[PATCH v2 0/3] MR255: vkd3d-shader: Scan for the maximum temporary register in d3dbc and emit a spirv temp array accordingly.

[Zebediah Figura (＠zfigura)]

```
On Margate Sands. 
I can connect
Nothing with nothing.
```

--
  v2: vkd3d-shader/d3dbc: Scan for the maximum temporary register index.
      vkd3d-shader: Record a global temporary count per sm4 shader.
      vkd3d-shader/dxbc: Remove redundant zero-initialization of the vkd3d_shader_desc structure.

https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/255

[PATCH v2 1/3] vkd3d-shader/dxbc: Remove redundant zero-initialization of the vkd3d_shader_desc structure.

[Zebediah Figura]

From: Zebediah Figura

---
 libs/vkd3d-shader/dxbc.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/libs/vkd3d-shader/dxbc.c b/libs/vkd3d-shader/dxbc.c
index 3c52e332c..716b7bdb7 100644
--- a/libs/vkd3d-shader/dxbc.c
+++ b/libs/vkd3d-shader/dxbc.c
@@ -531,12 +531,6 @@ int shader_extract_from_dxbc(const struct vkd3d_shader_code *dxbc,
 {
     int ret;

-    desc->byte_code = NULL;
-    desc->byte_code_size = 0;
-    memset(&desc->input_signature, 0, sizeof(desc->input_signature));
-    memset(&desc->output_signature, 0, sizeof(desc->output_signature));
-    memset(&desc->patch_constant_signature, 0, sizeof(desc->patch_constant_signature));
-
     ret = for_each_dxbc_section(dxbc, message_context, source_name, shdr_handler, desc);
     if (!desc->byte_code)
         ret = VKD3D_ERROR_INVALID_ARGUMENT;

[PATCH v2 2/3] vkd3d-shader: Record a global temporary count per sm4 shader.

[Zebediah Figura]

From: Zebediah Figura

Store it in the shader_desc, and declare temps from that when compiling SPIR-V,
instead of parsing dcl_instructions.

As part of this change, we declare a single, global temps array (with Private
scope instead of Function) which is as large as the maximum of all dcl_temps
instructions. It is not clear to me whether this will improve, hurt, or have no
significant effect on the lower-level compiler. An alternative is to still
redeclare a new temps array every time (although still with a smaller size).
---
 libs/vkd3d-shader/spirv.c                | 17 +++++++----------
 libs/vkd3d-shader/tpf.c                  |  2 ++
 libs/vkd3d-shader/vkd3d_shader_private.h |  2 ++
 3 files changed, 11 insertions(+), 10 deletions(-)

diff --git a/libs/vkd3d-shader/spirv.c b/libs/vkd3d-shader/spirv.c
index 3542b5fac..8c2ec132d 100644
--- a/libs/vkd3d-shader/spirv.c
+++ b/libs/vkd3d-shader/spirv.c
@@ -5258,8 +5258,7 @@ static void spirv_compiler_emit_dcl_global_flags(struct spirv_compiler *compiler
         WARN("Unhandled global flags %#x.\n", flags);
 }

-static void spirv_compiler_emit_dcl_temps(struct spirv_compiler *compiler,
-        const struct vkd3d_shader_instruction *instruction)
+static void spirv_compiler_emit_temps(struct spirv_compiler *compiler, uint32_t count)
 {
     struct vkd3d_spirv_builder *builder = &compiler->spirv_builder;
     size_t function_location;
@@ -5270,11 +5269,11 @@ static void spirv_compiler_emit_dcl_temps(struct spirv_compiler *compiler,
     vkd3d_spirv_begin_function_stream_insertion(builder, function_location);

assert(!compiler->temp_count);
-    compiler->temp_count = instruction->declaration.count;
+    compiler->temp_count = count;
     for (i = 0; i < compiler->temp_count; ++i)
     {
         id = spirv_compiler_emit_variable(compiler, &builder->function_stream,
-                SpvStorageClassFunction, VKD3D_SHADER_COMPONENT_FLOAT, VKD3D_VEC4_SIZE);
+                SpvStorageClassPrivate, VKD3D_SHADER_COMPONENT_FLOAT, VKD3D_VEC4_SIZE);
         if (!i)
             compiler->temp_id = id;
         assert(id == compiler->temp_id + i);
@@ -6236,9 +6235,6 @@ static void spirv_compiler_leave_shader_phase(struct spirv_compiler *compiler)

vkd3d_spirv_build_op_function_end(builder);

-    compiler->temp_id = 0;
-    compiler->temp_count = 0;
-
     if (is_in_control_point_phase(compiler))
     {
         if (compiler->epilogue_function_id)
@@ -9103,9 +9099,6 @@ static int spirv_compiler_handle_instruction(struct spirv_compiler *compiler,
         case VKD3DSIH_DCL_GLOBAL_FLAGS:
             spirv_compiler_emit_dcl_global_flags(compiler, instruction);
             break;
-        case VKD3DSIH_DCL_TEMPS:
-            spirv_compiler_emit_dcl_temps(compiler, instruction);
-            break;
         case VKD3DSIH_DCL_INDEXABLE_TEMP:
             spirv_compiler_emit_dcl_indexable_temp(compiler, instruction);
             break;
@@ -9426,6 +9419,7 @@ static int spirv_compiler_handle_instruction(struct spirv_compiler *compiler,
             spirv_compiler_emit_cut_stream(compiler, instruction);
             break;
         case VKD3DSIH_DCL_HS_MAX_TESSFACTOR:
+        case VKD3DSIH_DCL_TEMPS:
         case VKD3DSIH_HS_DECLS:
         case VKD3DSIH_NOP:
             /* nothing to do */
@@ -9448,6 +9442,9 @@ static int spirv_compiler_generate_spirv(struct spirv_compiler *compiler,
     enum vkd3d_result result = VKD3D_OK;
     unsigned int i;

+    if (parser->shader_desc.temp_count)
+        spirv_compiler_emit_temps(compiler, parser->shader_desc.temp_count);
+
     compiler->location.column = 0;
     compiler->location.line = 1;

diff --git a/libs/vkd3d-shader/tpf.c b/libs/vkd3d-shader/tpf.c
index 677243e15..31c43a5a4 100644
--- a/libs/vkd3d-shader/tpf.c
+++ b/libs/vkd3d-shader/tpf.c
@@ -989,6 +989,8 @@ static void shader_sm4_read_declaration_count(struct vkd3d_shader_instruction *i
         uint32_t opcode_token, const uint32_t *tokens, unsigned int token_count, struct vkd3d_shader_sm4_parser *priv)
 {
     ins->declaration.count = *tokens;
+    if (opcode == VKD3D_SM4_OP_DCL_TEMPS)
+        priv->p.shader_desc.temp_count = max(priv->p.shader_desc.temp_count, *tokens);
 }

static void shader_sm4_read_declaration_dst(struct vkd3d_shader_instruction *ins, uint32_t opcode,
diff --git a/libs/vkd3d-shader/vkd3d_shader_private.h b/libs/vkd3d-shader/vkd3d_shader_private.h
index 398139c3e..85fca9642 100644
--- a/libs/vkd3d-shader/vkd3d_shader_private.h
+++ b/libs/vkd3d-shader/vkd3d_shader_private.h
@@ -829,6 +829,8 @@ struct vkd3d_shader_desc
     struct shader_signature input_signature;
     struct shader_signature output_signature;
     struct shader_signature patch_constant_signature;
+
+    uint32_t temp_count;
 };

struct vkd3d_shader_register_semantic

[PATCH v2 3/3] vkd3d-shader/d3dbc: Scan for the maximum temporary register index.

[Zebediah Figura]

From: Zebediah Figura

---
 libs/vkd3d-shader/d3dbc.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/libs/vkd3d-shader/d3dbc.c b/libs/vkd3d-shader/d3dbc.c
index d2a4666a5..e8a6236bd 100644
--- a/libs/vkd3d-shader/d3dbc.c
+++ b/libs/vkd3d-shader/d3dbc.c
@@ -729,6 +729,16 @@ static bool add_signature_element_from_semantic(struct vkd3d_shader_sm1_parser *
             semantic->usage_idx, sysval, reg->idx[0].offset, true, mask);
 }

+static void shader_sm1_scan_register(struct vkd3d_shader_sm1_parser *sm1, const struct vkd3d_shader_register *reg, unsigned int mask)
+{
+    uint32_t register_index = reg->idx[0].offset;
+
+    if (reg->type == VKD3DSPR_TEMP)
+        sm1->p.shader_desc.temp_count = max(sm1->p.shader_desc.temp_count, register_index + 1);
+
+    add_signature_element_from_register(sm1, reg, false, mask);
+}
+
 /* Read a parameter token from the input stream, and possibly a relative
  * addressing token. */
 static void shader_sm1_read_param(struct vkd3d_shader_sm1_parser *sm1,
@@ -1083,7 +1093,7 @@ static void shader_sm1_read_instruction(struct vkd3d_shader_sm1_parser *sm1, str
         if (ins->dst_count)
         {
             shader_sm1_read_dst_param(sm1, &p, dst_param);
-            add_signature_element_from_register(sm1, &dst_param->reg, false, dst_param->write_mask);
+            shader_sm1_scan_register(sm1, &dst_param->reg, dst_param->write_mask);
         }

/* Predication token */
@@ -1094,8 +1104,7 @@ static void shader_sm1_read_instruction(struct vkd3d_shader_sm1_parser *sm1, str
         for (i = 0; i < ins->src_count; ++i)
         {
             shader_sm1_read_src_param(sm1, &p, &src_params[i]);
-            add_signature_element_from_register(sm1, &src_params[i].reg,
-                    false, mask_from_swizzle(src_params[i].swizzle));
+            shader_sm1_scan_register(sm1, &src_params[i].reg, mask_from_swizzle(src_params[i].swizzle));
         }
     }

[Zebediah Figura (＠zfigura)]

Pushed a v2 that does this.