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 "Symbols.h"
260b57cec5SDimitry Andric #include "SyntheticSections.h"
270b57cec5SDimitry Andric #include "llvm/ADT/MapVector.h"
280b57cec5SDimitry Andric #include "llvm/ADT/SetVector.h"
2981ad6265SDimitry Andric #include "llvm/ADT/SmallPtrSet.h"
3081ad6265SDimitry Andric #include "llvm/Support/FileSystem.h"
315ffd83dbSDimitry Andric #include "llvm/Support/Parallel.h"
32e8d8bef9SDimitry Andric #include "llvm/Support/TimeProfiler.h"
330b57cec5SDimitry Andric #include "llvm/Support/raw_ostream.h"
340b57cec5SDimitry Andric
350b57cec5SDimitry Andric using namespace llvm;
360b57cec5SDimitry Andric using namespace llvm::object;
375ffd83dbSDimitry Andric using namespace lld;
385ffd83dbSDimitry Andric using namespace lld::elf;
390b57cec5SDimitry Andric
4004eeddc0SDimitry Andric using SymbolMapTy = DenseMap<const SectionBase *,
4104eeddc0SDimitry Andric SmallVector<std::pair<Defined *, uint64_t>, 0>>;
420b57cec5SDimitry Andric
4385868e8aSDimitry Andric static constexpr char indent8[] = " "; // 8 spaces
4485868e8aSDimitry Andric static constexpr char indent16[] = " "; // 16 spaces
450b57cec5SDimitry Andric
460b57cec5SDimitry Andric // Print out the first three columns of a line.
writeHeader(raw_ostream & os,uint64_t vma,uint64_t lma,uint64_t size,uint64_t align)470b57cec5SDimitry Andric static void writeHeader(raw_ostream &os, uint64_t vma, uint64_t lma,
480b57cec5SDimitry Andric uint64_t size, uint64_t align) {
490b57cec5SDimitry Andric if (config->is64)
500b57cec5SDimitry Andric os << format("%16llx %16llx %8llx %5lld ", vma, lma, size, align);
510b57cec5SDimitry Andric else
520b57cec5SDimitry Andric os << format("%8llx %8llx %8llx %5lld ", vma, lma, size, align);
530b57cec5SDimitry Andric }
540b57cec5SDimitry Andric
550b57cec5SDimitry Andric // Returns a list of all symbols that we want to print out.
getSymbols()560b57cec5SDimitry Andric static std::vector<Defined *> getSymbols() {
570b57cec5SDimitry Andric std::vector<Defined *> v;
58bdd1243dSDimitry Andric for (ELFFileBase *file : ctx.objectFiles)
590b57cec5SDimitry Andric for (Symbol *b : file->getSymbols())
600b57cec5SDimitry Andric if (auto *dr = dyn_cast<Defined>(b))
610b57cec5SDimitry Andric if (!dr->isSection() && dr->section && dr->section->isLive() &&
62bdd1243dSDimitry Andric (dr->file == file || dr->hasFlag(NEEDS_COPY) || dr->section->bss))
630b57cec5SDimitry Andric v.push_back(dr);
640b57cec5SDimitry Andric return v;
650b57cec5SDimitry Andric }
660b57cec5SDimitry Andric
670b57cec5SDimitry Andric // Returns a map from sections to their symbols.
getSectionSyms(ArrayRef<Defined * > syms)680b57cec5SDimitry Andric static SymbolMapTy getSectionSyms(ArrayRef<Defined *> syms) {
690b57cec5SDimitry Andric SymbolMapTy ret;
700b57cec5SDimitry Andric for (Defined *dr : syms)
7104eeddc0SDimitry Andric ret[dr->section].emplace_back(dr, dr->getVA());
720b57cec5SDimitry Andric
730b57cec5SDimitry Andric // Sort symbols by address. We want to print out symbols in the
740b57cec5SDimitry Andric // order in the output file rather than the order they appeared
750b57cec5SDimitry Andric // in the input files.
760eae32dcSDimitry Andric SmallPtrSet<Defined *, 4> set;
770eae32dcSDimitry Andric for (auto &it : ret) {
780eae32dcSDimitry Andric // Deduplicate symbols which need a canonical PLT entry/copy relocation.
790eae32dcSDimitry Andric set.clear();
8004eeddc0SDimitry Andric llvm::erase_if(it.second, [&](std::pair<Defined *, uint64_t> a) {
8104eeddc0SDimitry Andric return !set.insert(a.first).second;
820b57cec5SDimitry Andric });
8304eeddc0SDimitry Andric
8404eeddc0SDimitry Andric llvm::stable_sort(it.second, llvm::less_second());
850eae32dcSDimitry Andric }
860b57cec5SDimitry Andric return ret;
870b57cec5SDimitry Andric }
880b57cec5SDimitry Andric
890b57cec5SDimitry Andric // Construct a map from symbols to their stringified representations.
900b57cec5SDimitry Andric // Demangling symbols (which is what toString() does) is slow, so
910b57cec5SDimitry Andric // we do that in batch using parallel-for.
920b57cec5SDimitry Andric static DenseMap<Symbol *, std::string>
getSymbolStrings(ArrayRef<Defined * > syms)930b57cec5SDimitry Andric getSymbolStrings(ArrayRef<Defined *> syms) {
9404eeddc0SDimitry Andric auto strs = std::make_unique<std::string[]>(syms.size());
9581ad6265SDimitry Andric parallelFor(0, syms.size(), [&](size_t i) {
9604eeddc0SDimitry Andric raw_string_ostream os(strs[i]);
970b57cec5SDimitry Andric OutputSection *osec = syms[i]->getOutputSection();
980b57cec5SDimitry Andric uint64_t vma = syms[i]->getVA();
990b57cec5SDimitry Andric uint64_t lma = osec ? osec->getLMA() + vma - osec->getVA(0) : 0;
1000b57cec5SDimitry Andric writeHeader(os, vma, lma, syms[i]->getSize(), 1);
1010b57cec5SDimitry Andric os << indent16 << toString(*syms[i]);
1020b57cec5SDimitry Andric });
1030b57cec5SDimitry Andric
1040b57cec5SDimitry Andric DenseMap<Symbol *, std::string> ret;
1050b57cec5SDimitry Andric for (size_t i = 0, e = syms.size(); i < e; ++i)
10604eeddc0SDimitry Andric ret[syms[i]] = std::move(strs[i]);
1070b57cec5SDimitry Andric return ret;
1080b57cec5SDimitry Andric }
1090b57cec5SDimitry Andric
1100b57cec5SDimitry Andric // Print .eh_frame contents. Since the section consists of EhSectionPieces,
1110b57cec5SDimitry Andric // we need a specialized printer for that section.
1120b57cec5SDimitry Andric //
1130b57cec5SDimitry Andric // .eh_frame tend to contain a lot of section pieces that are contiguous
1140b57cec5SDimitry Andric // both in input file and output file. Such pieces are squashed before
1150b57cec5SDimitry Andric // being displayed to make output compact.
printEhFrame(raw_ostream & os,const EhFrameSection * sec)1160b57cec5SDimitry Andric static void printEhFrame(raw_ostream &os, const EhFrameSection *sec) {
1170b57cec5SDimitry Andric std::vector<EhSectionPiece> pieces;
1180b57cec5SDimitry Andric
1190b57cec5SDimitry Andric auto add = [&](const EhSectionPiece &p) {
1200b57cec5SDimitry Andric // If P is adjacent to Last, squash the two.
1210b57cec5SDimitry Andric if (!pieces.empty()) {
1220b57cec5SDimitry Andric EhSectionPiece &last = pieces.back();
1230b57cec5SDimitry Andric if (last.sec == p.sec && last.inputOff + last.size == p.inputOff &&
124*0fca6ea1SDimitry Andric last.outputOff + last.size == (unsigned)p.outputOff) {
1250b57cec5SDimitry Andric last.size += p.size;
1260b57cec5SDimitry Andric return;
1270b57cec5SDimitry Andric }
1280b57cec5SDimitry Andric }
1290b57cec5SDimitry Andric pieces.push_back(p);
1300b57cec5SDimitry Andric };
1310b57cec5SDimitry Andric
1320b57cec5SDimitry Andric // Gather section pieces.
1330b57cec5SDimitry Andric for (const CieRecord *rec : sec->getCieRecords()) {
1340b57cec5SDimitry Andric add(*rec->cie);
1350b57cec5SDimitry Andric for (const EhSectionPiece *fde : rec->fdes)
1360b57cec5SDimitry Andric add(*fde);
1370b57cec5SDimitry Andric }
1380b57cec5SDimitry Andric
1390b57cec5SDimitry Andric // Print out section pieces.
1400b57cec5SDimitry Andric const OutputSection *osec = sec->getOutputSection();
1410b57cec5SDimitry Andric for (EhSectionPiece &p : pieces) {
1420b57cec5SDimitry Andric writeHeader(os, osec->addr + p.outputOff, osec->getLMA() + p.outputOff,
1430b57cec5SDimitry Andric p.size, 1);
1440b57cec5SDimitry Andric os << indent8 << toString(p.sec->file) << ":(" << p.sec->name << "+0x"
1450b57cec5SDimitry Andric << Twine::utohexstr(p.inputOff) + ")\n";
1460b57cec5SDimitry Andric }
1470b57cec5SDimitry Andric }
1480b57cec5SDimitry Andric
writeMapFile(raw_fd_ostream & os)1494824e7fdSDimitry Andric static void writeMapFile(raw_fd_ostream &os) {
1500b57cec5SDimitry Andric // Collect symbol info that we want to print out.
1510b57cec5SDimitry Andric std::vector<Defined *> syms = getSymbols();
1520b57cec5SDimitry Andric SymbolMapTy sectionSyms = getSectionSyms(syms);
1530b57cec5SDimitry Andric DenseMap<Symbol *, std::string> symStr = getSymbolStrings(syms);
1540b57cec5SDimitry Andric
1550b57cec5SDimitry Andric // Print out the header line.
1560b57cec5SDimitry Andric int w = config->is64 ? 16 : 8;
1570b57cec5SDimitry Andric os << right_justify("VMA", w) << ' ' << right_justify("LMA", w)
1580b57cec5SDimitry Andric << " Size Align Out In Symbol\n";
1590b57cec5SDimitry Andric
1600b57cec5SDimitry Andric OutputSection *osec = nullptr;
1614824e7fdSDimitry Andric for (SectionCommand *cmd : script->sectionCommands) {
1624824e7fdSDimitry Andric if (auto *assign = dyn_cast<SymbolAssignment>(cmd)) {
1634824e7fdSDimitry Andric if (assign->provide && !assign->sym)
1640b57cec5SDimitry Andric continue;
1654824e7fdSDimitry Andric uint64_t lma = osec ? osec->getLMA() + assign->addr - osec->getVA(0) : 0;
1664824e7fdSDimitry Andric writeHeader(os, assign->addr, lma, assign->size, 1);
1674824e7fdSDimitry Andric os << assign->commandString << '\n';
1680b57cec5SDimitry Andric continue;
1690b57cec5SDimitry Andric }
1700b57cec5SDimitry Andric
17181ad6265SDimitry Andric osec = &cast<OutputDesc>(cmd)->osec;
172bdd1243dSDimitry Andric writeHeader(os, osec->addr, osec->getLMA(), osec->size, osec->addralign);
1730b57cec5SDimitry Andric os << osec->name << '\n';
1740b57cec5SDimitry Andric
1750b57cec5SDimitry Andric // Dump symbols for each input section.
1764824e7fdSDimitry Andric for (SectionCommand *subCmd : osec->commands) {
1774824e7fdSDimitry Andric if (auto *isd = dyn_cast<InputSectionDescription>(subCmd)) {
1780b57cec5SDimitry Andric for (InputSection *isec : isd->sections) {
1790b57cec5SDimitry Andric if (auto *ehSec = dyn_cast<EhFrameSection>(isec)) {
1800b57cec5SDimitry Andric printEhFrame(os, ehSec);
1810b57cec5SDimitry Andric continue;
1820b57cec5SDimitry Andric }
1830b57cec5SDimitry Andric
1844824e7fdSDimitry Andric writeHeader(os, isec->getVA(), osec->getLMA() + isec->outSecOff,
185bdd1243dSDimitry Andric isec->getSize(), isec->addralign);
1860b57cec5SDimitry Andric os << indent8 << toString(isec) << '\n';
18704eeddc0SDimitry Andric for (Symbol *sym : llvm::make_first_range(sectionSyms[isec]))
1880b57cec5SDimitry Andric os << symStr[sym] << '\n';
1890b57cec5SDimitry Andric }
1900b57cec5SDimitry Andric continue;
1910b57cec5SDimitry Andric }
1920b57cec5SDimitry Andric
1934824e7fdSDimitry Andric if (auto *data = dyn_cast<ByteCommand>(subCmd)) {
1944824e7fdSDimitry Andric writeHeader(os, osec->addr + data->offset,
1954824e7fdSDimitry Andric osec->getLMA() + data->offset, data->size, 1);
1964824e7fdSDimitry Andric os << indent8 << data->commandString << '\n';
1970b57cec5SDimitry Andric continue;
1980b57cec5SDimitry Andric }
1990b57cec5SDimitry Andric
2004824e7fdSDimitry Andric if (auto *assign = dyn_cast<SymbolAssignment>(subCmd)) {
2014824e7fdSDimitry Andric if (assign->provide && !assign->sym)
2020b57cec5SDimitry Andric continue;
2034824e7fdSDimitry Andric writeHeader(os, assign->addr,
2044824e7fdSDimitry Andric osec->getLMA() + assign->addr - osec->getVA(0),
2054824e7fdSDimitry Andric assign->size, 1);
2064824e7fdSDimitry Andric os << indent8 << assign->commandString << '\n';
2070b57cec5SDimitry Andric continue;
2080b57cec5SDimitry Andric }
2090b57cec5SDimitry Andric }
2100b57cec5SDimitry Andric }
2110b57cec5SDimitry Andric }
2120b57cec5SDimitry Andric
2130b57cec5SDimitry Andric // Output a cross reference table to stdout. This is for --cref.
2140b57cec5SDimitry Andric //
2150b57cec5SDimitry Andric // For each global symbol, we print out a file that defines the symbol
2160b57cec5SDimitry Andric // followed by files that uses that symbol. Here is an example.
2170b57cec5SDimitry Andric //
2180b57cec5SDimitry Andric // strlen /lib/x86_64-linux-gnu/libc.so.6
2190b57cec5SDimitry Andric // tools/lld/tools/lld/CMakeFiles/lld.dir/lld.cpp.o
2200b57cec5SDimitry Andric // lib/libLLVMSupport.a(PrettyStackTrace.cpp.o)
2210b57cec5SDimitry Andric //
2220b57cec5SDimitry Andric // In this case, strlen is defined by libc.so.6 and used by other two
2230b57cec5SDimitry Andric // files.
writeCref(raw_fd_ostream & os)2244824e7fdSDimitry Andric static void writeCref(raw_fd_ostream &os) {
2250b57cec5SDimitry Andric // Collect symbols and files.
2260b57cec5SDimitry Andric MapVector<Symbol *, SetVector<InputFile *>> map;
227bdd1243dSDimitry Andric for (ELFFileBase *file : ctx.objectFiles) {
2280b57cec5SDimitry Andric for (Symbol *sym : file->getSymbols()) {
2290b57cec5SDimitry Andric if (isa<SharedSymbol>(sym))
2300b57cec5SDimitry Andric map[sym].insert(file);
2310b57cec5SDimitry Andric if (auto *d = dyn_cast<Defined>(sym))
2325f757f3fSDimitry Andric if (!d->isLocal())
2330b57cec5SDimitry Andric map[d].insert(file);
2340b57cec5SDimitry Andric }
2350b57cec5SDimitry Andric }
2360b57cec5SDimitry Andric
2374824e7fdSDimitry Andric auto print = [&](StringRef a, StringRef b) {
2384824e7fdSDimitry Andric os << left_justify(a, 49) << ' ' << b << '\n';
2394824e7fdSDimitry Andric };
2404824e7fdSDimitry Andric
2414824e7fdSDimitry Andric // Print a blank line and a header. The format matches GNU ld.
2424824e7fdSDimitry Andric os << "\nCross Reference Table\n\n";
2430b57cec5SDimitry Andric print("Symbol", "File");
2440b57cec5SDimitry Andric
2450b57cec5SDimitry Andric // Print out a table.
2460b57cec5SDimitry Andric for (auto kv : map) {
2470b57cec5SDimitry Andric Symbol *sym = kv.first;
2480b57cec5SDimitry Andric SetVector<InputFile *> &files = kv.second;
2490b57cec5SDimitry Andric
2500b57cec5SDimitry Andric print(toString(*sym), toString(sym->file));
2510b57cec5SDimitry Andric for (InputFile *file : files)
2520b57cec5SDimitry Andric if (file != sym->file)
2530b57cec5SDimitry Andric print("", toString(file));
2540b57cec5SDimitry Andric }
2550b57cec5SDimitry Andric }
25685868e8aSDimitry Andric
writeMapAndCref()2574824e7fdSDimitry Andric void elf::writeMapAndCref() {
2584824e7fdSDimitry Andric if (config->mapFile.empty() && !config->cref)
2594824e7fdSDimitry Andric return;
2604824e7fdSDimitry Andric
2614824e7fdSDimitry Andric llvm::TimeTraceScope timeScope("Write map file");
2624824e7fdSDimitry Andric
2634824e7fdSDimitry Andric // Open a map file for writing.
2644824e7fdSDimitry Andric std::error_code ec;
2654824e7fdSDimitry Andric StringRef mapFile = config->mapFile.empty() ? "-" : config->mapFile;
26606c3fb27SDimitry Andric raw_fd_ostream os = ctx.openAuxiliaryFile(mapFile, ec);
2674824e7fdSDimitry Andric if (ec) {
2684824e7fdSDimitry Andric error("cannot open " + mapFile + ": " + ec.message());
2694824e7fdSDimitry Andric return;
2704824e7fdSDimitry Andric }
2714824e7fdSDimitry Andric
2724824e7fdSDimitry Andric if (!config->mapFile.empty())
2734824e7fdSDimitry Andric writeMapFile(os);
2744824e7fdSDimitry Andric if (config->cref)
2754824e7fdSDimitry Andric writeCref(os);
2764824e7fdSDimitry Andric }
277