ocfs2: Wakeup the downconvert thread after a successful cancel convert

When two nodes holding PR locks on a resource concurrently attempt to
upconvert the locks to EX, the master sends a BAST to one of the nodes. This
message tells that node to first cancel convert the upconvert request,
followed by downconvert to a NL. Only when this lock is downconverted to NL,
can the master upconvert the first node's lock to EX.

While the fs was doing the cancel convert, it was forgetting to wake up the
dc thread after a successful cancel, leading to a deadlock.

Reported-and-Tested-by: David Teigland <teigland@redhat.com>
Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>

diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index b0c4cadd4c45181f7d8d02f1af11a182e2f380dd..206a2370876a8c82f4c6ad04f101a7d801be6e5b 100644 (file)
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -2860,6 +2860,10 @@ static void ocfs2_unlock_ast(void *opaque, int error)
        case OCFS2_UNLOCK_CANCEL_CONVERT:
                mlog(0, "Cancel convert success for %s\n", lockres->l_name);
                lockres->l_action = OCFS2_AST_INVALID;
+               /* Downconvert thread may have requeued this lock, we
+                * need to wake it. */
+               if (lockres->l_flags & OCFS2_LOCK_BLOCKED)
+                       ocfs2_wake_downconvert_thread(ocfs2_get_lockres_osb(lockres));
                break;
        case OCFS2_UNLOCK_DROP_LOCK:
                lockres->l_level = DLM_LOCK_IV;