1//===- MapFile.cpp --------------------------------------------------------===//
2//
3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4// See https://llvm.org/LICENSE.txt for license information.
5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6//
7//===----------------------------------------------------------------------===//
8//
9// This file implements the -Map option. It shows lists in order and
10// hierarchically the output sections, input sections, input files and
11// symbol:
12//
13// Address Size Align Out In Symbol
14// 00201000 00000015 4 .text
15// 00201000 0000000e 4 test.o:(.text)
16// 0020100e 00000000 0 local
17// 00201005 00000000 0 f(int)
18//
19//===----------------------------------------------------------------------===//
20
21#include "MapFile.h"
22#include "InputFiles.h"
23#include "LinkerScript.h"
24#include "OutputSections.h"
25#include "Symbols.h"
26#include "SyntheticSections.h"
27#include "llvm/ADT/MapVector.h"
28#include "llvm/ADT/SetVector.h"
29#include "llvm/ADT/SmallPtrSet.h"
30#include "llvm/Support/FileSystem.h"
31#include "llvm/Support/Parallel.h"
32#include "llvm/Support/TimeProfiler.h"
33#include "llvm/Support/raw_ostream.h"
34
35using namespace llvm;
36using namespace llvm::object;
37using namespace lld;
38using namespace lld::elf;
39
40using SymbolMapTy = DenseMap<const SectionBase *,
41 SmallVector<std::pair<Defined *, uint64_t>, 0>>;
42
43static constexpr char indent8[] = " "; // 8 spaces
44static constexpr char indent16[] = " "; // 16 spaces
45
46// Print out the first three columns of a line.
47static void writeHeader(raw_ostream &os, uint64_t vma, uint64_t lma,
48 uint64_t size, uint64_t align) {
49 if (config->is64)
50 os << format(Fmt: "%16llx %16llx %8llx %5lld ", Vals: vma, Vals: lma, Vals: size, Vals: align);
51 else
52 os << format(Fmt: "%8llx %8llx %8llx %5lld ", Vals: vma, Vals: lma, Vals: size, Vals: align);
53}
54
55// Returns a list of all symbols that we want to print out.
56static std::vector<Defined *> getSymbols() {
57 std::vector<Defined *> v;
58 for (ELFFileBase *file : ctx.objectFiles)
59 for (Symbol *b : file->getSymbols())
60 if (auto *dr = dyn_cast<Defined>(Val: b))
61 if (!dr->isSection() && dr->section && dr->section->isLive() &&
62 (dr->file == file || dr->hasFlag(bit: NEEDS_COPY) || dr->section->bss))
63 v.push_back(x: dr);
64 return v;
65}
66
67// Returns a map from sections to their symbols.
68static SymbolMapTy getSectionSyms(ArrayRef<Defined *> syms) {
69 SymbolMapTy ret;
70 for (Defined *dr : syms)
71 ret[dr->section].emplace_back(Args&: dr, Args: dr->getVA());
72
73 // Sort symbols by address. We want to print out symbols in the
74 // order in the output file rather than the order they appeared
75 // in the input files.
76 SmallPtrSet<Defined *, 4> set;
77 for (auto &it : ret) {
78 // Deduplicate symbols which need a canonical PLT entry/copy relocation.
79 set.clear();
80 llvm::erase_if(C&: it.second, P: [&](std::pair<Defined *, uint64_t> a) {
81 return !set.insert(Ptr: a.first).second;
82 });
83
84 llvm::stable_sort(Range&: it.second, C: llvm::less_second());
85 }
86 return ret;
87}
88
89// Construct a map from symbols to their stringified representations.
90// Demangling symbols (which is what toString() does) is slow, so
91// we do that in batch using parallel-for.
92static DenseMap<Symbol *, std::string>
93getSymbolStrings(ArrayRef<Defined *> syms) {
94 auto strs = std::make_unique<std::string[]>(num: syms.size());
95 parallelFor(Begin: 0, End: syms.size(), Fn: [&](size_t i) {
96 raw_string_ostream os(strs[i]);
97 OutputSection *osec = syms[i]->getOutputSection();
98 uint64_t vma = syms[i]->getVA();
99 uint64_t lma = osec ? osec->getLMA() + vma - osec->getVA(offset: 0) : 0;
100 writeHeader(os, vma, lma, size: syms[i]->getSize(), align: 1);
101 os << indent16 << toString(*syms[i]);
102 });
103
104 DenseMap<Symbol *, std::string> ret;
105 for (size_t i = 0, e = syms.size(); i < e; ++i)
106 ret[syms[i]] = std::move(strs[i]);
107 return ret;
108}
109
110// Print .eh_frame contents. Since the section consists of EhSectionPieces,
111// we need a specialized printer for that section.
112//
113// .eh_frame tend to contain a lot of section pieces that are contiguous
114// both in input file and output file. Such pieces are squashed before
115// being displayed to make output compact.
116static void printEhFrame(raw_ostream &os, const EhFrameSection *sec) {
117 std::vector<EhSectionPiece> pieces;
118
119 auto add = [&](const EhSectionPiece &p) {
120 // If P is adjacent to Last, squash the two.
121 if (!pieces.empty()) {
122 EhSectionPiece &last = pieces.back();
123 if (last.sec == p.sec && last.inputOff + last.size == p.inputOff &&
124 last.outputOff + last.size == (unsigned)p.outputOff) {
125 last.size += p.size;
126 return;
127 }
128 }
129 pieces.push_back(x: p);
130 };
131
132 // Gather section pieces.
133 for (const CieRecord *rec : sec->getCieRecords()) {
134 add(*rec->cie);
135 for (const EhSectionPiece *fde : rec->fdes)
136 add(*fde);
137 }
138
139 // Print out section pieces.
140 const OutputSection *osec = sec->getOutputSection();
141 for (EhSectionPiece &p : pieces) {
142 writeHeader(os, vma: osec->addr + p.outputOff, lma: osec->getLMA() + p.outputOff,
143 size: p.size, align: 1);
144 os << indent8 << toString(f: p.sec->file) << ":(" << p.sec->name << "+0x"
145 << Twine::utohexstr(Val: p.inputOff) + ")\n";
146 }
147}
148
149static void writeMapFile(raw_fd_ostream &os) {
150 // Collect symbol info that we want to print out.
151 std::vector<Defined *> syms = getSymbols();
152 SymbolMapTy sectionSyms = getSectionSyms(syms);
153 DenseMap<Symbol *, std::string> symStr = getSymbolStrings(syms);
154
155 // Print out the header line.
156 int w = config->is64 ? 16 : 8;
157 os << right_justify(Str: "VMA", Width: w) << ' ' << right_justify(Str: "LMA", Width: w)
158 << " Size Align Out In Symbol\n";
159
160 OutputSection *osec = nullptr;
161 for (SectionCommand *cmd : script->sectionCommands) {
162 if (auto *assign = dyn_cast<SymbolAssignment>(Val: cmd)) {
163 if (assign->provide && !assign->sym)
164 continue;
165 uint64_t lma = osec ? osec->getLMA() + assign->addr - osec->getVA(offset: 0) : 0;
166 writeHeader(os, vma: assign->addr, lma, size: assign->size, align: 1);
167 os << assign->commandString << '\n';
168 continue;
169 }
170
171 osec = &cast<OutputDesc>(Val: cmd)->osec;
172 writeHeader(os, vma: osec->addr, lma: osec->getLMA(), size: osec->size, align: osec->addralign);
173 os << osec->name << '\n';
174
175 // Dump symbols for each input section.
176 for (SectionCommand *subCmd : osec->commands) {
177 if (auto *isd = dyn_cast<InputSectionDescription>(Val: subCmd)) {
178 for (InputSection *isec : isd->sections) {
179 if (auto *ehSec = dyn_cast<EhFrameSection>(Val: isec)) {
180 printEhFrame(os, sec: ehSec);
181 continue;
182 }
183
184 writeHeader(os, vma: isec->getVA(), lma: osec->getLMA() + isec->outSecOff,
185 size: isec->getSize(), align: isec->addralign);
186 os << indent8 << toString(isec) << '\n';
187 for (Symbol *sym : llvm::make_first_range(c&: sectionSyms[isec]))
188 os << symStr[sym] << '\n';
189 }
190 continue;
191 }
192
193 if (auto *data = dyn_cast<ByteCommand>(Val: subCmd)) {
194 writeHeader(os, vma: osec->addr + data->offset,
195 lma: osec->getLMA() + data->offset, size: data->size, align: 1);
196 os << indent8 << data->commandString << '\n';
197 continue;
198 }
199
200 if (auto *assign = dyn_cast<SymbolAssignment>(Val: subCmd)) {
201 if (assign->provide && !assign->sym)
202 continue;
203 writeHeader(os, vma: assign->addr,
204 lma: osec->getLMA() + assign->addr - osec->getVA(offset: 0),
205 size: assign->size, align: 1);
206 os << indent8 << assign->commandString << '\n';
207 continue;
208 }
209 }
210 }
211}
212
213// Output a cross reference table to stdout. This is for --cref.
214//
215// For each global symbol, we print out a file that defines the symbol
216// followed by files that uses that symbol. Here is an example.
217//
218// strlen /lib/x86_64-linux-gnu/libc.so.6
219// tools/lld/tools/lld/CMakeFiles/lld.dir/lld.cpp.o
220// lib/libLLVMSupport.a(PrettyStackTrace.cpp.o)
221//
222// In this case, strlen is defined by libc.so.6 and used by other two
223// files.
224static void writeCref(raw_fd_ostream &os) {
225 // Collect symbols and files.
226 MapVector<Symbol *, SetVector<InputFile *>> map;
227 for (ELFFileBase *file : ctx.objectFiles) {
228 for (Symbol *sym : file->getSymbols()) {
229 if (isa<SharedSymbol>(Val: sym))
230 map[sym].insert(X: file);
231 if (auto *d = dyn_cast<Defined>(Val: sym))
232 if (!d->isLocal())
233 map[d].insert(X: file);
234 }
235 }
236
237 auto print = [&](StringRef a, StringRef b) {
238 os << left_justify(Str: a, Width: 49) << ' ' << b << '\n';
239 };
240
241 // Print a blank line and a header. The format matches GNU ld.
242 os << "\nCross Reference Table\n\n";
243 print("Symbol", "File");
244
245 // Print out a table.
246 for (auto kv : map) {
247 Symbol *sym = kv.first;
248 SetVector<InputFile *> &files = kv.second;
249
250 print(toString(*sym), toString(f: sym->file));
251 for (InputFile *file : files)
252 if (file != sym->file)
253 print("", toString(f: file));
254 }
255}
256
257void elf::writeMapAndCref() {
258 if (config->mapFile.empty() && !config->cref)
259 return;
260
261 llvm::TimeTraceScope timeScope("Write map file");
262
263 // Open a map file for writing.
264 std::error_code ec;
265 StringRef mapFile = config->mapFile.empty() ? "-" : config->mapFile;
266 raw_fd_ostream os = ctx.openAuxiliaryFile(mapFile, ec);
267 if (ec) {
268 error(msg: "cannot open " + mapFile + ": " + ec.message());
269 return;
270 }
271
272 if (!config->mapFile.empty())
273 writeMapFile(os);
274 if (config->cref)
275 writeCref(os);
276}
277