rbd: call rbd_dev_unprobe() after unwatching and flushing notifies

[ Upstream commit 952c48b0ed18919bff7528501e9a3fff8a24f8cd ]

rbd_dev_unprobe() is supposed to undo most of rbd_dev_image_probe(),
including rbd_dev_header_info(), which means that rbd_dev_header_info()
isn't supposed to be called after rbd_dev_unprobe().

However, rbd_dev_image_release() calls rbd_dev_unprobe() before
rbd_unregister_watch().  This is racy because a header update notify
can sneak in:

  "rbd unmap" thread                   ceph-watch-notify worker

  rbd_dev_image_release()
    rbd_dev_unprobe()
      free and zero out header
                                       rbd_watch_cb()
                                         rbd_dev_refresh()
                                           rbd_dev_header_info()
                                             read in header

The same goes for "rbd map" because rbd_dev_image_probe() calls
rbd_dev_unprobe() on errors.  In both cases this results in a memory
leak.

Fixes: fd22aef8b47c ("rbd: move rbd_unregister_watch() call into rbd_dev_image_release()")
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index bab9c54..274beda 100644 (file)
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -6933,9 +6933,10 @@ static int rbd_dev_header_name(struct rbd_device *rbd_dev)
 
 static void rbd_dev_image_release(struct rbd_device *rbd_dev)
 {
-       rbd_dev_unprobe(rbd_dev);
        if (rbd_dev->opts)
                rbd_unregister_watch(rbd_dev);
+
+       rbd_dev_unprobe(rbd_dev);
        rbd_dev->image_format = 0;
        kfree(rbd_dev->spec->image_id);
        rbd_dev->spec->image_id = NULL;
@@ -6986,7 +6987,7 @@ static int rbd_dev_image_probe(struct rbd_device *rbd_dev, int depth)
 
        ret = rbd_dev_header_info(rbd_dev);
        if (ret)
-               goto err_out_watch;
+               goto err_out_probe;
 
        /*
         * If this image is the one being mapped, we have pool name and
@@ -7035,12 +7036,11 @@ static int rbd_dev_image_probe(struct rbd_device *rbd_dev, int depth)
        return 0;
 
 err_out_probe:
-       rbd_dev_unprobe(rbd_dev);
-err_out_watch:
        if (!depth)
                up_write(&rbd_dev->header_rwsem);
        if (!depth)
                rbd_unregister_watch(rbd_dev);
+       rbd_dev_unprobe(rbd_dev);
 err_out_format:
        rbd_dev->image_format = 0;
        kfree(rbd_dev->spec->image_id);