Chris Cappuccio schreef op do 21-10-2021 om 07:56 [-0700]: > This appears to be a totally different failure than found by Avon. > It's also on a different class of hardware.
Is it? The errors seem very similar. Also, Avon's dmesg suggests he's using amdgpu as well. I wonder why he deletes all the amdgpu firmwares only to copy over the radeon firmwares. For me, that would break xenocara. > My suggestion to try an older firmware was based on a specific bug > report where someone 1. using Linux had a very similar crash 2. on > very similar hardware, and decided to 3. downgrade firmware and saw > some sort of improvements. Yeah, I tried that. Same result. > > Oct 18 01:52:28 polaris /bsd: [drm] *ERROR* Illegal register access > > in > > command stream > > Oct 18 01:52:28 polaris /bsd: [drm] *ERROR* ring gfx_0.0.0 timeout, > > signaled seq=129772, emitted seq=129774 > > Oct 18 01:52:28 polaris /bsd: [drm] *ERROR* Process information: > > process pid 0 thread Xorg pid 84104 > > > > So, what is this? Is this from before you altered the firmware > version? > Or after? Straight from the most recent snapshot as of the 18th. But I saw the same errors ith downgraded firmware. No need to spam the list with syslog spam since it's the same. > You should report issues like this with debugging information that > represents the stock installation first. I've seen the same error with 7.0 stock, snapshot and snapshot with downgraded hardware. I'm just trying to help pin this down. Right now OpenBSD isn't really useful as a desktop. Regards, Emiel
