840c..3ee85c4e65d8 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1136,7 +1136,8 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, clnt->cl_softrtry = 1; clp->cl_rpcclient = clnt; memcpy(clp->cl_ipaddr, server->ip_addr, sizeof(clp->cl_ipaddr)); - nfs_idmap_new(clp); + if (nfs_idmap_new(clp) < 0) + goto out_fail; } list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks); clnt = rpc_clone_client(clp->cl_rpcclient); diff --git a/include/linux/nfs_idmap.h b/include/linux/nfs_idmap.h index 678fe68982ef..15a9f3b7289a 100644 --- a/include/linux/nfs_idmap.h +++ b/include/linux/nfs_idmap.h @@ -64,7 +64,7 @@ struct idmap_msg { /* Forward declaration to make this header independent of others */ struct nfs_client; -void nfs_idmap_new(struct nfs_client *); +int nfs_idmap_new(struct nfs_client *); void nfs_idmap_delete(struct nfs_client *); int nfs_map_name_to_uid(struct nfs_client *, const char *, size_t, __u32 *); From 2b3de4411b3ccaeb00018c99d1bbe7203554cf7f Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:09 -0400 Subject: [PATCH 11/74] NFS: Add a lookupfh NFS RPC op Add a lookup filehandle NFS RPC op so that a file handle can be looked up without requiring dentries and inodes and other VFS stuff when doing an NFS4 pathwalk during mounting. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 47 +++++++++++++++++++++++++++++++++++++++++ include/linux/nfs_xdr.h | 3 +++ 2 files changed, 50 insertions(+) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index b46597fc81e1..de2006f754ef 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -1583,6 +1583,52 @@ nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr, return status; } +static int _nfs4_proc_lookupfh(struct nfs_server *server, struct nfs_fh *dirfh, + struct qstr *name, struct nfs_fh *fhandle, + struct nfs_fattr *fattr) +{ + int status; + struct nfs4_lookup_arg args = { + .bitmask = server->attr_bitmask, + .dir_fh = dirfh, + .name = name, + }; + struct nfs4_lookup_res res = { + .server = server, + .fattr = fattr, + .fh = fhandle, + }; + struct rpc_message msg = { + .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP], + .rpc_argp = &args, + .rpc_resp = &res, + }; + + nfs_fattr_init(fattr); + + dprintk("NFS call lookupfh %s\n", name->name); + status = rpc_call_sync(server->client, &msg, 0); + dprintk("NFS reply lookupfh: %d\n", status); + if (status == -NFS4ERR_MOVED) + status = -EREMOTE; + return status; +} + +static int nfs4_proc_lookupfh(struct nfs_server *server, struct nfs_fh *dirfh, + struct qstr *name, struct nfs_fh *fhandle, + struct nfs_fattr *fattr) +{ + struct nfs4_exception exception = { }; + int err; + do { + err = nfs4_handle_exception(server, + _nfs4_proc_lookupfh(server, dirfh, name, + fhandle, fattr), + &exception); + } while (exception.retry); + return err; +} + static int _nfs4_proc_lookup(struct inode *dir, struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr) { @@ -3723,6 +3769,7 @@ struct nfs_rpc_ops nfs_v4_clientops = { .getroot = nfs4_proc_get_root, .getattr = nfs4_proc_getattr, .setattr = nfs4_proc_setattr, + .lookupfh = nfs4_proc_lookupfh, .lookup = nfs4_proc_lookup, .access = nfs4_proc_access, .readlink = nfs4_proc_readlink, diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 41e5a19199e9..26879771831d 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -770,6 +770,9 @@ struct nfs_rpc_ops { int (*getroot) (struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *); + int (*lookupfh)(struct nfs_server *, struct nfs_fh *, + struct qstr *, struct nfs_fh *, + struct nfs_fattr *); int (*getattr) (struct nfs_server *, struct nfs_fh *, struct nfs_fattr *); int (*setattr) (struct dentry *, struct nfs_fattr *, From e9326dcab413848e70ab746c7c5363da13e5f801 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:10 -0400 Subject: [PATCH 12/74] NFS: Add a server capabilities NFS RPC op Add a set_capabilities NFS RPC op so that the server capabilities can be set. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 1 + include/linux/nfs_xdr.h | 1 + 2 files changed, 2 insertions(+) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index de2006f754ef..850f0851023a 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3790,6 +3790,7 @@ struct nfs_rpc_ops nfs_v4_clientops = { .statfs = nfs4_proc_statfs, .fsinfo = nfs4_proc_fsinfo, .pathconf = nfs4_proc_pathconf, + .set_capabilities = nfs4_server_capabilities, .decode_dirent = nfs4_decode_dirent, .read_setup = nfs4_proc_read_setup, .read_done = nfs4_read_done, diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 26879771831d..dd9ae6761f71 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -809,6 +809,7 @@ struct nfs_rpc_ops { struct nfs_fsinfo *); int (*pathconf) (struct nfs_server *, struct nfs_fh *, struct nfs_pathconf *); + int (*set_capabilities)(struct nfs_server *, struct nfs_fh *); u32 * (*decode_dirent)(u32 *, struct nfs_entry *, int plus); void (*read_setup) (struct nfs_read_data *); int (*read_done) (struct rpc_task *, struct nfs_read_data *); From 24c8dbbb5f777187d660393599641ab3307b4b97 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:10 -0400 Subject: [PATCH 13/74] NFS: Generalise the nfs_client structure Generalise the nfs_client structure by: (1) Moving nfs_client to a more general place (nfs_fs_sb.h). (2) Renaming its maintenance routines to be non-NFS4 specific. (3) Move those maintenance routines to a new non-NFS4 specific file (client.c) and move the declarations to internal.h. (4) Make nfs_find/get_client() take a full sockaddr_in to include the port number (will be required for NFS2/3). (5) Make nfs_find/get_client() take the NFS protocol version (again will be required to differentiate NFS2, 3 & 4 client records). Also: (6) Make nfs_client construction proceed akin to inodes, marking them as under construction and providing a function to indicate completion. (7) Make nfs_get_client() wait interruptibly if it finds a client that it can share, but that client is currently being constructed. (8) Make nfs4_create_client() use (6) and (7) instead of locking cl_sem. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/Makefile | 6 +- fs/nfs/callback.c | 9 +- fs/nfs/callback_proc.c | 9 +- fs/nfs/client.c | 312 ++++++++++++++++++++++++++++++++++++++ fs/nfs/delegation.c | 9 +- fs/nfs/internal.h | 6 + fs/nfs/nfs4_fs.h | 52 ------- fs/nfs/nfs4proc.c | 2 +- fs/nfs/nfs4state.c | 128 +--------------- fs/nfs/super.c | 53 +++---- include/linux/nfs_fs.h | 1 + include/linux/nfs_fs_sb.h | 60 ++++++++ 12 files changed, 425 insertions(+), 222 deletions(-) create mode 100644 fs/nfs/client.c diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile index 0b572a0c1967..3b993a6f8163 100644 --- a/fs/nfs/Makefile +++ b/fs/nfs/Makefile @@ -4,9 +4,9 @@ obj-$(CONFIG_NFS_FS) += nfs.o -nfs-y := dir.o file.o inode.o super.o nfs2xdr.o pagelist.o \ - proc.o read.o symlink.o unlink.o write.o \ - namespace.o +nfs-y := client.o dir.o file.o inode.o super.o nfs2xdr.o \ + pagelist.o proc.o read.o symlink.o unlink.o \ + write.o namespace.o nfs-$(CONFIG_ROOT_NFS) += nfsroot.o mount_clnt.o nfs-$(CONFIG_NFS_V3) += nfs3proc.o nfs3xdr.o nfs-$(CONFIG_NFS_V3_ACL) += nfs3acl.o diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index 1b596b6d9dc2..a3ee11364db0 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -19,6 +19,7 @@ #include "nfs4_fs.h" #include "callback.h" +#include "internal.h" #define NFSDBG_FACILITY NFSDBG_CALLBACK @@ -166,15 +167,15 @@ void nfs_callback_down(void) static int nfs_callback_authenticate(struct svc_rqst *rqstp) { - struct in_addr *addr = &rqstp->rq_addr.sin_addr; + struct sockaddr_in *addr = &rqstp->rq_addr; struct nfs_client *clp; /* Don't talk to strangers */ - clp = nfs4_find_client(addr); + clp = nfs_find_client(addr, 4); if (clp == NULL) return SVC_DROP; - dprintk("%s: %u.%u.%u.%u NFSv4 callback!\n", __FUNCTION__, NIPQUAD(addr)); - nfs4_put_client(clp); + dprintk("%s: %u.%u.%u.%u NFSv4 callback!\n", __FUNCTION__, NIPQUAD(addr->sin_addr)); + nfs_put_client(clp); switch (rqstp->rq_authop->flavour) { case RPC_AUTH_NULL: if (rqstp->rq_proc != CB_NULL) diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 55d6e2ec157f..97cf8f71451f 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -10,6 +10,7 @@ #include "nfs4_fs.h" #include "callback.h" #include "delegation.h" +#include "internal.h" #define NFSDBG_FACILITY NFSDBG_CALLBACK @@ -22,7 +23,7 @@ unsigned nfs4_callback_getattr(struct cb_getattrargs *args, struct cb_getattrres res->bitmap[0] = res->bitmap[1] = 0; res->status = htonl(NFS4ERR_BADHANDLE); - clp = nfs4_find_client(&args->addr->sin_addr); + clp = nfs_find_client(args->addr, 4); if (clp == NULL) goto out; inode = nfs_delegation_find_inode(clp, &args->fh); @@ -48,7 +49,7 @@ unsigned nfs4_callback_getattr(struct cb_getattrargs *args, struct cb_getattrres up_read(&nfsi->rwsem); iput(inode); out_putclient: - nfs4_put_client(clp); + nfs_put_client(clp); out: dprintk("%s: exit with status = %d\n", __FUNCTION__, ntohl(res->status)); return res->status; @@ -61,7 +62,7 @@ unsigned nfs4_callback_recall(struct cb_recallargs *args, void *dummy) unsigned res; res = htonl(NFS4ERR_BADHANDLE); - clp = nfs4_find_client(&args->addr->sin_addr); + clp = nfs_find_client(args->addr, 4); if (clp == NULL) goto out; inode = nfs_delegation_find_inode(clp, &args->fh); @@ -80,7 +81,7 @@ unsigned nfs4_callback_recall(struct cb_recallargs *args, void *dummy) } iput(inode); out_putclient: - nfs4_put_client(clp); + nfs_put_client(clp); out: dprintk("%s: exit with status = %d\n", __FUNCTION__, ntohl(res)); return res; diff --git a/fs/nfs/client.c b/fs/nfs/client.c new file mode 100644 index 000000000000..cb5e92463bdb --- /dev/null +++ b/fs/nfs/client.c @@ -0,0 +1,312 @@ +/* client.c: NFS client sharing and management code + * + * Copyright (C) 2006 Red Hat, Inc. All Rights Reserved. + * Written by David Howells (dhowells@redhat.com) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + + +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#include "nfs4_fs.h" +#include "callback.h" +#include "delegation.h" +#include "iostat.h" +#include "internal.h" + +#define NFSDBG_FACILITY NFSDBG_CLIENT + +static DEFINE_SPINLOCK(nfs_client_lock); +static LIST_HEAD(nfs_client_list); +static DECLARE_WAIT_QUEUE_HEAD(nfs_client_active_wq); + +/* + * Allocate a shared client record + * + * Since these are allocated/deallocated very rarely, we don't + * bother putting them in a slab cache... + */ +static struct nfs_client *nfs_alloc_client(const char *hostname, + const struct sockaddr_in *addr, + int nfsversion) +{ + struct nfs_client *clp; + int error; + + if ((clp = kzalloc(sizeof(*clp), GFP_KERNEL)) == NULL) + goto error_0; + + error = rpciod_up(); + if (error < 0) { + dprintk("%s: couldn't start rpciod! Error = %d\n", + __FUNCTION__, error); + __set_bit(NFS_CS_RPCIOD, &clp->cl_res_state); + goto error_1; + } + + if (nfsversion == 4) { + if (nfs_callback_up() < 0) + goto error_2; + __set_bit(NFS_CS_CALLBACK, &clp->cl_res_state); + } + + atomic_set(&clp->cl_count, 1); + clp->cl_cons_state = NFS_CS_INITING; + + clp->cl_nfsversion = nfsversion; + memcpy(&clp->cl_addr, addr, sizeof(clp->cl_addr)); + + if (hostname) { + clp->cl_hostname = kstrdup(hostname, GFP_KERNEL); + if (!clp->cl_hostname) + goto error_3; + } + + INIT_LIST_HEAD(&clp->cl_superblocks); + clp->cl_rpcclient = ERR_PTR(-EINVAL); + +#ifdef CONFIG_NFS_V4 + init_rwsem(&clp->cl_sem); + INIT_LIST_HEAD(&clp->cl_delegations); + INIT_LIST_HEAD(&clp->cl_state_owners); + INIT_LIST_HEAD(&clp->cl_unused); + spin_lock_init(&clp->cl_lock); + INIT_WORK(&clp->cl_renewd, nfs4_renew_state, clp); + rpc_init_wait_queue(&clp->cl_rpcwaitq, "NFS client"); + clp->cl_boot_time = CURRENT_TIME; + clp->cl_state = 1 << NFS4CLNT_LEASE_EXPIRED; +#endif + + return clp; + +error_3: + nfs_callback_down(); + __clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state); +error_2: + rpciod_down(); + __clear_bit(NFS_CS_RPCIOD, &clp->cl_res_state); +error_1: + kfree(clp); +error_0: + return NULL; +} + +/* + * Destroy a shared client record + */ +static void nfs_free_client(struct nfs_client *clp) +{ + dprintk("--> nfs_free_client(%d)\n", clp->cl_nfsversion); + +#ifdef CONFIG_NFS_V4 + if (__test_and_clear_bit(NFS_CS_IDMAP, &clp->cl_res_state)) { + while (!list_empty(&clp->cl_unused)) { + struct nfs4_state_owner *sp; + + sp = list_entry(clp->cl_unused.next, + struct nfs4_state_owner, + so_list); + list_del(&sp->so_list); + kfree(sp); + } + BUG_ON(!list_empty(&clp->cl_state_owners)); + nfs_idmap_delete(clp); + } +#endif + + /* -EIO all pending I/O */ + if (!IS_ERR(clp->cl_rpcclient)) + rpc_shutdown_client(clp->cl_rpcclient); + + if (__test_and_clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state)) + nfs_callback_down(); + + if (__test_and_clear_bit(NFS_CS_RPCIOD, &clp->cl_res_state)) + rpciod_down(); + + kfree(clp->cl_hostname); + kfree(clp); + + dprintk("<-- nfs_free_client()\n"); +} + +/* + * Release a reference to a shared client record + */ +void nfs_put_client(struct nfs_client *clp) +{ + dprintk("--> nfs_put_client({%d})\n", atomic_read(&clp->cl_count)); + + if (atomic_dec_and_lock(&clp->cl_count, &nfs_client_lock)) { + list_del(&clp->cl_share_link); + spin_unlock(&nfs_client_lock); + + BUG_ON(!list_empty(&clp->cl_superblocks)); + + nfs_free_client(clp); + } +} + +/* + * Find a client by address + * - caller must hold nfs_client_lock + */ +static struct nfs_client *__nfs_find_client(const struct sockaddr_in *addr, int nfsversion) +{ + struct nfs_client *clp; + + list_for_each_entry(clp, &nfs_client_list, cl_share_link) { + /* Different NFS versions cannot share the same nfs_client */ + if (clp->cl_nfsversion != nfsversion) + continue; + + if (memcmp(&clp->cl_addr.sin_addr, &addr->sin_addr, + sizeof(clp->cl_addr.sin_addr)) != 0) + continue; + + if (clp->cl_addr.sin_port == addr->sin_port) + goto found; + } + + return NULL; + +found: + atomic_inc(&clp->cl_count); + return clp; +} + +/* + * Find a client by IP address and protocol version + * - returns NULL if no such client + */ +struct nfs_client *nfs_find_client(const struct sockaddr_in *addr, int nfsversion) +{ + struct nfs_client *clp; + + spin_lock(&nfs_client_lock); + clp = __nfs_find_client(addr, nfsversion); + spin_unlock(&nfs_client_lock); + + BUG_ON(clp->cl_cons_state == 0); + + return clp; +} + +/* + * Look up a client by IP address and protocol version + * - creates a new record if one doesn't yet exist + */ +struct nfs_client *nfs_get_client(const char *hostname, + const struct sockaddr_in *addr, + int nfsversion) +{ + struct nfs_client *clp, *new = NULL; + int error; + + dprintk("--> nfs_get_client(%s,"NIPQUAD_FMT":%d,%d)\n", + hostname ?: "", NIPQUAD(addr->sin_addr), + addr->sin_port, nfsversion); + + /* see if the client already exists */ + do { + spin_lock(&nfs_client_lock); + + clp = __nfs_find_client(addr, nfsversion); + if (clp) + goto found_client; + if (new) + goto install_client; + + spin_unlock(&nfs_client_lock); + + new = nfs_alloc_client(hostname, addr, nfsversion); + } while (new); + + return ERR_PTR(-ENOMEM); + + /* install a new client and return with it unready */ +install_client: + clp = new; + list_add(&clp->cl_share_link, &nfs_client_list); + spin_unlock(&nfs_client_lock); + dprintk("--> nfs_get_client() = %p [new]\n", clp); + return clp; + + /* found an existing client + * - make sure it's ready before returning + */ +found_client: + spin_unlock(&nfs_client_lock); + + if (new) + nfs_free_client(new); + + if (clp->cl_cons_state == NFS_CS_INITING) { + DECLARE_WAITQUEUE(myself, current); + + add_wait_queue(&nfs_client_active_wq, &myself); + + for (;;) { + set_current_state(TASK_INTERRUPTIBLE); + if (signal_pending(current) || + clp->cl_cons_state > NFS_CS_READY) + break; + schedule(); + } + + remove_wait_queue(&nfs_client_active_wq, &myself); + + if (signal_pending(current)) { + nfs_put_client(clp); + return ERR_PTR(-ERESTARTSYS); + } + } + + if (clp->cl_cons_state < NFS_CS_READY) { + error = clp->cl_cons_state; + nfs_put_client(clp); + return ERR_PTR(error); + } + + dprintk("--> nfs_get_client() = %p [share]\n", clp); + return clp; +} + +/* + * Mark a server as ready or failed + */ +void nfs_mark_client_ready(struct nfs_client *clp, int state) +{ + clp->cl_cons_state = state; + wake_up_all(&nfs_client_active_wq); +} diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index cfe239736ac0..57133678db16 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -18,6 +18,7 @@ #include "nfs4_fs.h" #include "delegation.h" +#include "internal.h" static struct nfs_delegation *nfs_alloc_delegation(void) { @@ -145,7 +146,7 @@ int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct sizeof(delegation->stateid)) != 0 || delegation->type != nfsi->delegation->type) { printk("%s: server %u.%u.%u.%u, handed out a duplicate delegation!\n", - __FUNCTION__, NIPQUAD(clp->cl_addr)); + __FUNCTION__, NIPQUAD(clp->cl_addr.sin_addr)); status = -EIO; } } @@ -254,7 +255,7 @@ int nfs_do_expire_all_delegations(void *ptr) } out: spin_unlock(&clp->cl_lock); - nfs4_put_client(clp); + nfs_put_client(clp); module_put_and_exit(0); } @@ -266,10 +267,10 @@ void nfs_expire_all_delegations(struct nfs_client *clp) atomic_inc(&clp->cl_count); task = kthread_run(nfs_do_expire_all_delegations, clp, "%u.%u.%u.%u-delegreturn", - NIPQUAD(clp->cl_addr)); + NIPQUAD(clp->cl_addr.sin_addr)); if (!IS_ERR(task)) return; - nfs4_put_client(clp); + nfs_put_client(clp); module_put(THIS_MODULE); } diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 4802157963f8..ac370d5d4494 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -15,6 +15,12 @@ struct nfs_clone_mount { rpc_authflavor_t authflavor; }; +/* client.c */ +extern void nfs_put_client(struct nfs_client *); +extern struct nfs_client *nfs_find_client(const struct sockaddr_in *, int); +extern struct nfs_client *nfs_get_client(const char *, const struct sockaddr_in *, int); +extern void nfs_mark_client_ready(struct nfs_client *, int); + /* nfs4namespace.c */ #ifdef CONFIG_NFS_V4 extern struct vfsmount *nfs_do_refmount(const struct vfsmount *mnt_parent, struct dentry *dentry); diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 4e334cb48498..e7879245361e 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -42,55 +42,6 @@ enum nfs4_client_state { NFS4CLNT_LEASE_EXPIRED, }; -/* - * The nfs_client identifies our client state to the server. - */ -struct nfs_client { - struct list_head cl_servers; /* Global list of servers */ - struct in_addr cl_addr; /* Server identifier */ - u64 cl_clientid; /* constant */ - nfs4_verifier cl_confirm; - unsigned long cl_state; - - u32 cl_lockowner_id; - - /* - * The following rwsem ensures exclusive access to the server - * while we recover the state following a lease expiration. - */ - struct rw_semaphore cl_sem; - - struct list_head cl_delegations; - struct list_head cl_state_owners; - struct list_head cl_unused; - int cl_nunused; - spinlock_t cl_lock; - atomic_t cl_count; - - struct rpc_clnt * cl_rpcclient; - - struct list_head cl_superblocks; /* List of nfs_server structs */ - - unsigned long cl_lease_time; - unsigned long cl_last_renewal; - struct work_struct cl_renewd; - struct work_struct cl_recoverd; - - struct rpc_wait_queue cl_rpcwaitq; - - /* used for the setclientid verifier */ - struct timespec cl_boot_time; - - /* idmapper */ - struct idmap * cl_idmap; - - /* Our own IP address, as a null-terminated string. - * This is used to generate the clientid, and the callback address. - */ - char cl_ipaddr[16]; - unsigned char cl_id_uniquifier; -}; - /* * struct rpc_sequence ensures that RPC calls are sent in the exact * order that they appear on the list. @@ -239,9 +190,6 @@ extern void nfs4_renew_state(void *); /* nfs4state.c */ extern void init_nfsv4_state(struct nfs_server *); extern void destroy_nfsv4_state(struct nfs_server *); -extern struct nfs_client *nfs4_get_client(struct in_addr *); -extern void nfs4_put_client(struct nfs_client *clp); -extern struct nfs_client *nfs4_find_client(struct in_addr *); struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp); extern u32 nfs4_alloc_lockowner_id(struct nfs_client *); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 850f0851023a..803c31b88bb5 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2968,7 +2968,7 @@ int nfs4_proc_setclientid(struct nfs_client *clp, u32 program, unsigned short po for(;;) { setclientid.sc_name_len = scnprintf(setclientid.sc_name, sizeof(setclientid.sc_name), "%s/%u.%u.%u.%u %s %u", - clp->cl_ipaddr, NIPQUAD(clp->cl_addr.s_addr), + clp->cl_ipaddr, NIPQUAD(clp->cl_addr.sin_addr), cred->cr_ops->cr_name, clp->cl_id_uniquifier); setclientid.sc_netid_len = scnprintf(setclientid.sc_netid, diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index fa51a7d4c022..058811e39555 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -50,12 +50,12 @@ #include "nfs4_fs.h" #include "callback.h" #include "delegation.h" +#include "internal.h" #define OPENOWNER_POOL_SIZE 8 const nfs4_stateid zero_stateid; -static DEFINE_SPINLOCK(state_spinlock); static LIST_HEAD(nfs4_clientid_list); void @@ -71,127 +71,11 @@ destroy_nfsv4_state(struct nfs_server *server) kfree(server->mnt_path); server->mnt_path = NULL; if (server->nfs_client) { - nfs4_put_client(server->nfs_client); + nfs_put_client(server->nfs_client); server->nfs_client = NULL; } } -/* - * nfs4_get_client(): returns an empty client structure - * nfs4_put_client(): drops reference to client structure - * - * Since these are allocated/deallocated very rarely, we don't - * bother putting them in a slab cache... - */ -static struct nfs_client * -nfs4_alloc_client(struct in_addr *addr) -{ - struct nfs_client *clp; - - if (nfs_callback_up() < 0) - return NULL; - if ((clp = kzalloc(sizeof(*clp), GFP_KERNEL)) == NULL) { - nfs_callback_down(); - return NULL; - } - memcpy(&clp->cl_addr, addr, sizeof(clp->cl_addr)); - init_rwsem(&clp->cl_sem); - INIT_LIST_HEAD(&clp->cl_delegations); - INIT_LIST_HEAD(&clp->cl_state_owners); - INIT_LIST_HEAD(&clp->cl_unused); - spin_lock_init(&clp->cl_lock); - atomic_set(&clp->cl_count, 1); - INIT_WORK(&clp->cl_renewd, nfs4_renew_state, clp); - INIT_LIST_HEAD(&clp->cl_superblocks); - rpc_init_wait_queue(&clp->cl_rpcwaitq, "NFS4 client"); - clp->cl_rpcclient = ERR_PTR(-EINVAL); - clp->cl_boot_time = CURRENT_TIME; - clp->cl_state = 1 << NFS4CLNT_LEASE_EXPIRED; - return clp; -} - -static void -nfs4_free_client(struct nfs_client *clp) -{ - struct nfs4_state_owner *sp; - - while (!list_empty(&clp->cl_unused)) { - sp = list_entry(clp->cl_unused.next, - struct nfs4_state_owner, - so_list); - list_del(&sp->so_list); - kfree(sp); - } - BUG_ON(!list_empty(&clp->cl_state_owners)); - nfs_idmap_delete(clp); - if (!IS_ERR(clp->cl_rpcclient)) - rpc_shutdown_client(clp->cl_rpcclient); - kfree(clp); - nfs_callback_down(); -} - -static struct nfs_client *__nfs4_find_client(struct in_addr *addr) -{ - struct nfs_client *clp; - list_for_each_entry(clp, &nfs4_clientid_list, cl_servers) { - if (memcmp(&clp->cl_addr, addr, sizeof(clp->cl_addr)) == 0) { - atomic_inc(&clp->cl_count); - return clp; - } - } - return NULL; -} - -struct nfs_client *nfs4_find_client(struct in_addr *addr) -{ - struct nfs_client *clp; - spin_lock(&state_spinlock); - clp = __nfs4_find_client(addr); - spin_unlock(&state_spinlock); - return clp; -} - -struct nfs_client * -nfs4_get_client(struct in_addr *addr) -{ - struct nfs_client *clp, *new = NULL; - - spin_lock(&state_spinlock); - for (;;) { - clp = __nfs4_find_client(addr); - if (clp != NULL) - break; - clp = new; - if (clp != NULL) { - list_add(&clp->cl_servers, &nfs4_clientid_list); - new = NULL; - break; - } - spin_unlock(&state_spinlock); - new = nfs4_alloc_client(addr); - spin_lock(&state_spinlock); - if (new == NULL) - break; - } - spin_unlock(&state_spinlock); - if (new) - nfs4_free_client(new); - return clp; -} - -void -nfs4_put_client(struct nfs_client *clp) -{ - if (!atomic_dec_and_lock(&clp->cl_count, &state_spinlock)) - return; - list_del(&clp->cl_servers); - spin_unlock(&state_spinlock); - BUG_ON(!list_empty(&clp->cl_superblocks)); - rpc_wake_up(&clp->cl_rpcwaitq); - nfs4_kill_renewd(clp); - nfs4_free_client(clp); -} - static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred) { int status = nfs4_proc_setclientid(clp, NFS4_CALLBACK, @@ -771,11 +655,11 @@ static void nfs4_recover_state(struct nfs_client *clp) __module_get(THIS_MODULE); atomic_inc(&clp->cl_count); task = kthread_run(reclaimer, clp, "%u.%u.%u.%u-reclaim", - NIPQUAD(clp->cl_addr)); + NIPQUAD(clp->cl_addr.sin_addr)); if (!IS_ERR(task)) return; nfs4_clear_recover_bit(clp); - nfs4_put_client(clp); + nfs_put_client(clp); module_put(THIS_MODULE); } @@ -970,12 +854,12 @@ static int reclaimer(void *ptr) if (status == -NFS4ERR_CB_PATH_DOWN) nfs_handle_cb_pathdown(clp); nfs4_clear_recover_bit(clp); - nfs4_put_client(clp); + nfs_put_client(clp); module_put_and_exit(0); return 0; out_error: printk(KERN_WARNING "Error: state recovery failed on NFSv4 server %u.%u.%u.%u with error %d\n", - NIPQUAD(clp->cl_addr.s_addr), -status); + NIPQUAD(clp->cl_addr.sin_addr), -status); set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state); goto out; } diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 3ee85c4e65d8..f97d7d9c5c32 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1104,47 +1104,46 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, struct rpc_clnt *clnt = NULL; int err = -EIO; - clp = nfs4_get_client(&server->addr.sin_addr); + clp = nfs_get_client(server->hostname, &server->addr, 4); if (!clp) { dprintk("%s: failed to create NFS4 client.\n", __FUNCTION__); return ERR_PTR(err); } /* Now create transport and client */ - down_write(&clp->cl_sem); - if (IS_ERR(clp->cl_rpcclient)) { + if (clp->cl_cons_state == NFS_CS_INITING) { xprt = xprt_create_proto(proto, &server->addr, timeparms); if (IS_ERR(xprt)) { - up_write(&clp->cl_sem); err = PTR_ERR(xprt); dprintk("%s: cannot create RPC transport. Error = %d\n", __FUNCTION__, err); - goto out_fail; + goto client_init_error; } /* Bind to a reserved port! */ xprt->resvport = 1; clnt = rpc_create_client(xprt, server->hostname, &nfs_program, server->rpc_ops->version, flavor); if (IS_ERR(clnt)) { - up_write(&clp->cl_sem); err = PTR_ERR(clnt); dprintk("%s: cannot create RPC client. Error = %d\n", __FUNCTION__, err); - goto out_fail; + goto client_init_error; } clnt->cl_intr = 1; clnt->cl_softrtry = 1; clp->cl_rpcclient = clnt; memcpy(clp->cl_ipaddr, server->ip_addr, sizeof(clp->cl_ipaddr)); - if (nfs_idmap_new(clp) < 0) - goto out_fail; + err = nfs_idmap_new(clp); + if (err < 0) { + dprintk("%s: failed to create idmapper.\n", + __FUNCTION__); + goto client_init_error; + } + __set_bit(NFS_CS_IDMAP, &clp->cl_res_state); + nfs_mark_client_ready(clp, 0); } - list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks); + clnt = rpc_clone_client(clp->cl_rpcclient); - if (!IS_ERR(clnt)) - server->nfs_client = clp; - up_write(&clp->cl_sem); - clp = NULL; if (IS_ERR(clnt)) { dprintk("%s: cannot create RPC client. Error = %d\n", @@ -1152,11 +1151,6 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, return clnt; } - if (server->nfs_client->cl_idmap == NULL) { - dprintk("%s: failed to create idmapper.\n", __FUNCTION__); - return ERR_PTR(-ENOMEM); - } - if (clnt->cl_auth->au_flavor != flavor) { struct rpc_auth *auth; @@ -1166,11 +1160,16 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, return (struct rpc_clnt *)auth; } } + + server->nfs_client = clp; + down_write(&clp->cl_sem); + list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks); + up_write(&clp->cl_sem); return clnt; - out_fail: - if (clp) - nfs4_put_client(clp); +client_init_error: + nfs_mark_client_ready(clp, err); + nfs_put_client(clp); return ERR_PTR(err); } @@ -1329,14 +1328,6 @@ static int nfs4_get_sb(struct file_system_type *fs_type, goto out_free; } - /* Fire up rpciod if not yet running */ - error = rpciod_up(); - if (error < 0) { - dprintk("%s: couldn't start rpciod! Error = %d\n", - __FUNCTION__, error); - goto out_free; - } - s = sget(fs_type, nfs4_compare_super, nfs_set_super, server); if (IS_ERR(s)) { error = PTR_ERR(s); @@ -1383,8 +1374,6 @@ static void nfs4_kill_super(struct super_block *sb) destroy_nfsv4_state(server); - rpciod_down(); - nfs_free_iostats(server->io_stats); kfree(server->hostname); kfree(server); diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index a36e01cd6321..70e1dc9162e2 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -586,6 +586,7 @@ extern void * nfs_root_data(void); #define NFSDBG_FILE 0x0040 #define NFSDBG_ROOT 0x0080 #define NFSDBG_CALLBACK 0x0100 +#define NFSDBG_CLIENT 0x0200 #define NFSDBG_ALL 0xFFFF #ifdef __KERNEL__ diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index fc20d6b934fb..a727657e0ad3 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -6,6 +6,66 @@ struct nfs_iostats; +/* + * The nfs_client identifies our client state to the server. + */ +struct nfs_client { + atomic_t cl_count; + int cl_cons_state; /* current construction state (-ve: init error) */ +#define NFS_CS_READY 0 /* ready to be used */ +#define NFS_CS_INITING 1 /* busy initialising */ + int cl_nfsversion; /* NFS protocol version */ + unsigned long cl_res_state; /* NFS resources state */ +#define NFS_CS_RPCIOD 0 /* - rpciod started */ +#define NFS_CS_CALLBACK 1 /* - callback started */ +#define NFS_CS_IDMAP 2 /* - idmap started */ + struct sockaddr_in cl_addr; /* server identifier */ + char * cl_hostname; /* hostname of server */ + struct list_head cl_share_link; /* link in global client list */ + struct list_head cl_superblocks; /* List of nfs_server structs */ + + struct rpc_clnt * cl_rpcclient; + +#ifdef CONFIG_NFS_V4 + u64 cl_clientid; /* constant */ + nfs4_verifier cl_confirm; + unsigned long cl_state; + + u32 cl_lockowner_id; + + /* + * The following rwsem ensures exclusive access to the server + * while we recover the state following a lease expiration. + */ + struct rw_semaphore cl_sem; + + struct list_head cl_delegations; + struct list_head cl_state_owners; + struct list_head cl_unused; + int cl_nunused; + spinlock_t cl_lock; + + unsigned long cl_lease_time; + unsigned long cl_last_renewal; + struct work_struct cl_renewd; + struct work_struct cl_recoverd; + + struct rpc_wait_queue cl_rpcwaitq; + + /* used for the setclientid verifier */ + struct timespec cl_boot_time; + + /* idmapper */ + struct idmap * cl_idmap; + + /* Our own IP address, as a null-terminated string. + * This is used to generate the clientid, and the callback address. + */ + char cl_ipaddr[16]; + unsigned char cl_id_uniquifier; +#endif +}; + /* * NFS client parameters stored in the superblock. */ From 0c7d90cfed91a283228017ba6faf37ee0bcd32b1 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:10 -0400 Subject: [PATCH 14/74] NFS: Use the dentry superblock directly in nfs_statfs() Use the nominated dentry's superblock directly in the NFS statfs() op to get a file handle, rather than using s_root (which will become a dummy dentry in a future patch). Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index f97d7d9c5c32..a41d516ed595 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -240,11 +240,10 @@ void __exit unregister_nfs_fs(void) */ static int nfs_statfs(struct dentry *dentry, struct kstatfs *buf) { - struct super_block *sb = dentry->d_sb; - struct nfs_server *server = NFS_SB(sb); + struct nfs_server *server = NFS_SB(dentry->d_sb); unsigned char blockbits; unsigned long blockres; - struct nfs_fh *rootfh = NFS_FH(sb->s_root->d_inode); + struct nfs_fh *fh = NFS_FH(dentry->d_inode); struct nfs_fattr fattr; struct nfs_fsstat res = { .fattr = &fattr, @@ -253,7 +252,7 @@ static int nfs_statfs(struct dentry *dentry, struct kstatfs *buf) lock_kernel(); - error = server->rpc_ops->statfs(server, rootfh, &res); + error = server->rpc_ops->statfs(server, fh, &res); buf->f_type = NFS_SUPER_MAGIC; if (error < 0) goto out_err; @@ -263,7 +262,7 @@ static int nfs_statfs(struct dentry *dentry, struct kstatfs *buf) * case where f_frsize != f_bsize. Eventually we want to * report the value of wtmult in this field. */ - buf->f_frsize = sb->s_blocksize; + buf->f_frsize = dentry->d_sb->s_blocksize; /* * On most *nix systems, f_blocks, f_bfree, and f_bavail @@ -272,8 +271,8 @@ static int nfs_statfs(struct dentry *dentry, struct kstatfs *buf) * thus historically Linux's sys_statfs reports these * fields in units of f_bsize. */ - buf->f_bsize = sb->s_blocksize; - blockbits = sb->s_blocksize_bits; + buf->f_bsize = dentry->d_sb->s_blocksize; + blockbits = dentry->d_sb->s_blocksize_bits; blockres = (1 << blockbits) - 1; buf->f_blocks = (res.tbytes + blockres) >> blockbits; buf->f_bfree = (res.fbytes + blockres) >> blockbits; From 509de8111656a7d89b4a1a5f430f4460ce510f0f Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:11 -0400 Subject: [PATCH 15/74] NFS: Add extra const qualifiers Add some extra const qualifiers into NFS. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/namespace.c | 3 ++- fs/nfs/nfs3proc.c | 2 +- fs/nfs/nfs4namespace.c | 8 ++++---- fs/nfs/nfs4proc.c | 2 +- fs/nfs/proc.c | 2 +- fs/nfs/super.c | 10 +++++----- include/linux/nfs_fs_sb.h | 2 +- include/linux/nfs_xdr.h | 6 +++--- 8 files changed, 18 insertions(+), 17 deletions(-) diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c index 86b3169c8cac..85d9ed1dcf42 100644 --- a/fs/nfs/namespace.c +++ b/fs/nfs/namespace.c @@ -172,7 +172,8 @@ void nfs_release_automount_timer(void) /* * Clone a mountpoint of the appropriate type */ -static struct vfsmount *nfs_do_clone_mount(struct nfs_server *server, char *devname, +static struct vfsmount *nfs_do_clone_mount(struct nfs_server *server, + const char *devname, struct nfs_clone_mount *mountdata) { #ifdef CONFIG_NFS_V4 diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c index 7143b1f82cea..3e5371241cea 100644 --- a/fs/nfs/nfs3proc.c +++ b/fs/nfs/nfs3proc.c @@ -886,7 +886,7 @@ nfs3_proc_lock(struct file *filp, int cmd, struct file_lock *fl) return nlmclnt_proc(filp->f_dentry->d_inode, cmd, fl); } -struct nfs_rpc_ops nfs_v3_clientops = { +const struct nfs_rpc_ops nfs_v3_clientops = { .version = 3, /* protocol version */ .dentry_ops = &nfs_dentry_operations, .dir_inode_ops = &nfs3_dir_inode_operations, diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c index ea38d27b74e6..faed9bcba50f 100644 --- a/fs/nfs/nfs4namespace.c +++ b/fs/nfs/nfs4namespace.c @@ -23,7 +23,7 @@ /* * Check if fs_root is valid */ -static inline char *nfs4_pathname_string(struct nfs4_pathname *pathname, +static inline char *nfs4_pathname_string(const struct nfs4_pathname *pathname, char *buffer, ssize_t buflen) { char *end = buffer + buflen; @@ -34,7 +34,7 @@ static inline char *nfs4_pathname_string(struct nfs4_pathname *pathname, n = pathname->ncomponents; while (--n >= 0) { - struct nfs4_string *component = &pathname->components[n]; + const struct nfs4_string *component = &pathname->components[n]; buflen -= component->len + 1; if (buflen < 0) goto Elong; @@ -60,7 +60,7 @@ static inline char *nfs4_pathname_string(struct nfs4_pathname *pathname, */ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent, const struct dentry *dentry, - struct nfs4_fs_locations *locations) + const struct nfs4_fs_locations *locations) { struct vfsmount *mnt = ERR_PTR(-ENOENT); struct nfs_clone_mount mountdata = { @@ -108,7 +108,7 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent, loc = 0; while (loc < locations->nlocations && IS_ERR(mnt)) { - struct nfs4_fs_location *location = &locations->locations[loc]; + const struct nfs4_fs_location *location = &locations->locations[loc]; char *mnt_path; if (location == NULL || location->nservers <= 0 || diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 803c31b88bb5..061be713b206 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3761,7 +3761,7 @@ static struct inode_operations nfs4_file_inode_operations = { .listxattr = nfs4_listxattr, }; -struct nfs_rpc_ops nfs_v4_clientops = { +const struct nfs_rpc_ops nfs_v4_clientops = { .version = 4, /* protocol version */ .dentry_ops = &nfs4_dentry_operations, .dir_inode_ops = &nfs4_dir_inode_operations, diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c index b3899ea3229e..77676903e0f5 100644 --- a/fs/nfs/proc.c +++ b/fs/nfs/proc.c @@ -671,7 +671,7 @@ nfs_proc_lock(struct file *filp, int cmd, struct file_lock *fl) } -struct nfs_rpc_ops nfs_v2_clientops = { +const struct nfs_rpc_ops nfs_v2_clientops = { .version = 2, /* protocol version */ .dentry_ops = &nfs_dentry_operations, .dir_inode_ops = &nfs_dir_inode_operations, diff --git a/fs/nfs/super.c b/fs/nfs/super.c index a41d516ed595..c97f30967955 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -329,10 +329,10 @@ static const char *nfs_pseudoflavour_to_name(rpc_authflavor_t flavour) */ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, int showdefaults) { - static struct proc_nfs_info { + static const struct proc_nfs_info { int flag; - char *str; - char *nostr; + const char *str; + const char *nostr; } nfs_info[] = { { NFS_MOUNT_SOFT, ",soft", ",hard" }, { NFS_MOUNT_INTR, ",intr", "" }, @@ -342,9 +342,9 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, { NFS_MOUNT_NOACL, ",noacl", "" }, { 0, NULL, NULL } }; - struct proc_nfs_info *nfs_infop; + const struct proc_nfs_info *nfs_infop; char buf[12]; - char *proto; + const char *proto; seq_printf(m, ",vers=%d", nfss->rpc_ops->version); seq_printf(m, ",rsize=%d", nfss->rsize); diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index a727657e0ad3..95f32d5f6e9c 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -73,7 +73,7 @@ struct nfs_server { struct rpc_clnt * client; /* RPC client handle */ struct rpc_clnt * client_sys; /* 2nd handle for FSINFO */ struct rpc_clnt * client_acl; /* ACL RPC client handle */ - struct nfs_rpc_ops * rpc_ops; /* NFS protocol vector */ + const struct nfs_rpc_ops *rpc_ops; /* NFS protocol vector */ struct nfs_iostats * io_stats; /* I/O statistics */ struct backing_dev_info backing_dev_info; int flags; /* various flags */ diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index dd9ae6761f71..2426b11b6cce 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -833,9 +833,9 @@ struct nfs_rpc_ops { /* * Function vectors etc. for the NFS client */ -extern struct nfs_rpc_ops nfs_v2_clientops; -extern struct nfs_rpc_ops nfs_v3_clientops; -extern struct nfs_rpc_ops nfs_v4_clientops; +extern const struct nfs_rpc_ops nfs_v2_clientops; +extern const struct nfs_rpc_ops nfs_v3_clientops; +extern const struct nfs_rpc_ops nfs_v4_clientops; extern struct rpc_version nfs_version2; extern struct rpc_version nfs_version3; extern struct rpc_version nfs_version4; From 27951bd26031f6c27d38df9e94623bbe208a2464 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:11 -0400 Subject: [PATCH 16/74] NFS: Maintain a common server record for NFS2/3 as well as for NFS4 Maintain a common server record for NFS2/3 as well as for NFS4 so that common stuff can be moved there from struct nfs_server. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 21 ++++++++++++++++++++- include/linux/nfs_fs_sb.h | 2 +- 2 files changed, 21 insertions(+), 2 deletions(-) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index c97f30967955..d1b4a5b36e33 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -658,11 +658,19 @@ static void nfs_init_timeout_values(struct rpc_timeout *to, int proto, unsigned static struct rpc_clnt * nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) { + struct nfs_client *clp; struct rpc_timeout timeparms; struct rpc_xprt *xprt = NULL; struct rpc_clnt *clnt = NULL; int proto = (data->flags & NFS_MOUNT_TCP) ? IPPROTO_TCP : IPPROTO_UDP; + clp = nfs_get_client(server->hostname, &server->addr, + server->rpc_ops->version); + if (!clp) { + dprintk("%s: failed to create NFS4 client.\n", __FUNCTION__); + return ERR_PTR(PTR_ERR(clp)); + } + nfs_init_timeout_values(&timeparms, proto, data->timeo, data->retrans); server->retrans_timeo = timeparms.to_initval; @@ -673,6 +681,8 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) if (IS_ERR(xprt)) { dprintk("%s: cannot create RPC transport. Error = %ld\n", __FUNCTION__, PTR_ERR(xprt)); + nfs_mark_client_ready(clp, PTR_ERR(xprt)); + nfs_put_client(clp); return (struct rpc_clnt *)xprt; } clnt = rpc_create_client(xprt, server->hostname, &nfs_program, @@ -686,9 +696,13 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) clnt->cl_intr = 1; clnt->cl_softrtry = 1; + nfs_mark_client_ready(clp, 0); + server->nfs_client = clp; return clnt; out_fail: + nfs_mark_client_ready(clp, PTR_ERR(xprt)); + nfs_put_client(clp); return clnt; } @@ -764,6 +778,7 @@ static int nfs_clone_generic_sb(struct nfs_clone_mount *data, if (server == NULL) goto out_err; memcpy(server, parent, sizeof(*server)); + atomic_inc(&server->nfs_client->cl_count); hostname = (data->hostname != NULL) ? data->hostname : parent->hostname; len = strlen(hostname) + 1; server->hostname = kmalloc(len, GFP_KERNEL); @@ -796,6 +811,7 @@ static int nfs_clone_generic_sb(struct nfs_clone_mount *data, out_rpciod_down: rpciod_down(); kfree(server->hostname); + nfs_put_client(server->nfs_client); kfree(server); return simple_set_mnt(mnt, sb); kill_rpciod: @@ -803,6 +819,7 @@ static int nfs_clone_generic_sb(struct nfs_clone_mount *data, free_hostname: kfree(server->hostname); free_server: + nfs_put_client(server->nfs_client); kfree(server); out_err: return error; @@ -1071,6 +1088,7 @@ static void nfs_kill_super(struct super_block *s) nfs_free_iostats(server->io_stats); kfree(server->hostname); + nfs_put_client(server->nfs_client); kfree(server); nfs_release_automount_timer(); } @@ -1421,7 +1439,6 @@ static struct super_block *nfs4_clone_sb(struct nfs_server *server, struct nfs_c nfs4_server_capabilities(server, &server->fh); down_write(&clp->cl_sem); - atomic_inc(&clp->cl_count); list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks); up_write(&clp->cl_sem); return sb; @@ -1476,6 +1493,8 @@ static struct nfs_server *nfs4_referral_server(struct super_block *sb, struct nf retrans = 1; nfs_init_timeout_values(&timeparms, proto, timeo, retrans); + nfs_put_client(server->nfs_client); + server->nfs_client = NULL; server->client = nfs4_create_client(server, &timeparms, proto, data->authflavor); if (IS_ERR((err = server->client))) goto out_err; diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 95f32d5f6e9c..e7d7662f51fd 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -70,6 +70,7 @@ struct nfs_client { * NFS client parameters stored in the superblock. */ struct nfs_server { + struct nfs_client * nfs_client; /* shared client and NFS4 state */ struct rpc_clnt * client; /* RPC client handle */ struct rpc_clnt * client_sys; /* 2nd handle for FSINFO */ struct rpc_clnt * client_acl; /* ACL RPC client handle */ @@ -103,7 +104,6 @@ struct nfs_server { */ char ip_addr[16]; char * mnt_path; - struct nfs_client * nfs_client; /* all NFSv4 state starts here */ struct list_head nfs4_siblings; /* List of other nfs_server structs * that share the same clientid */ From 1f163415dc05983830bcc47b33c155b2528b1574 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:11 -0400 Subject: [PATCH 17/74] NFS: Make better use of inode* dereferencing macros Make better use of inode* dereferencing macros to hide dereferencing chains (including NFS_PROTO and NFS_CLIENT). Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 2 +- fs/nfs/nfs4proc.c | 10 +++++----- 2 files changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 48e892880d5b..a146ed338534 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -111,7 +111,7 @@ nfs_file_open(struct inode *inode, struct file *filp) nfs_inc_stats(inode, NFSIOS_VFSOPEN); lock_kernel(); - res = NFS_SERVER(inode)->rpc_ops->file_open(inode, filp); + res = NFS_PROTO(inode)->file_open(inode, filp); unlock_kernel(); return res; } diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 061be713b206..b731b1945270 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -1268,7 +1268,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd) BUG_ON(nd->intent.open.flags & O_CREAT); } - cred = rpcauth_lookupcred(NFS_SERVER(dir)->client->cl_auth, 0); + cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0); if (IS_ERR(cred)) return (struct dentry *)cred; state = nfs4_do_open(dir, dentry, nd->intent.open.flags, &attr, cred); @@ -1291,7 +1291,7 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st struct rpc_cred *cred; struct nfs4_state *state; - cred = rpcauth_lookupcred(NFS_SERVER(dir)->client->cl_auth, 0); + cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0); if (IS_ERR(cred)) return PTR_ERR(cred); state = nfs4_open_delegated(dentry->d_inode, openflags, cred); @@ -1565,7 +1565,7 @@ nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr, nfs_fattr_init(fattr); - cred = rpcauth_lookupcred(NFS_SERVER(inode)->client->cl_auth, 0); + cred = rpcauth_lookupcred(NFS_CLIENT(inode)->cl_auth, 0); if (IS_ERR(cred)) return PTR_ERR(cred); @@ -1927,7 +1927,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr, struct rpc_cred *cred; int status = 0; - cred = rpcauth_lookupcred(NFS_SERVER(dir)->client->cl_auth, 0); + cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0); if (IS_ERR(cred)) { status = PTR_ERR(cred); goto out; @@ -2816,7 +2816,7 @@ static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t bufl return -EOPNOTSUPP; nfs_inode_return_delegation(inode); buf_to_pages(buf, buflen, arg.acl_pages, &arg.acl_pgbase); - ret = rpc_call_sync(NFS_SERVER(inode)->client, &msg, 0); + ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0); if (ret == 0) nfs4_write_cached_acl(inode, buf, buflen); return ret; From 8fa5c000d7f986ef9cdc6d95f9f7fcee20e0a7d6 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:12 -0400 Subject: [PATCH 18/74] NFS: Move rpc_ops from nfs_server to nfs_client Move the rpc_ops from the nfs_server struct to the nfs_client struct as they're common to all server records of a particular NFS protocol version. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 2 +- fs/nfs/inode.c | 4 +-- fs/nfs/namespace.c | 6 ++-- fs/nfs/nfs4proc.c | 2 +- fs/nfs/super.c | 59 ++++++++++++++++++++++----------------- include/linux/nfs_fs.h | 2 +- include/linux/nfs_fs_sb.h | 2 +- 7 files changed, 43 insertions(+), 34 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 067d144d141b..19362712452f 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1147,7 +1147,7 @@ int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle, } if (!(fattr->valid & NFS_ATTR_FATTR)) { struct nfs_server *server = NFS_SB(dentry->d_sb); - error = server->rpc_ops->getattr(server, fhandle, fattr); + error = server->nfs_client->rpc_ops->getattr(server, fhandle, fattr); if (error < 0) goto out_err; } diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 6ed018c9aad2..771c3b833757 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -237,13 +237,13 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr) /* Why so? Because we want revalidate for devices/FIFOs, and * that's precisely what we have in nfs_file_inode_operations. */ - inode->i_op = NFS_SB(sb)->rpc_ops->file_inode_ops; + inode->i_op = NFS_SB(sb)->nfs_client->rpc_ops->file_inode_ops; if (S_ISREG(inode->i_mode)) { inode->i_fop = &nfs_file_operations; inode->i_data.a_ops = &nfs_file_aops; inode->i_data.backing_dev_info = &NFS_SB(sb)->backing_dev_info; } else if (S_ISDIR(inode->i_mode)) { - inode->i_op = NFS_SB(sb)->rpc_ops->dir_inode_ops; + inode->i_op = NFS_SB(sb)->nfs_client->rpc_ops->dir_inode_ops; inode->i_fop = &nfs_dir_operations; if (nfs_server_capable(inode, NFS_CAP_READDIRPLUS) && fattr->size <= NFS_LIMIT_READDIRPLUS) diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c index 85d9ed1dcf42..d8b8d56266cb 100644 --- a/fs/nfs/namespace.c +++ b/fs/nfs/namespace.c @@ -104,7 +104,9 @@ static void * nfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd) goto out_follow; /* Look it up again */ parent = dget_parent(nd->dentry); - err = server->rpc_ops->lookup(parent->d_inode, &nd->dentry->d_name, &fh, &fattr); + err = server->nfs_client->rpc_ops->lookup(parent->d_inode, + &nd->dentry->d_name, + &fh, &fattr); dput(parent); if (err != 0) goto out_err; @@ -178,7 +180,7 @@ static struct vfsmount *nfs_do_clone_mount(struct nfs_server *server, { #ifdef CONFIG_NFS_V4 struct vfsmount *mnt = NULL; - switch (server->rpc_ops->version) { + switch (server->nfs_client->cl_nfsversion) { case 2: case 3: mnt = vfs_kern_mount(&clone_nfs_fs_type, 0, devname, mountdata); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index b731b1945270..1573eeb07ce1 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -758,7 +758,7 @@ static int _nfs4_proc_open(struct nfs4_opendata *data) } nfs_confirm_seqid(&data->owner->so_seqid, 0); if (!(o_res->f_attr->valid & NFS_ATTR_FATTR)) - return server->rpc_ops->getattr(server, &o_res->fh, o_res->f_attr); + return server->nfs_client->rpc_ops->getattr(server, &o_res->fh, o_res->f_attr); return 0; } diff --git a/fs/nfs/super.c b/fs/nfs/super.c index d1b4a5b36e33..e1e5eab0259b 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -252,7 +252,7 @@ static int nfs_statfs(struct dentry *dentry, struct kstatfs *buf) lock_kernel(); - error = server->rpc_ops->statfs(server, fh, &res); + error = server->nfs_client->rpc_ops->statfs(server, fh, &res); buf->f_type = NFS_SUPER_MAGIC; if (error < 0) goto out_err; @@ -343,10 +343,11 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, { 0, NULL, NULL } }; const struct proc_nfs_info *nfs_infop; + struct nfs_client *clp = nfss->nfs_client; char buf[12]; const char *proto; - seq_printf(m, ",vers=%d", nfss->rpc_ops->version); + seq_printf(m, ",vers=%d", clp->rpc_ops->version); seq_printf(m, ",rsize=%d", nfss->rsize); seq_printf(m, ",wsize=%d", nfss->wsize); if (nfss->acregmin != 3*HZ || showdefaults) @@ -427,7 +428,7 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt) seq_printf(m, ",namelen=%d", nfss->namelen); #ifdef CONFIG_NFS_V4 - if (nfss->rpc_ops->version == 4) { + if (nfss->nfs_client->cl_nfsversion == 4) { seq_printf(m, "\n\tnfsv4:\t"); seq_printf(m, "bm0=0x%x", nfss->attr_bitmask[0]); seq_printf(m, ",bm1=0x%x", nfss->attr_bitmask[1]); @@ -503,7 +504,7 @@ nfs_get_root(struct super_block *sb, struct nfs_fh *rootfh, struct nfs_fsinfo *f struct nfs_server *server = NFS_SB(sb); int error; - error = server->rpc_ops->getroot(server, rootfh, fsinfo); + error = server->nfs_client->rpc_ops->getroot(server, rootfh, fsinfo); if (error < 0) { dprintk("nfs_get_root: getattr error = %d\n", -error); return ERR_PTR(error); @@ -553,14 +554,14 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor) no_root_error = -ENOMEM; goto out_no_root; } - sb->s_root->d_op = server->rpc_ops->dentry_ops; + sb->s_root->d_op = server->nfs_client->rpc_ops->dentry_ops; /* mount time stamp, in seconds */ server->mount_time = jiffies; /* Get some general file system info */ if (server->namelen == 0 && - server->rpc_ops->pathconf(server, &server->fh, &pathinfo) >= 0) + server->nfs_client->rpc_ops->pathconf(server, &server->fh, &pathinfo) >= 0) server->namelen = pathinfo.max_namelen; /* Work out a lot of parameters */ if (server->rsize == 0) @@ -663,9 +664,14 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) struct rpc_xprt *xprt = NULL; struct rpc_clnt *clnt = NULL; int proto = (data->flags & NFS_MOUNT_TCP) ? IPPROTO_TCP : IPPROTO_UDP; + int nfsversion = 2; - clp = nfs_get_client(server->hostname, &server->addr, - server->rpc_ops->version); +#ifdef CONFIG_NFS_V3 + if (server->flags & NFS_MOUNT_VER3) + nfsversion = 3; +#endif + + clp = nfs_get_client(server->hostname, &server->addr, nfsversion); if (!clp) { dprintk("%s: failed to create NFS4 client.\n", __FUNCTION__); return ERR_PTR(PTR_ERR(clp)); @@ -676,6 +682,19 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) server->retrans_timeo = timeparms.to_initval; server->retrans_count = timeparms.to_retries; + /* Check NFS protocol revision and initialize RPC op vector + * and file handle pool. */ +#ifdef CONFIG_NFS_V3 + if (nfsversion == 3) { + clp->rpc_ops = &nfs_v3_clientops; + server->caps |= NFS_CAP_READDIRPLUS; + } else { + clp->rpc_ops = &nfs_v2_clientops; + } +#else + clp->rpc_ops = &nfs_v2_clientops; +#endif + /* create transport and client */ xprt = xprt_create_proto(proto, &server->addr, &timeparms); if (IS_ERR(xprt)) { @@ -686,7 +705,7 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) return (struct rpc_clnt *)xprt; } clnt = rpc_create_client(xprt, server->hostname, &nfs_program, - server->rpc_ops->version, data->pseudoflavor); + clp->cl_nfsversion, data->pseudoflavor); if (IS_ERR(clnt)) { dprintk("%s: cannot create RPC client. Error = %ld\n", __FUNCTION__, PTR_ERR(xprt)); @@ -750,7 +769,7 @@ static struct nfs_server *nfs_clone_server(struct super_block *sb, struct nfs_cl fsinfo.fattr = data->fattr; if (NFS_PROTO(root_inode)->fsinfo(server, data->fh, &fsinfo) == 0) nfs_super_set_maxbytes(sb, fsinfo.maxfilesize); - sb->s_root->d_op = server->rpc_ops->dentry_ops; + sb->s_root->d_op = server->nfs_client->rpc_ops->dentry_ops; sb->s_flags |= MS_ACTIVE; return server; out_put_root: @@ -865,19 +884,6 @@ nfs_fill_super(struct super_block *sb, struct nfs_mount_data *data, int silent) return -ENOMEM; strcpy(server->hostname, data->hostname); - /* Check NFS protocol revision and initialize RPC op vector - * and file handle pool. */ -#ifdef CONFIG_NFS_V3 - if (server->flags & NFS_MOUNT_VER3) { - server->rpc_ops = &nfs_v3_clientops; - server->caps |= NFS_CAP_READDIRPLUS; - } else { - server->rpc_ops = &nfs_v2_clientops; - } -#else - server->rpc_ops = &nfs_v2_clientops; -#endif - /* Fill in pseudoflavor for mount version < 5 */ if (!(data->flags & NFS_MOUNT_SECFLAVOUR)) data->pseudoflavor = RPC_AUTH_UNIX; @@ -888,6 +894,7 @@ nfs_fill_super(struct super_block *sb, struct nfs_mount_data *data, int silent) server->client = nfs_create_client(server, data); if (IS_ERR(server->client)) return PTR_ERR(server->client); + /* RFC 2623, sec 2.3.2 */ if (authflavor != RPC_AUTH_UNIX) { struct rpc_auth *auth; @@ -1129,6 +1136,8 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, /* Now create transport and client */ if (clp->cl_cons_state == NFS_CS_INITING) { + clp->rpc_ops = &nfs_v4_clientops; + xprt = xprt_create_proto(proto, &server->addr, timeparms); if (IS_ERR(xprt)) { err = PTR_ERR(xprt); @@ -1139,7 +1148,7 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, /* Bind to a reserved port! */ xprt->resvport = 1; clnt = rpc_create_client(xprt, server->hostname, &nfs_program, - server->rpc_ops->version, flavor); + clp->cl_nfsversion, flavor); if (IS_ERR(clnt)) { err = PTR_ERR(clnt); dprintk("%s: cannot create RPC client. Error = %d\n", @@ -1215,8 +1224,6 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data, server->acdirmin = data->acdirmin*HZ; server->acdirmax = data->acdirmax*HZ; - server->rpc_ops = &nfs_v4_clientops; - nfs_init_timeout_values(&timeparms, data->proto, data->timeo, data->retrans); server->retrans_timeo = timeparms.to_initval; diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 70e1dc9162e2..51e9bd90dedc 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -215,7 +215,7 @@ static inline struct nfs_inode *NFS_I(struct inode *inode) #define NFS_FH(inode) (&NFS_I(inode)->fh) #define NFS_SERVER(inode) (NFS_SB(inode->i_sb)) #define NFS_CLIENT(inode) (NFS_SERVER(inode)->client) -#define NFS_PROTO(inode) (NFS_SERVER(inode)->rpc_ops) +#define NFS_PROTO(inode) (NFS_SERVER(inode)->nfs_client->rpc_ops) #define NFS_ADDR(inode) (RPC_PEERADDR(NFS_CLIENT(inode))) #define NFS_COOKIEVERF(inode) (NFS_I(inode)->cookieverf) #define NFS_READTIME(inode) (NFS_I(inode)->read_cache_jiffies) diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index e7d7662f51fd..aae7c117597a 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -25,6 +25,7 @@ struct nfs_client { struct list_head cl_superblocks; /* List of nfs_server structs */ struct rpc_clnt * cl_rpcclient; + const struct nfs_rpc_ops *rpc_ops; /* NFS protocol vector */ #ifdef CONFIG_NFS_V4 u64 cl_clientid; /* constant */ @@ -74,7 +75,6 @@ struct nfs_server { struct rpc_clnt * client; /* RPC client handle */ struct rpc_clnt * client_sys; /* 2nd handle for FSINFO */ struct rpc_clnt * client_acl; /* ACL RPC client handle */ - const struct nfs_rpc_ops *rpc_ops; /* NFS protocol vector */ struct nfs_iostats * io_stats; /* I/O statistics */ struct backing_dev_info backing_dev_info; int flags; /* various flags */ From 5006a76cca8f86c6975c16fcf67e83b8b0eee2b6 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:12 -0400 Subject: [PATCH 19/74] NFS: Eliminate client_sys in favour of cl_rpcclient Eliminate nfs_server::client_sys in favour of nfs_client::cl_rpcclient as we only really need one per server that we're talking to since it doesn't have any security on it. The retransmission management variables are also moved to the common struct as they're required to set up the cl_rpcclient connection. The NFS2/3 client and client_acl connections are thenceforth derived by cloning the cl_rpcclient connection and post-applying the authorisation flavour. The code for setting up the initial common connection has been moved to client.c as nfs_create_rpc_client(). All the NFS program definition tables are also moved there as that's where they're now required rather than super.c. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 119 ++++++++++++++++++++ fs/nfs/internal.h | 2 + fs/nfs/nfs3proc.c | 6 +- fs/nfs/proc.c | 4 +- fs/nfs/super.c | 222 +++++++++----------------------------- include/linux/nfs_fs_sb.h | 5 +- 6 files changed, 179 insertions(+), 179 deletions(-) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index cb5e92463bdb..c08cab935ad5 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -50,6 +50,48 @@ static DEFINE_SPINLOCK(nfs_client_lock); static LIST_HEAD(nfs_client_list); static DECLARE_WAIT_QUEUE_HEAD(nfs_client_active_wq); +/* + * RPC cruft for NFS + */ +static struct rpc_version *nfs_version[5] = { + [2] = &nfs_version2, +#ifdef CONFIG_NFS_V3 + [3] = &nfs_version3, +#endif +#ifdef CONFIG_NFS_V4 + [4] = &nfs_version4, +#endif +}; + +struct rpc_program nfs_program = { + .name = "nfs", + .number = NFS_PROGRAM, + .nrvers = ARRAY_SIZE(nfs_version), + .version = nfs_version, + .stats = &nfs_rpcstat, + .pipe_dir_name = "/nfs", +}; + +struct rpc_stat nfs_rpcstat = { + .program = &nfs_program +}; + + +#ifdef CONFIG_NFS_V3_ACL +static struct rpc_stat nfsacl_rpcstat = { &nfsacl_program }; +static struct rpc_version * nfsacl_version[] = { + [3] = &nfsacl_version3, +}; + +struct rpc_program nfsacl_program = { + .name = "nfsacl", + .number = NFS_ACL_PROGRAM, + .nrvers = ARRAY_SIZE(nfsacl_version), + .version = nfsacl_version, + .stats = &nfsacl_rpcstat, +}; +#endif /* CONFIG_NFS_V3_ACL */ + /* * Allocate a shared client record * @@ -310,3 +352,80 @@ void nfs_mark_client_ready(struct nfs_client *clp, int state) clp->cl_cons_state = state; wake_up_all(&nfs_client_active_wq); } + +/* + * Initialise the timeout values for a connection + */ +static void nfs_init_timeout_values(struct rpc_timeout *to, int proto, + unsigned int timeo, unsigned int retrans) +{ + to->to_initval = timeo * HZ / 10; + to->to_retries = retrans; + if (!to->to_retries) + to->to_retries = 2; + + switch (proto) { + case IPPROTO_TCP: + if (!to->to_initval) + to->to_initval = 60 * HZ; + if (to->to_initval > NFS_MAX_TCP_TIMEOUT) + to->to_initval = NFS_MAX_TCP_TIMEOUT; + to->to_increment = to->to_initval; + to->to_maxval = to->to_initval + (to->to_increment * to->to_retries); + to->to_exponential = 0; + break; + case IPPROTO_UDP: + default: + if (!to->to_initval) + to->to_initval = 11 * HZ / 10; + if (to->to_initval > NFS_MAX_UDP_TIMEOUT) + to->to_initval = NFS_MAX_UDP_TIMEOUT; + to->to_maxval = NFS_MAX_UDP_TIMEOUT; + to->to_exponential = 1; + break; + } +} + +/* + * Create an RPC client handle + */ +int nfs_create_rpc_client(struct nfs_client *clp, int proto, + unsigned int timeo, + unsigned int retrans, + rpc_authflavor_t flavor) +{ + struct rpc_timeout timeparms; + struct rpc_xprt *xprt = NULL; + struct rpc_clnt *clnt = NULL; + + if (!IS_ERR(clp->cl_rpcclient)) + return 0; + + nfs_init_timeout_values(&timeparms, proto, timeo, retrans); + clp->retrans_timeo = timeparms.to_initval; + clp->retrans_count = timeparms.to_retries; + + /* create transport and client */ + xprt = xprt_create_proto(proto, &clp->cl_addr, &timeparms); + if (IS_ERR(xprt)) { + dprintk("%s: cannot create RPC transport. Error = %ld\n", + __FUNCTION__, PTR_ERR(xprt)); + return PTR_ERR(xprt); + } + + /* Bind to a reserved port! */ + xprt->resvport = 1; + /* Create the client RPC handle */ + clnt = rpc_create_client(xprt, clp->cl_hostname, &nfs_program, + clp->rpc_ops->version, RPC_AUTH_UNIX); + if (IS_ERR(clnt)) { + dprintk("%s: cannot create RPC client. Error = %ld\n", + __FUNCTION__, PTR_ERR(clnt)); + return PTR_ERR(clnt); + } + + clnt->cl_intr = 1; + clnt->cl_softrtry = 1; + clp->cl_rpcclient = clnt; + return 0; +} diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index ac370d5d4494..2f3aa52fbefc 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -20,6 +20,8 @@ extern void nfs_put_client(struct nfs_client *); extern struct nfs_client *nfs_find_client(const struct sockaddr_in *, int); extern struct nfs_client *nfs_get_client(const char *, const struct sockaddr_in *, int); extern void nfs_mark_client_ready(struct nfs_client *, int); +extern int nfs_create_rpc_client(struct nfs_client *, int, unsigned int, + unsigned int, rpc_authflavor_t); /* nfs4namespace.c */ #ifdef CONFIG_NFS_V4 diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c index 3e5371241cea..0622af0122be 100644 --- a/fs/nfs/nfs3proc.c +++ b/fs/nfs/nfs3proc.c @@ -90,8 +90,8 @@ nfs3_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle, int status; status = do_proc_get_root(server->client, fhandle, info); - if (status && server->client_sys != server->client) - status = do_proc_get_root(server->client_sys, fhandle, info); + if (status && server->nfs_client->cl_rpcclient != server->client) + status = do_proc_get_root(server->nfs_client->cl_rpcclient, fhandle, info); return status; } @@ -785,7 +785,7 @@ nfs3_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, dprintk("NFS call fsinfo\n"); nfs_fattr_init(info->fattr); - status = rpc_call_sync(server->client_sys, &msg, 0); + status = rpc_call_sync(server->nfs_client->cl_rpcclient, &msg, 0); dprintk("NFS reply fsinfo: %d\n", status); return status; } diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c index 77676903e0f5..5a8b9407ee9a 100644 --- a/fs/nfs/proc.c +++ b/fs/nfs/proc.c @@ -66,14 +66,14 @@ nfs_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle, dprintk("%s: call getattr\n", __FUNCTION__); nfs_fattr_init(fattr); - status = rpc_call_sync(server->client_sys, &msg, 0); + status = rpc_call_sync(server->nfs_client->cl_rpcclient, &msg, 0); dprintk("%s: reply getattr: %d\n", __FUNCTION__, status); if (status) return status; dprintk("%s: call statfs\n", __FUNCTION__); msg.rpc_proc = &nfs_procedures[NFSPROC_STATFS]; msg.rpc_resp = &fsinfo; - status = rpc_call_sync(server->client_sys, &msg, 0); + status = rpc_call_sync(server->nfs_client->cl_rpcclient, &msg, 0); dprintk("%s: reply statfs: %d\n", __FUNCTION__, status); if (status) return status; diff --git a/fs/nfs/super.c b/fs/nfs/super.c index e1e5eab0259b..85583414a3ca 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -60,52 +60,6 @@ */ #define NFS_MAX_READAHEAD (RPC_DEF_SLOT_TABLE - 1) -/* - * RPC cruft for NFS - */ -static struct rpc_version * nfs_version[] = { - NULL, - NULL, - &nfs_version2, -#if defined(CONFIG_NFS_V3) - &nfs_version3, -#elif defined(CONFIG_NFS_V4) - NULL, -#endif -#if defined(CONFIG_NFS_V4) - &nfs_version4, -#endif -}; - -static struct rpc_program nfs_program = { - .name = "nfs", - .number = NFS_PROGRAM, - .nrvers = ARRAY_SIZE(nfs_version), - .version = nfs_version, - .stats = &nfs_rpcstat, - .pipe_dir_name = "/nfs", -}; - -struct rpc_stat nfs_rpcstat = { - .program = &nfs_program -}; - - -#ifdef CONFIG_NFS_V3_ACL -static struct rpc_stat nfsacl_rpcstat = { &nfsacl_program }; -static struct rpc_version * nfsacl_version[] = { - [3] = &nfsacl_version3, -}; - -struct rpc_program nfsacl_program = { - .name = "nfsacl", - .number = NFS_ACL_PROGRAM, - .nrvers = ARRAY_SIZE(nfsacl_version), - .version = nfsacl_version, - .stats = &nfsacl_rpcstat, -}; -#endif /* CONFIG_NFS_V3_ACL */ - static void nfs_umount_begin(struct vfsmount *, int); static int nfs_statfs(struct dentry *, struct kstatfs *); static int nfs_show_options(struct seq_file *, struct vfsmount *); @@ -376,8 +330,8 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, proto = buf; } seq_printf(m, ",proto=%s", proto); - seq_printf(m, ",timeo=%lu", 10U * nfss->retrans_timeo / HZ); - seq_printf(m, ",retrans=%u", nfss->retrans_count); + seq_printf(m, ",timeo=%lu", 10U * clp->retrans_timeo / HZ); + seq_printf(m, ",retrans=%u", clp->retrans_count); seq_printf(m, ",sec=%s", nfs_pseudoflavour_to_name(nfss->client->cl_auth->au_flavor)); } @@ -621,38 +575,6 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor) return no_root_error; } -/* - * Initialise the timeout values for a connection - */ -static void nfs_init_timeout_values(struct rpc_timeout *to, int proto, unsigned int timeo, unsigned int retrans) -{ - to->to_initval = timeo * HZ / 10; - to->to_retries = retrans; - if (!to->to_retries) - to->to_retries = 2; - - switch (proto) { - case IPPROTO_TCP: - if (!to->to_initval) - to->to_initval = 60 * HZ; - if (to->to_initval > NFS_MAX_TCP_TIMEOUT) - to->to_initval = NFS_MAX_TCP_TIMEOUT; - to->to_increment = to->to_initval; - to->to_maxval = to->to_initval + (to->to_increment * to->to_retries); - to->to_exponential = 0; - break; - case IPPROTO_UDP: - default: - if (!to->to_initval) - to->to_initval = 11 * HZ / 10; - if (to->to_initval > NFS_MAX_UDP_TIMEOUT) - to->to_initval = NFS_MAX_UDP_TIMEOUT; - to->to_maxval = NFS_MAX_UDP_TIMEOUT; - to->to_exponential = 1; - break; - } -} - /* * Create an RPC client handle. */ @@ -660,11 +582,10 @@ static struct rpc_clnt * nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) { struct nfs_client *clp; - struct rpc_timeout timeparms; - struct rpc_xprt *xprt = NULL; - struct rpc_clnt *clnt = NULL; + struct rpc_clnt *clnt; int proto = (data->flags & NFS_MOUNT_TCP) ? IPPROTO_TCP : IPPROTO_UDP; int nfsversion = 2; + int err; #ifdef CONFIG_NFS_V3 if (server->flags & NFS_MOUNT_VER3) @@ -677,52 +598,54 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) return ERR_PTR(PTR_ERR(clp)); } - nfs_init_timeout_values(&timeparms, proto, data->timeo, data->retrans); - - server->retrans_timeo = timeparms.to_initval; - server->retrans_count = timeparms.to_retries; - - /* Check NFS protocol revision and initialize RPC op vector - * and file handle pool. */ + if (clp->cl_cons_state == NFS_CS_INITING) { + /* Check NFS protocol revision and initialize RPC op + * vector and file handle pool. */ #ifdef CONFIG_NFS_V3 - if (nfsversion == 3) { - clp->rpc_ops = &nfs_v3_clientops; - server->caps |= NFS_CAP_READDIRPLUS; - } else { - clp->rpc_ops = &nfs_v2_clientops; - } + if (nfsversion == 3) { + clp->rpc_ops = &nfs_v3_clientops; + server->caps |= NFS_CAP_READDIRPLUS; + } else { + clp->rpc_ops = &nfs_v2_clientops; + } #else - clp->rpc_ops = &nfs_v2_clientops; + clp->rpc_ops = &nfs_v2_clientops; #endif - /* create transport and client */ - xprt = xprt_create_proto(proto, &server->addr, &timeparms); - if (IS_ERR(xprt)) { - dprintk("%s: cannot create RPC transport. Error = %ld\n", - __FUNCTION__, PTR_ERR(xprt)); - nfs_mark_client_ready(clp, PTR_ERR(xprt)); - nfs_put_client(clp); - return (struct rpc_clnt *)xprt; + /* create transport and client */ + err = nfs_create_rpc_client(clp, proto, data->timeo, + data->retrans, RPC_AUTH_UNIX); + if (err < 0) + goto client_init_error; + + nfs_mark_client_ready(clp, 0); } - clnt = rpc_create_client(xprt, server->hostname, &nfs_program, - clp->cl_nfsversion, data->pseudoflavor); + + /* create an nfs_server-specific client */ + clnt = rpc_clone_client(clp->cl_rpcclient); if (IS_ERR(clnt)) { - dprintk("%s: cannot create RPC client. Error = %ld\n", - __FUNCTION__, PTR_ERR(xprt)); - goto out_fail; + dprintk("%s: couldn't create rpc_client!\n", __FUNCTION__); + nfs_put_client(clp); + return ERR_PTR(PTR_ERR(clnt)); } - clnt->cl_intr = 1; - clnt->cl_softrtry = 1; + if (data->pseudoflavor != clp->cl_rpcclient->cl_auth->au_flavor) { + struct rpc_auth *auth; + + auth = rpcauth_create(data->pseudoflavor, server->client); + if (IS_ERR(auth)) { + dprintk("%s: couldn't create credcache!\n", __FUNCTION__); + return ERR_PTR(PTR_ERR(auth)); + } + } - nfs_mark_client_ready(clp, 0); server->nfs_client = clp; return clnt; -out_fail: - nfs_mark_client_ready(clp, PTR_ERR(xprt)); +client_init_error: + nfs_mark_client_ready(clp, err); nfs_put_client(clp); - return clnt; + return ERR_PTR(err); } /* @@ -741,7 +664,7 @@ static struct nfs_server *nfs_clone_server(struct super_block *sb, struct nfs_cl sb->s_blocksize_bits = data->sb->s_blocksize_bits; sb->s_maxbytes = data->sb->s_maxbytes; - server->client_sys = server->client_acl = ERR_PTR(-EINVAL); + server->client_acl = ERR_PTR(-EINVAL); server->io_stats = nfs_alloc_iostats(); if (server->io_stats == NULL) goto out; @@ -750,11 +673,6 @@ static struct nfs_server *nfs_clone_server(struct super_block *sb, struct nfs_cl if (IS_ERR((err = server->client))) goto out; - if (!IS_ERR(parent->client_sys)) { - server->client_sys = rpc_clone_client(parent->client_sys); - if (IS_ERR((err = server->client_sys))) - goto out; - } if (!IS_ERR(parent->client_acl)) { server->client_acl = rpc_clone_client(parent->client_acl); if (IS_ERR((err = server->client_acl))) @@ -813,7 +731,7 @@ static int nfs_clone_generic_sb(struct nfs_clone_mount *data, error = PTR_ERR(sb); goto kill_rpciod; } - + if (sb->s_root) goto out_rpciod_down; @@ -896,19 +814,6 @@ nfs_fill_super(struct super_block *sb, struct nfs_mount_data *data, int silent) return PTR_ERR(server->client); /* RFC 2623, sec 2.3.2 */ - if (authflavor != RPC_AUTH_UNIX) { - struct rpc_auth *auth; - - server->client_sys = rpc_clone_client(server->client); - if (IS_ERR(server->client_sys)) - return PTR_ERR(server->client_sys); - auth = rpcauth_create(RPC_AUTH_UNIX, server->client_sys); - if (IS_ERR(auth)) - return PTR_ERR(auth); - } else { - atomic_inc(&server->client->cl_count); - server->client_sys = server->client; - } if (server->flags & NFS_MOUNT_VER3) { #ifdef CONFIG_NFS_V3_ACL if (!(server->flags & NFS_MOUNT_NOACL)) { @@ -1012,7 +917,7 @@ static int nfs_get_sb(struct file_system_type *fs_type, goto out_err_noserver; /* Zero out the NFS state stuff */ init_nfsv4_state(server); - server->client = server->client_sys = server->client_acl = ERR_PTR(-EINVAL); + server->client = server->client_acl = ERR_PTR(-EINVAL); root = &server->fh; if (data->flags & NFS_MOUNT_VER3) @@ -1083,8 +988,6 @@ static void nfs_kill_super(struct super_block *s) if (!IS_ERR(server->client)) rpc_shutdown_client(server->client); - if (!IS_ERR(server->client_sys)) - rpc_shutdown_client(server->client_sys); if (!IS_ERR(server->client_acl)) rpc_shutdown_client(server->client_acl); @@ -1121,10 +1024,9 @@ static int nfs_clone_nfs_sb(struct file_system_type *fs_type, #ifdef CONFIG_NFS_V4 static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, - struct rpc_timeout *timeparms, int proto, rpc_authflavor_t flavor) + int timeo, int retrans, int proto, rpc_authflavor_t flavor) { struct nfs_client *clp; - struct rpc_xprt *xprt = NULL; struct rpc_clnt *clnt = NULL; int err = -EIO; @@ -1138,26 +1040,10 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, if (clp->cl_cons_state == NFS_CS_INITING) { clp->rpc_ops = &nfs_v4_clientops; - xprt = xprt_create_proto(proto, &server->addr, timeparms); - if (IS_ERR(xprt)) { - err = PTR_ERR(xprt); - dprintk("%s: cannot create RPC transport. Error = %d\n", - __FUNCTION__, err); + err = nfs_create_rpc_client(clp, proto, timeo, retrans, flavor); + if (err < 0) goto client_init_error; - } - /* Bind to a reserved port! */ - xprt->resvport = 1; - clnt = rpc_create_client(xprt, server->hostname, &nfs_program, - clp->cl_nfsversion, flavor); - if (IS_ERR(clnt)) { - err = PTR_ERR(clnt); - dprintk("%s: cannot create RPC client. Error = %d\n", - __FUNCTION__, err); - goto client_init_error; - } - clnt->cl_intr = 1; - clnt->cl_softrtry = 1; - clp->cl_rpcclient = clnt; + memcpy(clp->cl_ipaddr, server->ip_addr, sizeof(clp->cl_ipaddr)); err = nfs_idmap_new(clp); if (err < 0) { @@ -1205,7 +1091,6 @@ static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data, int silent) { struct nfs_server *server; - struct rpc_timeout timeparms; rpc_authflavor_t authflavour; int err = -EIO; @@ -1224,11 +1109,6 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data, server->acdirmin = data->acdirmin*HZ; server->acdirmax = data->acdirmax*HZ; - nfs_init_timeout_values(&timeparms, data->proto, data->timeo, data->retrans); - - server->retrans_timeo = timeparms.to_initval; - server->retrans_count = timeparms.to_retries; - /* Now create transport and client */ authflavour = RPC_AUTH_UNIX; if (data->auth_flavourlen != 0) { @@ -1244,7 +1124,8 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data, } } - server->client = nfs4_create_client(server, &timeparms, data->proto, authflavour); + server->client = nfs4_create_client(server, data->timeo, data->retrans, + data->proto, authflavour); if (IS_ERR(server->client)) { err = PTR_ERR(server->client); dprintk("%s: cannot create RPC client. Error = %d\n", @@ -1318,7 +1199,7 @@ static int nfs4_get_sb(struct file_system_type *fs_type, return -ENOMEM; /* Zero out the NFS state stuff */ init_nfsv4_state(server); - server->client = server->client_sys = server->client_acl = ERR_PTR(-EINVAL); + server->client = server->client_acl = ERR_PTR(-EINVAL); p = nfs_copy_user_string(NULL, &data->hostname, 256); if (IS_ERR(p)) @@ -1489,7 +1370,6 @@ static struct super_block *nfs4_referral_sb(struct nfs_server *server, struct nf static struct nfs_server *nfs4_referral_server(struct super_block *sb, struct nfs_clone_mount *data) { struct nfs_server *server = NFS_SB(sb); - struct rpc_timeout timeparms; int proto, timeo, retrans; void *err; @@ -1498,11 +1378,11 @@ static struct nfs_server *nfs4_referral_server(struct super_block *sb, struct nf set the timeouts and retries to low values */ timeo = 2; retrans = 1; - nfs_init_timeout_values(&timeparms, proto, timeo, retrans); nfs_put_client(server->nfs_client); server->nfs_client = NULL; - server->client = nfs4_create_client(server, &timeparms, proto, data->authflavor); + server->client = nfs4_create_client(server, timeo, retrans, proto, + data->authflavor); if (IS_ERR((err = server->client))) goto out_err; diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index aae7c117597a..d404ceca9168 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -26,6 +26,8 @@ struct nfs_client { struct rpc_clnt * cl_rpcclient; const struct nfs_rpc_ops *rpc_ops; /* NFS protocol vector */ + unsigned long retrans_timeo; /* retransmit timeout */ + unsigned int retrans_count; /* number of retransmit tries */ #ifdef CONFIG_NFS_V4 u64 cl_clientid; /* constant */ @@ -73,7 +75,6 @@ struct nfs_client { struct nfs_server { struct nfs_client * nfs_client; /* shared client and NFS4 state */ struct rpc_clnt * client; /* RPC client handle */ - struct rpc_clnt * client_sys; /* 2nd handle for FSINFO */ struct rpc_clnt * client_acl; /* ACL RPC client handle */ struct nfs_iostats * io_stats; /* I/O statistics */ struct backing_dev_info backing_dev_info; @@ -90,8 +91,6 @@ struct nfs_server { unsigned int acregmax; unsigned int acdirmin; unsigned int acdirmax; - unsigned long retrans_timeo; /* retransmit timeout */ - unsigned int retrans_count; /* number of retransmit tries */ unsigned int namelen; char * hostname; /* remote hostname */ struct nfs_fh fh; From cf6d7b5de8535a9f0088c5cc28ee2dae87371b4a Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:12 -0400 Subject: [PATCH 20/74] NFS: Start rpciod in server common management Start rpciod in the server common (nfs_client struct) management code rather than in the superblock management code. This means we only need to "start" it once per server instead of once per superblock. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 31 ++++++------------------------- 1 file changed, 6 insertions(+), 25 deletions(-) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 85583414a3ca..5842d510d732 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -722,18 +722,15 @@ static int nfs_clone_generic_sb(struct nfs_clone_mount *data, if (server->hostname == NULL) goto free_server; memcpy(server->hostname, hostname, len); - error = rpciod_up(); - if (error != 0) - goto free_hostname; sb = fill_sb(server, data); if (IS_ERR(sb)) { error = PTR_ERR(sb); - goto kill_rpciod; + goto free_hostname; } if (sb->s_root) - goto out_rpciod_down; + goto out_share; server = fill_server(sb, data); if (IS_ERR(server)) { @@ -745,14 +742,11 @@ static int nfs_clone_generic_sb(struct nfs_clone_mount *data, up_write(&sb->s_umount); deactivate_super(sb); return error; -out_rpciod_down: - rpciod_down(); +out_share: kfree(server->hostname); nfs_put_client(server->nfs_client); kfree(server); return simple_set_mnt(mnt, sb); -kill_rpciod: - rpciod_down(); free_hostname: kfree(server->hostname); free_server: @@ -939,22 +933,14 @@ static int nfs_get_sb(struct file_system_type *fs_type, goto out_err; } - /* Fire up rpciod if not yet running */ - error = rpciod_up(); - if (error < 0) { - dprintk("%s: couldn't start rpciod! Error = %d\n", - __FUNCTION__, error); - goto out_err; - } - s = sget(fs_type, nfs_compare_super, nfs_set_super, server); if (IS_ERR(s)) { error = PTR_ERR(s); - goto out_err_rpciod; + goto out_err; } if (s->s_root) - goto out_rpciod_down; + goto out_share; s->s_flags = flags; @@ -967,13 +953,10 @@ static int nfs_get_sb(struct file_system_type *fs_type, s->s_flags |= MS_ACTIVE; return simple_set_mnt(mnt, s); -out_rpciod_down: - rpciod_down(); +out_share: kfree(server); return simple_set_mnt(mnt, s); -out_err_rpciod: - rpciod_down(); out_err: kfree(server); out_err_noserver: @@ -994,8 +977,6 @@ static void nfs_kill_super(struct super_block *s) if (!(server->flags & NFS_MOUNT_NONLM)) lockd_down(); /* release rpc.lockd */ - rpciod_down(); /* release rpciod */ - nfs_free_iostats(server->io_stats); kfree(server->hostname); nfs_put_client(server->nfs_client); From 54ceac4515986030c2502960be620198dd8fe25b Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:13 -0400 Subject: [PATCH 21/74] NFS: Share NFS superblocks per-protocol per-server per-FSID The attached patch makes NFS share superblocks between mounts from the same server and FSID over the same protocol. It does this by creating each superblock with a false root and returning the real root dentry in the vfsmount presented by get_sb(). The root dentry set starts off as an anonymous dentry if we don't already have the dentry for its inode, otherwise it simply returns the dentry we already have. We may thus end up with several trees of dentries in the superblock, and if at some later point one of anonymous tree roots is discovered by normal filesystem activity to be located in another tree within the superblock, the anonymous root is named and materialises attached to the second tree at the appropriate point. Why do it this way? Why not pass an extra argument to the mount() syscall to indicate the subpath and then pathwalk from the server root to the desired directory? You can't guarantee this will work for two reasons: (1) The root and intervening nodes may not be accessible to the client. With NFS2 and NFS3, for instance, mountd is called on the server to get the filehandle for the tip of a path. mountd won't give us handles for anything we don't have permission to access, and so we can't set up NFS inodes for such nodes, and so can't easily set up dentries (we'd have to have ghost inodes or something). With this patch we don't actually create dentries until we get handles from the server that we can use to set up their inodes, and we don't actually bind them into the tree until we know for sure where they go. (2) Inaccessible symbolic links. If we're asked to mount two exports from the server, eg: mount warthog:/warthog/aaa/xxx /mmm mount warthog:/warthog/bbb/yyy /nnn We may not be able to access anything nearer the root than xxx and yyy, but we may find out later that /mmm/www/yyy, say, is actually the same directory as the one mounted on /nnn. What we might then find out, for example, is that /warthog/bbb was actually a symbolic link to /warthog/aaa/xxx/www, but we can't actually determine that by talking to the server until /warthog is made available by NFS. This would lead to having constructed an errneous dentry tree which we can't easily fix. We can end up with a dentry marked as a directory when it should actually be a symlink, or we could end up with an apparently hardlinked directory. With this patch we need not make assumptions about the type of a dentry for which we can't retrieve information, nor need we assume we know its place in the grand scheme of things until we actually see that place. This patch reduces the possibility of aliasing in the inode and page caches for inodes that may be accessed by more than one NFS export. It also reduces the number of superblocks required for NFS where there are many NFS exports being used from a server (home directory server + autofs for example). This in turn makes it simpler to do local caching of network filesystems, as it can then be guaranteed that there won't be links from multiple inodes in separate superblocks to the same cache file. Obviously, cache aliasing between different levels of NFS protocol could still be a problem, but at least that gives us another key to use when indexing the cache. This patch makes the following changes: (1) The server record construction/destruction has been abstracted out into its own set of functions to make things easier to get right. These have been moved into fs/nfs/client.c. All the code in fs/nfs/client.c has to do with the management of connections to servers, and doesn't touch superblocks in any way; the remaining code in fs/nfs/super.c has to do with VFS superblock management. (2) The sequence of events undertaken by NFS mount is now reordered: (a) A volume representation (struct nfs_server) is allocated. (b) A server representation (struct nfs_client) is acquired. This may be allocated or shared, and is keyed on server address, port and NFS version. (c) If allocated, the client representation is initialised. The state member variable of nfs_client is used to prevent a race during initialisation from two mounts. (d) For NFS4 a simple pathwalk is performed, walking from FH to FH to find the root filehandle for the mount (fs/nfs/getroot.c). For NFS2/3 we are given the root FH in advance. (e) The volume FSID is probed for on the root FH. (f) The volume representation is initialised from the FSINFO record retrieved on the root FH. (g) sget() is called to acquire a superblock. This may be allocated or shared, keyed on client pointer and FSID. (h) If allocated, the superblock is initialised. (i) If the superblock is shared, then the new nfs_server record is discarded. (j) The root dentry for this mount is looked up from the root FH. (k) The root dentry for this mount is assigned to the vfsmount. (3) nfs_readdir_lookup() creates dentries for each of the entries readdir() returns; this function now attaches disconnected trees from alternate roots that happen to be discovered attached to a directory being read (in the same way nfs_lookup() is made to do for lookup ops). The new d_materialise_unique() function is now used to do this, thus permitting the whole thing to be done under one set of locks, and thus avoiding any race between mount and lookup operations on the same directory. (4) The client management code uses a new debug facility: NFSDBG_CLIENT which is set by echoing 1024 to /proc/net/sunrpc/nfs_debug. (5) Clone mounts are now called xdev mounts. (6) Use the dentry passed to the statfs() op as the handle for retrieving fs statistics rather than the root dentry of the superblock (which is now a dummy). Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/Makefile | 2 +- fs/nfs/client.c | 735 ++++++++++++++++++++- fs/nfs/dir.c | 16 +- fs/nfs/getroot.c | 306 +++++++++ fs/nfs/idmap.c | 3 +- fs/nfs/inode.c | 2 +- fs/nfs/internal.h | 82 ++- fs/nfs/namespace.c | 25 +- fs/nfs/nfs3proc.c | 2 +- fs/nfs/nfs4_fs.h | 6 - fs/nfs/nfs4namespace.c | 110 +++- fs/nfs/nfs4proc.c | 59 +- fs/nfs/nfs4renewd.c | 13 - fs/nfs/nfs4state.c | 18 - fs/nfs/read.c | 2 +- fs/nfs/super.c | 1297 ++++++++++++++----------------------- fs/nfs/write.c | 2 +- include/linux/nfs_fs_sb.h | 21 +- 18 files changed, 1700 insertions(+), 1001 deletions(-) create mode 100644 fs/nfs/getroot.c diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile index 3b993a6f8163..f4580b44eef4 100644 --- a/fs/nfs/Makefile +++ b/fs/nfs/Makefile @@ -4,7 +4,7 @@ obj-$(CONFIG_NFS_FS) += nfs.o -nfs-y := client.o dir.o file.o inode.o super.o nfs2xdr.o \ +nfs-y := client.o dir.o file.o getroot.o inode.o super.o nfs2xdr.o \ pagelist.o proc.o read.o symlink.o unlink.o \ write.o namespace.o nfs-$(CONFIG_ROOT_NFS) += nfsroot.o mount_clnt.o diff --git a/fs/nfs/client.c b/fs/nfs/client.c index c08cab935ad5..dafba608c0a0 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -48,6 +48,7 @@ static DEFINE_SPINLOCK(nfs_client_lock); static LIST_HEAD(nfs_client_list); +static LIST_HEAD(nfs_volume_list); static DECLARE_WAIT_QUEUE_HEAD(nfs_client_active_wq); /* @@ -268,9 +269,9 @@ struct nfs_client *nfs_find_client(const struct sockaddr_in *addr, int nfsversio * Look up a client by IP address and protocol version * - creates a new record if one doesn't yet exist */ -struct nfs_client *nfs_get_client(const char *hostname, - const struct sockaddr_in *addr, - int nfsversion) +static struct nfs_client *nfs_get_client(const char *hostname, + const struct sockaddr_in *addr, + int nfsversion) { struct nfs_client *clp, *new = NULL; int error; @@ -340,6 +341,8 @@ struct nfs_client *nfs_get_client(const char *hostname, return ERR_PTR(error); } + BUG_ON(clp->cl_cons_state != NFS_CS_READY); + dprintk("--> nfs_get_client() = %p [share]\n", clp); return clp; } @@ -347,7 +350,7 @@ struct nfs_client *nfs_get_client(const char *hostname, /* * Mark a server as ready or failed */ -void nfs_mark_client_ready(struct nfs_client *clp, int state) +static void nfs_mark_client_ready(struct nfs_client *clp, int state) { clp->cl_cons_state = state; wake_up_all(&nfs_client_active_wq); @@ -389,10 +392,10 @@ static void nfs_init_timeout_values(struct rpc_timeout *to, int proto, /* * Create an RPC client handle */ -int nfs_create_rpc_client(struct nfs_client *clp, int proto, - unsigned int timeo, - unsigned int retrans, - rpc_authflavor_t flavor) +static int nfs_create_rpc_client(struct nfs_client *clp, int proto, + unsigned int timeo, + unsigned int retrans, + rpc_authflavor_t flavor) { struct rpc_timeout timeparms; struct rpc_xprt *xprt = NULL; @@ -429,3 +432,719 @@ int nfs_create_rpc_client(struct nfs_client *clp, int proto, clp->cl_rpcclient = clnt; return 0; } + +/* + * Version 2 or 3 client destruction + */ +static void nfs_destroy_server(struct nfs_server *server) +{ + if (!IS_ERR(server->client_acl)) + rpc_shutdown_client(server->client_acl); + + if (!(server->flags & NFS_MOUNT_NONLM)) + lockd_down(); /* release rpc.lockd */ +} + +/* + * Version 2 or 3 lockd setup + */ +static int nfs_start_lockd(struct nfs_server *server) +{ + int error = 0; + + if (server->nfs_client->cl_nfsversion > 3) + goto out; + if (server->flags & NFS_MOUNT_NONLM) + goto out; + error = lockd_up(); + if (error < 0) + server->flags |= NFS_MOUNT_NONLM; + else + server->destroy = nfs_destroy_server; +out: + return error; +} + +/* + * Initialise an NFSv3 ACL client connection + */ +#ifdef CONFIG_NFS_V3_ACL +static void nfs_init_server_aclclient(struct nfs_server *server) +{ + if (server->nfs_client->cl_nfsversion != 3) + goto out_noacl; + if (server->flags & NFS_MOUNT_NOACL) + goto out_noacl; + + server->client_acl = rpc_bind_new_program(server->client, &nfsacl_program, 3); + if (IS_ERR(server->client_acl)) + goto out_noacl; + + /* No errors! Assume that Sun nfsacls are supported */ + server->caps |= NFS_CAP_ACLS; + return; + +out_noacl: + server->caps &= ~NFS_CAP_ACLS; +} +#else +static inline void nfs_init_server_aclclient(struct nfs_server *server) +{ + server->flags &= ~NFS_MOUNT_NOACL; + server->caps &= ~NFS_CAP_ACLS; +} +#endif + +/* + * Create a general RPC client + */ +static int nfs_init_server_rpcclient(struct nfs_server *server, rpc_authflavor_t pseudoflavour) +{ + struct nfs_client *clp = server->nfs_client; + + server->client = rpc_clone_client(clp->cl_rpcclient); + if (IS_ERR(server->client)) { + dprintk("%s: couldn't create rpc_client!\n", __FUNCTION__); + return PTR_ERR(server->client); + } + + if (pseudoflavour != clp->cl_rpcclient->cl_auth->au_flavor) { + struct rpc_auth *auth; + + auth = rpcauth_create(pseudoflavour, server->client); + if (IS_ERR(auth)) { + dprintk("%s: couldn't create credcache!\n", __FUNCTION__); + return PTR_ERR(auth); + } + } + server->client->cl_softrtry = 0; + if (server->flags & NFS_MOUNT_SOFT) + server->client->cl_softrtry = 1; + + server->client->cl_intr = 0; + if (server->flags & NFS4_MOUNT_INTR) + server->client->cl_intr = 1; + + return 0; +} + +/* + * Initialise an NFS2 or NFS3 client + */ +static int nfs_init_client(struct nfs_client *clp, const struct nfs_mount_data *data) +{ + int proto = (data->flags & NFS_MOUNT_TCP) ? IPPROTO_TCP : IPPROTO_UDP; + int error; + + if (clp->cl_cons_state == NFS_CS_READY) { + /* the client is already initialised */ + dprintk("<-- nfs_init_client() = 0 [already %p]\n", clp); + return 0; + } + + /* Check NFS protocol revision and initialize RPC op vector */ + clp->rpc_ops = &nfs_v2_clientops; +#ifdef CONFIG_NFS_V3 + if (clp->cl_nfsversion == 3) + clp->rpc_ops = &nfs_v3_clientops; +#endif + /* + * Create a client RPC handle for doing FSSTAT with UNIX auth only + * - RFC 2623, sec 2.3.2 + */ + error = nfs_create_rpc_client(clp, proto, data->timeo, data->retrans, + RPC_AUTH_UNIX); + if (error < 0) + goto error; + nfs_mark_client_ready(clp, NFS_CS_READY); + return 0; + +error: + nfs_mark_client_ready(clp, error); + dprintk("<-- nfs_init_client() = xerror %d\n", error); + return error; +} + +/* + * Create a version 2 or 3 client + */ +static int nfs_init_server(struct nfs_server *server, const struct nfs_mount_data *data) +{ + struct nfs_client *clp; + int error, nfsvers = 2; + + dprintk("--> nfs_init_server()\n"); + +#ifdef CONFIG_NFS_V3 + if (data->flags & NFS_MOUNT_VER3) + nfsvers = 3; +#endif + + /* Allocate or find a client reference we can use */ + clp = nfs_get_client(data->hostname, &data->addr, nfsvers); + if (IS_ERR(clp)) { + dprintk("<-- nfs_init_server() = error %ld\n", PTR_ERR(clp)); + return PTR_ERR(clp); + } + + error = nfs_init_client(clp, data); + if (error < 0) + goto error; + + server->nfs_client = clp; + + /* Initialise the client representation from the mount data */ + server->flags = data->flags & NFS_MOUNT_FLAGMASK; + + if (data->rsize) + server->rsize = nfs_block_size(data->rsize, NULL); + if (data->wsize) + server->wsize = nfs_block_size(data->wsize, NULL); + + server->acregmin = data->acregmin * HZ; + server->acregmax = data->acregmax * HZ; + server->acdirmin = data->acdirmin * HZ; + server->acdirmax = data->acdirmax * HZ; + + /* Start lockd here, before we might error out */ + error = nfs_start_lockd(server); + if (error < 0) + goto error; + + error = nfs_init_server_rpcclient(server, data->pseudoflavor); + if (error < 0) + goto error; + + server->namelen = data->namlen; + /* Create a client RPC handle for the NFSv3 ACL management interface */ + nfs_init_server_aclclient(server); + if (clp->cl_nfsversion == 3) { + if (server->namelen == 0 || server->namelen > NFS3_MAXNAMLEN) + server->namelen = NFS3_MAXNAMLEN; + server->caps |= NFS_CAP_READDIRPLUS; + } else { + if (server->namelen == 0 || server->namelen > NFS2_MAXNAMLEN) + server->namelen = NFS2_MAXNAMLEN; + } + + dprintk("<-- nfs_init_server() = 0 [new %p]\n", clp); + return 0; + +error: + server->nfs_client = NULL; + nfs_put_client(clp); + dprintk("<-- nfs_init_server() = xerror %d\n", error); + return error; +} + +/* + * Load up the server record from information gained in an fsinfo record + */ +static void nfs_server_set_fsinfo(struct nfs_server *server, struct nfs_fsinfo *fsinfo) +{ + unsigned long max_rpc_payload; + + /* Work out a lot of parameters */ + if (server->rsize == 0) + server->rsize = nfs_block_size(fsinfo->rtpref, NULL); + if (server->wsize == 0) + server->wsize = nfs_block_size(fsinfo->wtpref, NULL); + + if (fsinfo->rtmax >= 512 && server->rsize > fsinfo->rtmax) + server->rsize = nfs_block_size(fsinfo->rtmax, NULL); + if (fsinfo->wtmax >= 512 && server->wsize > fsinfo->wtmax) + server->wsize = nfs_block_size(fsinfo->wtmax, NULL); + + max_rpc_payload = nfs_block_size(rpc_max_payload(server->client), NULL); + if (server->rsize > max_rpc_payload) + server->rsize = max_rpc_payload; + if (server->rsize > NFS_MAX_FILE_IO_SIZE) + server->rsize = NFS_MAX_FILE_IO_SIZE; + server->rpages = (server->rsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + server->backing_dev_info.ra_pages = server->rpages * NFS_MAX_READAHEAD; + + if (server->wsize > max_rpc_payload) + server->wsize = max_rpc_payload; + if (server->wsize > NFS_MAX_FILE_IO_SIZE) + server->wsize = NFS_MAX_FILE_IO_SIZE; + server->wpages = (server->wsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + server->wtmult = nfs_block_bits(fsinfo->wtmult, NULL); + + server->dtsize = nfs_block_size(fsinfo->dtpref, NULL); + if (server->dtsize > PAGE_CACHE_SIZE) + server->dtsize = PAGE_CACHE_SIZE; + if (server->dtsize > server->rsize) + server->dtsize = server->rsize; + + if (server->flags & NFS_MOUNT_NOAC) { + server->acregmin = server->acregmax = 0; + server->acdirmin = server->acdirmax = 0; + } + + server->maxfilesize = fsinfo->maxfilesize; + + /* We're airborne Set socket buffersize */ + rpc_setbufsize(server->client, server->wsize + 100, server->rsize + 100); +} + +/* + * Probe filesystem information, including the FSID on v2/v3 + */ +static int nfs_probe_fsinfo(struct nfs_server *server, struct nfs_fh *mntfh, struct nfs_fattr *fattr) +{ + struct nfs_fsinfo fsinfo; + struct nfs_client *clp = server->nfs_client; + int error; + + dprintk("--> nfs_probe_fsinfo()\n"); + + if (clp->rpc_ops->set_capabilities != NULL) { + error = clp->rpc_ops->set_capabilities(server, mntfh); + if (error < 0) + goto out_error; + } + + fsinfo.fattr = fattr; + nfs_fattr_init(fattr); + error = clp->rpc_ops->fsinfo(server, mntfh, &fsinfo); + if (error < 0) + goto out_error; + + nfs_server_set_fsinfo(server, &fsinfo); + + /* Get some general file system info */ + if (server->namelen == 0) { + struct nfs_pathconf pathinfo; + + pathinfo.fattr = fattr; + nfs_fattr_init(fattr); + + if (clp->rpc_ops->pathconf(server, mntfh, &pathinfo) >= 0) + server->namelen = pathinfo.max_namelen; + } + + dprintk("<-- nfs_probe_fsinfo() = 0\n"); + return 0; + +out_error: + dprintk("nfs_probe_fsinfo: error = %d\n", -error); + return error; +} + +/* + * Copy useful information when duplicating a server record + */ +static void nfs_server_copy_userdata(struct nfs_server *target, struct nfs_server *source) +{ + target->flags = source->flags; + target->acregmin = source->acregmin; + target->acregmax = source->acregmax; + target->acdirmin = source->acdirmin; + target->acdirmax = source->acdirmax; + target->caps = source->caps; +} + +/* + * Allocate and initialise a server record + */ +static struct nfs_server *nfs_alloc_server(void) +{ + struct nfs_server *server; + + server = kzalloc(sizeof(struct nfs_server), GFP_KERNEL); + if (!server) + return NULL; + + server->client = server->client_acl = ERR_PTR(-EINVAL); + + /* Zero out the NFS state stuff */ + INIT_LIST_HEAD(&server->client_link); + INIT_LIST_HEAD(&server->master_link); + + server->io_stats = nfs_alloc_iostats(); + if (!server->io_stats) { + kfree(server); + return NULL; + } + + return server; +} + +/* + * Free up a server record + */ +void nfs_free_server(struct nfs_server *server) +{ + dprintk("--> nfs_free_server()\n"); + + spin_lock(&nfs_client_lock); + list_del(&server->client_link); + list_del(&server->master_link); + spin_unlock(&nfs_client_lock); + + if (server->destroy != NULL) + server->destroy(server); + if (!IS_ERR(server->client)) + rpc_shutdown_client(server->client); + + nfs_put_client(server->nfs_client); + + nfs_free_iostats(server->io_stats); + kfree(server); + nfs_release_automount_timer(); + dprintk("<-- nfs_free_server()\n"); +} + +/* + * Create a version 2 or 3 volume record + * - keyed on server and FSID + */ +struct nfs_server *nfs_create_server(const struct nfs_mount_data *data, + struct nfs_fh *mntfh) +{ + struct nfs_server *server; + struct nfs_fattr fattr; + int error; + + server = nfs_alloc_server(); + if (!server) + return ERR_PTR(-ENOMEM); + + /* Get a client representation */ + error = nfs_init_server(server, data); + if (error < 0) + goto error; + + BUG_ON(!server->nfs_client); + BUG_ON(!server->nfs_client->rpc_ops); + BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops); + + /* Probe the root fh to retrieve its FSID */ + error = nfs_probe_fsinfo(server, mntfh, &fattr); + if (error < 0) + goto error; + if (!(fattr.valid & NFS_ATTR_FATTR)) { + error = server->nfs_client->rpc_ops->getattr(server, mntfh, &fattr); + if (error < 0) { + dprintk("nfs_create_server: getattr error = %d\n", -error); + goto error; + } + } + memcpy(&server->fsid, &fattr.fsid, sizeof(server->fsid)); + + dprintk("Server FSID: %llx:%llx\n", server->fsid.major, server->fsid.minor); + + BUG_ON(!server->nfs_client); + BUG_ON(!server->nfs_client->rpc_ops); + BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops); + + spin_lock(&nfs_client_lock); + list_add_tail(&server->client_link, &server->nfs_client->cl_superblocks); + list_add_tail(&server->master_link, &nfs_volume_list); + spin_unlock(&nfs_client_lock); + + server->mount_time = jiffies; + return server; + +error: + nfs_free_server(server); + return ERR_PTR(error); +} + +#ifdef CONFIG_NFS_V4 +/* + * Initialise an NFS4 client record + */ +static int nfs4_init_client(struct nfs_client *clp, + int proto, int timeo, int retrans, + rpc_authflavor_t authflavour) +{ + int error; + + if (clp->cl_cons_state == NFS_CS_READY) { + /* the client is initialised already */ + dprintk("<-- nfs4_init_client() = 0 [already %p]\n", clp); + return 0; + } + + /* Check NFS protocol revision and initialize RPC op vector */ + clp->rpc_ops = &nfs_v4_clientops; + + error = nfs_create_rpc_client(clp, proto, timeo, retrans, authflavour); + if (error < 0) + goto error; + + error = nfs_idmap_new(clp); + if (error < 0) { + dprintk("%s: failed to create idmapper. Error = %d\n", + __FUNCTION__, error); + __set_bit(NFS_CS_IDMAP, &clp->cl_res_state); + goto error; + } + + nfs_mark_client_ready(clp, NFS_CS_READY); + return 0; + +error: + nfs_mark_client_ready(clp, error); + dprintk("<-- nfs4_init_client() = xerror %d\n", error); + return error; +} + +/* + * Set up an NFS4 client + */ +static int nfs4_set_client(struct nfs_server *server, + const char *hostname, const struct sockaddr_in *addr, + rpc_authflavor_t authflavour, + int proto, int timeo, int retrans) +{ + struct nfs_client *clp; + int error; + + dprintk("--> nfs4_set_client()\n"); + + /* Allocate or find a client reference we can use */ + clp = nfs_get_client(hostname, addr, 4); + if (IS_ERR(clp)) { + error = PTR_ERR(clp); + goto error; + } + error = nfs4_init_client(clp, proto, timeo, retrans, authflavour); + if (error < 0) + goto error_put; + + server->nfs_client = clp; + dprintk("<-- nfs4_set_client() = 0 [new %p]\n", clp); + return 0; + +error_put: + nfs_put_client(clp); +error: + dprintk("<-- nfs4_set_client() = xerror %d\n", error); + return error; +} + +/* + * Create a version 4 volume record + */ +static int nfs4_init_server(struct nfs_server *server, + const struct nfs4_mount_data *data, rpc_authflavor_t authflavour) +{ + int error; + + dprintk("--> nfs4_init_server()\n"); + + /* Initialise the client representation from the mount data */ + server->flags = data->flags & NFS_MOUNT_FLAGMASK; + server->caps |= NFS_CAP_ATOMIC_OPEN; + + if (data->rsize) + server->rsize = nfs_block_size(data->rsize, NULL); + if (data->wsize) + server->wsize = nfs_block_size(data->wsize, NULL); + + server->acregmin = data->acregmin * HZ; + server->acregmax = data->acregmax * HZ; + server->acdirmin = data->acdirmin * HZ; + server->acdirmax = data->acdirmax * HZ; + + error = nfs_init_server_rpcclient(server, authflavour); + + /* Done */ + dprintk("<-- nfs4_init_server() = %d\n", error); + return error; +} + +/* + * Create a version 4 volume record + * - keyed on server and FSID + */ +struct nfs_server *nfs4_create_server(const struct nfs4_mount_data *data, + const char *hostname, + const struct sockaddr_in *addr, + const char *mntpath, + const char *ip_addr, + rpc_authflavor_t authflavour, + struct nfs_fh *mntfh) +{ + struct nfs_fattr fattr; + struct nfs_server *server; + int error; + + dprintk("--> nfs4_create_server()\n"); + + server = nfs_alloc_server(); + if (!server) + return ERR_PTR(-ENOMEM); + + /* Get a client record */ + error = nfs4_set_client(server, hostname, addr, authflavour, + data->proto, data->timeo, data->retrans); + if (error < 0) + goto error; + + /* set up the general RPC client */ + error = nfs4_init_server(server, data, authflavour); + if (error < 0) + goto error; + + BUG_ON(!server->nfs_client); + BUG_ON(!server->nfs_client->rpc_ops); + BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops); + + /* Probe the root fh to retrieve its FSID */ + error = nfs4_path_walk(server, mntfh, mntpath); + if (error < 0) + goto error; + + dprintk("Server FSID: %llx:%llx\n", server->fsid.major, server->fsid.minor); + dprintk("Mount FH: %d\n", mntfh->size); + + error = nfs_probe_fsinfo(server, mntfh, &fattr); + if (error < 0) + goto error; + + BUG_ON(!server->nfs_client); + BUG_ON(!server->nfs_client->rpc_ops); + BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops); + + spin_lock(&nfs_client_lock); + list_add_tail(&server->client_link, &server->nfs_client->cl_superblocks); + list_add_tail(&server->master_link, &nfs_volume_list); + spin_unlock(&nfs_client_lock); + + server->mount_time = jiffies; + dprintk("<-- nfs4_create_server() = %p\n", server); + return server; + +error: + nfs_free_server(server); + dprintk("<-- nfs4_create_server() = error %d\n", error); + return ERR_PTR(error); +} + +/* + * Create an NFS4 referral server record + */ +struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data, + struct nfs_fh *fh) +{ + struct nfs_client *parent_client; + struct nfs_server *server, *parent_server; + struct nfs_fattr fattr; + int error; + + dprintk("--> nfs4_create_referral_server()\n"); + + server = nfs_alloc_server(); + if (!server) + return ERR_PTR(-ENOMEM); + + parent_server = NFS_SB(data->sb); + parent_client = parent_server->nfs_client; + + /* Get a client representation. + * Note: NFSv4 always uses TCP, */ + error = nfs4_set_client(server, data->hostname, data->addr, + data->authflavor, + parent_server->client->cl_xprt->prot, + parent_client->retrans_timeo, + parent_client->retrans_count); + + /* Initialise the client representation from the parent server */ + nfs_server_copy_userdata(server, parent_server); + server->caps |= NFS_CAP_ATOMIC_OPEN; + + error = nfs_init_server_rpcclient(server, data->authflavor); + if (error < 0) + goto error; + + BUG_ON(!server->nfs_client); + BUG_ON(!server->nfs_client->rpc_ops); + BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops); + + /* probe the filesystem info for this server filesystem */ + error = nfs_probe_fsinfo(server, fh, &fattr); + if (error < 0) + goto error; + + dprintk("Referral FSID: %llx:%llx\n", + server->fsid.major, server->fsid.minor); + + spin_lock(&nfs_client_lock); + list_add_tail(&server->client_link, &server->nfs_client->cl_superblocks); + list_add_tail(&server->master_link, &nfs_volume_list); + spin_unlock(&nfs_client_lock); + + server->mount_time = jiffies; + + dprintk("<-- nfs_create_referral_server() = %p\n", server); + return server; + +error: + nfs_free_server(server); + dprintk("<-- nfs4_create_referral_server() = error %d\n", error); + return ERR_PTR(error); +} + +#endif /* CONFIG_NFS_V4 */ + +/* + * Clone an NFS2, NFS3 or NFS4 server record + */ +struct nfs_server *nfs_clone_server(struct nfs_server *source, + struct nfs_fh *fh, + struct nfs_fattr *fattr) +{ + struct nfs_server *server; + struct nfs_fattr fattr_fsinfo; + int error; + + dprintk("--> nfs_clone_server(,%llx:%llx,)\n", + fattr->fsid.major, fattr->fsid.minor); + + server = nfs_alloc_server(); + if (!server) + return ERR_PTR(-ENOMEM); + + /* Copy data from the source */ + server->nfs_client = source->nfs_client; + atomic_inc(&server->nfs_client->cl_count); + nfs_server_copy_userdata(server, source); + + server->fsid = fattr->fsid; + + error = nfs_init_server_rpcclient(server, source->client->cl_auth->au_flavor); + if (error < 0) + goto out_free_server; + if (!IS_ERR(source->client_acl)) + nfs_init_server_aclclient(server); + + /* probe the filesystem info for this server filesystem */ + error = nfs_probe_fsinfo(server, fh, &fattr_fsinfo); + if (error < 0) + goto out_free_server; + + dprintk("Cloned FSID: %llx:%llx\n", + server->fsid.major, server->fsid.minor); + + error = nfs_start_lockd(server); + if (error < 0) + goto out_free_server; + + spin_lock(&nfs_client_lock); + list_add_tail(&server->client_link, &server->nfs_client->cl_superblocks); + list_add_tail(&server->master_link, &nfs_volume_list); + spin_unlock(&nfs_client_lock); + + server->mount_time = jiffies; + + dprintk("<-- nfs_clone_server() = %p\n", server); + return server; + +out_free_server: + nfs_free_server(server); + dprintk("<-- nfs_clone_server() = error %d\n", error); + return ERR_PTR(error); +} diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 19362712452f..9b496ef4abea 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -31,6 +31,7 @@ #include #include #include +#include #include "nfs4_fs.h" #include "delegation.h" @@ -870,14 +871,14 @@ int nfs_is_exclusive_create(struct inode *dir, struct nameidata *nd) return (nd->intent.open.flags & O_EXCL) != 0; } -static inline int nfs_reval_fsid(struct inode *dir, - struct nfs_fh *fh, struct nfs_fattr *fattr) +static inline int nfs_reval_fsid(struct vfsmount *mnt, struct inode *dir, + struct nfs_fh *fh, struct nfs_fattr *fattr) { struct nfs_server *server = NFS_SERVER(dir); if (!nfs_fsid_equal(&server->fsid, &fattr->fsid)) /* Revalidate fsid on root dir */ - return __nfs_revalidate_inode(server, dir->i_sb->s_root->d_inode); + return __nfs_revalidate_inode(server, mnt->mnt_root->d_inode); return 0; } @@ -913,7 +914,7 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru res = ERR_PTR(error); goto out_unlock; } - error = nfs_reval_fsid(dir, &fhandle, &fattr); + error = nfs_reval_fsid(nd->mnt, dir, &fhandle, &fattr); if (error < 0) { res = ERR_PTR(error); goto out_unlock; @@ -922,8 +923,9 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru res = (struct dentry *)inode; if (IS_ERR(res)) goto out_unlock; + no_entry: - res = d_add_unique(dentry, inode); + res = d_materialise_unique(dentry, inode); if (res != NULL) dentry = res; nfs_renew_times(dentry); @@ -1117,11 +1119,13 @@ static struct dentry *nfs_readdir_lookup(nfs_readdir_descriptor_t *desc) dput(dentry); return NULL; } - alias = d_add_unique(dentry, inode); + + alias = d_materialise_unique(dentry, inode); if (alias != NULL) { dput(dentry); dentry = alias; } + nfs_renew_times(dentry); nfs_set_verifier(dentry, nfs_save_change_attribute(dir)); return dentry; diff --git a/fs/nfs/getroot.c b/fs/nfs/getroot.c new file mode 100644 index 000000000000..977e59088eeb --- /dev/null +++ b/fs/nfs/getroot.c @@ -0,0 +1,306 @@ +/* getroot.c: get the root dentry for an NFS mount + * + * Copyright (C) 2006 Red Hat, Inc. All Rights Reserved. + * Written by David Howells (dhowells@redhat.com) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +#include "nfs4_fs.h" +#include "delegation.h" +#include "internal.h" + +#define NFSDBG_FACILITY NFSDBG_CLIENT +#define NFS_PARANOIA 1 + +/* + * get an NFS2/NFS3 root dentry from the root filehandle + */ +struct dentry *nfs_get_root(struct super_block *sb, struct nfs_fh *mntfh) +{ + struct nfs_server *server = NFS_SB(sb); + struct nfs_fsinfo fsinfo; + struct nfs_fattr fattr; + struct dentry *mntroot; + struct inode *inode; + int error; + + /* create a dummy root dentry with dummy inode for this superblock */ + if (!sb->s_root) { + struct nfs_fh dummyfh; + struct dentry *root; + struct inode *iroot; + + memset(&dummyfh, 0, sizeof(dummyfh)); + memset(&fattr, 0, sizeof(fattr)); + nfs_fattr_init(&fattr); + fattr.valid = NFS_ATTR_FATTR; + fattr.type = NFDIR; + fattr.mode = S_IFDIR | S_IRUSR | S_IWUSR; + fattr.nlink = 2; + + iroot = nfs_fhget(sb, &dummyfh, &fattr); + if (IS_ERR(iroot)) + return ERR_PTR(PTR_ERR(iroot)); + + root = d_alloc_root(iroot); + if (!root) { + iput(iroot); + return ERR_PTR(-ENOMEM); + } + + sb->s_root = root; + } + + /* get the actual root for this mount */ + fsinfo.fattr = &fattr; + + error = server->nfs_client->rpc_ops->getroot(server, mntfh, &fsinfo); + if (error < 0) { + dprintk("nfs_get_root: getattr error = %d\n", -error); + return ERR_PTR(error); + } + + inode = nfs_fhget(sb, mntfh, fsinfo.fattr); + if (IS_ERR(inode)) { + dprintk("nfs_get_root: get root inode failed\n"); + return ERR_PTR(PTR_ERR(inode)); + } + + /* root dentries normally start off anonymous and get spliced in later + * if the dentry tree reaches them; however if the dentry already + * exists, we'll pick it up at this point and use it as the root + */ + mntroot = d_alloc_anon(inode); + if (!mntroot) { + iput(inode); + dprintk("nfs_get_root: get root dentry failed\n"); + return ERR_PTR(-ENOMEM); + } + + if (!mntroot->d_op) + mntroot->d_op = server->nfs_client->rpc_ops->dentry_ops; + + return mntroot; +} + +#ifdef CONFIG_NFS_V4 + +/* + * Do a simple pathwalk from the root FH of the server to the nominated target + * of the mountpoint + * - give error on symlinks + * - give error on ".." occurring in the path + * - follow traversals + */ +int nfs4_path_walk(struct nfs_server *server, + struct nfs_fh *mntfh, + const char *path) +{ + struct nfs_fsinfo fsinfo; + struct nfs_fattr fattr; + struct nfs_fh lastfh; + struct qstr name; + int ret; + //int referral_count = 0; + + dprintk("--> nfs4_path_walk(,,%s)\n", path); + + fsinfo.fattr = &fattr; + nfs_fattr_init(&fattr); + + if (*path++ != '/') { + dprintk("nfs4_get_root: Path does not begin with a slash\n"); + return -EINVAL; + } + + /* Start by getting the root filehandle from the server */ + ret = server->nfs_client->rpc_ops->getroot(server, mntfh, &fsinfo); + if (ret < 0) { + dprintk("nfs4_get_root: getroot error = %d\n", -ret); + return ret; + } + + if (fattr.type != NFDIR) { + printk(KERN_ERR "nfs4_get_root:" + " getroot encountered non-directory\n"); + return -ENOTDIR; + } + + if (fattr.valid & NFS_ATTR_FATTR_V4_REFERRAL) { + printk(KERN_ERR "nfs4_get_root:" + " getroot obtained referral\n"); + return -EREMOTE; + } + +next_component: + dprintk("Next: %s\n", path); + + /* extract the next bit of the path */ + if (!*path) + goto path_walk_complete; + + name.name = path; + while (*path && *path != '/') + path++; + name.len = path - (const char *) name.name; + +eat_dot_dir: + while (*path == '/') + path++; + + if (path[0] == '.' && (path[1] == '/' || !path[1])) { + path += 2; + goto eat_dot_dir; + } + + if (path[0] == '.' && path[1] == '.' && (path[2] == '/' || !path[2]) + ) { + printk(KERN_ERR "nfs4_get_root:" + " Mount path contains reference to \"..\"\n"); + return -EINVAL; + } + + /* lookup the next FH in the sequence */ + memcpy(&lastfh, mntfh, sizeof(lastfh)); + + dprintk("LookupFH: %*.*s [%s]\n", name.len, name.len, name.name, path); + + ret = server->nfs_client->rpc_ops->lookupfh(server, &lastfh, &name, + mntfh, &fattr); + if (ret < 0) { + dprintk("nfs4_get_root: getroot error = %d\n", -ret); + return ret; + } + + if (fattr.type != NFDIR) { + printk(KERN_ERR "nfs4_get_root:" + " lookupfh encountered non-directory\n"); + return -ENOTDIR; + } + + if (fattr.valid & NFS_ATTR_FATTR_V4_REFERRAL) { + printk(KERN_ERR "nfs4_get_root:" + " lookupfh obtained referral\n"); + return -EREMOTE; + } + + goto next_component; + +path_walk_complete: + memcpy(&server->fsid, &fattr.fsid, sizeof(server->fsid)); + dprintk("<-- nfs4_path_walk() = 0\n"); + return 0; +} + +/* + * get an NFS4 root dentry from the root filehandle + */ +struct dentry *nfs4_get_root(struct super_block *sb, struct nfs_fh *mntfh) +{ + struct nfs_server *server = NFS_SB(sb); + struct nfs_fattr fattr; + struct dentry *mntroot; + struct inode *inode; + int error; + + dprintk("--> nfs4_get_root()\n"); + + /* create a dummy root dentry with dummy inode for this superblock */ + if (!sb->s_root) { + struct nfs_fh dummyfh; + struct dentry *root; + struct inode *iroot; + + memset(&dummyfh, 0, sizeof(dummyfh)); + memset(&fattr, 0, sizeof(fattr)); + nfs_fattr_init(&fattr); + fattr.valid = NFS_ATTR_FATTR; + fattr.type = NFDIR; + fattr.mode = S_IFDIR | S_IRUSR | S_IWUSR; + fattr.nlink = 2; + + iroot = nfs_fhget(sb, &dummyfh, &fattr); + if (IS_ERR(iroot)) + return ERR_PTR(PTR_ERR(iroot)); + + root = d_alloc_root(iroot); + if (!root) { + iput(iroot); + return ERR_PTR(-ENOMEM); + } + + sb->s_root = root; + } + + /* get the info about the server and filesystem */ + error = nfs4_server_capabilities(server, mntfh); + if (error < 0) { + dprintk("nfs_get_root: getcaps error = %d\n", + -error); + return ERR_PTR(error); + } + + /* get the actual root for this mount */ + error = server->nfs_client->rpc_ops->getattr(server, mntfh, &fattr); + if (error < 0) { + dprintk("nfs_get_root: getattr error = %d\n", -error); + return ERR_PTR(error); + } + + inode = nfs_fhget(sb, mntfh, &fattr); + if (IS_ERR(inode)) { + dprintk("nfs_get_root: get root inode failed\n"); + return ERR_PTR(PTR_ERR(inode)); + } + + /* root dentries normally start off anonymous and get spliced in later + * if the dentry tree reaches them; however if the dentry already + * exists, we'll pick it up at this point and use it as the root + */ + mntroot = d_alloc_anon(inode); + if (!mntroot) { + iput(inode); + dprintk("nfs_get_root: get root dentry failed\n"); + return ERR_PTR(-ENOMEM); + } + + if (!mntroot->d_op) + mntroot->d_op = server->nfs_client->rpc_ops->dentry_ops; + + dprintk("<-- nfs4_get_root()\n"); + return mntroot; +} + +#endif /* CONFIG_NFS_V4 */ diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c index 231c20ffc0ff..f96dfac7dc9a 100644 --- a/fs/nfs/idmap.c +++ b/fs/nfs/idmap.c @@ -114,8 +114,7 @@ nfs_idmap_new(struct nfs_client *clp) struct idmap *idmap; int error; - if (clp->cl_idmap != NULL) - return 0; + BUG_ON(clp->cl_idmap != NULL); if ((idmap = kzalloc(sizeof(*idmap), GFP_KERNEL)) == NULL) return -ENOMEM; diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 771c3b833757..a547c58a83e6 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1020,7 +1020,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) out_fileid: printk(KERN_ERR "NFS: server %s error: fileid changed\n" "fsid %s: expected fileid 0x%Lx, got 0x%Lx\n", - NFS_SERVER(inode)->hostname, inode->i_sb->s_id, + NFS_SERVER(inode)->nfs_client->cl_hostname, inode->i_sb->s_id, (long long)nfsi->fileid, (long long)fattr->fileid); goto out_err; } diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 2f3aa52fbefc..e73ba4f1052a 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -4,6 +4,18 @@ #include +struct nfs_string; +struct nfs_mount_data; +struct nfs4_mount_data; + +/* Maximum number of readahead requests + * FIXME: this should really be a sysctl so that users may tune it to suit + * their needs. People that do NFS over a slow network, might for + * instance want to reduce it to something closer to 1 for improved + * interactive response. + */ +#define NFS_MAX_READAHEAD (RPC_DEF_SLOT_TABLE - 1) + struct nfs_clone_mount { const struct super_block *sb; const struct dentry *dentry; @@ -16,12 +28,25 @@ struct nfs_clone_mount { }; /* client.c */ +extern struct rpc_program nfs_program; + extern void nfs_put_client(struct nfs_client *); extern struct nfs_client *nfs_find_client(const struct sockaddr_in *, int); -extern struct nfs_client *nfs_get_client(const char *, const struct sockaddr_in *, int); -extern void nfs_mark_client_ready(struct nfs_client *, int); -extern int nfs_create_rpc_client(struct nfs_client *, int, unsigned int, - unsigned int, rpc_authflavor_t); +extern struct nfs_server *nfs_create_server(const struct nfs_mount_data *, + struct nfs_fh *); +extern struct nfs_server *nfs4_create_server(const struct nfs4_mount_data *, + const char *, + const struct sockaddr_in *, + const char *, + const char *, + rpc_authflavor_t, + struct nfs_fh *); +extern struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *, + struct nfs_fh *); +extern void nfs_free_server(struct nfs_server *server); +extern struct nfs_server *nfs_clone_server(struct nfs_server *, + struct nfs_fh *, + struct nfs_fattr *); /* nfs4namespace.c */ #ifdef CONFIG_NFS_V4 @@ -89,10 +114,10 @@ extern void nfs4_clear_inode(struct inode *); #endif /* super.c */ -extern struct file_system_type nfs_referral_nfs4_fs_type; -extern struct file_system_type clone_nfs_fs_type; +extern struct file_system_type nfs_xdev_fs_type; #ifdef CONFIG_NFS_V4 -extern struct file_system_type clone_nfs4_fs_type; +extern struct file_system_type nfs4_xdev_fs_type; +extern struct file_system_type nfs4_referral_fs_type; #endif extern struct rpc_stat nfs_rpcstat; @@ -101,28 +126,30 @@ extern int __init register_nfs_fs(void); extern void __exit unregister_nfs_fs(void); /* namespace.c */ -extern char *nfs_path(const char *base, const struct dentry *dentry, +extern char *nfs_path(const char *base, + const struct dentry *droot, + const struct dentry *dentry, char *buffer, ssize_t buflen); -/* - * Determine the mount path as a string - */ +/* getroot.c */ +extern struct dentry *nfs_get_root(struct super_block *, struct nfs_fh *); #ifdef CONFIG_NFS_V4 -static inline char * -nfs4_path(const struct dentry *dentry, char *buffer, ssize_t buflen) -{ - return nfs_path(NFS_SB(dentry->d_sb)->mnt_path, dentry, buffer, buflen); -} +extern struct dentry *nfs4_get_root(struct super_block *, struct nfs_fh *); + +extern int nfs4_path_walk(struct nfs_server *server, + struct nfs_fh *mntfh, + const char *path); #endif /* * Determine the device name as a string */ static inline char *nfs_devname(const struct vfsmount *mnt_parent, - const struct dentry *dentry, - char *buffer, ssize_t buflen) + const struct dentry *dentry, + char *buffer, ssize_t buflen) { - return nfs_path(mnt_parent->mnt_devname, dentry, buffer, buflen); + return nfs_path(mnt_parent->mnt_devname, mnt_parent->mnt_root, + dentry, buffer, buflen); } /* @@ -178,20 +205,3 @@ void nfs_super_set_maxbytes(struct super_block *sb, __u64 maxfilesize) if (sb->s_maxbytes > MAX_LFS_FILESIZE || sb->s_maxbytes <= 0) sb->s_maxbytes = MAX_LFS_FILESIZE; } - -/* - * Check if the string represents a "valid" IPv4 address - */ -static inline int valid_ipaddr4(const char *buf) -{ - int rc, count, in[4]; - - rc = sscanf(buf, "%d.%d.%d.%d", &in[0], &in[1], &in[2], &in[3]); - if (rc != 4) - return -EINVAL; - for (count = 0; count < 4; count++) { - if (in[count] > 255) - return -EINVAL; - } - return 0; -} diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c index d8b8d56266cb..77b00684894d 100644 --- a/fs/nfs/namespace.c +++ b/fs/nfs/namespace.c @@ -2,6 +2,7 @@ * linux/fs/nfs/namespace.c * * Copyright (C) 2005 Trond Myklebust + * - Modified by David Howells * * NFS namespace */ @@ -28,6 +29,7 @@ int nfs_mountpoint_expiry_timeout = 500 * HZ; /* * nfs_path - reconstruct the path given an arbitrary dentry * @base - arbitrary string to prepend to the path + * @droot - pointer to root dentry for mountpoint * @dentry - pointer to dentry * @buffer - result buffer * @buflen - length of buffer @@ -38,7 +40,9 @@ int nfs_mountpoint_expiry_timeout = 500 * HZ; * This is mainly for use in figuring out the path on the * server side when automounting on top of an existing partition. */ -char *nfs_path(const char *base, const struct dentry *dentry, +char *nfs_path(const char *base, + const struct dentry *droot, + const struct dentry *dentry, char *buffer, ssize_t buflen) { char *end = buffer+buflen; @@ -47,7 +51,7 @@ char *nfs_path(const char *base, const struct dentry *dentry, *--end = '\0'; buflen--; spin_lock(&dcache_lock); - while (!IS_ROOT(dentry)) { + while (!IS_ROOT(dentry) && dentry != droot) { namelen = dentry->d_name.len; buflen -= namelen + 1; if (buflen < 0) @@ -96,12 +100,13 @@ static void * nfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd) struct nfs_fattr fattr; int err; + dprintk("--> nfs_follow_mountpoint()\n"); + BUG_ON(IS_ROOT(dentry)); dprintk("%s: enter\n", __FUNCTION__); dput(nd->dentry); nd->dentry = dget(dentry); - if (d_mountpoint(nd->dentry)) - goto out_follow; + /* Look it up again */ parent = dget_parent(nd->dentry); err = server->nfs_client->rpc_ops->lookup(parent->d_inode, @@ -134,6 +139,8 @@ static void * nfs_follow_mountpoint(struct dentry *dentry, struct nameidata *nd) schedule_delayed_work(&nfs_automount_task, nfs_mountpoint_expiry_timeout); out: dprintk("%s: done, returned %d\n", __FUNCTION__, err); + + dprintk("<-- nfs_follow_mountpoint() = %d\n", err); return ERR_PTR(err); out_err: path_release(nd); @@ -183,14 +190,14 @@ static struct vfsmount *nfs_do_clone_mount(struct nfs_server *server, switch (server->nfs_client->cl_nfsversion) { case 2: case 3: - mnt = vfs_kern_mount(&clone_nfs_fs_type, 0, devname, mountdata); + mnt = vfs_kern_mount(&nfs_xdev_fs_type, 0, devname, mountdata); break; case 4: - mnt = vfs_kern_mount(&clone_nfs4_fs_type, 0, devname, mountdata); + mnt = vfs_kern_mount(&nfs4_xdev_fs_type, 0, devname, mountdata); } return mnt; #else - return vfs_kern_mount(&clone_nfs_fs_type, 0, devname, mountdata); + return vfs_kern_mount(&nfs_xdev_fs_type, 0, devname, mountdata); #endif } @@ -216,6 +223,8 @@ struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent, char *page = (char *) __get_free_page(GFP_USER); char *devname; + dprintk("--> nfs_do_submount()\n"); + dprintk("%s: submounting on %s/%s\n", __FUNCTION__, dentry->d_parent->d_name.name, dentry->d_name.name); @@ -230,5 +239,7 @@ struct vfsmount *nfs_do_submount(const struct vfsmount *mnt_parent, free_page((unsigned long)page); out: dprintk("%s: done\n", __FUNCTION__); + + dprintk("<-- nfs_do_submount() = %p\n", mnt); return mnt; } diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c index 0622af0122be..9e8258ece6fd 100644 --- a/fs/nfs/nfs3proc.c +++ b/fs/nfs/nfs3proc.c @@ -81,7 +81,7 @@ do_proc_get_root(struct rpc_clnt *client, struct nfs_fh *fhandle, } /* - * Bare-bones access to getattr: this is for nfs_read_super. + * Bare-bones access to getattr: this is for nfs_get_root/nfs_get_sb */ static int nfs3_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle, diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index e7879245361e..61095fe4b5ca 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -188,8 +188,6 @@ extern void nfs4_kill_renewd(struct nfs_client *); extern void nfs4_renew_state(void *); /* nfs4state.c */ -extern void init_nfsv4_state(struct nfs_server *); -extern void destroy_nfsv4_state(struct nfs_server *); struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp); extern u32 nfs4_alloc_lockowner_id(struct nfs_client *); @@ -224,10 +222,6 @@ extern struct svc_version nfs4_callback_version1; #else -#define init_nfsv4_state(server) do { } while (0) -#define destroy_nfsv4_state(server) do { } while (0) -#define nfs4_put_state_owner(inode, owner) do { } while (0) -#define nfs4_put_open_state(state) do { } while (0) #define nfs4_close_state(a, b) do { } while (0) #endif /* CONFIG_NFS_V4 */ diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c index faed9bcba50f..24e47f3bbd17 100644 --- a/fs/nfs/nfs4namespace.c +++ b/fs/nfs/nfs4namespace.c @@ -2,6 +2,7 @@ * linux/fs/nfs/nfs4namespace.c * * Copyright (C) 2005 Trond Myklebust + * - Modified by David Howells * * NFSv4 namespace */ @@ -47,6 +48,68 @@ static inline char *nfs4_pathname_string(const struct nfs4_pathname *pathname, return ERR_PTR(-ENAMETOOLONG); } +/* + * Determine the mount path as a string + */ +static char *nfs4_path(const struct vfsmount *mnt_parent, + const struct dentry *dentry, + char *buffer, ssize_t buflen) +{ + const char *srvpath; + + srvpath = strchr(mnt_parent->mnt_devname, ':'); + if (srvpath) + srvpath++; + else + srvpath = mnt_parent->mnt_devname; + + return nfs_path(srvpath, mnt_parent->mnt_root, dentry, buffer, buflen); +} + +/* + * Check that fs_locations::fs_root [RFC3530 6.3] is a prefix for what we + * believe to be the server path to this dentry + */ +static int nfs4_validate_fspath(const struct vfsmount *mnt_parent, + const struct dentry *dentry, + const struct nfs4_fs_locations *locations, + char *page, char *page2) +{ + const char *path, *fs_path; + + path = nfs4_path(mnt_parent, dentry, page, PAGE_SIZE); + if (IS_ERR(path)) + return PTR_ERR(path); + + fs_path = nfs4_pathname_string(&locations->fs_path, page2, PAGE_SIZE); + if (IS_ERR(fs_path)) + return PTR_ERR(fs_path); + + if (strncmp(path, fs_path, strlen(fs_path)) != 0) { + dprintk("%s: path %s does not begin with fsroot %s\n", + __FUNCTION__, path, fs_path); + return -ENOENT; + } + + return 0; +} + +/* + * Check if the string represents a "valid" IPv4 address + */ +static inline int valid_ipaddr4(const char *buf) +{ + int rc, count, in[4]; + + rc = sscanf(buf, "%d.%d.%d.%d", &in[0], &in[1], &in[2], &in[3]); + if (rc != 4) + return -EINVAL; + for (count = 0; count < 4; count++) { + if (in[count] > 255) + return -EINVAL; + } + return 0; +} /** * nfs_follow_referral - set up mountpoint when hitting a referral on moved error @@ -68,10 +131,9 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent, .dentry = dentry, .authflavor = NFS_SB(mnt_parent->mnt_sb)->client->cl_auth->au_flavor, }; - char *page, *page2; - char *path, *fs_path; + char *page = NULL, *page2 = NULL; char *devname; - int loc, s; + int loc, s, error; if (locations == NULL || locations->nlocations <= 0) goto out; @@ -79,31 +141,25 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent, dprintk("%s: referral at %s/%s\n", __FUNCTION__, dentry->d_parent->d_name.name, dentry->d_name.name); - /* Ensure fs path is a prefix of current dentry path */ page = (char *) __get_free_page(GFP_USER); - if (page == NULL) + if (!page) goto out; + page2 = (char *) __get_free_page(GFP_USER); - if (page2 == NULL) + if (!page2) goto out; - path = nfs4_path(dentry, page, PAGE_SIZE); - if (IS_ERR(path)) - goto out_free; - - fs_path = nfs4_pathname_string(&locations->fs_path, page2, PAGE_SIZE); - if (IS_ERR(fs_path)) - goto out_free; - - if (strncmp(path, fs_path, strlen(fs_path)) != 0) { - dprintk("%s: path %s does not begin with fsroot %s\n", __FUNCTION__, path, fs_path); - goto out_free; + /* Ensure fs path is a prefix of current dentry path */ + error = nfs4_validate_fspath(mnt_parent, dentry, locations, page, page2); + if (error < 0) { + mnt = ERR_PTR(error); + goto out; } devname = nfs_devname(mnt_parent, dentry, page, PAGE_SIZE); if (IS_ERR(devname)) { mnt = (struct vfsmount *)devname; - goto out_free; + goto out; } loc = 0; @@ -140,7 +196,7 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent, addr.sin_port = htons(NFS_PORT); mountdata.addr = &addr; - mnt = vfs_kern_mount(&nfs_referral_nfs4_fs_type, 0, devname, &mountdata); + mnt = vfs_kern_mount(&nfs4_referral_fs_type, 0, devname, &mountdata); if (!IS_ERR(mnt)) { break; } @@ -149,10 +205,9 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent, loc++; } -out_free: - free_page((unsigned long)page); - free_page((unsigned long)page2); out: + free_page((unsigned long) page); + free_page((unsigned long) page2); dprintk("%s: done\n", __FUNCTION__); return mnt; } @@ -165,7 +220,7 @@ static struct vfsmount *nfs_follow_referral(const struct vfsmount *mnt_parent, */ struct vfsmount *nfs_do_refmount(const struct vfsmount *mnt_parent, struct dentry *dentry) { - struct vfsmount *mnt = ERR_PTR(-ENOENT); + struct vfsmount *mnt = ERR_PTR(-ENOMEM); struct dentry *parent; struct nfs4_fs_locations *fs_locations = NULL; struct page *page; @@ -183,11 +238,16 @@ struct vfsmount *nfs_do_refmount(const struct vfsmount *mnt_parent, struct dentr goto out_free; /* Get locations */ + mnt = ERR_PTR(-ENOENT); + parent = dget_parent(dentry); - dprintk("%s: getting locations for %s/%s\n", __FUNCTION__, parent->d_name.name, dentry->d_name.name); + dprintk("%s: getting locations for %s/%s\n", + __FUNCTION__, parent->d_name.name, dentry->d_name.name); + err = nfs4_proc_fs_locations(parent->d_inode, dentry, fs_locations, page); dput(parent); - if (err != 0 || fs_locations->nlocations <= 0 || + if (err != 0 || + fs_locations->nlocations <= 0 || fs_locations->fs_path.ncomponents <= 0) goto out_free; diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 1573eeb07ce1..a825547e8214 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -1393,70 +1393,19 @@ static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle, return err; } +/* + * get the file handle for the "/" directory on the server + */ static int nfs4_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle, - struct nfs_fsinfo *info) + struct nfs_fsinfo *info) { - struct nfs_fattr * fattr = info->fattr; - unsigned char * p; - struct qstr q; - struct nfs4_lookup_arg args = { - .dir_fh = fhandle, - .name = &q, - .bitmask = nfs4_fattr_bitmap, - }; - struct nfs4_lookup_res res = { - .server = server, - .fattr = fattr, - .fh = fhandle, - }; - struct rpc_message msg = { - .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP], - .rpc_argp = &args, - .rpc_resp = &res, - }; int status; - /* - * Now we do a separate LOOKUP for each component of the mount path. - * The LOOKUPs are done separately so that we can conveniently - * catch an ERR_WRONGSEC if it occurs along the way... - */ status = nfs4_lookup_root(server, fhandle, info); - if (status) - goto out; - - p = server->mnt_path; - for (;;) { - struct nfs4_exception exception = { }; - - while (*p == '/') - p++; - if (!*p) - break; - q.name = p; - while (*p && (*p != '/')) - p++; - q.len = p - q.name; - - do { - nfs_fattr_init(fattr); - status = nfs4_handle_exception(server, - rpc_call_sync(server->client, &msg, 0), - &exception); - } while (exception.retry); - if (status == 0) - continue; - if (status == -ENOENT) { - printk(KERN_NOTICE "NFS: mount path %s does not exist!\n", server->mnt_path); - printk(KERN_NOTICE "NFS: suggestion: try mounting '/' instead.\n"); - } - break; - } if (status == 0) status = nfs4_server_capabilities(server, fhandle); if (status == 0) status = nfs4_do_fsinfo(server, fhandle, info); -out: return nfs4_map_errors(status); } diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c index ff947ecb8b81..f2c893690ac4 100644 --- a/fs/nfs/nfs4renewd.c +++ b/fs/nfs/nfs4renewd.c @@ -127,26 +127,13 @@ nfs4_schedule_state_renewal(struct nfs_client *clp) void nfs4_renewd_prepare_shutdown(struct nfs_server *server) { - struct nfs_client *clp = server->nfs_client; - - if (!clp) - return; flush_scheduled_work(); - down_write(&clp->cl_sem); - if (!list_empty(&server->nfs4_siblings)) - list_del_init(&server->nfs4_siblings); - up_write(&clp->cl_sem); } -/* Must be called with clp->cl_sem locked for writes */ void nfs4_kill_renewd(struct nfs_client *clp) { down_read(&clp->cl_sem); - if (!list_empty(&clp->cl_superblocks)) { - up_read(&clp->cl_sem); - return; - } cancel_delayed_work(&clp->cl_renewd); up_read(&clp->cl_sem); flush_scheduled_work(); diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 058811e39555..5fffbdfa971f 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -58,24 +58,6 @@ const nfs4_stateid zero_stateid; static LIST_HEAD(nfs4_clientid_list); -void -init_nfsv4_state(struct nfs_server *server) -{ - server->nfs_client = NULL; - INIT_LIST_HEAD(&server->nfs4_siblings); -} - -void -destroy_nfsv4_state(struct nfs_server *server) -{ - kfree(server->mnt_path); - server->mnt_path = NULL; - if (server->nfs_client) { - nfs_put_client(server->nfs_client); - server->nfs_client = NULL; - } -} - static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred) { int status = nfs4_proc_setclientid(clp, NFS4_CALLBACK, diff --git a/fs/nfs/read.c b/fs/nfs/read.c index f0aff824a291..dae33c1e8a77 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -171,7 +171,7 @@ static int nfs_readpage_sync(struct nfs_open_context *ctx, struct inode *inode, rdata->args.offset = page_offset(page) + rdata->args.pgbase; dprintk("NFS: nfs_proc_read(%s, (%s/%Ld), %Lu, %u)\n", - NFS_SERVER(inode)->hostname, + NFS_SERVER(inode)->nfs_client->cl_hostname, inode->i_sb->s_id, (long long)NFS_FILEID(inode), (unsigned long long)rdata->args.pgbase, diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 5842d510d732..867b5dcd3a40 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -13,6 +13,11 @@ * * Split from inode.c by David Howells * + * - superblocks are indexed on server only - all inodes, dentries, etc. associated with a + * particular server are held in the same superblock + * - NFS superblocks can have several effective roots to the dentry tree + * - directory type roots are spliced into the tree when a path from one root reaches the root + * of another (see nfs_lookup()) */ #include @@ -52,20 +57,12 @@ #define NFSDBG_FACILITY NFSDBG_VFS -/* Maximum number of readahead requests - * FIXME: this should really be a sysctl so that users may tune it to suit - * their needs. People that do NFS over a slow network, might for - * instance want to reduce it to something closer to 1 for improved - * interactive response. - */ -#define NFS_MAX_READAHEAD (RPC_DEF_SLOT_TABLE - 1) - static void nfs_umount_begin(struct vfsmount *, int); static int nfs_statfs(struct dentry *, struct kstatfs *); static int nfs_show_options(struct seq_file *, struct vfsmount *); static int nfs_show_stats(struct seq_file *, struct vfsmount *); static int nfs_get_sb(struct file_system_type *, int, const char *, void *, struct vfsmount *); -static int nfs_clone_nfs_sb(struct file_system_type *fs_type, +static int nfs_xdev_get_sb(struct file_system_type *fs_type, int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt); static void nfs_kill_super(struct super_block *); @@ -77,10 +74,10 @@ static struct file_system_type nfs_fs_type = { .fs_flags = FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA, }; -struct file_system_type clone_nfs_fs_type = { +struct file_system_type nfs_xdev_fs_type = { .owner = THIS_MODULE, .name = "nfs", - .get_sb = nfs_clone_nfs_sb, + .get_sb = nfs_xdev_get_sb, .kill_sb = nfs_kill_super, .fs_flags = FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA, }; @@ -99,10 +96,10 @@ static struct super_operations nfs_sops = { #ifdef CONFIG_NFS_V4 static int nfs4_get_sb(struct file_system_type *fs_type, int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt); -static int nfs_clone_nfs4_sb(struct file_system_type *fs_type, - int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt); -static int nfs_referral_nfs4_sb(struct file_system_type *fs_type, - int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt); +static int nfs4_xdev_get_sb(struct file_system_type *fs_type, + int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt); +static int nfs4_referral_get_sb(struct file_system_type *fs_type, + int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt); static void nfs4_kill_super(struct super_block *sb); static struct file_system_type nfs4_fs_type = { @@ -113,18 +110,18 @@ static struct file_system_type nfs4_fs_type = { .fs_flags = FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA, }; -struct file_system_type clone_nfs4_fs_type = { +struct file_system_type nfs4_xdev_fs_type = { .owner = THIS_MODULE, .name = "nfs4", - .get_sb = nfs_clone_nfs4_sb, + .get_sb = nfs4_xdev_get_sb, .kill_sb = nfs4_kill_super, .fs_flags = FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA, }; -struct file_system_type nfs_referral_nfs4_fs_type = { +struct file_system_type nfs4_referral_fs_type = { .owner = THIS_MODULE, .name = "nfs4", - .get_sb = nfs_referral_nfs4_sb, + .get_sb = nfs4_referral_get_sb, .kill_sb = nfs4_kill_super, .fs_flags = FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA, }; @@ -345,7 +342,7 @@ static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt) nfs_show_mount_options(m, nfss, 0); seq_puts(m, ",addr="); - seq_escape(m, nfss->hostname, " \t\n\\"); + seq_escape(m, nfss->nfs_client->cl_hostname, " \t\n\\"); return 0; } @@ -429,449 +426,30 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt) /* * Begin unmount by attempting to remove all automounted mountpoints we added - * in response to traversals + * in response to xdev traversals and referrals */ static void nfs_umount_begin(struct vfsmount *vfsmnt, int flags) { - struct nfs_server *server; - struct rpc_clnt *rpc; - shrink_submounts(vfsmnt, &nfs_automount_list); - if (!(flags & MNT_FORCE)) - return; - /* -EIO all pending I/O */ - server = NFS_SB(vfsmnt->mnt_sb); - rpc = server->client; - if (!IS_ERR(rpc)) - rpc_killall_tasks(rpc); - rpc = server->client_acl; - if (!IS_ERR(rpc)) - rpc_killall_tasks(rpc); } /* - * Obtain the root inode of the file system. + * Validate the NFS2/NFS3 mount data + * - fills in the mount root filehandle */ -static struct inode * -nfs_get_root(struct super_block *sb, struct nfs_fh *rootfh, struct nfs_fsinfo *fsinfo) +static int nfs_validate_mount_data(struct nfs_mount_data *data, + struct nfs_fh *mntfh) { - struct nfs_server *server = NFS_SB(sb); - int error; - - error = server->nfs_client->rpc_ops->getroot(server, rootfh, fsinfo); - if (error < 0) { - dprintk("nfs_get_root: getattr error = %d\n", -error); - return ERR_PTR(error); - } - - server->fsid = fsinfo->fattr->fsid; - return nfs_fhget(sb, rootfh, fsinfo->fattr); -} - -/* - * Do NFS version-independent mount processing, and sanity checking - */ -static int -nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor) -{ - struct nfs_server *server; - struct inode *root_inode; - struct nfs_fattr fattr; - struct nfs_fsinfo fsinfo = { - .fattr = &fattr, - }; - struct nfs_pathconf pathinfo = { - .fattr = &fattr, - }; - int no_root_error = 0; - unsigned long max_rpc_payload; - - /* We probably want something more informative here */ - snprintf(sb->s_id, sizeof(sb->s_id), "%x:%x", MAJOR(sb->s_dev), MINOR(sb->s_dev)); - - server = NFS_SB(sb); - - sb->s_magic = NFS_SUPER_MAGIC; - - server->io_stats = nfs_alloc_iostats(); - if (server->io_stats == NULL) - return -ENOMEM; - - root_inode = nfs_get_root(sb, &server->fh, &fsinfo); - /* Did getting the root inode fail? */ - if (IS_ERR(root_inode)) { - no_root_error = PTR_ERR(root_inode); - goto out_no_root; - } - sb->s_root = d_alloc_root(root_inode); - if (!sb->s_root) { - no_root_error = -ENOMEM; - goto out_no_root; - } - sb->s_root->d_op = server->nfs_client->rpc_ops->dentry_ops; - - /* mount time stamp, in seconds */ - server->mount_time = jiffies; - - /* Get some general file system info */ - if (server->namelen == 0 && - server->nfs_client->rpc_ops->pathconf(server, &server->fh, &pathinfo) >= 0) - server->namelen = pathinfo.max_namelen; - /* Work out a lot of parameters */ - if (server->rsize == 0) - server->rsize = nfs_block_size(fsinfo.rtpref, NULL); - if (server->wsize == 0) - server->wsize = nfs_block_size(fsinfo.wtpref, NULL); - - if (fsinfo.rtmax >= 512 && server->rsize > fsinfo.rtmax) - server->rsize = nfs_block_size(fsinfo.rtmax, NULL); - if (fsinfo.wtmax >= 512 && server->wsize > fsinfo.wtmax) - server->wsize = nfs_block_size(fsinfo.wtmax, NULL); - - max_rpc_payload = nfs_block_size(rpc_max_payload(server->client), NULL); - if (server->rsize > max_rpc_payload) - server->rsize = max_rpc_payload; - if (server->rsize > NFS_MAX_FILE_IO_SIZE) - server->rsize = NFS_MAX_FILE_IO_SIZE; - server->rpages = (server->rsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; - - if (server->wsize > max_rpc_payload) - server->wsize = max_rpc_payload; - if (server->wsize > NFS_MAX_FILE_IO_SIZE) - server->wsize = NFS_MAX_FILE_IO_SIZE; - server->wpages = (server->wsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; - - if (sb->s_blocksize == 0) - sb->s_blocksize = nfs_block_bits(server->wsize, - &sb->s_blocksize_bits); - server->wtmult = nfs_block_bits(fsinfo.wtmult, NULL); - - server->dtsize = nfs_block_size(fsinfo.dtpref, NULL); - if (server->dtsize > PAGE_CACHE_SIZE) - server->dtsize = PAGE_CACHE_SIZE; - if (server->dtsize > server->rsize) - server->dtsize = server->rsize; - - if (server->flags & NFS_MOUNT_NOAC) { - server->acregmin = server->acregmax = 0; - server->acdirmin = server->acdirmax = 0; - sb->s_flags |= MS_SYNCHRONOUS; - } - server->backing_dev_info.ra_pages = server->rpages * NFS_MAX_READAHEAD; - - nfs_super_set_maxbytes(sb, fsinfo.maxfilesize); - - server->client->cl_intr = (server->flags & NFS_MOUNT_INTR) ? 1 : 0; - server->client->cl_softrtry = (server->flags & NFS_MOUNT_SOFT) ? 1 : 0; - - /* We're airborne Set socket buffersize */ - rpc_setbufsize(server->client, server->wsize + 100, server->rsize + 100); - return 0; - /* Yargs. It didn't work out. */ -out_no_root: - dprintk("nfs_sb_init: get root inode failed: errno %d\n", -no_root_error); - if (!IS_ERR(root_inode)) - iput(root_inode); - return no_root_error; -} - -/* - * Create an RPC client handle. - */ -static struct rpc_clnt * -nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data) -{ - struct nfs_client *clp; - struct rpc_clnt *clnt; - int proto = (data->flags & NFS_MOUNT_TCP) ? IPPROTO_TCP : IPPROTO_UDP; - int nfsversion = 2; - int err; - -#ifdef CONFIG_NFS_V3 - if (server->flags & NFS_MOUNT_VER3) - nfsversion = 3; -#endif - - clp = nfs_get_client(server->hostname, &server->addr, nfsversion); - if (!clp) { - dprintk("%s: failed to create NFS4 client.\n", __FUNCTION__); - return ERR_PTR(PTR_ERR(clp)); - } - - if (clp->cl_cons_state == NFS_CS_INITING) { - /* Check NFS protocol revision and initialize RPC op - * vector and file handle pool. */ -#ifdef CONFIG_NFS_V3 - if (nfsversion == 3) { - clp->rpc_ops = &nfs_v3_clientops; - server->caps |= NFS_CAP_READDIRPLUS; - } else { - clp->rpc_ops = &nfs_v2_clientops; - } -#else - clp->rpc_ops = &nfs_v2_clientops; -#endif - - /* create transport and client */ - err = nfs_create_rpc_client(clp, proto, data->timeo, - data->retrans, RPC_AUTH_UNIX); - if (err < 0) - goto client_init_error; - - nfs_mark_client_ready(clp, 0); - } - - /* create an nfs_server-specific client */ - clnt = rpc_clone_client(clp->cl_rpcclient); - if (IS_ERR(clnt)) { - dprintk("%s: couldn't create rpc_client!\n", __FUNCTION__); - nfs_put_client(clp); - return ERR_PTR(PTR_ERR(clnt)); - } - - if (data->pseudoflavor != clp->cl_rpcclient->cl_auth->au_flavor) { - struct rpc_auth *auth; - - auth = rpcauth_create(data->pseudoflavor, server->client); - if (IS_ERR(auth)) { - dprintk("%s: couldn't create credcache!\n", __FUNCTION__); - return ERR_PTR(PTR_ERR(auth)); - } - } - - server->nfs_client = clp; - return clnt; - -client_init_error: - nfs_mark_client_ready(clp, err); - nfs_put_client(clp); - return ERR_PTR(err); -} - -/* - * Clone a server record - */ -static struct nfs_server *nfs_clone_server(struct super_block *sb, struct nfs_clone_mount *data) -{ - struct nfs_server *server = NFS_SB(sb); - struct nfs_server *parent = NFS_SB(data->sb); - struct inode *root_inode; - struct nfs_fsinfo fsinfo; - void *err = ERR_PTR(-ENOMEM); - - sb->s_op = data->sb->s_op; - sb->s_blocksize = data->sb->s_blocksize; - sb->s_blocksize_bits = data->sb->s_blocksize_bits; - sb->s_maxbytes = data->sb->s_maxbytes; - - server->client_acl = ERR_PTR(-EINVAL); - server->io_stats = nfs_alloc_iostats(); - if (server->io_stats == NULL) - goto out; - - server->client = rpc_clone_client(parent->client); - if (IS_ERR((err = server->client))) - goto out; - - if (!IS_ERR(parent->client_acl)) { - server->client_acl = rpc_clone_client(parent->client_acl); - if (IS_ERR((err = server->client_acl))) - goto out; - } - root_inode = nfs_fhget(sb, data->fh, data->fattr); - if (!root_inode) - goto out; - sb->s_root = d_alloc_root(root_inode); - if (!sb->s_root) - goto out_put_root; - fsinfo.fattr = data->fattr; - if (NFS_PROTO(root_inode)->fsinfo(server, data->fh, &fsinfo) == 0) - nfs_super_set_maxbytes(sb, fsinfo.maxfilesize); - sb->s_root->d_op = server->nfs_client->rpc_ops->dentry_ops; - sb->s_flags |= MS_ACTIVE; - return server; -out_put_root: - iput(root_inode); -out: - return err; -} - -/* - * Copy an existing superblock and attach revised data - */ -static int nfs_clone_generic_sb(struct nfs_clone_mount *data, - struct super_block *(*fill_sb)(struct nfs_server *, struct nfs_clone_mount *), - struct nfs_server *(*fill_server)(struct super_block *, struct nfs_clone_mount *), - struct vfsmount *mnt) -{ - struct nfs_server *server; - struct nfs_server *parent = NFS_SB(data->sb); - struct super_block *sb = ERR_PTR(-EINVAL); - char *hostname; - int error = -ENOMEM; - int len; - - server = kmalloc(sizeof(struct nfs_server), GFP_KERNEL); - if (server == NULL) - goto out_err; - memcpy(server, parent, sizeof(*server)); - atomic_inc(&server->nfs_client->cl_count); - hostname = (data->hostname != NULL) ? data->hostname : parent->hostname; - len = strlen(hostname) + 1; - server->hostname = kmalloc(len, GFP_KERNEL); - if (server->hostname == NULL) - goto free_server; - memcpy(server->hostname, hostname, len); - - sb = fill_sb(server, data); - if (IS_ERR(sb)) { - error = PTR_ERR(sb); - goto free_hostname; - } - - if (sb->s_root) - goto out_share; - - server = fill_server(sb, data); - if (IS_ERR(server)) { - error = PTR_ERR(server); - goto out_deactivate; - } - return simple_set_mnt(mnt, sb); -out_deactivate: - up_write(&sb->s_umount); - deactivate_super(sb); - return error; -out_share: - kfree(server->hostname); - nfs_put_client(server->nfs_client); - kfree(server); - return simple_set_mnt(mnt, sb); -free_hostname: - kfree(server->hostname); -free_server: - nfs_put_client(server->nfs_client); - kfree(server); -out_err: - return error; -} - -/* - * Set up an NFS2/3 superblock - * - * The way this works is that the mount process passes a structure - * in the data argument which contains the server's IP address - * and the root file handle obtained from the server's mount - * daemon. We stash these away in the private superblock fields. - */ -static int -nfs_fill_super(struct super_block *sb, struct nfs_mount_data *data, int silent) -{ - struct nfs_server *server; - rpc_authflavor_t authflavor; - - server = NFS_SB(sb); - sb->s_blocksize_bits = 0; - sb->s_blocksize = 0; - if (data->bsize) - sb->s_blocksize = nfs_block_size(data->bsize, &sb->s_blocksize_bits); - if (data->rsize) - server->rsize = nfs_block_size(data->rsize, NULL); - if (data->wsize) - server->wsize = nfs_block_size(data->wsize, NULL); - server->flags = data->flags & NFS_MOUNT_FLAGMASK; - - server->acregmin = data->acregmin*HZ; - server->acregmax = data->acregmax*HZ; - server->acdirmin = data->acdirmin*HZ; - server->acdirmax = data->acdirmax*HZ; - - /* Start lockd here, before we might error out */ - if (!(server->flags & NFS_MOUNT_NONLM)) - lockd_up(); - - server->namelen = data->namlen; - server->hostname = kmalloc(strlen(data->hostname) + 1, GFP_KERNEL); - if (!server->hostname) - return -ENOMEM; - strcpy(server->hostname, data->hostname); - - /* Fill in pseudoflavor for mount version < 5 */ - if (!(data->flags & NFS_MOUNT_SECFLAVOUR)) - data->pseudoflavor = RPC_AUTH_UNIX; - authflavor = data->pseudoflavor; /* save for sb_init() */ - /* XXX maybe we want to add a server->pseudoflavor field */ - - /* Create RPC client handles */ - server->client = nfs_create_client(server, data); - if (IS_ERR(server->client)) - return PTR_ERR(server->client); - - /* RFC 2623, sec 2.3.2 */ - if (server->flags & NFS_MOUNT_VER3) { -#ifdef CONFIG_NFS_V3_ACL - if (!(server->flags & NFS_MOUNT_NOACL)) { - server->client_acl = rpc_bind_new_program(server->client, &nfsacl_program, 3); - /* No errors! Assume that Sun nfsacls are supported */ - if (!IS_ERR(server->client_acl)) - server->caps |= NFS_CAP_ACLS; - } -#else - server->flags &= ~NFS_MOUNT_NOACL; -#endif /* CONFIG_NFS_V3_ACL */ - /* - * The VFS shouldn't apply the umask to mode bits. We will - * do so ourselves when necessary. - */ - sb->s_flags |= MS_POSIXACL; - if (server->namelen == 0 || server->namelen > NFS3_MAXNAMLEN) - server->namelen = NFS3_MAXNAMLEN; - sb->s_time_gran = 1; - } else { - if (server->namelen == 0 || server->namelen > NFS2_MAXNAMLEN) - server->namelen = NFS2_MAXNAMLEN; - } - - sb->s_op = &nfs_sops; - return nfs_sb_init(sb, authflavor); -} - -static int nfs_set_super(struct super_block *s, void *data) -{ - s->s_fs_info = data; - return set_anon_super(s, data); -} - -static int nfs_compare_super(struct super_block *sb, void *data) -{ - struct nfs_server *server = data; - struct nfs_server *old = NFS_SB(sb); - - if (old->addr.sin_addr.s_addr != server->addr.sin_addr.s_addr) - return 0; - if (old->addr.sin_port != server->addr.sin_port) - return 0; - return !nfs_compare_fh(&old->fh, &server->fh); -} - -static int nfs_get_sb(struct file_system_type *fs_type, - int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt) -{ - int error; - struct nfs_server *server = NULL; - struct super_block *s; - struct nfs_fh *root; - struct nfs_mount_data *data = raw_data; - - error = -EINVAL; if (data == NULL) { dprintk("%s: missing data argument\n", __FUNCTION__); - goto out_err_noserver; + return -EINVAL; } + if (data->version <= 0 || data->version > NFS_MOUNT_VERSION) { dprintk("%s: bad mount version\n", __FUNCTION__); - goto out_err_noserver; + return -EINVAL; } + switch (data->version) { case 1: data->namlen = 0; @@ -882,7 +460,7 @@ static int nfs_get_sb(struct file_system_type *fs_type, dprintk("%s: mount structure version %d does not support NFSv3\n", __FUNCTION__, data->version); - goto out_err_noserver; + return -EINVAL; } data->root.size = NFS2_FHSIZE; memcpy(data->root.data, data->old_root.data, NFS2_FHSIZE); @@ -891,252 +469,308 @@ static int nfs_get_sb(struct file_system_type *fs_type, dprintk("%s: mount structure version %d does not support strong security\n", __FUNCTION__, data->version); - goto out_err_noserver; + return -EINVAL; } + /* Fill in pseudoflavor for mount version < 5 */ + data->pseudoflavor = RPC_AUTH_UNIX; case 5: memset(data->context, 0, sizeof(data->context)); } + #ifndef CONFIG_NFS_V3 /* If NFSv3 is not compiled in, return -EPROTONOSUPPORT */ - error = -EPROTONOSUPPORT; if (data->flags & NFS_MOUNT_VER3) { dprintk("%s: NFSv3 not compiled into kernel\n", __FUNCTION__); - goto out_err_noserver; + return -EPROTONOSUPPORT; } #endif /* CONFIG_NFS_V3 */ - error = -ENOMEM; - server = kzalloc(sizeof(struct nfs_server), GFP_KERNEL); - if (!server) - goto out_err_noserver; - /* Zero out the NFS state stuff */ - init_nfsv4_state(server); - server->client = server->client_acl = ERR_PTR(-EINVAL); - - root = &server->fh; - if (data->flags & NFS_MOUNT_VER3) - root->size = data->root.size; - else - root->size = NFS2_FHSIZE; - error = -EINVAL; - if (root->size > sizeof(root->data)) { - dprintk("%s: invalid root filehandle\n", __FUNCTION__); - goto out_err; - } - memcpy(root->data, data->root.data, root->size); - /* We now require that the mount process passes the remote address */ - memcpy(&server->addr, &data->addr, sizeof(server->addr)); - if (server->addr.sin_addr.s_addr == INADDR_ANY) { + if (data->addr.sin_addr.s_addr == INADDR_ANY) { dprintk("%s: mount program didn't pass remote address!\n", - __FUNCTION__); - goto out_err; + __FUNCTION__); + return -EINVAL; } + /* Prepare the root filehandle */ + if (data->flags & NFS_MOUNT_VER3) + mntfh->size = data->root.size; + else + mntfh->size = NFS2_FHSIZE; + + if (mntfh->size > sizeof(mntfh->data)) { + dprintk("%s: invalid root filehandle\n", __FUNCTION__); + return -EINVAL; + } + + memcpy(mntfh->data, data->root.data, mntfh->size); + if (mntfh->size < sizeof(mntfh->data)) + memset(mntfh->data + mntfh->size, 0, + sizeof(mntfh->data) - mntfh->size); + + return 0; +} + +/* + * Initialise the common bits of the superblock + */ +static inline void nfs_initialise_sb(struct super_block *sb) +{ + struct nfs_server *server = NFS_SB(sb); + + sb->s_magic = NFS_SUPER_MAGIC; + + /* We probably want something more informative here */ + snprintf(sb->s_id, sizeof(sb->s_id), + "%x:%x", MAJOR(sb->s_dev), MINOR(sb->s_dev)); + + if (sb->s_blocksize == 0) + sb->s_blocksize = nfs_block_bits(server->wsize, + &sb->s_blocksize_bits); + + if (server->flags & NFS_MOUNT_NOAC) + sb->s_flags |= MS_SYNCHRONOUS; + + nfs_super_set_maxbytes(sb, server->maxfilesize); +} + +/* + * Finish setting up an NFS2/3 superblock + */ +static void nfs_fill_super(struct super_block *sb, struct nfs_mount_data *data) +{ + struct nfs_server *server = NFS_SB(sb); + + sb->s_blocksize_bits = 0; + sb->s_blocksize = 0; + if (data->bsize) + sb->s_blocksize = nfs_block_size(data->bsize, &sb->s_blocksize_bits); + + if (server->flags & NFS_MOUNT_VER3) { + /* The VFS shouldn't apply the umask to mode bits. We will do + * so ourselves when necessary. + */ + sb->s_flags |= MS_POSIXACL; + sb->s_time_gran = 1; + } + + sb->s_op = &nfs_sops; + nfs_initialise_sb(sb); +} + +/* + * Finish setting up a cloned NFS2/3 superblock + */ +static void nfs_clone_super(struct super_block *sb, + const struct super_block *old_sb) +{ + struct nfs_server *server = NFS_SB(sb); + + sb->s_blocksize_bits = old_sb->s_blocksize_bits; + sb->s_blocksize = old_sb->s_blocksize; + sb->s_maxbytes = old_sb->s_maxbytes; + + if (server->flags & NFS_MOUNT_VER3) { + /* The VFS shouldn't apply the umask to mode bits. We will do + * so ourselves when necessary. + */ + sb->s_flags |= MS_POSIXACL; + sb->s_time_gran = 1; + } + + sb->s_op = old_sb->s_op; + nfs_initialise_sb(sb); +} + +static int nfs_set_super(struct super_block *s, void *_server) +{ + struct nfs_server *server = _server; + int ret; + + s->s_fs_info = server; + ret = set_anon_super(s, server); + if (ret == 0) + server->s_dev = s->s_dev; + return ret; +} + +static int nfs_compare_super(struct super_block *sb, void *data) +{ + struct nfs_server *server = data, *old = NFS_SB(sb); + + if (old->nfs_client != server->nfs_client) + return 0; + if (memcmp(&old->fsid, &server->fsid, sizeof(old->fsid)) != 0) + return 0; + return 1; +} + +static int nfs_get_sb(struct file_system_type *fs_type, + int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt) +{ + struct nfs_server *server = NULL; + struct super_block *s; + struct nfs_fh mntfh; + struct nfs_mount_data *data = raw_data; + struct dentry *mntroot; + int error; + + /* Validate the mount data */ + error = nfs_validate_mount_data(data, &mntfh); + if (error < 0) + return error; + + /* Get a volume representation */ + server = nfs_create_server(data, &mntfh); + if (IS_ERR(server)) { + error = PTR_ERR(server); + goto out_err_noserver; + } + + /* Get a superblock - note that we may end up sharing one that already exists */ s = sget(fs_type, nfs_compare_super, nfs_set_super, server); if (IS_ERR(s)) { error = PTR_ERR(s); - goto out_err; + goto out_err_nosb; } - if (s->s_root) - goto out_share; - - s->s_flags = flags; - - error = nfs_fill_super(s, data, flags & MS_SILENT ? 1 : 0); - if (error) { - up_write(&s->s_umount); - deactivate_super(s); - return error; + if (s->s_fs_info != server) { + nfs_free_server(server); + server = NULL; } + + if (!s->s_root) { + /* initial superblock/root creation */ + s->s_flags = flags; + nfs_fill_super(s, data); + } + + mntroot = nfs_get_root(s, &mntfh); + if (IS_ERR(mntroot)) { + error = PTR_ERR(mntroot); + goto error_splat_super; + } + s->s_flags |= MS_ACTIVE; - return simple_set_mnt(mnt, s); + mnt->mnt_sb = s; + mnt->mnt_root = mntroot; + return 0; -out_share: - kfree(server); - return simple_set_mnt(mnt, s); - -out_err: - kfree(server); +out_err_nosb: + nfs_free_server(server); out_err_noserver: return error; + +error_splat_super: + up_write(&s->s_umount); + deactivate_super(s); + return error; } +/* + * Destroy an NFS2/3 superblock + */ static void nfs_kill_super(struct super_block *s) { struct nfs_server *server = NFS_SB(s); kill_anon_super(s); - - if (!IS_ERR(server->client)) - rpc_shutdown_client(server->client); - if (!IS_ERR(server->client_acl)) - rpc_shutdown_client(server->client_acl); - - if (!(server->flags & NFS_MOUNT_NONLM)) - lockd_down(); /* release rpc.lockd */ - - nfs_free_iostats(server->io_stats); - kfree(server->hostname); - nfs_put_client(server->nfs_client); - kfree(server); - nfs_release_automount_timer(); + nfs_free_server(server); } -static struct super_block *nfs_clone_sb(struct nfs_server *server, struct nfs_clone_mount *data) -{ - struct super_block *sb; - - server->fsid = data->fattr->fsid; - nfs_copy_fh(&server->fh, data->fh); - sb = sget(&nfs_fs_type, nfs_compare_super, nfs_set_super, server); - if (!IS_ERR(sb) && sb->s_root == NULL && !(server->flags & NFS_MOUNT_NONLM)) - lockd_up(); - return sb; -} - -static int nfs_clone_nfs_sb(struct file_system_type *fs_type, - int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt) +/* + * Clone an NFS2/3 server record on xdev traversal (FSID-change) + */ +static int nfs_xdev_get_sb(struct file_system_type *fs_type, int flags, + const char *dev_name, void *raw_data, + struct vfsmount *mnt) { struct nfs_clone_mount *data = raw_data; - return nfs_clone_generic_sb(data, nfs_clone_sb, nfs_clone_server, mnt); + struct super_block *s; + struct nfs_server *server; + struct dentry *mntroot; + int error; + + dprintk("--> nfs_xdev_get_sb()\n"); + + /* create a new volume representation */ + server = nfs_clone_server(NFS_SB(data->sb), data->fh, data->fattr); + if (IS_ERR(server)) { + error = PTR_ERR(server); + goto out_err_noserver; + } + + /* Get a superblock - note that we may end up sharing one that already exists */ + s = sget(&nfs_fs_type, nfs_compare_super, nfs_set_super, server); + if (IS_ERR(s)) { + error = PTR_ERR(s); + goto out_err_nosb; + } + + if (s->s_fs_info != server) { + nfs_free_server(server); + server = NULL; + } + + if (!s->s_root) { + /* initial superblock/root creation */ + s->s_flags = flags; + nfs_clone_super(s, data->sb); + } + + mntroot = nfs_get_root(s, data->fh); + if (IS_ERR(mntroot)) { + error = PTR_ERR(mntroot); + goto error_splat_super; + } + + s->s_flags |= MS_ACTIVE; + mnt->mnt_sb = s; + mnt->mnt_root = mntroot; + + dprintk("<-- nfs_xdev_get_sb() = 0\n"); + return 0; + +out_err_nosb: + nfs_free_server(server); +out_err_noserver: + dprintk("<-- nfs_xdev_get_sb() = %d [error]\n", error); + return error; + +error_splat_super: + up_write(&s->s_umount); + deactivate_super(s); + dprintk("<-- nfs_xdev_get_sb() = %d [splat]\n", error); + return error; } #ifdef CONFIG_NFS_V4 -static struct rpc_clnt *nfs4_create_client(struct nfs_server *server, - int timeo, int retrans, int proto, rpc_authflavor_t flavor) + +/* + * Finish setting up a cloned NFS4 superblock + */ +static void nfs4_clone_super(struct super_block *sb, + const struct super_block *old_sb) { - struct nfs_client *clp; - struct rpc_clnt *clnt = NULL; - int err = -EIO; - - clp = nfs_get_client(server->hostname, &server->addr, 4); - if (!clp) { - dprintk("%s: failed to create NFS4 client.\n", __FUNCTION__); - return ERR_PTR(err); - } - - /* Now create transport and client */ - if (clp->cl_cons_state == NFS_CS_INITING) { - clp->rpc_ops = &nfs_v4_clientops; - - err = nfs_create_rpc_client(clp, proto, timeo, retrans, flavor); - if (err < 0) - goto client_init_error; - - memcpy(clp->cl_ipaddr, server->ip_addr, sizeof(clp->cl_ipaddr)); - err = nfs_idmap_new(clp); - if (err < 0) { - dprintk("%s: failed to create idmapper.\n", - __FUNCTION__); - goto client_init_error; - } - __set_bit(NFS_CS_IDMAP, &clp->cl_res_state); - nfs_mark_client_ready(clp, 0); - } - - clnt = rpc_clone_client(clp->cl_rpcclient); - - if (IS_ERR(clnt)) { - dprintk("%s: cannot create RPC client. Error = %d\n", - __FUNCTION__, err); - return clnt; - } - - if (clnt->cl_auth->au_flavor != flavor) { - struct rpc_auth *auth; - - auth = rpcauth_create(flavor, clnt); - if (IS_ERR(auth)) { - dprintk("%s: couldn't create credcache!\n", __FUNCTION__); - return (struct rpc_clnt *)auth; - } - } - - server->nfs_client = clp; - down_write(&clp->cl_sem); - list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks); - up_write(&clp->cl_sem); - return clnt; - -client_init_error: - nfs_mark_client_ready(clp, err); - nfs_put_client(clp); - return ERR_PTR(err); + sb->s_blocksize_bits = old_sb->s_blocksize_bits; + sb->s_blocksize = old_sb->s_blocksize; + sb->s_maxbytes = old_sb->s_maxbytes; + sb->s_time_gran = 1; + sb->s_op = old_sb->s_op; + nfs_initialise_sb(sb); } /* * Set up an NFS4 superblock */ -static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data, int silent) +static void nfs4_fill_super(struct super_block *sb) { - struct nfs_server *server; - rpc_authflavor_t authflavour; - int err = -EIO; - - sb->s_blocksize_bits = 0; - sb->s_blocksize = 0; - server = NFS_SB(sb); - if (data->rsize != 0) - server->rsize = nfs_block_size(data->rsize, NULL); - if (data->wsize != 0) - server->wsize = nfs_block_size(data->wsize, NULL); - server->flags = data->flags & NFS_MOUNT_FLAGMASK; - server->caps = NFS_CAP_ATOMIC_OPEN; - - server->acregmin = data->acregmin*HZ; - server->acregmax = data->acregmax*HZ; - server->acdirmin = data->acdirmin*HZ; - server->acdirmax = data->acdirmax*HZ; - - /* Now create transport and client */ - authflavour = RPC_AUTH_UNIX; - if (data->auth_flavourlen != 0) { - if (data->auth_flavourlen != 1) { - dprintk("%s: Invalid number of RPC auth flavours %d.\n", - __FUNCTION__, data->auth_flavourlen); - err = -EINVAL; - goto out_fail; - } - if (copy_from_user(&authflavour, data->auth_flavours, sizeof(authflavour))) { - err = -EFAULT; - goto out_fail; - } - } - - server->client = nfs4_create_client(server, data->timeo, data->retrans, - data->proto, authflavour); - if (IS_ERR(server->client)) { - err = PTR_ERR(server->client); - dprintk("%s: cannot create RPC client. Error = %d\n", - __FUNCTION__, err); - goto out_fail; - } - sb->s_time_gran = 1; - sb->s_op = &nfs4_sops; - err = nfs_sb_init(sb, authflavour); - - out_fail: - return err; + nfs_initialise_sb(sb); } -static int nfs4_compare_super(struct super_block *sb, void *data) -{ - struct nfs_server *server = data; - struct nfs_server *old = NFS_SB(sb); - - if (strcmp(server->hostname, old->hostname) != 0) - return 0; - if (strcmp(server->mnt_path, old->mnt_path) != 0) - return 0; - return 1; -} - -static void * -nfs_copy_user_string(char *dst, struct nfs_string *src, int maxlen) +static void *nfs_copy_user_string(char *dst, struct nfs_string *src, int maxlen) { void *p = NULL; @@ -1157,14 +791,22 @@ nfs_copy_user_string(char *dst, struct nfs_string *src, int maxlen) return dst; } +/* + * Get the superblock for an NFS4 mountpoint + */ static int nfs4_get_sb(struct file_system_type *fs_type, int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt) { - int error; - struct nfs_server *server; - struct super_block *s; struct nfs4_mount_data *data = raw_data; + struct super_block *s; + struct nfs_server *server; + struct sockaddr_in addr; + rpc_authflavor_t authflavour; + struct nfs_fh mntfh; + struct dentry *mntroot; + char *mntpath = NULL, *hostname = NULL, ip_addr[16]; void *p; + int error; if (data == NULL) { dprintk("%s: missing data argument\n", __FUNCTION__); @@ -1175,75 +817,107 @@ static int nfs4_get_sb(struct file_system_type *fs_type, return -EINVAL; } - server = kzalloc(sizeof(struct nfs_server), GFP_KERNEL); - if (!server) - return -ENOMEM; - /* Zero out the NFS state stuff */ - init_nfsv4_state(server); - server->client = server->client_acl = ERR_PTR(-EINVAL); + /* We now require that the mount process passes the remote address */ + if (data->host_addrlen != sizeof(addr)) + return -EINVAL; + + if (copy_from_user(&addr, data->host_addr, sizeof(addr))) + return -EFAULT; + + if (addr.sin_family != AF_INET || + addr.sin_addr.s_addr == INADDR_ANY + ) { + dprintk("%s: mount program didn't pass remote IP address!\n", + __FUNCTION__); + return -EINVAL; + } + + /* Grab the authentication type */ + authflavour = RPC_AUTH_UNIX; + if (data->auth_flavourlen != 0) { + if (data->auth_flavourlen != 1) { + dprintk("%s: Invalid number of RPC auth flavours %d.\n", + __FUNCTION__, data->auth_flavourlen); + error = -EINVAL; + goto out_err_noserver; + } + + if (copy_from_user(&authflavour, data->auth_flavours, + sizeof(authflavour))) { + error = -EFAULT; + goto out_err_noserver; + } + } p = nfs_copy_user_string(NULL, &data->hostname, 256); if (IS_ERR(p)) goto out_err; - server->hostname = p; + hostname = p; p = nfs_copy_user_string(NULL, &data->mnt_path, 1024); if (IS_ERR(p)) goto out_err; - server->mnt_path = p; + mntpath = p; - p = nfs_copy_user_string(server->ip_addr, &data->client_addr, - sizeof(server->ip_addr) - 1); + dprintk("MNTPATH: %s\n", mntpath); + + p = nfs_copy_user_string(ip_addr, &data->client_addr, + sizeof(ip_addr) - 1); if (IS_ERR(p)) goto out_err; - /* We now require that the mount process passes the remote address */ - if (data->host_addrlen != sizeof(server->addr)) { - error = -EINVAL; - goto out_free; - } - if (copy_from_user(&server->addr, data->host_addr, sizeof(server->addr))) { - error = -EFAULT; - goto out_free; - } - if (server->addr.sin_family != AF_INET || - server->addr.sin_addr.s_addr == INADDR_ANY) { - dprintk("%s: mount program didn't pass remote IP address!\n", - __FUNCTION__); - error = -EINVAL; - goto out_free; + /* Get a volume representation */ + server = nfs4_create_server(data, hostname, &addr, mntpath, ip_addr, + authflavour, &mntfh); + if (IS_ERR(server)) { + error = PTR_ERR(server); + goto out_err_noserver; } - s = sget(fs_type, nfs4_compare_super, nfs_set_super, server); + /* Get a superblock - note that we may end up sharing one that already exists */ + s = sget(fs_type, nfs_compare_super, nfs_set_super, server); if (IS_ERR(s)) { error = PTR_ERR(s); goto out_free; } - if (s->s_root) { - kfree(server->mnt_path); - kfree(server->hostname); - kfree(server); - return simple_set_mnt(mnt, s); + if (!s->s_root) { + /* initial superblock/root creation */ + s->s_flags = flags; + + nfs4_fill_super(s); + } else { + nfs_free_server(server); } - s->s_flags = flags; - - error = nfs4_fill_super(s, data, flags & MS_SILENT ? 1 : 0); - if (error) { - up_write(&s->s_umount); - deactivate_super(s); - return error; + mntroot = nfs4_get_root(s, &mntfh); + if (IS_ERR(mntroot)) { + error = PTR_ERR(mntroot); + goto error_splat_super; } + s->s_flags |= MS_ACTIVE; - return simple_set_mnt(mnt, s); + mnt->mnt_sb = s; + mnt->mnt_root = mntroot; + kfree(mntpath); + kfree(hostname); + return 0; + out_err: error = PTR_ERR(p); + goto out_err_noserver; + out_free: - kfree(server->mnt_path); - kfree(server->hostname); - kfree(server); + nfs_free_server(server); +out_err_noserver: + kfree(mntpath); + kfree(hostname); return error; + +error_splat_super: + up_write(&s->s_umount); + deactivate_super(s); + goto out_err_noserver; } static void nfs4_kill_super(struct super_block *sb) @@ -1254,133 +928,140 @@ static void nfs4_kill_super(struct super_block *sb) kill_anon_super(sb); nfs4_renewd_prepare_shutdown(server); - - if (server->client != NULL && !IS_ERR(server->client)) - rpc_shutdown_client(server->client); - - destroy_nfsv4_state(server); - - nfs_free_iostats(server->io_stats); - kfree(server->hostname); - kfree(server); - nfs_release_automount_timer(); + nfs_free_server(server); } /* - * Constructs the SERVER-side path + * Clone an NFS4 server record on xdev traversal (FSID-change) */ -static inline char *nfs4_dup_path(const struct dentry *dentry) -{ - char *page = (char *) __get_free_page(GFP_USER); - char *path; - - path = nfs4_path(dentry, page, PAGE_SIZE); - if (!IS_ERR(path)) { - int len = PAGE_SIZE + page - path; - char *tmp = path; - - path = kmalloc(len, GFP_KERNEL); - if (path) - memcpy(path, tmp, len); - else - path = ERR_PTR(-ENOMEM); - } - free_page((unsigned long)page); - return path; -} - -static struct super_block *nfs4_clone_sb(struct nfs_server *server, struct nfs_clone_mount *data) -{ - const struct dentry *dentry = data->dentry; - struct nfs_client *clp = server->nfs_client; - struct super_block *sb; - - server->fsid = data->fattr->fsid; - nfs_copy_fh(&server->fh, data->fh); - server->mnt_path = nfs4_dup_path(dentry); - if (IS_ERR(server->mnt_path)) { - sb = (struct super_block *)server->mnt_path; - goto err; - } - sb = sget(&nfs4_fs_type, nfs4_compare_super, nfs_set_super, server); - if (IS_ERR(sb) || sb->s_root) - goto free_path; - nfs4_server_capabilities(server, &server->fh); - - down_write(&clp->cl_sem); - list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks); - up_write(&clp->cl_sem); - return sb; -free_path: - kfree(server->mnt_path); -err: - server->mnt_path = NULL; - return sb; -} - -static int nfs_clone_nfs4_sb(struct file_system_type *fs_type, - int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt) +static int nfs4_xdev_get_sb(struct file_system_type *fs_type, int flags, + const char *dev_name, void *raw_data, + struct vfsmount *mnt) { struct nfs_clone_mount *data = raw_data; - return nfs_clone_generic_sb(data, nfs4_clone_sb, nfs_clone_server, mnt); + struct super_block *s; + struct nfs_server *server; + struct dentry *mntroot; + int error; + + dprintk("--> nfs4_xdev_get_sb()\n"); + + /* create a new volume representation */ + server = nfs_clone_server(NFS_SB(data->sb), data->fh, data->fattr); + if (IS_ERR(server)) { + error = PTR_ERR(server); + goto out_err_noserver; + } + + /* Get a superblock - note that we may end up sharing one that already exists */ + s = sget(&nfs_fs_type, nfs_compare_super, nfs_set_super, server); + if (IS_ERR(s)) { + error = PTR_ERR(s); + goto out_err_nosb; + } + + if (s->s_fs_info != server) { + nfs_free_server(server); + server = NULL; + } + + if (!s->s_root) { + /* initial superblock/root creation */ + s->s_flags = flags; + nfs4_clone_super(s, data->sb); + } + + mntroot = nfs4_get_root(s, data->fh); + if (IS_ERR(mntroot)) { + error = PTR_ERR(mntroot); + goto error_splat_super; + } + + s->s_flags |= MS_ACTIVE; + mnt->mnt_sb = s; + mnt->mnt_root = mntroot; + + dprintk("<-- nfs4_xdev_get_sb() = 0\n"); + return 0; + +out_err_nosb: + nfs_free_server(server); +out_err_noserver: + dprintk("<-- nfs4_xdev_get_sb() = %d [error]\n", error); + return error; + +error_splat_super: + up_write(&s->s_umount); + deactivate_super(s); + dprintk("<-- nfs4_xdev_get_sb() = %d [splat]\n", error); + return error; } -static struct super_block *nfs4_referral_sb(struct nfs_server *server, struct nfs_clone_mount *data) -{ - struct super_block *sb = ERR_PTR(-ENOMEM); - int len; - - len = strlen(data->mnt_path) + 1; - server->mnt_path = kmalloc(len, GFP_KERNEL); - if (server->mnt_path == NULL) - goto err; - memcpy(server->mnt_path, data->mnt_path, len); - memcpy(&server->addr, data->addr, sizeof(struct sockaddr_in)); - - sb = sget(&nfs4_fs_type, nfs4_compare_super, nfs_set_super, server); - if (IS_ERR(sb) || sb->s_root) - goto free_path; - return sb; -free_path: - kfree(server->mnt_path); -err: - server->mnt_path = NULL; - return sb; -} - -static struct nfs_server *nfs4_referral_server(struct super_block *sb, struct nfs_clone_mount *data) -{ - struct nfs_server *server = NFS_SB(sb); - int proto, timeo, retrans; - void *err; - - proto = IPPROTO_TCP; - /* Since we are following a referral and there may be alternatives, - set the timeouts and retries to low values */ - timeo = 2; - retrans = 1; - - nfs_put_client(server->nfs_client); - server->nfs_client = NULL; - server->client = nfs4_create_client(server, timeo, retrans, proto, - data->authflavor); - if (IS_ERR((err = server->client))) - goto out_err; - - sb->s_time_gran = 1; - sb->s_op = &nfs4_sops; - err = ERR_PTR(nfs_sb_init(sb, data->authflavor)); - if (!IS_ERR(err)) - return server; -out_err: - return (struct nfs_server *)err; -} - -static int nfs_referral_nfs4_sb(struct file_system_type *fs_type, - int flags, const char *dev_name, void *raw_data, struct vfsmount *mnt) +/* + * Create an NFS4 server record on referral traversal + */ +static int nfs4_referral_get_sb(struct file_system_type *fs_type, int flags, + const char *dev_name, void *raw_data, + struct vfsmount *mnt) { struct nfs_clone_mount *data = raw_data; - return nfs_clone_generic_sb(data, nfs4_referral_sb, nfs4_referral_server, mnt); + struct super_block *s; + struct nfs_server *server; + struct dentry *mntroot; + struct nfs_fh mntfh; + int error; + + dprintk("--> nfs4_referral_get_sb()\n"); + + /* create a new volume representation */ + server = nfs4_create_referral_server(data, &mntfh); + if (IS_ERR(server)) { + error = PTR_ERR(server); + goto out_err_noserver; + } + + /* Get a superblock - note that we may end up sharing one that already exists */ + s = sget(&nfs_fs_type, nfs_compare_super, nfs_set_super, server); + if (IS_ERR(s)) { + error = PTR_ERR(s); + goto out_err_nosb; + } + + if (s->s_fs_info != server) { + nfs_free_server(server); + server = NULL; + } + + if (!s->s_root) { + /* initial superblock/root creation */ + s->s_flags = flags; + nfs4_fill_super(s); + } + + mntroot = nfs4_get_root(s, data->fh); + if (IS_ERR(mntroot)) { + error = PTR_ERR(mntroot); + goto error_splat_super; + } + + s->s_flags |= MS_ACTIVE; + mnt->mnt_sb = s; + mnt->mnt_root = mntroot; + + dprintk("<-- nfs4_referral_get_sb() = 0\n"); + return 0; + +out_err_nosb: + nfs_free_server(server); +out_err_noserver: + dprintk("<-- nfs4_referral_get_sb() = %d [error]\n", error); + return error; + +error_splat_super: + up_write(&s->s_umount); + deactivate_super(s); + dprintk("<-- nfs4_referral_get_sb() = %d [splat]\n", error); + return error; } -#endif +#endif /* CONFIG_NFS_V4 */ diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 7084ac9a6455..453d44666ea5 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1273,7 +1273,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data) if (time_before(complain, jiffies)) { dprintk("NFS: faulty NFS server %s:" " (committed = %d) != (stable = %d)\n", - NFS_SERVER(data->inode)->hostname, + NFS_SERVER(data->inode)->nfs_client->cl_hostname, resp->verf->committed, argp->stable); complain = jiffies + 300 * HZ; } diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index d404ceca9168..6d0be0efd1b5 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -51,7 +51,6 @@ struct nfs_client { unsigned long cl_lease_time; unsigned long cl_last_renewal; struct work_struct cl_renewd; - struct work_struct cl_recoverd; struct rpc_wait_queue cl_rpcwaitq; @@ -74,6 +73,10 @@ struct nfs_client { */ struct nfs_server { struct nfs_client * nfs_client; /* shared client and NFS4 state */ + struct list_head client_link; /* List of other nfs_server structs + * that share the same client + */ + struct list_head master_link; /* link in master servers list */ struct rpc_clnt * client; /* RPC client handle */ struct rpc_clnt * client_acl; /* ACL RPC client handle */ struct nfs_iostats * io_stats; /* I/O statistics */ @@ -92,20 +95,13 @@ struct nfs_server { unsigned int acdirmin; unsigned int acdirmax; unsigned int namelen; - char * hostname; /* remote hostname */ - struct nfs_fh fh; - struct sockaddr_in addr; + struct nfs_fsid fsid; + __u64 maxfilesize; /* maximum file size */ unsigned long mount_time; /* when this fs was mounted */ + dev_t s_dev; /* superblock dev numbers */ + #ifdef CONFIG_NFS_V4 - /* Our own IP address, as a null-terminated string. - * This is used to generate the clientid, and the callback address. - */ - char ip_addr[16]; - char * mnt_path; - struct list_head nfs4_siblings; /* List of other nfs_server structs - * that share the same clientid - */ u32 attr_bitmask[2];/* V4 bitmask representing the set of attributes supported on this filesystem */ @@ -113,6 +109,7 @@ struct nfs_server { that are supported on this filesystem */ #endif + void (*destroy)(struct nfs_server *); }; /* Server capabilities */ From 6aaca566503296a73f956908ec98173946134fe2 Mon Sep 17 00:00:00 2001 From: David Howells Date: Tue, 22 Aug 2006 20:06:13 -0400 Subject: [PATCH 22/74] NFS: Add server and volume lists to /proc Make two new proc files available: /proc/fs/nfsfs/servers /proc/fs/nfsfs/volumes The first lists the servers with which we are currently dealing (struct nfs_client), and the second lists the volumes we have on those servers (struct nfs_server). Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 284 ++++++++++++++++++++++++++++++++++++++++++++++ fs/nfs/inode.c | 7 ++ fs/nfs/internal.h | 12 ++ 3 files changed, 303 insertions(+) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index dafba608c0a0..27f64781444a 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -1148,3 +1148,287 @@ struct nfs_server *nfs_clone_server(struct nfs_server *source, dprintk("<-- nfs_clone_server() = error %d\n", error); return ERR_PTR(error); } + +#ifdef CONFIG_PROC_FS +static struct proc_dir_entry *proc_fs_nfs; + +static int nfs_server_list_open(struct inode *inode, struct file *file); +static void *nfs_server_list_start(struct seq_file *p, loff_t *pos); +static void *nfs_server_list_next(struct seq_file *p, void *v, loff_t *pos); +static void nfs_server_list_stop(struct seq_file *p, void *v); +static int nfs_server_list_show(struct seq_file *m, void *v); + +static struct seq_operations nfs_server_list_ops = { + .start = nfs_server_list_start, + .next = nfs_server_list_next, + .stop = nfs_server_list_stop, + .show = nfs_server_list_show, +}; + +static struct file_operations nfs_server_list_fops = { + .open = nfs_server_list_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +static int nfs_volume_list_open(struct inode *inode, struct file *file); +static void *nfs_volume_list_start(struct seq_file *p, loff_t *pos); +static void *nfs_volume_list_next(struct seq_file *p, void *v, loff_t *pos); +static void nfs_volume_list_stop(struct seq_file *p, void *v); +static int nfs_volume_list_show(struct seq_file *m, void *v); + +static struct seq_operations nfs_volume_list_ops = { + .start = nfs_volume_list_start, + .next = nfs_volume_list_next, + .stop = nfs_volume_list_stop, + .show = nfs_volume_list_show, +}; + +static struct file_operations nfs_volume_list_fops = { + .open = nfs_volume_list_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +/* + * open "/proc/fs/nfsfs/servers" which provides a summary of servers with which + * we're dealing + */ +static int nfs_server_list_open(struct inode *inode, struct file *file) +{ + struct seq_file *m; + int ret; + + ret = seq_open(file, &nfs_server_list_ops); + if (ret < 0) + return ret; + + m = file->private_data; + m->private = PDE(inode)->data; + + return 0; +} + +/* + * set up the iterator to start reading from the server list and return the first item + */ +static void *nfs_server_list_start(struct seq_file *m, loff_t *_pos) +{ + struct list_head *_p; + loff_t pos = *_pos; + + /* lock the list against modification */ + spin_lock(&nfs_client_lock); + + /* allow for the header line */ + if (!pos) + return SEQ_START_TOKEN; + pos--; + + /* find the n'th element in the list */ + list_for_each(_p, &nfs_client_list) + if (!pos--) + break; + + return _p != &nfs_client_list ? _p : NULL; +} + +/* + * move to next server + */ +static void *nfs_server_list_next(struct seq_file *p, void *v, loff_t *pos) +{ + struct list_head *_p; + + (*pos)++; + + _p = v; + _p = (v == SEQ_START_TOKEN) ? nfs_client_list.next : _p->next; + + return _p != &nfs_client_list ? _p : NULL; +} + +/* + * clean up after reading from the transports list + */ +static void nfs_server_list_stop(struct seq_file *p, void *v) +{ + spin_unlock(&nfs_client_lock); +} + +/* + * display a header line followed by a load of call lines + */ +static int nfs_server_list_show(struct seq_file *m, void *v) +{ + struct nfs_client *clp; + + /* display header on line 1 */ + if (v == SEQ_START_TOKEN) { + seq_puts(m, "NV SERVER PORT USE HOSTNAME\n"); + return 0; + } + + /* display one transport per line on subsequent lines */ + clp = list_entry(v, struct nfs_client, cl_share_link); + + seq_printf(m, "v%d %02x%02x%02x%02x %4hx %3d %s\n", + clp->cl_nfsversion, + NIPQUAD(clp->cl_addr.sin_addr), + ntohs(clp->cl_addr.sin_port), + atomic_read(&clp->cl_count), + clp->cl_hostname); + + return 0; +} + +/* + * open "/proc/fs/nfsfs/volumes" which provides a summary of extant volumes + */ +static int nfs_volume_list_open(struct inode *inode, struct file *file) +{ + struct seq_file *m; + int ret; + + ret = seq_open(file, &nfs_volume_list_ops); + if (ret < 0) + return ret; + + m = file->private_data; + m->private = PDE(inode)->data; + + return 0; +} + +/* + * set up the iterator to start reading from the volume list and return the first item + */ +static void *nfs_volume_list_start(struct seq_file *m, loff_t *_pos) +{ + struct list_head *_p; + loff_t pos = *_pos; + + /* lock the list against modification */ + spin_lock(&nfs_client_lock); + + /* allow for the header line */ + if (!pos) + return SEQ_START_TOKEN; + pos--; + + /* find the n'th element in the list */ + list_for_each(_p, &nfs_volume_list) + if (!pos--) + break; + + return _p != &nfs_volume_list ? _p : NULL; +} + +/* + * move to next volume + */ +static void *nfs_volume_list_next(struct seq_file *p, void *v, loff_t *pos) +{ + struct list_head *_p; + + (*pos)++; + + _p = v; + _p = (v == SEQ_START_TOKEN) ? nfs_volume_list.next : _p->next; + + return _p != &nfs_volume_list ? _p : NULL; +} + +/* + * clean up after reading from the transports list + */ +static void nfs_volume_list_stop(struct seq_file *p, void *v) +{ + spin_unlock(&nfs_client_lock); +} + +/* + * display a header line followed by a load of call lines + */ +static int nfs_volume_list_show(struct seq_file *m, void *v) +{ + struct nfs_server *server; + struct nfs_client *clp; + char dev[8], fsid[17]; + + /* display header on line 1 */ + if (v == SEQ_START_TOKEN) { + seq_puts(m, "NV SERVER PORT DEV FSID\n"); + return 0; + } + /* display one transport per line on subsequent lines */ + server = list_entry(v, struct nfs_server, master_link); + clp = server->nfs_client; + + snprintf(dev, 8, "%u:%u", + MAJOR(server->s_dev), MINOR(server->s_dev)); + + snprintf(fsid, 17, "%llx:%llx", + server->fsid.major, server->fsid.minor); + + seq_printf(m, "v%d %02x%02x%02x%02x %4hx %-7s %-17s\n", + clp->cl_nfsversion, + NIPQUAD(clp->cl_addr.sin_addr), + ntohs(clp->cl_addr.sin_port), + dev, + fsid); + + return 0; +} + +/* + * initialise the /proc/fs/nfsfs/ directory + */ +int __init nfs_fs_proc_init(void) +{ + struct proc_dir_entry *p; + + proc_fs_nfs = proc_mkdir("nfsfs", proc_root_fs); + if (!proc_fs_nfs) + goto error_0; + + proc_fs_nfs->owner = THIS_MODULE; + + /* a file of servers with which we're dealing */ + p = create_proc_entry("servers", S_IFREG|S_IRUGO, proc_fs_nfs); + if (!p) + goto error_1; + + p->proc_fops = &nfs_server_list_fops; + p->owner = THIS_MODULE; + + /* a file of volumes that we have mounted */ + p = create_proc_entry("volumes", S_IFREG|S_IRUGO, proc_fs_nfs); + if (!p) + goto error_2; + + p->proc_fops = &nfs_volume_list_fops; + p->owner = THIS_MODULE; + return 0; + +error_2: + remove_proc_entry("servers", proc_fs_nfs); +error_1: + remove_proc_entry("nfsfs", proc_root_fs); +error_0: + return -ENOMEM; +} + +/* + * clean up the /proc/fs/nfsfs/ directory + */ +void nfs_fs_proc_exit(void) +{ + remove_proc_entry("volumes", proc_fs_nfs); + remove_proc_entry("servers", proc_fs_nfs); + remove_proc_entry("nfsfs", proc_root_fs); +} + +#endif /* CONFIG_PROC_FS */ diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index a547c58a83e6..cb5c65f0bc12 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1141,6 +1141,10 @@ static int __init init_nfs_fs(void) { int err; + err = nfs_fs_proc_init(); + if (err) + goto out5; + err = nfs_init_nfspagecache(); if (err) goto out4; @@ -1181,6 +1185,8 @@ static int __init init_nfs_fs(void) out3: nfs_destroy_nfspagecache(); out4: + nfs_fs_proc_exit(); +out5: return err; } @@ -1195,6 +1201,7 @@ static void __exit exit_nfs_fs(void) rpc_proc_unregister("nfs"); #endif unregister_nfs_fs(); + nfs_fs_proc_exit(); } /* Not quite true; I just maintain it */ diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index e73ba4f1052a..bea0b016bd70 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -47,6 +47,18 @@ extern void nfs_free_server(struct nfs_server *server); extern struct nfs_server *nfs_clone_server(struct nfs_server *, struct nfs_fh *, struct nfs_fattr *); +#ifdef CONFIG_PROC_FS +extern int __init nfs_fs_proc_init(void); +extern void nfs_fs_proc_exit(void); +#else +static inline int nfs_fs_proc_init(void) +{ + return 0; +} +static inline void nfs_fs_proc_exit(void) +{ +} +#endif /* nfs4namespace.c */ #ifdef CONFIG_NFS_V4 From 27ba851244f627a302d0fc6469d1ad413fc34fcb Mon Sep 17 00:00:00 2001 From: David Howells Date: Sun, 30 Jul 2006 14:40:56 -0400 Subject: [PATCH 23/74] NFS: Fix error handling Fix an error handling problem: nfs_put_client() can be given a NULL pointer if nfs_free_server() is asked to destroy a partially initialised record. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 27f64781444a..700bd5801223 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -208,6 +208,9 @@ static void nfs_free_client(struct nfs_client *clp) */ void nfs_put_client(struct nfs_client *clp) { + if (!clp) + return; + dprintk("--> nfs_put_client({%d})\n", atomic_read(&clp->cl_count)); if (atomic_dec_and_lock(&clp->cl_count, &nfs_client_lock)) { From 738a35195941ecf604d3070e2a053e1df3de350b Mon Sep 17 00:00:00 2001 From: David Howells Date: Sun, 30 Jul 2006 14:58:27 -0400 Subject: [PATCH 24/74] NFS: Secure the roots of the NFS subtrees in a shared superblock Invoke security_d_instantiate() on root dentries after allocating them with dentry_alloc_anon(). Normally dentry_alloc_root() would do that, but we don't call that as we don't want to assign a name to the root dentry at this point (we may discover the real name later). Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/getroot.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/fs/nfs/getroot.c b/fs/nfs/getroot.c index 977e59088eeb..76b08ae9ed82 100644 --- a/fs/nfs/getroot.c +++ b/fs/nfs/getroot.c @@ -33,6 +33,7 @@ #include #include #include +#include #include #include @@ -109,6 +110,8 @@ struct dentry *nfs_get_root(struct super_block *sb, struct nfs_fh *mntfh) return ERR_PTR(-ENOMEM); } + security_d_instantiate(mntroot, inode); + if (!mntroot->d_op) mntroot->d_op = server->nfs_client->rpc_ops->dentry_ops; @@ -296,6 +299,8 @@ struct dentry *nfs4_get_root(struct super_block *sb, struct nfs_fh *mntfh) return ERR_PTR(-ENOMEM); } + security_d_instantiate(mntroot, inode); + if (!mntroot->d_op) mntroot->d_op = server->nfs_client->rpc_ops->dentry_ops; From 36b15c54cd0d6f707a3ac03e4a2a60bb530a95b9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 22 Aug 2006 20:06:14 -0400 Subject: [PATCH 25/74] NFS: Ensure NFSv2/v3 mounts respect the NFS_MOUNT_SECFLAVOUR flag Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 867b5dcd3a40..97cfb143e09f 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -471,12 +471,14 @@ static int nfs_validate_mount_data(struct nfs_mount_data *data, data->version); return -EINVAL; } - /* Fill in pseudoflavor for mount version < 5 */ - data->pseudoflavor = RPC_AUTH_UNIX; case 5: memset(data->context, 0, sizeof(data->context)); } + /* Set the pseudoflavor */ + if (!(data->flags & NFS_MOUNT_SECFLAVOUR)) + data->pseudoflavor = RPC_AUTH_UNIX; + #ifndef CONFIG_NFS_V3 /* If NFSv3 is not compiled in, return -EPROTONOSUPPORT */ if (data->flags & NFS_MOUNT_VER3) { From 9c5bf38d85a31b946664bcc21078ef5bb10672f7 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 22 Aug 2006 20:06:14 -0400 Subject: [PATCH 26/74] NFS: Fix nfs_alloc_client() The scheme to indicate which services have been started up appears to be seriously broken. Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 700bd5801223..471d975e63c3 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -113,9 +113,9 @@ static struct nfs_client *nfs_alloc_client(const char *hostname, if (error < 0) { dprintk("%s: couldn't start rpciod! Error = %d\n", __FUNCTION__, error); - __set_bit(NFS_CS_RPCIOD, &clp->cl_res_state); goto error_1; } + __set_bit(NFS_CS_RPCIOD, &clp->cl_res_state); if (nfsversion == 4) { if (nfs_callback_up() < 0) @@ -153,8 +153,8 @@ static struct nfs_client *nfs_alloc_client(const char *hostname, return clp; error_3: - nfs_callback_down(); - __clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state); + if (__test_and_clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state)) + nfs_callback_down(); error_2: rpciod_down(); __clear_bit(NFS_CS_RPCIOD, &clp->cl_res_state); @@ -195,7 +195,7 @@ static void nfs_free_client(struct nfs_client *clp) nfs_callback_down(); if (__test_and_clear_bit(NFS_CS_RPCIOD, &clp->cl_res_state)) - rpciod_down(); + rpciod_down(); kfree(clp->cl_hostname); kfree(clp); @@ -881,9 +881,9 @@ static int nfs4_init_client(struct nfs_client *clp, if (error < 0) { dprintk("%s: failed to create idmapper. Error = %d\n", __FUNCTION__, error); - __set_bit(NFS_CS_IDMAP, &clp->cl_res_state); goto error; } + __set_bit(NFS_CS_IDMAP, &clp->cl_res_state); nfs_mark_client_ready(clp, NFS_CS_READY); return 0; From ec739ef03dc926d05051c8c5838971445504470a Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:15 -0400 Subject: [PATCH 27/74] SUNRPC: Create a helper to tell whether a transport is bound Hide the contents and format of xprt->addr by eliminating direct uses of the xprt->addr.sin_port field. This change is required to support alternate RPC host address formats (eg IPv6). Test-plan: Destructive testing (unplugging the network temporarily). Repeated runs of Connectathon locking suite with UDP and TCP. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprt.h | 16 ++++++++++++++++ net/sunrpc/clnt.c | 10 +++++----- net/sunrpc/pmap_clnt.c | 5 ++++- net/sunrpc/xprt.c | 2 +- net/sunrpc/xprtsock.c | 14 ++++++++++---- 5 files changed, 36 insertions(+), 11 deletions(-) diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 3a0cca255b76..a71106723d71 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -269,6 +269,7 @@ int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to); #define XPRT_CONNECTED (1) #define XPRT_CONNECTING (2) #define XPRT_CLOSE_WAIT (3) +#define XPRT_BOUND (4) static inline void xprt_set_connected(struct rpc_xprt *xprt) { @@ -312,6 +313,21 @@ static inline int xprt_test_and_set_connecting(struct rpc_xprt *xprt) return test_and_set_bit(XPRT_CONNECTING, &xprt->state); } +static inline void xprt_set_bound(struct rpc_xprt *xprt) +{ + test_and_set_bit(XPRT_BOUND, &xprt->state); +} + +static inline int xprt_bound(struct rpc_xprt *xprt) +{ + return test_bit(XPRT_BOUND, &xprt->state); +} + +static inline void xprt_clear_bound(struct rpc_xprt *xprt) +{ + clear_bit(XPRT_BOUND, &xprt->state); +} + #endif /* __KERNEL__*/ #endif /* _LINUX_SUNRPC_XPRT_H */ diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 3e19d321067a..0b8d03d08561 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -148,7 +148,6 @@ rpc_new_client(struct rpc_xprt *xprt, char *servname, clnt->cl_maxproc = version->nrprocs; clnt->cl_protname = program->name; clnt->cl_pmap = &clnt->cl_pmap_default; - clnt->cl_port = xprt->addr.sin_port; clnt->cl_prog = program->number; clnt->cl_vers = version->number; clnt->cl_prot = xprt->prot; @@ -156,7 +155,7 @@ rpc_new_client(struct rpc_xprt *xprt, char *servname, clnt->cl_metrics = rpc_alloc_iostats(clnt); rpc_init_wait_queue(&clnt->cl_pmap_default.pm_bindwait, "bindwait"); - if (!clnt->cl_port) + if (!xprt_bound(clnt->cl_xprt)) clnt->cl_autobind = 1; clnt->cl_rtt = &clnt->cl_rtt_default; @@ -570,7 +569,7 @@ EXPORT_SYMBOL(rpc_max_payload); void rpc_force_rebind(struct rpc_clnt *clnt) { if (clnt->cl_autobind) - clnt->cl_port = 0; + xprt_clear_bound(clnt->cl_xprt); } EXPORT_SYMBOL(rpc_force_rebind); @@ -782,14 +781,15 @@ static void call_bind(struct rpc_task *task) { struct rpc_clnt *clnt = task->tk_client; + struct rpc_xprt *xprt = task->tk_xprt; dprintk("RPC: %4d call_bind (status %d)\n", task->tk_pid, task->tk_status); task->tk_action = call_connect; - if (!clnt->cl_port) { + if (!xprt_bound(xprt)) { task->tk_action = call_bind_status; - task->tk_timeout = task->tk_xprt->bind_timeout; + task->tk_timeout = xprt->bind_timeout; rpc_getport(task, clnt); } } diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index 623180f224c9..209ffdfee10b 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -142,15 +142,17 @@ pmap_getport_done(struct rpc_task *task) dprintk("RPC: %4d pmap_getport_done(status %d, port %d)\n", task->tk_pid, task->tk_status, clnt->cl_port); - xprt->ops->set_port(xprt, 0); if (task->tk_status < 0) { /* Make the calling task exit with an error */ + xprt->ops->set_port(xprt, 0); task->tk_action = rpc_exit_task; } else if (clnt->cl_port == 0) { /* Program not registered */ + xprt->ops->set_port(xprt, 0); rpc_exit(task, -EACCES); } else { xprt->ops->set_port(xprt, clnt->cl_port); + xprt_set_bound(xprt); clnt->cl_port = htons(clnt->cl_port); } spin_lock(&pmap_lock); @@ -218,6 +220,7 @@ pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileg if (IS_ERR(xprt)) return (struct rpc_clnt *)xprt; xprt->ops->set_port(xprt, RPC_PMAP_PORT); + xprt_set_bound(xprt); if (!privileged) xprt->resvport = 0; diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index e8c2bc4977f3..e239ef985ef7 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -534,7 +534,7 @@ void xprt_connect(struct rpc_task *task) dprintk("RPC: %4d xprt_connect xprt %p %s connected\n", task->tk_pid, xprt, (xprt_connected(xprt) ? "is" : "is not")); - if (!xprt->addr.sin_port) { + if (!xprt_bound(xprt)) { task->tk_status = -EIO; return; } diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 441bd53f5eca..123ac1e5ba15 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -1016,7 +1016,7 @@ static void xs_udp_connect_worker(void *args) struct socket *sock = xprt->sock; int err, status = -EIO; - if (xprt->shutdown || xprt->addr.sin_port == 0) + if (xprt->shutdown || !xprt_bound(xprt)) goto out; dprintk("RPC: xs_udp_connect_worker for xprt %p\n", xprt); @@ -1099,7 +1099,7 @@ static void xs_tcp_connect_worker(void *args) struct socket *sock = xprt->sock; int err, status = -EIO; - if (xprt->shutdown || xprt->addr.sin_port == 0) + if (xprt->shutdown || !xprt_bound(xprt)) goto out; dprintk("RPC: xs_tcp_connect_worker for xprt %p\n", xprt); @@ -1307,8 +1307,11 @@ int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to) if (xprt->slot == NULL) return -ENOMEM; - xprt->prot = IPPROTO_UDP; + if (ntohs(xprt->addr.sin_port) != 0) + xprt_set_bound(xprt); xprt->port = xs_get_random_port(); + + xprt->prot = IPPROTO_UDP; xprt->tsh_size = 0; xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0; /* XXX: header size can vary due to auth type, IPv6, etc. */ @@ -1348,8 +1351,11 @@ int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to) if (xprt->slot == NULL) return -ENOMEM; - xprt->prot = IPPROTO_TCP; + if (ntohs(xprt->addr.sin_port) != 0) + xprt_set_bound(xprt); xprt->port = xs_get_random_port(); + + xprt->prot = IPPROTO_TCP; xprt->tsh_size = sizeof(rpc_fraghdr) / sizeof(u32); xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0; xprt->max_payload = RPC_MAX_FRAGMENT_SIZE; From 4a68179d38874c37be2802442a71b847f5d1a2a9 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:15 -0400 Subject: [PATCH 28/74] SUNRPC: Make RPC portmapper use per-transport storage Move connection and bind state that was maintained in the rpc_clnt structure to the rpc_xprt structure. This will allow the creation of a clean API for plugging in different types of bind mechanisms. This brings improvements such as the elimination of a single spin lock to control serialization for all in-kernel RPC binding. A set of per-xprt bitops is used to serialize tasks during RPC binding, just like it now works for making RPC transport connections. Test-plan: Destructive testing (unplugging the network temporarily). Connectathon with UDP and TCP. NFSv2/3 and NFSv4 mounting should be carefully checked. Probably need to rig a server where certain services aren't running, or that returns an error for some typical operation. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/clnt.h | 23 +----- include/linux/sunrpc/xprt.h | 14 ++++ net/sunrpc/clnt.c | 8 +- net/sunrpc/pmap_clnt.c | 158 ++++++++++++++++++++++++------------ net/sunrpc/xprt.c | 1 + 5 files changed, 123 insertions(+), 81 deletions(-) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 8fe9f35eba31..00e9dbaec9c5 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -18,18 +18,6 @@ #include #include -/* - * This defines an RPC port mapping - */ -struct rpc_portmap { - __u32 pm_prog; - __u32 pm_vers; - __u32 pm_prot; - __u16 pm_port; - unsigned char pm_binding : 1; /* doing a getport() */ - struct rpc_wait_queue pm_bindwait; /* waiting on getport() */ -}; - struct rpc_inode; /* @@ -40,7 +28,9 @@ struct rpc_clnt { atomic_t cl_users; /* number of references */ struct rpc_xprt * cl_xprt; /* transport */ struct rpc_procinfo * cl_procinfo; /* procedure info */ - u32 cl_maxproc; /* max procedure number */ + u32 cl_prog, /* RPC program number */ + cl_vers, /* RPC version number */ + cl_maxproc; /* max procedure number */ char * cl_server; /* server machine name */ char * cl_protname; /* protocol name */ @@ -55,7 +45,6 @@ struct rpc_clnt { cl_dead : 1;/* abandoned */ struct rpc_rtt * cl_rtt; /* RTO estimator data */ - struct rpc_portmap * cl_pmap; /* port mapping */ int cl_nodelen; /* nodename length */ char cl_nodename[UNX_MAXNODENAME]; @@ -64,14 +53,8 @@ struct rpc_clnt { struct dentry * cl_dentry; /* inode */ struct rpc_clnt * cl_parent; /* Points to parent of clones */ struct rpc_rtt cl_rtt_default; - struct rpc_portmap cl_pmap_default; char cl_inline_name[32]; }; -#define cl_timeout cl_xprt->timeout -#define cl_prog cl_pmap->pm_prog -#define cl_vers cl_pmap->pm_vers -#define cl_port cl_pmap->pm_port -#define cl_prot cl_pmap->pm_prot /* * General RPC program info diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index a71106723d71..4ce82616873d 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -138,6 +138,7 @@ struct rpc_xprt { unsigned int tsh_size; /* size of transport specific header */ + struct rpc_wait_queue binding; /* requests waiting on rpcbind */ struct rpc_wait_queue sending; /* requests waiting to send */ struct rpc_wait_queue resend; /* requests waiting to resend */ struct rpc_wait_queue pending; /* requests in flight */ @@ -270,6 +271,7 @@ int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to); #define XPRT_CONNECTING (2) #define XPRT_CLOSE_WAIT (3) #define XPRT_BOUND (4) +#define XPRT_BINDING (5) static inline void xprt_set_connected(struct rpc_xprt *xprt) { @@ -328,6 +330,18 @@ static inline void xprt_clear_bound(struct rpc_xprt *xprt) clear_bit(XPRT_BOUND, &xprt->state); } +static inline void xprt_clear_binding(struct rpc_xprt *xprt) +{ + smp_mb__before_clear_bit(); + clear_bit(XPRT_BINDING, &xprt->state); + smp_mb__after_clear_bit(); +} + +static inline int xprt_test_and_set_binding(struct rpc_xprt *xprt) +{ + return test_and_set_bit(XPRT_BINDING, &xprt->state); +} + #endif /* __KERNEL__*/ #endif /* _LINUX_SUNRPC_XPRT_H */ diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 0b8d03d08561..cee504162a3f 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -147,13 +147,10 @@ rpc_new_client(struct rpc_xprt *xprt, char *servname, clnt->cl_procinfo = version->procs; clnt->cl_maxproc = version->nrprocs; clnt->cl_protname = program->name; - clnt->cl_pmap = &clnt->cl_pmap_default; clnt->cl_prog = program->number; clnt->cl_vers = version->number; - clnt->cl_prot = xprt->prot; clnt->cl_stats = program->stats; clnt->cl_metrics = rpc_alloc_iostats(clnt); - rpc_init_wait_queue(&clnt->cl_pmap_default.pm_bindwait, "bindwait"); if (!xprt_bound(clnt->cl_xprt)) clnt->cl_autobind = 1; @@ -243,8 +240,6 @@ rpc_clone_client(struct rpc_clnt *clnt) atomic_set(&new->cl_users, 0); new->cl_parent = clnt; atomic_inc(&clnt->cl_count); - /* Duplicate portmapper */ - rpc_init_wait_queue(&new->cl_pmap_default.pm_bindwait, "bindwait"); /* Turn off autobind on clones */ new->cl_autobind = 0; new->cl_oneshot = 0; @@ -254,8 +249,7 @@ rpc_clone_client(struct rpc_clnt *clnt) rpc_init_rtt(&new->cl_rtt_default, clnt->cl_xprt->timeout.to_initval); if (new->cl_auth) atomic_inc(&new->cl_auth->au_count); - new->cl_pmap = &new->cl_pmap_default; - new->cl_metrics = rpc_alloc_iostats(clnt); + new->cl_metrics = rpc_alloc_iostats(clnt); return new; out_no_clnt: printk(KERN_INFO "RPC: out of memory in %s\n", __FUNCTION__); diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index 209ffdfee10b..59d542436ca9 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -24,11 +24,57 @@ #define PMAP_UNSET 2 #define PMAP_GETPORT 3 +struct portmap_args { + u32 pm_prog; + u32 pm_vers; + u32 pm_prot; + unsigned short pm_port; + struct rpc_task * pm_task; +}; + static struct rpc_procinfo pmap_procedures[]; static struct rpc_clnt * pmap_create(char *, struct sockaddr_in *, int, int); -static void pmap_getport_done(struct rpc_task *); +static void pmap_getport_done(struct rpc_task *, void *); static struct rpc_program pmap_program; -static DEFINE_SPINLOCK(pmap_lock); + +static void pmap_getport_prepare(struct rpc_task *task, void *calldata) +{ + struct portmap_args *map = calldata; + struct rpc_message msg = { + .rpc_proc = &pmap_procedures[PMAP_GETPORT], + .rpc_argp = map, + .rpc_resp = &map->pm_port, + }; + + rpc_call_setup(task, &msg, 0); +} + +static inline struct portmap_args *pmap_map_alloc(void) +{ + return kmalloc(sizeof(struct portmap_args), GFP_NOFS); +} + +static inline void pmap_map_free(struct portmap_args *map) +{ + kfree(map); +} + +static void pmap_map_release(void *data) +{ + pmap_map_free(data); +} + +static const struct rpc_call_ops pmap_getport_ops = { + .rpc_call_prepare = pmap_getport_prepare, + .rpc_call_done = pmap_getport_done, + .rpc_release = pmap_map_release, +}; + +static inline void pmap_wake_portmap_waiters(struct rpc_xprt *xprt) +{ + xprt_clear_binding(xprt); + rpc_wake_up(&xprt->binding); +} /* * Obtain the port for a given RPC service on a given host. This one can @@ -37,67 +83,71 @@ static DEFINE_SPINLOCK(pmap_lock); void rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt) { - struct rpc_portmap *map = clnt->cl_pmap; - struct sockaddr_in *sap = &clnt->cl_xprt->addr; - struct rpc_message msg = { - .rpc_proc = &pmap_procedures[PMAP_GETPORT], - .rpc_argp = map, - .rpc_resp = &clnt->cl_port, - .rpc_cred = NULL - }; + struct rpc_xprt *xprt = task->tk_xprt; + struct sockaddr_in *sap = &xprt->addr; + struct portmap_args *map; struct rpc_clnt *pmap_clnt; - struct rpc_task *child; + struct rpc_task *child; - dprintk("RPC: %4d rpc_getport(%s, %d, %d, %d)\n", + dprintk("RPC: %4d rpc_getport(%s, %u, %u, %d)\n", task->tk_pid, clnt->cl_server, - map->pm_prog, map->pm_vers, map->pm_prot); + clnt->cl_prog, clnt->cl_vers, xprt->prot); /* Autobind on cloned rpc clients is discouraged */ BUG_ON(clnt->cl_parent != clnt); - spin_lock(&pmap_lock); - if (map->pm_binding) { - rpc_sleep_on(&map->pm_bindwait, task, NULL, NULL); - spin_unlock(&pmap_lock); + if (xprt_test_and_set_binding(xprt)) { + task->tk_status = -EACCES; /* tell caller to check again */ + rpc_sleep_on(&xprt->binding, task, NULL, NULL); return; } - map->pm_binding = 1; - spin_unlock(&pmap_lock); + + /* Someone else may have bound if we slept */ + if (xprt_bound(xprt)) { + task->tk_status = 0; + goto bailout_nofree; + } + + map = pmap_map_alloc(); + if (!map) { + task->tk_status = -ENOMEM; + goto bailout_nofree; + } + map->pm_prog = clnt->cl_prog; + map->pm_vers = clnt->cl_vers; + map->pm_prot = xprt->prot; + map->pm_port = 0; + map->pm_task = task; pmap_clnt = pmap_create(clnt->cl_server, sap, map->pm_prot, 0); if (IS_ERR(pmap_clnt)) { task->tk_status = PTR_ERR(pmap_clnt); goto bailout; } - task->tk_status = 0; - /* - * Note: rpc_new_child will release client after a failure. - */ - if (!(child = rpc_new_child(pmap_clnt, task))) + child = rpc_run_task(pmap_clnt, RPC_TASK_ASYNC, &pmap_getport_ops, map); + if (IS_ERR(child)) { + task->tk_status = -EIO; goto bailout; + } + rpc_release_task(child); - /* Setup the call info struct */ - rpc_call_setup(child, &msg, 0); + rpc_sleep_on(&xprt->binding, task, NULL, NULL); - /* ... and run the child task */ task->tk_xprt->stat.bind_count++; - rpc_run_child(task, child, pmap_getport_done); return; bailout: - spin_lock(&pmap_lock); - map->pm_binding = 0; - rpc_wake_up(&map->pm_bindwait); - spin_unlock(&pmap_lock); - rpc_exit(task, -EIO); + pmap_map_free(map); +bailout_nofree: + pmap_wake_portmap_waiters(xprt); } #ifdef CONFIG_ROOT_NFS int rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot) { - struct rpc_portmap map = { + struct portmap_args map = { .pm_prog = prog, .pm_vers = vers, .pm_prot = prot, @@ -133,32 +183,32 @@ rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot) #endif static void -pmap_getport_done(struct rpc_task *task) +pmap_getport_done(struct rpc_task *child, void *data) { - struct rpc_clnt *clnt = task->tk_client; + struct portmap_args *map = data; + struct rpc_task *task = map->pm_task; struct rpc_xprt *xprt = task->tk_xprt; - struct rpc_portmap *map = clnt->cl_pmap; + int status = child->tk_status; - dprintk("RPC: %4d pmap_getport_done(status %d, port %d)\n", - task->tk_pid, task->tk_status, clnt->cl_port); - - if (task->tk_status < 0) { - /* Make the calling task exit with an error */ + if (status < 0) { + /* Portmapper not available */ xprt->ops->set_port(xprt, 0); - task->tk_action = rpc_exit_task; - } else if (clnt->cl_port == 0) { - /* Program not registered */ + task->tk_status = status; + } else if (map->pm_port == 0) { + /* Requested RPC service wasn't registered */ xprt->ops->set_port(xprt, 0); - rpc_exit(task, -EACCES); + task->tk_status = -EACCES; } else { - xprt->ops->set_port(xprt, clnt->cl_port); + /* Succeeded */ + xprt->ops->set_port(xprt, map->pm_port); xprt_set_bound(xprt); - clnt->cl_port = htons(clnt->cl_port); + task->tk_status = 0; } - spin_lock(&pmap_lock); - map->pm_binding = 0; - rpc_wake_up(&map->pm_bindwait); - spin_unlock(&pmap_lock); + + dprintk("RPC: %4d pmap_getport_done(status %d, port %u)\n", + child->tk_pid, child->tk_status, map->pm_port); + + pmap_wake_portmap_waiters(xprt); } /* @@ -172,7 +222,7 @@ rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay) .sin_family = AF_INET, .sin_addr.s_addr = htonl(INADDR_LOOPBACK), }; - struct rpc_portmap map = { + struct portmap_args map = { .pm_prog = prog, .pm_vers = vers, .pm_prot = prot, @@ -239,7 +289,7 @@ pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileg * XDR encode/decode functions for PMAP */ static int -xdr_encode_mapping(struct rpc_rqst *req, u32 *p, struct rpc_portmap *map) +xdr_encode_mapping(struct rpc_rqst *req, u32 *p, struct portmap_args *map) { dprintk("RPC: xdr_encode_mapping(%d, %d, %d, %d)\n", map->pm_prog, map->pm_vers, map->pm_prot, map->pm_port); diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index e239ef985ef7..b45abd0743cb 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -928,6 +928,7 @@ static struct rpc_xprt *xprt_setup(int proto, struct sockaddr_in *ap, struct rpc xprt->last_used = jiffies; xprt->cwnd = RPC_INITCWND; + rpc_init_wait_queue(&xprt->binding, "xprt_binding"); rpc_init_wait_queue(&xprt->pending, "xprt_pending"); rpc_init_wait_queue(&xprt->sending, "xprt_sending"); rpc_init_wait_queue(&xprt->resend, "xprt_resend"); From c4a5692fb83f23008c720fe84454d5603e80b103 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:16 -0400 Subject: [PATCH 29/74] SUNRPC: Clean-up after recent changes to sunrpc/pmap_clnt.c Add comments for external functions, use modern function definition style, and fix up dprintk formatting. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/pmap_clnt.c | 70 +++++++++++++++++++++++++----------------- 1 file changed, 42 insertions(+), 28 deletions(-) diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index 59d542436ca9..0efcbf1302a2 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -1,7 +1,9 @@ /* - * linux/net/sunrpc/pmap.c + * linux/net/sunrpc/pmap_clnt.c * - * Portmapper client. + * In-kernel RPC portmapper client. + * + * Portmapper supports version 2 of the rpcbind protocol (RFC 1833). * * Copyright (C) 1996, Olaf Kirch */ @@ -76,12 +78,15 @@ static inline void pmap_wake_portmap_waiters(struct rpc_xprt *xprt) rpc_wake_up(&xprt->binding); } -/* - * Obtain the port for a given RPC service on a given host. This one can - * be called for an ongoing RPC request. +/** + * rpc_getport - obtain the port for a given RPC service on a given host + * @task: task that is waiting for portmapper request + * @clnt: controlling rpc_clnt + * + * This one can be called for an ongoing RPC request, and can be used in + * an async (rpciod) context. */ -void -rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt) +void rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt) { struct rpc_xprt *xprt = task->tk_xprt; struct sockaddr_in *sap = &xprt->addr; @@ -144,8 +149,16 @@ rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt) } #ifdef CONFIG_ROOT_NFS -int -rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot) +/** + * rpc_getport_external - obtain the port for a given RPC service on a given host + * @sin: address of remote peer + * @prog: RPC program number to bind + * @vers: RPC version number to bind + * @prot: transport protocol to use to make this request + * + * This one is called from outside the RPC client in a synchronous task context. + */ +int rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot) { struct portmap_args map = { .pm_prog = prog, @@ -162,7 +175,7 @@ rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot) char hostname[32]; int status; - dprintk("RPC: rpc_getport_external(%u.%u.%u.%u, %d, %d, %d)\n", + dprintk("RPC: rpc_getport_external(%u.%u.%u.%u, %u, %u, %d)\n", NIPQUAD(sin->sin_addr.s_addr), prog, vers, prot); sprintf(hostname, "%u.%u.%u.%u", NIPQUAD(sin->sin_addr.s_addr)); @@ -182,8 +195,10 @@ rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot) } #endif -static void -pmap_getport_done(struct rpc_task *child, void *data) +/* + * Portmapper child task invokes this callback via tk_exit. + */ +static void pmap_getport_done(struct rpc_task *child, void *data) { struct portmap_args *map = data; struct rpc_task *task = map->pm_task; @@ -211,12 +226,17 @@ pmap_getport_done(struct rpc_task *child, void *data) pmap_wake_portmap_waiters(xprt); } -/* - * Set or unset a port registration with the local portmapper. +/** + * rpc_register - set or unset a port registration with the local portmapper + * @prog: RPC program number to bind + * @vers: RPC version number to bind + * @prot: transport protocol to use to make this request + * @port: port value to register + * @okay: result code + * * port == 0 means unregister, port != 0 means register. */ -int -rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay) +int rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay) { struct sockaddr_in sin = { .sin_family = AF_INET, @@ -236,7 +256,7 @@ rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay) struct rpc_clnt *pmap_clnt; int error = 0; - dprintk("RPC: registering (%d, %d, %d, %d) with portmapper.\n", + dprintk("RPC: registering (%u, %u, %d, %u) with portmapper.\n", prog, vers, prot, port); pmap_clnt = pmap_create("localhost", &sin, IPPROTO_UDP, 1); @@ -259,13 +279,11 @@ rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay) return error; } -static struct rpc_clnt * -pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileged) +static struct rpc_clnt *pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileged) { struct rpc_xprt *xprt; struct rpc_clnt *clnt; - /* printk("pmap: create xprt\n"); */ xprt = xprt_create_proto(proto, srvaddr, NULL); if (IS_ERR(xprt)) return (struct rpc_clnt *)xprt; @@ -274,7 +292,6 @@ pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileg if (!privileged) xprt->resvport = 0; - /* printk("pmap: create clnt\n"); */ clnt = rpc_new_client(xprt, hostname, &pmap_program, RPC_PMAP_VERSION, RPC_AUTH_UNIX); @@ -288,10 +305,9 @@ pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileg /* * XDR encode/decode functions for PMAP */ -static int -xdr_encode_mapping(struct rpc_rqst *req, u32 *p, struct portmap_args *map) +static int xdr_encode_mapping(struct rpc_rqst *req, u32 *p, struct portmap_args *map) { - dprintk("RPC: xdr_encode_mapping(%d, %d, %d, %d)\n", + dprintk("RPC: xdr_encode_mapping(%u, %u, %u, %u)\n", map->pm_prog, map->pm_vers, map->pm_prot, map->pm_port); *p++ = htonl(map->pm_prog); *p++ = htonl(map->pm_vers); @@ -302,15 +318,13 @@ xdr_encode_mapping(struct rpc_rqst *req, u32 *p, struct portmap_args *map) return 0; } -static int -xdr_decode_port(struct rpc_rqst *req, u32 *p, unsigned short *portp) +static int xdr_decode_port(struct rpc_rqst *req, u32 *p, unsigned short *portp) { *portp = (unsigned short) ntohl(*p++); return 0; } -static int -xdr_decode_bool(struct rpc_rqst *req, u32 *p, unsigned int *boolp) +static int xdr_decode_bool(struct rpc_rqst *req, u32 *p, unsigned int *boolp) { *boolp = (unsigned int) ntohl(*p++); return 0; From 5b1eacbcd78930d976eb50a93f1779d311b553d1 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:16 -0400 Subject: [PATCH 30/74] SUNRPC: Support for RPC child tasks no longer needed The previous patches removed the last user of RPC child tasks, so we can remove support for child tasks from net/sunrpc/sched.c now. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/sched.h | 5 --- net/sunrpc/sched.c | 82 ------------------------------------ 2 files changed, 87 deletions(-) diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h index 82a91bb22362..f399c138f79d 100644 --- a/include/linux/sunrpc/sched.h +++ b/include/linux/sunrpc/sched.h @@ -127,7 +127,6 @@ struct rpc_call_ops { */ #define RPC_TASK_ASYNC 0x0001 /* is an async task */ #define RPC_TASK_SWAPPER 0x0002 /* is swapping in/out */ -#define RPC_TASK_CHILD 0x0008 /* is child of other task */ #define RPC_CALL_MAJORSEEN 0x0020 /* major timeout seen */ #define RPC_TASK_ROOTCREDS 0x0040 /* force root creds */ #define RPC_TASK_DYNAMIC 0x0080 /* task was kmalloc'ed */ @@ -136,7 +135,6 @@ struct rpc_call_ops { #define RPC_TASK_NOINTR 0x0400 /* uninterruptible task */ #define RPC_IS_ASYNC(t) ((t)->tk_flags & RPC_TASK_ASYNC) -#define RPC_IS_CHILD(t) ((t)->tk_flags & RPC_TASK_CHILD) #define RPC_IS_SWAPPER(t) ((t)->tk_flags & RPC_TASK_SWAPPER) #define RPC_DO_ROOTOVERRIDE(t) ((t)->tk_flags & RPC_TASK_ROOTCREDS) #define RPC_ASSASSINATED(t) ((t)->tk_flags & RPC_TASK_KILLED) @@ -253,7 +251,6 @@ struct rpc_task *rpc_new_task(struct rpc_clnt *, int flags, const struct rpc_call_ops *ops, void *data); struct rpc_task *rpc_run_task(struct rpc_clnt *clnt, int flags, const struct rpc_call_ops *ops, void *data); -struct rpc_task *rpc_new_child(struct rpc_clnt *, struct rpc_task *parent); void rpc_init_task(struct rpc_task *task, struct rpc_clnt *clnt, int flags, const struct rpc_call_ops *ops, void *data); @@ -261,8 +258,6 @@ void rpc_release_task(struct rpc_task *); void rpc_exit_task(struct rpc_task *); void rpc_killall_tasks(struct rpc_clnt *); int rpc_execute(struct rpc_task *); -void rpc_run_child(struct rpc_task *parent, struct rpc_task *child, - rpc_action action); void rpc_init_priority_wait_queue(struct rpc_wait_queue *, const char *); void rpc_init_wait_queue(struct rpc_wait_queue *, const char *); void rpc_sleep_on(struct rpc_wait_queue *, struct rpc_task *, diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c index 5c3eee768504..015ffe423a2f 100644 --- a/net/sunrpc/sched.c +++ b/net/sunrpc/sched.c @@ -44,12 +44,6 @@ static void __rpc_default_timer(struct rpc_task *task); static void rpciod_killall(void); static void rpc_async_schedule(void *); -/* - * RPC tasks that create another task (e.g. for contacting the portmapper) - * will wait on this queue for their child's completion - */ -static RPC_WAITQ(childq, "childq"); - /* * RPC tasks sit here while waiting for conditions to improve. */ @@ -323,16 +317,6 @@ static void rpc_make_runnable(struct rpc_task *task) wake_up_bit(&task->tk_runstate, RPC_TASK_QUEUED); } -/* - * Place a newly initialized task on the workqueue. - */ -static inline void -rpc_schedule_run(struct rpc_task *task) -{ - rpc_set_active(task); - rpc_make_runnable(task); -} - /* * Prepare for sleeping on a wait queue. * By always appending tasks to the list we ensure FIFO behavior. @@ -933,72 +917,6 @@ struct rpc_task *rpc_run_task(struct rpc_clnt *clnt, int flags, } EXPORT_SYMBOL(rpc_run_task); -/** - * rpc_find_parent - find the parent of a child task. - * @child: child task - * @parent: parent task - * - * Checks that the parent task is still sleeping on the - * queue 'childq'. If so returns a pointer to the parent. - * Upon failure returns NULL. - * - * Caller must hold childq.lock - */ -static inline struct rpc_task *rpc_find_parent(struct rpc_task *child, struct rpc_task *parent) -{ - struct rpc_task *task; - struct list_head *le; - - task_for_each(task, le, &childq.tasks[0]) - if (task == parent) - return parent; - - return NULL; -} - -static void rpc_child_exit(struct rpc_task *child, void *calldata) -{ - struct rpc_task *parent; - - spin_lock_bh(&childq.lock); - if ((parent = rpc_find_parent(child, calldata)) != NULL) { - parent->tk_status = child->tk_status; - __rpc_wake_up_task(parent); - } - spin_unlock_bh(&childq.lock); -} - -static const struct rpc_call_ops rpc_child_ops = { - .rpc_call_done = rpc_child_exit, -}; - -/* - * Note: rpc_new_task releases the client after a failure. - */ -struct rpc_task * -rpc_new_child(struct rpc_clnt *clnt, struct rpc_task *parent) -{ - struct rpc_task *task; - - task = rpc_new_task(clnt, RPC_TASK_ASYNC | RPC_TASK_CHILD, &rpc_child_ops, parent); - if (!task) - goto fail; - return task; - -fail: - parent->tk_status = -ENOMEM; - return NULL; -} - -void rpc_run_child(struct rpc_task *task, struct rpc_task *child, rpc_action func) -{ - spin_lock_bh(&childq.lock); - /* N.B. Is it possible for the child to have already finished? */ - __rpc_sleep_on(&childq, task, func, NULL); - rpc_schedule_run(child); - spin_unlock_bh(&childq.lock); -} - /* * Kill all tasks for the given client. * XXX: kill their descendants as well? From bbf7c1dd2ae2b4040b41b1065ee9b1b6905b1605 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:16 -0400 Subject: [PATCH 31/74] SUNRPC: Introduce transport switch callout for pluggable rpcbind Introduce a clean transport switch API for plugging in different types of rpcbind mechanisms. For instance, rpcbind can cleanly replace the existing portmapper client, or a transport can choose to implement RPC binding any way it likes. Test plan: Destructive testing (unplugging the network temporarily). Connectathon with UDP and TCP. NFSv2/3 and NFSv4 mounting should be carefully checked. Probably need to rig a server where certain services aren't running, or that returns an error for some typical operation. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/clnt.h | 2 +- include/linux/sunrpc/xprt.h | 1 + net/sunrpc/clnt.c | 3 +-- net/sunrpc/pmap_clnt.c | 4 ++-- net/sunrpc/xprtsock.c | 2 ++ 5 files changed, 7 insertions(+), 5 deletions(-) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 00e9dbaec9c5..2e68ac0aa022 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -106,7 +106,7 @@ struct rpc_clnt *rpc_clone_client(struct rpc_clnt *); int rpc_shutdown_client(struct rpc_clnt *); int rpc_destroy_client(struct rpc_clnt *); void rpc_release_client(struct rpc_clnt *); -void rpc_getport(struct rpc_task *, struct rpc_clnt *); +void rpc_getport(struct rpc_task *); int rpc_register(u32, u32, int, unsigned short, int *); void rpc_call_setup(struct rpc_task *, struct rpc_message *, int); diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 4ce82616873d..84122559fa17 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -105,6 +105,7 @@ struct rpc_xprt_ops { void (*set_buffer_size)(struct rpc_xprt *xprt, size_t sndsize, size_t rcvsize); int (*reserve_xprt)(struct rpc_task *task); void (*release_xprt)(struct rpc_xprt *xprt, struct rpc_task *task); + void (*rpcbind)(struct rpc_task *task); void (*set_port)(struct rpc_xprt *xprt, unsigned short port); void (*connect)(struct rpc_task *task); void * (*buf_alloc)(struct rpc_task *task, size_t size); diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index cee504162a3f..d003c2f5688f 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -774,7 +774,6 @@ call_encode(struct rpc_task *task) static void call_bind(struct rpc_task *task) { - struct rpc_clnt *clnt = task->tk_client; struct rpc_xprt *xprt = task->tk_xprt; dprintk("RPC: %4d call_bind (status %d)\n", @@ -784,7 +783,7 @@ call_bind(struct rpc_task *task) if (!xprt_bound(xprt)) { task->tk_action = call_bind_status; task->tk_timeout = xprt->bind_timeout; - rpc_getport(task, clnt); + xprt->ops->rpcbind(task); } } diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index 0efcbf1302a2..f7b279a63baa 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -81,13 +81,13 @@ static inline void pmap_wake_portmap_waiters(struct rpc_xprt *xprt) /** * rpc_getport - obtain the port for a given RPC service on a given host * @task: task that is waiting for portmapper request - * @clnt: controlling rpc_clnt * * This one can be called for an ongoing RPC request, and can be used in * an async (rpciod) context. */ -void rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt) +void rpc_getport(struct rpc_task *task) { + struct rpc_clnt *clnt = task->tk_client; struct rpc_xprt *xprt = task->tk_xprt; struct sockaddr_in *sap = &xprt->addr; struct portmap_args *map; diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 123ac1e5ba15..4c98b89a5b48 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -1262,6 +1262,7 @@ static struct rpc_xprt_ops xs_udp_ops = { .set_buffer_size = xs_udp_set_buffer_size, .reserve_xprt = xprt_reserve_xprt_cong, .release_xprt = xprt_release_xprt_cong, + .rpcbind = rpc_getport, .set_port = xs_set_port, .connect = xs_connect, .buf_alloc = rpc_malloc, @@ -1278,6 +1279,7 @@ static struct rpc_xprt_ops xs_udp_ops = { static struct rpc_xprt_ops xs_tcp_ops = { .reserve_xprt = xprt_reserve_xprt, .release_xprt = xs_tcp_release_xprt, + .rpcbind = rpc_getport, .set_port = xs_set_port, .connect = xs_connect, .buf_alloc = rpc_malloc, From ed39440a2573abc926f230267000f21fa5a87822 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:17 -0400 Subject: [PATCH 32/74] SUNRPC: create API for getting remote peer address Provide an API for retrieving the remote peer address without allowing direct access to the rpc_xprt struct. Test-plan: Compile kernel with CONFIG_NFS enabled. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/clnt.h | 1 + net/sunrpc/clnt.c | 21 +++++++++++++++++++++ 2 files changed, 22 insertions(+) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 2e68ac0aa022..65196b03f0ab 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -123,6 +123,7 @@ void rpc_setbufsize(struct rpc_clnt *, unsigned int, unsigned int); size_t rpc_max_payload(struct rpc_clnt *); void rpc_force_rebind(struct rpc_clnt *); int rpc_ping(struct rpc_clnt *clnt, int flags); +size_t rpc_peeraddr(struct rpc_clnt *, struct sockaddr *, size_t); /* * Helper function for NFSroot support diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index d003c2f5688f..94768cf5fd5b 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -533,6 +533,27 @@ rpc_call_setup(struct rpc_task *task, struct rpc_message *msg, int flags) task->tk_action = rpc_exit_task; } +/** + * rpc_peeraddr - extract remote peer address from clnt's xprt + * @clnt: RPC client structure + * @buf: target buffer + * @size: length of target buffer + * + * Returns the number of bytes that are actually in the stored address. + */ +size_t rpc_peeraddr(struct rpc_clnt *clnt, struct sockaddr *buf, size_t bufsize) +{ + size_t bytes; + struct rpc_xprt *xprt = clnt->cl_xprt; + + bytes = sizeof(xprt->addr); + if (bytes > bufsize) + bytes = bufsize; + memcpy(buf, &clnt->cl_xprt->addr, bytes); + return sizeof(xprt->addr); +} +EXPORT_SYMBOL(rpc_peeraddr); + void rpc_setbufsize(struct rpc_clnt *clnt, unsigned int sndsize, unsigned int rcvsize) { From 44c31be261540acf66ddd730631ead8009cc361d Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:17 -0400 Subject: [PATCH 33/74] LOCKD: Teach lockd to use the new rpc_peeraddr() API Hide the details of how the RPC client stores remote peer addresses from the Network Lock Manager. Test plan: Destructive testing (unplugging the network temporarily). Connectathon with UDP and TCP. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index 89ba0df14c22..50dbb67ae0c4 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -151,11 +151,13 @@ static void nlmclnt_release_lockargs(struct nlm_rqst *req) int nlmclnt_proc(struct inode *inode, int cmd, struct file_lock *fl) { + struct rpc_clnt *client = NFS_CLIENT(inode); + struct sockaddr_in addr; struct nlm_host *host; struct nlm_rqst *call; sigset_t oldset; unsigned long flags; - int status, proto, vers; + int status, vers; vers = (NFS_PROTO(inode)->version == 3) ? 4 : 1; if (NFS_PROTO(inode)->version > 3) { @@ -163,10 +165,8 @@ nlmclnt_proc(struct inode *inode, int cmd, struct file_lock *fl) return -ENOLCK; } - /* Retrieve transport protocol from NFS client */ - proto = NFS_CLIENT(inode)->cl_xprt->prot; - - host = nlmclnt_lookup_host(NFS_ADDR(inode), proto, vers); + rpc_peeraddr(client, (struct sockaddr *) &addr, sizeof(addr)); + host = nlmclnt_lookup_host(&addr, client->cl_xprt->prot, vers); if (host == NULL) return -ENOLCK; From 081f79a9b09b634f0dc08ed014e0195464d52535 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:17 -0400 Subject: [PATCH 34/74] SUNRPC: Teach the RPC portmapper to use the new rpc_peeraddr() API. Hide the details of how the RPC client stores remote peer addresses from the RPC portmapper. Test plan: Destructive testing (unplugging the network temporarily). Connectathon with UDP and TCP. NFSv2/3 and NFSv4 mounting should be carefully checked. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/pmap_clnt.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index f7b279a63baa..3eee8e907275 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -89,7 +89,7 @@ void rpc_getport(struct rpc_task *task) { struct rpc_clnt *clnt = task->tk_client; struct rpc_xprt *xprt = task->tk_xprt; - struct sockaddr_in *sap = &xprt->addr; + struct sockaddr_in addr; struct portmap_args *map; struct rpc_clnt *pmap_clnt; struct rpc_task *child; @@ -124,7 +124,8 @@ void rpc_getport(struct rpc_task *task) map->pm_port = 0; map->pm_task = task; - pmap_clnt = pmap_create(clnt->cl_server, sap, map->pm_prot, 0); + rpc_peeraddr(clnt, (struct sockaddr *) &addr, sizeof(addr)); + pmap_clnt = pmap_create(clnt->cl_server, &addr, map->pm_prot, 0); if (IS_ERR(pmap_clnt)) { task->tk_status = PTR_ERR(pmap_clnt); goto bailout; From 39d7bbcb5ba5e9d8d658b70903dd7939400e57db Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:18 -0400 Subject: [PATCH 35/74] SUNRPC: remove extraneous header inclusions include/linux/sunrpc/clnt.h already includes include/linux/sunrpc/xprt.h. We can remove xprt.h from source files that already include clnt.h. Likewise include/linux/sunrpc/timer.h. Test plan: Compile kernel with CONFIG_NFS enabled. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/mount_clnt.c | 1 - include/linux/nfs_xdr.h | 1 - net/sunrpc/pmap_clnt.c | 1 - net/sunrpc/sched.c | 1 - net/sunrpc/timer.c | 2 -- 5 files changed, 6 deletions(-) diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c index 445abb4d4214..41274874b9a5 100644 --- a/fs/nfs/mount_clnt.c +++ b/fs/nfs/mount_clnt.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 2426b11b6cce..0f33e621892f 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1,7 +1,6 @@ #ifndef _LINUX_NFS_XDR_H #define _LINUX_NFS_XDR_H -#include #include /* diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index 3eee8e907275..523f0e825dea 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -15,7 +15,6 @@ #include #include #include -#include #include #ifdef RPC_DEBUG diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c index 015ffe423a2f..ecf366351bf7 100644 --- a/net/sunrpc/sched.c +++ b/net/sunrpc/sched.c @@ -21,7 +21,6 @@ #include #include -#include #ifdef RPC_DEBUG #define RPCDBG_FACILITY RPCDBG_SCHED diff --git a/net/sunrpc/timer.c b/net/sunrpc/timer.c index bcbdf6430d5c..8142fdb8a930 100644 --- a/net/sunrpc/timer.c +++ b/net/sunrpc/timer.c @@ -19,8 +19,6 @@ #include #include -#include -#include #define RPC_RTO_MAX (60*HZ) #define RPC_RTO_INIT (HZ/5) From edb267a688fcee5335d596752f117a30c7152e44 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:18 -0400 Subject: [PATCH 36/74] SUNRPC: add xprt switch API for printing formatted remote peer addresses Add a new method to the transport switch API to provide a way to convert the opaque contents of xprt->addr to a human-readable string. Test plan: Compile kernel with CONFIG_NFS enabled. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprt.h | 11 ++++++ net/sunrpc/xprtsock.c | 79 +++++++++++++++++++++++++++++++++---- 2 files changed, 82 insertions(+), 8 deletions(-) diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 84122559fa17..8372ab8fc9b5 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -51,6 +51,14 @@ struct rpc_timeout { unsigned char to_exponential; }; +enum rpc_display_format_t { + RPC_DISPLAY_ADDR = 0, + RPC_DISPLAY_PORT, + RPC_DISPLAY_PROTO, + RPC_DISPLAY_ALL, + RPC_DISPLAY_MAX, +}; + struct rpc_task; struct rpc_xprt; struct seq_file; @@ -103,6 +111,7 @@ struct rpc_rqst { struct rpc_xprt_ops { void (*set_buffer_size)(struct rpc_xprt *xprt, size_t sndsize, size_t rcvsize); + char * (*print_addr)(struct rpc_xprt *xprt, enum rpc_display_format_t format); int (*reserve_xprt)(struct rpc_task *task); void (*release_xprt)(struct rpc_xprt *xprt, struct rpc_task *task); void (*rpcbind)(struct rpc_task *task); @@ -207,6 +216,8 @@ struct rpc_xprt { void (*old_data_ready)(struct sock *, int); void (*old_state_change)(struct sock *); void (*old_write_space)(struct sock *); + + char * address_strings[RPC_DISPLAY_MAX]; }; #define XPRT_LAST_FRAG (1 << 0) diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 4c98b89a5b48..cb8e6c34e12f 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -125,6 +125,47 @@ static inline void xs_pktdump(char *msg, u32 *packet, unsigned int count) } #endif +static void xs_format_peer_addresses(struct rpc_xprt *xprt) +{ + struct sockaddr_in *addr = (struct sockaddr_in *) &xprt->addr; + char *buf; + + buf = kzalloc(20, GFP_KERNEL); + if (buf) { + snprintf(buf, 20, "%u.%u.%u.%u", + NIPQUAD(addr->sin_addr.s_addr)); + } + xprt->address_strings[RPC_DISPLAY_ADDR] = buf; + + buf = kzalloc(8, GFP_KERNEL); + if (buf) { + snprintf(buf, 8, "%u", + ntohs(addr->sin_port)); + } + xprt->address_strings[RPC_DISPLAY_PORT] = buf; + + if (xprt->prot == IPPROTO_UDP) + xprt->address_strings[RPC_DISPLAY_PROTO] = "udp"; + else + xprt->address_strings[RPC_DISPLAY_PROTO] = "tcp"; + + buf = kzalloc(48, GFP_KERNEL); + if (buf) { + snprintf(buf, 48, "addr=%u.%u.%u.%u port=%u proto=%s", + NIPQUAD(addr->sin_addr.s_addr), + ntohs(addr->sin_port), + xprt->prot == IPPROTO_UDP ? "udp" : "tcp"); + } + xprt->address_strings[RPC_DISPLAY_ALL] = buf; +} + +static void xs_free_peer_addresses(struct rpc_xprt *xprt) +{ + kfree(xprt->address_strings[RPC_DISPLAY_ADDR]); + kfree(xprt->address_strings[RPC_DISPLAY_PORT]); + kfree(xprt->address_strings[RPC_DISPLAY_ALL]); +} + #define XS_SENDMSG_FLAGS (MSG_DONTWAIT | MSG_NOSIGNAL) static inline int xs_send_head(struct socket *sock, struct sockaddr *addr, int addrlen, struct xdr_buf *xdr, unsigned int base, unsigned int len) @@ -490,6 +531,7 @@ static void xs_destroy(struct rpc_xprt *xprt) xprt_disconnect(xprt); xs_close(xprt); + xs_free_peer_addresses(xprt); kfree(xprt->slot); } @@ -964,6 +1006,19 @@ static unsigned short xs_get_random_port(void) return rand + xprt_min_resvport; } +/** + * xs_print_peer_address - format an IPv4 address for printing + * @xprt: generic transport + * @format: flags field indicating which parts of the address to render + */ +static char *xs_print_peer_address(struct rpc_xprt *xprt, enum rpc_display_format_t format) +{ + if (xprt->address_strings[format] != NULL) + return xprt->address_strings[format]; + else + return "unprintable"; +} + /** * xs_set_port - reset the port number in the remote endpoint address * @xprt: generic transport @@ -1019,8 +1074,6 @@ static void xs_udp_connect_worker(void *args) if (xprt->shutdown || !xprt_bound(xprt)) goto out; - dprintk("RPC: xs_udp_connect_worker for xprt %p\n", xprt); - /* Start by resetting any existing state */ xs_close(xprt); @@ -1034,6 +1087,9 @@ static void xs_udp_connect_worker(void *args) goto out; } + dprintk("RPC: worker connecting xprt %p to address: %s\n", + xprt, xs_print_peer_address(xprt, RPC_DISPLAY_ALL)); + if (!xprt->inet) { struct sock *sk = sock->sk; @@ -1102,8 +1158,6 @@ static void xs_tcp_connect_worker(void *args) if (xprt->shutdown || !xprt_bound(xprt)) goto out; - dprintk("RPC: xs_tcp_connect_worker for xprt %p\n", xprt); - if (!xprt->sock) { /* start from scratch */ if ((err = sock_create_kern(PF_INET, SOCK_STREAM, IPPROTO_TCP, &sock)) < 0) { @@ -1119,6 +1173,9 @@ static void xs_tcp_connect_worker(void *args) /* "close" the socket, preserving the local port */ xs_tcp_reuse_connection(xprt); + dprintk("RPC: worker connecting xprt %p to address: %s\n", + xprt, xs_print_peer_address(xprt, RPC_DISPLAY_ALL)); + if (!xprt->inet) { struct sock *sk = sock->sk; @@ -1260,6 +1317,7 @@ static void xs_tcp_print_stats(struct rpc_xprt *xprt, struct seq_file *seq) static struct rpc_xprt_ops xs_udp_ops = { .set_buffer_size = xs_udp_set_buffer_size, + .print_addr = xs_print_peer_address, .reserve_xprt = xprt_reserve_xprt_cong, .release_xprt = xprt_release_xprt_cong, .rpcbind = rpc_getport, @@ -1277,6 +1335,7 @@ static struct rpc_xprt_ops xs_udp_ops = { }; static struct rpc_xprt_ops xs_tcp_ops = { + .print_addr = xs_print_peer_address, .reserve_xprt = xprt_reserve_xprt, .release_xprt = xs_tcp_release_xprt, .rpcbind = rpc_getport, @@ -1301,8 +1360,6 @@ int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to) { size_t slot_table_size; - dprintk("RPC: setting up udp-ipv4 transport...\n"); - xprt->max_reqs = xprt_udp_slot_table_entries; slot_table_size = xprt->max_reqs * sizeof(xprt->slot[0]); xprt->slot = kzalloc(slot_table_size, GFP_KERNEL); @@ -1332,6 +1389,10 @@ int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to) else xprt_set_timeout(&xprt->timeout, 5, 5 * HZ); + xs_format_peer_addresses(xprt); + dprintk("RPC: set up transport to address %s\n", + xs_print_peer_address(xprt, RPC_DISPLAY_ALL)); + return 0; } @@ -1345,8 +1406,6 @@ int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to) { size_t slot_table_size; - dprintk("RPC: setting up tcp-ipv4 transport...\n"); - xprt->max_reqs = xprt_tcp_slot_table_entries; slot_table_size = xprt->max_reqs * sizeof(xprt->slot[0]); xprt->slot = kzalloc(slot_table_size, GFP_KERNEL); @@ -1375,5 +1434,9 @@ int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to) else xprt_set_timeout(&xprt->timeout, 2, 60 * HZ); + xs_format_peer_addresses(xprt); + dprintk("RPC: set up transport to address %s\n", + xs_print_peer_address(xprt, RPC_DISPLAY_ALL)); + return 0; } From f425eba437f0051bde979ea2eef8bc875a77cd00 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:18 -0400 Subject: [PATCH 37/74] SUNRPC: Create API for displaying remote peer address Provide an API for formatting the remote peer address for printing without exposing its internal structure. The address could be dynamic, so we support a function call to get the address rather than reading it straight out of a structure. Test-plan: Destructive testing (unplugging the network temporarily). Probably need to rig a server where certain services aren't running, or that returns an error for some typical operation. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/clnt.h | 1 + net/sunrpc/clnt.c | 13 +++++++++++++ 2 files changed, 14 insertions(+) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 65196b03f0ab..b7d47f018353 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -124,6 +124,7 @@ size_t rpc_max_payload(struct rpc_clnt *); void rpc_force_rebind(struct rpc_clnt *); int rpc_ping(struct rpc_clnt *clnt, int flags); size_t rpc_peeraddr(struct rpc_clnt *, struct sockaddr *, size_t); +char * rpc_peeraddr2str(struct rpc_clnt *, enum rpc_display_format_t); /* * Helper function for NFSroot support diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 94768cf5fd5b..e5b19e348d88 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -554,6 +554,19 @@ size_t rpc_peeraddr(struct rpc_clnt *clnt, struct sockaddr *buf, size_t bufsize) } EXPORT_SYMBOL(rpc_peeraddr); +/** + * rpc_peeraddr2str - return remote peer address in printable format + * @clnt: RPC client structure + * @format: address format + * + */ +char *rpc_peeraddr2str(struct rpc_clnt *clnt, enum rpc_display_format_t format) +{ + struct rpc_xprt *xprt = clnt->cl_xprt; + return xprt->ops->print_addr(xprt, format); +} +EXPORT_SYMBOL(rpc_peeraddr2str); + void rpc_setbufsize(struct rpc_clnt *clnt, unsigned int sndsize, unsigned int rcvsize) { From e7f7865743fff3d3938ec7540e5a784d662426da Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:19 -0400 Subject: [PATCH 38/74] SUNRPC: Teach rpc_pipe.c to use new rpc_peeraddr() API Hide the details of how the RPC client stores remote peer addresses from the RPC pipefs implementation. Test plan: Connectathon with Kerberos 5 authentication. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/rpc_pipe.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c index 0b1a1ac8a4bc..c21dc07f2a8c 100644 --- a/net/sunrpc/rpc_pipe.c +++ b/net/sunrpc/rpc_pipe.c @@ -327,10 +327,8 @@ rpc_show_info(struct seq_file *m, void *v) seq_printf(m, "RPC server: %s\n", clnt->cl_server); seq_printf(m, "service: %s (%d) version %d\n", clnt->cl_protname, clnt->cl_prog, clnt->cl_vers); - seq_printf(m, "address: %u.%u.%u.%u\n", - NIPQUAD(clnt->cl_xprt->addr.sin_addr.s_addr)); - seq_printf(m, "protocol: %s\n", - clnt->cl_xprt->prot == IPPROTO_UDP ? "udp" : "tcp"); + seq_printf(m, "address: %s\n", rpc_peeraddr2str(clnt, RPC_DISPLAY_ADDR)); + seq_printf(m, "protocol: %s\n", rpc_peeraddr2str(clnt, RPC_DISPLAY_PROTO)); return 0; } From c4efcb1d3e0bc76aeb9ca6301d19a5079893c6c9 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:19 -0400 Subject: [PATCH 39/74] SUNRPC: Use "sockaddr_storage" for storing RPC client's remote peer address IPv6 addresses are big (128 bytes). Now that no RPC client consumers treat the addr field in rpc_xprt structs as an opaque, and access it only via the API calls, we can safely widen the field in the rpc_xprt struct to accomodate larger addresses. Test plan: Compile kernel with CONFIG_NFS enabled. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprt.h | 3 ++- net/sunrpc/clnt.c | 2 +- net/sunrpc/xprt.c | 3 ++- net/sunrpc/xprtsock.c | 15 ++++++++++----- 4 files changed, 15 insertions(+), 8 deletions(-) diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 8372ab8fc9b5..fc05cfbd5805 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -134,7 +134,8 @@ struct rpc_xprt { struct sock * inet; /* INET layer */ struct rpc_timeout timeout; /* timeout parms */ - struct sockaddr_in addr; /* server address */ + struct sockaddr_storage addr; /* server address */ + size_t addrlen; /* size of server address */ int prot; /* IP protocol */ unsigned long cong; /* current congestion */ diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index e5b19e348d88..ff1e90fd81ab 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -550,7 +550,7 @@ size_t rpc_peeraddr(struct rpc_clnt *clnt, struct sockaddr *buf, size_t bufsize) if (bytes > bufsize) bytes = bufsize; memcpy(buf, &clnt->cl_xprt->addr, bytes); - return sizeof(xprt->addr); + return xprt->addrlen; } EXPORT_SYMBOL(rpc_peeraddr); diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index b45abd0743cb..4987517cc74b 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -896,7 +896,8 @@ static struct rpc_xprt *xprt_setup(int proto, struct sockaddr_in *ap, struct rpc if ((xprt = kzalloc(sizeof(struct rpc_xprt), GFP_KERNEL)) == NULL) return ERR_PTR(-ENOMEM); - xprt->addr = *ap; + memcpy(&xprt->addr, ap, sizeof(*ap)); + xprt->addrlen = sizeof(*ap); switch (proto) { case IPPROTO_UDP: diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index cb8e6c34e12f..17179aa4c207 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -341,7 +341,7 @@ static int xs_udp_send_request(struct rpc_task *task) req->rq_xtime = jiffies; status = xs_sendpages(xprt->sock, (struct sockaddr *) &xprt->addr, - sizeof(xprt->addr), xdr, req->rq_bytes_sent); + xprt->addrlen, xdr, req->rq_bytes_sent); dprintk("RPC: xs_udp_send_request(%u) = %d\n", xdr->len - req->rq_bytes_sent, status); @@ -1027,8 +1027,11 @@ static char *xs_print_peer_address(struct rpc_xprt *xprt, enum rpc_display_forma */ static void xs_set_port(struct rpc_xprt *xprt, unsigned short port) { + struct sockaddr_in *sap = (struct sockaddr_in *) &xprt->addr; + dprintk("RPC: setting port for xprt %p to %u\n", xprt, port); - xprt->addr.sin_port = htons(port); + + sap->sin_port = htons(port); } static int xs_bindresvport(struct rpc_xprt *xprt, struct socket *sock) @@ -1209,7 +1212,7 @@ static void xs_tcp_connect_worker(void *args) xprt->stat.connect_count++; xprt->stat.connect_start = jiffies; status = sock->ops->connect(sock, (struct sockaddr *) &xprt->addr, - sizeof(xprt->addr), O_NONBLOCK); + xprt->addrlen, O_NONBLOCK); dprintk("RPC: %p connect status %d connected %d sock state %d\n", xprt, -status, xprt_connected(xprt), sock->sk->sk_state); if (status < 0) { @@ -1359,6 +1362,7 @@ static struct rpc_xprt_ops xs_tcp_ops = { int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to) { size_t slot_table_size; + struct sockaddr_in *addr = (struct sockaddr_in *) &xprt->addr; xprt->max_reqs = xprt_udp_slot_table_entries; slot_table_size = xprt->max_reqs * sizeof(xprt->slot[0]); @@ -1366,7 +1370,7 @@ int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to) if (xprt->slot == NULL) return -ENOMEM; - if (ntohs(xprt->addr.sin_port) != 0) + if (ntohs(addr->sin_port != 0)) xprt_set_bound(xprt); xprt->port = xs_get_random_port(); @@ -1405,6 +1409,7 @@ int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to) int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to) { size_t slot_table_size; + struct sockaddr_in *addr = (struct sockaddr_in *) &xprt->addr; xprt->max_reqs = xprt_tcp_slot_table_entries; slot_table_size = xprt->max_reqs * sizeof(xprt->slot[0]); @@ -1412,7 +1417,7 @@ int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to) if (xprt->slot == NULL) return -ENOMEM; - if (ntohs(xprt->addr.sin_port) != 0) + if (ntohs(addr->sin_port) != 0) xprt_set_bound(xprt); xprt->port = xs_get_random_port(); From 6ca948238724c945bd353f51d54ae7d285f3889f Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:19 -0400 Subject: [PATCH 40/74] SUNRPC: Clean-up after previous patches. Remove some unused macros related to accessing an RPC peer address Test plan: Compile kernel with CONFIG_NFS option enabled. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/host.c | 1 - include/linux/nfs_fs.h | 1 - include/linux/sunrpc/clnt.h | 3 --- 3 files changed, 5 deletions(-) diff --git a/fs/lockd/host.c b/fs/lockd/host.c index 38b0e8a1aec0..a516a01561b8 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -26,7 +26,6 @@ #define NLM_HOST_REBIND (60 * HZ) #define NLM_HOST_EXPIRE ((nrhosts > NLM_HOST_MAX)? 300 * HZ : 120 * HZ) #define NLM_HOST_COLLECT ((nrhosts > NLM_HOST_MAX)? 120 * HZ : 60 * HZ) -#define NLM_HOST_ADDR(sv) (&(sv)->s_nlmclnt->cl_xprt->addr) static struct nlm_host * nlm_hosts[NLM_HOST_NRHASH]; static unsigned long next_gc; diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 51e9bd90dedc..3b5b04193fee 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -216,7 +216,6 @@ static inline struct nfs_inode *NFS_I(struct inode *inode) #define NFS_SERVER(inode) (NFS_SB(inode->i_sb)) #define NFS_CLIENT(inode) (NFS_SERVER(inode)->client) #define NFS_PROTO(inode) (NFS_SERVER(inode)->nfs_client->rpc_ops) -#define NFS_ADDR(inode) (RPC_PEERADDR(NFS_CLIENT(inode))) #define NFS_COOKIEVERF(inode) (NFS_I(inode)->cookieverf) #define NFS_READTIME(inode) (NFS_I(inode)->read_cache_jiffies) #define NFS_CHANGE_ATTR(inode) (NFS_I(inode)->change_attr) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index b7d47f018353..a26d69583c7a 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -89,9 +89,6 @@ struct rpc_procinfo { char * p_name; /* name of procedure */ }; -#define RPC_CONGESTED(clnt) (RPCXPRT_CONGESTED((clnt)->cl_xprt)) -#define RPC_PEERADDR(clnt) (&(clnt)->cl_xprt->addr) - #ifdef __KERNEL__ struct rpc_clnt *rpc_create_client(struct rpc_xprt *xprt, char *servname, From c2866763b4029411d166040306691773c12d4caf Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:20 -0400 Subject: [PATCH 41/74] SUNRPC: use sockaddr + size when creating remote transport endpoints Prepare for more generic transport endpoint handling needed by transports that might use different forms of addressing, such as IPv6. Introduce a single function call to replace the two-call xprt_create_proto/rpc_create_client API. Define a new rpc_create_args structure that allows callers to pass in remote endpoint addresses of varying length. Test-plan: Compile kernel with CONFIG_NFS enabled. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/clnt.h | 22 +++++++++++ include/linux/sunrpc/xprt.h | 1 + net/sunrpc/clnt.c | 61 ++++++++++++++++++++++++++++++ net/sunrpc/xprt.c | 75 +++++++++++++++++++++++++++++++++++++ 4 files changed, 159 insertions(+) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index a26d69583c7a..7817ba82f1b2 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -97,6 +97,28 @@ struct rpc_clnt *rpc_create_client(struct rpc_xprt *xprt, char *servname, struct rpc_clnt *rpc_new_client(struct rpc_xprt *xprt, char *servname, struct rpc_program *info, u32 version, rpc_authflavor_t authflavor); + +struct rpc_create_args { + int protocol; + struct sockaddr *address; + size_t addrsize; + struct rpc_timeout *timeout; + char *servername; + struct rpc_program *program; + u32 version; + rpc_authflavor_t authflavor; + unsigned long flags; +}; + +/* Values for "flags" field */ +#define RPC_CLNT_CREATE_HARDRTRY (1UL << 0) +#define RPC_CLNT_CREATE_INTR (1UL << 1) +#define RPC_CLNT_CREATE_AUTOBIND (1UL << 2) +#define RPC_CLNT_CREATE_ONESHOT (1UL << 3) +#define RPC_CLNT_CREATE_NONPRIVPORT (1UL << 4) +#define RPC_CLNT_CREATE_NOPING (1UL << 5) + +struct rpc_clnt *rpc_create(struct rpc_create_args *args); struct rpc_clnt *rpc_bind_new_program(struct rpc_clnt *, struct rpc_program *, int); struct rpc_clnt *rpc_clone_client(struct rpc_clnt *); diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index fc05cfbd5805..bc80fcfdd892 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -237,6 +237,7 @@ void xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long /* * Generic internal transport functions */ +struct rpc_xprt * xprt_create_transport(int proto, struct sockaddr *addr, size_t size, struct rpc_timeout *toparms); void xprt_connect(struct rpc_task *task); void xprt_reserve(struct rpc_task *task); int xprt_reserve_xprt(struct rpc_task *task); diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index ff1e90fd81ab..dbb93bdf6cc9 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -192,6 +192,67 @@ rpc_new_client(struct rpc_xprt *xprt, char *servname, return ERR_PTR(err); } +/* + * rpc_create - create an RPC client and transport with one call + * @args: rpc_clnt create argument structure + * + * Creates and initializes an RPC transport and an RPC client. + * + * It can ping the server in order to determine if it is up, and to see if + * it supports this program and version. RPC_CLNT_CREATE_NOPING disables + * this behavior so asynchronous tasks can also use rpc_create. + */ +struct rpc_clnt *rpc_create(struct rpc_create_args *args) +{ + struct rpc_xprt *xprt; + struct rpc_clnt *clnt; + + xprt = xprt_create_transport(args->protocol, args->address, + args->addrsize, args->timeout); + if (IS_ERR(xprt)) + return (struct rpc_clnt *)xprt; + + /* + * By default, kernel RPC client connects from a reserved port. + * CAP_NET_BIND_SERVICE will not be set for unprivileged requesters, + * but it is always enabled for rpciod, which handles the connect + * operation. + */ + xprt->resvport = 1; + if (args->flags & RPC_CLNT_CREATE_NONPRIVPORT) + xprt->resvport = 0; + + dprintk("RPC: creating %s client for %s (xprt %p)\n", + args->program->name, args->servername, xprt); + + clnt = rpc_new_client(xprt, args->servername, args->program, + args->version, args->authflavor); + if (IS_ERR(clnt)) + return clnt; + + if (!(args->flags & RPC_CLNT_CREATE_NOPING)) { + int err = rpc_ping(clnt, RPC_TASK_SOFT|RPC_TASK_NOINTR); + if (err != 0) { + rpc_shutdown_client(clnt); + return ERR_PTR(err); + } + } + + clnt->cl_softrtry = 1; + if (args->flags & RPC_CLNT_CREATE_HARDRTRY) + clnt->cl_softrtry = 0; + + if (args->flags & RPC_CLNT_CREATE_INTR) + clnt->cl_intr = 1; + if (args->flags & RPC_CLNT_CREATE_AUTOBIND) + clnt->cl_autobind = 1; + if (args->flags & RPC_CLNT_CREATE_ONESHOT) + clnt->cl_oneshot = 1; + + return clnt; +} +EXPORT_SYMBOL(rpc_create); + /** * Create an RPC client * @xprt - pointer to xprt struct diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index 4987517cc74b..17f56cfe2412 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -887,6 +887,81 @@ void xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long i to->to_exponential = 0; } +/** + * xprt_create_transport - create an RPC transport + * @proto: requested transport protocol + * @ap: remote peer address + * @size: length of address + * @to: timeout parameters + * + */ +struct rpc_xprt *xprt_create_transport(int proto, struct sockaddr *ap, size_t size, struct rpc_timeout *to) +{ + int result; + struct rpc_xprt *xprt; + struct rpc_rqst *req; + + if ((xprt = kzalloc(sizeof(struct rpc_xprt), GFP_KERNEL)) == NULL) { + dprintk("RPC: xprt_create_transport: no memory\n"); + return ERR_PTR(-ENOMEM); + } + if (size <= sizeof(xprt->addr)) { + memcpy(&xprt->addr, ap, size); + xprt->addrlen = size; + } else { + kfree(xprt); + dprintk("RPC: xprt_create_transport: address too large\n"); + return ERR_PTR(-EBADF); + } + + switch (proto) { + case IPPROTO_UDP: + result = xs_setup_udp(xprt, to); + break; + case IPPROTO_TCP: + result = xs_setup_tcp(xprt, to); + break; + default: + printk(KERN_ERR "RPC: unrecognized transport protocol: %d\n", + proto); + return ERR_PTR(-EIO); + } + if (result) { + kfree(xprt); + dprintk("RPC: xprt_create_transport: failed, %d\n", result); + return ERR_PTR(result); + } + + spin_lock_init(&xprt->transport_lock); + spin_lock_init(&xprt->reserve_lock); + + INIT_LIST_HEAD(&xprt->free); + INIT_LIST_HEAD(&xprt->recv); + INIT_WORK(&xprt->task_cleanup, xprt_autoclose, xprt); + init_timer(&xprt->timer); + xprt->timer.function = xprt_init_autodisconnect; + xprt->timer.data = (unsigned long) xprt; + xprt->last_used = jiffies; + xprt->cwnd = RPC_INITCWND; + + rpc_init_wait_queue(&xprt->binding, "xprt_binding"); + rpc_init_wait_queue(&xprt->pending, "xprt_pending"); + rpc_init_wait_queue(&xprt->sending, "xprt_sending"); + rpc_init_wait_queue(&xprt->resend, "xprt_resend"); + rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog"); + + /* initialize free list */ + for (req = &xprt->slot[xprt->max_reqs-1]; req >= &xprt->slot[0]; req--) + list_add(&req->rq_list, &xprt->free); + + xprt_init_xid(xprt); + + dprintk("RPC: created transport %p with %u slots\n", xprt, + xprt->max_reqs); + + return xprt; +} + static struct rpc_xprt *xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to) { int result; From e1ec78928b4d5a31b7a847e65c6009f4229f7c0f Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:20 -0400 Subject: [PATCH 42/74] LOCKD: Convert to use new rpc_create() API Replace xprt_create_proto/rpc_create_client with new rpc_create() interface in the Network Lock Manager. Note that the semantics of NLM transports is now "hard" instead of "soft" to provide a better guarantee that lock requests will get to the server. Test plan: Repeated runs of Connectathon locking suite. Check network trace to ensure NLM requests are working correctly. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/host.c | 48 ++++++++++++++++++++++++++---------------------- fs/lockd/mon.c | 39 ++++++++++++++++----------------------- 2 files changed, 42 insertions(+), 45 deletions(-) diff --git a/fs/lockd/host.c b/fs/lockd/host.c index a516a01561b8..703fb038c813 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -166,7 +166,6 @@ struct rpc_clnt * nlm_bind_host(struct nlm_host *host) { struct rpc_clnt *clnt; - struct rpc_xprt *xprt; dprintk("lockd: nlm_bind_host(%08x)\n", (unsigned)ntohl(host->h_addr.sin_addr.s_addr)); @@ -178,7 +177,6 @@ nlm_bind_host(struct nlm_host *host) * RPC rebind is required */ if ((clnt = host->h_rpcclnt) != NULL) { - xprt = clnt->cl_xprt; if (time_after_eq(jiffies, host->h_nextrebind)) { rpc_force_rebind(clnt); host->h_nextrebind = jiffies + NLM_HOST_REBIND; @@ -186,31 +184,37 @@ nlm_bind_host(struct nlm_host *host) host->h_nextrebind - jiffies); } } else { - xprt = xprt_create_proto(host->h_proto, &host->h_addr, NULL); - if (IS_ERR(xprt)) - goto forgetit; + unsigned long increment = nlmsvc_timeout * HZ; + struct rpc_timeout timeparms = { + .to_initval = increment, + .to_increment = increment, + .to_maxval = increment * 6UL, + .to_retries = 5U, + }; + struct rpc_create_args args = { + .protocol = host->h_proto, + .address = (struct sockaddr *)&host->h_addr, + .addrsize = sizeof(host->h_addr), + .timeout = &timeparms, + .servername = host->h_name, + .program = &nlm_program, + .version = host->h_version, + .authflavor = RPC_AUTH_UNIX, + .flags = (RPC_CLNT_CREATE_HARDRTRY | + RPC_CLNT_CREATE_AUTOBIND), + }; - xprt_set_timeout(&xprt->timeout, 5, nlmsvc_timeout); - xprt->resvport = 1; /* NLM requires a reserved port */ - - /* Existing NLM servers accept AUTH_UNIX only */ - clnt = rpc_new_client(xprt, host->h_name, &nlm_program, - host->h_version, RPC_AUTH_UNIX); - if (IS_ERR(clnt)) - goto forgetit; - clnt->cl_autobind = 1; /* turn on pmap queries */ - clnt->cl_softrtry = 1; /* All queries are soft */ - - host->h_rpcclnt = clnt; + clnt = rpc_create(&args); + if (!IS_ERR(clnt)) + host->h_rpcclnt = clnt; + else { + printk("lockd: couldn't create RPC handle for %s\n", host->h_name); + clnt = NULL; + } } mutex_unlock(&host->h_mutex); return clnt; - -forgetit: - printk("lockd: couldn't create RPC handle for %s\n", host->h_name); - mutex_unlock(&host->h_mutex); - return NULL; } /* diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index 3fc683f46b3e..5954dcb497e4 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -109,30 +109,23 @@ nsm_unmonitor(struct nlm_host *host) static struct rpc_clnt * nsm_create(void) { - struct rpc_xprt *xprt; - struct rpc_clnt *clnt; - struct sockaddr_in sin; + struct sockaddr_in sin = { + .sin_family = AF_INET, + .sin_addr.s_addr = htonl(INADDR_LOOPBACK), + .sin_port = 0, + }; + struct rpc_create_args args = { + .protocol = IPPROTO_UDP, + .address = (struct sockaddr *)&sin, + .addrsize = sizeof(sin), + .servername = "localhost", + .program = &nsm_program, + .version = SM_VERSION, + .authflavor = RPC_AUTH_NULL, + .flags = (RPC_CLNT_CREATE_ONESHOT), + }; - sin.sin_family = AF_INET; - sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK); - sin.sin_port = 0; - - xprt = xprt_create_proto(IPPROTO_UDP, &sin, NULL); - if (IS_ERR(xprt)) - return (struct rpc_clnt *)xprt; - xprt->resvport = 1; /* NSM requires a reserved port */ - - clnt = rpc_create_client(xprt, "localhost", - &nsm_program, SM_VERSION, - RPC_AUTH_NULL); - if (IS_ERR(clnt)) - goto out_err; - clnt->cl_softrtry = 1; - clnt->cl_oneshot = 1; - return clnt; - -out_err: - return clnt; + return rpc_create(&args); } /* From 41877d207c46f050b709f452703ade20c3b4a096 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:20 -0400 Subject: [PATCH 43/74] NFS: Convert NFS client to use new rpc_create() API Convert NFS client mount logic to use rpc_create() instead of the old xprt_create_proto/rpc_create_client API. Test plan: Mount stress tests. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 27 +++++++++++---------------- 1 file changed, 11 insertions(+), 16 deletions(-) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 471d975e63c3..12941a8a6d75 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -401,8 +401,17 @@ static int nfs_create_rpc_client(struct nfs_client *clp, int proto, rpc_authflavor_t flavor) { struct rpc_timeout timeparms; - struct rpc_xprt *xprt = NULL; struct rpc_clnt *clnt = NULL; + struct rpc_create_args args = { + .protocol = proto, + .address = (struct sockaddr *)&clp->cl_addr, + .addrsize = sizeof(clp->cl_addr), + .timeout = &timeparms, + .servername = clp->cl_hostname, + .program = &nfs_program, + .version = clp->rpc_ops->version, + .authflavor = flavor, + }; if (!IS_ERR(clp->cl_rpcclient)) return 0; @@ -411,27 +420,13 @@ static int nfs_create_rpc_client(struct nfs_client *clp, int proto, clp->retrans_timeo = timeparms.to_initval; clp->retrans_count = timeparms.to_retries; - /* create transport and client */ - xprt = xprt_create_proto(proto, &clp->cl_addr, &timeparms); - if (IS_ERR(xprt)) { - dprintk("%s: cannot create RPC transport. Error = %ld\n", - __FUNCTION__, PTR_ERR(xprt)); - return PTR_ERR(xprt); - } - - /* Bind to a reserved port! */ - xprt->resvport = 1; - /* Create the client RPC handle */ - clnt = rpc_create_client(xprt, clp->cl_hostname, &nfs_program, - clp->rpc_ops->version, RPC_AUTH_UNIX); + clnt = rpc_create(&args); if (IS_ERR(clnt)) { dprintk("%s: cannot create RPC client. Error = %ld\n", __FUNCTION__, PTR_ERR(clnt)); return PTR_ERR(clnt); } - clnt->cl_intr = 1; - clnt->cl_softrtry = 1; clp->cl_rpcclient = clnt; return 0; } From ae5c79476f36512d1100e162606bb5691f2cce5a Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:21 -0400 Subject: [PATCH 44/74] NFSD: Convert NFS server callback logic to use new rpc_create API Replace xprt_create_proto/rpc_create_client call in NFS server callback functions to use new rpc_create() API. Test plan: NFSv4 delegation functionality tests. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfsd/nfs4callback.c | 64 +++++++++++++++++------------------------- 1 file changed, 26 insertions(+), 38 deletions(-) diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 54b37b1d2e3a..8583d99ee740 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -375,16 +375,28 @@ nfsd4_probe_callback(struct nfs4_client *clp) { struct sockaddr_in addr; struct nfs4_callback *cb = &clp->cl_callback; - struct rpc_timeout timeparms; - struct rpc_xprt * xprt; + struct rpc_timeout timeparms = { + .to_initval = (NFSD_LEASE_TIME/4) * HZ, + .to_retries = 5, + .to_maxval = (NFSD_LEASE_TIME/2) * HZ, + .to_exponential = 1, + }; struct rpc_program * program = &cb->cb_program; - struct rpc_stat * stat = &cb->cb_stat; - struct rpc_clnt * clnt; + struct rpc_create_args args = { + .protocol = IPPROTO_TCP, + .address = (struct sockaddr *)&addr, + .addrsize = sizeof(addr), + .timeout = &timeparms, + .servername = clp->cl_name.data, + .program = program, + .version = nfs_cb_version[1]->number, + .authflavor = RPC_AUTH_UNIX, /* XXX: need AUTH_GSS... */ + .flags = (RPC_CLNT_CREATE_NOPING), + }; struct rpc_message msg = { .rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL], .rpc_argp = clp, }; - char hostname[32]; int status; if (atomic_read(&cb->cb_set)) @@ -396,51 +408,27 @@ nfsd4_probe_callback(struct nfs4_client *clp) addr.sin_port = htons(cb->cb_port); addr.sin_addr.s_addr = htonl(cb->cb_addr); - /* Initialize timeout */ - timeparms.to_initval = (NFSD_LEASE_TIME/4) * HZ; - timeparms.to_retries = 0; - timeparms.to_maxval = (NFSD_LEASE_TIME/2) * HZ; - timeparms.to_exponential = 1; - - /* Create RPC transport */ - xprt = xprt_create_proto(IPPROTO_TCP, &addr, &timeparms); - if (IS_ERR(xprt)) { - dprintk("NFSD: couldn't create callback transport!\n"); - goto out_err; - } - /* Initialize rpc_program */ program->name = "nfs4_cb"; program->number = cb->cb_prog; program->nrvers = ARRAY_SIZE(nfs_cb_version); program->version = nfs_cb_version; - program->stats = stat; + program->stats = &cb->cb_stat; /* Initialize rpc_stat */ - memset(stat, 0, sizeof(struct rpc_stat)); - stat->program = program; + memset(program->stats, 0, sizeof(cb->cb_stat)); + program->stats->program = program; - /* Create RPC client - * - * XXX AUTH_UNIX only - need AUTH_GSS.... - */ - sprintf(hostname, "%u.%u.%u.%u", NIPQUAD(addr.sin_addr.s_addr)); - clnt = rpc_new_client(xprt, hostname, program, 1, RPC_AUTH_UNIX); - if (IS_ERR(clnt)) { + /* Create RPC client */ + cb->cb_client = rpc_create(&args); + if (!cb->cb_client) { dprintk("NFSD: couldn't create callback client\n"); goto out_err; } - clnt->cl_intr = 0; - clnt->cl_softrtry = 1; /* Kick rpciod, put the call on the wire. */ - - if (rpciod_up() != 0) { - dprintk("nfsd: couldn't start rpciod for callbacks!\n"); + if (rpciod_up() != 0) goto out_clnt; - } - - cb->cb_client = clnt; /* the task holds a reference to the nfs4_client struct */ atomic_inc(&clp->cl_count); @@ -448,7 +436,7 @@ nfsd4_probe_callback(struct nfs4_client *clp) msg.rpc_cred = nfsd4_lookupcred(clp,0); if (IS_ERR(msg.rpc_cred)) goto out_rpciod; - status = rpc_call_async(clnt, &msg, RPC_TASK_ASYNC, &nfs4_cb_null_ops, NULL); + status = rpc_call_async(cb->cb_client, &msg, RPC_TASK_ASYNC, &nfs4_cb_null_ops, NULL); put_rpccred(msg.rpc_cred); if (status != 0) { @@ -462,7 +450,7 @@ nfsd4_probe_callback(struct nfs4_client *clp) rpciod_down(); cb->cb_client = NULL; out_clnt: - rpc_shutdown_client(clnt); + rpc_shutdown_client(cb->cb_client); out_err: dprintk("NFSD: warning: no callback path to client %.*s\n", (int)clp->cl_name.len, clp->cl_name.data); From 9e1968c58d72c4b85d8a69bda1e194f9701fb224 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:21 -0400 Subject: [PATCH 45/74] SUNRPC: Convert RPC portmapper to use new rpc_create() API Replace xprt_create_proto/rpc_create_client calls in pmap_clnt.c with new rpc_create() API. Test plan: Repeated runs of Connectathon locking suite. Check network trace for proper PMAP calls and replies. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/pmap_clnt.c | 31 ++++++++++++++----------------- 1 file changed, 14 insertions(+), 17 deletions(-) diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index 523f0e825dea..f476f4df0f48 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -281,25 +281,22 @@ int rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay) static struct rpc_clnt *pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileged) { - struct rpc_xprt *xprt; - struct rpc_clnt *clnt; + struct rpc_create_args args = { + .protocol = proto, + .address = (struct sockaddr *)srvaddr, + .addrsize = sizeof(*srvaddr), + .servername = hostname, + .program = &pmap_program, + .version = RPC_PMAP_VERSION, + .authflavor = RPC_AUTH_UNIX, + .flags = (RPC_CLNT_CREATE_ONESHOT | + RPC_CLNT_CREATE_NOPING), + }; - xprt = xprt_create_proto(proto, srvaddr, NULL); - if (IS_ERR(xprt)) - return (struct rpc_clnt *)xprt; - xprt->ops->set_port(xprt, RPC_PMAP_PORT); - xprt_set_bound(xprt); + srvaddr->sin_port = htons(RPC_PMAP_PORT); if (!privileged) - xprt->resvport = 0; - - clnt = rpc_new_client(xprt, hostname, - &pmap_program, RPC_PMAP_VERSION, - RPC_AUTH_UNIX); - if (!IS_ERR(clnt)) { - clnt->cl_softrtry = 1; - clnt->cl_oneshot = 1; - } - return clnt; + args.flags |= RPC_CLNT_CREATE_NONPRIVPORT; + return rpc_create(&args); } /* From ff9aa5e56df60cc8565a93cc868fe25ae3f20e49 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:21 -0400 Subject: [PATCH 46/74] SUNRPC: Eliminate xprt_create_proto and rpc_create_client The two function call API for creating a new RPC client is now obsolete. Remove it. Also, remove an unnecessary check to see whether the caller is capable of using privileged network services. The kernel RPC client always uses a privileged ephemeral port by default; callers are responsible for checking the authority of users to make use of any RPC service, or for specifying that a nonprivileged port is acceptable. Test plan: Repeated runs of Connectathon locking suite. Check network trace to ensure correctness of NLM requests and replies. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- include/linux/sunrpc/clnt.h | 7 ---- include/linux/sunrpc/xprt.h | 1 - net/sunrpc/clnt.c | 42 +------------------- net/sunrpc/sunrpc_syms.c | 3 -- net/sunrpc/xprt.c | 79 ------------------------------------- net/sunrpc/xprtsock.c | 2 - 6 files changed, 1 insertion(+), 133 deletions(-) diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h index 7817ba82f1b2..f6d1d646ce05 100644 --- a/include/linux/sunrpc/clnt.h +++ b/include/linux/sunrpc/clnt.h @@ -91,13 +91,6 @@ struct rpc_procinfo { #ifdef __KERNEL__ -struct rpc_clnt *rpc_create_client(struct rpc_xprt *xprt, char *servname, - struct rpc_program *info, - u32 version, rpc_authflavor_t authflavor); -struct rpc_clnt *rpc_new_client(struct rpc_xprt *xprt, char *servname, - struct rpc_program *info, - u32 version, rpc_authflavor_t authflavor); - struct rpc_create_args { int protocol; struct sockaddr *address; diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index bc80fcfdd892..de4efea7c856 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -231,7 +231,6 @@ struct rpc_xprt { /* * Transport operations used by ULPs */ -struct rpc_xprt * xprt_create_proto(int proto, struct sockaddr_in *addr, struct rpc_timeout *to); void xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long incr); /* diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index dbb93bdf6cc9..428704dd5b3e 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -97,17 +97,7 @@ rpc_setup_pipedir(struct rpc_clnt *clnt, char *dir_name) } } -/* - * Create an RPC client - * FIXME: This should also take a flags argument (as in task->tk_flags). - * It's called (among others) from pmap_create_client, which may in - * turn be called by an async task. In this case, rpciod should not be - * made to sleep too long. - */ -struct rpc_clnt * -rpc_new_client(struct rpc_xprt *xprt, char *servname, - struct rpc_program *program, u32 vers, - rpc_authflavor_t flavor) +static struct rpc_clnt * rpc_new_client(struct rpc_xprt *xprt, char *servname, struct rpc_program *program, u32 vers, rpc_authflavor_t flavor) { struct rpc_version *version; struct rpc_clnt *clnt = NULL; @@ -253,36 +243,6 @@ struct rpc_clnt *rpc_create(struct rpc_create_args *args) } EXPORT_SYMBOL(rpc_create); -/** - * Create an RPC client - * @xprt - pointer to xprt struct - * @servname - name of server - * @info - rpc_program - * @version - rpc_program version - * @authflavor - rpc_auth flavour to use - * - * Creates an RPC client structure, then pings the server in order to - * determine if it is up, and if it supports this program and version. - * - * This function should never be called by asynchronous tasks such as - * the portmapper. - */ -struct rpc_clnt *rpc_create_client(struct rpc_xprt *xprt, char *servname, - struct rpc_program *info, u32 version, rpc_authflavor_t authflavor) -{ - struct rpc_clnt *clnt; - int err; - - clnt = rpc_new_client(xprt, servname, info, version, authflavor); - if (IS_ERR(clnt)) - return clnt; - err = rpc_ping(clnt, RPC_TASK_SOFT|RPC_TASK_NOINTR); - if (err == 0) - return clnt; - rpc_shutdown_client(clnt); - return ERR_PTR(err); -} - /* * This function clones the RPC client structure. It allows us to share the * same transport while varying parameters such as the authentication diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c index f38f939ce95f..26c0531d7e25 100644 --- a/net/sunrpc/sunrpc_syms.c +++ b/net/sunrpc/sunrpc_syms.c @@ -36,8 +36,6 @@ EXPORT_SYMBOL(rpc_wake_up_status); EXPORT_SYMBOL(rpc_release_task); /* RPC client functions */ -EXPORT_SYMBOL(rpc_create_client); -EXPORT_SYMBOL(rpc_new_client); EXPORT_SYMBOL(rpc_clone_client); EXPORT_SYMBOL(rpc_bind_new_program); EXPORT_SYMBOL(rpc_destroy_client); @@ -57,7 +55,6 @@ EXPORT_SYMBOL(rpc_queue_upcall); EXPORT_SYMBOL(rpc_mkpipe); /* Client transport */ -EXPORT_SYMBOL(xprt_create_proto); EXPORT_SYMBOL(xprt_set_timeout); /* Client credential cache */ diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index 17f56cfe2412..e4f64fb58ff2 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -962,85 +962,6 @@ struct rpc_xprt *xprt_create_transport(int proto, struct sockaddr *ap, size_t si return xprt; } -static struct rpc_xprt *xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to) -{ - int result; - struct rpc_xprt *xprt; - struct rpc_rqst *req; - - if ((xprt = kzalloc(sizeof(struct rpc_xprt), GFP_KERNEL)) == NULL) - return ERR_PTR(-ENOMEM); - - memcpy(&xprt->addr, ap, sizeof(*ap)); - xprt->addrlen = sizeof(*ap); - - switch (proto) { - case IPPROTO_UDP: - result = xs_setup_udp(xprt, to); - break; - case IPPROTO_TCP: - result = xs_setup_tcp(xprt, to); - break; - default: - printk(KERN_ERR "RPC: unrecognized transport protocol: %d\n", - proto); - result = -EIO; - break; - } - if (result) { - kfree(xprt); - return ERR_PTR(result); - } - - spin_lock_init(&xprt->transport_lock); - spin_lock_init(&xprt->reserve_lock); - - INIT_LIST_HEAD(&xprt->free); - INIT_LIST_HEAD(&xprt->recv); - INIT_WORK(&xprt->task_cleanup, xprt_autoclose, xprt); - init_timer(&xprt->timer); - xprt->timer.function = xprt_init_autodisconnect; - xprt->timer.data = (unsigned long) xprt; - xprt->last_used = jiffies; - xprt->cwnd = RPC_INITCWND; - - rpc_init_wait_queue(&xprt->binding, "xprt_binding"); - rpc_init_wait_queue(&xprt->pending, "xprt_pending"); - rpc_init_wait_queue(&xprt->sending, "xprt_sending"); - rpc_init_wait_queue(&xprt->resend, "xprt_resend"); - rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog"); - - /* initialize free list */ - for (req = &xprt->slot[xprt->max_reqs-1]; req >= &xprt->slot[0]; req--) - list_add(&req->rq_list, &xprt->free); - - xprt_init_xid(xprt); - - dprintk("RPC: created transport %p with %u slots\n", xprt, - xprt->max_reqs); - - return xprt; -} - -/** - * xprt_create_proto - create an RPC client transport - * @proto: requested transport protocol - * @sap: remote peer's address - * @to: timeout parameters for new transport - * - */ -struct rpc_xprt *xprt_create_proto(int proto, struct sockaddr_in *sap, struct rpc_timeout *to) -{ - struct rpc_xprt *xprt; - - xprt = xprt_setup(proto, sap, to); - if (IS_ERR(xprt)) - dprintk("RPC: xprt_create_proto failed\n"); - else - dprintk("RPC: xprt_create_proto created xprt %p\n", xprt); - return xprt; -} - /** * xprt_destroy - destroy an RPC transport, killing off all requests. * @xprt: transport to destroy diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 17179aa4c207..0b84fab68d7e 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -1376,7 +1376,6 @@ int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to) xprt->prot = IPPROTO_UDP; xprt->tsh_size = 0; - xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0; /* XXX: header size can vary due to auth type, IPv6, etc. */ xprt->max_payload = (1U << 16) - (MAX_HEADER << 3); @@ -1423,7 +1422,6 @@ int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to) xprt->prot = IPPROTO_TCP; xprt->tsh_size = sizeof(rpc_fraghdr) / sizeof(u32); - xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0; xprt->max_payload = RPC_MAX_FRAGMENT_SIZE; INIT_WORK(&xprt->connect_worker, xs_tcp_connect_worker, xprt); From b86acd501a34227e0ed2b2d54dc8002c1701ce17 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:22 -0400 Subject: [PATCH 47/74] SUNRPC: export new RPC client functions with _GPL This patch is optional. It has been suggested that the RPC client internal functions used by upper layer protocols (such as NFS) be exported via EXPORT_SYMBOL_GPL. This patch does that. Test plan: Compile kernel with CONFIG_NFS enabled as a module. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- net/sunrpc/clnt.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 428704dd5b3e..87efcd207f23 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -241,7 +241,7 @@ struct rpc_clnt *rpc_create(struct rpc_create_args *args) return clnt; } -EXPORT_SYMBOL(rpc_create); +EXPORT_SYMBOL_GPL(rpc_create); /* * This function clones the RPC client structure. It allows us to share the @@ -573,7 +573,7 @@ size_t rpc_peeraddr(struct rpc_clnt *clnt, struct sockaddr *buf, size_t bufsize) memcpy(buf, &clnt->cl_xprt->addr, bytes); return xprt->addrlen; } -EXPORT_SYMBOL(rpc_peeraddr); +EXPORT_SYMBOL_GPL(rpc_peeraddr); /** * rpc_peeraddr2str - return remote peer address in printable format @@ -586,7 +586,7 @@ char *rpc_peeraddr2str(struct rpc_clnt *clnt, enum rpc_display_format_t format) struct rpc_xprt *xprt = clnt->cl_xprt; return xprt->ops->print_addr(xprt, format); } -EXPORT_SYMBOL(rpc_peeraddr2str); +EXPORT_SYMBOL_GPL(rpc_peeraddr2str); void rpc_setbufsize(struct rpc_clnt *clnt, unsigned int sndsize, unsigned int rcvsize) @@ -608,7 +608,7 @@ size_t rpc_max_payload(struct rpc_clnt *clnt) { return clnt->cl_xprt->max_payload; } -EXPORT_SYMBOL(rpc_max_payload); +EXPORT_SYMBOL_GPL(rpc_max_payload); /** * rpc_force_rebind - force transport to check that remote port is unchanged @@ -620,7 +620,7 @@ void rpc_force_rebind(struct rpc_clnt *clnt) if (clnt->cl_autobind) xprt_clear_bound(clnt->cl_xprt); } -EXPORT_SYMBOL(rpc_force_rebind); +EXPORT_SYMBOL_GPL(rpc_force_rebind); /* * Restart an (async) RPC call. Usually called from within the From d3db90e270791b21cd00d3c094884bffa907cc9e Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:22 -0400 Subject: [PATCH 48/74] NFS: remove a no-longer-needed error check in nfs_symlink() In the early days of NFS, there was no duplicate reply cache on the server. Thus retransmitted non-idempotent requests often found that the request had already completed on the server. To avoid passing an unanticipated return code to unsuspecting applications, NFS clients would often shunt error codes that implied the request had been retried but already completed. Thanks to NFS over TCP, duplicate reply caches on the server, and network performance and reliability improvements, it is safe to remove such checks. Test plan: None. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 9b496ef4abea..084e8cb41c84 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1476,14 +1476,10 @@ dentry->d_parent->d_name.name, dentry->d_name.name); error = NFS_PROTO(dir)->symlink(dir, &dentry->d_name, &qsymname, &attr, &sym_fh, &sym_attr); nfs_end_data_update(dir); - if (!error) { + if (!error) error = nfs_instantiate(dentry, &sym_fh, &sym_attr); - } else { - if (error == -EEXIST) - printk("nfs_proc_symlink: %s/%s already exists??\n", - dentry->d_parent->d_name.name, dentry->d_name.name); + else d_drop(dentry); - } unlock_kernel(); return error; } From 4f390c152bc87165da4b1f5b7d870b46fb106d4e Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:22 -0400 Subject: [PATCH 49/74] NFS: Fix double d_drop in nfs_instantiate() error path If the LOOKUP or GETATTR in nfs_instantiate fail, nfs_instantiate will do a d_drop before returning. But some callers already do a d_drop in the case of an error return. Make certain we do only one d_drop in all error paths. This issue was introduced because over time, the symlink proc API diverged slightly from the create/mkdir/mknod proc API. To prevent other coding mistakes of this type, change the symlink proc API to be more like create/mkdir/mknod and move the nfs_instantiate call into the symlink proc routines so it is used in exactly the same way for create, mkdir, mknod, and symlink. Test plan: Connectathon, all versions of NFS. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 16 ++++------------ fs/nfs/nfs3proc.c | 26 ++++++++++++++++---------- fs/nfs/nfs4proc.c | 31 ++++++++++++++++--------------- fs/nfs/proc.c | 29 +++++++++++++++++++++-------- include/linux/nfs_xdr.h | 5 ++--- 5 files changed, 59 insertions(+), 48 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 084e8cb41c84..affd3ae52e55 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1147,23 +1147,20 @@ int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle, struct inode *dir = dentry->d_parent->d_inode; error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr); if (error) - goto out_err; + return error; } if (!(fattr->valid & NFS_ATTR_FATTR)) { struct nfs_server *server = NFS_SB(dentry->d_sb); error = server->nfs_client->rpc_ops->getattr(server, fhandle, fattr); if (error < 0) - goto out_err; + return error; } inode = nfs_fhget(dentry->d_sb, fhandle, fattr); error = PTR_ERR(inode); if (IS_ERR(inode)) - goto out_err; + return error; d_instantiate(dentry, inode); return 0; -out_err: - d_drop(dentry); - return error; } /* @@ -1448,8 +1445,6 @@ static int nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname) { struct iattr attr; - struct nfs_fattr sym_attr; - struct nfs_fh sym_fh; struct qstr qsymname; int error; @@ -1473,12 +1468,9 @@ dentry->d_parent->d_name.name, dentry->d_name.name); lock_kernel(); nfs_begin_data_update(dir); - error = NFS_PROTO(dir)->symlink(dir, &dentry->d_name, &qsymname, - &attr, &sym_fh, &sym_attr); + error = NFS_PROTO(dir)->symlink(dir, dentry, &qsymname, &attr); nfs_end_data_update(dir); if (!error) - error = nfs_instantiate(dentry, &sym_fh, &sym_attr); - else d_drop(dentry); unlock_kernel(); return error; diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c index 9e8258ece6fd..d85ac427c326 100644 --- a/fs/nfs/nfs3proc.c +++ b/fs/nfs/nfs3proc.c @@ -544,23 +544,23 @@ nfs3_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) } static int -nfs3_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path, - struct iattr *sattr, struct nfs_fh *fhandle, - struct nfs_fattr *fattr) +nfs3_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, + struct iattr *sattr) { - struct nfs_fattr dir_attr; + struct nfs_fh fhandle; + struct nfs_fattr fattr, dir_attr; struct nfs3_symlinkargs arg = { .fromfh = NFS_FH(dir), - .fromname = name->name, - .fromlen = name->len, + .fromname = dentry->d_name.name, + .fromlen = dentry->d_name.len, .topath = path->name, .tolen = path->len, .sattr = sattr }; struct nfs3_diropres res = { .dir_attr = &dir_attr, - .fh = fhandle, - .fattr = fattr + .fh = &fhandle, + .fattr = &fattr }; struct rpc_message msg = { .rpc_proc = &nfs3_procedures[NFS3PROC_SYMLINK], @@ -571,11 +571,17 @@ nfs3_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path, if (path->len > NFS3_MAXPATHLEN) return -ENAMETOOLONG; - dprintk("NFS call symlink %s -> %s\n", name->name, path->name); + + dprintk("NFS call symlink %s -> %s\n", dentry->d_name.name, + path->name); nfs_fattr_init(&dir_attr); - nfs_fattr_init(fattr); + nfs_fattr_init(&fattr); status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0); nfs_post_op_update_inode(dir, &dir_attr); + if (status != 0) + goto out; + status = nfs_instantiate(dentry, &fhandle, &fattr); +out: dprintk("NFS reply symlink: %d\n", status); return status; } diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index a825547e8214..2d18eac6bee5 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2084,24 +2084,24 @@ static int nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *n return err; } -static int _nfs4_proc_symlink(struct inode *dir, struct qstr *name, - struct qstr *path, struct iattr *sattr, struct nfs_fh *fhandle, - struct nfs_fattr *fattr) +static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry, + struct qstr *path, struct iattr *sattr) { struct nfs_server *server = NFS_SERVER(dir); - struct nfs_fattr dir_fattr; + struct nfs_fh fhandle; + struct nfs_fattr fattr, dir_fattr; struct nfs4_create_arg arg = { .dir_fh = NFS_FH(dir), .server = server, - .name = name, + .name = &dentry->d_name, .attrs = sattr, .ftype = NF4LNK, .bitmask = server->attr_bitmask, }; struct nfs4_create_res res = { .server = server, - .fh = fhandle, - .fattr = fattr, + .fh = &fhandle, + .fattr = &fattr, .dir_fattr = &dir_fattr, }; struct rpc_message msg = { @@ -2113,27 +2113,28 @@ static int _nfs4_proc_symlink(struct inode *dir, struct qstr *name, if (path->len > NFS4_MAXPATHLEN) return -ENAMETOOLONG; + arg.u.symlink = path; - nfs_fattr_init(fattr); + nfs_fattr_init(&fattr); nfs_fattr_init(&dir_fattr); status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0); - if (!status) + if (!status) { update_changeattr(dir, &res.dir_cinfo); - nfs_post_op_update_inode(dir, res.dir_fattr); + nfs_post_op_update_inode(dir, res.dir_fattr); + status = nfs_instantiate(dentry, &fhandle, &fattr); + } return status; } -static int nfs4_proc_symlink(struct inode *dir, struct qstr *name, - struct qstr *path, struct iattr *sattr, struct nfs_fh *fhandle, - struct nfs_fattr *fattr) +static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry, + struct qstr *path, struct iattr *sattr) { struct nfs4_exception exception = { }; int err; do { err = nfs4_handle_exception(NFS_SERVER(dir), - _nfs4_proc_symlink(dir, name, path, sattr, - fhandle, fattr), + _nfs4_proc_symlink(dir, dentry, path, sattr), &exception); } while (exception.retry); return err; diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c index 5a8b9407ee9a..0b507bf0f330 100644 --- a/fs/nfs/proc.c +++ b/fs/nfs/proc.c @@ -425,14 +425,15 @@ nfs_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) } static int -nfs_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path, - struct iattr *sattr, struct nfs_fh *fhandle, - struct nfs_fattr *fattr) +nfs_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, + struct iattr *sattr) { + struct nfs_fh fhandle; + struct nfs_fattr fattr; struct nfs_symlinkargs arg = { .fromfh = NFS_FH(dir), - .fromname = name->name, - .fromlen = name->len, + .fromname = dentry->d_name.name, + .fromlen = dentry->d_name.len, .topath = path->name, .tolen = path->len, .sattr = sattr @@ -445,11 +446,23 @@ nfs_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path, if (path->len > NFS2_MAXPATHLEN) return -ENAMETOOLONG; - dprintk("NFS call symlink %s -> %s\n", name->name, path->name); - nfs_fattr_init(fattr); - fhandle->size = 0; + + dprintk("NFS call symlink %s -> %s\n", dentry->d_name.name, + path->name); status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0); nfs_mark_for_revalidate(dir); + + /* + * V2 SYMLINK requests don't return any attributes. Setting the + * filehandle size to zero indicates to nfs_instantiate that it + * should fill in the data with a LOOKUP call on the wire. + */ + if (status == 0) { + nfs_fattr_init(&fattr); + fhandle.size = 0; + status = nfs_instantiate(dentry, &fhandle, &fattr); + } + dprintk("NFS reply symlink: %d\n", status); return status; } diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index 0f33e621892f..ddf5d75e97a2 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -793,9 +793,8 @@ struct nfs_rpc_ops { int (*rename) (struct inode *, struct qstr *, struct inode *, struct qstr *); int (*link) (struct inode *, struct inode *, struct qstr *); - int (*symlink) (struct inode *, struct qstr *, struct qstr *, - struct iattr *, struct nfs_fh *, - struct nfs_fattr *); + int (*symlink) (struct inode *, struct dentry *, struct qstr *, + struct iattr *); int (*mkdir) (struct inode *, struct dentry *, struct iattr *); int (*rmdir) (struct inode *, struct qstr *); int (*readdir) (struct dentry *, struct rpc_cred *, From 873101b33776780d32610fc4c90c7358a5e98f51 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:23 -0400 Subject: [PATCH 50/74] NFS: copy symlinks into page cache before sending NFS SYMLINK request Currently the NFS client does not cache symlinks it creates. They get cached only when the NFS client reads them back from the server. Copy the symlink into the page cache before sending it. Test plan: Connectathon, all NFS versions. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 86 +++++++++++++++++++++++++++++++++++++++++----------- 1 file changed, 68 insertions(+), 18 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index affd3ae52e55..b483e5d206cb 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -30,6 +30,7 @@ #include #include #include +#include #include #include @@ -1441,39 +1442,88 @@ static int nfs_unlink(struct inode *dir, struct dentry *dentry) return error; } -static int -nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname) +/* + * To create a symbolic link, most file systems instantiate a new inode, + * add a page to it containing the path, then write it out to the disk + * using prepare_write/commit_write. + * + * Unfortunately the NFS client can't create the in-core inode first + * because it needs a file handle to create an in-core inode (see + * fs/nfs/inode.c:nfs_fhget). We only have a file handle *after* the + * symlink request has completed on the server. + * + * So instead we allocate a raw page, copy the symname into it, then do + * the SYMLINK request with the page as the buffer. If it succeeds, we + * now have a new file handle and can instantiate an in-core NFS inode + * and move the raw page into its mapping. + */ +static int nfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname) { + struct pagevec lru_pvec; + struct page *page; + char *kaddr; struct iattr attr; - struct qstr qsymname; + unsigned int pathlen = strlen(symname); + struct qstr qsymname = { + .name = symname, + .len = pathlen, + }; int error; dfprintk(VFS, "NFS: symlink(%s/%ld, %s, %s)\n", dir->i_sb->s_id, dir->i_ino, dentry->d_name.name, symname); -#ifdef NFS_PARANOIA -if (dentry->d_inode) -printk("nfs_proc_symlink: %s/%s not negative!\n", -dentry->d_parent->d_name.name, dentry->d_name.name); -#endif - /* - * Fill in the sattr for the call. - * Note: SunOS 4.1.2 crashes if the mode isn't initialized! - */ - attr.ia_valid = ATTR_MODE; - attr.ia_mode = S_IFLNK | S_IRWXUGO; + if (pathlen > PAGE_SIZE) + return -ENAMETOOLONG; - qsymname.name = symname; - qsymname.len = strlen(symname); + attr.ia_mode = S_IFLNK | S_IRWXUGO; + attr.ia_valid = ATTR_MODE; lock_kernel(); + + page = alloc_page(GFP_KERNEL); + if (!page) { + unlock_kernel(); + return -ENOMEM; + } + + kaddr = kmap_atomic(page, KM_USER0); + memcpy(kaddr, symname, pathlen); + if (pathlen < PAGE_SIZE) + memset(kaddr + pathlen, 0, PAGE_SIZE - pathlen); + kunmap_atomic(kaddr, KM_USER0); + + /* XXX: eventually this will pass in {page, pathlen}, + * instead of qsymname; need XDR changes for that */ nfs_begin_data_update(dir); error = NFS_PROTO(dir)->symlink(dir, dentry, &qsymname, &attr); nfs_end_data_update(dir); - if (!error) + if (error != 0) { + dfprintk(VFS, "NFS: symlink(%s/%ld, %s, %s) error %d\n", + dir->i_sb->s_id, dir->i_ino, + dentry->d_name.name, symname, error); d_drop(dentry); + __free_page(page); + unlock_kernel(); + return error; + } + + /* + * No big deal if we can't add this page to the page cache here. + * READLINK will get the missing page from the server if needed. + */ + pagevec_init(&lru_pvec, 0); + if (!add_to_page_cache(page, dentry->d_inode->i_mapping, 0, + GFP_KERNEL)) { + if (!pagevec_add(&lru_pvec, page)) + __pagevec_lru_add(&lru_pvec); + SetPageUptodate(page); + unlock_page(page); + } else + __free_page(page); + unlock_kernel(); - return error; + return 0; } static int From 94a6d75320b3681e6e728b70e18bd186cb55e682 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Tue, 22 Aug 2006 20:06:23 -0400 Subject: [PATCH 51/74] NFS: Use cached page as buffer for NFS symlink requests Now that we have a copy of the symlink path in the page cache, we can pass a struct page down to the XDR routines instead of a string buffer. Test plan: Connectathon, all NFS versions. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 8 +------- fs/nfs/nfs2xdr.c | 21 ++++++++++++++++++--- fs/nfs/nfs3proc.c | 14 +++++++------- fs/nfs/nfs3xdr.c | 7 +++++-- fs/nfs/nfs4proc.c | 12 +++++++----- fs/nfs/nfs4xdr.c | 8 ++++---- fs/nfs/proc.c | 14 +++++++------- include/linux/nfs_xdr.h | 17 ++++++++++------- 8 files changed, 59 insertions(+), 42 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index b483e5d206cb..51328ae640dd 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1464,10 +1464,6 @@ static int nfs_symlink(struct inode *dir, struct dentry *dentry, const char *sym char *kaddr; struct iattr attr; unsigned int pathlen = strlen(symname); - struct qstr qsymname = { - .name = symname, - .len = pathlen, - }; int error; dfprintk(VFS, "NFS: symlink(%s/%ld, %s, %s)\n", dir->i_sb->s_id, @@ -1493,10 +1489,8 @@ static int nfs_symlink(struct inode *dir, struct dentry *dentry, const char *sym memset(kaddr + pathlen, 0, PAGE_SIZE - pathlen); kunmap_atomic(kaddr, KM_USER0); - /* XXX: eventually this will pass in {page, pathlen}, - * instead of qsymname; need XDR changes for that */ nfs_begin_data_update(dir); - error = NFS_PROTO(dir)->symlink(dir, dentry, &qsymname, &attr); + error = NFS_PROTO(dir)->symlink(dir, dentry, page, pathlen, &attr); nfs_end_data_update(dir); if (error != 0) { dfprintk(VFS, "NFS: symlink(%s/%ld, %s, %s) error %d\n", diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c index 67391eef6b93..b49501fc0a79 100644 --- a/fs/nfs/nfs2xdr.c +++ b/fs/nfs/nfs2xdr.c @@ -51,7 +51,7 @@ #define NFS_createargs_sz (NFS_diropargs_sz+NFS_sattr_sz) #define NFS_renameargs_sz (NFS_diropargs_sz+NFS_diropargs_sz) #define NFS_linkargs_sz (NFS_fhandle_sz+NFS_diropargs_sz) -#define NFS_symlinkargs_sz (NFS_diropargs_sz+NFS_path_sz+NFS_sattr_sz) +#define NFS_symlinkargs_sz (NFS_diropargs_sz+1+NFS_sattr_sz) #define NFS_readdirargs_sz (NFS_fhandle_sz+2) #define NFS_attrstat_sz (1+NFS_fattr_sz) @@ -351,11 +351,26 @@ nfs_xdr_linkargs(struct rpc_rqst *req, u32 *p, struct nfs_linkargs *args) static int nfs_xdr_symlinkargs(struct rpc_rqst *req, u32 *p, struct nfs_symlinkargs *args) { + struct xdr_buf *sndbuf = &req->rq_snd_buf; + size_t pad; + p = xdr_encode_fhandle(p, args->fromfh); p = xdr_encode_array(p, args->fromname, args->fromlen); - p = xdr_encode_array(p, args->topath, args->tolen); + *p++ = htonl(args->pathlen); + sndbuf->len = xdr_adjust_iovec(sndbuf->head, p); + + xdr_encode_pages(sndbuf, args->pages, 0, args->pathlen); + + /* + * xdr_encode_pages may have added a few bytes to ensure the + * pathname ends on a 4-byte boundary. Start encoding the + * attributes after the pad bytes. + */ + pad = sndbuf->tail->iov_len; + if (pad > 0) + p++; p = xdr_encode_sattr(p, args->sattr); - req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); + sndbuf->len += xdr_adjust_iovec(sndbuf->tail, p) - pad; return 0; } diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c index d85ac427c326..f8688eaa0001 100644 --- a/fs/nfs/nfs3proc.c +++ b/fs/nfs/nfs3proc.c @@ -544,8 +544,8 @@ nfs3_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) } static int -nfs3_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, - struct iattr *sattr) +nfs3_proc_symlink(struct inode *dir, struct dentry *dentry, struct page *page, + unsigned int len, struct iattr *sattr) { struct nfs_fh fhandle; struct nfs_fattr fattr, dir_attr; @@ -553,8 +553,8 @@ nfs3_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, .fromfh = NFS_FH(dir), .fromname = dentry->d_name.name, .fromlen = dentry->d_name.len, - .topath = path->name, - .tolen = path->len, + .pages = &page, + .pathlen = len, .sattr = sattr }; struct nfs3_diropres res = { @@ -569,11 +569,11 @@ nfs3_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, }; int status; - if (path->len > NFS3_MAXPATHLEN) + if (len > NFS3_MAXPATHLEN) return -ENAMETOOLONG; - dprintk("NFS call symlink %s -> %s\n", dentry->d_name.name, - path->name); + dprintk("NFS call symlink %s\n", dentry->d_name.name); + nfs_fattr_init(&dir_attr); nfs_fattr_init(&fattr); status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0); diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c index 0250269e9753..16556fa4effb 100644 --- a/fs/nfs/nfs3xdr.c +++ b/fs/nfs/nfs3xdr.c @@ -56,7 +56,7 @@ #define NFS3_writeargs_sz (NFS3_fh_sz+5) #define NFS3_createargs_sz (NFS3_diropargs_sz+NFS3_sattr_sz) #define NFS3_mkdirargs_sz (NFS3_diropargs_sz+NFS3_sattr_sz) -#define NFS3_symlinkargs_sz (NFS3_diropargs_sz+NFS3_path_sz+NFS3_sattr_sz) +#define NFS3_symlinkargs_sz (NFS3_diropargs_sz+1+NFS3_sattr_sz) #define NFS3_mknodargs_sz (NFS3_diropargs_sz+2+NFS3_sattr_sz) #define NFS3_renameargs_sz (NFS3_diropargs_sz+NFS3_diropargs_sz) #define NFS3_linkargs_sz (NFS3_fh_sz+NFS3_diropargs_sz) @@ -398,8 +398,11 @@ nfs3_xdr_symlinkargs(struct rpc_rqst *req, u32 *p, struct nfs3_symlinkargs *args p = xdr_encode_fhandle(p, args->fromfh); p = xdr_encode_array(p, args->fromname, args->fromlen); p = xdr_encode_sattr(p, args->sattr); - p = xdr_encode_array(p, args->topath, args->tolen); + *p++ = htonl(args->pathlen); req->rq_slen = xdr_adjust_iovec(req->rq_svec, p); + + /* Copy the page */ + xdr_encode_pages(&req->rq_snd_buf, args->pages, 0, args->pathlen); return 0; } diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 2d18eac6bee5..7f60beb40df3 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2085,7 +2085,7 @@ static int nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *n } static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry, - struct qstr *path, struct iattr *sattr) + struct page *page, unsigned int len, struct iattr *sattr) { struct nfs_server *server = NFS_SERVER(dir); struct nfs_fh fhandle; @@ -2111,10 +2111,11 @@ static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry, }; int status; - if (path->len > NFS4_MAXPATHLEN) + if (len > NFS4_MAXPATHLEN) return -ENAMETOOLONG; - arg.u.symlink = path; + arg.u.symlink.pages = &page; + arg.u.symlink.len = len; nfs_fattr_init(&fattr); nfs_fattr_init(&dir_fattr); @@ -2128,13 +2129,14 @@ static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry, } static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry, - struct qstr *path, struct iattr *sattr) + struct page *page, unsigned int len, struct iattr *sattr) { struct nfs4_exception exception = { }; int err; do { err = nfs4_handle_exception(NFS_SERVER(dir), - _nfs4_proc_symlink(dir, dentry, path, sattr), + _nfs4_proc_symlink(dir, dentry, page, + len, sattr), &exception); } while (exception.retry); return err; diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 99926067eca4..3dd413f52da1 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -128,7 +128,7 @@ static int nfs4_stat_to_errno(int); #define decode_link_maxsz (op_decode_hdr_maxsz + 5) #define encode_symlink_maxsz (op_encode_hdr_maxsz + \ 1 + nfs4_name_maxsz + \ - nfs4_path_maxsz + \ + 1 + \ nfs4_fattr_maxsz) #define decode_symlink_maxsz (op_decode_hdr_maxsz + 8) #define encode_create_maxsz (op_encode_hdr_maxsz + \ @@ -673,9 +673,9 @@ static int encode_create(struct xdr_stream *xdr, const struct nfs4_create_arg *c switch (create->ftype) { case NF4LNK: - RESERVE_SPACE(4 + create->u.symlink->len); - WRITE32(create->u.symlink->len); - WRITEMEM(create->u.symlink->name, create->u.symlink->len); + RESERVE_SPACE(4); + WRITE32(create->u.symlink.len); + xdr_write_pages(xdr, create->u.symlink.pages, 0, create->u.symlink.len); break; case NF4BLK: case NF4CHR: diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c index 0b507bf0f330..630e50647bbb 100644 --- a/fs/nfs/proc.c +++ b/fs/nfs/proc.c @@ -425,8 +425,8 @@ nfs_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) } static int -nfs_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, - struct iattr *sattr) +nfs_proc_symlink(struct inode *dir, struct dentry *dentry, struct page *page, + unsigned int len, struct iattr *sattr) { struct nfs_fh fhandle; struct nfs_fattr fattr; @@ -434,8 +434,8 @@ nfs_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, .fromfh = NFS_FH(dir), .fromname = dentry->d_name.name, .fromlen = dentry->d_name.len, - .topath = path->name, - .tolen = path->len, + .pages = &page, + .pathlen = len, .sattr = sattr }; struct rpc_message msg = { @@ -444,11 +444,11 @@ nfs_proc_symlink(struct inode *dir, struct dentry *dentry, struct qstr *path, }; int status; - if (path->len > NFS2_MAXPATHLEN) + if (len > NFS2_MAXPATHLEN) return -ENAMETOOLONG; - dprintk("NFS call symlink %s -> %s\n", dentry->d_name.name, - path->name); + dprintk("NFS call symlink %s\n", dentry->d_name.name); + status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0); nfs_mark_for_revalidate(dir); diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index ddf5d75e97a2..dc5397d9d23c 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -358,8 +358,8 @@ struct nfs_symlinkargs { struct nfs_fh * fromfh; const char * fromname; unsigned int fromlen; - const char * topath; - unsigned int tolen; + struct page ** pages; + unsigned int pathlen; struct iattr * sattr; }; @@ -434,8 +434,8 @@ struct nfs3_symlinkargs { struct nfs_fh * fromfh; const char * fromname; unsigned int fromlen; - const char * topath; - unsigned int tolen; + struct page ** pages; + unsigned int pathlen; struct iattr * sattr; }; @@ -533,7 +533,10 @@ struct nfs4_accessres { struct nfs4_create_arg { u32 ftype; union { - struct qstr * symlink; /* NF4LNK */ + struct { + struct page ** pages; + unsigned int len; + } symlink; /* NF4LNK */ struct { u32 specdata1; u32 specdata2; @@ -793,8 +796,8 @@ struct nfs_rpc_ops { int (*rename) (struct inode *, struct qstr *, struct inode *, struct qstr *); int (*link) (struct inode *, struct inode *, struct qstr *); - int (*symlink) (struct inode *, struct dentry *, struct qstr *, - struct iattr *); + int (*symlink) (struct inode *, struct dentry *, struct page *, + unsigned int, struct iattr *); int (*mkdir) (struct inode *, struct dentry *, struct iattr *); int (*rmdir) (struct inode *, struct qstr *); int (*readdir) (struct dentry *, struct rpc_cred *, From 275a082fe9308e710324e26ccb5363c53d8fd45f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 22 Aug 2006 20:06:24 -0400 Subject: [PATCH 52/74] Add a real API for dealing with blk_congestion_wait() Signed-off-by: Trond Myklebust --- block/ll_rw_blk.c | 12 ++++++++++++ fs/nfs/write.c | 1 + include/linux/blkdev.h | 1 + include/linux/writeback.h | 1 + mm/page-writeback.c | 9 +++++++++ 5 files changed, 24 insertions(+) diff --git a/block/ll_rw_blk.c b/block/ll_rw_blk.c index ddd9253f9d55..dcbd6ff1fa33 100644 --- a/block/ll_rw_blk.c +++ b/block/ll_rw_blk.c @@ -2734,6 +2734,18 @@ long blk_congestion_wait(int rw, long timeout) EXPORT_SYMBOL(blk_congestion_wait); +/** + * blk_congestion_end - wake up sleepers on a congestion queue + * @rw: READ or WRITE + */ +void blk_congestion_end(int rw) +{ + wait_queue_head_t *wqh = &congestion_wqh[rw]; + + if (waitqueue_active(wqh)) + wake_up(wqh); +} + /* * Has to be called with the request spinlock acquired */ diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 453d44666ea5..38ba5c09af08 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -396,6 +396,7 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc) out: clear_bit(BDI_write_congested, &bdi->state); wake_up_all(&nfs_write_congestion); + writeback_congestion_end(); return err; } diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index aafe82788b4e..96c9040c00a8 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -746,6 +746,7 @@ extern void blk_queue_free_tags(request_queue_t *); extern int blk_queue_resize_tags(request_queue_t *, int); extern void blk_queue_invalidate_tags(request_queue_t *); extern long blk_congestion_wait(int rw, long timeout); +extern void blk_congestion_end(int rw); extern void blk_rq_bio_prep(request_queue_t *, struct request *, struct bio *); extern int blkdev_issue_flush(struct block_device *, sector_t *); diff --git a/include/linux/writeback.h b/include/linux/writeback.h index 9e38b566d0e7..0422036af4eb 100644 --- a/include/linux/writeback.h +++ b/include/linux/writeback.h @@ -85,6 +85,7 @@ int wakeup_pdflush(long nr_pages); void laptop_io_completion(void); void laptop_sync_completion(void); void throttle_vm_writeout(void); +void writeback_congestion_end(void); /* These are exported to sysctl. */ extern int dirty_background_ratio; diff --git a/mm/page-writeback.c b/mm/page-writeback.c index e630188ccc40..77a0bc4e261a 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -802,6 +802,15 @@ int test_set_page_writeback(struct page *page) } EXPORT_SYMBOL(test_set_page_writeback); +/* + * Wakes up tasks that are being throttled due to writeback congestion + */ +void writeback_congestion_end(void) +{ + blk_congestion_end(WRITE); +} +EXPORT_SYMBOL(writeback_congestion_end); + /* * Return true if any of the pages in the mapping are marged with the * passed tag. From 5dd3177ae5012c1e2ad7a9ffdbd0e0d0de2f60e4 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 24 Aug 2006 01:03:05 -0400 Subject: [PATCH 53/74] NFSv4: Fix a use-after-free issue with the nfs server. Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 36 +++++++++++++++++++++--------------- fs/nfs/nfs4renewd.c | 1 + fs/nfs/super.c | 8 +++++--- include/linux/nfs_fs_sb.h | 1 + 4 files changed, 28 insertions(+), 18 deletions(-) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 12941a8a6d75..f1ff2aec2ca5 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -164,6 +164,26 @@ static struct nfs_client *nfs_alloc_client(const char *hostname, return NULL; } +static void nfs4_shutdown_client(struct nfs_client *clp) +{ +#ifdef CONFIG_NFS_V4 + if (__test_and_clear_bit(NFS_CS_RENEWD, &clp->cl_res_state)) + nfs4_kill_renewd(clp); + while (!list_empty(&clp->cl_unused)) { + struct nfs4_state_owner *sp; + + sp = list_entry(clp->cl_unused.next, + struct nfs4_state_owner, + so_list); + list_del(&sp->so_list); + kfree(sp); + } + BUG_ON(!list_empty(&clp->cl_state_owners)); + if (__test_and_clear_bit(NFS_CS_IDMAP, &clp->cl_res_state)) + nfs_idmap_delete(clp); +#endif +} + /* * Destroy a shared client record */ @@ -171,21 +191,7 @@ static void nfs_free_client(struct nfs_client *clp) { dprintk("--> nfs_free_client(%d)\n", clp->cl_nfsversion); -#ifdef CONFIG_NFS_V4 - if (__test_and_clear_bit(NFS_CS_IDMAP, &clp->cl_res_state)) { - while (!list_empty(&clp->cl_unused)) { - struct nfs4_state_owner *sp; - - sp = list_entry(clp->cl_unused.next, - struct nfs4_state_owner, - so_list); - list_del(&sp->so_list); - kfree(sp); - } - BUG_ON(!list_empty(&clp->cl_state_owners)); - nfs_idmap_delete(clp); - } -#endif + nfs4_shutdown_client(clp); /* -EIO all pending I/O */ if (!IS_ERR(clp->cl_rpcclient)) diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c index f2c893690ac4..7b6df1852e75 100644 --- a/fs/nfs/nfs4renewd.c +++ b/fs/nfs/nfs4renewd.c @@ -121,6 +121,7 @@ nfs4_schedule_state_renewal(struct nfs_client *clp) __FUNCTION__, (timeout + HZ - 1) / HZ); cancel_delayed_work(&clp->cl_renewd); schedule_delayed_work(&clp->cl_renewd, timeout); + set_bit(NFS_CS_RENEWD, &clp->cl_res_state); spin_unlock(&clp->cl_lock); } diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 97cfb143e09f..665949d27798 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -883,13 +883,15 @@ static int nfs4_get_sb(struct file_system_type *fs_type, goto out_free; } + if (s->s_fs_info != server) { + nfs_free_server(server); + server = NULL; + } + if (!s->s_root) { /* initial superblock/root creation */ s->s_flags = flags; - nfs4_fill_super(s); - } else { - nfs_free_server(server); } mntroot = nfs4_get_root(s, &mntfh); diff --git a/include/linux/nfs_fs_sb.h b/include/linux/nfs_fs_sb.h index 6d0be0efd1b5..7ccfc7ef0a83 100644 --- a/include/linux/nfs_fs_sb.h +++ b/include/linux/nfs_fs_sb.h @@ -19,6 +19,7 @@ struct nfs_client { #define NFS_CS_RPCIOD 0 /* - rpciod started */ #define NFS_CS_CALLBACK 1 /* - callback started */ #define NFS_CS_IDMAP 2 /* - idmap started */ +#define NFS_CS_RENEWD 3 /* - renewd started */ struct sockaddr_in cl_addr; /* server identifier */ char * cl_hostname; /* hostname of server */ struct list_head cl_share_link; /* link in global client list */ From 158998b6fe36f6acef087f574c96d44713499cc9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 24 Aug 2006 01:03:17 -0400 Subject: [PATCH 54/74] SUNRPC: Make rpc_mkpipe() take the parent dentry as an argument Signed-off-by: Trond Myklebust --- fs/nfs/idmap.c | 6 +---- include/linux/sunrpc/rpc_pipe_fs.h | 2 +- net/sunrpc/auth_gss/auth_gss.c | 7 ++---- net/sunrpc/rpc_pipe.c | 38 ++++++++++++++++++------------ 4 files changed, 27 insertions(+), 26 deletions(-) diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c index f96dfac7dc9a..82ad7110a1c0 100644 --- a/fs/nfs/idmap.c +++ b/fs/nfs/idmap.c @@ -84,7 +84,6 @@ struct idmap_hashtable { }; struct idmap { - char idmap_path[48]; struct dentry *idmap_dentry; wait_queue_head_t idmap_wq; struct idmap_msg idmap_im; @@ -119,10 +118,7 @@ nfs_idmap_new(struct nfs_client *clp) if ((idmap = kzalloc(sizeof(*idmap), GFP_KERNEL)) == NULL) return -ENOMEM; - snprintf(idmap->idmap_path, sizeof(idmap->idmap_path), - "%s/idmap", clp->cl_rpcclient->cl_pathname); - - idmap->idmap_dentry = rpc_mkpipe(idmap->idmap_path, + idmap->idmap_dentry = rpc_mkpipe(clp->cl_rpcclient->cl_dentry, "idmap", idmap, &idmap_upcall_ops, 0); if (IS_ERR(idmap->idmap_dentry)) { error = PTR_ERR(idmap->idmap_dentry); diff --git a/include/linux/sunrpc/rpc_pipe_fs.h b/include/linux/sunrpc/rpc_pipe_fs.h index a481472c9484..a2eb9b4a9de3 100644 --- a/include/linux/sunrpc/rpc_pipe_fs.h +++ b/include/linux/sunrpc/rpc_pipe_fs.h @@ -43,7 +43,7 @@ extern int rpc_queue_upcall(struct inode *, struct rpc_pipe_msg *); extern struct dentry *rpc_mkdir(char *, struct rpc_clnt *); extern int rpc_rmdir(struct dentry *); -extern struct dentry *rpc_mkpipe(char *, void *, struct rpc_pipe_ops *, int flags); +extern struct dentry *rpc_mkpipe(struct dentry *, const char *, void *, struct rpc_pipe_ops *, int flags); extern int rpc_unlink(struct dentry *); extern struct vfsmount *rpc_get_mount(void); extern void rpc_put_mount(void); diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c index ef1cf5b476c8..6eed3e166ba3 100644 --- a/net/sunrpc/auth_gss/auth_gss.c +++ b/net/sunrpc/auth_gss/auth_gss.c @@ -88,7 +88,6 @@ struct gss_auth { struct list_head upcalls; struct rpc_clnt *client; struct dentry *dentry; - char path[48]; spinlock_t lock; }; @@ -690,10 +689,8 @@ gss_create(struct rpc_clnt *clnt, rpc_authflavor_t flavor) if (err) goto err_put_mech; - snprintf(gss_auth->path, sizeof(gss_auth->path), "%s/%s", - clnt->cl_pathname, - gss_auth->mech->gm_name); - gss_auth->dentry = rpc_mkpipe(gss_auth->path, clnt, &gss_upcall_ops, RPC_PIPE_WAIT_FOR_OPEN); + gss_auth->dentry = rpc_mkpipe(clnt->cl_dentry, gss_auth->mech->gm_name, + clnt, &gss_upcall_ops, RPC_PIPE_WAIT_FOR_OPEN); if (IS_ERR(gss_auth->dentry)) { err = PTR_ERR(gss_auth->dentry); goto err_put_mech; diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c index c21dc07f2a8c..11ec12a09d70 100644 --- a/net/sunrpc/rpc_pipe.c +++ b/net/sunrpc/rpc_pipe.c @@ -621,17 +621,13 @@ __rpc_rmdir(struct inode *dir, struct dentry *dentry) } static struct dentry * -rpc_lookup_negative(char *path, struct nameidata *nd) +rpc_lookup_create(struct dentry *parent, const char *name, int len) { + struct inode *dir = parent->d_inode; struct dentry *dentry; - struct inode *dir; - int error; - if ((error = rpc_lookup_parent(path, nd)) != 0) - return ERR_PTR(error); - dir = nd->dentry->d_inode; mutex_lock_nested(&dir->i_mutex, I_MUTEX_PARENT); - dentry = lookup_one_len(nd->last.name, nd->dentry, nd->last.len); + dentry = lookup_one_len(name, parent, len); if (IS_ERR(dentry)) goto out_err; if (dentry->d_inode) { @@ -642,7 +638,20 @@ rpc_lookup_negative(char *path, struct nameidata *nd) return dentry; out_err: mutex_unlock(&dir->i_mutex); - rpc_release_path(nd); + return dentry; +} + +static struct dentry * +rpc_lookup_negative(char *path, struct nameidata *nd) +{ + struct dentry *dentry; + int error; + + if ((error = rpc_lookup_parent(path, nd)) != 0) + return ERR_PTR(error); + dentry = rpc_lookup_create(nd->dentry, nd->last.name, nd->last.len); + if (IS_ERR(dentry)) + rpc_release_path(nd); return dentry; } @@ -701,17 +710,16 @@ rpc_rmdir(struct dentry *dentry) } struct dentry * -rpc_mkpipe(char *path, void *private, struct rpc_pipe_ops *ops, int flags) +rpc_mkpipe(struct dentry *parent, const char *name, void *private, struct rpc_pipe_ops *ops, int flags) { - struct nameidata nd; struct dentry *dentry; struct inode *dir, *inode; struct rpc_inode *rpci; - dentry = rpc_lookup_negative(path, &nd); + dentry = rpc_lookup_create(parent, name, strlen(name)); if (IS_ERR(dentry)) return dentry; - dir = nd.dentry->d_inode; + dir = parent->d_inode; inode = rpc_get_inode(dir->i_sb, S_IFSOCK | S_IRUSR | S_IWUSR); if (!inode) goto err_dput; @@ -726,13 +734,13 @@ rpc_mkpipe(char *path, void *private, struct rpc_pipe_ops *ops, int flags) dget(dentry); out: mutex_unlock(&dir->i_mutex); - rpc_release_path(&nd); return dentry; err_dput: dput(dentry); dentry = ERR_PTR(-ENOMEM); - printk(KERN_WARNING "%s: %s() failed to create pipe %s (errno = %d)\n", - __FILE__, __FUNCTION__, path, -ENOMEM); + printk(KERN_WARNING "%s: %s() failed to create pipe %s/%s (errno = %d)\n", + __FILE__, __FUNCTION__, parent->d_name.name, name, + -ENOMEM); goto out; } From 6daabf1b04c89f1fbd8eab5450261360943c8e20 Mon Sep 17 00:00:00 2001 From: David Howells Date: Thu, 24 Aug 2006 15:44:16 -0400 Subject: [PATCH 55/74] NFS: Fix up compiler warnings on 64-bit platforms in client.c Fix up warnings from compiling on ppc64. Signed-Off-By: David Howells Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index f1ff2aec2ca5..a4aa47913a5c 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -836,7 +836,9 @@ struct nfs_server *nfs_create_server(const struct nfs_mount_data *data, } memcpy(&server->fsid, &fattr.fsid, sizeof(server->fsid)); - dprintk("Server FSID: %llx:%llx\n", server->fsid.major, server->fsid.minor); + dprintk("Server FSID: %llx:%llx\n", + (unsigned long long) server->fsid.major, + (unsigned long long) server->fsid.minor); BUG_ON(!server->nfs_client); BUG_ON(!server->nfs_client->rpc_ops); @@ -1002,7 +1004,9 @@ struct nfs_server *nfs4_create_server(const struct nfs4_mount_data *data, if (error < 0) goto error; - dprintk("Server FSID: %llx:%llx\n", server->fsid.major, server->fsid.minor); + dprintk("Server FSID: %llx:%llx\n", + (unsigned long long) server->fsid.major, + (unsigned long long) server->fsid.minor); dprintk("Mount FH: %d\n", mntfh->size); error = nfs_probe_fsinfo(server, mntfh, &fattr); @@ -1074,7 +1078,8 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data, goto error; dprintk("Referral FSID: %llx:%llx\n", - server->fsid.major, server->fsid.minor); + (unsigned long long) server->fsid.major, + (unsigned long long) server->fsid.minor); spin_lock(&nfs_client_lock); list_add_tail(&server->client_link, &server->nfs_client->cl_superblocks); @@ -1106,7 +1111,8 @@ struct nfs_server *nfs_clone_server(struct nfs_server *source, int error; dprintk("--> nfs_clone_server(,%llx:%llx,)\n", - fattr->fsid.major, fattr->fsid.minor); + (unsigned long long) fattr->fsid.major, + (unsigned long long) fattr->fsid.minor); server = nfs_alloc_server(); if (!server) @@ -1131,7 +1137,8 @@ struct nfs_server *nfs_clone_server(struct nfs_server *source, goto out_free_server; dprintk("Cloned FSID: %llx:%llx\n", - server->fsid.major, server->fsid.minor); + (unsigned long long) server->fsid.major, + (unsigned long long) server->fsid.minor); error = nfs_start_lockd(server); if (error < 0) @@ -1375,7 +1382,8 @@ static int nfs_volume_list_show(struct seq_file *m, void *v) MAJOR(server->s_dev), MINOR(server->s_dev)); snprintf(fsid, 17, "%llx:%llx", - server->fsid.major, server->fsid.minor); + (unsigned long long) server->fsid.major, + (unsigned long long) server->fsid.minor); seq_printf(m, "v%d %02x%02x%02x%02x %4hx %-7s %-17s\n", clp->cl_nfsversion, From 058ad9cbf14b3c7480d01b20280cb4d5858f7a50 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 27 Aug 2006 17:23:53 -0400 Subject: [PATCH 56/74] NFS: NFS_ROOT should use the new rpc_create API Teach NFS_ROOT to use the new rpc_create API instead of the old two-call API for creating an RPC transport. Test plan: Compile the kernel with the NFS client build-in, and set CONFIG_NFS_ROOT. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/mount_clnt.c | 27 ++++++++++++--------------- 1 file changed, 12 insertions(+), 15 deletions(-) diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c index 41274874b9a5..d507b021207f 100644 --- a/fs/nfs/mount_clnt.c +++ b/fs/nfs/mount_clnt.c @@ -76,22 +76,19 @@ static struct rpc_clnt * mnt_create(char *hostname, struct sockaddr_in *srvaddr, int version, int protocol) { - struct rpc_xprt *xprt; - struct rpc_clnt *clnt; + struct rpc_create_args args = { + .protocol = protocol, + .address = (struct sockaddr *)srvaddr, + .addrsize = sizeof(*srvaddr), + .servername = hostname, + .program = &mnt_program, + .version = version, + .authflavor = RPC_AUTH_UNIX, + .flags = (RPC_CLNT_CREATE_ONESHOT | + RPC_CLNT_CREATE_INTR), + }; - xprt = xprt_create_proto(protocol, srvaddr, NULL); - if (IS_ERR(xprt)) - return (struct rpc_clnt *)xprt; - - clnt = rpc_create_client(xprt, hostname, - &mnt_program, version, - RPC_AUTH_UNIX); - if (!IS_ERR(clnt)) { - clnt->cl_softrtry = 1; - clnt->cl_oneshot = 1; - clnt->cl_intr = 1; - } - return clnt; + return rpc_create(&args); } /* From 297de4f65698ee1e1c75e27d57933b5fa8227e72 Mon Sep 17 00:00:00 2001 From: "andros@citi.umich.edu" Date: Tue, 29 Aug 2006 12:19:41 -0400 Subject: [PATCH 57/74] Fix a referral error Oops Fix an oops when the referral server is not responding. Check the error return from nfs4_set_client() in nfs4_create_referral_server. Signed-off-by: Andy Adamson Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index a4aa47913a5c..110f80e7bd4c 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -1059,6 +1059,8 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data, parent_server->client->cl_xprt->prot, parent_client->retrans_timeo, parent_client->retrans_count); + if (error < 0) + goto error; /* Initialise the client representation from the parent server */ nfs_server_copy_userdata(server, parent_server); From 8014793b1b2869445adfe678d64cdacd10e99d53 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 31 Aug 2006 18:24:08 -0400 Subject: [PATCH 58/74] SUNRPC: rpc_delay() should not clobber the rpc_task->tk_status Doing so prevents stuff like call_encode() from working correctly. Signed-off-by: Trond Myklebust --- net/sunrpc/sched.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c index ecf366351bf7..6390461a9756 100644 --- a/net/sunrpc/sched.c +++ b/net/sunrpc/sched.c @@ -542,24 +542,20 @@ void rpc_wake_up_status(struct rpc_wait_queue *queue, int status) spin_unlock_bh(&queue->lock); } +static void __rpc_atrun(struct rpc_task *task) +{ + rpc_wake_up_task(task); +} + /* * Run a task at a later time */ -static void __rpc_atrun(struct rpc_task *); -void -rpc_delay(struct rpc_task *task, unsigned long delay) +void rpc_delay(struct rpc_task *task, unsigned long delay) { task->tk_timeout = delay; rpc_sleep_on(&delay_queue, task, NULL, __rpc_atrun); } -static void -__rpc_atrun(struct rpc_task *task) -{ - task->tk_status = 0; - rpc_wake_up_task(task); -} - /* * Helper to call task->tk_ops->rpc_call_prepare */ From 76303992b4701124f4cd0791ae2049ab4332f02c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 30 Aug 2006 14:32:49 -0400 Subject: [PATCH 59/74] SUNRPC: Handle ENETUNREACH, EHOSTUNREACH and EHOSTDOWN socket errors In case of any of the above errors occuring, delay for 3 seconds, then handle as if it were a timeout error. Signed-off-by: Trond Myklebust --- net/sunrpc/clnt.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 87efcd207f23..355e7863c0aa 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -1030,6 +1030,14 @@ call_status(struct rpc_task *task) task->tk_status = 0; switch(status) { + case -EHOSTDOWN: + case -EHOSTUNREACH: + case -ENETUNREACH: + /* + * Delay any retries for 3 seconds, then handle as if it + * were a timeout. + */ + rpc_delay(task, 3*HZ); case -ETIMEDOUT: task->tk_action = call_timeout; break; From da45828e2835057045150b318c4fbe9bb91f18dd Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 31 Aug 2006 15:44:52 -0400 Subject: [PATCH 60/74] SUNRPC: Clean up soft task error handling - Ensure that the task aborts the RPC call only when it has actually timed out. - Ensure that req->rq_majortimeo is initialised correctly. Signed-off-by: Trond Myklebust --- net/sunrpc/clnt.c | 34 ++++++++++++++-------------------- net/sunrpc/xprt.c | 8 +------- 2 files changed, 15 insertions(+), 27 deletions(-) diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 355e7863c0aa..ceadb728f0da 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -863,15 +863,11 @@ call_bind_status(struct rpc_task *task) dprintk("RPC: %4d remote rpcbind: RPC program/version unavailable\n", task->tk_pid); rpc_delay(task, 3*HZ); - goto retry_bind; + goto retry_timeout; case -ETIMEDOUT: dprintk("RPC: %4d rpcbind request timed out\n", task->tk_pid); - if (RPC_IS_SOFT(task)) { - status = -EIO; - break; - } - goto retry_bind; + goto retry_timeout; case -EPFNOSUPPORT: dprintk("RPC: %4d remote rpcbind service unavailable\n", task->tk_pid); @@ -884,16 +880,13 @@ call_bind_status(struct rpc_task *task) dprintk("RPC: %4d unrecognized rpcbind error (%d)\n", task->tk_pid, -task->tk_status); status = -EIO; - break; } rpc_exit(task, status); return; -retry_bind: - task->tk_status = 0; - task->tk_action = call_bind; - return; +retry_timeout: + task->tk_action = call_timeout; } /* @@ -941,14 +934,16 @@ call_connect_status(struct rpc_task *task) switch (status) { case -ENOTCONN: - case -ETIMEDOUT: case -EAGAIN: task->tk_action = call_bind; - break; - default: - rpc_exit(task, -EIO); - break; + if (!RPC_IS_SOFT(task)) + return; + /* if soft mounted, test if we've timed out */ + case -ETIMEDOUT: + task->tk_action = call_timeout; + return; } + rpc_exit(task, -EIO); } /* @@ -1057,7 +1052,6 @@ call_status(struct rpc_task *task) printk("%s: RPC call returned error %d\n", clnt->cl_protname, -status); rpc_exit(task, status); - break; } } @@ -1125,10 +1119,10 @@ call_decode(struct rpc_task *task) clnt->cl_stats->rpcretrans++; goto out_retry; } - printk(KERN_WARNING "%s: too small RPC reply size (%d bytes)\n", + dprintk("%s: too small RPC reply size (%d bytes)\n", clnt->cl_protname, task->tk_status); - rpc_exit(task, -EIO); - return; + task->tk_action = call_timeout; + goto out_retry; } /* diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index e4f64fb58ff2..a85f82baefc1 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -585,13 +585,6 @@ static void xprt_connect_status(struct rpc_task *task) task->tk_pid, -task->tk_status, task->tk_client->cl_server); xprt_release_write(xprt, task); task->tk_status = -EIO; - return; - } - - /* if soft mounted, just cause this RPC to fail */ - if (RPC_IS_SOFT(task)) { - xprt_release_write(xprt, task); - task->tk_status = -EIO; } } @@ -829,6 +822,7 @@ static void xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt) req->rq_bufsize = 0; req->rq_xid = xprt_alloc_xid(xprt); req->rq_release_snd_buf = NULL; + xprt_reset_majortimeo(req); dprintk("RPC: %4d reserved req %p xid %08x\n", task->tk_pid, req, ntohl(req->rq_xid)); } From 6b6ca86b77b62b798cf9ca2599036420abce7796 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 5 Sep 2006 12:55:57 -0400 Subject: [PATCH 61/74] SUNRPC: Add refcounting to the struct rpc_xprt In a subsequent patch, this will allow the portmapper to take a reference to the rpc_xprt for which it is updating the port number, fixing an Oops. Signed-off-by: Trond Myklebust --- include/linux/sunrpc/xprt.h | 5 ++++- net/sunrpc/clnt.c | 8 +++----- net/sunrpc/xprt.c | 30 ++++++++++++++++++++++++++---- 3 files changed, 33 insertions(+), 10 deletions(-) diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index de4efea7c856..bdeba8538c71 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -12,6 +12,7 @@ #include #include #include +#include #include #include @@ -129,6 +130,7 @@ struct rpc_xprt_ops { }; struct rpc_xprt { + struct kref kref; /* Reference count */ struct rpc_xprt_ops * ops; /* transport methods */ struct socket * sock; /* BSD socket layer */ struct sock * inet; /* INET layer */ @@ -248,7 +250,8 @@ int xprt_adjust_timeout(struct rpc_rqst *req); void xprt_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task); void xprt_release_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task); void xprt_release(struct rpc_task *task); -int xprt_destroy(struct rpc_xprt *xprt); +struct rpc_xprt * xprt_get(struct rpc_xprt *xprt); +void xprt_put(struct rpc_xprt *xprt); static inline u32 *xprt_skip_transport_header(struct rpc_xprt *xprt, u32 *p) { diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index ceadb728f0da..084a0ad5c64e 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -177,7 +177,7 @@ static struct rpc_clnt * rpc_new_client(struct rpc_xprt *xprt, char *servname, s kfree(clnt->cl_server); kfree(clnt); out_err: - xprt_destroy(xprt); + xprt_put(xprt); out_no_xprt: return ERR_PTR(err); } @@ -261,6 +261,7 @@ rpc_clone_client(struct rpc_clnt *clnt) atomic_set(&new->cl_users, 0); new->cl_parent = clnt; atomic_inc(&clnt->cl_count); + new->cl_xprt = xprt_get(clnt->cl_xprt); /* Turn off autobind on clones */ new->cl_autobind = 0; new->cl_oneshot = 0; @@ -337,15 +338,12 @@ rpc_destroy_client(struct rpc_clnt *clnt) rpc_rmdir(clnt->cl_dentry); rpc_put_mount(); } - if (clnt->cl_xprt) { - xprt_destroy(clnt->cl_xprt); - clnt->cl_xprt = NULL; - } if (clnt->cl_server != clnt->cl_inline_name) kfree(clnt->cl_server); out_free: rpc_free_iostats(clnt->cl_metrics); clnt->cl_metrics = NULL; + xprt_put(clnt->cl_xprt); kfree(clnt); return 0; } diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index a85f82baefc1..1f786f68729d 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -926,6 +926,7 @@ struct rpc_xprt *xprt_create_transport(int proto, struct sockaddr *ap, size_t si return ERR_PTR(result); } + kref_init(&xprt->kref); spin_lock_init(&xprt->transport_lock); spin_lock_init(&xprt->reserve_lock); @@ -958,16 +959,37 @@ struct rpc_xprt *xprt_create_transport(int proto, struct sockaddr *ap, size_t si /** * xprt_destroy - destroy an RPC transport, killing off all requests. - * @xprt: transport to destroy + * @kref: kref for the transport to destroy * */ -int xprt_destroy(struct rpc_xprt *xprt) +static void xprt_destroy(struct kref *kref) { + struct rpc_xprt *xprt = container_of(kref, struct rpc_xprt, kref); + dprintk("RPC: destroying transport %p\n", xprt); xprt->shutdown = 1; del_timer_sync(&xprt->timer); xprt->ops->destroy(xprt); kfree(xprt); - - return 0; +} + +/** + * xprt_put - release a reference to an RPC transport. + * @xprt: pointer to the transport + * + */ +void xprt_put(struct rpc_xprt *xprt) +{ + kref_put(&xprt->kref, xprt_destroy); +} + +/** + * xprt_get - return a reference to an RPC transport. + * @xprt: pointer to the transport + * + */ +struct rpc_xprt *xprt_get(struct rpc_xprt *xprt) +{ + kref_get(&xprt->kref); + return xprt; } From 762d4527c2fc19d821a13d9a3455ccc2d4073731 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 3 Sep 2006 00:51:55 -0400 Subject: [PATCH 62/74] SUNRPC: Fix Oops in pmap_getport_done There is no guarantee that the parent task still exists when we exit from the portmapper. Save the xprt instead. Signed-off-by: Trond Myklebust --- net/sunrpc/pmap_clnt.c | 46 ++++++++++++++++++++---------------------- 1 file changed, 22 insertions(+), 24 deletions(-) diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c index f476f4df0f48..c04609d3476a 100644 --- a/net/sunrpc/pmap_clnt.c +++ b/net/sunrpc/pmap_clnt.c @@ -30,7 +30,7 @@ struct portmap_args { u32 pm_vers; u32 pm_prot; unsigned short pm_port; - struct rpc_task * pm_task; + struct rpc_xprt * pm_xprt; }; static struct rpc_procinfo pmap_procedures[]; @@ -71,10 +71,10 @@ static const struct rpc_call_ops pmap_getport_ops = { .rpc_release = pmap_map_release, }; -static inline void pmap_wake_portmap_waiters(struct rpc_xprt *xprt) +static inline void pmap_wake_portmap_waiters(struct rpc_xprt *xprt, int status) { xprt_clear_binding(xprt); - rpc_wake_up(&xprt->binding); + rpc_wake_up_status(&xprt->binding, status); } /** @@ -92,6 +92,7 @@ void rpc_getport(struct rpc_task *task) struct portmap_args *map; struct rpc_clnt *pmap_clnt; struct rpc_task *child; + int status; dprintk("RPC: %4d rpc_getport(%s, %u, %u, %d)\n", task->tk_pid, clnt->cl_server, @@ -107,34 +108,30 @@ void rpc_getport(struct rpc_task *task) } /* Someone else may have bound if we slept */ - if (xprt_bound(xprt)) { - task->tk_status = 0; + status = 0; + if (xprt_bound(xprt)) goto bailout_nofree; - } + status = -ENOMEM; map = pmap_map_alloc(); - if (!map) { - task->tk_status = -ENOMEM; + if (!map) goto bailout_nofree; - } map->pm_prog = clnt->cl_prog; map->pm_vers = clnt->cl_vers; map->pm_prot = xprt->prot; map->pm_port = 0; - map->pm_task = task; + map->pm_xprt = xprt_get(xprt); rpc_peeraddr(clnt, (struct sockaddr *) &addr, sizeof(addr)); pmap_clnt = pmap_create(clnt->cl_server, &addr, map->pm_prot, 0); - if (IS_ERR(pmap_clnt)) { - task->tk_status = PTR_ERR(pmap_clnt); + status = PTR_ERR(pmap_clnt); + if (IS_ERR(pmap_clnt)) goto bailout; - } + status = -EIO; child = rpc_run_task(pmap_clnt, RPC_TASK_ASYNC, &pmap_getport_ops, map); - if (IS_ERR(child)) { - task->tk_status = -EIO; + if (IS_ERR(child)) goto bailout; - } rpc_release_task(child); rpc_sleep_on(&xprt->binding, task, NULL, NULL); @@ -144,8 +141,10 @@ void rpc_getport(struct rpc_task *task) bailout: pmap_map_free(map); + xprt_put(xprt); bailout_nofree: - pmap_wake_portmap_waiters(xprt); + task->tk_status = status; + pmap_wake_portmap_waiters(xprt, status); } #ifdef CONFIG_ROOT_NFS @@ -201,29 +200,28 @@ int rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int pr static void pmap_getport_done(struct rpc_task *child, void *data) { struct portmap_args *map = data; - struct rpc_task *task = map->pm_task; - struct rpc_xprt *xprt = task->tk_xprt; + struct rpc_xprt *xprt = map->pm_xprt; int status = child->tk_status; if (status < 0) { /* Portmapper not available */ xprt->ops->set_port(xprt, 0); - task->tk_status = status; } else if (map->pm_port == 0) { /* Requested RPC service wasn't registered */ xprt->ops->set_port(xprt, 0); - task->tk_status = -EACCES; + status = -EACCES; } else { /* Succeeded */ xprt->ops->set_port(xprt, map->pm_port); xprt_set_bound(xprt); - task->tk_status = 0; + status = 0; } dprintk("RPC: %4d pmap_getport_done(status %d, port %u)\n", - child->tk_pid, child->tk_status, map->pm_port); + child->tk_pid, status, map->pm_port); - pmap_wake_portmap_waiters(xprt); + pmap_wake_portmap_waiters(xprt, status); + xprt_put(xprt); } /** From fd6840714d9cf6e93f1d42b904860a94df316b85 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 5 Sep 2006 12:27:44 -0400 Subject: [PATCH 63/74] NFS: nfs_lookup - don't hash dentry when optimising away the lookup If the open intents tell us that a given lookup is going to result in a, exclusive create, we currently optimize away the lookup call itself. The reason is that the lookup would not be atomic with the create RPC call, so why do it in the first place? A problem occurs, however, if the VFS aborts the exclusive create operation after the lookup, but before the call to create the file/directory: in this case we will end up with a hashed negative dentry in the dcache that has never been looked up. Fix this by only actually hashing the dentry once the create operation has been successfully completed. Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 51328ae640dd..3419c2da9ba9 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -904,9 +904,15 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru lock_kernel(); - /* If we're doing an exclusive create, optimize away the lookup */ - if (nfs_is_exclusive_create(dir, nd)) - goto no_entry; + /* + * If we're doing an exclusive create, optimize away the lookup + * but don't hash the dentry. + */ + if (nfs_is_exclusive_create(dir, nd)) { + d_instantiate(dentry, NULL); + res = NULL; + goto out_unlock; + } error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, &fhandle, &fattr); if (error == -ENOENT) @@ -1161,6 +1167,8 @@ int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle, if (IS_ERR(inode)) return error; d_instantiate(dentry, inode); + if (d_unhashed(dentry)) + d_rehash(dentry); return 0; } From 2dec51466a08ac1c67da41bfd0518d43d983a2eb Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 12 Sep 2006 11:53:23 -0400 Subject: [PATCH 64/74] NFSv4: It's perfectly legal for clp to be NULL here.... Signed-off-by: J. Bruce Fields Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 110f80e7bd4c..ec1938d4b814 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -269,7 +269,7 @@ struct nfs_client *nfs_find_client(const struct sockaddr_in *addr, int nfsversio clp = __nfs_find_client(addr, nfsversion); spin_unlock(&nfs_client_lock); - BUG_ON(clp->cl_cons_state == 0); + BUG_ON(clp && clp->cl_cons_state == 0); return clp; } From 5f004cf2aa8494708fd8d78e78142b7b2748e765 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Sep 2006 14:03:14 -0400 Subject: [PATCH 65/74] NFS: Make read() return an ESTALE if the file has been deleted Currently, a read() request will return EIO even if the file has been deleted on the server, simply because that is what the VM will return if the call to readpage() fails to update the page. Ensure that readpage() marks the inode as stale if it receives an ESTALE. Then return that error to userland. Signed-off-by: Trond Myklebust --- fs/nfs/read.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/fs/nfs/read.c b/fs/nfs/read.c index dae33c1e8a77..69f1549da2b9 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -568,8 +568,13 @@ int nfs_readpage_result(struct rpc_task *task, struct nfs_read_data *data) nfs_add_stats(data->inode, NFSIOS_SERVERREADBYTES, resp->count); - /* Is this a short read? */ - if (task->tk_status >= 0 && resp->count < argp->count && !resp->eof) { + if (task->tk_status < 0) { + if (task->tk_status == -ESTALE) { + set_bit(NFS_INO_STALE, &NFS_FLAGS(data->inode)); + nfs_mark_for_revalidate(data->inode); + } + } else if (resp->count < argp->count && !resp->eof) { + /* This is a short read! */ nfs_inc_stats(data->inode, NFSIOS_SHORTREAD); /* Has the server at least made some progress? */ if (resp->count != 0) { @@ -616,6 +621,10 @@ int nfs_readpage(struct file *file, struct page *page) if (error) goto out_error; + error = -ESTALE; + if (NFS_STALE(inode)) + goto out_error; + if (file == NULL) { ctx = nfs_find_open_context(inode, NULL, FMODE_READ); if (ctx == NULL) @@ -678,7 +687,7 @@ int nfs_readpages(struct file *filp, struct address_space *mapping, }; struct inode *inode = mapping->host; struct nfs_server *server = NFS_SERVER(inode); - int ret; + int ret = -ESTALE; dprintk("NFS: nfs_readpages (%s/%Ld %d)\n", inode->i_sb->s_id, @@ -686,6 +695,9 @@ int nfs_readpages(struct file *filp, struct address_space *mapping, nr_pages); nfs_inc_stats(inode, NFSIOS_VFSREADPAGES); + if (NFS_STALE(inode)) + goto out; + if (filp == NULL) { desc.ctx = nfs_find_open_context(inode, NULL, FMODE_READ); if (desc.ctx == NULL) @@ -701,6 +713,7 @@ int nfs_readpages(struct file *filp, struct address_space *mapping, ret = err; } put_nfs_open_context(desc.ctx); +out: return ret; } From 97db8f41792839a6912fd21be8b61dd6c50db58f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Sep 2006 14:03:14 -0400 Subject: [PATCH 66/74] NFS: Don't invalidate the symlink we just stuffed into the cache And slight optimisation of nfs_end_data_update(): directories never have delegations anyway. Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index cb5c65f0bc12..a56add0bc5b7 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -717,13 +717,11 @@ void nfs_end_data_update(struct inode *inode) { struct nfs_inode *nfsi = NFS_I(inode); - if (!nfs_have_delegation(inode, FMODE_READ)) { - /* Directories and symlinks: invalidate page cache */ - if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode)) { - spin_lock(&inode->i_lock); - nfsi->cache_validity |= NFS_INO_INVALID_DATA; - spin_unlock(&inode->i_lock); - } + /* Directories: invalidate page cache */ + if (S_ISDIR(inode->i_mode)) { + spin_lock(&inode->i_lock); + nfsi->cache_validity |= NFS_INO_INVALID_DATA; + spin_unlock(&inode->i_lock); } nfsi->cache_change_attribute = jiffies; atomic_dec(&nfsi->data_updates); From 6b30954ebb569fa1b2abdb21f2f4290eec76bf80 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Sep 2006 14:03:14 -0400 Subject: [PATCH 67/74] NFSv4: Retry lease recovery if it failed during a synchronous operation. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 7f60beb40df3..c218cc450b92 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -793,10 +793,17 @@ static int _nfs4_do_access(struct inode *inode, struct rpc_cred *cred, int openf int nfs4_recover_expired_lease(struct nfs_server *server) { struct nfs_client *clp = server->nfs_client; + int ret; - if (test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) + for (;;) { + ret = nfs4_wait_clnt_recover(server->client, clp); + if (ret != 0) + return ret; + if (!test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) + break; nfs4_schedule_state_recovery(clp); - return nfs4_wait_clnt_recover(server->client, clp); + } + return 0; } /* From c514983d8d2260020543a81589a2b8c7d4bdab4e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 15 Sep 2006 08:25:04 -0400 Subject: [PATCH 68/74] NFSv4: Handle the condition NFS4ERR_FILE_OPEN Retry a few times before we give up: the error is usually due to ordering issues with asynchronous RPC calls. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index c218cc450b92..c49ac3ea1dc3 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2891,6 +2891,7 @@ int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct if (ret == 0) exception->retry = 1; break; + case -NFS4ERR_FILE_OPEN: case -NFS4ERR_GRACE: case -NFS4ERR_DELAY: ret = nfs4_delay(server->client, &exception->timeout); From 2066fe89b459c3c787c811b3369df191cddd93d8 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 15 Sep 2006 08:30:46 -0400 Subject: [PATCH 69/74] NFSv4: Poll more aggressively when handling NFS4ERR_DELAY Change the initial retry delay from 1s to 0.1s (and then back off exponentially). Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index c49ac3ea1dc3..47c7e6e3910d 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -55,7 +55,7 @@ #define NFSDBG_FACILITY NFSDBG_PROC -#define NFS4_POLL_RETRY_MIN (1*HZ) +#define NFS4_POLL_RETRY_MIN (HZ/10) #define NFS4_POLL_RETRY_MAX (15*HZ) struct nfs4_opendata; From 51b6ded4d9a94a61035deba1d8f51a54e3a3dd86 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 15 Sep 2006 16:31:56 -0400 Subject: [PATCH 70/74] NFSv4: When mounting with a port=0 argument, substitute port=2049 RFC3530 states that the registered port 2049 for the NFS protocol should be the default configuration in order to allow clients not to use the RPC binding protocols. If the mount program sends us a port=0, we therefore substitute port=2049. Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 665949d27798..b99113b0f65f 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -833,6 +833,9 @@ static int nfs4_get_sb(struct file_system_type *fs_type, __FUNCTION__); return -EINVAL; } + /* RFC3530: The default port for NFS is 2049 */ + if (addr.sin_port == 0) + addr.sin_port = NFS_PORT; /* Grab the authentication type */ authflavour = RPC_AUTH_UNIX; From aec5e175288c711cbe44750276f61efa3fa3d370 Mon Sep 17 00:00:00 2001 From: Josef 'Jeff' Sipek Date: Sat, 16 Sep 2006 21:09:32 -0400 Subject: [PATCH 71/74] NFS: Use SEEK_END instead of hardcoded value Signed-off-by: Josef 'Jeff' Sipek Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/nfs/file.c b/fs/nfs/file.c index a146ed338534..be997d649127 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -157,7 +157,7 @@ static int nfs_revalidate_file_size(struct inode *inode, struct file *filp) static loff_t nfs_file_llseek(struct file *filp, loff_t offset, int origin) { /* origin == SEEK_END => we must revalidate the cached file length */ - if (origin == 2) { + if (origin == SEEK_END) { struct inode *inode = filp->f_mapping->host; int retval = nfs_revalidate_file_size(inode, filp); if (retval < 0) From a53a3c58fd83e572a7c768d88b4c4e9840a57e82 Mon Sep 17 00:00:00 2001 From: Steve Dickson Date: Wed, 6 Sep 2006 11:51:21 -0400 Subject: [PATCH 72/74] NFSv4: rpc_mkpipe creating socket inodes w/out sk buffers This patch stop rpc_mkpipe from create S_IFSOCK nodes what don't have associated sk buffers attached (which causes SELinux to oops during NFSv4 mounts). Instead the S_IFIFO mode bit is set which probably make more sense and seems to work just fine during my connectathon and fsx testing... Signed-off-by: Steve Dickson Signed-off-by: Trond Myklebust --- net/sunrpc/rpc_pipe.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c index 11ec12a09d70..dfa504fe383f 100644 --- a/net/sunrpc/rpc_pipe.c +++ b/net/sunrpc/rpc_pipe.c @@ -720,7 +720,7 @@ rpc_mkpipe(struct dentry *parent, const char *name, void *private, struct rpc_pi if (IS_ERR(dentry)) return dentry; dir = parent->d_inode; - inode = rpc_get_inode(dir->i_sb, S_IFSOCK | S_IRUSR | S_IWUSR); + inode = rpc_get_inode(dir->i_sb, S_IFIFO | S_IRUSR | S_IWUSR); if (!inode) goto err_dput; inode->i_ino = iunique(dir->i_sb, 100); From f551e44ff11d3e2ec8f37907bb88ec2433cc8b74 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 20 Sep 2006 14:33:04 -0400 Subject: [PATCH 73/74] NFS: add comments clarifying the use of nfs_post_op_update() Comments-only change to clarify a detail of the NFS protocol and how it is implemented in Linux. Test plan: None. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 6 ++++++ fs/nfs/write.c | 8 +++++++- 2 files changed, 13 insertions(+), 1 deletion(-) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index a56add0bc5b7..e8c143d182c4 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -840,6 +840,12 @@ int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr) * * After an operation that has changed the inode metadata, mark the * attribute cache as being invalid, then try to update it. + * + * NB: if the server didn't return any post op attributes, this + * function will force the retrieval of attributes before the next + * NFS request. Thus it should be used only for operations that + * are expected to change one or more attributes, to avoid + * unnecessary NFS requests and trips through nfs_update_inode(). */ int nfs_post_op_update_inode(struct inode *inode, struct nfs_fattr *fattr) { diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 38ba5c09af08..c12effb46fe5 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1253,7 +1253,13 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data) dprintk("NFS: %4d nfs_writeback_done (status %d)\n", task->tk_pid, task->tk_status); - /* Call the NFS version-specific code */ + /* + * ->write_done will attempt to use post-op attributes to detect + * conflicting writes by other clients. A strict interpretation + * of close-to-open would allow us to continue caching even if + * another writer had changed the file, but some applications + * depend on tighter cache coherency when writing. + */ status = NFS_PROTO(data->inode)->write_done(task, data); if (status != 0) return status; From 026ed5c9185dcc4b2df92e98c3d61a01cea19cbf Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 20 Sep 2006 14:33:07 -0400 Subject: [PATCH 74/74] NFS: unmark NFS direct I/O as experimental Remove the EXPERIMENTAL flag from the NFS_DIRECTIO option. Test plan: Unset the EXPERIMENTAL kernel build option and check to see that the NFS direct I/O option is still available. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/Kconfig b/fs/Kconfig index 530581628311..a27002668bd3 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1471,8 +1471,8 @@ config NFS_V4 If unsure, say N. config NFS_DIRECTIO - bool "Allow direct I/O on NFS files (EXPERIMENTAL)" - depends on NFS_FS && EXPERIMENTAL + bool "Allow direct I/O on NFS files" + depends on NFS_FS help This option enables applications to perform uncached I/O on files in NFS file systems using the O_DIRECT open() flag. When O_DIRECT