Mesa (staging/22.1): aco: fix load_barycentric_at_{sample,offset} on GFX6-7

[GitLab Mirror]

Module: Mesa
Branch: staging/22.1
Commit: 2379bc19a8b7d2fd034cd72b10107756d9bce13b
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=2379bc19a8b7d2fd034cd72b10107756d9bce13b

Author: Samuel Pitoiset <samuel.pitoiset at gmail.com>
Date:   Fri Apr 22 15:15:33 2022 +0200

aco: fix load_barycentric_at_{sample,offset} on GFX6-7

The computation was wrong.

Fixes dEQP-GLES31.functional.shaders.multisample_interpolation.interpolate_at_offset.*
with Zink on GFX6 (Pitcairn).

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset at gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02 at gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16099>
(cherry picked from commit 6873da0e42182520a38a6f88abb3717eea12d774)

---

 .pick_status.json                              | 2 +-
 src/amd/compiler/aco_instruction_selection.cpp | 9 +++++----
 2 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/.pick_status.json b/.pick_status.json
index f5334d3fd1d..57a83ba900c 100644
--- a/.pick_status.json
+++ b/.pick_status.json
@@ -652,7 +652,7 @@
         "description": "aco: fix load_barycentric_at_{sample,offset} on GFX6-7",
         "nominated": true,
         "nomination_type": 0,
-        "resolution": 0,
+        "resolution": 1,
         "main_sha": null,
         "because_sha": null
     },
diff --git a/src/amd/compiler/aco_instruction_selection.cpp b/src/amd/compiler/aco_instruction_selection.cpp
index ac1a53bf5c5..7b0a43f0667 100644
--- a/src/amd/compiler/aco_instruction_selection.cpp
+++ b/src/amd/compiler/aco_instruction_selection.cpp
@@ -7843,11 +7843,12 @@ emit_interp_center(isel_context* ctx, Temp dst, Temp bary, Temp pos1, Temp pos2)
       Temp tl_1 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p1, (1 << 15) | dpp_ctrl0);
       ddx_1 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p1, (1 << 15) | dpp_ctrl1);
       ddx_1 = bld.vop2(aco_opcode::v_sub_f32, bld.def(v1), ddx_1, tl_1);
-      ddx_2 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p1, (1 << 15) | dpp_ctrl2);
-      ddx_2 = bld.vop2(aco_opcode::v_sub_f32, bld.def(v1), ddx_2, tl_1);
+      ddy_1 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p1, (1 << 15) | dpp_ctrl2);
+      ddy_1 = bld.vop2(aco_opcode::v_sub_f32, bld.def(v1), ddy_1, tl_1);
+
       Temp tl_2 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p2, (1 << 15) | dpp_ctrl0);
-      ddy_1 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p2, (1 << 15) | dpp_ctrl1);
-      ddy_1 = bld.vop2(aco_opcode::v_sub_f32, bld.def(v1), ddy_1, tl_2);
+      ddx_2 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p2, (1 << 15) | dpp_ctrl1);
+      ddx_2 = bld.vop2(aco_opcode::v_sub_f32, bld.def(v1), ddx_2, tl_2);
       ddy_2 = bld.ds(aco_opcode::ds_swizzle_b32, bld.def(v1), p2, (1 << 15) | dpp_ctrl2);
       ddy_2 = bld.vop2(aco_opcode::v_sub_f32, bld.def(v1), ddy_2, tl_2);
    }