[PATCH v3 0/2] MR728: vkd3d-shader/dxil: Implement DX intrinsics Dot2, Dot3 and Dot4.

[Conor McCarthy (＠cmccarthy)]

--
  v3: vkd3d-shader/dxil: Implement DX intrinsics Dot2, Dot3 and Dot4.
      vkd3d-shader/dxil: Emit an error if a constant code is unhandled.

https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/728

[PATCH v3 1/2] vkd3d-shader/dxil: Emit an error if a constant code is unhandled.

[Conor McCarthy]

From: Conor McCarthy

---
 libs/vkd3d-shader/dxil.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/libs/vkd3d-shader/dxil.c b/libs/vkd3d-shader/dxil.c
index 7f9a74fa7..b3e42c158 100644
--- a/libs/vkd3d-shader/dxil.c
+++ b/libs/vkd3d-shader/dxil.c
@@ -2960,6 +2960,8 @@ static enum vkd3d_result sm6_parser_constants_init(struct sm6_parser *sm6, const

default:
                 FIXME("Unhandled constant code %u.\n", record->code);
+                vkd3d_shader_parser_error(&sm6->p, VKD3D_SHADER_ERROR_DXIL_INVALID_OPERAND,
+                        "Constant code %u is unhandled.", record->code);
                 dst->u.reg.type = VKD3DSPR_UNDEF;
                 break;
         }

[PATCH v3 2/2] vkd3d-shader/dxil: Implement DX intrinsics Dot2, Dot3 and Dot4.

[Conor McCarthy]

From: Conor McCarthy

---
 libs/vkd3d-shader/dxil.c         | 48 ++++++++++++++++++++++++++++++++
 tests/hlsl/dot.shader_test       |  8 +++---
 tests/hlsl/normalize.shader_test |  6 ++--
 tests/hlsl/reflect.shader_test   | 10 +++----
 4 files changed, 60 insertions(+), 12 deletions(-)

diff --git a/libs/vkd3d-shader/dxil.c b/libs/vkd3d-shader/dxil.c
index b3e42c158..a2bf55859 100644
--- a/libs/vkd3d-shader/dxil.c
+++ b/libs/vkd3d-shader/dxil.c
@@ -382,6 +382,9 @@ enum dx_intrinsic_opcode
     DX_UMAD                         =  49,
     DX_IBFE                         =  51,
     DX_UBFE                         =  52,
+    DX_DOT2                         =  54,
+    DX_DOT3                         =  55,
+    DX_DOT4                         =  56,
     DX_CREATE_HANDLE                =  57,
     DX_CBUFFER_LOAD_LEGACY          =  59,
     DX_SAMPLE                       =  60,
@@ -4275,6 +4278,48 @@ static void sm6_parser_emit_dx_create_handle(struct sm6_parser *sm6, enum dx_int
     ins->handler_idx = VKD3DSIH_NOP;
 }

+static void sm6_parser_emit_dx_dot(struct sm6_parser *sm6, enum dx_intrinsic_opcode op,
+        const struct sm6_value **operands, struct function_emission_state *state)
+{
+    struct vkd3d_shader_src_param *src_params;
+    struct vkd3d_shader_instruction *ins;
+    struct vkd3d_shader_register regs[2];
+    enum vkd3d_shader_opcode handler_idx;
+    unsigned int component_count;
+
+    switch (op)
+    {
+        case DX_DOT2:
+            handler_idx = VKD3DSIH_DP2;
+            component_count = 2;
+            break;
+        case DX_DOT3:
+            handler_idx = VKD3DSIH_DP3;
+            component_count = 3;
+            break;
+        case DX_DOT4:
+            handler_idx = VKD3DSIH_DP4;
+            component_count = 4;
+            break;
+        default:
+            vkd3d_unreachable();
+    }
+
+    if (!sm6_parser_emit_composite_construct(sm6, &operands[0], component_count, state, &regs[0]))
+        return;
+    if (!sm6_parser_emit_composite_construct(sm6, &operands[component_count], component_count, state, &regs[1]))
+        return;
+
+    ins = state->ins;
+    vsir_instruction_init(ins, &sm6->p.location, handler_idx);
+    if (!(src_params = instruction_src_params_alloc(ins, 2, sm6)))
+        return;
+    src_param_init_vector_from_reg(&src_params[0], &regs[0]);
+    src_param_init_vector_from_reg(&src_params[1], &regs[1]);
+
+    instruction_dst_param_init_ssa_scalar(ins, sm6);
+}
+
 static void sm6_parser_emit_dx_fabs(struct sm6_parser *sm6, enum dx_intrinsic_opcode op,
         const struct sm6_value **operands, struct function_emission_state *state)
 {
@@ -5088,6 +5133,9 @@ static const struct sm6_dx_opcode_info sm6_dx_op_table[] =
     [DX_DERIV_COARSEY                 ] = {"e", "R",    sm6_parser_emit_dx_unary},
     [DX_DERIV_FINEX                   ] = {"e", "R",    sm6_parser_emit_dx_unary},
     [DX_DERIV_FINEY                   ] = {"e", "R",    sm6_parser_emit_dx_unary},
+    [DX_DOT2                          ] = {"g", "RRRR", sm6_parser_emit_dx_dot},
+    [DX_DOT3                          ] = {"g", "RRRRRR", sm6_parser_emit_dx_dot},
+    [DX_DOT4                          ] = {"g", "RRRRRRRR", sm6_parser_emit_dx_dot},
     [DX_EXP                           ] = {"g", "R",    sm6_parser_emit_dx_unary},
     [DX_FABS                          ] = {"g", "R",    sm6_parser_emit_dx_fabs},
     [DX_FIRST_BIT_HI                  ] = {"i", "m",    sm6_parser_emit_dx_unary},
diff --git a/tests/hlsl/dot.shader_test b/tests/hlsl/dot.shader_test
index c6a96cc1e..73c29512d 100644
--- a/tests/hlsl/dot.shader_test
+++ b/tests/hlsl/dot.shader_test
@@ -10,7 +10,7 @@ float4 main() : SV_TARGET
 [test]
 uniform 0 float4 2.0 3.0 4.0 5.0
 uniform 4 float4 10.0 11.0 12.0 13.0
-todo(sm>=6) draw quad
+draw quad
 probe all rgba (166.0, 166.0, 166.0, 166.0)

[pixel shader]
@@ -25,7 +25,7 @@ float4 main() : SV_TARGET
 [test]
 uniform 0 float4 2.0 3.0 0.0 0.0
 uniform 4 float4 10.0 11.0 12.0 13.0
-todo(sm<4 | sm>=6) draw quad
+todo(sm<4) draw quad
 probe all rgba (53.0, 53.0, 53.0, 53.0)

[pixel shader]
@@ -40,7 +40,7 @@ float4 main() : SV_TARGET
 [test]
 uniform 0 float4 2.0 0.0 0.0 0.0
 uniform 4 float4 10.0 11.0 12.0 13.0
-todo(sm>=6) draw quad
+draw quad
 probe all rgba (92.0, 92.0, 92.0, 92.0)

[pixel shader]
@@ -55,7 +55,7 @@ float4 main() : SV_TARGET
 [test]
 uniform 0 float4 10.0 11.0 12.0 13.0
 uniform 4 float4 2.0 0.0 0.0 0.0
-todo(sm>=6) draw quad
+draw quad
  probe all rgba (92.0, 92.0, 92.0, 92.0)

[pixel shader]
diff --git a/tests/hlsl/normalize.shader_test b/tests/hlsl/normalize.shader_test
index 9f350a7ee..1712b639a 100644
--- a/tests/hlsl/normalize.shader_test
+++ b/tests/hlsl/normalize.shader_test
@@ -8,7 +8,7 @@ float4 main() : SV_TARGET

[test]
 uniform 0 float4 2.0 3.0 4.0 5.0
-todo(sm>=6) draw quad
+draw quad
 probe all rgba (0.272165537, 0.408248305, 0.544331074, 0.680413842) 2

[pixel shader]
@@ -21,7 +21,7 @@ float4 main() : SV_TARGET

[test]
 uniform 0 float4 2.0 3.0 4.0 0.0
-todo(sm>=6) draw quad
+draw quad
 probe all rgba (0.371390700, 0.557086051, 0.742781401, 0.0) 1

[pixel shader]
@@ -34,7 +34,7 @@ float4 main() : SV_TARGET

[test]
 uniform 0 float4 2.0 3.0 0.0 0.0
-todo(sm<4 | sm>=6) draw quad
+todo(sm<4) draw quad
 probe all rgba (0.554700196, 0.832050323, 0.0, 0.0) 1

[pixel shader]
diff --git a/tests/hlsl/reflect.shader_test b/tests/hlsl/reflect.shader_test
index 03772dfa9..34b800f2b 100644
--- a/tests/hlsl/reflect.shader_test
+++ b/tests/hlsl/reflect.shader_test
@@ -10,7 +10,7 @@ float4 main() : sv_target
 [test]
 uniform 0 float4 0.5 -0.1 0.2 0.3
 uniform 4 float4 0.6 0.4 -0.3 1.0
-todo(sm>=6) draw quad
+draw quad
 probe all rgba (-0.1, -0.5, 0.5, -0.7) 4

[pixel shader]
@@ -27,7 +27,7 @@ float4 main() : sv_target
 [test]
 uniform 0 float4 0.5 0.0 0.0 0.0
 uniform 4 float4 0.6 0.4 -0.3 1.0
-todo(sm>=6) draw quad
+draw quad
 probe all rgba (-0.52, -0.18, 1.01, -1.2) 4

[pixel shader]
@@ -44,7 +44,7 @@ float4 main() : sv_target
 [test]
 uniform 0 float4 0.5 -0.1 0.2 0.3
 uniform 4 float4 0.6 0.0 0.0 0.0
-todo(sm>=6) draw quad
+draw quad
 probe all rgba (-0.148, -0.748, -0.448, -0.348) 4

[pixel shader]
@@ -79,7 +79,7 @@ float4 main() : sv_target
 [test]
 uniform 0 float4 0.5 -0.1 0.0 0.0
 uniform 4 float4 0.6 0.4 -0.3 1.0
-todo(sm<4 | sm>=6) draw quad
+todo(sm<4) draw quad
 probe all rgba (0.188, -0.308, 0.0, 0.0) 4

[pixel shader]
@@ -97,5 +97,5 @@ float4 main() : sv_target
 [test]
 uniform 0 float4 0.5 -0.1 0.2 0.0
 uniform 4 float4 0.6 0.4 0.0 0.0
-todo(sm<4 | sm>=6) draw quad
+todo(sm<4) draw quad
 probe all rgba (0.188, -0.308, 0.0, 0.0) 4

[PATCH v3 0/2] MR728: vkd3d-shader/dxil: Implement DX intrinsics Dot2, Dot3 and Dot4. - approved

[Giovanni Mascellani (＠giomasce)]

This merge request was approved by Giovanni Mascellani.

[Conor McCarthy (＠cmccarthy)]

The first commit is to prevent ShadowCI crashing because of an issue uncovered by adding Dot support, or perhaps something already upstream.