[Bug 20053] New: wined3d - TF2 suffers from black menu and missing UI textures in dx9 mode
http://bugs.winehq.org/show_bug.cgi?id=20053
Summary: wined3d - TF2 suffers from black menu and missing UI textures in dx9 mode Product: Wine Version: 1.1.29 Platform: PC OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: directx-d3d AssignedTo: wine-bugs@winehq.org ReportedBy: Nephyrin@nephyrin.net
When starting TF2 in directx9 mode (that is, starting it normally, out of the box), the menu is black (no background image) and there are no HUD elements in game. Other than this, it appears to work fine and looks great in Dx9 mode, and would be platinum status if you could see your UI.
Disabling pixelshaders in winecfg causes the game to load with a visible background on the main menu, but the game immediately bails telling you your graphics card is not able to run the game.
-dxlevel 81 forces the game into dx8 mode where everything works, but this is a major visual quality regression.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #1 from Nephyrin zey Nephyrin@nephyrin.net 2009-09-14 16:32:53 --- Created an attachment (id=23598) --> (http://bugs.winehq.org/attachment.cgi?id=23598) Screenshot showing rendering working properly, but blank and black UI elements
http://bugs.winehq.org/show_bug.cgi?id=20053
Nephyrin zey Nephyrin@nephyrin.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |Nephyrin@nephyrin.net
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #2 from Nephyrin zey Nephyrin@nephyrin.net 2009-09-14 17:40:40 --- Possibly relevant d3d output
fixme:d3d:IWineD3DDeviceImpl_SetRenderTarget (0x14b760)Trying to set the render target to a surface(0x1826d218) that wasn't created with a usage of WINED3DUSAGE_RENDERTARGET fixme:d3d:IWineD3DDeviceImpl_SetRenderTarget (0x14b760)Trying to set the render target to a surface(0x181201c0) that wasn't created with a usage of WINED3DUSAGE_RENDERTARGET
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #3 from Vitaliy Margolen vitaliy@kievinfo.com 2009-09-14 20:58:40 --- What' video card and driver version? Do you have 3d desktop effects disabled?
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #4 from Nephyrin zey Nephyrin@nephyrin.net 2009-09-14 21:08:51 --- I'm using nvidia 190.32 with a GTX295 at the moment, but this issue has existed for me for a long time, at least since 185.* drivers. I'll try some older drivers/wine to see if i can pinpoint a regression, but I'm not sure its ever worked in recent memory.
I've tried different settings, windows versions, an isolated wine prefix, disabling compiz entirely, and virtual desktop mode. None seem to affect it. This issue also occurs on my other system using an 8800GTX.
The appdb page comments has many people referencing dx9 mode simply not working, and advises everyone to use dxlevel 81. However, in my testing, dx9 mode works great and gives me 20-30 FPS on high settings. multisampling being broken (in dx8 mode as well though) and this bug are the only remaining issues. If this bug were fixed TF2 would work very well out of the box. (the current platinum rating is bogus, even the reviewer mentions he used dx8).
I should also mention, if i force multisampling in the nvidia-settings utility it does give me multisampling, but then i get only a first person view (my hand and gun) and the world is black. It really seems like offscreen rendering is just borked.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #5 from Henri Verbeet hverbeet@gmail.com 2009-09-15 02:03:38 --- The problem is that recent versions of the nvidia drivers generate NaN instead of infinity on division be zero. I think the spec allows that, but I'm not sure if it's intentional. 180.60 should still work with TF2.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #6 from Nephyrin zey Nephyrin@nephyrin.net 2009-09-17 16:55:56 --- Assuming it to be intentional, what would the recourse be? Is there one or two spots in the d3d API where it would need to handle and translate this, or is it a much more involved change?
I would be willing to look into hacking on it, but I haven't worked with wine much and have little idea where to start
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #7 from Henri Verbeet hverbeet@gmail.com 2009-09-18 03:17:12 --- Created an attachment (id=23642) --> (http://bugs.winehq.org/attachment.cgi?id=23642) hack
Something like the attached hack should work for TF2. It isn't entirely correct though.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #8 from Nephyrin zey Nephyrin@nephyrin.net 2009-09-18 19:21:12 --- This patch does indeed let TF2 work in directx9 mode out of the box. In fact, it runs really well. The multisampling issue remains, and looks like its actually a FBO interaction issue. Using backbuffer offscreen rendering allowed me to use full DX9 max settings with multisampling at ~20FPS, which is pretty good, though I get about 120 in windows. I can get about 30 if i use FBO and no multisampling. Few other game settings appear to affect it, it looks like its CPU bound.
http://www.infernalsoul.net/Junk/Dump/tf1.png http://www.infernalsoul.net/Junk/Dump/tf2.png http://www.infernalsoul.net/Junk/Dump/tf3.png
Wine gets better by the day!
If you can point me in the direction of what would need to be done to make this a proper patch, I can look into it.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #9 from Henri Verbeet hverbeet@gmail.com 2009-09-19 13:41:59 --- (In reply to comment #8)
If you can point me in the direction of what would need to be done to make this a proper patch, I can look into it.
A proper patch would generate +INF instead of NaN for rcp 0. My preferred way of fixing that would be by giving us some control over which floating point specials are generated, but I don't think the current GLSL spec has anything to allow that.
http://bugs.winehq.org/show_bug.cgi?id=20053
msn@gaiatools.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW Ever Confirmed|0 |1
--- Comment #10 from msn@gaiatools.com 2009-10-07 23:23:12 --- *** This bug has been confirmed by popular vote. ***
http://bugs.winehq.org/show_bug.cgi?id=20053
Karsten Elfenbein kelfe@gmx.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |kelfe@gmx.de
http://bugs.winehq.org/show_bug.cgi?id=20053
msn@gaiatools.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |msn@gaiatools.com
http://bugs.winehq.org/show_bug.cgi?id=20053
Jan Kalab pitel@nomi.cz changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |pitel@nomi.cz
--- Comment #11 from Jan Kalab pitel@nomi.cz 2009-11-16 05:52:10 --- I just got the same problem. I was ok When I used Ubuntu 9.04, but after upgrade to 9.10, this bug appeared.
http://bugs.winehq.org/show_bug.cgi?id=20053
Pavel Valach valach.pavel@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |valach.pavel@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=20053
byteframe byteframe@yahoo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |byteframe@yahoo.com
http://bugs.winehq.org/show_bug.cgi?id=20053
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|wined3d - TF2 suffers from |Team-Fortress 2 and other |black menu and missing UI |source games suffer from |textures in dx9 mode |black menu and missing UI | |textures in dx9 mode
http://bugs.winehq.org/show_bug.cgi?id=20053
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |webmobster@gmail.com
--- Comment #12 from Vitaliy Margolen vitaliy@kievinfo.com 2009-12-20 12:52:25 --- *** Bug 21093 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=20053
mokrad mokrad@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |mokrad@gmail.com
--- Comment #13 from mokrad mokrad@gmail.com 2010-01-03 03:31:15 --- Hi, Do you have any new about this bug ? I known that it's possible to play on TF2 using -dxlevel 90 mode with enough fps but we have black menu and missing UI with this mode. For me, solving this bug is the only thing to add in wine to put a platinum rating at TF2.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #14 from byteframe byteframe@yahoo.com 2010-01-03 03:53:23 --- (In reply to comment #13)
Hi, Do you have any new about this bug ? I known that it's possible to play on TF2 using -dxlevel 90 mode with enough fps but we have black menu and missing UI with this mode. For me, solving this bug is the only thing to add in wine to put a platinum rating at TF2.
You will have to use nvidia drivers 180.60 at the present time to avoid the black background problem.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #15 from Nephyrin zey Nephyrin@nephyrin.net 2010-01-03 13:15:23 --- (In reply to comment #14)
(In reply to comment #13)
Hi, Do you have any new about this bug ? I known that it's possible to play on TF2 using -dxlevel 90 mode with enough fps but we have black menu and missing UI with this mode. For me, solving this bug is the only thing to add in wine to put a platinum rating at TF2.
You will have to use nvidia drivers 180.60 at the present time to avoid the black background problem.
With the hack posted by henri, it runs flawlessly on newer drivers, if you're capable of compiling wine yourself. I might add a PKGBUILD with this fix to AUR if you're on Arch
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #16 from mokrad mokrad@gmail.com 2010-01-04 06:18:39 --- 1 . To reply to byteframe : could you explain to me why a older driver is better than a newer ?
2. To reply to Nephyrin zey : I compiled wine 1.1.35 with the hack. then TF2 seems to run fine (with UI) in dxlevel 90 mode. (around 25-30 fps with screen :1920*1200). Could you explain why this hack is not include directly in wine ?
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #17 from Nephyrin zey Nephyrin@nephyrin.net 2010-01-04 12:10:05 --- (In reply to comment #16)
1 . To reply to byteframe : could you explain to me why a older driver is better than a newer ?
- To reply to Nephyrin zey : I compiled wine 1.1.35 with the hack. then TF2
seems to run fine (with UI) in dxlevel 90 mode. (around 25-30 fps with screen :1920*1200). Could you explain why this hack is not include directly in wine ?
As far as I understand it: Drivers up to 180.60 returned INF on divide by zero in shaders. Newer ones return NaN, which directx does not expect, causing issues. This hack changes the particular offending shader path to check for div by zero and make sure it generates INF. It does not fix the problem in general, just masks it, hence, a hack.
I'm not sure what the proper way to fix this would be, my guess is it would require going through and ensuring ambiguous divide by zeros dont occur in all directx -> glsl situations, but how feasible that is is beyond me. Henri would probably be the one to decide how to go about properly remedying it, if at all.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #18 from Henri Verbeet hverbeet@gmail.com 2010-01-04 12:39:16 --- (In reply to comment #17)
As far as I understand it: Drivers up to 180.60 returned INF on divide by zero in shaders. Newer ones return NaN, which directx does not expect, causing issues. This hack changes the particular offending shader path to check for div by zero and make sure it generates INF. It does not fix the problem in general, just masks it, hence, a hack.
The problem with the hack is mostly that it returns 0.0 instead of +INF. The problem with writing a proper fix is that I don't know a practical/reliable way to generate +INF in GLSL.
For what it's worth, the main difference between NaN and +INF that TF2 cares about is that NaN keeps propagating while e.g. "clamp(+INF, 0.0, 1.0);" returns 1.0.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #19 from mokrad mokrad@gmail.com 2010-01-04 13:34:31 --- It's not very clear for me too much technical. But thanks to make your best in wine development. I hope you will find the solution. The only think I can do to help you is to test some path or hack.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #20 from Nephyrin zey Nephyrin@nephyrin.net 2010-01-04 14:25:20 --- (In reply to comment #18)
(In reply to comment #17)
As far as I understand it: Drivers up to 180.60 returned INF on divide by zero in shaders. Newer ones return NaN, which directx does not expect, causing issues. This hack changes the particular offending shader path to check for div by zero and make sure it generates INF. It does not fix the problem in general, just masks it, hence, a hack.
The problem with the hack is mostly that it returns 0.0 instead of +INF. The problem with writing a proper fix is that I don't know a practical/reliable way to generate +INF in GLSL.
For what it's worth, the main difference between NaN and +INF that TF2 cares about is that NaN keeps propagating while e.g. "clamp(+INF, 0.0, 1.0);" returns 1.0.
Perhaps wine should do some tests on the implementation at startup to determine the behavior of undefineds like div 0 and log 0, and assuming at least one +INF generating special is available, store it in a constant? Or is this way too hacky as well? It seems like implementing dx shaders with a defined floating point model on GLSL without one is basically not possible without manually resolving any ambiguous operations.
http://bugs.winehq.org/show_bug.cgi?id=20053
William wwarsin@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |wwarsin@gmail.com
http://bugs.winehq.org/show_bug.cgi?id=20053
JP St. Pierre jstpierre@mecheye.net changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jstpierre@mecheye.net
--- Comment #21 from JP St. Pierre jstpierre@mecheye.net 2010-01-11 17:40:02 --- Hmm... I'm not sure if I should file a new bug, but I'm seeing the black textures and missing HUD even with the patch applied and -dxlevel 80 or -dxlevel 81. Any info you want I'll be able to give. I'm using latest nVidia too, but it's quite an old card (GeForce 6200). Arch Linux, wine built from git yesterday.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #22 from Nephyrin zey Nephyrin@nephyrin.net 2010-01-11 17:43:13 --- (In reply to comment #21)
Hmm... I'm not sure if I should file a new bug, but I'm seeing the black textures and missing HUD even with the patch applied and -dxlevel 80 or -dxlevel 81. Any info you want I'll be able to give. I'm using latest nVidia too, but it's quite an old card (GeForce 6200). Arch Linux, wine built from git yesterday.
The patch shouldn't be necessary with dxlevel 80/1. If you're seeing black still its a separate bug, make sure you dont have multisampling+fbo enabled, as that will cause a unrelated issue. (use multisampling + backbuffer instead, or no multisampling)
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #23 from JP St. Pierre jstpierre@mecheye.net 2010-01-11 18:04:07 --- I'm not using multisampling, but I am using fbo. I'll try next using multisampling+backbuffer.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #24 from JP St. Pierre jstpierre@mecheye.net 2010-01-11 18:08:12 --- And using multisampling+backbuffer makes no difference.
I'm always in #winehq on Freenode as "magcius", if you want to reach me there.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #25 from Henri Verbeet hverbeet@gmail.com 2010-02-01 09:30:07 --- Should be fixed by ef82681d59ec264b981bb3653ea774044f02fa9c.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #26 from mokrad mokrad@gmail.com 2010-02-01 14:22:21 --- Created an attachment (id=26011) --> (http://bugs.winehq.org/attachment.cgi?id=26011) this menu should not be transparent
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #27 from mokrad mokrad@gmail.com 2010-02-01 14:23:35 --- Created an attachment (id=26012) --> (http://bugs.winehq.org/attachment.cgi?id=26012) the doors are not draw
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #28 from mokrad mokrad@gmail.com 2010-02-01 14:26:10 --- (In reply to comment #25)
Should be fixed by ef82681d59ec264b981bb3653ea774044f02fa9c.
i think it's semi fixed because the UI is not missing. But there is others graphical problems. See the 2 previous attached screenshot.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #29 from Henri Verbeet hverbeet@gmail.com 2010-02-02 05:40:14 --- (In reply to comment #28)
(In reply to comment #25)
Should be fixed by ef82681d59ec264b981bb3653ea774044f02fa9c.
i think it's semi fixed because the UI is not missing. But there is others graphical problems. See the 2 previous attached screenshot.
Did the hack attached to this bug fix those, or is this a different issue?
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #30 from mokrad mokrad@gmail.com 2010-02-02 05:46:14 --- (In reply to comment #29)
(In reply to comment #28)
(In reply to comment #25)
Should be fixed by ef82681d59ec264b981bb3653ea774044f02fa9c.
i think it's semi fixed because the UI is not missing. But there is others graphical problems. See the 2 previous attached screenshot.
Did the hack attached to this bug fix those, or is this a different issue?
This hack seems to fix the inital bug. But It create a different issue (see Created an attachment (id=26012) --> (http://bugs.winehq.org/attachment.cgi?id=26012) and Created an attachment (id=26011) --> (http://bugs.winehq.org/attachment.cgi?id=26011))
http://bugs.winehq.org/show_bug.cgi?id=20053
Daniel Harvison sirbubbles01@yahoo.com.au changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |sirbubbles01@yahoo.com.au
--- Comment #31 from Daniel Harvison sirbubbles01@yahoo.com.au 2010-02-06 05:08:42 --- I'm using wine 1.1.38 and nvidia drivers 195.36.03 from the website. Team Fortress 2 and Half Life 2 Episode 2 both appear to be working properly now. If anyone could verify with wine 1.1.38 and various nvidia drivers, that would confirm this bug is fixed beyond doubt.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #32 from msn@gaiatools.com 2010-02-06 21:36:40 --- Sorry, but it is still not fixed for me... Using wine 1.1.38 and nvidia drivers v195, the HUD appears on screen during the first few seconds, then disappears (after like five seconds of walking around).
It's still an improvement though!
I am using vanilla Wine built from source today + the winepulse patches (but I unchecked all audio drivers in winecfg while testing)
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #33 from Daniel Harvison sirbubbles01@yahoo.com.au 2010-02-07 04:57:33 --- Try the drivers from these links. If you need any help installing official nvidia binary drivers, I can provide instructions.
http://www.nvnews.net/vbulletin/showthread.php?p=2178512
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #34 from mokrad mokrad@gmail.com 2010-02-07 05:13:41 --- (In reply to comment #31)
I'm using wine 1.1.38 and nvidia drivers 195.36.03 from the website. Team Fortress 2 and Half Life 2 Episode 2 both appear to be working properly now. If anyone could verify with wine 1.1.38 and various nvidia drivers, that would confirm this bug is fixed beyond doubt.
No change for me too. (wine 1.1.38 and nvidia drivers 195.36.03, ubuntu 9.10). 2 big problem: - It create a different issue (see Created an attachment (id=26012) --> (http://bugs.winehq.org/attachment.cgi?id=26012) [details] and Created an attachment (id=26011) --> (http://bugs.winehq.org/attachment.cgi?id=26011) [details]) - This bug is always very critical (dxlevel 81 and dxlevel 90) http://bugs.winehq.org/show_bug.cgi?id=21534
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #35 from Daniel Harvison sirbubbles01@yahoo.com.au 2010-02-09 10:16:40 --- I have half life 2 episode 2 running mostly well with UseGLSL set to enabled in the wine registry. However, now there's another issue with strange graphical anomalies showing up all over the place.
http://bugs.winehq.org/show_bug.cgi?id=20053
edward savage epssyis@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |epssyis@gmail.com
--- Comment #36 from edward savage epssyis@gmail.com 2010-05-03 04:27:53 --- A default install of TF2 worked 'out of the box' for me with a fresh Ubuntu and Steam install with only 'surface' water effects not being rendered correctly. I have had absolutely no menu or UI problems. Frame rate was actually acceptable and I played for a couple of hours. I haven't bothered to set Directx back to 8.1.
Bumping all graphics settings to full and enabling Shadows and HDR works how ever the game runs at about 10 FPS for me. No notable graphical artifacts though I didn't find any water.
I do have the ClientScheme.res workaround in place and MOTD disabled via the in-game options.
TF2 run with launch options -width 1680 -height 1050 -novid using a Virtual Desktop of 1680x1050 on Ubuntu 10.04 and an nVidia 9600GT with 195.36.15 drivers.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #37 from Nephyrin zey Nephyrin@nephyrin.net 2010-05-03 05:11:39 --- The issue for this bug, missing UI caused by the nvidia/GLSL divide-by-zero issue, is fixed - further issues should probably be put in other bugs, or this bug should be renamed to be the TF2 meta-bug.
TF2 now runs out of the box on modern nvidia drivers without issue. The game is not quite bug free yet, but works amazingly well for a modern game on my system.
Here's some screenshots i just took: http://www.infernalsoul.net/Junk/Dump/tf2-1.png http://www.infernalsoul.net/Junk/Dump/tf2-2.png http://www.infernalsoul.net/Junk/Dump/tf2-3.png
If you enable multisampling, you have to also enable 'backbuffer' for OffscreenRenderingMode (at least on nvidia), but this is a known bug (and is listed on UsefulRegistryKeys as one of the reasons for multisampling being off by default).
Outstanding TF2 issues i noticed playing with it: - The first time you start it, you need to use something like: "-windowed -height 1200 -width 1920 -noborder" to set it to a good 'fullscreen windowed' configuration. Once you've done this, hit 'apply' in TF2 settings, and they'll be saved to the config (no need to do it multiple times). This might be necessary for me because it tries to go to 3600x1200 resolution (dual head). - Some basic sprite/halos around lights clip through walls. You can see this in my screenshots above. - The biggest issue - the first game you join will stutter madly for upwards of two minutes before becoming smooth, while the console complaints about expired RtlpWaitForCriticalSection timeouts. After it stops doing this, it will still stutter every now and then (seemingly when loading new textures... maybe) - HDR/bloom cannot be turned on, or your screen will be 100% white on most maps. This is a regression (though might be due to a driver upgrade)
Other than those caveats, it works out of the box on max settings. I get an average of 15-20FPS in a 32slot server, whereas in windows I get ~120. Not great, but for such a good looking game running in DX9, still pretty impressive.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #38 from edward savage epssyis@gmail.com 2010-05-03 05:24:48 --- (In reply to comment #37)
The issue for this bug, missing UI caused by the nvidia/GLSL divide-by-zero issue, is fixed - further issues should probably be put in other bugs, or this bug should be renamed to be the TF2 meta-bug.
Agreed.
Outstanding TF2 issues i noticed playing with it:
- The biggest issue - the first game you join will stutter madly for upwards
of two minutes before becoming smooth, while the console complaints about expired RtlpWaitForCriticalSection timeouts. After it stops doing this, it will still stutter every now and then (seemingly when loading new textures... maybe)
I don't think this is a universal problem? I know three TF2 players who do not experience it including myself. We have MOTD disabled how ever if that has some thing to do with it?
- HDR/bloom cannot be turned on, or your screen will be 100% white on most
maps. This is a regression (though might be due to a driver upgrade)
I have not tried enabling Bloom how ever HDR does not effect my ability to play the game in this way (I didn't check to see how it effected FPS or if it actually did anything).
I'd like to add to your detailed list: - Enabling Steam Cloud resets TF2 configuration options (graphics, voice, bindings) to fresh install defaults after every launch. Disabling it stops this. - Alt-tabbing causes the game to crash in a repeatable fashion. While this is a Windows bug it isn't consistent under Windows while it is under Wine. Using a Virtual Desktop works around this issue.
http://bugs.winehq.org/show_bug.cgi?id=20053
DL taedium_vitae@eml.cc changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |taedium_vitae@eml.cc
--- Comment #39 from DL taedium_vitae@eml.cc 2010-05-07 01:57:34 --- I'm still having missing HUD issues with Half-Life 2 : Episode Two and wine-1.1.43. The game starts off with the HUD working fine, but at a certain point early in the game, it just stops working. Sometimes it will start working again; it seems to be quite intermittent. I'm not sure it's the same bug, or something new, though.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #40 from William wwarsin@gmail.com 2010-05-09 23:13:40 --- I am currently not having any HUD/UI problems in CS:S or Garry's mod, with: -dxlevel 95 Wine version 1.1.42 ATI radeon hd 4770 (ATI drivers with Catalyst)
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #41 from DL taedium_vitae@eml.cc 2010-05-10 00:33:16 --- It just seems to be an issue with HL2 : EP2; it doesn't occur with any of the other HL2 games, nor does it occur with portal. I'll have to get around to installing TF2 and CS:S to see if it is EP2 specific.
I tested EP2 in native windows as well, BTW, and it works fine. So it's not a corrupted installation.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #42 from byteframe byteframe@gmail.com 2010-05-10 01:18:30 --- It's been a release or two since I checked, and I haven't been following this bug very well, but at the moment wine-1.1.44 on nvidia-195.36.15 is fixed for me! I can recall the problems as shown in the attachments, and they are all gone now.
http://bugs.winehq.org/show_bug.cgi?id=20053
Vitaliy Margolen vitaliy@kievinfo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |i386dx@hotmail.com
--- Comment #43 from Vitaliy Margolen vitaliy@kievinfo.com 2010-05-10 21:59:28 --- *** Bug 22637 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #44 from Daniel Harvison sirbubbles01@yahoo.com.au 2010-05-14 07:15:54 --- not resolved so far as hl2 episode 2 is concerned. There's a weird graphical glitch where shining a torch/flashlight is concerned, and some graphical effects, such as a glowing blue device worn by Alyx is not rendered correctly. Apart from that, it seems to work okay, as long as you're using GLSL shaders (UseGLSL=enabled).
http://bugs.winehq.org/show_bug.cgi?id=20053
quaker quaker66@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |quaker66@gmail.com
--- Comment #45 from quaker quaker66@gmail.com 2010-08-09 18:23:47 --- tried with 1.3git and episode two from steam. On places where it was unfixed it's fixed now, HUD is always showing, played several levels. So at least for Episode Two, I can confirm this to be fixed :)
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #46 from byteframe byteframe@gmail.com 2010-08-22 22:47:06 --- For TF2 this seems fixed, but a few other source game/mods still have this problem.
Starting Resistance and Liberation on dxlevel 95 (default d3d9 mode) has the black background and invisible text issues. Also probably related is the whole of Zombie Panic: Source turning black when you play as a zombie.
http://www.resistanceandliberation.com/ http://store.steampowered.com/app/17500/
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #47 from byteframe byteframe@gmail.com 2010-09-03 18:08:34 CDT --- To follow up on zombie panic source, I'm attaching two sets of output. One is for starting the game via dxlevel 81 (older mode that works), and the other is for dxlevel 95 which has the black menu background, black zombie vision (whole screen goes black when you're a zombie) missing crosshair, (some) missing hud text, missing server browser text, etc.
This is on wine-1.3.2, nvidia-256.35. A quick glance suggests that the multiple instances of "fixme:d3d:IWineD3DDeviceImpl_SetRenderTarget Surface 0xXXXXX doesn't have render target usage." in the dxlevel 95 log might be important.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #48 from byteframe byteframe@gmail.com 2010-09-03 18:10:04 CDT --- Created an attachment (id=30538) --> (http://bugs.winehq.org/attachment.cgi?id=30538) Output from starting zps in dxlevel95, in which black screen appears
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #49 from byteframe byteframe@gmail.com 2010-09-03 18:11:00 CDT --- Created an attachment (id=30539) --> (http://bugs.winehq.org/attachment.cgi?id=30539) Output from zombie panic in dxlevel 81, in which everything is fine
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #50 from Nephyrin zey Nephyrin@nephyrin.net 2010-09-03 20:46:20 CDT --- The original bug, having to do with nvidia, was fixed ages ago - but various black-screen issues exist in source games.
TF2 no longer broken, but some settings + maps (draw cart outline + pl_thundermountain) still are mostly black.
Transparent effects in source engine games seem to be blended with black instead of the world. (invisible spies in TF2 are black outlines) Turning on backbuffer mode fixes this, but seems to kill framerates whenever these transparency effects are on screen.
http://bugs.winehq.org/show_bug.cgi?id=20053
hiroe ashadocat@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ashadocat@gmail.com
--- Comment #51 from hiroe ashadocat@gmail.com 2011-01-03 00:27:19 CST --- This bug is in full affect in dystopia and several of the implants don't render properly in dx8 mode. Dystopia's a free source mod so it should be easy enough to reproduce. This appears to be a regression as I remember it working fine in dx 9 mode a couple of months ago, I upgrade and it's broken.
http://bugs.winehq.org/show_bug.cgi?id=20053
Patrick Rudolph patrick1804@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |patrick1804@web.de
--- Comment #52 from Patrick Rudolph patrick1804@web.de 2011-10-13 02:10:16 CDT --- Hi I investigated some time and found a fix. I noticed that the game was running without errors inside my Virtual Machine (Virtualbox). Virtualbox does not provide as much GL calls as my host does.
I added gl_info->supported[ARB_SHADING_LANGUAGE_100] = FALSE; gl_info->supported[NV_FRAGMENT_PROGRAM2] = FALSE; to line 2385 in wined3d:directx.c:wined3d_adapter_init_gl_caps. This disables GLSL and SM 3.0 using ARB.
But this is not a good solution, because GLSL isn't working and SM 3.0 too.
Next I tried was so force PixelShader 2.0 instead of PixelShader 3.0. This can be done by if (caps->PixelShaderVersion == WINED3DPS_VERSION(3,0)) caps->PixelShaderVersion = WINED3DPS_VERSION(2,0); in line 4750 in wined3d:directx.c:wined3d_get_device_caps
This removed all graphical problems, but PixelShader 3.0 won't work.
Then I found the problem: caps->PS20Caps.NumInstructionSlots = WINED3DPS20_MAX_NUMINSTRUCTIONSLOTS; has to be changed to caps->PS20Caps.NumInstructionSlots = WINED3DPS20_MAX_NUMINSTRUCTIONSLOTS-1; in line 4813 in wined3d:directx.c:wined3d_get_device_caps.
This is a hack, MSDN says: SM3.0 has to support WINED3DPS20_MAX_NUMINSTRUCTIONSLOTS NumInstructionSlots.
I'm not sure what this value is good for, I guess some PS instruction queue length ?
With this setting I have no graphical problems, performance is quite well ~70fps with MSAA off and AF off.
Tested on: 2.6.38-11-generic x86_64 wine1.3.30 nvidia driver: 280.13 Geforce 8600GTS 512MB HL2 DirectX9.0 mode
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #53 from Patrick Rudolph patrick1804@web.de 2011-10-14 11:03:06 CDT --- Hi, I investigated some more time and found this:
PS 2.0 = DirectX 9.0 PS 2.1 = DirectX 9.0a/b PS 3.0 = DirectX 9.0c
by setting caps->PS20Caps.NumInstructionSlots = 512; HL2 detects PS3 by setting caps->PS20Caps.NumInstructionSlots = 511; HL2 detects PS2.0
actually HL2 uses PS2.1 whenever PS3 is detected (because if PS3 is supported, PS2.1 has to be supported too).
Have a look at WINEDEBUG=+d3d_shader
The problem is somewhere inside the PS2.1 code.
The only way to force PS2.0 are the hacks recently posted.
Patrick
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #54 from Patrick Rudolph patrick1804@web.de 2011-10-23 13:13:31 CDT --- It looks like this a a vendor specific bug, I've tested with ATI graphic card to, and there' re no graphical problems with PS2.1.
Maybe this has something to do with Nvidia specific GLX extensions, which aren't available using an ATI card. Please report your card vendor, too.
Tested with: wine1.3.31 ATI Mobility Radeon HD 4500 Series Driver version 8.78-3-100920a 2.6.35-30-generic x86_64
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #55 from Luke wolf@wolfcrafting.com 2012-02-11 22:26:22 CST --- Created attachment 38811 --> http://bugs.winehq.org/attachment.cgi?id=38811 TF2 Working -dxlevel 95
Proof of aforementioned patch working.
http://bugs.winehq.org/show_bug.cgi?id=20053
--- Comment #56 from byteframe byteframe@gmail.com 2012-03-17 17:43:54 CDT --- As of 1.5.0 or before, I no longer experience the dxlevel 9 black background/ui issue with the two source game mods that we're exhibiting it for me (zombie panic source, and resistance and liberation). Yay.
http://bugs.winehq.org/show_bug.cgi?id=20053
Jerome Leclanche adys.wh@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |adys.wh@gmail.com Resolution| |FIXED
--- Comment #57 from Jerome Leclanche adys.wh@gmail.com 2012-03-25 11:43:23 CDT --- Reported fixed. Reopen if not.
http://bugs.winehq.org/show_bug.cgi?id=20053
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #58 from Alexandre Julliard julliard@winehq.org 2012-03-30 13:36:02 CDT --- Closing bugs fixed in 1.5.1.
-
wine-bugs＠winehq.org