1*0b57cec5SDimitry Andric //===- MapFile.cpp --------------------------------------------------------===//
2*0b57cec5SDimitry Andric //
3*0b57cec5SDimitry Andric // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4*0b57cec5SDimitry Andric // See https://llvm.org/LICENSE.txt for license information.
5*0b57cec5SDimitry Andric // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6*0b57cec5SDimitry Andric //
7*0b57cec5SDimitry Andric //===----------------------------------------------------------------------===//
8*0b57cec5SDimitry Andric //
9*0b57cec5SDimitry Andric // This file implements the -Map option. It shows lists in order and
10*0b57cec5SDimitry Andric // hierarchically the output sections, input sections, input files and
11*0b57cec5SDimitry Andric // symbol:
12*0b57cec5SDimitry Andric //
13*0b57cec5SDimitry Andric //   Address  Size     Align Out     In      Symbol
14*0b57cec5SDimitry Andric //   00201000 00000015     4 .text
15*0b57cec5SDimitry Andric //   00201000 0000000e     4         test.o:(.text)
16*0b57cec5SDimitry Andric //   0020100e 00000000     0                 local
17*0b57cec5SDimitry Andric //   00201005 00000000     0                 f(int)
18*0b57cec5SDimitry Andric //
19*0b57cec5SDimitry Andric //===----------------------------------------------------------------------===//
20*0b57cec5SDimitry Andric 
21*0b57cec5SDimitry Andric #include "MapFile.h"
22*0b57cec5SDimitry Andric #include "InputFiles.h"
23*0b57cec5SDimitry Andric #include "LinkerScript.h"
24*0b57cec5SDimitry Andric #include "OutputSections.h"
25*0b57cec5SDimitry Andric #include "SymbolTable.h"
26*0b57cec5SDimitry Andric #include "Symbols.h"
27*0b57cec5SDimitry Andric #include "SyntheticSections.h"
28*0b57cec5SDimitry Andric #include "lld/Common/Strings.h"
29*0b57cec5SDimitry Andric #include "llvm/ADT/MapVector.h"
30*0b57cec5SDimitry Andric #include "llvm/ADT/SetVector.h"
315ffd83dbSDimitry Andric #include "llvm/Support/Parallel.h"
32e8d8bef9SDimitry Andric #include "llvm/Support/TimeProfiler.h"
33*0b57cec5SDimitry Andric #include "llvm/Support/raw_ostream.h"
34*0b57cec5SDimitry Andric 
35*0b57cec5SDimitry Andric using namespace llvm;
36*0b57cec5SDimitry Andric using namespace llvm::object;
375ffd83dbSDimitry Andric using namespace lld;
385ffd83dbSDimitry Andric using namespace lld::elf;
39*0b57cec5SDimitry Andric 
40*0b57cec5SDimitry Andric using SymbolMapTy = DenseMap<const SectionBase *, SmallVector<Defined *, 4>>;
41*0b57cec5SDimitry Andric 
4285868e8aSDimitry Andric static constexpr char indent8[] = "        ";          // 8 spaces
4385868e8aSDimitry Andric static constexpr char indent16[] = "                "; // 16 spaces
44*0b57cec5SDimitry Andric 
45*0b57cec5SDimitry Andric // Print out the first three columns of a line.
46*0b57cec5SDimitry Andric static void writeHeader(raw_ostream &os, uint64_t vma, uint64_t lma,
47*0b57cec5SDimitry Andric                         uint64_t size, uint64_t align) {
48*0b57cec5SDimitry Andric   if (config->is64)
49*0b57cec5SDimitry Andric     os << format("%16llx %16llx %8llx %5lld ", vma, lma, size, align);
50*0b57cec5SDimitry Andric   else
51*0b57cec5SDimitry Andric     os << format("%8llx %8llx %8llx %5lld ", vma, lma, size, align);
52*0b57cec5SDimitry Andric }
53*0b57cec5SDimitry Andric 
54*0b57cec5SDimitry Andric // Returns a list of all symbols that we want to print out.
55*0b57cec5SDimitry Andric static std::vector<Defined *> getSymbols() {
56*0b57cec5SDimitry Andric   std::vector<Defined *> v;
57*0b57cec5SDimitry Andric   for (InputFile *file : objectFiles)
58*0b57cec5SDimitry Andric     for (Symbol *b : file->getSymbols())
59*0b57cec5SDimitry Andric       if (auto *dr = dyn_cast<Defined>(b))
60*0b57cec5SDimitry Andric         if (!dr->isSection() && dr->section && dr->section->isLive() &&
61*0b57cec5SDimitry Andric             (dr->file == file || dr->needsPltAddr || dr->section->bss))
62*0b57cec5SDimitry Andric           v.push_back(dr);
63*0b57cec5SDimitry Andric   return v;
64*0b57cec5SDimitry Andric }
65*0b57cec5SDimitry Andric 
66*0b57cec5SDimitry Andric // Returns a map from sections to their symbols.
67*0b57cec5SDimitry Andric static SymbolMapTy getSectionSyms(ArrayRef<Defined *> syms) {
68*0b57cec5SDimitry Andric   SymbolMapTy ret;
69*0b57cec5SDimitry Andric   for (Defined *dr : syms)
70*0b57cec5SDimitry Andric     ret[dr->section].push_back(dr);
71*0b57cec5SDimitry Andric 
72*0b57cec5SDimitry Andric   // Sort symbols by address. We want to print out symbols in the
73*0b57cec5SDimitry Andric   // order in the output file rather than the order they appeared
74*0b57cec5SDimitry Andric   // in the input files.
75*0b57cec5SDimitry Andric   for (auto &it : ret)
76*0b57cec5SDimitry Andric     llvm::stable_sort(it.second, [](Defined *a, Defined *b) {
77*0b57cec5SDimitry Andric       return a->getVA() < b->getVA();
78*0b57cec5SDimitry Andric     });
79*0b57cec5SDimitry Andric   return ret;
80*0b57cec5SDimitry Andric }
81*0b57cec5SDimitry Andric 
82*0b57cec5SDimitry Andric // Construct a map from symbols to their stringified representations.
83*0b57cec5SDimitry Andric // Demangling symbols (which is what toString() does) is slow, so
84*0b57cec5SDimitry Andric // we do that in batch using parallel-for.
85*0b57cec5SDimitry Andric static DenseMap<Symbol *, std::string>
86*0b57cec5SDimitry Andric getSymbolStrings(ArrayRef<Defined *> syms) {
87*0b57cec5SDimitry Andric   std::vector<std::string> str(syms.size());
88*0b57cec5SDimitry Andric   parallelForEachN(0, syms.size(), [&](size_t i) {
89*0b57cec5SDimitry Andric     raw_string_ostream os(str[i]);
90*0b57cec5SDimitry Andric     OutputSection *osec = syms[i]->getOutputSection();
91*0b57cec5SDimitry Andric     uint64_t vma = syms[i]->getVA();
92*0b57cec5SDimitry Andric     uint64_t lma = osec ? osec->getLMA() + vma - osec->getVA(0) : 0;
93*0b57cec5SDimitry Andric     writeHeader(os, vma, lma, syms[i]->getSize(), 1);
94*0b57cec5SDimitry Andric     os << indent16 << toString(*syms[i]);
95*0b57cec5SDimitry Andric   });
96*0b57cec5SDimitry Andric 
97*0b57cec5SDimitry Andric   DenseMap<Symbol *, std::string> ret;
98*0b57cec5SDimitry Andric   for (size_t i = 0, e = syms.size(); i < e; ++i)
99*0b57cec5SDimitry Andric     ret[syms[i]] = std::move(str[i]);
100*0b57cec5SDimitry Andric   return ret;
101*0b57cec5SDimitry Andric }
102*0b57cec5SDimitry Andric 
103*0b57cec5SDimitry Andric // Print .eh_frame contents. Since the section consists of EhSectionPieces,
104*0b57cec5SDimitry Andric // we need a specialized printer for that section.
105*0b57cec5SDimitry Andric //
106*0b57cec5SDimitry Andric // .eh_frame tend to contain a lot of section pieces that are contiguous
107*0b57cec5SDimitry Andric // both in input file and output file. Such pieces are squashed before
108*0b57cec5SDimitry Andric // being displayed to make output compact.
109*0b57cec5SDimitry Andric static void printEhFrame(raw_ostream &os, const EhFrameSection *sec) {
110*0b57cec5SDimitry Andric   std::vector<EhSectionPiece> pieces;
111*0b57cec5SDimitry Andric 
112*0b57cec5SDimitry Andric   auto add = [&](const EhSectionPiece &p) {
113*0b57cec5SDimitry Andric     // If P is adjacent to Last, squash the two.
114*0b57cec5SDimitry Andric     if (!pieces.empty()) {
115*0b57cec5SDimitry Andric       EhSectionPiece &last = pieces.back();
116*0b57cec5SDimitry Andric       if (last.sec == p.sec && last.inputOff + last.size == p.inputOff &&
117*0b57cec5SDimitry Andric           last.outputOff + last.size == p.outputOff) {
118*0b57cec5SDimitry Andric         last.size += p.size;
119*0b57cec5SDimitry Andric         return;
120*0b57cec5SDimitry Andric       }
121*0b57cec5SDimitry Andric     }
122*0b57cec5SDimitry Andric     pieces.push_back(p);
123*0b57cec5SDimitry Andric   };
124*0b57cec5SDimitry Andric 
125*0b57cec5SDimitry Andric   // Gather section pieces.
126*0b57cec5SDimitry Andric   for (const CieRecord *rec : sec->getCieRecords()) {
127*0b57cec5SDimitry Andric     add(*rec->cie);
128*0b57cec5SDimitry Andric     for (const EhSectionPiece *fde : rec->fdes)
129*0b57cec5SDimitry Andric       add(*fde);
130*0b57cec5SDimitry Andric   }
131*0b57cec5SDimitry Andric 
132*0b57cec5SDimitry Andric   // Print out section pieces.
133*0b57cec5SDimitry Andric   const OutputSection *osec = sec->getOutputSection();
134*0b57cec5SDimitry Andric   for (EhSectionPiece &p : pieces) {
135*0b57cec5SDimitry Andric     writeHeader(os, osec->addr + p.outputOff, osec->getLMA() + p.outputOff,
136*0b57cec5SDimitry Andric                 p.size, 1);
137*0b57cec5SDimitry Andric     os << indent8 << toString(p.sec->file) << ":(" << p.sec->name << "+0x"
138*0b57cec5SDimitry Andric        << Twine::utohexstr(p.inputOff) + ")\n";
139*0b57cec5SDimitry Andric   }
140*0b57cec5SDimitry Andric }
141*0b57cec5SDimitry Andric 
1424824e7fdSDimitry Andric static void writeMapFile(raw_fd_ostream &os) {
143*0b57cec5SDimitry Andric   // Collect symbol info that we want to print out.
144*0b57cec5SDimitry Andric   std::vector<Defined *> syms = getSymbols();
145*0b57cec5SDimitry Andric   SymbolMapTy sectionSyms = getSectionSyms(syms);
146*0b57cec5SDimitry Andric   DenseMap<Symbol *, std::string> symStr = getSymbolStrings(syms);
147*0b57cec5SDimitry Andric 
148*0b57cec5SDimitry Andric   // Print out the header line.
149*0b57cec5SDimitry Andric   int w = config->is64 ? 16 : 8;
150*0b57cec5SDimitry Andric   os << right_justify("VMA", w) << ' ' << right_justify("LMA", w)
151*0b57cec5SDimitry Andric      << "     Size Align Out     In      Symbol\n";
152*0b57cec5SDimitry Andric 
153*0b57cec5SDimitry Andric   OutputSection* osec = nullptr;
1544824e7fdSDimitry Andric   for (SectionCommand *cmd : script->sectionCommands) {
1554824e7fdSDimitry Andric     if (auto *assign = dyn_cast<SymbolAssignment>(cmd)) {
1564824e7fdSDimitry Andric       if (assign->provide && !assign->sym)
157*0b57cec5SDimitry Andric         continue;
1584824e7fdSDimitry Andric       uint64_t lma = osec ? osec->getLMA() + assign->addr - osec->getVA(0) : 0;
1594824e7fdSDimitry Andric       writeHeader(os, assign->addr, lma, assign->size, 1);
1604824e7fdSDimitry Andric       os << assign->commandString << '\n';
161*0b57cec5SDimitry Andric       continue;
162*0b57cec5SDimitry Andric     }
163*0b57cec5SDimitry Andric 
1644824e7fdSDimitry Andric     osec = cast<OutputSection>(cmd);
165*0b57cec5SDimitry Andric     writeHeader(os, osec->addr, osec->getLMA(), osec->size, osec->alignment);
166*0b57cec5SDimitry Andric     os << osec->name << '\n';
167*0b57cec5SDimitry Andric 
168*0b57cec5SDimitry Andric     // Dump symbols for each input section.
1694824e7fdSDimitry Andric     for (SectionCommand *subCmd : osec->commands) {
1704824e7fdSDimitry Andric       if (auto *isd = dyn_cast<InputSectionDescription>(subCmd)) {
171*0b57cec5SDimitry Andric         for (InputSection *isec : isd->sections) {
172*0b57cec5SDimitry Andric           if (auto *ehSec = dyn_cast<EhFrameSection>(isec)) {
173*0b57cec5SDimitry Andric             printEhFrame(os, ehSec);
174*0b57cec5SDimitry Andric             continue;
175*0b57cec5SDimitry Andric           }
176*0b57cec5SDimitry Andric 
1774824e7fdSDimitry Andric           writeHeader(os, isec->getVA(), osec->getLMA() + isec->outSecOff,
178*0b57cec5SDimitry Andric                       isec->getSize(), isec->alignment);
179*0b57cec5SDimitry Andric           os << indent8 << toString(isec) << '\n';
180*0b57cec5SDimitry Andric           for (Symbol *sym : sectionSyms[isec])
181*0b57cec5SDimitry Andric             os << symStr[sym] << '\n';
182*0b57cec5SDimitry Andric         }
183*0b57cec5SDimitry Andric         continue;
184*0b57cec5SDimitry Andric       }
185*0b57cec5SDimitry Andric 
1864824e7fdSDimitry Andric       if (auto *data = dyn_cast<ByteCommand>(subCmd)) {
1874824e7fdSDimitry Andric         writeHeader(os, osec->addr + data->offset,
1884824e7fdSDimitry Andric                     osec->getLMA() + data->offset, data->size, 1);
1894824e7fdSDimitry Andric         os << indent8 << data->commandString << '\n';
190*0b57cec5SDimitry Andric         continue;
191*0b57cec5SDimitry Andric       }
192*0b57cec5SDimitry Andric 
1934824e7fdSDimitry Andric       if (auto *assign = dyn_cast<SymbolAssignment>(subCmd)) {
1944824e7fdSDimitry Andric         if (assign->provide && !assign->sym)
195*0b57cec5SDimitry Andric           continue;
1964824e7fdSDimitry Andric         writeHeader(os, assign->addr,
1974824e7fdSDimitry Andric                     osec->getLMA() + assign->addr - osec->getVA(0),
1984824e7fdSDimitry Andric                     assign->size, 1);
1994824e7fdSDimitry Andric         os << indent8 << assign->commandString << '\n';
200*0b57cec5SDimitry Andric         continue;
201*0b57cec5SDimitry Andric       }
202*0b57cec5SDimitry Andric     }
203*0b57cec5SDimitry Andric   }
204*0b57cec5SDimitry Andric }
205*0b57cec5SDimitry Andric 
206349cc55cSDimitry Andric void elf::writeWhyExtract() {
207349cc55cSDimitry Andric   if (config->whyExtract.empty())
208349cc55cSDimitry Andric     return;
209349cc55cSDimitry Andric 
210349cc55cSDimitry Andric   std::error_code ec;
211349cc55cSDimitry Andric   raw_fd_ostream os(config->whyExtract, ec, sys::fs::OF_None);
212349cc55cSDimitry Andric   if (ec) {
213349cc55cSDimitry Andric     error("cannot open --why-extract= file " + config->whyExtract + ": " +
214349cc55cSDimitry Andric           ec.message());
215349cc55cSDimitry Andric     return;
216349cc55cSDimitry Andric   }
217349cc55cSDimitry Andric 
218349cc55cSDimitry Andric   os << "reference\textracted\tsymbol\n";
219349cc55cSDimitry Andric   for (auto &entry : whyExtract) {
220349cc55cSDimitry Andric     os << std::get<0>(entry) << '\t' << toString(std::get<1>(entry)) << '\t'
221349cc55cSDimitry Andric        << toString(std::get<2>(entry)) << '\n';
222349cc55cSDimitry Andric   }
223349cc55cSDimitry Andric }
224349cc55cSDimitry Andric 
225*0b57cec5SDimitry Andric // Output a cross reference table to stdout. This is for --cref.
226*0b57cec5SDimitry Andric //
227*0b57cec5SDimitry Andric // For each global symbol, we print out a file that defines the symbol
228*0b57cec5SDimitry Andric // followed by files that uses that symbol. Here is an example.
229*0b57cec5SDimitry Andric //
230*0b57cec5SDimitry Andric //     strlen     /lib/x86_64-linux-gnu/libc.so.6
231*0b57cec5SDimitry Andric //                tools/lld/tools/lld/CMakeFiles/lld.dir/lld.cpp.o
232*0b57cec5SDimitry Andric //                lib/libLLVMSupport.a(PrettyStackTrace.cpp.o)
233*0b57cec5SDimitry Andric //
234*0b57cec5SDimitry Andric // In this case, strlen is defined by libc.so.6 and used by other two
235*0b57cec5SDimitry Andric // files.
2364824e7fdSDimitry Andric static void writeCref(raw_fd_ostream &os) {
237*0b57cec5SDimitry Andric   // Collect symbols and files.
238*0b57cec5SDimitry Andric   MapVector<Symbol *, SetVector<InputFile *>> map;
239*0b57cec5SDimitry Andric   for (InputFile *file : objectFiles) {
240*0b57cec5SDimitry Andric     for (Symbol *sym : file->getSymbols()) {
241*0b57cec5SDimitry Andric       if (isa<SharedSymbol>(sym))
242*0b57cec5SDimitry Andric         map[sym].insert(file);
243*0b57cec5SDimitry Andric       if (auto *d = dyn_cast<Defined>(sym))
244*0b57cec5SDimitry Andric         if (!d->isLocal() && (!d->section || d->section->isLive()))
245*0b57cec5SDimitry Andric           map[d].insert(file);
246*0b57cec5SDimitry Andric     }
247*0b57cec5SDimitry Andric   }
248*0b57cec5SDimitry Andric 
2494824e7fdSDimitry Andric   auto print = [&](StringRef a, StringRef b) {
2504824e7fdSDimitry Andric     os << left_justify(a, 49) << ' ' << b << '\n';
2514824e7fdSDimitry Andric   };
2524824e7fdSDimitry Andric 
2534824e7fdSDimitry Andric   // Print a blank line and a header. The format matches GNU ld.
2544824e7fdSDimitry Andric   os << "\nCross Reference Table\n\n";
255*0b57cec5SDimitry Andric   print("Symbol", "File");
256*0b57cec5SDimitry Andric 
257*0b57cec5SDimitry Andric   // Print out a table.
258*0b57cec5SDimitry Andric   for (auto kv : map) {
259*0b57cec5SDimitry Andric     Symbol *sym = kv.first;
260*0b57cec5SDimitry Andric     SetVector<InputFile *> &files = kv.second;
261*0b57cec5SDimitry Andric 
262*0b57cec5SDimitry Andric     print(toString(*sym), toString(sym->file));
263*0b57cec5SDimitry Andric     for (InputFile *file : files)
264*0b57cec5SDimitry Andric       if (file != sym->file)
265*0b57cec5SDimitry Andric         print("", toString(file));
266*0b57cec5SDimitry Andric   }
267*0b57cec5SDimitry Andric }
26885868e8aSDimitry Andric 
2694824e7fdSDimitry Andric void elf::writeMapAndCref() {
2704824e7fdSDimitry Andric   if (config->mapFile.empty() && !config->cref)
2714824e7fdSDimitry Andric     return;
2724824e7fdSDimitry Andric 
2734824e7fdSDimitry Andric   llvm::TimeTraceScope timeScope("Write map file");
2744824e7fdSDimitry Andric 
2754824e7fdSDimitry Andric   // Open a map file for writing.
2764824e7fdSDimitry Andric   std::error_code ec;
2774824e7fdSDimitry Andric   StringRef mapFile = config->mapFile.empty() ? "-" : config->mapFile;
2784824e7fdSDimitry Andric   raw_fd_ostream os(mapFile, ec, sys::fs::OF_None);
2794824e7fdSDimitry Andric   if (ec) {
2804824e7fdSDimitry Andric     error("cannot open " + mapFile + ": " + ec.message());
2814824e7fdSDimitry Andric     return;
2824824e7fdSDimitry Andric   }
2834824e7fdSDimitry Andric 
2844824e7fdSDimitry Andric   if (!config->mapFile.empty())
2854824e7fdSDimitry Andric     writeMapFile(os);
2864824e7fdSDimitry Andric   if (config->cref)
2874824e7fdSDimitry Andric     writeCref(os);
2884824e7fdSDimitry Andric }
2894824e7fdSDimitry Andric 
2905ffd83dbSDimitry Andric void elf::writeArchiveStats() {
2915ffd83dbSDimitry Andric   if (config->printArchiveStats.empty())
2925ffd83dbSDimitry Andric     return;
2935ffd83dbSDimitry Andric 
2945ffd83dbSDimitry Andric   std::error_code ec;
2955ffd83dbSDimitry Andric   raw_fd_ostream os(config->printArchiveStats, ec, sys::fs::OF_None);
2965ffd83dbSDimitry Andric   if (ec) {
2975ffd83dbSDimitry Andric     error("--print-archive-stats=: cannot open " + config->printArchiveStats +
2985ffd83dbSDimitry Andric           ": " + ec.message());
2995ffd83dbSDimitry Andric     return;
3005ffd83dbSDimitry Andric   }
3015ffd83dbSDimitry Andric 
3024824e7fdSDimitry Andric   os << "members\textracted\tarchive\n";
3035ffd83dbSDimitry Andric   for (const ArchiveFile *f : archiveFiles)
3044824e7fdSDimitry Andric     os << f->getMemberCount() << '\t' << f->getExtractedMemberCount() << '\t'
3055ffd83dbSDimitry Andric        << f->getName() << '\n';
3065ffd83dbSDimitry Andric }
307