[Bug 201957] amdgpu: ring gfx timeout

[bugzilla-daemon at kernel.org]

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

--- Comment #96 from KC (kcohar at gmail.com) ---
The default is now -1.
https://unix.stackexchange.com/questions/756281/kernel-6-5-2-seems-to-have-amdgpu-crash-on-no-retry-page-fault
https://www.kernel.org/doc/html/v6.5/gpu/amdgpu/module-parameters.html

I set it to zero and I haven't had a single crash since (Fedora 39 beta,
Linux 6.5.5).
This one parameter change made my system entirely unusable (it would crash
very quickly after booting).


On Sat, Sep 30, 2023 at 9:35 PM <bugzilla-daemon at kernel.org> wrote:

> https://bugzilla.kernel.org/show_bug.cgi?id=201957
>
> --- Comment #95 from aspicer at gmail.com ---
> (In reply to KC from comment #94)
>
> Did you have it set to 1 previously? If not, I'm not sure if that was the
> silver bullet, because it looks like it defaults to 0.
> https://dri.freedesktop.org/docs/drm/gpu/amdgpu.html
>
> mcbp (int)
>
> It is used to enable mid command buffer preemption. (0 = disabled
> (default), 1
> = enabled)
>
> --
> You may reply to this email to add a comment.
>
> You are receiving this mail because:
> You are on the CC list for the bug.

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

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