Skip to content

Commit

Permalink
drbd: fix race between deconfiguring and reconfiguring network
Browse files Browse the repository at this point in the history
If a drbd_nl_net_conf hits the small window between the state change
to C_STANDALONE and the corresponding cleanup in after_state_ch,
that cleanup would throw away stuff we now need again,
and later trigger BUG_ON()s.

Fixed by properly serializing the new config request with
any pending cleanup.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
  • Loading branch information
Lars Ellenberg authored and Philipp Reisner committed Oct 14, 2010
1 parent 0778286 commit c518d04
Showing 1 changed file with 4 additions and 2 deletions.
6 changes: 4 additions & 2 deletions drivers/block/drbd/drbd_nl.c
Original file line number Diff line number Diff line change
Expand Up @@ -750,14 +750,16 @@ void drbd_setup_queue_param(struct drbd_conf *mdev, unsigned int max_seg_s) __mu
/* serialize deconfig (worker exiting, doing cleanup)
* and reconfig (drbdsetup disk, drbdsetup net)
*
* wait for a potentially exiting worker, then restart it,
* or start a new one.
* Wait for a potentially exiting worker, then restart it,
* or start a new one. Flush any pending work, there may still be an
* after_state_change queued.
*/
static void drbd_reconfig_start(struct drbd_conf *mdev)
{
wait_event(mdev->state_wait, !test_and_set_bit(CONFIG_PENDING, &mdev->flags));
wait_event(mdev->state_wait, !test_bit(DEVICE_DYING, &mdev->flags));
drbd_thread_start(&mdev->worker);
drbd_flush_workqueue(mdev);
}

/* if still unconfigured, stops worker again.
Expand Down

0 comments on commit c518d04

Please sign in to comment.