[Cluster-devel] [DLM][PATCH] another semaphore lockup fix
Josef Bacik
jwhiter at redhat.com
Fri Apr 20 20:16:00 UTC 2007
Hello,
Should have caught this yesterday but I didn't test until today. We already
hold the connection semaphore when we get here, so we need to use __nodeid2con
instead of nodeid2con. Thank you,
Josef
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 2b32f3c..6a4577d 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -1351,7 +1351,7 @@ static void clean_writequeues(void)
int nodeid;
for (nodeid = 1; nodeid < max_nodeid; nodeid++) {
- struct connection *con = nodeid2con(nodeid, 0);
+ struct connection *con = __nodeid2con(nodeid, 0);
if (con)
clean_one_writequeue(con);
More information about the Cluster-devel
mailing list