Summary: | VERDE lockup with Unigine Valley/Heaven if ARB_sample_shading is enabled | ||
---|---|---|---|
Product: | Mesa | Reporter: | Alexander Tsoy <alexander> |
Component: | Drivers/Gallium/radeonsi | Assignee: | Default DRI bug account <dri-devel> |
Status: | RESOLVED FIXED | QA Contact: | |
Severity: | normal | ||
Priority: | medium | ||
Version: | unspecified | ||
Hardware: | Other | ||
OS: | All | ||
Whiteboard: | |||
i915 platform: | i915 features: | ||
Attachments: | dmesg output |
Seems to work fine on my Cape Verde. What brand and model do you have specifically? Can you try a newer kernel? (In reply to Michel Dänzer from comment #1) > Seems to work fine on my Cape Verde. What brand and model do you have > specifically? Gigabyte GV-R775OC-2GI. And if that matters, it's a revision with Micron memory, flashed with latest available BIOS version F3. > > Can you try a newer kernel? I'll try 3.16 and 3.17 kernels next week. I think you forgot to install drirc from Mesa. What happens if you set this env var? allow_glsl_extension_directive_midshader=true (In reply to Marek Olšák from comment #3) > I think you forgot to install drirc from Mesa. What happens if you set this > env var? No, I have correct drirc. > > allow_glsl_extension_directive_midshader=true Setting this env variable doesn't help. (In reply to Michel Dänzer from comment #1) > Can you try a newer kernel? I just tried 3.17.0: - GPU still hangs if ARB_sample_shading is not disabled. - It seems lockup detection timeout was increased significantly: kernel: radeon 0000:01:00.0: ring 0 stalled for more than 91985msec kernel: radeon 0000:01:00.0: GPU lockup (waiting for 0x000000000000e90a last fence id 0x000000000000e8fb on ring 0) - Unrelated to this bug, but.. animation is VERY choppy. I constantly see several-seconds-long freezes, and sometimes >10 sec long freezes. (In reply to Alexander Tsoy from comment #4) > - Unrelated to this bug, but.. animation is VERY choppy. I constantly see > several-seconds-long freezes, and sometimes >10 sec long freezes. Should be better with current Mesa Git master, see bug 84662. Not sure what fixed this, but there is no GPU hang anymore. I also disabled the drirc workaround, so that Heaven would fail to compile its shaders. Tested on my VERDE. (In reply to Alexander Tsoy from comment #2) > (In reply to Michel Dänzer from comment #1) > > Seems to work fine on my Cape Verde. What brand and model do you have > > specifically? > > Gigabyte GV-R775OC-2GI. JFYI: This card was unstable when operating at 850Mhz sclk (same problem on Windows). After flashing VBIOS from GV-R775D3-2GI (same card physically, but with 800MHz sclk) it works fine now. So this looks like a faulty vendor overclock. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.
Created attachment 107614 [details] dmesg output Unigine Valley 1.0 and Heaven 4.0 cause VERDE GPU lockup when ARB_sample_shading is enabled. With kernel 3.14.19 I got immediate lockup. 3.14.20 allows Valley and Heaven to run for up to 1 min. GPU fails to resume after lockup, screen becomes corrupted and I need to reboot the system. If I disable ARB_sample_shading (via env variable MESA_EXTENSION_OVERRIDE=-GL_ARB_sample_shading) then VERDE GPU do not hang at all (currently valley is running on my system for more than two hours without lockups). I also have BONAIRE GPU and it does not hang. Software versions: kernel-3.14.20 mesa-10.3.0 llvm-3.5 xf86-video-ati 7.4.0 libdrm-2.4.56