yuzu/src/video_core/shader/control_flow.cpp
2019-10-25 09:01:29 -04:00

692 lines
23 KiB
C++

// Copyright 2019 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include <list>
#include <map>
#include <set>
#include <stack>
#include <unordered_map>
#include <vector>
#include "common/assert.h"
#include "common/common_types.h"
#include "video_core/shader/ast.h"
#include "video_core/shader/control_flow.h"
#include "video_core/shader/shader_ir.h"
namespace VideoCommon::Shader {
namespace {
using Tegra::Shader::Instruction;
using Tegra::Shader::OpCode;
constexpr s32 unassigned_branch = -2;
struct Query {
u32 address{};
std::stack<u32> ssy_stack{};
std::stack<u32> pbk_stack{};
};
struct BlockStack {
BlockStack() = default;
explicit BlockStack(const Query& q) : ssy_stack{q.ssy_stack}, pbk_stack{q.pbk_stack} {}
std::stack<u32> ssy_stack{};
std::stack<u32> pbk_stack{};
};
struct BlockBranchInfo {
Condition condition{};
s32 address{exit_branch};
bool kill{};
bool is_sync{};
bool is_brk{};
bool ignore{};
};
struct BlockInfo {
u32 start{};
u32 end{};
bool visited{};
BlockBranchInfo branch{};
bool IsInside(const u32 address) const {
return start <= address && address <= end;
}
};
struct CFGRebuildState {
explicit CFGRebuildState(const ProgramCode& program_code, const std::size_t program_size,
const u32 start)
: start{start}, program_code{program_code}, program_size{program_size} {}
u32 start{};
std::vector<BlockInfo> block_info{};
std::list<u32> inspect_queries{};
std::list<Query> queries{};
std::unordered_map<u32, u32> registered{};
std::set<u32> labels{};
std::map<u32, u32> ssy_labels{};
std::map<u32, u32> pbk_labels{};
std::unordered_map<u32, BlockStack> stacks{};
const ProgramCode& program_code;
const std::size_t program_size;
ASTManager* manager;
};
enum class BlockCollision : u32 { None, Found, Inside };
std::pair<BlockCollision, u32> TryGetBlock(CFGRebuildState& state, u32 address) {
const auto& blocks = state.block_info;
for (u32 index = 0; index < blocks.size(); index++) {
if (blocks[index].start == address) {
return {BlockCollision::Found, index};
}
if (blocks[index].IsInside(address)) {
return {BlockCollision::Inside, index};
}
}
return {BlockCollision::None, 0xFFFFFFFF};
}
struct ParseInfo {
BlockBranchInfo branch_info{};
u32 end_address{};
};
BlockInfo& CreateBlockInfo(CFGRebuildState& state, u32 start, u32 end) {
auto& it = state.block_info.emplace_back();
it.start = start;
it.end = end;
const u32 index = static_cast<u32>(state.block_info.size() - 1);
state.registered.insert({start, index});
return it;
}
Pred GetPredicate(u32 index, bool negated) {
return static_cast<Pred>(index + (negated ? 8 : 0));
}
/**
* Returns whether the instruction at the specified offset is a 'sched' instruction.
* Sched instructions always appear before a sequence of 3 instructions.
*/
constexpr bool IsSchedInstruction(u32 offset, u32 main_offset) {
constexpr u32 SchedPeriod = 4;
u32 absolute_offset = offset - main_offset;
return (absolute_offset % SchedPeriod) == 0;
}
enum class ParseResult : u32 {
ControlCaught,
BlockEnd,
AbnormalFlow,
};
struct BranchIndirectInfo {
u32 buffer{};
u32 offset{};
u32 entries{};
s32 relative_position{};
};
std::optional<BranchIndirectInfo> TrackBranchIndirectInfo(const CFGRebuildState& state,
u32 start_address, u32 current_position) {
const u32 shader_start = state.start;
u32 pos = current_position;
BranchIndirectInfo result{};
u64 track_register = 0;
// Step 0 Get BRX Info
const Instruction instr = {state.program_code[pos]};
const auto opcode = OpCode::Decode(instr);
if (opcode->get().GetId() != OpCode::Id::BRX) {
return {};
}
if (instr.brx.constant_buffer != 0) {
return {};
}
track_register = instr.gpr8.Value();
result.relative_position = instr.brx.GetBranchExtend();
pos--;
bool found_track = false;
// Step 1 Track LDC
while (pos >= shader_start) {
if (IsSchedInstruction(pos, shader_start)) {
pos--;
continue;
}
const Instruction instr = {state.program_code[pos]};
const auto opcode = OpCode::Decode(instr);
if (opcode->get().GetId() == OpCode::Id::LD_C) {
if (instr.gpr0.Value() == track_register &&
instr.ld_c.type.Value() == Tegra::Shader::UniformType::Single) {
result.buffer = instr.cbuf36.index;
result.offset = instr.cbuf36.GetOffset();
track_register = instr.gpr8.Value();
pos--;
found_track = true;
break;
}
}
pos--;
}
if (!found_track) {
return {};
}
found_track = false;
// Step 2 Track SHL
while (pos >= shader_start) {
if (IsSchedInstruction(pos, shader_start)) {
pos--;
continue;
}
const Instruction instr = {state.program_code[pos]};
const auto opcode = OpCode::Decode(instr);
if (opcode->get().GetId() == OpCode::Id::SHL_IMM) {
if (instr.gpr0.Value() == track_register) {
track_register = instr.gpr8.Value();
pos--;
found_track = true;
break;
}
}
pos--;
}
if (!found_track) {
return {};
}
found_track = false;
// Step 3 Track IMNMX
while (pos >= shader_start) {
if (IsSchedInstruction(pos, shader_start)) {
pos--;
continue;
}
const Instruction instr = {state.program_code[pos]};
const auto opcode = OpCode::Decode(instr);
if (opcode->get().GetId() == OpCode::Id::IMNMX_IMM) {
if (instr.gpr0.Value() == track_register) {
track_register = instr.gpr8.Value();
result.entries = instr.alu.GetSignedImm20_20();
pos--;
found_track = true;
break;
}
}
pos--;
}
if (!found_track) {
return {};
}
return {result};
}
std::pair<ParseResult, ParseInfo> ParseCode(CFGRebuildState& state, u32 address) {
u32 offset = static_cast<u32>(address);
const u32 end_address = static_cast<u32>(state.program_size / sizeof(Instruction));
ParseInfo parse_info{};
const auto insert_label = [](CFGRebuildState& state, u32 address) {
const auto pair = state.labels.emplace(address);
if (pair.second) {
state.inspect_queries.push_back(address);
}
};
while (true) {
if (offset >= end_address) {
// ASSERT_OR_EXECUTE can't be used, as it ignores the break
ASSERT_MSG(false, "Shader passed the current limit!");
parse_info.branch_info.address = exit_branch;
parse_info.branch_info.ignore = false;
break;
}
if (state.registered.count(offset) != 0) {
parse_info.branch_info.address = offset;
parse_info.branch_info.ignore = true;
break;
}
if (IsSchedInstruction(offset, state.start)) {
offset++;
continue;
}
const Instruction instr = {state.program_code[offset]};
const auto opcode = OpCode::Decode(instr);
if (!opcode || opcode->get().GetType() != OpCode::Type::Flow) {
offset++;
continue;
}
switch (opcode->get().GetId()) {
case OpCode::Id::EXIT: {
const auto pred_index = static_cast<u32>(instr.pred.pred_index);
parse_info.branch_info.condition.predicate =
GetPredicate(pred_index, instr.negate_pred != 0);
if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
offset++;
continue;
}
const ConditionCode cc = instr.flow_condition_code;
parse_info.branch_info.condition.cc = cc;
if (cc == ConditionCode::F) {
offset++;
continue;
}
parse_info.branch_info.address = exit_branch;
parse_info.branch_info.kill = false;
parse_info.branch_info.is_sync = false;
parse_info.branch_info.is_brk = false;
parse_info.branch_info.ignore = false;
parse_info.end_address = offset;
return {ParseResult::ControlCaught, parse_info};
}
case OpCode::Id::BRA: {
if (instr.bra.constant_buffer != 0) {
return {ParseResult::AbnormalFlow, parse_info};
}
const auto pred_index = static_cast<u32>(instr.pred.pred_index);
parse_info.branch_info.condition.predicate =
GetPredicate(pred_index, instr.negate_pred != 0);
if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
offset++;
continue;
}
const ConditionCode cc = instr.flow_condition_code;
parse_info.branch_info.condition.cc = cc;
if (cc == ConditionCode::F) {
offset++;
continue;
}
const u32 branch_offset = offset + instr.bra.GetBranchTarget();
if (branch_offset == 0) {
parse_info.branch_info.address = exit_branch;
} else {
parse_info.branch_info.address = branch_offset;
}
insert_label(state, branch_offset);
parse_info.branch_info.kill = false;
parse_info.branch_info.is_sync = false;
parse_info.branch_info.is_brk = false;
parse_info.branch_info.ignore = false;
parse_info.end_address = offset;
return {ParseResult::ControlCaught, parse_info};
}
case OpCode::Id::SYNC: {
const auto pred_index = static_cast<u32>(instr.pred.pred_index);
parse_info.branch_info.condition.predicate =
GetPredicate(pred_index, instr.negate_pred != 0);
if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
offset++;
continue;
}
const ConditionCode cc = instr.flow_condition_code;
parse_info.branch_info.condition.cc = cc;
if (cc == ConditionCode::F) {
offset++;
continue;
}
parse_info.branch_info.address = unassigned_branch;
parse_info.branch_info.kill = false;
parse_info.branch_info.is_sync = true;
parse_info.branch_info.is_brk = false;
parse_info.branch_info.ignore = false;
parse_info.end_address = offset;
return {ParseResult::ControlCaught, parse_info};
}
case OpCode::Id::BRK: {
const auto pred_index = static_cast<u32>(instr.pred.pred_index);
parse_info.branch_info.condition.predicate =
GetPredicate(pred_index, instr.negate_pred != 0);
if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
offset++;
continue;
}
const ConditionCode cc = instr.flow_condition_code;
parse_info.branch_info.condition.cc = cc;
if (cc == ConditionCode::F) {
offset++;
continue;
}
parse_info.branch_info.address = unassigned_branch;
parse_info.branch_info.kill = false;
parse_info.branch_info.is_sync = false;
parse_info.branch_info.is_brk = true;
parse_info.branch_info.ignore = false;
parse_info.end_address = offset;
return {ParseResult::ControlCaught, parse_info};
}
case OpCode::Id::KIL: {
const auto pred_index = static_cast<u32>(instr.pred.pred_index);
parse_info.branch_info.condition.predicate =
GetPredicate(pred_index, instr.negate_pred != 0);
if (parse_info.branch_info.condition.predicate == Pred::NeverExecute) {
offset++;
continue;
}
const ConditionCode cc = instr.flow_condition_code;
parse_info.branch_info.condition.cc = cc;
if (cc == ConditionCode::F) {
offset++;
continue;
}
parse_info.branch_info.address = exit_branch;
parse_info.branch_info.kill = true;
parse_info.branch_info.is_sync = false;
parse_info.branch_info.is_brk = false;
parse_info.branch_info.ignore = false;
parse_info.end_address = offset;
return {ParseResult::ControlCaught, parse_info};
}
case OpCode::Id::SSY: {
const u32 target = offset + instr.bra.GetBranchTarget();
insert_label(state, target);
state.ssy_labels.emplace(offset, target);
break;
}
case OpCode::Id::PBK: {
const u32 target = offset + instr.bra.GetBranchTarget();
insert_label(state, target);
state.pbk_labels.emplace(offset, target);
break;
}
case OpCode::Id::BRX: {
auto tmp = TrackBranchIndirectInfo(state, address, offset);
if (tmp) {
auto result = *tmp;
LOG_CRITICAL(HW_GPU, "Track Successful, BRX: buffer:{}, offset:{}, entries:{}",
result.buffer, result.offset, result.entries);
} else {
LOG_CRITICAL(HW_GPU, "Track Unsuccesful");
}
return {ParseResult::AbnormalFlow, parse_info};
}
default:
break;
}
offset++;
}
parse_info.branch_info.kill = false;
parse_info.branch_info.is_sync = false;
parse_info.branch_info.is_brk = false;
parse_info.end_address = offset - 1;
return {ParseResult::BlockEnd, parse_info};
}
bool TryInspectAddress(CFGRebuildState& state) {
if (state.inspect_queries.empty()) {
return false;
}
const u32 address = state.inspect_queries.front();
state.inspect_queries.pop_front();
const auto [result, block_index] = TryGetBlock(state, address);
switch (result) {
case BlockCollision::Found: {
return true;
}
case BlockCollision::Inside: {
// This case is the tricky one:
// We need to Split the block in 2 sepparate blocks
const u32 end = state.block_info[block_index].end;
BlockInfo& new_block = CreateBlockInfo(state, address, end);
BlockInfo& current_block = state.block_info[block_index];
current_block.end = address - 1;
new_block.branch = current_block.branch;
BlockBranchInfo forward_branch{};
forward_branch.address = address;
forward_branch.ignore = true;
current_block.branch = forward_branch;
return true;
}
default:
break;
}
const auto [parse_result, parse_info] = ParseCode(state, address);
if (parse_result == ParseResult::AbnormalFlow) {
// if it's AbnormalFlow, we end it as false, ending the CFG reconstruction
return false;
}
BlockInfo& block_info = CreateBlockInfo(state, address, parse_info.end_address);
block_info.branch = parse_info.branch_info;
if (parse_info.branch_info.condition.IsUnconditional()) {
return true;
}
const u32 fallthrough_address = parse_info.end_address + 1;
state.inspect_queries.push_front(fallthrough_address);
return true;
}
bool TryQuery(CFGRebuildState& state) {
const auto gather_labels = [](std::stack<u32>& cc, std::map<u32, u32>& labels,
BlockInfo& block) {
auto gather_start = labels.lower_bound(block.start);
const auto gather_end = labels.upper_bound(block.end);
while (gather_start != gather_end) {
cc.push(gather_start->second);
++gather_start;
}
};
if (state.queries.empty()) {
return false;
}
Query& q = state.queries.front();
const u32 block_index = state.registered[q.address];
BlockInfo& block = state.block_info[block_index];
// If the block is visited, check if the stacks match, else gather the ssy/pbk
// labels into the current stack and look if the branch at the end of the block
// consumes a label. Schedule new queries accordingly
if (block.visited) {
BlockStack& stack = state.stacks[q.address];
const bool all_okay = (stack.ssy_stack.empty() || q.ssy_stack == stack.ssy_stack) &&
(stack.pbk_stack.empty() || q.pbk_stack == stack.pbk_stack);
state.queries.pop_front();
return all_okay;
}
block.visited = true;
state.stacks.insert_or_assign(q.address, BlockStack{q});
Query q2(q);
state.queries.pop_front();
gather_labels(q2.ssy_stack, state.ssy_labels, block);
gather_labels(q2.pbk_stack, state.pbk_labels, block);
if (!block.branch.condition.IsUnconditional()) {
q2.address = block.end + 1;
state.queries.push_back(q2);
}
Query conditional_query{q2};
if (block.branch.is_sync) {
if (block.branch.address == unassigned_branch) {
block.branch.address = conditional_query.ssy_stack.top();
}
conditional_query.ssy_stack.pop();
}
if (block.branch.is_brk) {
if (block.branch.address == unassigned_branch) {
block.branch.address = conditional_query.pbk_stack.top();
}
conditional_query.pbk_stack.pop();
}
conditional_query.address = block.branch.address;
state.queries.push_back(std::move(conditional_query));
return true;
}
} // Anonymous namespace
void InsertBranch(ASTManager& mm, const BlockBranchInfo& branch) {
const auto get_expr = ([&](const Condition& cond) -> Expr {
Expr result{};
if (cond.cc != ConditionCode::T) {
result = MakeExpr<ExprCondCode>(cond.cc);
}
if (cond.predicate != Pred::UnusedIndex) {
u32 pred = static_cast<u32>(cond.predicate);
bool negate = false;
if (pred > 7) {
negate = true;
pred -= 8;
}
Expr extra = MakeExpr<ExprPredicate>(pred);
if (negate) {
extra = MakeExpr<ExprNot>(extra);
}
if (result) {
return MakeExpr<ExprAnd>(extra, result);
}
return extra;
}
if (result) {
return result;
}
return MakeExpr<ExprBoolean>(true);
});
if (branch.address < 0) {
if (branch.kill) {
mm.InsertReturn(get_expr(branch.condition), true);
return;
}
mm.InsertReturn(get_expr(branch.condition), false);
return;
}
mm.InsertGoto(get_expr(branch.condition), branch.address);
}
void DecompileShader(CFGRebuildState& state) {
state.manager->Init();
for (auto label : state.labels) {
state.manager->DeclareLabel(label);
}
for (auto& block : state.block_info) {
if (state.labels.count(block.start) != 0) {
state.manager->InsertLabel(block.start);
}
u32 end = block.branch.ignore ? block.end + 1 : block.end;
state.manager->InsertBlock(block.start, end);
if (!block.branch.ignore) {
InsertBranch(*state.manager, block.branch);
}
}
state.manager->Decompile();
}
std::unique_ptr<ShaderCharacteristics> ScanFlow(const ProgramCode& program_code,
std::size_t program_size, u32 start_address,
const CompilerSettings& settings) {
auto result_out = std::make_unique<ShaderCharacteristics>();
if (settings.depth == CompileDepth::BruteForce) {
result_out->settings.depth = CompileDepth::BruteForce;
return result_out;
}
CFGRebuildState state{program_code, program_size, start_address};
// Inspect Code and generate blocks
state.labels.clear();
state.labels.emplace(start_address);
state.inspect_queries.push_back(state.start);
while (!state.inspect_queries.empty()) {
if (!TryInspectAddress(state)) {
result_out->settings.depth = CompileDepth::BruteForce;
return result_out;
}
}
bool use_flow_stack = true;
bool decompiled = false;
if (settings.depth != CompileDepth::FlowStack) {
// Decompile Stacks
state.queries.push_back(Query{state.start, {}, {}});
decompiled = true;
while (!state.queries.empty()) {
if (!TryQuery(state)) {
decompiled = false;
break;
}
}
}
use_flow_stack = !decompiled;
// Sort and organize results
std::sort(state.block_info.begin(), state.block_info.end(),
[](const BlockInfo& a, const BlockInfo& b) -> bool { return a.start < b.start; });
if (decompiled && settings.depth != CompileDepth::NoFlowStack) {
ASTManager manager{settings.depth != CompileDepth::DecompileBackwards,
settings.disable_else_derivation};
state.manager = &manager;
DecompileShader(state);
decompiled = state.manager->IsFullyDecompiled();
if (!decompiled) {
if (settings.depth == CompileDepth::FullDecompile) {
LOG_CRITICAL(HW_GPU, "Failed to remove all the gotos!:");
} else {
LOG_CRITICAL(HW_GPU, "Failed to remove all backward gotos!:");
}
state.manager->ShowCurrentState("Of Shader");
state.manager->Clear();
} else {
auto characteristics = std::make_unique<ShaderCharacteristics>();
characteristics->start = start_address;
characteristics->settings.depth = settings.depth;
characteristics->manager = std::move(manager);
characteristics->end = state.block_info.back().end + 1;
return characteristics;
}
}
result_out->start = start_address;
result_out->settings.depth =
use_flow_stack ? CompileDepth::FlowStack : CompileDepth::NoFlowStack;
result_out->blocks.clear();
for (auto& block : state.block_info) {
ShaderBlock new_block{};
new_block.start = block.start;
new_block.end = block.end;
new_block.ignore_branch = block.branch.ignore;
if (!new_block.ignore_branch) {
new_block.branch.cond = block.branch.condition;
new_block.branch.kills = block.branch.kill;
new_block.branch.address = block.branch.address;
}
result_out->end = std::max(result_out->end, block.end);
result_out->blocks.push_back(new_block);
}
if (!use_flow_stack) {
result_out->labels = std::move(state.labels);
return result_out;
}
auto back = result_out->blocks.begin();
auto next = std::next(back);
while (next != result_out->blocks.end()) {
if (state.labels.count(next->start) == 0 && next->start == back->end + 1) {
back->end = next->end;
next = result_out->blocks.erase(next);
continue;
}
back = next;
++next;
}
return result_out;
}
} // namespace VideoCommon::Shader