[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: [e-mail]
        ReportedBy: [e-mail]
                CC: [e-mail]
    Classification: Unclassified

Created attachment 43810
  --> http://bugs.winehq.org/attachment.cgi?id=43810
log

Launched on Nvidia GT 540M with 310.32 drivers.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #1 from Andrey Gusev  2013-03-05 04:02:21 CST ---
Created attachment 43811
  --> http://bugs.winehq.org/attachment.cgi?id=43811
screen

screen 1

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #2 from Andrey Gusev  2013-03-05 04:02:56 CST ---
Created attachment 43812
  --> http://bugs.winehq.org/attachment.cgi?id=43812
screen

screen 2

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #3 from Andrey Gusev  2013-03-05 04:03:17 CST ---
Created attachment 43813
  --> http://bugs.winehq.org/attachment.cgi?id=43813
screen

screen 3

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

jiveaxe  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #4 from jiveaxe  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

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #5 from Andrey Gusev  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?

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #6 from jiveaxe  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #7 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

Anthony Mattheakakis  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #8 from Andrey Gusev  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?

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #9 from jiveaxe  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?

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #10 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

sacrediou  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

Jack Peirce  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

Felix Hellmann  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #11 from jiveaxe  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #12 from Andrey Gusev  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. )

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #13 from Andrey Gusev  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 ?

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #14 from jiveaxe  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

Andras Kovacs  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #15 from Andras Kovacs  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)

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #16 from Andras Kovacs  2013-03-06 06:37:14 CST ---
Created attachment 43832
  --> http://bugs.winehq.org/attachment.cgi?id=43832
screen (GeForce 8500GT)

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #17 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #18 from Felix Hellmann  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!

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #19 from Felix Hellmann  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

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #20 from Bruno Jesus  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

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

[e-mail] changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #21 from [e-mail] 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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #22 from Felix Hellmann  2013-03-06 19:27:58 CST ---
Ok I found the guilty one:

4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a is the first bad commit
commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a
Author: Alexandre Julliard 
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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

Andras Kovacs  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #23 from Andrey Gusev  2013-03-07 02:43:53 CST ---
What paths for fix in the next release, cut commit or refine it?

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #24 from Nikolay Sivov  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #25 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #26 from Felix Hellmann  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #27 from Felix Hellmann  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #28 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #29 from Felix Hellmann  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?

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #30 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #31 from Felix Hellmann  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #32 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) distortions during loading

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #33 from Andrey Gusev  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

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Andrey Gusev  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Tomb Raider (2013)          |Tomb Raider (2013) screen
                   |distortions during loading  |distortions and game
                   |                            |freezes

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

thanoulas  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #34 from thanoulas  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?

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #35 from thanoulas  2013-03-09 06:31:00 CST ---
So, reverting commit 4adfb787f4e8c36a37ce1d53a7e6df16d03ecd8a does absolutely
no difference to me.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #36 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #37 from Felix Hellmann  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).

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #38 from thanoulas  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?

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #39 from Felix Hellmann  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #40 from thanoulas  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #41 from Felix Hellmann  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #42 from Alexandre Julliard  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #43 from Felix Hellmann  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?

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #44 from Alexandre Julliard  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #45 from Felix Hellmann  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #46 from thanoulas  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

[e-mail] changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Jonathan Strander  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #47 from Felix Hellmann  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++.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #48 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Anthony Mattheakakis  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|[e-mail]         |

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #49 from Jonathan Strander  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #50 from Andrey Gusev  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.

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #51 from Andrey Gusev  2013-03-17 14:42:23 CDT ---
Continues on 1.5.26

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #52 from Andrey Gusev  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?

[Bug 33125] Tomb Raider (2013) screen distortions and game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #53 from Felix Hellmann  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[...]

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Andrey Gusev  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Tomb Raider (2013) screen   |Tomb Raider (2013) game
                   |distortions and game        |freezes
                   |freezes                     |

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Felix Hädicke  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #54 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #55 from Felix Hädicke  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)

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #56 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #57 from Jonathan Strander  2013-04-07 12:45:41 CDT ---
I still get graphical distortions without StrictDrawOrdering, however your
modification to sync.c does prevent the crash.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #58 from thanoulas  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?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #59 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #60 from thanoulas  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #61 from thanoulas  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?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #62 from thanoulas  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #63 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #64 from thanoulas  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

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #65 from thanoulas  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?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

[e-mail] changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #66 from [e-mail] 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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #67 from Andrey Gusev  2013-04-08 04:48:12 CDT ---
What a solution for Linux users? Updating GCC? Now I have version 4.7.2

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #68 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #69 from Andrey Gusev  2013-04-08 16:19:31 CDT ---
Felix, thanks for patch. Guys, please, post your results with this one.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #70 from Alexandre Julliard  2013-04-08 16:32:54 CDT ---
Does it work if you remove the IsBadReadPtr call instead?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #71 from Felix Hädicke  2013-04-08 16:55:52 CDT ---
No, removing the IsBadReadPtr call does unfortunately not help.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #72 from thanoulas  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #73 from aaronchi  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #74 from thanoulas  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

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #75 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #76 from Alexandre Julliard  2013-04-10 16:06:56 CDT ---
It would be interesting to try if making them DECLSPEC_HOTPATCH makes a
difference.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #77 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #78 from thanoulas  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

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #79 from Andrey Gusev  2013-04-11 15:03:15 CDT ---
I see patch was committed in git d8c72febc823c6076a58a7d1bb166ff240db8326

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #80 from Felix Hädicke  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #81 from Andrey Gusev  2013-04-13 02:46:52 CDT ---
Nothing changes on 1.5.28

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #82 from Jonathan Strander  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #83 from Andrey Gusev  2013-04-19 06:13:11 CDT ---
So, it's not quite a Wine bug? Something mutual?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

TestSubject  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #84 from TestSubject  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #85 from [e-mail] 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

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Max Mustermann  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #86 from Max Mustermann  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?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #87 from Austin English  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #88 from Max Mustermann  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?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Travis Greenman  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |cartman-2000@cartman-2000.n
                   |                            |et

--- Comment #89 from Travis Greenman  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #90 from Andrey Gusev  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 )

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Jeff Cook  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #91 from Jeff Cook  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #92 from Max Mustermann  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?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Bruno Jesus  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1

--- Comment #93 from Bruno Jesus  2013-11-27 05:19:23 CST ---
Confirmed by multiple users.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Berillions  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #94 from Berillions  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 ?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #95 from TestSubject  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

[e-mail]  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #96 from [e-mail]  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

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #97 from Berillions  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"

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #98 from Max Mustermann  2013-12-08 14:10:05 CST ---
Or compile it with clang.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Anastasius Focht  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #99 from Anastasius Focht  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

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #100 from [e-mail]  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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #101 from Andrey Gusev  ---
> 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?

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #102 from [e-mail]  ---
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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #103 from Adam Bolte  ---
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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Adam Bolte  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Wylda  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #104 from Wylda  ---
(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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #105 from Adam Bolte  ---
> Another one could be bug 35408 as that was also new in 1.7.11.

Correct, it was bug 35408. I just commented there. :)

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Matěj Týč  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #106 from Matěj Týč  ---
I experience the freeze even in Wine 1.7.12

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Jarkko K  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #107 from Jarkko K  ---
"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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Dooglio  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #108 from Dooglio  ---
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.

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

hNarval  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

pasa  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 33125] Tomb Raider (2013) game freezes

http://bugs.winehq.org/show_bug.cgi?id=33125

Rui Caridade  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #109 from Rui Caridade  ---
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?

[Bug 33125] Tomb Raider (2013) freezes (bug in game engine, semaphore object handle value must be zero on all error paths)

http://bugs.winehq.org/show_bug.cgi?id=33125

Anastasius Focht  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  ---
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

[Bug 33125] Tomb Raider (2013) freezes (bug in game engine, semaphore object handle value must be zero on all error paths)

http://bugs.winehq.org/show_bug.cgi?id=33125

Anastasius Focht  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
      Fixed by SHA1|                            |149d60318fac6314726f256c1cc
                   |                            |a1e8c4806202b
             Status|NEW                         |RESOLVED
          Component|-unknown                    |kernel32
         Resolution|---                         |FIXED

--- Comment #111 from Anastasius Focht  ---
Hello folks,

this is fixed by commit
http://source.winehq.org/git/wine.git/commitdiff/149d60318fac6314726f256c1cc...

Thanks Austin.

Regards

[Bug 33125] Tomb Raider (2013) freezes (bug in game engine, semaphore object handle value must be zero on all error paths)

https://bugs.winehq.org/show_bug.cgi?id=33125

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #112 from Alexandre Julliard  ---
Closing bugs fixed in 1.7.22.

[Bug 33125] Tomb Raider (2013) freezes (bug in game engine, semaphore object handle value must be zero on all error paths)

http://bugs.winehq.org/show_bug.cgi?id=33125

[e-mail] changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #113 from [e-mail] ---
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?

[Bug 33125] Tomb Raider (2013) freezes (bug in game engine, semaphore object handle value must be zero on all error paths)

http://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #114 from Berillions  ---
(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

[Bug 33125] Tomb Raider (2013) freezes (bug in game engine, semaphore object handle value must be zero on all error paths)

https://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #115 from Jerome Leclanche  ---
(In reply to Berillions from comment #114)
According to a user it still happens on 1.7.23. Can anyone confirm?

[Bug 33125] Tomb Raider (2013) freezes (bug in game engine, semaphore object handle value must be zero on all error paths)

https://bugs.winehq.org/show_bug.cgi?id=33125

--- Comment #116 from [e-mail]  ---
I just tested it with 1.7.23 and I can access to the aim menu without freeze.