http://bugs.winehq.org/show_bug.cgi?id=22252
Summary: Rhiannon installer hangs: possibly caused by a regression Product: Wine Version: 1.1.42 Platform: x86 URL: http://www.gamershell.com/download_33704.shtml OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: winmm&mci AssignedTo: wine-bugs@winehq.org ReportedBy: gyebro69@gmail.com CC: m.b.lankhorst@gmail.com
Created an attachment (id=27174) --> (http://bugs.winehq.org/attachment.cgi?id=27174) terminal output
Using Wine-1.1.42, Rhiannon: The Curse of the Four Branches installer hangs after starting up. Both the demo and the retail version is affected.
Howto reproduce the bug: Start the installer: the company logo is shown for a couple of seconds, then Wine hangs. Eventually an exception is generated by Wine and winedbg starts.
Bisecting between 1.1.41 and 1.1.42 resulted the following commit:
4f4dd855c47cca17d4b750ca118826614afdc20c is the first bad commit commit 4f4dd855c47cca17d4b750ca118826614afdc20c Author: Maarten Lankhorst m.b.lankhorst@gmail.com Date: Wed Mar 31 20:19:35 2010 +0200
winmm: Defer loading drivers until they are requested.
:040000 040000 f3db7429aa6c8f870e6abd1f6328c5412731dbe1 3bd406034df5d9e5ddd9b2cdf8b8a51af36926ea M dlls
However, I couldn't cleanly revert the patch on top of 1.1.42:
patching file dlls/winmm/lolvldrv.c Hunk #1 succeeded at 58 with fuzz 2 (offset 11 lines). Hunk #2 succeeded at 63 with fuzz 2 (offset -3 lines). Hunk #3 succeeded at 70 (offset -3 lines). Hunk #4 FAILED at 109. Hunk #5 succeeded at 243 (offset -21 lines). Hunk #6 succeeded at 500 (offset -27 lines). Hunk #7 FAILED at 563. Hunk #8 succeeded at 560 (offset -33 lines). 2 out of 8 hunks FAILED -- saving rejects to file dlls/winmm/lolvldrv.c.rej patching file dlls/winmm/winemm.h Hunk #1 succeeded at 133 (offset -4 lines). patching file dlls/winmm/winmm.c
Despite the rejected lines, the source re-compiled cleanly, and the installer worked again.
Fedora 12 x86 Wine-1.1.42 compiled from source, using CFLAGS="-gdwarf-2"
Link to the demo added to url. Author of the patch added to CC.
http://bugs.winehq.org/show_bug.cgi?id=22252
GyB gyebro69@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, regression
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #1 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-03 03:55:17 --- Can you use a version of wine with debug symbols? The output is useless without it.
Also can I get the output of you running it with WINEDEBUG=+module,+winmm so I can have a better idea of what is going on?
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #2 from GyB gyebro69@gmail.com 2010-04-03 04:16:50 --- Created an attachment (id=27175) --> (http://bugs.winehq.org/attachment.cgi?id=27175) output using +module,+winmm debug channels
Attached a log showing the terminal output when starting with WINEDEBUG=+module,+winmm.
As for the debugging symbols: I always compile Wine from source and use "gdwarf-2" CFLAGS, as mentioned at the bottom of this page (I'm on Fedora): http://wiki.winehq.org/Backtraces
In the meantime I'll recompile 1.1.42 without "gdwarf-2": without that cflag Wine probably produces a more useful backtrace.
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #3 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-03 04:51:15 --- Hello,
The wiki seems to be wrong there then, -gdwarf-2 is correct, but it just says if debugging is specified, use the dwarf-2 format, it doesn't actually enable it, so you would need -g -gdwarf-2
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #4 from GyB gyebro69@gmail.com 2010-04-03 07:49:02 --- Thanks for your correction on the wiki page.
I re-compiled 1.1.42 with CFLAGS="-g -gdwarf-2"...but still no joy. The output shows the same, except there are no fixmes this time. Tested a couple of apps/games that known to generate a crash dialog in Wine: all of them produced a proper backtrace (including line numbers, function names)... except this game installer. Sorry for my incompetence, i'm just a casual user.
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #5 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-03 08:10:54 --- Maybe you have a version of wine installed through your package manager also and it finds that version of wine first?
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #6 from GyB gyebro69@gmail.com 2010-04-03 13:05:38 --- I have no any versions installed by the package manager and have no further ideas what to do. In the following days someone else will probably identify the same issue using an other app/game and hopefully can get a proper backtrace. Thanks for your time.
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #7 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-05 14:40:19 --- Hello,
You could try running it in winedbg, instead of 'wine XXX.exe' you would do 'wine winedbg XXX.exe', then start the app with 'run'
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #8 from GyB gyebro69@gmail.com 2010-04-06 10:22:40 --- Created an attachment (id=27240) --> (http://bugs.winehq.org/attachment.cgi?id=27240) winedbg backtrace log
Hello,
Running 'wine winedbg installer.exe', then 'run', resulted almost nothing in the terminal, the prompt appeared after 'run' and I didn't know how to continue.
Next try, I followed this guide: http://wiki.winehq.org/Backtraces
In the attachment you can see the result, here is what i did:
1. started the installer with 'wine winedbg Rhiannon_Demo_C5_setup.exe' 2. typed 'set $BreakOnFirstChance=0' 3. here i typed a 'bt all': as you can see in the log, all other running wine processes produced a usable backtrace, except the installer itself. 4. typed 'cont' >>here comes the trouble, the installer shows the logo then 'err:ntdll:RtlpWaitForCriticalSection blah-blah' is printed on the console then 'process has terminated'.
Is it more usable for you?
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #9 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-06 11:29:29 --- Yeah, but oddly enough the threads that are most interesting are missing.
Can you try debugging again, and wait till it says:
err:ntdll:RtlpWaitForCriticalSection section 0x7bca9ae4 "loader.c: loader_section" wait timed out in thread 001e, blocked by 001a, retrying (60 sec)
The 2 important threads here are 0x001e and 0x001a, so on the run you would do 'bt 0x001e' followed by 'bt 0x001a'
Of course they can vary per run, so you would have to fill in the correct values based on the err message.
Cheers, Maarten
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #10 from GyB gyebro69@gmail.com 2010-04-06 11:56:08 --- Created an attachment (id=27243) --> (http://bugs.winehq.org/attachment.cgi?id=27243) backtrace, including 2 threads
I hope I did it well: I had to break (CTRL+C) winedbg, during the 60 sec waiting time and backtrace the 2 threads? If so, here's the result.
http://bugs.winehq.org/show_bug.cgi?id=22252
Maarten Lankhorst m.b.lankhorst@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #11 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-06 17:09:41 --- Ok, so the problem is that winmm is loaded, a sound is played and winmm is unloaded after that. The timer thread is still running and is killed in DllMain, this will cause a deadlock. I wrote a patch for it, should hopefully be merged soon.
http://bugs.winehq.org/show_bug.cgi?id=22252
Maarten Lankhorst m.b.lankhorst@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|wine-bugs@winehq.org |m.b.lankhorst@gmail.com
--- Comment #12 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-09 15:08:22 --- Created an attachment (id=27305) --> (http://bugs.winehq.org/attachment.cgi?id=27305) winmm: Rewrite time.c using windows timer queues
This should fix things, won't be accepted into wine yet, aj wants me to fix other things first. :)
http://bugs.winehq.org/show_bug.cgi?id=22252
Juan Lang juan_lang@yahoo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |patch
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #13 from GyB gyebro69@gmail.com 2010-04-09 22:55:56 --- Your patch works, applied to wine-1.1.42-182-g8f77dd8. Now the game can be installed again. Thank you and keep up the good work, Maarten.
http://bugs.winehq.org/show_bug.cgi?id=22252
Maarten Lankhorst m.b.lankhorst@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |willemmaster@hotmail.com
--- Comment #14 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-04-17 19:45:30 --- *** Bug 22368 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=22252
Maarten Lankhorst m.b.lankhorst@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |cerebro.alexiel@gmail.com
--- Comment #15 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-05-17 12:56:41 --- *** Bug 22619 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=22252
Andrew Nguyen arethusa26@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |christophe@wk3.org
--- Comment #16 from Andrew Nguyen arethusa26@gmail.com 2010-05-19 09:15:47 --- *** Bug 22760 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=22252
Maarten Lankhorst m.b.lankhorst@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED
--- Comment #17 from Maarten Lankhorst m.b.lankhorst@gmail.com 2010-05-19 11:26:50 --- Should be fixed by commit 00eaa92945
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #18 from Alex cerebro.alexiel@gmail.com 2010-05-19 13:24:18 --- I can confirm : it's now resolved. There is just a err:mmtime:TIME_MMTimeStop Timer still active?! but without any consequence.
Can we help you to determine if the timer is still active ?
http://bugs.winehq.org/show_bug.cgi?id=22252
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #19 from Alexandre Julliard julliard@winehq.org 2010-05-21 14:39:45 --- Closing bugs fixed in 1.2-rc1.
http://bugs.winehq.org/show_bug.cgi?id=22252
wentao Dstorm.xu@dextrys.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |Dstorm.xu@dextrys.com
--- Comment #20 from wentao Dstorm.xu@dextrys.com 2010-05-24 04:46:35 --- Very useful ,but not fixed in 1.2-rc1.
Thanks!
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #21 from Alex cerebro.alexiel@gmail.com 2010-05-24 05:05:35 --- ... What ? Are we talking about the same bug ? Did you bisect'ed and found the same commit ? You'd better open a new bug report and post all relevant informations (console output...etc)
@GyB: You openned this bug. Can you tell us if it's fixed for you ?
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #22 from GyB gyebro69@gmail.com 2010-05-24 05:28:42 --- (In reply to comment #21)
... What ? Are we talking about the same bug ? Did you bisect'ed and found the same commit ? You'd better open a new bug report and post all relevant informations (console output...etc)
@GyB: You openned this bug. Can you tell us if it's fixed for you ?
Yes, it is fixed in 1.2-rc1.
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #23 from wentao Dstorm.xu@dextrys.com 2010-05-24 20:48:19 --- Sorry about this ,I run a game launcher in(not linux,sorry): MAC OS: Version 10.6.1 (Build 10A2062) X11BundleVersion: 2.3.4 wineversion 1.2 rc1(I checked) Endof game launcher ,the console out put: 2010-05-25 09:00:11 135 err:ole:CoInitializeEx Attempt to change threading model of this apartment from multi-threaded to apartment threaded 2010-05-25 09:00:32 135 fixme:winmm:MMDRV_Exit Closing while ll-driver open 2010-05-25 09:00:32 135 err:mmtime:TIME_MMTimeStop Timer still active?! ... I want to debug it by myself,may be it not a bug,It's my mistake!
http://bugs.winehq.org/show_bug.cgi?id=22252
--- Comment #24 from Andrew Nguyen arethusa26@gmail.com 2010-05-24 21:05:37 --- (In reply to comment #23)
Sorry about this ,I run a game launcher in(not linux,sorry): MAC OS: Version 10.6.1 (Build 10A2062) X11BundleVersion: 2.3.4 wineversion 1.2 rc1(I checked) Endof game launcher ,the console out put: 2010-05-25 09:00:11 135 err:ole:CoInitializeEx Attempt to change threading model of this apartment from multi-threaded to apartment threaded 2010-05-25 09:00:32 135 fixme:winmm:MMDRV_Exit Closing while ll-driver open 2010-05-25 09:00:32 135 err:mmtime:TIME_MMTimeStop Timer still active?! ... I want to debug it by myself,may be it not a bug,It's my mistake!
Okay, so is there any actual hang/crash with the game launcher? If all that's happening is the output you see, then there isn't an actual problem.
http://bugs.winehq.org/show_bug.cgi?id=22252
Wylda wylda@volny.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |00eaa92945 CC| |wylda@volny.cz Regression SHA1| |4f4dd855c47cca17d4b750ca118 | |826614afdc20c
http://bugs.winehq.org/show_bug.cgi?id=22252
Wylda wylda@volny.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1|00eaa92945 |00eaa9294559bb75cf5cbf7f1b6 | |1c3e239d08d62
-
wine-bugs＠winehq.org