It seems like this still doesn't quite fix things - have a look at
fbo-mrt-alphatest in piglit. I had assumed this change would do the
trick, but no such luck.
On Wed, Nov 9, 2016 at 10:18 PM, Tim Rowley wrote:
> Alpha from render target 0 should always be used for alpha test for all
> render tar
Right you are - I must have imagined the third place.
On Thu, Nov 10, 2016 at 1:06 AM, Rowley, Timothy O
wrote:
> Ah, yes, this patch missed the 8x2 tile path - I’ve fixed that now. I don’t
> see another path to using the blend jit functions.
>
> Thanks.
>
> On Nov 9, 2016, at 10:44 PM, Ilia Mir
Ah, yes, this patch missed the 8x2 tile path - I’ve fixed that now. I don’t
see another path to using the blend jit functions.
Thanks.
On Nov 9, 2016, at 10:44 PM, Ilia Mirkin
mailto:imir...@alum.mit.edu>> wrote:
I think a second instance of the blend func being called was missed in
backend
I think a second instance of the blend func being called was missed in
backend.h (the avx512 one). Also iirc there was so.e third place where it
was called... Should grep for pfnBlendFunc and see if any other instances
pop up.
On Nov 9, 2016 10:19 PM, "Tim Rowley" wrote:
> Alpha from render targ
Alpha from render target 0 should always be used for alpha test for all
render targets, according to GL and DX9 specs. Previously we were using
alpha from the current render target.
---
src/gallium/drivers/swr/rasterizer/core/backend.h | 1 +
src/gallium/drivers/swr/rasterizer/core/state.h
