[Bug 201957] amdgpu: ring gfx timeout

[bugzilla-daemon at bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=201957

--- Comment #42 from MajorGonzo (majorgonzo at juno.com) ---
I made a change a while back.  I added:
amdgpu.gpu_recovery=1 
as a grub parameter.  I have no other (of the many suggested) parameters set:
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash amdgpu.ppfeaturemask=0xfffd7fff
amdgpu.gpu_recovery=1"

The feature mask was used to enable reducing the top speed of my video card to
reduce heating, and I was using corectrl for that.  However, it was something I
had to set manually after each boot.  Of course, I forgot to do so, and yet it
still stopped occurring.  So in reality, I don't think I need that anymore,
either.

Just checked my linux logs grepping for "ring gfx".  Before the change, I had
multiples each day up to Dec 10th.  Since then, I've had 3.

Also of note - for the last two, it was when I WASN'T playing.  Well, I was
playing a game, but I was AFK.  It seemed when I returned and did something, it
went black then.

Lastly, just to confirm, I checked my change log (my own log), and I did,
indeed, make that change on 10 Dec.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.