Lime3DS/src/video_core/shader/shader_interpreter.cpp
2017-01-25 18:53:25 -08:00

689 lines
28 KiB
C++

// Copyright 2014 Citra Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <algorithm>
#include <array>
#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"
#include "video_core/shader/shader.h"
#include "video_core/shader/shader_interpreter.h"
using nihstro::OpCode;
using nihstro::Instruction;
using nihstro::RegisterType;
using nihstro::SourceRegister;
using nihstro::SwizzlePattern;
namespace Pica {
namespace Shader {
struct CallStackElement {
u32 final_address; // Address upon which we jump to return_address
u32 return_address; // Where to jump when leaving scope
u8 repeat_counter; // How often to repeat until this call stack element is removed
u8 loop_increment; // Which value to add to the loop counter after an iteration
// TODO: Should this be a signed value? Does it even matter?
u32 loop_address; // The address where we'll return to after each loop iteration
};
template <bool Debug>
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
program_counter = offset - 1;
ASSERT(call_stack.size() < call_stack.capacity());
call_stack.push_back(
{offset + num_instructions, return_offset, repeat_count, loop_increment, offset});
};
auto evaluate_condition = [&state](Instruction::FlowControlType flow_control) {
using Op = Instruction::FlowControlType::Op;
bool result_x = flow_control.refx.Value() == state.conditional_code[0];
bool result_y = flow_control.refy.Value() == state.conditional_code[1];
switch (flow_control.op) {
case Op::Or:
return result_x || result_y;
case Op::And:
return result_x && result_y;
case Op::JustX:
return result_x;
case Op::JustY:
return result_y;
default:
UNREACHABLE();
return false;
}
};
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];
unsigned iteration = 0;
bool exit_loop = false;
while (!exit_loop) {
if (!call_stack.empty()) {
auto& top = call_stack.back();
if (program_counter == top.final_address) {
state.address_registers[2] += top.loop_increment;
if (top.repeat_counter-- == 0) {
program_counter = top.return_address;
call_stack.pop_back();
} else {
program_counter = top.loop_address;
}
// TODO: Is "trying again" accurate to hardware?
continue;
}
}
const Instruction instr = {program_code[program_counter]};
const SwizzlePattern swizzle = {swizzle_data[instr.common.operand_desc_id]};
Record<DebugDataRecord::CUR_INSTR>(debug_data, iteration, program_counter);
if (iteration > 0)
Record<DebugDataRecord::NEXT_INSTR>(debug_data, iteration - 1, program_counter);
debug_data.max_offset = std::max<u32>(debug_data.max_offset, 1 + program_counter);
auto LookupSourceRegister = [&](const SourceRegister& source_reg) -> const float24* {
switch (source_reg.GetRegisterType()) {
case RegisterType::Input:
return &state.registers.input[source_reg.GetIndex()].x;
case RegisterType::Temporary:
return &state.registers.temporary[source_reg.GetIndex()].x;
case RegisterType::FloatUniform:
return &uniforms.f[source_reg.GetIndex()].x;
default:
return dummy_vec4_float24;
}
};
switch (instr.opcode.Value().GetInfo().type) {
case OpCode::Type::Arithmetic: {
const bool is_inverted =
(0 != (instr.opcode.Value().GetInfo().subtype & OpCode::Info::SrcInversed));
const int address_offset =
(instr.common.address_register_index == 0)
? 0
: state.address_registers[instr.common.address_register_index - 1];
const float24* src1_ = LookupSourceRegister(instr.common.GetSrc1(is_inverted) +
(is_inverted ? 0 : address_offset));
const float24* src2_ = LookupSourceRegister(instr.common.GetSrc2(is_inverted) +
(is_inverted ? address_offset : 0));
const bool negate_src1 = ((bool)swizzle.negate_src1 != false);
const bool negate_src2 = ((bool)swizzle.negate_src2 != false);
float24 src1[4] = {
src1_[(int)swizzle.src1_selector_0.Value()],
src1_[(int)swizzle.src1_selector_1.Value()],
src1_[(int)swizzle.src1_selector_2.Value()],
src1_[(int)swizzle.src1_selector_3.Value()],
};
if (negate_src1) {
src1[0] = -src1[0];
src1[1] = -src1[1];
src1[2] = -src1[2];
src1[3] = -src1[3];
}
float24 src2[4] = {
src2_[(int)swizzle.src2_selector_0.Value()],
src2_[(int)swizzle.src2_selector_1.Value()],
src2_[(int)swizzle.src2_selector_2.Value()],
src2_[(int)swizzle.src2_selector_3.Value()],
};
if (negate_src2) {
src2[0] = -src2[0];
src2[1] = -src2[1];
src2[2] = -src2[2];
src2[3] = -src2[3];
}
float24* dest =
(instr.common.dest.Value() < 0x10)
? &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;
debug_data.max_opdesc_id =
std::max<u32>(debug_data.max_opdesc_id, 1 + instr.common.operand_desc_id);
switch (instr.opcode.Value().EffectiveOpCode()) {
case OpCode::Id::ADD: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = src1[i] + src2[i];
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
case OpCode::Id::MUL: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = src1[i] * src2[i];
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
case OpCode::Id::FLR:
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = float24::FromFloat32(std::floor(src1[i].ToFloat32()));
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
case OpCode::Id::MAX:
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
// NOTE: Exact form required to match NaN semantics to hardware:
// max(0, NaN) -> NaN
// max(NaN, 0) -> 0
dest[i] = (src1[i] > src2[i]) ? src1[i] : src2[i];
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
case OpCode::Id::MIN:
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
// NOTE: Exact form required to match NaN semantics to hardware:
// min(0, NaN) -> NaN
// min(NaN, 0) -> 0
dest[i] = (src1[i] < src2[i]) ? src1[i] : src2[i];
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
case OpCode::Id::DP3:
case OpCode::Id::DP4:
case OpCode::Id::DPH:
case OpCode::Id::DPHI: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
OpCode::Id opcode = instr.opcode.Value().EffectiveOpCode();
if (opcode == OpCode::Id::DPH || opcode == OpCode::Id::DPHI)
src1[3] = float24::FromFloat32(1.0f);
int num_components = (opcode == OpCode::Id::DP3) ? 3 : 4;
float24 dot = std::inner_product(src1, src1 + num_components, src2,
float24::FromFloat32(0.f));
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = dot;
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
// Reciprocal
case OpCode::Id::RCP: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
float24 rcp_res = float24::FromFloat32(1.0f / src1[0].ToFloat32());
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = rcp_res;
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
// Reciprocal Square Root
case OpCode::Id::RSQ: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
float24 rsq_res = float24::FromFloat32(1.0f / std::sqrt(src1[0].ToFloat32()));
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = rsq_res;
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
case OpCode::Id::MOVA: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
for (int i = 0; i < 2; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
// TODO: Figure out how the rounding is done on hardware
state.address_registers[i] = static_cast<s32>(src1[i].ToFloat32());
}
Record<DebugDataRecord::ADDR_REG_OUT>(debug_data, iteration,
state.address_registers);
break;
}
case OpCode::Id::MOV: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = src1[i];
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
case OpCode::Id::SGE:
case OpCode::Id::SGEI:
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = (src1[i] >= src2[i]) ? float24::FromFloat32(1.0f)
: float24::FromFloat32(0.0f);
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
case OpCode::Id::SLT:
case OpCode::Id::SLTI:
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = (src1[i] < src2[i]) ? float24::FromFloat32(1.0f)
: float24::FromFloat32(0.0f);
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
case OpCode::Id::CMP:
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
for (int i = 0; i < 2; ++i) {
// TODO: Can you restrict to one compare via dest masking?
auto compare_op = instr.common.compare_op;
auto op = (i == 0) ? compare_op.x.Value() : compare_op.y.Value();
switch (op) {
case Instruction::Common::CompareOpType::Equal:
state.conditional_code[i] = (src1[i] == src2[i]);
break;
case Instruction::Common::CompareOpType::NotEqual:
state.conditional_code[i] = (src1[i] != src2[i]);
break;
case Instruction::Common::CompareOpType::LessThan:
state.conditional_code[i] = (src1[i] < src2[i]);
break;
case Instruction::Common::CompareOpType::LessEqual:
state.conditional_code[i] = (src1[i] <= src2[i]);
break;
case Instruction::Common::CompareOpType::GreaterThan:
state.conditional_code[i] = (src1[i] > src2[i]);
break;
case Instruction::Common::CompareOpType::GreaterEqual:
state.conditional_code[i] = (src1[i] >= src2[i]);
break;
default:
LOG_ERROR(HW_GPU, "Unknown compare mode %x", static_cast<int>(op));
break;
}
}
Record<DebugDataRecord::CMP_RESULT>(debug_data, iteration, state.conditional_code);
break;
case OpCode::Id::EX2: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
// EX2 only takes first component exp2 and writes it to all dest components
float24 ex2_res = float24::FromFloat32(std::exp2(src1[0].ToFloat32()));
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = ex2_res;
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
case OpCode::Id::LG2: {
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
// LG2 only takes the first component log2 and writes it to all dest components
float24 lg2_res = float24::FromFloat32(std::log2(src1[0].ToFloat32()));
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = lg2_res;
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
break;
}
default:
LOG_ERROR(HW_GPU, "Unhandled arithmetic instruction: 0x%02x (%s): 0x%08x",
(int)instr.opcode.Value().EffectiveOpCode(),
instr.opcode.Value().GetInfo().name, instr.hex);
DEBUG_ASSERT(false);
break;
}
break;
}
case OpCode::Type::MultiplyAdd: {
if ((instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MAD) ||
(instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI)) {
const SwizzlePattern& swizzle = *reinterpret_cast<const SwizzlePattern*>(
&swizzle_data[instr.mad.operand_desc_id]);
bool is_inverted = (instr.opcode.Value().EffectiveOpCode() == OpCode::Id::MADI);
const int address_offset =
(instr.mad.address_register_index == 0)
? 0
: state.address_registers[instr.mad.address_register_index - 1];
const float24* src1_ = LookupSourceRegister(instr.mad.GetSrc1(is_inverted));
const float24* src2_ = LookupSourceRegister(instr.mad.GetSrc2(is_inverted) +
(!is_inverted * address_offset));
const float24* src3_ = LookupSourceRegister(instr.mad.GetSrc3(is_inverted) +
(is_inverted * address_offset));
const bool negate_src1 = ((bool)swizzle.negate_src1 != false);
const bool negate_src2 = ((bool)swizzle.negate_src2 != false);
const bool negate_src3 = ((bool)swizzle.negate_src3 != false);
float24 src1[4] = {
src1_[(int)swizzle.src1_selector_0.Value()],
src1_[(int)swizzle.src1_selector_1.Value()],
src1_[(int)swizzle.src1_selector_2.Value()],
src1_[(int)swizzle.src1_selector_3.Value()],
};
if (negate_src1) {
src1[0] = -src1[0];
src1[1] = -src1[1];
src1[2] = -src1[2];
src1[3] = -src1[3];
}
float24 src2[4] = {
src2_[(int)swizzle.src2_selector_0.Value()],
src2_[(int)swizzle.src2_selector_1.Value()],
src2_[(int)swizzle.src2_selector_2.Value()],
src2_[(int)swizzle.src2_selector_3.Value()],
};
if (negate_src2) {
src2[0] = -src2[0];
src2[1] = -src2[1];
src2[2] = -src2[2];
src2[3] = -src2[3];
}
float24 src3[4] = {
src3_[(int)swizzle.src3_selector_0.Value()],
src3_[(int)swizzle.src3_selector_1.Value()],
src3_[(int)swizzle.src3_selector_2.Value()],
src3_[(int)swizzle.src3_selector_3.Value()],
};
if (negate_src3) {
src3[0] = -src3[0];
src3[1] = -src3[1];
src3[2] = -src3[2];
src3[3] = -src3[3];
}
float24* dest =
(instr.mad.dest.Value() < 0x10)
? &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;
Record<DebugDataRecord::SRC1>(debug_data, iteration, src1);
Record<DebugDataRecord::SRC2>(debug_data, iteration, src2);
Record<DebugDataRecord::SRC3>(debug_data, iteration, src3);
Record<DebugDataRecord::DEST_IN>(debug_data, iteration, dest);
for (int i = 0; i < 4; ++i) {
if (!swizzle.DestComponentEnabled(i))
continue;
dest[i] = src1[i] * src2[i] + src3[i];
}
Record<DebugDataRecord::DEST_OUT>(debug_data, iteration, dest);
} else {
LOG_ERROR(HW_GPU, "Unhandled multiply-add instruction: 0x%02x (%s): 0x%08x",
(int)instr.opcode.Value().EffectiveOpCode(),
instr.opcode.Value().GetInfo().name, instr.hex);
}
break;
}
default: {
// Handle each instruction on its own
switch (instr.opcode.Value()) {
case OpCode::Id::END:
exit_loop = true;
break;
case OpCode::Id::JMPC:
Record<DebugDataRecord::COND_CMP_IN>(debug_data, iteration, state.conditional_code);
if (evaluate_condition(instr.flow_control)) {
program_counter = instr.flow_control.dest_offset - 1;
}
break;
case OpCode::Id::JMPU:
Record<DebugDataRecord::COND_BOOL_IN>(
debug_data, iteration, uniforms.b[instr.flow_control.bool_uniform_id]);
if (uniforms.b[instr.flow_control.bool_uniform_id] ==
!(instr.flow_control.num_instructions & 1)) {
program_counter = instr.flow_control.dest_offset - 1;
}
break;
case OpCode::Id::CALL:
call(instr.flow_control.dest_offset, instr.flow_control.num_instructions,
program_counter + 1, 0, 0);
break;
case OpCode::Id::CALLU:
Record<DebugDataRecord::COND_BOOL_IN>(
debug_data, iteration, uniforms.b[instr.flow_control.bool_uniform_id]);
if (uniforms.b[instr.flow_control.bool_uniform_id]) {
call(instr.flow_control.dest_offset, instr.flow_control.num_instructions,
program_counter + 1, 0, 0);
}
break;
case OpCode::Id::CALLC:
Record<DebugDataRecord::COND_CMP_IN>(debug_data, iteration, state.conditional_code);
if (evaluate_condition(instr.flow_control)) {
call(instr.flow_control.dest_offset, instr.flow_control.num_instructions,
program_counter + 1, 0, 0);
}
break;
case OpCode::Id::NOP:
break;
case OpCode::Id::IFU:
Record<DebugDataRecord::COND_BOOL_IN>(
debug_data, iteration, uniforms.b[instr.flow_control.bool_uniform_id]);
if (uniforms.b[instr.flow_control.bool_uniform_id]) {
call(program_counter + 1, instr.flow_control.dest_offset - program_counter - 1,
instr.flow_control.dest_offset + instr.flow_control.num_instructions, 0,
0);
} else {
call(instr.flow_control.dest_offset, instr.flow_control.num_instructions,
instr.flow_control.dest_offset + instr.flow_control.num_instructions, 0,
0);
}
break;
case OpCode::Id::IFC: {
// TODO: Do we need to consider swizzlers here?
Record<DebugDataRecord::COND_CMP_IN>(debug_data, iteration, state.conditional_code);
if (evaluate_condition(instr.flow_control)) {
call(program_counter + 1, instr.flow_control.dest_offset - program_counter - 1,
instr.flow_control.dest_offset + instr.flow_control.num_instructions, 0,
0);
} else {
call(instr.flow_control.dest_offset, instr.flow_control.num_instructions,
instr.flow_control.dest_offset + instr.flow_control.num_instructions, 0,
0);
}
break;
}
case OpCode::Id::LOOP: {
Math::Vec4<u8> loop_param(uniforms.i[instr.flow_control.int_uniform_id].x,
uniforms.i[instr.flow_control.int_uniform_id].y,
uniforms.i[instr.flow_control.int_uniform_id].z,
uniforms.i[instr.flow_control.int_uniform_id].w);
state.address_registers[2] = loop_param.y;
Record<DebugDataRecord::LOOP_INT_IN>(debug_data, iteration, loop_param);
call(program_counter + 1, instr.flow_control.dest_offset - program_counter + 1,
instr.flow_control.dest_offset + 1, loop_param.x, loop_param.z);
break;
}
default:
LOG_ERROR(HW_GPU, "Unhandled instruction: 0x%02x (%s): 0x%08x",
(int)instr.opcode.Value().EffectiveOpCode(),
instr.opcode.Value().GetInfo().name, instr.hex);
break;
}
break;
}
}
++program_counter;
++iteration;
}
}
void InterpreterEngine::SetupBatch(const ShaderSetup* setup_) {
setup = setup_;
}
MICROPROFILE_DECLARE(GPU_Shader);
void InterpreterEngine::Run(UnitState& state, unsigned int entry_point) const {
ASSERT(setup != nullptr);
ASSERT(entry_point < 1024);
MICROPROFILE_SCOPE(GPU_Shader);
DebugData<false> dummy_debug_data;
RunInterpreter(*setup, state, dummy_debug_data, entry_point);
}
DebugData<true> InterpreterEngine::ProduceDebugInfo(const InputVertex& input, int num_attributes,
unsigned int entry_point) const {
ASSERT(setup != nullptr);
ASSERT(entry_point < 1024);
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, entry_point);
return debug_data;
}
} // namespace
} // namespace