Eric Blake wrote:
On 05/17/2010 11:40 AM, Jim Meyering wrote:
> Eric Blake wrote:
>> On 05/17/2010 11:22 AM, Jim Meyering wrote:
>>> This addresses another coverity-spotted "flaw".
>>> However, since "cgroup" is never NULL after that initial
"if" stmt,
>>> the only penalty is that the useless cleanup test would make a reviewer
>>> try to figure out how cgroup could be NULL there.
>>
>> ACK.
>
> Thanks.
>
>>> cleanup:
>>> - if (cgroup)
>>> - virCgroupFree(&cgroup);
>>> + virCgroupFree(&cgroup);
Hmm, on re-reading this, why is virCgroupFree taking a pointer to a
virCgroupPtr, then blindly dereferencing it throughout the cleanup? It
almost seems like we have the wrong signature, and should be using:
virCgroupFree(virCgroupPtr group)
instead of
virCgroupFree(virCgroupPtr *group)
and adjust all callers.
Almost.
virCgroupPtr takes an address of a pointer so it can set the caller's
pointer to NULL (via its VIR_FREE use). Otherwise, in order to retain
existing semantics, some callers would have to change from this:
virCgroupFree(&p);
to this:
virCgroupFree(p);
p = NULL;
> This makes the useless-if-before-free test in maint.mk spot
> uses of virCgroupFree just like it does for free and the other
> listed functions.
> * cfg.mk (useless_free_options): Add virCgroupFree.
> Prompted by suggestion from Eric Blake.
ACK, given the current semantics of virCgroupFree.
Thanks.