[PATCH vkd3d 06/12] vkd3d-shader/hlsl: Split matrix copies.

[Giovanni Mascellani]

Signed-off-by: Giovanni Mascellani <gmascellani at codeweavers.com>
---
 libs/vkd3d-shader/hlsl_codegen.c       | 52 ++++++++++++++++++++++++++
 tests/hlsl-matrix-indexing.shader_test | 17 +++++++++
 2 files changed, 69 insertions(+)

diff --git a/libs/vkd3d-shader/hlsl_codegen.c b/libs/vkd3d-shader/hlsl_codegen.c
index 72c00430..73e3b73f 100644
--- a/libs/vkd3d-shader/hlsl_codegen.c
+++ b/libs/vkd3d-shader/hlsl_codegen.c
@@ -672,6 +672,57 @@ static bool split_struct_copies(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr
     return true;
 }
 
+static unsigned int minor_size(const struct hlsl_type *type)
+{
+    if (type->type == HLSL_CLASS_VECTOR || type->modifiers & HLSL_MODIFIER_ROW_MAJOR)
+        return type->dimx;
+    else
+        return type->dimy;
+}
+
+static unsigned int major_size(const struct hlsl_type *type)
+{
+    if (type->type == HLSL_CLASS_VECTOR || type->modifiers & HLSL_MODIFIER_ROW_MAJOR)
+        return type->dimy;
+    else
+        return type->dimx;
+}
+
+static bool split_matrix_copies(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
+{
+    const struct hlsl_ir_node *rhs;
+    struct hlsl_type *element_type;
+    const struct hlsl_type *type;
+    unsigned int i;
+    struct hlsl_ir_store *store;
+
+    if (instr->type != HLSL_IR_STORE)
+        return false;
+
+    store = hlsl_ir_store(instr);
+    rhs = store->rhs.node;
+    type = rhs->data_type;
+    if (type->type != HLSL_CLASS_MATRIX)
+        return false;
+    element_type = hlsl_get_vector_type(ctx, type->base_type, minor_size(type));
+
+    if (rhs->type != HLSL_IR_LOAD)
+    {
+        hlsl_fixme(ctx, &instr->loc, "Copying from unsupported node type.\n");
+        return false;
+    }
+
+    for (i = 0; i < major_size(type); ++i)
+    {
+        if (!split_copy(ctx, store, hlsl_ir_load(rhs), 4 * i, element_type))
+            return false;
+    }
+
+    list_remove(&store->node.entry);
+    hlsl_free_instr(&store->node);
+    return true;
+}
+
 static bool lower_narrowing_casts(struct hlsl_ctx *ctx, struct hlsl_ir_node *instr, void *context)
 {
     const struct hlsl_type *src_type, *dst_type;
@@ -1685,6 +1736,7 @@ int hlsl_emit_bytecode(struct hlsl_ctx *ctx, struct hlsl_ir_function_decl *entry
         progress |= transform_ir(ctx, split_struct_copies, body, NULL);
     }
     while (progress);
+    transform_ir(ctx, split_matrix_copies, body, NULL);
     transform_ir(ctx, lower_narrowing_casts, body, NULL);
     transform_ir(ctx, lower_casts_to_bool, body, NULL);
     do
diff --git a/tests/hlsl-matrix-indexing.shader_test b/tests/hlsl-matrix-indexing.shader_test
index 1d444ffa..e2103fe2 100644
--- a/tests/hlsl-matrix-indexing.shader_test
+++ b/tests/hlsl-matrix-indexing.shader_test
@@ -45,3 +45,20 @@ uniform 8 float4 9.0 10.0 11.0 12.0
 uniform 12 float4 13.0 14.0 15.0 16.0
 todo draw quad
 probe all rgba (1.0, 5.0, 7.0, 12.0)
+
+[pixel shader]
+uniform float4x4 m;
+
+float4 main() : SV_TARGET
+{
+    float4x4 m2 = m;
+    return float4(m2[0][0], m2[1][0], m2[1][2], m2[2][3]);
+}
+
+[test]
+uniform 0 float4 1.0 2.0 3.0 4.0
+uniform 4 float4 5.0 6.0 7.0 8.0
+uniform 8 float4 9.0 10.0 11.0 12.0
+uniform 12 float4 13.0 14.0 15.0 16.0
+todo draw quad
+probe all rgba (1.0, 2.0, 10.0, 15.0)
-- 
2.35.2