https://bugs.winehq.org/show_bug.cgi?id=42509
Bug ID: 42509 Summary: The Witcher 3 hangs the system on startup Product: Wine-staging Version: 2.2 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: shtetldik@gmail.com CC: erich.e.hoover@wine-staging.com, michael@fds-team.de, sebastian@fds-team.de Distribution: ---
Created attachment 57398 --> https://bugs.winehq.org/attachment.cgi?id=57398 Wine log (The Witcher 3 hangs system)
Since update of wine-staging to 2.2, Witcher 3 started hanging hard on startup. The screen goes black, and keyboard becomes completely unresponsive. The system can still be rebooted if accessed for example from another machine over ssh.
OS: Debian testing, x86-64, KDE Plasma 5.8.4. Graphics: AMD RX480, Mesa / Gallium 0.4 on AMD POLARIS10 (DRM 3.8.0/4.9.0-1-amd64, LLVM 3.9.1)
See attached Wine log.
https://bugs.winehq.org/show_bug.cgi?id=42509
--- Comment #1 from Shmerl shtetldik@gmail.com --- My Mesa is Mesa 13.0.4, but same happens with more recent Mesa as well.
https://bugs.winehq.org/show_bug.cgi?id=42509
--- Comment #2 from Sebastian Lackner sebastian@fds-team.de --- Could you please test if the issue is related to CSMT, or does it occur always? Also, if possible please try to check (for example via SSH) if the system is running out of CPU or GPU memory.
https://bugs.winehq.org/show_bug.cgi?id=42509
--- Comment #3 from Shmerl shtetldik@gmail.com --- It occurs both with and without CSMT. I logged in remotely, to check the system. CPU / RAM usage were normal. Any suggestion how to check VRAM utilization?
I tried using something like:
GALLIUM_HUD_DUMP_DIR=$HOME/tmp GALLIUM_HUD=VRAM-usage ...
To dump data to a log, but it hangs before Mesa starts logging anything apparently (the log file is created, but it remains empty). radeontop doesn't like Debian testing kernel, as they note:
For the amdgpu driver, only the mem path is currently supported - for those cards, this means radeontop won't run on the default Ubuntu kernels that block /dev/mem
so it can't show VRAM utilization.
https://bugs.winehq.org/show_bug.cgi?id=42509
Józef Kucia joseph.kucia@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|-unknown |-unknown Product|Wine-staging |Wine CC| |joseph.kucia@gmail.com
--- Comment #4 from Józef Kucia joseph.kucia@gmail.com --- This is likely caused by an infinite loop in a shader. I don't think the bug is specific to wine-staging. A +d3d_shader log might give some pointers. Anyway, this should be fixed soon when some missing shader instructions are implemented.
https://bugs.winehq.org/show_bug.cgi?id=42509
--- Comment #5 from Shmerl shtetldik@gmail.com --- Created attachment 57404 --> https://bugs.winehq.org/attachment.cgi?id=57404 Witcher 3 hanging run with +d3d_shader
Attaching the hanging Wine run log with +d3d_shader (using latest build of Wine from git master: wine-2.2-159-g0dc677c05f).
https://bugs.winehq.org/show_bug.cgi?id=42509
Guillaume Charifi guillaume.charifi@sfr.fr changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |guillaume.charifi@sfr.fr
--- Comment #6 from Guillaume Charifi guillaume.charifi@sfr.fr --- I was having this problem too, and I solved it by commenting the glDispatchCompute() call in wined3d/drawprim.c. I re-enabled it yesterday, and it's not crashing anymore. So I guess it was a problem in the Wine compute shader implementation, that seems fixed right now.
https://bugs.winehq.org/show_bug.cgi?id=42509
Józef Kucia joseph.kucia@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|-unknown |directx-d3d Status|UNCONFIRMED |RESOLVED Resolution|--- |FIXED
--- Comment #7 from Józef Kucia joseph.kucia@gmail.com --- Marking fixed.
https://bugs.winehq.org/show_bug.cgi?id=42509
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #8 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 2.3.