From 2cdc98aaf072d573df10c503d3b3b0b74e2a6d06 Mon Sep 17 00:00:00 2001 From: David Teigland Date: Tue, 31 Oct 2006 11:56:08 -0600 Subject: [DLM] fix stopping unstarted recovery Red Hat BZ 211914 When many nodes are joining a lockspace simultaneously, the dlm gets a quick sequence of stop/start events, a pair for adding each node. dlm_controld in user space sends dlm_recoverd in the kernel each stop and start event. dlm_controld will sometimes send the stop before dlm_recoverd has had a chance to take up the previously queued start. The stop aborts the processing of the previous start by setting the RECOVERY_STOP flag. dlm_recoverd is erroneously clearing this flag and ignoring the stop/abort if it happens to take up the start after the stop meant to abort it. The fix is to check the sequence number that's incremented for each stop/start before clearing the flag. Signed-off-by: David Teigland Signed-off-by: Steven Whitehouse --- fs/dlm/recoverd.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'fs/dlm') diff --git a/fs/dlm/recoverd.c b/fs/dlm/recoverd.c index 4a1d6023fd9..6e4ee94ce7d 100644 --- a/fs/dlm/recoverd.c +++ b/fs/dlm/recoverd.c @@ -219,6 +219,10 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv) return error; } +/* The dlm_ls_start() that created the rv we take here may already have been + stopped via dlm_ls_stop(); in that case we need to leave the RECOVERY_STOP + flag set. */ + static void do_ls_recovery(struct dlm_ls *ls) { struct dlm_recover *rv = NULL; @@ -226,7 +230,8 @@ static void do_ls_recovery(struct dlm_ls *ls) spin_lock(&ls->ls_recover_lock); rv = ls->ls_recover_args; ls->ls_recover_args = NULL; - clear_bit(LSFL_RECOVERY_STOP, &ls->ls_flags); + if (rv && ls->ls_recover_seq == rv->seq) + clear_bit(LSFL_RECOVERY_STOP, &ls->ls_flags); spin_unlock(&ls->ls_recover_lock); if (rv) { -- cgit v1.2.3-18-g5258