http://bugs.winehq.org/show_bug.cgi?id=22467
Summary: Natural Selection 2 crashes on start up Product: Wine Version: 1.1.43 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: epssyis@gmail.com
Created an attachment (id=27523) --> (http://bugs.winehq.org/attachment.cgi?id=27523) NS2 console output, clipped
The Natural Selection 2 Test Engine crashes when attempting to start just after sound has started to play and a black window has spawned. I have been given a key by one of the developers to debug this issue and if you require it please contact me privately.
This application did function correctly as of Wine ~1.1.33 and stopped working shortly after. Regression testing is not possible as Steam no longer starts with these older versions of Wine and NS2 crashes with the first version where they both work.
NS2 requires Steam be running in the background to function. There are two sections of note in the attached console log.
The first appeared at the same time the game stopped functioning so is most likely related:
err:d3d9:device_parent_CreateSurface (0x1280e4) CreateSurface failed, returning 0x8876017c fixme:d3d_texture:texture_init Failed to create surface 0xc4f8790, hr 0x8876017c err:d3d:resource_init Out of adapter memory
This function was defined by Henri in commit a966293f593e8b75e796a0a923626a1e73403a20 on the 16th of January which fits the time line roughly.
The second was introduced in 41 as part of a patch set by Henri to add texture compression support last month:
err:d3d:IWineD3DDeviceImpl_UpdateSurface >>>>>>>>>>>>>>>>> GL_INVALID_OPERATION (0x502) from glCompressedTexSubImage2DARB @ device.c / 5298
bb94bc382329f4c9950aad1aa6f32fe0514f1dd9 eced5fe8ec16c9bc502f293754774d1763eef004 61ca407151649dc763950414e076ea7c640afee9
http://bugs.winehq.org/show_bug.cgi?id=22467
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |hverbeet@gmail.com
--- Comment #1 from Henri Verbeet hverbeet@gmail.com 2010-04-23 11:32:56 --- Please attach a "+seh,+tid,+d3d9,+d3d,+d3d_caps,+d3d_surface,+d3d_texture" log.
(In reply to comment #0)
This function was defined by Henri in commit a966293f593e8b75e796a0a923626a1e73403a20 on the 16th of January which fits the time line roughly.
That's Jan 16 2009, not this year. The release that commit first appeared in was 1.1.13.
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #2 from edward savage epssyis@gmail.com 2010-04-23 11:51:48 --- Created an attachment (id=27525) --> (http://bugs.winehq.org/attachment.cgi?id=27525) ns2 debug output sed,tid,d3d-various
(In reply to comment #1)
Please attach a "+seh,+tid,+d3d9,+d3d,+d3d_caps,+d3d_surface,+d3d_texture" log.
Attached.
That's Jan 16 2009, not this year. The release that commit first appeared in was 1.1.13.
I made the assumption that it was 2010, my mistake.
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #3 from edward savage epssyis@gmail.com 2010-04-23 12:28:18 --- At the behest of Thunderbird I tried a much higher VideoMemorySize value and found that 1024+ cleaned up the first memory error (at least for the few minutes I kept the game open). The other error remained so I compiled wine 1.1.39 and this error vanished as well. I will do a regression test to isolate exactly which patch is causing the issue later today to see if it is those listed in the bug report.
The game does not run beyond 1 fps in 39 due to some other errors that may have already been addressed. For example:
fixme:d3d_surface:IWineD3DSurfaceImpl_LoadLocation 0x105fdfc8: Downloading rgb texture to reload it as srgb error in call_function: 'monitorPuck' is not a function
http://bugs.winehq.org/show_bug.cgi?id=22467
maxim.cournoyer@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |maxim.cournoyer@gmail.com
--- Comment #4 from maxim.cournoyer@gmail.com 2010-04-26 11:36:02 --- (In reply to comment #0)
This application did function correctly as of Wine ~1.1.33 and stopped working shortly after. Regression testing is not possible as Steam no longer starts with these older versions of Wine and NS2 crashes with the first version where they both work.
I don't understand this. I've been installing Steam on Wine 1.1.33, 1.1.34, 1.1.35 and 1.1.36 and using it with NS2 with success. I can do this regression testing, if it's needed. Using Play On Linux for it's convenience for managing WINE versions, I had already discovered that it worked from 1.1.33 (not tested with older versions) to 1.1.35 (with the VideoMemorySize set). The 1.1.35 version had some issues which I don't remember. 1.1.36 was the same as 1.1.42 and crashed instantly on game creation.
It even worked with CrossOver Games 8.1.4, which if I'm correct, uses WINE 1.1.25 as its engine. The viewer only worked correctly with 1.1.33 and 1.1.34.
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #5 from Henri Verbeet hverbeet@gmail.com 2010-04-26 11:38:31 --- (In reply to comment #4)
I don't understand this. I've been installing Steam on Wine 1.1.33, 1.1.34, 1.1.35 and 1.1.36 and using it with NS2 with success. I can do this regression testing, if it's needed.
A regression test would be useful, yes.
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #6 from Apteryx maxim.cournoyer@gmail.com 2010-04-26 14:07:00 ---
A regression test would be useful, yes.
I have been following this: http://wiki.winehq.org/RegressionTesting to make a regression test, and I think I now have a grasp of how it works, and have successfully compiled the "good" wine-1.1.34 (in the intent of testing it against 1.1.35). However, I'm now at making fresh installs of NS2 and Steam, and NS2 installer does not even want to start: ~/wine-git/wine ns2_setup_build140.exe wine: Call from 0x7b843cd2 to unimplemented function user32.dll.UserRegisterWowHandlers, aborting err:module:attach_process_dlls "user32.dll" failed to initialize, aborting err:module:LdrInitializeThunk Main exe initialization for L"[...]\ns2_setup_build140.exe" failed, status 80000100
If you know how to resolve this, I will be able to continue this regression work.
This is happenning with wine version: wine-1.1.34-317-g8b226ac. Should I just skip that one?
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #7 from Henri Verbeet hverbeet@gmail.com 2010-04-26 14:22:23 --- (In reply to comment #6)
~/wine-git/wine ns2_setup_build140.exe wine: Call from 0x7b843cd2 to unimplemented function user32.dll.UserRegisterWowHandlers, aborting err:module:attach_process_dlls "user32.dll" failed to initialize, aborting err:module:LdrInitializeThunk Main exe initialization for L"[...]\ns2_setup_build140.exe" failed, status 80000100
If you know how to resolve this, I will be able to continue this regression work.
Did you try running "make clean" before building? I think the build system may not have picked up some changes.
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #8 from Apteryx maxim.cournoyer@gmail.com 2010-04-26 18:14:42 ---
Did you try running "make clean" before building? I think the build system may not have picked up some changes.
Yes I did. I used make -j5 to use my 4 cpus, but I don't think this would make any difference. I decided to give a try to the current (master) branch, and at least I could install, so the problem might have been specific to the first builds of 1.1.33 branch. Anyway, the new Steam version has been launched and now I think I'm convinced we can't do regression testing with older WINE versions anymore.
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #9 from edward savage epssyis@gmail.com 2010-04-27 00:06:12 --- (In reply to comment #6)
~/wine-git/wine ns2_setup_build140.exe wine: Call from 0x7b843cd2 to unimplemented function user32.dll.UserRegisterWowHandlers, aborting err:module:attach_process_dlls "user32.dll" failed to initialize, aborting err:module:LdrInitializeThunk Main exe initialization for L"[...]\ns2_setup_build140.exe" failed, status 80000100
This is the reason I said we cannot test wine back that far.
The regression we are most interested in at the moment is the regression between ~39 and ~43 that stops the game being able to draw leaving a black game window. I've had a lot of university work to do so I haven't had a chance yet.
Once that regression has been identified we can start working on the performance slowdown using the most up to date wine.
http://bugs.winehq.org/show_bug.cgi?id=22467
--- Comment #10 from edward savage epssyis@gmail.com 2010-05-05 05:07:54 --- Right, using the current git you can get NS2 test engine running by setting your VideoMemorySize to an insane value. Via regedit add HKEY Current User -> Software -> Wine -> Direct3D and a key of VideoMemorySize and a value of upwards of a GB or more. NS2 will crash complaining of lack of memory if I have this value any less than ~800mb while my card only physically has 256mb.
I don't know how much luck you will have with 1.1.43 but the mmdevapi issue appears to have cleaned up some time between 41 and current git.
There are a couple of changes to Wine between the known good case that ran well (5 FPS+) and it running worse that make finding any regression really time consuming. On top of this there are many issues with the Engine Test itself. I would suggest a better approach would be to isolate bugs with the graphics only and attempt to fix those and wait for the next Engine Test release.
With the graphics bugs addressed I'm pretty confident an optimised NS2 will run rather well. I will update this bug when I've had a chance to track down the graphics errors that are being generated.
http://bugs.winehq.org/show_bug.cgi?id=22467
booiiing booiiing@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #11 from booiiing booiiing@gmail.com 2010-07-27 02:48:10 --- *** This bug has been confirmed by popular vote. ***
http://bugs.winehq.org/show_bug.cgi?id=22467
Scott Ritchie scott@open-vote.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |scott@open-vote.org
--- Comment #12 from Scott Ritchie scott@open-vote.org 2010-08-09 04:14:49 --- Is this the same issue as http://bugs.winehq.org/show_bug.cgi?id=23807 ? More modern logs are there.
http://bugs.winehq.org/show_bug.cgi?id=22467
joaopa jeremielapuree@yahoo.fr changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jeremielapuree@yahoo.fr
--- Comment #13 from joaopa jeremielapuree@yahoo.fr 2011-07-09 23:46:01 CDT --- 1 year without answer. Could be marked as ABANDONED.
http://bugs.winehq.org/show_bug.cgi?id=22467
Scott Ritchie scott@open-vote.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |ABANDONED
--- Comment #14 from Scott Ritchie scott@open-vote.org 2011-12-28 16:32:33 CST --- I don't think the engine test is available any more, and there is active testing of the actual game, so I'm going to mark this as abandoned.
http://bugs.winehq.org/show_bug.cgi?id=22467
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #15 from Austin English austinenglish@gmail.com 2012-01-23 23:55:00 CST --- Closing.
-
wine-bugs＠winehq.org