http://bugs.winehq.org/show_bug.cgi?id=28276
Summary: Crysis2 crashes on startup since commit 006413afe495fbf73542154462885e4fdbf12a2a Product: Wine Version: 1.3.27 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: ntdll AssignedTo: wine-bugs@winehq.org ReportedBy: rockorequin@hotmail.com
Crysis 2 crashes shortly after the startup screen appears ever since wine 1.3.26-213.
Bisecting shows this as the problem:
006413afe495fbf73542154462885e4fdbf12a2a is the first bad commit commit 006413afe495fbf73542154462885e4fdbf12a2a Author: Grazvydas Ignotas notasas@gmail.com Date: Thu Aug 4 23:24:49 2011 +0300
ntdll: Use swapping method to return . and .. as first entries.
:040000 040000 cb445dd153f9b57d3a2e7cc71206441d0530fdf3 5e84ecb0723072147d39a7be912139f962db85df M dlls
Reverting this commit fixes the problem in the latest wine wine-1.3.27-330-g6802bfc.
http://bugs.winehq.org/show_bug.cgi?id=28276
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression CC| |notasas@gmail.com Summary|Crysis2 crashes on startup |Crysis2 crashes on startup |since commit | |006413afe495fbf735421544628 | |85e4fdbf12a2a | Regression SHA1| |006413afe495fbf735421544628 | |85e4fdbf12a2a
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #1 from Grazvydas Ignotas notasas@gmail.com 2011-09-06 05:02:51 CDT --- Can you get a backtrace?
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #2 from rocko rockorequin@hotmail.com 2011-09-07 08:20:26 CDT --- I don't get a crash with backtrace, just an error:
wine client error:2e: write: Bad file descriptor
Which debug channel should I turn on to get more information?
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #3 from Grazvydas Ignotas notasas@gmail.com 2011-09-09 08:16:21 CDT --- Could you run it under winedbg instead? It basically works the same as wine, you just use "winedbg something.exe", then it should stop in a prompt, where you can type "c" to continue. After some time it should either print a backtrace for you, or prompt again where you can enter "bt" to get the backtrace.
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #4 from Grazvydas Ignotas notasas@gmail.com 2011-09-09 08:35:09 CDT --- Created an attachment (id=36292) --> (http://bugs.winehq.org/attachment.cgi?id=36292) log
This patch should produce a log that should help to understand what's going on. Run it without any special debug channels, just don't disable default fixme channel.
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #5 from rocko rockorequin@hotmail.com 2011-09-09 20:55:18 CDT --- Created an attachment (id=36302) --> (http://bugs.winehq.org/attachment.cgi?id=36302) requested log output
Here's the log output with that patch applied.
http://bugs.winehq.org/show_bug.cgi?id=28276
Grazvydas Ignotas notasas@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #36292|0 |1 is obsolete| |
--- Comment #6 from Grazvydas Ignotas notasas@gmail.com 2011-09-10 11:06:56 CDT --- Created an attachment (id=36318) --> (http://bugs.winehq.org/attachment.cgi?id=36318) log2
Thanks, it seems ".." is returned twice on your machine, but it's still unclear why. Here is more verbose patch, could you run it with this patch instead?
Also, was this really the game being run? "wine client error" message you reported before is missing, for example.
It would also help if you tried to get a backtrace, as described in comment 3.
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #7 from rocko rockorequin@hotmail.com 2011-09-11 03:06:59 CDT --- Created an attachment (id=36324) --> (http://bugs.winehq.org/attachment.cgi?id=36324) more detailed crash log
Here's a crash log running with the second patch applied. This time I haven't edited out the wine client error line.
The underlying filesystem is btrfs. Could that make a difference?
I tried using winedbg, but now the application is just exiting instead of saying that Crysis2 has crashed and must be terminated. So now winedbg says that it is no longer attached to a process and can't generate a backtrace.
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #8 from Grazvydas Ignotas notasas@gmail.com 2011-09-11 19:29:13 CDT --- Created an attachment (id=36332) --> (http://bugs.winehq.org/attachment.cgi?id=36332) btrfs workaround
Thanks, it seems to be btrfs bug, I've sent a fix for btrfs people: http://marc.info/?t=131577377600003&r=1&w=2
Meanwhile you can use this wine workaround, but it's not really suitable for upstream..
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #9 from rocko rockorequin@hotmail.com 2011-09-11 19:42:29 CDT --- I can confirm the workaround works fine, thanks!
I guess that this means it won't be fixed in btrfs until at least kernel 3.2?
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #10 from Austin English austinenglish@gmail.com 2011-09-12 12:49:51 CDT --- FWIW, a user in #winehq had a similar problem with Dead Island/brtfs and the patch here helped them as well.
http://bugs.winehq.org/show_bug.cgi?id=28276
Frédéric Delanoy frederic.delanoy@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |wolfgang.scheicher@gmail.co | |m
--- Comment #11 from Frédéric Delanoy frederic.delanoy@gmail.com 2011-09-17 08:56:00 CDT --- *** Bug 28410 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=28276
--- Comment #12 from Saulius K. saulius2@gmail.com 2011-10-04 14:49:03 CDT --- I'd say this isn't a regression, but rather BTRFS bug report triggered by completion of NtQueryDirectoryFile() implementation.
So I'd mark this as RESOLVED to INVALID (since I see no more appropriate keyword like a non-existent NOTAWINEBUG, eg.)
http://bugs.winehq.org/show_bug.cgi?id=28276
Saulius K. saulius2@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |saulius2@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=28276
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords|regression | Status|UNCONFIRMED |RESOLVED Resolution| |INVALID
--- Comment #13 from Austin English austinenglish@gmail.com 2011-10-04 15:25:18 CDT --- (In reply to comment #12)
I'd say this isn't a regression, but rather BTRFS bug report triggered by completion of NtQueryDirectoryFile() implementation.
Yes, not a regression.
So I'd mark this as RESOLVED to INVALID (since I see no more appropriate keyword like a non-existent NOTAWINEBUG, eg.)
Yeah, INVALID.
http://bugs.winehq.org/show_bug.cgi?id=28276
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1|006413afe495fbf735421544628 | |85e4fdbf12a2a |
http://bugs.winehq.org/show_bug.cgi?id=28276
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #14 from Austin English austinenglish@gmail.com 2011-10-04 15:25:54 CDT --- Closing.