1317025Sdim//===- MapFile.cpp --------------------------------------------------------===//
2317025Sdim//
3353358Sdim// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4353358Sdim// See https://llvm.org/LICENSE.txt for license information.
5353358Sdim// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6317025Sdim//
7317025Sdim//===----------------------------------------------------------------------===//
8317025Sdim//
9317025Sdim// This file implements the -Map option. It shows lists in order and
10317025Sdim// hierarchically the output sections, input sections, input files and
11317025Sdim// symbol:
12317025Sdim//
13317778Sdim//   Address  Size     Align Out     In      Symbol
14317778Sdim//   00201000 00000015     4 .text
15317778Sdim//   00201000 0000000e     4         test.o:(.text)
16317778Sdim//   0020100e 00000000     0                 local
17317778Sdim//   00201005 00000000     0                 f(int)
18317025Sdim//
19317025Sdim//===----------------------------------------------------------------------===//
20317025Sdim
21317025Sdim#include "MapFile.h"
22317025Sdim#include "InputFiles.h"
23318681Sdim#include "LinkerScript.h"
24318681Sdim#include "OutputSections.h"
25317778Sdim#include "SymbolTable.h"
26327952Sdim#include "Symbols.h"
27327952Sdim#include "SyntheticSections.h"
28341825Sdim#include "lld/Common/Strings.h"
29327952Sdim#include "lld/Common/Threads.h"
30341825Sdim#include "llvm/ADT/MapVector.h"
31341825Sdim#include "llvm/ADT/SetVector.h"
32317025Sdim#include "llvm/Support/raw_ostream.h"
33317025Sdim
34317025Sdimusing namespace llvm;
35317025Sdimusing namespace llvm::object;
36317025Sdim
37360784Sdimnamespace lld {
38360784Sdimnamespace elf {
39353358Sdimusing SymbolMapTy = DenseMap<const SectionBase *, SmallVector<Defined *, 4>>;
40317025Sdim
41360784Sdimstatic constexpr char indent8[] = "        ";          // 8 spaces
42360784Sdimstatic constexpr char indent16[] = "                "; // 16 spaces
43341825Sdim
44317778Sdim// Print out the first three columns of a line.
45353358Sdimstatic void writeHeader(raw_ostream &os, uint64_t vma, uint64_t lma,
46353358Sdim                        uint64_t size, uint64_t align) {
47353358Sdim  if (config->is64)
48353358Sdim    os << format("%16llx %16llx %8llx %5lld ", vma, lma, size, align);
49341825Sdim  else
50353358Sdim    os << format("%8llx %8llx %8llx %5lld ", vma, lma, size, align);
51317025Sdim}
52317025Sdim
53317778Sdim// Returns a list of all symbols that we want to print out.
54341825Sdimstatic std::vector<Defined *> getSymbols() {
55353358Sdim  std::vector<Defined *> v;
56353358Sdim  for (InputFile *file : objectFiles)
57353358Sdim    for (Symbol *b : file->getSymbols())
58353358Sdim      if (auto *dr = dyn_cast<Defined>(b))
59353358Sdim        if (!dr->isSection() && dr->section && dr->section->isLive() &&
60353358Sdim            (dr->file == file || dr->needsPltAddr || dr->section->bss))
61353358Sdim          v.push_back(dr);
62353358Sdim  return v;
63317025Sdim}
64317025Sdim
65317778Sdim// Returns a map from sections to their symbols.
66353358Sdimstatic SymbolMapTy getSectionSyms(ArrayRef<Defined *> syms) {
67353358Sdim  SymbolMapTy ret;
68353358Sdim  for (Defined *dr : syms)
69353358Sdim    ret[dr->section].push_back(dr);
70317025Sdim
71317778Sdim  // Sort symbols by address. We want to print out symbols in the
72317778Sdim  // order in the output file rather than the order they appeared
73317778Sdim  // in the input files.
74353358Sdim  for (auto &it : ret)
75353358Sdim    llvm::stable_sort(it.second, [](Defined *a, Defined *b) {
76353358Sdim      return a->getVA() < b->getVA();
77341825Sdim    });
78353358Sdim  return ret;
79317025Sdim}
80317025Sdim
81317778Sdim// Construct a map from symbols to their stringified representations.
82317778Sdim// Demangling symbols (which is what toString() does) is slow, so
83317778Sdim// we do that in batch using parallel-for.
84327952Sdimstatic DenseMap<Symbol *, std::string>
85353358SdimgetSymbolStrings(ArrayRef<Defined *> syms) {
86353358Sdim  std::vector<std::string> str(syms.size());
87353358Sdim  parallelForEachN(0, syms.size(), [&](size_t i) {
88353358Sdim    raw_string_ostream os(str[i]);
89353358Sdim    OutputSection *osec = syms[i]->getOutputSection();
90353358Sdim    uint64_t vma = syms[i]->getVA();
91353358Sdim    uint64_t lma = osec ? osec->getLMA() + vma - osec->getVA(0) : 0;
92353358Sdim    writeHeader(os, vma, lma, syms[i]->getSize(), 1);
93353358Sdim    os << indent16 << toString(*syms[i]);
94317778Sdim  });
95317025Sdim
96353358Sdim  DenseMap<Symbol *, std::string> ret;
97353358Sdim  for (size_t i = 0, e = syms.size(); i < e; ++i)
98353358Sdim    ret[syms[i]] = std::move(str[i]);
99353358Sdim  return ret;
100317025Sdim}
101317025Sdim
102341825Sdim// Print .eh_frame contents. Since the section consists of EhSectionPieces,
103341825Sdim// we need a specialized printer for that section.
104341825Sdim//
105341825Sdim// .eh_frame tend to contain a lot of section pieces that are contiguous
106341825Sdim// both in input file and output file. Such pieces are squashed before
107341825Sdim// being displayed to make output compact.
108353358Sdimstatic void printEhFrame(raw_ostream &os, const EhFrameSection *sec) {
109353358Sdim  std::vector<EhSectionPiece> pieces;
110341825Sdim
111353358Sdim  auto add = [&](const EhSectionPiece &p) {
112341825Sdim    // If P is adjacent to Last, squash the two.
113353358Sdim    if (!pieces.empty()) {
114353358Sdim      EhSectionPiece &last = pieces.back();
115353358Sdim      if (last.sec == p.sec && last.inputOff + last.size == p.inputOff &&
116353358Sdim          last.outputOff + last.size == p.outputOff) {
117353358Sdim        last.size += p.size;
118341825Sdim        return;
119341825Sdim      }
120341825Sdim    }
121353358Sdim    pieces.push_back(p);
122341825Sdim  };
123341825Sdim
124341825Sdim  // Gather section pieces.
125353358Sdim  for (const CieRecord *rec : sec->getCieRecords()) {
126353358Sdim    add(*rec->cie);
127353358Sdim    for (const EhSectionPiece *fde : rec->fdes)
128353358Sdim      add(*fde);
129341825Sdim  }
130341825Sdim
131341825Sdim  // Print out section pieces.
132353358Sdim  const OutputSection *osec = sec->getOutputSection();
133353358Sdim  for (EhSectionPiece &p : pieces) {
134353358Sdim    writeHeader(os, osec->addr + p.outputOff, osec->getLMA() + p.outputOff,
135353358Sdim                p.size, 1);
136353358Sdim    os << indent8 << toString(p.sec->file) << ":(" << p.sec->name << "+0x"
137353358Sdim       << Twine::utohexstr(p.inputOff) + ")\n";
138341825Sdim  }
139341825Sdim}
140341825Sdim
141360784Sdimvoid writeMapFile() {
142353358Sdim  if (config->mapFile.empty())
143317025Sdim    return;
144317025Sdim
145317778Sdim  // Open a map file for writing.
146353358Sdim  std::error_code ec;
147360784Sdim  raw_fd_ostream os(config->mapFile, ec, sys::fs::OF_None);
148353358Sdim  if (ec) {
149353358Sdim    error("cannot open " + config->mapFile + ": " + ec.message());
150317778Sdim    return;
151317778Sdim  }
152317778Sdim
153317778Sdim  // Collect symbol info that we want to print out.
154353358Sdim  std::vector<Defined *> syms = getSymbols();
155353358Sdim  SymbolMapTy sectionSyms = getSectionSyms(syms);
156353358Sdim  DenseMap<Symbol *, std::string> symStr = getSymbolStrings(syms);
157317778Sdim
158317778Sdim  // Print out the header line.
159353358Sdim  int w = config->is64 ? 16 : 8;
160353358Sdim  os << right_justify("VMA", w) << ' ' << right_justify("LMA", w)
161341825Sdim     << "     Size Align Out     In      Symbol\n";
162317778Sdim
163353358Sdim  OutputSection* osec = nullptr;
164353358Sdim  for (BaseCommand *base : script->sectionCommands) {
165353358Sdim    if (auto *cmd = dyn_cast<SymbolAssignment>(base)) {
166353358Sdim      if (cmd->provide && !cmd->sym)
167341825Sdim        continue;
168353358Sdim      uint64_t lma = osec ? osec->getLMA() + cmd->addr - osec->getVA(0) : 0;
169353358Sdim      writeHeader(os, cmd->addr, lma, cmd->size, 1);
170353358Sdim      os << cmd->commandString << '\n';
171341825Sdim      continue;
172341825Sdim    }
173341825Sdim
174353358Sdim    osec = cast<OutputSection>(base);
175353358Sdim    writeHeader(os, osec->addr, osec->getLMA(), osec->size, osec->alignment);
176353358Sdim    os << osec->name << '\n';
177317778Sdim
178317778Sdim    // Dump symbols for each input section.
179353358Sdim    for (BaseCommand *base : osec->sectionCommands) {
180353358Sdim      if (auto *isd = dyn_cast<InputSectionDescription>(base)) {
181353358Sdim        for (InputSection *isec : isd->sections) {
182353358Sdim          if (auto *ehSec = dyn_cast<EhFrameSection>(isec)) {
183353358Sdim            printEhFrame(os, ehSec);
184341825Sdim            continue;
185341825Sdim          }
186341825Sdim
187353358Sdim          writeHeader(os, isec->getVA(0), osec->getLMA() + isec->getOffset(0),
188353358Sdim                      isec->getSize(), isec->alignment);
189353358Sdim          os << indent8 << toString(isec) << '\n';
190353358Sdim          for (Symbol *sym : sectionSyms[isec])
191353358Sdim            os << symStr[sym] << '\n';
192341825Sdim        }
193319164Sdim        continue;
194319164Sdim      }
195341825Sdim
196353358Sdim      if (auto *cmd = dyn_cast<ByteCommand>(base)) {
197353358Sdim        writeHeader(os, osec->addr + cmd->offset, osec->getLMA() + cmd->offset,
198353358Sdim                    cmd->size, 1);
199353358Sdim        os << indent8 << cmd->commandString << '\n';
200341825Sdim        continue;
201341825Sdim      }
202341825Sdim
203353358Sdim      if (auto *cmd = dyn_cast<SymbolAssignment>(base)) {
204353358Sdim        if (cmd->provide && !cmd->sym)
205341825Sdim          continue;
206353358Sdim        writeHeader(os, cmd->addr, osec->getLMA() + cmd->addr - osec->getVA(0),
207353358Sdim                    cmd->size, 1);
208353358Sdim        os << indent8 << cmd->commandString << '\n';
209341825Sdim        continue;
210341825Sdim      }
211317778Sdim    }
212317778Sdim  }
213317025Sdim}
214341825Sdim
215353358Sdimstatic void print(StringRef a, StringRef b) {
216360784Sdim  lld::outs() << left_justify(a, 49) << " " << b << "\n";
217341825Sdim}
218341825Sdim
219341825Sdim// Output a cross reference table to stdout. This is for --cref.
220341825Sdim//
221341825Sdim// For each global symbol, we print out a file that defines the symbol
222341825Sdim// followed by files that uses that symbol. Here is an example.
223341825Sdim//
224341825Sdim//     strlen     /lib/x86_64-linux-gnu/libc.so.6
225341825Sdim//                tools/lld/tools/lld/CMakeFiles/lld.dir/lld.cpp.o
226341825Sdim//                lib/libLLVMSupport.a(PrettyStackTrace.cpp.o)
227341825Sdim//
228341825Sdim// In this case, strlen is defined by libc.so.6 and used by other two
229341825Sdim// files.
230360784Sdimvoid writeCrossReferenceTable() {
231353358Sdim  if (!config->cref)
232341825Sdim    return;
233341825Sdim
234341825Sdim  // Collect symbols and files.
235353358Sdim  MapVector<Symbol *, SetVector<InputFile *>> map;
236353358Sdim  for (InputFile *file : objectFiles) {
237353358Sdim    for (Symbol *sym : file->getSymbols()) {
238353358Sdim      if (isa<SharedSymbol>(sym))
239353358Sdim        map[sym].insert(file);
240353358Sdim      if (auto *d = dyn_cast<Defined>(sym))
241353358Sdim        if (!d->isLocal() && (!d->section || d->section->isLive()))
242353358Sdim          map[d].insert(file);
243341825Sdim    }
244341825Sdim  }
245341825Sdim
246341825Sdim  // Print out a header.
247360784Sdim  lld::outs() << "Cross Reference Table\n\n";
248341825Sdim  print("Symbol", "File");
249341825Sdim
250341825Sdim  // Print out a table.
251353358Sdim  for (auto kv : map) {
252353358Sdim    Symbol *sym = kv.first;
253353358Sdim    SetVector<InputFile *> &files = kv.second;
254341825Sdim
255353358Sdim    print(toString(*sym), toString(sym->file));
256353358Sdim    for (InputFile *file : files)
257353358Sdim      if (file != sym->file)
258353358Sdim        print("", toString(file));
259341825Sdim  }
260341825Sdim}
261360784Sdim
262360784Sdim} // namespace elf
263360784Sdim} // namespace lld
264