Signed-off-by: Francisco Casas fcasas@codeweavers.com --- Makefile.am | 4 + libs/vkd3d-shader/hlsl.c | 4 + libs/vkd3d-shader/hlsl.h | 4 + libs/vkd3d-shader/hlsl.y | 106 ++++++++++++++++++++++++++ libs/vkd3d-shader/hlsl_sm4.c | 61 +++++++++++++++ tests/hlsl-gather-offset.shader_test | 93 +++++++++++++++++++++++ tests/hlsl-gather.shader_test | 107 +++++++++++++++++++++++++++ 7 files changed, 379 insertions(+) create mode 100644 tests/hlsl-gather-offset.shader_test create mode 100644 tests/hlsl-gather.shader_test
diff --git a/Makefile.am b/Makefile.am index 16848f5a..749df363 100644 --- a/Makefile.am +++ b/Makefile.am @@ -68,6 +68,8 @@ vkd3d_shader_tests = \ tests/hlsl-duplicate-modifiers.shader_test \ tests/hlsl-for.shader_test \ tests/hlsl-function-overload.shader_test \ + tests/hlsl-gather-offset.shader_test \ + tests/hlsl-gather.shader_test \ tests/hlsl-invalid.shader_test \ tests/hlsl-majority-pragma.shader_test \ tests/hlsl-majority-typedef.shader_test \ @@ -294,6 +296,8 @@ XFAIL_TESTS = \ tests/hlsl-duplicate-modifiers.shader_test \ tests/hlsl-for.shader_test \ tests/hlsl-function-overload.shader_test \ + tests/hlsl-gather-offset.shader_test \ + tests/hlsl-gather.shader_test \ tests/hlsl-majority-pragma.shader_test \ tests/hlsl-majority-typedef.shader_test \ tests/hlsl-nested-arrays.shader_test \ diff --git a/libs/vkd3d-shader/hlsl.c b/libs/vkd3d-shader/hlsl.c index b426e461..3ace7462 100644 --- a/libs/vkd3d-shader/hlsl.c +++ b/libs/vkd3d-shader/hlsl.c @@ -1277,6 +1277,10 @@ static void dump_ir_resource_load(struct vkd3d_string_buffer *buffer, const stru { [HLSL_RESOURCE_LOAD] = "load_resource", [HLSL_RESOURCE_SAMPLE] = "sample", + [HLSL_RESOURCE_GATHER_RED] = "gather_red", + [HLSL_RESOURCE_GATHER_GREEN] = "gather_green", + [HLSL_RESOURCE_GATHER_BLUE] = "gather_blue", + [HLSL_RESOURCE_GATHER_ALPHA] = "gather_alpha", };
vkd3d_string_buffer_printf(buffer, "%s(resource = ", type_names[load->load_type]); diff --git a/libs/vkd3d-shader/hlsl.h b/libs/vkd3d-shader/hlsl.h index 2396adb4..49fa8d9d 100644 --- a/libs/vkd3d-shader/hlsl.h +++ b/libs/vkd3d-shader/hlsl.h @@ -378,6 +378,10 @@ enum hlsl_resource_load_type { HLSL_RESOURCE_LOAD, HLSL_RESOURCE_SAMPLE, + HLSL_RESOURCE_GATHER_RED, + HLSL_RESOURCE_GATHER_GREEN, + HLSL_RESOURCE_GATHER_BLUE, + HLSL_RESOURCE_GATHER_ALPHA, };
struct hlsl_ir_resource_load diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 3a0ddb93..a321789c 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -1930,6 +1930,112 @@ 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_node *offset = NULL; + struct hlsl_ir_load *sampler_load; + struct hlsl_type *result_type; + struct hlsl_ir_node *coords; + int status_out_arg; + int read_channel; + + 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 + { + load_type = HLSL_RESOURCE_GATHER_RED; + read_channel = 0; + } + + + if (!strcmp(name, "Gather")) + { + if (params->args_count != 2 && params->args_count != 3) + { + hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT, + "Wrong number of arguments to method 'Gather': expected 2 or 3, but got %u.", params->args_count); + return false; + } + } + else if (params->args_count < 2 || params->args_count == 5 || params->args_count > 7) + { + hlsl_error(ctx, loc, VKD3D_SHADER_ERROR_HLSL_WRONG_PARAMETER_COUNT, + "Wrong number of arguments to method '%s': expected 2,3,4,6 or 7, but got %u.", name, params->args_count); + return false; + } + + status_out_arg = -1; + if (params->args_count == 4) + status_out_arg = 3; + if (params->args_count == 7) + status_out_arg = 6; + if (status_out_arg != -1) + hlsl_fixme(ctx, loc, "Ignoring 'status' output parameter.\n"); + + if (params->args_count == 6 || params->args_count == 7) + hlsl_fixme(ctx, loc, "Ignoring multiple offset parameters.\n"); + + if (params->args_count == 3 || params->args_count == 4) + { + if (!(offset = add_implicit_conversion(ctx, instrs, params->args[2], + hlsl_get_vector_type(ctx, HLSL_TYPE_INT, sampler_dim), loc))) + return false; + } + + 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))) + return false; + + 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, offset, 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 d55815f1..61247031 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -1661,6 +1661,50 @@ 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, + unsigned int swizzle, const struct hlsl_ir_node *texel_offset) +{ + struct sm4_instruction instr; + unsigned int writemask; + + memset(&instr, 0, sizeof(instr)); + + /* TODO: possible optimization, if the offset ranges from -8 to 7, an VKD3D_SM4_OP_GATHER4 with + an aoffimmi modifier can be used. */ + instr.opcode = texel_offset? VKD3D_SM5_OP_GATHER4_PO : VKD3D_SM4_OP_GATHER4; + + sm4_register_from_node(&instr.dsts[0].reg, &instr.dsts[0].writemask, NULL, dst); + instr.dst_count = 1; + + instr.src_count = (instr.opcode == VKD3D_SM5_OP_GATHER4_PO)? 4 : 3; + + sm4_register_from_node(&instr.srcs[0].reg, &writemask, &instr.srcs[0].swizzle_type, coords); + instr.srcs[0].swizzle = hlsl_swizzle_from_writemask(writemask); + + if (texel_offset) + { + sm4_register_from_node(&instr.srcs[1].reg, &writemask, &instr.srcs[1].swizzle_type, texel_offset); + instr.srcs[1].swizzle = hlsl_swizzle_from_writemask(writemask); + } + + sm4_register_from_deref(ctx, &instr.srcs[instr.src_count-2].reg, &writemask, + &instr.srcs[instr.src_count-2].swizzle_type, resource, resource_type); + instr.srcs[instr.src_count-2].swizzle = hlsl_map_swizzle(hlsl_swizzle_from_writemask(writemask), instr.dsts[0].writemask); + + sm4_register_from_deref(ctx, &instr.srcs[instr.src_count-1].reg, &writemask, NULL, + sampler, sampler->var->data_type); + instr.srcs[instr.src_count-1].reg.type = VKD3D_SM4_RT_SAMPLER; + instr.srcs[instr.src_count-1].reg.dim = VKD3D_SM4_DIMENSION_VEC4; + instr.srcs[instr.src_count-1].swizzle_type = VKD3D_SM4_SWIZZLE_SCALAR; + instr.srcs[instr.src_count-1].swizzle = swizzle; + + 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) { @@ -1701,6 +1745,23 @@ static void write_sm4_resource_load(struct hlsl_ctx *ctx, write_sm4_sample(ctx, buffer, resource_type, &load->node, &load->resource, &load->sampler, coords, texel_offset); break; + + case HLSL_RESOURCE_GATHER_RED: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(X, X, X, X), texel_offset); + break; + case HLSL_RESOURCE_GATHER_GREEN: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(Y, Y, Y, Y), texel_offset); + break; + case HLSL_RESOURCE_GATHER_BLUE: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(Z, Z, Z, Z), texel_offset); + break; + case HLSL_RESOURCE_GATHER_ALPHA: + write_sm4_gather(ctx, buffer, resource_type, &load->node, &load->resource, + &load->sampler, coords, HLSL_SWIZZLE(W, W, W, W), texel_offset); + break; } }
diff --git a/tests/hlsl-gather-offset.shader_test b/tests/hlsl-gather-offset.shader_test new file mode 100644 index 00000000..d7e1a233 --- /dev/null +++ b/tests/hlsl-gather-offset.shader_test @@ -0,0 +1,93 @@ +[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(1, 0)); +} + +[test] +draw quad +probe all rgba (0.1, 0.2, 0.2, 0.1) + + +[pixel shader] +sampler s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherRed(s, float2(0.6, 0.6), int2(-1, 0)); +} + +[test] +draw quad +probe all rgba (0.0, 0.1, 0.1, 0.0) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherGreen(s, float2(0.2, 0.2), int2(0, 1)); +} + +[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(0.8, 0.8), int2(-1, -1)); +} + +[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(1, 0)); +} + +[test] +draw quad +probe all rgba (0.5, 0.0, 0.5, 0.0) + + +[pixel shader] +SamplerState s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherAlpha(s, float2(0.2, 0.8), int2(0, -1)); +} + +[test] +draw quad +probe all rgba (0.0, 0.4, 0.0, 0.4) diff --git a/tests/hlsl-gather.shader_test b/tests/hlsl-gather.shader_test new file mode 100644 index 00000000..2682c6ef --- /dev/null +++ b/tests/hlsl-gather.shader_test @@ -0,0 +1,107 @@ +[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)); +} + +[test] +draw quad +probe all rgba (0.0, 0.1, 0.1, 0.0) + + +[pixel shader] +sampler s; +Texture2D t; + +float4 main() : sv_target +{ + return t.GatherRed(s, float2(0.6, 0.6), int2(0, 0)); +} + +[test] +draw quad +probe all rgba (0.1, 0.2, 0.2, 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.GatherGreen(s, float2(0.8, 0.8)); +} + +[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), int2(0, 0), int2(0, 0), 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)
On Tue, Dec 21, 2021 at 2:04 PM Francisco Casas fcasas@codeweavers.com wrote:
diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c index d55815f1..61247031 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -1661,6 +1661,50 @@ 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,unsigned int swizzle, const struct hlsl_ir_node *texel_offset)+{
- struct sm4_instruction instr;
- unsigned int writemask;
- memset(&instr, 0, sizeof(instr));
- /* TODO: possible optimization, if the offset ranges from -8 to 7, an VKD3D_SM4_OP_GATHER4 with
- an aoffimmi modifier can be used. */
- instr.opcode = texel_offset? VKD3D_SM5_OP_GATHER4_PO : VKD3D_SM4_OP_GATHER4;
Not just an optimization, since the _PO variant is SM5 which is problematic if we're compiling for a 4.0 profile.
It's okay to have this as a (temporary) TODO though.
-
Francisco Casas -
Matteo Bruni