[PULL: xserver master] misc. bug fixes & cleanups, including CVE-2013-4396

[Keith Packard]

Alan Coopersmith <alan.coopersmith at oracle.com> writes:

> On 10/16/13 01:38 PM, Keith Packard wrote:
>> Alan Coopersmith <alan.coopersmith at oracle.com> writes:
>>
>>> Alan Coopersmith (4):
>>>         Skip damage calls if DamageCreate fails in exa functions
>>
>>          Seems like this is going to break stuff badly when this
>>          happens. Is there some way to recover from this allocation
>>          failure?
>
> Sorry, no idea - I was just mimicing the code in other places that calls
> DamageCreate to make the static analyzer happy, and haven't had time to
> dig any deeper.

I've gone ahead and merged this change; missing rendering seems a far
better situation than dereferencing a null pointer. If someone using EXA
wants to improve that by tracking these failures and recovering, that'd
be great.

A simple solution could be to allocate the damage object when the pixmap
is created; that'd avoid the failure case here and allow the server to
report BadAlloc at pixmap creation time where the application can be
expected to deal with it.

-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 827 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg-devel/attachments/20131018/ed141497/attachment.pgp>