[PATCH 0/1] MR79: vkd3d-shader/hlsl: Handle saturate() for SM1.

[Nikolay Sivov (＠nsivov)]

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com>

-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79

[PATCH 1/1] vkd3d-shader/hlsl: Handle saturate() for SM1.

[Nikolay Sivov]

From: Nikolay Sivov <nsivov(a)codeweavers.com>

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com>
---
 libs/vkd3d-shader/hlsl_sm1.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/libs/vkd3d-shader/hlsl_sm1.c b/libs/vkd3d-shader/hlsl_sm1.c
index a09d7566..60b54a42 100644
--- a/libs/vkd3d-shader/hlsl_sm1.c
+++ b/libs/vkd3d-shader/hlsl_sm1.c
@@ -512,13 +512,14 @@ static void write_sm1_binary_op(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buff
 
 static void write_sm1_unary_op(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *buffer,
         D3DSHADER_INSTRUCTION_OPCODE_TYPE opcode, const struct hlsl_reg *dst,
-        const struct hlsl_reg *src, D3DSHADER_PARAM_SRCMOD_TYPE src_mod)
+        const struct hlsl_reg *src, D3DSHADER_PARAM_SRCMOD_TYPE src_mod, D3DSHADER_PARAM_DSTMOD_TYPE dst_mod)
 {
     const struct sm1_instruction instr =
     {
         .opcode = opcode,
 
         .dst.type = D3DSPR_TEMP,
+        .dst.mod = dst_mod,
         .dst.writemask = dst->writemask,
         .dst.reg = dst->id,
         .has_dst = 1,
@@ -649,7 +650,7 @@ static void write_sm1_per_component_unary_op(struct hlsl_ctx *ctx, struct vkd3d_
 
         src.writemask = hlsl_combine_writemasks(src.writemask, 1u << i);
         dst.writemask = hlsl_combine_writemasks(dst.writemask, 1u << i);
-        write_sm1_unary_op(ctx, buffer, opcode, &dst, &src, 0);
+        write_sm1_unary_op(ctx, buffer, opcode, &dst, &src, 0, 0);
     }
 }
 
@@ -676,7 +677,11 @@ static void write_sm1_expr(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *b
             break;
 
         case HLSL_OP1_NEG:
-            write_sm1_unary_op(ctx, buffer, D3DSIO_MOV, &instr->reg, &arg1->reg, D3DSPSM_NEG);
+            write_sm1_unary_op(ctx, buffer, D3DSIO_MOV, &instr->reg, &arg1->reg, D3DSPSM_NEG, 0);
+            break;
+
+        case HLSL_OP1_SAT:
+            write_sm1_unary_op(ctx, buffer, D3DSIO_MOV, &instr->reg, &arg1->reg, 0, D3DSPDM_SATURATE);
             break;
 
         case HLSL_OP1_RCP:
@@ -704,7 +709,7 @@ static void write_sm1_expr(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *b
             break;
 
         case HLSL_OP1_FRACT:
-            write_sm1_unary_op(ctx, buffer, D3DSIO_FRC, &instr->reg, &arg1->reg, D3DSPSM_NONE);
+            write_sm1_unary_op(ctx, buffer, D3DSIO_FRC, &instr->reg, &arg1->reg, D3DSPSM_NONE, 0);
             break;
 
         case HLSL_OP2_DOT:
-- 
GitLab

https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79

[PATCH 0/1] MR79: vkd3d-shader/hlsl: Handle saturate() for SM1. - approved

[Zebediah Figura (＠zfigura)]

This merge request was approved by Zebediah Figura.
-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79

[Giovanni Mascellani (＠giomasce)]

Since we have names for `D3DSPDM_NONE` and `D3DSPSM_NONE`, it would probably be a good idea to use them instead of zeroes? Otherwise, that seems fine.

-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79#note_23077

[Zebediah Figura (＠zfigura)]

> Since we have names for `D3DSPDM_NONE` and `D3DSPSM_NONE`, it would probably be a good idea to use them instead of zeroes? Otherwise, that seems fine.

It seems unnecessary (perhaps even counterproductive) when dealing with flags arguments, but I don't feel strongly about it.

-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79#note_23109

[PATCH 0/1] MR79: vkd3d-shader/hlsl: Handle saturate() for SM1. - approved

[Francisco Casas (＠fcasas)]

This merge request was approved by Francisco Casas.
-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79

[Giovanni Mascellani (＠giomasce)]

On Tue Feb  7 12:26:41 2023 +0000, Zebediah Figura wrote:
> > Since we have names for `D3DSPDM_NONE` and `D3DSPSM_NONE`, it would
> > probably be a good idea to use them instead of zeroes? Otherwise, that
> > seems fine.
> > It seems unnecessary (perhaps even counterproductive) when dealing with
> > flags arguments, but I don't feel strongly about it.
> > Ah, wait, but `D3DSHADER_PARAM_SRCMOD_TYPE` and `D3DSHADER_PARAM_DSTMOD_TYPE` behave differently. The former doesn't seem to behave like a bitmask, the latter does. So my personal suggestion is that we should use `D3DSPSM_NONE` rather than `0`, and `0` rather than `D3DSPDM_NONE`.

But I'm not going to die on that hill.

-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79#note_23269

[PATCH 0/1] MR79: vkd3d-shader/hlsl: Handle saturate() for SM1. - approved

[Giovanni Mascellani (＠giomasce)]

This merge request was approved by Giovanni Mascellani.
-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79

[PATCH 0/1] MR79: vkd3d-shader/hlsl: Handle saturate() for SM1. - approved

[Henri Verbeet (＠hverbeet)]

This merge request was approved by Henri Verbeet.
-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79

[Zebediah Figura (＠zfigura)]

On Tue Feb  7 12:26:41 2023 +0000, Giovanni Mascellani wrote:
> Ah, wait, but `D3DSHADER_PARAM_SRCMOD_TYPE` and
> `D3DSHADER_PARAM_DSTMOD_TYPE` behave differently. The former doesn't
> seem to behave like a bitmask, the latter does. So my personal
> suggestion is that we should use `D3DSPSM_NONE` rather than `0`, and `0`
> rather than `D3DSPDM_NONE`.
> But I'm not going to die on that hill.
> Ah, I forgot that source modifiers are not a bitmask in sm1. Yes, that probably makes sense. Not worth rerolling this right now, though.

-- 
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/79#note_23311