[Mesa-dev] Mesa (master): draw: fix flat shading and screen-space linear interpolation in clipper

[Marek Olšák]

On Wed, Jul 11, 2012 at 5:26 PM, Jose Fonseca <jfonseca at vmware.com> wrote:
>
>
> ----- Original Message -----
>> On Wed, Jul 11, 2012 at 4:23 PM, Jose Fonseca <jfonseca at vmware.com>
>> wrote:
>> >
>> > When NativeIntegers is FALSE, Mesa generates this VS:
>> >
>> >   DCL SV[0], INSTANCEID
>> >   ...
>> >   ARL ADDR[0].x, SV[0].xxxx
>> >
>> > If INSTANCEID is an integer this is wrong, as it is reading a float
>> > where there's an integer. (e.g., for INSTANCEID =1 ARRD[0] would
>> > have round(0.0000000001) zero).
>> >
>> > But, if I modify mesa statetracker to insert the U2F as necessary
>> > evertyhing will work as expected
>> >
>> >   DCL SV[0], INSTANCEID
>> >   ...
>> >   U2F TEMP[0].x, SV[0].xxxx
>> >   ARL ADDR[0].x, TEMP[0].xxxx
>> >
>> > This is essencially what Olivier patch did, but without guessing.
>> > Mesa state tracker can known everything it needs to do the "right
>> > thing", i.e., INSTANCEID is integer.
>>
>> I see. However if PIPE_SHADER_CAP_INTEGERS is 0, I2F and U2F are
>> unsupported instructions.
>
> No.
>
>   PIPE_SHADER_CAP_INTEGERS is 0 for *PS* (i915)
>
>   PIPE_SHADER_CAP_INTEGERS is 1 for *VS* (draw)
>
> Therefore a I2F/U2F is not an unsupported function in draw.
>
> Everything's just fine AFAICT.
>
> I really don't want TGSI semantics to vary (having draw calling PIPE_SHADER_CAP_INTEGERS to understand the meaning of INSTANCEID).

Ah, I see. It's fine then.

Marek