[Bug 102177] [SKL] ES31-CTS.core.sepshaderobjs.StateInteraction fails sporadically

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Mon Aug 14 04:10:31 UTC 2017


https://bugs.freedesktop.org/show_bug.cgi?id=102177

--- Comment #3 from Kenneth Graunke <kenneth at whitecape.org> ---
This isn't Tim's fault.

The problem is that the test tries to re-link a program which is currently
bound for rendering...but with a bogus shader, which causes a link failure. 
The GL specification (I can't remember where exactly) says that in this case,
the previous program remains bound for rendering and is not destroyed until
it's unbound.

When re-linking, we immediately trash the program...but because linking fails,
and we don't get a new one...we end up leaving the old freed program bound.  So
when you go to draw...you hit a use-after-free...which may work or fail
depending on the malloc calls in the meantime.

You can easily see this problem by using 'valgrind'.  This is a nasty issue,
and I'm not exactly sure what to do about it.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20170814/a4f8974d/attachment.html>


More information about the intel-3d-bugs mailing list