10b57cec5SDimitry Andric //===- Writer.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 #include "Writer.h"
10349cc55cSDimitry Andric #include "COFFLinkerContext.h"
11e8d8bef9SDimitry Andric #include "CallGraphSort.h"
120b57cec5SDimitry Andric #include "Config.h"
130b57cec5SDimitry Andric #include "DLL.h"
140b57cec5SDimitry Andric #include "InputFiles.h"
155ffd83dbSDimitry Andric #include "LLDMapFile.h"
160b57cec5SDimitry Andric #include "MapFile.h"
170b57cec5SDimitry Andric #include "PDB.h"
180b57cec5SDimitry Andric #include "SymbolTable.h"
190b57cec5SDimitry Andric #include "Symbols.h"
200b57cec5SDimitry Andric #include "lld/Common/ErrorHandler.h"
210b57cec5SDimitry Andric #include "lld/Common/Memory.h"
220b57cec5SDimitry Andric #include "lld/Common/Timer.h"
230b57cec5SDimitry Andric #include "llvm/ADT/DenseMap.h"
240b57cec5SDimitry Andric #include "llvm/ADT/STLExtras.h"
25480093f4SDimitry Andric #include "llvm/ADT/StringSet.h"
26bdd1243dSDimitry Andric #include "llvm/BinaryFormat/COFF.h"
270b57cec5SDimitry Andric #include "llvm/Support/BinaryStreamReader.h"
280b57cec5SDimitry Andric #include "llvm/Support/Debug.h"
290b57cec5SDimitry Andric #include "llvm/Support/Endian.h"
300b57cec5SDimitry Andric #include "llvm/Support/FileOutputBuffer.h"
310b57cec5SDimitry Andric #include "llvm/Support/Parallel.h"
320b57cec5SDimitry Andric #include "llvm/Support/Path.h"
330b57cec5SDimitry Andric #include "llvm/Support/RandomNumberGenerator.h"
345f757f3fSDimitry Andric #include "llvm/Support/TimeProfiler.h"
350b57cec5SDimitry Andric #include "llvm/Support/xxhash.h"
360b57cec5SDimitry Andric #include <algorithm>
370b57cec5SDimitry Andric #include <cstdio>
380b57cec5SDimitry Andric #include <map>
390b57cec5SDimitry Andric #include <memory>
400b57cec5SDimitry Andric #include <utility>
410b57cec5SDimitry Andric
420b57cec5SDimitry Andric using namespace llvm;
430b57cec5SDimitry Andric using namespace llvm::COFF;
440b57cec5SDimitry Andric using namespace llvm::object;
450b57cec5SDimitry Andric using namespace llvm::support;
460b57cec5SDimitry Andric using namespace llvm::support::endian;
475ffd83dbSDimitry Andric using namespace lld;
485ffd83dbSDimitry Andric using namespace lld::coff;
490b57cec5SDimitry Andric
500b57cec5SDimitry Andric /* To re-generate DOSProgram:
510b57cec5SDimitry Andric $ cat > /tmp/DOSProgram.asm
520b57cec5SDimitry Andric org 0
530b57cec5SDimitry Andric ; Copy cs to ds.
540b57cec5SDimitry Andric push cs
550b57cec5SDimitry Andric pop ds
560b57cec5SDimitry Andric ; Point ds:dx at the $-terminated string.
570b57cec5SDimitry Andric mov dx, str
580b57cec5SDimitry Andric ; Int 21/AH=09h: Write string to standard output.
590b57cec5SDimitry Andric mov ah, 0x9
600b57cec5SDimitry Andric int 0x21
610b57cec5SDimitry Andric ; Int 21/AH=4Ch: Exit with return code (in AL).
620b57cec5SDimitry Andric mov ax, 0x4C01
630b57cec5SDimitry Andric int 0x21
640b57cec5SDimitry Andric str:
650b57cec5SDimitry Andric db 'This program cannot be run in DOS mode.$'
660b57cec5SDimitry Andric align 8, db 0
670b57cec5SDimitry Andric $ nasm -fbin /tmp/DOSProgram.asm -o /tmp/DOSProgram.bin
680b57cec5SDimitry Andric $ xxd -i /tmp/DOSProgram.bin
690b57cec5SDimitry Andric */
700b57cec5SDimitry Andric static unsigned char dosProgram[] = {
710b57cec5SDimitry Andric 0x0e, 0x1f, 0xba, 0x0e, 0x00, 0xb4, 0x09, 0xcd, 0x21, 0xb8, 0x01, 0x4c,
720b57cec5SDimitry Andric 0xcd, 0x21, 0x54, 0x68, 0x69, 0x73, 0x20, 0x70, 0x72, 0x6f, 0x67, 0x72,
730b57cec5SDimitry Andric 0x61, 0x6d, 0x20, 0x63, 0x61, 0x6e, 0x6e, 0x6f, 0x74, 0x20, 0x62, 0x65,
740b57cec5SDimitry Andric 0x20, 0x72, 0x75, 0x6e, 0x20, 0x69, 0x6e, 0x20, 0x44, 0x4f, 0x53, 0x20,
750b57cec5SDimitry Andric 0x6d, 0x6f, 0x64, 0x65, 0x2e, 0x24, 0x00, 0x00
760b57cec5SDimitry Andric };
770b57cec5SDimitry Andric static_assert(sizeof(dosProgram) % 8 == 0,
780b57cec5SDimitry Andric "DOSProgram size must be multiple of 8");
790b57cec5SDimitry Andric
800b57cec5SDimitry Andric static const int dosStubSize = sizeof(dos_header) + sizeof(dosProgram);
810b57cec5SDimitry Andric static_assert(dosStubSize % 8 == 0, "DOSStub size must be multiple of 8");
820b57cec5SDimitry Andric
830b57cec5SDimitry Andric static const int numberOfDataDirectory = 16;
840b57cec5SDimitry Andric
850b57cec5SDimitry Andric namespace {
860b57cec5SDimitry Andric
870b57cec5SDimitry Andric class DebugDirectoryChunk : public NonSectionChunk {
880b57cec5SDimitry Andric public:
DebugDirectoryChunk(const COFFLinkerContext & c,const std::vector<std::pair<COFF::DebugType,Chunk * >> & r,bool writeRepro)89bdd1243dSDimitry Andric DebugDirectoryChunk(const COFFLinkerContext &c,
90349cc55cSDimitry Andric const std::vector<std::pair<COFF::DebugType, Chunk *>> &r,
915ffd83dbSDimitry Andric bool writeRepro)
92349cc55cSDimitry Andric : records(r), writeRepro(writeRepro), ctx(c) {}
930b57cec5SDimitry Andric
getSize() const940b57cec5SDimitry Andric size_t getSize() const override {
950b57cec5SDimitry Andric return (records.size() + int(writeRepro)) * sizeof(debug_directory);
960b57cec5SDimitry Andric }
970b57cec5SDimitry Andric
writeTo(uint8_t * b) const980b57cec5SDimitry Andric void writeTo(uint8_t *b) const override {
990b57cec5SDimitry Andric auto *d = reinterpret_cast<debug_directory *>(b);
1000b57cec5SDimitry Andric
1015ffd83dbSDimitry Andric for (const std::pair<COFF::DebugType, Chunk *>& record : records) {
1025ffd83dbSDimitry Andric Chunk *c = record.second;
103bdd1243dSDimitry Andric const OutputSection *os = ctx.getOutputSection(c);
1045ffd83dbSDimitry Andric uint64_t offs = os->getFileOff() + (c->getRVA() - os->getRVA());
1055ffd83dbSDimitry Andric fillEntry(d, record.first, c->getSize(), c->getRVA(), offs);
1060b57cec5SDimitry Andric ++d;
1070b57cec5SDimitry Andric }
1080b57cec5SDimitry Andric
1090b57cec5SDimitry Andric if (writeRepro) {
1100b57cec5SDimitry Andric // FIXME: The COFF spec allows either a 0-sized entry to just say
1110b57cec5SDimitry Andric // "the timestamp field is really a hash", or a 4-byte size field
1120b57cec5SDimitry Andric // followed by that many bytes containing a longer hash (with the
1130b57cec5SDimitry Andric // lowest 4 bytes usually being the timestamp in little-endian order).
11406c3fb27SDimitry Andric // Consider storing the full 8 bytes computed by xxh3_64bits here.
1150b57cec5SDimitry Andric fillEntry(d, COFF::IMAGE_DEBUG_TYPE_REPRO, 0, 0, 0);
1160b57cec5SDimitry Andric }
1170b57cec5SDimitry Andric }
1180b57cec5SDimitry Andric
setTimeDateStamp(uint32_t timeDateStamp)1190b57cec5SDimitry Andric void setTimeDateStamp(uint32_t timeDateStamp) {
1200b57cec5SDimitry Andric for (support::ulittle32_t *tds : timeDateStamps)
1210b57cec5SDimitry Andric *tds = timeDateStamp;
1220b57cec5SDimitry Andric }
1230b57cec5SDimitry Andric
1240b57cec5SDimitry Andric private:
fillEntry(debug_directory * d,COFF::DebugType debugType,size_t size,uint64_t rva,uint64_t offs) const1250b57cec5SDimitry Andric void fillEntry(debug_directory *d, COFF::DebugType debugType, size_t size,
1260b57cec5SDimitry Andric uint64_t rva, uint64_t offs) const {
1270b57cec5SDimitry Andric d->Characteristics = 0;
1280b57cec5SDimitry Andric d->TimeDateStamp = 0;
1290b57cec5SDimitry Andric d->MajorVersion = 0;
1300b57cec5SDimitry Andric d->MinorVersion = 0;
1310b57cec5SDimitry Andric d->Type = debugType;
1320b57cec5SDimitry Andric d->SizeOfData = size;
1330b57cec5SDimitry Andric d->AddressOfRawData = rva;
1340b57cec5SDimitry Andric d->PointerToRawData = offs;
1350b57cec5SDimitry Andric
1360b57cec5SDimitry Andric timeDateStamps.push_back(&d->TimeDateStamp);
1370b57cec5SDimitry Andric }
1380b57cec5SDimitry Andric
1390b57cec5SDimitry Andric mutable std::vector<support::ulittle32_t *> timeDateStamps;
1405ffd83dbSDimitry Andric const std::vector<std::pair<COFF::DebugType, Chunk *>> &records;
1410b57cec5SDimitry Andric bool writeRepro;
142bdd1243dSDimitry Andric const COFFLinkerContext &ctx;
1430b57cec5SDimitry Andric };
1440b57cec5SDimitry Andric
1450b57cec5SDimitry Andric class CVDebugRecordChunk : public NonSectionChunk {
1460b57cec5SDimitry Andric public:
CVDebugRecordChunk(const COFFLinkerContext & c)147bdd1243dSDimitry Andric CVDebugRecordChunk(const COFFLinkerContext &c) : ctx(c) {}
148bdd1243dSDimitry Andric
getSize() const1490b57cec5SDimitry Andric size_t getSize() const override {
150bdd1243dSDimitry Andric return sizeof(codeview::DebugInfo) + ctx.config.pdbAltPath.size() + 1;
1510b57cec5SDimitry Andric }
1520b57cec5SDimitry Andric
writeTo(uint8_t * b) const1530b57cec5SDimitry Andric void writeTo(uint8_t *b) const override {
1540b57cec5SDimitry Andric // Save off the DebugInfo entry to backfill the file signature (build id)
1550b57cec5SDimitry Andric // in Writer::writeBuildId
1560b57cec5SDimitry Andric buildId = reinterpret_cast<codeview::DebugInfo *>(b);
1570b57cec5SDimitry Andric
1580b57cec5SDimitry Andric // variable sized field (PDB Path)
1590b57cec5SDimitry Andric char *p = reinterpret_cast<char *>(b + sizeof(*buildId));
160bdd1243dSDimitry Andric if (!ctx.config.pdbAltPath.empty())
161bdd1243dSDimitry Andric memcpy(p, ctx.config.pdbAltPath.data(), ctx.config.pdbAltPath.size());
162bdd1243dSDimitry Andric p[ctx.config.pdbAltPath.size()] = '\0';
1630b57cec5SDimitry Andric }
1640b57cec5SDimitry Andric
1650b57cec5SDimitry Andric mutable codeview::DebugInfo *buildId = nullptr;
166bdd1243dSDimitry Andric
167bdd1243dSDimitry Andric private:
168bdd1243dSDimitry Andric const COFFLinkerContext &ctx;
1690b57cec5SDimitry Andric };
1700b57cec5SDimitry Andric
1715ffd83dbSDimitry Andric class ExtendedDllCharacteristicsChunk : public NonSectionChunk {
1725ffd83dbSDimitry Andric public:
ExtendedDllCharacteristicsChunk(uint32_t c)1735ffd83dbSDimitry Andric ExtendedDllCharacteristicsChunk(uint32_t c) : characteristics(c) {}
1745ffd83dbSDimitry Andric
getSize() const1755ffd83dbSDimitry Andric size_t getSize() const override { return 4; }
1765ffd83dbSDimitry Andric
writeTo(uint8_t * buf) const1775ffd83dbSDimitry Andric void writeTo(uint8_t *buf) const override { write32le(buf, characteristics); }
1785ffd83dbSDimitry Andric
1795ffd83dbSDimitry Andric uint32_t characteristics = 0;
1805ffd83dbSDimitry Andric };
1815ffd83dbSDimitry Andric
1820b57cec5SDimitry Andric // PartialSection represents a group of chunks that contribute to an
1830b57cec5SDimitry Andric // OutputSection. Collating a collection of PartialSections of same name and
1840b57cec5SDimitry Andric // characteristics constitutes the OutputSection.
1850b57cec5SDimitry Andric class PartialSectionKey {
1860b57cec5SDimitry Andric public:
1870b57cec5SDimitry Andric StringRef name;
1880b57cec5SDimitry Andric unsigned characteristics;
1890b57cec5SDimitry Andric
operator <(const PartialSectionKey & other) const1900b57cec5SDimitry Andric bool operator<(const PartialSectionKey &other) const {
1910b57cec5SDimitry Andric int c = name.compare(other.name);
192bdd1243dSDimitry Andric if (c > 0)
1930b57cec5SDimitry Andric return false;
1940b57cec5SDimitry Andric if (c == 0)
1950b57cec5SDimitry Andric return characteristics < other.characteristics;
1960b57cec5SDimitry Andric return true;
1970b57cec5SDimitry Andric }
1980b57cec5SDimitry Andric };
1990b57cec5SDimitry Andric
2005f757f3fSDimitry Andric struct ChunkRange {
2015f757f3fSDimitry Andric Chunk *first = nullptr, *last;
2025f757f3fSDimitry Andric };
2035f757f3fSDimitry Andric
2040b57cec5SDimitry Andric // The writer writes a SymbolTable result to a file.
2050b57cec5SDimitry Andric class Writer {
2060b57cec5SDimitry Andric public:
Writer(COFFLinkerContext & c)207bdd1243dSDimitry Andric Writer(COFFLinkerContext &c)
208bdd1243dSDimitry Andric : buffer(errorHandler().outputBuffer), delayIdata(c), edata(c), ctx(c) {}
2090b57cec5SDimitry Andric void run();
2100b57cec5SDimitry Andric
2110b57cec5SDimitry Andric private:
2120b57cec5SDimitry Andric void createSections();
2130b57cec5SDimitry Andric void createMiscChunks();
2140b57cec5SDimitry Andric void createImportTables();
2150b57cec5SDimitry Andric void appendImportThunks();
2160b57cec5SDimitry Andric void locateImportTables();
2170b57cec5SDimitry Andric void createExportTable();
2180b57cec5SDimitry Andric void mergeSections();
2195f757f3fSDimitry Andric void sortECChunks();
2200b57cec5SDimitry Andric void removeUnusedSections();
2210b57cec5SDimitry Andric void assignAddresses();
222bdd1243dSDimitry Andric bool isInRange(uint16_t relType, uint64_t s, uint64_t p, int margin);
223bdd1243dSDimitry Andric std::pair<Defined *, bool> getThunk(DenseMap<uint64_t, Defined *> &lastThunks,
224bdd1243dSDimitry Andric Defined *target, uint64_t p,
225bdd1243dSDimitry Andric uint16_t type, int margin);
226bdd1243dSDimitry Andric bool createThunks(OutputSection *os, int margin);
227bdd1243dSDimitry Andric bool verifyRanges(const std::vector<Chunk *> chunks);
2285f757f3fSDimitry Andric void createECCodeMap();
2290b57cec5SDimitry Andric void finalizeAddresses();
2300b57cec5SDimitry Andric void removeEmptySections();
2310b57cec5SDimitry Andric void assignOutputSectionIndices();
2320b57cec5SDimitry Andric void createSymbolAndStringTable();
2330b57cec5SDimitry Andric void openFile(StringRef outputPath);
2340b57cec5SDimitry Andric template <typename PEHeaderTy> void writeHeader();
2350b57cec5SDimitry Andric void createSEHTable();
2360b57cec5SDimitry Andric void createRuntimePseudoRelocs();
2375f757f3fSDimitry Andric void createECChunks();
2380b57cec5SDimitry Andric void insertCtorDtorSymbols();
239bdd1243dSDimitry Andric void markSymbolsWithRelocations(ObjFile *file, SymbolRVASet &usedSymbols);
2400b57cec5SDimitry Andric void createGuardCFTables();
2410b57cec5SDimitry Andric void markSymbolsForRVATable(ObjFile *file,
2420b57cec5SDimitry Andric ArrayRef<SectionChunk *> symIdxChunks,
2430b57cec5SDimitry Andric SymbolRVASet &tableSymbols);
244e8d8bef9SDimitry Andric void getSymbolsFromSections(ObjFile *file,
245e8d8bef9SDimitry Andric ArrayRef<SectionChunk *> symIdxChunks,
246e8d8bef9SDimitry Andric std::vector<Symbol *> &symbols);
2470b57cec5SDimitry Andric void maybeAddRVATable(SymbolRVASet tableSymbols, StringRef tableSym,
248fe6060f1SDimitry Andric StringRef countSym, bool hasFlag=false);
2490b57cec5SDimitry Andric void setSectionPermissions();
2505f757f3fSDimitry Andric void setECSymbols();
2510b57cec5SDimitry Andric void writeSections();
2520b57cec5SDimitry Andric void writeBuildId();
253bdd1243dSDimitry Andric void writePEChecksum();
254e8d8bef9SDimitry Andric void sortSections();
2555f757f3fSDimitry Andric template <typename T> void sortExceptionTable(ChunkRange &exceptionTable);
2565f757f3fSDimitry Andric void sortExceptionTables();
2570b57cec5SDimitry Andric void sortCRTSectionChunks(std::vector<Chunk *> &chunks);
2580b57cec5SDimitry Andric void addSyntheticIdata();
259bdd1243dSDimitry Andric void sortBySectionOrder(std::vector<Chunk *> &chunks);
2600b57cec5SDimitry Andric void fixPartialSectionChars(StringRef name, uint32_t chars);
2610b57cec5SDimitry Andric bool fixGnuImportChunks();
262e8d8bef9SDimitry Andric void fixTlsAlignment();
2630b57cec5SDimitry Andric PartialSection *createPartialSection(StringRef name, uint32_t outChars);
2640b57cec5SDimitry Andric PartialSection *findPartialSection(StringRef name, uint32_t outChars);
2650b57cec5SDimitry Andric
266bdd1243dSDimitry Andric std::optional<coff_symbol16> createSymbol(Defined *d);
2670b57cec5SDimitry Andric size_t addEntryToStringTable(StringRef str);
2680b57cec5SDimitry Andric
2690b57cec5SDimitry Andric OutputSection *findSection(StringRef name);
2700b57cec5SDimitry Andric void addBaserels();
2710b57cec5SDimitry Andric void addBaserelBlocks(std::vector<Baserel> &v);
2720b57cec5SDimitry Andric
2730b57cec5SDimitry Andric uint32_t getSizeOfInitializedData();
2740b57cec5SDimitry Andric
2755f757f3fSDimitry Andric void prepareLoadConfig();
2765f757f3fSDimitry Andric template <typename T> void prepareLoadConfig(T *loadConfig);
277bdd1243dSDimitry Andric template <typename T> void checkLoadConfigGuardData(const T *loadConfig);
278bdd1243dSDimitry Andric
2790b57cec5SDimitry Andric std::unique_ptr<FileOutputBuffer> &buffer;
2800b57cec5SDimitry Andric std::map<PartialSectionKey, PartialSection *> partialSections;
2810b57cec5SDimitry Andric std::vector<char> strtab;
2820b57cec5SDimitry Andric std::vector<llvm::object::coff_symbol16> outputSymtab;
2835f757f3fSDimitry Andric std::vector<ECCodeMapEntry> codeMap;
2840b57cec5SDimitry Andric IdataContents idata;
2850b57cec5SDimitry Andric Chunk *importTableStart = nullptr;
2860b57cec5SDimitry Andric uint64_t importTableSize = 0;
28785868e8aSDimitry Andric Chunk *edataStart = nullptr;
28885868e8aSDimitry Andric Chunk *edataEnd = nullptr;
2890b57cec5SDimitry Andric Chunk *iatStart = nullptr;
2900b57cec5SDimitry Andric uint64_t iatSize = 0;
2910b57cec5SDimitry Andric DelayLoadContents delayIdata;
2920b57cec5SDimitry Andric EdataContents edata;
2930b57cec5SDimitry Andric bool setNoSEHCharacteristic = false;
294e8d8bef9SDimitry Andric uint32_t tlsAlignment = 0;
2950b57cec5SDimitry Andric
2960b57cec5SDimitry Andric DebugDirectoryChunk *debugDirectory = nullptr;
2975ffd83dbSDimitry Andric std::vector<std::pair<COFF::DebugType, Chunk *>> debugRecords;
2980b57cec5SDimitry Andric CVDebugRecordChunk *buildId = nullptr;
2990b57cec5SDimitry Andric ArrayRef<uint8_t> sectionTable;
3000b57cec5SDimitry Andric
3010b57cec5SDimitry Andric uint64_t fileSize;
3020b57cec5SDimitry Andric uint32_t pointerToSymbolTable = 0;
3030b57cec5SDimitry Andric uint64_t sizeOfImage;
3040b57cec5SDimitry Andric uint64_t sizeOfHeaders;
3050b57cec5SDimitry Andric
3060b57cec5SDimitry Andric OutputSection *textSec;
3070b57cec5SDimitry Andric OutputSection *rdataSec;
3080b57cec5SDimitry Andric OutputSection *buildidSec;
3090b57cec5SDimitry Andric OutputSection *dataSec;
3100b57cec5SDimitry Andric OutputSection *pdataSec;
3110b57cec5SDimitry Andric OutputSection *idataSec;
3120b57cec5SDimitry Andric OutputSection *edataSec;
3130b57cec5SDimitry Andric OutputSection *didatSec;
3140b57cec5SDimitry Andric OutputSection *rsrcSec;
3150b57cec5SDimitry Andric OutputSection *relocSec;
3160b57cec5SDimitry Andric OutputSection *ctorsSec;
3170b57cec5SDimitry Andric OutputSection *dtorsSec;
3185f757f3fSDimitry Andric // Either .rdata section or .buildid section.
3195f757f3fSDimitry Andric OutputSection *debugInfoSec;
3200b57cec5SDimitry Andric
3215f757f3fSDimitry Andric // The range of .pdata sections in the output file.
3220b57cec5SDimitry Andric //
3230b57cec5SDimitry Andric // We need to keep track of the location of .pdata in whichever section it
3240b57cec5SDimitry Andric // gets merged into so that we can sort its contents and emit a correct data
3250b57cec5SDimitry Andric // directory entry for the exception table. This is also the case for some
3260b57cec5SDimitry Andric // other sections (such as .edata) but because the contents of those sections
3270b57cec5SDimitry Andric // are entirely linker-generated we can keep track of their locations using
3280b57cec5SDimitry Andric // the chunks that the linker creates. All .pdata chunks come from input
3290b57cec5SDimitry Andric // files, so we need to keep track of them separately.
3305f757f3fSDimitry Andric ChunkRange pdata;
3315f757f3fSDimitry Andric
3325f757f3fSDimitry Andric // x86_64 .pdata sections on ARM64EC/ARM64X targets.
3335f757f3fSDimitry Andric ChunkRange hybridPdata;
334349cc55cSDimitry Andric
335349cc55cSDimitry Andric COFFLinkerContext &ctx;
3360b57cec5SDimitry Andric };
3370b57cec5SDimitry Andric } // anonymous namespace
3380b57cec5SDimitry Andric
writeResult(COFFLinkerContext & ctx)3395f757f3fSDimitry Andric void lld::coff::writeResult(COFFLinkerContext &ctx) {
3405f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Write output(s)");
3415f757f3fSDimitry Andric Writer(ctx).run();
3425f757f3fSDimitry Andric }
3430b57cec5SDimitry Andric
addChunk(Chunk * c)3440b57cec5SDimitry Andric void OutputSection::addChunk(Chunk *c) {
3450b57cec5SDimitry Andric chunks.push_back(c);
3460b57cec5SDimitry Andric }
3470b57cec5SDimitry Andric
insertChunkAtStart(Chunk * c)3480b57cec5SDimitry Andric void OutputSection::insertChunkAtStart(Chunk *c) {
3490b57cec5SDimitry Andric chunks.insert(chunks.begin(), c);
3500b57cec5SDimitry Andric }
3510b57cec5SDimitry Andric
setPermissions(uint32_t c)3520b57cec5SDimitry Andric void OutputSection::setPermissions(uint32_t c) {
3530b57cec5SDimitry Andric header.Characteristics &= ~permMask;
3540b57cec5SDimitry Andric header.Characteristics |= c;
3550b57cec5SDimitry Andric }
3560b57cec5SDimitry Andric
merge(OutputSection * other)3570b57cec5SDimitry Andric void OutputSection::merge(OutputSection *other) {
3580b57cec5SDimitry Andric chunks.insert(chunks.end(), other->chunks.begin(), other->chunks.end());
3590b57cec5SDimitry Andric other->chunks.clear();
3600b57cec5SDimitry Andric contribSections.insert(contribSections.end(), other->contribSections.begin(),
3610b57cec5SDimitry Andric other->contribSections.end());
3620b57cec5SDimitry Andric other->contribSections.clear();
3635f757f3fSDimitry Andric
3645f757f3fSDimitry Andric // MS link.exe compatibility: when merging a code section into a data section,
3655f757f3fSDimitry Andric // mark the target section as a code section.
3665f757f3fSDimitry Andric if (other->header.Characteristics & IMAGE_SCN_CNT_CODE) {
3675f757f3fSDimitry Andric header.Characteristics |= IMAGE_SCN_CNT_CODE;
3685f757f3fSDimitry Andric header.Characteristics &=
3695f757f3fSDimitry Andric ~(IMAGE_SCN_CNT_INITIALIZED_DATA | IMAGE_SCN_CNT_UNINITIALIZED_DATA);
3705f757f3fSDimitry Andric }
3710b57cec5SDimitry Andric }
3720b57cec5SDimitry Andric
3730b57cec5SDimitry Andric // Write the section header to a given buffer.
writeHeaderTo(uint8_t * buf,bool isDebug)374bdd1243dSDimitry Andric void OutputSection::writeHeaderTo(uint8_t *buf, bool isDebug) {
3750b57cec5SDimitry Andric auto *hdr = reinterpret_cast<coff_section *>(buf);
3760b57cec5SDimitry Andric *hdr = header;
3770b57cec5SDimitry Andric if (stringTableOff) {
3780b57cec5SDimitry Andric // If name is too long, write offset into the string table as a name.
37981ad6265SDimitry Andric encodeSectionName(hdr->Name, stringTableOff);
3800b57cec5SDimitry Andric } else {
381bdd1243dSDimitry Andric assert(!isDebug || name.size() <= COFF::NameSize ||
3820b57cec5SDimitry Andric (hdr->Characteristics & IMAGE_SCN_MEM_DISCARDABLE) == 0);
3830b57cec5SDimitry Andric strncpy(hdr->Name, name.data(),
3840b57cec5SDimitry Andric std::min(name.size(), (size_t)COFF::NameSize));
3850b57cec5SDimitry Andric }
3860b57cec5SDimitry Andric }
3870b57cec5SDimitry Andric
addContributingPartialSection(PartialSection * sec)3880b57cec5SDimitry Andric void OutputSection::addContributingPartialSection(PartialSection *sec) {
3890b57cec5SDimitry Andric contribSections.push_back(sec);
3900b57cec5SDimitry Andric }
3910b57cec5SDimitry Andric
3920b57cec5SDimitry Andric // Check whether the target address S is in range from a relocation
3930b57cec5SDimitry Andric // of type relType at address P.
isInRange(uint16_t relType,uint64_t s,uint64_t p,int margin)394bdd1243dSDimitry Andric bool Writer::isInRange(uint16_t relType, uint64_t s, uint64_t p, int margin) {
395bdd1243dSDimitry Andric if (ctx.config.machine == ARMNT) {
3960b57cec5SDimitry Andric int64_t diff = AbsoluteDifference(s, p + 4) + margin;
3970b57cec5SDimitry Andric switch (relType) {
3980b57cec5SDimitry Andric case IMAGE_REL_ARM_BRANCH20T:
3990b57cec5SDimitry Andric return isInt<21>(diff);
4000b57cec5SDimitry Andric case IMAGE_REL_ARM_BRANCH24T:
4010b57cec5SDimitry Andric case IMAGE_REL_ARM_BLX23T:
4020b57cec5SDimitry Andric return isInt<25>(diff);
4030b57cec5SDimitry Andric default:
4040b57cec5SDimitry Andric return true;
4050b57cec5SDimitry Andric }
406bdd1243dSDimitry Andric } else if (ctx.config.machine == ARM64) {
4070b57cec5SDimitry Andric int64_t diff = AbsoluteDifference(s, p) + margin;
4080b57cec5SDimitry Andric switch (relType) {
4090b57cec5SDimitry Andric case IMAGE_REL_ARM64_BRANCH26:
4100b57cec5SDimitry Andric return isInt<28>(diff);
4110b57cec5SDimitry Andric case IMAGE_REL_ARM64_BRANCH19:
4120b57cec5SDimitry Andric return isInt<21>(diff);
4130b57cec5SDimitry Andric case IMAGE_REL_ARM64_BRANCH14:
4140b57cec5SDimitry Andric return isInt<16>(diff);
4150b57cec5SDimitry Andric default:
4160b57cec5SDimitry Andric return true;
4170b57cec5SDimitry Andric }
4180b57cec5SDimitry Andric } else {
4190b57cec5SDimitry Andric llvm_unreachable("Unexpected architecture");
4200b57cec5SDimitry Andric }
4210b57cec5SDimitry Andric }
4220b57cec5SDimitry Andric
4230b57cec5SDimitry Andric // Return the last thunk for the given target if it is in range,
4240b57cec5SDimitry Andric // or create a new one.
425bdd1243dSDimitry Andric std::pair<Defined *, bool>
getThunk(DenseMap<uint64_t,Defined * > & lastThunks,Defined * target,uint64_t p,uint16_t type,int margin)426bdd1243dSDimitry Andric Writer::getThunk(DenseMap<uint64_t, Defined *> &lastThunks, Defined *target,
427bdd1243dSDimitry Andric uint64_t p, uint16_t type, int margin) {
4280b57cec5SDimitry Andric Defined *&lastThunk = lastThunks[target->getRVA()];
4290b57cec5SDimitry Andric if (lastThunk && isInRange(type, lastThunk->getRVA(), p, margin))
4300b57cec5SDimitry Andric return {lastThunk, false};
4310b57cec5SDimitry Andric Chunk *c;
432bdd1243dSDimitry Andric switch (ctx.config.machine) {
4330b57cec5SDimitry Andric case ARMNT:
434bdd1243dSDimitry Andric c = make<RangeExtensionThunkARM>(ctx, target);
4350b57cec5SDimitry Andric break;
4360b57cec5SDimitry Andric case ARM64:
437bdd1243dSDimitry Andric c = make<RangeExtensionThunkARM64>(ctx, target);
4380b57cec5SDimitry Andric break;
4390b57cec5SDimitry Andric default:
4400b57cec5SDimitry Andric llvm_unreachable("Unexpected architecture");
4410b57cec5SDimitry Andric }
4426246ae0bSDimitry Andric Defined *d = make<DefinedSynthetic>("range_extension_thunk", c);
4430b57cec5SDimitry Andric lastThunk = d;
4440b57cec5SDimitry Andric return {d, true};
4450b57cec5SDimitry Andric }
4460b57cec5SDimitry Andric
4470b57cec5SDimitry Andric // This checks all relocations, and for any relocation which isn't in range
4480b57cec5SDimitry Andric // it adds a thunk after the section chunk that contains the relocation.
4490b57cec5SDimitry Andric // If the latest thunk for the specific target is in range, that is used
4500b57cec5SDimitry Andric // instead of creating a new thunk. All range checks are done with the
4510b57cec5SDimitry Andric // specified margin, to make sure that relocations that originally are in
4520b57cec5SDimitry Andric // range, but only barely, also get thunks - in case other added thunks makes
4530b57cec5SDimitry Andric // the target go out of range.
4540b57cec5SDimitry Andric //
4550b57cec5SDimitry Andric // After adding thunks, we verify that all relocations are in range (with
4560b57cec5SDimitry Andric // no extra margin requirements). If this failed, we restart (throwing away
4570b57cec5SDimitry Andric // the previously created thunks) and retry with a wider margin.
createThunks(OutputSection * os,int margin)458bdd1243dSDimitry Andric bool Writer::createThunks(OutputSection *os, int margin) {
4590b57cec5SDimitry Andric bool addressesChanged = false;
4600b57cec5SDimitry Andric DenseMap<uint64_t, Defined *> lastThunks;
4610b57cec5SDimitry Andric DenseMap<std::pair<ObjFile *, Defined *>, uint32_t> thunkSymtabIndices;
4620b57cec5SDimitry Andric size_t thunksSize = 0;
4630b57cec5SDimitry Andric // Recheck Chunks.size() each iteration, since we can insert more
4640b57cec5SDimitry Andric // elements into it.
4650b57cec5SDimitry Andric for (size_t i = 0; i != os->chunks.size(); ++i) {
4660b57cec5SDimitry Andric SectionChunk *sc = dyn_cast_or_null<SectionChunk>(os->chunks[i]);
4670b57cec5SDimitry Andric if (!sc)
4680b57cec5SDimitry Andric continue;
4690b57cec5SDimitry Andric size_t thunkInsertionSpot = i + 1;
4700b57cec5SDimitry Andric
4710b57cec5SDimitry Andric // Try to get a good enough estimate of where new thunks will be placed.
4720b57cec5SDimitry Andric // Offset this by the size of the new thunks added so far, to make the
4730b57cec5SDimitry Andric // estimate slightly better.
4740b57cec5SDimitry Andric size_t thunkInsertionRVA = sc->getRVA() + sc->getSize() + thunksSize;
4750b57cec5SDimitry Andric ObjFile *file = sc->file;
4760b57cec5SDimitry Andric std::vector<std::pair<uint32_t, uint32_t>> relocReplacements;
4770b57cec5SDimitry Andric ArrayRef<coff_relocation> originalRelocs =
4780b57cec5SDimitry Andric file->getCOFFObj()->getRelocations(sc->header);
4790b57cec5SDimitry Andric for (size_t j = 0, e = originalRelocs.size(); j < e; ++j) {
4800b57cec5SDimitry Andric const coff_relocation &rel = originalRelocs[j];
4810b57cec5SDimitry Andric Symbol *relocTarget = file->getSymbol(rel.SymbolTableIndex);
4820b57cec5SDimitry Andric
4830b57cec5SDimitry Andric // The estimate of the source address P should be pretty accurate,
4840b57cec5SDimitry Andric // but we don't know whether the target Symbol address should be
4850b57cec5SDimitry Andric // offset by thunksSize or not (or by some of thunksSize but not all of
4860b57cec5SDimitry Andric // it), giving us some uncertainty once we have added one thunk.
4870b57cec5SDimitry Andric uint64_t p = sc->getRVA() + rel.VirtualAddress + thunksSize;
4880b57cec5SDimitry Andric
4890b57cec5SDimitry Andric Defined *sym = dyn_cast_or_null<Defined>(relocTarget);
4900b57cec5SDimitry Andric if (!sym)
4910b57cec5SDimitry Andric continue;
4920b57cec5SDimitry Andric
4930b57cec5SDimitry Andric uint64_t s = sym->getRVA();
4940b57cec5SDimitry Andric
4950b57cec5SDimitry Andric if (isInRange(rel.Type, s, p, margin))
4960b57cec5SDimitry Andric continue;
4970b57cec5SDimitry Andric
498bdd1243dSDimitry Andric // If the target isn't in range, hook it up to an existing or new thunk.
499bdd1243dSDimitry Andric auto [thunk, wasNew] = getThunk(lastThunks, sym, p, rel.Type, margin);
5000b57cec5SDimitry Andric if (wasNew) {
5010b57cec5SDimitry Andric Chunk *thunkChunk = thunk->getChunk();
5020b57cec5SDimitry Andric thunkChunk->setRVA(
5030b57cec5SDimitry Andric thunkInsertionRVA); // Estimate of where it will be located.
5040b57cec5SDimitry Andric os->chunks.insert(os->chunks.begin() + thunkInsertionSpot, thunkChunk);
5050b57cec5SDimitry Andric thunkInsertionSpot++;
5060b57cec5SDimitry Andric thunksSize += thunkChunk->getSize();
5070b57cec5SDimitry Andric thunkInsertionRVA += thunkChunk->getSize();
5080b57cec5SDimitry Andric addressesChanged = true;
5090b57cec5SDimitry Andric }
5100b57cec5SDimitry Andric
5110b57cec5SDimitry Andric // To redirect the relocation, add a symbol to the parent object file's
5120b57cec5SDimitry Andric // symbol table, and replace the relocation symbol table index with the
5130b57cec5SDimitry Andric // new index.
5140b57cec5SDimitry Andric auto insertion = thunkSymtabIndices.insert({{file, thunk}, ~0U});
5150b57cec5SDimitry Andric uint32_t &thunkSymbolIndex = insertion.first->second;
5160b57cec5SDimitry Andric if (insertion.second)
5170b57cec5SDimitry Andric thunkSymbolIndex = file->addRangeThunkSymbol(thunk);
51806c3fb27SDimitry Andric relocReplacements.emplace_back(j, thunkSymbolIndex);
5190b57cec5SDimitry Andric }
5200b57cec5SDimitry Andric
5210b57cec5SDimitry Andric // Get a writable copy of this section's relocations so they can be
5220b57cec5SDimitry Andric // modified. If the relocations point into the object file, allocate new
5230b57cec5SDimitry Andric // memory. Otherwise, this must be previously allocated memory that can be
5240b57cec5SDimitry Andric // modified in place.
5250b57cec5SDimitry Andric ArrayRef<coff_relocation> curRelocs = sc->getRelocs();
5260b57cec5SDimitry Andric MutableArrayRef<coff_relocation> newRelocs;
5270b57cec5SDimitry Andric if (originalRelocs.data() == curRelocs.data()) {
528bdd1243dSDimitry Andric newRelocs = MutableArrayRef(
52904eeddc0SDimitry Andric bAlloc().Allocate<coff_relocation>(originalRelocs.size()),
5300b57cec5SDimitry Andric originalRelocs.size());
5310b57cec5SDimitry Andric } else {
532bdd1243dSDimitry Andric newRelocs = MutableArrayRef(
5330b57cec5SDimitry Andric const_cast<coff_relocation *>(curRelocs.data()), curRelocs.size());
5340b57cec5SDimitry Andric }
5350b57cec5SDimitry Andric
5360b57cec5SDimitry Andric // Copy each relocation, but replace the symbol table indices which need
5370b57cec5SDimitry Andric // thunks.
5380b57cec5SDimitry Andric auto nextReplacement = relocReplacements.begin();
5390b57cec5SDimitry Andric auto endReplacement = relocReplacements.end();
5400b57cec5SDimitry Andric for (size_t i = 0, e = originalRelocs.size(); i != e; ++i) {
5410b57cec5SDimitry Andric newRelocs[i] = originalRelocs[i];
5420b57cec5SDimitry Andric if (nextReplacement != endReplacement && nextReplacement->first == i) {
5430b57cec5SDimitry Andric newRelocs[i].SymbolTableIndex = nextReplacement->second;
5440b57cec5SDimitry Andric ++nextReplacement;
5450b57cec5SDimitry Andric }
5460b57cec5SDimitry Andric }
5470b57cec5SDimitry Andric
5480b57cec5SDimitry Andric sc->setRelocs(newRelocs);
5490b57cec5SDimitry Andric }
5500b57cec5SDimitry Andric return addressesChanged;
5510b57cec5SDimitry Andric }
5520b57cec5SDimitry Andric
5535f757f3fSDimitry Andric // Create a code map for CHPE metadata.
createECCodeMap()5545f757f3fSDimitry Andric void Writer::createECCodeMap() {
5555f757f3fSDimitry Andric if (!isArm64EC(ctx.config.machine))
5565f757f3fSDimitry Andric return;
5575f757f3fSDimitry Andric
5585f757f3fSDimitry Andric // Clear the map in case we were're recomputing the map after adding
5595f757f3fSDimitry Andric // a range extension thunk.
5605f757f3fSDimitry Andric codeMap.clear();
5615f757f3fSDimitry Andric
5625f757f3fSDimitry Andric std::optional<chpe_range_type> lastType;
563647cbc5dSDimitry Andric Chunk *first, *last;
5645f757f3fSDimitry Andric
5655f757f3fSDimitry Andric auto closeRange = [&]() {
5665f757f3fSDimitry Andric if (lastType) {
5675f757f3fSDimitry Andric codeMap.push_back({first, last, *lastType});
5685f757f3fSDimitry Andric lastType.reset();
5695f757f3fSDimitry Andric }
5705f757f3fSDimitry Andric };
5715f757f3fSDimitry Andric
5725f757f3fSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
5735f757f3fSDimitry Andric for (Chunk *c : sec->chunks) {
5745f757f3fSDimitry Andric // Skip empty section chunks. MS link.exe does not seem to do that and
5755f757f3fSDimitry Andric // generates empty code ranges in some cases.
5765f757f3fSDimitry Andric if (isa<SectionChunk>(c) && !c->getSize())
5775f757f3fSDimitry Andric continue;
5785f757f3fSDimitry Andric
5795f757f3fSDimitry Andric std::optional<chpe_range_type> chunkType = c->getArm64ECRangeType();
5805f757f3fSDimitry Andric if (chunkType != lastType) {
5815f757f3fSDimitry Andric closeRange();
5825f757f3fSDimitry Andric first = c;
5835f757f3fSDimitry Andric lastType = chunkType;
5845f757f3fSDimitry Andric }
5855f757f3fSDimitry Andric last = c;
5865f757f3fSDimitry Andric }
5875f757f3fSDimitry Andric }
5885f757f3fSDimitry Andric
5895f757f3fSDimitry Andric closeRange();
5905f757f3fSDimitry Andric
5915f757f3fSDimitry Andric Symbol *tableCountSym = ctx.symtab.findUnderscore("__hybrid_code_map_count");
5925f757f3fSDimitry Andric cast<DefinedAbsolute>(tableCountSym)->setVA(codeMap.size());
5935f757f3fSDimitry Andric }
5945f757f3fSDimitry Andric
5950b57cec5SDimitry Andric // Verify that all relocations are in range, with no extra margin requirements.
verifyRanges(const std::vector<Chunk * > chunks)596bdd1243dSDimitry Andric bool Writer::verifyRanges(const std::vector<Chunk *> chunks) {
5970b57cec5SDimitry Andric for (Chunk *c : chunks) {
5980b57cec5SDimitry Andric SectionChunk *sc = dyn_cast_or_null<SectionChunk>(c);
5990b57cec5SDimitry Andric if (!sc)
6000b57cec5SDimitry Andric continue;
6010b57cec5SDimitry Andric
6020b57cec5SDimitry Andric ArrayRef<coff_relocation> relocs = sc->getRelocs();
60306c3fb27SDimitry Andric for (const coff_relocation &rel : relocs) {
6040b57cec5SDimitry Andric Symbol *relocTarget = sc->file->getSymbol(rel.SymbolTableIndex);
6050b57cec5SDimitry Andric
6060b57cec5SDimitry Andric Defined *sym = dyn_cast_or_null<Defined>(relocTarget);
6070b57cec5SDimitry Andric if (!sym)
6080b57cec5SDimitry Andric continue;
6090b57cec5SDimitry Andric
6100b57cec5SDimitry Andric uint64_t p = sc->getRVA() + rel.VirtualAddress;
6110b57cec5SDimitry Andric uint64_t s = sym->getRVA();
6120b57cec5SDimitry Andric
6130b57cec5SDimitry Andric if (!isInRange(rel.Type, s, p, 0))
6140b57cec5SDimitry Andric return false;
6150b57cec5SDimitry Andric }
6160b57cec5SDimitry Andric }
6170b57cec5SDimitry Andric return true;
6180b57cec5SDimitry Andric }
6190b57cec5SDimitry Andric
6200b57cec5SDimitry Andric // Assign addresses and add thunks if necessary.
finalizeAddresses()6210b57cec5SDimitry Andric void Writer::finalizeAddresses() {
6220b57cec5SDimitry Andric assignAddresses();
623bdd1243dSDimitry Andric if (ctx.config.machine != ARMNT && ctx.config.machine != ARM64)
6240b57cec5SDimitry Andric return;
6250b57cec5SDimitry Andric
6260b57cec5SDimitry Andric size_t origNumChunks = 0;
627349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
6280b57cec5SDimitry Andric sec->origChunks = sec->chunks;
6290b57cec5SDimitry Andric origNumChunks += sec->chunks.size();
6300b57cec5SDimitry Andric }
6310b57cec5SDimitry Andric
6320b57cec5SDimitry Andric int pass = 0;
6330b57cec5SDimitry Andric int margin = 1024 * 100;
6340b57cec5SDimitry Andric while (true) {
6355f757f3fSDimitry Andric llvm::TimeTraceScope timeScope2("Add thunks pass");
6365f757f3fSDimitry Andric
6370b57cec5SDimitry Andric // First check whether we need thunks at all, or if the previous pass of
6380b57cec5SDimitry Andric // adding them turned out ok.
6390b57cec5SDimitry Andric bool rangesOk = true;
6400b57cec5SDimitry Andric size_t numChunks = 0;
6415f757f3fSDimitry Andric {
6425f757f3fSDimitry Andric llvm::TimeTraceScope timeScope3("Verify ranges");
643349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
6440b57cec5SDimitry Andric if (!verifyRanges(sec->chunks)) {
6450b57cec5SDimitry Andric rangesOk = false;
6460b57cec5SDimitry Andric break;
6470b57cec5SDimitry Andric }
6480b57cec5SDimitry Andric numChunks += sec->chunks.size();
6490b57cec5SDimitry Andric }
6505f757f3fSDimitry Andric }
6510b57cec5SDimitry Andric if (rangesOk) {
6520b57cec5SDimitry Andric if (pass > 0)
6530b57cec5SDimitry Andric log("Added " + Twine(numChunks - origNumChunks) + " thunks with " +
6540b57cec5SDimitry Andric "margin " + Twine(margin) + " in " + Twine(pass) + " passes");
6550b57cec5SDimitry Andric return;
6560b57cec5SDimitry Andric }
6570b57cec5SDimitry Andric
6580b57cec5SDimitry Andric if (pass >= 10)
6590b57cec5SDimitry Andric fatal("adding thunks hasn't converged after " + Twine(pass) + " passes");
6600b57cec5SDimitry Andric
6610b57cec5SDimitry Andric if (pass > 0) {
6620b57cec5SDimitry Andric // If the previous pass didn't work out, reset everything back to the
6630b57cec5SDimitry Andric // original conditions before retrying with a wider margin. This should
6640b57cec5SDimitry Andric // ideally never happen under real circumstances.
665349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections)
6660b57cec5SDimitry Andric sec->chunks = sec->origChunks;
6670b57cec5SDimitry Andric margin *= 2;
6680b57cec5SDimitry Andric }
6690b57cec5SDimitry Andric
6700b57cec5SDimitry Andric // Try adding thunks everywhere where it is needed, with a margin
6710b57cec5SDimitry Andric // to avoid things going out of range due to the added thunks.
6720b57cec5SDimitry Andric bool addressesChanged = false;
6735f757f3fSDimitry Andric {
6745f757f3fSDimitry Andric llvm::TimeTraceScope timeScope3("Create thunks");
675349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections)
6760b57cec5SDimitry Andric addressesChanged |= createThunks(sec, margin);
6775f757f3fSDimitry Andric }
6780b57cec5SDimitry Andric // If the verification above thought we needed thunks, we should have
6790b57cec5SDimitry Andric // added some.
6800b57cec5SDimitry Andric assert(addressesChanged);
681fe6060f1SDimitry Andric (void)addressesChanged;
6820b57cec5SDimitry Andric
6830b57cec5SDimitry Andric // Recalculate the layout for the whole image (and verify the ranges at
6840b57cec5SDimitry Andric // the start of the next round).
6850b57cec5SDimitry Andric assignAddresses();
6860b57cec5SDimitry Andric
6870b57cec5SDimitry Andric pass++;
6880b57cec5SDimitry Andric }
6890b57cec5SDimitry Andric }
6900b57cec5SDimitry Andric
writePEChecksum()691bdd1243dSDimitry Andric void Writer::writePEChecksum() {
692bdd1243dSDimitry Andric if (!ctx.config.writeCheckSum) {
693bdd1243dSDimitry Andric return;
694bdd1243dSDimitry Andric }
695bdd1243dSDimitry Andric
6965f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("PE checksum");
6975f757f3fSDimitry Andric
698bdd1243dSDimitry Andric // https://docs.microsoft.com/en-us/windows/win32/debug/pe-format#checksum
699bdd1243dSDimitry Andric uint32_t *buf = (uint32_t *)buffer->getBufferStart();
700bdd1243dSDimitry Andric uint32_t size = (uint32_t)(buffer->getBufferSize());
701bdd1243dSDimitry Andric
702bdd1243dSDimitry Andric coff_file_header *coffHeader =
703bdd1243dSDimitry Andric (coff_file_header *)((uint8_t *)buf + dosStubSize + sizeof(PEMagic));
704bdd1243dSDimitry Andric pe32_header *peHeader =
705bdd1243dSDimitry Andric (pe32_header *)((uint8_t *)coffHeader + sizeof(coff_file_header));
706bdd1243dSDimitry Andric
707bdd1243dSDimitry Andric uint64_t sum = 0;
708bdd1243dSDimitry Andric uint32_t count = size;
709bdd1243dSDimitry Andric ulittle16_t *addr = (ulittle16_t *)buf;
710bdd1243dSDimitry Andric
711bdd1243dSDimitry Andric // The PE checksum algorithm, implemented as suggested in RFC1071
712bdd1243dSDimitry Andric while (count > 1) {
713bdd1243dSDimitry Andric sum += *addr++;
714bdd1243dSDimitry Andric count -= 2;
715bdd1243dSDimitry Andric }
716bdd1243dSDimitry Andric
717bdd1243dSDimitry Andric // Add left-over byte, if any
718bdd1243dSDimitry Andric if (count > 0)
719bdd1243dSDimitry Andric sum += *(unsigned char *)addr;
720bdd1243dSDimitry Andric
721bdd1243dSDimitry Andric // Fold 32-bit sum to 16 bits
722bdd1243dSDimitry Andric while (sum >> 16) {
723bdd1243dSDimitry Andric sum = (sum & 0xffff) + (sum >> 16);
724bdd1243dSDimitry Andric }
725bdd1243dSDimitry Andric
726bdd1243dSDimitry Andric sum += size;
727bdd1243dSDimitry Andric peHeader->CheckSum = sum;
728bdd1243dSDimitry Andric }
729bdd1243dSDimitry Andric
7300b57cec5SDimitry Andric // The main function of the writer.
run()7310b57cec5SDimitry Andric void Writer::run() {
7325f757f3fSDimitry Andric {
7335f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Write PE");
734349cc55cSDimitry Andric ScopedTimer t1(ctx.codeLayoutTimer);
7350b57cec5SDimitry Andric
7360b57cec5SDimitry Andric createImportTables();
7370b57cec5SDimitry Andric createSections();
7380b57cec5SDimitry Andric appendImportThunks();
7395f757f3fSDimitry Andric // Import thunks must be added before the Control Flow Guard tables are
7405f757f3fSDimitry Andric // added.
741e8d8bef9SDimitry Andric createMiscChunks();
7420b57cec5SDimitry Andric createExportTable();
7430b57cec5SDimitry Andric mergeSections();
7445f757f3fSDimitry Andric sortECChunks();
7450b57cec5SDimitry Andric removeUnusedSections();
7460b57cec5SDimitry Andric finalizeAddresses();
7470b57cec5SDimitry Andric removeEmptySections();
7480b57cec5SDimitry Andric assignOutputSectionIndices();
7490b57cec5SDimitry Andric setSectionPermissions();
7505f757f3fSDimitry Andric setECSymbols();
7510b57cec5SDimitry Andric createSymbolAndStringTable();
7520b57cec5SDimitry Andric
7530b57cec5SDimitry Andric if (fileSize > UINT32_MAX)
7540b57cec5SDimitry Andric fatal("image size (" + Twine(fileSize) + ") " +
7550b57cec5SDimitry Andric "exceeds maximum allowable size (" + Twine(UINT32_MAX) + ")");
7560b57cec5SDimitry Andric
757bdd1243dSDimitry Andric openFile(ctx.config.outputFile);
758bdd1243dSDimitry Andric if (ctx.config.is64()) {
7590b57cec5SDimitry Andric writeHeader<pe32plus_header>();
7600b57cec5SDimitry Andric } else {
7610b57cec5SDimitry Andric writeHeader<pe32_header>();
7620b57cec5SDimitry Andric }
7630b57cec5SDimitry Andric writeSections();
7645f757f3fSDimitry Andric prepareLoadConfig();
7655f757f3fSDimitry Andric sortExceptionTables();
7660b57cec5SDimitry Andric
767e8d8bef9SDimitry Andric // Fix up the alignment in the TLS Directory's characteristic field,
768e8d8bef9SDimitry Andric // if a specific alignment value is needed
769e8d8bef9SDimitry Andric if (tlsAlignment)
770e8d8bef9SDimitry Andric fixTlsAlignment();
7715f757f3fSDimitry Andric }
7720b57cec5SDimitry Andric
773bdd1243dSDimitry Andric if (!ctx.config.pdbPath.empty() && ctx.config.debug) {
7740b57cec5SDimitry Andric assert(buildId);
775349cc55cSDimitry Andric createPDB(ctx, sectionTable, buildId->buildId);
7760b57cec5SDimitry Andric }
7770b57cec5SDimitry Andric writeBuildId();
7780b57cec5SDimitry Andric
779349cc55cSDimitry Andric writeLLDMapFile(ctx);
780349cc55cSDimitry Andric writeMapFile(ctx);
7810b57cec5SDimitry Andric
782bdd1243dSDimitry Andric writePEChecksum();
783bdd1243dSDimitry Andric
7840b57cec5SDimitry Andric if (errorCount())
7850b57cec5SDimitry Andric return;
7860b57cec5SDimitry Andric
7875f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Commit PE to disk");
788349cc55cSDimitry Andric ScopedTimer t2(ctx.outputCommitTimer);
7890b57cec5SDimitry Andric if (auto e = buffer->commit())
790bdd1243dSDimitry Andric fatal("failed to write output '" + buffer->getPath() +
791bdd1243dSDimitry Andric "': " + toString(std::move(e)));
7920b57cec5SDimitry Andric }
7930b57cec5SDimitry Andric
getOutputSectionName(StringRef name)7940b57cec5SDimitry Andric static StringRef getOutputSectionName(StringRef name) {
7950b57cec5SDimitry Andric StringRef s = name.split('$').first;
7960b57cec5SDimitry Andric
7970b57cec5SDimitry Andric // Treat a later period as a separator for MinGW, for sections like
7980b57cec5SDimitry Andric // ".ctors.01234".
7990b57cec5SDimitry Andric return s.substr(0, s.find('.', 1));
8000b57cec5SDimitry Andric }
8010b57cec5SDimitry Andric
8020b57cec5SDimitry Andric // For /order.
sortBySectionOrder(std::vector<Chunk * > & chunks)803bdd1243dSDimitry Andric void Writer::sortBySectionOrder(std::vector<Chunk *> &chunks) {
804bdd1243dSDimitry Andric auto getPriority = [&ctx = ctx](const Chunk *c) {
8050b57cec5SDimitry Andric if (auto *sec = dyn_cast<SectionChunk>(c))
8060b57cec5SDimitry Andric if (sec->sym)
807bdd1243dSDimitry Andric return ctx.config.order.lookup(sec->sym->getName());
8080b57cec5SDimitry Andric return 0;
8090b57cec5SDimitry Andric };
8100b57cec5SDimitry Andric
8110b57cec5SDimitry Andric llvm::stable_sort(chunks, [=](const Chunk *a, const Chunk *b) {
8120b57cec5SDimitry Andric return getPriority(a) < getPriority(b);
8130b57cec5SDimitry Andric });
8140b57cec5SDimitry Andric }
8150b57cec5SDimitry Andric
8160b57cec5SDimitry Andric // Change the characteristics of existing PartialSections that belong to the
8170b57cec5SDimitry Andric // section Name to Chars.
fixPartialSectionChars(StringRef name,uint32_t chars)8180b57cec5SDimitry Andric void Writer::fixPartialSectionChars(StringRef name, uint32_t chars) {
8190b57cec5SDimitry Andric for (auto it : partialSections) {
8200b57cec5SDimitry Andric PartialSection *pSec = it.second;
8210b57cec5SDimitry Andric StringRef curName = pSec->name;
8220b57cec5SDimitry Andric if (!curName.consume_front(name) ||
82306c3fb27SDimitry Andric (!curName.empty() && !curName.starts_with("$")))
8240b57cec5SDimitry Andric continue;
8250b57cec5SDimitry Andric if (pSec->characteristics == chars)
8260b57cec5SDimitry Andric continue;
8270b57cec5SDimitry Andric PartialSection *destSec = createPartialSection(pSec->name, chars);
8280b57cec5SDimitry Andric destSec->chunks.insert(destSec->chunks.end(), pSec->chunks.begin(),
8290b57cec5SDimitry Andric pSec->chunks.end());
8300b57cec5SDimitry Andric pSec->chunks.clear();
8310b57cec5SDimitry Andric }
8320b57cec5SDimitry Andric }
8330b57cec5SDimitry Andric
8340b57cec5SDimitry Andric // Sort concrete section chunks from GNU import libraries.
8350b57cec5SDimitry Andric //
8360b57cec5SDimitry Andric // GNU binutils doesn't use short import files, but instead produces import
8370b57cec5SDimitry Andric // libraries that consist of object files, with section chunks for the .idata$*
8380b57cec5SDimitry Andric // sections. These are linked just as regular static libraries. Each import
8390b57cec5SDimitry Andric // library consists of one header object, one object file for every imported
8400b57cec5SDimitry Andric // symbol, and one trailer object. In order for the .idata tables/lists to
8410b57cec5SDimitry Andric // be formed correctly, the section chunks within each .idata$* section need
8420b57cec5SDimitry Andric // to be grouped by library, and sorted alphabetically within each library
8430b57cec5SDimitry Andric // (which makes sure the header comes first and the trailer last).
fixGnuImportChunks()8440b57cec5SDimitry Andric bool Writer::fixGnuImportChunks() {
8450b57cec5SDimitry Andric uint32_t rdata = IMAGE_SCN_CNT_INITIALIZED_DATA | IMAGE_SCN_MEM_READ;
8460b57cec5SDimitry Andric
8470b57cec5SDimitry Andric // Make sure all .idata$* section chunks are mapped as RDATA in order to
8480b57cec5SDimitry Andric // be sorted into the same sections as our own synthesized .idata chunks.
8490b57cec5SDimitry Andric fixPartialSectionChars(".idata", rdata);
8500b57cec5SDimitry Andric
8510b57cec5SDimitry Andric bool hasIdata = false;
8520b57cec5SDimitry Andric // Sort all .idata$* chunks, grouping chunks from the same library,
85381ad6265SDimitry Andric // with alphabetical ordering of the object files within a library.
8540b57cec5SDimitry Andric for (auto it : partialSections) {
8550b57cec5SDimitry Andric PartialSection *pSec = it.second;
85606c3fb27SDimitry Andric if (!pSec->name.starts_with(".idata"))
8570b57cec5SDimitry Andric continue;
8580b57cec5SDimitry Andric
8590b57cec5SDimitry Andric if (!pSec->chunks.empty())
8600b57cec5SDimitry Andric hasIdata = true;
8610b57cec5SDimitry Andric llvm::stable_sort(pSec->chunks, [&](Chunk *s, Chunk *t) {
8620b57cec5SDimitry Andric SectionChunk *sc1 = dyn_cast_or_null<SectionChunk>(s);
8630b57cec5SDimitry Andric SectionChunk *sc2 = dyn_cast_or_null<SectionChunk>(t);
8640b57cec5SDimitry Andric if (!sc1 || !sc2) {
8650b57cec5SDimitry Andric // if SC1, order them ascending. If SC2 or both null,
8660b57cec5SDimitry Andric // S is not less than T.
8670b57cec5SDimitry Andric return sc1 != nullptr;
8680b57cec5SDimitry Andric }
8690b57cec5SDimitry Andric // Make a string with "libraryname/objectfile" for sorting, achieving
8700b57cec5SDimitry Andric // both grouping by library and sorting of objects within a library,
8710b57cec5SDimitry Andric // at once.
8720b57cec5SDimitry Andric std::string key1 =
8730b57cec5SDimitry Andric (sc1->file->parentName + "/" + sc1->file->getName()).str();
8740b57cec5SDimitry Andric std::string key2 =
8750b57cec5SDimitry Andric (sc2->file->parentName + "/" + sc2->file->getName()).str();
8760b57cec5SDimitry Andric return key1 < key2;
8770b57cec5SDimitry Andric });
8780b57cec5SDimitry Andric }
8790b57cec5SDimitry Andric return hasIdata;
8800b57cec5SDimitry Andric }
8810b57cec5SDimitry Andric
8820b57cec5SDimitry Andric // Add generated idata chunks, for imported symbols and DLLs, and a
8830b57cec5SDimitry Andric // terminator in .idata$2.
addSyntheticIdata()8840b57cec5SDimitry Andric void Writer::addSyntheticIdata() {
8850b57cec5SDimitry Andric uint32_t rdata = IMAGE_SCN_CNT_INITIALIZED_DATA | IMAGE_SCN_MEM_READ;
886bdd1243dSDimitry Andric idata.create(ctx);
8870b57cec5SDimitry Andric
8880b57cec5SDimitry Andric // Add the .idata content in the right section groups, to allow
8890b57cec5SDimitry Andric // chunks from other linked in object files to be grouped together.
8900b57cec5SDimitry Andric // See Microsoft PE/COFF spec 5.4 for details.
8910b57cec5SDimitry Andric auto add = [&](StringRef n, std::vector<Chunk *> &v) {
8920b57cec5SDimitry Andric PartialSection *pSec = createPartialSection(n, rdata);
8930b57cec5SDimitry Andric pSec->chunks.insert(pSec->chunks.end(), v.begin(), v.end());
8940b57cec5SDimitry Andric };
8950b57cec5SDimitry Andric
8960b57cec5SDimitry Andric // The loader assumes a specific order of data.
8970b57cec5SDimitry Andric // Add each type in the correct order.
8980b57cec5SDimitry Andric add(".idata$2", idata.dirs);
8990b57cec5SDimitry Andric add(".idata$4", idata.lookups);
9000b57cec5SDimitry Andric add(".idata$5", idata.addresses);
90185868e8aSDimitry Andric if (!idata.hints.empty())
9020b57cec5SDimitry Andric add(".idata$6", idata.hints);
9030b57cec5SDimitry Andric add(".idata$7", idata.dllNames);
9040b57cec5SDimitry Andric }
9050b57cec5SDimitry Andric
9060b57cec5SDimitry Andric // Locate the first Chunk and size of the import directory list and the
9070b57cec5SDimitry Andric // IAT.
locateImportTables()9080b57cec5SDimitry Andric void Writer::locateImportTables() {
9090b57cec5SDimitry Andric uint32_t rdata = IMAGE_SCN_CNT_INITIALIZED_DATA | IMAGE_SCN_MEM_READ;
9100b57cec5SDimitry Andric
9110b57cec5SDimitry Andric if (PartialSection *importDirs = findPartialSection(".idata$2", rdata)) {
9120b57cec5SDimitry Andric if (!importDirs->chunks.empty())
9130b57cec5SDimitry Andric importTableStart = importDirs->chunks.front();
9140b57cec5SDimitry Andric for (Chunk *c : importDirs->chunks)
9150b57cec5SDimitry Andric importTableSize += c->getSize();
9160b57cec5SDimitry Andric }
9170b57cec5SDimitry Andric
9180b57cec5SDimitry Andric if (PartialSection *importAddresses = findPartialSection(".idata$5", rdata)) {
9190b57cec5SDimitry Andric if (!importAddresses->chunks.empty())
9200b57cec5SDimitry Andric iatStart = importAddresses->chunks.front();
9210b57cec5SDimitry Andric for (Chunk *c : importAddresses->chunks)
9220b57cec5SDimitry Andric iatSize += c->getSize();
9230b57cec5SDimitry Andric }
9240b57cec5SDimitry Andric }
9250b57cec5SDimitry Andric
9260b57cec5SDimitry Andric // Return whether a SectionChunk's suffix (the dollar and any trailing
9270b57cec5SDimitry Andric // suffix) should be removed and sorted into the main suffixless
9280b57cec5SDimitry Andric // PartialSection.
shouldStripSectionSuffix(SectionChunk * sc,StringRef name,bool isMinGW)929bdd1243dSDimitry Andric static bool shouldStripSectionSuffix(SectionChunk *sc, StringRef name,
930bdd1243dSDimitry Andric bool isMinGW) {
9310b57cec5SDimitry Andric // On MinGW, comdat groups are formed by putting the comdat group name
9320b57cec5SDimitry Andric // after the '$' in the section name. For .eh_frame$<symbol>, that must
9330b57cec5SDimitry Andric // still be sorted before the .eh_frame trailer from crtend.o, thus just
9340b57cec5SDimitry Andric // strip the section name trailer. For other sections, such as
9350b57cec5SDimitry Andric // .tls$$<symbol> (where non-comdat .tls symbols are otherwise stored in
9360b57cec5SDimitry Andric // ".tls$"), they must be strictly sorted after .tls. And for the
9370b57cec5SDimitry Andric // hypothetical case of comdat .CRT$XCU, we definitely need to keep the
9380b57cec5SDimitry Andric // suffix for sorting. Thus, to play it safe, only strip the suffix for
9390b57cec5SDimitry Andric // the standard sections.
940bdd1243dSDimitry Andric if (!isMinGW)
9410b57cec5SDimitry Andric return false;
9420b57cec5SDimitry Andric if (!sc || !sc->isCOMDAT())
9430b57cec5SDimitry Andric return false;
94406c3fb27SDimitry Andric return name.starts_with(".text$") || name.starts_with(".data$") ||
94506c3fb27SDimitry Andric name.starts_with(".rdata$") || name.starts_with(".pdata$") ||
94606c3fb27SDimitry Andric name.starts_with(".xdata$") || name.starts_with(".eh_frame$");
9470b57cec5SDimitry Andric }
9480b57cec5SDimitry Andric
sortSections()949e8d8bef9SDimitry Andric void Writer::sortSections() {
950bdd1243dSDimitry Andric if (!ctx.config.callGraphProfile.empty()) {
951349cc55cSDimitry Andric DenseMap<const SectionChunk *, int> order =
952349cc55cSDimitry Andric computeCallGraphProfileOrder(ctx);
953e8d8bef9SDimitry Andric for (auto it : order) {
954e8d8bef9SDimitry Andric if (DefinedRegular *sym = it.first->sym)
955bdd1243dSDimitry Andric ctx.config.order[sym->getName()] = it.second;
956e8d8bef9SDimitry Andric }
957e8d8bef9SDimitry Andric }
958bdd1243dSDimitry Andric if (!ctx.config.order.empty())
959e8d8bef9SDimitry Andric for (auto it : partialSections)
960e8d8bef9SDimitry Andric sortBySectionOrder(it.second->chunks);
961e8d8bef9SDimitry Andric }
962e8d8bef9SDimitry Andric
9630b57cec5SDimitry Andric // Create output section objects and add them to OutputSections.
createSections()9640b57cec5SDimitry Andric void Writer::createSections() {
9655f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Output sections");
9660b57cec5SDimitry Andric // First, create the builtin sections.
9670b57cec5SDimitry Andric const uint32_t data = IMAGE_SCN_CNT_INITIALIZED_DATA;
9680b57cec5SDimitry Andric const uint32_t bss = IMAGE_SCN_CNT_UNINITIALIZED_DATA;
9690b57cec5SDimitry Andric const uint32_t code = IMAGE_SCN_CNT_CODE;
9700b57cec5SDimitry Andric const uint32_t discardable = IMAGE_SCN_MEM_DISCARDABLE;
9710b57cec5SDimitry Andric const uint32_t r = IMAGE_SCN_MEM_READ;
9720b57cec5SDimitry Andric const uint32_t w = IMAGE_SCN_MEM_WRITE;
9730b57cec5SDimitry Andric const uint32_t x = IMAGE_SCN_MEM_EXECUTE;
9740b57cec5SDimitry Andric
9750b57cec5SDimitry Andric SmallDenseMap<std::pair<StringRef, uint32_t>, OutputSection *> sections;
9760b57cec5SDimitry Andric auto createSection = [&](StringRef name, uint32_t outChars) {
9770b57cec5SDimitry Andric OutputSection *&sec = sections[{name, outChars}];
9780b57cec5SDimitry Andric if (!sec) {
9790b57cec5SDimitry Andric sec = make<OutputSection>(name, outChars);
980349cc55cSDimitry Andric ctx.outputSections.push_back(sec);
9810b57cec5SDimitry Andric }
9820b57cec5SDimitry Andric return sec;
9830b57cec5SDimitry Andric };
9840b57cec5SDimitry Andric
9850b57cec5SDimitry Andric // Try to match the section order used by link.exe.
9860b57cec5SDimitry Andric textSec = createSection(".text", code | r | x);
9870b57cec5SDimitry Andric createSection(".bss", bss | r | w);
9880b57cec5SDimitry Andric rdataSec = createSection(".rdata", data | r);
9890b57cec5SDimitry Andric buildidSec = createSection(".buildid", data | r);
9900b57cec5SDimitry Andric dataSec = createSection(".data", data | r | w);
9910b57cec5SDimitry Andric pdataSec = createSection(".pdata", data | r);
9920b57cec5SDimitry Andric idataSec = createSection(".idata", data | r);
9930b57cec5SDimitry Andric edataSec = createSection(".edata", data | r);
9940b57cec5SDimitry Andric didatSec = createSection(".didat", data | r);
9950b57cec5SDimitry Andric rsrcSec = createSection(".rsrc", data | r);
9960b57cec5SDimitry Andric relocSec = createSection(".reloc", data | discardable | r);
9970b57cec5SDimitry Andric ctorsSec = createSection(".ctors", data | r | w);
9980b57cec5SDimitry Andric dtorsSec = createSection(".dtors", data | r | w);
9990b57cec5SDimitry Andric
10000b57cec5SDimitry Andric // Then bin chunks by name and output characteristics.
1001349cc55cSDimitry Andric for (Chunk *c : ctx.symtab.getChunks()) {
10020b57cec5SDimitry Andric auto *sc = dyn_cast<SectionChunk>(c);
10030b57cec5SDimitry Andric if (sc && !sc->live) {
1004bdd1243dSDimitry Andric if (ctx.config.verbose)
10050b57cec5SDimitry Andric sc->printDiscardedMessage();
10060b57cec5SDimitry Andric continue;
10070b57cec5SDimitry Andric }
10080b57cec5SDimitry Andric StringRef name = c->getSectionName();
1009bdd1243dSDimitry Andric if (shouldStripSectionSuffix(sc, name, ctx.config.mingw))
10100b57cec5SDimitry Andric name = name.split('$').first;
1011e8d8bef9SDimitry Andric
101206c3fb27SDimitry Andric if (name.starts_with(".tls"))
1013e8d8bef9SDimitry Andric tlsAlignment = std::max(tlsAlignment, c->getAlignment());
1014e8d8bef9SDimitry Andric
10150b57cec5SDimitry Andric PartialSection *pSec = createPartialSection(name,
10160b57cec5SDimitry Andric c->getOutputCharacteristics());
10170b57cec5SDimitry Andric pSec->chunks.push_back(c);
10180b57cec5SDimitry Andric }
10190b57cec5SDimitry Andric
10200b57cec5SDimitry Andric fixPartialSectionChars(".rsrc", data | r);
102185868e8aSDimitry Andric fixPartialSectionChars(".edata", data | r);
10220b57cec5SDimitry Andric // Even in non MinGW cases, we might need to link against GNU import
10230b57cec5SDimitry Andric // libraries.
10240b57cec5SDimitry Andric bool hasIdata = fixGnuImportChunks();
10250b57cec5SDimitry Andric if (!idata.empty())
10260b57cec5SDimitry Andric hasIdata = true;
10270b57cec5SDimitry Andric
10280b57cec5SDimitry Andric if (hasIdata)
10290b57cec5SDimitry Andric addSyntheticIdata();
10300b57cec5SDimitry Andric
1031e8d8bef9SDimitry Andric sortSections();
10320b57cec5SDimitry Andric
10330b57cec5SDimitry Andric if (hasIdata)
10340b57cec5SDimitry Andric locateImportTables();
10350b57cec5SDimitry Andric
10360b57cec5SDimitry Andric // Then create an OutputSection for each section.
10370b57cec5SDimitry Andric // '$' and all following characters in input section names are
10380b57cec5SDimitry Andric // discarded when determining output section. So, .text$foo
10390b57cec5SDimitry Andric // contributes to .text, for example. See PE/COFF spec 3.2.
10400b57cec5SDimitry Andric for (auto it : partialSections) {
10410b57cec5SDimitry Andric PartialSection *pSec = it.second;
10420b57cec5SDimitry Andric StringRef name = getOutputSectionName(pSec->name);
10430b57cec5SDimitry Andric uint32_t outChars = pSec->characteristics;
10440b57cec5SDimitry Andric
10450b57cec5SDimitry Andric if (name == ".CRT") {
10460b57cec5SDimitry Andric // In link.exe, there is a special case for the I386 target where .CRT
10470b57cec5SDimitry Andric // sections are treated as if they have output characteristics DATA | R if
10480b57cec5SDimitry Andric // their characteristics are DATA | R | W. This implements the same
10490b57cec5SDimitry Andric // special case for all architectures.
10500b57cec5SDimitry Andric outChars = data | r;
10510b57cec5SDimitry Andric
10520b57cec5SDimitry Andric log("Processing section " + pSec->name + " -> " + name);
10530b57cec5SDimitry Andric
10540b57cec5SDimitry Andric sortCRTSectionChunks(pSec->chunks);
10550b57cec5SDimitry Andric }
10560b57cec5SDimitry Andric
10570b57cec5SDimitry Andric OutputSection *sec = createSection(name, outChars);
10580b57cec5SDimitry Andric for (Chunk *c : pSec->chunks)
10590b57cec5SDimitry Andric sec->addChunk(c);
10600b57cec5SDimitry Andric
10610b57cec5SDimitry Andric sec->addContributingPartialSection(pSec);
10620b57cec5SDimitry Andric }
10630b57cec5SDimitry Andric
10640b57cec5SDimitry Andric // Finally, move some output sections to the end.
10650b57cec5SDimitry Andric auto sectionOrder = [&](const OutputSection *s) {
10660b57cec5SDimitry Andric // Move DISCARDABLE (or non-memory-mapped) sections to the end of file
10670b57cec5SDimitry Andric // because the loader cannot handle holes. Stripping can remove other
10680b57cec5SDimitry Andric // discardable ones than .reloc, which is first of them (created early).
1069fb03ea46SDimitry Andric if (s->header.Characteristics & IMAGE_SCN_MEM_DISCARDABLE) {
1070fb03ea46SDimitry Andric // Move discardable sections named .debug_ to the end, after other
1071fb03ea46SDimitry Andric // discardable sections. Stripping only removes the sections named
1072fb03ea46SDimitry Andric // .debug_* - thus try to avoid leaving holes after stripping.
107306c3fb27SDimitry Andric if (s->name.starts_with(".debug_"))
1074fb03ea46SDimitry Andric return 3;
10750b57cec5SDimitry Andric return 2;
1076fb03ea46SDimitry Andric }
10770b57cec5SDimitry Andric // .rsrc should come at the end of the non-discardable sections because its
10780b57cec5SDimitry Andric // size may change by the Win32 UpdateResources() function, causing
10790b57cec5SDimitry Andric // subsequent sections to move (see https://crbug.com/827082).
10800b57cec5SDimitry Andric if (s == rsrcSec)
10810b57cec5SDimitry Andric return 1;
10820b57cec5SDimitry Andric return 0;
10830b57cec5SDimitry Andric };
1084349cc55cSDimitry Andric llvm::stable_sort(ctx.outputSections,
10850b57cec5SDimitry Andric [&](const OutputSection *s, const OutputSection *t) {
10860b57cec5SDimitry Andric return sectionOrder(s) < sectionOrder(t);
10870b57cec5SDimitry Andric });
10880b57cec5SDimitry Andric }
10890b57cec5SDimitry Andric
createMiscChunks()10900b57cec5SDimitry Andric void Writer::createMiscChunks() {
10915f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Misc chunks");
1092bdd1243dSDimitry Andric Configuration *config = &ctx.config;
1093bdd1243dSDimitry Andric
1094349cc55cSDimitry Andric for (MergeChunk *p : ctx.mergeChunkInstances) {
10950b57cec5SDimitry Andric if (p) {
10960b57cec5SDimitry Andric p->finalizeContents();
10970b57cec5SDimitry Andric rdataSec->addChunk(p);
10980b57cec5SDimitry Andric }
10990b57cec5SDimitry Andric }
11000b57cec5SDimitry Andric
11010b57cec5SDimitry Andric // Create thunks for locally-dllimported symbols.
1102349cc55cSDimitry Andric if (!ctx.symtab.localImportChunks.empty()) {
1103349cc55cSDimitry Andric for (Chunk *c : ctx.symtab.localImportChunks)
11040b57cec5SDimitry Andric rdataSec->addChunk(c);
11050b57cec5SDimitry Andric }
11060b57cec5SDimitry Andric
11070b57cec5SDimitry Andric // Create Debug Information Chunks
11085f757f3fSDimitry Andric debugInfoSec = config->mingw ? buildidSec : rdataSec;
11095f757f3fSDimitry Andric if (config->buildIDHash != BuildIDHash::None || config->debug ||
11105f757f3fSDimitry Andric config->repro || config->cetCompat) {
1111349cc55cSDimitry Andric debugDirectory =
1112349cc55cSDimitry Andric make<DebugDirectoryChunk>(ctx, debugRecords, config->repro);
11135ffd83dbSDimitry Andric debugDirectory->setAlignment(4);
11140b57cec5SDimitry Andric debugInfoSec->addChunk(debugDirectory);
11150b57cec5SDimitry Andric }
11160b57cec5SDimitry Andric
11175f757f3fSDimitry Andric if (config->debug || config->buildIDHash != BuildIDHash::None) {
11180b57cec5SDimitry Andric // Make a CVDebugRecordChunk even when /DEBUG:CV is not specified. We
11190b57cec5SDimitry Andric // output a PDB no matter what, and this chunk provides the only means of
11200b57cec5SDimitry Andric // allowing a debugger to match a PDB and an executable. So we need it even
11210b57cec5SDimitry Andric // if we're ultimately not going to write CodeView data to the PDB.
1122bdd1243dSDimitry Andric buildId = make<CVDebugRecordChunk>(ctx);
112306c3fb27SDimitry Andric debugRecords.emplace_back(COFF::IMAGE_DEBUG_TYPE_CODEVIEW, buildId);
11245f757f3fSDimitry Andric if (Symbol *buildidSym = ctx.symtab.findUnderscore("__buildid"))
11255f757f3fSDimitry Andric replaceSymbol<DefinedSynthetic>(buildidSym, buildidSym->getName(),
11265f757f3fSDimitry Andric buildId, 4);
11275ffd83dbSDimitry Andric }
11280b57cec5SDimitry Andric
11295ffd83dbSDimitry Andric if (config->cetCompat) {
113006c3fb27SDimitry Andric debugRecords.emplace_back(COFF::IMAGE_DEBUG_TYPE_EX_DLLCHARACTERISTICS,
11315ffd83dbSDimitry Andric make<ExtendedDllCharacteristicsChunk>(
113206c3fb27SDimitry Andric IMAGE_DLL_CHARACTERISTICS_EX_CET_COMPAT));
11335ffd83dbSDimitry Andric }
11345ffd83dbSDimitry Andric
1135e8d8bef9SDimitry Andric // Align and add each chunk referenced by the debug data directory.
1136e8d8bef9SDimitry Andric for (std::pair<COFF::DebugType, Chunk *> r : debugRecords) {
1137e8d8bef9SDimitry Andric r.second->setAlignment(4);
11385ffd83dbSDimitry Andric debugInfoSec->addChunk(r.second);
11390b57cec5SDimitry Andric }
11400b57cec5SDimitry Andric
11410b57cec5SDimitry Andric // Create SEH table. x86-only.
11420b57cec5SDimitry Andric if (config->safeSEH)
11430b57cec5SDimitry Andric createSEHTable();
11440b57cec5SDimitry Andric
11450b57cec5SDimitry Andric // Create /guard:cf tables if requested.
11460b57cec5SDimitry Andric if (config->guardCF != GuardCFLevel::Off)
11470b57cec5SDimitry Andric createGuardCFTables();
11480b57cec5SDimitry Andric
11495f757f3fSDimitry Andric if (isArm64EC(config->machine))
11505f757f3fSDimitry Andric createECChunks();
11515f757f3fSDimitry Andric
11525ffd83dbSDimitry Andric if (config->autoImport)
11530b57cec5SDimitry Andric createRuntimePseudoRelocs();
11540b57cec5SDimitry Andric
11555ffd83dbSDimitry Andric if (config->mingw)
11560b57cec5SDimitry Andric insertCtorDtorSymbols();
11570b57cec5SDimitry Andric }
11580b57cec5SDimitry Andric
11590b57cec5SDimitry Andric // Create .idata section for the DLL-imported symbol table.
11600b57cec5SDimitry Andric // The format of this section is inherently Windows-specific.
11610b57cec5SDimitry Andric // IdataContents class abstracted away the details for us,
11620b57cec5SDimitry Andric // so we just let it create chunks and add them to the section.
createImportTables()11630b57cec5SDimitry Andric void Writer::createImportTables() {
11645f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Import tables");
11650b57cec5SDimitry Andric // Initialize DLLOrder so that import entries are ordered in
11660b57cec5SDimitry Andric // the same order as in the command line. (That affects DLL
11670b57cec5SDimitry Andric // initialization order, and this ordering is MSVC-compatible.)
1168349cc55cSDimitry Andric for (ImportFile *file : ctx.importFileInstances) {
11690b57cec5SDimitry Andric if (!file->live)
11700b57cec5SDimitry Andric continue;
11710b57cec5SDimitry Andric
11720b57cec5SDimitry Andric std::string dll = StringRef(file->dllName).lower();
1173bdd1243dSDimitry Andric if (ctx.config.dllOrder.count(dll) == 0)
1174bdd1243dSDimitry Andric ctx.config.dllOrder[dll] = ctx.config.dllOrder.size();
11750b57cec5SDimitry Andric
11760b57cec5SDimitry Andric if (file->impSym && !isa<DefinedImportData>(file->impSym))
1177bdd1243dSDimitry Andric fatal(toString(ctx, *file->impSym) + " was replaced");
11780b57cec5SDimitry Andric DefinedImportData *impSym = cast_or_null<DefinedImportData>(file->impSym);
1179bdd1243dSDimitry Andric if (ctx.config.delayLoads.count(StringRef(file->dllName).lower())) {
11800b57cec5SDimitry Andric if (!file->thunkSym)
11810b57cec5SDimitry Andric fatal("cannot delay-load " + toString(file) +
1182bdd1243dSDimitry Andric " due to import of data: " + toString(ctx, *impSym));
11830b57cec5SDimitry Andric delayIdata.add(impSym);
11840b57cec5SDimitry Andric } else {
11850b57cec5SDimitry Andric idata.add(impSym);
11860b57cec5SDimitry Andric }
11870b57cec5SDimitry Andric }
11880b57cec5SDimitry Andric }
11890b57cec5SDimitry Andric
appendImportThunks()11900b57cec5SDimitry Andric void Writer::appendImportThunks() {
1191349cc55cSDimitry Andric if (ctx.importFileInstances.empty())
11920b57cec5SDimitry Andric return;
11930b57cec5SDimitry Andric
11945f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Import thunks");
1195349cc55cSDimitry Andric for (ImportFile *file : ctx.importFileInstances) {
11960b57cec5SDimitry Andric if (!file->live)
11970b57cec5SDimitry Andric continue;
11980b57cec5SDimitry Andric
11990b57cec5SDimitry Andric if (!file->thunkSym)
12000b57cec5SDimitry Andric continue;
12010b57cec5SDimitry Andric
12020b57cec5SDimitry Andric if (!isa<DefinedImportThunk>(file->thunkSym))
1203bdd1243dSDimitry Andric fatal(toString(ctx, *file->thunkSym) + " was replaced");
12040b57cec5SDimitry Andric DefinedImportThunk *thunk = cast<DefinedImportThunk>(file->thunkSym);
12050b57cec5SDimitry Andric if (file->thunkLive)
12060b57cec5SDimitry Andric textSec->addChunk(thunk->getChunk());
12070b57cec5SDimitry Andric }
12080b57cec5SDimitry Andric
12090b57cec5SDimitry Andric if (!delayIdata.empty()) {
1210bdd1243dSDimitry Andric Defined *helper = cast<Defined>(ctx.config.delayLoadHelper);
1211bdd1243dSDimitry Andric delayIdata.create(helper);
12120b57cec5SDimitry Andric for (Chunk *c : delayIdata.getChunks())
12130b57cec5SDimitry Andric didatSec->addChunk(c);
12140b57cec5SDimitry Andric for (Chunk *c : delayIdata.getDataChunks())
12150b57cec5SDimitry Andric dataSec->addChunk(c);
12160b57cec5SDimitry Andric for (Chunk *c : delayIdata.getCodeChunks())
12170b57cec5SDimitry Andric textSec->addChunk(c);
1218bdd1243dSDimitry Andric for (Chunk *c : delayIdata.getCodePData())
1219bdd1243dSDimitry Andric pdataSec->addChunk(c);
1220bdd1243dSDimitry Andric for (Chunk *c : delayIdata.getCodeUnwindInfo())
1221bdd1243dSDimitry Andric rdataSec->addChunk(c);
12220b57cec5SDimitry Andric }
12230b57cec5SDimitry Andric }
12240b57cec5SDimitry Andric
createExportTable()12250b57cec5SDimitry Andric void Writer::createExportTable() {
12265f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Export table");
122785868e8aSDimitry Andric if (!edataSec->chunks.empty()) {
122885868e8aSDimitry Andric // Allow using a custom built export table from input object files, instead
122985868e8aSDimitry Andric // of having the linker synthesize the tables.
1230bdd1243dSDimitry Andric if (ctx.config.hadExplicitExports)
123185868e8aSDimitry Andric warn("literal .edata sections override exports");
1232bdd1243dSDimitry Andric } else if (!ctx.config.exports.empty()) {
12330b57cec5SDimitry Andric for (Chunk *c : edata.chunks)
12340b57cec5SDimitry Andric edataSec->addChunk(c);
12350b57cec5SDimitry Andric }
123685868e8aSDimitry Andric if (!edataSec->chunks.empty()) {
123785868e8aSDimitry Andric edataStart = edataSec->chunks.front();
123885868e8aSDimitry Andric edataEnd = edataSec->chunks.back();
123985868e8aSDimitry Andric }
1240fe6060f1SDimitry Andric // Warn on exported deleting destructor.
1241bdd1243dSDimitry Andric for (auto e : ctx.config.exports)
124206c3fb27SDimitry Andric if (e.sym && e.sym->getName().starts_with("??_G"))
1243bdd1243dSDimitry Andric warn("export of deleting dtor: " + toString(ctx, *e.sym));
124485868e8aSDimitry Andric }
12450b57cec5SDimitry Andric
removeUnusedSections()12460b57cec5SDimitry Andric void Writer::removeUnusedSections() {
12475f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Remove unused sections");
12480b57cec5SDimitry Andric // Remove sections that we can be sure won't get content, to avoid
12490b57cec5SDimitry Andric // allocating space for their section headers.
12500b57cec5SDimitry Andric auto isUnused = [this](OutputSection *s) {
12510b57cec5SDimitry Andric if (s == relocSec)
12520b57cec5SDimitry Andric return false; // This section is populated later.
12530b57cec5SDimitry Andric // MergeChunks have zero size at this point, as their size is finalized
12540b57cec5SDimitry Andric // later. Only remove sections that have no Chunks at all.
12550b57cec5SDimitry Andric return s->chunks.empty();
12560b57cec5SDimitry Andric };
1257349cc55cSDimitry Andric llvm::erase_if(ctx.outputSections, isUnused);
12580b57cec5SDimitry Andric }
12590b57cec5SDimitry Andric
12600b57cec5SDimitry Andric // The Windows loader doesn't seem to like empty sections,
12610b57cec5SDimitry Andric // so we remove them if any.
removeEmptySections()12620b57cec5SDimitry Andric void Writer::removeEmptySections() {
12635f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Remove empty sections");
12640b57cec5SDimitry Andric auto isEmpty = [](OutputSection *s) { return s->getVirtualSize() == 0; };
1265349cc55cSDimitry Andric llvm::erase_if(ctx.outputSections, isEmpty);
12660b57cec5SDimitry Andric }
12670b57cec5SDimitry Andric
assignOutputSectionIndices()12680b57cec5SDimitry Andric void Writer::assignOutputSectionIndices() {
12695f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Output sections indices");
12700b57cec5SDimitry Andric // Assign final output section indices, and assign each chunk to its output
12710b57cec5SDimitry Andric // section.
12720b57cec5SDimitry Andric uint32_t idx = 1;
1273349cc55cSDimitry Andric for (OutputSection *os : ctx.outputSections) {
12740b57cec5SDimitry Andric os->sectionIndex = idx;
12750b57cec5SDimitry Andric for (Chunk *c : os->chunks)
12760b57cec5SDimitry Andric c->setOutputSectionIdx(idx);
12770b57cec5SDimitry Andric ++idx;
12780b57cec5SDimitry Andric }
12790b57cec5SDimitry Andric
12800b57cec5SDimitry Andric // Merge chunks are containers of chunks, so assign those an output section
12810b57cec5SDimitry Andric // too.
1282349cc55cSDimitry Andric for (MergeChunk *mc : ctx.mergeChunkInstances)
12830b57cec5SDimitry Andric if (mc)
12840b57cec5SDimitry Andric for (SectionChunk *sc : mc->sections)
12850b57cec5SDimitry Andric if (sc && sc->live)
12860b57cec5SDimitry Andric sc->setOutputSectionIdx(mc->getOutputSectionIdx());
12870b57cec5SDimitry Andric }
12880b57cec5SDimitry Andric
addEntryToStringTable(StringRef str)12890b57cec5SDimitry Andric size_t Writer::addEntryToStringTable(StringRef str) {
12900b57cec5SDimitry Andric assert(str.size() > COFF::NameSize);
12910b57cec5SDimitry Andric size_t offsetOfEntry = strtab.size() + 4; // +4 for the size field
12920b57cec5SDimitry Andric strtab.insert(strtab.end(), str.begin(), str.end());
12930b57cec5SDimitry Andric strtab.push_back('\0');
12940b57cec5SDimitry Andric return offsetOfEntry;
12950b57cec5SDimitry Andric }
12960b57cec5SDimitry Andric
createSymbol(Defined * def)1297bdd1243dSDimitry Andric std::optional<coff_symbol16> Writer::createSymbol(Defined *def) {
12980b57cec5SDimitry Andric coff_symbol16 sym;
12990b57cec5SDimitry Andric switch (def->kind()) {
1300bdd1243dSDimitry Andric case Symbol::DefinedAbsoluteKind: {
1301bdd1243dSDimitry Andric auto *da = dyn_cast<DefinedAbsolute>(def);
1302bdd1243dSDimitry Andric // Note: COFF symbol can only store 32-bit values, so 64-bit absolute
1303bdd1243dSDimitry Andric // values will be truncated.
1304bdd1243dSDimitry Andric sym.Value = da->getVA();
13050b57cec5SDimitry Andric sym.SectionNumber = IMAGE_SYM_ABSOLUTE;
13060b57cec5SDimitry Andric break;
1307bdd1243dSDimitry Andric }
13080b57cec5SDimitry Andric default: {
13090b57cec5SDimitry Andric // Don't write symbols that won't be written to the output to the symbol
13100b57cec5SDimitry Andric // table.
1311bdd1243dSDimitry Andric // We also try to write DefinedSynthetic as a normal symbol. Some of these
1312bdd1243dSDimitry Andric // symbols do point to an actual chunk, like __safe_se_handler_table. Others
1313bdd1243dSDimitry Andric // like __ImageBase are outside of sections and thus cannot be represented.
13140b57cec5SDimitry Andric Chunk *c = def->getChunk();
13150b57cec5SDimitry Andric if (!c)
1316bdd1243dSDimitry Andric return std::nullopt;
1317349cc55cSDimitry Andric OutputSection *os = ctx.getOutputSection(c);
13180b57cec5SDimitry Andric if (!os)
1319bdd1243dSDimitry Andric return std::nullopt;
13200b57cec5SDimitry Andric
13210b57cec5SDimitry Andric sym.Value = def->getRVA() - os->getRVA();
13220b57cec5SDimitry Andric sym.SectionNumber = os->sectionIndex;
13230b57cec5SDimitry Andric break;
13240b57cec5SDimitry Andric }
13250b57cec5SDimitry Andric }
13260b57cec5SDimitry Andric
13270b57cec5SDimitry Andric // Symbols that are runtime pseudo relocations don't point to the actual
13280b57cec5SDimitry Andric // symbol data itself (as they are imported), but points to the IAT entry
13290b57cec5SDimitry Andric // instead. Avoid emitting them to the symbol table, as they can confuse
13300b57cec5SDimitry Andric // debuggers.
13310b57cec5SDimitry Andric if (def->isRuntimePseudoReloc)
1332bdd1243dSDimitry Andric return std::nullopt;
13330b57cec5SDimitry Andric
13340b57cec5SDimitry Andric StringRef name = def->getName();
13350b57cec5SDimitry Andric if (name.size() > COFF::NameSize) {
13360b57cec5SDimitry Andric sym.Name.Offset.Zeroes = 0;
13370b57cec5SDimitry Andric sym.Name.Offset.Offset = addEntryToStringTable(name);
13380b57cec5SDimitry Andric } else {
13390b57cec5SDimitry Andric memset(sym.Name.ShortName, 0, COFF::NameSize);
13400b57cec5SDimitry Andric memcpy(sym.Name.ShortName, name.data(), name.size());
13410b57cec5SDimitry Andric }
13420b57cec5SDimitry Andric
13430b57cec5SDimitry Andric if (auto *d = dyn_cast<DefinedCOFF>(def)) {
13440b57cec5SDimitry Andric COFFSymbolRef ref = d->getCOFFSymbol();
13450b57cec5SDimitry Andric sym.Type = ref.getType();
13460b57cec5SDimitry Andric sym.StorageClass = ref.getStorageClass();
1347bdd1243dSDimitry Andric } else if (def->kind() == Symbol::DefinedImportThunkKind) {
1348bdd1243dSDimitry Andric sym.Type = (IMAGE_SYM_DTYPE_FUNCTION << SCT_COMPLEX_TYPE_SHIFT) |
1349bdd1243dSDimitry Andric IMAGE_SYM_TYPE_NULL;
1350bdd1243dSDimitry Andric sym.StorageClass = IMAGE_SYM_CLASS_EXTERNAL;
13510b57cec5SDimitry Andric } else {
13520b57cec5SDimitry Andric sym.Type = IMAGE_SYM_TYPE_NULL;
13530b57cec5SDimitry Andric sym.StorageClass = IMAGE_SYM_CLASS_EXTERNAL;
13540b57cec5SDimitry Andric }
13550b57cec5SDimitry Andric sym.NumberOfAuxSymbols = 0;
13560b57cec5SDimitry Andric return sym;
13570b57cec5SDimitry Andric }
13580b57cec5SDimitry Andric
createSymbolAndStringTable()13590b57cec5SDimitry Andric void Writer::createSymbolAndStringTable() {
13605f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Symbol and string table");
13610b57cec5SDimitry Andric // PE/COFF images are limited to 8 byte section names. Longer names can be
13620b57cec5SDimitry Andric // supported by writing a non-standard string table, but this string table is
13630b57cec5SDimitry Andric // not mapped at runtime and the long names will therefore be inaccessible.
13640b57cec5SDimitry Andric // link.exe always truncates section names to 8 bytes, whereas binutils always
13650b57cec5SDimitry Andric // preserves long section names via the string table. LLD adopts a hybrid
13660b57cec5SDimitry Andric // solution where discardable sections have long names preserved and
13670b57cec5SDimitry Andric // non-discardable sections have their names truncated, to ensure that any
13680b57cec5SDimitry Andric // section which is mapped at runtime also has its name mapped at runtime.
1369349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
13700b57cec5SDimitry Andric if (sec->name.size() <= COFF::NameSize)
13710b57cec5SDimitry Andric continue;
13720b57cec5SDimitry Andric if ((sec->header.Characteristics & IMAGE_SCN_MEM_DISCARDABLE) == 0)
13730b57cec5SDimitry Andric continue;
1374bdd1243dSDimitry Andric if (ctx.config.warnLongSectionNames) {
1375480093f4SDimitry Andric warn("section name " + sec->name +
1376480093f4SDimitry Andric " is longer than 8 characters and will use a non-standard string "
1377480093f4SDimitry Andric "table");
1378480093f4SDimitry Andric }
13790b57cec5SDimitry Andric sec->setStringTableOff(addEntryToStringTable(sec->name));
13800b57cec5SDimitry Andric }
13810b57cec5SDimitry Andric
13825f757f3fSDimitry Andric if (ctx.config.writeSymtab) {
1383349cc55cSDimitry Andric for (ObjFile *file : ctx.objFileInstances) {
13840b57cec5SDimitry Andric for (Symbol *b : file->getSymbols()) {
13850b57cec5SDimitry Andric auto *d = dyn_cast_or_null<Defined>(b);
13860b57cec5SDimitry Andric if (!d || d->writtenToSymtab)
13870b57cec5SDimitry Andric continue;
13880b57cec5SDimitry Andric d->writtenToSymtab = true;
13894824e7fdSDimitry Andric if (auto *dc = dyn_cast_or_null<DefinedCOFF>(d)) {
13904824e7fdSDimitry Andric COFFSymbolRef symRef = dc->getCOFFSymbol();
13914824e7fdSDimitry Andric if (symRef.isSectionDefinition() ||
13924824e7fdSDimitry Andric symRef.getStorageClass() == COFF::IMAGE_SYM_CLASS_LABEL)
13934824e7fdSDimitry Andric continue;
13944824e7fdSDimitry Andric }
13950b57cec5SDimitry Andric
1396bdd1243dSDimitry Andric if (std::optional<coff_symbol16> sym = createSymbol(d))
13970b57cec5SDimitry Andric outputSymtab.push_back(*sym);
1398bdd1243dSDimitry Andric
1399bdd1243dSDimitry Andric if (auto *dthunk = dyn_cast<DefinedImportThunk>(d)) {
1400bdd1243dSDimitry Andric if (!dthunk->wrappedSym->writtenToSymtab) {
1401bdd1243dSDimitry Andric dthunk->wrappedSym->writtenToSymtab = true;
1402bdd1243dSDimitry Andric if (std::optional<coff_symbol16> sym =
1403bdd1243dSDimitry Andric createSymbol(dthunk->wrappedSym))
1404bdd1243dSDimitry Andric outputSymtab.push_back(*sym);
1405bdd1243dSDimitry Andric }
1406bdd1243dSDimitry Andric }
14070b57cec5SDimitry Andric }
14080b57cec5SDimitry Andric }
14090b57cec5SDimitry Andric }
14100b57cec5SDimitry Andric
14110b57cec5SDimitry Andric if (outputSymtab.empty() && strtab.empty())
14120b57cec5SDimitry Andric return;
14130b57cec5SDimitry Andric
14140b57cec5SDimitry Andric // We position the symbol table to be adjacent to the end of the last section.
14150b57cec5SDimitry Andric uint64_t fileOff = fileSize;
14160b57cec5SDimitry Andric pointerToSymbolTable = fileOff;
14170b57cec5SDimitry Andric fileOff += outputSymtab.size() * sizeof(coff_symbol16);
14180b57cec5SDimitry Andric fileOff += 4 + strtab.size();
1419bdd1243dSDimitry Andric fileSize = alignTo(fileOff, ctx.config.fileAlign);
14200b57cec5SDimitry Andric }
14210b57cec5SDimitry Andric
mergeSections()14220b57cec5SDimitry Andric void Writer::mergeSections() {
14235f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Merge sections");
14240b57cec5SDimitry Andric if (!pdataSec->chunks.empty()) {
14255f757f3fSDimitry Andric if (isArm64EC(ctx.config.machine)) {
14265f757f3fSDimitry Andric // On ARM64EC .pdata may contain both ARM64 and X64 data. Split them by
14275f757f3fSDimitry Andric // sorting and store their regions separately.
14285f757f3fSDimitry Andric llvm::stable_sort(pdataSec->chunks, [=](const Chunk *a, const Chunk *b) {
14295f757f3fSDimitry Andric return (a->getMachine() == AMD64) < (b->getMachine() == AMD64);
14305f757f3fSDimitry Andric });
14315f757f3fSDimitry Andric
14325f757f3fSDimitry Andric for (auto chunk : pdataSec->chunks) {
14335f757f3fSDimitry Andric if (chunk->getMachine() == AMD64) {
14345f757f3fSDimitry Andric hybridPdata.first = chunk;
14355f757f3fSDimitry Andric hybridPdata.last = pdataSec->chunks.back();
14365f757f3fSDimitry Andric break;
14375f757f3fSDimitry Andric }
14385f757f3fSDimitry Andric
14395f757f3fSDimitry Andric if (!pdata.first)
14405f757f3fSDimitry Andric pdata.first = chunk;
14415f757f3fSDimitry Andric pdata.last = chunk;
14425f757f3fSDimitry Andric }
14435f757f3fSDimitry Andric } else {
14445f757f3fSDimitry Andric pdata.first = pdataSec->chunks.front();
14455f757f3fSDimitry Andric pdata.last = pdataSec->chunks.back();
14465f757f3fSDimitry Andric }
14470b57cec5SDimitry Andric }
14480b57cec5SDimitry Andric
1449bdd1243dSDimitry Andric for (auto &p : ctx.config.merge) {
14500b57cec5SDimitry Andric StringRef toName = p.second;
14510b57cec5SDimitry Andric if (p.first == toName)
14520b57cec5SDimitry Andric continue;
14530b57cec5SDimitry Andric StringSet<> names;
145404eeddc0SDimitry Andric while (true) {
14550b57cec5SDimitry Andric if (!names.insert(toName).second)
14560b57cec5SDimitry Andric fatal("/merge: cycle found for section '" + p.first + "'");
1457bdd1243dSDimitry Andric auto i = ctx.config.merge.find(toName);
1458bdd1243dSDimitry Andric if (i == ctx.config.merge.end())
14590b57cec5SDimitry Andric break;
14600b57cec5SDimitry Andric toName = i->second;
14610b57cec5SDimitry Andric }
14620b57cec5SDimitry Andric OutputSection *from = findSection(p.first);
14630b57cec5SDimitry Andric OutputSection *to = findSection(toName);
14640b57cec5SDimitry Andric if (!from)
14650b57cec5SDimitry Andric continue;
14660b57cec5SDimitry Andric if (!to) {
14670b57cec5SDimitry Andric from->name = toName;
14680b57cec5SDimitry Andric continue;
14690b57cec5SDimitry Andric }
14700b57cec5SDimitry Andric to->merge(from);
14710b57cec5SDimitry Andric }
14720b57cec5SDimitry Andric }
14730b57cec5SDimitry Andric
14745f757f3fSDimitry Andric // EC targets may have chunks of various architectures mixed together at this
14755f757f3fSDimitry Andric // point. Group code chunks of the same architecture together by sorting chunks
14765f757f3fSDimitry Andric // by their EC range type.
sortECChunks()14775f757f3fSDimitry Andric void Writer::sortECChunks() {
14785f757f3fSDimitry Andric if (!isArm64EC(ctx.config.machine))
14795f757f3fSDimitry Andric return;
14805f757f3fSDimitry Andric
14815f757f3fSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
14825f757f3fSDimitry Andric if (sec->isCodeSection())
14835f757f3fSDimitry Andric llvm::stable_sort(sec->chunks, [=](const Chunk *a, const Chunk *b) {
14845f757f3fSDimitry Andric std::optional<chpe_range_type> aType = a->getArm64ECRangeType(),
14855f757f3fSDimitry Andric bType = b->getArm64ECRangeType();
14865f757f3fSDimitry Andric return bType && (!aType || *aType < *bType);
14875f757f3fSDimitry Andric });
14885f757f3fSDimitry Andric }
14895f757f3fSDimitry Andric }
14905f757f3fSDimitry Andric
14910b57cec5SDimitry Andric // Visits all sections to assign incremental, non-overlapping RVAs and
14920b57cec5SDimitry Andric // file offsets.
assignAddresses()14930b57cec5SDimitry Andric void Writer::assignAddresses() {
14945f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Assign addresses");
1495bdd1243dSDimitry Andric Configuration *config = &ctx.config;
1496bdd1243dSDimitry Andric
14975f757f3fSDimitry Andric // We need to create EC code map so that ECCodeMapChunk knows its size.
14985f757f3fSDimitry Andric // We do it here to make sure that we account for range extension chunks.
14995f757f3fSDimitry Andric createECCodeMap();
15005f757f3fSDimitry Andric
15010b57cec5SDimitry Andric sizeOfHeaders = dosStubSize + sizeof(PEMagic) + sizeof(coff_file_header) +
15020b57cec5SDimitry Andric sizeof(data_directory) * numberOfDataDirectory +
1503349cc55cSDimitry Andric sizeof(coff_section) * ctx.outputSections.size();
15040b57cec5SDimitry Andric sizeOfHeaders +=
15050b57cec5SDimitry Andric config->is64() ? sizeof(pe32plus_header) : sizeof(pe32_header);
15060b57cec5SDimitry Andric sizeOfHeaders = alignTo(sizeOfHeaders, config->fileAlign);
15070b57cec5SDimitry Andric fileSize = sizeOfHeaders;
15080b57cec5SDimitry Andric
15090b57cec5SDimitry Andric // The first page is kept unmapped.
15100b57cec5SDimitry Andric uint64_t rva = alignTo(sizeOfHeaders, config->align);
15110b57cec5SDimitry Andric
1512349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
15135f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Section: ", sec->name);
15140b57cec5SDimitry Andric if (sec == relocSec)
15150b57cec5SDimitry Andric addBaserels();
15160b57cec5SDimitry Andric uint64_t rawSize = 0, virtualSize = 0;
15170b57cec5SDimitry Andric sec->header.VirtualAddress = rva;
15180b57cec5SDimitry Andric
15190b57cec5SDimitry Andric // If /FUNCTIONPADMIN is used, functions are padded in order to create a
15200b57cec5SDimitry Andric // hotpatchable image.
15215f757f3fSDimitry Andric uint32_t padding = sec->isCodeSection() ? config->functionPadMin : 0;
15225f757f3fSDimitry Andric std::optional<chpe_range_type> prevECRange;
15230b57cec5SDimitry Andric
15240b57cec5SDimitry Andric for (Chunk *c : sec->chunks) {
15255f757f3fSDimitry Andric // Alignment EC code range baudaries.
15265f757f3fSDimitry Andric if (isArm64EC(ctx.config.machine) && sec->isCodeSection()) {
15275f757f3fSDimitry Andric std::optional<chpe_range_type> rangeType = c->getArm64ECRangeType();
15285f757f3fSDimitry Andric if (rangeType != prevECRange) {
15295f757f3fSDimitry Andric virtualSize = alignTo(virtualSize, 4096);
15305f757f3fSDimitry Andric prevECRange = rangeType;
15315f757f3fSDimitry Andric }
15325f757f3fSDimitry Andric }
15330b57cec5SDimitry Andric if (padding && c->isHotPatchable())
15340b57cec5SDimitry Andric virtualSize += padding;
1535*0fca6ea1SDimitry Andric // If chunk has EC entry thunk, reserve a space for an offset to the
1536*0fca6ea1SDimitry Andric // thunk.
1537*0fca6ea1SDimitry Andric if (c->getEntryThunk())
1538*0fca6ea1SDimitry Andric virtualSize += sizeof(uint32_t);
15390b57cec5SDimitry Andric virtualSize = alignTo(virtualSize, c->getAlignment());
15400b57cec5SDimitry Andric c->setRVA(rva + virtualSize);
15410b57cec5SDimitry Andric virtualSize += c->getSize();
15420b57cec5SDimitry Andric if (c->hasData)
15430b57cec5SDimitry Andric rawSize = alignTo(virtualSize, config->fileAlign);
15440b57cec5SDimitry Andric }
15450b57cec5SDimitry Andric if (virtualSize > UINT32_MAX)
15460b57cec5SDimitry Andric error("section larger than 4 GiB: " + sec->name);
15470b57cec5SDimitry Andric sec->header.VirtualSize = virtualSize;
15480b57cec5SDimitry Andric sec->header.SizeOfRawData = rawSize;
15490b57cec5SDimitry Andric if (rawSize != 0)
15500b57cec5SDimitry Andric sec->header.PointerToRawData = fileSize;
15510b57cec5SDimitry Andric rva += alignTo(virtualSize, config->align);
15520b57cec5SDimitry Andric fileSize += alignTo(rawSize, config->fileAlign);
15530b57cec5SDimitry Andric }
15540b57cec5SDimitry Andric sizeOfImage = alignTo(rva, config->align);
15550b57cec5SDimitry Andric
15560b57cec5SDimitry Andric // Assign addresses to sections in MergeChunks.
1557349cc55cSDimitry Andric for (MergeChunk *mc : ctx.mergeChunkInstances)
15580b57cec5SDimitry Andric if (mc)
15590b57cec5SDimitry Andric mc->assignSubsectionRVAs();
15600b57cec5SDimitry Andric }
15610b57cec5SDimitry Andric
writeHeader()15620b57cec5SDimitry Andric template <typename PEHeaderTy> void Writer::writeHeader() {
15630b57cec5SDimitry Andric // Write DOS header. For backwards compatibility, the first part of a PE/COFF
15640b57cec5SDimitry Andric // executable consists of an MS-DOS MZ executable. If the executable is run
15650b57cec5SDimitry Andric // under DOS, that program gets run (usually to just print an error message).
15660b57cec5SDimitry Andric // When run under Windows, the loader looks at AddressOfNewExeHeader and uses
15670b57cec5SDimitry Andric // the PE header instead.
1568bdd1243dSDimitry Andric Configuration *config = &ctx.config;
15690b57cec5SDimitry Andric uint8_t *buf = buffer->getBufferStart();
15700b57cec5SDimitry Andric auto *dos = reinterpret_cast<dos_header *>(buf);
15710b57cec5SDimitry Andric buf += sizeof(dos_header);
15720b57cec5SDimitry Andric dos->Magic[0] = 'M';
15730b57cec5SDimitry Andric dos->Magic[1] = 'Z';
15740b57cec5SDimitry Andric dos->UsedBytesInTheLastPage = dosStubSize % 512;
15750b57cec5SDimitry Andric dos->FileSizeInPages = divideCeil(dosStubSize, 512);
15760b57cec5SDimitry Andric dos->HeaderSizeInParagraphs = sizeof(dos_header) / 16;
15770b57cec5SDimitry Andric
15780b57cec5SDimitry Andric dos->AddressOfRelocationTable = sizeof(dos_header);
15790b57cec5SDimitry Andric dos->AddressOfNewExeHeader = dosStubSize;
15800b57cec5SDimitry Andric
15810b57cec5SDimitry Andric // Write DOS program.
15820b57cec5SDimitry Andric memcpy(buf, dosProgram, sizeof(dosProgram));
15830b57cec5SDimitry Andric buf += sizeof(dosProgram);
15840b57cec5SDimitry Andric
15850b57cec5SDimitry Andric // Write PE magic
15860b57cec5SDimitry Andric memcpy(buf, PEMagic, sizeof(PEMagic));
15870b57cec5SDimitry Andric buf += sizeof(PEMagic);
15880b57cec5SDimitry Andric
15890b57cec5SDimitry Andric // Write COFF header
15900b57cec5SDimitry Andric auto *coff = reinterpret_cast<coff_file_header *>(buf);
15910b57cec5SDimitry Andric buf += sizeof(*coff);
159206c3fb27SDimitry Andric switch (config->machine) {
159306c3fb27SDimitry Andric case ARM64EC:
159406c3fb27SDimitry Andric coff->Machine = AMD64;
159506c3fb27SDimitry Andric break;
159606c3fb27SDimitry Andric case ARM64X:
159706c3fb27SDimitry Andric coff->Machine = ARM64;
159806c3fb27SDimitry Andric break;
159906c3fb27SDimitry Andric default:
16000b57cec5SDimitry Andric coff->Machine = config->machine;
160106c3fb27SDimitry Andric }
1602349cc55cSDimitry Andric coff->NumberOfSections = ctx.outputSections.size();
16030b57cec5SDimitry Andric coff->Characteristics = IMAGE_FILE_EXECUTABLE_IMAGE;
16040b57cec5SDimitry Andric if (config->largeAddressAware)
16050b57cec5SDimitry Andric coff->Characteristics |= IMAGE_FILE_LARGE_ADDRESS_AWARE;
16060b57cec5SDimitry Andric if (!config->is64())
16070b57cec5SDimitry Andric coff->Characteristics |= IMAGE_FILE_32BIT_MACHINE;
16080b57cec5SDimitry Andric if (config->dll)
16090b57cec5SDimitry Andric coff->Characteristics |= IMAGE_FILE_DLL;
1610480093f4SDimitry Andric if (config->driverUponly)
1611480093f4SDimitry Andric coff->Characteristics |= IMAGE_FILE_UP_SYSTEM_ONLY;
16120b57cec5SDimitry Andric if (!config->relocatable)
16130b57cec5SDimitry Andric coff->Characteristics |= IMAGE_FILE_RELOCS_STRIPPED;
16140b57cec5SDimitry Andric if (config->swaprunCD)
16150b57cec5SDimitry Andric coff->Characteristics |= IMAGE_FILE_REMOVABLE_RUN_FROM_SWAP;
16160b57cec5SDimitry Andric if (config->swaprunNet)
16170b57cec5SDimitry Andric coff->Characteristics |= IMAGE_FILE_NET_RUN_FROM_SWAP;
16180b57cec5SDimitry Andric coff->SizeOfOptionalHeader =
16190b57cec5SDimitry Andric sizeof(PEHeaderTy) + sizeof(data_directory) * numberOfDataDirectory;
16200b57cec5SDimitry Andric
16210b57cec5SDimitry Andric // Write PE header
16220b57cec5SDimitry Andric auto *pe = reinterpret_cast<PEHeaderTy *>(buf);
16230b57cec5SDimitry Andric buf += sizeof(*pe);
16240b57cec5SDimitry Andric pe->Magic = config->is64() ? PE32Header::PE32_PLUS : PE32Header::PE32;
16250b57cec5SDimitry Andric
16260b57cec5SDimitry Andric // If {Major,Minor}LinkerVersion is left at 0.0, then for some
16270b57cec5SDimitry Andric // reason signing the resulting PE file with Authenticode produces a
16280b57cec5SDimitry Andric // signature that fails to validate on Windows 7 (but is OK on 10).
16290b57cec5SDimitry Andric // Set it to 14.0, which is what VS2015 outputs, and which avoids
16300b57cec5SDimitry Andric // that problem.
16310b57cec5SDimitry Andric pe->MajorLinkerVersion = 14;
16320b57cec5SDimitry Andric pe->MinorLinkerVersion = 0;
16330b57cec5SDimitry Andric
16340b57cec5SDimitry Andric pe->ImageBase = config->imageBase;
16350b57cec5SDimitry Andric pe->SectionAlignment = config->align;
16360b57cec5SDimitry Andric pe->FileAlignment = config->fileAlign;
16370b57cec5SDimitry Andric pe->MajorImageVersion = config->majorImageVersion;
16380b57cec5SDimitry Andric pe->MinorImageVersion = config->minorImageVersion;
16390b57cec5SDimitry Andric pe->MajorOperatingSystemVersion = config->majorOSVersion;
16400b57cec5SDimitry Andric pe->MinorOperatingSystemVersion = config->minorOSVersion;
1641e8d8bef9SDimitry Andric pe->MajorSubsystemVersion = config->majorSubsystemVersion;
1642e8d8bef9SDimitry Andric pe->MinorSubsystemVersion = config->minorSubsystemVersion;
16430b57cec5SDimitry Andric pe->Subsystem = config->subsystem;
16440b57cec5SDimitry Andric pe->SizeOfImage = sizeOfImage;
16450b57cec5SDimitry Andric pe->SizeOfHeaders = sizeOfHeaders;
16460b57cec5SDimitry Andric if (!config->noEntry) {
16470b57cec5SDimitry Andric Defined *entry = cast<Defined>(config->entry);
16480b57cec5SDimitry Andric pe->AddressOfEntryPoint = entry->getRVA();
16490b57cec5SDimitry Andric // Pointer to thumb code must have the LSB set, so adjust it.
16500b57cec5SDimitry Andric if (config->machine == ARMNT)
16510b57cec5SDimitry Andric pe->AddressOfEntryPoint |= 1;
16520b57cec5SDimitry Andric }
16530b57cec5SDimitry Andric pe->SizeOfStackReserve = config->stackReserve;
16540b57cec5SDimitry Andric pe->SizeOfStackCommit = config->stackCommit;
16550b57cec5SDimitry Andric pe->SizeOfHeapReserve = config->heapReserve;
16560b57cec5SDimitry Andric pe->SizeOfHeapCommit = config->heapCommit;
16570b57cec5SDimitry Andric if (config->appContainer)
16580b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_APPCONTAINER;
1659480093f4SDimitry Andric if (config->driverWdm)
1660480093f4SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_WDM_DRIVER;
16610b57cec5SDimitry Andric if (config->dynamicBase)
16620b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_DYNAMIC_BASE;
16630b57cec5SDimitry Andric if (config->highEntropyVA)
16640b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_HIGH_ENTROPY_VA;
16650b57cec5SDimitry Andric if (!config->allowBind)
16660b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_NO_BIND;
16670b57cec5SDimitry Andric if (config->nxCompat)
16680b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_NX_COMPAT;
16690b57cec5SDimitry Andric if (!config->allowIsolation)
16700b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_NO_ISOLATION;
16710b57cec5SDimitry Andric if (config->guardCF != GuardCFLevel::Off)
16720b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_GUARD_CF;
16730b57cec5SDimitry Andric if (config->integrityCheck)
16740b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_FORCE_INTEGRITY;
1675979e22ffSDimitry Andric if (setNoSEHCharacteristic || config->noSEH)
16760b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_NO_SEH;
16770b57cec5SDimitry Andric if (config->terminalServerAware)
16780b57cec5SDimitry Andric pe->DLLCharacteristics |= IMAGE_DLL_CHARACTERISTICS_TERMINAL_SERVER_AWARE;
16790b57cec5SDimitry Andric pe->NumberOfRvaAndSize = numberOfDataDirectory;
16800b57cec5SDimitry Andric if (textSec->getVirtualSize()) {
16810b57cec5SDimitry Andric pe->BaseOfCode = textSec->getRVA();
16820b57cec5SDimitry Andric pe->SizeOfCode = textSec->getRawSize();
16830b57cec5SDimitry Andric }
16840b57cec5SDimitry Andric pe->SizeOfInitializedData = getSizeOfInitializedData();
16850b57cec5SDimitry Andric
16860b57cec5SDimitry Andric // Write data directory
16870b57cec5SDimitry Andric auto *dir = reinterpret_cast<data_directory *>(buf);
16880b57cec5SDimitry Andric buf += sizeof(*dir) * numberOfDataDirectory;
168985868e8aSDimitry Andric if (edataStart) {
169085868e8aSDimitry Andric dir[EXPORT_TABLE].RelativeVirtualAddress = edataStart->getRVA();
169185868e8aSDimitry Andric dir[EXPORT_TABLE].Size =
169285868e8aSDimitry Andric edataEnd->getRVA() + edataEnd->getSize() - edataStart->getRVA();
16930b57cec5SDimitry Andric }
16940b57cec5SDimitry Andric if (importTableStart) {
16950b57cec5SDimitry Andric dir[IMPORT_TABLE].RelativeVirtualAddress = importTableStart->getRVA();
16960b57cec5SDimitry Andric dir[IMPORT_TABLE].Size = importTableSize;
16970b57cec5SDimitry Andric }
16980b57cec5SDimitry Andric if (iatStart) {
16990b57cec5SDimitry Andric dir[IAT].RelativeVirtualAddress = iatStart->getRVA();
17000b57cec5SDimitry Andric dir[IAT].Size = iatSize;
17010b57cec5SDimitry Andric }
17020b57cec5SDimitry Andric if (rsrcSec->getVirtualSize()) {
17030b57cec5SDimitry Andric dir[RESOURCE_TABLE].RelativeVirtualAddress = rsrcSec->getRVA();
17040b57cec5SDimitry Andric dir[RESOURCE_TABLE].Size = rsrcSec->getVirtualSize();
17050b57cec5SDimitry Andric }
17065f757f3fSDimitry Andric // ARM64EC (but not ARM64X) contains x86_64 exception table in data directory.
17075f757f3fSDimitry Andric ChunkRange &exceptionTable =
17085f757f3fSDimitry Andric ctx.config.machine == ARM64EC ? hybridPdata : pdata;
17095f757f3fSDimitry Andric if (exceptionTable.first) {
17105f757f3fSDimitry Andric dir[EXCEPTION_TABLE].RelativeVirtualAddress =
17115f757f3fSDimitry Andric exceptionTable.first->getRVA();
17125f757f3fSDimitry Andric dir[EXCEPTION_TABLE].Size = exceptionTable.last->getRVA() +
17135f757f3fSDimitry Andric exceptionTable.last->getSize() -
17145f757f3fSDimitry Andric exceptionTable.first->getRVA();
17150b57cec5SDimitry Andric }
17160b57cec5SDimitry Andric if (relocSec->getVirtualSize()) {
17170b57cec5SDimitry Andric dir[BASE_RELOCATION_TABLE].RelativeVirtualAddress = relocSec->getRVA();
17180b57cec5SDimitry Andric dir[BASE_RELOCATION_TABLE].Size = relocSec->getVirtualSize();
17190b57cec5SDimitry Andric }
1720349cc55cSDimitry Andric if (Symbol *sym = ctx.symtab.findUnderscore("_tls_used")) {
17210b57cec5SDimitry Andric if (Defined *b = dyn_cast<Defined>(sym)) {
17220b57cec5SDimitry Andric dir[TLS_TABLE].RelativeVirtualAddress = b->getRVA();
17230b57cec5SDimitry Andric dir[TLS_TABLE].Size = config->is64()
17240b57cec5SDimitry Andric ? sizeof(object::coff_tls_directory64)
17250b57cec5SDimitry Andric : sizeof(object::coff_tls_directory32);
17260b57cec5SDimitry Andric }
17270b57cec5SDimitry Andric }
17280b57cec5SDimitry Andric if (debugDirectory) {
17290b57cec5SDimitry Andric dir[DEBUG_DIRECTORY].RelativeVirtualAddress = debugDirectory->getRVA();
17300b57cec5SDimitry Andric dir[DEBUG_DIRECTORY].Size = debugDirectory->getSize();
17310b57cec5SDimitry Andric }
1732349cc55cSDimitry Andric if (Symbol *sym = ctx.symtab.findUnderscore("_load_config_used")) {
17330b57cec5SDimitry Andric if (auto *b = dyn_cast<DefinedRegular>(sym)) {
17340b57cec5SDimitry Andric SectionChunk *sc = b->getChunk();
17350b57cec5SDimitry Andric assert(b->getRVA() >= sc->getRVA());
17360b57cec5SDimitry Andric uint64_t offsetInChunk = b->getRVA() - sc->getRVA();
17370b57cec5SDimitry Andric if (!sc->hasData || offsetInChunk + 4 > sc->getSize())
17380b57cec5SDimitry Andric fatal("_load_config_used is malformed");
17390b57cec5SDimitry Andric
17400b57cec5SDimitry Andric ArrayRef<uint8_t> secContents = sc->getContents();
17410b57cec5SDimitry Andric uint32_t loadConfigSize =
17420b57cec5SDimitry Andric *reinterpret_cast<const ulittle32_t *>(&secContents[offsetInChunk]);
17430b57cec5SDimitry Andric if (offsetInChunk + loadConfigSize > sc->getSize())
17440b57cec5SDimitry Andric fatal("_load_config_used is too large");
17450b57cec5SDimitry Andric dir[LOAD_CONFIG_TABLE].RelativeVirtualAddress = b->getRVA();
17460b57cec5SDimitry Andric dir[LOAD_CONFIG_TABLE].Size = loadConfigSize;
17470b57cec5SDimitry Andric }
17480b57cec5SDimitry Andric }
17490b57cec5SDimitry Andric if (!delayIdata.empty()) {
17500b57cec5SDimitry Andric dir[DELAY_IMPORT_DESCRIPTOR].RelativeVirtualAddress =
17510b57cec5SDimitry Andric delayIdata.getDirRVA();
17520b57cec5SDimitry Andric dir[DELAY_IMPORT_DESCRIPTOR].Size = delayIdata.getDirSize();
17530b57cec5SDimitry Andric }
17540b57cec5SDimitry Andric
17550b57cec5SDimitry Andric // Write section table
1756349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
1757bdd1243dSDimitry Andric sec->writeHeaderTo(buf, config->debug);
17580b57cec5SDimitry Andric buf += sizeof(coff_section);
17590b57cec5SDimitry Andric }
17600b57cec5SDimitry Andric sectionTable = ArrayRef<uint8_t>(
1761349cc55cSDimitry Andric buf - ctx.outputSections.size() * sizeof(coff_section), buf);
17620b57cec5SDimitry Andric
17630b57cec5SDimitry Andric if (outputSymtab.empty() && strtab.empty())
17640b57cec5SDimitry Andric return;
17650b57cec5SDimitry Andric
17660b57cec5SDimitry Andric coff->PointerToSymbolTable = pointerToSymbolTable;
17670b57cec5SDimitry Andric uint32_t numberOfSymbols = outputSymtab.size();
17680b57cec5SDimitry Andric coff->NumberOfSymbols = numberOfSymbols;
17690b57cec5SDimitry Andric auto *symbolTable = reinterpret_cast<coff_symbol16 *>(
17700b57cec5SDimitry Andric buffer->getBufferStart() + coff->PointerToSymbolTable);
17710b57cec5SDimitry Andric for (size_t i = 0; i != numberOfSymbols; ++i)
17720b57cec5SDimitry Andric symbolTable[i] = outputSymtab[i];
17730b57cec5SDimitry Andric // Create the string table, it follows immediately after the symbol table.
17740b57cec5SDimitry Andric // The first 4 bytes is length including itself.
17750b57cec5SDimitry Andric buf = reinterpret_cast<uint8_t *>(&symbolTable[numberOfSymbols]);
17760b57cec5SDimitry Andric write32le(buf, strtab.size() + 4);
17770b57cec5SDimitry Andric if (!strtab.empty())
17780b57cec5SDimitry Andric memcpy(buf + 4, strtab.data(), strtab.size());
17790b57cec5SDimitry Andric }
17800b57cec5SDimitry Andric
openFile(StringRef path)17810b57cec5SDimitry Andric void Writer::openFile(StringRef path) {
17820b57cec5SDimitry Andric buffer = CHECK(
17830b57cec5SDimitry Andric FileOutputBuffer::create(path, fileSize, FileOutputBuffer::F_executable),
17840b57cec5SDimitry Andric "failed to open " + path);
17850b57cec5SDimitry Andric }
17860b57cec5SDimitry Andric
createSEHTable()17870b57cec5SDimitry Andric void Writer::createSEHTable() {
17880b57cec5SDimitry Andric SymbolRVASet handlers;
1789349cc55cSDimitry Andric for (ObjFile *file : ctx.objFileInstances) {
17900b57cec5SDimitry Andric if (!file->hasSafeSEH())
17910b57cec5SDimitry Andric error("/safeseh: " + file->getName() + " is not compatible with SEH");
17920b57cec5SDimitry Andric markSymbolsForRVATable(file, file->getSXDataChunks(), handlers);
17930b57cec5SDimitry Andric }
17940b57cec5SDimitry Andric
17950b57cec5SDimitry Andric // Set the "no SEH" characteristic if there really were no handlers, or if
17960b57cec5SDimitry Andric // there is no load config object to point to the table of handlers.
17970b57cec5SDimitry Andric setNoSEHCharacteristic =
1798349cc55cSDimitry Andric handlers.empty() || !ctx.symtab.findUnderscore("_load_config_used");
17990b57cec5SDimitry Andric
18000b57cec5SDimitry Andric maybeAddRVATable(std::move(handlers), "__safe_se_handler_table",
18010b57cec5SDimitry Andric "__safe_se_handler_count");
18020b57cec5SDimitry Andric }
18030b57cec5SDimitry Andric
18040b57cec5SDimitry Andric // Add a symbol to an RVA set. Two symbols may have the same RVA, but an RVA set
18050b57cec5SDimitry Andric // cannot contain duplicates. Therefore, the set is uniqued by Chunk and the
18060b57cec5SDimitry Andric // symbol's offset into that Chunk.
addSymbolToRVASet(SymbolRVASet & rvaSet,Defined * s)18070b57cec5SDimitry Andric static void addSymbolToRVASet(SymbolRVASet &rvaSet, Defined *s) {
18080b57cec5SDimitry Andric Chunk *c = s->getChunk();
18097a6dacacSDimitry Andric if (!c)
18107a6dacacSDimitry Andric return;
18110b57cec5SDimitry Andric if (auto *sc = dyn_cast<SectionChunk>(c))
18120b57cec5SDimitry Andric c = sc->repl; // Look through ICF replacement.
18130b57cec5SDimitry Andric uint32_t off = s->getRVA() - (c ? c->getRVA() : 0);
18140b57cec5SDimitry Andric rvaSet.insert({c, off});
18150b57cec5SDimitry Andric }
18160b57cec5SDimitry Andric
18170b57cec5SDimitry Andric // Given a symbol, add it to the GFIDs table if it is a live, defined, function
18180b57cec5SDimitry Andric // symbol in an executable section.
maybeAddAddressTakenFunction(SymbolRVASet & addressTakenSyms,Symbol * s)18190b57cec5SDimitry Andric static void maybeAddAddressTakenFunction(SymbolRVASet &addressTakenSyms,
18200b57cec5SDimitry Andric Symbol *s) {
18210b57cec5SDimitry Andric if (!s)
18220b57cec5SDimitry Andric return;
18230b57cec5SDimitry Andric
18240b57cec5SDimitry Andric switch (s->kind()) {
18250b57cec5SDimitry Andric case Symbol::DefinedLocalImportKind:
18260b57cec5SDimitry Andric case Symbol::DefinedImportDataKind:
18270b57cec5SDimitry Andric // Defines an __imp_ pointer, so it is data, so it is ignored.
18280b57cec5SDimitry Andric break;
18290b57cec5SDimitry Andric case Symbol::DefinedCommonKind:
18300b57cec5SDimitry Andric // Common is always data, so it is ignored.
18310b57cec5SDimitry Andric break;
18320b57cec5SDimitry Andric case Symbol::DefinedAbsoluteKind:
18330b57cec5SDimitry Andric case Symbol::DefinedSyntheticKind:
18340b57cec5SDimitry Andric // Absolute is never code, synthetic generally isn't and usually isn't
18350b57cec5SDimitry Andric // determinable.
18360b57cec5SDimitry Andric break;
183785868e8aSDimitry Andric case Symbol::LazyArchiveKind:
183885868e8aSDimitry Andric case Symbol::LazyObjectKind:
1839fe6060f1SDimitry Andric case Symbol::LazyDLLSymbolKind:
18400b57cec5SDimitry Andric case Symbol::UndefinedKind:
18410b57cec5SDimitry Andric // Undefined symbols resolve to zero, so they don't have an RVA. Lazy
18420b57cec5SDimitry Andric // symbols shouldn't have relocations.
18430b57cec5SDimitry Andric break;
18440b57cec5SDimitry Andric
18450b57cec5SDimitry Andric case Symbol::DefinedImportThunkKind:
18460b57cec5SDimitry Andric // Thunks are always code, include them.
18470b57cec5SDimitry Andric addSymbolToRVASet(addressTakenSyms, cast<Defined>(s));
18480b57cec5SDimitry Andric break;
18490b57cec5SDimitry Andric
18500b57cec5SDimitry Andric case Symbol::DefinedRegularKind: {
18510b57cec5SDimitry Andric // This is a regular, defined, symbol from a COFF file. Mark the symbol as
18520b57cec5SDimitry Andric // address taken if the symbol type is function and it's in an executable
18530b57cec5SDimitry Andric // section.
18540b57cec5SDimitry Andric auto *d = cast<DefinedRegular>(s);
18550b57cec5SDimitry Andric if (d->getCOFFSymbol().getComplexType() == COFF::IMAGE_SYM_DTYPE_FUNCTION) {
18560b57cec5SDimitry Andric SectionChunk *sc = dyn_cast<SectionChunk>(d->getChunk());
18570b57cec5SDimitry Andric if (sc && sc->live &&
18580b57cec5SDimitry Andric sc->getOutputCharacteristics() & IMAGE_SCN_MEM_EXECUTE)
18590b57cec5SDimitry Andric addSymbolToRVASet(addressTakenSyms, d);
18600b57cec5SDimitry Andric }
18610b57cec5SDimitry Andric break;
18620b57cec5SDimitry Andric }
18630b57cec5SDimitry Andric }
18640b57cec5SDimitry Andric }
18650b57cec5SDimitry Andric
18660b57cec5SDimitry Andric // Visit all relocations from all section contributions of this object file and
18670b57cec5SDimitry Andric // mark the relocation target as address-taken.
markSymbolsWithRelocations(ObjFile * file,SymbolRVASet & usedSymbols)1868bdd1243dSDimitry Andric void Writer::markSymbolsWithRelocations(ObjFile *file,
18690b57cec5SDimitry Andric SymbolRVASet &usedSymbols) {
18700b57cec5SDimitry Andric for (Chunk *c : file->getChunks()) {
18710b57cec5SDimitry Andric // We only care about live section chunks. Common chunks and other chunks
18720b57cec5SDimitry Andric // don't generally contain relocations.
18730b57cec5SDimitry Andric SectionChunk *sc = dyn_cast<SectionChunk>(c);
18740b57cec5SDimitry Andric if (!sc || !sc->live)
18750b57cec5SDimitry Andric continue;
18760b57cec5SDimitry Andric
18770b57cec5SDimitry Andric for (const coff_relocation &reloc : sc->getRelocs()) {
1878bdd1243dSDimitry Andric if (ctx.config.machine == I386 &&
1879bdd1243dSDimitry Andric reloc.Type == COFF::IMAGE_REL_I386_REL32)
18800b57cec5SDimitry Andric // Ignore relative relocations on x86. On x86_64 they can't be ignored
18810b57cec5SDimitry Andric // since they're also used to compute absolute addresses.
18820b57cec5SDimitry Andric continue;
18830b57cec5SDimitry Andric
18840b57cec5SDimitry Andric Symbol *ref = sc->file->getSymbol(reloc.SymbolTableIndex);
18850b57cec5SDimitry Andric maybeAddAddressTakenFunction(usedSymbols, ref);
18860b57cec5SDimitry Andric }
18870b57cec5SDimitry Andric }
18880b57cec5SDimitry Andric }
18890b57cec5SDimitry Andric
18900b57cec5SDimitry Andric // Create the guard function id table. This is a table of RVAs of all
18910b57cec5SDimitry Andric // address-taken functions. It is sorted and uniqued, just like the safe SEH
18920b57cec5SDimitry Andric // table.
createGuardCFTables()18930b57cec5SDimitry Andric void Writer::createGuardCFTables() {
1894bdd1243dSDimitry Andric Configuration *config = &ctx.config;
1895bdd1243dSDimitry Andric
18960b57cec5SDimitry Andric SymbolRVASet addressTakenSyms;
1897e8d8bef9SDimitry Andric SymbolRVASet giatsRVASet;
1898e8d8bef9SDimitry Andric std::vector<Symbol *> giatsSymbols;
18990b57cec5SDimitry Andric SymbolRVASet longJmpTargets;
1900fe6060f1SDimitry Andric SymbolRVASet ehContTargets;
1901349cc55cSDimitry Andric for (ObjFile *file : ctx.objFileInstances) {
19020b57cec5SDimitry Andric // If the object was compiled with /guard:cf, the address taken symbols
190306c3fb27SDimitry Andric // are in .gfids$y sections, and the longjmp targets are in .gljmp$y
190406c3fb27SDimitry Andric // sections. If the object was not compiled with /guard:cf, we assume there
190506c3fb27SDimitry Andric // were no setjmp targets, and that all code symbols with relocations are
190606c3fb27SDimitry Andric // possibly address-taken.
19070b57cec5SDimitry Andric if (file->hasGuardCF()) {
19080b57cec5SDimitry Andric markSymbolsForRVATable(file, file->getGuardFidChunks(), addressTakenSyms);
1909e8d8bef9SDimitry Andric markSymbolsForRVATable(file, file->getGuardIATChunks(), giatsRVASet);
1910e8d8bef9SDimitry Andric getSymbolsFromSections(file, file->getGuardIATChunks(), giatsSymbols);
19110b57cec5SDimitry Andric markSymbolsForRVATable(file, file->getGuardLJmpChunks(), longJmpTargets);
19120b57cec5SDimitry Andric } else {
19130b57cec5SDimitry Andric markSymbolsWithRelocations(file, addressTakenSyms);
19140b57cec5SDimitry Andric }
191506c3fb27SDimitry Andric // If the object was compiled with /guard:ehcont, the ehcont targets are in
191606c3fb27SDimitry Andric // .gehcont$y sections.
191706c3fb27SDimitry Andric if (file->hasGuardEHCont())
191806c3fb27SDimitry Andric markSymbolsForRVATable(file, file->getGuardEHContChunks(), ehContTargets);
19190b57cec5SDimitry Andric }
19200b57cec5SDimitry Andric
19210b57cec5SDimitry Andric // Mark the image entry as address-taken.
19220b57cec5SDimitry Andric if (config->entry)
19230b57cec5SDimitry Andric maybeAddAddressTakenFunction(addressTakenSyms, config->entry);
19240b57cec5SDimitry Andric
19250b57cec5SDimitry Andric // Mark exported symbols in executable sections as address-taken.
19260b57cec5SDimitry Andric for (Export &e : config->exports)
19270b57cec5SDimitry Andric maybeAddAddressTakenFunction(addressTakenSyms, e.sym);
19280b57cec5SDimitry Andric
1929e8d8bef9SDimitry Andric // For each entry in the .giats table, check if it has a corresponding load
1930e8d8bef9SDimitry Andric // thunk (e.g. because the DLL that defines it will be delay-loaded) and, if
1931e8d8bef9SDimitry Andric // so, add the load thunk to the address taken (.gfids) table.
1932e8d8bef9SDimitry Andric for (Symbol *s : giatsSymbols) {
1933e8d8bef9SDimitry Andric if (auto *di = dyn_cast<DefinedImportData>(s)) {
1934e8d8bef9SDimitry Andric if (di->loadThunkSym)
1935e8d8bef9SDimitry Andric addSymbolToRVASet(addressTakenSyms, di->loadThunkSym);
1936e8d8bef9SDimitry Andric }
1937e8d8bef9SDimitry Andric }
1938e8d8bef9SDimitry Andric
19390b57cec5SDimitry Andric // Ensure sections referenced in the gfid table are 16-byte aligned.
19400b57cec5SDimitry Andric for (const ChunkAndOffset &c : addressTakenSyms)
19410b57cec5SDimitry Andric if (c.inputChunk->getAlignment() < 16)
19420b57cec5SDimitry Andric c.inputChunk->setAlignment(16);
19430b57cec5SDimitry Andric
19440b57cec5SDimitry Andric maybeAddRVATable(std::move(addressTakenSyms), "__guard_fids_table",
19450b57cec5SDimitry Andric "__guard_fids_count");
19460b57cec5SDimitry Andric
1947e8d8bef9SDimitry Andric // Add the Guard Address Taken IAT Entry Table (.giats).
1948e8d8bef9SDimitry Andric maybeAddRVATable(std::move(giatsRVASet), "__guard_iat_table",
1949e8d8bef9SDimitry Andric "__guard_iat_count");
1950e8d8bef9SDimitry Andric
19510b57cec5SDimitry Andric // Add the longjmp target table unless the user told us not to.
1952fe6060f1SDimitry Andric if (config->guardCF & GuardCFLevel::LongJmp)
19530b57cec5SDimitry Andric maybeAddRVATable(std::move(longJmpTargets), "__guard_longjmp_table",
19540b57cec5SDimitry Andric "__guard_longjmp_count");
19550b57cec5SDimitry Andric
1956fe6060f1SDimitry Andric // Add the ehcont target table unless the user told us not to.
1957fe6060f1SDimitry Andric if (config->guardCF & GuardCFLevel::EHCont)
1958fe6060f1SDimitry Andric maybeAddRVATable(std::move(ehContTargets), "__guard_eh_cont_table",
19594542f901SDimitry Andric "__guard_eh_cont_count");
1960fe6060f1SDimitry Andric
19610b57cec5SDimitry Andric // Set __guard_flags, which will be used in the load config to indicate that
19620b57cec5SDimitry Andric // /guard:cf was enabled.
1963bdd1243dSDimitry Andric uint32_t guardFlags = uint32_t(GuardFlags::CF_INSTRUMENTED) |
1964bdd1243dSDimitry Andric uint32_t(GuardFlags::CF_FUNCTION_TABLE_PRESENT);
1965fe6060f1SDimitry Andric if (config->guardCF & GuardCFLevel::LongJmp)
1966bdd1243dSDimitry Andric guardFlags |= uint32_t(GuardFlags::CF_LONGJUMP_TABLE_PRESENT);
1967fe6060f1SDimitry Andric if (config->guardCF & GuardCFLevel::EHCont)
1968bdd1243dSDimitry Andric guardFlags |= uint32_t(GuardFlags::EH_CONTINUATION_TABLE_PRESENT);
1969349cc55cSDimitry Andric Symbol *flagSym = ctx.symtab.findUnderscore("__guard_flags");
19700b57cec5SDimitry Andric cast<DefinedAbsolute>(flagSym)->setVA(guardFlags);
19710b57cec5SDimitry Andric }
19720b57cec5SDimitry Andric
19730b57cec5SDimitry Andric // Take a list of input sections containing symbol table indices and add those
1974e8d8bef9SDimitry Andric // symbols to a vector. The challenge is that symbol RVAs are not known and
19750b57cec5SDimitry Andric // depend on the table size, so we can't directly build a set of integers.
getSymbolsFromSections(ObjFile * file,ArrayRef<SectionChunk * > symIdxChunks,std::vector<Symbol * > & symbols)1976e8d8bef9SDimitry Andric void Writer::getSymbolsFromSections(ObjFile *file,
19770b57cec5SDimitry Andric ArrayRef<SectionChunk *> symIdxChunks,
1978e8d8bef9SDimitry Andric std::vector<Symbol *> &symbols) {
19790b57cec5SDimitry Andric for (SectionChunk *c : symIdxChunks) {
19800b57cec5SDimitry Andric // Skip sections discarded by linker GC. This comes up when a .gfids section
19810b57cec5SDimitry Andric // is associated with something like a vtable and the vtable is discarded.
19820b57cec5SDimitry Andric // In this case, the associated gfids section is discarded, and we don't
19830b57cec5SDimitry Andric // mark the virtual member functions as address-taken by the vtable.
19840b57cec5SDimitry Andric if (!c->live)
19850b57cec5SDimitry Andric continue;
19860b57cec5SDimitry Andric
19870b57cec5SDimitry Andric // Validate that the contents look like symbol table indices.
19880b57cec5SDimitry Andric ArrayRef<uint8_t> data = c->getContents();
19890b57cec5SDimitry Andric if (data.size() % 4 != 0) {
19900b57cec5SDimitry Andric warn("ignoring " + c->getSectionName() +
19910b57cec5SDimitry Andric " symbol table index section in object " + toString(file));
19920b57cec5SDimitry Andric continue;
19930b57cec5SDimitry Andric }
19940b57cec5SDimitry Andric
19950b57cec5SDimitry Andric // Read each symbol table index and check if that symbol was included in the
1996e8d8bef9SDimitry Andric // final link. If so, add it to the vector of symbols.
19970b57cec5SDimitry Andric ArrayRef<ulittle32_t> symIndices(
19980b57cec5SDimitry Andric reinterpret_cast<const ulittle32_t *>(data.data()), data.size() / 4);
19990b57cec5SDimitry Andric ArrayRef<Symbol *> objSymbols = file->getSymbols();
20000b57cec5SDimitry Andric for (uint32_t symIndex : symIndices) {
20010b57cec5SDimitry Andric if (symIndex >= objSymbols.size()) {
20020b57cec5SDimitry Andric warn("ignoring invalid symbol table index in section " +
20030b57cec5SDimitry Andric c->getSectionName() + " in object " + toString(file));
20040b57cec5SDimitry Andric continue;
20050b57cec5SDimitry Andric }
20060b57cec5SDimitry Andric if (Symbol *s = objSymbols[symIndex]) {
20070b57cec5SDimitry Andric if (s->isLive())
2008e8d8bef9SDimitry Andric symbols.push_back(cast<Symbol>(s));
2009e8d8bef9SDimitry Andric }
2010e8d8bef9SDimitry Andric }
2011e8d8bef9SDimitry Andric }
2012e8d8bef9SDimitry Andric }
2013e8d8bef9SDimitry Andric
2014e8d8bef9SDimitry Andric // Take a list of input sections containing symbol table indices and add those
2015e8d8bef9SDimitry Andric // symbols to an RVA table.
markSymbolsForRVATable(ObjFile * file,ArrayRef<SectionChunk * > symIdxChunks,SymbolRVASet & tableSymbols)2016e8d8bef9SDimitry Andric void Writer::markSymbolsForRVATable(ObjFile *file,
2017e8d8bef9SDimitry Andric ArrayRef<SectionChunk *> symIdxChunks,
2018e8d8bef9SDimitry Andric SymbolRVASet &tableSymbols) {
2019e8d8bef9SDimitry Andric std::vector<Symbol *> syms;
2020e8d8bef9SDimitry Andric getSymbolsFromSections(file, symIdxChunks, syms);
2021e8d8bef9SDimitry Andric
2022e8d8bef9SDimitry Andric for (Symbol *s : syms)
20230b57cec5SDimitry Andric addSymbolToRVASet(tableSymbols, cast<Defined>(s));
20240b57cec5SDimitry Andric }
20250b57cec5SDimitry Andric
20260b57cec5SDimitry Andric // Replace the absolute table symbol with a synthetic symbol pointing to
20270b57cec5SDimitry Andric // tableChunk so that we can emit base relocations for it and resolve section
20280b57cec5SDimitry Andric // relative relocations.
maybeAddRVATable(SymbolRVASet tableSymbols,StringRef tableSym,StringRef countSym,bool hasFlag)20290b57cec5SDimitry Andric void Writer::maybeAddRVATable(SymbolRVASet tableSymbols, StringRef tableSym,
2030fe6060f1SDimitry Andric StringRef countSym, bool hasFlag) {
20310b57cec5SDimitry Andric if (tableSymbols.empty())
20320b57cec5SDimitry Andric return;
20330b57cec5SDimitry Andric
2034fe6060f1SDimitry Andric NonSectionChunk *tableChunk;
2035fe6060f1SDimitry Andric if (hasFlag)
2036fe6060f1SDimitry Andric tableChunk = make<RVAFlagTableChunk>(std::move(tableSymbols));
2037fe6060f1SDimitry Andric else
2038fe6060f1SDimitry Andric tableChunk = make<RVATableChunk>(std::move(tableSymbols));
20390b57cec5SDimitry Andric rdataSec->addChunk(tableChunk);
20400b57cec5SDimitry Andric
2041349cc55cSDimitry Andric Symbol *t = ctx.symtab.findUnderscore(tableSym);
2042349cc55cSDimitry Andric Symbol *c = ctx.symtab.findUnderscore(countSym);
20430b57cec5SDimitry Andric replaceSymbol<DefinedSynthetic>(t, t->getName(), tableChunk);
2044fe6060f1SDimitry Andric cast<DefinedAbsolute>(c)->setVA(tableChunk->getSize() / (hasFlag ? 5 : 4));
20450b57cec5SDimitry Andric }
20460b57cec5SDimitry Andric
20475f757f3fSDimitry Andric // Create CHPE metadata chunks.
createECChunks()20485f757f3fSDimitry Andric void Writer::createECChunks() {
20495f757f3fSDimitry Andric auto codeMapChunk = make<ECCodeMapChunk>(codeMap);
20505f757f3fSDimitry Andric rdataSec->addChunk(codeMapChunk);
20515f757f3fSDimitry Andric Symbol *codeMapSym = ctx.symtab.findUnderscore("__hybrid_code_map");
20525f757f3fSDimitry Andric replaceSymbol<DefinedSynthetic>(codeMapSym, codeMapSym->getName(),
20535f757f3fSDimitry Andric codeMapChunk);
20545f757f3fSDimitry Andric }
20555f757f3fSDimitry Andric
20560b57cec5SDimitry Andric // MinGW specific. Gather all relocations that are imported from a DLL even
20570b57cec5SDimitry Andric // though the code didn't expect it to, produce the table that the runtime
20580b57cec5SDimitry Andric // uses for fixing them up, and provide the synthetic symbols that the
20590b57cec5SDimitry Andric // runtime uses for finding the table.
createRuntimePseudoRelocs()20600b57cec5SDimitry Andric void Writer::createRuntimePseudoRelocs() {
20610b57cec5SDimitry Andric std::vector<RuntimePseudoReloc> rels;
20620b57cec5SDimitry Andric
2063349cc55cSDimitry Andric for (Chunk *c : ctx.symtab.getChunks()) {
20640b57cec5SDimitry Andric auto *sc = dyn_cast<SectionChunk>(c);
20650b57cec5SDimitry Andric if (!sc || !sc->live)
20660b57cec5SDimitry Andric continue;
2067*0fca6ea1SDimitry Andric // Don't create pseudo relocations for sections that won't be
2068*0fca6ea1SDimitry Andric // mapped at runtime.
2069*0fca6ea1SDimitry Andric if (sc->header->Characteristics & IMAGE_SCN_MEM_DISCARDABLE)
2070*0fca6ea1SDimitry Andric continue;
20710b57cec5SDimitry Andric sc->getRuntimePseudoRelocs(rels);
20720b57cec5SDimitry Andric }
20730b57cec5SDimitry Andric
2074bdd1243dSDimitry Andric if (!ctx.config.pseudoRelocs) {
20755ffd83dbSDimitry Andric // Not writing any pseudo relocs; if some were needed, error out and
20765ffd83dbSDimitry Andric // indicate what required them.
20775ffd83dbSDimitry Andric for (const RuntimePseudoReloc &rpr : rels)
20785ffd83dbSDimitry Andric error("automatic dllimport of " + rpr.sym->getName() + " in " +
20795ffd83dbSDimitry Andric toString(rpr.target->file) + " requires pseudo relocations");
20805ffd83dbSDimitry Andric return;
20815ffd83dbSDimitry Andric }
20825ffd83dbSDimitry Andric
2083*0fca6ea1SDimitry Andric if (!rels.empty()) {
20840b57cec5SDimitry Andric log("Writing " + Twine(rels.size()) + " runtime pseudo relocations");
2085*0fca6ea1SDimitry Andric const char *symbolName = "_pei386_runtime_relocator";
2086*0fca6ea1SDimitry Andric Symbol *relocator = ctx.symtab.findUnderscore(symbolName);
2087*0fca6ea1SDimitry Andric if (!relocator)
2088*0fca6ea1SDimitry Andric error("output image has runtime pseudo relocations, but the function " +
2089*0fca6ea1SDimitry Andric Twine(symbolName) +
2090*0fca6ea1SDimitry Andric " is missing; it is needed for fixing the relocations at runtime");
2091*0fca6ea1SDimitry Andric }
2092*0fca6ea1SDimitry Andric
20930b57cec5SDimitry Andric PseudoRelocTableChunk *table = make<PseudoRelocTableChunk>(rels);
20940b57cec5SDimitry Andric rdataSec->addChunk(table);
20950b57cec5SDimitry Andric EmptyChunk *endOfList = make<EmptyChunk>();
20960b57cec5SDimitry Andric rdataSec->addChunk(endOfList);
20970b57cec5SDimitry Andric
2098349cc55cSDimitry Andric Symbol *headSym = ctx.symtab.findUnderscore("__RUNTIME_PSEUDO_RELOC_LIST__");
2099349cc55cSDimitry Andric Symbol *endSym =
2100349cc55cSDimitry Andric ctx.symtab.findUnderscore("__RUNTIME_PSEUDO_RELOC_LIST_END__");
21010b57cec5SDimitry Andric replaceSymbol<DefinedSynthetic>(headSym, headSym->getName(), table);
21020b57cec5SDimitry Andric replaceSymbol<DefinedSynthetic>(endSym, endSym->getName(), endOfList);
21030b57cec5SDimitry Andric }
21040b57cec5SDimitry Andric
21050b57cec5SDimitry Andric // MinGW specific.
21060b57cec5SDimitry Andric // The MinGW .ctors and .dtors lists have sentinels at each end;
21070b57cec5SDimitry Andric // a (uintptr_t)-1 at the start and a (uintptr_t)0 at the end.
21080b57cec5SDimitry Andric // There's a symbol pointing to the start sentinel pointer, __CTOR_LIST__
21090b57cec5SDimitry Andric // and __DTOR_LIST__ respectively.
insertCtorDtorSymbols()21100b57cec5SDimitry Andric void Writer::insertCtorDtorSymbols() {
2111bdd1243dSDimitry Andric AbsolutePointerChunk *ctorListHead = make<AbsolutePointerChunk>(ctx, -1);
2112bdd1243dSDimitry Andric AbsolutePointerChunk *ctorListEnd = make<AbsolutePointerChunk>(ctx, 0);
2113bdd1243dSDimitry Andric AbsolutePointerChunk *dtorListHead = make<AbsolutePointerChunk>(ctx, -1);
2114bdd1243dSDimitry Andric AbsolutePointerChunk *dtorListEnd = make<AbsolutePointerChunk>(ctx, 0);
21150b57cec5SDimitry Andric ctorsSec->insertChunkAtStart(ctorListHead);
21160b57cec5SDimitry Andric ctorsSec->addChunk(ctorListEnd);
21170b57cec5SDimitry Andric dtorsSec->insertChunkAtStart(dtorListHead);
21180b57cec5SDimitry Andric dtorsSec->addChunk(dtorListEnd);
21190b57cec5SDimitry Andric
2120349cc55cSDimitry Andric Symbol *ctorListSym = ctx.symtab.findUnderscore("__CTOR_LIST__");
2121349cc55cSDimitry Andric Symbol *dtorListSym = ctx.symtab.findUnderscore("__DTOR_LIST__");
21220b57cec5SDimitry Andric replaceSymbol<DefinedSynthetic>(ctorListSym, ctorListSym->getName(),
21230b57cec5SDimitry Andric ctorListHead);
21240b57cec5SDimitry Andric replaceSymbol<DefinedSynthetic>(dtorListSym, dtorListSym->getName(),
21250b57cec5SDimitry Andric dtorListHead);
21260b57cec5SDimitry Andric }
21270b57cec5SDimitry Andric
21280b57cec5SDimitry Andric // Handles /section options to allow users to overwrite
21290b57cec5SDimitry Andric // section attributes.
setSectionPermissions()21300b57cec5SDimitry Andric void Writer::setSectionPermissions() {
21315f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Sections permissions");
2132bdd1243dSDimitry Andric for (auto &p : ctx.config.section) {
21330b57cec5SDimitry Andric StringRef name = p.first;
21340b57cec5SDimitry Andric uint32_t perm = p.second;
2135349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections)
21360b57cec5SDimitry Andric if (sec->name == name)
21370b57cec5SDimitry Andric sec->setPermissions(perm);
21380b57cec5SDimitry Andric }
21390b57cec5SDimitry Andric }
21400b57cec5SDimitry Andric
21415f757f3fSDimitry Andric // Set symbols used by ARM64EC metadata.
setECSymbols()21425f757f3fSDimitry Andric void Writer::setECSymbols() {
21435f757f3fSDimitry Andric if (!isArm64EC(ctx.config.machine))
21445f757f3fSDimitry Andric return;
21455f757f3fSDimitry Andric
21465f757f3fSDimitry Andric Symbol *rfeTableSym = ctx.symtab.findUnderscore("__arm64x_extra_rfe_table");
21475f757f3fSDimitry Andric replaceSymbol<DefinedSynthetic>(rfeTableSym, "__arm64x_extra_rfe_table",
21485f757f3fSDimitry Andric pdata.first);
21495f757f3fSDimitry Andric
21505f757f3fSDimitry Andric if (pdata.first) {
21515f757f3fSDimitry Andric Symbol *rfeSizeSym =
21525f757f3fSDimitry Andric ctx.symtab.findUnderscore("__arm64x_extra_rfe_table_size");
21535f757f3fSDimitry Andric cast<DefinedAbsolute>(rfeSizeSym)
21545f757f3fSDimitry Andric ->setVA(pdata.last->getRVA() + pdata.last->getSize() -
21555f757f3fSDimitry Andric pdata.first->getRVA());
21565f757f3fSDimitry Andric }
21575f757f3fSDimitry Andric }
21585f757f3fSDimitry Andric
21590b57cec5SDimitry Andric // Write section contents to a mmap'ed file.
writeSections()21600b57cec5SDimitry Andric void Writer::writeSections() {
21615f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Write sections");
21620b57cec5SDimitry Andric uint8_t *buf = buffer->getBufferStart();
2163349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
21640b57cec5SDimitry Andric uint8_t *secBuf = buf + sec->getFileOff();
21650b57cec5SDimitry Andric // Fill gaps between functions in .text with INT3 instructions
21660b57cec5SDimitry Andric // instead of leaving as NUL bytes (which can be interpreted as
21675f757f3fSDimitry Andric // ADD instructions). Only fill the gaps between chunks. Most
21685f757f3fSDimitry Andric // chunks overwrite it anyway, but uninitialized data chunks
21695f757f3fSDimitry Andric // merged into a code section don't.
217006c3fb27SDimitry Andric if ((sec->header.Characteristics & IMAGE_SCN_CNT_CODE) &&
21715f757f3fSDimitry Andric (ctx.config.machine == AMD64 || ctx.config.machine == I386)) {
21725f757f3fSDimitry Andric uint32_t prevEnd = 0;
21735f757f3fSDimitry Andric for (Chunk *c : sec->chunks) {
21745f757f3fSDimitry Andric uint32_t off = c->getRVA() - sec->getRVA();
21755f757f3fSDimitry Andric memset(secBuf + prevEnd, 0xCC, off - prevEnd);
21765f757f3fSDimitry Andric prevEnd = off + c->getSize();
21775f757f3fSDimitry Andric }
21785f757f3fSDimitry Andric memset(secBuf + prevEnd, 0xCC, sec->getRawSize() - prevEnd);
21795f757f3fSDimitry Andric }
21805f757f3fSDimitry Andric
21810b57cec5SDimitry Andric parallelForEach(sec->chunks, [&](Chunk *c) {
21820b57cec5SDimitry Andric c->writeTo(secBuf + c->getRVA() - sec->getRVA());
21830b57cec5SDimitry Andric });
21840b57cec5SDimitry Andric }
21850b57cec5SDimitry Andric }
21860b57cec5SDimitry Andric
writeBuildId()21870b57cec5SDimitry Andric void Writer::writeBuildId() {
21885f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Write build ID");
21895f757f3fSDimitry Andric
21900b57cec5SDimitry Andric // There are two important parts to the build ID.
21910b57cec5SDimitry Andric // 1) If building with debug info, the COFF debug directory contains a
21920b57cec5SDimitry Andric // timestamp as well as a Guid and Age of the PDB.
21930b57cec5SDimitry Andric // 2) In all cases, the PE COFF file header also contains a timestamp.
21940b57cec5SDimitry Andric // For reproducibility, instead of a timestamp we want to use a hash of the
21950b57cec5SDimitry Andric // PE contents.
2196bdd1243dSDimitry Andric Configuration *config = &ctx.config;
21975f757f3fSDimitry Andric bool generateSyntheticBuildId = config->buildIDHash == BuildIDHash::Binary;
21985f757f3fSDimitry Andric if (generateSyntheticBuildId) {
21990b57cec5SDimitry Andric assert(buildId && "BuildId is not set!");
22000b57cec5SDimitry Andric // BuildId->BuildId was filled in when the PDB was written.
22010b57cec5SDimitry Andric }
22020b57cec5SDimitry Andric
22030b57cec5SDimitry Andric // At this point the only fields in the COFF file which remain unset are the
22040b57cec5SDimitry Andric // "timestamp" in the COFF file header, and the ones in the coff debug
22050b57cec5SDimitry Andric // directory. Now we can hash the file and write that hash to the various
22060b57cec5SDimitry Andric // timestamp fields in the file.
22070b57cec5SDimitry Andric StringRef outputFileData(
22080b57cec5SDimitry Andric reinterpret_cast<const char *>(buffer->getBufferStart()),
22090b57cec5SDimitry Andric buffer->getBufferSize());
22100b57cec5SDimitry Andric
22110b57cec5SDimitry Andric uint32_t timestamp = config->timestamp;
22120b57cec5SDimitry Andric uint64_t hash = 0;
22130b57cec5SDimitry Andric
22140b57cec5SDimitry Andric if (config->repro || generateSyntheticBuildId)
221506c3fb27SDimitry Andric hash = xxh3_64bits(outputFileData);
22160b57cec5SDimitry Andric
22170b57cec5SDimitry Andric if (config->repro)
22180b57cec5SDimitry Andric timestamp = static_cast<uint32_t>(hash);
22190b57cec5SDimitry Andric
22200b57cec5SDimitry Andric if (generateSyntheticBuildId) {
22210b57cec5SDimitry Andric buildId->buildId->PDB70.CVSignature = OMF::Signature::PDB70;
22220b57cec5SDimitry Andric buildId->buildId->PDB70.Age = 1;
22230b57cec5SDimitry Andric memcpy(buildId->buildId->PDB70.Signature, &hash, 8);
22240b57cec5SDimitry Andric // xxhash only gives us 8 bytes, so put some fixed data in the other half.
22250b57cec5SDimitry Andric memcpy(&buildId->buildId->PDB70.Signature[8], "LLD PDB.", 8);
22260b57cec5SDimitry Andric }
22270b57cec5SDimitry Andric
22280b57cec5SDimitry Andric if (debugDirectory)
22290b57cec5SDimitry Andric debugDirectory->setTimeDateStamp(timestamp);
22300b57cec5SDimitry Andric
22310b57cec5SDimitry Andric uint8_t *buf = buffer->getBufferStart();
22320b57cec5SDimitry Andric buf += dosStubSize + sizeof(PEMagic);
22330b57cec5SDimitry Andric object::coff_file_header *coffHeader =
22340b57cec5SDimitry Andric reinterpret_cast<coff_file_header *>(buf);
22350b57cec5SDimitry Andric coffHeader->TimeDateStamp = timestamp;
22360b57cec5SDimitry Andric }
22370b57cec5SDimitry Andric
22380b57cec5SDimitry Andric // Sort .pdata section contents according to PE/COFF spec 5.5.
22395f757f3fSDimitry Andric template <typename T>
sortExceptionTable(ChunkRange & exceptionTable)22405f757f3fSDimitry Andric void Writer::sortExceptionTable(ChunkRange &exceptionTable) {
22415f757f3fSDimitry Andric if (!exceptionTable.first)
22420b57cec5SDimitry Andric return;
22435f757f3fSDimitry Andric
22440b57cec5SDimitry Andric // We assume .pdata contains function table entries only.
22450b57cec5SDimitry Andric auto bufAddr = [&](Chunk *c) {
2246349cc55cSDimitry Andric OutputSection *os = ctx.getOutputSection(c);
22470b57cec5SDimitry Andric return buffer->getBufferStart() + os->getFileOff() + c->getRVA() -
22480b57cec5SDimitry Andric os->getRVA();
22490b57cec5SDimitry Andric };
22505f757f3fSDimitry Andric uint8_t *begin = bufAddr(exceptionTable.first);
22515f757f3fSDimitry Andric uint8_t *end = bufAddr(exceptionTable.last) + exceptionTable.last->getSize();
22525f757f3fSDimitry Andric if ((end - begin) % sizeof(T) != 0) {
22535ffd83dbSDimitry Andric fatal("unexpected .pdata size: " + Twine(end - begin) +
22545f757f3fSDimitry Andric " is not a multiple of " + Twine(sizeof(T)));
22555ffd83dbSDimitry Andric }
22565f757f3fSDimitry Andric
22575f757f3fSDimitry Andric parallelSort(MutableArrayRef<T>(reinterpret_cast<T *>(begin),
22585f757f3fSDimitry Andric reinterpret_cast<T *>(end)),
22595f757f3fSDimitry Andric [](const T &a, const T &b) { return a.begin < b.begin; });
22600b57cec5SDimitry Andric }
22615f757f3fSDimitry Andric
22625f757f3fSDimitry Andric // Sort .pdata section contents according to PE/COFF spec 5.5.
sortExceptionTables()22635f757f3fSDimitry Andric void Writer::sortExceptionTables() {
22645f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Sort exception table");
22655f757f3fSDimitry Andric
22665f757f3fSDimitry Andric struct EntryX64 {
22675f757f3fSDimitry Andric ulittle32_t begin, end, unwind;
22685f757f3fSDimitry Andric };
22695f757f3fSDimitry Andric struct EntryArm {
22705f757f3fSDimitry Andric ulittle32_t begin, unwind;
22715f757f3fSDimitry Andric };
22725f757f3fSDimitry Andric
22735f757f3fSDimitry Andric switch (ctx.config.machine) {
22745f757f3fSDimitry Andric case AMD64:
22755f757f3fSDimitry Andric sortExceptionTable<EntryX64>(pdata);
22765f757f3fSDimitry Andric break;
22775f757f3fSDimitry Andric case ARM64EC:
22785f757f3fSDimitry Andric case ARM64X:
22795f757f3fSDimitry Andric sortExceptionTable<EntryX64>(hybridPdata);
22805f757f3fSDimitry Andric [[fallthrough]];
22815f757f3fSDimitry Andric case ARMNT:
22825f757f3fSDimitry Andric case ARM64:
22835f757f3fSDimitry Andric sortExceptionTable<EntryArm>(pdata);
22845f757f3fSDimitry Andric break;
22855f757f3fSDimitry Andric default:
22865f757f3fSDimitry Andric if (pdata.first)
2287480093f4SDimitry Andric lld::errs() << "warning: don't know how to handle .pdata.\n";
22885f757f3fSDimitry Andric break;
22895f757f3fSDimitry Andric }
22900b57cec5SDimitry Andric }
22910b57cec5SDimitry Andric
22920b57cec5SDimitry Andric // The CRT section contains, among other things, the array of function
22930b57cec5SDimitry Andric // pointers that initialize every global variable that is not trivially
22940b57cec5SDimitry Andric // constructed. The CRT calls them one after the other prior to invoking
22950b57cec5SDimitry Andric // main().
22960b57cec5SDimitry Andric //
22970b57cec5SDimitry Andric // As per C++ spec, 3.6.2/2.3,
22980b57cec5SDimitry Andric // "Variables with ordered initialization defined within a single
22990b57cec5SDimitry Andric // translation unit shall be initialized in the order of their definitions
23000b57cec5SDimitry Andric // in the translation unit"
23010b57cec5SDimitry Andric //
23020b57cec5SDimitry Andric // It is therefore critical to sort the chunks containing the function
23030b57cec5SDimitry Andric // pointers in the order that they are listed in the object file (top to
23040b57cec5SDimitry Andric // bottom), otherwise global objects might not be initialized in the
23050b57cec5SDimitry Andric // correct order.
sortCRTSectionChunks(std::vector<Chunk * > & chunks)23060b57cec5SDimitry Andric void Writer::sortCRTSectionChunks(std::vector<Chunk *> &chunks) {
23070b57cec5SDimitry Andric auto sectionChunkOrder = [](const Chunk *a, const Chunk *b) {
23080b57cec5SDimitry Andric auto sa = dyn_cast<SectionChunk>(a);
23090b57cec5SDimitry Andric auto sb = dyn_cast<SectionChunk>(b);
23100b57cec5SDimitry Andric assert(sa && sb && "Non-section chunks in CRT section!");
23110b57cec5SDimitry Andric
23120b57cec5SDimitry Andric StringRef sAObj = sa->file->mb.getBufferIdentifier();
23130b57cec5SDimitry Andric StringRef sBObj = sb->file->mb.getBufferIdentifier();
23140b57cec5SDimitry Andric
23150b57cec5SDimitry Andric return sAObj == sBObj && sa->getSectionNumber() < sb->getSectionNumber();
23160b57cec5SDimitry Andric };
23170b57cec5SDimitry Andric llvm::stable_sort(chunks, sectionChunkOrder);
23180b57cec5SDimitry Andric
2319bdd1243dSDimitry Andric if (ctx.config.verbose) {
23200b57cec5SDimitry Andric for (auto &c : chunks) {
23210b57cec5SDimitry Andric auto sc = dyn_cast<SectionChunk>(c);
23220b57cec5SDimitry Andric log(" " + sc->file->mb.getBufferIdentifier().str() +
23230b57cec5SDimitry Andric ", SectionID: " + Twine(sc->getSectionNumber()));
23240b57cec5SDimitry Andric }
23250b57cec5SDimitry Andric }
23260b57cec5SDimitry Andric }
23270b57cec5SDimitry Andric
findSection(StringRef name)23280b57cec5SDimitry Andric OutputSection *Writer::findSection(StringRef name) {
2329349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections)
23300b57cec5SDimitry Andric if (sec->name == name)
23310b57cec5SDimitry Andric return sec;
23320b57cec5SDimitry Andric return nullptr;
23330b57cec5SDimitry Andric }
23340b57cec5SDimitry Andric
getSizeOfInitializedData()23350b57cec5SDimitry Andric uint32_t Writer::getSizeOfInitializedData() {
23360b57cec5SDimitry Andric uint32_t res = 0;
2337349cc55cSDimitry Andric for (OutputSection *s : ctx.outputSections)
23380b57cec5SDimitry Andric if (s->header.Characteristics & IMAGE_SCN_CNT_INITIALIZED_DATA)
23390b57cec5SDimitry Andric res += s->getRawSize();
23400b57cec5SDimitry Andric return res;
23410b57cec5SDimitry Andric }
23420b57cec5SDimitry Andric
23430b57cec5SDimitry Andric // Add base relocations to .reloc section.
addBaserels()23440b57cec5SDimitry Andric void Writer::addBaserels() {
2345bdd1243dSDimitry Andric if (!ctx.config.relocatable)
23460b57cec5SDimitry Andric return;
23470b57cec5SDimitry Andric relocSec->chunks.clear();
23480b57cec5SDimitry Andric std::vector<Baserel> v;
2349349cc55cSDimitry Andric for (OutputSection *sec : ctx.outputSections) {
23500b57cec5SDimitry Andric if (sec->header.Characteristics & IMAGE_SCN_MEM_DISCARDABLE)
23510b57cec5SDimitry Andric continue;
23525f757f3fSDimitry Andric llvm::TimeTraceScope timeScope("Base relocations: ", sec->name);
23530b57cec5SDimitry Andric // Collect all locations for base relocations.
23540b57cec5SDimitry Andric for (Chunk *c : sec->chunks)
23550b57cec5SDimitry Andric c->getBaserels(&v);
23560b57cec5SDimitry Andric // Add the addresses to .reloc section.
23570b57cec5SDimitry Andric if (!v.empty())
23580b57cec5SDimitry Andric addBaserelBlocks(v);
23590b57cec5SDimitry Andric v.clear();
23600b57cec5SDimitry Andric }
23610b57cec5SDimitry Andric }
23620b57cec5SDimitry Andric
23630b57cec5SDimitry Andric // Add addresses to .reloc section. Note that addresses are grouped by page.
addBaserelBlocks(std::vector<Baserel> & v)23640b57cec5SDimitry Andric void Writer::addBaserelBlocks(std::vector<Baserel> &v) {
23650b57cec5SDimitry Andric const uint32_t mask = ~uint32_t(pageSize - 1);
23660b57cec5SDimitry Andric uint32_t page = v[0].rva & mask;
23670b57cec5SDimitry Andric size_t i = 0, j = 1;
23680b57cec5SDimitry Andric for (size_t e = v.size(); j < e; ++j) {
23690b57cec5SDimitry Andric uint32_t p = v[j].rva & mask;
23700b57cec5SDimitry Andric if (p == page)
23710b57cec5SDimitry Andric continue;
23720b57cec5SDimitry Andric relocSec->addChunk(make<BaserelChunk>(page, &v[i], &v[0] + j));
23730b57cec5SDimitry Andric i = j;
23740b57cec5SDimitry Andric page = p;
23750b57cec5SDimitry Andric }
23760b57cec5SDimitry Andric if (i == j)
23770b57cec5SDimitry Andric return;
23780b57cec5SDimitry Andric relocSec->addChunk(make<BaserelChunk>(page, &v[i], &v[0] + j));
23790b57cec5SDimitry Andric }
23800b57cec5SDimitry Andric
createPartialSection(StringRef name,uint32_t outChars)23810b57cec5SDimitry Andric PartialSection *Writer::createPartialSection(StringRef name,
23820b57cec5SDimitry Andric uint32_t outChars) {
23830b57cec5SDimitry Andric PartialSection *&pSec = partialSections[{name, outChars}];
23840b57cec5SDimitry Andric if (pSec)
23850b57cec5SDimitry Andric return pSec;
23860b57cec5SDimitry Andric pSec = make<PartialSection>(name, outChars);
23870b57cec5SDimitry Andric return pSec;
23880b57cec5SDimitry Andric }
23890b57cec5SDimitry Andric
findPartialSection(StringRef name,uint32_t outChars)23900b57cec5SDimitry Andric PartialSection *Writer::findPartialSection(StringRef name, uint32_t outChars) {
23910b57cec5SDimitry Andric auto it = partialSections.find({name, outChars});
23920b57cec5SDimitry Andric if (it != partialSections.end())
23930b57cec5SDimitry Andric return it->second;
23940b57cec5SDimitry Andric return nullptr;
23950b57cec5SDimitry Andric }
2396e8d8bef9SDimitry Andric
fixTlsAlignment()2397e8d8bef9SDimitry Andric void Writer::fixTlsAlignment() {
2398e8d8bef9SDimitry Andric Defined *tlsSym =
2399349cc55cSDimitry Andric dyn_cast_or_null<Defined>(ctx.symtab.findUnderscore("_tls_used"));
2400e8d8bef9SDimitry Andric if (!tlsSym)
2401e8d8bef9SDimitry Andric return;
2402e8d8bef9SDimitry Andric
2403349cc55cSDimitry Andric OutputSection *sec = ctx.getOutputSection(tlsSym->getChunk());
2404e8d8bef9SDimitry Andric assert(sec && tlsSym->getRVA() >= sec->getRVA() &&
2405e8d8bef9SDimitry Andric "no output section for _tls_used");
2406e8d8bef9SDimitry Andric
2407e8d8bef9SDimitry Andric uint8_t *secBuf = buffer->getBufferStart() + sec->getFileOff();
2408e8d8bef9SDimitry Andric uint64_t tlsOffset = tlsSym->getRVA() - sec->getRVA();
2409bdd1243dSDimitry Andric uint64_t directorySize = ctx.config.is64()
2410e8d8bef9SDimitry Andric ? sizeof(object::coff_tls_directory64)
2411e8d8bef9SDimitry Andric : sizeof(object::coff_tls_directory32);
2412e8d8bef9SDimitry Andric
2413e8d8bef9SDimitry Andric if (tlsOffset + directorySize > sec->getRawSize())
2414e8d8bef9SDimitry Andric fatal("_tls_used sym is malformed");
2415e8d8bef9SDimitry Andric
2416bdd1243dSDimitry Andric if (ctx.config.is64()) {
2417e8d8bef9SDimitry Andric object::coff_tls_directory64 *tlsDir =
2418e8d8bef9SDimitry Andric reinterpret_cast<object::coff_tls_directory64 *>(&secBuf[tlsOffset]);
2419e8d8bef9SDimitry Andric tlsDir->setAlignment(tlsAlignment);
2420e8d8bef9SDimitry Andric } else {
2421e8d8bef9SDimitry Andric object::coff_tls_directory32 *tlsDir =
2422e8d8bef9SDimitry Andric reinterpret_cast<object::coff_tls_directory32 *>(&secBuf[tlsOffset]);
2423e8d8bef9SDimitry Andric tlsDir->setAlignment(tlsAlignment);
2424e8d8bef9SDimitry Andric }
2425e8d8bef9SDimitry Andric }
2426bdd1243dSDimitry Andric
prepareLoadConfig()24275f757f3fSDimitry Andric void Writer::prepareLoadConfig() {
2428bdd1243dSDimitry Andric Symbol *sym = ctx.symtab.findUnderscore("_load_config_used");
2429bdd1243dSDimitry Andric auto *b = cast_if_present<DefinedRegular>(sym);
2430bdd1243dSDimitry Andric if (!b) {
2431bdd1243dSDimitry Andric if (ctx.config.guardCF != GuardCFLevel::Off)
2432bdd1243dSDimitry Andric warn("Control Flow Guard is enabled but '_load_config_used' is missing");
2433bdd1243dSDimitry Andric return;
2434bdd1243dSDimitry Andric }
2435bdd1243dSDimitry Andric
2436bdd1243dSDimitry Andric OutputSection *sec = ctx.getOutputSection(b->getChunk());
2437bdd1243dSDimitry Andric uint8_t *buf = buffer->getBufferStart();
2438bdd1243dSDimitry Andric uint8_t *secBuf = buf + sec->getFileOff();
2439bdd1243dSDimitry Andric uint8_t *symBuf = secBuf + (b->getRVA() - sec->getRVA());
2440bdd1243dSDimitry Andric uint32_t expectedAlign = ctx.config.is64() ? 8 : 4;
2441bdd1243dSDimitry Andric if (b->getChunk()->getAlignment() < expectedAlign)
2442bdd1243dSDimitry Andric warn("'_load_config_used' is misaligned (expected alignment to be " +
2443bdd1243dSDimitry Andric Twine(expectedAlign) + " bytes, got " +
2444bdd1243dSDimitry Andric Twine(b->getChunk()->getAlignment()) + " instead)");
2445bdd1243dSDimitry Andric else if (!isAligned(Align(expectedAlign), b->getRVA()))
2446bdd1243dSDimitry Andric warn("'_load_config_used' is misaligned (RVA is 0x" +
2447bdd1243dSDimitry Andric Twine::utohexstr(b->getRVA()) + " not aligned to " +
2448bdd1243dSDimitry Andric Twine(expectedAlign) + " bytes)");
2449bdd1243dSDimitry Andric
2450bdd1243dSDimitry Andric if (ctx.config.is64())
24515f757f3fSDimitry Andric prepareLoadConfig(reinterpret_cast<coff_load_configuration64 *>(symBuf));
2452bdd1243dSDimitry Andric else
24535f757f3fSDimitry Andric prepareLoadConfig(reinterpret_cast<coff_load_configuration32 *>(symBuf));
24545f757f3fSDimitry Andric }
24555f757f3fSDimitry Andric
prepareLoadConfig(T * loadConfig)24565f757f3fSDimitry Andric template <typename T> void Writer::prepareLoadConfig(T *loadConfig) {
24575f757f3fSDimitry Andric if (ctx.config.dependentLoadFlags)
24585f757f3fSDimitry Andric loadConfig->DependentLoadFlags = ctx.config.dependentLoadFlags;
24595f757f3fSDimitry Andric
24605f757f3fSDimitry Andric checkLoadConfigGuardData(loadConfig);
2461bdd1243dSDimitry Andric }
2462bdd1243dSDimitry Andric
2463bdd1243dSDimitry Andric template <typename T>
checkLoadConfigGuardData(const T * loadConfig)2464bdd1243dSDimitry Andric void Writer::checkLoadConfigGuardData(const T *loadConfig) {
2465bdd1243dSDimitry Andric size_t loadConfigSize = loadConfig->Size;
2466bdd1243dSDimitry Andric
2467bdd1243dSDimitry Andric #define RETURN_IF_NOT_CONTAINS(field) \
2468bdd1243dSDimitry Andric if (loadConfigSize < offsetof(T, field) + sizeof(T::field)) { \
2469bdd1243dSDimitry Andric warn("'_load_config_used' structure too small to include " #field); \
2470bdd1243dSDimitry Andric return; \
2471bdd1243dSDimitry Andric }
2472bdd1243dSDimitry Andric
2473bdd1243dSDimitry Andric #define IF_CONTAINS(field) \
2474bdd1243dSDimitry Andric if (loadConfigSize >= offsetof(T, field) + sizeof(T::field))
2475bdd1243dSDimitry Andric
2476bdd1243dSDimitry Andric #define CHECK_VA(field, sym) \
2477bdd1243dSDimitry Andric if (auto *s = dyn_cast<DefinedSynthetic>(ctx.symtab.findUnderscore(sym))) \
2478bdd1243dSDimitry Andric if (loadConfig->field != ctx.config.imageBase + s->getRVA()) \
2479bdd1243dSDimitry Andric warn(#field " not set correctly in '_load_config_used'");
2480bdd1243dSDimitry Andric
2481bdd1243dSDimitry Andric #define CHECK_ABSOLUTE(field, sym) \
2482bdd1243dSDimitry Andric if (auto *s = dyn_cast<DefinedAbsolute>(ctx.symtab.findUnderscore(sym))) \
2483bdd1243dSDimitry Andric if (loadConfig->field != s->getVA()) \
2484bdd1243dSDimitry Andric warn(#field " not set correctly in '_load_config_used'");
2485bdd1243dSDimitry Andric
2486bdd1243dSDimitry Andric if (ctx.config.guardCF == GuardCFLevel::Off)
2487bdd1243dSDimitry Andric return;
2488bdd1243dSDimitry Andric RETURN_IF_NOT_CONTAINS(GuardFlags)
2489bdd1243dSDimitry Andric CHECK_VA(GuardCFFunctionTable, "__guard_fids_table")
2490bdd1243dSDimitry Andric CHECK_ABSOLUTE(GuardCFFunctionCount, "__guard_fids_count")
2491bdd1243dSDimitry Andric CHECK_ABSOLUTE(GuardFlags, "__guard_flags")
2492bdd1243dSDimitry Andric IF_CONTAINS(GuardAddressTakenIatEntryCount) {
2493bdd1243dSDimitry Andric CHECK_VA(GuardAddressTakenIatEntryTable, "__guard_iat_table")
2494bdd1243dSDimitry Andric CHECK_ABSOLUTE(GuardAddressTakenIatEntryCount, "__guard_iat_count")
2495bdd1243dSDimitry Andric }
2496bdd1243dSDimitry Andric
2497bdd1243dSDimitry Andric if (!(ctx.config.guardCF & GuardCFLevel::LongJmp))
2498bdd1243dSDimitry Andric return;
2499bdd1243dSDimitry Andric RETURN_IF_NOT_CONTAINS(GuardLongJumpTargetCount)
2500bdd1243dSDimitry Andric CHECK_VA(GuardLongJumpTargetTable, "__guard_longjmp_table")
2501bdd1243dSDimitry Andric CHECK_ABSOLUTE(GuardLongJumpTargetCount, "__guard_longjmp_count")
2502bdd1243dSDimitry Andric
2503bdd1243dSDimitry Andric if (!(ctx.config.guardCF & GuardCFLevel::EHCont))
2504bdd1243dSDimitry Andric return;
2505bdd1243dSDimitry Andric RETURN_IF_NOT_CONTAINS(GuardEHContinuationCount)
2506bdd1243dSDimitry Andric CHECK_VA(GuardEHContinuationTable, "__guard_eh_cont_table")
2507bdd1243dSDimitry Andric CHECK_ABSOLUTE(GuardEHContinuationCount, "__guard_eh_cont_count")
2508bdd1243dSDimitry Andric
2509bdd1243dSDimitry Andric #undef RETURN_IF_NOT_CONTAINS
2510bdd1243dSDimitry Andric #undef IF_CONTAINS
2511bdd1243dSDimitry Andric #undef CHECK_VA
2512bdd1243dSDimitry Andric #undef CHECK_ABSOLUTE
2513bdd1243dSDimitry Andric }
2514