On 11/22/21 9:28 AM, Francisco Casas wrote:
diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index b7e0409d..8a2c1dd9 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -1852,6 +1852,84 @@ static bool add_method_call(struct hlsl_ctx *ctx, struct list *instrs, struct hl list_add_tail(instrs, &load->node.entry); return true; }
- else if (!strcmp(name, "Gather") || !strcmp(name, "GatherRed") || !strcmp(name, "GatherBlue")
|| !strcmp(name, "GatherGreen") || !strcmp(name, "GatherAlpha"))- {
const unsigned int sampler_dim = sampler_dim_count(object_type->sampler_dim);enum hlsl_resource_load_type load_type;const struct hlsl_type *sampler_type;struct hlsl_ir_resource_load *load;struct hlsl_ir_load *sampler_load;struct hlsl_type *result_type;struct hlsl_ir_node *coords;int read_channel;if(!strcmp(name, "Gather") || !strcmp(name, "GatherRed")){load_type = HLSL_RESOURCE_GATHER_RED;read_channel = 0;}else if(!strcmp(name, "GatherGreen")){load_type = HLSL_RESOURCE_GATHER_GREEN;read_channel = 1;}else if(!strcmp(name, "GatherBlue")){load_type = HLSL_RESOURCE_GATHER_BLUE;read_channel = 2;}else if(!strcmp(name, "GatherAlpha")){load_type = HLSL_RESOURCE_GATHER_ALPHA;read_channel = 3;} else {assert(!"Unexpected Gather method.");}if (params->args_count != 3 && params->args_count != 4){hlsl_error(ctx, *loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT,"Wrong number of arguments to method '%s': expected 3 or 4, but got %u.", name, params->args_count);return false;}
It seems that it's legal to omit the offset parameter as well.
It's also apparently legal to specify a list of four different offsets, one for each component, for a total of six or seven arguments. That doesn't necessarily need to be handled here, but it might be a good idea to at least print a FIXME for it.
if (params->args_count >= 3)FIXME("Ignoring offset parameter.\n");if (params->args_count >= 4)FIXME("Ignoring 'status' output parameter.\n");sampler_type = params->args[0]->data_type;if (sampler_type->type != HLSL_CLASS_OBJECT || sampler_type->base_type != HLSL_TYPE_SAMPLER|| sampler_type->sampler_dim != HLSL_SAMPLER_DIM_GENERIC){struct vkd3d_string_buffer *string;if ((string = hlsl_type_to_string(ctx, sampler_type)))hlsl_error(ctx, *loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE,"Wrong type for argument 0 of %s(): expected 'sampler', but got '%s'.", name, string->buffer);hlsl_release_string_buffer(ctx, string);return false;}if(read_channel >= object_type->e.resource_format->dimx){hlsl_error(ctx, *loc, VKD3D_SHADER_ERROR_HLSL_INVALID_TYPE,"Method %s() requires at least %d channels.", name, read_channel+1);return false;}result_type = hlsl_get_vector_type(ctx, object_type->e.resource_format->base_type, 4);/* Only HLSL_IR_LOAD can return an object. */sampler_load = hlsl_ir_load(params->args[0]);if (!(coords = add_implicit_conversion(ctx, instrs, params->args[1],hlsl_get_vector_type(ctx, HLSL_TYPE_FLOAT, sampler_dim), loc)))coords = params->args[1];if (!(load = hlsl_new_resource_load(ctx, result_type,load_type, object_load->src.var, object_load->src.offset.node,sampler_load->src.var, sampler_load->src.offset.node, coords, loc)))return false;list_add_tail(instrs, &load->node.entry);return true;- } else { struct vkd3d_string_buffer *string;
diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c index a95503ee..5f290129 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -753,7 +753,7 @@ struct sm4_instruction struct sm4_register reg; enum vkd3d_sm4_swizzle_type swizzle_type; unsigned int swizzle;
- } srcs[2];
} srcs[3]; unsigned int src_count;
uint32_t idx[2];
@@ -1506,6 +1506,61 @@ static void write_sm4_loop(struct hlsl_ctx *ctx, write_sm4_instruction(buffer, &instr); }
+static void write_sm4_gather(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *buffer,
const struct hlsl_type *resource_type, const struct hlsl_ir_node *dst,const struct hlsl_deref *resource, const struct hlsl_deref *sampler, const struct hlsl_ir_node *coords,enum hlsl_resource_load_type load_type)+{
- struct sm4_instruction instr;
- unsigned int writemask;
- memset(&instr, 0, sizeof(instr));
- instr.opcode = VKD3D_SM4_OP_GATHER4;
- sm4_register_from_node(&instr.dsts[0].reg, &instr.dsts[0].writemask, NULL, dst);
- instr.dst_count = 1;
- sm4_register_from_node(&instr.srcs[0].reg, &writemask, &instr.srcs[0].swizzle_type, coords);
- instr.srcs[0].swizzle = hlsl_swizzle_from_writemask(writemask);
- sm4_register_from_deref(ctx, &instr.srcs[1].reg, &writemask, &instr.srcs[1].swizzle_type,
resource, resource_type);- instr.srcs[1].swizzle = hlsl_map_swizzle(hlsl_swizzle_from_writemask(writemask), instr.dsts[0].writemask);
- sm4_register_from_deref(ctx, &instr.srcs[2].reg, &writemask, NULL,
sampler, sampler->var->data_type);- instr.srcs[2].reg.type = VKD3D_SM4_RT_SAMPLER;
- instr.srcs[2].reg.dim = 1;
This should be a VKD3D_SM4_DIMENSION constant (for which I observe VEC4).
- instr.srcs[2].swizzle_type = VKD3D_SM4_SWIZZLE_VEC4;
I observe VKD3D_SM4_SWIZZLE_SCALAR from native here; do you have an example with VEC4?
- switch (load_type){
case HLSL_RESOURCE_GATHER_RED:instr.srcs[2].swizzle = HLSL_SWIZZLE(X, X, X, X);break;case HLSL_RESOURCE_GATHER_GREEN:instr.srcs[2].swizzle = HLSL_SWIZZLE(Y, Y, Y, Y);break;case HLSL_RESOURCE_GATHER_BLUE:instr.srcs[2].swizzle = HLSL_SWIZZLE(Z, Z, Z, Z);break;case HLSL_RESOURCE_GATHER_ALPHA:instr.srcs[2].swizzle = HLSL_SWIZZLE(W, W, W, W);break;default:assert(!"Invalid hlsl_resource_load_type.");break;- }
It might be simpler code just to pass the swizzle (component) as an argument to this function.
- instr.src_count = 3;
- write_sm4_instruction(buffer, &instr);
+}
- static void write_sm4_resource_load(struct hlsl_ctx *ctx, struct vkd3d_bytecode_buffer *buffer, const struct hlsl_ir_resource_load *load) {
@@ -1527,6 +1582,15 @@ static void write_sm4_resource_load(struct hlsl_ctx *ctx, case HLSL_RESOURCE_SAMPLE: hlsl_fixme(ctx, load->node.loc, "Resource sample instruction."); break;
case HLSL_RESOURCE_GATHER_RED:case HLSL_RESOURCE_GATHER_GREEN:case HLSL_RESOURCE_GATHER_BLUE:case HLSL_RESOURCE_GATHER_ALPHA:assert(load->sampler.var);write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource,&load->sampler, coords, load->load_type);break; }diff --git a/tests/hlsl-gather.shader_test b/tests/hlsl-gather.shader_test new file mode 100644 index 00000000..6731b77c --- /dev/null +++ b/tests/hlsl-gather.shader_test @@ -0,0 +1,106 @@ +[sampler 0] +filter linear linear linear +address clamp clamp clamp
+[texture 0] +size (3, 3) +0.0 0.0 0.0 0.4 0.1 0.0 0.5 0.0 0.2 0.0 0.0 0.4 +0.0 0.1 0.5 0.0 0.1 0.1 0.0 0.4 0.2 0.1 0.5 0.0 +0.0 0.2 0.0 0.4 0.1 0.2 0.5 0.0 0.2 0.2 0.0 0.4
+[pixel shader] +sampler s; +Texture2D t;
+float4 main() : sv_target +{
- return t.Gather(s, float2(0.2, 0.2), int2(0 , 0));
+}
+[test] +draw quad +probe all rgba (0.0, 0.1, 0.1, 0.0)
Can you please add a call to GatherRed()?
+[pixel shader] +SamplerState s; +Texture2D t;
+float4 main() : sv_target +{
- return t.GatherGreen(s, float2(0.2, 0.2), int2(0 , 0));
+}
+[test] +draw quad +probe all rgba (0.1, 0.1, 0.0, 0.0)
+[pixel shader] +SamplerState s; +Texture2D t;
+float4 main() : sv_target +{
- return t.GatherGreen(s, float2(0.8, 0.8), int2(0 , 0));
+}
+[test] +draw quad +probe all rgba (0.2, 0.2, 0.1, 0.1)
+[pixel shader] +SamplerState s; +Texture2D t;
+float4 main() : sv_target +{
- return t.GatherGreen(s, float2(0.2, 0.2), int2(0 , 0));
+}
+[test] +draw quad +probe all rgba (0.1, 0.1, 0.0, 0.0)
+[pixel shader] +SamplerState s; +Texture2D t;
+float4 main() : sv_target +{
- return t.GatherBlue(s, float2(0.2, 0.8), int2(0 , 0));
+}
+[test] +draw quad +probe all rgba (0.0, 0.5, 0.0, 0.5)
+[pixel shader] +SamplerState s; +Texture2D t;
+float4 main() : sv_target +{
- return t.GatherAlpha(s, float2(0.2, 0.8), int2(0 , 0));
+}
+[test] +draw quad +probe all rgba (0.4, 0.0, 0.4, 0.0)