[Mesa-dev] [PATCH 11/33] intel/blorp: Use fake_interleaved_msaa in retile_w_to_y

[Pohjolainen, Topi]

On Thu, Sep 01, 2016 at 12:02:17PM -0700, Jason Ekstrand wrote:
>    On Thu, Sep 1, 2016 at 12:09 AM, Pohjolainen, Topi
>    <[1]topi.pohjolainen at gmail.com> wrote:
> 
>      On Wed, Aug 31, 2016 at 02:22:30PM -0700, Jason Ekstrand wrote:
>      > ---
>      >  src/intel/blorp/blorp_blit.c | 4 +---
>      >  1 file changed, 1 insertion(+), 3 deletions(-)
>      >
>      > diff --git a/src/intel/blorp/blorp_blit.c
>      b/src/intel/blorp/blorp_blit.c
>      > index 05fad8f..2838a26 100644
>      > --- a/src/intel/blorp/blorp_blit.c
>      > +++ b/src/intel/blorp/blorp_blit.c
>      > @@ -1342,9 +1342,7 @@ surf_retile_w_to_y(const struct isl_device
>      *isl_dev,
>      >      */
>      >     if (isl_dev->info->gen > 6 &&
>      >         info->surf.msaa_layout == ISL_MSAA_LAYOUT_INTERLEAVED) {
>      > -      info->surf.logical_level0_px = info->surf.phys_level0_sa;
>      > -      info->surf.samples = 1;
>      > -      info->surf.msaa_layout = ISL_MSAA_LAYOUT_NONE;
>      > +      surf_fake_interleaved_msaa(isl_dev, info);
>      So now we start calling surf_convert_to_single_slice() twice,
>      surf_retile_w_to_y() itself calls it and then
>      surf_fake_interleaved_msaa()
>      again. I'll read ahead a little how this evolves...
> 
>    Yes, we do.  That *was* supposed to be safe, but actually wasn't due to
>    a mistake made a while ago.  As of an earlier patch in the series, it
>    is safe.

Could we this instead, or do we really need surf_convert_to_single_slice()
to be called twice?

     if (isl_dev->info->gen > 6 &&
         info->surf.msaa_layout == ISL_MSAA_LAYOUT_INTERLEAVED) {
        surf_fake_interleaved_msaa(isl_dev, info); 
     } else {
        surf_convert_to_single_slice(isl_dev, info);
     }

>      >     }
>      >
>      >     if (isl_dev->info->gen == 6) {
>      > --
>      > 2.5.0.400.gff86faf
>      >
>      > _______________________________________________
>      > mesa-dev mailing list
>      > [2]mesa-dev at lists.freedesktop.org
>      > [3]https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> 
> References
> 
>    1. mailto:topi.pohjolainen at gmail.com
>    2. mailto:mesa-dev at lists.freedesktop.org
>    3. https://lists.freedesktop.org/mailman/listinfo/mesa-dev