shader: Properly manage attributes not written from previous stages

src/shader_recompiler/backend/glsl/emit_context.cpp

@@ -327,11 +327,12 @@ EmitContext::EmitContext(IR::Program& program, Bindings& bindings, const Profile
 
     for (size_t index = 0; index < info.input_generics.size(); ++index) {
         const auto& generic{info.input_generics[index]};
-        if (generic.used) {
-            header += fmt::format("layout(location={}){}in vec4 in_attr{}{};", index,
-                                  InterpDecorator(generic.interpolation), index,
-                                  InputArrayDecorator(stage));
+        if (!generic.used || !runtime_info.previous_stage_stores_generic[index]) {
+            continue;
         }
+        header +=
+            fmt::format("layout(location={}){}in vec4 in_attr{}{};", index,
+                        InterpDecorator(generic.interpolation), index, InputArrayDecorator(stage));
     }
     for (size_t index = 0; index < info.uses_patches.size(); ++index) {
         if (!info.uses_patches[index]) {
@@ -349,10 +350,10 @@ EmitContext::EmitContext(IR::Program& program, Bindings& bindings, const Profile
         }
     }
     for (size_t index = 0; index < info.stores_generics.size(); ++index) {
-        // TODO: Properly resolve attribute issues
-        if (info.stores_generics[index] || StageInitializesVaryings()) {
-            DefineGenericOutput(index, program.invocations);
+        if (!info.stores_generics[index]) {
+            continue;
         }
+        DefineGenericOutput(index, program.invocations);
     }
     DefineConstantBuffers(bindings);
     DefineStorageBuffers(bindings);
@@ -362,17 +363,6 @@ EmitContext::EmitContext(IR::Program& program, Bindings& bindings, const Profile
     DefineConstants();
 }
 
-bool EmitContext::StageInitializesVaryings() const noexcept {
-    switch (stage) {
-    case Stage::VertexA:
-    case Stage::VertexB:
-    case Stage::Geometry:
-        return true;
-    default:
-        return false;
-    }
-}
-
 void EmitContext::SetupExtensions() {
     if (info.uses_shadow_lod && profile.support_gl_texture_shadow_lod) {
         header += "#extension GL_EXT_texture_shadow_lod : enable\n";

src/shader_recompiler/backend/glsl/emit_context.h

@@ -136,8 +136,6 @@ public:
         code += '\n';
     }
 
-    [[nodiscard]] bool StageInitializesVaryings() const noexcept;
-
     std::string header;
     std::string code;
     VarAlloc var_alloc;

src/shader_recompiler/backend/glsl/emit_glsl_context_get_set.cpp

@@ -8,6 +8,7 @@
 #include "shader_recompiler/backend/glsl/emit_glsl_instructions.h"
 #include "shader_recompiler/frontend/ir/value.h"
 #include "shader_recompiler/profile.h"
+#include "shader_recompiler/runtime_info.h"
 
 namespace Shader::Backend::GLSL {
 namespace {
@@ -179,6 +180,10 @@ void EmitGetAttribute(EmitContext& ctx, IR::Inst& inst, IR::Attribute attr,
     const char swizzle{"xyzw"[element]};
     if (IR::IsGeneric(attr)) {
         const u32 index{IR::GenericAttributeIndex(attr)};
+        if (!ctx.runtime_info.previous_stage_stores_generic[index]) {
+            ctx.AddF32("{}=0.f;", inst, attr);
+            return;
+        }
         ctx.AddF32("{}=in_attr{}{}.{};", inst, index, InputVertexIndex(ctx, vertex), swizzle);
         return;
     }

src/shader_recompiler/backend/glsl/emit_glsl_special.cpp

@@ -12,11 +12,12 @@
 
 namespace Shader::Backend::GLSL {
 namespace {
-void InitializeVaryings(EmitContext& ctx) {
-    ctx.Add("gl_Position=vec4(0,0,0,1);");
-    // TODO: Properly resolve attribute issues
-    for (size_t index = 0; index < ctx.info.stores_generics.size() / 2; ++index) {
-        if (!ctx.info.stores_generics[index]) {
+void InitializeOutputVaryings(EmitContext& ctx) {
+    if (ctx.stage == Stage::VertexB || ctx.stage == Stage::Geometry) {
+        ctx.Add("gl_Position=vec4(0,0,0,1);");
+    }
+    for (size_t index = 0; index < 16; ++index) {
+        if (ctx.info.stores_generics[index]) {
             ctx.Add("out_attr{}=vec4(0,0,0,1);", index);
         }
     }
@@ -56,9 +57,8 @@ void EmitPhiMove(EmitContext& ctx, const IR::Value& phi_value, const IR::Value&
 }
 
 void EmitPrologue(EmitContext& ctx) {
-    if (ctx.StageInitializesVaryings()) {
-        InitializeVaryings(ctx);
-    }
+    InitializeOutputVaryings(ctx);
+
     if (ctx.stage == Stage::Fragment && ctx.profile.need_declared_frag_colors) {
         for (size_t index = 0; index < ctx.info.stores_frag_color.size(); ++index) {
             if (ctx.info.stores_frag_color[index]) {
@@ -73,7 +73,7 @@ void EmitEpilogue(EmitContext&) {}
 
 void EmitEmitVertex(EmitContext& ctx, const IR::Value& stream) {
     ctx.Add("EmitStreamVertex(int({}));", ctx.var_alloc.Consume(stream));
-    InitializeVaryings(ctx);
+    InitializeOutputVaryings(ctx);
 }
 
 void EmitEndPrimitive(EmitContext& ctx, const IR::Value& stream) {

src/shader_recompiler/backend/spirv/emit_context.cpp

@@ -1209,6 +1209,9 @@ void EmitContext::DefineInputs(const Info& info) {
         tess_coord = DefineInput(*this, F32[3], false, spv::BuiltIn::TessCoord);
     }
     for (size_t index = 0; index < info.input_generics.size(); ++index) {
+        if (!runtime_info.previous_stage_stores_generic[index]) {
+            continue;
+        }
         const InputVarying generic{info.input_generics[index]};
         if (!generic.used) {
             continue;

src/shader_recompiler/backend/spirv/emit_spirv_context_get_set.cpp

@@ -286,7 +286,7 @@ Id EmitGetAttribute(EmitContext& ctx, IR::Attribute attr, Id vertex) {
     if (IR::IsGeneric(attr)) {
         const u32 index{IR::GenericAttributeIndex(attr)};
         const std::optional<AttrInfo> type{AttrTypes(ctx, index)};
-        if (!type) {
+        if (!type || !ctx.runtime_info.previous_stage_stores_generic[index]) {
             // Attribute is disabled
             return ctx.Const(0.0f);
         }

src/shader_recompiler/frontend/maxwell/translate_program.cpp

@@ -192,7 +192,9 @@ IR::Program MergeDualVertexPrograms(IR::Program& vertex_a, IR::Program& vertex_b
     result.local_memory_size = std::max(vertex_a.local_memory_size, vertex_b.local_memory_size);
     for (size_t index = 0; index < 32; ++index) {
         result.info.input_generics[index].used |= vertex_b.info.input_generics[index].used;
-        result.info.stores_generics[index] |= vertex_b.info.stores_generics[index];
+        if (vertex_b.info.stores_generics[index]) {
+            result.info.stores_generics[index] = true;
+        }
     }
     Optimization::JoinTextureInfo(result.info, vertex_b.info);
     Optimization::JoinStorageInfo(result.info, vertex_b.info);

src/shader_recompiler/ir_opt/collect_shader_info_pass.cpp

@@ -79,7 +79,7 @@ void GetAttribute(Info& info, IR::Attribute attr) {
 
 void SetAttribute(Info& info, IR::Attribute attr) {
     if (IR::IsGeneric(attr)) {
-        info.stores_generics.at(IR::GenericAttributeIndex(attr)) = true;
+        info.stores_generics[IR::GenericAttributeIndex(attr)] = true;
         return;
     }
     if (attr >= IR::Attribute::FixedFncTexture0S && attr <= IR::Attribute::FixedFncTexture9Q) {
@@ -956,7 +956,9 @@ void GatherInfoFromHeader(Environment& env, Info& info) {
     }
     if (info.stores_indexed_attributes) {
         for (size_t i = 0; i < info.stores_generics.size(); i++) {
-            info.stores_generics[i] |= header.vtg.IsOutputGenericVectorActive(i);
+            if (header.vtg.IsOutputGenericVectorActive(i)) {
+                info.stores_generics[i] = true;
+            }
         }
         info.stores_clip_distance |= header.vtg.omap_systemc.clip_distances != 0;
         info.stores_position |= header.vtg.omap_systemb.position != 0;

src/shader_recompiler/runtime_info.h

@@ -5,6 +5,7 @@
 #pragma once
 
 #include <array>
+#include <bitset>
 #include <optional>
 #include <vector>
 
@@ -59,6 +60,8 @@ struct TransformFeedbackVarying {
 
 struct RuntimeInfo {
     std::array<AttributeType, 32> generic_input_types{};
+    std::bitset<32> previous_stage_stores_generic{};
+
     bool convert_depth_mode{};
     bool force_early_z{};
 
@@ -72,11 +75,12 @@ struct RuntimeInfo {
     std::optional<CompareFunction> alpha_test_func;
     float alpha_test_reference{};
 
-    // Static y negate value
+    /// Static Y negate value
     bool y_negate{};
-    // Use storage buffers instead of global pointers on GLASM
+    /// Use storage buffers instead of global pointers on GLASM
     bool glasm_use_storage_buffers{};
 
+    /// Transform feedback state for each varying
     std::vector<TransformFeedbackVarying> xfb_varyings;
 };
 

src/shader_recompiler/shader_info.h

@@ -140,7 +140,7 @@ struct Info {
     bool stores_sample_mask{};
     bool stores_frag_depth{};
 
-    std::array<bool, 32> stores_generics{};
+    std::bitset<32> stores_generics{};
     bool stores_layer{};
     bool stores_viewport_index{};
     bool stores_point_size{};

src/video_core/renderer_opengl/gl_shader_cache.cpp

@@ -58,8 +58,15 @@ auto MakeSpan(Container& container) {
 
 Shader::RuntimeInfo MakeRuntimeInfo(const GraphicsPipelineKey& key,
                                     const Shader::IR::Program& program,
+                                    const Shader::IR::Program* previous_program,
                                     bool glasm_use_storage_buffers, bool use_assembly_shaders) {
     Shader::RuntimeInfo info;
+    if (previous_program) {
+        info.previous_stage_stores_generic = previous_program->info.stores_generics;
+    } else {
+        // Mark all stores as available
+        info.previous_stage_stores_generic.flip();
+    }
     switch (program.stage) {
     case Shader::Stage::VertexB:
     case Shader::Stage::Geometry:
@@ -400,6 +407,7 @@ std::unique_ptr<GraphicsPipeline> ShaderCache::CreateGraphicsPipeline(
     OGLProgram source_program;
     std::array<std::string, 5> sources;
     Shader::Backend::Bindings binding;
+    Shader::IR::Program* previous_program{};
     const bool use_glasm{device.UseAssemblyShaders()};
     const size_t first_index = uses_vertex_a && uses_vertex_b ? 1 : 0;
     for (size_t index = first_index; index < Maxwell::MaxShaderProgram; ++index) {
@@ -413,12 +421,13 @@ std::unique_ptr<GraphicsPipeline> ShaderCache::CreateGraphicsPipeline(
         infos[stage_index] = &program.info;
 
         const auto runtime_info{
-            MakeRuntimeInfo(key, program, glasm_use_storage_buffers, use_glasm)};
+            MakeRuntimeInfo(key, program, previous_program, glasm_use_storage_buffers, use_glasm)};
         if (use_glasm) {
             sources[stage_index] = EmitGLASM(profile, runtime_info, program, binding);
         } else {
             sources[stage_index] = EmitGLSL(profile, runtime_info, program, binding);
         }
+        previous_program = &program;
     }
     auto* const thread_worker{build_in_parallel ? workers.get() : nullptr};
     VideoCore::ShaderNotify* const notify{build_in_parallel ? &shader_notify : nullptr};

src/video_core/renderer_vulkan/vk_pipeline_cache.cpp

@@ -90,7 +90,7 @@ Shader::CompareFunction MaxwellToCompareFunction(Maxwell::ComparisonOp compariso
     return {};
 }
 
-static Shader::AttributeType CastAttributeType(const FixedPipelineState::VertexAttribute& attr) {
+Shader::AttributeType CastAttributeType(const FixedPipelineState::VertexAttribute& attr) {
     if (attr.enabled == 0) {
         return Shader::AttributeType::Disabled;
     }
@@ -124,9 +124,15 @@ Shader::AttributeType AttributeType(const FixedPipelineState& state, size_t inde
 }
 
 Shader::RuntimeInfo MakeRuntimeInfo(const GraphicsPipelineCacheKey& key,
-                                    const Shader::IR::Program& program) {
+                                    const Shader::IR::Program& program,
+                                    const Shader::IR::Program* previous_program) {
     Shader::RuntimeInfo info;
-
+    if (previous_program) {
+        info.previous_stage_stores_generic = previous_program->info.stores_generics;
+    } else {
+        // Mark all stores as available
+        info.previous_stage_stores_generic.flip();
+    }
     const Shader::Stage stage{program.stage};
     const bool has_geometry{key.unique_hashes[4] != 0};
     const bool gl_ndc{key.state.ndc_minus_one_to_one != 0};
@@ -499,6 +505,7 @@ std::unique_ptr<GraphicsPipeline> PipelineCache::CreateGraphicsPipeline(
     std::array<const Shader::Info*, Maxwell::MaxShaderStage> infos{};
     std::array<vk::ShaderModule, Maxwell::MaxShaderStage> modules;
 
+    const Shader::IR::Program* previous_stage{};
     Shader::Backend::Bindings binding;
     for (size_t index = uses_vertex_a && uses_vertex_b ? 1 : 0; index < Maxwell::MaxShaderProgram;
          ++index) {
@@ -511,7 +518,7 @@ std::unique_ptr<GraphicsPipeline> PipelineCache::CreateGraphicsPipeline(
         const size_t stage_index{index - 1};
         infos[stage_index] = &program.info;
 
-        const Shader::RuntimeInfo runtime_info{MakeRuntimeInfo(key, program)};
+        const Shader::RuntimeInfo runtime_info{MakeRuntimeInfo(key, program, previous_stage)};
         const std::vector<u32> code{EmitSPIRV(profile, runtime_info, program, binding)};
         device.SaveShader(code);
         modules[stage_index] = BuildShader(device, code);
@@ -519,6 +526,7 @@ std::unique_ptr<GraphicsPipeline> PipelineCache::CreateGraphicsPipeline(
             const std::string name{fmt::format("Shader {:016x}", key.unique_hashes[index])};
             modules[stage_index].SetObjectNameEXT(name.c_str());
         }
+        previous_stage = &program;
     }
     Common::ThreadWorker* const thread_worker{build_in_parallel ? &workers : nullptr};
     VideoCore::ShaderNotify* const notify{build_in_parallel ? &shader_notify : nullptr};