xref: /llvm-project-15.0.7/lld/wasm/Writer.cpp (revision ef2cdfe3)
1 //===- Writer.cpp ---------------------------------------------------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 
9 #include "Writer.h"
10 #include "Config.h"
11 #include "InputChunks.h"
12 #include "InputElement.h"
13 #include "MapFile.h"
14 #include "OutputSections.h"
15 #include "OutputSegment.h"
16 #include "Relocations.h"
17 #include "SymbolTable.h"
18 #include "SyntheticSections.h"
19 #include "WriterUtils.h"
20 #include "lld/Common/ErrorHandler.h"
21 #include "lld/Common/Memory.h"
22 #include "lld/Common/Strings.h"
23 #include "llvm/ADT/DenseSet.h"
24 #include "llvm/ADT/SmallSet.h"
25 #include "llvm/ADT/SmallVector.h"
26 #include "llvm/ADT/StringMap.h"
27 #include "llvm/BinaryFormat/Wasm.h"
28 #include "llvm/BinaryFormat/WasmTraits.h"
29 #include "llvm/Support/FileOutputBuffer.h"
30 #include "llvm/Support/Format.h"
31 #include "llvm/Support/FormatVariadic.h"
32 #include "llvm/Support/LEB128.h"
33 #include "llvm/Support/Parallel.h"
34 
35 #include <cstdarg>
36 #include <map>
37 
38 #define DEBUG_TYPE "lld"
39 
40 using namespace llvm;
41 using namespace llvm::wasm;
42 
43 namespace lld {
44 namespace wasm {
45 static constexpr int stackAlignment = 16;
46 static constexpr int heapAlignment = 16;
47 
48 namespace {
49 
50 // The writer writes a SymbolTable result to a file.
51 class Writer {
52 public:
53   void run();
54 
55 private:
56   void openFile();
57 
58   bool needsPassiveInitialization(const OutputSegment *segment);
59   bool hasPassiveInitializedSegments();
60 
61   void createSyntheticInitFunctions();
62   void createInitMemoryFunction();
63   void createStartFunction();
64   void createApplyDataRelocationsFunction();
65   void createApplyGlobalRelocationsFunction();
66   void createCallCtorsFunction();
67   void createInitTLSFunction();
68   void createCommandExportWrappers();
69   void createCommandExportWrapper(uint32_t functionIndex, DefinedFunction *f);
70 
71   void assignIndexes();
72   void populateSymtab();
73   void populateProducers();
74   void populateTargetFeatures();
75   void calculateInitFunctions();
76   void calculateImports();
77   void calculateExports();
78   void calculateCustomSections();
79   void calculateTypes();
80   void createOutputSegments();
81   OutputSegment *createOutputSegment(StringRef name);
82   void combineOutputSegments();
83   void layoutMemory();
84   void createHeader();
85 
86   void addSection(OutputSection *sec);
87 
88   void addSections();
89 
90   void createCustomSections();
91   void createSyntheticSections();
92   void createSyntheticSectionsPostLayout();
93   void finalizeSections();
94 
95   // Custom sections
96   void createRelocSections();
97 
98   void writeHeader();
99   void writeSections();
100 
101   uint64_t fileSize = 0;
102 
103   std::vector<WasmInitEntry> initFunctions;
104   llvm::StringMap<std::vector<InputChunk *>> customSectionMapping;
105 
106   // Stable storage for command export wrapper function name strings.
107   std::list<std::string> commandExportWrapperNames;
108 
109   // Elements that are used to construct the final output
110   std::string header;
111   std::vector<OutputSection *> outputSections;
112 
113   std::unique_ptr<FileOutputBuffer> buffer;
114 
115   std::vector<OutputSegment *> segments;
116   llvm::SmallDenseMap<StringRef, OutputSegment *> segmentMap;
117 };
118 
119 } // anonymous namespace
120 
121 void Writer::calculateCustomSections() {
122   log("calculateCustomSections");
123   bool stripDebug = config->stripDebug || config->stripAll;
124   for (ObjFile *file : symtab->objectFiles) {
125     for (InputChunk *section : file->customSections) {
126       // Exclude COMDAT sections that are not selected for inclusion
127       if (section->discarded)
128         continue;
129       StringRef name = section->getName();
130       // These custom sections are known the linker and synthesized rather than
131       // blindly copied.
132       if (name == "linking" || name == "name" || name == "producers" ||
133           name == "target_features" || name.startswith("reloc."))
134         continue;
135       // These custom sections are generated by `clang -fembed-bitcode`.
136       // These are used by the rust toolchain to ship LTO data along with
137       // compiled object code, but they don't want this included in the linker
138       // output.
139       if (name == ".llvmbc" || name == ".llvmcmd")
140         continue;
141       // Strip debug section in that option was specified.
142       if (stripDebug && name.startswith(".debug_"))
143         continue;
144       // Otherwise include custom sections by default and concatenate their
145       // contents.
146       customSectionMapping[name].push_back(section);
147     }
148   }
149 }
150 
151 void Writer::createCustomSections() {
152   log("createCustomSections");
153   for (auto &pair : customSectionMapping) {
154     StringRef name = pair.first();
155     LLVM_DEBUG(dbgs() << "createCustomSection: " << name << "\n");
156 
157     OutputSection *sec = make<CustomSection>(std::string(name), pair.second);
158     if (config->relocatable || config->emitRelocs) {
159       auto *sym = make<OutputSectionSymbol>(sec);
160       out.linkingSec->addToSymtab(sym);
161       sec->sectionSym = sym;
162     }
163     addSection(sec);
164   }
165 }
166 
167 // Create relocations sections in the final output.
168 // These are only created when relocatable output is requested.
169 void Writer::createRelocSections() {
170   log("createRelocSections");
171   // Don't use iterator here since we are adding to OutputSection
172   size_t origSize = outputSections.size();
173   for (size_t i = 0; i < origSize; i++) {
174     LLVM_DEBUG(dbgs() << "check section " << i << "\n");
175     OutputSection *sec = outputSections[i];
176 
177     // Count the number of needed sections.
178     uint32_t count = sec->getNumRelocations();
179     if (!count)
180       continue;
181 
182     StringRef name;
183     if (sec->type == WASM_SEC_DATA)
184       name = "reloc.DATA";
185     else if (sec->type == WASM_SEC_CODE)
186       name = "reloc.CODE";
187     else if (sec->type == WASM_SEC_CUSTOM)
188       name = saver.save("reloc." + sec->name);
189     else
190       llvm_unreachable(
191           "relocations only supported for code, data, or custom sections");
192 
193     addSection(make<RelocSection>(name, sec));
194   }
195 }
196 
197 void Writer::populateProducers() {
198   for (ObjFile *file : symtab->objectFiles) {
199     const WasmProducerInfo &info = file->getWasmObj()->getProducerInfo();
200     out.producersSec->addInfo(info);
201   }
202 }
203 
204 void Writer::writeHeader() {
205   memcpy(buffer->getBufferStart(), header.data(), header.size());
206 }
207 
208 void Writer::writeSections() {
209   uint8_t *buf = buffer->getBufferStart();
210   parallelForEach(outputSections, [buf](OutputSection *s) {
211     assert(s->isNeeded());
212     s->writeTo(buf);
213   });
214 }
215 
216 static void setGlobalPtr(DefinedGlobal *g, uint64_t memoryPtr) {
217   g->global->setPointerValue(memoryPtr);
218 }
219 
220 // Fix the memory layout of the output binary.  This assigns memory offsets
221 // to each of the input data sections as well as the explicit stack region.
222 // The default memory layout is as follows, from low to high.
223 //
224 //  - initialized data (starting at Config->globalBase)
225 //  - BSS data (not currently implemented in llvm)
226 //  - explicit stack (Config->ZStackSize)
227 //  - heap start / unallocated
228 //
229 // The --stack-first option means that stack is placed before any static data.
230 // This can be useful since it means that stack overflow traps immediately
231 // rather than overwriting global data, but also increases code size since all
232 // static data loads and stores requires larger offsets.
233 void Writer::layoutMemory() {
234   uint64_t memoryPtr = 0;
235 
236   auto placeStack = [&]() {
237     if (config->relocatable || config->isPic)
238       return;
239     memoryPtr = alignTo(memoryPtr, stackAlignment);
240     if (config->zStackSize != alignTo(config->zStackSize, stackAlignment))
241       error("stack size must be " + Twine(stackAlignment) + "-byte aligned");
242     log("mem: stack size  = " + Twine(config->zStackSize));
243     log("mem: stack base  = " + Twine(memoryPtr));
244     memoryPtr += config->zStackSize;
245     setGlobalPtr(cast<DefinedGlobal>(WasmSym::stackPointer), memoryPtr);
246     log("mem: stack top   = " + Twine(memoryPtr));
247   };
248 
249   if (config->stackFirst) {
250     placeStack();
251   } else {
252     memoryPtr = config->globalBase;
253     log("mem: global base = " + Twine(config->globalBase));
254   }
255 
256   if (WasmSym::globalBase)
257     WasmSym::globalBase->setVA(memoryPtr);
258 
259   uint64_t dataStart = memoryPtr;
260 
261   // Arbitrarily set __dso_handle handle to point to the start of the data
262   // segments.
263   if (WasmSym::dsoHandle)
264     WasmSym::dsoHandle->setVA(dataStart);
265 
266   out.dylinkSec->memAlign = 0;
267   for (OutputSegment *seg : segments) {
268     out.dylinkSec->memAlign = std::max(out.dylinkSec->memAlign, seg->alignment);
269     memoryPtr = alignTo(memoryPtr, 1ULL << seg->alignment);
270     seg->startVA = memoryPtr;
271     log(formatv("mem: {0,-15} offset={1,-8} size={2,-8} align={3}", seg->name,
272                 memoryPtr, seg->size, seg->alignment));
273 
274     if (!config->relocatable && seg->isTLS()) {
275       if (config->sharedMemory) {
276         auto *tlsSize = cast<DefinedGlobal>(WasmSym::tlsSize);
277         setGlobalPtr(tlsSize, seg->size);
278 
279         auto *tlsAlign = cast<DefinedGlobal>(WasmSym::tlsAlign);
280         setGlobalPtr(tlsAlign, int64_t{1} << seg->alignment);
281       } else {
282         auto *tlsBase = cast<DefinedGlobal>(WasmSym::tlsBase);
283         setGlobalPtr(tlsBase, memoryPtr);
284       }
285     }
286 
287     memoryPtr += seg->size;
288   }
289 
290   // Make space for the memory initialization flag
291   if (config->sharedMemory && hasPassiveInitializedSegments()) {
292     memoryPtr = alignTo(memoryPtr, 4);
293     WasmSym::initMemoryFlag = symtab->addSyntheticDataSymbol(
294         "__wasm_init_memory_flag", WASM_SYMBOL_VISIBILITY_HIDDEN);
295     WasmSym::initMemoryFlag->markLive();
296     WasmSym::initMemoryFlag->setVA(memoryPtr);
297     log(formatv("mem: {0,-15} offset={1,-8} size={2,-8} align={3}",
298                 "__wasm_init_memory_flag", memoryPtr, 4, 4));
299     memoryPtr += 4;
300   }
301 
302   if (WasmSym::dataEnd)
303     WasmSym::dataEnd->setVA(memoryPtr);
304 
305   uint64_t staticDataSize = memoryPtr - dataStart;
306   log("mem: static data = " + Twine(staticDataSize));
307   if (config->isPic)
308     out.dylinkSec->memSize = staticDataSize;
309 
310   if (!config->stackFirst)
311     placeStack();
312 
313   if (WasmSym::heapBase) {
314     // Set `__heap_base` to follow the end of the stack or global data. The
315     // fact that this comes last means that a malloc/brk implementation can
316     // grow the heap at runtime.
317     // We'll align the heap base here because memory allocators might expect
318     // __heap_base to be aligned already.
319     memoryPtr = alignTo(memoryPtr, heapAlignment);
320     log("mem: heap base   = " + Twine(memoryPtr));
321     WasmSym::heapBase->setVA(memoryPtr);
322   }
323 
324   uint64_t maxMemorySetting = 1ULL
325                               << (config->is64.getValueOr(false) ? 48 : 32);
326 
327   if (config->initialMemory != 0) {
328     if (config->initialMemory != alignTo(config->initialMemory, WasmPageSize))
329       error("initial memory must be " + Twine(WasmPageSize) + "-byte aligned");
330     if (memoryPtr > config->initialMemory)
331       error("initial memory too small, " + Twine(memoryPtr) + " bytes needed");
332     if (config->initialMemory > maxMemorySetting)
333       error("initial memory too large, cannot be greater than " +
334             Twine(maxMemorySetting));
335     memoryPtr = config->initialMemory;
336   }
337   out.memorySec->numMemoryPages =
338       alignTo(memoryPtr, WasmPageSize) / WasmPageSize;
339   log("mem: total pages = " + Twine(out.memorySec->numMemoryPages));
340 
341   if (config->maxMemory != 0) {
342     if (config->maxMemory != alignTo(config->maxMemory, WasmPageSize))
343       error("maximum memory must be " + Twine(WasmPageSize) + "-byte aligned");
344     if (memoryPtr > config->maxMemory)
345       error("maximum memory too small, " + Twine(memoryPtr) + " bytes needed");
346     if (config->maxMemory > maxMemorySetting)
347       error("maximum memory too large, cannot be greater than " +
348             Twine(maxMemorySetting));
349   }
350 
351   // Check max if explicitly supplied or required by shared memory
352   if (config->maxMemory != 0 || config->sharedMemory) {
353     uint64_t max = config->maxMemory;
354     if (max == 0) {
355       // If no maxMemory config was supplied but we are building with
356       // shared memory, we need to pick a sensible upper limit.
357       if (config->isPic)
358         max = maxMemorySetting;
359       else
360         max = alignTo(memoryPtr, WasmPageSize);
361     }
362     out.memorySec->maxMemoryPages = max / WasmPageSize;
363     log("mem: max pages   = " + Twine(out.memorySec->maxMemoryPages));
364   }
365 }
366 
367 void Writer::addSection(OutputSection *sec) {
368   if (!sec->isNeeded())
369     return;
370   log("addSection: " + toString(*sec));
371   sec->sectionIndex = outputSections.size();
372   outputSections.push_back(sec);
373 }
374 
375 // If a section name is valid as a C identifier (which is rare because of
376 // the leading '.'), linkers are expected to define __start_<secname> and
377 // __stop_<secname> symbols. They are at beginning and end of the section,
378 // respectively. This is not requested by the ELF standard, but GNU ld and
379 // gold provide the feature, and used by many programs.
380 static void addStartStopSymbols(const OutputSegment *seg) {
381   StringRef name = seg->name;
382   if (!isValidCIdentifier(name))
383     return;
384   LLVM_DEBUG(dbgs() << "addStartStopSymbols: " << name << "\n");
385   uint64_t start = seg->startVA;
386   uint64_t stop = start + seg->size;
387   symtab->addOptionalDataSymbol(saver.save("__start_" + name), start);
388   symtab->addOptionalDataSymbol(saver.save("__stop_" + name), stop);
389 }
390 
391 void Writer::addSections() {
392   addSection(out.dylinkSec);
393   addSection(out.typeSec);
394   addSection(out.importSec);
395   addSection(out.functionSec);
396   addSection(out.tableSec);
397   addSection(out.memorySec);
398   addSection(out.tagSec);
399   addSection(out.globalSec);
400   addSection(out.exportSec);
401   addSection(out.startSec);
402   addSection(out.elemSec);
403   addSection(out.dataCountSec);
404 
405   addSection(make<CodeSection>(out.functionSec->inputFunctions));
406   addSection(make<DataSection>(segments));
407 
408   createCustomSections();
409 
410   addSection(out.linkingSec);
411   if (config->emitRelocs || config->relocatable) {
412     createRelocSections();
413   }
414 
415   addSection(out.nameSec);
416   addSection(out.producersSec);
417   addSection(out.targetFeaturesSec);
418 }
419 
420 void Writer::finalizeSections() {
421   for (OutputSection *s : outputSections) {
422     s->setOffset(fileSize);
423     s->finalizeContents();
424     fileSize += s->getSize();
425   }
426 }
427 
428 void Writer::populateTargetFeatures() {
429   StringMap<std::string> used;
430   StringMap<std::string> required;
431   StringMap<std::string> disallowed;
432   SmallSet<std::string, 8> &allowed = out.targetFeaturesSec->features;
433   bool tlsUsed = false;
434 
435   // Only infer used features if user did not specify features
436   bool inferFeatures = !config->features.hasValue();
437 
438   if (!inferFeatures) {
439     auto &explicitFeatures = config->features.getValue();
440     allowed.insert(explicitFeatures.begin(), explicitFeatures.end());
441     if (!config->checkFeatures)
442       return;
443   }
444 
445   // Find the sets of used, required, and disallowed features
446   for (ObjFile *file : symtab->objectFiles) {
447     StringRef fileName(file->getName());
448     for (auto &feature : file->getWasmObj()->getTargetFeatures()) {
449       switch (feature.Prefix) {
450       case WASM_FEATURE_PREFIX_USED:
451         used.insert({feature.Name, std::string(fileName)});
452         break;
453       case WASM_FEATURE_PREFIX_REQUIRED:
454         used.insert({feature.Name, std::string(fileName)});
455         required.insert({feature.Name, std::string(fileName)});
456         break;
457       case WASM_FEATURE_PREFIX_DISALLOWED:
458         disallowed.insert({feature.Name, std::string(fileName)});
459         break;
460       default:
461         error("Unrecognized feature policy prefix " +
462               std::to_string(feature.Prefix));
463       }
464     }
465 
466     // Find TLS data segments
467     auto isTLS = [](InputChunk *segment) {
468       return segment->live && segment->isTLS();
469     };
470     tlsUsed = tlsUsed ||
471               std::any_of(file->segments.begin(), file->segments.end(), isTLS);
472   }
473 
474   if (inferFeatures)
475     for (const auto &key : used.keys())
476       allowed.insert(std::string(key));
477 
478   if (!config->checkFeatures)
479     return;
480 
481   if (!config->relocatable && allowed.count("mutable-globals") == 0) {
482     for (const Symbol *sym : out.importSec->importedSymbols) {
483       if (auto *global = dyn_cast<GlobalSymbol>(sym)) {
484         if (global->getGlobalType()->Mutable) {
485           error(Twine("mutable global imported but 'mutable-globals' feature "
486                       "not present in inputs: `") +
487                 toString(*sym) + "`. Use --no-check-features to suppress.");
488         }
489       }
490     }
491     for (const Symbol *sym : out.exportSec->exportedSymbols) {
492       if (isa<GlobalSymbol>(sym)) {
493         error(Twine("mutable global exported but 'mutable-globals' feature "
494                     "not present in inputs: `") +
495               toString(*sym) + "`. Use --no-check-features to suppress.");
496       }
497     }
498   }
499 
500   if (config->sharedMemory) {
501     if (disallowed.count("shared-mem"))
502       error("--shared-memory is disallowed by " + disallowed["shared-mem"] +
503             " because it was not compiled with 'atomics' or 'bulk-memory' "
504             "features.");
505 
506     for (auto feature : {"atomics", "bulk-memory"})
507       if (!allowed.count(feature))
508         error(StringRef("'") + feature +
509               "' feature must be used in order to use shared memory");
510   }
511 
512   if (tlsUsed) {
513     for (auto feature : {"atomics", "bulk-memory"})
514       if (!allowed.count(feature))
515         error(StringRef("'") + feature +
516               "' feature must be used in order to use thread-local storage");
517   }
518 
519   // Validate that used features are allowed in output
520   if (!inferFeatures) {
521     for (const auto &feature : used.keys()) {
522       if (!allowed.count(std::string(feature)))
523         error(Twine("Target feature '") + feature + "' used by " +
524               used[feature] + " is not allowed.");
525     }
526   }
527 
528   // Validate the required and disallowed constraints for each file
529   for (ObjFile *file : symtab->objectFiles) {
530     StringRef fileName(file->getName());
531     SmallSet<std::string, 8> objectFeatures;
532     for (const auto &feature : file->getWasmObj()->getTargetFeatures()) {
533       if (feature.Prefix == WASM_FEATURE_PREFIX_DISALLOWED)
534         continue;
535       objectFeatures.insert(feature.Name);
536       if (disallowed.count(feature.Name))
537         error(Twine("Target feature '") + feature.Name + "' used in " +
538               fileName + " is disallowed by " + disallowed[feature.Name] +
539               ". Use --no-check-features to suppress.");
540     }
541     for (const auto &feature : required.keys()) {
542       if (!objectFeatures.count(std::string(feature)))
543         error(Twine("Missing target feature '") + feature + "' in " + fileName +
544               ", required by " + required[feature] +
545               ". Use --no-check-features to suppress.");
546     }
547   }
548 }
549 
550 static bool shouldImport(Symbol *sym) {
551   // We don't generate imports for data symbols. They however can be imported
552   // as GOT entries.
553   if (isa<DataSymbol>(sym))
554     return false;
555   if (!sym->isLive())
556     return false;
557   if (!sym->isUsedInRegularObj)
558     return false;
559 
560   // When a symbol is weakly defined in a shared library we need to allow
561   // it to be overridden by another module so need to both import
562   // and export the symbol.
563   if (config->shared && sym->isDefined() && sym->isWeak())
564     return true;
565   if (!sym->isUndefined())
566     return false;
567   if (sym->isWeak() && !config->relocatable && !config->isPic)
568     return false;
569 
570   // In PIC mode we only need to import functions when they are called directly.
571   // Indirect usage all goes via GOT imports.
572   if (config->isPic) {
573     if (auto *f = dyn_cast<UndefinedFunction>(sym))
574       if (!f->isCalledDirectly)
575         return false;
576   }
577 
578   if (config->isPic || config->relocatable || config->importUndefined)
579     return true;
580   if (config->allowUndefinedSymbols.count(sym->getName()) != 0)
581     return true;
582 
583   return sym->importName.hasValue();
584 }
585 
586 void Writer::calculateImports() {
587   // Some inputs require that the indirect function table be assigned to table
588   // number 0, so if it is present and is an import, allocate it before any
589   // other tables.
590   if (WasmSym::indirectFunctionTable &&
591       shouldImport(WasmSym::indirectFunctionTable))
592     out.importSec->addImport(WasmSym::indirectFunctionTable);
593 
594   for (Symbol *sym : symtab->getSymbols()) {
595     if (!shouldImport(sym))
596       continue;
597     if (sym == WasmSym::indirectFunctionTable)
598       continue;
599     LLVM_DEBUG(dbgs() << "import: " << sym->getName() << "\n");
600     out.importSec->addImport(sym);
601   }
602 }
603 
604 void Writer::calculateExports() {
605   if (config->relocatable)
606     return;
607 
608   if (!config->relocatable && !config->importMemory)
609     out.exportSec->exports.push_back(
610         WasmExport{"memory", WASM_EXTERNAL_MEMORY, 0});
611 
612   unsigned globalIndex =
613       out.importSec->getNumImportedGlobals() + out.globalSec->numGlobals();
614 
615   for (Symbol *sym : symtab->getSymbols()) {
616     if (!sym->isExported())
617       continue;
618     if (!sym->isLive())
619       continue;
620 
621     StringRef name = sym->getName();
622     WasmExport export_;
623     if (auto *f = dyn_cast<DefinedFunction>(sym)) {
624       if (Optional<StringRef> exportName = f->function->getExportName()) {
625         name = *exportName;
626       }
627       export_ = {name, WASM_EXTERNAL_FUNCTION, f->getExportedFunctionIndex()};
628     } else if (auto *g = dyn_cast<DefinedGlobal>(sym)) {
629       if (g->getGlobalType()->Mutable && !g->getFile() && !g->forceExport) {
630         // Avoid exporting mutable globals are linker synthesized (e.g.
631         // __stack_pointer or __tls_base) unless they are explicitly exported
632         // from the command line.
633         // Without this check `--export-all` would cause any program using the
634         // stack pointer to export a mutable global even if none of the input
635         // files were built with the `mutable-globals` feature.
636         continue;
637       }
638       export_ = {name, WASM_EXTERNAL_GLOBAL, g->getGlobalIndex()};
639     } else if (auto *t = dyn_cast<DefinedTag>(sym)) {
640       export_ = {name, WASM_EXTERNAL_TAG, t->getTagIndex()};
641     } else if (auto *d = dyn_cast<DefinedData>(sym)) {
642       if (d->segment && d->segment->isTLS()) {
643         // We can't currenly export TLS data symbols.
644         if (sym->isExportedExplicit())
645           error("TLS symbols cannot yet be exported: `" + toString(*sym) + "`");
646         continue;
647       }
648       out.globalSec->dataAddressGlobals.push_back(d);
649       export_ = {name, WASM_EXTERNAL_GLOBAL, globalIndex++};
650     } else {
651       auto *t = cast<DefinedTable>(sym);
652       export_ = {name, WASM_EXTERNAL_TABLE, t->getTableNumber()};
653     }
654 
655     LLVM_DEBUG(dbgs() << "Export: " << name << "\n");
656     out.exportSec->exports.push_back(export_);
657     out.exportSec->exportedSymbols.push_back(sym);
658   }
659 }
660 
661 void Writer::populateSymtab() {
662   if (!config->relocatable && !config->emitRelocs)
663     return;
664 
665   for (Symbol *sym : symtab->getSymbols())
666     if (sym->isUsedInRegularObj && sym->isLive())
667       out.linkingSec->addToSymtab(sym);
668 
669   for (ObjFile *file : symtab->objectFiles) {
670     LLVM_DEBUG(dbgs() << "Local symtab entries: " << file->getName() << "\n");
671     for (Symbol *sym : file->getSymbols())
672       if (sym->isLocal() && !isa<SectionSymbol>(sym) && sym->isLive())
673         out.linkingSec->addToSymtab(sym);
674   }
675 }
676 
677 void Writer::calculateTypes() {
678   // The output type section is the union of the following sets:
679   // 1. Any signature used in the TYPE relocation
680   // 2. The signatures of all imported functions
681   // 3. The signatures of all defined functions
682   // 4. The signatures of all imported tags
683   // 5. The signatures of all defined tags
684 
685   for (ObjFile *file : symtab->objectFiles) {
686     ArrayRef<WasmSignature> types = file->getWasmObj()->types();
687     for (uint32_t i = 0; i < types.size(); i++)
688       if (file->typeIsUsed[i])
689         file->typeMap[i] = out.typeSec->registerType(types[i]);
690   }
691 
692   for (const Symbol *sym : out.importSec->importedSymbols) {
693     if (auto *f = dyn_cast<FunctionSymbol>(sym))
694       out.typeSec->registerType(*f->signature);
695     else if (auto *t = dyn_cast<TagSymbol>(sym))
696       out.typeSec->registerType(*t->signature);
697   }
698 
699   for (const InputFunction *f : out.functionSec->inputFunctions)
700     out.typeSec->registerType(f->signature);
701 
702   for (const InputTag *t : out.tagSec->inputTags)
703     out.typeSec->registerType(t->signature);
704 }
705 
706 // In a command-style link, create a wrapper for each exported symbol
707 // which calls the constructors and destructors.
708 void Writer::createCommandExportWrappers() {
709   // This logic doesn't currently support Emscripten-style PIC mode.
710   assert(!config->isPic);
711 
712   // If there are no ctors and there's no libc `__wasm_call_dtors` to
713   // call, don't wrap the exports.
714   if (initFunctions.empty() && WasmSym::callDtors == NULL)
715     return;
716 
717   std::vector<DefinedFunction *> toWrap;
718 
719   for (Symbol *sym : symtab->getSymbols())
720     if (sym->isExported())
721       if (auto *f = dyn_cast<DefinedFunction>(sym))
722         toWrap.push_back(f);
723 
724   for (auto *f : toWrap) {
725     auto funcNameStr = (f->getName() + ".command_export").str();
726     commandExportWrapperNames.push_back(funcNameStr);
727     const std::string &funcName = commandExportWrapperNames.back();
728 
729     auto func = make<SyntheticFunction>(*f->getSignature(), funcName);
730     if (f->function->getExportName().hasValue())
731       func->setExportName(f->function->getExportName()->str());
732     else
733       func->setExportName(f->getName().str());
734 
735     DefinedFunction *def =
736         symtab->addSyntheticFunction(funcName, f->flags, func);
737     def->markLive();
738 
739     def->flags |= WASM_SYMBOL_EXPORTED;
740     def->flags &= ~WASM_SYMBOL_VISIBILITY_HIDDEN;
741     def->forceExport = f->forceExport;
742 
743     f->flags |= WASM_SYMBOL_VISIBILITY_HIDDEN;
744     f->flags &= ~WASM_SYMBOL_EXPORTED;
745     f->forceExport = false;
746 
747     out.functionSec->addFunction(func);
748 
749     createCommandExportWrapper(f->getFunctionIndex(), def);
750   }
751 }
752 
753 static void finalizeIndirectFunctionTable() {
754   if (!WasmSym::indirectFunctionTable)
755     return;
756 
757   if (shouldImport(WasmSym::indirectFunctionTable) &&
758       !WasmSym::indirectFunctionTable->hasTableNumber()) {
759     // Processing -Bsymbolic relocations resulted in a late requirement that the
760     // indirect function table be present, and we are running in --import-table
761     // mode.  Add the table now to the imports section.  Otherwise it will be
762     // added to the tables section later in assignIndexes.
763     out.importSec->addImport(WasmSym::indirectFunctionTable);
764   }
765 
766   uint32_t tableSize = config->tableBase + out.elemSec->numEntries();
767   WasmLimits limits = {0, tableSize, 0};
768   if (WasmSym::indirectFunctionTable->isDefined() && !config->growableTable) {
769     limits.Flags |= WASM_LIMITS_FLAG_HAS_MAX;
770     limits.Maximum = limits.Minimum;
771   }
772   WasmSym::indirectFunctionTable->setLimits(limits);
773 }
774 
775 static void scanRelocations() {
776   for (ObjFile *file : symtab->objectFiles) {
777     LLVM_DEBUG(dbgs() << "scanRelocations: " << file->getName() << "\n");
778     for (InputChunk *chunk : file->functions)
779       scanRelocations(chunk);
780     for (InputChunk *chunk : file->segments)
781       scanRelocations(chunk);
782     for (auto &p : file->customSections)
783       scanRelocations(p);
784   }
785 }
786 
787 void Writer::assignIndexes() {
788   // Seal the import section, since other index spaces such as function and
789   // global are effected by the number of imports.
790   out.importSec->seal();
791 
792   for (InputFunction *func : symtab->syntheticFunctions)
793     out.functionSec->addFunction(func);
794 
795   for (ObjFile *file : symtab->objectFiles) {
796     LLVM_DEBUG(dbgs() << "Functions: " << file->getName() << "\n");
797     for (InputFunction *func : file->functions)
798       out.functionSec->addFunction(func);
799   }
800 
801   for (InputGlobal *global : symtab->syntheticGlobals)
802     out.globalSec->addGlobal(global);
803 
804   for (ObjFile *file : symtab->objectFiles) {
805     LLVM_DEBUG(dbgs() << "Globals: " << file->getName() << "\n");
806     for (InputGlobal *global : file->globals)
807       out.globalSec->addGlobal(global);
808   }
809 
810   for (ObjFile *file : symtab->objectFiles) {
811     LLVM_DEBUG(dbgs() << "Tags: " << file->getName() << "\n");
812     for (InputTag *tag : file->tags)
813       out.tagSec->addTag(tag);
814   }
815 
816   for (ObjFile *file : symtab->objectFiles) {
817     LLVM_DEBUG(dbgs() << "Tables: " << file->getName() << "\n");
818     for (InputTable *table : file->tables)
819       out.tableSec->addTable(table);
820   }
821 
822   for (InputTable *table : symtab->syntheticTables)
823     out.tableSec->addTable(table);
824 
825   out.globalSec->assignIndexes();
826   out.tableSec->assignIndexes();
827 }
828 
829 static StringRef getOutputDataSegmentName(const InputChunk &seg) {
830   // We always merge .tbss and .tdata into a single TLS segment so all TLS
831   // symbols are be relative to single __tls_base.
832   if (seg.isTLS())
833     return ".tdata";
834   StringRef name = seg.getName();
835   if (!config->mergeDataSegments)
836     return name;
837   if (name.startswith(".text."))
838     return ".text";
839   if (name.startswith(".data."))
840     return ".data";
841   if (name.startswith(".bss."))
842     return ".bss";
843   if (name.startswith(".rodata."))
844     return ".rodata";
845   return name;
846 }
847 
848 OutputSegment *Writer::createOutputSegment(StringRef name) {
849   LLVM_DEBUG(dbgs() << "new segment: " << name << "\n");
850   OutputSegment *s = make<OutputSegment>(name);
851   if (config->sharedMemory)
852     s->initFlags = WASM_DATA_SEGMENT_IS_PASSIVE;
853   // Exported memories are guaranteed to be zero-initialized, so no need
854   // to emit data segments for bss sections.
855   // TODO: consider initializing bss sections with memory.fill
856   // instructions when memory is imported and bulk-memory is available.
857   if (!config->importMemory && !config->relocatable && name.startswith(".bss"))
858     s->isBss = true;
859   segments.push_back(s);
860   return s;
861 }
862 
863 void Writer::createOutputSegments() {
864   for (ObjFile *file : symtab->objectFiles) {
865     for (InputChunk *segment : file->segments) {
866       if (!segment->live)
867         continue;
868       StringRef name = getOutputDataSegmentName(*segment);
869       OutputSegment *s = nullptr;
870       // When running in relocatable mode we can't merge segments that are part
871       // of comdat groups since the ultimate linker needs to be able exclude or
872       // include them individually.
873       if (config->relocatable && !segment->getComdatName().empty()) {
874         s = createOutputSegment(name);
875       } else {
876         if (segmentMap.count(name) == 0)
877           segmentMap[name] = createOutputSegment(name);
878         s = segmentMap[name];
879       }
880       s->addInputSegment(segment);
881     }
882   }
883 
884   // Sort segments by type, placing .bss last
885   std::stable_sort(segments.begin(), segments.end(),
886                    [](const OutputSegment *a, const OutputSegment *b) {
887                      auto order = [](StringRef name) {
888                        return StringSwitch<int>(name)
889                            .StartsWith(".tdata", 0)
890                            .StartsWith(".rodata", 1)
891                            .StartsWith(".data", 2)
892                            .StartsWith(".bss", 4)
893                            .Default(3);
894                      };
895                      return order(a->name) < order(b->name);
896                    });
897 
898   for (size_t i = 0; i < segments.size(); ++i)
899     segments[i]->index = i;
900 
901   // Merge MergeInputSections into a single MergeSyntheticSection.
902   LLVM_DEBUG(dbgs() << "-- finalize input semgments\n");
903   for (OutputSegment *seg : segments)
904     seg->finalizeInputSegments();
905 }
906 
907 void Writer::combineOutputSegments() {
908   // With PIC code we currently only support a single active data segment since
909   // we only have a single __memory_base to use as our base address.  This pass
910   // combines all data segments into a single .data segment.
911   // This restructions can be relaxed once we have extended constant
912   // expressions available:
913   // https://github.com/WebAssembly/extended-const
914   assert(config->isPic && !config->sharedMemory);
915   if (segments.size() <= 1)
916     return;
917   OutputSegment *combined = make<OutputSegment>(".data");
918   combined->startVA = segments[0]->startVA;
919   for (OutputSegment *s : segments) {
920     bool first = true;
921     for (InputChunk *inSeg : s->inputSegments) {
922       if (first)
923         inSeg->alignment = std::max(inSeg->alignment, s->alignment);
924       first = false;
925 #ifndef NDEBUG
926       uint64_t oldVA = inSeg->getVA();
927 #endif
928       combined->addInputSegment(inSeg);
929 #ifndef NDEBUG
930       uint64_t newVA = inSeg->getVA();
931       LLVM_DEBUG(dbgs() << "added input segment. name=" << inSeg->getName()
932                         << " oldVA=" << oldVA << " newVA=" << newVA << "\n");
933       assert(oldVA == newVA);
934 #endif
935     }
936   }
937 
938   segments = {combined};
939 }
940 
941 static void createFunction(DefinedFunction *func, StringRef bodyContent) {
942   std::string functionBody;
943   {
944     raw_string_ostream os(functionBody);
945     writeUleb128(os, bodyContent.size(), "function size");
946     os << bodyContent;
947   }
948   ArrayRef<uint8_t> body = arrayRefFromStringRef(saver.save(functionBody));
949   cast<SyntheticFunction>(func->function)->setBody(body);
950 }
951 
952 bool Writer::needsPassiveInitialization(const OutputSegment *segment) {
953   return segment->initFlags & WASM_DATA_SEGMENT_IS_PASSIVE &&
954          !segment->isTLS() && !segment->isBss;
955 }
956 
957 bool Writer::hasPassiveInitializedSegments() {
958   return std::find_if(segments.begin(), segments.end(),
959                       [this](const OutputSegment *s) {
960                         return this->needsPassiveInitialization(s);
961                       }) != segments.end();
962 }
963 
964 void Writer::createSyntheticInitFunctions() {
965   if (config->relocatable)
966     return;
967 
968   static WasmSignature nullSignature = {{}, {}};
969 
970   // Passive segments are used to avoid memory being reinitialized on each
971   // thread's instantiation. These passive segments are initialized and
972   // dropped in __wasm_init_memory, which is registered as the start function
973   if (config->sharedMemory && hasPassiveInitializedSegments()) {
974     WasmSym::initMemory = symtab->addSyntheticFunction(
975         "__wasm_init_memory", WASM_SYMBOL_VISIBILITY_HIDDEN,
976         make<SyntheticFunction>(nullSignature, "__wasm_init_memory"));
977     WasmSym::initMemory->markLive();
978   }
979 
980   if (config->isPic) {
981     // For PIC code we create synthetic functions that apply relocations.
982     // These get called from __wasm_call_ctors before the user-level
983     // constructors.
984     WasmSym::applyDataRelocs = symtab->addSyntheticFunction(
985         "__wasm_apply_data_relocs", WASM_SYMBOL_VISIBILITY_HIDDEN,
986         make<SyntheticFunction>(nullSignature, "__wasm_apply_data_relocs"));
987     WasmSym::applyDataRelocs->markLive();
988 
989     if (out.globalSec->needsRelocations()) {
990       WasmSym::applyGlobalRelocs = symtab->addSyntheticFunction(
991           "__wasm_apply_global_relocs", WASM_SYMBOL_VISIBILITY_HIDDEN,
992           make<SyntheticFunction>(nullSignature, "__wasm_apply_global_relocs"));
993       WasmSym::applyGlobalRelocs->markLive();
994     }
995   }
996 
997   if (WasmSym::applyGlobalRelocs && WasmSym::initMemory) {
998     WasmSym::startFunction = symtab->addSyntheticFunction(
999         "__wasm_start", WASM_SYMBOL_VISIBILITY_HIDDEN,
1000         make<SyntheticFunction>(nullSignature, "__wasm_start"));
1001     WasmSym::startFunction->markLive();
1002   }
1003 }
1004 
1005 void Writer::createInitMemoryFunction() {
1006   LLVM_DEBUG(dbgs() << "createInitMemoryFunction\n");
1007   assert(WasmSym::initMemory);
1008   assert(WasmSym::initMemoryFlag);
1009   assert(hasPassiveInitializedSegments());
1010   uint64_t flagAddress = WasmSym::initMemoryFlag->getVA();
1011   bool is64 = config->is64.getValueOr(false);
1012   std::string bodyContent;
1013   {
1014     raw_string_ostream os(bodyContent);
1015     // Initialize memory in a thread-safe manner. The thread that successfully
1016     // increments the flag from 0 to 1 is is responsible for performing the
1017     // memory initialization. Other threads go sleep on the flag until the
1018     // first thread finishing initializing memory, increments the flag to 2,
1019     // and wakes all the other threads. Once the flag has been set to 2,
1020     // subsequently started threads will skip the sleep. All threads
1021     // unconditionally drop their passive data segments once memory has been
1022     // initialized. The generated code is as follows:
1023     //
1024     // (func $__wasm_init_memory
1025     //  (if
1026     //   (i32.atomic.rmw.cmpxchg align=2 offset=0
1027     //    (i32.const $__init_memory_flag)
1028     //    (i32.const 0)
1029     //    (i32.const 1)
1030     //   )
1031     //   (then
1032     //    (drop
1033     //     (i32.atomic.wait align=2 offset=0
1034     //      (i32.const $__init_memory_flag)
1035     //      (i32.const 1)
1036     //      (i32.const -1)
1037     //     )
1038     //    )
1039     //   )
1040     //   (else
1041     //    ( ... initialize data segments ... )
1042     //    (i32.atomic.store align=2 offset=0
1043     //     (i32.const $__init_memory_flag)
1044     //     (i32.const 2)
1045     //    )
1046     //    (drop
1047     //     (i32.atomic.notify align=2 offset=0
1048     //      (i32.const $__init_memory_flag)
1049     //      (i32.const -1u)
1050     //     )
1051     //    )
1052     //   )
1053     //  )
1054     //  ( ... drop data segments ... )
1055     // )
1056     //
1057     // When we are building with PIC, calculate the flag location using:
1058     //
1059     //    (global.get $__memory_base)
1060     //    (i32.const $__init_memory_flag)
1061     //    (i32.const 1)
1062 
1063     // With PIC code we cache the flag address in local 0
1064     if (config->isPic) {
1065       writeUleb128(os, 1, "num local decls");
1066       writeUleb128(os, 1, "local count");
1067       writeU8(os, is64 ? WASM_TYPE_I64 : WASM_TYPE_I32, "address type");
1068       writeU8(os, WASM_OPCODE_GLOBAL_GET, "GLOBAL_GET");
1069       writeUleb128(os, WasmSym::memoryBase->getGlobalIndex(), "memory_base");
1070       writePtrConst(os, flagAddress, is64, "flag address");
1071       writeU8(os, is64 ? WASM_OPCODE_I64_ADD : WASM_OPCODE_I32_ADD, "add");
1072       writeU8(os, WASM_OPCODE_LOCAL_SET, "local.set");
1073       writeUleb128(os, 0, "local 0");
1074     } else {
1075       writeUleb128(os, 0, "num locals");
1076     }
1077 
1078     auto writeGetFlagAddress = [&]() {
1079       if (config->isPic) {
1080         writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get");
1081         writeUleb128(os, 0, "local 0");
1082       } else {
1083         writePtrConst(os, flagAddress, is64, "flag address");
1084       }
1085     };
1086 
1087     // Atomically check whether this is the main thread.
1088     writeGetFlagAddress();
1089     writeI32Const(os, 0, "expected flag value");
1090     writeI32Const(os, 1, "flag value");
1091     writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix");
1092     writeUleb128(os, WASM_OPCODE_I32_RMW_CMPXCHG, "i32.atomic.rmw.cmpxchg");
1093     writeMemArg(os, 2, 0);
1094     writeU8(os, WASM_OPCODE_IF, "IF");
1095     writeU8(os, WASM_TYPE_NORESULT, "blocktype");
1096 
1097     // Did not increment 0, so wait for main thread to initialize memory
1098     writeGetFlagAddress();
1099     writeI32Const(os, 1, "expected flag value");
1100     writeI64Const(os, -1, "timeout");
1101 
1102     writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix");
1103     writeUleb128(os, WASM_OPCODE_I32_ATOMIC_WAIT, "i32.atomic.wait");
1104     writeMemArg(os, 2, 0);
1105     writeU8(os, WASM_OPCODE_DROP, "drop");
1106 
1107     writeU8(os, WASM_OPCODE_ELSE, "ELSE");
1108 
1109     // Did increment 0, so conditionally initialize passive data segments
1110     for (const OutputSegment *s : segments) {
1111       if (needsPassiveInitialization(s)) {
1112         // destination address
1113         writePtrConst(os, s->startVA, is64, "destination address");
1114         if (config->isPic) {
1115           writeU8(os, WASM_OPCODE_GLOBAL_GET, "GLOBAL_GET");
1116           writeUleb128(os, WasmSym::memoryBase->getGlobalIndex(),
1117                        "memory_base");
1118           writeU8(os, is64 ? WASM_OPCODE_I64_ADD : WASM_OPCODE_I32_ADD,
1119                   "i32.add");
1120         }
1121         // source segment offset
1122         writeI32Const(os, 0, "segment offset");
1123         // memory region size
1124         writeI32Const(os, s->size, "memory region size");
1125         // memory.init instruction
1126         writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix");
1127         writeUleb128(os, WASM_OPCODE_MEMORY_INIT, "memory.init");
1128         writeUleb128(os, s->index, "segment index immediate");
1129         writeU8(os, 0, "memory index immediate");
1130       }
1131     }
1132 
1133     // Set flag to 2 to mark end of initialization
1134     writeGetFlagAddress();
1135     writeI32Const(os, 2, "flag value");
1136     writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix");
1137     writeUleb128(os, WASM_OPCODE_I32_ATOMIC_STORE, "i32.atomic.store");
1138     writeMemArg(os, 2, 0);
1139 
1140     // Notify any waiters that memory initialization is complete
1141     writeGetFlagAddress();
1142     writeI32Const(os, -1, "number of waiters");
1143     writeU8(os, WASM_OPCODE_ATOMICS_PREFIX, "atomics prefix");
1144     writeUleb128(os, WASM_OPCODE_ATOMIC_NOTIFY, "atomic.notify");
1145     writeMemArg(os, 2, 0);
1146     writeU8(os, WASM_OPCODE_DROP, "drop");
1147 
1148     writeU8(os, WASM_OPCODE_END, "END");
1149 
1150     // Unconditionally drop passive data segments
1151     for (const OutputSegment *s : segments) {
1152       if (needsPassiveInitialization(s)) {
1153         // data.drop instruction
1154         writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix");
1155         writeUleb128(os, WASM_OPCODE_DATA_DROP, "data.drop");
1156         writeUleb128(os, s->index, "segment index immediate");
1157       }
1158     }
1159     writeU8(os, WASM_OPCODE_END, "END");
1160   }
1161 
1162   createFunction(WasmSym::initMemory, bodyContent);
1163 }
1164 
1165 void Writer::createStartFunction() {
1166   if (WasmSym::startFunction) {
1167     std::string bodyContent;
1168     {
1169       raw_string_ostream os(bodyContent);
1170       writeUleb128(os, 0, "num locals");
1171       writeU8(os, WASM_OPCODE_CALL, "CALL");
1172       writeUleb128(os, WasmSym::initMemory->getFunctionIndex(),
1173                    "function index");
1174       writeU8(os, WASM_OPCODE_CALL, "CALL");
1175       writeUleb128(os, WasmSym::applyGlobalRelocs->getFunctionIndex(),
1176                    "function index");
1177       writeU8(os, WASM_OPCODE_END, "END");
1178     }
1179     createFunction(WasmSym::startFunction, bodyContent);
1180   } else if (WasmSym::initMemory) {
1181     WasmSym::startFunction = WasmSym::initMemory;
1182   } else if (WasmSym::applyGlobalRelocs) {
1183     WasmSym::startFunction = WasmSym::applyGlobalRelocs;
1184   }
1185 }
1186 
1187 // For -shared (PIC) output, we create create a synthetic function which will
1188 // apply any relocations to the data segments on startup.  This function is
1189 // called `__wasm_apply_data_relocs` and is added at the beginning of
1190 // `__wasm_call_ctors` before any of the constructors run.
1191 void Writer::createApplyDataRelocationsFunction() {
1192   LLVM_DEBUG(dbgs() << "createApplyDataRelocationsFunction\n");
1193   // First write the body's contents to a string.
1194   std::string bodyContent;
1195   {
1196     raw_string_ostream os(bodyContent);
1197     writeUleb128(os, 0, "num locals");
1198     for (const OutputSegment *seg : segments)
1199       for (const InputChunk *inSeg : seg->inputSegments)
1200         inSeg->generateRelocationCode(os);
1201 
1202     writeU8(os, WASM_OPCODE_END, "END");
1203   }
1204 
1205   createFunction(WasmSym::applyDataRelocs, bodyContent);
1206 }
1207 
1208 // Similar to createApplyDataRelocationsFunction but generates relocation code
1209 // fro WebAssembly globals. Because these globals are not shared between threads
1210 // these relocation need to run on every thread.
1211 void Writer::createApplyGlobalRelocationsFunction() {
1212   // First write the body's contents to a string.
1213   std::string bodyContent;
1214   {
1215     raw_string_ostream os(bodyContent);
1216     writeUleb128(os, 0, "num locals");
1217     out.globalSec->generateRelocationCode(os);
1218     writeU8(os, WASM_OPCODE_END, "END");
1219   }
1220 
1221   createFunction(WasmSym::applyGlobalRelocs, bodyContent);
1222 }
1223 
1224 // Create synthetic "__wasm_call_ctors" function based on ctor functions
1225 // in input object.
1226 void Writer::createCallCtorsFunction() {
1227   // If __wasm_call_ctors isn't referenced, there aren't any ctors, and we
1228   // aren't calling `__wasm_apply_data_relocs` for Emscripten-style PIC, don't
1229   // define the `__wasm_call_ctors` function.
1230   if (!WasmSym::callCtors->isLive() && !WasmSym::applyDataRelocs &&
1231       initFunctions.empty())
1232     return;
1233 
1234   // First write the body's contents to a string.
1235   std::string bodyContent;
1236   {
1237     raw_string_ostream os(bodyContent);
1238     writeUleb128(os, 0, "num locals");
1239 
1240     if (WasmSym::applyDataRelocs) {
1241       writeU8(os, WASM_OPCODE_CALL, "CALL");
1242       writeUleb128(os, WasmSym::applyDataRelocs->getFunctionIndex(),
1243                    "function index");
1244     }
1245 
1246     // Call constructors
1247     for (const WasmInitEntry &f : initFunctions) {
1248       writeU8(os, WASM_OPCODE_CALL, "CALL");
1249       writeUleb128(os, f.sym->getFunctionIndex(), "function index");
1250       for (size_t i = 0; i < f.sym->signature->Returns.size(); i++) {
1251         writeU8(os, WASM_OPCODE_DROP, "DROP");
1252       }
1253     }
1254 
1255     writeU8(os, WASM_OPCODE_END, "END");
1256   }
1257 
1258   createFunction(WasmSym::callCtors, bodyContent);
1259 }
1260 
1261 // Create a wrapper around a function export which calls the
1262 // static constructors and destructors.
1263 void Writer::createCommandExportWrapper(uint32_t functionIndex,
1264                                         DefinedFunction *f) {
1265   // First write the body's contents to a string.
1266   std::string bodyContent;
1267   {
1268     raw_string_ostream os(bodyContent);
1269     writeUleb128(os, 0, "num locals");
1270 
1271     // Call `__wasm_call_ctors` which call static constructors (and
1272     // applies any runtime relocations in Emscripten-style PIC mode)
1273     if (WasmSym::callCtors->isLive()) {
1274       writeU8(os, WASM_OPCODE_CALL, "CALL");
1275       writeUleb128(os, WasmSym::callCtors->getFunctionIndex(),
1276                    "function index");
1277     }
1278 
1279     // Call the user's code, leaving any return values on the operand stack.
1280     for (size_t i = 0; i < f->signature->Params.size(); ++i) {
1281       writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get");
1282       writeUleb128(os, i, "local index");
1283     }
1284     writeU8(os, WASM_OPCODE_CALL, "CALL");
1285     writeUleb128(os, functionIndex, "function index");
1286 
1287     // Call the function that calls the destructors.
1288     if (DefinedFunction *callDtors = WasmSym::callDtors) {
1289       writeU8(os, WASM_OPCODE_CALL, "CALL");
1290       writeUleb128(os, callDtors->getFunctionIndex(), "function index");
1291     }
1292 
1293     // End the function, returning the return values from the user's code.
1294     writeU8(os, WASM_OPCODE_END, "END");
1295   }
1296 
1297   createFunction(f, bodyContent);
1298 }
1299 
1300 void Writer::createInitTLSFunction() {
1301   std::string bodyContent;
1302   {
1303     raw_string_ostream os(bodyContent);
1304 
1305     OutputSegment *tlsSeg = nullptr;
1306     for (auto *seg : segments) {
1307       if (seg->name == ".tdata") {
1308         tlsSeg = seg;
1309         break;
1310       }
1311     }
1312 
1313     writeUleb128(os, 0, "num locals");
1314     if (tlsSeg) {
1315       writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get");
1316       writeUleb128(os, 0, "local index");
1317 
1318       writeU8(os, WASM_OPCODE_GLOBAL_SET, "global.set");
1319       writeUleb128(os, WasmSym::tlsBase->getGlobalIndex(), "global index");
1320 
1321       // FIXME(wvo): this local needs to be I64 in wasm64, or we need an extend op.
1322       writeU8(os, WASM_OPCODE_LOCAL_GET, "local.get");
1323       writeUleb128(os, 0, "local index");
1324 
1325       writeI32Const(os, 0, "segment offset");
1326 
1327       writeI32Const(os, tlsSeg->size, "memory region size");
1328 
1329       writeU8(os, WASM_OPCODE_MISC_PREFIX, "bulk-memory prefix");
1330       writeUleb128(os, WASM_OPCODE_MEMORY_INIT, "MEMORY.INIT");
1331       writeUleb128(os, tlsSeg->index, "segment index immediate");
1332       writeU8(os, 0, "memory index immediate");
1333     }
1334     writeU8(os, WASM_OPCODE_END, "end function");
1335   }
1336 
1337   createFunction(WasmSym::initTLS, bodyContent);
1338 }
1339 
1340 // Populate InitFunctions vector with init functions from all input objects.
1341 // This is then used either when creating the output linking section or to
1342 // synthesize the "__wasm_call_ctors" function.
1343 void Writer::calculateInitFunctions() {
1344   if (!config->relocatable && !WasmSym::callCtors->isLive())
1345     return;
1346 
1347   for (ObjFile *file : symtab->objectFiles) {
1348     const WasmLinkingData &l = file->getWasmObj()->linkingData();
1349     for (const WasmInitFunc &f : l.InitFunctions) {
1350       FunctionSymbol *sym = file->getFunctionSymbol(f.Symbol);
1351       // comdat exclusions can cause init functions be discarded.
1352       if (sym->isDiscarded() || !sym->isLive())
1353         continue;
1354       if (sym->signature->Params.size() != 0)
1355         error("constructor functions cannot take arguments: " + toString(*sym));
1356       LLVM_DEBUG(dbgs() << "initFunctions: " << toString(*sym) << "\n");
1357       initFunctions.emplace_back(WasmInitEntry{sym, f.Priority});
1358     }
1359   }
1360 
1361   // Sort in order of priority (lowest first) so that they are called
1362   // in the correct order.
1363   llvm::stable_sort(initFunctions,
1364                     [](const WasmInitEntry &l, const WasmInitEntry &r) {
1365                       return l.priority < r.priority;
1366                     });
1367 }
1368 
1369 void Writer::createSyntheticSections() {
1370   out.dylinkSec = make<DylinkSection>();
1371   out.typeSec = make<TypeSection>();
1372   out.importSec = make<ImportSection>();
1373   out.functionSec = make<FunctionSection>();
1374   out.tableSec = make<TableSection>();
1375   out.memorySec = make<MemorySection>();
1376   out.tagSec = make<TagSection>();
1377   out.globalSec = make<GlobalSection>();
1378   out.exportSec = make<ExportSection>();
1379   out.startSec = make<StartSection>();
1380   out.elemSec = make<ElemSection>();
1381   out.producersSec = make<ProducersSection>();
1382   out.targetFeaturesSec = make<TargetFeaturesSection>();
1383 }
1384 
1385 void Writer::createSyntheticSectionsPostLayout() {
1386   out.dataCountSec = make<DataCountSection>(segments);
1387   out.linkingSec = make<LinkingSection>(initFunctions, segments);
1388   out.nameSec = make<NameSection>(segments);
1389 }
1390 
1391 void Writer::run() {
1392   if (config->relocatable || config->isPic)
1393     config->globalBase = 0;
1394 
1395   // For PIC code the table base is assigned dynamically by the loader.
1396   // For non-PIC, we start at 1 so that accessing table index 0 always traps.
1397   if (!config->isPic) {
1398     config->tableBase = 1;
1399     if (WasmSym::definedTableBase)
1400       WasmSym::definedTableBase->setVA(config->tableBase);
1401     if (WasmSym::definedTableBase32)
1402       WasmSym::definedTableBase32->setVA(config->tableBase);
1403   }
1404 
1405   log("-- createOutputSegments");
1406   createOutputSegments();
1407   log("-- createSyntheticSections");
1408   createSyntheticSections();
1409   log("-- layoutMemory");
1410   layoutMemory();
1411 
1412   if (!config->relocatable) {
1413     // Create linker synthesized __start_SECNAME/__stop_SECNAME symbols
1414     // This has to be done after memory layout is performed.
1415     for (const OutputSegment *seg : segments) {
1416       addStartStopSymbols(seg);
1417     }
1418   }
1419 
1420   for (auto &pair : config->exportedSymbols) {
1421     Symbol *sym = symtab->find(pair.first());
1422     if (sym && sym->isDefined())
1423       sym->forceExport = true;
1424   }
1425 
1426   // Delay reporting error about explict exports until after addStartStopSymbols
1427   // which can create optional symbols.
1428   for (auto &name : config->requiredExports) {
1429     Symbol *sym = symtab->find(name);
1430     if (!sym || !sym->isDefined()) {
1431       if (config->unresolvedSymbols == UnresolvedPolicy::ReportError)
1432         error(Twine("symbol exported via --export not found: ") + name);
1433       if (config->unresolvedSymbols == UnresolvedPolicy::Warn)
1434         warn(Twine("symbol exported via --export not found: ") + name);
1435     }
1436   }
1437 
1438   if (config->isPic && !config->sharedMemory) {
1439     // In shared memory mode all data segments are passive and initilized
1440     // via __wasm_init_memory.
1441     log("-- combineOutputSegments");
1442     combineOutputSegments();
1443   }
1444 
1445   log("-- createSyntheticSectionsPostLayout");
1446   createSyntheticSectionsPostLayout();
1447   log("-- populateProducers");
1448   populateProducers();
1449   log("-- calculateImports");
1450   calculateImports();
1451   log("-- scanRelocations");
1452   scanRelocations();
1453   log("-- finalizeIndirectFunctionTable");
1454   finalizeIndirectFunctionTable();
1455   log("-- createSyntheticInitFunctions");
1456   createSyntheticInitFunctions();
1457   log("-- assignIndexes");
1458   assignIndexes();
1459   log("-- calculateInitFunctions");
1460   calculateInitFunctions();
1461 
1462   if (!config->relocatable) {
1463     // Create linker synthesized functions
1464     if (WasmSym::applyDataRelocs)
1465       createApplyDataRelocationsFunction();
1466     if (WasmSym::applyGlobalRelocs)
1467       createApplyGlobalRelocationsFunction();
1468     if (WasmSym::initMemory)
1469       createInitMemoryFunction();
1470     createStartFunction();
1471 
1472     createCallCtorsFunction();
1473 
1474     // Create export wrappers for commands if needed.
1475     //
1476     // If the input contains a call to `__wasm_call_ctors`, either in one of
1477     // the input objects or an explicit export from the command-line, we
1478     // assume ctors and dtors are taken care of already.
1479     if (!config->relocatable && !config->isPic &&
1480         !WasmSym::callCtors->isUsedInRegularObj &&
1481         !WasmSym::callCtors->isExported()) {
1482       log("-- createCommandExportWrappers");
1483       createCommandExportWrappers();
1484     }
1485   }
1486 
1487   if (WasmSym::initTLS && WasmSym::initTLS->isLive())
1488     createInitTLSFunction();
1489 
1490   if (errorCount())
1491     return;
1492 
1493   log("-- calculateTypes");
1494   calculateTypes();
1495   log("-- calculateExports");
1496   calculateExports();
1497   log("-- calculateCustomSections");
1498   calculateCustomSections();
1499   log("-- populateSymtab");
1500   populateSymtab();
1501   log("-- populateTargetFeatures");
1502   populateTargetFeatures();
1503   log("-- addSections");
1504   addSections();
1505 
1506   if (errorHandler().verbose) {
1507     log("Defined Functions: " + Twine(out.functionSec->inputFunctions.size()));
1508     log("Defined Globals  : " + Twine(out.globalSec->numGlobals()));
1509     log("Defined Tags     : " + Twine(out.tagSec->inputTags.size()));
1510     log("Defined Tables   : " + Twine(out.tableSec->inputTables.size()));
1511     log("Function Imports : " +
1512         Twine(out.importSec->getNumImportedFunctions()));
1513     log("Global Imports   : " + Twine(out.importSec->getNumImportedGlobals()));
1514     log("Tag Imports      : " + Twine(out.importSec->getNumImportedTags()));
1515     log("Table Imports    : " + Twine(out.importSec->getNumImportedTables()));
1516     for (ObjFile *file : symtab->objectFiles)
1517       file->dumpInfo();
1518   }
1519 
1520   createHeader();
1521   log("-- finalizeSections");
1522   finalizeSections();
1523 
1524   log("-- writeMapFile");
1525   writeMapFile(outputSections);
1526 
1527   log("-- openFile");
1528   openFile();
1529   if (errorCount())
1530     return;
1531 
1532   writeHeader();
1533 
1534   log("-- writeSections");
1535   writeSections();
1536   if (errorCount())
1537     return;
1538 
1539   if (Error e = buffer->commit())
1540     fatal("failed to write the output file: " + toString(std::move(e)));
1541 }
1542 
1543 // Open a result file.
1544 void Writer::openFile() {
1545   log("writing: " + config->outputFile);
1546 
1547   Expected<std::unique_ptr<FileOutputBuffer>> bufferOrErr =
1548       FileOutputBuffer::create(config->outputFile, fileSize,
1549                                FileOutputBuffer::F_executable);
1550 
1551   if (!bufferOrErr)
1552     error("failed to open " + config->outputFile + ": " +
1553           toString(bufferOrErr.takeError()));
1554   else
1555     buffer = std::move(*bufferOrErr);
1556 }
1557 
1558 void Writer::createHeader() {
1559   raw_string_ostream os(header);
1560   writeBytes(os, WasmMagic, sizeof(WasmMagic), "wasm magic");
1561   writeU32(os, WasmVersion, "wasm version");
1562   os.flush();
1563   fileSize += header.size();
1564 }
1565 
1566 void writeResult() { Writer().run(); }
1567 
1568 } // namespace wasm
1569 } // namespace lld
1570