wine users - May 2009 - Running 3D in Virtual Desktop Mode freezes Nvidia hardware

Greetings.

This is a problem that just started 2 days ago. Neither Wine, nor any config has
changed. I even played back an old system backup from when it worked to no
avail.

Basically, the problem is this: When 2D apps run in Virtual Desktop Mode, they
run fine. As soon as the apps switches to 3D, or is a 3D app to begin with, the
following can be observed:

- app works fine for about 10 seconds, displaying normal 3D graphics
- Then, app freezes all Video adapter output. This does include the whole
screen, not just the contents of the Virtual Desktop!
- After about 20 more seconds, I can visibly move the mouse again. Everything
else is still frozen.
- After a few more seconds, everything seems to resume. The game continues to
display 3D content, but with graphical artifacts.
- Finally, the app freezes and the rest of the screen freezes again, never to
return.

At this point, I am usually hammering my Winekill keyboard shortcut (executes
wineserver -k), but to no avail.
The system is still responding tho. Magic SysRq keys work and I can log in via
SSH. I don't know about the other keys, as I only have frozen visual output.
Trying to switch the vt does not seem to have any effect.
The only thing I can do at this point is sync my harddrives and reboot.

I redirected the output from one app into a file, to see if there's any
suspicious logs. This is the result:

Code:
err:ddraw:PixelFormat_WineD3DtoDD Can't translate this Pixelformat 19
err:ddraw:PixelFormat_WineD3DtoDD Can't translate this Pixelformat 64
err:ddraw:PixelFormat_WineD3DtoDD Can't translate this Pixelformat 70
fixme:d3d_surface:flush_to_framebuffer_drawpixels
>>>>>>>>>>>>>>>>>
GL_INVALID_OPERATION (0x502) from glDrawBuffer(GL_BACK) @ surface.c / 1363
fixme:d3d:apply_draw_buffer
>>>>>>>>>>>>>>>>>
GL_INVALID_OPERATION (0x502) from glDrawBuffers() @ context.c / 1464
fixme:d3d:apply_draw_buffer
>>>>>>>>>>>>>>>>>
GL_INVALID_OPERATION (0x502) from glDrawBuffers() @ context.c / 1464
fixme:d3d:apply_draw_buffer
>>>>>>>>>>>>>>>>>
GL_INVALID_OPERATION (0x502) from glDrawBuffers() @ context.c / 1464
fixme:ddraw:IDirectDrawImpl_WaitForVerticalBlank (0x1379e8)->(1,(nil)): Stub
fixme:quartz:AVISplitter_InitializeStreams stream 1: frames found: 406272,
frames meant to be found: 203136
err:d3d:resource_init Out of memory!
err:ddraw:IDirectDrawImpl_CreateNewSurface IWineD3DDevice::CreateSurface failed.
hr = 8876017c
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 8876017c
err:d3d:resource_init Out of memory!
err:ddraw:IDirectDrawImpl_CreateNewSurface IWineD3DDevice::CreateSurface failed.
hr = 8876017c
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 8876017c
fixme:quartz:AVISplitter_InitializeStreams stream 1: frames found: 406272,
frames meant to be found: 203136
fixme:d3d:apply_draw_buffer
>>>>>>>>>>>>>>>>>
GL_INVALID_OPERATION (0x502) from glDrawBuffers() @ context.c / 1464



"out of memory" looks suspiciously like a Memory leak. The games I
tested are all pretty lightweight, and surely wouldn't fill the Video memory
under normal conditions.
The last GL_INVALID_OPERATION always seems to be present aswell.

Distro: Gentoo AMD64
Card: Geforce 8800GTS 320MB
Driver: 180.29

When playing native 3D games, this problem does not occur. Not even if I play
them in windowed mode.

Please don't tell me "not to use Virtual Desktop Mode then". I
have a similiar poblem when playing WINE games in normal fullscreen, but it
usually lasts a bit longer. I'm currently trying to gather some more info on
that and will post when I'm done.

Again, this does not seem to be a problem related to Software. A backup made 3
days ago (the problem only started appearing 2 days ago) results in the same
behavior. I already suspected that video card may be bonkers for a while now,
but I don't see why native games work just fine then...
Does anyone have any idea what could be the problem here? Any help will be
appreciated

Ok, the same game in Fullscreen results in similiar console log. Except in this
case, it won't freeze the Video card, but bothers to crash properly :)


Code:
err:ddraw:PixelFormat_WineD3DtoDD Can't translate this Pixelformat 64
err:ddraw:PixelFormat_WineD3DtoDD Can't translate this Pixelformat 70
fixme:ddraw:IDirectDrawImpl_WaitForVerticalBlank (0x137a88)->(1,(nil)): Stub
fixme:quartz:AVISplitter_InitializeStreams stream 1: frames found: 406272,
frames meant to be found: 203136
err:d3d:resource_init Out of memory!
err:ddraw:IDirectDrawImpl_CreateNewSurface IWineD3DDevice::CreateSurface failed.
hr = 8876017c
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 8876017c
err:d3d:resource_init Out of memory!
err:ddraw:IDirectDrawImpl_CreateNewSurface IWineD3DDevice::CreateSurface failed.
hr = 8876017c
err:ddraw:IDirectDrawImpl_CreateSurface IDirectDrawImpl_CreateNewSurface failed
with 8876017c
fixme:quartz:AVISplitter_InitializeStreams stream 1: frames found: 406272,
frames meant to be found: 203136
fixme:d3d_surface:surface_upload_data
>>>>>>>>>>>>>>>>>
GL_INVALID_VALUE (0x501) from glTexSubImage2D @ surface.c / 349
fixme:d3d_surface:surface_upload_data
>>>>>>>>>>>>>>>>>
GL_INVALID_VALUE (0x501) from glTexSubImage2D @ surface.c / 349
Xlib: sequence lost (0x10000 > 0xa925) in reply type 0x0!
X Error:  0
  Request Major code 0 ()
  Error Serial #0
  Current Serial #43301

haarp wrote:
> Driver: 180.29
These are known bad, upgrade.

Probably this bug: http://bugs.winehq.org/show_bug.cgi?id=13335

Apply latest mmap and malloc patches, which are attachments in that bug report.