[Bug 12557] New: System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Summary: System Shock 2 Demo: Starting game results in black
                    screen
           Product: Wine
           Version: unspecified
          Platform: Other
               URL: http://www.download.com/System-Shock-2-demo/3000-7539_4-
                    10031250.html
        OS/Version: other
            Status: UNCONFIRMED
          Severity: major
          Priority: P2
         Component: directx-d3d
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

Created an attachment (id=12130)
 --> (http://bugs.winehq.org/attachment.cgi?id=12130)
shock2.exe run with d3d tracing

Hi there,

I just installed the System Shock 2 demo to check if it still does not work
with wine (according to the AppDB it doesn't really work well).

And indeed, it doesn't even start. I get some very high pitched sound
distortions after the loading screen ends. The usual background hum continues,
but all I see is a black screen. Sometimes I see a dialogue box for some
milliseconds, but that's all.
Now wine consumes a huge amount of CPU and sometimes doesn't even let me switch
to other applications (currently virtual desktop is on) and even VT switch is
impossible. I have put wineserver -k on a ACPI event button, this is the only
way I can kill wine after SS2 fucks it up.

I post a very short console log here and a more detailed d3d trace as
attachment:
err:ddraw:PixelFormat_DD2WineD3D Unknown Pixelformat!
err:ddraw:IDirectDrawImpl_CreateNewSurface Unsupported / Unknown pixelformat
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 88760091
err:ddraw:PixelFormat_DD2WineD3D Unknown Pixelformat!
err:ddraw:IDirectDrawImpl_CreateNewSurface Unsupported / Unknown pixelformat
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 88760091

The thing is that the game should use D3D, but I can't even select it in the
options. The only thing I can select is a DirectDraw renderer. I'm going to
create some ddraw traces, maybe these have additional informations.

BUT the game seems to use d3d calls (the d3d trace grew quite large).

Oh yes, I have added the safe_texture_manager stuff to the cam.cfg, according
to board postings on the net this should reduce some problems on windows as
well. But it doesn't really make any difference here, I also get a black screen
without the config modification.

Greetings,
Tobias

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Tobias Jakobi  changed:

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

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Tobias Jakobi  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         OS/Version|other                       |Linux
           Platform|Other                       |PC
            Version|unspecified                 |0.9.59.

--- Comment #1 from Tobias Jakobi   2008-04-13 11:58:41 ---
Additional infos on the system:

lspci:
00:02.1 Display controller: Intel Corporation Mobile 915GM/GMS/910GML Express
Graphics Controller (rev 03)

mesa 7.0.2
xorg-server 1.4.0.90
xf86-video-intel 2.2.99.902

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #2 from Tobias Jakobi   2008-04-13 12:18:01 ---
Created an attachment (id=12134)
 --> (http://bugs.winehq.org/attachment.cgi?id=12134)
shock2.exe run with ddraw tracing

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |download

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Lei Zhang  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|major                       |normal

--- Comment #3 from Lei Zhang   2008-04-14 15:43:40 ---
Severity is not major.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #4 from Tobias Jakobi   2008-04-14 15:50:45 ---
I should state that this is the second DirectX6 game I have problems with. The
other one is Blood 2, which AFAIK also uses DX6 - this one is impossible to
play in D3D mode.

Can someone post some information about the DX6 implementation of wine? I
already searched the Wiki but found almost nothing about that version. Is it
even supported at all?

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #5 from Tobias Jakobi   2008-04-14 15:54:17 ---
http://www.winehq.org/pipermail/wine-users/2007-April/027117.html <- seems like
my problems are even more serious :-D

I forgot to add something: I asked my colleague to test the demo on his NV card
and it worked for him. I'm a bit surprised. I wasn't expecting uber-great
performance from my i915 but even CS 1.5 is running and SS2 should use a
similar amount of performance...

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Alexander Dorofeyev  changed:

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

--- Comment #6 from Alexander Dorofeyev   2008-04-14 16:43:00 ---
(In reply to comment #4)
> I should state that this is the second DirectX6 game I have problems with. The
> other one is Blood 2, which AFAIK also uses DX6 - this one is impossible to
> play in D3D mode.
>
> Can someone post some information about the DX6 implementation of wine? I
> already searched the Wiki but found almost nothing about that version. Is it
> even supported at all?

All of DX6 (and earlier) interfaces and all or most of their functionality is
there. But there are often many bugs. By selecting DirectDraw renderer you ARE
selecting direct3d. Old D3D was kind of extension of ddraw. But, you are not
the first person who mentions this issue in bug reports about old games (naming
of renderer), so, perhaps, Wine doesn't properly report the name of device or
something like that. Although it's non-essential and minor problem, to avoid
user confusion it would, of course, be better if in Wine these kinds of options
were presented identical to Windows. If you are absolutely sure it is "D3D" in
Windows and something else in Wine (in renderer choices), and can verify that,
then record exactly what it should be and what it is and/or make screenshots.
But, open a new bug for that if you do, because it's unrelated.

This game reportedly ran before, so you can try some earlier versions and maybe
do a regression test if they run SS2 better.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #7 from Tobias Jakobi   2008-04-14 17:18:46 ---
The reason I mentioned DDraw was that the Sacrifice demo also only reports
DirectDraw renderers. So something is definitely wrong there.

I can try a regression test, what wine version do you suggest? I suspect
however that building old wine versions is going to be complicated. I already
had a lot of trouble building versions for my last bisect, versions below
0.9.36 didn't really want to build on my 64bit system.

Oh yes, I forgot to mention that in the SS2 config menu I have exactly two
renderer from which I can choose. But both are named the same: "DirectDraw
blabla"

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #8 from Tobias Jakobi   2008-04-15 14:04:14 ---
I tried shock2.exe with both wine-0.9.44 and wine-0.9.36 - these versions
didn't even start the application. I just get nothing, an emtpy screen (when
virtual desktop is activated).

Should I downgrade even more?

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #9 from Alexander Dorofeyev   2008-04-16 01:32:48 ---
(In reply to comment #8)
> I tried shock2.exe with both wine-0.9.44 and wine-0.9.36 - these versions
> didn't even start the application. I just get nothing, an emtpy screen (when
> virtual desktop is activated).
>
> Should I downgrade even more?

I think, no. AppDB indicates that it likely has worked around 0.9.53. But maybe
on Intel it never did.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Jan Buecken  changed:

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

--- Comment #10 from Jan Buecken   2008-04-16 04:53:42 ---
*** This bug has been confirmed by popular vote. ***

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #11 from Tobias Jakobi   2008-04-16 13:59:49 ---
I wanted to add that Jan and I tested this on a Intel i945 chipset. Another
Intel that doesn't work. The test was done with 0.9.58, 0.9.44 and 0.9.36 - no
luck at all.

I can do a test on my NV35 card - sort of a reference check. I still hope to
get this working on the Intel cards. Any ideas what to do (and check)?

Greetings,
Tobias

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #12 from Tobias Jakobi   2008-04-16 15:53:00 ---
Tested on NV35 and it worked flawless. I only had some minor rendering issues,
but nothing that made the game unplayable.

I really have the impression that at least DX6 never really worked with Intel
cards. That's the same with Blood 2 for me. I can get it working somehow there,
but either it segfaults or if I can get it running it's horribly slow and
visuals are garbled.
Any ideas about DX6 games I can test? Games with demos are preferred :D

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #13 from Alexander Dorofeyev   2008-04-17 02:56:47 ---
(In reply to comment #12)
> Tested on NV35 and it worked flawless. I only had some minor rendering issues,
> but nothing that made the game unplayable.
>
> I really have the impression that at least DX6 never really worked with Intel
> cards. That's the same with Blood 2 for me. I can get it working somehow there,
> but either it segfaults or if I can get it running it's horribly slow and
> visuals are garbled.
> Any ideas about DX6 games I can test? Games with demos are preferred :D

Intel has many problems because of driver bugs, but usually of the sort of
crashing / memory corruption. I don't know if that could be related.

DX6 is a bit moot, because some games claim to require it during install, but
actually appear to use basic functionality that should've been available prior
to dx6. The code in wine is mostly common anyway. Some old demos that I know
that use d3d/ddraw of dx6 or prior and should be available for download:

Aliens vs Predator 1 demo
Forsaken demo
Moto Racer demo

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #14 from Alexander Dorofeyev   2008-04-18 14:57:37 ---
(In reply to comment #7)
> The reason I mentioned DDraw was that the Sacrifice demo also only reports
> DirectDraw renderers. So something is definitely wrong there.

I would appreciate if you create separate bug for that, with full info
included: what the game reports on native, what it reports on wine, screenshots
or exact output, the more games the better.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #15 from Tobias Jakobi   2008-04-20 13:58:53 ---
Info: The AvP1 demo works for me. I have problems with transparency there but I
get a somehow playable game.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #16 from Alexander Dorofeyev   2008-04-21 11:30:29 ---
Tried to analyze the log, but I think there isn't enough info to tell where
that strange error is coming from. Try creating more verbose log with
+ddraw,+ddraw_thunk,+d3d7,+d3d.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #17 from Tobias Jakobi   2008-04-27 16:01:30 ---
Created an attachment (id=12526)
 --> (http://bugs.winehq.org/attachment.cgi?id=12526)
shock2.exe run with WINEDEBUG=+ddraw,+ddraw_thunk,+d3d7,+d3d

simple run where I only switch into the config menu and into the graphics
device setup
maybe this also helps with the "only DirectDraw renderer showing up" problem

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #18 from Tobias Jakobi   2008-04-27 16:03:28 ---
Created an attachment (id=12527)
 --> (http://bugs.winehq.org/attachment.cgi?id=12527)
shock2.exe run with WINEDEBUG=+ddraw,+ddraw_thunk,+d3d7,+d3d

run with entering straight into the game
loading sequence ends, the HDA codec outputs some high pitched clicks, after
that I wait a few seconds, then finishing the process with "wineserver -k"

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #19 from Alexander Dorofeyev   2008-04-30 15:09:17 ---
BTW is there a chance you are using any direct3d registry overrides as
described in http://wiki.winehq.org/UsefulRegistryKeys? Probably not, as you
didn't mention it and also it works for you on nvidia, but just asking to be
sure.

I've found there's a bug with RenderTargetLockMode=readtex that affects Nvidia
cards and produces similar results (menu failing to render). It affects at
least Thief (The Dark Project) and System Shock 2. I'll file a bug for that.

I will also try to analyze your new log.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #20 from Alexander Dorofeyev   2008-04-30 15:27:35 ---
Ok, there's something interesting in the log. I compared to such log on my
machine, and it looks like the game wants 16-bit Z-BUFFER and on my machine the
app gets it from IDirect3DImpl_7_EnumZBufferFormats, cancels enumeration and
proceeds happily.

On your card it looks like D16 format is never returned by
IDirect3DImpl_7_EnumZBufferFormats. Possibly this is what confuses the game.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #21 from Tobias Jakobi   2008-04-30 16:23:34 ---
(In reply to comment #19)
> BTW is there a chance you are using any direct3d registry overrides as
> described in http://wiki.winehq.org/UsefulRegistryKeys? Probably not, as you
> didn't mention it and also it works for you on nvidia, but just asking to be
> sure.

No, I only have set GLSL to disabled since the i915 doesn't support it in the
first place. This may change with future driver releases, but currently it only
has ARB_fragment_program and ARB_vertex_program support.

(In reply to comment #20)
> Ok, there's something interesting in the log. I compared to such log on my
> machine, and it looks like the game wants 16-bit Z-BUFFER and on my machine the
> app gets it from IDirect3DImpl_7_EnumZBufferFormats, cancels enumeration and
> proceeds happily.
>
> On your card it looks like D16 format is never returned by
> IDirect3DImpl_7_EnumZBufferFormats. Possibly this is what confuses the game.

Oh no, not again.
You maybe wanna take at look at this bug
http://bugs.winehq.org/show_bug.cgi?id=11970

That's OpenGL but it shows the Intel misbehaving when it comes to z-buffer
formats. Probably these two are related, but as I have no deeper insight into
the wine source I can't say for sure.

And thanks for looking into that! Anything further I can do? I have currently
access to three Intel based systems. One i915 chipset on my laptop system (only
linux env, no Windows here), a i945 on my university machine (also linux env
only, Ubuntu based distro, rather old driver for the card) and a i965 (at least
that's what I remember) with a windows-only (XP distro) config.

Greetings,
Tobias

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Alexander Dorofeyev  changed:

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

--- Comment #22 from Alexander Dorofeyev   2008-04-30 17:00:44 ---
(In reply to comment #21)
> (In reply to comment #19)
> And thanks for looking into that! Anything further I can do? I have currently
> access to three Intel based systems. One i915 chipset on my laptop system (only
> linux env, no Windows here), a i945 on my university machine (also linux env
> only, Ubuntu based distro, rather old driver for the card) and a i965 (at least
> that's what I remember) with a windows-only (XP distro) config.

You can attach glxinfo output for non-working card(s).

This bug looks like something for Roderick, as afaik he's the man working on
this part. I'll add him to CC here. Not sure if this will notify him, so
contacting him on #winehackers may be a good idea too. I would tell him, but he
isn't on the channel atm.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #23 from Alexander Dorofeyev   2008-05-01 05:37:10 ---
Created an attachment (id=12614)
 --> (http://bugs.winehq.org/attachment.cgi?id=12614)
sshock log on a system where it works

For the record, adding first 70000 lines of the log on my system. The
interesting part is near EnumZBufferFormats.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #24 from Tobias Jakobi   2008-05-01 17:16:11 ---
Created an attachment (id=12628)
 --> (http://bugs.winehq.org/attachment.cgi?id=12628)
glxinfo on my i915 based system (non-working)

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #25 from Alexander Dorofeyev   2008-05-01 19:35:20 ---
Created an attachment (id=12631)
 --> (http://bugs.winehq.org/attachment.cgi?id=12631)
hack for intel

Can you try this hack, please. If it doesn't improve anything, post
+ddraw,+ddraw_thunk,+d3d7,+d3d,+d3d_caps log with this hack.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #26 from Roderick Colenbrander   2008-05-02 02:41:43 ---
Likely my latest patch
(http://www.winehq.org/pipermail/wine-patches/2008-May/054435.html) fixes the
issue. Basically DirectDraw never requested a depth buffer. Before it worked
because WineD3D was only able to use a single pixel format and on that one we
always had a depth buffer.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #27 from Alexander Dorofeyev   2008-05-02 08:35:05 ---
(In reply to comment #26)
> Likely my latest patch
> (http://www.winehq.org/pipermail/wine-patches/2008-May/054435.html) fixes the
> issue. Basically DirectDraw never requested a depth buffer. Before it worked
> because WineD3D was only able to use a single pixel format and on that one we
> always had a depth buffer.

I think this could be a separate issue - CheckDeviceFormat failing for D16
because only 24/8 depth/stencil is available. Anyway, since your fix is in git
it should be easy to find out now - by simply updating git.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #28 from Roderick Colenbrander   2008-05-02 08:49:26 ---
I haven't checked the log but in general when no 16-bit depth format is around
we accept 24-bit (IsPixelFormatCompatibleWithDepthStencilFmt or whatever the
function is called does that).

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #29 from Tobias Jakobi   2008-05-02 08:50:08 ---
I'm gonna test git on the i945 system, if that does not work I check it again
with the patch applied.
I can do a check with my i915 system this evening.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #30 from Tobias Jakobi   2008-05-02 11:07:54 ---
Results from the i945 system. The problem is somehow still there, but it has
changed to some degree. I know can identify an error message box in the virtual
desktop. I'm gonna attach a screenshot so you can see what I mean.

I also get these outputs on console:
fixme:win:EnumDisplayDevicesW ((null),0,0x32f7dc,0x00000000), stub!
fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to
16
fixme:x11drv:X11DRV_desktop_SetCurrentMode Cannot change screen BPP from 32 to
16
err:d3d:WineD3D_ChoosePixelFormat Can't find a suitable iPixelFormat
err:ddraw:PixelFormat_DD2WineD3D Unknown Pixelformat!
err:ddraw:IDirectDrawImpl_CreateNewSurface Unsupported / Unknown pixelformat
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 88760091
fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x155b20) : stub
fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x155b20) : stub
fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x155b20) : stub
err:ddraw:PixelFormat_DD2WineD3D Unknown Pixelformat!
err:ddraw:IDirectDrawImpl_CreateNewSurface Unsupported / Unknown pixelformat
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 88760091
fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x155b20) : stub
fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x155b20) : stub
fixme:d3d:IWineD3DDeviceImpl_ValidateDevice (0x155b20) : stub

That's with current GIT, Alexander's patch was not applied.

Now doing a check with the patch...

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #31 from Tobias Jakobi   2008-05-02 11:10:06 ---
Created an attachment (id=12643)
 --> (http://bugs.winehq.org/attachment.cgi?id=12643)
screen after the game stops

test run on i945 system, no patch applied (just pure GIT)

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #32 from Tobias Jakobi   2008-05-02 12:08:07 ---
Created an attachment (id=12645)
 --> (http://bugs.winehq.org/attachment.cgi?id=12645)
shock2.exe segfaults with patched wine GIT on i945

after applying the patch the game now exits immediately with a segfault

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #33 from Tobias Jakobi   2008-05-02 12:14:12 ---
Created an attachment (id=12646)
 --> (http://bugs.winehq.org/attachment.cgi?id=12646)
shock2.exe on i945, patched GIT: crashing not predictable

W00t!

I started shock2.exe a second time and not it worked. I could actually enter
the game and move around. I didn't try out many things but tried to exit the
game again, triggering a segfault when pressing ESC.

Seems a bit unstable but at least it's now possible to get in-game.

Leaving campus now, I try it again on my home system.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #34 from Alexander Dorofeyev   2008-05-02 13:07:17 ---
Random-like crashing probably is that evil intel driver memory corruption bug,
there isn't much we can do about it :(.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #35 from Tobias Jakobi   2008-05-02 14:48:12 ---
I'm currently cloning GIT on my local i915 system. I have recent GPU drivers,
like mesa 7.0.3 and xf86-video-intel-2.3.0.
I hope this works, at least regular OpenGl applications are very stable with my
card.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #36 from Tobias Jakobi   2008-05-02 14:51:37 ---
Forgot:

The crashing does not seem to be random. It either happens when the game is
started, means right after the loading sequence ends and the engine gets
"in-game".
The other crash ocurred when pressing ESC. I still have to find out what should
happen there, but I suspect that the game changes to the regular 2D game menu.

So it's not totally random. You can trigger it with specific actions. Have to
look into this a bit more.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #37 from Roderick Colenbrander   2008-05-03 07:39:13 ---
Created an attachment (id=12661)
 --> (http://bugs.winehq.org/attachment.cgi?id=12661)
proposed fix

Try this patch.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #38 from Tobias Jakobi   2008-05-03 07:50:40 ---
Hi Roderick,

thanks for the patch. I gonna try it later.

Currently I'm still experimenting with the GIT snapshot and Alexander's hack.
It seems to work somehow, but the gfx are garbled (I'm gonny create some shots)
and it's really slow, like 1 frame every 2 seconds.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #39 from Roderick Colenbrander   2008-05-03 09:45:07 ---
The game is known to be slow. It uses z-buffer read backs a lot which are slow.
On nvidia hardware the game is fast these days for a big part due to the
availability of PBOs. If you want to play this game upgrade your videocard.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #40 from Tobias Jakobi   2008-05-03 10:45:03 ---
Created an attachment (id=12670)
 --> (http://bugs.winehq.org/attachment.cgi?id=12670)
shock2.exe run on i915 system

this is GIT patched with Alexander's hack
shock2.exe was run on the i915 system
notice the vertical lines, the pattern isn't static - sometimes it changes, but
usually when moving around (so when the screen content itself changes, maybe
this is related to framebuffer readbacks?)

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #41 from Tobias Jakobi   2008-05-03 10:46:10 ---
Created an attachment (id=12671)
 --> (http://bugs.winehq.org/attachment.cgi?id=12671)
shock2.exe run on i915 (another shot)

another screenshot
this time you can see a bit more, even status messages are displayed :-)

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #42 from Roderick Colenbrander   2008-05-03 10:50:18 ---
I really recommend you to try my patch as it another big issue not fixed by the
previous hack (actually the hack is correct, I did something similar). It might
result in a better pixel format..

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #43 from Tobias Jakobi   2008-05-03 10:58:19 ---
Additional informations:

I updated my in-portage wine version to 0.9.61 and checked the game with this
version. It does not work and seems to display the same error message box like
on the other screenshot I attached.
The thing is that on the i945 system the message box (even if it's not drawn
correctly) stays, so I can take a screenshot. On the i915 system it only shows
up for some milliseconds and is gone after that, resulting in a fully black
screen.

Alexander's hack is working partially on i915. I tried starting the game
multiple times and it was possible to get in-game every time. So this seems to
be stable. I did however NOT try to exit the game via ESC, I always used the
wineserver -k method. This however did not produce segfaults or any other error
messages.

Next thing I'm gonna do is to check (and record) what happens in case I exit
the game with ESC (logs are coming...)
After that I try Roderick's patch on a new GIT snapshot and check back here.

@pixel buffer objects:
I was under the impression that the i915tex driver (which is now renamed to
i915, the previous i915 mesa driver was dropped) DOES support PBO, at least I
can read something about i915 and PBO on the mesa bugzilla:
http://bugs.freedesktop.org/show_bug.cgi?id=13747

I really don't know whats happening there.

According to the mesa changelog
http://www.mesa3d.org/relnotes-6.5.2.html
the i915tex driver should have these capabilities. But you're right, I don't
see them exposed in the glxinfo extension string...

Gonna ask on the freedesktop ml for help.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #44 from Tobias Jakobi   2008-05-03 11:03:35 ---
Created an attachment (id=12673)
 --> (http://bugs.winehq.org/attachment.cgi?id=12673)
shock2.exe run on i915, pressing ESC results in big trouble

using GIT snapshot + Alexander's hack
getting into game works as usual, I'm walking around a bit (to me it seems like
the status display in the middle of the screen is drawn OVER the vertical
lines, or not??) and then pressing ESC

@Roderick: That's the next thing I try :-)

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #45 from Roderick Colenbrander   2008-05-03 11:13:55 ---
PBOs are around if GL_ARB_pixel_buffer_object is shown by glxinfo.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #46 from Tobias Jakobi   2008-05-03 11:15:40 ---
Yep, I know that (I programmed a lot in OpenGL in the past) - but sadly that's
not the case here... :-(

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #47 from Tobias Jakobi   2008-05-03 12:35:02 ---
Behaviour with your patch Roderick isn't as "stable" as with Alexander's patch.

It sometimes works to get into the game, sometimes the screen stays black. I'm
currently experimenting inside my XFCE environment, going to do some test in a
failsafe environment without window manager.

The visual errors (vertical colored lines) stay however, IF I manage to get
in-game.

I have posted a message to the mesa3d-users ml asking them for help about the
PBO "problem".

Any traces I can do?

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #48 from Roderick Colenbrander   2008-05-03 12:59:24 ---
Alexander his patch only lets IsPixelFormatCompatibleWithDepthFmt continue but
doesn't change WineD3D_ChoosePixelFormat which is needed. (Well it isn't really
needed as without it would fall back to plain ChoosePixelFormat but in general
it is better not to call it as it wouldn't allow you multisampling and other
extended pixel format features)

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #49 from Tobias Jakobi   2008-05-03 14:49:54 ---
I have quite some problems getting sane results with the failsafe terminal.

I have virtual desktop enabled and if I don't pipe (via &>) all console output
into a file and just start wine with:
~/wine-git/wine shock2.exe
The process crashes with a flood of logging output to the console as soon as
the loading sequence ends.

Still trying to capture the log somehow.

If I pipe the output into a file the process also crashes, but the log is quite
small. It's not the same crashing behaviour.

I also tested the AvP1 demo with wine-0.9.61 and z-buffer problems are gone (I
had portions of the scene, almost only dynamic structures like water or doors
rendered wrong with wine-0.9.60). But I'm also experiencing the vertical lines
there, BUT not in-game. Only during the starting sequence and the ending
sequence (as bunch of pictures is shown when you quit the game, these pictures
are affected by the vertical line bug).

I'm going to make some checks for shock2 without vdesktop.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #50 from Tobias Jakobi   2008-05-03 15:18:46 ---
Created an attachment (id=12677)
 --> (http://bugs.winehq.org/attachment.cgi?id=12677)
shock2.exe runs on failsafe terminal X environment

i made multiple runs, testing some settings

runs are documented inside the log, interested is a change in crash behaviour
when ingame resolution is changed.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #51 from Tobias Jakobi   2008-05-03 17:14:48 ---
In can confirm that setting RenderTargetLockMode to textex also (partially)
solves the problem with System Shock 2.
The gfx issues are gone and the performance is much better (still bad, but a
big improvement when comparing to readdraw mode).

I still get these a lot:
fixme:d3d_surface:IWineD3DSurfaceImpl_LockRect Reading from render target with
a texture isn't implemented yet, falling back to framebuffer reading

Pressing ESC also works now, but not with the desired effect. Upon pressing ESC
rendering stops, BUT the application does not crash. The menu that should show
up does not. Pressing ESC again lets the "ingame rendering" continue and I can
play again. Exiting via Alt+F4 is possible and works with problems.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #52 from Roderick Colenbrander   2008-05-04 03:27:22 ---
Just set it to readtex. If someone can test on Nvidia if the issues are there
too (most likely they aren't) then you know whether it is a driver bug or not.
I believe Stefan tried it on his Geforce7 less than a month ago and it worked
great.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #53 from Alexander Dorofeyev   2008-05-04 04:14:42 ---
(In reply to comment #52)
> Just set it to readtex. If someone can test on Nvidia if the issues are there
> too (most likely they aren't) then you know whether it is a driver bug or not.
> I believe Stefan tried it on his Geforce7 less than a month ago and it worked
> great.

On nvidia i've issues with readtex in these games
(http://bugs.winehq.org/show_bug.cgi?id=12890). With default settings it works
visually ok on geforce 2 mx and geforce 6100, although on geforce2 mx
performance is poor (just few fps vs what feels like at least 20-30 fps on
windows xp in 1024x768).

Don't know about the color lines. Considering the effect of RT lock mode
setting, i guess it's glDrawPixels that produces issues? I guess it may be
worth seeking some non-wine opengl tests or demos that do read/draw pixels a
lot. If same thing shows up there then obviously it's not wine's problem.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #54 from Alexander Dorofeyev   2008-05-04 04:19:16 ---
Hmm. BTW, this is somewhat similar:
http://bugs.winehq.org/show_bug.cgi?id=12092 (reportedly on radeon with open
source drivers).

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #55 from Austin English   2008-05-04 04:43:44 ---
(In reply to comment #52)
> Just set it to readtex. If someone can test on Nvidia if the issues are there
> too (most likely they aren't) then you know whether it is a driver bug or not.
> I believe Stefan tried it on his Geforce7 less than a month ago and it worked
> great.

Works fine with nvidia geforce fx 5200 / 169.12 drivers.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #56 from Tobias Jakobi   2008-05-04 08:11:36 ---
(In reply to comment #53)
> On nvidia i've issues with readtex in these games
> (http://bugs.winehq.org/show_bug.cgi?id=12890). With default settings it works
> visually ok on geforce 2 mx and geforce 6100, although on geforce2 mx
> performance is poor (just few fps vs what feels like at least 20-30 fps on
> windows xp in 1024x768).

Does wine really create such high amount of overhead?
> Don't know about the color lines. Considering the effect of RT lock mode
> setting, i guess it's glDrawPixels that produces issues? I guess it may be
> worth seeking some non-wine opengl tests or demos that do read/draw pixels a
> lot. If same thing shows up there then obviously it's not wine's problem.

Any application/demo you can suggest?

I'm going to file a bugreport for mesa nonetheless. If the radeon OS driver is
also affected this may be a bug that's affecting more than one card.

Still no answer from the mesa3d-users ml yet. I'm going to try bumping some of
my gfx components to GIT, maybe I can get the features this way. Currently I'm
suspecting that the mesa build system still creates the ordinary "old" i915
driver.
Why do I think that? I can find 4 libs in /usr/lib/dri:
i810_dri.so  i915_dri.so  i915tex_dri.so  i965_dri.so

According to my knowledge about the mesa i915 driver development the
i915tex_dri.so should not exist at all. There should be only one lib, the
i915_dri.so - so I tried to symlink i915tex_dri.so to i915_dri.so, resulting in
indirect rendering. Reason was a too old DRM interface (LIBGL_DEBUG=verbose).

My DRM interface version is 1.6.0 and the i915tex driver wants a 1.7.x
interface. So the i915 driver seems to be satisfied with only the old 1.6.0
interface. Another proof more me that I'm not using the modern driver rewrite,
since this one uses TTM and the more recent goodies from the DRM tree.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #57 from Roderick Colenbrander   2008-05-04 09:54:13 ---
The main limitation for older ddraw / d3d apps is the poor LockRect /
UnlockRect performance. These calls on Windows give you direct access to e.g.
the framebuffer (they can give you a direct pointer to the video memory).
OpenGL doesn't provide this and we need to read the data from the framebuffer
on a LockRect operation and need to write the result back to screen during an
UnlockRect operation. On systems with PBOs performance is a lot better because
they give you asynchronous reads / writes which results in that you don't have
to wait before you can continue with other tasks.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #58 from Tobias Jakobi   2008-05-04 17:13:10 ---
I was right about the driver fiasco. mesa was indeed also still building the
older driver components (and X was using it). After massive GIT upgrades
(kernel DRM, libdrm, mesa, xorg-server, dri2proto, etc.) I can finally use the
new i915tex codebase.
And it's really MUCH faster than the old driver, I have doubled my framerate in
quake3.

New glxinfo output:
name of display: :0.0
display: :0  screen: 0
direct rendering: Yes
server glx vendor string: SGI
server glx version string: 1.2
server glx extensions:
    GLX_ARB_multisample, GLX_EXT_import_context, GLX_EXT_texture_from_pixmap, 
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_copy_sub_buffer, 
    GLX_OML_swap_method, GLX_SGI_swap_control, GLX_SGIS_multisample, 
    GLX_SGIX_fbconfig, GLX_SGIX_visual_select_group
client glx vendor string: SGI
client glx version string: 1.4
client glx extensions:
    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context, 
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_allocate_memory, 
    GLX_MESA_copy_sub_buffer, GLX_MESA_swap_control, 
    GLX_MESA_swap_frame_usage, GLX_OML_swap_method, GLX_OML_sync_control, 
    GLX_SGI_make_current_read, GLX_SGI_swap_control, GLX_SGI_video_sync, 
    GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer, 
    GLX_SGIX_visual_select_group, GLX_EXT_texture_from_pixmap
GLX version: 1.2
GLX extensions:
    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context, 
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_swap_control, 
    GLX_MESA_swap_frame_usage, GLX_OML_swap_method, GLX_SGI_swap_control, 
    GLX_SGI_video_sync, GLX_SGIS_multisample, GLX_SGIX_fbconfig, 
    GLX_SGIX_visual_select_group, GLX_EXT_texture_from_pixmap
OpenGL vendor string: Tungsten Graphics, Inc
OpenGL renderer string: Mesa DRI Intel(R) 915GM 20061102 x86/MMX/SSE2
OpenGL version string: 1.4 Mesa 7.1
OpenGL extensions:
    GL_ARB_depth_texture, GL_ARB_fragment_program, GL_ARB_multisample, 
    GL_ARB_multitexture, GL_ARB_pixel_buffer_object, GL_ARB_point_parameters, 
    GL_ARB_shadow, GL_ARB_texture_border_clamp, GL_ARB_texture_compression, 
    GL_ARB_texture_cube_map, GL_ARB_texture_env_add, 
    GL_ARB_texture_env_combine, GL_ARB_texture_env_crossbar, 
    GL_ARB_texture_env_dot3, GL_ARB_texture_mirrored_repeat, 
    GL_ARB_texture_non_power_of_two, GL_ARB_texture_rectangle, 
    GL_ARB_transpose_matrix, GL_ARB_vertex_buffer_object, 
    GL_ARB_vertex_program, GL_ARB_window_pos, GL_EXT_abgr, GL_EXT_bgra, 
    GL_EXT_blend_color, GL_EXT_blend_equation_separate, 
    GL_EXT_blend_func_separate, GL_EXT_blend_logic_op, GL_EXT_blend_minmax, 
    GL_EXT_blend_subtract, GL_EXT_clip_volume_hint, GL_EXT_cull_vertex, 
    GL_EXT_compiled_vertex_array, GL_EXT_copy_texture, 
    GL_EXT_draw_range_elements, GL_EXT_framebuffer_object, GL_EXT_fog_coord, 
    GL_EXT_multi_draw_arrays, GL_EXT_packed_depth_stencil, 
    GL_EXT_packed_pixels, GL_EXT_pixel_buffer_object, GL_EXT_point_parameters, 
    GL_EXT_polygon_offset, GL_EXT_rescale_normal, GL_EXT_secondary_color, 
    GL_EXT_separate_specular_color, GL_EXT_shadow_funcs, GL_EXT_stencil_wrap, 
    GL_EXT_subtexture, GL_EXT_texture, GL_EXT_texture3D, 
    GL_EXT_texture_edge_clamp, GL_EXT_texture_env_add, 
    GL_EXT_texture_env_combine, GL_EXT_texture_env_dot3, 
    GL_EXT_texture_filter_anisotropic, GL_EXT_texture_lod_bias, 
    GL_EXT_texture_object, GL_EXT_texture_rectangle, GL_EXT_vertex_array, 
    GL_3DFX_texture_compression_FXT1, GL_APPLE_client_storage, 
    GL_APPLE_packed_pixels, GL_ATI_blend_equation_separate, 
    GL_ATI_separate_stencil, GL_IBM_rasterpos_clip, 
    GL_IBM_texture_mirrored_repeat, GL_INGR_blend_func_separate, 
    GL_MESA_pack_invert, GL_MESA_ycbcr_texture, GL_MESA_window_pos, 
    GL_NV_blend_square, GL_NV_light_max_exponent, GL_NV_point_sprite, 
    GL_NV_texture_rectangle, GL_NV_texgen_reflection, GL_NV_vertex_program, 
    GL_NV_vertex_program1_1, GL_OES_read_format, GL_SGIS_generate_mipmap, 
    GL_SGIS_texture_border_clamp, GL_SGIS_texture_edge_clamp, 
    GL_SGIS_texture_lod, GL_SGIX_depth_texture, GL_SUN_multi_draw_arrays

visual  x  bf lv rg d st colorbuffer ax dp st accumbuffer  ms  cav
 id dep cl sp sz l  ci b ro  r  g  b  a bf th cl  r  g  b  a ns b eat
----------------------------------------------------------------------
0x21 24 tc  0 32  0 r  y  .  8  8  8  8  0 24  8  0  0  0  0  0 0 None
0x22 24 dc  0 32  0 r  y  .  8  8  8  8  0 24  8  0  0  0  0  0 0 None
0x55 32 tc  0 32  0 r  .  .  8  8  8  8  0  0  0  0  0  0  0  0 0 None

As you can see there is finally FBO and PBO support. Also glDrawPixels and
glReadPixels performance has increased. Driver behaviour hasn't changed though.

I still can't get into game with wine-0.9.61, but this time the black error
message box stays on screen. With Roderick's patch it's possible to get ingame.
With RenderTargetLockMode set to auto performance is still not acceptable for
real playing, but with readtex it's OK. Not fast but a GREAT improvement
compared to the older driver.

Does wine use PBO and FBO now by default, or do I have to somehow force the
usage, Roderick?

Vertical colored lines are gone, this also applies to AvP1. I still have the
problem that the menu doesn't appear when pressing ESC. It's simply not
rendered, but I can hear that it's there (the ambient sound changes).
That's with readtex.

With auto everything is correct, menu is rendered like it should.

Should I file a different bugreport for that one?

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #59 from Alexander Dorofeyev   2008-05-04 17:25:04 ---
(In reply to comment #58)
> Does wine use PBO and FBO now by default, or do I have to somehow force the
> usage, Roderick?

PBO - yes, FBO - no (there's OffscreenRenderingMode registry key for that).
> Vertical colored lines are gone, this also applies to AvP1. I still have the
> problem that the menu doesn't appear when pressing ESC. It's simply not
> rendered, but I can hear that it's there (the ambient sound changes).
> That's with readtex.
>
> With auto everything is correct, menu is rendered like it should.
>
> Should I file a different bugreport for that one?

I believe this is http://bugs.winehq.org/show_bug.cgi?id=12890. I think Thief
probably shares some code with SS2, as afaik same developers participated in
both. I've the same thing on nvidia, btw.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #60 from Roderick Colenbrander   2008-05-05 11:40:32 ---
Does it work properly using my patch? If so I'm going to submit it.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Tobias Jakobi  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #61 from Tobias Jakobi   2008-05-05 11:44:12 ---
Yes, the initial problem (black screen after loading sequence, or better
"invisible error message box") is gone.
What stays is the menu problem with readtex.

I'm going to set this one to fixed.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #62 from Roderick Colenbrander   2008-05-05 11:48:16 ---
Ok, I will submit it then. Note that normally a bug is marked as fixed after
the fix is in ... (just for next time)

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

--- Comment #63 from Roderick Colenbrander   2008-05-05 14:29:43 ---
It is in now.

[Bug 12557] System Shock 2 Demo: Starting game results in black screen

http://bugs.winehq.org/show_bug.cgi?id=12557

Roderick Colenbrander  changed:

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

--- Comment #64 from Roderick Colenbrander   2008-05-08 14:31:09 ---
Closing because the bug has been fixed.