[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 12/28] libxl: only free cputopology if it was allocated in libxl__get_numa_candidate

On Wed, Sep 18, 2013 at 9:10 PM, Dario Faggioli
<dario.faggioli@xxxxxxxxxx> wrote:
> On mer, 2013-09-18 at 20:42 +1200, Matthew Daley wrote:
>> On Wed, Sep 18, 2013 at 7:57 PM, Dario Faggioli
>> > Also, if something like that would be necessary, why doing it only for
>> > tinfo and not for ninfo as well? I'm not that into coverity, but I don't
>> > think I see why it treats the two of them differently... :-O
>> Not sure. Probably because if ninfo == NULL, we return directly out of
>> the function instead of goto-ing the cleanup.
> Right, which, now that I've seen it, calls for a patch making it go
> through cleanup too! :-P
> Anyway, for libxl_cputopology_list_free() (and also for
> libxl_numainfo_list_free() if I really decide to patch it to make it
> more "libxl style"), especially considering that something similar to
> this happens in other places (as we're seeing in 13/28), is there
> anything we can do for preventing this to come up again in future scans?

Simply marking the Defects in Coverity as "False Positive" (which I
just did before) prevents them from bugging us in the future.

A quick grep shows that the functions are only called in ~11 spots
AFAICT; that's still small enough to sensibly allow manual overriding
if any of the other invocations are complained about. However, I can't
see any others from a quick look (I haven't figured out a way, if
there is one, to go from a function *out* to its invocations which are
tagged, unfortunately.)

- Matthew

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.