On Thu, Mar 8, 2018 at 9:45 PM, Varadarajan, Charulatha
<charulatha.varadarajan(a)intel.com> wrote:
Hi all,
We are seeing an issue in Kabylake Chrome systems where during a 6hrs google
hangout test, we see a system reboot due to GPU hang after some hours.
I am investigating this on Chrome, once it can be reproduced and
validated against libva-2.1.0, I'll post further details once found.
I see the below error in console ramops during the gpu hang.
[drm] GPU HANG: ecode 9:2:0xa8dfbffd, in chrome
According to Chris Wilson in
https://bugs.freedesktop.org/show_bug.cgi?id=103920,
“The first thing to do is check libva/intel-driver. The error state
indicates a userspace error.”
This fdo bug was apparently reported for X11 backend and a different
codec. At this point although a gpu hang signature might look similar
it is not the same use case.
Thanks.
--
Daniel.
A segfault is also observed during the issue.
Watchdog[1381]: segfault at 0 ip 00005c2fd3761a97 sp 00007ae71f305a00 error
6 in chrome[5c2fcf86b000+8601000]
Kindly suggest how to interpret the error GPU HANG: ecode 9:2:0xa8dfbffd.
Thanks
Charulatha
_______________________________________________
intel-vaapi-media mailing list
intel-vaapi-media(a)lists.01.org
https://lists.01.org/mailman/listinfo/intel-vaapi-media