[Bug 46435] New: F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

Bug ID: 46435
           Summary: F1 2011 crashes with a messed up rendering in DX9 mode
           Product: Wine
           Version: 4.0-rc5
          Hardware: x86-64
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: -unknown
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

Created attachment 63213
  --> https://bugs.winehq.org/attachment.cgi?id=63213
backtraces

d3dx9_43 - builtin
fixme:d3dx:D3DXGetShaderConstantTableEx Flags (0x20000) are not handled, yet!

Crashes immediately after race starts. Backtrace is different each time.

Rendering is normal and no crash with d3dx9_43 native,builtin

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

--- Comment #1 from Andrey Gusev  ---
Created attachment 63214
  --> https://bugs.winehq.org/attachment.cgi?id=63214
screenshot

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

Paul Gofman  changed:

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

--- Comment #2 from Paul Gofman  ---
Created attachment 63230
  --> https://bugs.winehq.org/attachment.cgi?id=63230
Fix comment length in get_instr_length()

I tested the game and could reproduce broken rendering with builtin d3dx9_43
but not the crash. Broken rendering is due to incorrect
D3DXGetShaderInputSemantics() results. I am attaching the patch which fixes
this issue. Can you test it with this patch to see if the crash goes away too?
In case the crash is still there, and is not there with native d3dx9_43, can
you please upload WINEDEBUG=+d3dx log?

As a separate note, fixme from D3DXGetShaderConstantTableEx() is not related to
this problem. The flag it complains about (the only one documented to this
function) is D3DXCONSTTABLE_LARGEADDRESSAWARE. I don't think this flag should
actually affect this function behavior, my bet the fixme should be downgraded
to warn.

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

Andrey Gusev  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|-unknown                    |directx-d3dx9
           See Also|                            |https://bugs.winehq.org/sho
                   |                            |w_bug.cgi?id=46007

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

--- Comment #3 from Matteo Bruni  ---
Thank you Paul for the patch, I just resent it unchanged except for the new
standard "Wine-Bug:" tag.

WRT the FIXME, I don't think it's useless. We don't have tests for it at the
moment but I expect that functions like GetConstantDesc() shouldn't accept
strings as handles when the ID3DXConstantTable object is created with the
D3DXCONSTTABLE_LARGEADDRESSAWARE flag, in the same vein as the effect
framework.

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

Andrey Gusev  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
      Fixed by SHA1|                            |807e5fc04af7a7ea563af1e7da6
                   |                            |ebe6662536e6b
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #4 from Andrey Gusev  ---
Fixed by
https://source.winehq.org/git/wine.git/commit/807e5fc04af7a7ea563af1e7da6ebe...

Thanks, Paul.

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

--- Comment #5 from Paul Gofman  ---
(In reply to Matteo Bruni from comment #3)
> Thank you Paul for the patch, I just resent it unchanged except for the new
> standard "Wine-Bug:" tag.

Sorry forgot that.
> WRT the FIXME, I don't think it's useless. We don't have tests for it at the
> moment but I expect that functions like GetConstantDesc() shouldn't accept
> strings as handles when the ID3DXConstantTable object is created with the
> D3DXCONSTTABLE_LARGEADDRESSAWARE flag, in the same vein as the effect
> framework.

Oh yeah, I missed that ID3DXConstantTable has this string / reference dualism
too.

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

Alexandre Julliard  changed:

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

--- Comment #6 from Alexandre Julliard  ---
Closing bugs fixed in 4.0-rc6.

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

Michael Stefaniuc  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |3.0.x

[Bug 46435] F1 2011 crashes with a messed up rendering in DX9 mode

https://bugs.winehq.org/show_bug.cgi?id=46435

Michael Stefaniuc  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.0.x                       |---

--- Comment #7 from Michael Stefaniuc  ---
Removing the 3.0.x milestone from bug fixes included in 3.0.5.