AMDGPU Reset Recovery To Be Flipped On By Default For Newer Radeon GPUs

[mv.gavrilov]

100% reproduction case GPU hang for Vega 56/64:
Launch The Evil Within Demo https://store.steampowered.com/app/2...e_Evil_Within/ via steam play and watch first cut scene.

[obedlink]

Have you had any Radeon GPU hangs under Linux recently? Or any particular Radeon Linux driver bugs still biting you?

every time I start the saga metro 2033, it does not matter if they are the original version or redux.

after the introduction video the screen turns black and nothing can be done anymore, that happens to me from Metro was launched in linux, and no matter what version of GNU / Linux or MESA3D I use, the same thing always happens.

now I'm with ubuntu 18.10 and the repositories of padoka and the problem persists to this day, it always happens to me with my Radeon R9 290.

[Zyklon]

…what will eventually be either Linux 4.21 or Linux 5.1

Or 5.0.

[Med_]

What is your system specs? Do you measure temps and do you have good airflow in the case? Is your computer clean? That reset patch does not solve anything, tested with RX560. When X freezes, it does not help X to recover.

RX480. I will try to log temperature to see whether that happens on possible temperature spikes. However the GPU fans are not at max speed when this occurs (in fact as soon as there is the hang, they do spin up to max speed). The case is clean and the airflow should be OK. I can increase a bit the case fan speed (linked to CPU fan speed) to see whether that help.

One maybe interesting bit, is that I nearly get a checkered pattern on the screen when the hang occurs.

[M@GOid]

What is your system specs? Do you measure temps and do you have good airflow in the case? Is your computer clean?

That is a very important aspect on a system performance and stability, that generally goes under people's radar. Laptops can be particularly sensible to dirt and poor air circulation. PCs need hot air to be taken out of the case. I personally have cut fan grids out of 2 cases of mine to improve efficiency, since things get hot in the summer where I live and air conditioning is not a option.

[boxie]

It would be worth a try. How do you alter that with AMDGPU?

I used a pwm fan control app that does it for me. Not currently in front of my machine atm, but a quick Google search should yield what you want

[M@GOid]

The hangs I have is while gaming. Desktop under KDE 5 is very stable.

On games it generally is on the bleeding edge side of drivers and poor ported Linux games. Steam Proton have some nasty hangs, but that's expected since it's highly experimental on today's stage.

[Med_]

What GPU?

RX480.

[Med_]

I have found the only reason my GPU hangs (vega64) is because it gets too hot. happens under both windows and linux - so, I have got my fan profiles cranked up. I can get many hour gaming sessions in, albeit with more fan noise.

It would be worth a try. How do you alter that with AMDGPU?

[puleglot]

This is good news. I consistently get hangs with games. Typically once every few hours. I do not bother reporting them as I cannot reproduce on demand and the bug tracker is full of them with similar logs ([drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, last signaled seq=145101, last emitted seq=145103). I have activated the option, we will see whether that at least prevents the power button treatment.

What GPU?