Commit 01b172b7b10146cf5f02604047bee065cfb49946

Authored by Bob Peterson
Committed by Steven Whitehouse
1 parent 48f8f711ed

GFS2: Ensure workqueue is scheduled after noexp request

This patch closes a small timing window whereby a request to hold the
transaction glock can get stuck. The problem is that after the DLM has
granted the lock, it can get into a state whereby it doesn't transition
the glock to a held state, due to not having requeued the glock state
machine to finish the transition.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

Showing 1 changed file with 6 additions and 2 deletions Side-by-side Diff

... ... @@ -1047,9 +1047,13 @@
1047 1047  
1048 1048 spin_lock(&gl->gl_spin);
1049 1049 add_to_queue(gh);
1050   - if ((LM_FLAG_NOEXP & gh->gh_flags) &&
1051   - test_and_clear_bit(GLF_FROZEN, &gl->gl_flags))
  1050 + if (unlikely((LM_FLAG_NOEXP & gh->gh_flags) &&
  1051 + test_and_clear_bit(GLF_FROZEN, &gl->gl_flags))) {
1052 1052 set_bit(GLF_REPLY_PENDING, &gl->gl_flags);
  1053 + gl->gl_lockref.count++;
  1054 + if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0)
  1055 + gl->gl_lockref.count--;
  1056 + }
1053 1057 run_queue(gl, 1);
1054 1058 spin_unlock(&gl->gl_spin);
1055 1059