15ffd83dbSDimitry Andric //===- ObjectFileTransformer.cpp --------------------------------*- C++ -*-===//
25ffd83dbSDimitry Andric //
35ffd83dbSDimitry Andric // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
45ffd83dbSDimitry Andric // See https://llvm.org/LICENSE.txt for license information.
55ffd83dbSDimitry Andric // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
65ffd83dbSDimitry Andric //
75ffd83dbSDimitry Andric //===----------------------------------------------------------------------===//
85ffd83dbSDimitry Andric
95ffd83dbSDimitry Andric #include <unordered_set>
105ffd83dbSDimitry Andric
115ffd83dbSDimitry Andric #include "llvm/Object/ELFObjectFile.h"
125ffd83dbSDimitry Andric #include "llvm/Object/MachOUniversal.h"
135ffd83dbSDimitry Andric #include "llvm/Object/ObjectFile.h"
145ffd83dbSDimitry Andric #include "llvm/Support/DataExtractor.h"
155ffd83dbSDimitry Andric #include "llvm/Support/raw_ostream.h"
165ffd83dbSDimitry Andric
175ffd83dbSDimitry Andric #include "llvm/DebugInfo/GSYM/ObjectFileTransformer.h"
185ffd83dbSDimitry Andric #include "llvm/DebugInfo/GSYM/GsymCreator.h"
195ffd83dbSDimitry Andric
205ffd83dbSDimitry Andric using namespace llvm;
215ffd83dbSDimitry Andric using namespace gsym;
225ffd83dbSDimitry Andric
235ffd83dbSDimitry Andric constexpr uint32_t NT_GNU_BUILD_ID_TAG = 0x03;
245ffd83dbSDimitry Andric
getUUID(const object::ObjectFile & Obj)255ffd83dbSDimitry Andric static std::vector<uint8_t> getUUID(const object::ObjectFile &Obj) {
265ffd83dbSDimitry Andric // Extract the UUID from the object file
275ffd83dbSDimitry Andric std::vector<uint8_t> UUID;
285ffd83dbSDimitry Andric if (auto *MachO = dyn_cast<object::MachOObjectFile>(&Obj)) {
295ffd83dbSDimitry Andric const ArrayRef<uint8_t> MachUUID = MachO->getUuid();
305ffd83dbSDimitry Andric if (!MachUUID.empty())
315ffd83dbSDimitry Andric UUID.assign(MachUUID.data(), MachUUID.data() + MachUUID.size());
325ffd83dbSDimitry Andric } else if (isa<object::ELFObjectFileBase>(&Obj)) {
335ffd83dbSDimitry Andric const StringRef GNUBuildID(".note.gnu.build-id");
345ffd83dbSDimitry Andric for (const object::SectionRef &Sect : Obj.sections()) {
355ffd83dbSDimitry Andric Expected<StringRef> SectNameOrErr = Sect.getName();
365ffd83dbSDimitry Andric if (!SectNameOrErr) {
375ffd83dbSDimitry Andric consumeError(SectNameOrErr.takeError());
385ffd83dbSDimitry Andric continue;
395ffd83dbSDimitry Andric }
405ffd83dbSDimitry Andric StringRef SectName(*SectNameOrErr);
415ffd83dbSDimitry Andric if (SectName != GNUBuildID)
425ffd83dbSDimitry Andric continue;
435ffd83dbSDimitry Andric StringRef BuildIDData;
445ffd83dbSDimitry Andric Expected<StringRef> E = Sect.getContents();
455ffd83dbSDimitry Andric if (E)
465ffd83dbSDimitry Andric BuildIDData = *E;
475ffd83dbSDimitry Andric else {
485ffd83dbSDimitry Andric consumeError(E.takeError());
495ffd83dbSDimitry Andric continue;
505ffd83dbSDimitry Andric }
515ffd83dbSDimitry Andric DataExtractor Decoder(BuildIDData, Obj.makeTriple().isLittleEndian(), 8);
525ffd83dbSDimitry Andric uint64_t Offset = 0;
535ffd83dbSDimitry Andric const uint32_t NameSize = Decoder.getU32(&Offset);
545ffd83dbSDimitry Andric const uint32_t PayloadSize = Decoder.getU32(&Offset);
555ffd83dbSDimitry Andric const uint32_t PayloadType = Decoder.getU32(&Offset);
565ffd83dbSDimitry Andric StringRef Name(Decoder.getFixedLengthString(&Offset, NameSize));
575ffd83dbSDimitry Andric if (Name == "GNU" && PayloadType == NT_GNU_BUILD_ID_TAG) {
585ffd83dbSDimitry Andric Offset = alignTo(Offset, 4);
595ffd83dbSDimitry Andric StringRef UUIDBytes(Decoder.getBytes(&Offset, PayloadSize));
605ffd83dbSDimitry Andric if (!UUIDBytes.empty()) {
615ffd83dbSDimitry Andric auto Ptr = reinterpret_cast<const uint8_t *>(UUIDBytes.data());
625ffd83dbSDimitry Andric UUID.assign(Ptr, Ptr + UUIDBytes.size());
635ffd83dbSDimitry Andric }
645ffd83dbSDimitry Andric }
655ffd83dbSDimitry Andric }
665ffd83dbSDimitry Andric }
675ffd83dbSDimitry Andric return UUID;
685ffd83dbSDimitry Andric }
695ffd83dbSDimitry Andric
convert(const object::ObjectFile & Obj,raw_ostream * Log,GsymCreator & Gsym)705ffd83dbSDimitry Andric llvm::Error ObjectFileTransformer::convert(const object::ObjectFile &Obj,
71*c9157d92SDimitry Andric raw_ostream *Log,
725ffd83dbSDimitry Andric GsymCreator &Gsym) {
735ffd83dbSDimitry Andric using namespace llvm::object;
745ffd83dbSDimitry Andric
755ffd83dbSDimitry Andric const bool IsMachO = isa<MachOObjectFile>(&Obj);
765ffd83dbSDimitry Andric const bool IsELF = isa<ELFObjectFileBase>(&Obj);
775ffd83dbSDimitry Andric
785ffd83dbSDimitry Andric // Read build ID.
795ffd83dbSDimitry Andric Gsym.setUUID(getUUID(Obj));
805ffd83dbSDimitry Andric
815ffd83dbSDimitry Andric // Parse the symbol table.
825ffd83dbSDimitry Andric size_t NumBefore = Gsym.getNumFunctionInfos();
835ffd83dbSDimitry Andric for (const object::SymbolRef &Sym : Obj.symbols()) {
845ffd83dbSDimitry Andric Expected<SymbolRef::Type> SymType = Sym.getType();
855ffd83dbSDimitry Andric if (!SymType) {
865ffd83dbSDimitry Andric consumeError(SymType.takeError());
875ffd83dbSDimitry Andric continue;
885ffd83dbSDimitry Andric }
895ffd83dbSDimitry Andric Expected<uint64_t> AddrOrErr = Sym.getValue();
905ffd83dbSDimitry Andric if (!AddrOrErr)
915ffd83dbSDimitry Andric // TODO: Test this error.
925ffd83dbSDimitry Andric return AddrOrErr.takeError();
935ffd83dbSDimitry Andric
945ffd83dbSDimitry Andric if (SymType.get() != SymbolRef::Type::ST_Function ||
95*c9157d92SDimitry Andric !Gsym.IsValidTextAddress(*AddrOrErr))
965ffd83dbSDimitry Andric continue;
975ffd83dbSDimitry Andric // Function size for MachO files will be 0
985ffd83dbSDimitry Andric constexpr bool NoCopy = false;
995ffd83dbSDimitry Andric const uint64_t size = IsELF ? ELFSymbolRef(Sym).getSize() : 0;
1005ffd83dbSDimitry Andric Expected<StringRef> Name = Sym.getName();
1015ffd83dbSDimitry Andric if (!Name) {
102*c9157d92SDimitry Andric if (Log)
103*c9157d92SDimitry Andric logAllUnhandledErrors(Name.takeError(), *Log,
104*c9157d92SDimitry Andric "ObjectFileTransformer: ");
105*c9157d92SDimitry Andric else
106*c9157d92SDimitry Andric consumeError(Name.takeError());
1075ffd83dbSDimitry Andric continue;
1085ffd83dbSDimitry Andric }
1095ffd83dbSDimitry Andric // Remove the leading '_' character in any symbol names if there is one
1105ffd83dbSDimitry Andric // for mach-o files.
1115ffd83dbSDimitry Andric if (IsMachO)
1125ffd83dbSDimitry Andric Name->consume_front("_");
1135ffd83dbSDimitry Andric Gsym.addFunctionInfo(
1145ffd83dbSDimitry Andric FunctionInfo(*AddrOrErr, size, Gsym.insertString(*Name, NoCopy)));
1155ffd83dbSDimitry Andric }
1165ffd83dbSDimitry Andric size_t FunctionsAddedCount = Gsym.getNumFunctionInfos() - NumBefore;
117*c9157d92SDimitry Andric if (Log)
118*c9157d92SDimitry Andric *Log << "Loaded " << FunctionsAddedCount
119*c9157d92SDimitry Andric << " functions from symbol table.\n";
1205ffd83dbSDimitry Andric return Error::success();
1215ffd83dbSDimitry Andric }
122