[Cluster-devel] [DLM] Fix dlm_lowcoms_stop hang [28/34]

Steven Whitehouse swhiteho at redhat.com
Tue May 1 10:21:30 UTC 2007


>From 2439fe50724e8693e8b933b3f8125d870bfbdb25 Mon Sep 17 00:00:00 2001
From: Josef Bacik <jwhiter at redhat.com>
Date: Thu, 19 Apr 2007 17:59:05 -0400
Subject: [PATCH] [DLM] Fix dlm_lowcoms_stop hang

When you attempt to release a lockspace in DLM, it will hang trying to down a
semaphore that has already been downed.  The attached patch fixes the problem.

Signed-off-by: Josef Bacik <jwhiter at redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho at redhat.com>
Cc: Patrick Caulfield <pcaulfie at redhat.com>

diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 76399b7..2b32f3c 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -1407,7 +1407,7 @@ void dlm_lowcomms_stop(void)
 	clean_writequeues();
 
 	for (i = 0; i < max_nodeid; i++) {
-		con = nodeid2con(i, 0);
+		con = __nodeid2con(i, 0);
 		if (con) {
 			close_connection(con, true);
 			if (con->othercon)
-- 
1.5.1.2






More information about the Cluster-devel mailing list