http://bugs.winehq.org/show_bug.cgi?id=15297
Summary: WoW either fails at startup, runs fine, or crashes the XServer, all at random. Product: Wine Version: CVS/GIT Platform: PC OS/Version: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: opengl AssignedTo: wine-bugs@winehq.org ReportedBy: rankincj@yahoo.com
Created an attachment (id=16112) --> (http://bugs.winehq.org/attachment.cgi?id=16112) WINEDEBUG=warn+opengl,+wgl,+x11drv fails with a GLX error
I've noticed this bug for several months now, and can't remember exactly when it started. Basically, I am running World of Warcraft using both Wine and Mesa from their respective git repositories. My PC is a dual P4 Xeon machine, with hyperthreading enabled, so that I have four virtual CPUs. I am using a Radeon 9550 graphics card.
When I run WoW in OpenGL mode, then one of three things happens:
a) WoW completely fails to start, and reports this error: X Error of failed request: GLXBadDrawable Major opcode of failed request: 156 (GLX) Minor opcode of failed request: 5 (X_GLXMakeCurrent) Serial number of failed request: 586 Current serial number in output stream: 586
b) WoW runs fine(-ish), albeit slowly. (But maybe that's just Mesa's R300 support)
c) The entire X server crashes, and the Xorg.0.log file reports a SIGSEGV while calling __glxDeassociateContext.
So assuming that I don't crash the X server first, I can eventually get WoW to start just by trying to run it again and again. It sounds like there might be some hideous sort of race condition in the context-handling code.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #1 from rankincj@yahoo.com 2008-09-15 07:59:13 --- Created an attachment (id=16113) --> (http://bugs.winehq.org/attachment.cgi?id=16113) WINEDEBUG=warn+opengl,+wgl,+x11drv runs successfully
WoW has started successfully this time. Notice how the OpenGL debug statement says that Direct Rendering is *not* enabled. I don't understand how that can be true either, because it's not *that* slow. Unless AIGLX is doing all the work?
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #2 from rankincj@yahoo.com 2008-09-15 08:00:08 --- Created an attachment (id=16114) --> (http://bugs.winehq.org/attachment.cgi?id=16114) WINEDEBUG=warn+opengl,+wgl,+x11drv when the Xserver crashes
This is the debug log from a time that brought the Xserver down in flames.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #3 from rankincj@yahoo.com 2008-09-15 08:01:40 --- Created an attachment (id=16115) --> (http://bugs.winehq.org/attachment.cgi?id=16115) The Xorg.0.log file from a WoW crash.
This wasn't the same WoW crash that I have the WINEDEBUG log for, but I have no reason to suspect that the two would be any different.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #4 from rankincj@yahoo.com 2008-09-15 08:05:10 --- Created an attachment (id=16116) --> (http://bugs.winehq.org/attachment.cgi?id=16116) Test program, to see if I can get a Direct Rendered GLX context
This program is based heavily on Wine's X11DRV_WineGL_InitOpenglInfo() function, and this is its output:
GLX Version: 1.3 Mesa 7.3-devel GLX Renderer: Mesa DRI R300 20060815 AGP 8x x86/MMX/SSE2 TCL GLX Extensions: GL_ARB_depth_texture GL_ARB_fragment_program GL_ARB_imaging GL_ARB_multisample GL_ARB_multitexture GL_ARB_point_parameters GL_ARB_shadow GL_ARB_shadow_ambient ... etc ... GL_SGIS_texture_lod GL_SGIX_depth_texture GL_SGIX_shadow_ambient GL_SUN_multi_draw_arrays GL_S3_s3tc Direct Rendering: Yes
So my Mesa installation seems OK.
http://bugs.winehq.org/show_bug.cgi?id=15297
Dmitry Timoshkov dmitry@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|major |normal
--- Comment #5 from Dmitry Timoshkov dmitry@codeweavers.com 2008-09-15 08:46:08 ---
c) The entire X server crashes, and the Xorg.0.log file reports a SIGSEGV while calling __glxDeassociateContext.
That's a sign of broken video/opengl drivers.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #6 from rankincj@yahoo.com 2008-09-15 08:51:56 --- Created an attachment (id=16117) --> (http://bugs.winehq.org/attachment.cgi?id=16117) WoW configuration file.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #7 from rankincj@yahoo.com 2008-09-15 09:05:51 --- (In reply to comment #5)
That's a sign of broken video/opengl drivers.
If that's true then they're only broken when Wine/WoW tries to use them. I also don't see how you can rule out Wine passing the Xserver either a garbage pointer, or a pointer to garbage memory.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #8 from Dmitry Timoshkov dmitry@codeweavers.com 2008-09-15 09:23:29 --- (In reply to comment #7)
(In reply to comment #5)
That's a sign of broken video/opengl drivers.
If that's true then they're only broken when Wine/WoW tries to use them.
Wine is known to provoke bugs since it uses a very large range of system libraries.
I also don't see how you can rule out Wine passing the Xserver either a garbage pointer, or a pointer to garbage memory.
If the X server crashes that's by definition its fault. No application should be able to crash the X server no matter how hard it tries.
http://bugs.winehq.org/show_bug.cgi?id=15297
Juan Lang juan_lang@yahoo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |INVALID
--- Comment #9 from Juan Lang juan_lang@yahoo.com 2008-09-15 10:05:57 --- Agreed, that's a driver bug. You need to take it up with the driver vendor.
http://bugs.winehq.org/show_bug.cgi?id=15297
James Hawkins truiken@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #10 from James Hawkins truiken@gmail.com 2008-09-15 10:16:26 --- Closing.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #11 from rankincj@yahoo.com 2008-09-15 11:20:43 ---
(In reply to comment #7) If the X server crashes that's by definition its fault. No application should be able to crash the X server no matter how hard it tries.
That does NOT exclude the possibility the incorrect Wine behaviour is triggering a separate bug in the X server.
Did anyone even look at the WINEDEBUG traces, or has this just been dismissed out-of-hand?
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #12 from Vitaliy Margolen vitaliy@kievinfo.com 2008-09-15 12:21:18 --- See bug 12929. All these are well known ATI driver problems.
http://bugs.winehq.org/show_bug.cgi?id=15297
--- Comment #13 from rankincj@yahoo.com 2008-09-15 12:41:00 --- (In reply to comment #12)
See bug 12929. All these are well known ATI driver problems.
Except that I am using the Mesa driver, not fglrx's ATI driver.
http://bugs.winehq.org/show_bug.cgi?id=15297
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|CVS/GIT |unspecified