ocfs2_dlm: Dlm dispatch was stopping too early
authorKurt Hackel <kurt.hackel@oracle.com>
Wed, 17 Jan 2007 22:57:50 +0000 (14:57 -0800)
committerMark Fasheh <mark.fasheh@oracle.com>
Wed, 7 Feb 2007 20:04:27 +0000 (12:04 -0800)
dlm_dispatch_work was not processing the queued up tasks at
the first sign of the node leaving the domain leading to not
only incompleted tasks but also a mismatch in the dlm refcnt.

Signed-off-by: Kurt Hackel <kurt.hackel@oracle.com>
Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
fs/ocfs2/dlm/dlmrecovery.c

index 2e32fe65c6c212aa1c5f967ae0db59538b887522..8c60ccc7460cfdc020390035b3a82ee5d2543755 100644 (file)
@@ -163,9 +163,6 @@ void dlm_dispatch_work(struct work_struct *work)
        dlm_workfunc_t *workfunc;
        int tot=0;
 
-       if (!dlm_joined(dlm))
-               return;
-
        spin_lock(&dlm->work_lock);
        list_splice_init(&dlm->work_list, &tmp_list);
        spin_unlock(&dlm->work_lock);