[Bug 33125] New: Tomb Raider (2013) distortions during loading
http://bugs.winehq.org/show_bug.cgi?id=33125 Bug #: 33125 Summary: Tomb Raider (2013) distortions during loading Product: Wine Version: 1.5.25 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs(a)winehq.org ReportedBy: andrey.goosev(a)gmail.com CC: andrey.goosev(a)gmail.com Classification: Unclassified Created attachment 43810 --> http://bugs.winehq.org/attachment.cgi?id=43810 log Launched on Nvidia GT 540M with 310.32 drivers. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #1 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-05 04:02:21 CST --- Created attachment 43811 --> http://bugs.winehq.org/attachment.cgi?id=43811 screen screen 1 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #2 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-05 04:02:56 CST --- Created attachment 43812 --> http://bugs.winehq.org/attachment.cgi?id=43812 screen screen 2 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #3 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-05 04:03:17 CST --- Created attachment 43813 --> http://bugs.winehq.org/attachment.cgi?id=43813 screen screen 3 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 jiveaxe <jiveaxe(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jiveaxe(a)gmail.com --- Comment #4 from jiveaxe <jiveaxe(a)gmail.com> 2013-03-05 07:53:24 CST --- I confirm it on Chakra 64bit, kernel 3.7.6, nvidia 313.18, GTX-560Ti, kde 4.10, wine 1.5.24 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #5 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-05 08:05:33 CST --- (In reply to comment #4)
I confirm it on Chakra 64bit, kernel 3.7.6, nvidia 313.18, GTX-560Ti, kde 4.10, wine 1.5.24
Seems it's not a drivers. Can you create ingame profile or it just loading before this step and than nothing? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #6 from jiveaxe <jiveaxe(a)gmail.com> 2013-03-05 08:25:32 CST --- Some advancement: enabling "StrictDrawOrdering" in registry reduces flickering but after loading, perhaps just before menu loading, game freezes. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #7 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-05 08:45:22 CST --- (In reply to comment #6)
Some advancement: enabling "StrictDrawOrdering" in registry reduces flickering but after loading, perhaps just before menu loading, game freezes.
Yes, enabling function gives more better result. Thanks. But still can't create a profile. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Anthony Mattheakakis <antony256(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |antony256(a)gmail.com -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #8 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-05 14:01:32 CST --- Think with distortion/flickering is solved. Second side of the problem is freezing game before creating a profile (or maybe entering). Need a new report? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #9 from jiveaxe <jiveaxe(a)gmail.com> 2013-03-05 15:36:14 CST --- I dunno if the flickering trick is fine; initially, when flickering, I was able to read some text like game tips, bottom right; "StrictDrawOrdering" eliminated flickering and texts. Anyone noticed this? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #10 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-05 16:00:39 CST --- (In reply to comment #9)
I dunno if the flickering trick is fine; initially, when flickering, I was able to read some text like game tips, bottom right; "StrictDrawOrdering" eliminated flickering and texts. Anyone noticed this?
Look this video: youtube.com/watch?v= 0DMf9r6uz74 (just an example) When I set "StrictDrawOrdering" flickering (like on the screens on the top) is gone. Same trick is discribed for Deus Ex: Human Revolution. Look before 00:26 - after "profile" text game is freezes and it's a main clue. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 sacrediou <sacrediou(a)yahoo.fr> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sacrediou(a)yahoo.fr -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Jack Peirce <jpeirce(a)sourcecode.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jpeirce(a)sourcecode.com -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Felix Hellmann <privat(a)cirk2.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |privat(a)cirk2.de -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #11 from jiveaxe <jiveaxe(a)gmail.com> 2013-03-06 03:25:21 CST --- (In reply to comment #10)
Look this video: youtube.com/watch?v= 0DMf9r6uz74 (just an example) When I set "StrictDrawOrdering" flickering (like on the screens on the top) is gone. Same trick is discribed for Deus Ex: Human Revolution. Look before 00:26 - after "profile" text game is freezes and it's a main clue.
So, also windows version shows glitches on loading... funny. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #12 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-06 03:56:01 CST --- (In reply to comment #11)
(In reply to comment #10) So, also windows version shows glitches on loading... funny.
Perhaps it's a developers concept. ) -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #13 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-06 04:25:08 CST --- Solution for starting is downgrade Wine version. Thanks to youtube.com/watch? v=syvn-k2oxY0 and youtube.com/watch? v=-HBO5IVOg40 What's wrong with the latest 1.5.23-25 ? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #14 from jiveaxe <jiveaxe(a)gmail.com> 2013-03-06 04:45:39 CST --- (In reply to comment #13)
Solution for starting is downgrade Wine version.
Great catch. Thanks you. Have installed TR in playonlinux, selected wine 1.5.9-raw3 (instead of system) and TR finally works. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Andras Kovacs <andras(a)sth.sze.hu> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |andras(a)sth.sze.hu --- Comment #15 from Andras Kovacs <andras(a)sth.sze.hu> 2013-03-06 06:29:50 CST --- With the latest wine (1.5.25) on Ubuntu 12.10 32 bit with GeForce 8500GT (driver 310.14) I have no problem at all, it runs slow (but I think this is OK because of old graphics card) -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #16 from Andras Kovacs <andras(a)sth.sze.hu> 2013-03-06 06:37:14 CST --- Created attachment 43832 --> http://bugs.winehq.org/attachment.cgi?id=43832 screen (GeForce 8500GT) -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #17 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-06 11:09:02 CST --- (In reply to comment #15)
With the latest wine (1.5.25) on Ubuntu 12.10 32 bit with GeForce 8500GT (driver 310.14) I have no problem at all, it runs slow (but I think this is OK because of old graphics card)
I'm using 12.10 (32-bit) too with kernel 3.5.0-25 Very strange why some players can runs without issues while other players haven't. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #18 from Felix Hellmann <privat(a)cirk2.de> 2013-03-06 17:09:51 CST --- I can confirm that StrictDrawOrdering fixes the graphical distortions with a wine version 1.5.17. On a actual git build (as of 07.03) the distortions are not fixed with StrictDrawOrdering and the game crashes after the loading screen. Strangely with a build from 06.03 the game didn't crash, but the distortions also where not fixed with StrictDrawOrdering (There I could observe that the same distortions also occur ingame). So this bug is at least partly regression! -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #19 from Felix Hellmann <privat(a)cirk2.de> 2013-03-06 17:14:37 CST --- (In reply to comment #18) Oh I missed the date change at midnight, the crashing build is from 06.03 and the non crashing one from 05.03 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #20 from Bruno Jesus <00cpxxx(a)gmail.com> 2013-03-06 17:18:24 CST --- You could run a regression test to find the commit that caused the problem: http://wiki.winehq.org/RegressionTesting -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 papon.jonathan(a)laposte.net changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |papon.jonathan(a)laposte.net --- Comment #21 from papon.jonathan(a)laposte.net 2013-03-06 18:09:24 CST --- It seem this bug first appeard in 1.5.13. The 1.5.12 doesn't suffer the crash at launch, nor the distortions bug (with or without StrictDrawOrdering registry fix) I am currently trying to do a regression test, but I am running through compilation issues. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #22 from Felix Hellmann <privat(a)cirk2.de> 2013-03-06 19:27:58 CST --- Ok I found the guilty one: 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a is the first bad commit commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a Author: Alexandre Julliard <julliard(a)winehq.org> Date: Sat Dec 15 16:49:47 2012 +0100 include: Force stack alignment also on Linux to work around the ABI breakage. :040000 040000 77ef795291e351dff842cfa08a77c48ee9a4c5c7 e29359eaeee7b50d6aec06ff4a4b55e280c8fe0e M include I already tested it by reverting the commit on the current head with git show 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a | patch -p1 -R Also disregard my comment regarding the distortions persisting after enabling StrictDrawOrder, this was caused by the glBufferSubDataARB patch (http://bugs.winehq.org/show_bug.cgi?id=11674#c263). I've done the Regression test without any patches. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Andras Kovacs <andras(a)sth.sze.hu> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |julliard(a)winehq.org -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #23 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-07 02:43:53 CST --- What paths for fix in the next release, cut commit or refine it? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #24 from Nikolay Sivov <bunglehead(a)gmail.com> 2013-03-07 02:54:13 CST --- Does it have something to do with d3d11 used by this game? You could try to disable is in winecfg. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #25 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-07 03:21:06 CST --- (In reply to comment #24)
Does it have something to do with d3d11 used by this game? You could try to disable is in winecfg.
For me it's: fixme:d3d11:D3D11CreateDevice stub: adapter (nil), driver_type D3D_DRIVER_TYPE_HARDWARE, swrast (nil), flags 0, feature_levels 3341084, levels 0x2, sdk_version 7, device 0x32fb14, feature_level 0x32fb18, context 0x32fb10 but force disabling is not necessarily. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #26 from Felix Hellmann <privat(a)cirk2.de> 2013-03-07 06:09:52 CST --- (In reply to comment #24)
Does it have something to do with d3d11 used by this game? You could try to disable is in winecfg.
Disabling d3d11 does not change the outcome. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #27 from Felix Hellmann <privat(a)cirk2.de> 2013-03-07 08:10:00 CST --- (In reply to comment #23)
What paths for fix in the next release, cut commit or refine it?
I can't really grasp the ramification in this commit. Aparently gcc version 4.2 added some function to align function entry points to allow old applications to use these entry points. On the other hand this commit had only been introduced in December 2012, maybe it was a bugfix? Alexandre Julliard should be able to help us with that. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #28 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-07 08:23:44 CST --- (In reply to comment #27) Felix, yes Julliard might be help with it, but how be with Andras comment #15 ? A strange situation indeed. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #29 from Felix Hellmann <privat(a)cirk2.de> 2013-03-07 09:14:35 CST --- (In reply to comment #15)
With the latest wine (1.5.25) on Ubuntu 12.10 32 bit with GeForce 8500GT (driver 310.14) I have no problem at all, it runs slow (but I think this is OK because of old graphics card)
can you try to upgrade to Nvidia version 310.32 or higher? Maybe the regression only kicks in on those newer Driver version? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #30 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-07 09:29:49 CST --- Felix, I don't think the problem in driver, because of launching on Intel HD 3000 (IGP) give the same result. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #31 from Felix Hellmann <privat(a)cirk2.de> 2013-03-07 10:00:25 CST --- (In reply to comment #30)
Felix, I don't think the problem in driver, because of launching on Intel HD 3000 (IGP) give the same result.
Ok only saw that he was the only one with a Nvidia driver before yours. Do both of you use the Wine Packages? if not do you have the same gcc version? I must admit it is only guessed, but since the commit only changed a compile option i think this is the direction to guess toward. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #32 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-07 10:25:20 CST --- (In reply to comment #31)
Ok only saw that he was the only one with a Nvidia driver before yours.
Do both of you use the Wine Packages? if not do you have the same gcc version? I must admit it is only guessed, but since the commit only changed a compile option i think this is the direction to guess toward.
I'm using Ubuntu package from Wine PPA, so I'm didn't compile source by myself or change it somehow. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #33 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-08 08:32:40 CST --- 1.5.19 is the latest version with which game is working. It can be dependent to commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Andrey Gusev <andrey.goosev(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|Tomb Raider (2013) |Tomb Raider (2013) screen |distortions during loading |distortions and game | |freezes -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 thanoulas <thanoulas(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |thanoulas(a)gmail.com -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #34 from thanoulas <thanoulas(a)gmail.com> 2013-03-09 06:18:17 CST --- On a Mac here, OS X 10.8.2, trying with different wine versions, all behave the same. I tried the following: 1.5.9, 1.5.11, 1.5.12, 1.5.13, 1.5.19, 1.5.25, and latest from git. Game launching on all, screen is all glitched, but I can see Loading and then goes black when loading is done. Not quitting, or crashing. StrictDrawOrdering enabled, 1.5.9 never even goes in game, all the rest look fine without glitches, then I see profile and then just the background and the game stops. I am now building the latest got without commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a, I'll let you know how it goes. So, have you managed to start the game with older versions, or not? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #35 from thanoulas <thanoulas(a)gmail.com> 2013-03-09 06:31:00 CST --- So, reverting commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a does absolutely no difference to me. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #36 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-09 07:31:10 CST --- (In reply to comment #35)
So, reverting commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a does absolutely no difference to me.
I can't to affirm on 100% that problem in this commit, but last working version (perhaps for me) is 1.5.19. Next version and all after has this commit and not working. At the end it can be a dependency tree. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #37 from Felix Hellmann <privat(a)cirk2.de> 2013-03-09 07:51:00 CST --- (In reply to comment #35)
So, reverting commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a does absolutely no difference to me.
The commit had no effect on OSX. It only affects systems witch GCC >= 4.2 that are not OSX. (Have a look in the diff (git show 4adfb...) of the commit its pretty self explanatory, even if you don't know C). -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #38 from thanoulas <thanoulas(a)gmail.com> 2013-03-09 08:28:53 CST --- (In reply to comment #37)
(In reply to comment #35)
So, reverting commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a does absolutely no difference to me.
The commit had no effect on OSX. It only affects systems witch GCC >= 4.2 that are not OSX. (Have a look in the diff (git show 4adfb...) of the commit its pretty self explanatory, even if you don't know C).
I do know C, but I didn't read the code in the commit. Now that I did I can understand why I see no difference. So what exactly does reversing this commit fix in Linux? Does this fix the glitches, the freezing, or both? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #39 from Felix Hellmann <privat(a)cirk2.de> 2013-03-09 08:33:32 CST --- (In reply to comment #38)
I do know C, but I didn't read the code in the commit. Now that I did I can understand why I see no difference. So what exactly does reversing this commit fix in Linux? Does this fix the glitches, the freezing, or both?
The freezing, the glitches are fixed with "StrictDrawOrdering" enabled. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #40 from thanoulas <thanoulas(a)gmail.com> 2013-03-09 08:41:46 CST --- Ok. When I enable StrictDrawOrdering on mac, I no longer see the Loading screen at the beginning, but when loading is finished I see the profile screen for a split second (without glitches), then the screen is emptied and left with only the background image (where it freezes). Is this the case in Linux, too? (only reverting the commit does fix the freezing for you). If that's the case, I wonder what the result on mac would be if I also removed the #ifdef __APPLE__ from that same line. I am going to give it a try. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #41 from Felix Hellmann <privat(a)cirk2.de> 2013-03-09 08:47:10 CST --- (In reply to comment #40)
Ok. When I enable StrictDrawOrdering on mac, I no longer see the Loading screen at the beginning, but when loading is finished I see the profile screen for a split second (without glitches), then the screen is emptied and left with only the background image (where it freezes). Is this the case in Linux, too? (only reverting the commit does fix the freezing for you).
If that's the case, I wonder what the result on mac would be if I also removed the #ifdef __APPLE__ from that same line. I am going to give it a try.
The exact point of freezing varies a bit (had plenty of occasions while regression testing). I suspect it to freeze during engine initalisation which is happening n the lst frames of the Loading/intro screen, the "Profile" frames are the last ones, I had occasions where I could hear the beginning of the waves from the menu etc. Would be interesting to see your result of changing the alignment in that header. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #42 from Alexandre Julliard <julliard(a)winehq.org> 2013-03-09 08:51:52 CST --- You can't change it on Mac OS, it will crash and burn. If you want to try to track it down (on Linux) you should compile individual dlls with the patch reverted until you find the one that makes a difference. Then repeat for individual files in that dll, and finally function by function in that file to find the offending code. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #43 from Felix Hellmann <privat(a)cirk2.de> 2013-03-09 08:56:31 CST --- (In reply to comment #42)
You can't change it on Mac OS, it will crash and burn.
If you want to try to track it down (on Linux) you should compile individual dlls with the patch reverted until you find the one that makes a difference. Then repeat for individual files in that dll, and finally function by function in that file to find the offending code.
How is that workflow for such a process? Copy the compiled dll in the prefix? And how do I compile single functions? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #44 from Alexandre Julliard <julliard(a)winehq.org> 2013-03-09 09:00:03 CST --- You can simply rebuild individual dlls by running make in the dll directory. For individual functions you'll have to put a #undef/#define __stdcall at various places in the file. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #45 from Felix Hellmann <privat(a)cirk2.de> 2013-03-09 09:06:42 CST --- I'm not at home for the Weekend, I'll try to track down the dll when I'm back. Maybe I can figure out the compiling of single functions, I'll have a read on the backgrounds of #define __stdcall. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #46 from thanoulas <thanoulas(a)gmail.com> 2013-03-09 11:40:32 CST --- (In reply to comment #42)
You can't change it on Mac OS, it will crash and burn.
If you want to try to track it down (on Linux) you should compile individual dlls with the patch reverted until you find the one that makes a difference. Then repeat for individual files in that dll, and finally function by function in that file to find the offending code.
Well it didn't crash or burn, but it didn't compile either. Getting errors. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 rico.tilgner(a)gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rico.tilgner(a)gmail.com -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Jonathan Strander <mblackwell1024(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mblackwell1024(a)gmail.com -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #47 from Felix Hellmann <privat(a)cirk2.de> 2013-03-12 21:04:32 CDT --- Ok I found the guilty dll, and as I feared it is one of the "big" ones: kernel32 now how exactly does the compiling of single Functions work? I have no idea of the win32 api and only limited experience in c and c++. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #48 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-13 06:31:22 CDT --- (In reply to comment #47)
Ok I found the guilty dll, and as I feared it is one of the "big" ones: kernel32 now how exactly does the compiling of single Functions work? I have no idea of the win32 api and only limited experience in c and c++.
I guess, now we need to find someone who can work with it. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Anthony Mattheakakis <antony256(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|antony256(a)gmail.com | -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #49 from Jonathan Strander <mblackwell1024(a)gmail.com> 2013-03-13 23:04:33 CDT --- Aren't these really two separate bugs? One is a graphical distortion (fixed by StrictDrawOrdering, which is a workaround not a fix of course), and the other is a crash. Anyway, I've tested this now with 1.5.23 and 1.5.16 (in a clean prefix). 1.5.23 had graphical glitches (no matter the settings) and would display only the background after the initial loading and then appear to freeze (but not crash). 1.5.16 had slight graphical glitches during the initial load (the copyright credits to be precise), but the menu and game all played normally... as long as StrictDrawOrdering was enabled. Otherwise it had wild distortions. On Ubuntu Linux 12.10, using an Intel Q9550 and a GTX 560Ti. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #50 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-14 07:55:46 CDT --- (In reply to comment #49) Yes, these bugs are separate. And problem with glitches can be "solved" by workaround for now. About freezes - it's a big problem, because each user have a different reaction to it. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #51 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-03-17 14:42:23 CDT --- Continues on 1.5.26 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #52 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-04-05 05:07:57 CDT --- Retesting in 1.5.27 no shows any flickering problem. Clean wineprefix. Game with latest update 1.01.732.1. Nvidia driver 313.30. Freezes still exist. Can someone try to yourself? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #53 from Felix Hellmann <privat(a)cirk2.de> 2013-04-05 06:27:29 CDT --- I can confir that the distortions during the loading are gone, even when disabling strickDrawOrder. But it now freezes even when reverting the commit 4adfb787f4e8c36a37[...] -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Andrey Gusev <andrey.goosev(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|Tomb Raider (2013) screen |Tomb Raider (2013) game |distortions and game |freezes |freezes | -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Felix Hädicke <felixhaedicke(a)web.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |felixhaedicke(a)web.de --- Comment #54 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-06 20:47:49 CDT --- For me, The freeze is gone in Wine 1.5.27 (Linux) after compiling with one of the following changes: * revert commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a * unmodified 1.5.27 sources, but without optimizations: CFLAGS="-O0" The latter could be an option to try on Mac OS X (I don't have a Mac, so I cannot try it myself), because reverting the commit is useless on Mac OS X. But it is of course not a real solution for this game, because missing optimization leads to a serious performance impact for particularly for DirectX software. Another attemt was to compile Wine with the reverted commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a and CFLAGS="-O2 -mstackrealign". I had expected that the deadlock would also occur, because from my understanding the compiler option "-mstackrealign" has the same effect for all functions as the function attribute "force_align_arg_pointer" (which the commit adds for __stdcall / __cdecl functions). More information about stack alignment issues can be found in Wine Bug 22316 and the following GCC bug entry: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40838. But I do not have the knowledge to judge if our bug is caused by a GCC bug or if it is a bug in Wine. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #55 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-07 06:44:19 CDT --- Created attachment 44108 --> http://bugs.winehq.org/attachment.cgi?id=44108 backtraces in freezed state winebd "bt all" backtrace for all processes / threads when the freeze occured. Used Wine 1.5.27 (unmodified sources) with CFLAGS="-g -O2" (could not use -O0, because the freeze does not occur without optimization) -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #56 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-07 09:08:04 CDT --- It is just dlls/kernel32/sync.c which is broken with compiler optimizations. So disabling optimizations for this file only, by adding a line "#pragma GCC optimize ("O0")" on top of this source file, and compiling with default CFLAGS, fixes the freeze for me. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #57 from Jonathan Strander <mblackwell1024(a)gmail.com> 2013-04-07 12:45:41 CDT --- I still get graphical distortions without StrictDrawOrdering, however your modification to sync.c does prevent the crash. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #58 from thanoulas <thanoulas(a)gmail.com> 2013-04-07 13:45:47 CDT --- I tried building wine with the #pragma directive, but it makes no difference on OS X 10.8.3, I'm afraid. I still get to the "Profiles" screen, then I hear a selection sound, then just shows the game's background (something that looks like a map). Can I somehow trace the frozen processes like you did? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #59 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-07 14:25:40 CDT --- Have you tried to compile whole Wine without optimizations? CFLAGS="-g -O0" This would also help to get a better backtrace if the freeze still occurs. It could be interesting to compare the Mac OS X backtraces with mine. To get backtraces for all threads, start the wine process as you always do, wait until it freezes, start winedbg (in the same WINEPREFIX) and execute the command "bt all". To redirect the backtraces to a file, I started winedbg using "winedbg > tr-deadlock-trace.txt" and typed "bt all" without seeing the winedbg command line which is also redirected then. But there is maybe a more elegant way. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #60 from thanoulas <thanoulas(a)gmail.com> 2013-04-07 14:29:15 CDT --- Thanks I'll try that. I'm now trying to rebuild wine with the #pragma after I do a make clean first, because I suspect that kernel32 was already built in the folder and that the compilation never happened. If that doesn't work, I'll try building it without optimizations. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #61 from thanoulas <thanoulas(a)gmail.com> 2013-04-07 15:03:53 CDT --- Ok, good news, compiling with -O0 makes the game go past the freezing point, though it's very slow and I get some graphical glitches on screen, even with strictdrawordering enabled. I'll try and get a backtrace now, but I'll have to rebuild with -g -02, is that correct? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #62 from thanoulas <thanoulas(a)gmail.com> 2013-04-07 15:34:52 CDT --- Created attachment 44114 --> http://bugs.winehq.org/attachment.cgi?id=44114 OS X 10.8.3 - Wine 1.5.27 Winedbg Backtrace Here's the backtrace for when the game is stuck at the profile page with winedbg bt all Engine is wine from git, built with CFLAGS="-g -O2" Unfortunately the output is not showing function names for some reason, like the linux backtrace, so I can't really tell what's going on. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #63 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-07 16:09:00 CDT --- Ok, the winedbg backtrace without function names is not really helpful. Could that be Bug 22384 ? Including "-gstabs+" in CFLAGS and maybe using "-O1" instead of "-O2" could help. The freeze is also there with -O1 on Linux. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #64 from thanoulas <thanoulas(a)gmail.com> 2013-04-07 17:48:39 CDT --- "-gstabs+" and -01 still don't generate function names in the backtrace. I'm thinking of disabling optimizations in the kernel32 .c files one by one until the game works, this is the only way I can think of now to find the file where this bug exists in OS X -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #65 from thanoulas <thanoulas(a)gmail.com> 2013-04-07 18:11:37 CDT --- Ok, from http://gcc.gnu.org/onlinedocs/gcc/Function-Specific-Option-Pragmas.html "The ‘#pragma GCC optimize’ pragma is not implemented in GCC versions earlier than 4.4." My OS X: gcc --version i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00) So #pragma GCC optimize is not supported! Anything equivalent for older GCC versions, or a way to update my GCC on OS X? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 aaronchi(a)gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |aaronchi(a)gmail.com --- Comment #66 from aaronchi(a)gmail.com 2013-04-07 18:29:48 CDT --- I've been following this thread as I'm having the same issues with Tomb Raider on a mac. I tried the #pragma GCC optimize that didn't work either. Going to try now with an updated gcc: brew tap homebrew/versions brew install gcc48 Thanks for working on this. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #67 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-04-08 04:48:12 CDT --- What a solution for Linux users? Updating GCC? Now I have version 4.7.2 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #68 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-08 14:31:34 CDT --- Created attachment 44122 --> http://bugs.winehq.org/attachment.cgi?id=44122 Patch which disables optimization for CreateEventExW It seems to be the "CreateEventExW" function which gets broken in optimized builds. This patch disables optimization for this function only. Sorry, it uses a function attribute which only recent GCC version (probably >= 4.4) understand. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #69 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-04-08 16:19:31 CDT --- Felix, thanks for patch. Guys, please, post your results with this one. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #70 from Alexandre Julliard <julliard(a)winehq.org> 2013-04-08 16:32:54 CDT --- Does it work if you remove the IsBadReadPtr call instead? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #71 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-08 16:55:52 CDT --- No, removing the IsBadReadPtr call does unfortunately not help. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #72 from thanoulas <thanoulas(a)gmail.com> 2013-04-10 10:40:21 CDT --- (In reply to comment #68)
Created attachment 44122 [details] Patch which disables optimization for CreateEventExW
It seems to be the "CreateEventExW" function which gets broken in optimized builds. This patch disables optimization for this function only. Sorry, it uses a function attribute which only recent GCC version (probably >= 4.4) understand.
So, after a LONG time, trying to setup GCC 4.8 on my OS X 10.8, that can successfully build wine, I can confirm that this patch fixes the freeze on OS X, too. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #73 from aaronchi <aaronchi(a)gmail.com> 2013-04-10 12:59:58 CDT --- @thanoulas Do you have any documentation on the flags you used to install gcc 4.8 on OSX? I was having issues as well. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #74 from thanoulas <thanoulas(a)gmail.com> 2013-04-10 13:07:39 CDT --- I'm using macports, so you need to get gcc48 and any required dependencies with the +universal flag, or you'll be missing the required i386 arch in the dylibs. I originally had Homebrew as my package manager, but I had to change that because they don't offer a universal GCC you will also need to change the Makefile.in in the folder /dlls/winemac.drv and add CC=clang I'm sure I had more issues, but they were just missing dependencies, some trial and error will get you there, eventually. Make sure to build with -m32 and the MacOSX10.8.sdk in your DYLD path -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #75 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-10 15:50:25 CDT --- @thanoulas Your description confuses me a little. Have you used clang (3.2) or GCC (4.8) - these are different compilers - also for dlls/kernel32/sync.c and the rest, or did you just use clang for winemac.drv? I tried to compile only the unmodified dlls/kernel32/sync.c with clang 3.0 instead of GCC (the rest of Wine with GCC 4.7.2) with -O1 and again with -O3 and the game did not freeze. Another thing I tried was to compile Wine without optimizations, only dlls/kernel32/sync.c with optimization turned on. The bug was still there. Then I turned on optimization only for CreateEventExW and the game did not freeze as I had expected. What I found out in my further investigations is that enabling optimizations (-O1 is enough) for the following thre functions only (and not for all the rest of Wine) makes the bug occur: * CreateEventExW * CreateSemaphoreW * CreateSemaphoreExW And if optimization is disabled for only one of these three functions, the freeze is gone. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #76 from Alexandre Julliard <julliard(a)winehq.org> 2013-04-10 16:06:56 CDT --- It would be interesting to try if making them DECLSPEC_HOTPATCH makes a difference. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #77 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-10 16:46:18 CDT --- Created attachment 44141 --> http://bugs.winehq.org/attachment.cgi?id=44141 Patch: Add DECLSPEC_HOTPATCH to CreateSemaphoreExW Adding DECLSPEC_HOTPATCH to CreateSemaphoreExW does fix the problem. It makes no difference for me if the attribute is added to CreateEventExW and / or CreateSemaphoreW as well or not. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #78 from thanoulas <thanoulas(a)gmail.com> 2013-04-10 17:13:46 CDT --- (In reply to comment #75)
@thanoulas Your description confuses me a little. Have you used clang (3.2) or GCC (4.8) - these are different compilers - also for dlls/kernel32/sync.c and the rest, or did you just use clang for winemac.drv?
Compile everything with gcc48 and winemac.drv with clang Doing the change I described in the Makefile.in of winemac.drv will force the main make command to compile winemac.drv with clang instead of gcc -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #79 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-04-11 15:03:15 CDT --- I see patch was committed in git d8c72febc823c6076a58a7d1bb166ff240db8326 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #80 from Felix Hädicke <felixhaedicke(a)web.de> 2013-04-11 15:22:23 CDT --- I've recompiled wine 1.5.27 with the DECLSPEC_HOTFIX patch and noticed that it does not fix it any longer for me. The same for recent git version with commit d8c72febc823c6076a58a7d1bb166ff240db8326. I don't know yet what was different when I tested it yesterday. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #81 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-04-13 02:46:52 CDT --- Nothing changes on 1.5.28 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #82 from Jonathan Strander <mblackwell1024(a)gmail.com> 2013-04-13 16:43:55 CDT --- I can confirm that I still require at least the modification from comment #68 (with some changes obviously so it patches correctly) in order to not freeze with 1.5.28. And also I still require StrictDrawOrdering to be enabled. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #83 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-04-19 06:13:11 CDT --- So, it's not quite a Wine bug? Something mutual? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 TestSubject <sirbubbles01(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sirbubbles01(a)gmail.com --- Comment #84 from TestSubject <sirbubbles01(a)gmail.com> 2013-04-22 14:16:04 CDT --- (In reply to comment #82)
I can confirm that I still require at least the modification from comment #68 (with some changes obviously so it patches correctly) in order to not freeze with 1.5.28.
And also I still require StrictDrawOrdering to be enabled.
I don't know what everyone else's experience has been with the git version of wine, but I've found that simply reverting patch 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a is enough to allow the game to run. I've also found that StrictDrawOrdering is not necessary for me. I've got nvidia driver version 319.12, so I don't know if that's having some sort of effect. But I have had no issues whatsoever getting the game to load by simply reverting this patch. I have an amd 955 cpu, nvidia gtx660, and Ubuntu 12.10 x64. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #85 from sylvain.pare(a)gmail.com 2013-05-01 13:07:58 CDT --- Created attachment 44345 --> http://bugs.winehq.org/attachment.cgi?id=44345 tomb raider crash wine 1.5.29 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Max Mustermann <maxmusterm(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |maxmusterm(a)gmail.com --- Comment #86 from Max Mustermann <maxmusterm(a)gmail.com> 2013-05-08 15:10:00 CDT --- Would it be bad to add a gcc workaround in the wine codebase? Or do you think the bug should only be fixed in gcc itself? What is btw the status of wine with clang?The bugs mentioned here http://wiki.winehq.org/Clang seem all only to apply to older clang versions or don't they? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #87 from Austin English <austinenglish(a)gmail.com> 2013-05-08 17:06:28 CDT --- (In reply to comment #86)
Would it be bad to add a gcc workaround in the wine codebase? Or do you think the bug should only be fixed in gcc itself? What is btw the status of wine with clang?The bugs mentioned here http://wiki.winehq.org/Clang seem all only to apply to older clang versions or don't they?
From that page: Using Wine 1.5.20+ and Clang 3.2, the Wine ConformanceTests behaves the same as gcc (tested using 4.7.2 on Fedora 18, x86_64).
I can't comment on how it runs on OSX, but at least on Linux, things seem to work. I haven't tried using it for my system installed wine, so there may be other bugs that the test suite doesn't expose. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #88 from Max Mustermann <maxmusterm(a)gmail.com> 2013-06-05 13:00:48 CDT --- Does anyone know which gcc optimisation breaks the code and did anyone open a bug report for gcc? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Travis Greenman <cartman-2000(a)cartman-2000.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |cartman-2000(a)cartman-2000.n | |et --- Comment #89 from Travis Greenman <cartman-2000(a)cartman-2000.net> 2013-07-14 19:38:30 CDT --- (In reply to comment #68)
Created attachment 44122 [details] Patch which disables optimization for CreateEventExW
It seems to be the "CreateEventExW" function which gets broken in optimized builds. This patch disables optimization for this function only. Sorry, it uses a function attribute which only recent GCC version (probably >= 4.4) understand.
This also fixes the freeze for me on wine 1.6_rc3. I'm using 64bit gentoo with the 3.8.13 kernel, gcc 4.6.3. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #90 from Andrey Gusev <andrey.goosev(a)gmail.com> 2013-09-15 05:04:49 CDT --- Updating for Felix Hädicke's attachment 44122 for Wine 1.7.2 HANDLE WINAPI __attribute__((optimize("-O0"))) DECLSPEC_HOTPATCH CreateEventExW( SECURITY_ATTRIBUTES *sa, LPCWSTR name, DWORD flags, DWORD access ) -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Jeff Cook <jeff(a)deseret-tech.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jeff(a)deseret-tech.com --- Comment #91 from Jeff Cook <jeff(a)deseret-tech.com> 2013-10-06 00:34:43 CDT --- Experiencing freeze on the profile screen with 6d03ce8d08e68d, even after modifying the posted patch to be applicable in consideration of d8c72febc823c607 which makes the function hotpatchable. Have done a full clean and recompile and still experience the issue. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #92 from Max Mustermann <maxmusterm(a)gmail.com> 2013-11-27 05:14:37 CST --- This bug was confirmed by many peers and there even are a few solutions (and proposals), yet this bug is still marked as unconfirmed? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Bruno Jesus <00cpxxx(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |patch Status|UNCONFIRMED |NEW Ever Confirmed|0 |1 --- Comment #93 from Bruno Jesus <00cpxxx(a)gmail.com> 2013-11-27 05:19:23 CST --- Confirmed by multiple users. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Berillions <berillions(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |berillions(a)gmail.com --- Comment #94 from Berillions <berillions(a)gmail.com> 2013-12-08 03:40:34 CST --- The game still crash after the login profil for me. All patch in this bug report does not work for me. I must to play at this game on Windows ? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #95 from TestSubject <sirbubbles01(a)gmail.com> 2013-12-08 07:12:27 CST --- If you happen to be building wine from git, you can get around the freezing problem by reverting this git commit. 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a. I've been running the game for a long time via this method. Shouldn't have to revert this git and others for basic functionality, but there you go. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 legluondunet(a)free.fr <legluondunet(a)free.fr> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |legluondunet(a)free.fr --- Comment #96 from legluondunet(a)free.fr <legluondunet(a)free.fr> 2013-12-08 13:00:25 CST --- I just tried this game, Steam version ,and after asked me my mail, the game freeze! Thank you for your help. LGDN. Config: Ubuntu 13.10 wine 1.7.8 32 bits AMD radeon 7850 with Ubuntu Catalyst drivers 13.101 version Processor Intel i5-3570 3,4 Ghz -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #97 from Berillions <berillions(a)gmail.com> 2013-12-08 13:05:41 CST --- Same thing for me ans there are two solutions : 1- If you use wine-git, apply this patch http://bugs.winehq.org/attachment.cgi?id=44122&action=diff 2- Or configure Wine with CFLAGS="-g -O0" -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #98 from Max Mustermann <maxmusterm(a)gmail.com> 2013-12-08 14:10:05 CST --- Or compile it with clang. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Anastasius Focht <focht(a)gmx.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |focht(a)gmx.net --- Comment #99 from Anastasius Focht <focht(a)gmx.net> 2013-12-08 17:03:24 CST --- Hello folks, I can't confirm this. Neither the freeze nor this 'flickering' effect is present here. The game works out of the box, I played until first savepoint. Using the Steam version (2013-12-08), 'windowed' mode. Specs: * Fedora 19 64-bit/32-bit * gcc 4.8.2 20131017 (Red Hat 4.8.2-1) * Wine 1.7.8-128-g37460b6 (64/32-bit, CFLAGS: -O2 -g) * Steam + game installed in 32-bit WINEPREFIX, dwrite disabled * NVIDIA GeForce GT 425M Regards -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #100 from legluondunet(a)free.fr <legluondunet(a)free.fr> 2013-12-09 04:07:08 CST --- Finally this game does not freeze if I launch it in full screen, without a wine virtual destop. Same problem for other Tomb Raider title like Tomb Raider last revelations and Chronicles. It seems AMD Catalyst drivers does not appreciate wine virtual desktop.I will open a bug report on AMD. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #101 from Andrey Gusev <andrey.goosev(a)gmail.com> ---
From changes list for gcc 4.8.2
"A new general optimization level, -Og, has been introduced. It addresses the need for fast compilation and a superior debugging experience while providing a reasonable level of runtime performance. Overall experience for development should be better than the default optimization level -O0." Upstream? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #102 from legluondunet(a)free.fr <legluondunet(a)free.fr> --- Sorry, wrong comment 100: the games still freeze with last 1.7.x wine version. But not with wine 1.5.17 and 1.5.19. I will try to patch the last git with http://bugs.winehq.org/attachment.cgi?id=44122&action=diff for test. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #103 from Adam Bolte <abolte(a)systemsaviour.com> --- Created attachment 47293 --> http://bugs.winehq.org/attachment.cgi?id=47293 Patch which disables optimization for CreateEventExW Updated for Wine 1.7.10+. Note that Tomb Raider is completely broken in 1.7.11. Will probably run a regression test soon, but I suspect it's bug 35388. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Adam Bolte <abolte(a)systemsaviour.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |abolte(a)systemsaviour.com -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Wylda <wylda(a)volny.cz> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |wylda(a)volny.cz --- Comment #104 from Wylda <wylda(a)volny.cz> --- (In reply to comment #103)
regression test soon, but I suspect it's bug 35388.
Another one could be bug 35408 as that was also new in 1.7.11. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #105 from Adam Bolte <abolte(a)systemsaviour.com> ---
Another one could be bug 35408 as that was also new in 1.7.11.
Correct, it was bug 35408. I just commented there. :) -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Matěj Týč <matej.tyc(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |matej.tyc(a)gmail.com --- Comment #106 from Matěj Týč <matej.tyc(a)gmail.com> --- I experience the freeze even in Wine 1.7.12 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Jarkko K <jarkko_korpi(a)hotmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jarkko_korpi(a)hotmail.com --- Comment #107 from Jarkko K <jarkko_korpi(a)hotmail.com> --- "fixme:d3d11:D3D11CreateDevice stub: adapter (nil), driver_type D3D_DRIVER_TYPE_HARDWARE, swrast (nil), flags 0, feature_levels 3406616, levels 0x2, sdk_version 7, device 0x33fb10, " Someone had that in his backlog. It means basically that the game tries to run dx11. Please disable directx 11 and 10 related libraries while trying to run the game. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Dooglio <doug(a)dooglio.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |doug(a)dooglio.net --- Comment #108 from Dooglio <doug(a)dooglio.net> --- I was experiencing a crash when I touched any key at all. It would play up to the point of when Laura is first hanging from the ceiling, but when I used the arrow keys, it would immediately lock up. I applied the last patch in the list above (http://bugs.winehq.org/attachment.cgi?id=47293) to wine 1.7 and it works like a champ. I even made a PPA for it if you run Ubuntu Saucy (13.10): https://launchpad.net/~doug/+archive/steam-wine-fix To use, add the following PPA: sudo apt-add-repository ppa:doug/steam-wine-fix sudo apt-get update sudo apt-get upgrade # This should force wine-1.7 to upgrade, if you have it installed. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 hNarval <drummaniac47(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |drummaniac47(a)gmail.com -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 pasa <psnirvana.lviv(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |psnirvana.lviv(a)gmail.com -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Rui Caridade <rcaridade145(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rcaridade145(a)gmail.com --- Comment #109 from Rui Caridade <rcaridade145(a)gmail.com> --- I am experiencing this problem on ubuntu 14.04 wine 1.7.20. Any chance this patch gets commited or better to talk to the ubuntu maintainers? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Anastasius Focht <focht(a)gmx.net> changed: What |Removed |Added ---------------------------------------------------------------------------- URL| |http://store.steampowered.c | |om/app/203160/ Summary|Tomb Raider (2013) game |Tomb Raider (2013) freezes |freezes |(bug in game engine, | |semaphore object handle | |value must be zero on all | |error paths) --- Comment #110 from Anastasius Focht <focht(a)gmx.net> --- Hello folks, I could reproduce the problem. It's actually buggy code in the game engine that just works by chance. I looked at the disassembly from different gcc optimization levels and could not really spot a problem. There is some questionable (redundant) code in the function prolog at O1+ but still no overly optimized code that breaks something. Running with +relay avoided the freeze so stack layout/values certainly had an impact. A +server trace along with +msvcrt (msvcrt = good search patterns to correlate with working a +relay trace): --- snip --- ... 0063: create_semaphore( access=001f0003, attributes=00000080, initial=00000000, max=7fffffff, objattr={rootdir=0000,sd={},name=L""} ) 0063: create_semaphore() = 0 { handle=011c } ... 0063: create_semaphore( access=001f0003, attributes=00000080, initial=00000000, max=7fffffff, objattr={rootdir=0000,sd={},name=L""} ) 0063: create_semaphore() = 0 { handle=0120 } ... 0063: create_event( access=001f0003, attributes=00000000, manual_reset=0, initial_state=0, objattr={rootdir=0000,sd={},name=L""} ) 0063: create_event() = 0 { handle=0208 } 0063: create_event( access=001f0003, attributes=00000000, manual_reset=0, initial_state=0, objattr={rootdir=0000,sd={},name=L""} ) 0063: create_event() = 0 { handle=020c } 0063:trace:msvcrt:pf_printf_a Format is: "%s%s.drm" ... 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33ece4 2000 "InitializePanel: 0" 18) 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33eae4 2000 "ShowPanel(0)" 12) 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33eae4 2000 "Prev ID 0" 9) 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33ece4 2000 "this[nav_list_item_8_mc].GetState() 2" 37) ... 0063: release_semaphore() = 0 { prev_count=00000000 } 0063: release_semaphore( handle=011c, count=00000001 ) 0063: release_semaphore() = 0 { prev_count=00000000 } 0066: release_semaphore( handle=0208, count=00000001 ) 0066: release_semaphore() = OBJECT_TYPE_MISMATCH { prev_count=00000000 } 0063: release_semaphore( handle=011c, count=00000001 ) 0063: release_semaphore() = 0 { prev_count=00000001 } 0066: select( flags=2, cookie=053fe4dc, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={011c,0120}} ) 0066: select() = 0 { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0063: select( flags=2, cookie=0033f09c, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={0208}} ) 0063: select() = PENDING { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0066: release_semaphore( handle=0208, count=00000001 ) 0066: release_semaphore() = OBJECT_TYPE_MISMATCH { prev_count=00000000 } 0066: select( flags=2, cookie=053fe4dc, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={011c,0120}} ) 0066: select() = 0 { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0066: release_semaphore( handle=0208, count=00000001 ) 0066: release_semaphore() = OBJECT_TYPE_MISMATCH { prev_count=00000000 } 0066: event_op( handle=0208, op=1 ) 0063: *wakeup* signaled=0 0066: event_op() = 0 0063: select( flags=2, cookie=0033f09c, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={0208}} ) 0063: select() = PENDING { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0066: select( flags=2, cookie=053fe4dc, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={011c,0120}} ) 0066: select() = PENDING { timeout=infinite, call={APC_NONE}, apc_handle=0000 } ... <threads 00063 and 0066 wait blocked> --- snip --- The 'OBJECT_TYPE_MISMATCH' error catched my attention. Object/handle 0x208 was certainly valid but not a semaphore. That value was leaked through earlier call(s) which incidentally mapped to a valid handle (event object). Because the event was waitable but not signaled, the thread assumed it waited on a counted semaphore got blocked forever. An unoptimized build of Wine gave this trace: --- snip --- ... 0062: create_semaphore( access=001f0003, attributes=00000080, initial=00000000, max=7fffffff, objattr={rootdir=0000,sd={},name=L""} ) 0062: create_semaphore() = 0 { handle=011c } 0062: create_semaphore( access=001f0003, attributes=00000080, initial=00000000, max=7fffffff, objattr={rootdir=0000,sd={},name=L""} ) 0062: create_semaphore() = 0 { handle=0120 } ... 0062: create_event( access=001f0003, attributes=00000080, manual_reset=0, initial_state=0, objattr={rootdir=0000,sd={},name=L""} ) 0062: create_event() = 0 { handle=0208 } 0062: create_event( access=001f0003, attributes=00000080, manual_reset=0, initial_state=0, objattr={rootdir=0000,sd={},name=L""} ) 0062: create_event() = 0 { handle=020c } 0062:trace:msvcrt:pf_printf_a Format is: "%s%s.drm" ... 0062:trace:msvcrt:MSVCRT_strncpy_s (0x33eca4 2000 "InitializePanel: 0" 18) 0062:trace:msvcrt:MSVCRT_strncpy_s (0x33eaa4 2000 "ShowPanel(0)" 12) 0062:trace:msvcrt:MSVCRT_strncpy_s (0x33eaa4 2000 "Prev ID 0" 9) 0062:trace:msvcrt:MSVCRT_strncpy_s (0x33eca4 2000 "this[nav_list_item_8_mc].GetState() 2" 37) ... 0062: release_semaphore( handle=011c, count=00000001 ) 0065: *wakeup* signaled=0 0062: release_semaphore() = 0 { prev_count=00000000 } ... 0062: release_semaphore( handle=011c, count=00000001 ) 0065: *wakeup* signaled=0 0062: release_semaphore() = 0 { prev_count=00000000 } 0062: release_semaphore( handle=011c, count=00000001 ) 0062: release_semaphore() = 0 { prev_count=00000000 } 0065: release_semaphore( handle=7b8bc000, count=00000001 ) 0065: release_semaphore() = INVALID_HANDLE { prev_count=00000000 } 0062: release_semaphore( handle=011c, count=00000001 ) 0062: release_semaphore() = 0 { prev_count=00000001 } 0065: select( flags=2, cookie=053fe5ec, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={011c,0120}} ) 0065: select() = 0 { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0062: select( flags=2, cookie=0033f15c, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={7b8bc000}} ) 0062: select() = INVALID_HANDLE { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0065: release_semaphore( handle=7b8bc000, count=00000001 ) 0065: release_semaphore() = INVALID_HANDLE { prev_count=00000000 } 0062: select( flags=2, cookie=0033f15c, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={7b8bc000}} ) 0062: select() = INVALID_HANDLE { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0065: select( flags=2, cookie=053fe5ec, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={011c,0120}} ) 0065: select() = 0 { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0062: select( flags=2, cookie=0033f15c, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={7b8bc000}} ) 0062: select() = INVALID_HANDLE { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0065: release_semaphore( handle=7b8bc000, count=00000001 ) 0065: release_semaphore() = INVALID_HANDLE { prev_count=00000000 } 0062: select( flags=2, cookie=0033f15c, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={0208}} ) 0062: select() = PENDING { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0065: event_op( handle=0208, op=1 ) 0062: *wakeup* signaled=0 0065: event_op() = 0 0065: select( flags=2, cookie=053fe5ec, timeout=infinite, prev_apc=0000, result={}, data={WAIT,handles={011c,0120}} ) 0065: select() = PENDING { timeout=infinite, call={APC_NONE}, apc_handle=0000 } 0062: release_semaphore( handle=011c, count=00000001 ) 0065: *wakeup* signaled=0 ... --- snip --- The handle value now changed to 0x7b8bc000, which works "better" on sync APIs because it immediately fails them, giving 'INVALID_HANDLE'. Mapped base address of kernel32, leaked through local stack area. With +relay,+server together it become more obvious: --- snip --- ... 0063:Call KERNEL32.CreateSemaphoreW(00000000,00000000,7fffffff,00000000) ret=005283d1 0063: create_semaphore( access=001f0003, attributes=00000080, initial=00000000, max=7fffffff, objattr={rootdir=0000,sd={},name=L""} ) 0063: create_semaphore() = 0 { handle=011c } 0063:Ret KERNEL32.CreateSemaphoreW() retval=0000011c ret=005283d1 ... 0063:Call KERNEL32.CreateSemaphoreW(00000000,00000000,7fffffff,00000000) ret=005283e0 0063: create_semaphore( access=001f0003, attributes=00000080, initial=00000000, max=7fffffff, objattr={rootdir=0000,sd={},name=L""} ) 0063: create_semaphore() = 0 { handle=0120 } 0063:Ret KERNEL32.CreateSemaphoreW() retval=00000120 ret=005283e0 0063:Call msvcr90._beginthreadex(00000000,00020000,00527fd0,00000000,00000000,00000000) ret=00528406 ... 0063:Call KERNEL32.CreateEventW(00000000,00000000,00000000,00000000) ret=00527298 0063: create_event( access=001f0003, attributes=00000000, manual_reset=0, initial_state=0, objattr={rootdir=0000,sd={},name=L""} ) 0063: create_event() = 0 { handle=0208 } 0063:Ret KERNEL32.CreateEventW() retval=00000208 ret=00527298 ... 0063:Call KERNEL32.CreateSemaphoreW(00000000,00000000,ffffffff,00000000) ret=0064381d 0063:Ret KERNEL32.CreateSemaphoreW() retval=7bcde220 ret=0064381d ... 0063:trace:msvcrt:pf_printf_a Format is: "%s%s.drm" 0063:Call msvcr90._stricmp(00dd3b48 "generalbank",129ee9f9 "ac_bendy_01_medium_shrub") ret=0055b09c 0063:Ret msvcr90._stricmp() retval=00000001 ret=0055b09c ... 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33ece4 2000 "InitializePanel: 0" 18) ... 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33eae4 2000 "ShowPanel(0)" 12) ... 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33eae4 2000 "Prev ID 0" 9) ... 0063:trace:msvcrt:MSVCRT_strncpy_s (0x33ece4 2000 "this[nav_list_item_8_mc].GetState() 2" 37) ... --- snip --- 'TombRaider.exe' disassembly, the culprit: --- snip --- 00643804 33C0 XOR EAX,EAX 00643806 53 PUSH EBX ; lpName = NULL 00643807 6A FF PUSH -1 ; lMaximumCount = 0xffffffff (!) 00643809 53 PUSH EBX ; lInitialCount = 0 0064380A 53 PUSH EBX ; lpSemaphoreAttributes = NULL 0064380B 8946 08 MOV DWORD PTR DS:[ESI+8],EAX 0064380E 895E 0C MOV DWORD PTR DS:[ESI+C],EBX 00643811 895E 10 MOV DWORD PTR DS:[ESI+10],EBX 00643814 895E 14 MOV DWORD PTR DS:[ESI+14],EBX 00643817 FF15 FC70D800 CALL DWORD PTR DS:[<&KERNEL32.CreateSemaphoreW>] 0064381D 8906 MOV DWORD PTR DS:[ESI],EAX 0064381F 8BC6 MOV EAX,ESI 00643821 5E POP ESI 00643822 5B POP EBX 00643823 C2 0400 RETN 4 --- snip --- 'lMaximumCount' < 0 (LONG) can't obviously work. The code here (including caller) doesn't check the returned handle value and stores it right away for later use. Wine doesn't initialize the local variable that ought to carry the return value, which is passed as 'out' parameter to NtXXX API. Depending on the stack layout/optimizations, different stack values get leaked as handle values. Source: http://source.winehq.org/git/wine.git/blob/0e1b381503751294c6e6925b9ba1eeb0a... --- snip --- 777 HANDLE WINAPI DECLSPEC_HOTPATCH CreateSemaphoreExW( SECURITY_ATTRIBUTES *sa, LONG initial, LONG max, 778 LPCWSTR name, DWORD flags, DWORD access ) 779 { 780 HANDLE ret; 781 UNICODE_STRING nameW; 782 OBJECT_ATTRIBUTES attr; 783 NTSTATUS status; ... 798 status = NtCreateSemaphore( &ret, access, &attr, initial, max ); 799 if (status == STATUS_OBJECT_NAME_EXISTS) 800 SetLastError( ERROR_ALREADY_EXISTS ); 801 else 802 SetLastError( RtlNtStatusToDosError(status) ); 803 return ret; 804 } --- snip --- You need to explicitly zero-init 'ret' here or have 'NtCreateSemaphore' zero-init the 'out' handle parameter. With that part fixed, the freeze is gone. Despite the failure of 'CreateSemaphoreExW' to create a sync object the game works (profile -> main screen cut scene threading synchronization). Regards -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 Anastasius Focht <focht(a)gmx.net> changed: What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |149d60318fac6314726f256c1cc | |a1e8c4806202b Status|NEW |RESOLVED Component|-unknown |kernel32 Resolution|--- |FIXED --- Comment #111 from Anastasius Focht <focht(a)gmx.net> --- Hello folks, this is fixed by commit http://source.winehq.org/git/wine.git/commitdiff/149d60318fac6314726f256c1cc... Thanks Austin. Regards -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=33125 Alexandre Julliard <julliard(a)winehq.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #112 from Alexandre Julliard <julliard(a)winehq.org> --- Closing bugs fixed in 1.7.22. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 fusion.lightcat(a)gmail.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |fusion.lightcat(a)gmail.com --- Comment #113 from fusion.lightcat(a)gmail.com --- When i try to patch the attachment number 47293 (disable optimization) with "patch -p1 < patch.diff" it says: patching file dlls/kernel32/sync.c Hunk #1 FAILED at 456. 1 out of 1 hunk FAILED -- saving rejects to file dlls/kernel32/sync.c.rej How can i fix this? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #114 from Berillions <berillions(a)gmail.com> --- (In reply to fusion.lightcat from comment #113)
When i try to patch the attachment number 47293 (disable optimization) with "patch -p1 < patch.diff" it says:
patching file dlls/kernel32/sync.c Hunk #1 FAILED at 456. 1 out of 1 hunk FAILED -- saving rejects to file dlls/kernel32/sync.c.rej
How can i fix this?
The bug is fixed in wine 1.7.22 ... read the AJ's message -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #115 from Jerome Leclanche <adys.wh(a)gmail.com> --- (In reply to Berillions from comment #114) According to a user it still happens on 1.7.23. Can anyone confirm? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=33125 --- Comment #116 from legluondunet(a)free.fr <legluondunet(a)free.fr> --- I just tested it with 1.7.23 and I can access to the aim menu without freeze. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
participants (1)
-
wine-bugs＠winehq.org