[Bug 38545] New: Calculate surface size is wrong sometimes

https://bugs.winehq.org/show_bug.cgi?id=38545

Bug ID: 38545
           Summary: Calculate surface size is wrong sometimes
           Product: Wine
           Version: 1.7.42
          Hardware: x86
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: directx-d3dx9
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

Bug 38178 is closed resolved by prohibit DXTn texture. 
The real reason for the game crash is wrong sizes calculation in case of
compressed textures.
Even if we are not supporting DXTn compress/decompress we should allow the
application do this by own libraries and we must correctly calculate sizes in
those cases.
First what I see is the procedure calculate_dds_surface_size(..)
It adopted to block based textures but not to compressed textures
According to MSDN
https://msdn.microsoft.com/en-us/library/windows/desktop/bb205578(v=vs.85).a...
the formula should be
max(1, ( (width + 3) / 4 ) ) x max(1, ( (height + 3) / 4 ) ) x 8(DXT1) or
16(DXT2-5)

But I see there are more caveats...

[Bug 38545] Calculate surface size is wrong sometimes

https://bugs.winehq.org/show_bug.cgi?id=38545

Józef Kucia  changed:

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

--- Comment #1 from Józef Kucia  ---
(In reply to Sergey Isakov from comment #0)
> First what I see is the procedure calculate_dds_surface_size(..)
> It adopted to block based textures but not to compressed textures
> According to MSDN
> https://msdn.microsoft.com/en-us/library/windows/desktop/bb205578(v=vs.85).
> aspx
> the formula should be
> max(1, ( (width + 3) / 4 ) ) x max(1, ( (height + 3) / 4 ) ) x 8(DXT1) or
> 16(DXT2-5)

AFAIK calculate_dds_surface_size() and calculate_dds_file_size() should return
the correct size for all D3DFMT formats. These functions also handle block
compression formats, i.e. DXTn. Could you provide an example when the
calculated size is wrong?

[Bug 38545] Calculate surface size is wrong sometimes

https://bugs.winehq.org/show_bug.cgi?id=38545

--- Comment #2 from Sergey Isakov  ---
Agree, the procedures are correct. The formulae is different but the result is
good.

[Bug 38545] Calculate surface size is wrong sometimes

https://bugs.winehq.org/show_bug.cgi?id=38545

--- Comment #3 from Józef Kucia  ---
(In reply to Sergey Isakov from comment #2)
> Agree, the procedures are correct. The formulae is different but the result
> is good.

We should close this bug as invalid, then.

[Bug 38545] Calculate surface size is wrong sometimes

https://bugs.winehq.org/show_bug.cgi?id=38545

Matteo Bruni  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |INVALID

--- Comment #4 from Matteo Bruni  ---
(In reply to Józef Kucia from comment #3)
> (In reply to Sergey Isakov from comment #2)
> > Agree, the procedures are correct. The formulae is different but the result
> > is good.
>
> We should close this bug as invalid, then.

Resolving INVALID.

[Bug 38545] Calculate surface size is wrong sometimes

https://bugs.winehq.org/show_bug.cgi?id=38545

Bruno Jesus  changed:

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

--- Comment #5 from Bruno Jesus  ---
Closing invalid bugs.