http://bugs.winehq.org/show_bug.cgi?id=33964
Bug #: 33964 Summary: GTAVC hangs on loading saved game on intel video Product: Wine Version: 1.6-rc4 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: b7.10110111@gmail.com Classification: Unclassified
I start GTAVC, then try to load a saved game. It shows progress bar going to its end, then displays the message "Loading game, please wait", produces a sound telling that game is loaded, and hangs with this message on screen. Alt+F4 doesn't close it (working in wine desktop). This only happens on intel video, namely i915 and N10 graphics controllers. This is a regression: 10d46bc042696bb9e5c45ac88af332c957bbff03 is the first bad commit commit 10d46bc042696bb9e5c45ac88af332c957bbff03 Author: Henri Verbeet hverbeet@codeweavers.com Date: Sun Nov 11 18:52:38 2012 +0100
wined3d: Add support for GLSL based fixed function fragment shaders.
:040000 040000 091c0723beec13a7d04542a59eb7586b01627f4b 8d78954f05c51057c6cc1c259be552fba1ee0f08 M dlls
http://bugs.winehq.org/show_bug.cgi?id=33964
Ruslan Kabatsayev b7.10110111@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression CC| |hverbeet@gmail.com Version|1.6-rc4 |1.5.18 Regression SHA1| |10d46bc042696bb9e5c45ac88af | |332c957bbff03
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #1 from Ruslan Kabatsayev b7.10110111@gmail.com 2013-07-05 14:01:21 CDT --- In fact, after looking at commit description, it may just be very slow to render since these intel chips don't quite support shaders (only ARB asm with severely limited resources, and vertex shaders are executed on CPU).
http://bugs.winehq.org/show_bug.cgi?id=33964
Brandon Corujo haku08879@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |haku08879@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #2 from Henri Verbeet hverbeet@gmail.com 2013-08-22 07:55:14 CDT --- It does sound a bit like a software fallback somewhere, but I'm not sure the limited hardware capabilities are enough to explain that by itself. Before the commit in question it would use ARB_fragment_program based shaders, which should have had much of the same issues in that case. Does running with INTEL_DEBUG=perf give any interesting output? Also, could you attach the output of "glxinfo -l", just for reference?
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #3 from Ruslan Kabatsayev b7.10110111@gmail.com 2013-08-22 08:42:54 CDT --- Created attachment 45690 --> http://bugs.winehq.org/attachment.cgi?id=45690 INTEL_DEBUG=perf log
Here's the log. The interesting part, appearing at the moment of hang, is this:
i915_program_error: Bad source->Index: 3 ENTER FALLBACK 10000: Program
Also, there's a "Mapping a busy BO, causing a stall on the GPU." line, but it also appears with previous wine version, where the game works normally.
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #4 from Ruslan Kabatsayev b7.10110111@gmail.com 2013-08-22 08:43:54 CDT --- Created attachment 45691 --> http://bugs.winehq.org/attachment.cgi?id=45691 glxinfo -l
Here's "glxinfo -l" output.
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #5 from Henri Verbeet hverbeet@gmail.com 2013-08-22 09:33:27 CDT --- Could you attach a log with WINEDEBUG=+d3d_shader and INTEL_DEBUG=perf? It does look like a driver issue though, so it may be a good idea to test if it's any better with a newer Mesa version, even though the i915 driver probably doesn't see a lot of development these days.
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #6 from Ruslan Kabatsayev b7.10110111@gmail.com 2013-08-22 09:47:01 CDT --- Created attachment 45693 --> http://bugs.winehq.org/attachment.cgi?id=45693 +d3d_shader and perf log
Here it is
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #7 from Henri Verbeet hverbeet@gmail.com --- Created attachment 46841 --> http://bugs.winehq.org/attachment.cgi?id=46841&action=edit patch
Does the attached patch make it any better? Looking at the driver source a bit, it looks like reads from gl_FragData[0] may cause a software fallback.
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #8 from Ruslan Kabatsayev b7.10110111@gmail.com --- (In reply to comment #7)
Does the attached patch make it any better?
No, still the same symptoms.
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #9 from Henri Verbeet hverbeet@gmail.com --- In that case, could you please attach a new log (WINEDEBUG=+d3d_shader INTEL_DEBUG=perf)?
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #10 from Ruslan Kabatsayev b7.10110111@gmail.com --- Created attachment 46849 --> http://bugs.winehq.org/attachment.cgi?id=46849 New log
Here it is
http://bugs.winehq.org/show_bug.cgi?id=33964
--- Comment #11 from Henri Verbeet hverbeet@gmail.com --- (In reply to comment #10)
trace:d3d_shader:shader_vaddline GL HW (12, 254) : gl_FragData[0] = gl_SecondaryColor * specular_enable + ret; trace:d3d_shader:shader_vaddline GL HW (13, 314) : float Fog = (gl_Fog.end - gl_FogFragCoord) * gl_Fog.scale; trace:d3d_shader:shader_vaddline GL HW (14, 373) : gl_FragData[0].xyz = mix(gl_Fog.color.xyz, gl_FragData[0].xyz, clamp(Fog, 0.0, 1.0)); trace:d3d_shader:shader_vaddline GL HW (15, 459) : }
It looks like something went wrong with applying the patch, that log shows the original code being used.
http://bugs.winehq.org/show_bug.cgi?id=33964
Ruslan Kabatsayev b7.10110111@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #46849|0 |1 is obsolete| |
--- Comment #12 from Ruslan Kabatsayev b7.10110111@gmail.com --- Created attachment 46850 --> http://bugs.winehq.org/attachment.cgi?id=46850 Log with patch
Sorry, you're right. Now it indeed works much better, although I see some artifacts, and performance is not that good as it seems it was before regression, but this might be unrelated.
https://bugs.winehq.org/show_bug.cgi?id=33964
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution|--- |UPSTREAM
--- Comment #13 from Henri Verbeet hverbeet@gmail.com --- Should be fixed by Mesa commit b094b3b9f4c7b40056c31e3480ab7dc530da56e7.
https://bugs.winehq.org/show_bug.cgi?id=33964
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #14 from Henri Verbeet hverbeet@gmail.com --- Closing.