[PATCH vkd3d 3/5] vkd3d-shader/hlsl: Write SM4 negation instructions.

Zebediah Figura (she/her) zfigura at codeweavers.com
Wed Sep 1 09:16:03 CDT 2021


On 9/1/21 3:57 AM, Matteo Bruni wrote:
> On Wed, Sep 1, 2021 at 10:40 AM Henri Verbeet <hverbeet at gmail.com> wrote:
>>
>> On Wed, 1 Sept 2021 at 10:27, Matteo Bruni <matteo.mystral at gmail.com> wrote:
>>> On Wed, Sep 1, 2021 at 3:24 AM Zebediah Figura (she/her)
>>> <zfigura at codeweavers.com> wrote:
>>>> I recently discovered that the SM4 format is effectively documented in
>>>> headers released with the DDK, namely "d3d10tokenizedprogramformat.hpp"
>>>> and similar. Anyway, their version of SM4_REGISTER_ORDER_* is
>>>> D3D10_SB_OPERAND_INDEX_DIMENSION_*.
>>>
>>> Oh, excellent! I'll have a look at some point.
>>> No reason to change everything now, but going forward it's probably a
>>> good idea to make more informed decisions so to use the same naming
>>> principles as the original.
>>>
>> This has been around for a long time. We have generally avoided
>> looking at the DDK, and I can't think of any cases where looking at it
>> would have helped much.
> 
> If it's dangerous then sure. I'm certain we don't NEED anything that
> might be in there.
> 
> Otherwise, I don't see the problem with at least using the correct
> names. It might just be that I recently discovered the effect
> framework's "preshaders" are supposed to be called FXLVM and I liked
> it a bit too much...
> 

For what it's worth, the d3d12tokenizedprogramformat.hpp also from the 
DDK has the same definitions, and seems to be the same file as is 
shipped with the DXC source [1], and that at least is explicitly under 
an open license.

[1] 
https://github.com/microsoft/DirectXShaderCompiler/blob/cb485263b75495015e2a77cc438e72dc813ea287/include/dxc/Support/d3d12TokenizedProgramFormat.hpp



More information about the wine-devel mailing list