Re: [libvirt] [PATCH v2] Remove qemuDriverLock from almost everywhere

On 02/11/2013 09:47 AM, Daniel P. Berrange wrote: > From: "Daniel P. Berrange" <berrange(a)redhat.com&gt; > > With the majority of fields in the virQEMUDriverPtr struct > now immutable or self-locking, there is no need for practically > any methods to be using the QEMU driver lock. Only a handful > of helper APIs in qemu_conf.c now need it I agree with squashing in Martin's cleanups. Beyond that, the locking changes looked sane. ACK. Let's get this in and stress-tested before the 1.0.3 release. > +++ b/src/qemu/qemu_driver.c > @@ -153,16 +153,9 @@ virQEMUDriverPtr qemu_driver = NULL; > > > static void > -qemuVMDriverLock(void) { > - qemuDriverLock(qemu_driver); > -}; > - > - > +qemuVMDriverLock(void) {} > static void > -qemuVMDriverUnlock(void) { > - qemuDriverUnlock(qemu_driver); > -}; > - > +qemuVMDriverUnlock(void) {} Do we need to keep these functions any more, or can we get rid of them in a followup patch?

> cleanup: > - if (watchdogEvent || lifecycleEvent) { > - qemuDriverLock(driver); > - if (watchdogEvent) > - qemuDomainEventQueue(driver, watchdogEvent); > - if (lifecycleEvent) > - qemuDomainEventQueue(driver, lifecycleEvent); > - qemuDriverUnlock(driver); > - } > + if (watchdogEvent) > + qemuDomainEventQueue(driver, watchdogEvent); > + if (lifecycleEvent) > + qemuDomainEventQueue(driver, lifecycleEvent); Is it worth a followup patch to qemuDomainEventQueue that becomes a no-op when passed a NULL event, so we can let the callers avoid the 'if'?