VFS: allow ->d_manage() to declare -EISDIR in rcu_walk mode.

In REF-walk mode, ->d_manage can return -EISDIR to indicate
that the dentry is not really a mount trap (or even a mount point)
and that any mounts or any DCACHE_NEED_AUTOMOUNT flag should be
ignored.

RCU-walk mode doesn't currently support this, so if there is a dentry
with DCACHE_NEED_AUTOMOUNT set but which shouldn't be a mount-trap,
lookup_fast() will always drop in REF-walk mode.

With this patch, an -EISDIR from ->d_manage will always cause mounts
and automounts to be ignored, both in REF-walk and RCU-walk.

Bug-fixed-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Ian Kent <raven@themaw.net>
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

2 files changed, 18 insertions, 12 deletions

diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt
index a1d0d7a3016..61d65cc65c5 100644
--- a/Documentation/filesystems/vfs.txt
+++ b/Documentation/filesystems/vfs.txt
@@ -1053,7 +1053,8 @@ struct dentry_operations {
 	If the 'rcu_walk' parameter is true, then the caller is doing a
 	pathwalk in RCU-walk mode.  Sleeping is not permitted in this mode,
 	and the caller can be asked to leave it and call again by returning
-	-ECHILD.
+	-ECHILD.  -EISDIR may also be returned to tell pathwalk to
+	ignore d_automount or any mounts.
 
 	This function is only used if DCACHE_MANAGE_TRANSIT is set on the
 	dentry being transited from.
diff --git a/fs/namei.c b/fs/namei.c
index 0ff23cecb1b..8a217c48f6d 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -1091,10 +1091,10 @@ int follow_down_one(struct path *path)
 }
 EXPORT_SYMBOL(follow_down_one);
 
-static inline bool managed_dentry_might_block(struct dentry *dentry)
+static inline int managed_dentry_rcu(struct dentry *dentry)
 {
-	return (dentry->d_flags & DCACHE_MANAGE_TRANSIT &&
-		dentry->d_op->d_manage(dentry, true) < 0);
+	return (dentry->d_flags & DCACHE_MANAGE_TRANSIT) ?
+		dentry->d_op->d_manage(dentry, true) : 0;
 }
 
 /*
@@ -1110,11 +1110,18 @@ static bool __follow_mount_rcu(struct nameidata *nd, struct path *path,
 		 * Don't forget we might have a non-mountpoint managed dentry
 		 * that wants to block transit.
 		 */
-		if (unlikely(managed_dentry_might_block(path->dentry)))
+		switch (managed_dentry_rcu(path->dentry)) {
+		case -ECHILD:
+		default:
 			return false;
+		case -EISDIR:
+			return true;
+		case 0:
+			break;
+		}
 
 		if (!d_mountpoint(path->dentry))
-			return true;
+			return !(path->dentry->d_flags & DCACHE_NEED_AUTOMOUNT);
 
 		mounted = __lookup_mnt(path->mnt, path->dentry);
 		if (!mounted)
@@ -1130,7 +1137,8 @@ static bool __follow_mount_rcu(struct nameidata *nd, struct path *path,
 		 */
 		*inode = path->dentry->d_inode;
 	}
-	return read_seqretry(&mount_lock, nd->m_seq);
+	return read_seqretry(&mount_lock, nd->m_seq) &&
+		!(path->dentry->d_flags & DCACHE_NEED_AUTOMOUNT);
 }
 
 static int follow_dotdot_rcu(struct nameidata *nd)
@@ -1402,11 +1410,8 @@ static int lookup_fast(struct nameidata *nd,
 		}
 		path->mnt = mnt;
 		path->dentry = dentry;
-		if (unlikely(!__follow_mount_rcu(nd, path, inode)))
-			goto unlazy;
-		if (unlikely(path->dentry->d_flags & DCACHE_NEED_AUTOMOUNT))
-			goto unlazy;
-		return 0;
+		if (likely(__follow_mount_rcu(nd, path, inode)))
+			return 0;
 unlazy:
 		if (unlazy_walk(nd, dentry))
 			return -ECHILD;