On Tue, Dec 02, 2008 at 11:06:38AM +0000, Daniel P. Berrange wrote:
On Tue, Dec 02, 2008 at 12:05:10PM +0100, Daniel Veillard wrote:
On Sun, Nov 30, 2008 at 11:53:17PM +0000, Daniel P. Berrange wrote:
This patch makes the UML driver thread safe
[...]
@@ -1395,21 +1485,27 @@ static int umlListDefinedDomains(virConn struct uml_driver *driver = conn->privateData; int got = 0, i;
+ umlDriverLock(driver); for (i = 0 ; i < driver->domains.count && got < nnames ; i++) { + virDomainObjLock(driver->domains.objs[i]); if (!virDomainIsActive(driver->domains.objs[i])) { if (!(names[got++] = strdup(driver->domains.objs[i]->def->name))) { umlReportError(conn, NULL, NULL, VIR_ERR_NO_MEMORY, "%s", _("failed to allocate space for VM name string")); + virDomainObjUnlock(driver->domains.objs[i]); goto cleanup; } } + virDomainObjUnlock(driver->domains.objs[i]); } + umlDriverUnlock(driver);
return got;
cleanup:
since this is an error code path, I would rather change the label to error: ...
That's a good idea - we should define a standard naming for this usage
'cleanup' to be used where error & normal conditions shared the same exit path of a method.
'error' to be used where error path is completely separate from the normal exit path.
yup since we're going to have labels on a lot of functions for exit we should try to standardize their use and name. The set of patches does this mostly, just a small incremental improvement, and then we can document it in the HACKING file. 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/