http://bugs.winehq.org/show_bug.cgi?id=21540
Summary: Half-Life 2 crashes after start Product: Wine Version: 1.1.37 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: monraaf@gmail.com CC: hverbeet@gmail.com
Games such as Half-Life2, Portal and F.E.A.R. all crash after starting when using the default fbo as OffscreenRenderingMode:
Unhandled exception: page fault on read access to 0x00000014 in 32-bit code (0x2096a8fc) [...] Backtrace: =>0 0x2096a8fc in fglrx_dri.so (+0x8978fc) (0x00000000)
This is on a Radeon HD 5750 tested with Catalyst 9.12 and 10.1. With wine 1.1.31 there are no crashes. I've bisected it down to this commit:
--------------------------------------------------------------- 94d33d3e861d8e01b4edcb940e8ab1aa98359520 is first bad commit commit 94d33d3e861d8e01b4edcb940e8ab1aa98359520 Author: Henri Verbeet hverbeet@codeweavers.com Date: Tue Oct 20 11:05:06 2009 +0200
wined3d: Destroy FBO entries from the context that created them.
EXT_framebuffer_object doesn't specify if FBOs are shareable between GL contexts, but ARB_framebuffer_object explicitly prohibits it.
:040000 040000 74fa45b1fdd2fed692657fe264f0a3fa716a2a6b d784dc56faeadb10a9dec8507b2f25b9dff88294 M dlls
http://bugs.winehq.org/show_bug.cgi?id=21540
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
http://bugs.winehq.org/show_bug.cgi?id=21540
--- Comment #1 from Henri Verbeet hverbeet@gmail.com 2010-01-31 09:20:22 --- What does a +d3d log look like?
http://bugs.winehq.org/show_bug.cgi?id=21540
--- Comment #2 from Rafael Monica monraaf@gmail.com 2010-01-31 11:02:19 --- Created an attachment (id=25984) --> (http://bugs.winehq.org/attachment.cgi?id=25984) Wine log +d3d
More specifically, the crash occurs on the line:
gl_info->fbo_ops.glFramebufferTexture2D(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0 + i, GL_TEXTURE_2D, 0, 0);
in context_clean_fbo_attachments. If I comment out the call to context_clean_fbo_attachments in context_destroy_fbo there's no crash.
http://bugs.winehq.org/show_bug.cgi?id=21540
--- Comment #3 from Henri Verbeet hverbeet@gmail.com 2010-02-05 12:39:28 --- I don't see wined3d doing anything suspicious in that log, as far as the spec is concerned. I could imagine that the texture being deleted before it is detached from the FBO triggers a bug in the driver though. Is AMD aware of this bug?
http://bugs.winehq.org/show_bug.cgi?id=21540
--- Comment #4 from Rafael Monica monraaf@gmail.com 2010-02-05 12:56:33 --- I wasn't sure if this was a Wine bug or the changes exposed some bug in fglrx. And yes there's a bug report at the unofficial ATI bug tracker:
http://ati.cchtml.com/show_bug.cgi?id=1724
Unfortunately ATI doesn't have a good track record of fixing their bugs.
http://bugs.winehq.org/show_bug.cgi?id=21540
Grazvydas Ignotas notasas@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |notasas@gmail.com
--- Comment #5 from Grazvydas Ignotas notasas@gmail.com 2010-02-05 15:10:18 --- I've filled similar bug a while ago too: http://bugs.winehq.org/show_bug.cgi?id=20749
I've hacked fglrx_dri.so with a NULL pointer check (found some spare bytes), works for me: http://notaz.gp2x.de/misc/fglrx_dri_hack_10.1.zip
http://bugs.winehq.org/show_bug.cgi?id=21540
--- Comment #6 from Henri Verbeet hverbeet@gmail.com 2010-03-08 15:04:55 --- For what it's worth, the AMD bug report got a reply saying they were able to reproduce and identify the issue. I assume that means the bug will be fixed in some future fglrx release.
http://bugs.winehq.org/show_bug.cgi?id=21540
--- Comment #7 from Jeff Zaroyko jeffz@jeffz.name 2010-03-08 18:21:32 --- *** Bug 20749 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=21540
--- Comment #8 from Henri Verbeet hverbeet@gmail.com 2010-04-06 03:26:27 --- According to http://ati.cchtml.com/show_bug.cgi?id=1724 this is fixed in the GL4 beta fglrx driver. I think this bug can be resolved INVALID.
http://bugs.winehq.org/show_bug.cgi?id=21540
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC|hverbeet@gmail.com |
http://bugs.winehq.org/show_bug.cgi?id=21540
xpmlbs@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |xpmlbs@gmail.com
--- Comment #9 from xpmlbs@gmail.com 2010-05-21 10:40:28 --- How do I set offscreen to backbuffer?
http://bugs.winehq.org/show_bug.cgi?id=21540
Rafael Monica monraaf@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |INVALID
--- Comment #10 from Rafael Monica monraaf@gmail.com 2010-05-26 19:30:27 --- This was actually a bug in the ATI catalyst drivers, and has been fixed in their 10.5 release.
http://bugs.winehq.org/show_bug.cgi?id=21540
Jeff Zaroyko jeffz@jeffz.name changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED Component|directx-d3d |-unknown
--- Comment #11 from Jeff Zaroyko jeffz@jeffz.name 2010-05-26 19:48:14 --- Closing invalid.
http://bugs.winehq.org/show_bug.cgi?id=21540
Jeff Zaroyko jeffz@jeffz.name changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |wine@eternaldusk.com
--- Comment #12 from Jeff Zaroyko jeffz@jeffz.name 2010-05-26 19:50:19 --- *** Bug 22534 has been marked as a duplicate of this bug. ***
-
wine-bugs＠winehq.org