Francisco Casas : vkd3d-shader/hlsl: Fix SM1 dp2add swizzles.

[Alexandre Julliard]

Module: vkd3d
Branch: master
Commit: 902ddee557f3b2c0e5bb668c7330b79b7954ebce
URL:    https://gitlab.winehq.org/wine/vkd3d/-/commit/902ddee557f3b2c0e5bb668c7330b7...

Author: Francisco Casas 
Date:   Fri Feb  3 12:17:26 2023 -0300

vkd3d-shader/hlsl: Fix SM1 dp2add swizzles.

SM1 dp2add doesn't map src swizzles to the dst writemask, also it
expects the last argument to have a replicate swizzle.

Before this patch we were writing the operation as:
```
dp2add r0.x, r1.x, r0.x, r2.x
```

and now it is:
```
dp2add r0.x, r1.xyxx, r0.xyxx, r2.x
```

dp2add now has its own function, write_sm1_dp2add(), since it seems to
be the only instruction with this structure.

Ideally we would be using the default swizzles for the first two src
arguments:
```
dp2add r0.x, r1, r0, r2.x
```
since, according to native's documentation, these are supported for all
sm < 4.

But this change -- along with following the convention of repeating the
last component of the swizzle when fewer than 4 components are to be
specified -- would require more global changes, probably in
hlsl_swizzle_from_writemask() and hlsl_map_swizzle().

---

libs/vkd3d-shader/hlsl_sm1.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/libs/vkd3d-shader/hlsl_sm1.c b/libs/vkd3d-shader/hlsl_sm1.c
index ebeaadb1..34bead54 100644
--- a/libs/vkd3d-shader/hlsl_sm1.c
+++ b/libs/vkd3d-shader/hlsl_sm1.c
@@ -462,13 +462,13 @@ static void sm1_map_src_swizzle(struct sm1_src_register *src, unsigned int map_w
     src->swizzle = hlsl_map_swizzle(src->swizzle, map_writemask);
 }

-static void write_sm1_ternary_op(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *buffer,
-        D3DSHADER_INSTRUCTION_OPCODE_TYPE opcode, const struct hlsl_reg *dst,
-        const struct hlsl_reg *src1, const struct hlsl_reg *src2, const struct hlsl_reg *src3)
+static void write_sm1_dp2add(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *buffer,
+        const struct hlsl_reg *dst, const struct hlsl_reg *src1, const struct hlsl_reg *src2,
+        const struct hlsl_reg *src3)
 {
     struct sm1_instruction instr =
     {
-        .opcode = opcode,
+        .opcode = D3DSIO_DP2ADD,

.dst.type = D3DSPR_TEMP,
         .dst.writemask = dst->writemask,
@@ -487,9 +487,6 @@ static void write_sm1_ternary_op(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buf
         .src_count = 3,
     };

-    sm1_map_src_swizzle(&instr.srcs[0], instr.dst.writemask);
-    sm1_map_src_swizzle(&instr.srcs[1], instr.dst.writemask);
-    sm1_map_src_swizzle(&instr.srcs[2], instr.dst.writemask);
     write_sm1_instruction(ctx, buffer, &instr);
 }

@@ -742,7 +739,7 @@ static void write_sm1_expr(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *b
             break;

case HLSL_OP3_DP2ADD:
-            write_sm1_ternary_op(ctx, buffer, D3DSIO_DP2ADD, &instr->reg, &arg1->reg, &arg2->reg, &arg3->reg);
+            write_sm1_dp2add(ctx, buffer, &instr->reg, &arg1->reg, &arg2->reg, &arg3->reg);
             break;

default: