http://bugs.winehq.org/show_bug.cgi?id=11826
Summary: World of Warcraft: White Minimap in OpenGL Mode Product: Wine Version: 0.9.56. Platform: PC OS/Version: Linux Status: UNCONFIRMED Severity: minor Priority: P2 Component: opengl AssignedTo: wine-bugs@winehq.org ReportedBy: zak.mayer@gmail.com
Since version 0.9.48 to current 0.9.56 if you use the switch from command line -opengl OR set the World of Warcraft (WoW) Config.wtf SET gxApi "opengl" the in game minimap turns white.
With these commands off (assumed D3D) the minimap is fine, however the game runs 15-50 fps less.
With it on you get the white minimap and 15-50fps more.
ATI 8.1 and 8.2 Catalyst driver w/ 9600XT on Ubuntu 7.10
Another bug report mentions: http://bugs.winehq.org/show_bug.cgi?id=6255
But the Text issue mentioned in that post has not applied since version 0.9.53.
http://bugs.winehq.org/show_bug.cgi?id=11826
Zak Mayer zak.mayer@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Alias|Resonance378 |minimap
http://bugs.winehq.org/show_bug.cgi?id=11826
Stefan Dösinger stefandoesinger@gmx.at changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |stefandoesinger@gmx.at
--- Comment #1 from Stefan Dösinger stefandoesinger@gmx.at 2008-03-03 11:16:35 --- Did it work before 0.9.48? Does it still work if you go back? It might be an ATI driver bug that breaks the minimap, that wouldn't be the first time.
Currently the minimap is broken on r200 cards with both fglrx(not supported any more since quite some time), and all radeon cards with the open source driver(no GL_SGIX_pbuffer support, and no GL_EXT_framebuffer_object support). I don't know the fglrx state for sure, but it wouldn't be the first time this is broken.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #2 from Zak Mayer zak.mayer@gmail.com 2008-03-03 11:22:55 --- I have no real experience with running Wine, WoW, and ATI Prior to the posted versions.
Do any of these GL keys you mention provide a work around in the registry ??
I don't doubt that it could be an ATI driver issue but I thought I'd get started here as this community is much more active in bug hunting.
-Zak
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #3 from Stefan Dösinger stefandoesinger@gmx.at 2008-03-03 11:37:23 --- You said "Since version 0.9.48". That made me think that this is a regression, ie, it worked in 0.9.47. Did you test earlier versions?
Note that if this is an ATI driver then the active bughunting community won't be able to fix it. All in All Wine bugs in the OpenGL area are rather rare, since the platform independent part is just an 1:1 forward. Wine can't do anything wrong there, and in the same way we cannot workaround game<->driver incompatibilities. The only case where Wine could do something wrong is the pixelformat setup.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #4 from Roderick Colenbrander thunderbird2k@gmx.net 2008-03-03 12:20:44 --- The only thing you can try is to try WoW using the latest GIT version of Wine. It lifts the OpenGL pixel format limitation we had in Wine. If that doesn't work in GL mode then there is nothing we can do as Wine's OpenGL directly maps to Linux OpenGL.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #5 from Zak Mayer zak.mayer@gmail.com 2008-03-03 14:21:56 --- @Stefan: "You said "Since version 0.9.48". That made me think that this is a regression, ie, it worked in 0.9.47. Did you test earlier versions?"
Yes I had tested 0.9.46 but getting the ATI driver installed was half of the problem and I never did get the game to run (my fault).
There appear to be a lot of recalls to this issue in various forums prior to the 8.1 Catalyst series so I'm beginning to lean toward it being the ATI driver.
@Roderick "GIT version of Wine"- What? Sorry just not familiar with the lingo around here yet and Acronym speak loses me quickly.
General: Are there any Wine registry keys I can use to check or troubleshoot this issue so that I can write a better bug report when I go to ATI with this? Since you do a 1:1 conversion on calls it's likely that disabling or enable a mode or feature will point to the culprit. As it looks the minimap border(frame) and buttons all render fine, but the minimap itself is just a white disc. The image here: http://ubuntuforums.org/showthread.php?p=4371103 is from someone having the same issue. The white minimap - that is exactly how it appears. This thread is also a good historical context of people using WoW and Wine and their issues past and present. Not the best bug reporting in all but it certainly runs the gambit of issues. 1000+ posts...
Looks like some sort of Alpha overlay issue. Come to think of it the driver install wiki has people setting the environment to --overlay=Xv where opengl or disable are other options. I'll have to test with --overlay=opengl and follow up. http://wiki.cchtml.com/index.php/Ubuntu_Gutsy_Installation_Guide#Configure_t...
Thoughts?
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #6 from Roderick Colenbrander thunderbird2k@gmx.net 2008-03-04 08:22:00 --- The latest Wine source version is found in GIT (something similar to CVS or SVN). See http://www.winehq.org/site/git. Use the info from there to obtain the latest source and install that (remove your old wine first).
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #7 from Zak Mayer zak.mayer@gmail.com 2008-03-04 21:27:12 --- Created an attachment (id=11117) --> (http://bugs.winehq.org/attachment.cgi?id=11117) Commandline Output
Commandline output running wine Wow.exe -opengl
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #8 from Zak Mayer zak.mayer@gmail.com 2008-03-04 21:31:00 --- Addendum - the white minimap only happens when 'indoors' in game.
At this time what happens with the minimap is a zoom/transition on the minimap and it goes to white. As soon as you move the avatar outdoors the minimap comes back.
@Roderick Using the GIT is way beyond my technical expertise at this time as well as overall comfort level with Wine.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #9 from Roderick Colenbrander thunderbird2k@gmx.net 2008-03-07 16:17:01 --- Try again using Wine 0.9.57.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #10 from Zak Mayer zak.mayer@gmail.com 2008-03-08 15:19:10 --- Created an attachment (id=11231) --> (http://bugs.winehq.org/attachment.cgi?id=11231) D3D Err messages
These ERR messages appear in the command line when not using the Wine switch -opengl or not using gxAPI "opengl" in the WoW Config.wtf
http://bugs.winehq.org/show_bug.cgi?id=11826
Zak Mayer zak.mayer@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|0.9.56. |0.9.57.
--- Comment #11 from Zak Mayer zak.mayer@gmail.com 2008-03-08 15:32:51 --- (In reply to comment #9)
Try again using Wine 0.9.57.
Ok removed 0.9.56 and install 0.9.57.
The minimap issue remains in OpenGL when going indoors.
I also had an issue with game text flying around on the screen but resolved that by re-adding to the Wine Registry the tweak found here: http://www.wowwiki.com/Linux/Wine#Registry_Tweak_for_FPS_Boost
I noted some 'err' messages in D3D mode which may or may not be related to frame rates permanently stuck under 10fps unless motionless and looking at the ground.
Attaching that information.
I've noted that ATI has released another version of Catalyst (8.3) but I have not updated due to people mentioning that 3D doesn't work well with X 7.1 ... could just be hearsay though.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #12 from Stefan Dösinger stefandoesinger@gmx.at 2008-03-08 16:39:56 --- The issue is I think that fglrx does not advertize (or support, even?) pbuffers, but wow needs them for the indoor minimap in opengl mode
http://bugs.winehq.org/show_bug.cgi?id=11826
Lei Zhang thestig@google.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|0.9.57. |0.9.56.
--- Comment #13 from Lei Zhang thestig@google.com 2008-03-08 22:13:53 --- please don't change the original reported version
http://bugs.winehq.org/show_bug.cgi?id=11826
Roderick Colenbrander thunderbird2k@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |FIXED
--- Comment #14 from Roderick Colenbrander thunderbird2k@gmx.net 2008-03-09 05:51:16 --- As I said as of Wine 0.9.57 wine's opengl32 is more or less bugfree (there are some small issues left but nothing major.) This means that the issue is either an ATI driver bug or a feature not exported by your drivers.
Like Stefan said the issue is likely related to the lack of pbuffers. Basically all windows opengl drivers offer it but most linux ones except nvidia don't offer it. Older ATI drivers offered it too but those were buggy and slow.
There is nothing which we can do about the lack of pbuffers. I fear your only option is to use direct3d.
http://bugs.winehq.org/show_bug.cgi?id=11826
Stefan Dösinger stefandoesinger@gmx.at changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|FIXED |WONTFIX
--- Comment #15 from Stefan Dösinger stefandoesinger@gmx.at 2008-03-09 06:40:13 --- It's not fixed, but we can't fix it. Changing resolution to WONTFIX
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #16 from Zak Mayer zak.mayer@gmail.com 2008-03-10 09:34:14 --- Thank you all for your time in looking after this issue and stewarding it properly.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #17 from Zak Mayer zak.mayer@gmail.com 2008-03-11 09:50:14 ---
From the FGLRX page: http://debian.wgdd.de/temp/fglrx_man/fglrx.4x.html
Option "PBuffer" "integer"
... Parameter PBuffer area size "0" PBuffer area disabled. "1" Size of PBuffer area 1280x1024. "2" Size of PBuffer area 1600x1200. "3" Size of PBuffer area 1920x1200. "4" Size of PBuffer area 2048x1536.
My guess is that PBuffer is disabled by default. If that is the case what sized should be used or will it be trial and error?
However, these look like old pages (2007) pre Catalyst 8.0+ drivers. http://debian.wgdd.de/temp/fglrx_man/
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #18 from Zak Mayer zak.mayer@gmail.com 2008-03-17 15:56:48 --- Additionally I've seen this old post: http://www.winehq.org/pipermail/wine-patches/2006-December/033938.html
Although I'm not certain of its relevancy to this situation. I don't see how pbuffer support would be pulled from a driver only to re-implement it down the road, but one never can tell.
Additionally in gdi32 the call for the stub in the gdi32 spec is commented out: # @ stub GdiSwapBuffers Obviously I have no clue what this means for the overall pbuffer support but I've been able to see that it and several other buffer objects(methods,variables) are working with pbuffer support in Wine. OpenGl.spec calls - @ stdcall wglSwapBuffers(long) gdi32.SwapBuffers Which is @ stdcall SwapBuffers(long) in the gdi32.spec
I haven't been able to review the code beyond this and if it isn't obvious already I am quit lost as to what is getting called when and where. However given past reading and examples that I've found throughout the Wine project it looks as if pbuffer is and has been and continues to be a pita.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #19 from Stefan Dösinger stefandoesinger@gmx.at 2008-03-17 16:13:16 --- Pbuffers are a pain in the ass, that's why they have been replaced with frame buffer objects soon. On GLX(X11's openGL interface), PBuffers never made it beyong an experimental extension, which is widely supported and used though, but it is still named GLX_SGIX_pbuffer, the X in SGIX indicates it is not finalized.
Generally games nowadays use frame buffer objects, they are platform independent and do not require context switches. However, there are a few games that use pbuffers. It might be helpful to emulate pbuffers using fbos, but we currently do not have the time and manpower to do this, but we accept patches ;-)
GdiSwapBuffers is not related to buffers. It is a opengl 1.0 level function which swaps the contents of the front and back buffer after the game finished rendering a frame.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #20 from Zak Mayer zak.mayer@gmail.com 2008-03-17 16:57:27 --- Ah many new things then but in summation:
pbuffers bad, frame buffer good
I think I have it and I will look to direct my inquisitive efforts in that direction.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #21 from Zak Mayer zak.mayer@gmail.com 2008-03-17 23:47:42 --- Funny now that I should think of this. I reviewed what I know about ATI and their support for OpenGL and recalled a feature of OpenGL (and an option) called TripleBuffer.
Sure enough, the World of Warcraft - Config.wtf file has an option for triple buffering.
SET gxTripleBuffer "1"
And there is no longer a white minimap when indoors.
Additionally this has introduced being able to see the buffer or frames update every 3-10 seconds when indoors in the lower left hand corner of the screen. You see a little black box pop up behind the text window (almost the same size as) and then away it goes.
Wish I had some way of getting video of it.
In the greater scheme of things it's highly annoying but livable.
Still interested in seeing what can be done with Frame Buffers though.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #22 from Zak Mayer zak.mayer@gmail.com 2008-03-17 23:50:59 --- (In reply to comment #21)
Funny now that I should think of this. I reviewed what I know about ATI and their support for OpenGL and recalled a feature of OpenGL (and an option) called TripleBuffer.
Sure enough, the World of Warcraft - Config.wtf file has an option for triple buffering.
SET gxTripleBuffer "1"
And there is no longer a white minimap when indoors.
Additionally this has introduced being able to see the buffer or frames update every 3-10 seconds when indoors in the lower left hand corner of the screen. You see a little black box pop up behind the text window (almost the same size as) and then away it goes.
Wish I had some way of getting video of it.
In the greater scheme of things it's highly annoying but livable.
Still interested in seeing what can be done with Frame Buffers though.
*sigh* never mind... never test late at night... I was not running in opengl. the white minimap still persists.
OpenGL got a huge jump in framerates though. Went from ~15-20 outdoors to ~40-60 and from 25-35 indoors to ~90-110.
http://bugs.winehq.org/show_bug.cgi?id=11826
--- Comment #23 from Zak Mayer zak.mayer@gmail.com 2008-03-17 23:52:43 ---
Additionally the command line has this message in it:
fixme:d3d:test_pbo_functionality >>>>>>>>>>>>>>>>> GL_INVALID_OPERATION (0x502) from Loading the PBO test texture @ directx.c / 3004
http://bugs.winehq.org/show_bug.cgi?id=11826
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #24 from Austin English austinenglish@gmail.com 2008-09-17 08:51:35 --- Closing WONTFIX.
http://bugs.winehq.org/show_bug.cgi?id=11826
Jerome Leclanche adys.wh@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |adys.wh@gmail.com Resolution|WONTFIX |DUPLICATE
--- Comment #25 from Jerome Leclanche adys.wh@gmail.com 2012-02-24 06:02:57 CST --- Duplicate
*** This bug has been marked as a duplicate of bug 12300 ***