//===- SampleProfReader.cpp - Read LLVM sample profile data ---------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception // //===----------------------------------------------------------------------===// // // This file implements the class that reads LLVM sample profiles. It // supports three file formats: text, binary and gcov. // // The textual representation is useful for debugging and testing purposes. The // binary representation is more compact, resulting in smaller file sizes. // // The gcov encoding is the one generated by GCC's AutoFDO profile creation // tool (https://github.com/google/autofdo) // // All three encodings can be used interchangeably as an input sample profile. // //===----------------------------------------------------------------------===// #include "llvm/ProfileData/SampleProfReader.h" #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/StringRef.h" #include "llvm/IR/ProfileSummary.h" #include "llvm/ProfileData/ProfileCommon.h" #include "llvm/ProfileData/SampleProf.h" #include "llvm/Support/Compression.h" #include "llvm/Support/ErrorOr.h" #include "llvm/Support/LEB128.h" #include "llvm/Support/LineIterator.h" #include "llvm/Support/MD5.h" #include "llvm/Support/MemoryBuffer.h" #include "llvm/Support/raw_ostream.h" #include #include #include #include #include #include #include using namespace llvm; using namespace sampleprof; /// Dump the function profile for \p FName. /// /// \param FName Name of the function to print. /// \param OS Stream to emit the output to. void SampleProfileReader::dumpFunctionProfile(StringRef FName, raw_ostream &OS) { OS << "Function: " << FName << ": " << Profiles[FName]; } /// Dump all the function profiles found on stream \p OS. void SampleProfileReader::dump(raw_ostream &OS) { for (const auto &I : Profiles) dumpFunctionProfile(I.getKey(), OS); } /// Parse \p Input as function head. /// /// Parse one line of \p Input, and update function name in \p FName, /// function's total sample count in \p NumSamples, function's entry /// count in \p NumHeadSamples. /// /// \returns true if parsing is successful. static bool ParseHead(const StringRef &Input, StringRef &FName, uint64_t &NumSamples, uint64_t &NumHeadSamples) { if (Input[0] == ' ') return false; size_t n2 = Input.rfind(':'); size_t n1 = Input.rfind(':', n2 - 1); FName = Input.substr(0, n1); if (Input.substr(n1 + 1, n2 - n1 - 1).getAsInteger(10, NumSamples)) return false; if (Input.substr(n2 + 1).getAsInteger(10, NumHeadSamples)) return false; return true; } /// Returns true if line offset \p L is legal (only has 16 bits). static bool isOffsetLegal(unsigned L) { return (L & 0xffff) == L; } /// Parse \p Input that contains metadata. /// Possible metadata: /// - CFG Checksum information: /// !CFGChecksum: 12345 /// Stores the FunctionHash (a.k.a. CFG Checksum) into \p FunctionHash. static bool parseMetadata(const StringRef &Input, uint64_t &FunctionHash) { if (!Input.startswith("!CFGChecksum:")) return false; StringRef CFGInfo = Input.substr(strlen("!CFGChecksum:")).trim(); return !CFGInfo.getAsInteger(10, FunctionHash); } enum class LineType { CallSiteProfile, BodyProfile, Metadata, }; /// Parse \p Input as line sample. /// /// \param Input input line. /// \param LineTy Type of this line. /// \param Depth the depth of the inline stack. /// \param NumSamples total samples of the line/inlined callsite. /// \param LineOffset line offset to the start of the function. /// \param Discriminator discriminator of the line. /// \param TargetCountMap map from indirect call target to count. /// \param FunctionHash the function's CFG hash, used by pseudo probe. /// /// returns true if parsing is successful. static bool ParseLine(const StringRef &Input, LineType &LineTy, uint32_t &Depth, uint64_t &NumSamples, uint32_t &LineOffset, uint32_t &Discriminator, StringRef &CalleeName, DenseMap &TargetCountMap, uint64_t &FunctionHash) { for (Depth = 0; Input[Depth] == ' '; Depth++) ; if (Depth == 0) return false; if (Depth == 1 && Input[Depth] == '!') { LineTy = LineType::Metadata; return parseMetadata(Input.substr(Depth), FunctionHash); } size_t n1 = Input.find(':'); StringRef Loc = Input.substr(Depth, n1 - Depth); size_t n2 = Loc.find('.'); if (n2 == StringRef::npos) { if (Loc.getAsInteger(10, LineOffset) || !isOffsetLegal(LineOffset)) return false; Discriminator = 0; } else { if (Loc.substr(0, n2).getAsInteger(10, LineOffset)) return false; if (Loc.substr(n2 + 1).getAsInteger(10, Discriminator)) return false; } StringRef Rest = Input.substr(n1 + 2); if (isDigit(Rest[0])) { LineTy = LineType::BodyProfile; size_t n3 = Rest.find(' '); if (n3 == StringRef::npos) { if (Rest.getAsInteger(10, NumSamples)) return false; } else { if (Rest.substr(0, n3).getAsInteger(10, NumSamples)) return false; } // Find call targets and their sample counts. // Note: In some cases, there are symbols in the profile which are not // mangled. To accommodate such cases, use colon + integer pairs as the // anchor points. // An example: // _M_construct:1000 string_view >:437 // ":1000" and ":437" are used as anchor points so the string above will // be interpreted as // target: _M_construct // count: 1000 // target: string_view > // count: 437 while (n3 != StringRef::npos) { n3 += Rest.substr(n3).find_first_not_of(' '); Rest = Rest.substr(n3); n3 = Rest.find_first_of(':'); if (n3 == StringRef::npos || n3 == 0) return false; StringRef Target; uint64_t count, n4; while (true) { // Get the segment after the current colon. StringRef AfterColon = Rest.substr(n3 + 1); // Get the target symbol before the current colon. Target = Rest.substr(0, n3); // Check if the word after the current colon is an integer. n4 = AfterColon.find_first_of(' '); n4 = (n4 != StringRef::npos) ? n3 + n4 + 1 : Rest.size(); StringRef WordAfterColon = Rest.substr(n3 + 1, n4 - n3 - 1); if (!WordAfterColon.getAsInteger(10, count)) break; // Try to find the next colon. uint64_t n5 = AfterColon.find_first_of(':'); if (n5 == StringRef::npos) return false; n3 += n5 + 1; } // An anchor point is found. Save the {target, count} pair TargetCountMap[Target] = count; if (n4 == Rest.size()) break; // Change n3 to the next blank space after colon + integer pair. n3 = n4; } } else { LineTy = LineType::CallSiteProfile; size_t n3 = Rest.find_last_of(':'); CalleeName = Rest.substr(0, n3); if (Rest.substr(n3 + 1).getAsInteger(10, NumSamples)) return false; } return true; } /// Load samples from a text file. /// /// See the documentation at the top of the file for an explanation of /// the expected format. /// /// \returns true if the file was loaded successfully, false otherwise. std::error_code SampleProfileReaderText::readImpl() { line_iterator LineIt(*Buffer, /*SkipBlanks=*/true, '#'); sampleprof_error Result = sampleprof_error::success; InlineCallStack InlineStack; uint32_t ProbeProfileCount = 0; // SeenMetadata tracks whether we have processed metadata for the current // top-level function profile. bool SeenMetadata = false; for (; !LineIt.is_at_eof(); ++LineIt) { if ((*LineIt)[(*LineIt).find_first_not_of(' ')] == '#') continue; // Read the header of each function. // // Note that for function identifiers we are actually expecting // mangled names, but we may not always get them. This happens when // the compiler decides not to emit the function (e.g., it was inlined // and removed). In this case, the binary will not have the linkage // name for the function, so the profiler will emit the function's // unmangled name, which may contain characters like ':' and '>' in its // name (member functions, templates, etc). // // The only requirement we place on the identifier, then, is that it // should not begin with a number. if ((*LineIt)[0] != ' ') { uint64_t NumSamples, NumHeadSamples; StringRef FName; if (!ParseHead(*LineIt, FName, NumSamples, NumHeadSamples)) { reportError(LineIt.line_number(), "Expected 'mangled_name:NUM:NUM', found " + *LineIt); return sampleprof_error::malformed; } SeenMetadata = false; SampleContext FContext(FName); if (FContext.hasContext()) ++CSProfileCount; Profiles[FContext] = FunctionSamples(); FunctionSamples &FProfile = Profiles[FContext]; FProfile.setName(FContext.getNameWithoutContext()); FProfile.setContext(FContext); MergeResult(Result, FProfile.addTotalSamples(NumSamples)); MergeResult(Result, FProfile.addHeadSamples(NumHeadSamples)); InlineStack.clear(); InlineStack.push_back(&FProfile); } else { uint64_t NumSamples; StringRef FName; DenseMap TargetCountMap; uint32_t Depth, LineOffset, Discriminator; LineType LineTy; uint64_t FunctionHash; if (!ParseLine(*LineIt, LineTy, Depth, NumSamples, LineOffset, Discriminator, FName, TargetCountMap, FunctionHash)) { reportError(LineIt.line_number(), "Expected 'NUM[.NUM]: NUM[ mangled_name:NUM]*', found " + *LineIt); return sampleprof_error::malformed; } if (SeenMetadata && LineTy != LineType::Metadata) { // Metadata must be put at the end of a function profile. reportError(LineIt.line_number(), "Found non-metadata after metadata: " + *LineIt); return sampleprof_error::malformed; } while (InlineStack.size() > Depth) { InlineStack.pop_back(); } switch (LineTy) { case LineType::CallSiteProfile: { FunctionSamples &FSamples = InlineStack.back()->functionSamplesAt( LineLocation(LineOffset, Discriminator))[std::string(FName)]; FSamples.setName(FName); MergeResult(Result, FSamples.addTotalSamples(NumSamples)); InlineStack.push_back(&FSamples); break; } case LineType::BodyProfile: { while (InlineStack.size() > Depth) { InlineStack.pop_back(); } FunctionSamples &FProfile = *InlineStack.back(); for (const auto &name_count : TargetCountMap) { MergeResult(Result, FProfile.addCalledTargetSamples( LineOffset, Discriminator, name_count.first, name_count.second)); } MergeResult(Result, FProfile.addBodySamples(LineOffset, Discriminator, NumSamples)); break; } case LineType::Metadata: { FunctionSamples &FProfile = *InlineStack.back(); FProfile.setFunctionHash(FunctionHash); ++ProbeProfileCount; SeenMetadata = true; break; } } } } assert((CSProfileCount == 0 || CSProfileCount == Profiles.size()) && "Cannot have both context-sensitive and regular profile"); ProfileIsCS = (CSProfileCount > 0); assert((ProbeProfileCount == 0 || ProbeProfileCount == Profiles.size()) && "Cannot have both probe-based profiles and regular profiles"); ProfileIsProbeBased = (ProbeProfileCount > 0); FunctionSamples::ProfileIsProbeBased = ProfileIsProbeBased; FunctionSamples::ProfileIsCS = ProfileIsCS; if (Result == sampleprof_error::success) computeSummary(); return Result; } bool SampleProfileReaderText::hasFormat(const MemoryBuffer &Buffer) { bool result = false; // Check that the first non-comment line is a valid function header. line_iterator LineIt(Buffer, /*SkipBlanks=*/true, '#'); if (!LineIt.is_at_eof()) { if ((*LineIt)[0] != ' ') { uint64_t NumSamples, NumHeadSamples; StringRef FName; result = ParseHead(*LineIt, FName, NumSamples, NumHeadSamples); } } return result; } template ErrorOr SampleProfileReaderBinary::readNumber() { unsigned NumBytesRead = 0; std::error_code EC; uint64_t Val = decodeULEB128(Data, &NumBytesRead); if (Val > std::numeric_limits::max()) EC = sampleprof_error::malformed; else if (Data + NumBytesRead > End) EC = sampleprof_error::truncated; else EC = sampleprof_error::success; if (EC) { reportError(0, EC.message()); return EC; } Data += NumBytesRead; return static_cast(Val); } ErrorOr SampleProfileReaderBinary::readString() { std::error_code EC; StringRef Str(reinterpret_cast(Data)); if (Data + Str.size() + 1 > End) { EC = sampleprof_error::truncated; reportError(0, EC.message()); return EC; } Data += Str.size() + 1; return Str; } template ErrorOr SampleProfileReaderBinary::readUnencodedNumber() { std::error_code EC; if (Data + sizeof(T) > End) { EC = sampleprof_error::truncated; reportError(0, EC.message()); return EC; } using namespace support; T Val = endian::readNext(Data); return Val; } template inline ErrorOr SampleProfileReaderBinary::readStringIndex(T &Table) { std::error_code EC; auto Idx = readNumber(); if (std::error_code EC = Idx.getError()) return EC; if (*Idx >= Table.size()) return sampleprof_error::truncated_name_table; return *Idx; } ErrorOr SampleProfileReaderBinary::readStringFromTable() { auto Idx = readStringIndex(NameTable); if (std::error_code EC = Idx.getError()) return EC; return NameTable[*Idx]; } ErrorOr SampleProfileReaderExtBinaryBase::readStringFromTable() { if (!FixedLengthMD5) return SampleProfileReaderBinary::readStringFromTable(); // read NameTable index. auto Idx = readStringIndex(NameTable); if (std::error_code EC = Idx.getError()) return EC; // Check whether the name to be accessed has been accessed before, // if not, read it from memory directly. StringRef &SR = NameTable[*Idx]; if (SR.empty()) { const uint8_t *SavedData = Data; Data = MD5NameMemStart + ((*Idx) * sizeof(uint64_t)); auto FID = readUnencodedNumber(); if (std::error_code EC = FID.getError()) return EC; // Save the string converted from uint64_t in MD5StringBuf. All the // references to the name are all StringRefs refering to the string // in MD5StringBuf. MD5StringBuf->push_back(std::to_string(*FID)); SR = MD5StringBuf->back(); Data = SavedData; } return SR; } ErrorOr SampleProfileReaderCompactBinary::readStringFromTable() { auto Idx = readStringIndex(NameTable); if (std::error_code EC = Idx.getError()) return EC; return StringRef(NameTable[*Idx]); } std::error_code SampleProfileReaderBinary::readProfile(FunctionSamples &FProfile) { auto NumSamples = readNumber(); if (std::error_code EC = NumSamples.getError()) return EC; FProfile.addTotalSamples(*NumSamples); // Read the samples in the body. auto NumRecords = readNumber(); if (std::error_code EC = NumRecords.getError()) return EC; for (uint32_t I = 0; I < *NumRecords; ++I) { auto LineOffset = readNumber(); if (std::error_code EC = LineOffset.getError()) return EC; if (!isOffsetLegal(*LineOffset)) { return std::error_code(); } auto Discriminator = readNumber(); if (std::error_code EC = Discriminator.getError()) return EC; auto NumSamples = readNumber(); if (std::error_code EC = NumSamples.getError()) return EC; auto NumCalls = readNumber(); if (std::error_code EC = NumCalls.getError()) return EC; for (uint32_t J = 0; J < *NumCalls; ++J) { auto CalledFunction(readStringFromTable()); if (std::error_code EC = CalledFunction.getError()) return EC; auto CalledFunctionSamples = readNumber(); if (std::error_code EC = CalledFunctionSamples.getError()) return EC; FProfile.addCalledTargetSamples(*LineOffset, *Discriminator, *CalledFunction, *CalledFunctionSamples); } FProfile.addBodySamples(*LineOffset, *Discriminator, *NumSamples); } // Read all the samples for inlined function calls. auto NumCallsites = readNumber(); if (std::error_code EC = NumCallsites.getError()) return EC; for (uint32_t J = 0; J < *NumCallsites; ++J) { auto LineOffset = readNumber(); if (std::error_code EC = LineOffset.getError()) return EC; auto Discriminator = readNumber(); if (std::error_code EC = Discriminator.getError()) return EC; auto FName(readStringFromTable()); if (std::error_code EC = FName.getError()) return EC; FunctionSamples &CalleeProfile = FProfile.functionSamplesAt( LineLocation(*LineOffset, *Discriminator))[std::string(*FName)]; CalleeProfile.setName(*FName); if (std::error_code EC = readProfile(CalleeProfile)) return EC; } return sampleprof_error::success; } std::error_code SampleProfileReaderBinary::readFuncProfile(const uint8_t *Start) { Data = Start; auto NumHeadSamples = readNumber(); if (std::error_code EC = NumHeadSamples.getError()) return EC; auto FName(readStringFromTable()); if (std::error_code EC = FName.getError()) return EC; SampleContext FContext(*FName); Profiles[FContext] = FunctionSamples(); FunctionSamples &FProfile = Profiles[FContext]; FProfile.setName(FContext.getNameWithoutContext()); FProfile.setContext(FContext); FProfile.addHeadSamples(*NumHeadSamples); if (FContext.hasContext()) CSProfileCount++; if (std::error_code EC = readProfile(FProfile)) return EC; return sampleprof_error::success; } std::error_code SampleProfileReaderBinary::readImpl() { while (!at_eof()) { if (std::error_code EC = readFuncProfile(Data)) return EC; } return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readOneSection( const uint8_t *Start, uint64_t Size, const SecHdrTableEntry &Entry) { Data = Start; End = Start + Size; switch (Entry.Type) { case SecProfSummary: if (std::error_code EC = readSummary()) return EC; if (hasSecFlag(Entry, SecProfSummaryFlags::SecFlagPartial)) Summary->setPartialProfile(true); break; case SecNameTable: { FixedLengthMD5 = hasSecFlag(Entry, SecNameTableFlags::SecFlagFixedLengthMD5); bool UseMD5 = hasSecFlag(Entry, SecNameTableFlags::SecFlagMD5Name); assert((!FixedLengthMD5 || UseMD5) && "If FixedLengthMD5 is true, UseMD5 has to be true"); if (std::error_code EC = readNameTableSec(UseMD5)) return EC; break; } case SecLBRProfile: if (std::error_code EC = readFuncProfiles()) return EC; break; case SecFuncOffsetTable: if (std::error_code EC = readFuncOffsetTable()) return EC; break; case SecFuncMetadata: ProfileIsProbeBased = hasSecFlag(Entry, SecFuncMetadataFlags::SecFlagIsProbeBased); FunctionSamples::ProfileIsProbeBased = ProfileIsProbeBased; if (std::error_code EC = readFuncMetadata()) return EC; break; case SecProfileSymbolList: if (std::error_code EC = readProfileSymbolList()) return EC; break; default: if (std::error_code EC = readCustomSection(Entry)) return EC; break; } return sampleprof_error::success; } void SampleProfileReaderExtBinaryBase::collectFuncsFrom(const Module &M) { UseAllFuncs = false; FuncsToUse.clear(); for (auto &F : M) FuncsToUse.insert(FunctionSamples::getCanonicalFnName(F)); } std::error_code SampleProfileReaderExtBinaryBase::readFuncOffsetTable() { // If there are more than one FuncOffsetTable, the profile read associated // with previous FuncOffsetTable has to be done before next FuncOffsetTable // is read. FuncOffsetTable.clear(); auto Size = readNumber(); if (std::error_code EC = Size.getError()) return EC; FuncOffsetTable.reserve(*Size); for (uint32_t I = 0; I < *Size; ++I) { auto FName(readStringFromTable()); if (std::error_code EC = FName.getError()) return EC; auto Offset = readNumber(); if (std::error_code EC = Offset.getError()) return EC; FuncOffsetTable[*FName] = *Offset; } return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readFuncProfiles() { const uint8_t *Start = Data; if (UseAllFuncs) { while (Data < End) { if (std::error_code EC = readFuncProfile(Data)) return EC; } assert(Data == End && "More data is read than expected"); } else { if (Remapper) { for (auto Name : FuncsToUse) { Remapper->insert(Name); } } if (useMD5()) { for (auto Name : FuncsToUse) { auto GUID = std::to_string(MD5Hash(Name)); auto iter = FuncOffsetTable.find(StringRef(GUID)); if (iter == FuncOffsetTable.end()) continue; const uint8_t *FuncProfileAddr = Start + iter->second; assert(FuncProfileAddr < End && "out of LBRProfile section"); if (std::error_code EC = readFuncProfile(FuncProfileAddr)) return EC; } } else { for (auto NameOffset : FuncOffsetTable) { SampleContext FContext(NameOffset.first); auto FuncName = FContext.getNameWithoutContext(); if (!FuncsToUse.count(FuncName) && (!Remapper || !Remapper->exist(FuncName))) continue; const uint8_t *FuncProfileAddr = Start + NameOffset.second; assert(FuncProfileAddr < End && "out of LBRProfile section"); if (std::error_code EC = readFuncProfile(FuncProfileAddr)) return EC; } } Data = End; } assert((CSProfileCount == 0 || CSProfileCount == Profiles.size()) && "Cannot have both context-sensitive and regular profile"); ProfileIsCS = (CSProfileCount > 0); FunctionSamples::ProfileIsCS = ProfileIsCS; return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readProfileSymbolList() { if (!ProfSymList) ProfSymList = std::make_unique(); if (std::error_code EC = ProfSymList->read(Data, End - Data)) return EC; Data = End; return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::decompressSection( const uint8_t *SecStart, const uint64_t SecSize, const uint8_t *&DecompressBuf, uint64_t &DecompressBufSize) { Data = SecStart; End = SecStart + SecSize; auto DecompressSize = readNumber(); if (std::error_code EC = DecompressSize.getError()) return EC; DecompressBufSize = *DecompressSize; auto CompressSize = readNumber(); if (std::error_code EC = CompressSize.getError()) return EC; if (!llvm::zlib::isAvailable()) return sampleprof_error::zlib_unavailable; StringRef CompressedStrings(reinterpret_cast(Data), *CompressSize); char *Buffer = Allocator.Allocate(DecompressBufSize); size_t UCSize = DecompressBufSize; llvm::Error E = zlib::uncompress(CompressedStrings, Buffer, UCSize); if (E) return sampleprof_error::uncompress_failed; DecompressBuf = reinterpret_cast(Buffer); return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readImpl() { const uint8_t *BufStart = reinterpret_cast(Buffer->getBufferStart()); for (auto &Entry : SecHdrTable) { // Skip empty section. if (!Entry.Size) continue; // Skip sections without context when SkipFlatProf is true. if (SkipFlatProf && hasSecFlag(Entry, SecCommonFlags::SecFlagFlat)) continue; const uint8_t *SecStart = BufStart + Entry.Offset; uint64_t SecSize = Entry.Size; // If the section is compressed, decompress it into a buffer // DecompressBuf before reading the actual data. The pointee of // 'Data' will be changed to buffer hold by DecompressBuf // temporarily when reading the actual data. bool isCompressed = hasSecFlag(Entry, SecCommonFlags::SecFlagCompress); if (isCompressed) { const uint8_t *DecompressBuf; uint64_t DecompressBufSize; if (std::error_code EC = decompressSection( SecStart, SecSize, DecompressBuf, DecompressBufSize)) return EC; SecStart = DecompressBuf; SecSize = DecompressBufSize; } if (std::error_code EC = readOneSection(SecStart, SecSize, Entry)) return EC; if (Data != SecStart + SecSize) return sampleprof_error::malformed; // Change the pointee of 'Data' from DecompressBuf to original Buffer. if (isCompressed) { Data = BufStart + Entry.Offset; End = BufStart + Buffer->getBufferSize(); } } return sampleprof_error::success; } std::error_code SampleProfileReaderCompactBinary::readImpl() { std::vector OffsetsToUse; if (UseAllFuncs) { for (auto FuncEntry : FuncOffsetTable) { OffsetsToUse.push_back(FuncEntry.second); } } else { for (auto Name : FuncsToUse) { auto GUID = std::to_string(MD5Hash(Name)); auto iter = FuncOffsetTable.find(StringRef(GUID)); if (iter == FuncOffsetTable.end()) continue; OffsetsToUse.push_back(iter->second); } } for (auto Offset : OffsetsToUse) { const uint8_t *SavedData = Data; if (std::error_code EC = readFuncProfile( reinterpret_cast(Buffer->getBufferStart()) + Offset)) return EC; Data = SavedData; } return sampleprof_error::success; } std::error_code SampleProfileReaderRawBinary::verifySPMagic(uint64_t Magic) { if (Magic == SPMagic()) return sampleprof_error::success; return sampleprof_error::bad_magic; } std::error_code SampleProfileReaderExtBinary::verifySPMagic(uint64_t Magic) { if (Magic == SPMagic(SPF_Ext_Binary)) return sampleprof_error::success; return sampleprof_error::bad_magic; } std::error_code SampleProfileReaderCompactBinary::verifySPMagic(uint64_t Magic) { if (Magic == SPMagic(SPF_Compact_Binary)) return sampleprof_error::success; return sampleprof_error::bad_magic; } std::error_code SampleProfileReaderBinary::readNameTable() { auto Size = readNumber(); if (std::error_code EC = Size.getError()) return EC; NameTable.reserve(*Size + NameTable.size()); for (uint32_t I = 0; I < *Size; ++I) { auto Name(readString()); if (std::error_code EC = Name.getError()) return EC; NameTable.push_back(*Name); } return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readMD5NameTable() { auto Size = readNumber(); if (std::error_code EC = Size.getError()) return EC; MD5StringBuf = std::make_unique>(); MD5StringBuf->reserve(*Size); if (FixedLengthMD5) { // Preallocate and initialize NameTable so we can check whether a name // index has been read before by checking whether the element in the // NameTable is empty, meanwhile readStringIndex can do the boundary // check using the size of NameTable. NameTable.resize(*Size + NameTable.size()); MD5NameMemStart = Data; Data = Data + (*Size) * sizeof(uint64_t); return sampleprof_error::success; } NameTable.reserve(*Size); for (uint32_t I = 0; I < *Size; ++I) { auto FID = readNumber(); if (std::error_code EC = FID.getError()) return EC; MD5StringBuf->push_back(std::to_string(*FID)); // NameTable is a vector of StringRef. Here it is pushing back a // StringRef initialized with the last string in MD5stringBuf. NameTable.push_back(MD5StringBuf->back()); } return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readNameTableSec(bool IsMD5) { if (IsMD5) return readMD5NameTable(); return SampleProfileReaderBinary::readNameTable(); } std::error_code SampleProfileReaderExtBinaryBase::readFuncMetadata() { if (!ProfileIsProbeBased) return sampleprof_error::success; while (Data < End) { auto FName(readStringFromTable()); if (std::error_code EC = FName.getError()) return EC; auto Checksum = readNumber(); if (std::error_code EC = Checksum.getError()) return EC; SampleContext FContext(*FName); // No need to load metadata for profiles that are not loaded in the current // module. if (Profiles.count(FContext)) Profiles[FContext].setFunctionHash(*Checksum); } assert(Data == End && "More data is read than expected"); return sampleprof_error::success; } std::error_code SampleProfileReaderCompactBinary::readNameTable() { auto Size = readNumber(); if (std::error_code EC = Size.getError()) return EC; NameTable.reserve(*Size); for (uint32_t I = 0; I < *Size; ++I) { auto FID = readNumber(); if (std::error_code EC = FID.getError()) return EC; NameTable.push_back(std::to_string(*FID)); } return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readSecHdrTableEntry(uint32_t Idx) { SecHdrTableEntry Entry; auto Type = readUnencodedNumber(); if (std::error_code EC = Type.getError()) return EC; Entry.Type = static_cast(*Type); auto Flags = readUnencodedNumber(); if (std::error_code EC = Flags.getError()) return EC; Entry.Flags = *Flags; auto Offset = readUnencodedNumber(); if (std::error_code EC = Offset.getError()) return EC; Entry.Offset = *Offset; auto Size = readUnencodedNumber(); if (std::error_code EC = Size.getError()) return EC; Entry.Size = *Size; Entry.LayoutIndex = Idx; SecHdrTable.push_back(std::move(Entry)); return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readSecHdrTable() { auto EntryNum = readUnencodedNumber(); if (std::error_code EC = EntryNum.getError()) return EC; for (uint32_t i = 0; i < (*EntryNum); i++) if (std::error_code EC = readSecHdrTableEntry(i)) return EC; return sampleprof_error::success; } std::error_code SampleProfileReaderExtBinaryBase::readHeader() { const uint8_t *BufStart = reinterpret_cast(Buffer->getBufferStart()); Data = BufStart; End = BufStart + Buffer->getBufferSize(); if (std::error_code EC = readMagicIdent()) return EC; if (std::error_code EC = readSecHdrTable()) return EC; return sampleprof_error::success; } uint64_t SampleProfileReaderExtBinaryBase::getSectionSize(SecType Type) { uint64_t Size = 0; for (auto &Entry : SecHdrTable) { if (Entry.Type == Type) Size += Entry.Size; } return Size; } uint64_t SampleProfileReaderExtBinaryBase::getFileSize() { // Sections in SecHdrTable is not necessarily in the same order as // sections in the profile because section like FuncOffsetTable needs // to be written after section LBRProfile but needs to be read before // section LBRProfile, so we cannot simply use the last entry in // SecHdrTable to calculate the file size. uint64_t FileSize = 0; for (auto &Entry : SecHdrTable) { FileSize = std::max(Entry.Offset + Entry.Size, FileSize); } return FileSize; } static std::string getSecFlagsStr(const SecHdrTableEntry &Entry) { std::string Flags; if (hasSecFlag(Entry, SecCommonFlags::SecFlagCompress)) Flags.append("{compressed,"); else Flags.append("{"); if (hasSecFlag(Entry, SecCommonFlags::SecFlagFlat)) Flags.append("flat,"); switch (Entry.Type) { case SecNameTable: if (hasSecFlag(Entry, SecNameTableFlags::SecFlagFixedLengthMD5)) Flags.append("fixlenmd5,"); else if (hasSecFlag(Entry, SecNameTableFlags::SecFlagMD5Name)) Flags.append("md5,"); break; case SecProfSummary: if (hasSecFlag(Entry, SecProfSummaryFlags::SecFlagPartial)) Flags.append("partial,"); break; default: break; } char &last = Flags.back(); if (last == ',') last = '}'; else Flags.append("}"); return Flags; } bool SampleProfileReaderExtBinaryBase::dumpSectionInfo(raw_ostream &OS) { uint64_t TotalSecsSize = 0; for (auto &Entry : SecHdrTable) { OS << getSecName(Entry.Type) << " - Offset: " << Entry.Offset << ", Size: " << Entry.Size << ", Flags: " << getSecFlagsStr(Entry) << "\n"; ; TotalSecsSize += Entry.Size; } uint64_t HeaderSize = SecHdrTable.front().Offset; assert(HeaderSize + TotalSecsSize == getFileSize() && "Size of 'header + sections' doesn't match the total size of profile"); OS << "Header Size: " << HeaderSize << "\n"; OS << "Total Sections Size: " << TotalSecsSize << "\n"; OS << "File Size: " << getFileSize() << "\n"; return true; } std::error_code SampleProfileReaderBinary::readMagicIdent() { // Read and check the magic identifier. auto Magic = readNumber(); if (std::error_code EC = Magic.getError()) return EC; else if (std::error_code EC = verifySPMagic(*Magic)) return EC; // Read the version number. auto Version = readNumber(); if (std::error_code EC = Version.getError()) return EC; else if (*Version != SPVersion()) return sampleprof_error::unsupported_version; return sampleprof_error::success; } std::error_code SampleProfileReaderBinary::readHeader() { Data = reinterpret_cast(Buffer->getBufferStart()); End = Data + Buffer->getBufferSize(); if (std::error_code EC = readMagicIdent()) return EC; if (std::error_code EC = readSummary()) return EC; if (std::error_code EC = readNameTable()) return EC; return sampleprof_error::success; } std::error_code SampleProfileReaderCompactBinary::readHeader() { SampleProfileReaderBinary::readHeader(); if (std::error_code EC = readFuncOffsetTable()) return EC; return sampleprof_error::success; } std::error_code SampleProfileReaderCompactBinary::readFuncOffsetTable() { auto TableOffset = readUnencodedNumber(); if (std::error_code EC = TableOffset.getError()) return EC; const uint8_t *SavedData = Data; const uint8_t *TableStart = reinterpret_cast(Buffer->getBufferStart()) + *TableOffset; Data = TableStart; auto Size = readNumber(); if (std::error_code EC = Size.getError()) return EC; FuncOffsetTable.reserve(*Size); for (uint32_t I = 0; I < *Size; ++I) { auto FName(readStringFromTable()); if (std::error_code EC = FName.getError()) return EC; auto Offset = readNumber(); if (std::error_code EC = Offset.getError()) return EC; FuncOffsetTable[*FName] = *Offset; } End = TableStart; Data = SavedData; return sampleprof_error::success; } void SampleProfileReaderCompactBinary::collectFuncsFrom(const Module &M) { UseAllFuncs = false; FuncsToUse.clear(); for (auto &F : M) FuncsToUse.insert(FunctionSamples::getCanonicalFnName(F)); } std::error_code SampleProfileReaderBinary::readSummaryEntry( std::vector &Entries) { auto Cutoff = readNumber(); if (std::error_code EC = Cutoff.getError()) return EC; auto MinBlockCount = readNumber(); if (std::error_code EC = MinBlockCount.getError()) return EC; auto NumBlocks = readNumber(); if (std::error_code EC = NumBlocks.getError()) return EC; Entries.emplace_back(*Cutoff, *MinBlockCount, *NumBlocks); return sampleprof_error::success; } std::error_code SampleProfileReaderBinary::readSummary() { auto TotalCount = readNumber(); if (std::error_code EC = TotalCount.getError()) return EC; auto MaxBlockCount = readNumber(); if (std::error_code EC = MaxBlockCount.getError()) return EC; auto MaxFunctionCount = readNumber(); if (std::error_code EC = MaxFunctionCount.getError()) return EC; auto NumBlocks = readNumber(); if (std::error_code EC = NumBlocks.getError()) return EC; auto NumFunctions = readNumber(); if (std::error_code EC = NumFunctions.getError()) return EC; auto NumSummaryEntries = readNumber(); if (std::error_code EC = NumSummaryEntries.getError()) return EC; std::vector Entries; for (unsigned i = 0; i < *NumSummaryEntries; i++) { std::error_code EC = readSummaryEntry(Entries); if (EC != sampleprof_error::success) return EC; } Summary = std::make_unique( ProfileSummary::PSK_Sample, Entries, *TotalCount, *MaxBlockCount, 0, *MaxFunctionCount, *NumBlocks, *NumFunctions); return sampleprof_error::success; } bool SampleProfileReaderRawBinary::hasFormat(const MemoryBuffer &Buffer) { const uint8_t *Data = reinterpret_cast(Buffer.getBufferStart()); uint64_t Magic = decodeULEB128(Data); return Magic == SPMagic(); } bool SampleProfileReaderExtBinary::hasFormat(const MemoryBuffer &Buffer) { const uint8_t *Data = reinterpret_cast(Buffer.getBufferStart()); uint64_t Magic = decodeULEB128(Data); return Magic == SPMagic(SPF_Ext_Binary); } bool SampleProfileReaderCompactBinary::hasFormat(const MemoryBuffer &Buffer) { const uint8_t *Data = reinterpret_cast(Buffer.getBufferStart()); uint64_t Magic = decodeULEB128(Data); return Magic == SPMagic(SPF_Compact_Binary); } std::error_code SampleProfileReaderGCC::skipNextWord() { uint32_t dummy; if (!GcovBuffer.readInt(dummy)) return sampleprof_error::truncated; return sampleprof_error::success; } template ErrorOr SampleProfileReaderGCC::readNumber() { if (sizeof(T) <= sizeof(uint32_t)) { uint32_t Val; if (GcovBuffer.readInt(Val) && Val <= std::numeric_limits::max()) return static_cast(Val); } else if (sizeof(T) <= sizeof(uint64_t)) { uint64_t Val; if (GcovBuffer.readInt64(Val) && Val <= std::numeric_limits::max()) return static_cast(Val); } std::error_code EC = sampleprof_error::malformed; reportError(0, EC.message()); return EC; } ErrorOr SampleProfileReaderGCC::readString() { StringRef Str; if (!GcovBuffer.readString(Str)) return sampleprof_error::truncated; return Str; } std::error_code SampleProfileReaderGCC::readHeader() { // Read the magic identifier. if (!GcovBuffer.readGCDAFormat()) return sampleprof_error::unrecognized_format; // Read the version number. Note - the GCC reader does not validate this // version, but the profile creator generates v704. GCOV::GCOVVersion version; if (!GcovBuffer.readGCOVVersion(version)) return sampleprof_error::unrecognized_format; if (version != GCOV::V407) return sampleprof_error::unsupported_version; // Skip the empty integer. if (std::error_code EC = skipNextWord()) return EC; return sampleprof_error::success; } std::error_code SampleProfileReaderGCC::readSectionTag(uint32_t Expected) { uint32_t Tag; if (!GcovBuffer.readInt(Tag)) return sampleprof_error::truncated; if (Tag != Expected) return sampleprof_error::malformed; if (std::error_code EC = skipNextWord()) return EC; return sampleprof_error::success; } std::error_code SampleProfileReaderGCC::readNameTable() { if (std::error_code EC = readSectionTag(GCOVTagAFDOFileNames)) return EC; uint32_t Size; if (!GcovBuffer.readInt(Size)) return sampleprof_error::truncated; for (uint32_t I = 0; I < Size; ++I) { StringRef Str; if (!GcovBuffer.readString(Str)) return sampleprof_error::truncated; Names.push_back(std::string(Str)); } return sampleprof_error::success; } std::error_code SampleProfileReaderGCC::readFunctionProfiles() { if (std::error_code EC = readSectionTag(GCOVTagAFDOFunction)) return EC; uint32_t NumFunctions; if (!GcovBuffer.readInt(NumFunctions)) return sampleprof_error::truncated; InlineCallStack Stack; for (uint32_t I = 0; I < NumFunctions; ++I) if (std::error_code EC = readOneFunctionProfile(Stack, true, 0)) return EC; computeSummary(); return sampleprof_error::success; } std::error_code SampleProfileReaderGCC::readOneFunctionProfile( const InlineCallStack &InlineStack, bool Update, uint32_t Offset) { uint64_t HeadCount = 0; if (InlineStack.size() == 0) if (!GcovBuffer.readInt64(HeadCount)) return sampleprof_error::truncated; uint32_t NameIdx; if (!GcovBuffer.readInt(NameIdx)) return sampleprof_error::truncated; StringRef Name(Names[NameIdx]); uint32_t NumPosCounts; if (!GcovBuffer.readInt(NumPosCounts)) return sampleprof_error::truncated; uint32_t NumCallsites; if (!GcovBuffer.readInt(NumCallsites)) return sampleprof_error::truncated; FunctionSamples *FProfile = nullptr; if (InlineStack.size() == 0) { // If this is a top function that we have already processed, do not // update its profile again. This happens in the presence of // function aliases. Since these aliases share the same function // body, there will be identical replicated profiles for the // original function. In this case, we simply not bother updating // the profile of the original function. FProfile = &Profiles[Name]; FProfile->addHeadSamples(HeadCount); if (FProfile->getTotalSamples() > 0) Update = false; } else { // Otherwise, we are reading an inlined instance. The top of the // inline stack contains the profile of the caller. Insert this // callee in the caller's CallsiteMap. FunctionSamples *CallerProfile = InlineStack.front(); uint32_t LineOffset = Offset >> 16; uint32_t Discriminator = Offset & 0xffff; FProfile = &CallerProfile->functionSamplesAt( LineLocation(LineOffset, Discriminator))[std::string(Name)]; } FProfile->setName(Name); for (uint32_t I = 0; I < NumPosCounts; ++I) { uint32_t Offset; if (!GcovBuffer.readInt(Offset)) return sampleprof_error::truncated; uint32_t NumTargets; if (!GcovBuffer.readInt(NumTargets)) return sampleprof_error::truncated; uint64_t Count; if (!GcovBuffer.readInt64(Count)) return sampleprof_error::truncated; // The line location is encoded in the offset as: // high 16 bits: line offset to the start of the function. // low 16 bits: discriminator. uint32_t LineOffset = Offset >> 16; uint32_t Discriminator = Offset & 0xffff; InlineCallStack NewStack; NewStack.push_back(FProfile); llvm::append_range(NewStack, InlineStack); if (Update) { // Walk up the inline stack, adding the samples on this line to // the total sample count of the callers in the chain. for (auto CallerProfile : NewStack) CallerProfile->addTotalSamples(Count); // Update the body samples for the current profile. FProfile->addBodySamples(LineOffset, Discriminator, Count); } // Process the list of functions called at an indirect call site. // These are all the targets that a function pointer (or virtual // function) resolved at runtime. for (uint32_t J = 0; J < NumTargets; J++) { uint32_t HistVal; if (!GcovBuffer.readInt(HistVal)) return sampleprof_error::truncated; if (HistVal != HIST_TYPE_INDIR_CALL_TOPN) return sampleprof_error::malformed; uint64_t TargetIdx; if (!GcovBuffer.readInt64(TargetIdx)) return sampleprof_error::truncated; StringRef TargetName(Names[TargetIdx]); uint64_t TargetCount; if (!GcovBuffer.readInt64(TargetCount)) return sampleprof_error::truncated; if (Update) FProfile->addCalledTargetSamples(LineOffset, Discriminator, TargetName, TargetCount); } } // Process all the inlined callers into the current function. These // are all the callsites that were inlined into this function. for (uint32_t I = 0; I < NumCallsites; I++) { // The offset is encoded as: // high 16 bits: line offset to the start of the function. // low 16 bits: discriminator. uint32_t Offset; if (!GcovBuffer.readInt(Offset)) return sampleprof_error::truncated; InlineCallStack NewStack; NewStack.push_back(FProfile); llvm::append_range(NewStack, InlineStack); if (std::error_code EC = readOneFunctionProfile(NewStack, Update, Offset)) return EC; } return sampleprof_error::success; } /// Read a GCC AutoFDO profile. /// /// This format is generated by the Linux Perf conversion tool at /// https://github.com/google/autofdo. std::error_code SampleProfileReaderGCC::readImpl() { // Read the string table. if (std::error_code EC = readNameTable()) return EC; // Read the source profile. if (std::error_code EC = readFunctionProfiles()) return EC; return sampleprof_error::success; } bool SampleProfileReaderGCC::hasFormat(const MemoryBuffer &Buffer) { StringRef Magic(reinterpret_cast(Buffer.getBufferStart())); return Magic == "adcg*704"; } void SampleProfileReaderItaniumRemapper::applyRemapping(LLVMContext &Ctx) { // If the reader uses MD5 to represent string, we can't remap it because // we don't know what the original function names were. if (Reader.useMD5()) { Ctx.diagnose(DiagnosticInfoSampleProfile( Reader.getBuffer()->getBufferIdentifier(), "Profile data remapping cannot be applied to profile data " "in compact format (original mangled names are not available).", DS_Warning)); return; } // CSSPGO-TODO: Remapper is not yet supported. // We will need to remap the entire context string. assert(Remappings && "should be initialized while creating remapper"); for (auto &Sample : Reader.getProfiles()) { DenseSet NamesInSample; Sample.second.findAllNames(NamesInSample); for (auto &Name : NamesInSample) if (auto Key = Remappings->insert(Name)) NameMap.insert({Key, Name}); } RemappingApplied = true; } Optional SampleProfileReaderItaniumRemapper::lookUpNameInProfile(StringRef Fname) { if (auto Key = Remappings->lookup(Fname)) return NameMap.lookup(Key); return None; } /// Prepare a memory buffer for the contents of \p Filename. /// /// \returns an error code indicating the status of the buffer. static ErrorOr> setupMemoryBuffer(const Twine &Filename) { auto BufferOrErr = MemoryBuffer::getFileOrSTDIN(Filename); if (std::error_code EC = BufferOrErr.getError()) return EC; auto Buffer = std::move(BufferOrErr.get()); // Sanity check the file. if (uint64_t(Buffer->getBufferSize()) > std::numeric_limits::max()) return sampleprof_error::too_large; return std::move(Buffer); } /// Create a sample profile reader based on the format of the input file. /// /// \param Filename The file to open. /// /// \param C The LLVM context to use to emit diagnostics. /// /// \param RemapFilename The file used for profile remapping. /// /// \returns an error code indicating the status of the created reader. ErrorOr> SampleProfileReader::create(const std::string Filename, LLVMContext &C, const std::string RemapFilename) { auto BufferOrError = setupMemoryBuffer(Filename); if (std::error_code EC = BufferOrError.getError()) return EC; return create(BufferOrError.get(), C, RemapFilename); } /// Create a sample profile remapper from the given input, to remap the /// function names in the given profile data. /// /// \param Filename The file to open. /// /// \param Reader The profile reader the remapper is going to be applied to. /// /// \param C The LLVM context to use to emit diagnostics. /// /// \returns an error code indicating the status of the created reader. ErrorOr> SampleProfileReaderItaniumRemapper::create(const std::string Filename, SampleProfileReader &Reader, LLVMContext &C) { auto BufferOrError = setupMemoryBuffer(Filename); if (std::error_code EC = BufferOrError.getError()) return EC; return create(BufferOrError.get(), Reader, C); } /// Create a sample profile remapper from the given input, to remap the /// function names in the given profile data. /// /// \param B The memory buffer to create the reader from (assumes ownership). /// /// \param C The LLVM context to use to emit diagnostics. /// /// \param Reader The profile reader the remapper is going to be applied to. /// /// \returns an error code indicating the status of the created reader. ErrorOr> SampleProfileReaderItaniumRemapper::create(std::unique_ptr &B, SampleProfileReader &Reader, LLVMContext &C) { auto Remappings = std::make_unique(); if (Error E = Remappings->read(*B.get())) { handleAllErrors( std::move(E), [&](const SymbolRemappingParseError &ParseError) { C.diagnose(DiagnosticInfoSampleProfile(B->getBufferIdentifier(), ParseError.getLineNum(), ParseError.getMessage())); }); return sampleprof_error::malformed; } return std::make_unique( std::move(B), std::move(Remappings), Reader); } /// Create a sample profile reader based on the format of the input data. /// /// \param B The memory buffer to create the reader from (assumes ownership). /// /// \param C The LLVM context to use to emit diagnostics. /// /// \param RemapFilename The file used for profile remapping. /// /// \returns an error code indicating the status of the created reader. ErrorOr> SampleProfileReader::create(std::unique_ptr &B, LLVMContext &C, const std::string RemapFilename) { std::unique_ptr Reader; if (SampleProfileReaderRawBinary::hasFormat(*B)) Reader.reset(new SampleProfileReaderRawBinary(std::move(B), C)); else if (SampleProfileReaderExtBinary::hasFormat(*B)) Reader.reset(new SampleProfileReaderExtBinary(std::move(B), C)); else if (SampleProfileReaderCompactBinary::hasFormat(*B)) Reader.reset(new SampleProfileReaderCompactBinary(std::move(B), C)); else if (SampleProfileReaderGCC::hasFormat(*B)) Reader.reset(new SampleProfileReaderGCC(std::move(B), C)); else if (SampleProfileReaderText::hasFormat(*B)) Reader.reset(new SampleProfileReaderText(std::move(B), C)); else return sampleprof_error::unrecognized_format; if (!RemapFilename.empty()) { auto ReaderOrErr = SampleProfileReaderItaniumRemapper::create(RemapFilename, *Reader, C); if (std::error_code EC = ReaderOrErr.getError()) { std::string Msg = "Could not create remapper: " + EC.message(); C.diagnose(DiagnosticInfoSampleProfile(RemapFilename, Msg)); return EC; } Reader->Remapper = std::move(ReaderOrErr.get()); } FunctionSamples::Format = Reader->getFormat(); if (std::error_code EC = Reader->readHeader()) { return EC; } return std::move(Reader); } // For text and GCC file formats, we compute the summary after reading the // profile. Binary format has the profile summary in its header. void SampleProfileReader::computeSummary() { SampleProfileSummaryBuilder Builder(ProfileSummaryBuilder::DefaultCutoffs); Summary = Builder.computeSummaryForProfiles(Profiles); }