| 1 | //===- bolt/Passes/StokeInfo.cpp ------------------------------------------===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | // |
| 9 | // This file implements the StokeInfo class. |
| 10 | // |
| 11 | //===----------------------------------------------------------------------===// |
| 12 | |
| 13 | #include "bolt/Passes/StokeInfo.h" |
| 14 | #include "bolt/Core/BinaryFunctionCallGraph.h" |
| 15 | #include "bolt/Passes/DataflowInfoManager.h" |
| 16 | #include "llvm/Support/CommandLine.h" |
| 17 | |
| 18 | #define DEBUG_TYPE "stoke" |
| 19 | |
| 20 | using namespace llvm; |
| 21 | using namespace bolt; |
| 22 | |
| 23 | namespace opts { |
| 24 | static cl::OptionCategory StokeOptCategory("STOKE pass options" ); |
| 25 | |
| 26 | static cl::opt<std::string> |
| 27 | StokeOutputDataFilename("stoke-out" , |
| 28 | cl::desc("output data (.csv) for Stoke's use" ), |
| 29 | cl::Optional, |
| 30 | cl::cat(StokeOptCategory)); |
| 31 | } |
| 32 | |
| 33 | namespace llvm { |
| 34 | namespace bolt { |
| 35 | |
| 36 | void getRegNameFromBitVec(const BinaryContext &BC, const BitVector &RegV, |
| 37 | std::set<std::string> *NameVec = nullptr) { |
| 38 | for (int RegIdx : RegV.set_bits()) { |
| 39 | LLVM_DEBUG(dbgs() << BC.MRI->getName(RegIdx) << " " ); |
| 40 | if (NameVec) |
| 41 | NameVec->insert(x: std::string(BC.MRI->getName(RegNo: RegIdx))); |
| 42 | } |
| 43 | LLVM_DEBUG(dbgs() << "\n" ); |
| 44 | } |
| 45 | |
| 46 | void StokeInfo::checkInstr(const BinaryFunction &BF, StokeFuncInfo &FuncInfo) { |
| 47 | MCPlusBuilder *MIB = BF.getBinaryContext().MIB.get(); |
| 48 | BitVector RegV(NumRegs, false); |
| 49 | for (const BinaryBasicBlock *BB : BF.getLayout().blocks()) { |
| 50 | if (BB->empty()) |
| 51 | continue; |
| 52 | |
| 53 | // Skip function with exception handling. |
| 54 | if (BB->throw_size() || BB->lp_size()) { |
| 55 | FuncInfo.Omitted = true; |
| 56 | return; |
| 57 | } |
| 58 | |
| 59 | for (const MCInst &It : *BB) { |
| 60 | if (MIB->isPseudo(Inst: It)) |
| 61 | continue; |
| 62 | // skip function with exception handling yet |
| 63 | if (MIB->isInvoke(Inst: It)) { |
| 64 | FuncInfo.Omitted = true; |
| 65 | return; |
| 66 | } |
| 67 | // check if this function contains call instruction |
| 68 | if (MIB->isCall(Inst: It)) { |
| 69 | FuncInfo.HasCall = true; |
| 70 | const MCSymbol *TargetSymbol = MIB->getTargetSymbol(Inst: It); |
| 71 | // if it is an indirect call, skip |
| 72 | if (TargetSymbol == nullptr) { |
| 73 | FuncInfo.Omitted = true; |
| 74 | return; |
| 75 | } |
| 76 | } |
| 77 | // check if this function modify stack or heap |
| 78 | // TODO: more accurate analysis |
| 79 | bool IsPush = MIB->isPush(Inst: It); |
| 80 | bool IsRipAddr = MIB->hasPCRelOperand(Inst: It); |
| 81 | if (IsPush) |
| 82 | FuncInfo.StackOut = true; |
| 83 | |
| 84 | if (MIB->mayStore(Inst: It) && !IsPush && !IsRipAddr) |
| 85 | FuncInfo.HeapOut = true; |
| 86 | |
| 87 | if (IsRipAddr) |
| 88 | FuncInfo.HasRipAddr = true; |
| 89 | } // end of for (auto &It : ...) |
| 90 | } // end of for (auto *BB : ...) |
| 91 | } |
| 92 | |
| 93 | bool StokeInfo::checkFunction(BinaryFunction &BF, DataflowInfoManager &DInfo, |
| 94 | RegAnalysis &RA, StokeFuncInfo &FuncInfo) { |
| 95 | |
| 96 | std::string Name = BF.getSymbol()->getName().str(); |
| 97 | |
| 98 | if (!BF.isSimple() || BF.isMultiEntry() || BF.empty()) |
| 99 | return false; |
| 100 | BF.getBinaryContext().outs() |
| 101 | << " STOKE-INFO: analyzing function " << Name << "\n" ; |
| 102 | |
| 103 | FuncInfo.FuncName = Name; |
| 104 | FuncInfo.Offset = BF.getFileOffset(); |
| 105 | FuncInfo.Size = BF.getMaxSize(); |
| 106 | FuncInfo.NumInstrs = BF.getNumNonPseudos(); |
| 107 | FuncInfo.NumBlocks = BF.size(); |
| 108 | // early stop for large functions |
| 109 | if (FuncInfo.NumInstrs > 500) |
| 110 | return false; |
| 111 | |
| 112 | FuncInfo.IsLoopFree = BF.isLoopFree(); |
| 113 | if (!FuncInfo.IsLoopFree) { |
| 114 | const BinaryLoopInfo &BLI = BF.getLoopInfo(); |
| 115 | FuncInfo.NumLoops = BLI.OuterLoops; |
| 116 | FuncInfo.MaxLoopDepth = BLI.MaximumDepth; |
| 117 | } |
| 118 | |
| 119 | FuncInfo.HotSize = BF.estimateHotSize(); |
| 120 | FuncInfo.TotalSize = BF.estimateSize(); |
| 121 | FuncInfo.Score = BF.getFunctionScore(); |
| 122 | |
| 123 | checkInstr(BF, FuncInfo); |
| 124 | |
| 125 | // register analysis |
| 126 | BinaryBasicBlock &EntryBB = BF.front(); |
| 127 | assert(EntryBB.isEntryPoint() && "Weird, this should be the entry block!" ); |
| 128 | |
| 129 | MCInst *FirstNonPseudo = EntryBB.getFirstNonPseudoInstr(); |
| 130 | if (!FirstNonPseudo) |
| 131 | return false; |
| 132 | |
| 133 | LLVM_DEBUG(dbgs() << "\t [DefIn]\n\t " ); |
| 134 | BitVector LiveInBV = |
| 135 | *(DInfo.getLivenessAnalysis().getStateAt(Point: FirstNonPseudo)); |
| 136 | LiveInBV &= DefaultDefInMask; |
| 137 | getRegNameFromBitVec(BC: BF.getBinaryContext(), RegV: LiveInBV, NameVec: &FuncInfo.DefIn); |
| 138 | |
| 139 | LLVM_DEBUG(dbgs() << "\t [LiveOut]\n\t " ); |
| 140 | BitVector LiveOutBV = RA.getFunctionClobberList(Func: &BF); |
| 141 | LiveOutBV &= DefaultLiveOutMask; |
| 142 | getRegNameFromBitVec(BC: BF.getBinaryContext(), RegV: LiveOutBV, NameVec: &FuncInfo.LiveOut); |
| 143 | |
| 144 | BF.getBinaryContext().outs() << " STOKE-INFO: end function \n" ; |
| 145 | return true; |
| 146 | } |
| 147 | |
| 148 | Error StokeInfo::runOnFunctions(BinaryContext &BC) { |
| 149 | BC.outs() << "STOKE-INFO: begin of stoke pass\n" ; |
| 150 | |
| 151 | std::ofstream Outfile; |
| 152 | if (!opts::StokeOutputDataFilename.empty()) { |
| 153 | Outfile.open(s: opts::StokeOutputDataFilename); |
| 154 | } else { |
| 155 | BC.errs() << "STOKE-INFO: output file is required\n" ; |
| 156 | return Error::success(); |
| 157 | } |
| 158 | |
| 159 | // check some context meta data |
| 160 | LLVM_DEBUG(dbgs() << "\tTarget: " << BC.TheTarget->getName() << "\n" ); |
| 161 | LLVM_DEBUG(dbgs() << "\tTripleName " << BC.TripleName << "\n" ); |
| 162 | LLVM_DEBUG(dbgs() << "\tgetNumRegs " << BC.MRI->getNumRegs() << "\n" ); |
| 163 | |
| 164 | BinaryFunctionCallGraph CG = buildCallGraph(BC); |
| 165 | RegAnalysis RA(BC, &BC.getBinaryFunctions(), &CG); |
| 166 | |
| 167 | NumRegs = BC.MRI->getNumRegs(); |
| 168 | assert(NumRegs > 0 && "STOKE-INFO: the target register number is incorrect!" ); |
| 169 | |
| 170 | DefaultDefInMask.resize(N: NumRegs, t: false); |
| 171 | DefaultLiveOutMask.resize(N: NumRegs, t: false); |
| 172 | |
| 173 | BC.MIB->getDefaultDefIn(Regs&: DefaultDefInMask); |
| 174 | BC.MIB->getDefaultLiveOut(Regs&: DefaultLiveOutMask); |
| 175 | |
| 176 | getRegNameFromBitVec(BC, RegV: DefaultDefInMask); |
| 177 | getRegNameFromBitVec(BC, RegV: DefaultLiveOutMask); |
| 178 | |
| 179 | StokeFuncInfo FuncInfo; |
| 180 | // analyze all functions |
| 181 | FuncInfo.printCsvHeader(Outfile); |
| 182 | for (auto &BF : BC.getBinaryFunctions()) { |
| 183 | DataflowInfoManager DInfo(BF.second, &RA, nullptr); |
| 184 | FuncInfo.reset(); |
| 185 | if (checkFunction(BF&: BF.second, DInfo, RA, FuncInfo)) |
| 186 | FuncInfo.printData(Outfile); |
| 187 | } |
| 188 | |
| 189 | BC.outs() << "STOKE-INFO: end of stoke pass\n" ; |
| 190 | return Error::success(); |
| 191 | } |
| 192 | |
| 193 | } // namespace bolt |
| 194 | } // namespace llvm |
| 195 | |