Re: [Xen-devel] [PATCH] passthrough/amd: avoid reading an uninitialized variable.

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 16.04.15 at 19:19, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 16/04/15 17:44, Tim Deegan wrote:
>> update_intremap_entry_from_msi() doesn't write to its data pointer on
>> some error paths, so we copying that variable into the msg would count
>> as undefined behaviour.
>>
>> Signed-off-by: Tim Deegan <tim@xxxxxxx>
>> Cc: Suravee Suthikulpanit <suravee.suthikulpanit@xxxxxxx>
>> Cc: Aravind Gopalakrishnan <Aravind.Gopalakrishnan@xxxxxxx>
> 
> This looks curiously familiar.
> 
> http://lists.xen.org/archives/html/xen-devel/2015-03/msg03679.html 
> 
> I was meaning to get around to submitting it as a formal patch, but you
> have definitely beaten me to it :)

But sadly it didn't help with the issue Sander reported (and I also
didn't expect it to, as errors here should cause the caller to not
use msg at all, and the regressing patch didn't alter anything that
could have caused errors to surface where none where seen
before). Still the patch is a good one as correcting a language
spec violation.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel