On 12/11/2013 05:41 AM, Michal Privoznik wrote:
>>> +static int
>>> +netcfInterfaceObjIsActive(struct netcf_if *iface,
>>> + bool *active)
>>> +{
>>> + int ret = -1;
>>> + unsigned int flags = 0;
>>> +
>>> + virObjectRef(driverState);
>>
>> What's the ref / unref of driverState for ? The code you're
>> replacing doesn't do that, and AFAICT this shouldn't be
>> required since we're always calling this from public API
>> context
>
> Well, for now it is not required, true. But I was thinking if somebody
> rewrites something this part wouldn't need any change as it's already
> foolproof. But I don't mind dropping ref and unref.
Spurious locking now for future code patches is counter-productive; if
someone adds code that needs locking, we can add the locking at that
point (and ideally keeping things lockless is the better approach). So
yes, let's drop the ref and unref.
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
Attachment:
signature.asc
Description: OpenPGP digital signature
-- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list