btrfs: fix a deadlock in btrfs_scan_one_device()

pathname resolution under a global mutex, taken on some paths in ->mount()
is a Bad Idea(tm) - think what happens if said pathname resolution triggers
automount of some btrfs instance and walks into attempt to grab the same
mutex.  Deadlock - we are waiting for daemon to finish walking the path,
daemon is waiting for us to release the mutex...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index f4b839fd3c9dd5cd854cb7bada4e3831d8ea1713..96bb3c0a79b08da610f41ed7442cec92cd8973c2 100644 (file)
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -706,8 +706,6 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
        u64 devid;
        u64 transid;
 
-       mutex_lock(&uuid_mutex);
-
        flags |= FMODE_EXCL;
        bdev = blkdev_get_by_path(path, flags, holder);
 
@@ -716,6 +714,7 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
                goto error;
        }
 
+       mutex_lock(&uuid_mutex);
        ret = set_blocksize(bdev, 4096);
        if (ret)
                goto error_close;
@@ -737,9 +736,9 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 
        brelse(bh);
 error_close:
+       mutex_unlock(&uuid_mutex);
        blkdev_put(bdev, flags);
 error:
-       mutex_unlock(&uuid_mutex);
        return ret;
 }