LLVM 19.0.0git
WasmObjectWriter.cpp
Go to the documentation of this file.
1//===- lib/MC/WasmObjectWriter.cpp - Wasm File Writer ---------------------===//
2//
3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4// See https://llvm.org/LICENSE.txt for license information.
5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6//
7//===----------------------------------------------------------------------===//
8//
9// This file implements Wasm object file writer information.
10//
11//===----------------------------------------------------------------------===//
12
13#include "llvm/ADT/STLExtras.h"
16#include "llvm/Config/llvm-config.h"
18#include "llvm/MC/MCAsmLayout.h"
19#include "llvm/MC/MCAssembler.h"
20#include "llvm/MC/MCContext.h"
21#include "llvm/MC/MCExpr.h"
26#include "llvm/MC/MCValue.h"
29#include "llvm/Support/Debug.h"
32#include "llvm/Support/LEB128.h"
33#include <vector>
34
35using namespace llvm;
36
37#define DEBUG_TYPE "mc"
38
39namespace {
40
41// When we create the indirect function table we start at 1, so that there is
42// and empty slot at 0 and therefore calling a null function pointer will trap.
43static const uint32_t InitialTableOffset = 1;
44
45// For patching purposes, we need to remember where each section starts, both
46// for patching up the section size field, and for patching up references to
47// locations within the section.
48struct SectionBookkeeping {
49 // Where the size of the section is written.
50 uint64_t SizeOffset;
51 // Where the section header ends (without custom section name).
52 uint64_t PayloadOffset;
53 // Where the contents of the section starts.
54 uint64_t ContentsOffset;
56};
57
58// A wasm data segment. A wasm binary contains only a single data section
59// but that can contain many segments, each with their own virtual location
60// in memory. Each MCSection data created by llvm is modeled as its own
61// wasm data segment.
62struct WasmDataSegment {
63 MCSectionWasm *Section;
64 StringRef Name;
65 uint32_t InitFlags;
66 uint64_t Offset;
67 uint32_t Alignment;
68 uint32_t LinkingFlags;
70};
71
72// A wasm function to be written into the function section.
73struct WasmFunction {
74 uint32_t SigIndex;
75 MCSection *Section;
76};
77
78// A wasm global to be written into the global section.
79struct WasmGlobal {
81 uint64_t InitialValue;
82};
83
84// Information about a single item which is part of a COMDAT. For each data
85// segment or function which is in the COMDAT, there is a corresponding
86// WasmComdatEntry.
87struct WasmComdatEntry {
88 unsigned Kind;
90};
91
92// Information about a single relocation.
93struct WasmRelocationEntry {
94 uint64_t Offset; // Where is the relocation.
95 const MCSymbolWasm *Symbol; // The symbol to relocate with.
96 int64_t Addend; // A value to add to the symbol.
97 unsigned Type; // The type of the relocation.
98 const MCSectionWasm *FixupSection; // The section the relocation is targeting.
99
100 WasmRelocationEntry(uint64_t Offset, const MCSymbolWasm *Symbol,
101 int64_t Addend, unsigned Type,
102 const MCSectionWasm *FixupSection)
103 : Offset(Offset), Symbol(Symbol), Addend(Addend), Type(Type),
104 FixupSection(FixupSection) {}
105
106 bool hasAddend() const { return wasm::relocTypeHasAddend(Type); }
107
108 void print(raw_ostream &Out) const {
109 Out << wasm::relocTypetoString(Type) << " Off=" << Offset
110 << ", Sym=" << *Symbol << ", Addend=" << Addend
111 << ", FixupSection=" << FixupSection->getName();
112 }
113
114#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
115 LLVM_DUMP_METHOD void dump() const { print(dbgs()); }
116#endif
117};
118
119static const uint32_t InvalidIndex = -1;
120
121struct WasmCustomSection {
122
123 StringRef Name;
124 MCSectionWasm *Section;
125
126 uint32_t OutputContentsOffset = 0;
127 uint32_t OutputIndex = InvalidIndex;
128
129 WasmCustomSection(StringRef Name, MCSectionWasm *Section)
130 : Name(Name), Section(Section) {}
131};
132
133#if !defined(NDEBUG)
134raw_ostream &operator<<(raw_ostream &OS, const WasmRelocationEntry &Rel) {
135 Rel.print(OS);
136 return OS;
137}
138#endif
139
140// Write Value as an (unsigned) LEB value at offset Offset in Stream, padded
141// to allow patching.
142template <typename T, int W>
143void writePatchableULEB(raw_pwrite_stream &Stream, T Value, uint64_t Offset) {
144 uint8_t Buffer[W];
145 unsigned SizeLen = encodeULEB128(Value, Buffer, W);
146 assert(SizeLen == W);
147 Stream.pwrite((char *)Buffer, SizeLen, Offset);
148}
149
150// Write Value as an signed LEB value at offset Offset in Stream, padded
151// to allow patching.
152template <typename T, int W>
153void writePatchableSLEB(raw_pwrite_stream &Stream, T Value, uint64_t Offset) {
154 uint8_t Buffer[W];
155 unsigned SizeLen = encodeSLEB128(Value, Buffer, W);
156 assert(SizeLen == W);
157 Stream.pwrite((char *)Buffer, SizeLen, Offset);
158}
159
160static void writePatchableU32(raw_pwrite_stream &Stream, uint32_t Value,
162 writePatchableULEB<uint32_t, 5>(Stream, Value, Offset);
163}
164
165static void writePatchableS32(raw_pwrite_stream &Stream, int32_t Value,
167 writePatchableSLEB<int32_t, 5>(Stream, Value, Offset);
168}
169
170static void writePatchableU64(raw_pwrite_stream &Stream, uint64_t Value,
172 writePatchableSLEB<uint64_t, 10>(Stream, Value, Offset);
173}
174
175static void writePatchableS64(raw_pwrite_stream &Stream, int64_t Value,
177 writePatchableSLEB<int64_t, 10>(Stream, Value, Offset);
178}
179
180// Write Value as a plain integer value at offset Offset in Stream.
181static void patchI32(raw_pwrite_stream &Stream, uint32_t Value,
183 uint8_t Buffer[4];
185 Stream.pwrite((char *)Buffer, sizeof(Buffer), Offset);
186}
187
188static void patchI64(raw_pwrite_stream &Stream, uint64_t Value,
190 uint8_t Buffer[8];
192 Stream.pwrite((char *)Buffer, sizeof(Buffer), Offset);
193}
194
195bool isDwoSection(const MCSection &Sec) {
196 return Sec.getName().ends_with(".dwo");
197}
198
199class WasmObjectWriter : public MCObjectWriter {
200 support::endian::Writer *W = nullptr;
201
202 /// The target specific Wasm writer instance.
203 std::unique_ptr<MCWasmObjectTargetWriter> TargetObjectWriter;
204
205 // Relocations for fixing up references in the code section.
206 std::vector<WasmRelocationEntry> CodeRelocations;
207 // Relocations for fixing up references in the data section.
208 std::vector<WasmRelocationEntry> DataRelocations;
209
210 // Index values to use for fixing up call_indirect type indices.
211 // Maps function symbols to the index of the type of the function
213 // Maps function symbols to the table element index space. Used
214 // for TABLE_INDEX relocation types (i.e. address taken functions).
216 // Maps function/global/table symbols to the
217 // function/global/table/tag/section index space.
220 // Maps data symbols to the Wasm segment and offset/size with the segment.
222
223 // Stores output data (index, relocations, content offset) for custom
224 // section.
225 std::vector<WasmCustomSection> CustomSections;
226 std::unique_ptr<WasmCustomSection> ProducersSection;
227 std::unique_ptr<WasmCustomSection> TargetFeaturesSection;
228 // Relocations for fixing up references in the custom sections.
230 CustomSectionsRelocations;
231
232 // Map from section to defining function symbol.
234
238 unsigned NumFunctionImports = 0;
239 unsigned NumGlobalImports = 0;
240 unsigned NumTableImports = 0;
241 unsigned NumTagImports = 0;
242 uint32_t SectionCount = 0;
243
244 enum class DwoMode {
245 AllSections,
246 NonDwoOnly,
247 DwoOnly,
248 };
249 bool IsSplitDwarf = false;
250 raw_pwrite_stream *OS = nullptr;
251 raw_pwrite_stream *DwoOS = nullptr;
252
253 // TargetObjectWriter wranppers.
254 bool is64Bit() const { return TargetObjectWriter->is64Bit(); }
255 bool isEmscripten() const { return TargetObjectWriter->isEmscripten(); }
256
257 void startSection(SectionBookkeeping &Section, unsigned SectionId);
258 void startCustomSection(SectionBookkeeping &Section, StringRef Name);
259 void endSection(SectionBookkeeping &Section);
260
261public:
262 WasmObjectWriter(std::unique_ptr<MCWasmObjectTargetWriter> MOTW,
264 : TargetObjectWriter(std::move(MOTW)), OS(&OS_) {}
265
266 WasmObjectWriter(std::unique_ptr<MCWasmObjectTargetWriter> MOTW,
268 : TargetObjectWriter(std::move(MOTW)), IsSplitDwarf(true), OS(&OS_),
269 DwoOS(&DwoOS_) {}
270
271private:
272 void reset() override {
273 CodeRelocations.clear();
274 DataRelocations.clear();
275 TypeIndices.clear();
276 WasmIndices.clear();
277 GOTIndices.clear();
278 TableIndices.clear();
279 DataLocations.clear();
280 CustomSections.clear();
281 ProducersSection.reset();
282 TargetFeaturesSection.reset();
283 CustomSectionsRelocations.clear();
284 SignatureIndices.clear();
285 Signatures.clear();
286 DataSegments.clear();
287 SectionFunctions.clear();
288 NumFunctionImports = 0;
289 NumGlobalImports = 0;
290 NumTableImports = 0;
292 }
293
294 void writeHeader(const MCAssembler &Asm);
295
296 void recordRelocation(MCAssembler &Asm, const MCAsmLayout &Layout,
297 const MCFragment *Fragment, const MCFixup &Fixup,
298 MCValue Target, uint64_t &FixedValue) override;
299
301 const MCAsmLayout &Layout) override;
302 void prepareImports(SmallVectorImpl<wasm::WasmImport> &Imports,
303 MCAssembler &Asm, const MCAsmLayout &Layout);
304 uint64_t writeObject(MCAssembler &Asm, const MCAsmLayout &Layout) override;
305
306 uint64_t writeOneObject(MCAssembler &Asm, const MCAsmLayout &Layout,
307 DwoMode Mode);
308
309 void writeString(const StringRef Str) {
310 encodeULEB128(Str.size(), W->OS);
311 W->OS << Str;
312 }
313
314 void writeStringWithAlignment(const StringRef Str, unsigned Alignment);
315
316 void writeI32(int32_t val) {
317 char Buffer[4];
318 support::endian::write32le(Buffer, val);
319 W->OS.write(Buffer, sizeof(Buffer));
320 }
321
322 void writeI64(int64_t val) {
323 char Buffer[8];
324 support::endian::write64le(Buffer, val);
325 W->OS.write(Buffer, sizeof(Buffer));
326 }
327
328 void writeValueType(wasm::ValType Ty) { W->OS << static_cast<char>(Ty); }
329
330 void writeTypeSection(ArrayRef<wasm::WasmSignature> Signatures);
331 void writeImportSection(ArrayRef<wasm::WasmImport> Imports, uint64_t DataSize,
332 uint32_t NumElements);
333 void writeFunctionSection(ArrayRef<WasmFunction> Functions);
334 void writeExportSection(ArrayRef<wasm::WasmExport> Exports);
335 void writeElemSection(const MCSymbolWasm *IndirectFunctionTable,
336 ArrayRef<uint32_t> TableElems);
337 void writeDataCountSection();
338 uint32_t writeCodeSection(const MCAssembler &Asm, const MCAsmLayout &Layout,
339 ArrayRef<WasmFunction> Functions);
340 uint32_t writeDataSection(const MCAsmLayout &Layout);
341 void writeTagSection(ArrayRef<uint32_t> TagTypes);
342 void writeGlobalSection(ArrayRef<wasm::WasmGlobal> Globals);
343 void writeTableSection(ArrayRef<wasm::WasmTable> Tables);
344 void writeRelocSection(uint32_t SectionIndex, StringRef Name,
345 std::vector<WasmRelocationEntry> &Relocations);
346 void writeLinkingMetaDataSection(
348 ArrayRef<std::pair<uint16_t, uint32_t>> InitFuncs,
349 const std::map<StringRef, std::vector<WasmComdatEntry>> &Comdats);
350 void writeCustomSection(WasmCustomSection &CustomSection,
351 const MCAssembler &Asm, const MCAsmLayout &Layout);
352 void writeCustomRelocSections();
353
354 uint64_t getProvisionalValue(const WasmRelocationEntry &RelEntry,
355 const MCAsmLayout &Layout);
356 void applyRelocations(ArrayRef<WasmRelocationEntry> Relocations,
357 uint64_t ContentsOffset, const MCAsmLayout &Layout);
358
359 uint32_t getRelocationIndexValue(const WasmRelocationEntry &RelEntry);
360 uint32_t getFunctionType(const MCSymbolWasm &Symbol);
361 uint32_t getTagType(const MCSymbolWasm &Symbol);
362 void registerFunctionType(const MCSymbolWasm &Symbol);
363 void registerTagType(const MCSymbolWasm &Symbol);
364};
365
366} // end anonymous namespace
367
368// Write out a section header and a patchable section size field.
369void WasmObjectWriter::startSection(SectionBookkeeping &Section,
370 unsigned SectionId) {
371 LLVM_DEBUG(dbgs() << "startSection " << SectionId << "\n");
372 W->OS << char(SectionId);
373
374 Section.SizeOffset = W->OS.tell();
375
376 // The section size. We don't know the size yet, so reserve enough space
377 // for any 32-bit value; we'll patch it later.
378 encodeULEB128(0, W->OS, 5);
379
380 // The position where the section starts, for measuring its size.
381 Section.ContentsOffset = W->OS.tell();
382 Section.PayloadOffset = W->OS.tell();
383 Section.Index = SectionCount++;
384}
385
386// Write a string with extra paddings for trailing alignment
387// TODO: support alignment at asm and llvm level?
388void WasmObjectWriter::writeStringWithAlignment(const StringRef Str,
389 unsigned Alignment) {
390
391 // Calculate the encoded size of str length and add pads based on it and
392 // alignment.
393 raw_null_ostream NullOS;
394 uint64_t StrSizeLength = encodeULEB128(Str.size(), NullOS);
395 uint64_t Offset = W->OS.tell() + StrSizeLength + Str.size();
396 uint64_t Paddings = offsetToAlignment(Offset, Align(Alignment));
397 Offset += Paddings;
398
399 // LEB128 greater than 5 bytes is invalid
400 assert((StrSizeLength + Paddings) <= 5 && "too long string to align");
401
402 encodeSLEB128(Str.size(), W->OS, StrSizeLength + Paddings);
403 W->OS << Str;
404
405 assert(W->OS.tell() == Offset && "invalid padding");
406}
407
408void WasmObjectWriter::startCustomSection(SectionBookkeeping &Section,
409 StringRef Name) {
410 LLVM_DEBUG(dbgs() << "startCustomSection " << Name << "\n");
411 startSection(Section, wasm::WASM_SEC_CUSTOM);
412
413 // The position where the section header ends, for measuring its size.
414 Section.PayloadOffset = W->OS.tell();
415
416 // Custom sections in wasm also have a string identifier.
417 if (Name != "__clangast") {
418 writeString(Name);
419 } else {
420 // The on-disk hashtable in clangast needs to be aligned by 4 bytes.
421 writeStringWithAlignment(Name, 4);
422 }
423
424 // The position where the custom section starts.
425 Section.ContentsOffset = W->OS.tell();
426}
427
428// Now that the section is complete and we know how big it is, patch up the
429// section size field at the start of the section.
430void WasmObjectWriter::endSection(SectionBookkeeping &Section) {
431 uint64_t Size = W->OS.tell();
432 // /dev/null doesn't support seek/tell and can report offset of 0.
433 // Simply skip this patching in that case.
434 if (!Size)
435 return;
436
437 Size -= Section.PayloadOffset;
438 if (uint32_t(Size) != Size)
439 report_fatal_error("section size does not fit in a uint32_t");
440
441 LLVM_DEBUG(dbgs() << "endSection size=" << Size << "\n");
442
443 // Write the final section size to the payload_len field, which follows
444 // the section id byte.
445 writePatchableU32(static_cast<raw_pwrite_stream &>(W->OS), Size,
446 Section.SizeOffset);
447}
448
449// Emit the Wasm header.
450void WasmObjectWriter::writeHeader(const MCAssembler &Asm) {
451 W->OS.write(wasm::WasmMagic, sizeof(wasm::WasmMagic));
453}
454
455void WasmObjectWriter::executePostLayoutBinding(MCAssembler &Asm,
456 const MCAsmLayout &Layout) {
457 // Some compilation units require the indirect function table to be present
458 // but don't explicitly reference it. This is the case for call_indirect
459 // without the reference-types feature, and also function bitcasts in all
460 // cases. In those cases the __indirect_function_table has the
461 // WASM_SYMBOL_NO_STRIP attribute. Here we make sure this symbol makes it to
462 // the assembler, if needed.
463 if (auto *Sym = Asm.getContext().lookupSymbol("__indirect_function_table")) {
464 const auto *WasmSym = static_cast<const MCSymbolWasm *>(Sym);
465 if (WasmSym->isNoStrip())
466 Asm.registerSymbol(*Sym);
467 }
468
469 // Build a map of sections to the function that defines them, for use
470 // in recordRelocation.
471 for (const MCSymbol &S : Asm.symbols()) {
472 const auto &WS = static_cast<const MCSymbolWasm &>(S);
473 if (WS.isDefined() && WS.isFunction() && !WS.isVariable()) {
474 const auto &Sec = static_cast<const MCSectionWasm &>(S.getSection());
475 auto Pair = SectionFunctions.insert(std::make_pair(&Sec, &S));
476 if (!Pair.second)
477 report_fatal_error("section already has a defining function: " +
478 Sec.getName());
479 }
480 }
481}
482
483void WasmObjectWriter::recordRelocation(MCAssembler &Asm,
484 const MCAsmLayout &Layout,
485 const MCFragment *Fragment,
486 const MCFixup &Fixup, MCValue Target,
487 uint64_t &FixedValue) {
488 // The WebAssembly backend should never generate FKF_IsPCRel fixups
489 assert(!(Asm.getBackend().getFixupKindInfo(Fixup.getKind()).Flags &
491
492 const auto &FixupSection = cast<MCSectionWasm>(*Fragment->getParent());
493 uint64_t C = Target.getConstant();
494 uint64_t FixupOffset = Layout.getFragmentOffset(Fragment) + Fixup.getOffset();
495 MCContext &Ctx = Asm.getContext();
496 bool IsLocRel = false;
497
498 if (const MCSymbolRefExpr *RefB = Target.getSymB()) {
499
500 const auto &SymB = cast<MCSymbolWasm>(RefB->getSymbol());
501
502 if (FixupSection.isText()) {
503 Ctx.reportError(Fixup.getLoc(),
504 Twine("symbol '") + SymB.getName() +
505 "' unsupported subtraction expression used in "
506 "relocation in code section.");
507 return;
508 }
509
510 if (SymB.isUndefined()) {
511 Ctx.reportError(Fixup.getLoc(),
512 Twine("symbol '") + SymB.getName() +
513 "' can not be undefined in a subtraction expression");
514 return;
515 }
516 const MCSection &SecB = SymB.getSection();
517 if (&SecB != &FixupSection) {
518 Ctx.reportError(Fixup.getLoc(),
519 Twine("symbol '") + SymB.getName() +
520 "' can not be placed in a different section");
521 return;
522 }
523 IsLocRel = true;
524 C += FixupOffset - Layout.getSymbolOffset(SymB);
525 }
526
527 // We either rejected the fixup or folded B into C at this point.
528 const MCSymbolRefExpr *RefA = Target.getSymA();
529 const auto *SymA = cast<MCSymbolWasm>(&RefA->getSymbol());
530
531 // The .init_array isn't translated as data, so don't do relocations in it.
532 if (FixupSection.getName().starts_with(".init_array")) {
533 SymA->setUsedInInitArray();
534 return;
535 }
536
537 if (SymA->isVariable()) {
538 const MCExpr *Expr = SymA->getVariableValue();
539 if (const auto *Inner = dyn_cast<MCSymbolRefExpr>(Expr))
540 if (Inner->getKind() == MCSymbolRefExpr::VK_WEAKREF)
541 llvm_unreachable("weakref used in reloc not yet implemented");
542 }
543
544 // Put any constant offset in an addend. Offsets can be negative, and
545 // LLVM expects wrapping, in contrast to wasm's immediates which can't
546 // be negative and don't wrap.
547 FixedValue = 0;
548
549 unsigned Type =
550 TargetObjectWriter->getRelocType(Target, Fixup, FixupSection, IsLocRel);
551
552 // Absolute offset within a section or a function.
553 // Currently only supported for metadata sections.
554 // See: test/MC/WebAssembly/blockaddress.ll
555 if ((Type == wasm::R_WASM_FUNCTION_OFFSET_I32 ||
556 Type == wasm::R_WASM_FUNCTION_OFFSET_I64 ||
557 Type == wasm::R_WASM_SECTION_OFFSET_I32) &&
558 SymA->isDefined()) {
559 // SymA can be a temp data symbol that represents a function (in which case
560 // it needs to be replaced by the section symbol), [XXX and it apparently
561 // later gets changed again to a func symbol?] or it can be a real
562 // function symbol, in which case it can be left as-is.
563
564 if (!FixupSection.isMetadata())
565 report_fatal_error("relocations for function or section offsets are "
566 "only supported in metadata sections");
567
568 const MCSymbol *SectionSymbol = nullptr;
569 const MCSection &SecA = SymA->getSection();
570 if (SecA.isText()) {
571 auto SecSymIt = SectionFunctions.find(&SecA);
572 if (SecSymIt == SectionFunctions.end())
573 report_fatal_error("section doesn\'t have defining symbol");
574 SectionSymbol = SecSymIt->second;
575 } else {
576 SectionSymbol = SecA.getBeginSymbol();
577 }
578 if (!SectionSymbol)
579 report_fatal_error("section symbol is required for relocation");
580
581 C += Layout.getSymbolOffset(*SymA);
582 SymA = cast<MCSymbolWasm>(SectionSymbol);
583 }
584
585 if (Type == wasm::R_WASM_TABLE_INDEX_REL_SLEB ||
586 Type == wasm::R_WASM_TABLE_INDEX_REL_SLEB64 ||
587 Type == wasm::R_WASM_TABLE_INDEX_SLEB ||
588 Type == wasm::R_WASM_TABLE_INDEX_SLEB64 ||
589 Type == wasm::R_WASM_TABLE_INDEX_I32 ||
590 Type == wasm::R_WASM_TABLE_INDEX_I64) {
591 // TABLE_INDEX relocs implicitly use the default indirect function table.
592 // We require the function table to have already been defined.
593 auto TableName = "__indirect_function_table";
594 MCSymbolWasm *Sym = cast_or_null<MCSymbolWasm>(Ctx.lookupSymbol(TableName));
595 if (!Sym) {
596 report_fatal_error("missing indirect function table symbol");
597 } else {
598 if (!Sym->isFunctionTable())
599 report_fatal_error("__indirect_function_table symbol has wrong type");
600 // Ensure that __indirect_function_table reaches the output.
601 Sym->setNoStrip();
602 Asm.registerSymbol(*Sym);
603 }
604 }
605
606 // Relocation other than R_WASM_TYPE_INDEX_LEB are required to be
607 // against a named symbol.
608 if (Type != wasm::R_WASM_TYPE_INDEX_LEB) {
609 if (SymA->getName().empty())
610 report_fatal_error("relocations against un-named temporaries are not yet "
611 "supported by wasm");
612
613 SymA->setUsedInReloc();
614 }
615
616 switch (RefA->getKind()) {
619 SymA->setUsedInGOT();
620 break;
621 default:
622 break;
623 }
624
625 WasmRelocationEntry Rec(FixupOffset, SymA, C, Type, &FixupSection);
626 LLVM_DEBUG(dbgs() << "WasmReloc: " << Rec << "\n");
627
628 if (FixupSection.isWasmData()) {
629 DataRelocations.push_back(Rec);
630 } else if (FixupSection.isText()) {
631 CodeRelocations.push_back(Rec);
632 } else if (FixupSection.isMetadata()) {
633 CustomSectionsRelocations[&FixupSection].push_back(Rec);
634 } else {
635 llvm_unreachable("unexpected section type");
636 }
637}
638
639// Compute a value to write into the code at the location covered
640// by RelEntry. This value isn't used by the static linker; it just serves
641// to make the object format more readable and more likely to be directly
642// useable.
644WasmObjectWriter::getProvisionalValue(const WasmRelocationEntry &RelEntry,
645 const MCAsmLayout &Layout) {
646 if ((RelEntry.Type == wasm::R_WASM_GLOBAL_INDEX_LEB ||
647 RelEntry.Type == wasm::R_WASM_GLOBAL_INDEX_I32) &&
648 !RelEntry.Symbol->isGlobal()) {
649 assert(GOTIndices.count(RelEntry.Symbol) > 0 && "symbol not found in GOT index space");
650 return GOTIndices[RelEntry.Symbol];
651 }
652
653 switch (RelEntry.Type) {
654 case wasm::R_WASM_TABLE_INDEX_REL_SLEB:
655 case wasm::R_WASM_TABLE_INDEX_REL_SLEB64:
656 case wasm::R_WASM_TABLE_INDEX_SLEB:
657 case wasm::R_WASM_TABLE_INDEX_SLEB64:
658 case wasm::R_WASM_TABLE_INDEX_I32:
659 case wasm::R_WASM_TABLE_INDEX_I64: {
660 // Provisional value is table address of the resolved symbol itself
661 const MCSymbolWasm *Base =
662 cast<MCSymbolWasm>(Layout.getBaseSymbol(*RelEntry.Symbol));
663 assert(Base->isFunction());
664 if (RelEntry.Type == wasm::R_WASM_TABLE_INDEX_REL_SLEB ||
665 RelEntry.Type == wasm::R_WASM_TABLE_INDEX_REL_SLEB64)
666 return TableIndices[Base] - InitialTableOffset;
667 else
668 return TableIndices[Base];
669 }
670 case wasm::R_WASM_TYPE_INDEX_LEB:
671 // Provisional value is same as the index
672 return getRelocationIndexValue(RelEntry);
673 case wasm::R_WASM_FUNCTION_INDEX_LEB:
674 case wasm::R_WASM_FUNCTION_INDEX_I32:
675 case wasm::R_WASM_GLOBAL_INDEX_LEB:
676 case wasm::R_WASM_GLOBAL_INDEX_I32:
677 case wasm::R_WASM_TAG_INDEX_LEB:
678 case wasm::R_WASM_TABLE_NUMBER_LEB:
679 // Provisional value is function/global/tag Wasm index
680 assert(WasmIndices.count(RelEntry.Symbol) > 0 && "symbol not found in wasm index space");
681 return WasmIndices[RelEntry.Symbol];
682 case wasm::R_WASM_FUNCTION_OFFSET_I32:
683 case wasm::R_WASM_FUNCTION_OFFSET_I64:
684 case wasm::R_WASM_SECTION_OFFSET_I32: {
685 if (!RelEntry.Symbol->isDefined())
686 return 0;
687 const auto &Section =
688 static_cast<const MCSectionWasm &>(RelEntry.Symbol->getSection());
689 return Section.getSectionOffset() + RelEntry.Addend;
690 }
691 case wasm::R_WASM_MEMORY_ADDR_LEB:
692 case wasm::R_WASM_MEMORY_ADDR_LEB64:
693 case wasm::R_WASM_MEMORY_ADDR_SLEB:
694 case wasm::R_WASM_MEMORY_ADDR_SLEB64:
695 case wasm::R_WASM_MEMORY_ADDR_REL_SLEB:
696 case wasm::R_WASM_MEMORY_ADDR_REL_SLEB64:
697 case wasm::R_WASM_MEMORY_ADDR_I32:
698 case wasm::R_WASM_MEMORY_ADDR_I64:
699 case wasm::R_WASM_MEMORY_ADDR_TLS_SLEB:
700 case wasm::R_WASM_MEMORY_ADDR_TLS_SLEB64:
701 case wasm::R_WASM_MEMORY_ADDR_LOCREL_I32: {
702 // Provisional value is address of the global plus the offset
703 // For undefined symbols, use zero
704 if (!RelEntry.Symbol->isDefined())
705 return 0;
706 const wasm::WasmDataReference &SymRef = DataLocations[RelEntry.Symbol];
707 const WasmDataSegment &Segment = DataSegments[SymRef.Segment];
708 // Ignore overflow. LLVM allows address arithmetic to silently wrap.
709 return Segment.Offset + SymRef.Offset + RelEntry.Addend;
710 }
711 default:
712 llvm_unreachable("invalid relocation type");
713 }
714}
715
716static void addData(SmallVectorImpl<char> &DataBytes,
717 MCSectionWasm &DataSection) {
718 LLVM_DEBUG(errs() << "addData: " << DataSection.getName() << "\n");
719
720 DataBytes.resize(alignTo(DataBytes.size(), DataSection.getAlign()));
721
722 for (const MCFragment &Frag : DataSection) {
723 if (Frag.hasInstructions())
724 report_fatal_error("only data supported in data sections");
725
726 if (auto *Align = dyn_cast<MCAlignFragment>(&Frag)) {
727 if (Align->getValueSize() != 1)
728 report_fatal_error("only byte values supported for alignment");
729 // If nops are requested, use zeros, as this is the data section.
730 uint8_t Value = Align->hasEmitNops() ? 0 : Align->getValue();
731 uint64_t Size =
732 std::min<uint64_t>(alignTo(DataBytes.size(), Align->getAlignment()),
733 DataBytes.size() + Align->getMaxBytesToEmit());
734 DataBytes.resize(Size, Value);
735 } else if (auto *Fill = dyn_cast<MCFillFragment>(&Frag)) {
736 int64_t NumValues;
737 if (!Fill->getNumValues().evaluateAsAbsolute(NumValues))
738 llvm_unreachable("The fill should be an assembler constant");
739 DataBytes.insert(DataBytes.end(), Fill->getValueSize() * NumValues,
740 Fill->getValue());
741 } else if (auto *LEB = dyn_cast<MCLEBFragment>(&Frag)) {
742 const SmallVectorImpl<char> &Contents = LEB->getContents();
743 llvm::append_range(DataBytes, Contents);
744 } else {
745 const auto &DataFrag = cast<MCDataFragment>(Frag);
746 const SmallVectorImpl<char> &Contents = DataFrag.getContents();
747 llvm::append_range(DataBytes, Contents);
748 }
749 }
750
751 LLVM_DEBUG(dbgs() << "addData -> " << DataBytes.size() << "\n");
752}
753
755WasmObjectWriter::getRelocationIndexValue(const WasmRelocationEntry &RelEntry) {
756 if (RelEntry.Type == wasm::R_WASM_TYPE_INDEX_LEB) {
757 if (!TypeIndices.count(RelEntry.Symbol))
758 report_fatal_error("symbol not found in type index space: " +
759 RelEntry.Symbol->getName());
760 return TypeIndices[RelEntry.Symbol];
761 }
762
763 return RelEntry.Symbol->getIndex();
764}
765
766// Apply the portions of the relocation records that we can handle ourselves
767// directly.
768void WasmObjectWriter::applyRelocations(
769 ArrayRef<WasmRelocationEntry> Relocations, uint64_t ContentsOffset,
770 const MCAsmLayout &Layout) {
771 auto &Stream = static_cast<raw_pwrite_stream &>(W->OS);
772 for (const WasmRelocationEntry &RelEntry : Relocations) {
773 uint64_t Offset = ContentsOffset +
774 RelEntry.FixupSection->getSectionOffset() +
775 RelEntry.Offset;
776
777 LLVM_DEBUG(dbgs() << "applyRelocation: " << RelEntry << "\n");
778 uint64_t Value = getProvisionalValue(RelEntry, Layout);
779
780 switch (RelEntry.Type) {
781 case wasm::R_WASM_FUNCTION_INDEX_LEB:
782 case wasm::R_WASM_TYPE_INDEX_LEB:
783 case wasm::R_WASM_GLOBAL_INDEX_LEB:
784 case wasm::R_WASM_MEMORY_ADDR_LEB:
785 case wasm::R_WASM_TAG_INDEX_LEB:
786 case wasm::R_WASM_TABLE_NUMBER_LEB:
787 writePatchableU32(Stream, Value, Offset);
788 break;
789 case wasm::R_WASM_MEMORY_ADDR_LEB64:
790 writePatchableU64(Stream, Value, Offset);
791 break;
792 case wasm::R_WASM_TABLE_INDEX_I32:
793 case wasm::R_WASM_MEMORY_ADDR_I32:
794 case wasm::R_WASM_FUNCTION_OFFSET_I32:
795 case wasm::R_WASM_FUNCTION_INDEX_I32:
796 case wasm::R_WASM_SECTION_OFFSET_I32:
797 case wasm::R_WASM_GLOBAL_INDEX_I32:
798 case wasm::R_WASM_MEMORY_ADDR_LOCREL_I32:
799 patchI32(Stream, Value, Offset);
800 break;
801 case wasm::R_WASM_TABLE_INDEX_I64:
802 case wasm::R_WASM_MEMORY_ADDR_I64:
803 case wasm::R_WASM_FUNCTION_OFFSET_I64:
804 patchI64(Stream, Value, Offset);
805 break;
806 case wasm::R_WASM_TABLE_INDEX_SLEB:
807 case wasm::R_WASM_TABLE_INDEX_REL_SLEB:
808 case wasm::R_WASM_MEMORY_ADDR_SLEB:
809 case wasm::R_WASM_MEMORY_ADDR_REL_SLEB:
810 case wasm::R_WASM_MEMORY_ADDR_TLS_SLEB:
811 writePatchableS32(Stream, Value, Offset);
812 break;
813 case wasm::R_WASM_TABLE_INDEX_SLEB64:
814 case wasm::R_WASM_TABLE_INDEX_REL_SLEB64:
815 case wasm::R_WASM_MEMORY_ADDR_SLEB64:
816 case wasm::R_WASM_MEMORY_ADDR_REL_SLEB64:
817 case wasm::R_WASM_MEMORY_ADDR_TLS_SLEB64:
818 writePatchableS64(Stream, Value, Offset);
819 break;
820 default:
821 llvm_unreachable("invalid relocation type");
822 }
823 }
824}
825
826void WasmObjectWriter::writeTypeSection(
828 if (Signatures.empty())
829 return;
830
831 SectionBookkeeping Section;
832 startSection(Section, wasm::WASM_SEC_TYPE);
833
834 encodeULEB128(Signatures.size(), W->OS);
835
836 for (const wasm::WasmSignature &Sig : Signatures) {
838 encodeULEB128(Sig.Params.size(), W->OS);
839 for (wasm::ValType Ty : Sig.Params)
840 writeValueType(Ty);
841 encodeULEB128(Sig.Returns.size(), W->OS);
842 for (wasm::ValType Ty : Sig.Returns)
843 writeValueType(Ty);
844 }
845
846 endSection(Section);
847}
848
849void WasmObjectWriter::writeImportSection(ArrayRef<wasm::WasmImport> Imports,
850 uint64_t DataSize,
851 uint32_t NumElements) {
852 if (Imports.empty())
853 return;
854
855 uint64_t NumPages = (DataSize + wasm::WasmPageSize - 1) / wasm::WasmPageSize;
856
857 SectionBookkeeping Section;
858 startSection(Section, wasm::WASM_SEC_IMPORT);
859
860 encodeULEB128(Imports.size(), W->OS);
861 for (const wasm::WasmImport &Import : Imports) {
862 writeString(Import.Module);
863 writeString(Import.Field);
864 W->OS << char(Import.Kind);
865
866 switch (Import.Kind) {
868 encodeULEB128(Import.SigIndex, W->OS);
869 break;
871 W->OS << char(Import.Global.Type);
872 W->OS << char(Import.Global.Mutable ? 1 : 0);
873 break;
875 encodeULEB128(Import.Memory.Flags, W->OS);
876 encodeULEB128(NumPages, W->OS); // initial
877 break;
879 W->OS << char(Import.Table.ElemType);
880 encodeULEB128(Import.Table.Limits.Flags, W->OS);
881 encodeULEB128(NumElements, W->OS); // initial
882 break;
884 W->OS << char(0); // Reserved 'attribute' field
885 encodeULEB128(Import.SigIndex, W->OS);
886 break;
887 default:
888 llvm_unreachable("unsupported import kind");
889 }
890 }
891
892 endSection(Section);
893}
894
895void WasmObjectWriter::writeFunctionSection(ArrayRef<WasmFunction> Functions) {
896 if (Functions.empty())
897 return;
898
899 SectionBookkeeping Section;
900 startSection(Section, wasm::WASM_SEC_FUNCTION);
901
902 encodeULEB128(Functions.size(), W->OS);
903 for (const WasmFunction &Func : Functions)
904 encodeULEB128(Func.SigIndex, W->OS);
905
906 endSection(Section);
907}
908
909void WasmObjectWriter::writeTagSection(ArrayRef<uint32_t> TagTypes) {
910 if (TagTypes.empty())
911 return;
912
913 SectionBookkeeping Section;
914 startSection(Section, wasm::WASM_SEC_TAG);
915
916 encodeULEB128(TagTypes.size(), W->OS);
917 for (uint32_t Index : TagTypes) {
918 W->OS << char(0); // Reserved 'attribute' field
919 encodeULEB128(Index, W->OS);
920 }
921
922 endSection(Section);
923}
924
925void WasmObjectWriter::writeGlobalSection(ArrayRef<wasm::WasmGlobal> Globals) {
926 if (Globals.empty())
927 return;
928
929 SectionBookkeeping Section;
930 startSection(Section, wasm::WASM_SEC_GLOBAL);
931
932 encodeULEB128(Globals.size(), W->OS);
933 for (const wasm::WasmGlobal &Global : Globals) {
934 encodeULEB128(Global.Type.Type, W->OS);
935 W->OS << char(Global.Type.Mutable);
936 if (Global.InitExpr.Extended) {
937 llvm_unreachable("extected init expressions not supported");
938 } else {
939 W->OS << char(Global.InitExpr.Inst.Opcode);
940 switch (Global.Type.Type) {
942 encodeSLEB128(0, W->OS);
943 break;
945 encodeSLEB128(0, W->OS);
946 break;
948 writeI32(0);
949 break;
951 writeI64(0);
952 break;
954 writeValueType(wasm::ValType::EXTERNREF);
955 break;
956 default:
957 llvm_unreachable("unexpected type");
958 }
959 }
961 }
962
963 endSection(Section);
964}
965
966void WasmObjectWriter::writeTableSection(ArrayRef<wasm::WasmTable> Tables) {
967 if (Tables.empty())
968 return;
969
970 SectionBookkeeping Section;
971 startSection(Section, wasm::WASM_SEC_TABLE);
972
973 encodeULEB128(Tables.size(), W->OS);
974 for (const wasm::WasmTable &Table : Tables) {
975 assert(Table.Type.ElemType != wasm::ValType::OTHERREF &&
976 "Cannot encode general ref-typed tables");
977 encodeULEB128((uint32_t)Table.Type.ElemType, W->OS);
978 encodeULEB128(Table.Type.Limits.Flags, W->OS);
979 encodeULEB128(Table.Type.Limits.Minimum, W->OS);
980 if (Table.Type.Limits.Flags & wasm::WASM_LIMITS_FLAG_HAS_MAX)
981 encodeULEB128(Table.Type.Limits.Maximum, W->OS);
982 }
983 endSection(Section);
984}
985
986void WasmObjectWriter::writeExportSection(ArrayRef<wasm::WasmExport> Exports) {
987 if (Exports.empty())
988 return;
989
990 SectionBookkeeping Section;
991 startSection(Section, wasm::WASM_SEC_EXPORT);
992
993 encodeULEB128(Exports.size(), W->OS);
994 for (const wasm::WasmExport &Export : Exports) {
995 writeString(Export.Name);
996 W->OS << char(Export.Kind);
997 encodeULEB128(Export.Index, W->OS);
998 }
999
1000 endSection(Section);
1001}
1002
1003void WasmObjectWriter::writeElemSection(
1004 const MCSymbolWasm *IndirectFunctionTable, ArrayRef<uint32_t> TableElems) {
1005 if (TableElems.empty())
1006 return;
1007
1008 assert(IndirectFunctionTable);
1009
1010 SectionBookkeeping Section;
1011 startSection(Section, wasm::WASM_SEC_ELEM);
1012
1013 encodeULEB128(1, W->OS); // number of "segments"
1014
1015 assert(WasmIndices.count(IndirectFunctionTable));
1016 uint32_t TableNumber = WasmIndices.find(IndirectFunctionTable)->second;
1017 uint32_t Flags = 0;
1018 if (TableNumber)
1020 encodeULEB128(Flags, W->OS);
1022 encodeULEB128(TableNumber, W->OS); // the table number
1023
1024 // init expr for starting offset
1027 encodeSLEB128(InitialTableOffset, W->OS);
1029
1031 // We only write active function table initializers, for which the elem kind
1032 // is specified to be written as 0x00 and interpreted to mean "funcref".
1033 const uint8_t ElemKind = 0;
1034 W->OS << ElemKind;
1035 }
1036
1037 encodeULEB128(TableElems.size(), W->OS);
1038 for (uint32_t Elem : TableElems)
1039 encodeULEB128(Elem, W->OS);
1040
1041 endSection(Section);
1042}
1043
1044void WasmObjectWriter::writeDataCountSection() {
1045 if (DataSegments.empty())
1046 return;
1047
1048 SectionBookkeeping Section;
1049 startSection(Section, wasm::WASM_SEC_DATACOUNT);
1050 encodeULEB128(DataSegments.size(), W->OS);
1051 endSection(Section);
1052}
1053
1054uint32_t WasmObjectWriter::writeCodeSection(const MCAssembler &Asm,
1055 const MCAsmLayout &Layout,
1056 ArrayRef<WasmFunction> Functions) {
1057 if (Functions.empty())
1058 return 0;
1059
1060 SectionBookkeeping Section;
1061 startSection(Section, wasm::WASM_SEC_CODE);
1062
1063 encodeULEB128(Functions.size(), W->OS);
1064
1065 for (const WasmFunction &Func : Functions) {
1066 auto *FuncSection = static_cast<MCSectionWasm *>(Func.Section);
1067
1068 int64_t Size = Layout.getSectionAddressSize(FuncSection);
1069 encodeULEB128(Size, W->OS);
1070 FuncSection->setSectionOffset(W->OS.tell() - Section.ContentsOffset);
1071 Asm.writeSectionData(W->OS, FuncSection, Layout);
1072 }
1073
1074 // Apply fixups.
1075 applyRelocations(CodeRelocations, Section.ContentsOffset, Layout);
1076
1077 endSection(Section);
1078 return Section.Index;
1079}
1080
1081uint32_t WasmObjectWriter::writeDataSection(const MCAsmLayout &Layout) {
1082 if (DataSegments.empty())
1083 return 0;
1084
1085 SectionBookkeeping Section;
1086 startSection(Section, wasm::WASM_SEC_DATA);
1087
1088 encodeULEB128(DataSegments.size(), W->OS); // count
1089
1090 for (const WasmDataSegment &Segment : DataSegments) {
1091 encodeULEB128(Segment.InitFlags, W->OS); // flags
1092 if (Segment.InitFlags & wasm::WASM_DATA_SEGMENT_HAS_MEMINDEX)
1093 encodeULEB128(0, W->OS); // memory index
1094 if ((Segment.InitFlags & wasm::WASM_DATA_SEGMENT_IS_PASSIVE) == 0) {
1097 encodeSLEB128(Segment.Offset, W->OS); // offset
1099 }
1100 encodeULEB128(Segment.Data.size(), W->OS); // size
1101 Segment.Section->setSectionOffset(W->OS.tell() - Section.ContentsOffset);
1102 W->OS << Segment.Data; // data
1103 }
1104
1105 // Apply fixups.
1106 applyRelocations(DataRelocations, Section.ContentsOffset, Layout);
1107
1108 endSection(Section);
1109 return Section.Index;
1110}
1111
1112void WasmObjectWriter::writeRelocSection(
1113 uint32_t SectionIndex, StringRef Name,
1114 std::vector<WasmRelocationEntry> &Relocs) {
1115 // See: https://github.com/WebAssembly/tool-conventions/blob/main/Linking.md
1116 // for descriptions of the reloc sections.
1117
1118 if (Relocs.empty())
1119 return;
1120
1121 // First, ensure the relocations are sorted in offset order. In general they
1122 // should already be sorted since `recordRelocation` is called in offset
1123 // order, but for the code section we combine many MC sections into single
1124 // wasm section, and this order is determined by the order of Asm.Symbols()
1125 // not the sections order.
1127 Relocs, [](const WasmRelocationEntry &A, const WasmRelocationEntry &B) {
1128 return (A.Offset + A.FixupSection->getSectionOffset()) <
1129 (B.Offset + B.FixupSection->getSectionOffset());
1130 });
1131
1132 SectionBookkeeping Section;
1133 startCustomSection(Section, std::string("reloc.") + Name.str());
1134
1135 encodeULEB128(SectionIndex, W->OS);
1136 encodeULEB128(Relocs.size(), W->OS);
1137 for (const WasmRelocationEntry &RelEntry : Relocs) {
1139 RelEntry.Offset + RelEntry.FixupSection->getSectionOffset();
1140 uint32_t Index = getRelocationIndexValue(RelEntry);
1141
1142 W->OS << char(RelEntry.Type);
1143 encodeULEB128(Offset, W->OS);
1144 encodeULEB128(Index, W->OS);
1145 if (RelEntry.hasAddend())
1146 encodeSLEB128(RelEntry.Addend, W->OS);
1147 }
1148
1149 endSection(Section);
1150}
1151
1152void WasmObjectWriter::writeCustomRelocSections() {
1153 for (const auto &Sec : CustomSections) {
1154 auto &Relocations = CustomSectionsRelocations[Sec.Section];
1155 writeRelocSection(Sec.OutputIndex, Sec.Name, Relocations);
1156 }
1157}
1158
1159void WasmObjectWriter::writeLinkingMetaDataSection(
1161 ArrayRef<std::pair<uint16_t, uint32_t>> InitFuncs,
1162 const std::map<StringRef, std::vector<WasmComdatEntry>> &Comdats) {
1163 SectionBookkeeping Section;
1164 startCustomSection(Section, "linking");
1166
1167 SectionBookkeeping SubSection;
1168 if (SymbolInfos.size() != 0) {
1169 startSection(SubSection, wasm::WASM_SYMBOL_TABLE);
1170 encodeULEB128(SymbolInfos.size(), W->OS);
1171 for (const wasm::WasmSymbolInfo &Sym : SymbolInfos) {
1172 encodeULEB128(Sym.Kind, W->OS);
1173 encodeULEB128(Sym.Flags, W->OS);
1174 switch (Sym.Kind) {
1179 encodeULEB128(Sym.ElementIndex, W->OS);
1180 if ((Sym.Flags & wasm::WASM_SYMBOL_UNDEFINED) == 0 ||
1181 (Sym.Flags & wasm::WASM_SYMBOL_EXPLICIT_NAME) != 0)
1182 writeString(Sym.Name);
1183 break;
1185 writeString(Sym.Name);
1186 if ((Sym.Flags & wasm::WASM_SYMBOL_UNDEFINED) == 0) {
1187 encodeULEB128(Sym.DataRef.Segment, W->OS);
1188 encodeULEB128(Sym.DataRef.Offset, W->OS);
1189 encodeULEB128(Sym.DataRef.Size, W->OS);
1190 }
1191 break;
1193 const uint32_t SectionIndex =
1194 CustomSections[Sym.ElementIndex].OutputIndex;
1195 encodeULEB128(SectionIndex, W->OS);
1196 break;
1197 }
1198 default:
1199 llvm_unreachable("unexpected kind");
1200 }
1201 }
1202 endSection(SubSection);
1203 }
1204
1205 if (DataSegments.size()) {
1206 startSection(SubSection, wasm::WASM_SEGMENT_INFO);
1207 encodeULEB128(DataSegments.size(), W->OS);
1208 for (const WasmDataSegment &Segment : DataSegments) {
1209 writeString(Segment.Name);
1210 encodeULEB128(Segment.Alignment, W->OS);
1211 encodeULEB128(Segment.LinkingFlags, W->OS);
1212 }
1213 endSection(SubSection);
1214 }
1215
1216 if (!InitFuncs.empty()) {
1217 startSection(SubSection, wasm::WASM_INIT_FUNCS);
1218 encodeULEB128(InitFuncs.size(), W->OS);
1219 for (auto &StartFunc : InitFuncs) {
1220 encodeULEB128(StartFunc.first, W->OS); // priority
1221 encodeULEB128(StartFunc.second, W->OS); // function index
1222 }
1223 endSection(SubSection);
1224 }
1225
1226 if (Comdats.size()) {
1227 startSection(SubSection, wasm::WASM_COMDAT_INFO);
1228 encodeULEB128(Comdats.size(), W->OS);
1229 for (const auto &C : Comdats) {
1230 writeString(C.first);
1231 encodeULEB128(0, W->OS); // flags for future use
1232 encodeULEB128(C.second.size(), W->OS);
1233 for (const WasmComdatEntry &Entry : C.second) {
1234 encodeULEB128(Entry.Kind, W->OS);
1235 encodeULEB128(Entry.Index, W->OS);
1236 }
1237 }
1238 endSection(SubSection);
1239 }
1240
1241 endSection(Section);
1242}
1243
1244void WasmObjectWriter::writeCustomSection(WasmCustomSection &CustomSection,
1245 const MCAssembler &Asm,
1246 const MCAsmLayout &Layout) {
1247 SectionBookkeeping Section;
1248 auto *Sec = CustomSection.Section;
1249 startCustomSection(Section, CustomSection.Name);
1250
1251 Sec->setSectionOffset(W->OS.tell() - Section.ContentsOffset);
1252 Asm.writeSectionData(W->OS, Sec, Layout);
1253
1254 CustomSection.OutputContentsOffset = Section.ContentsOffset;
1255 CustomSection.OutputIndex = Section.Index;
1256
1257 endSection(Section);
1258
1259 // Apply fixups.
1260 auto &Relocations = CustomSectionsRelocations[CustomSection.Section];
1261 applyRelocations(Relocations, CustomSection.OutputContentsOffset, Layout);
1262}
1263
1264uint32_t WasmObjectWriter::getFunctionType(const MCSymbolWasm &Symbol) {
1265 assert(Symbol.isFunction());
1266 assert(TypeIndices.count(&Symbol));
1267 return TypeIndices[&Symbol];
1268}
1269
1270uint32_t WasmObjectWriter::getTagType(const MCSymbolWasm &Symbol) {
1271 assert(Symbol.isTag());
1272 assert(TypeIndices.count(&Symbol));
1273 return TypeIndices[&Symbol];
1274}
1275
1276void WasmObjectWriter::registerFunctionType(const MCSymbolWasm &Symbol) {
1277 assert(Symbol.isFunction());
1278
1280
1281 if (auto *Sig = Symbol.getSignature()) {
1282 S.Returns = Sig->Returns;
1283 S.Params = Sig->Params;
1284 }
1285
1286 auto Pair = SignatureIndices.insert(std::make_pair(S, Signatures.size()));
1287 if (Pair.second)
1288 Signatures.push_back(S);
1289 TypeIndices[&Symbol] = Pair.first->second;
1290
1291 LLVM_DEBUG(dbgs() << "registerFunctionType: " << Symbol
1292 << " new:" << Pair.second << "\n");
1293 LLVM_DEBUG(dbgs() << " -> type index: " << Pair.first->second << "\n");
1294}
1295
1296void WasmObjectWriter::registerTagType(const MCSymbolWasm &Symbol) {
1297 assert(Symbol.isTag());
1298
1299 // TODO Currently we don't generate imported exceptions, but if we do, we
1300 // should have a way of infering types of imported exceptions.
1302 if (auto *Sig = Symbol.getSignature()) {
1303 S.Returns = Sig->Returns;
1304 S.Params = Sig->Params;
1305 }
1306
1307 auto Pair = SignatureIndices.insert(std::make_pair(S, Signatures.size()));
1308 if (Pair.second)
1309 Signatures.push_back(S);
1310 TypeIndices[&Symbol] = Pair.first->second;
1311
1312 LLVM_DEBUG(dbgs() << "registerTagType: " << Symbol << " new:" << Pair.second
1313 << "\n");
1314 LLVM_DEBUG(dbgs() << " -> type index: " << Pair.first->second << "\n");
1315}
1316
1317static bool isInSymtab(const MCSymbolWasm &Sym) {
1318 if (Sym.isUsedInReloc() || Sym.isUsedInInitArray())
1319 return true;
1320
1321 if (Sym.isComdat() && !Sym.isDefined())
1322 return false;
1323
1324 if (Sym.isTemporary())
1325 return false;
1326
1327 if (Sym.isSection())
1328 return false;
1329
1330 if (Sym.omitFromLinkingSection())
1331 return false;
1332
1333 return true;
1334}
1335
1336void WasmObjectWriter::prepareImports(
1338 const MCAsmLayout &Layout) {
1339 // For now, always emit the memory import, since loads and stores are not
1340 // valid without it. In the future, we could perhaps be more clever and omit
1341 // it if there are no loads or stores.
1342 wasm::WasmImport MemImport;
1343 MemImport.Module = "env";
1344 MemImport.Field = "__linear_memory";
1345 MemImport.Kind = wasm::WASM_EXTERNAL_MEMORY;
1348 Imports.push_back(MemImport);
1349
1350 // Populate SignatureIndices, and Imports and WasmIndices for undefined
1351 // symbols. This must be done before populating WasmIndices for defined
1352 // symbols.
1353 for (const MCSymbol &S : Asm.symbols()) {
1354 const auto &WS = static_cast<const MCSymbolWasm &>(S);
1355
1356 // Register types for all functions, including those with private linkage
1357 // (because wasm always needs a type signature).
1358 if (WS.isFunction()) {
1359 const auto *BS = Layout.getBaseSymbol(S);
1360 if (!BS)
1361 report_fatal_error(Twine(S.getName()) +
1362 ": absolute addressing not supported!");
1363 registerFunctionType(*cast<MCSymbolWasm>(BS));
1364 }
1365
1366 if (WS.isTag())
1367 registerTagType(WS);
1368
1369 if (WS.isTemporary())
1370 continue;
1371
1372 // If the symbol is not defined in this translation unit, import it.
1373 if (!WS.isDefined() && !WS.isComdat()) {
1374 if (WS.isFunction()) {
1376 Import.Module = WS.getImportModule();
1377 Import.Field = WS.getImportName();
1379 Import.SigIndex = getFunctionType(WS);
1380 Imports.push_back(Import);
1381 assert(WasmIndices.count(&WS) == 0);
1382 WasmIndices[&WS] = NumFunctionImports++;
1383 } else if (WS.isGlobal()) {
1384 if (WS.isWeak())
1385 report_fatal_error("undefined global symbol cannot be weak");
1386
1388 Import.Field = WS.getImportName();
1390 Import.Module = WS.getImportModule();
1391 Import.Global = WS.getGlobalType();
1392 Imports.push_back(Import);
1393 assert(WasmIndices.count(&WS) == 0);
1394 WasmIndices[&WS] = NumGlobalImports++;
1395 } else if (WS.isTag()) {
1396 if (WS.isWeak())
1397 report_fatal_error("undefined tag symbol cannot be weak");
1398
1400 Import.Module = WS.getImportModule();
1401 Import.Field = WS.getImportName();
1403 Import.SigIndex = getTagType(WS);
1404 Imports.push_back(Import);
1405 assert(WasmIndices.count(&WS) == 0);
1406 WasmIndices[&WS] = NumTagImports++;
1407 } else if (WS.isTable()) {
1408 if (WS.isWeak())
1409 report_fatal_error("undefined table symbol cannot be weak");
1410
1412 Import.Module = WS.getImportModule();
1413 Import.Field = WS.getImportName();
1415 Import.Table = WS.getTableType();
1416 Imports.push_back(Import);
1417 assert(WasmIndices.count(&WS) == 0);
1418 WasmIndices[&WS] = NumTableImports++;
1419 }
1420 }
1421 }
1422
1423 // Add imports for GOT globals
1424 for (const MCSymbol &S : Asm.symbols()) {
1425 const auto &WS = static_cast<const MCSymbolWasm &>(S);
1426 if (WS.isUsedInGOT()) {
1428 if (WS.isFunction())
1429 Import.Module = "GOT.func";
1430 else
1431 Import.Module = "GOT.mem";
1432 Import.Field = WS.getName();
1434 Import.Global = {wasm::WASM_TYPE_I32, true};
1435 Imports.push_back(Import);
1436 assert(GOTIndices.count(&WS) == 0);
1437 GOTIndices[&WS] = NumGlobalImports++;
1438 }
1439 }
1440}
1441
1442uint64_t WasmObjectWriter::writeObject(MCAssembler &Asm,
1443 const MCAsmLayout &Layout) {
1445 W = &MainWriter;
1446 if (IsSplitDwarf) {
1447 uint64_t TotalSize = writeOneObject(Asm, Layout, DwoMode::NonDwoOnly);
1448 assert(DwoOS);
1450 W = &DwoWriter;
1451 return TotalSize + writeOneObject(Asm, Layout, DwoMode::DwoOnly);
1452 } else {
1453 return writeOneObject(Asm, Layout, DwoMode::AllSections);
1454 }
1455}
1456
1457uint64_t WasmObjectWriter::writeOneObject(MCAssembler &Asm,
1458 const MCAsmLayout &Layout,
1459 DwoMode Mode) {
1460 uint64_t StartOffset = W->OS.tell();
1461 SectionCount = 0;
1462 CustomSections.clear();
1463
1464 LLVM_DEBUG(dbgs() << "WasmObjectWriter::writeObject\n");
1465
1466 // Collect information from the available symbols.
1468 SmallVector<uint32_t, 4> TableElems;
1471 SmallVector<uint32_t, 2> TagTypes;
1476 std::map<StringRef, std::vector<WasmComdatEntry>> Comdats;
1477 uint64_t DataSize = 0;
1478 if (Mode != DwoMode::DwoOnly) {
1479 prepareImports(Imports, Asm, Layout);
1480 }
1481
1482 // Populate DataSegments and CustomSections, which must be done before
1483 // populating DataLocations.
1484 for (MCSection &Sec : Asm) {
1485 auto &Section = static_cast<MCSectionWasm &>(Sec);
1486 StringRef SectionName = Section.getName();
1487
1488 if (Mode == DwoMode::NonDwoOnly && isDwoSection(Sec))
1489 continue;
1490 if (Mode == DwoMode::DwoOnly && !isDwoSection(Sec))
1491 continue;
1492
1493 LLVM_DEBUG(dbgs() << "Processing Section " << SectionName << " group "
1494 << Section.getGroup() << "\n";);
1495
1496 // .init_array sections are handled specially elsewhere.
1497 if (SectionName.starts_with(".init_array"))
1498 continue;
1499
1500 // Code is handled separately
1501 if (Section.isText())
1502 continue;
1503
1504 if (Section.isWasmData()) {
1505 uint32_t SegmentIndex = DataSegments.size();
1506 DataSize = alignTo(DataSize, Section.getAlign());
1507 DataSegments.emplace_back();
1508 WasmDataSegment &Segment = DataSegments.back();
1509 Segment.Name = SectionName;
1510 Segment.InitFlags = Section.getPassive()
1512 : 0;
1513 Segment.Offset = DataSize;
1514 Segment.Section = &Section;
1515 addData(Segment.Data, Section);
1516 Segment.Alignment = Log2(Section.getAlign());
1517 Segment.LinkingFlags = Section.getSegmentFlags();
1518 DataSize += Segment.Data.size();
1519 Section.setSegmentIndex(SegmentIndex);
1520
1521 if (const MCSymbolWasm *C = Section.getGroup()) {
1522 Comdats[C->getName()].emplace_back(
1523 WasmComdatEntry{wasm::WASM_COMDAT_DATA, SegmentIndex});
1524 }
1525 } else {
1526 // Create custom sections
1527 assert(Section.isMetadata());
1528
1530
1531 // For user-defined custom sections, strip the prefix
1532 Name.consume_front(".custom_section.");
1533
1534 MCSymbol *Begin = Sec.getBeginSymbol();
1535 if (Begin) {
1536 assert(WasmIndices.count(cast<MCSymbolWasm>(Begin)) == 0);
1537 WasmIndices[cast<MCSymbolWasm>(Begin)] = CustomSections.size();
1538 }
1539
1540 // Separate out the producers and target features sections
1541 if (Name == "producers") {
1542 ProducersSection = std::make_unique<WasmCustomSection>(Name, &Section);
1543 continue;
1544 }
1545 if (Name == "target_features") {
1546 TargetFeaturesSection =
1547 std::make_unique<WasmCustomSection>(Name, &Section);
1548 continue;
1549 }
1550
1551 // Custom sections can also belong to COMDAT groups. In this case the
1552 // decriptor's "index" field is the section index (in the final object
1553 // file), but that is not known until after layout, so it must be fixed up
1554 // later
1555 if (const MCSymbolWasm *C = Section.getGroup()) {
1556 Comdats[C->getName()].emplace_back(
1557 WasmComdatEntry{wasm::WASM_COMDAT_SECTION,
1558 static_cast<uint32_t>(CustomSections.size())});
1559 }
1560
1561 CustomSections.emplace_back(Name, &Section);
1562 }
1563 }
1564
1565 if (Mode != DwoMode::DwoOnly) {
1566 // Populate WasmIndices and DataLocations for defined symbols.
1567 for (const MCSymbol &S : Asm.symbols()) {
1568 // Ignore unnamed temporary symbols, which aren't ever exported, imported,
1569 // or used in relocations.
1570 if (S.isTemporary() && S.getName().empty())
1571 continue;
1572
1573 const auto &WS = static_cast<const MCSymbolWasm &>(S);
1574 LLVM_DEBUG(
1575 dbgs() << "MCSymbol: "
1576 << toString(WS.getType().value_or(wasm::WASM_SYMBOL_TYPE_DATA))
1577 << " '" << S << "'"
1578 << " isDefined=" << S.isDefined() << " isExternal="
1579 << S.isExternal() << " isTemporary=" << S.isTemporary()
1580 << " isWeak=" << WS.isWeak() << " isHidden=" << WS.isHidden()
1581 << " isVariable=" << WS.isVariable() << "\n");
1582
1583 if (WS.isVariable())
1584 continue;
1585 if (WS.isComdat() && !WS.isDefined())
1586 continue;
1587
1588 if (WS.isFunction()) {
1589 unsigned Index;
1590 if (WS.isDefined()) {
1591 if (WS.getOffset() != 0)
1593 "function sections must contain one function each");
1594
1595 // A definition. Write out the function body.
1596 Index = NumFunctionImports + Functions.size();
1597 WasmFunction Func;
1598 Func.SigIndex = getFunctionType(WS);
1599 Func.Section = &WS.getSection();
1600 assert(WasmIndices.count(&WS) == 0);
1601 WasmIndices[&WS] = Index;
1602 Functions.push_back(Func);
1603
1604 auto &Section = static_cast<MCSectionWasm &>(WS.getSection());
1605 if (const MCSymbolWasm *C = Section.getGroup()) {
1606 Comdats[C->getName()].emplace_back(
1607 WasmComdatEntry{wasm::WASM_COMDAT_FUNCTION, Index});
1608 }
1609
1610 if (WS.hasExportName()) {
1612 Export.Name = WS.getExportName();
1614 Export.Index = Index;
1615 Exports.push_back(Export);
1616 }
1617 } else {
1618 // An import; the index was assigned above.
1619 Index = WasmIndices.find(&WS)->second;
1620 }
1621
1622 LLVM_DEBUG(dbgs() << " -> function index: " << Index << "\n");
1623
1624 } else if (WS.isData()) {
1625 if (!isInSymtab(WS))
1626 continue;
1627
1628 if (!WS.isDefined()) {
1629 LLVM_DEBUG(dbgs() << " -> segment index: -1"
1630 << "\n");
1631 continue;
1632 }
1633
1634 if (!WS.getSize())
1635 report_fatal_error("data symbols must have a size set with .size: " +
1636 WS.getName());
1637
1638 int64_t Size = 0;
1639 if (!WS.getSize()->evaluateAsAbsolute(Size, Layout))
1640 report_fatal_error(".size expression must be evaluatable");
1641
1642 auto &DataSection = static_cast<MCSectionWasm &>(WS.getSection());
1643 if (!DataSection.isWasmData())
1644 report_fatal_error("data symbols must live in a data section: " +
1645 WS.getName());
1646
1647 // For each data symbol, export it in the symtab as a reference to the
1648 // corresponding Wasm data segment.
1650 DataSection.getSegmentIndex(), Layout.getSymbolOffset(WS),
1651 static_cast<uint64_t>(Size)};
1652 assert(DataLocations.count(&WS) == 0);
1653 DataLocations[&WS] = Ref;
1654 LLVM_DEBUG(dbgs() << " -> segment index: " << Ref.Segment << "\n");
1655
1656 } else if (WS.isGlobal()) {
1657 // A "true" Wasm global (currently just __stack_pointer)
1658 if (WS.isDefined()) {
1660 Global.Type = WS.getGlobalType();
1661 Global.Index = NumGlobalImports + Globals.size();
1662 Global.InitExpr.Extended = false;
1663 switch (Global.Type.Type) {
1665 Global.InitExpr.Inst.Opcode = wasm::WASM_OPCODE_I32_CONST;
1666 break;
1668 Global.InitExpr.Inst.Opcode = wasm::WASM_OPCODE_I64_CONST;
1669 break;
1671 Global.InitExpr.Inst.Opcode = wasm::WASM_OPCODE_F32_CONST;
1672 break;
1674 Global.InitExpr.Inst.Opcode = wasm::WASM_OPCODE_F64_CONST;
1675 break;
1677 Global.InitExpr.Inst.Opcode = wasm::WASM_OPCODE_REF_NULL;
1678 break;
1679 default:
1680 llvm_unreachable("unexpected type");
1681 }
1682 assert(WasmIndices.count(&WS) == 0);
1683 WasmIndices[&WS] = Global.Index;
1684 Globals.push_back(Global);
1685 } else {
1686 // An import; the index was assigned above
1687 LLVM_DEBUG(dbgs() << " -> global index: "
1688 << WasmIndices.find(&WS)->second << "\n");
1689 }
1690 } else if (WS.isTable()) {
1691 if (WS.isDefined()) {
1692 wasm::WasmTable Table;
1693 Table.Index = NumTableImports + Tables.size();
1694 Table.Type = WS.getTableType();
1695 assert(WasmIndices.count(&WS) == 0);
1696 WasmIndices[&WS] = Table.Index;
1697 Tables.push_back(Table);
1698 }
1699 LLVM_DEBUG(dbgs() << " -> table index: "
1700 << WasmIndices.find(&WS)->second << "\n");
1701 } else if (WS.isTag()) {
1702 // C++ exception symbol (__cpp_exception) or longjmp symbol
1703 // (__c_longjmp)
1704 unsigned Index;
1705 if (WS.isDefined()) {
1706 Index = NumTagImports + TagTypes.size();
1707 uint32_t SigIndex = getTagType(WS);
1708 assert(WasmIndices.count(&WS) == 0);
1709 WasmIndices[&WS] = Index;
1710 TagTypes.push_back(SigIndex);
1711 } else {
1712 // An import; the index was assigned above.
1713 assert(WasmIndices.count(&WS) > 0);
1714 }
1715 LLVM_DEBUG(dbgs() << " -> tag index: " << WasmIndices.find(&WS)->second
1716 << "\n");
1717
1718 } else {
1719 assert(WS.isSection());
1720 }
1721 }
1722
1723 // Populate WasmIndices and DataLocations for aliased symbols. We need to
1724 // process these in a separate pass because we need to have processed the
1725 // target of the alias before the alias itself and the symbols are not
1726 // necessarily ordered in this way.
1727 for (const MCSymbol &S : Asm.symbols()) {
1728 if (!S.isVariable())
1729 continue;
1730
1731 assert(S.isDefined());
1732
1733 const auto *BS = Layout.getBaseSymbol(S);
1734 if (!BS)
1735 report_fatal_error(Twine(S.getName()) +
1736 ": absolute addressing not supported!");
1737 const MCSymbolWasm *Base = cast<MCSymbolWasm>(BS);
1738
1739 // Find the target symbol of this weak alias and export that index
1740 const auto &WS = static_cast<const MCSymbolWasm &>(S);
1741 LLVM_DEBUG(dbgs() << WS.getName() << ": weak alias of '" << *Base
1742 << "'\n");
1743
1744 if (Base->isFunction()) {
1745 assert(WasmIndices.count(Base) > 0);
1746 uint32_t WasmIndex = WasmIndices.find(Base)->second;
1747 assert(WasmIndices.count(&WS) == 0);
1748 WasmIndices[&WS] = WasmIndex;
1749 LLVM_DEBUG(dbgs() << " -> index:" << WasmIndex << "\n");
1750 } else if (Base->isData()) {
1751 auto &DataSection = static_cast<MCSectionWasm &>(WS.getSection());
1752 uint64_t Offset = Layout.getSymbolOffset(S);
1753 int64_t Size = 0;
1754 // For data symbol alias we use the size of the base symbol as the
1755 // size of the alias. When an offset from the base is involved this
1756 // can result in a offset + size goes past the end of the data section
1757 // which out object format doesn't support. So we must clamp it.
1758 if (!Base->getSize()->evaluateAsAbsolute(Size, Layout))
1759 report_fatal_error(".size expression must be evaluatable");
1760 const WasmDataSegment &Segment =
1761 DataSegments[DataSection.getSegmentIndex()];
1762 Size =
1763 std::min(static_cast<uint64_t>(Size), Segment.Data.size() - Offset);
1765 DataSection.getSegmentIndex(),
1766 static_cast<uint32_t>(Layout.getSymbolOffset(S)),
1767 static_cast<uint32_t>(Size)};
1768 DataLocations[&WS] = Ref;
1769 LLVM_DEBUG(dbgs() << " -> index:" << Ref.Segment << "\n");
1770 } else {
1771 report_fatal_error("don't yet support global/tag aliases");
1772 }
1773 }
1774 }
1775
1776 // Finally, populate the symbol table itself, in its "natural" order.
1777 for (const MCSymbol &S : Asm.symbols()) {
1778 const auto &WS = static_cast<const MCSymbolWasm &>(S);
1779 if (!isInSymtab(WS)) {
1781 continue;
1782 }
1783 LLVM_DEBUG(dbgs() << "adding to symtab: " << WS << "\n");
1784
1785 uint32_t Flags = 0;
1786 if (WS.isWeak())
1788 if (WS.isHidden())
1790 if (!WS.isExternal() && WS.isDefined())
1792 if (WS.isUndefined())
1794 if (WS.isNoStrip()) {
1796 if (isEmscripten()) {
1798 }
1799 }
1800 if (WS.hasImportName())
1802 if (WS.hasExportName())
1804 if (WS.isTLS())
1806
1808 Info.Name = WS.getName();
1809 Info.Kind = WS.getType().value_or(wasm::WASM_SYMBOL_TYPE_DATA);
1810 Info.Flags = Flags;
1811 if (!WS.isData()) {
1812 assert(WasmIndices.count(&WS) > 0);
1813 Info.ElementIndex = WasmIndices.find(&WS)->second;
1814 } else if (WS.isDefined()) {
1815 assert(DataLocations.count(&WS) > 0);
1816 Info.DataRef = DataLocations.find(&WS)->second;
1817 }
1818 WS.setIndex(SymbolInfos.size());
1819 SymbolInfos.emplace_back(Info);
1820 }
1821
1822 {
1823 auto HandleReloc = [&](const WasmRelocationEntry &Rel) {
1824 // Functions referenced by a relocation need to put in the table. This is
1825 // purely to make the object file's provisional values readable, and is
1826 // ignored by the linker, which re-calculates the relocations itself.
1827 if (Rel.Type != wasm::R_WASM_TABLE_INDEX_I32 &&
1828 Rel.Type != wasm::R_WASM_TABLE_INDEX_I64 &&
1829 Rel.Type != wasm::R_WASM_TABLE_INDEX_SLEB &&
1830 Rel.Type != wasm::R_WASM_TABLE_INDEX_SLEB64 &&
1831 Rel.Type != wasm::R_WASM_TABLE_INDEX_REL_SLEB &&
1832 Rel.Type != wasm::R_WASM_TABLE_INDEX_REL_SLEB64)
1833 return;
1834 assert(Rel.Symbol->isFunction());
1835 const MCSymbolWasm *Base =
1836 cast<MCSymbolWasm>(Layout.getBaseSymbol(*Rel.Symbol));
1837 uint32_t FunctionIndex = WasmIndices.find(Base)->second;
1838 uint32_t TableIndex = TableElems.size() + InitialTableOffset;
1839 if (TableIndices.try_emplace(Base, TableIndex).second) {
1840 LLVM_DEBUG(dbgs() << " -> adding " << Base->getName()
1841 << " to table: " << TableIndex << "\n");
1842 TableElems.push_back(FunctionIndex);
1843 registerFunctionType(*Base);
1844 }
1845 };
1846
1847 for (const WasmRelocationEntry &RelEntry : CodeRelocations)
1848 HandleReloc(RelEntry);
1849 for (const WasmRelocationEntry &RelEntry : DataRelocations)
1850 HandleReloc(RelEntry);
1851 }
1852
1853 // Translate .init_array section contents into start functions.
1854 for (const MCSection &S : Asm) {
1855 const auto &WS = static_cast<const MCSectionWasm &>(S);
1856 if (WS.getName().starts_with(".fini_array"))
1857 report_fatal_error(".fini_array sections are unsupported");
1858 if (!WS.getName().starts_with(".init_array"))
1859 continue;
1860 auto IT = WS.begin();
1861 if (IT == WS.end())
1862 continue;
1863 const MCFragment &EmptyFrag = *IT;
1864 if (EmptyFrag.getKind() != MCFragment::FT_Data)
1865 report_fatal_error(".init_array section should be aligned");
1866
1867 const MCFragment &AlignFrag = *EmptyFrag.getNext();
1868 if (AlignFrag.getKind() != MCFragment::FT_Align)
1869 report_fatal_error(".init_array section should be aligned");
1870 if (cast<MCAlignFragment>(AlignFrag).getAlignment() !=
1871 Align(is64Bit() ? 8 : 4))
1872 report_fatal_error(".init_array section should be aligned for pointers");
1873
1874 const MCFragment &Frag = *AlignFrag.getNext();
1875 if (Frag.hasInstructions() || Frag.getKind() != MCFragment::FT_Data)
1876 report_fatal_error("only data supported in .init_array section");
1877
1878 uint16_t Priority = UINT16_MAX;
1879 unsigned PrefixLength = strlen(".init_array");
1880 if (WS.getName().size() > PrefixLength) {
1881 if (WS.getName()[PrefixLength] != '.')
1883 ".init_array section priority should start with '.'");
1884 if (WS.getName().substr(PrefixLength + 1).getAsInteger(10, Priority))
1885 report_fatal_error("invalid .init_array section priority");
1886 }
1887 const auto &DataFrag = cast<MCDataFragment>(Frag);
1888 const SmallVectorImpl<char> &Contents = DataFrag.getContents();
1889 for (const uint8_t *
1890 P = (const uint8_t *)Contents.data(),
1891 *End = (const uint8_t *)Contents.data() + Contents.size();
1892 P != End; ++P) {
1893 if (*P != 0)
1894 report_fatal_error("non-symbolic data in .init_array section");
1895 }
1896 for (const MCFixup &Fixup : DataFrag.getFixups()) {
1897 assert(Fixup.getKind() ==
1898 MCFixup::getKindForSize(is64Bit() ? 8 : 4, false));
1899 const MCExpr *Expr = Fixup.getValue();
1900 auto *SymRef = dyn_cast<MCSymbolRefExpr>(Expr);
1901 if (!SymRef)
1902 report_fatal_error("fixups in .init_array should be symbol references");
1903 const auto &TargetSym = cast<const MCSymbolWasm>(SymRef->getSymbol());
1904 if (TargetSym.getIndex() == InvalidIndex)
1905 report_fatal_error("symbols in .init_array should exist in symtab");
1906 if (!TargetSym.isFunction())
1907 report_fatal_error("symbols in .init_array should be for functions");
1908 InitFuncs.push_back(
1909 std::make_pair(Priority, TargetSym.getIndex()));
1910 }
1911 }
1912
1913 // Write out the Wasm header.
1914 writeHeader(Asm);
1915
1916 uint32_t CodeSectionIndex, DataSectionIndex;
1917 if (Mode != DwoMode::DwoOnly) {
1918 writeTypeSection(Signatures);
1919 writeImportSection(Imports, DataSize, TableElems.size());
1920 writeFunctionSection(Functions);
1921 writeTableSection(Tables);
1922 // Skip the "memory" section; we import the memory instead.
1923 writeTagSection(TagTypes);
1924 writeGlobalSection(Globals);
1925 writeExportSection(Exports);
1926 const MCSymbol *IndirectFunctionTable =
1927 Asm.getContext().lookupSymbol("__indirect_function_table");
1928 writeElemSection(cast_or_null<const MCSymbolWasm>(IndirectFunctionTable),
1929 TableElems);
1930 writeDataCountSection();
1931
1932 CodeSectionIndex = writeCodeSection(Asm, Layout, Functions);
1933 DataSectionIndex = writeDataSection(Layout);
1934 }
1935
1936 // The Sections in the COMDAT list have placeholder indices (their index among
1937 // custom sections, rather than among all sections). Fix them up here.
1938 for (auto &Group : Comdats) {
1939 for (auto &Entry : Group.second) {
1940 if (Entry.Kind == wasm::WASM_COMDAT_SECTION) {
1941 Entry.Index += SectionCount;
1942 }
1943 }
1944 }
1945 for (auto &CustomSection : CustomSections)
1946 writeCustomSection(CustomSection, Asm, Layout);
1947
1948 if (Mode != DwoMode::DwoOnly) {
1949 writeLinkingMetaDataSection(SymbolInfos, InitFuncs, Comdats);
1950
1951 writeRelocSection(CodeSectionIndex, "CODE", CodeRelocations);
1952 writeRelocSection(DataSectionIndex, "DATA", DataRelocations);
1953 }
1954 writeCustomRelocSections();
1955 if (ProducersSection)
1956 writeCustomSection(*ProducersSection, Asm, Layout);
1957 if (TargetFeaturesSection)
1958 writeCustomSection(*TargetFeaturesSection, Asm, Layout);
1959
1960 // TODO: Translate the .comment section to the output.
1961 return W->OS.tell() - StartOffset;
1962}
1963
1964std::unique_ptr<MCObjectWriter>
1965llvm::createWasmObjectWriter(std::unique_ptr<MCWasmObjectTargetWriter> MOTW,
1967 return std::make_unique<WasmObjectWriter>(std::move(MOTW), OS);
1968}
1969
1970std::unique_ptr<MCObjectWriter>
1971llvm::createWasmDwoObjectWriter(std::unique_ptr<MCWasmObjectTargetWriter> MOTW,
1973 raw_pwrite_stream &DwoOS) {
1974 return std::make_unique<WasmObjectWriter>(std::move(MOTW), OS, DwoOS);
1975}
static cl::opt< ITMode > IT(cl::desc("IT block support"), cl::Hidden, cl::init(DefaultIT), cl::values(clEnumValN(DefaultIT, "arm-default-it", "Generate any type of IT block"), clEnumValN(RestrictedIT, "arm-restrict-it", "Disallow complex IT blocks")))
static void print(raw_ostream &Out, object::Archive::Kind Kind, T Val)
static GCRegistry::Add< OcamlGC > B("ocaml", "ocaml 3.10-compatible GC")
static GCRegistry::Add< ErlangGC > A("erlang", "erlang-compatible garbage collector")
Analysis containing CSE Info
Definition: CSEInfo.cpp:27
#define LLVM_DUMP_METHOD
Mark debug helper function definitions like dump() that should not be stripped from debug builds.
Definition: Compiler.h:537
#define LLVM_DEBUG(X)
Definition: Debug.h:101
std::string Name
uint64_t Size
bool End
Definition: ELF_riscv.cpp:480
Symbol * Sym
Definition: ELF_riscv.cpp:479
#define P(N)
PowerPC TLS Dynamic Call Fixup
assert(ImpDefSCC.getReg()==AMDGPU::SCC &&ImpDefSCC.isDef())
This file contains some templates that are useful if you are working with the STL at all.
raw_pwrite_stream & OS
static const FuncProtoTy Signatures[]
static const unsigned InvalidIndex
static void addData(SmallVectorImpl< char > &DataBytes, MCSectionWasm &DataSection)
static bool isInSymtab(const MCSymbolWasm &Sym)
static bool isDwoSection(const MCSection &Sec)
static bool is64Bit(const char *name)
ArrayRef - Represent a constant reference to an array (0 or more elements consecutively in memory),...
Definition: ArrayRef.h:41
size_t size() const
size - Get the array size.
Definition: ArrayRef.h:165
bool empty() const
empty - Check if the array is empty.
Definition: ArrayRef.h:160
Encapsulates the layout of an assembly file at a particular point in time.
Definition: MCAsmLayout.h:28
const MCSymbol * getBaseSymbol(const MCSymbol &Symbol) const
If this symbol is equivalent to A + Constant, return A.
Definition: MCFragment.cpp:107
uint64_t getSectionAddressSize(const MCSection *Sec) const
Get the address space size of the given section, as it effects layout.
Definition: MCFragment.cpp:143
bool getSymbolOffset(const MCSymbol &S, uint64_t &Val) const
Get the offset of the given symbol, as computed in the current layout.
Definition: MCFragment.cpp:97
uint64_t getFragmentOffset(const MCFragment *F) const
Get the offset of the given fragment inside its containing section.
Context object for machine code objects.
Definition: MCContext.h:83
MCSymbol * lookupSymbol(const Twine &Name) const
Get the symbol for Name, or null.
Definition: MCContext.cpp:407
void reportError(SMLoc L, const Twine &Msg)
Definition: MCContext.cpp:1070
Base class for the full range of assembler expressions which are needed for parsing.
Definition: MCExpr.h:35
Encode information on a single operation to perform on a byte sequence (e.g., an encoded instruction)...
Definition: MCFixup.h:71
static MCFixupKind getKindForSize(unsigned Size, bool IsPCRel)
Return the generic fixup kind for a value with the given size.
Definition: MCFixup.h:109
FragmentType getKind() const
Definition: MCFragment.h:92
MCSection * getParent() const
Definition: MCFragment.h:94
MCFragment * getNext() const
Definition: MCFragment.h:90
bool hasInstructions() const
Does this fragment have instructions emitted into it? By default this is false, but specific fragment...
Definition: MCFragment.h:104
Defines the object file and target independent interfaces used by the assembler backend to write nati...
virtual uint64_t writeObject(MCAssembler &Asm, const MCAsmLayout &Layout)=0
Write the object file and returns the number of bytes written.
virtual void executePostLayoutBinding(MCAssembler &Asm, const MCAsmLayout &Layout)=0
Perform any late binding of symbols (for example, to assign symbol indices for use when generating re...
virtual void reset()
lifetime management
virtual void recordRelocation(MCAssembler &Asm, const MCAsmLayout &Layout, const MCFragment *Fragment, const MCFixup &Fixup, MCValue Target, uint64_t &FixedValue)=0
Record a relocation entry.
This represents a section on wasm.
Definition: MCSectionWasm.h:26
Instances of this class represent a uniqued identifier for a section in the current translation unit.
Definition: MCSection.h:36
Align getAlign() const
Definition: MCSection.h:158
bool isText() const
Definition: MCSection.h:143
StringRef getName() const
Definition: MCSection.h:142
MCSymbol * getBeginSymbol()
Definition: MCSection.h:147
Represent a reference to a symbol from inside an expression.
Definition: MCExpr.h:192
const MCSymbol & getSymbol() const
Definition: MCExpr.h:410
VariantKind getKind() const
Definition: MCExpr.h:412
MCSymbol - Instances of this class represent a symbol name in the MC file, and MCSymbols are created ...
Definition: MCSymbol.h:41
void setIndex(uint32_t Value) const
Set the (implementation defined) index.
Definition: MCSymbol.h:321
This represents an "assembler immediate".
Definition: MCValue.h:36
size_t size() const
Definition: SmallVector.h:91
This class consists of common code factored out of the SmallVector class to reduce code duplication b...
Definition: SmallVector.h:586
reference emplace_back(ArgTypes &&... Args)
Definition: SmallVector.h:950
iterator insert(iterator I, T &&Elt)
Definition: SmallVector.h:818
void resize(size_type N)
Definition: SmallVector.h:651
void push_back(const T &Elt)
Definition: SmallVector.h:426
pointer data()
Return a pointer to the vector's buffer, even if empty().
Definition: SmallVector.h:299
This is a 'vector' (really, a variable-sized array), optimized for the case when the array is small.
Definition: SmallVector.h:1209
StringRef - Represent a constant reference to a string, i.e.
Definition: StringRef.h:50
bool ends_with(StringRef Suffix) const
Check if this string ends with the given Suffix.
Definition: StringRef.h:270
Target - Wrapper for Target specific information.
Twine - A lightweight data structure for efficiently representing the concatenation of temporary valu...
Definition: Twine.h:81
The instances of the Type class are immutable: once they are created, they are never changed.
Definition: Type.h:45
LLVM Value Representation.
Definition: Value.h:74
A raw_ostream that discards all output.
Definition: raw_ostream.h:731
This class implements an extremely fast bulk output stream that can only output to a stream.
Definition: raw_ostream.h:52
An abstract base class for streams implementations that also support a pwrite operation.
Definition: raw_ostream.h:445
void pwrite(const char *Ptr, size_t Size, uint64_t Offset)
Definition: raw_ostream.h:453
#define llvm_unreachable(msg)
Marks that the current location is not supposed to be reachable.
@ Entry
Definition: COFF.h:811
@ C
The default llvm calling convention, compatible with C.
Definition: CallingConv.h:34
std::optional< const char * > toString(const std::optional< DWARFFormValue > &V)
Take an optional DWARFFormValue and try to extract a string value from it.
NodeAddr< FuncNode * > Func
Definition: RDFGraph.h:393
void write64le(void *P, uint64_t V)
Definition: Endian.h:471
void write32le(void *P, uint32_t V)
Definition: Endian.h:468
@ WASM_ELEM_SEGMENT_HAS_TABLE_NUMBER
Definition: Wasm.h:162
const unsigned WASM_SYMBOL_UNDEFINED
Definition: Wasm.h:236
const uint32_t WasmPageSize
Definition: Wasm.h:32
@ WASM_LIMITS_FLAG_HAS_MAX
Definition: Wasm.h:149
@ WASM_LIMITS_FLAG_IS_64
Definition: Wasm.h:151
@ WASM_LIMITS_FLAG_NONE
Definition: Wasm.h:148
const unsigned WASM_SYMBOL_NO_STRIP
Definition: Wasm.h:239
@ WASM_EXTERNAL_TABLE
Definition: Wasm.h:85
@ WASM_EXTERNAL_FUNCTION
Definition: Wasm.h:84
@ WASM_EXTERNAL_TAG
Definition: Wasm.h:88
@ WASM_EXTERNAL_MEMORY
Definition: Wasm.h:86
@ WASM_EXTERNAL_GLOBAL
Definition: Wasm.h:87
const char WasmMagic[]
Definition: Wasm.h:26
@ WASM_TYPE_I64
Definition: Wasm.h:55
@ WASM_TYPE_F64
Definition: Wasm.h:57
@ WASM_TYPE_EXTERNREF
Definition: Wasm.h:64
@ WASM_TYPE_FUNC
Definition: Wasm.h:73
@ WASM_TYPE_I32
Definition: Wasm.h:54
@ WASM_TYPE_F32
Definition: Wasm.h:56
@ WASM_DATA_SEGMENT_IS_PASSIVE
Definition: Wasm.h:155
@ WASM_DATA_SEGMENT_HAS_MEMINDEX
Definition: Wasm.h:156
const unsigned WASM_ELEM_SEGMENT_MASK_HAS_ELEM_KIND
Definition: Wasm.h:165
const unsigned WASM_SYMBOL_TLS
Definition: Wasm.h:240
@ WASM_OPCODE_F64_CONST
Definition: Wasm.h:105
@ WASM_OPCODE_END
Definition: Wasm.h:93
@ WASM_OPCODE_REF_NULL
Definition: Wasm.h:112
@ WASM_OPCODE_F32_CONST
Definition: Wasm.h:104
@ WASM_OPCODE_I64_CONST
Definition: Wasm.h:103
@ WASM_OPCODE_I32_CONST
Definition: Wasm.h:102
const uint32_t WasmMetadataVersion
Definition: Wasm.h:30
const unsigned WASM_SYMBOL_BINDING_WEAK
Definition: Wasm.h:232
const unsigned WASM_SYMBOL_BINDING_LOCAL
Definition: Wasm.h:233
@ WASM_SYMBOL_TYPE_GLOBAL
Definition: Wasm.h:211
@ WASM_SYMBOL_TYPE_DATA
Definition: Wasm.h:210
@ WASM_SYMBOL_TYPE_TAG
Definition: Wasm.h:213
@ WASM_SYMBOL_TYPE_TABLE
Definition: Wasm.h:214
@ WASM_SYMBOL_TYPE_SECTION
Definition: Wasm.h:212
@ WASM_SYMBOL_TYPE_FUNCTION
Definition: Wasm.h:209
const uint32_t WasmVersion
Definition: Wasm.h:28
const unsigned WASM_SYMBOL_EXPORTED
Definition: Wasm.h:237
bool relocTypeHasAddend(uint32_t type)
Definition: Wasm.cpp:66
@ WASM_INIT_FUNCS
Definition: Wasm.h:186
@ WASM_COMDAT_INFO
Definition: Wasm.h:187
@ WASM_SEGMENT_INFO
Definition: Wasm.h:185
@ WASM_SYMBOL_TABLE
Definition: Wasm.h:188
@ WASM_COMDAT_SECTION
Definition: Wasm.h:204
@ WASM_COMDAT_FUNCTION
Definition: Wasm.h:202
@ WASM_COMDAT_DATA
Definition: Wasm.h:201
@ WASM_SEC_CODE
Definition: Wasm.h:45
@ WASM_SEC_IMPORT
Definition: Wasm.h:37
@ WASM_SEC_EXPORT
Definition: Wasm.h:42
@ WASM_SEC_DATACOUNT
Definition: Wasm.h:47
@ WASM_SEC_CUSTOM
Definition: Wasm.h:35
@ WASM_SEC_FUNCTION
Definition: Wasm.h:38
@ WASM_SEC_ELEM
Definition: Wasm.h:44
@ WASM_SEC_TABLE
Definition: Wasm.h:39
@ WASM_SEC_TYPE
Definition: Wasm.h:36
@ WASM_SEC_TAG
Definition: Wasm.h:48
@ WASM_SEC_GLOBAL
Definition: Wasm.h:41
@ WASM_SEC_DATA
Definition: Wasm.h:46
const unsigned WASM_SYMBOL_EXPLICIT_NAME
Definition: Wasm.h:238
const unsigned WASM_SYMBOL_VISIBILITY_HIDDEN
Definition: Wasm.h:235
llvm::StringRef relocTypetoString(uint32_t type)
Definition: Wasm.cpp:29
This is an optimization pass for GlobalISel generic memory operations.
Definition: AddressRanges.h:18
void dump(const SparseBitVector< ElementSize > &LHS, raw_ostream &out)
@ Offset
Definition: DWP.cpp:480
void stable_sort(R &&Range)
Definition: STLExtras.h:1995
@ Export
Export information to summary.
@ Import
Import information from summary.
void append_range(Container &C, Range &&R)
Wrapper function to append range R to container C.
Definition: STLExtras.h:2067
std::unique_ptr< MCObjectWriter > createWasmObjectWriter(std::unique_ptr< MCWasmObjectTargetWriter > MOTW, raw_pwrite_stream &OS)
Construct a new Wasm writer instance.
raw_ostream & dbgs()
dbgs() - This returns a reference to a raw_ostream for debugging messages.
Definition: Debug.cpp:163
void report_fatal_error(Error Err, bool gen_crash_diag=true)
Report a serious error, calling any installed error handler.
Definition: Error.cpp:167
uint64_t offsetToAlignment(uint64_t Value, Align Alignment)
Returns the offset to the next integer (mod 2**64) that is greater than or equal to Value and is a mu...
Definition: Alignment.h:197
raw_fd_ostream & errs()
This returns a reference to a raw_ostream for standard error.
@ Global
Append to llvm.global_dtors.
@ Ref
The access may reference the value stored in memory.
uint64_t alignTo(uint64_t Size, Align A)
Returns a multiple of A needed to store Size bytes.
Definition: Alignment.h:155
raw_ostream & operator<<(raw_ostream &OS, const APFixedPoint &FX)
Definition: APFixedPoint.h:293
unsigned encodeSLEB128(int64_t Value, raw_ostream &OS, unsigned PadTo=0)
Utility function to encode a SLEB128 value to an output stream.
Definition: LEB128.h:23
unsigned encodeULEB128(uint64_t Value, raw_ostream &OS, unsigned PadTo=0)
Utility function to encode a ULEB128 value to an output stream.
Definition: LEB128.h:80
unsigned Log2(Align A)
Returns the log2 of the alignment.
Definition: Alignment.h:208
std::unique_ptr< MCObjectWriter > createWasmDwoObjectWriter(std::unique_ptr< MCWasmObjectTargetWriter > MOTW, raw_pwrite_stream &OS, raw_pwrite_stream &DwoOS)
This struct is a compact representation of a valid (non-zero power of two) alignment.
Definition: Alignment.h:39
@ FKF_IsPCRel
Is this fixup kind PCrelative? This is used by the assembler backend to evaluate fixup values in a ta...
Adapter to write values to a stream in a particular byte order.
Definition: EndianStream.h:67
WasmLimits Memory
Definition: Wasm.h:374
StringRef Field
Definition: Wasm.h:368
StringRef Module
Definition: Wasm.h:367
SmallVector< ValType, 1 > Returns
Definition: Wasm.h:488
SmallVector< ValType, 4 > Params
Definition: Wasm.h:489
WasmTableType Type
Definition: Wasm.h:322
uint32_t Index
Definition: Wasm.h:321