10b57cec5SDimitry Andric //===- MapFile.cpp --------------------------------------------------------===// 20b57cec5SDimitry Andric // 30b57cec5SDimitry Andric // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. 40b57cec5SDimitry Andric // See https://llvm.org/LICENSE.txt for license information. 50b57cec5SDimitry Andric // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception 60b57cec5SDimitry Andric // 70b57cec5SDimitry Andric //===----------------------------------------------------------------------===// 80b57cec5SDimitry Andric // 90b57cec5SDimitry Andric // This file implements the -Map option. It shows lists in order and 100b57cec5SDimitry Andric // hierarchically the output sections, input sections, input files and 110b57cec5SDimitry Andric // symbol: 120b57cec5SDimitry Andric // 130b57cec5SDimitry Andric // Address Size Align Out In Symbol 140b57cec5SDimitry Andric // 00201000 00000015 4 .text 150b57cec5SDimitry Andric // 00201000 0000000e 4 test.o:(.text) 160b57cec5SDimitry Andric // 0020100e 00000000 0 local 170b57cec5SDimitry Andric // 00201005 00000000 0 f(int) 180b57cec5SDimitry Andric // 190b57cec5SDimitry Andric //===----------------------------------------------------------------------===// 200b57cec5SDimitry Andric 210b57cec5SDimitry Andric #include "MapFile.h" 220b57cec5SDimitry Andric #include "InputFiles.h" 230b57cec5SDimitry Andric #include "LinkerScript.h" 240b57cec5SDimitry Andric #include "OutputSections.h" 250b57cec5SDimitry Andric #include "SymbolTable.h" 260b57cec5SDimitry Andric #include "Symbols.h" 270b57cec5SDimitry Andric #include "SyntheticSections.h" 280b57cec5SDimitry Andric #include "lld/Common/Strings.h" 290b57cec5SDimitry Andric #include "llvm/ADT/MapVector.h" 300b57cec5SDimitry Andric #include "llvm/ADT/SetVector.h" 31*5ffd83dbSDimitry Andric #include "llvm/Support/Parallel.h" 320b57cec5SDimitry Andric #include "llvm/Support/raw_ostream.h" 330b57cec5SDimitry Andric 340b57cec5SDimitry Andric using namespace llvm; 350b57cec5SDimitry Andric using namespace llvm::object; 36*5ffd83dbSDimitry Andric using namespace lld; 37*5ffd83dbSDimitry Andric using namespace lld::elf; 380b57cec5SDimitry Andric 390b57cec5SDimitry Andric using SymbolMapTy = DenseMap<const SectionBase *, SmallVector<Defined *, 4>>; 400b57cec5SDimitry Andric 4185868e8aSDimitry Andric static constexpr char indent8[] = " "; // 8 spaces 4285868e8aSDimitry Andric static constexpr char indent16[] = " "; // 16 spaces 430b57cec5SDimitry Andric 440b57cec5SDimitry Andric // Print out the first three columns of a line. 450b57cec5SDimitry Andric static void writeHeader(raw_ostream &os, uint64_t vma, uint64_t lma, 460b57cec5SDimitry Andric uint64_t size, uint64_t align) { 470b57cec5SDimitry Andric if (config->is64) 480b57cec5SDimitry Andric os << format("%16llx %16llx %8llx %5lld ", vma, lma, size, align); 490b57cec5SDimitry Andric else 500b57cec5SDimitry Andric os << format("%8llx %8llx %8llx %5lld ", vma, lma, size, align); 510b57cec5SDimitry Andric } 520b57cec5SDimitry Andric 530b57cec5SDimitry Andric // Returns a list of all symbols that we want to print out. 540b57cec5SDimitry Andric static std::vector<Defined *> getSymbols() { 550b57cec5SDimitry Andric std::vector<Defined *> v; 560b57cec5SDimitry Andric for (InputFile *file : objectFiles) 570b57cec5SDimitry Andric for (Symbol *b : file->getSymbols()) 580b57cec5SDimitry Andric if (auto *dr = dyn_cast<Defined>(b)) 590b57cec5SDimitry Andric if (!dr->isSection() && dr->section && dr->section->isLive() && 600b57cec5SDimitry Andric (dr->file == file || dr->needsPltAddr || dr->section->bss)) 610b57cec5SDimitry Andric v.push_back(dr); 620b57cec5SDimitry Andric return v; 630b57cec5SDimitry Andric } 640b57cec5SDimitry Andric 650b57cec5SDimitry Andric // Returns a map from sections to their symbols. 660b57cec5SDimitry Andric static SymbolMapTy getSectionSyms(ArrayRef<Defined *> syms) { 670b57cec5SDimitry Andric SymbolMapTy ret; 680b57cec5SDimitry Andric for (Defined *dr : syms) 690b57cec5SDimitry Andric ret[dr->section].push_back(dr); 700b57cec5SDimitry Andric 710b57cec5SDimitry Andric // Sort symbols by address. We want to print out symbols in the 720b57cec5SDimitry Andric // order in the output file rather than the order they appeared 730b57cec5SDimitry Andric // in the input files. 740b57cec5SDimitry Andric for (auto &it : ret) 750b57cec5SDimitry Andric llvm::stable_sort(it.second, [](Defined *a, Defined *b) { 760b57cec5SDimitry Andric return a->getVA() < b->getVA(); 770b57cec5SDimitry Andric }); 780b57cec5SDimitry Andric return ret; 790b57cec5SDimitry Andric } 800b57cec5SDimitry Andric 810b57cec5SDimitry Andric // Construct a map from symbols to their stringified representations. 820b57cec5SDimitry Andric // Demangling symbols (which is what toString() does) is slow, so 830b57cec5SDimitry Andric // we do that in batch using parallel-for. 840b57cec5SDimitry Andric static DenseMap<Symbol *, std::string> 850b57cec5SDimitry Andric getSymbolStrings(ArrayRef<Defined *> syms) { 860b57cec5SDimitry Andric std::vector<std::string> str(syms.size()); 870b57cec5SDimitry Andric parallelForEachN(0, syms.size(), [&](size_t i) { 880b57cec5SDimitry Andric raw_string_ostream os(str[i]); 890b57cec5SDimitry Andric OutputSection *osec = syms[i]->getOutputSection(); 900b57cec5SDimitry Andric uint64_t vma = syms[i]->getVA(); 910b57cec5SDimitry Andric uint64_t lma = osec ? osec->getLMA() + vma - osec->getVA(0) : 0; 920b57cec5SDimitry Andric writeHeader(os, vma, lma, syms[i]->getSize(), 1); 930b57cec5SDimitry Andric os << indent16 << toString(*syms[i]); 940b57cec5SDimitry Andric }); 950b57cec5SDimitry Andric 960b57cec5SDimitry Andric DenseMap<Symbol *, std::string> ret; 970b57cec5SDimitry Andric for (size_t i = 0, e = syms.size(); i < e; ++i) 980b57cec5SDimitry Andric ret[syms[i]] = std::move(str[i]); 990b57cec5SDimitry Andric return ret; 1000b57cec5SDimitry Andric } 1010b57cec5SDimitry Andric 1020b57cec5SDimitry Andric // Print .eh_frame contents. Since the section consists of EhSectionPieces, 1030b57cec5SDimitry Andric // we need a specialized printer for that section. 1040b57cec5SDimitry Andric // 1050b57cec5SDimitry Andric // .eh_frame tend to contain a lot of section pieces that are contiguous 1060b57cec5SDimitry Andric // both in input file and output file. Such pieces are squashed before 1070b57cec5SDimitry Andric // being displayed to make output compact. 1080b57cec5SDimitry Andric static void printEhFrame(raw_ostream &os, const EhFrameSection *sec) { 1090b57cec5SDimitry Andric std::vector<EhSectionPiece> pieces; 1100b57cec5SDimitry Andric 1110b57cec5SDimitry Andric auto add = [&](const EhSectionPiece &p) { 1120b57cec5SDimitry Andric // If P is adjacent to Last, squash the two. 1130b57cec5SDimitry Andric if (!pieces.empty()) { 1140b57cec5SDimitry Andric EhSectionPiece &last = pieces.back(); 1150b57cec5SDimitry Andric if (last.sec == p.sec && last.inputOff + last.size == p.inputOff && 1160b57cec5SDimitry Andric last.outputOff + last.size == p.outputOff) { 1170b57cec5SDimitry Andric last.size += p.size; 1180b57cec5SDimitry Andric return; 1190b57cec5SDimitry Andric } 1200b57cec5SDimitry Andric } 1210b57cec5SDimitry Andric pieces.push_back(p); 1220b57cec5SDimitry Andric }; 1230b57cec5SDimitry Andric 1240b57cec5SDimitry Andric // Gather section pieces. 1250b57cec5SDimitry Andric for (const CieRecord *rec : sec->getCieRecords()) { 1260b57cec5SDimitry Andric add(*rec->cie); 1270b57cec5SDimitry Andric for (const EhSectionPiece *fde : rec->fdes) 1280b57cec5SDimitry Andric add(*fde); 1290b57cec5SDimitry Andric } 1300b57cec5SDimitry Andric 1310b57cec5SDimitry Andric // Print out section pieces. 1320b57cec5SDimitry Andric const OutputSection *osec = sec->getOutputSection(); 1330b57cec5SDimitry Andric for (EhSectionPiece &p : pieces) { 1340b57cec5SDimitry Andric writeHeader(os, osec->addr + p.outputOff, osec->getLMA() + p.outputOff, 1350b57cec5SDimitry Andric p.size, 1); 1360b57cec5SDimitry Andric os << indent8 << toString(p.sec->file) << ":(" << p.sec->name << "+0x" 1370b57cec5SDimitry Andric << Twine::utohexstr(p.inputOff) + ")\n"; 1380b57cec5SDimitry Andric } 1390b57cec5SDimitry Andric } 1400b57cec5SDimitry Andric 141*5ffd83dbSDimitry Andric void elf::writeMapFile() { 1420b57cec5SDimitry Andric if (config->mapFile.empty()) 1430b57cec5SDimitry Andric return; 1440b57cec5SDimitry Andric 1450b57cec5SDimitry Andric // Open a map file for writing. 1460b57cec5SDimitry Andric std::error_code ec; 14785868e8aSDimitry Andric raw_fd_ostream os(config->mapFile, ec, sys::fs::OF_None); 1480b57cec5SDimitry Andric if (ec) { 1490b57cec5SDimitry Andric error("cannot open " + config->mapFile + ": " + ec.message()); 1500b57cec5SDimitry Andric return; 1510b57cec5SDimitry Andric } 1520b57cec5SDimitry Andric 1530b57cec5SDimitry Andric // Collect symbol info that we want to print out. 1540b57cec5SDimitry Andric std::vector<Defined *> syms = getSymbols(); 1550b57cec5SDimitry Andric SymbolMapTy sectionSyms = getSectionSyms(syms); 1560b57cec5SDimitry Andric DenseMap<Symbol *, std::string> symStr = getSymbolStrings(syms); 1570b57cec5SDimitry Andric 1580b57cec5SDimitry Andric // Print out the header line. 1590b57cec5SDimitry Andric int w = config->is64 ? 16 : 8; 1600b57cec5SDimitry Andric os << right_justify("VMA", w) << ' ' << right_justify("LMA", w) 1610b57cec5SDimitry Andric << " Size Align Out In Symbol\n"; 1620b57cec5SDimitry Andric 1630b57cec5SDimitry Andric OutputSection* osec = nullptr; 1640b57cec5SDimitry Andric for (BaseCommand *base : script->sectionCommands) { 1650b57cec5SDimitry Andric if (auto *cmd = dyn_cast<SymbolAssignment>(base)) { 1660b57cec5SDimitry Andric if (cmd->provide && !cmd->sym) 1670b57cec5SDimitry Andric continue; 1680b57cec5SDimitry Andric uint64_t lma = osec ? osec->getLMA() + cmd->addr - osec->getVA(0) : 0; 1690b57cec5SDimitry Andric writeHeader(os, cmd->addr, lma, cmd->size, 1); 1700b57cec5SDimitry Andric os << cmd->commandString << '\n'; 1710b57cec5SDimitry Andric continue; 1720b57cec5SDimitry Andric } 1730b57cec5SDimitry Andric 1740b57cec5SDimitry Andric osec = cast<OutputSection>(base); 1750b57cec5SDimitry Andric writeHeader(os, osec->addr, osec->getLMA(), osec->size, osec->alignment); 1760b57cec5SDimitry Andric os << osec->name << '\n'; 1770b57cec5SDimitry Andric 1780b57cec5SDimitry Andric // Dump symbols for each input section. 1790b57cec5SDimitry Andric for (BaseCommand *base : osec->sectionCommands) { 1800b57cec5SDimitry Andric if (auto *isd = dyn_cast<InputSectionDescription>(base)) { 1810b57cec5SDimitry Andric for (InputSection *isec : isd->sections) { 1820b57cec5SDimitry Andric if (auto *ehSec = dyn_cast<EhFrameSection>(isec)) { 1830b57cec5SDimitry Andric printEhFrame(os, ehSec); 1840b57cec5SDimitry Andric continue; 1850b57cec5SDimitry Andric } 1860b57cec5SDimitry Andric 1870b57cec5SDimitry Andric writeHeader(os, isec->getVA(0), osec->getLMA() + isec->getOffset(0), 1880b57cec5SDimitry Andric isec->getSize(), isec->alignment); 1890b57cec5SDimitry Andric os << indent8 << toString(isec) << '\n'; 1900b57cec5SDimitry Andric for (Symbol *sym : sectionSyms[isec]) 1910b57cec5SDimitry Andric os << symStr[sym] << '\n'; 1920b57cec5SDimitry Andric } 1930b57cec5SDimitry Andric continue; 1940b57cec5SDimitry Andric } 1950b57cec5SDimitry Andric 1960b57cec5SDimitry Andric if (auto *cmd = dyn_cast<ByteCommand>(base)) { 1970b57cec5SDimitry Andric writeHeader(os, osec->addr + cmd->offset, osec->getLMA() + cmd->offset, 1980b57cec5SDimitry Andric cmd->size, 1); 1990b57cec5SDimitry Andric os << indent8 << cmd->commandString << '\n'; 2000b57cec5SDimitry Andric continue; 2010b57cec5SDimitry Andric } 2020b57cec5SDimitry Andric 2030b57cec5SDimitry Andric if (auto *cmd = dyn_cast<SymbolAssignment>(base)) { 2040b57cec5SDimitry Andric if (cmd->provide && !cmd->sym) 2050b57cec5SDimitry Andric continue; 2060b57cec5SDimitry Andric writeHeader(os, cmd->addr, osec->getLMA() + cmd->addr - osec->getVA(0), 2070b57cec5SDimitry Andric cmd->size, 1); 2080b57cec5SDimitry Andric os << indent8 << cmd->commandString << '\n'; 2090b57cec5SDimitry Andric continue; 2100b57cec5SDimitry Andric } 2110b57cec5SDimitry Andric } 2120b57cec5SDimitry Andric } 2130b57cec5SDimitry Andric } 2140b57cec5SDimitry Andric 2150b57cec5SDimitry Andric static void print(StringRef a, StringRef b) { 216480093f4SDimitry Andric lld::outs() << left_justify(a, 49) << " " << b << "\n"; 2170b57cec5SDimitry Andric } 2180b57cec5SDimitry Andric 2190b57cec5SDimitry Andric // Output a cross reference table to stdout. This is for --cref. 2200b57cec5SDimitry Andric // 2210b57cec5SDimitry Andric // For each global symbol, we print out a file that defines the symbol 2220b57cec5SDimitry Andric // followed by files that uses that symbol. Here is an example. 2230b57cec5SDimitry Andric // 2240b57cec5SDimitry Andric // strlen /lib/x86_64-linux-gnu/libc.so.6 2250b57cec5SDimitry Andric // tools/lld/tools/lld/CMakeFiles/lld.dir/lld.cpp.o 2260b57cec5SDimitry Andric // lib/libLLVMSupport.a(PrettyStackTrace.cpp.o) 2270b57cec5SDimitry Andric // 2280b57cec5SDimitry Andric // In this case, strlen is defined by libc.so.6 and used by other two 2290b57cec5SDimitry Andric // files. 230*5ffd83dbSDimitry Andric void elf::writeCrossReferenceTable() { 2310b57cec5SDimitry Andric if (!config->cref) 2320b57cec5SDimitry Andric return; 2330b57cec5SDimitry Andric 2340b57cec5SDimitry Andric // Collect symbols and files. 2350b57cec5SDimitry Andric MapVector<Symbol *, SetVector<InputFile *>> map; 2360b57cec5SDimitry Andric for (InputFile *file : objectFiles) { 2370b57cec5SDimitry Andric for (Symbol *sym : file->getSymbols()) { 2380b57cec5SDimitry Andric if (isa<SharedSymbol>(sym)) 2390b57cec5SDimitry Andric map[sym].insert(file); 2400b57cec5SDimitry Andric if (auto *d = dyn_cast<Defined>(sym)) 2410b57cec5SDimitry Andric if (!d->isLocal() && (!d->section || d->section->isLive())) 2420b57cec5SDimitry Andric map[d].insert(file); 2430b57cec5SDimitry Andric } 2440b57cec5SDimitry Andric } 2450b57cec5SDimitry Andric 2460b57cec5SDimitry Andric // Print out a header. 247480093f4SDimitry Andric lld::outs() << "Cross Reference Table\n\n"; 2480b57cec5SDimitry Andric print("Symbol", "File"); 2490b57cec5SDimitry Andric 2500b57cec5SDimitry Andric // Print out a table. 2510b57cec5SDimitry Andric for (auto kv : map) { 2520b57cec5SDimitry Andric Symbol *sym = kv.first; 2530b57cec5SDimitry Andric SetVector<InputFile *> &files = kv.second; 2540b57cec5SDimitry Andric 2550b57cec5SDimitry Andric print(toString(*sym), toString(sym->file)); 2560b57cec5SDimitry Andric for (InputFile *file : files) 2570b57cec5SDimitry Andric if (file != sym->file) 2580b57cec5SDimitry Andric print("", toString(file)); 2590b57cec5SDimitry Andric } 2600b57cec5SDimitry Andric } 26185868e8aSDimitry Andric 262*5ffd83dbSDimitry Andric void elf::writeArchiveStats() { 263*5ffd83dbSDimitry Andric if (config->printArchiveStats.empty()) 264*5ffd83dbSDimitry Andric return; 265*5ffd83dbSDimitry Andric 266*5ffd83dbSDimitry Andric std::error_code ec; 267*5ffd83dbSDimitry Andric raw_fd_ostream os(config->printArchiveStats, ec, sys::fs::OF_None); 268*5ffd83dbSDimitry Andric if (ec) { 269*5ffd83dbSDimitry Andric error("--print-archive-stats=: cannot open " + config->printArchiveStats + 270*5ffd83dbSDimitry Andric ": " + ec.message()); 271*5ffd83dbSDimitry Andric return; 272*5ffd83dbSDimitry Andric } 273*5ffd83dbSDimitry Andric 274*5ffd83dbSDimitry Andric os << "members\tfetched\tarchive\n"; 275*5ffd83dbSDimitry Andric for (const ArchiveFile *f : archiveFiles) 276*5ffd83dbSDimitry Andric os << f->getMemberCount() << '\t' << f->getFetchedMemberCount() << '\t' 277*5ffd83dbSDimitry Andric << f->getName() << '\n'; 278*5ffd83dbSDimitry Andric } 279