[Cluster-devel] [PATCH] rgmanager: Fix short read handling while loading resource rules
Lon Hohberger <[email protected]>
| Newsgroups | com.redhat.cluster-devel |
|---|---|
| Message-ID | <[email protected]> |
If a short read occurs (maybe due to a race between the parent/child
processes in SMP systems?) while reading from the pipe while using
resource rules, we break the loop early even if there's more data to be
had at a later time.
It fixed the problem for the reporter.
-- Lon
Index: resrules.c
===================================================================
RCS file: /cvs/cluster/cluster/rgmanager/src/daemons/resrules.c,v
retrieving revision 1.16.2.9
diff -u -r1.16.2.9 resrules.c
--- resrules.c 18 Dec 2007 17:52:56 -0000 1.16.2.9
+++ resrules.c 30 Jan 2008 18:35:14 -0000
@@ -906,7 +906,7 @@
if (n == 0 && (!*length))
return 0;
- if (n != sizeof(buf)) {
+ if (n == 0) {
done = 1;
}