http://bugs.winehq.org/show_bug.cgi?id=27879
Summary: Zombie Driver becomes unresponsive mid-game with shadows enabled Product: Wine Version: 1.3.25 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: gyebro69@gmail.com CC: miegalius@gmail.com
Created an attachment (id=35669) --> (http://bugs.winehq.org/attachment.cgi?id=35669) terminal output
In Wine-1.3.25 Zombie Driver stalls in a matter of seconds after a new game was started.
Symptom: sound begins to stutter and framerate drops dramatically in a couple of seconds mid-game, the game hardly responses to keypresses. The game doesn't turn into "zombie" completely, if I press <Esc> several times I'm able to go back to the menus where the framerate and mouse responsiveness are okay. As soon as I resume the game, the issue is there again.
Workaround: disable shadows (in the launcher). The game offers 3 choices for shadows: none, normal, soft.
I can reproduce the problem with the following game versions: 1.1.2, 1.1.4, 1.2.2, 1.2.6 The demo, still at version 1.0.6 is not affected.
Regression testing pointed to:
6c5e109af7184ee557aa24c676cc78f3e92e6782 is the first bad commit commit 6c5e109af7184ee557aa24c676cc78f3e92e6782 Author: Ričardas Barkauskas rbarkauskas@codeweavers.com Date: Fri Jul 8 22:11:29 2011 +0300
wined3d: Introduce quirk for proper RGBA16 support.
:040000 040000 1fc08fe5e9148a2a9699c6c6edb6c042192a5072 4e9e3555ccd4a5892a60c2a4aff1d28ea2229a9e M dlls
The patch can be reverted cleanly on 1.3.25 and that fixed the problem. Author of the patch added to CC.
I'd like to add that this was the commit that fixed a long-standing graphical problem in the game: dark, diagonal lines appeared across the screen when shadows was enabled.
Fedora 15 Nvidia GeForce 250 / driver 275.21
http://bugs.winehq.org/show_bug.cgi?id=27879
GyB gyebro69@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |regression
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #1 from GyB gyebro69@gmail.com 2011-07-24 00:58:33 CDT --- I forgot to say that the following native components were installed by the game: vcrun2008, Physx, Openal, d3dx9 libraries.
http://bugs.winehq.org/show_bug.cgi?id=27879
GyB gyebro69@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1| |6c5e109af7184ee557aa24c676c | |c78f3e92e6782
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #2 from GyB gyebro69@gmail.com 2012-01-13 23:55:37 CST --- Still present in Wine-1.3.37 and with Nvidia driver 290.10.
The demo available on Steam is the same version as the full game and the problem can be reproduced with that: http://store.steampowered.com/app/31410/
Steps to reproduce the problem with the demo (Steam): 1. Install the demo via Steam. The game needs certain native components (see comment #1). You can install them via winetricks or let the Steam installscript do the job. 2. Set openal32.dll to native (bug #20799 ?) 3. In the launcher make sure that shadows are enabled. 4. Start a new game and begin driving. If you drive straight ahead you will soon cross some railway tracks and you can notice that the game begins to slow down noticeably. At some point the game may hang completely.
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #3 from Henri Verbeet hverbeet@gmail.com 2012-02-20 13:10:35 CST --- Created attachment 38991 --> http://bugs.winehq.org/attachment.cgi?id=38991 patch
It looks like the game enables blending on WINED3DFMT_R16G16B16A16_UNORM render targets, and the driver hits a software callback. The attached patch helps for me. I'm not sure if there's a good way to detect hardware format blending capabilities.
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #4 from GyB gyebro69@gmail.com 2012-02-21 13:57:27 CST --- (In reply to comment #3)
Created attachment 38991 [details] patch
It looks like the game enables blending on WINED3DFMT_R16G16B16A16_UNORM render targets, and the driver hits a software callback. The attached patch helps for me. I'm not sure if there's a good way to detect hardware format blending capabilities.
Thanks, the patch fixes the problem here as well.
http://bugs.winehq.org/show_bug.cgi?id=27879
Bruno Jesus 00cpxxx@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |patch CC| |00cpxxx@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=27879
GyB gyebro69@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jorgitorrr@gmail.com
--- Comment #5 from GyB gyebro69@gmail.com 2012-03-17 15:23:56 CDT --- *** Bug 29517 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=27879
Mateusz Stachowski mateusz.stachowski@wp.pl changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |mateusz.stachowski@wp.pl
--- Comment #6 from Mateusz Stachowski mateusz.stachowski@wp.pl 2012-05-27 04:01:18 CDT --- The patch attached by Henri Verbeet is also necessary for the Hard Reset. Without that patch I can't get that game to run it freezes on loading screens before reaching to menu (tested on wine-1.5.5).
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #7 from Bruno Jesus 00cpxxx@gmail.com 2013-01-20 07:25:49 CST --- It looks like the patch was not commited. Can anyone update the status of this bug?
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #8 from GyB gyebro69@gmail.com 2013-01-20 07:38:13 CST --- (In reply to comment #7)
It looks like the patch was not commited. Can anyone update the status of this bug?
Yep, it's still present in Wine 1.5.22. Nvidia binary drivers 304.64. Orm=backbuffer is a workaround to the slowdown: but shadows are not rendered correctly with that setting (basically it's the same as before the regression, dark, diagonal lines appear instead of shadows, but game speed is okay).
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #9 from Henri Verbeet hverbeet@gmail.com 2013-01-20 14:52:27 CST --- (In reply to comment #7)
It looks like the patch was not commited. Can anyone update the status of this bug?
Yes, that patch is just a hack. We probably need the ARB_interformat_query2 extension to properly fix this.
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #10 from Matteo Bruni matteo.mystral@gmail.com 2013-02-17 11:20:52 CST --- Created attachment 43595 --> http://bugs.winehq.org/attachment.cgi?id=43595 patch1
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #11 from Matteo Bruni matteo.mystral@gmail.com 2013-02-17 11:25:52 CST --- Created attachment 43596 --> http://bugs.winehq.org/attachment.cgi?id=43596 patch2
Can you please try the patches I just attached? You need ARB_internalformat_query2 support for these patches to work, which means Nvidia drivers 310.xx or newer IIRC.
If these patches don't help, can you please attach a +d3d log? If it gets too big to attach, the first few thousand lines are probably enough.
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #12 from GyB gyebro69@gmail.com 2013-02-17 12:24:28 CST --- Created attachment 43597 --> http://bugs.winehq.org/attachment.cgi?id=43597 7zipped +d3d log (uncompressed 250 MB)
The patches don't seem to help at all, tested with Nvidia binary drivers 313.18
http://bugs.winehq.org/show_bug.cgi?id=27879
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |matteo.mystral@gmail.com
--- Comment #13 from Matteo Bruni matteo.mystral@gmail.com 2013-02-17 13:13:10 CST --- Maybe my patch is wrong (extra sets of eyes are appreciated), but I find it suspect that every texture format, including those not supporting FBO attachment, are claimed to support blending. ARB_internalformat_query2 spec text pretty much implies that FRAMEBUFFER_BLEND should be a stricter requirement than FRAMEBUFFER_RENDERABLE.
So (guess what) I'm blaming the drivers. I'd be happy to be proved wrong.
http://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #14 from Henri Verbeet hverbeet@gmail.com 2013-04-04 06:04:01 CDT --- (In reply to comment #13)
Maybe my patch is wrong (extra sets of eyes are appreciated), but I find it suspect that every texture format, including those not supporting FBO attachment, are claimed to support blending. ARB_internalformat_query2 spec text pretty much implies that FRAMEBUFFER_BLEND should be a stricter requirement than FRAMEBUFFER_RENDERABLE.
So (guess what) I'm blaming the drivers. I'd be happy to be proved wrong.
I can confirm GL_RGBA16 returns FULL_SUPPORT for FRAMEBUFFER_BLEND on my GF9600M with 313.26. I can also confirm that Nouveau lists the format as supporting blending only on Fermi and later at http://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/nv50/nv50_for...
So yeah, I'd consider that a bug in nvidia's internalformat_query2 implementation. I think we'll still want internalformat_query2 support, but once that's in this will probably be resolved UPSTREAM.
http://bugs.winehq.org/show_bug.cgi?id=27879
Henri Verbeet hverbeet@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |UPSTREAM
--- Comment #15 from Henri Verbeet hverbeet@gmail.com 2013-04-23 04:55:29 CDT --- We now have support for internalformat_query2, since commit 6fa87980caabdeebd0a396ec458b7262f65be696. In principle it's now up to NVIDIA to correctly advertise GL_RGBA16 hardware limitations.
https://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #16 from Béla Gyebrószki gyebro69@gmail.com --- I still have this issue in Zombie Driver as well as in Hard Reset (bug #29517). Hard Reset hangs during the first intro video just as described in #29517. I have Nvidia binary drivers 340.76 installed.
Others with Nvidia cards + binary drivers seem to be able to start Hard Reset -> bug #35876
Does this bug affect only certain Nvidia cards or what? Mine is a venerable GeForce 250 GTS.
Wine 1.7.43
https://bugs.winehq.org/show_bug.cgi?id=27879
--- Comment #17 from Matteo Bruni matteo.mystral@gmail.com --- (In reply to Béla Gyebrószki from comment #16)
Does this bug affect only certain Nvidia cards or what? Mine is a venerable GeForce 250 GTS.
Yes, I think this only happens on NV50 (Tesla)-generation GPUs, as Henri mentioned in comment 14.
I guess the issue was never reported to Nvidia and I don't think they care now that they dropped support for pre-Fermi (GeForce 400+) GPUs in their current drivers.
Maybe you tested that already but does the same issue happen with the Mesa (nouveau) driver?
https://bugs.winehq.org/show_bug.cgi?id=27879
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #18 from Austin English austinenglish@gmail.com --- Closing.
https://bugs.winehq.org/show_bug.cgi?id=27879
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|CLOSED |RESOLVED
--- Comment #19 from Austin English austinenglish@gmail.com --- This was inadvertently caught up in my unclosed bugs filter. NOTOURBUG should only be closed when fixed upstream.
Setting back to RESOLVED NOTOURBUG.
Sorry for the spam.