http://bugs.winehq.org/show_bug.cgi?id=22406
Summary: Star Wars DroidWorks crashes on exit Product: Wine Version: 1.1.43 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: minor Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: matt@mattmccutchen.net
Created an attachment (id=27425) --> (http://bugs.winehq.org/attachment.cgi?id=27425) Log showing the crash
I'm running Star Wars DroidWorks on Fedora 12 using the nvidia proprietary video driver and OpenGL library. When Star Wars DroidWorks exits, it crashes inside libglcore. This is a minor annoyance and does not impact playing the game.
I'm not sure if the crash is Wine's fault, but it did not occur when I played the game on Windows 95 many years ago, so I'm entering a bug. Unfortunately, the game doesn't run on Windows Vista at all, even in compatibility mode, so it isn't possible to do a comparison on a modern Windows system.
http://bugs.winehq.org/show_bug.cgi?id=22406
GyB gyebro69@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |gyebro69@gmail.com
--- Comment #1 from GyB gyebro69@gmail.com 2010-04-18 03:52:59 --- Confirming the bug in the demo, too.
A possible workaround: setting "OffscreenRenderingMode" to "backbuffer" in the registry. The crash only occurs for me with the default "fbo" setting.
I don't know what's the cause of the crash, but I know at least one another game that crashes on exit and changing ORM to "backbuffer" solved the problem.
I'm on Fedora 12, too, but I couldn't produce a usable backtrace when Droidworks.exe is crashing: all wine processes contain debug information except the game itself. Is it a special case (the crash on exit in Wine), because generally I can get good backtraces when a program crashes in Wine, but not in this case?
Fedora 12 x86 Wine-1.1.43 Nvidia 7xxx card with driver 195.36.15
http://bugs.winehq.org/show_bug.cgi?id=22406
Andrew Nguyen arethusa26@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #27425|application/octet-stream |text/plain mime type| |
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #2 from Matt McCutchen matt@mattmccutchen.net 2010-04-18 04:18:11 --- (In reply to comment #1)
A possible workaround: setting "OffscreenRenderingMode" to "backbuffer" in the registry. The crash only occurs for me with the default "fbo" setting.
Confirmed, the workaround avoids the crash for me too. I hoped it might fix big 22407 too, but apparently not.
I'm on Fedora 12, too, but I couldn't produce a usable backtrace when Droidworks.exe is crashing: all wine processes contain debug information except the game itself. Is it a special case (the crash on exit in Wine), because generally I can get good backtraces when a program crashes in Wine, but not in this case?
The nvidia proprietary libglcore.so doesn't contain debug info either. I would still expect to see some Wine Direct3D frames between DroidWorks.exe and libglcore.so, but maybe libglcore.so was built with some fancy optimization that makes it impossible for the debugger to traverse the stack. :(
http://bugs.winehq.org/show_bug.cgi?id=22406
Roderick Colenbrander thunderbird2k@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |thunderbird2k@gmail.com
--- Comment #3 from Roderick Colenbrander thunderbird2k@gmail.com 2010-04-18 09:43:16 --- Did the game work before? If so could you perform a regression test, see http://wiki.winehq.org/RegressionTesting
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #4 from Matt McCutchen matt@mattmccutchen.net 2010-04-18 09:48:44 --- (In reply to comment #3)
Did the game work before?
No, I am not aware of any Wine version in which the game runs without this crash. I only tested with Wine 1.1.42 and 1.1.43, and the game crashes on exit in both of those versions. So I have no "good" endpoint for a bisection.
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #5 from Roderick Colenbrander thunderbird2k@gmail.com 2010-04-18 10:21:54 --- Try running using WINEDEBUG=+d3d,+d3d_surface it might show the calls which are made before the crash. That way we can identify where it goes wrong in Wine (it might be that we are passing incorrect data around or so).
http://bugs.winehq.org/show_bug.cgi?id=22406
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download Status|UNCONFIRMED |NEW Component|directx-d3d |directx-ddraw Ever Confirmed|0 |1
--- Comment #6 from Vitaliy Margolen vitaliy@kievinfo.com 2010-04-18 10:46:53 --- Confirming per comment 1. Have a demo link?
http://bugs.winehq.org/show_bug.cgi?id=22406
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- URL| |http://remote.newswall.org. | |uk/swgamer/files/demos/droi | |dworks.exe
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #7 from Matt McCutchen matt@mattmccutchen.net 2010-04-18 12:14:59 --- I captured a log with WINEDEBUG=+d3d,+d3d_surface but the bug tracker did not let me attach it because it is too big, so instead I posted it at:
http://mattmccutchen.net/private/droidworks-debug-d3d.log
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #8 from Roderick Colenbrander thunderbird2k@gmail.com 2010-04-18 13:18:25 --- A hint for next time. Gzip or even better bzip2 can compress these files heavily. I wouldn't be surprised if bzip2 could compress it to less than 100kB or perhaps even less. I remember it compressing some 1GB+ logs to such sizes.
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #9 from Roderick Colenbrander thunderbird2k@gmail.com 2010-04-18 13:33:01 --- I have tried the demo on my laptop which uses a Radeon HD3450 + fglrx and here I don't see any crash. The demo runs fine. Can you check if it also occurs on the demo?
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #10 from Matt McCutchen matt@mattmccutchen.net 2010-04-18 13:49:31 --- Created an attachment (id=27442) --> (http://bugs.winehq.org/attachment.cgi?id=27442) Log with WINEDEBUG=+d3d,+d3d_surface (xz compressed)
(In reply to comment #8)
A hint for next time. Gzip or even better bzip2 can compress these files heavily.
Duh. But I can one-up that: xz.
3479438 droidworks-debug-d3d.log 34898 droidworks-debug-d3d.log.bz2 98414 droidworks-debug-d3d.log.gz 16884 droidworks-debug-d3d.log.xz
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #11 from Matt McCutchen matt@mattmccutchen.net 2010-04-18 14:00:42 --- Created an attachment (id=27443) --> (http://bugs.winehq.org/attachment.cgi?id=27443) Log of the demo with WINEDEBUG=+d3d,+d3d_surface
(In reply to comment #9)
Can you check if it also occurs on the demo?
Yes, the demo crashes for me too.
http://bugs.winehq.org/show_bug.cgi?id=22406
--- Comment #12 from GyB gyebro69@gmail.com 2010-05-06 13:47:10 --- Just tried with wine-1.1.43-602-g2a9b4e0, it doesn't crash on exit any more. The commit that fixed the issue should be:
commit 7b47dabd0951a3eeaca38d1997419a836b158b6d Author: Henri Verbeet hverbeet@codeweavers.com Date: Tue Apr 20 22:38:41 2010 +0200
wined3d: Avoid some unnecessary invalidations of the the onscreen depth buffer.
This can happen for example when the application switches to a different depth/stencil for offscreen rendering and back again.
That was the result of a regression test I made to discover which commit might have fixed the issue.
Additionally, a very similar 'crash on exit when fbo is used' is also fixed in the game Vampire: The Masquerade - Redemption now (there was no bugreport about that). Thanks, anyway.
http://bugs.winehq.org/show_bug.cgi?id=22406
Matt McCutchen matt@mattmccutchen.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #13 from Matt McCutchen matt@mattmccutchen.net 2010-05-06 14:27:08 --- Confirmed, the crash no longer occurs.
http://bugs.winehq.org/show_bug.cgi?id=22406
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #14 from Alexandre Julliard julliard@winehq.org 2010-05-07 13:30:22 --- Closing bugs fixed in 1.1.44.
http://bugs.winehq.org/show_bug.cgi?id=22406
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|directx-ddraw |directx-d3d
-
wine-bugs＠winehq.org