vkd3d-shader/hlsl: Implement texture gather methods.

Signed-off-by: Francisco Casas <fcasas@codeweavers.com>
Signed-off-by: Zebediah Figura <zfigura@codeweavers.com>
Signed-off-by: Giovanni Mascellani <gmascellani@codeweavers.com>
Signed-off-by: Matteo Bruni <mbruni@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>

libs/vkd3d-shader/hlsl.c

@@ -1278,8 +1278,13 @@ 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",
     };
 
+    assert(load->load_type < ARRAY_SIZE(type_names));
     vkd3d_string_buffer_printf(buffer, "%s(resource = ", type_names[load->load_type]);
     dump_deref(buffer, &load->resource);
     vkd3d_string_buffer_printf(buffer, ", sampler = ");

libs/vkd3d-shader/hlsl.h

@@ -376,6 +376,10 @@ 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

libs/vkd3d-shader/hlsl.y

@@ -1930,6 +1930,106 @@ 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;
+        unsigned 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;
+        }
+
+        if (params->args_count == 4 || params->args_count == 7)
+            hlsl_fixme(ctx, loc, "Tiled resource status argument.");
+
+        if (params->args_count == 6 || params->args_count == 7)
+            hlsl_fixme(ctx, loc, "Multiple Gather() offset parameters.");
+
+        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 1 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 %u 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;

libs/vkd3d-shader/hlsl_sm4.c

@@ -794,7 +794,7 @@ struct sm4_instruction
         struct sm4_register reg;
         enum vkd3d_sm4_swizzle_type swizzle_type;
         unsigned int swizzle;
-    } srcs[3];
+    } srcs[4];
     unsigned int src_count;
 
     uint32_t idx[2];
@@ -1771,6 +1771,41 @@ 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_src_register *src;
+    struct sm4_instruction instr;
+
+    memset(&instr, 0, sizeof(instr));
+
+    instr.opcode = VKD3D_SM4_OP_GATHER4;
+
+    sm4_dst_from_node(&instr.dsts[0], dst);
+    instr.dst_count = 1;
+
+    sm4_src_from_node(&instr.srcs[instr.src_count++], coords, VKD3DSP_WRITEMASK_ALL);
+
+    /* FIXME: Use an aoffimmi modifier if possible. */
+    if (texel_offset)
+    {
+        instr.opcode = VKD3D_SM5_OP_GATHER4_PO;
+        sm4_src_from_node(&instr.srcs[instr.src_count++], texel_offset, VKD3DSP_WRITEMASK_ALL);
+    }
+
+    sm4_src_from_deref(ctx, &instr.srcs[instr.src_count++], resource, resource_type, instr.dsts[0].writemask);
+
+    src = &instr.srcs[instr.src_count++];
+    sm4_src_from_deref(ctx, src, sampler, sampler->var->data_type, VKD3DSP_WRITEMASK_ALL);
+    src->reg.dim = VKD3D_SM4_DIMENSION_VEC4;
+    src->swizzle_type = VKD3D_SM4_SWIZZLE_SCALAR;
+    src->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)
 {
@@ -1811,6 +1846,26 @@ 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;
     }
 }