black screen/death on X

[csiga77]

Hi,

What happens is, randomly my monitors go black, and never wakes up. Not even C+A+F1 (textmode terminal) works. Looks like the video card goes "mad" .
Fortunately I can shut it down via ssh (init 0), it shuts down VERY slowly. After that, the PC doesn't start. I have to cut the power for 5 minutes (which we HW pros call a "hardest reset", then it works. No hardware problems here I guarantee, being in the HW business for 30+ years. This never happened.

I played through the game with no big errors which is a congrats to openmw, knowing the game since it came out. I suppose it started when I started bloodmoon, but I don't see a feasible relation. Going through the main and side quests took xx hours, and it didn't happen. _Strange_.

(At a point I feared I have a serious videocard/ps/mb problem. I have a ps tester which said OK, stress-tested the mb/RAM, and switched to a similar videocard. No errors before or since outside of openmw. And I game a lot.)

openmw 0.48.0 windowed
linux mint 20.3 kernel 5.4.0-128-generic x86/64
amd/ati 6770 (/5770)

Tried to find a related topic here and on gitlab/issues, no luck.

[AnyOldName3]

There's a decent chance this is down to a bug in the graphics driver, and it might have shown up when using OpenMW just because OpenMW happens to do a sequence of calls that none of the other applications you normally run ever do. If you report the problem to Mesa, they might have some idea of troubleshooting steps to isolate the problem and hopefully fix it, although it might take a while as not many people still use the AMD Radeon HD 6000 series any more - it's so old that AMD felt it safe to release the AMD Radeon RX 6000 series without anyone getting confused.

[csiga77]

Thanks,

I will plug in a new card (recently unused for the bcoin falls reason), and I'll report back. I don't think ati/amd will be interested in this rare shite.

I use this particular HW for one reason: old games. OpenMW still runs on this with 300+fps, I capped it at my refrate: 150, why would I stress it otherwise...

Really thanks for this new releases, I will test it with the amazing works of SureAI too - If I can finish Bloodmoon.

(One question, why can't I sub on feature requests? Double map window and double keybinds would it be.)

[AnyOldName3]

I don't think the people maintaining the part of Mesa that deals with cards this old are actually AMD employees, so it might not matter that AMD doesn't care as someone else might. In fact, I think AMD's contributions to the open-source drivers for their cards didn't start for years after these cards came out, so the relevant parts of the driver are entirely third-party.

By the way, ATi stopped existing before the cards we're discussing came into existence, so it's no more sensible to call them ATi/AMD than it is to call them Tseng Labs/ArtX/Terayon/ATi/AMD, which is obviously a silly thing to do.

As for subscribing to notifications for feature requests on GitLab, it should be as simple as clicking the Notifications switch on the right sidebar on the issue's page while you're signed into GitLab.

[csiga77]

"Tseng Labs/ArtX/Terayon" wow some names I didn't hear for a time...

Still I will report back how it goes with the RX5800 on the same rig, just to be sure.

(I backed up a server in which the PCIe ports died. With what? A Tseng ET6000PCI. It worked through until we switched the motherboard. What old stuff can be handy on The Ashlands )

[csiga77]

The new card fixed the thing. I think the other older cards worth considered somewhat incompatible with the new OpenMW versions.

I played an older version with the same cards and no problems, about 18 months ago.