| Summary: | [regression] mostly blank graphics on Faeria | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Timo Aaltonen <tjaalton> |
| Component: | Drivers/Gallium/radeonsi | Assignee: | Default DRI bug account <dri-devel> |
| Status: | RESOLVED WORKSFORME | QA Contact: | Default DRI bug account <dri-devel> |
| Severity: | normal | ||
| Priority: | medium | ||
| Version: | 17.0 | ||
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
Timo Aaltonen
2017-04-26 21:06:42 UTC
still there in 17.1-rc2 happens on debian stretch too with mesa 13.0.6 Ran more tests and tried to bisect mesa, but the end result is something weird that I can't figure out anymore where the bug is. Using whatever combination of kernel, llvm, mesa, or the radeon X driver on Ubuntu 16.04 seems to work, but fails on 17.04. Took an apitrace of running Faeria: https://aaltoset.kapsi.fi/tmp/Faeria.x86_64.trace.xz Backporting the full X stack (including kernel 4.10, xorg 1.19.3) from 17.04 to 16.04 didn't break it, so looks like the breakage is due to some compiler issue on 17.04? That sounds like a treat to track down.. FWIW, the apitrace plays back fine for me on Kaveri, with radeonsi from Mesa Git master built using -fno-math-errno -fno-trapping-math -O3 -fno-omit-frame-pointer -ftree-vectorize -march=amdfam10 -mtune=native for optimization with gcc version 6.3.0 20170415 (Debian 6.3.0-14) Same here, the trace replays fine with Mesa/LLVM from today on RX480. Is this still a problem for you will more recent Mesa/LLVM versions? No reply so assuming regression was fixed and closing. |
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.