Created attachment 28679 [details]
dmesg captured from terminal using ssh.
I was running modified version of mesa when I got GPU hang. (My changes may have broken state)
But effects of gpu hang were fatal:
-dmesg has message that radeon was trying to reset GPU
-ring test failed so reset failed
-dmesg is spammed with messages that scheduling IB failed
-whole computer is unresponsive localy
-open ssh connection works for only few commands (luckily dmesg did work)
-It seems like no disk access can happen after hang
-No new ssh connection are possible
-Old ssh connection did freeze soon so not much useful info there
After reboot logs didn't include any info from hang or after it.
I have seen before similar hang with M9+ card but no ssh connected then so no debug info.
Created attachment 28680 [details]
lspci for hardware details.
kernel is vanila 2.6.31-rc6
mesa master from yesterday with my patches applied+minor hacking
libdrm master 1d465178fbab77a9c
xf86-video-ati master cd99d9f0
PS. Afre ssh freeze I did try Sysrq+S,K,R and Then just B. Nothing did happend to first 3 but reboot worked.
Created attachment 28681 [details]
dmesg from 2nd hang
ok. I can reproduce this hang with my broken mesa.
This time I got IB info over ssh before I lost control.
Created attachment 28682 [details]
all IB buffers dumped to single file
-- GitLab Migration Automatic Message --
This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.
You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/amd/issues/65.