X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;ds=sidebyside;f=fs%2Fdcache.c;fp=fs%2Fdcache.c;h=cbf76e692013281de5852d569896ac1625bd96fb;hb=16c70f8c1b54b61c3b951b6fb220df250fe09b32;hp=5e279e6dfafc0e50bdb03bee048a66405e84e8c2;hpb=4e76c8a9fa413ccc09d3f7f664183dcce3555d57;p=linux-2.6.git diff --git a/fs/dcache.c b/fs/dcache.c index 5e279e6df..cbf76e692 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -14,7 +14,6 @@ * the dcache entry is deleted or garbage collected. */ -#include #include #include #include @@ -39,7 +38,7 @@ int sysctl_vfs_cache_pressure __read_mostly = 100; EXPORT_SYMBOL_GPL(sysctl_vfs_cache_pressure); __cacheline_aligned_in_smp DEFINE_SPINLOCK(dcache_lock); -static seqlock_t rename_lock __cacheline_aligned_in_smp = SEQLOCK_UNLOCKED; +static __cacheline_aligned_in_smp DEFINE_SEQLOCK(rename_lock); EXPORT_SYMBOL(dcache_lock); @@ -363,12 +362,13 @@ restart: } /* - * Throw away a dentry - free the inode, dput the parent. - * This requires that the LRU list has already been - * removed. + * Throw away a dentry - free the inode, dput the parent. This requires that + * the LRU list has already been removed. + * * Called with dcache_lock, drops it and then regains. + * Called with dentry->d_lock held, drops it. */ -static inline void prune_one_dentry(struct dentry * dentry) +static void prune_one_dentry(struct dentry * dentry) { struct dentry * parent; @@ -386,6 +386,8 @@ static inline void prune_one_dentry(struct dentry * dentry) /** * prune_dcache - shrink the dcache * @count: number of entries to try and free + * @sb: if given, ignore dentries for other superblocks + * which are being unmounted. * * Shrink the dcache. This is done when we need * more memory, or simply when we need to unmount @@ -396,16 +398,29 @@ static inline void prune_one_dentry(struct dentry * dentry) * all the dentries are in use. */ -static void prune_dcache(int count) +static void prune_dcache(int count, struct super_block *sb) { spin_lock(&dcache_lock); for (; count ; count--) { struct dentry *dentry; struct list_head *tmp; + struct rw_semaphore *s_umount; cond_resched_lock(&dcache_lock); tmp = dentry_unused.prev; + if (sb) { + /* Try to find a dentry for this sb, but don't try + * too hard, if they aren't near the tail they will + * be moved down again soon + */ + int skip = count; + while (skip && tmp != &dentry_unused && + list_entry(tmp, struct dentry, d_lru)->d_sb != sb) { + skip--; + tmp = tmp->prev; + } + } if (tmp == &dentry_unused) break; list_del_init(tmp); @@ -431,7 +446,45 @@ static void prune_dcache(int count) spin_unlock(&dentry->d_lock); continue; } - prune_one_dentry(dentry); + /* + * If the dentry is not DCACHED_REFERENCED, it is time + * to remove it from the dcache, provided the super block is + * NULL (which means we are trying to reclaim memory) + * or this dentry belongs to the same super block that + * we want to shrink. + */ + /* + * If this dentry is for "my" filesystem, then I can prune it + * without taking the s_umount lock (I already hold it). + */ + if (sb && dentry->d_sb == sb) { + prune_one_dentry(dentry); + continue; + } + /* + * ...otherwise we need to be sure this filesystem isn't being + * unmounted, otherwise we could race with + * generic_shutdown_super(), and end up holding a reference to + * an inode while the filesystem is unmounted. + * So we try to get s_umount, and make sure s_root isn't NULL. + * (Take a local copy of s_umount to avoid a use-after-free of + * `dentry'). + */ + s_umount = &dentry->d_sb->s_umount; + if (down_read_trylock(s_umount)) { + if (dentry->d_sb->s_root != NULL) { + prune_one_dentry(dentry); + up_read(s_umount); + continue; + } + up_read(s_umount); + } + spin_unlock(&dentry->d_lock); + /* Cannot remove the first dentry, and it isn't appropriate + * to move it to the head of the list, so give up, and try + * later + */ + break; } spin_unlock(&dcache_lock); } @@ -472,8 +525,7 @@ void shrink_dcache_sb(struct super_block * sb) dentry = list_entry(tmp, struct dentry, d_lru); if (dentry->d_sb != sb) continue; - list_del(tmp); - list_add(tmp, &dentry_unused); + list_move(tmp, &dentry_unused); } /* @@ -498,6 +550,136 @@ repeat: spin_unlock(&dcache_lock); } +/* + * destroy a single subtree of dentries for unmount + * - see the comments on shrink_dcache_for_umount() for a description of the + * locking + */ +static void shrink_dcache_for_umount_subtree(struct dentry *dentry) +{ + struct dentry *parent; + + BUG_ON(!IS_ROOT(dentry)); + + /* detach this root from the system */ + spin_lock(&dcache_lock); + if (!list_empty(&dentry->d_lru)) { + dentry_stat.nr_unused--; + list_del_init(&dentry->d_lru); + } + __d_drop(dentry); + spin_unlock(&dcache_lock); + + for (;;) { + /* descend to the first leaf in the current subtree */ + while (!list_empty(&dentry->d_subdirs)) { + struct dentry *loop; + + /* this is a branch with children - detach all of them + * from the system in one go */ + spin_lock(&dcache_lock); + list_for_each_entry(loop, &dentry->d_subdirs, + d_u.d_child) { + if (!list_empty(&loop->d_lru)) { + dentry_stat.nr_unused--; + list_del_init(&loop->d_lru); + } + + __d_drop(loop); + cond_resched_lock(&dcache_lock); + } + spin_unlock(&dcache_lock); + + /* move to the first child */ + dentry = list_entry(dentry->d_subdirs.next, + struct dentry, d_u.d_child); + } + + /* consume the dentries from this leaf up through its parents + * until we find one with children or run out altogether */ + do { + struct inode *inode; + + if (atomic_read(&dentry->d_count) != 0) { + printk(KERN_ERR + "BUG: Dentry %p{i=%lx,n=%s}" + " still in use (%d)" + " [unmount of %s %s]\n", + dentry, + dentry->d_inode ? + dentry->d_inode->i_ino : 0UL, + dentry->d_name.name, + atomic_read(&dentry->d_count), + dentry->d_sb->s_type->name, + dentry->d_sb->s_id); + BUG(); + } + + parent = dentry->d_parent; + if (parent == dentry) + parent = NULL; + else + atomic_dec(&parent->d_count); + + list_del(&dentry->d_u.d_child); + dentry_stat.nr_dentry--; /* For d_free, below */ + + inode = dentry->d_inode; + if (inode) { + dentry->d_inode = NULL; + list_del_init(&dentry->d_alias); + if (dentry->d_op && dentry->d_op->d_iput) + dentry->d_op->d_iput(dentry, inode); + else + iput(inode); + } + + d_free(dentry); + + /* finished when we fall off the top of the tree, + * otherwise we ascend to the parent and move to the + * next sibling if there is one */ + if (!parent) + return; + + dentry = parent; + + } while (list_empty(&dentry->d_subdirs)); + + dentry = list_entry(dentry->d_subdirs.next, + struct dentry, d_u.d_child); + } +} + +/* + * destroy the dentries attached to a superblock on unmounting + * - we don't need to use dentry->d_lock, and only need dcache_lock when + * removing the dentry from the system lists and hashes because: + * - the superblock is detached from all mountings and open files, so the + * dentry trees will not be rearranged by the VFS + * - s_umount is write-locked, so the memory pressure shrinker will ignore + * any dentries belonging to this superblock that it comes across + * - the filesystem itself is no longer permitted to rearrange the dentries + * in this superblock + */ +void shrink_dcache_for_umount(struct super_block *sb) +{ + struct dentry *dentry; + + if (down_read_trylock(&sb->s_umount)) + BUG(); + + dentry = sb->s_root; + sb->s_root = NULL; + atomic_dec(&dentry->d_count); + shrink_dcache_for_umount_subtree(dentry); + + while (!hlist_empty(&sb->s_anon)) { + dentry = hlist_entry(sb->s_anon.first, struct dentry, d_hash); + shrink_dcache_for_umount_subtree(dentry); + } +} + /* * Search for at least 1 mount point in the dentry's subdirs. * We descend to the next level whenever the d_subdirs @@ -588,7 +770,7 @@ resume: * of the unused list for prune_dcache */ if (!atomic_read(&dentry->d_count)) { - list_add(&dentry->d_lru, dentry_unused.prev); + list_add_tail(&dentry->d_lru, &dentry_unused); dentry_stat.nr_unused++; found++; } @@ -634,46 +816,7 @@ void shrink_dcache_parent(struct dentry * parent) int found; while ((found = select_parent(parent)) != 0) - prune_dcache(found); -} - -/** - * shrink_dcache_anon - further prune the cache - * @head: head of d_hash list of dentries to prune - * - * Prune the dentries that are anonymous - * - * parsing d_hash list does not hlist_for_each_entry_rcu() as it - * done under dcache_lock. - * - */ -void shrink_dcache_anon(struct hlist_head *head) -{ - struct hlist_node *lp; - int found; - do { - found = 0; - spin_lock(&dcache_lock); - hlist_for_each(lp, head) { - struct dentry *this = hlist_entry(lp, struct dentry, d_hash); - if (!list_empty(&this->d_lru)) { - dentry_stat.nr_unused--; - list_del_init(&this->d_lru); - } - - /* - * move only zero ref count dentries to the end - * of the unused list for prune_dcache - */ - if (!atomic_read(&this->d_count)) { - list_add_tail(&this->d_lru, &dentry_unused); - dentry_stat.nr_unused++; - found++; - } - } - spin_unlock(&dcache_lock); - prune_dcache(found); - } while(found); + prune_dcache(found, parent->d_sb); } /* @@ -693,7 +836,7 @@ static int shrink_dcache_memory(int nr, gfp_t gfp_mask) if (nr) { if (!(gfp_mask & __GFP_FS)) return -1; - prune_dcache(nr); + prune_dcache(nr, NULL); } return (dentry_stat.nr_unused / 100) * sysctl_vfs_cache_pressure; } @@ -820,17 +963,19 @@ 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; @@ -843,19 +988,35 @@ 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); /** @@ -1227,6 +1388,11 @@ 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 @@ -1236,11 +1402,9 @@ 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); } @@ -1331,10 +1495,10 @@ void d_move(struct dentry * dentry, struct dentry * target) */ if (target < dentry) { spin_lock(&target->d_lock); - spin_lock(&dentry->d_lock); + spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED); } else { spin_lock(&dentry->d_lock); - spin_lock(&target->d_lock); + spin_lock_nested(&target->d_lock, DENTRY_D_LOCK_NESTED); } /* Move the dentry to the target hash queue, if on different bucket */ @@ -1359,7 +1523,7 @@ already_unhashed: kfree(target->d_extra_attributes); target->d_extra_attributes = NULL; } - + list_del(&dentry->d_u.d_child); list_del(&target->d_u.d_child); @@ -1388,6 +1552,120 @@ already_unhashed: 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 @@ -1805,6 +2083,7 @@ 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);