aboutsummaryrefslogtreecommitdiff
path: root/src/video_core/shader
diff options
context:
space:
mode:
Diffstat (limited to 'src/video_core/shader')
-rw-r--r--src/video_core/shader/ast.h25
-rw-r--r--src/video_core/shader/async_shaders.h11
-rw-r--r--src/video_core/shader/control_flow.cpp27
-rw-r--r--src/video_core/shader/decode/arithmetic_half.cpp3
-rw-r--r--src/video_core/shader/decode/arithmetic_integer_immediate.cpp35
-rw-r--r--src/video_core/shader/decode/image.cpp15
-rw-r--r--src/video_core/shader/decode/texture.cpp2
-rw-r--r--src/video_core/shader/registry.cpp50
-rw-r--r--src/video_core/shader/registry.h2
-rw-r--r--src/video_core/shader/track.cpp4
10 files changed, 96 insertions, 78 deletions
diff --git a/src/video_core/shader/ast.h b/src/video_core/shader/ast.h
index cca13bcde..8e5a22ab3 100644
--- a/src/video_core/shader/ast.h
+++ b/src/video_core/shader/ast.h
@@ -199,55 +199,48 @@ public:
}
std::optional<u32> GetGotoLabel() const {
- auto inner = std::get_if<ASTGoto>(&data);
- if (inner) {
+ if (const auto* inner = std::get_if<ASTGoto>(&data)) {
return {inner->label};
}
- return {};
+ return std::nullopt;
}
Expr GetGotoCondition() const {
- auto inner = std::get_if<ASTGoto>(&data);
- if (inner) {
+ if (const auto* inner = std::get_if<ASTGoto>(&data)) {
return inner->condition;
}
return nullptr;
}
void MarkLabelUnused() {
- auto inner = std::get_if<ASTLabel>(&data);
- if (inner) {
+ if (auto* inner = std::get_if<ASTLabel>(&data)) {
inner->unused = true;
}
}
bool IsLabelUnused() const {
- auto inner = std::get_if<ASTLabel>(&data);
- if (inner) {
+ if (const auto* inner = std::get_if<ASTLabel>(&data)) {
return inner->unused;
}
return true;
}
std::optional<u32> GetLabelIndex() const {
- auto inner = std::get_if<ASTLabel>(&data);
- if (inner) {
+ if (const auto* inner = std::get_if<ASTLabel>(&data)) {
return {inner->index};
}
- return {};
+ return std::nullopt;
}
Expr GetIfCondition() const {
- auto inner = std::get_if<ASTIfThen>(&data);
- if (inner) {
+ if (const auto* inner = std::get_if<ASTIfThen>(&data)) {
return inner->condition;
}
return nullptr;
}
void SetGotoCondition(Expr new_condition) {
- auto inner = std::get_if<ASTGoto>(&data);
- if (inner) {
+ if (auto* inner = std::get_if<ASTGoto>(&data)) {
inner->condition = std::move(new_condition);
}
}
diff --git a/src/video_core/shader/async_shaders.h b/src/video_core/shader/async_shaders.h
index 7cf8d994c..7a99e1dc5 100644
--- a/src/video_core/shader/async_shaders.h
+++ b/src/video_core/shader/async_shaders.h
@@ -9,6 +9,17 @@
#include <shared_mutex>
#include <thread>
+// This header includes both Vulkan and OpenGL headers, this has to be fixed
+// Unfortunately, including OpenGL will include Windows.h that defines macros that can cause issues.
+// Forcefully include glad early and undefine macros
+#include <glad/glad.h>
+#ifdef CreateEvent
+#undef CreateEvent
+#endif
+#ifdef CreateSemaphore
+#undef CreateSemaphore
+#endif
+
#include "common/common_types.h"
#include "video_core/renderer_opengl/gl_device.h"
#include "video_core/renderer_opengl/gl_resource_manager.h"
diff --git a/src/video_core/shader/control_flow.cpp b/src/video_core/shader/control_flow.cpp
index 336397cdb..4c8971615 100644
--- a/src/video_core/shader/control_flow.cpp
+++ b/src/video_core/shader/control_flow.cpp
@@ -547,13 +547,13 @@ bool TryQuery(CFGRebuildState& state) {
gather_labels(q2.ssy_stack, state.ssy_labels, block);
gather_labels(q2.pbk_stack, state.pbk_labels, block);
if (std::holds_alternative<SingleBranch>(*block.branch)) {
- const auto branch = std::get_if<SingleBranch>(block.branch.get());
+ auto* branch = std::get_if<SingleBranch>(block.branch.get());
if (!branch->condition.IsUnconditional()) {
q2.address = block.end + 1;
state.queries.push_back(q2);
}
- Query conditional_query{q2};
+ auto& conditional_query = state.queries.emplace_back(q2);
if (branch->is_sync) {
if (branch->address == unassigned_branch) {
branch->address = conditional_query.ssy_stack.top();
@@ -567,21 +567,21 @@ bool TryQuery(CFGRebuildState& state) {
conditional_query.pbk_stack.pop();
}
conditional_query.address = branch->address;
- state.queries.push_back(std::move(conditional_query));
return true;
}
- const auto multi_branch = std::get_if<MultiBranch>(block.branch.get());
+
+ const auto* multi_branch = std::get_if<MultiBranch>(block.branch.get());
for (const auto& branch_case : multi_branch->branches) {
- Query conditional_query{q2};
+ auto& conditional_query = state.queries.emplace_back(q2);
conditional_query.address = branch_case.address;
- state.queries.push_back(std::move(conditional_query));
}
+
return true;
}
void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) {
- const auto get_expr = ([&](const Condition& cond) -> Expr {
- Expr result{};
+ const auto get_expr = [](const Condition& cond) -> Expr {
+ Expr result;
if (cond.cc != ConditionCode::T) {
result = MakeExpr<ExprCondCode>(cond.cc);
}
@@ -594,10 +594,10 @@ void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) {
}
Expr extra = MakeExpr<ExprPredicate>(pred);
if (negate) {
- extra = MakeExpr<ExprNot>(extra);
+ extra = MakeExpr<ExprNot>(std::move(extra));
}
if (result) {
- return MakeExpr<ExprAnd>(extra, result);
+ return MakeExpr<ExprAnd>(std::move(extra), std::move(result));
}
return extra;
}
@@ -605,9 +605,10 @@ void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) {
return result;
}
return MakeExpr<ExprBoolean>(true);
- });
+ };
+
if (std::holds_alternative<SingleBranch>(*branch_info)) {
- const auto branch = std::get_if<SingleBranch>(branch_info.get());
+ const auto* branch = std::get_if<SingleBranch>(branch_info.get());
if (branch->address < 0) {
if (branch->kill) {
mm.InsertReturn(get_expr(branch->condition), true);
@@ -619,7 +620,7 @@ void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch_info) {
mm.InsertGoto(get_expr(branch->condition), branch->address);
return;
}
- const auto multi_branch = std::get_if<MultiBranch>(branch_info.get());
+ const auto* multi_branch = std::get_if<MultiBranch>(branch_info.get());
for (const auto& branch_case : multi_branch->branches) {
mm.InsertGoto(MakeExpr<ExprGprEqual>(multi_branch->gpr, branch_case.cmp_value),
branch_case.address);
diff --git a/src/video_core/shader/decode/arithmetic_half.cpp b/src/video_core/shader/decode/arithmetic_half.cpp
index a276aee44..88103fede 100644
--- a/src/video_core/shader/decode/arithmetic_half.cpp
+++ b/src/video_core/shader/decode/arithmetic_half.cpp
@@ -53,6 +53,9 @@ u32 ShaderIR::DecodeArithmeticHalf(NodeBlock& bb, u32 pc) {
absolute_a = ((instr.value >> 44) & 1) != 0;
absolute_b = ((instr.value >> 54) & 1) != 0;
break;
+ default:
+ UNREACHABLE();
+ break;
}
Node op_a = UnpackHalfFloat(GetRegister(instr.gpr8), instr.alu_half.type_a);
diff --git a/src/video_core/shader/decode/arithmetic_integer_immediate.cpp b/src/video_core/shader/decode/arithmetic_integer_immediate.cpp
index 73880db0e..2a30aab2b 100644
--- a/src/video_core/shader/decode/arithmetic_integer_immediate.cpp
+++ b/src/video_core/shader/decode/arithmetic_integer_immediate.cpp
@@ -28,23 +28,26 @@ u32 ShaderIR::DecodeArithmeticIntegerImmediate(NodeBlock& bb, u32 pc) {
case OpCode::Id::IADD32I: {
UNIMPLEMENTED_IF_MSG(instr.iadd32i.saturate, "IADD32I saturation is not implemented");
- op_a = GetOperandAbsNegInteger(op_a, false, instr.iadd32i.negate_a, true);
+ op_a = GetOperandAbsNegInteger(std::move(op_a), false, instr.iadd32i.negate_a != 0, true);
- const Node value = Operation(OperationCode::IAdd, PRECISE, op_a, op_b);
+ Node value = Operation(OperationCode::IAdd, PRECISE, std::move(op_a), std::move(op_b));
- SetInternalFlagsFromInteger(bb, value, instr.op_32.generates_cc);
- SetRegister(bb, instr.gpr0, value);
+ SetInternalFlagsFromInteger(bb, value, instr.op_32.generates_cc != 0);
+ SetRegister(bb, instr.gpr0, std::move(value));
break;
}
case OpCode::Id::LOP32I: {
- if (instr.alu.lop32i.invert_a)
- op_a = Operation(OperationCode::IBitwiseNot, NO_PRECISE, op_a);
+ if (instr.alu.lop32i.invert_a) {
+ op_a = Operation(OperationCode::IBitwiseNot, NO_PRECISE, std::move(op_a));
+ }
- if (instr.alu.lop32i.invert_b)
- op_b = Operation(OperationCode::IBitwiseNot, NO_PRECISE, op_b);
+ if (instr.alu.lop32i.invert_b) {
+ op_b = Operation(OperationCode::IBitwiseNot, NO_PRECISE, std::move(op_b));
+ }
- WriteLogicOperation(bb, instr.gpr0, instr.alu.lop32i.operation, op_a, op_b,
- PredicateResultMode::None, Pred::UnusedIndex, instr.op_32.generates_cc);
+ WriteLogicOperation(bb, instr.gpr0, instr.alu.lop32i.operation, std::move(op_a),
+ std::move(op_b), PredicateResultMode::None, Pred::UnusedIndex,
+ instr.op_32.generates_cc != 0);
break;
}
default:
@@ -58,14 +61,14 @@ u32 ShaderIR::DecodeArithmeticIntegerImmediate(NodeBlock& bb, u32 pc) {
void ShaderIR::WriteLogicOperation(NodeBlock& bb, Register dest, LogicOperation logic_op, Node op_a,
Node op_b, PredicateResultMode predicate_mode, Pred predicate,
bool sets_cc) {
- const Node result = [&]() {
+ Node result = [&] {
switch (logic_op) {
case LogicOperation::And:
- return Operation(OperationCode::IBitwiseAnd, PRECISE, op_a, op_b);
+ return Operation(OperationCode::IBitwiseAnd, PRECISE, std::move(op_a), std::move(op_b));
case LogicOperation::Or:
- return Operation(OperationCode::IBitwiseOr, PRECISE, op_a, op_b);
+ return Operation(OperationCode::IBitwiseOr, PRECISE, std::move(op_a), std::move(op_b));
case LogicOperation::Xor:
- return Operation(OperationCode::IBitwiseXor, PRECISE, op_a, op_b);
+ return Operation(OperationCode::IBitwiseXor, PRECISE, std::move(op_a), std::move(op_b));
case LogicOperation::PassB:
return op_b;
default:
@@ -84,8 +87,8 @@ void ShaderIR::WriteLogicOperation(NodeBlock& bb, Register dest, LogicOperation
return;
case PredicateResultMode::NotZero: {
// Set the predicate to true if the result is not zero.
- const Node compare = Operation(OperationCode::LogicalINotEqual, result, Immediate(0));
- SetPredicate(bb, static_cast<u64>(predicate), compare);
+ Node compare = Operation(OperationCode::LogicalINotEqual, std::move(result), Immediate(0));
+ SetPredicate(bb, static_cast<u64>(predicate), std::move(compare));
break;
}
default:
diff --git a/src/video_core/shader/decode/image.cpp b/src/video_core/shader/decode/image.cpp
index e75ca4fdb..618d309d2 100644
--- a/src/video_core/shader/decode/image.cpp
+++ b/src/video_core/shader/decode/image.cpp
@@ -119,6 +119,8 @@ ComponentType GetComponentType(Tegra::Engines::SamplerDescriptor descriptor,
return descriptor.r_type;
}
break;
+ default:
+ break;
}
UNIMPLEMENTED_MSG("Texture format not implemented={}", format);
return ComponentType::FLOAT;
@@ -220,9 +222,10 @@ u32 GetComponentSize(TextureFormat format, std::size_t component) {
return (component == 0 || component == 1) ? 8 : 0;
case TextureFormat::G4R4:
return (component == 0 || component == 1) ? 4 : 0;
+ default:
+ UNIMPLEMENTED_MSG("Texture format not implemented={}", format);
+ return 0;
}
- UNIMPLEMENTED_MSG("Texture format not implemented={}", format);
- return 0;
}
std::size_t GetImageComponentMask(TextureFormat format) {
@@ -257,9 +260,10 @@ std::size_t GetImageComponentMask(TextureFormat format) {
case TextureFormat::R8:
case TextureFormat::R1:
return std::size_t{R};
+ default:
+ UNIMPLEMENTED_MSG("Texture format not implemented={}", format);
+ return std::size_t{R | G | B | A};
}
- UNIMPLEMENTED_MSG("Texture format not implemented={}", format);
- return std::size_t{R | G | B | A};
}
std::size_t GetImageTypeNumCoordinates(Tegra::Shader::ImageType image_type) {
@@ -463,7 +467,10 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) {
return OperationCode::AtomicImageXor;
case Tegra::Shader::ImageAtomicOperation::Exch:
return OperationCode::AtomicImageExchange;
+ default:
+ break;
}
+ break;
default:
break;
}
diff --git a/src/video_core/shader/decode/texture.cpp b/src/video_core/shader/decode/texture.cpp
index 29ebf65ba..a03b50e39 100644
--- a/src/video_core/shader/decode/texture.cpp
+++ b/src/video_core/shader/decode/texture.cpp
@@ -763,7 +763,7 @@ Node4 ShaderIR::GetTld4Code(Instruction instr, TextureType texture_type, bool de
Node4 ShaderIR::GetTldCode(Tegra::Shader::Instruction instr) {
const auto texture_type{instr.tld.texture_type};
- const bool is_array{instr.tld.is_array};
+ const bool is_array{instr.tld.is_array != 0};
const bool lod_enabled{instr.tld.GetTextureProcessMode() == TextureProcessMode::LL};
const std::size_t coord_count{GetCoordCount(texture_type)};
diff --git a/src/video_core/shader/registry.cpp b/src/video_core/shader/registry.cpp
index cdf274e54..148d91fcb 100644
--- a/src/video_core/shader/registry.cpp
+++ b/src/video_core/shader/registry.cpp
@@ -24,44 +24,45 @@ GraphicsInfo MakeGraphicsInfo(ShaderType shader_stage, ConstBufferEngineInterfac
if (shader_stage == ShaderType::Compute) {
return {};
}
- auto& graphics = static_cast<Tegra::Engines::Maxwell3D&>(engine);
-
- GraphicsInfo info;
- info.tfb_layouts = graphics.regs.tfb_layouts;
- info.tfb_varying_locs = graphics.regs.tfb_varying_locs;
- info.primitive_topology = graphics.regs.draw.topology;
- info.tessellation_primitive = graphics.regs.tess_mode.prim;
- info.tessellation_spacing = graphics.regs.tess_mode.spacing;
- info.tfb_enabled = graphics.regs.tfb_enabled;
- info.tessellation_clockwise = graphics.regs.tess_mode.cw;
- return info;
+
+ auto& graphics = dynamic_cast<Tegra::Engines::Maxwell3D&>(engine);
+
+ return {
+ .tfb_layouts = graphics.regs.tfb_layouts,
+ .tfb_varying_locs = graphics.regs.tfb_varying_locs,
+ .primitive_topology = graphics.regs.draw.topology,
+ .tessellation_primitive = graphics.regs.tess_mode.prim,
+ .tessellation_spacing = graphics.regs.tess_mode.spacing,
+ .tfb_enabled = graphics.regs.tfb_enabled != 0,
+ .tessellation_clockwise = graphics.regs.tess_mode.cw.Value() != 0,
+ };
}
ComputeInfo MakeComputeInfo(ShaderType shader_stage, ConstBufferEngineInterface& engine) {
if (shader_stage != ShaderType::Compute) {
return {};
}
- auto& compute = static_cast<Tegra::Engines::KeplerCompute&>(engine);
+
+ auto& compute = dynamic_cast<Tegra::Engines::KeplerCompute&>(engine);
const auto& launch = compute.launch_description;
- ComputeInfo info;
- info.workgroup_size = {launch.block_dim_x, launch.block_dim_y, launch.block_dim_z};
- info.local_memory_size_in_words = launch.local_pos_alloc;
- info.shared_memory_size_in_words = launch.shared_alloc;
- return info;
+ return {
+ .workgroup_size = {launch.block_dim_x, launch.block_dim_y, launch.block_dim_z},
+ .shared_memory_size_in_words = launch.shared_alloc,
+ .local_memory_size_in_words = launch.local_pos_alloc,
+ };
}
} // Anonymous namespace
-Registry::Registry(Tegra::Engines::ShaderType shader_stage, const SerializedRegistryInfo& info)
+Registry::Registry(ShaderType shader_stage, const SerializedRegistryInfo& info)
: stage{shader_stage}, stored_guest_driver_profile{info.guest_driver_profile},
bound_buffer{info.bound_buffer}, graphics_info{info.graphics}, compute_info{info.compute} {}
-Registry::Registry(Tegra::Engines::ShaderType shader_stage,
- Tegra::Engines::ConstBufferEngineInterface& engine)
- : stage{shader_stage}, engine{&engine}, bound_buffer{engine.GetBoundBuffer()},
- graphics_info{MakeGraphicsInfo(shader_stage, engine)}, compute_info{MakeComputeInfo(
- shader_stage, engine)} {}
+Registry::Registry(ShaderType shader_stage, ConstBufferEngineInterface& engine_)
+ : stage{shader_stage}, engine{&engine_}, bound_buffer{engine_.GetBoundBuffer()},
+ graphics_info{MakeGraphicsInfo(shader_stage, engine_)}, compute_info{MakeComputeInfo(
+ shader_stage, engine_)} {}
Registry::~Registry() = default;
@@ -113,8 +114,7 @@ std::optional<Tegra::Engines::SamplerDescriptor> Registry::ObtainSeparateSampler
return value;
}
-std::optional<Tegra::Engines::SamplerDescriptor> Registry::ObtainBindlessSampler(u32 buffer,
- u32 offset) {
+std::optional<SamplerDescriptor> Registry::ObtainBindlessSampler(u32 buffer, u32 offset) {
const std::pair key = {buffer, offset};
const auto iter = bindless_samplers.find(key);
if (iter != bindless_samplers.end()) {
diff --git a/src/video_core/shader/registry.h b/src/video_core/shader/registry.h
index 231206765..4bebefdde 100644
--- a/src/video_core/shader/registry.h
+++ b/src/video_core/shader/registry.h
@@ -94,7 +94,7 @@ public:
explicit Registry(Tegra::Engines::ShaderType shader_stage, const SerializedRegistryInfo& info);
explicit Registry(Tegra::Engines::ShaderType shader_stage,
- Tegra::Engines::ConstBufferEngineInterface& engine);
+ Tegra::Engines::ConstBufferEngineInterface& engine_);
~Registry();
diff --git a/src/video_core/shader/track.cpp b/src/video_core/shader/track.cpp
index d5ed81442..6be3ea92b 100644
--- a/src/video_core/shader/track.cpp
+++ b/src/video_core/shader/track.cpp
@@ -205,12 +205,12 @@ std::optional<u32> ShaderIR::TrackImmediate(Node tracked, const NodeBlock& code,
const auto result = TrackRegister(&std::get<GprNode>(*tracked), code, cursor - 1);
const auto& found = result.first;
if (!found) {
- return {};
+ return std::nullopt;
}
if (const auto immediate = std::get_if<ImmediateNode>(&*found)) {
return immediate->GetValue();
}
- return {};
+ return std::nullopt;
}
std::pair<Node, s64> ShaderIR::TrackRegister(const GprNode* tracked, const NodeBlock& code,
generated by cgit v1.2.3 (git 2.25.1) at 2025-12-29 16:14:10 +0000