On Wed, Feb 03, 2010 at 07:57:42AM -0500, Cole Robinson wrote:
On 02/03/2010 03:33 AM, Jiri Denemark wrote:
...
diff --git a/src/util/util.c b/src/util/util.c index cf1290d..901c0d2 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -415,12 +415,19 @@ __virExec(virConnectPtr conn, childerr = null; }
+ /* Ensure we hold the logging lock, to protect child processes + * from deadlocking on another threads inheirited mutex state */ + virLogLock(); + if ((pid = fork()) < 0) { virReportSystemError(conn, errno, "%s", _("cannot fork child process")); goto cleanup; }
+ /* Unlock for both parent and child process */ + virLogUnlock(); + if (pid) { /* parent */ close(null); if (outfd && *outfd == -1) {
Hmm, shouldn't we virLogUnlock() even if fork() fails? That is, something like:
virLogLock(); pid = fork(); virLogUnlock();
if (pid < 0) error; else if (pid) parent; ...
Jirka
Argh, yes, good call. I'll update the patch and resend.
Thanks, Cole
Okay I did the small change and commited the patch accordingly, Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel@veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/