Mesa (main): Revert "radeonsi/gfx11: limit MSAA color buffers to the RGBA channel order"

GitLab Mirror gitlab-mirror at kemper.freedesktop.org
Tue May 17 10:58:38 UTC 2022


Module: Mesa
Branch: main
Commit: c74d85434866f6340866940d8a07db70a28de063
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=c74d85434866f6340866940d8a07db70a28de063

Author: Marek Olšák <marek.olsak at amd.com>
Date:   Sat May 14 03:55:34 2022 -0400

Revert "radeonsi/gfx11: limit MSAA color buffers to the RGBA channel order"

This reverts commit 54d85700a12aa33d185e147c2d5c794ba54b7a11.

It's an LLVM bug. If you disable AMDGPUImageIntrinsicOptimizer in LLVM,
MSAA is fixed. There is no LLVM command line option to disable it from Mesa.

Reviewed-by: Mihai Preda <mhpreda at gmail.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer at amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16509>

---

 src/gallium/drivers/radeonsi/si_state.c | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/src/gallium/drivers/radeonsi/si_state.c b/src/gallium/drivers/radeonsi/si_state.c
index 53dd6fcca02..015cb277895 100644
--- a/src/gallium/drivers/radeonsi/si_state.c
+++ b/src/gallium/drivers/radeonsi/si_state.c
@@ -2452,16 +2452,6 @@ static bool si_is_format_supported(struct pipe_screen *screen, enum pipe_format
          if (sample_count > max_eqaa_samples || storage_sample_count > max_samples)
             return false;
       }
-
-      /* Gfx11: BGRA doesn't work with samples >= 4. Only allow R/0/1 to be the first
-       * component for simplicity.
-       */
-      if (sscreen->info.gfx_level >= GFX11 &&
-          !util_format_is_depth_or_stencil(format) &&
-          util_format_description(format)->swizzle[0] != PIPE_SWIZZLE_X &&
-          util_format_description(format)->swizzle[0] != PIPE_SWIZZLE_0 &&
-          util_format_description(format)->swizzle[0] != PIPE_SWIZZLE_1)
-         return false;
    }
 
    if (usage & (PIPE_BIND_SAMPLER_VIEW | PIPE_BIND_SHADER_IMAGE)) {



More information about the mesa-commit mailing list