First, thanks your answer. On Wed, Dec 09, 2020 at 10:25:19PM +0100, Charlene Wendling wrote: > Hi, > > On Mon, 7 Dec 2020 18:16:03 +0100 > Alex Raschi wrote: > > > Sorry i was trying to reply to this thread: > > > > https://marc.info/?l=openbsd-bugs&m=159383256417523&w=2 > > > > On Mon, Dec 07, 2020 at 06:08:02PM +0100, Alex Raschi wrote: > > > Hi, > > > > > > I have a similar problem but with a RX 560, i got some logs for a > > > bug report then i saw this issue open. The screen becomes > > > unresponsive here too, with xf86-video-amdgpu the issue appears > > > quickly: sometimes running 2 games for some seconds is enough. > > I have a Sapphire RX 560 4GB (this one [0]). It used to have the same > behaviour you're describing. Post-release (october iirc), jsg@ brought > many fixes for amdgpu, and since then i met no issues. > > If you're not following -current, you may want to give it a try.
We have exactly the same gpu, i'm already on -current and i update it frequently, i did an upgrade yesterday too. The bug is still there unfortunately, so if you don't have this issue i think there is something different in your setup (i doubt that hardware other than the gpu is important here). Do you have some configuration in /etc/X11/xorg.conf.d? Compositor, Desktop Environment, ...? > > > But with xf86-video-modesetting the system is stabler, with this > > > driver usually i am able to use X11 for some hours or days (with > > > suspend). > > > > > > I don't know if it is related or not but with xf86-video-amdgpu > > > libGL is unable to use DRI3 and says: > > > > > > libGL: screen 0 does not appear to be DRI3 capable > > > ... > > > libGL: Using DRI2 for screen 0 > > > > > > With xf86-video-modesetting DRI3 works without issues. > > > Here are some examples of dmesg errors i get when X11 freezes: > > > > > > [drm] *ERROR* Waiting for fences timed out![drm] *ERROR* ring sdma1 > > > timeout, signaled seq=2236, emitted seq=2236 [drm] *ERROR* Process > > > information: process pid 0 thread pid 0 > > > > > > [drm] *ERROR* Waiting for fences timed out! > > > > > > [drm] *ERROR* ring gfx timeout, but soft recovered > > > [drm] *ERROR* ring sdma0 timeout, signaled seq=22318, emitted > > > seq=22318 [drm] *ERROR* Process information: process pid 0 thread > > > pid 0 > > > > > > [drm] *ERROR* ring sdma1 timeout, signaled seq=12928, emitted > > > seq=12928 [drm] *ERROR* Process information: process pid 0 thread > > > pid 0 > > > > > > [drm] *ERROR* ring sdma1 timeout, signaled seq=38, emitted seq=38 > > > [drm] *ERROR* Process information: process pid 0 thread pid 0 > > > > > > Should i open a new issue and send the complete logs? > > > > > > Cheers, Alex > > > > [0] https://www.sapphiretech.com/en/consumer/pulse-rx-560-4g-g5-14-cu
