Merge pull request #2346 from yuriks/shader-refactor2

More shader refactoring
master
Yuri Kunde Schlesner 2017-01-27 14:29:10 +07:00 committed by GitHub
commit bf14f4be22
13 changed files with 1192 additions and 1113 deletions

@ -18,7 +18,9 @@
#include "citra_qt/util/util.h"
#include "video_core/pica.h"
#include "video_core/pica_state.h"
#include "video_core/shader/debug_data.h"
#include "video_core/shader/shader.h"
#include "video_core/shader/shader_interpreter.h"
using nihstro::OpCode;
using nihstro::Instruction;
@ -518,8 +520,9 @@ void GraphicsVertexShaderWidget::Reload(bool replace_vertex_data, void* vertex_d
info.labels.insert({entry_point, "main"});
// Generate debug information
debug_data = Pica::g_state.vs.ProduceDebugInfo(input_vertex, num_attributes, shader_config,
shader_setup);
Pica::Shader::InterpreterEngine shader_engine;
shader_engine.SetupBatch(shader_setup, entry_point);
debug_data = shader_engine.ProduceDebugInfo(shader_setup, input_vertex, num_attributes);
// Reload widget state
for (int attr = 0; attr < num_attributes; ++attr) {

@ -8,6 +8,7 @@
#include <QTreeView>
#include "citra_qt/debugger/graphics/graphics_breakpoint_observer.h"
#include "nihstro/parser_shbin.h"
#include "video_core/shader/debug_data.h"
#include "video_core/shader/shader.h"
class QLabel;

@ -50,10 +50,12 @@ set(HEADERS
if(ARCHITECTURE_x86_64)
set(SRCS ${SRCS}
shader/shader_jit_x64.cpp)
shader/shader_jit_x64.cpp
shader/shader_jit_x64_compiler.cpp)
set(HEADERS ${HEADERS}
shader/shader_jit_x64.h)
shader/shader_jit_x64.h
shader/shader_jit_x64_compiler.h)
endif()
create_directory_groups(${SRCS} ${HEADERS})

@ -142,16 +142,18 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
MICROPROFILE_SCOPE(GPU_Drawing);
immediate_attribute_id = 0;
Shader::UnitState shader_unit;
g_state.vs.Setup();
auto* shader_engine = Shader::GetEngine();
shader_engine->SetupBatch(g_state.vs, regs.vs.main_offset);
// Send to vertex shader
if (g_debug_context)
g_debug_context->OnEvent(DebugContext::Event::VertexShaderInvocation,
static_cast<void*>(&immediate_input));
g_state.vs.Run(shader_unit, immediate_input, regs.vs.num_input_attributes + 1);
Shader::OutputVertex output_vertex =
shader_unit.output_registers.ToVertex(regs.vs);
Shader::UnitState shader_unit;
shader_unit.LoadInputVertex(immediate_input, regs.vs.num_input_attributes + 1);
shader_engine->Run(g_state.vs, shader_unit);
auto output_vertex = Shader::OutputVertex::FromRegisters(
shader_unit.registers.output, regs, regs.vs.output_mask);
// Send to renderer
using Pica::Shader::OutputVertex;
@ -243,8 +245,10 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
unsigned int vertex_cache_pos = 0;
vertex_cache_ids.fill(-1);
auto* shader_engine = Shader::GetEngine();
Shader::UnitState shader_unit;
g_state.vs.Setup();
shader_engine->SetupBatch(g_state.vs, regs.vs.main_offset);
for (unsigned int index = 0; index < regs.num_vertices; ++index) {
// Indexed rendering doesn't use the start offset
@ -283,10 +287,12 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
if (g_debug_context)
g_debug_context->OnEvent(DebugContext::Event::VertexShaderInvocation,
(void*)&input);
g_state.vs.Run(shader_unit, input, loader.GetNumTotalAttributes());
shader_unit.LoadInputVertex(input, loader.GetNumTotalAttributes());
shader_engine->Run(g_state.vs, shader_unit);
// Retrieve vertex from register data
output_vertex = shader_unit.output_registers.ToVertex(regs.vs);
output_vertex = Shader::OutputVertex::FromRegisters(shader_unit.registers.output,
regs, regs.vs.output_mask);
if (is_indexed) {
vertex_cache[vertex_cache_pos] = output_vertex;

@ -499,7 +499,7 @@ void Init() {
}
void Shutdown() {
Shader::ClearCache();
Shader::Shutdown();
}
template <typename T>

@ -2,14 +2,8 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <atomic>
#include <cmath>
#include <cstring>
#include <unordered_map>
#include <utility>
#include <boost/range/algorithm/fill.hpp>
#include "common/bit_field.h"
#include "common/hash.h"
#include "common/logging/log.h"
#include "common/microprofile.h"
#include "video_core/pica.h"
@ -25,7 +19,8 @@ namespace Pica {
namespace Shader {
OutputVertex OutputRegisters::ToVertex(const Regs::ShaderConfig& config) const {
OutputVertex OutputVertex::FromRegisters(Math::Vec4<float24> output_regs[16], const Regs& regs,
u32 output_mask) {
// Setup output data
OutputVertex ret;
// TODO(neobrain): Under some circumstances, up to 16 attributes may be output. We need to
@ -33,13 +28,13 @@ OutputVertex OutputRegisters::ToVertex(const Regs::ShaderConfig& config) const {
unsigned index = 0;
for (unsigned i = 0; i < 7; ++i) {
if (index >= g_state.regs.vs_output_total)
if (index >= regs.vs_output_total)
break;
if ((config.output_mask & (1 << i)) == 0)
if ((output_mask & (1 << i)) == 0)
continue;
const auto& output_register_map = g_state.regs.vs_output_attributes[index];
const auto& output_register_map = regs.vs_output_attributes[index];
u32 semantics[4] = {output_register_map.map_x, output_register_map.map_y,
output_register_map.map_z, output_register_map.map_w};
@ -47,7 +42,7 @@ OutputVertex OutputRegisters::ToVertex(const Regs::ShaderConfig& config) const {
for (unsigned comp = 0; comp < 4; ++comp) {
float24* out = ((float24*)&ret) + semantics[comp];
if (semantics[comp] != Regs::VSOutputAttributes::INVALID) {
*out = value[i][comp];
*out = output_regs[i][comp];
} else {
// Zero output so that attributes which aren't output won't have denormals in them,
// which would slow us down later.
@ -76,84 +71,39 @@ OutputVertex OutputRegisters::ToVertex(const Regs::ShaderConfig& config) const {
return ret;
}
#ifdef ARCHITECTURE_x86_64
static std::unordered_map<u64, std::unique_ptr<JitShader>> shader_map;
static const JitShader* jit_shader;
#endif // ARCHITECTURE_x86_64
void UnitState::LoadInputVertex(const InputVertex& input, int num_attributes) {
// Setup input register table
const auto& attribute_register_map = g_state.regs.vs.input_register_map;
void ClearCache() {
#ifdef ARCHITECTURE_x86_64
shader_map.clear();
#endif // ARCHITECTURE_x86_64
}
void ShaderSetup::Setup() {
#ifdef ARCHITECTURE_x86_64
if (VideoCore::g_shader_jit_enabled) {
u64 cache_key =
Common::ComputeHash64(&g_state.vs.program_code, sizeof(g_state.vs.program_code)) ^
Common::ComputeHash64(&g_state.vs.swizzle_data, sizeof(g_state.vs.swizzle_data));
auto iter = shader_map.find(cache_key);
if (iter != shader_map.end()) {
jit_shader = iter->second.get();
} else {
auto shader = std::make_unique<JitShader>();
shader->Compile();
jit_shader = shader.get();
shader_map[cache_key] = std::move(shader);
}
}
#endif // ARCHITECTURE_x86_64
for (int i = 0; i < num_attributes; i++)
registers.input[attribute_register_map.GetRegisterForAttribute(i)] = input.attr[i];
}
MICROPROFILE_DEFINE(GPU_Shader, "GPU", "Shader", MP_RGB(50, 50, 240));
void ShaderSetup::Run(UnitState& state, const InputVertex& input, int num_attributes) {
auto& config = g_state.regs.vs;
auto& setup = g_state.vs;
MICROPROFILE_SCOPE(GPU_Shader);
// Setup input register table
const auto& attribute_register_map = config.input_register_map;
for (int i = 0; i < num_attributes; i++)
state.registers.input[attribute_register_map.GetRegisterForAttribute(i)] = input.attr[i];
state.conditional_code[0] = false;
state.conditional_code[1] = false;
#ifdef ARCHITECTURE_x86_64
if (VideoCore::g_shader_jit_enabled) {
jit_shader->Run(setup, state, config.main_offset);
} else {
DebugData<false> dummy_debug_data;
RunInterpreter(setup, state, dummy_debug_data, config.main_offset);
}
#else
DebugData<false> dummy_debug_data;
RunInterpreter(setup, state, dummy_debug_data, config.main_offset);
static std::unique_ptr<JitX64Engine> jit_engine;
#endif // ARCHITECTURE_x86_64
static InterpreterEngine interpreter_engine;
ShaderEngine* GetEngine() {
#ifdef ARCHITECTURE_x86_64
// TODO(yuriks): Re-initialize on each change rather than being persistent
if (VideoCore::g_shader_jit_enabled) {
if (jit_engine == nullptr) {
jit_engine = std::make_unique<JitX64Engine>();
}
return jit_engine.get();
}
#endif // ARCHITECTURE_x86_64
return &interpreter_engine;
}
DebugData<true> ShaderSetup::ProduceDebugInfo(const InputVertex& input, int num_attributes,
const Regs::ShaderConfig& config,
const ShaderSetup& setup) {
UnitState state;
DebugData<true> debug_data;
// Setup input register table
boost::fill(state.registers.input, Math::Vec4<float24>::AssignToAll(float24::Zero()));
const auto& attribute_register_map = config.input_register_map;
for (int i = 0; i < num_attributes; i++)
state.registers.input[attribute_register_map.GetRegisterForAttribute(i)] = input.attr[i];
state.conditional_code[0] = false;
state.conditional_code[1] = false;
RunInterpreter(setup, state, debug_data, config.main_offset);
return debug_data;
void Shutdown() {
#ifdef ARCHITECTURE_x86_64
jit_engine = nullptr;
#endif // ARCHITECTURE_x86_64
}
} // namespace Shader

@ -6,7 +6,6 @@
#include <array>
#include <cstddef>
#include <memory>
#include <type_traits>
#include <nihstro/shader_bytecode.h>
#include "common/assert.h"
@ -15,7 +14,6 @@
#include "common/vector_math.h"
#include "video_core/pica.h"
#include "video_core/pica_types.h"
#include "video_core/shader/debug_data.h"
using nihstro::RegisterType;
using nihstro::SourceRegister;
@ -75,19 +73,13 @@ struct OutputVertex {
ret.Lerp(factor, v1);
return ret;
}
static OutputVertex FromRegisters(Math::Vec4<float24> output_regs[16], const Regs& regs,
u32 output_mask);
};
static_assert(std::is_pod<OutputVertex>::value, "Structure is not POD");
static_assert(sizeof(OutputVertex) == 32 * sizeof(float), "OutputVertex has invalid size");
struct OutputRegisters {
OutputRegisters() = default;
alignas(16) Math::Vec4<float24> value[16];
OutputVertex ToVertex(const Regs::ShaderConfig& config) const;
};
static_assert(std::is_pod<OutputRegisters>::value, "Structure is not POD");
/**
* This structure contains the state information that needs to be unique for a shader unit. The 3DS
* has four shader units that process shaders in parallel. At the present, Citra only implements a
@ -100,11 +92,10 @@ struct UnitState {
// required to be 16-byte aligned.
alignas(16) Math::Vec4<float24> input[16];
alignas(16) Math::Vec4<float24> temporary[16];
alignas(16) Math::Vec4<float24> output[16];
} registers;
static_assert(std::is_pod<Registers>::value, "Structure is not POD");
OutputRegisters output_registers;
bool conditional_code[2];
// Two Address registers and one loop counter
@ -130,7 +121,7 @@ struct UnitState {
static size_t OutputOffset(const DestRegister& reg) {
switch (reg.GetRegisterType()) {
case RegisterType::Output:
return offsetof(UnitState, output_registers.value) +
return offsetof(UnitState, registers.output) +
reg.GetIndex() * sizeof(Math::Vec4<float24>);
case RegisterType::Temporary:
@ -142,13 +133,17 @@ struct UnitState {
return 0;
}
}
/**
* Loads the unit state with an input vertex.
*
* @param input Input vertex into the shader
* @param num_attributes The number of vertex shader attributes to load
*/
void LoadInputVertex(const InputVertex& input, int num_attributes);
};
/// Clears the shader cache
void ClearCache();
struct ShaderSetup {
struct {
// The float uniforms are accessed by the shader JIT using SSE instructions, and are
// therefore required to be 16-byte aligned.
@ -173,32 +168,37 @@ struct ShaderSetup {
std::array<u32, 1024> program_code;
std::array<u32, 1024> swizzle_data;
/// Data private to ShaderEngines
struct EngineData {
unsigned int entry_point;
/// Used by the JIT, points to a compiled shader object.
const void* cached_shader = nullptr;
} engine_data;
};
class ShaderEngine {
public:
virtual ~ShaderEngine() = default;
/**
* Performs any shader unit setup that only needs to happen once per shader (as opposed to once
* per vertex, which would happen within the `Run` function).
*/
void Setup();
virtual void SetupBatch(ShaderSetup& setup, unsigned int entry_point) = 0;
/**
* Runs the currently setup shader
* @param state Shader unit state, must be setup per shader and per shader unit
* @param input Input vertex into the shader
* @param num_attributes The number of vertex shader attributes
* Runs the currently setup shader.
*
* @param setup Shader engine state, must be setup with SetupBatch on each shader change.
* @param state Shader unit state, must be setup with input data before each shader invocation.
*/
void Run(UnitState& state, const InputVertex& input, int num_attributes);
/**
* Produce debug information based on the given shader and input vertex
* @param input Input vertex into the shader
* @param num_attributes The number of vertex shader attributes
* @param config Configuration object for the shader pipeline
* @param setup Setup object for the shader pipeline
* @return Debug information for this shader with regards to the given vertex
*/
DebugData<true> ProduceDebugInfo(const InputVertex& input, int num_attributes,
const Regs::ShaderConfig& config, const ShaderSetup& setup);
virtual void Run(const ShaderSetup& setup, UnitState& state) const = 0;
};
// TODO(yuriks): Remove and make it non-global state somewhere
ShaderEngine* GetEngine();
void Shutdown();
} // namespace Shader
} // namespace Pica

@ -7,10 +7,12 @@
#include <cmath>
#include <numeric>
#include <boost/container/static_vector.hpp>
#include <boost/range/algorithm/fill.hpp>
#include <nihstro/shader_bytecode.h>
#include "common/assert.h"
#include "common/common_types.h"
#include "common/logging/log.h"
#include "common/microprofile.h"
#include "common/vector_math.h"
#include "video_core/pica_state.h"
#include "video_core/pica_types.h"
@ -37,12 +39,15 @@ struct CallStackElement {
};
template <bool Debug>
void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData<Debug>& debug_data,
unsigned offset) {
static void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData<Debug>& debug_data,
unsigned offset) {
// TODO: Is there a maximal size for this?
boost::container::static_vector<CallStackElement, 16> call_stack;
u32 program_counter = offset;
state.conditional_code[0] = false;
state.conditional_code[1] = false;
auto call = [&program_counter, &call_stack](u32 offset, u32 num_instructions, u32 return_offset,
u8 repeat_count, u8 loop_increment) {
// -1 to make sure when incrementing the PC we end up at the correct offset
@ -73,9 +78,9 @@ void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData<Debug>
}
};
const auto& uniforms = g_state.vs.uniforms;
const auto& swizzle_data = g_state.vs.swizzle_data;
const auto& program_code = g_state.vs.program_code;
const auto& uniforms = setup.uniforms;
const auto& swizzle_data = setup.swizzle_data;
const auto& program_code = setup.program_code;
// Placeholder for invalid inputs
static float24 dummy_vec4_float24[4];
@ -170,7 +175,7 @@ void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData<Debug>
float24* dest =
(instr.common.dest.Value() < 0x10)
? &state.output_registers.value[instr.common.dest.Value().GetIndex()][0]
? &state.registers.output[instr.common.dest.Value().GetIndex()][0]
: (instr.common.dest.Value() < 0x20)
? &state.registers.temporary[instr.common.dest.Value().GetIndex()][0]
: dummy_vec4_float24;
@ -513,7 +518,7 @@ void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData<Debug>
float24* dest =
(instr.mad.dest.Value() < 0x10)
? &state.output_registers.value[instr.mad.dest.Value().GetIndex()][0]
? &state.registers.output[instr.mad.dest.Value().GetIndex()][0]
: (instr.mad.dest.Value() < 0x20)
? &state.registers.temporary[instr.mad.dest.Value().GetIndex()][0]
: dummy_vec4_float24;
@ -647,9 +652,33 @@ void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData<Debug>
}
}
// Explicit instantiation
template void RunInterpreter(const ShaderSetup&, UnitState&, DebugData<false>&, unsigned offset);
template void RunInterpreter(const ShaderSetup&, UnitState&, DebugData<true>&, unsigned offset);
void InterpreterEngine::SetupBatch(ShaderSetup& setup, unsigned int entry_point) {
ASSERT(entry_point < 1024);
setup.engine_data.entry_point = entry_point;
}
MICROPROFILE_DECLARE(GPU_Shader);
void InterpreterEngine::Run(const ShaderSetup& setup, UnitState& state) const {
MICROPROFILE_SCOPE(GPU_Shader);
DebugData<false> dummy_debug_data;
RunInterpreter(setup, state, dummy_debug_data, setup.engine_data.entry_point);
}
DebugData<true> InterpreterEngine::ProduceDebugInfo(const ShaderSetup& setup,
const InputVertex& input,
int num_attributes) const {
UnitState state;
DebugData<true> debug_data;
// Setup input register table
boost::fill(state.registers.input, Math::Vec4<float24>::AssignToAll(float24::Zero()));
state.LoadInputVertex(input, num_attributes);
RunInterpreter(setup, state, debug_data, setup.engine_data.entry_point);
return debug_data;
}
} // namespace

@ -4,18 +4,28 @@
#pragma once
#include "video_core/shader/debug_data.h"
#include "video_core/shader/shader.h"
namespace Pica {
namespace Shader {
struct UnitState;
class InterpreterEngine final : public ShaderEngine {
public:
void SetupBatch(ShaderSetup& setup, unsigned int entry_point) override;
void Run(const ShaderSetup& setup, UnitState& state) const override;
template <bool Debug>
struct DebugData;
template <bool Debug>
void RunInterpreter(const ShaderSetup& setup, UnitState& state, DebugData<Debug>& debug_data,
unsigned offset);
/**
* Produce debug information based on the given shader and input vertex
* @param input Input vertex into the shader
* @param num_attributes The number of vertex shader attributes
* @param config Configuration object for the shader pipeline
* @return Debug information for this shader with regards to the given vertex
*/
DebugData<true> ProduceDebugInfo(const ShaderSetup& setup, const InputVertex& input,
int num_attributes) const;
};
} // namespace

@ -1,888 +1,48 @@
// Copyright 2015 Citra Emulator Project
// Copyright 2016 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <algorithm>
#include <cmath>
#include <cstdint>
#include <nihstro/shader_bytecode.h>
#include <smmintrin.h>
#include <xmmintrin.h>
#include "common/assert.h"
#include "common/logging/log.h"
#include "common/vector_math.h"
#include "common/x64/cpu_detect.h"
#include "common/x64/xbyak_abi.h"
#include "common/x64/xbyak_util.h"
#include "video_core/pica_state.h"
#include "video_core/pica_types.h"
#include "common/hash.h"
#include "common/microprofile.h"
#include "video_core/shader/shader.h"
#include "video_core/shader/shader_jit_x64.h"
using namespace Common::X64;
using namespace Xbyak::util;
using Xbyak::Label;
using Xbyak::Reg32;
using Xbyak::Reg64;
using Xbyak::Xmm;
#include "video_core/shader/shader_jit_x64_compiler.h"
namespace Pica {
namespace Shader {
typedef void (JitShader::*JitFunction)(Instruction instr);
JitX64Engine::JitX64Engine() = default;
JitX64Engine::~JitX64Engine() = default;
const JitFunction instr_table[64] = {
&JitShader::Compile_ADD, // add
&JitShader::Compile_DP3, // dp3
&JitShader::Compile_DP4, // dp4
&JitShader::Compile_DPH, // dph
nullptr, // unknown
&JitShader::Compile_EX2, // ex2
&JitShader::Compile_LG2, // lg2
nullptr, // unknown
&JitShader::Compile_MUL, // mul
&JitShader::Compile_SGE, // sge
&JitShader::Compile_SLT, // slt
&JitShader::Compile_FLR, // flr
&JitShader::Compile_MAX, // max
&JitShader::Compile_MIN, // min
&JitShader::Compile_RCP, // rcp
&JitShader::Compile_RSQ, // rsq
nullptr, // unknown
nullptr, // unknown
&JitShader::Compile_MOVA, // mova
&JitShader::Compile_MOV, // mov
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
&JitShader::Compile_DPH, // dphi
nullptr, // unknown
&JitShader::Compile_SGE, // sgei
&JitShader::Compile_SLT, // slti
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
&JitShader::Compile_NOP, // nop
&JitShader::Compile_END, // end
nullptr, // break
&JitShader::Compile_CALL, // call
&JitShader::Compile_CALLC, // callc
&JitShader::Compile_CALLU, // callu
&JitShader::Compile_IF, // ifu
&JitShader::Compile_IF, // ifc
&JitShader::Compile_LOOP, // loop
nullptr, // emit
nullptr, // sete
&JitShader::Compile_JMP, // jmpc
&JitShader::Compile_JMP, // jmpu
&JitShader::Compile_CMP, // cmp
&JitShader::Compile_CMP, // cmp
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
};
void JitX64Engine::SetupBatch(ShaderSetup& setup, unsigned int entry_point) {
ASSERT(entry_point < 1024);
setup.engine_data.entry_point = entry_point;
// The following is used to alias some commonly used registers. Generally, RAX-RDX and XMM0-XMM3 can
// be used as scratch registers within a compiler function. The other registers have designated
// purposes, as documented below:
u64 code_hash = Common::ComputeHash64(&setup.program_code, sizeof(setup.program_code));
u64 swizzle_hash = Common::ComputeHash64(&setup.swizzle_data, sizeof(setup.swizzle_data));
/// Pointer to the uniform memory
static const Reg64 SETUP = r9;
/// The two 32-bit VS address offset registers set by the MOVA instruction
static const Reg64 ADDROFFS_REG_0 = r10;
static const Reg64 ADDROFFS_REG_1 = r11;
/// VS loop count register (Multiplied by 16)
static const Reg32 LOOPCOUNT_REG = r12d;
/// Current VS loop iteration number (we could probably use LOOPCOUNT_REG, but this quicker)
static const Reg32 LOOPCOUNT = esi;
/// Number to increment LOOPCOUNT_REG by on each loop iteration (Multiplied by 16)
static const Reg32 LOOPINC = edi;
/// Result of the previous CMP instruction for the X-component comparison
static const Reg64 COND0 = r13;
/// Result of the previous CMP instruction for the Y-component comparison
static const Reg64 COND1 = r14;
/// Pointer to the UnitState instance for the current VS unit
static const Reg64 STATE = r15;
/// SIMD scratch register
static const Xmm SCRATCH = xmm0;
/// Loaded with the first swizzled source register, otherwise can be used as a scratch register
static const Xmm SRC1 = xmm1;
/// Loaded with the second swizzled source register, otherwise can be used as a scratch register
static const Xmm SRC2 = xmm2;
/// Loaded with the third swizzled source register, otherwise can be used as a scratch register
static const Xmm SRC3 = xmm3;
/// Additional scratch register
static const Xmm SCRATCH2 = xmm4;
/// Constant vector of [1.0f, 1.0f, 1.0f, 1.0f], used to efficiently set a vector to one
static const Xmm ONE = xmm14;
/// Constant vector of [-0.f, -0.f, -0.f, -0.f], used to efficiently negate a vector with XOR
static const Xmm NEGBIT = xmm15;
// State registers that must not be modified by external functions calls
// Scratch registers, e.g., SRC1 and SCRATCH, have to be saved on the side if needed
static const BitSet32 persistent_regs = BuildRegSet({
// Pointers to register blocks
SETUP, STATE,
// Cached registers
ADDROFFS_REG_0, ADDROFFS_REG_1, LOOPCOUNT_REG, COND0, COND1,
// Constants
ONE, NEGBIT,
});
/// Raw constant for the source register selector that indicates no swizzling is performed
static const u8 NO_SRC_REG_SWIZZLE = 0x1b;
/// Raw constant for the destination register enable mask that indicates all components are enabled
static const u8 NO_DEST_REG_MASK = 0xf;
/**
* Get the vertex shader instruction for a given offset in the current shader program
* @param offset Offset in the current shader program of the instruction
* @return Instruction at the specified offset
*/
static Instruction GetVertexShaderInstruction(size_t offset) {
return {g_state.vs.program_code[offset]};
}
static void LogCritical(const char* msg) {
LOG_CRITICAL(HW_GPU, "%s", msg);
}
void JitShader::Compile_Assert(bool condition, const char* msg) {
if (!condition) {
mov(ABI_PARAM1, reinterpret_cast<size_t>(msg));
CallFarFunction(*this, LogCritical);
}
}
/**
* Loads and swizzles a source register into the specified XMM register.
* @param instr VS instruction, used for determining how to load the source register
* @param src_num Number indicating which source register to load (1 = src1, 2 = src2, 3 = src3)
* @param src_reg SourceRegister object corresponding to the source register to load
* @param dest Destination XMM register to store the loaded, swizzled source register
*/
void JitShader::Compile_SwizzleSrc(Instruction instr, unsigned src_num, SourceRegister src_reg,
Xmm dest) {
Reg64 src_ptr;
size_t src_offset;
if (src_reg.GetRegisterType() == RegisterType::FloatUniform) {
src_ptr = SETUP;
src_offset = ShaderSetup::GetFloatUniformOffset(src_reg.GetIndex());
u64 cache_key = code_hash ^ swizzle_hash;
auto iter = cache.find(cache_key);
if (iter != cache.end()) {
setup.engine_data.cached_shader = iter->second.get();
} else {
src_ptr = STATE;
src_offset = UnitState::InputOffset(src_reg);
}
int src_offset_disp = (int)src_offset;
ASSERT_MSG(src_offset == src_offset_disp, "Source register offset too large for int type");
unsigned operand_desc_id;
const bool is_inverted =
(0 != (instr.opcode.Value().GetInfo().subtype & OpCode::Info::SrcInversed));
unsigned address_register_index;
unsigned offset_src;
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MAD ||
instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI) {
operand_desc_id = instr.mad.operand_desc_id;
offset_src = is_inverted ? 3 : 2;
address_register_index = instr.mad.address_register_index;
} else {
operand_desc_id = instr.common.operand_desc_id;
offset_src = is_inverted ? 2 : 1;
address_register_index = instr.common.address_register_index;
}
if (src_num == offset_src && address_register_index != 0) {
switch (address_register_index) {
case 1: // address offset 1
movaps(dest, xword[src_ptr + ADDROFFS_REG_0 + src_offset_disp]);
break;
case 2: // address offset 2
movaps(dest, xword[src_ptr + ADDROFFS_REG_1 + src_offset_disp]);
break;
case 3: // address offset 3
movaps(dest, xword[src_ptr + LOOPCOUNT_REG.cvt64() + src_offset_disp]);
break;
default:
UNREACHABLE();
break;
}
} else {
// Load the source
movaps(dest, xword[src_ptr + src_offset_disp]);
}
SwizzlePattern swiz = {g_state.vs.swizzle_data[operand_desc_id]};
// Generate instructions for source register swizzling as needed
u8 sel = swiz.GetRawSelector(src_num);
if (sel != NO_SRC_REG_SWIZZLE) {
// Selector component order needs to be reversed for the SHUFPS instruction
sel = ((sel & 0xc0) >> 6) | ((sel & 3) << 6) | ((sel & 0xc) << 2) | ((sel & 0x30) >> 2);
// Shuffle inputs for swizzle
shufps(dest, dest, sel);
}
// If the source register should be negated, flip the negative bit using XOR
const bool negate[] = {swiz.negate_src1, swiz.negate_src2, swiz.negate_src3};
if (negate[src_num - 1]) {
xorps(dest, NEGBIT);
auto shader = std::make_unique<JitShader>();
shader->Compile(&setup.program_code, &setup.swizzle_data);
setup.engine_data.cached_shader = shader.get();
cache.emplace_hint(iter, cache_key, std::move(shader));
}
}
void JitShader::Compile_DestEnable(Instruction instr, Xmm src) {
DestRegister dest;
unsigned operand_desc_id;
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MAD ||
instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI) {
operand_desc_id = instr.mad.operand_desc_id;
dest = instr.mad.dest.Value();
} else {
operand_desc_id = instr.common.operand_desc_id;
dest = instr.common.dest.Value();
}
MICROPROFILE_DECLARE(GPU_Shader);
SwizzlePattern swiz = {g_state.vs.swizzle_data[operand_desc_id]};
void JitX64Engine::Run(const ShaderSetup& setup, UnitState& state) const {
ASSERT(setup.engine_data.cached_shader != nullptr);
size_t dest_offset_disp = UnitState::OutputOffset(dest);
MICROPROFILE_SCOPE(GPU_Shader);
// If all components are enabled, write the result to the destination register
if (swiz.dest_mask == NO_DEST_REG_MASK) {
// Store dest back to memory
movaps(xword[STATE + dest_offset_disp], src);
} else {
// Not all components are enabled, so mask the result when storing to the destination
// register...
movaps(SCRATCH, xword[STATE + dest_offset_disp]);
if (Common::GetCPUCaps().sse4_1) {
u8 mask = ((swiz.dest_mask & 1) << 3) | ((swiz.dest_mask & 8) >> 3) |
((swiz.dest_mask & 2) << 1) | ((swiz.dest_mask & 4) >> 1);
blendps(SCRATCH, src, mask);
} else {
movaps(SCRATCH2, src);
unpckhps(SCRATCH2, SCRATCH); // Unpack X/Y components of source and destination
unpcklps(SCRATCH, src); // Unpack Z/W components of source and destination
// Compute selector to selectively copy source components to destination for SHUFPS
// instruction
u8 sel = ((swiz.DestComponentEnabled(0) ? 1 : 0) << 0) |
((swiz.DestComponentEnabled(1) ? 3 : 2) << 2) |
((swiz.DestComponentEnabled(2) ? 0 : 1) << 4) |
((swiz.DestComponentEnabled(3) ? 2 : 3) << 6);
shufps(SCRATCH, SCRATCH2, sel);
}
// Store dest back to memory
movaps(xword[STATE + dest_offset_disp], SCRATCH);
}
const JitShader* shader = static_cast<const JitShader*>(setup.engine_data.cached_shader);
shader->Run(setup, state, setup.engine_data.entry_point);
}
void JitShader::Compile_SanitizedMul(Xmm src1, Xmm src2, Xmm scratch) {
movaps(scratch, src1);
cmpordps(scratch, src2);
mulps(src1, src2);
movaps(src2, src1);
cmpunordps(src2, src2);
xorps(scratch, src2);
andps(src1, scratch);
}
void JitShader::Compile_EvaluateCondition(Instruction instr) {
// Note: NXOR is used below to check for equality
switch (instr.flow_control.op) {
case Instruction::FlowControlType::Or:
mov(eax, COND0);
mov(ebx, COND1);
xor(eax, (instr.flow_control.refx.Value() ^ 1));
xor(ebx, (instr.flow_control.refy.Value() ^ 1));
or (eax, ebx);
break;
case Instruction::FlowControlType::And:
mov(eax, COND0);
mov(ebx, COND1);
xor(eax, (instr.flow_control.refx.Value() ^ 1));
xor(ebx, (instr.flow_control.refy.Value() ^ 1));
and(eax, ebx);
break;
case Instruction::FlowControlType::JustX:
mov(eax, COND0);
xor(eax, (instr.flow_control.refx.Value() ^ 1));
break;
case Instruction::FlowControlType::JustY:
mov(eax, COND1);
xor(eax, (instr.flow_control.refy.Value() ^ 1));
break;
}
}
void JitShader::Compile_UniformCondition(Instruction instr) {
size_t offset = ShaderSetup::GetBoolUniformOffset(instr.flow_control.bool_uniform_id);
cmp(byte[SETUP + offset], 0);
}
BitSet32 JitShader::PersistentCallerSavedRegs() {
return persistent_regs & ABI_ALL_CALLER_SAVED;
}
void JitShader::Compile_ADD(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
addps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_DP3(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
movaps(SRC2, SRC1);
shufps(SRC2, SRC2, _MM_SHUFFLE(1, 1, 1, 1));
movaps(SRC3, SRC1);
shufps(SRC3, SRC3, _MM_SHUFFLE(2, 2, 2, 2));
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 0, 0, 0));
addps(SRC1, SRC2);
addps(SRC1, SRC3);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_DP4(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(2, 3, 0, 1)); // XYZW -> ZWXY
addps(SRC1, SRC2);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 1, 2, 3)); // XYZW -> WZYX
addps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_DPH(Instruction instr) {
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::DPHI) {
Compile_SwizzleSrc(instr, 1, instr.common.src1i, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2i, SRC2);
} else {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
}
if (Common::GetCPUCaps().sse4_1) {
// Set 4th component to 1.0
blendps(SRC1, ONE, 0b1000);
} else {
// Set 4th component to 1.0
movaps(SCRATCH, SRC1);
unpckhps(SCRATCH, ONE); // XYZW, 1111 -> Z1__
unpcklpd(SRC1, SCRATCH); // XYZW, Z1__ -> XYZ1
}
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(2, 3, 0, 1)); // XYZW -> ZWXY
addps(SRC1, SRC2);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 1, 2, 3)); // XYZW -> WZYX
addps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_EX2(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
movss(xmm0, SRC1); // ABI_PARAM1
ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
CallFarFunction(*this, exp2f);
ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
shufps(xmm0, xmm0, _MM_SHUFFLE(0, 0, 0, 0)); // ABI_RETURN
movaps(SRC1, xmm0);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_LG2(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
movss(xmm0, SRC1); // ABI_PARAM1
ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
CallFarFunction(*this, log2f);
ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
shufps(xmm0, xmm0, _MM_SHUFFLE(0, 0, 0, 0)); // ABI_RETURN
movaps(SRC1, xmm0);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MUL(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_SGE(Instruction instr) {
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::SGEI) {
Compile_SwizzleSrc(instr, 1, instr.common.src1i, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2i, SRC2);
} else {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
}
cmpleps(SRC2, SRC1);
andps(SRC2, ONE);
Compile_DestEnable(instr, SRC2);
}
void JitShader::Compile_SLT(Instruction instr) {
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::SLTI) {
Compile_SwizzleSrc(instr, 1, instr.common.src1i, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2i, SRC2);
} else {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
}
cmpltps(SRC1, SRC2);
andps(SRC1, ONE);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_FLR(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
if (Common::GetCPUCaps().sse4_1) {
roundps(SRC1, SRC1, _MM_FROUND_FLOOR);
} else {
cvttps2dq(SRC1, SRC1);
cvtdq2ps(SRC1, SRC1);
}
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MAX(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
// SSE semantics match PICA200 ones: In case of NaN, SRC2 is returned.
maxps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MIN(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
// SSE semantics match PICA200 ones: In case of NaN, SRC2 is returned.
minps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MOVA(Instruction instr) {
SwizzlePattern swiz = {g_state.vs.swizzle_data[instr.common.operand_desc_id]};
if (!swiz.DestComponentEnabled(0) && !swiz.DestComponentEnabled(1)) {
return; // NoOp
}
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
// Convert floats to integers using truncation (only care about X and Y components)
cvttps2dq(SRC1, SRC1);
// Get result
movq(rax, SRC1);
// Handle destination enable
if (swiz.DestComponentEnabled(0) && swiz.DestComponentEnabled(1)) {
// Move and sign-extend low 32 bits
movsxd(ADDROFFS_REG_0, eax);
// Move and sign-extend high 32 bits
shr(rax, 32);
movsxd(ADDROFFS_REG_1, eax);
// Multiply by 16 to be used as an offset later
shl(ADDROFFS_REG_0, 4);
shl(ADDROFFS_REG_1, 4);
} else {
if (swiz.DestComponentEnabled(0)) {
// Move and sign-extend low 32 bits
movsxd(ADDROFFS_REG_0, eax);
// Multiply by 16 to be used as an offset later
shl(ADDROFFS_REG_0, 4);
} else if (swiz.DestComponentEnabled(1)) {
// Move and sign-extend high 32 bits
shr(rax, 32);
movsxd(ADDROFFS_REG_1, eax);
// Multiply by 16 to be used as an offset later
shl(ADDROFFS_REG_1, 4);
}
}
}
void JitShader::Compile_MOV(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_RCP(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
// TODO(bunnei): RCPSS is a pretty rough approximation, this might cause problems if Pica
// performs this operation more accurately. This should be checked on hardware.
rcpss(SRC1, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 0, 0, 0)); // XYWZ -> XXXX
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_RSQ(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
// TODO(bunnei): RSQRTSS is a pretty rough approximation, this might cause problems if Pica
// performs this operation more accurately. This should be checked on hardware.
rsqrtss(SRC1, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 0, 0, 0)); // XYWZ -> XXXX
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_NOP(Instruction instr) {}
void JitShader::Compile_END(Instruction instr) {
ABI_PopRegistersAndAdjustStack(*this, ABI_ALL_CALLEE_SAVED, 8);
ret();
}
void JitShader::Compile_CALL(Instruction instr) {
// Push offset of the return
push(qword, (instr.flow_control.dest_offset + instr.flow_control.num_instructions));
// Call the subroutine
call(instruction_labels[instr.flow_control.dest_offset]);
// Skip over the return offset that's on the stack
add(rsp, 8);
}
void JitShader::Compile_CALLC(Instruction instr) {
Compile_EvaluateCondition(instr);
Label b;
jz(b);
Compile_CALL(instr);
L(b);
}
void JitShader::Compile_CALLU(Instruction instr) {
Compile_UniformCondition(instr);
Label b;
jz(b);
Compile_CALL(instr);
L(b);
}
void JitShader::Compile_CMP(Instruction instr) {
using Op = Instruction::Common::CompareOpType::Op;
Op op_x = instr.common.compare_op.x;
Op op_y = instr.common.compare_op.y;
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
// SSE doesn't have greater-than (GT) or greater-equal (GE) comparison operators. You need to
// emulate them by swapping the lhs and rhs and using LT and LE. NLT and NLE can't be used here
// because they don't match when used with NaNs.
static const u8 cmp[] = {CMP_EQ, CMP_NEQ, CMP_LT, CMP_LE, CMP_LT, CMP_LE};
bool invert_op_x = (op_x == Op::GreaterThan || op_x == Op::GreaterEqual);
Xmm lhs_x = invert_op_x ? SRC2 : SRC1;
Xmm rhs_x = invert_op_x ? SRC1 : SRC2;
if (op_x == op_y) {
// Compare X-component and Y-component together
cmpps(lhs_x, rhs_x, cmp[op_x]);
movq(COND0, lhs_x);
mov(COND1, COND0);
} else {
bool invert_op_y = (op_y == Op::GreaterThan || op_y == Op::GreaterEqual);
Xmm lhs_y = invert_op_y ? SRC2 : SRC1;
Xmm rhs_y = invert_op_y ? SRC1 : SRC2;
// Compare X-component
movaps(SCRATCH, lhs_x);
cmpss(SCRATCH, rhs_x, cmp[op_x]);
// Compare Y-component
cmpps(lhs_y, rhs_y, cmp[op_y]);
movq(COND0, SCRATCH);
movq(COND1, lhs_y);
}
shr(COND0.cvt32(), 31); // ignores upper 32 bits in source
shr(COND1, 63);
}
void JitShader::Compile_MAD(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.mad.src1, SRC1);
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI) {
Compile_SwizzleSrc(instr, 2, instr.mad.src2i, SRC2);
Compile_SwizzleSrc(instr, 3, instr.mad.src3i, SRC3);
} else {
Compile_SwizzleSrc(instr, 2, instr.mad.src2, SRC2);
Compile_SwizzleSrc(instr, 3, instr.mad.src3, SRC3);
}
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
addps(SRC1, SRC3);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_IF(Instruction instr) {
Compile_Assert(instr.flow_control.dest_offset >= program_counter,
"Backwards if-statements not supported");
Label l_else, l_endif;
// Evaluate the "IF" condition
if (instr.opcode.Value() == OpCode::Id::IFU) {
Compile_UniformCondition(instr);
} else if (instr.opcode.Value() == OpCode::Id::IFC) {
Compile_EvaluateCondition(instr);
}
jz(l_else, T_NEAR);
// Compile the code that corresponds to the condition evaluating as true
Compile_Block(instr.flow_control.dest_offset);
// If there isn't an "ELSE" condition, we are done here
if (instr.flow_control.num_instructions == 0) {
L(l_else);
return;
}
jmp(l_endif, T_NEAR);
L(l_else);
// This code corresponds to the "ELSE" condition
// Comple the code that corresponds to the condition evaluating as false
Compile_Block(instr.flow_control.dest_offset + instr.flow_control.num_instructions);
L(l_endif);
}
void JitShader::Compile_LOOP(Instruction instr) {
Compile_Assert(instr.flow_control.dest_offset >= program_counter,
"Backwards loops not supported");
Compile_Assert(!looping, "Nested loops not supported");
looping = true;
// This decodes the fields from the integer uniform at index instr.flow_control.int_uniform_id.
// The Y (LOOPCOUNT_REG) and Z (LOOPINC) component are kept multiplied by 16 (Left shifted by
// 4 bits) to be used as an offset into the 16-byte vector registers later
size_t offset = ShaderSetup::GetIntUniformOffset(instr.flow_control.int_uniform_id);
mov(LOOPCOUNT, dword[SETUP + offset]);
mov(LOOPCOUNT_REG, LOOPCOUNT);
shr(LOOPCOUNT_REG, 4);
and(LOOPCOUNT_REG, 0xFF0); // Y-component is the start
mov(LOOPINC, LOOPCOUNT);
shr(LOOPINC, 12);
and(LOOPINC, 0xFF0); // Z-component is the incrementer
movzx(LOOPCOUNT, LOOPCOUNT.cvt8()); // X-component is iteration count
add(LOOPCOUNT, 1); // Iteration count is X-component + 1
Label l_loop_start;
L(l_loop_start);
Compile_Block(instr.flow_control.dest_offset + 1);
add(LOOPCOUNT_REG, LOOPINC); // Increment LOOPCOUNT_REG by Z-component
sub(LOOPCOUNT, 1); // Increment loop count by 1
jnz(l_loop_start); // Loop if not equal
looping = false;
}
void JitShader::Compile_JMP(Instruction instr) {
if (instr.opcode.Value() == OpCode::Id::JMPC)
Compile_EvaluateCondition(instr);
else if (instr.opcode.Value() == OpCode::Id::JMPU)
Compile_UniformCondition(instr);
else
UNREACHABLE();
bool inverted_condition =
(instr.opcode.Value() == OpCode::Id::JMPU) && (instr.flow_control.num_instructions & 1);
Label& b = instruction_labels[instr.flow_control.dest_offset];
if (inverted_condition) {
jz(b, T_NEAR);
} else {
jnz(b, T_NEAR);
}
}
void JitShader::Compile_Block(unsigned end) {
while (program_counter < end) {
Compile_NextInstr();
}
}
void JitShader::Compile_Return() {
// Peek return offset on the stack and check if we're at that offset
mov(rax, qword[rsp + 8]);
cmp(eax, (program_counter));
// If so, jump back to before CALL
Label b;
jnz(b);
ret();
L(b);
}
void JitShader::Compile_NextInstr() {
if (std::binary_search(return_offsets.begin(), return_offsets.end(), program_counter)) {
Compile_Return();
}
L(instruction_labels[program_counter]);
Instruction instr = GetVertexShaderInstruction(program_counter++);
OpCode::Id opcode = instr.opcode.Value();
auto instr_func = instr_table[static_cast<unsigned>(opcode)];
if (instr_func) {
// JIT the instruction!
((*this).*instr_func)(instr);
} else {
// Unhandled instruction
LOG_CRITICAL(HW_GPU, "Unhandled instruction: 0x%02x (0x%08x)",
instr.opcode.Value().EffectiveOpCode(), instr.hex);
}
}
void JitShader::FindReturnOffsets() {
return_offsets.clear();
for (size_t offset = 0; offset < g_state.vs.program_code.size(); ++offset) {
Instruction instr = GetVertexShaderInstruction(offset);
switch (instr.opcode.Value()) {
case OpCode::Id::CALL:
case OpCode::Id::CALLC:
case OpCode::Id::CALLU:
return_offsets.push_back(instr.flow_control.dest_offset +
instr.flow_control.num_instructions);
break;
default:
break;
}
}
// Sort for efficient binary search later
std::sort(return_offsets.begin(), return_offsets.end());
}
void JitShader::Compile() {
// Reset flow control state
program = (CompiledShader*)getCurr();
program_counter = 0;
looping = false;
instruction_labels.fill(Xbyak::Label());
// Find all `CALL` instructions and identify return locations
FindReturnOffsets();
// The stack pointer is 8 modulo 16 at the entry of a procedure
ABI_PushRegistersAndAdjustStack(*this, ABI_ALL_CALLEE_SAVED, 8);
mov(SETUP, ABI_PARAM1);
mov(STATE, ABI_PARAM2);
// Zero address/loop registers
xor(ADDROFFS_REG_0.cvt32(), ADDROFFS_REG_0.cvt32());
xor(ADDROFFS_REG_1.cvt32(), ADDROFFS_REG_1.cvt32());
xor(LOOPCOUNT_REG, LOOPCOUNT_REG);
// Used to set a register to one
static const __m128 one = {1.f, 1.f, 1.f, 1.f};
mov(rax, reinterpret_cast<size_t>(&one));
movaps(ONE, xword[rax]);
// Used to negate registers
static const __m128 neg = {-0.f, -0.f, -0.f, -0.f};
mov(rax, reinterpret_cast<size_t>(&neg));
movaps(NEGBIT, xword[rax]);
// Jump to start of the shader program
jmp(ABI_PARAM3);
// Compile entire program
Compile_Block(static_cast<unsigned>(g_state.vs.program_code.size()));
// Free memory that's no longer needed
return_offsets.clear();
return_offsets.shrink_to_fit();
ready();
uintptr_t size = reinterpret_cast<uintptr_t>(getCurr()) - reinterpret_cast<uintptr_t>(program);
ASSERT_MSG(size <= MAX_SHADER_SIZE, "Compiled a shader that exceeds the allocated size!");
LOG_DEBUG(HW_GPU, "Compiled shader size=%lu", size);
}
JitShader::JitShader() : Xbyak::CodeGenerator(MAX_SHADER_SIZE) {}
} // namespace Shader
} // namespace Pica

@ -1,121 +1,30 @@
// Copyright 2015 Citra Emulator Project
// Copyright 2016 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <array>
#include <cstddef>
#include <utility>
#include <vector>
#include <nihstro/shader_bytecode.h>
#include <xbyak.h>
#include "common/bit_set.h"
#include <memory>
#include <unordered_map>
#include "common/common_types.h"
#include "common/x64/emitter.h"
#include "video_core/shader/shader.h"
using nihstro::Instruction;
using nihstro::OpCode;
using nihstro::SwizzlePattern;
namespace Pica {
namespace Shader {
/// Memory allocated for each compiled shader (64Kb)
constexpr size_t MAX_SHADER_SIZE = 1024 * 64;
class JitShader;
/**
* This class implements the shader JIT compiler. It recompiles a Pica shader program into x86_64
* code that can be executed on the host machine directly.
*/
class JitShader : public Xbyak::CodeGenerator {
class JitX64Engine final : public ShaderEngine {
public:
JitShader();
JitX64Engine();
~JitX64Engine() override;
void Run(const ShaderSetup& setup, UnitState& state, unsigned offset) const {
program(&setup, &state, instruction_labels[offset].getAddress());
}
void Compile();
void Compile_ADD(Instruction instr);
void Compile_DP3(Instruction instr);
void Compile_DP4(Instruction instr);
void Compile_DPH(Instruction instr);
void Compile_EX2(Instruction instr);
void Compile_LG2(Instruction instr);
void Compile_MUL(Instruction instr);
void Compile_SGE(Instruction instr);
void Compile_SLT(Instruction instr);
void Compile_FLR(Instruction instr);
void Compile_MAX(Instruction instr);
void Compile_MIN(Instruction instr);
void Compile_RCP(Instruction instr);
void Compile_RSQ(Instruction instr);
void Compile_MOVA(Instruction instr);
void Compile_MOV(Instruction instr);
void Compile_NOP(Instruction instr);
void Compile_END(Instruction instr);
void Compile_CALL(Instruction instr);
void Compile_CALLC(Instruction instr);
void Compile_CALLU(Instruction instr);
void Compile_IF(Instruction instr);
void Compile_LOOP(Instruction instr);
void Compile_JMP(Instruction instr);
void Compile_CMP(Instruction instr);
void Compile_MAD(Instruction instr);
void SetupBatch(ShaderSetup& setup, unsigned int entry_point) override;
void Run(const ShaderSetup& setup, UnitState& state) const override;
private:
void Compile_Block(unsigned end);
void Compile_NextInstr();
void Compile_SwizzleSrc(Instruction instr, unsigned src_num, SourceRegister src_reg,
Xbyak::Xmm dest);
void Compile_DestEnable(Instruction instr, Xbyak::Xmm dest);
/**
* Compiles a `MUL src1, src2` operation, properly handling the PICA semantics when multiplying
* zero by inf. Clobbers `src2` and `scratch`.
*/
void Compile_SanitizedMul(Xbyak::Xmm src1, Xbyak::Xmm src2, Xbyak::Xmm scratch);
void Compile_EvaluateCondition(Instruction instr);
void Compile_UniformCondition(Instruction instr);
/**
* Emits the code to conditionally return from a subroutine envoked by the `CALL` instruction.
*/
void Compile_Return();
BitSet32 PersistentCallerSavedRegs();
/**
* Assertion evaluated at compile-time, but only triggered if executed at runtime.
* @param msg Message to be logged if the assertion fails.
*/
void Compile_Assert(bool condition, const char* msg);
/**
* Analyzes the entire shader program for `CALL` instructions before emitting any code,
* identifying the locations where a return needs to be inserted.
*/
void FindReturnOffsets();
/// Mapping of Pica VS instructions to pointers in the emitted code
std::array<Xbyak::Label, 1024> instruction_labels;
/// Offsets in code where a return needs to be inserted
std::vector<unsigned> return_offsets;
unsigned program_counter = 0; ///< Offset of the next instruction to decode
bool looping = false; ///< True if compiling a loop, used to check for nested loops
using CompiledShader = void(const void* setup, void* state, const u8* start_addr);
CompiledShader* program = nullptr;
std::unordered_map<u64, std::unique_ptr<JitShader>> cache;
};
} // Shader
} // Pica
} // namespace Shader
} // namespace Pica

@ -0,0 +1,884 @@
// Copyright 2015 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <algorithm>
#include <cmath>
#include <cstdint>
#include <nihstro/shader_bytecode.h>
#include <smmintrin.h>
#include <xmmintrin.h>
#include "common/assert.h"
#include "common/logging/log.h"
#include "common/vector_math.h"
#include "common/x64/cpu_detect.h"
#include "common/x64/xbyak_abi.h"
#include "common/x64/xbyak_util.h"
#include "video_core/pica_state.h"
#include "video_core/pica_types.h"
#include "video_core/shader/shader.h"
#include "video_core/shader/shader_jit_x64_compiler.h"
using namespace Common::X64;
using namespace Xbyak::util;
using Xbyak::Label;
using Xbyak::Reg32;
using Xbyak::Reg64;
using Xbyak::Xmm;
namespace Pica {
namespace Shader {
typedef void (JitShader::*JitFunction)(Instruction instr);
const JitFunction instr_table[64] = {
&JitShader::Compile_ADD, // add
&JitShader::Compile_DP3, // dp3
&JitShader::Compile_DP4, // dp4
&JitShader::Compile_DPH, // dph
nullptr, // unknown
&JitShader::Compile_EX2, // ex2
&JitShader::Compile_LG2, // lg2
nullptr, // unknown
&JitShader::Compile_MUL, // mul
&JitShader::Compile_SGE, // sge
&JitShader::Compile_SLT, // slt
&JitShader::Compile_FLR, // flr
&JitShader::Compile_MAX, // max
&JitShader::Compile_MIN, // min
&JitShader::Compile_RCP, // rcp
&JitShader::Compile_RSQ, // rsq
nullptr, // unknown
nullptr, // unknown
&JitShader::Compile_MOVA, // mova
&JitShader::Compile_MOV, // mov
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
&JitShader::Compile_DPH, // dphi
nullptr, // unknown
&JitShader::Compile_SGE, // sgei
&JitShader::Compile_SLT, // slti
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
nullptr, // unknown
&JitShader::Compile_NOP, // nop
&JitShader::Compile_END, // end
nullptr, // break
&JitShader::Compile_CALL, // call
&JitShader::Compile_CALLC, // callc
&JitShader::Compile_CALLU, // callu
&JitShader::Compile_IF, // ifu
&JitShader::Compile_IF, // ifc
&JitShader::Compile_LOOP, // loop
nullptr, // emit
nullptr, // sete
&JitShader::Compile_JMP, // jmpc
&JitShader::Compile_JMP, // jmpu
&JitShader::Compile_CMP, // cmp
&JitShader::Compile_CMP, // cmp
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // madi
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
&JitShader::Compile_MAD, // mad
};
// The following is used to alias some commonly used registers. Generally, RAX-RDX and XMM0-XMM3 can
// be used as scratch registers within a compiler function. The other registers have designated
// purposes, as documented below:
/// Pointer to the uniform memory
static const Reg64 SETUP = r9;
/// The two 32-bit VS address offset registers set by the MOVA instruction
static const Reg64 ADDROFFS_REG_0 = r10;
static const Reg64 ADDROFFS_REG_1 = r11;
/// VS loop count register (Multiplied by 16)
static const Reg32 LOOPCOUNT_REG = r12d;
/// Current VS loop iteration number (we could probably use LOOPCOUNT_REG, but this quicker)
static const Reg32 LOOPCOUNT = esi;
/// Number to increment LOOPCOUNT_REG by on each loop iteration (Multiplied by 16)
static const Reg32 LOOPINC = edi;
/// Result of the previous CMP instruction for the X-component comparison
static const Reg64 COND0 = r13;
/// Result of the previous CMP instruction for the Y-component comparison
static const Reg64 COND1 = r14;
/// Pointer to the UnitState instance for the current VS unit
static const Reg64 STATE = r15;
/// SIMD scratch register
static const Xmm SCRATCH = xmm0;
/// Loaded with the first swizzled source register, otherwise can be used as a scratch register
static const Xmm SRC1 = xmm1;
/// Loaded with the second swizzled source register, otherwise can be used as a scratch register
static const Xmm SRC2 = xmm2;
/// Loaded with the third swizzled source register, otherwise can be used as a scratch register
static const Xmm SRC3 = xmm3;
/// Additional scratch register
static const Xmm SCRATCH2 = xmm4;
/// Constant vector of [1.0f, 1.0f, 1.0f, 1.0f], used to efficiently set a vector to one
static const Xmm ONE = xmm14;
/// Constant vector of [-0.f, -0.f, -0.f, -0.f], used to efficiently negate a vector with XOR
static const Xmm NEGBIT = xmm15;
// State registers that must not be modified by external functions calls
// Scratch registers, e.g., SRC1 and SCRATCH, have to be saved on the side if needed
static const BitSet32 persistent_regs = BuildRegSet({
// Pointers to register blocks
SETUP, STATE,
// Cached registers
ADDROFFS_REG_0, ADDROFFS_REG_1, LOOPCOUNT_REG, COND0, COND1,
// Constants
ONE, NEGBIT,
});
/// Raw constant for the source register selector that indicates no swizzling is performed
static const u8 NO_SRC_REG_SWIZZLE = 0x1b;
/// Raw constant for the destination register enable mask that indicates all components are enabled
static const u8 NO_DEST_REG_MASK = 0xf;
static void LogCritical(const char* msg) {
LOG_CRITICAL(HW_GPU, "%s", msg);
}
void JitShader::Compile_Assert(bool condition, const char* msg) {
if (!condition) {
mov(ABI_PARAM1, reinterpret_cast<size_t>(msg));
CallFarFunction(*this, LogCritical);
}
}
/**
* Loads and swizzles a source register into the specified XMM register.
* @param instr VS instruction, used for determining how to load the source register
* @param src_num Number indicating which source register to load (1 = src1, 2 = src2, 3 = src3)
* @param src_reg SourceRegister object corresponding to the source register to load
* @param dest Destination XMM register to store the loaded, swizzled source register
*/
void JitShader::Compile_SwizzleSrc(Instruction instr, unsigned src_num, SourceRegister src_reg,
Xmm dest) {
Reg64 src_ptr;
size_t src_offset;
if (src_reg.GetRegisterType() == RegisterType::FloatUniform) {
src_ptr = SETUP;
src_offset = ShaderSetup::GetFloatUniformOffset(src_reg.GetIndex());
} else {
src_ptr = STATE;
src_offset = UnitState::InputOffset(src_reg);
}
int src_offset_disp = (int)src_offset;
ASSERT_MSG(src_offset == src_offset_disp, "Source register offset too large for int type");
unsigned operand_desc_id;
const bool is_inverted =
(0 != (instr.opcode.Value().GetInfo().subtype & OpCode::Info::SrcInversed));
unsigned address_register_index;
unsigned offset_src;
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MAD ||
instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI) {
operand_desc_id = instr.mad.operand_desc_id;
offset_src = is_inverted ? 3 : 2;
address_register_index = instr.mad.address_register_index;
} else {
operand_desc_id = instr.common.operand_desc_id;
offset_src = is_inverted ? 2 : 1;
address_register_index = instr.common.address_register_index;
}
if (src_num == offset_src && address_register_index != 0) {
switch (address_register_index) {
case 1: // address offset 1
movaps(dest, xword[src_ptr + ADDROFFS_REG_0 + src_offset_disp]);
break;
case 2: // address offset 2
movaps(dest, xword[src_ptr + ADDROFFS_REG_1 + src_offset_disp]);
break;
case 3: // address offset 3
movaps(dest, xword[src_ptr + LOOPCOUNT_REG.cvt64() + src_offset_disp]);
break;
default:
UNREACHABLE();
break;
}
} else {
// Load the source
movaps(dest, xword[src_ptr + src_offset_disp]);
}
SwizzlePattern swiz = {(*swizzle_data)[operand_desc_id]};
// Generate instructions for source register swizzling as needed
u8 sel = swiz.GetRawSelector(src_num);
if (sel != NO_SRC_REG_SWIZZLE) {
// Selector component order needs to be reversed for the SHUFPS instruction
sel = ((sel & 0xc0) >> 6) | ((sel & 3) << 6) | ((sel & 0xc) << 2) | ((sel & 0x30) >> 2);
// Shuffle inputs for swizzle
shufps(dest, dest, sel);
}
// If the source register should be negated, flip the negative bit using XOR
const bool negate[] = {swiz.negate_src1, swiz.negate_src2, swiz.negate_src3};
if (negate[src_num - 1]) {
xorps(dest, NEGBIT);
}
}
void JitShader::Compile_DestEnable(Instruction instr, Xmm src) {
DestRegister dest;
unsigned operand_desc_id;
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MAD ||
instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI) {
operand_desc_id = instr.mad.operand_desc_id;
dest = instr.mad.dest.Value();
} else {
operand_desc_id = instr.common.operand_desc_id;
dest = instr.common.dest.Value();
}
SwizzlePattern swiz = {(*swizzle_data)[operand_desc_id]};
size_t dest_offset_disp = UnitState::OutputOffset(dest);
// If all components are enabled, write the result to the destination register
if (swiz.dest_mask == NO_DEST_REG_MASK) {
// Store dest back to memory
movaps(xword[STATE + dest_offset_disp], src);
} else {
// Not all components are enabled, so mask the result when storing to the destination
// register...
movaps(SCRATCH, xword[STATE + dest_offset_disp]);
if (Common::GetCPUCaps().sse4_1) {
u8 mask = ((swiz.dest_mask & 1) << 3) | ((swiz.dest_mask & 8) >> 3) |
((swiz.dest_mask & 2) << 1) | ((swiz.dest_mask & 4) >> 1);
blendps(SCRATCH, src, mask);
} else {
movaps(SCRATCH2, src);
unpckhps(SCRATCH2, SCRATCH); // Unpack X/Y components of source and destination
unpcklps(SCRATCH, src); // Unpack Z/W components of source and destination
// Compute selector to selectively copy source components to destination for SHUFPS
// instruction
u8 sel = ((swiz.DestComponentEnabled(0) ? 1 : 0) << 0) |
((swiz.DestComponentEnabled(1) ? 3 : 2) << 2) |
((swiz.DestComponentEnabled(2) ? 0 : 1) << 4) |
((swiz.DestComponentEnabled(3) ? 2 : 3) << 6);
shufps(SCRATCH, SCRATCH2, sel);
}
// Store dest back to memory
movaps(xword[STATE + dest_offset_disp], SCRATCH);
}
}
void JitShader::Compile_SanitizedMul(Xmm src1, Xmm src2, Xmm scratch) {
movaps(scratch, src1);
cmpordps(scratch, src2);
mulps(src1, src2);
movaps(src2, src1);
cmpunordps(src2, src2);
xorps(scratch, src2);
andps(src1, scratch);
}
void JitShader::Compile_EvaluateCondition(Instruction instr) {
// Note: NXOR is used below to check for equality
switch (instr.flow_control.op) {
case Instruction::FlowControlType::Or:
mov(eax, COND0);
mov(ebx, COND1);
xor(eax, (instr.flow_control.refx.Value() ^ 1));
xor(ebx, (instr.flow_control.refy.Value() ^ 1));
or (eax, ebx);
break;
case Instruction::FlowControlType::And:
mov(eax, COND0);
mov(ebx, COND1);
xor(eax, (instr.flow_control.refx.Value() ^ 1));
xor(ebx, (instr.flow_control.refy.Value() ^ 1));
and(eax, ebx);
break;
case Instruction::FlowControlType::JustX:
mov(eax, COND0);
xor(eax, (instr.flow_control.refx.Value() ^ 1));
break;
case Instruction::FlowControlType::JustY:
mov(eax, COND1);
xor(eax, (instr.flow_control.refy.Value() ^ 1));
break;
}
}
void JitShader::Compile_UniformCondition(Instruction instr) {
size_t offset = ShaderSetup::GetBoolUniformOffset(instr.flow_control.bool_uniform_id);
cmp(byte[SETUP + offset], 0);
}
BitSet32 JitShader::PersistentCallerSavedRegs() {
return persistent_regs & ABI_ALL_CALLER_SAVED;
}
void JitShader::Compile_ADD(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
addps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_DP3(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
movaps(SRC2, SRC1);
shufps(SRC2, SRC2, _MM_SHUFFLE(1, 1, 1, 1));
movaps(SRC3, SRC1);
shufps(SRC3, SRC3, _MM_SHUFFLE(2, 2, 2, 2));
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 0, 0, 0));
addps(SRC1, SRC2);
addps(SRC1, SRC3);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_DP4(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(2, 3, 0, 1)); // XYZW -> ZWXY
addps(SRC1, SRC2);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 1, 2, 3)); // XYZW -> WZYX
addps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_DPH(Instruction instr) {
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::DPHI) {
Compile_SwizzleSrc(instr, 1, instr.common.src1i, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2i, SRC2);
} else {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
}
if (Common::GetCPUCaps().sse4_1) {
// Set 4th component to 1.0
blendps(SRC1, ONE, 0b1000);
} else {
// Set 4th component to 1.0
movaps(SCRATCH, SRC1);
unpckhps(SCRATCH, ONE); // XYZW, 1111 -> Z1__
unpcklpd(SRC1, SCRATCH); // XYZW, Z1__ -> XYZ1
}
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(2, 3, 0, 1)); // XYZW -> ZWXY
addps(SRC1, SRC2);
movaps(SRC2, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 1, 2, 3)); // XYZW -> WZYX
addps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_EX2(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
movss(xmm0, SRC1); // ABI_PARAM1
ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
CallFarFunction(*this, exp2f);
ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
shufps(xmm0, xmm0, _MM_SHUFFLE(0, 0, 0, 0)); // ABI_RETURN
movaps(SRC1, xmm0);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_LG2(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
movss(xmm0, SRC1); // ABI_PARAM1
ABI_PushRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
CallFarFunction(*this, log2f);
ABI_PopRegistersAndAdjustStack(*this, PersistentCallerSavedRegs(), 0);
shufps(xmm0, xmm0, _MM_SHUFFLE(0, 0, 0, 0)); // ABI_RETURN
movaps(SRC1, xmm0);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MUL(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_SGE(Instruction instr) {
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::SGEI) {
Compile_SwizzleSrc(instr, 1, instr.common.src1i, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2i, SRC2);
} else {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
}
cmpleps(SRC2, SRC1);
andps(SRC2, ONE);
Compile_DestEnable(instr, SRC2);
}
void JitShader::Compile_SLT(Instruction instr) {
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::SLTI) {
Compile_SwizzleSrc(instr, 1, instr.common.src1i, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2i, SRC2);
} else {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
}
cmpltps(SRC1, SRC2);
andps(SRC1, ONE);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_FLR(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
if (Common::GetCPUCaps().sse4_1) {
roundps(SRC1, SRC1, _MM_FROUND_FLOOR);
} else {
cvttps2dq(SRC1, SRC1);
cvtdq2ps(SRC1, SRC1);
}
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MAX(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
// SSE semantics match PICA200 ones: In case of NaN, SRC2 is returned.
maxps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MIN(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
// SSE semantics match PICA200 ones: In case of NaN, SRC2 is returned.
minps(SRC1, SRC2);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_MOVA(Instruction instr) {
SwizzlePattern swiz = {(*swizzle_data)[instr.common.operand_desc_id]};
if (!swiz.DestComponentEnabled(0) && !swiz.DestComponentEnabled(1)) {
return; // NoOp
}
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
// Convert floats to integers using truncation (only care about X and Y components)
cvttps2dq(SRC1, SRC1);
// Get result
movq(rax, SRC1);
// Handle destination enable
if (swiz.DestComponentEnabled(0) && swiz.DestComponentEnabled(1)) {
// Move and sign-extend low 32 bits
movsxd(ADDROFFS_REG_0, eax);
// Move and sign-extend high 32 bits
shr(rax, 32);
movsxd(ADDROFFS_REG_1, eax);
// Multiply by 16 to be used as an offset later
shl(ADDROFFS_REG_0, 4);
shl(ADDROFFS_REG_1, 4);
} else {
if (swiz.DestComponentEnabled(0)) {
// Move and sign-extend low 32 bits
movsxd(ADDROFFS_REG_0, eax);
// Multiply by 16 to be used as an offset later
shl(ADDROFFS_REG_0, 4);
} else if (swiz.DestComponentEnabled(1)) {
// Move and sign-extend high 32 bits
shr(rax, 32);
movsxd(ADDROFFS_REG_1, eax);
// Multiply by 16 to be used as an offset later
shl(ADDROFFS_REG_1, 4);
}
}
}
void JitShader::Compile_MOV(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_RCP(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
// TODO(bunnei): RCPSS is a pretty rough approximation, this might cause problems if Pica
// performs this operation more accurately. This should be checked on hardware.
rcpss(SRC1, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 0, 0, 0)); // XYWZ -> XXXX
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_RSQ(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
// TODO(bunnei): RSQRTSS is a pretty rough approximation, this might cause problems if Pica
// performs this operation more accurately. This should be checked on hardware.
rsqrtss(SRC1, SRC1);
shufps(SRC1, SRC1, _MM_SHUFFLE(0, 0, 0, 0)); // XYWZ -> XXXX
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_NOP(Instruction instr) {}
void JitShader::Compile_END(Instruction instr) {
ABI_PopRegistersAndAdjustStack(*this, ABI_ALL_CALLEE_SAVED, 8);
ret();
}
void JitShader::Compile_CALL(Instruction instr) {
// Push offset of the return
push(qword, (instr.flow_control.dest_offset + instr.flow_control.num_instructions));
// Call the subroutine
call(instruction_labels[instr.flow_control.dest_offset]);
// Skip over the return offset that's on the stack
add(rsp, 8);
}
void JitShader::Compile_CALLC(Instruction instr) {
Compile_EvaluateCondition(instr);
Label b;
jz(b);
Compile_CALL(instr);
L(b);
}
void JitShader::Compile_CALLU(Instruction instr) {
Compile_UniformCondition(instr);
Label b;
jz(b);
Compile_CALL(instr);
L(b);
}
void JitShader::Compile_CMP(Instruction instr) {
using Op = Instruction::Common::CompareOpType::Op;
Op op_x = instr.common.compare_op.x;
Op op_y = instr.common.compare_op.y;
Compile_SwizzleSrc(instr, 1, instr.common.src1, SRC1);
Compile_SwizzleSrc(instr, 2, instr.common.src2, SRC2);
// SSE doesn't have greater-than (GT) or greater-equal (GE) comparison operators. You need to
// emulate them by swapping the lhs and rhs and using LT and LE. NLT and NLE can't be used here
// because they don't match when used with NaNs.
static const u8 cmp[] = {CMP_EQ, CMP_NEQ, CMP_LT, CMP_LE, CMP_LT, CMP_LE};
bool invert_op_x = (op_x == Op::GreaterThan || op_x == Op::GreaterEqual);
Xmm lhs_x = invert_op_x ? SRC2 : SRC1;
Xmm rhs_x = invert_op_x ? SRC1 : SRC2;
if (op_x == op_y) {
// Compare X-component and Y-component together
cmpps(lhs_x, rhs_x, cmp[op_x]);
movq(COND0, lhs_x);
mov(COND1, COND0);
} else {
bool invert_op_y = (op_y == Op::GreaterThan || op_y == Op::GreaterEqual);
Xmm lhs_y = invert_op_y ? SRC2 : SRC1;
Xmm rhs_y = invert_op_y ? SRC1 : SRC2;
// Compare X-component
movaps(SCRATCH, lhs_x);
cmpss(SCRATCH, rhs_x, cmp[op_x]);
// Compare Y-component
cmpps(lhs_y, rhs_y, cmp[op_y]);
movq(COND0, SCRATCH);
movq(COND1, lhs_y);
}
shr(COND0.cvt32(), 31); // ignores upper 32 bits in source
shr(COND1, 63);
}
void JitShader::Compile_MAD(Instruction instr) {
Compile_SwizzleSrc(instr, 1, instr.mad.src1, SRC1);
if (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI) {
Compile_SwizzleSrc(instr, 2, instr.mad.src2i, SRC2);
Compile_SwizzleSrc(instr, 3, instr.mad.src3i, SRC3);
} else {
Compile_SwizzleSrc(instr, 2, instr.mad.src2, SRC2);
Compile_SwizzleSrc(instr, 3, instr.mad.src3, SRC3);
}
Compile_SanitizedMul(SRC1, SRC2, SCRATCH);
addps(SRC1, SRC3);
Compile_DestEnable(instr, SRC1);
}
void JitShader::Compile_IF(Instruction instr) {
Compile_Assert(instr.flow_control.dest_offset >= program_counter,
"Backwards if-statements not supported");
Label l_else, l_endif;
// Evaluate the "IF" condition
if (instr.opcode.Value() == OpCode::Id::IFU) {
Compile_UniformCondition(instr);
} else if (instr.opcode.Value() == OpCode::Id::IFC) {
Compile_EvaluateCondition(instr);
}
jz(l_else, T_NEAR);
// Compile the code that corresponds to the condition evaluating as true
Compile_Block(instr.flow_control.dest_offset);
// If there isn't an "ELSE" condition, we are done here
if (instr.flow_control.num_instructions == 0) {
L(l_else);
return;
}
jmp(l_endif, T_NEAR);
L(l_else);
// This code corresponds to the "ELSE" condition
// Comple the code that corresponds to the condition evaluating as false
Compile_Block(instr.flow_control.dest_offset + instr.flow_control.num_instructions);
L(l_endif);
}
void JitShader::Compile_LOOP(Instruction instr) {
Compile_Assert(instr.flow_control.dest_offset >= program_counter,
"Backwards loops not supported");
Compile_Assert(!looping, "Nested loops not supported");
looping = true;
// This decodes the fields from the integer uniform at index instr.flow_control.int_uniform_id.
// The Y (LOOPCOUNT_REG) and Z (LOOPINC) component are kept multiplied by 16 (Left shifted by
// 4 bits) to be used as an offset into the 16-byte vector registers later
size_t offset = ShaderSetup::GetIntUniformOffset(instr.flow_control.int_uniform_id);
mov(LOOPCOUNT, dword[SETUP + offset]);
mov(LOOPCOUNT_REG, LOOPCOUNT);
shr(LOOPCOUNT_REG, 4);
and(LOOPCOUNT_REG, 0xFF0); // Y-component is the start
mov(LOOPINC, LOOPCOUNT);
shr(LOOPINC, 12);
and(LOOPINC, 0xFF0); // Z-component is the incrementer
movzx(LOOPCOUNT, LOOPCOUNT.cvt8()); // X-component is iteration count
add(LOOPCOUNT, 1); // Iteration count is X-component + 1
Label l_loop_start;
L(l_loop_start);
Compile_Block(instr.flow_control.dest_offset + 1);
add(LOOPCOUNT_REG, LOOPINC); // Increment LOOPCOUNT_REG by Z-component
sub(LOOPCOUNT, 1); // Increment loop count by 1
jnz(l_loop_start); // Loop if not equal
looping = false;
}
void JitShader::Compile_JMP(Instruction instr) {
if (instr.opcode.Value() == OpCode::Id::JMPC)
Compile_EvaluateCondition(instr);
else if (instr.opcode.Value() == OpCode::Id::JMPU)
Compile_UniformCondition(instr);
else
UNREACHABLE();
bool inverted_condition =
(instr.opcode.Value() == OpCode::Id::JMPU) && (instr.flow_control.num_instructions & 1);
Label& b = instruction_labels[instr.flow_control.dest_offset];
if (inverted_condition) {
jz(b, T_NEAR);
} else {
jnz(b, T_NEAR);
}
}
void JitShader::Compile_Block(unsigned end) {
while (program_counter < end) {
Compile_NextInstr();
}
}
void JitShader::Compile_Return() {
// Peek return offset on the stack and check if we're at that offset
mov(rax, qword[rsp + 8]);
cmp(eax, (program_counter));
// If so, jump back to before CALL
Label b;
jnz(b);
ret();
L(b);
}
void JitShader::Compile_NextInstr() {
if (std::binary_search(return_offsets.begin(), return_offsets.end(), program_counter)) {
Compile_Return();
}
L(instruction_labels[program_counter]);
Instruction instr = {(*program_code)[program_counter++]};
OpCode::Id opcode = instr.opcode.Value();
auto instr_func = instr_table[static_cast<unsigned>(opcode)];
if (instr_func) {
// JIT the instruction!
((*this).*instr_func)(instr);
} else {
// Unhandled instruction
LOG_CRITICAL(HW_GPU, "Unhandled instruction: 0x%02x (0x%08x)",
instr.opcode.Value().EffectiveOpCode(), instr.hex);
}
}
void JitShader::FindReturnOffsets() {
return_offsets.clear();
for (size_t offset = 0; offset < program_code->size(); ++offset) {
Instruction instr = {(*program_code)[offset]};
switch (instr.opcode.Value()) {
case OpCode::Id::CALL:
case OpCode::Id::CALLC:
case OpCode::Id::CALLU:
return_offsets.push_back(instr.flow_control.dest_offset +
instr.flow_control.num_instructions);
break;
default:
break;
}
}
// Sort for efficient binary search later
std::sort(return_offsets.begin(), return_offsets.end());
}
void JitShader::Compile(const std::array<u32, 1024>* program_code_,
const std::array<u32, 1024>* swizzle_data_) {
program_code = program_code_;
swizzle_data = swizzle_data_;
// Reset flow control state
program = (CompiledShader*)getCurr();
program_counter = 0;
looping = false;
instruction_labels.fill(Xbyak::Label());
// Find all `CALL` instructions and identify return locations
FindReturnOffsets();
// The stack pointer is 8 modulo 16 at the entry of a procedure
ABI_PushRegistersAndAdjustStack(*this, ABI_ALL_CALLEE_SAVED, 8);
mov(SETUP, ABI_PARAM1);
mov(STATE, ABI_PARAM2);
// Zero address/loop registers
xor(ADDROFFS_REG_0.cvt32(), ADDROFFS_REG_0.cvt32());
xor(ADDROFFS_REG_1.cvt32(), ADDROFFS_REG_1.cvt32());
xor(LOOPCOUNT_REG, LOOPCOUNT_REG);
// Used to set a register to one
static const __m128 one = {1.f, 1.f, 1.f, 1.f};
mov(rax, reinterpret_cast<size_t>(&one));
movaps(ONE, xword[rax]);
// Used to negate registers
static const __m128 neg = {-0.f, -0.f, -0.f, -0.f};
mov(rax, reinterpret_cast<size_t>(&neg));
movaps(NEGBIT, xword[rax]);
// Jump to start of the shader program
jmp(ABI_PARAM3);
// Compile entire program
Compile_Block(static_cast<unsigned>(program_code->size()));
// Free memory that's no longer needed
program_code = nullptr;
swizzle_data = nullptr;
return_offsets.clear();
return_offsets.shrink_to_fit();
ready();
ASSERT_MSG(getSize() <= MAX_SHADER_SIZE, "Compiled a shader that exceeds the allocated size!");
LOG_DEBUG(HW_GPU, "Compiled shader size=%lu", getSize());
}
JitShader::JitShader() : Xbyak::CodeGenerator(MAX_SHADER_SIZE) {}
} // namespace Shader
} // namespace Pica

@ -0,0 +1,125 @@
// Copyright 2015 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#pragma once
#include <array>
#include <cstddef>
#include <utility>
#include <vector>
#include <nihstro/shader_bytecode.h>
#include <xbyak.h>
#include "common/bit_set.h"
#include "common/common_types.h"
#include "common/x64/emitter.h"
#include "video_core/shader/shader.h"
using nihstro::Instruction;
using nihstro::OpCode;
using nihstro::SwizzlePattern;
namespace Pica {
namespace Shader {
/// Memory allocated for each compiled shader (64Kb)
constexpr size_t MAX_SHADER_SIZE = 1024 * 64;
/**
* This class implements the shader JIT compiler. It recompiles a Pica shader program into x86_64
* code that can be executed on the host machine directly.
*/
class JitShader : public Xbyak::CodeGenerator {
public:
JitShader();
void Run(const ShaderSetup& setup, UnitState& state, unsigned offset) const {
program(&setup, &state, instruction_labels[offset].getAddress());
}
void Compile(const std::array<u32, 1024>* program_code,
const std::array<u32, 1024>* swizzle_data);
void Compile_ADD(Instruction instr);
void Compile_DP3(Instruction instr);
void Compile_DP4(Instruction instr);
void Compile_DPH(Instruction instr);
void Compile_EX2(Instruction instr);
void Compile_LG2(Instruction instr);
void Compile_MUL(Instruction instr);
void Compile_SGE(Instruction instr);
void Compile_SLT(Instruction instr);
void Compile_FLR(Instruction instr);
void Compile_MAX(Instruction instr);
void Compile_MIN(Instruction instr);
void Compile_RCP(Instruction instr);
void Compile_RSQ(Instruction instr);
void Compile_MOVA(Instruction instr);
void Compile_MOV(Instruction instr);
void Compile_NOP(Instruction instr);
void Compile_END(Instruction instr);
void Compile_CALL(Instruction instr);
void Compile_CALLC(Instruction instr);
void Compile_CALLU(Instruction instr);
void Compile_IF(Instruction instr);
void Compile_LOOP(Instruction instr);
void Compile_JMP(Instruction instr);
void Compile_CMP(Instruction instr);
void Compile_MAD(Instruction instr);
private:
void Compile_Block(unsigned end);
void Compile_NextInstr();
void Compile_SwizzleSrc(Instruction instr, unsigned src_num, SourceRegister src_reg,
Xbyak::Xmm dest);
void Compile_DestEnable(Instruction instr, Xbyak::Xmm dest);
/**
* Compiles a `MUL src1, src2` operation, properly handling the PICA semantics when multiplying
* zero by inf. Clobbers `src2` and `scratch`.
*/
void Compile_SanitizedMul(Xbyak::Xmm src1, Xbyak::Xmm src2, Xbyak::Xmm scratch);
void Compile_EvaluateCondition(Instruction instr);
void Compile_UniformCondition(Instruction instr);
/**
* Emits the code to conditionally return from a subroutine envoked by the `CALL` instruction.
*/
void Compile_Return();
BitSet32 PersistentCallerSavedRegs();
/**
* Assertion evaluated at compile-time, but only triggered if executed at runtime.
* @param msg Message to be logged if the assertion fails.
*/
void Compile_Assert(bool condition, const char* msg);
/**
* Analyzes the entire shader program for `CALL` instructions before emitting any code,
* identifying the locations where a return needs to be inserted.
*/
void FindReturnOffsets();
const std::array<u32, 1024>* program_code = nullptr;
const std::array<u32, 1024>* swizzle_data = nullptr;
/// Mapping of Pica VS instructions to pointers in the emitted code
std::array<Xbyak::Label, 1024> instruction_labels;
/// Offsets in code where a return needs to be inserted
std::vector<unsigned> return_offsets;
unsigned program_counter = 0; ///< Offset of the next instruction to decode
bool looping = false; ///< True if compiling a loop, used to check for nested loops
using CompiledShader = void(const void* setup, void* state, const u8* start_addr);
CompiledShader* program = nullptr;
};
} // Shader
} // Pica