diff --git a/fs/dcache.c b/fs/dcache.c index 4d13bf50b7b1..d56a40b5a577 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1332,31 +1332,13 @@ EXPORT_SYMBOL(d_add_ci); * d_lookup - search for a dentry * @parent: parent dentry * @name: qstr of name we wish to find + * Returns: dentry, or NULL * - * Searches the children of the parent dentry for the name in question. If - * the dentry is found its reference count is incremented and the dentry - * is returned. The caller must use dput to free the entry when it has - * finished using it. %NULL is returned on failure. - * - * __d_lookup is dcache_lock free. The hash list is protected using RCU. - * Memory barriers are used while updating and doing lockless traversal. - * To avoid races with d_move while rename is happening, d_lock is used. - * - * Overflows in memcmp(), while d_move, are avoided by keeping the length - * and name pointer in one structure pointed by d_qstr. - * - * rcu_read_lock() and rcu_read_unlock() are used to disable preemption while - * lookup is going on. - * - * The dentry unused LRU is not updated even if lookup finds the required dentry - * in there. It is updated in places such as prune_dcache, shrink_dcache_sb, - * select_parent and __dget_locked. This laziness saves lookup from dcache_lock - * acquisition. - * - * d_lookup() is protected against the concurrent renames in some unrelated - * directory using the seqlockt_t rename_lock. + * d_lookup searches the children of the parent dentry for the name in + * question. If the dentry is found its reference count is incremented and the + * dentry is returned. The caller must use dput to free the entry when it has + * finished using it. %NULL is returned if the dentry does not exist. */ - struct dentry * d_lookup(struct dentry * parent, struct qstr * name) { struct dentry * dentry = NULL; @@ -1372,6 +1354,21 @@ struct dentry * d_lookup(struct dentry * parent, struct qstr * name) } EXPORT_SYMBOL(d_lookup); +/* + * __d_lookup - search for a dentry (racy) + * @parent: parent dentry + * @name: qstr of name we wish to find + * Returns: dentry, or NULL + * + * __d_lookup is like d_lookup, however it may (rarely) return a + * false-negative result due to unrelated rename activity. + * + * __d_lookup is slightly faster by avoiding rename_lock read seqlock, + * however it must be used carefully, eg. with a following d_lookup in + * the case of failure. + * + * __d_lookup callers must be commented. + */ struct dentry * __d_lookup(struct dentry * parent, struct qstr * name) { unsigned int len = name->len; @@ -1382,6 +1379,19 @@ struct dentry * __d_lookup(struct dentry * parent, struct qstr * name) struct hlist_node *node; struct dentry *dentry; + /* + * The hash list is protected using RCU. + * + * Take d_lock when comparing a candidate dentry, to avoid races + * with d_move(). + * + * It is possible that concurrent renames can mess up our list + * walk here and result in missing our dentry, resulting in the + * false-negative result. d_lookup() protects against concurrent + * renames using rename_lock seqlock. + * + * See Documentation/vfs/dcache-locking.txt for more details. + */ rcu_read_lock(); hlist_for_each_entry_rcu(dentry, node, head, d_hash) { @@ -1396,8 +1406,8 @@ struct dentry * __d_lookup(struct dentry * parent, struct qstr * name) /* * Recheck the dentry after taking the lock - d_move may have - * changed things. Don't bother checking the hash because we're - * about to compare the whole name anyway. + * changed things. Don't bother checking the hash because + * we're about to compare the whole name anyway. */ if (dentry->d_parent != parent) goto next; diff --git a/fs/namei.c b/fs/namei.c index b815a4d2e1d6..11de7c39ff76 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -735,6 +735,11 @@ static int do_lookup(struct nameidata *nd, struct qstr *name, return err; } + /* + * Rename seqlock is not required here because in the off chance + * of a false negative due to a concurrent rename, we're going to + * do the non-racy lookup, below. + */ dentry = __d_lookup(nd->path.dentry, name); if (!dentry) goto need_lookup; @@ -754,17 +759,13 @@ need_lookup: mutex_lock(&dir->i_mutex); /* * First re-do the cached lookup just in case it was created - * while we waited for the directory semaphore.. + * while we waited for the directory semaphore, or the first + * lookup failed due to an unrelated rename. * - * FIXME! This could use version numbering or similar to - * avoid unnecessary cache lookups. - * - * The "dcache_lock" is purely to protect the RCU list walker - * from concurrent renames at this point (we mustn't get false - * negatives from the RCU list walk here, unlike the optimistic - * fast walk). - * - * so doing d_lookup() (with seqlock), instead of lockfree __d_lookup + * This could use version numbering or similar to avoid unnecessary + * cache lookups, but then we'd have to do the first lookup in the + * non-racy way. However in the common case here, everything should + * be hot in cache, so would it be a big win? */ dentry = d_lookup(parent, name); if (likely(!dentry)) { @@ -1136,13 +1137,12 @@ static struct dentry *__lookup_hash(struct qstr *name, goto out; } - dentry = __d_lookup(base, name); - - /* lockess __d_lookup may fail due to concurrent d_move() - * in some unrelated directory, so try with d_lookup + /* + * Don't bother with __d_lookup: callers are for creat as + * well as unlink, so a lot of the time it would cost + * a double lookup. */ - if (!dentry) - dentry = d_lookup(base, name); + dentry = d_lookup(base, name); if (dentry && dentry->d_op && dentry->d_op->d_revalidate) dentry = do_revalidate(dentry, nd);