]> git.saurik.com Git - apple/ld64.git/blame - src/ld/parsers/lto_file.cpp
ld64-409.12.tar.gz
[apple/ld64.git] / src / ld / parsers / lto_file.cpp
CommitLineData
a645023d
A
1/* -*- mode: C++; c-basic-offset: 4; tab-width: 4 -*-
2 *
3 * Copyright (c) 2006-2010 Apple Inc. All rights reserved.
4 *
5 * @APPLE_LICENSE_HEADER_START@
6 *
7 * This file contains Original Code and/or Modifications of Original Code
8 * as defined in and that are subject to the Apple Public Source License
9 * Version 2.0 (the 'License'). You may not use this file except in
10 * compliance with the License. Please obtain a copy of the License at
11 * http://www.opensource.apple.com/apsl/ and read it before using this
12 * file.
13 *
14 * The Original Code and all software distributed under the License are
15 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
16 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
17 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
19 * Please see the License for the specific language governing rights and
20 * limitations under the License.
21 *
22 * @APPLE_LICENSE_HEADER_END@
23 */
24
25#ifndef __LTO_READER_H__
26#define __LTO_READER_H__
27
28#include <stdlib.h>
29#include <sys/param.h>
30#include <sys/fcntl.h>
82b4b32b 31#include <sys/mman.h>
a645023d
A
32#include <sys/stat.h>
33#include <errno.h>
d425e388 34#include <pthread.h>
a645023d
A
35#include <mach-o/dyld.h>
36#include <vector>
eaf282aa 37#include <map>
d425e388
A
38#include <unordered_set>
39#include <unordered_map>
0a8dc3df
A
40#include <iostream>
41#include <fstream>
a645023d
A
42
43#include "MachOFileAbstraction.hpp"
44#include "Architectures.hpp"
45#include "ld.hpp"
46#include "macho_relocatable_file.h"
47#include "lto_file.h"
48
49#include "llvm-c/lto.h"
50
a645023d
A
51namespace lto {
52
53
54//
55// ld64 only tracks non-internal symbols from an llvm bitcode file.
56// We model this by having an InternalAtom which represent all internal functions and data.
57// All non-interal symbols from a bitcode file are represented by an Atom
58// and each Atom has a reference to the InternalAtom. The InternalAtom
59// also has references to each symbol external to the bitcode file.
60//
61class InternalAtom : public ld::Atom
62{
63public:
64 InternalAtom(class File& f);
65 // overrides of ld::Atom
0a8dc3df
A
66 ld::File* file() const override { return &_file; }
67 const char* name() const override { return "import-atom"; }
68 uint64_t size() const override { return 0; }
69 uint64_t objectAddress() const override { return 0; }
70 void copyRawContent(uint8_t buffer[]) const override { }
71 ld::Fixup::iterator fixupsBegin() const override { return &_undefs[0]; }
72 ld::Fixup::iterator fixupsEnd() const override { return &_undefs[_undefs.size()]; }
a645023d
A
73
74 // for adding references to symbols outside bitcode file
75 void addReference(const char* nm)
76 { _undefs.push_back(ld::Fixup(0, ld::Fixup::k1of1,
599556ff 77 ld::Fixup::kindNone, false, strdup(nm))); }
a645023d
A
78private:
79
80 ld::File& _file;
81 mutable std::vector<ld::Fixup> _undefs;
82};
83
84
85//
86// LLVM bitcode file
87//
88class File : public ld::relocatable::File
89{
90public:
b1f7435d
A
91 File(const char* path, time_t mTime, ld::File::Ordinal ordinal,
92 const uint8_t* content, uint32_t contentLength, cpu_type_t arch);
0a8dc3df 93 ~File() override;
a645023d
A
94
95 // overrides of ld::File
0a8dc3df
A
96 bool forEachAtom(ld::File::AtomHandler&) const override;
97 bool justInTimeforEachAtom(const char* name, ld::File::AtomHandler&) const override
a645023d 98 { return false; }
0a8dc3df 99 uint32_t cpuSubType() const override { return _cpuSubType; }
a645023d
A
100
101 // overrides of ld::relocatable::File
0a8dc3df
A
102 DebugInfoKind debugInfo() const override { return _debugInfo; }
103 const char* debugInfoPath() const override { return _debugInfoPath; }
104 time_t debugInfoModificationTime() const override
a645023d 105 { return _debugInfoModTime; }
0a8dc3df
A
106 const std::vector<ld::relocatable::File::Stab>* stabs() const override { return NULL; }
107 bool canScatterAtoms() const override { return true; }
108 LinkerOptionsList* linkerOptions() const override { return NULL; }
bee7e226 109 const ToolVersionList& toolVersions() const override { return _toolVersions; }
0a8dc3df
A
110 bool isThinLTO() const { return _isThinLTO; }
111 void setIsThinLTO(bool ThinLTO) { _isThinLTO = ThinLTO; }
112 // fixme rdar://24734472 objCConstraint() and objcHasCategoryClassProperties()
599556ff 113 void release();
a645023d
A
114 lto_module_t module() { return _module; }
115 class InternalAtom& internalAtom() { return _internalAtom; }
116 void setDebugInfo(ld::relocatable::File::DebugInfoKind k,
117 const char* pth, time_t modTime, uint32_t subtype)
118 { _debugInfo = k;
119 _debugInfoPath = pth;
120 _debugInfoModTime = modTime;
121 _cpuSubType = subtype;}
122
eaf282aa
A
123 static bool sSupportsLocalContext;
124 static bool sHasTriedLocalContext;
125 bool mergeIntoGenerator(lto_code_gen_t generator, bool useSetModule);
0a8dc3df 126#if LTO_API_VERSION >= 18
82b4b32b 127 void addToThinGenerator(thinlto_code_gen_t generator, int id);
0a8dc3df 128#endif
a645023d
A
129private:
130 friend class Atom;
131 friend class InternalAtom;
132 friend class Parser;
0a8dc3df
A
133
134 bool _isThinLTO;
a645023d
A
135 cpu_type_t _architecture;
136 class InternalAtom _internalAtom;
137 class Atom* _atomArray;
138 uint32_t _atomArrayCount;
139 lto_module_t _module;
eaf282aa
A
140 const char* _path;
141 const uint8_t* _content;
142 uint32_t _contentLength;
a645023d
A
143 const char* _debugInfoPath;
144 time_t _debugInfoModTime;
145 ld::Section _section;
146 ld::Fixup _fixupToInternal;
147 ld::relocatable::File::DebugInfoKind _debugInfo;
148 uint32_t _cpuSubType;
bee7e226 149 ToolVersionList _toolVersions; // unused, may some day contain version of clang the created bitcode
a645023d
A
150};
151
152//
153// Atom acts as a proxy Atom for the symbols that are exported by LLVM bitcode file. Initially,
154// Reader creates Atoms to allow linker proceed with usual symbol resolution phase. After
155// optimization is performed, real Atoms are created for these symobls. However these real Atoms
156// are not inserted into global symbol table. Atom holds real Atom and forwards appropriate
157// methods to real atom.
158//
159class Atom : public ld::Atom
160{
161public:
162 Atom(File& f, const char* name, ld::Atom::Scope s,
b2fa67a8 163 ld::Atom::Definition d, ld::Atom::Combine c, ld::Atom::Alignment a, bool ah);
a645023d
A
164
165 // overrides of ld::Atom
0a8dc3df
A
166 const ld::File* file() const override { return (_compiledAtom ? _compiledAtom->file() : &_file ); }
167 const ld::File* originalFile() const override { return &_file; }
168 const char* translationUnitSource() const override
b1f7435d 169 { return (_compiledAtom ? _compiledAtom->translationUnitSource() : NULL); }
0a8dc3df
A
170 const char* name() const override { return _name; }
171 uint64_t size() const override { return (_compiledAtom ? _compiledAtom->size() : 0); }
172 uint64_t objectAddress() const override { return (_compiledAtom ? _compiledAtom->objectAddress() : 0); }
173 void copyRawContent(uint8_t buffer[]) const override
a645023d 174 { if (_compiledAtom) _compiledAtom->copyRawContent(buffer); }
0a8dc3df 175 const uint8_t* rawContentPointer() const override
a645023d 176 { return (_compiledAtom ? _compiledAtom->rawContentPointer() : NULL); }
0a8dc3df 177 unsigned long contentHash(const class ld::IndirectBindingTable& ibt) const override
a645023d 178 { return (_compiledAtom ? _compiledAtom->contentHash(ibt) : 0); }
0a8dc3df 179 bool canCoalesceWith(const ld::Atom& rhs, const class ld::IndirectBindingTable& ibt) const override
a645023d 180 { return (_compiledAtom ? _compiledAtom->canCoalesceWith(rhs,ibt) : false); }
0a8dc3df 181 ld::Fixup::iterator fixupsBegin() const override
a645023d 182 { return (_compiledAtom ? _compiledAtom->fixupsBegin() : (ld::Fixup*)&_file._fixupToInternal); }
0a8dc3df 183 ld::Fixup::iterator fixupsEnd() const override
a645023d 184 { return (_compiledAtom ? _compiledAtom->fixupsEnd() : &((ld::Fixup*)&_file._fixupToInternal)[1]); }
0a8dc3df 185 ld::Atom::UnwindInfo::iterator beginUnwind() const override
a645023d 186 { return (_compiledAtom ? _compiledAtom->beginUnwind() : NULL); }
0a8dc3df 187 ld::Atom::UnwindInfo::iterator endUnwind() const override
a645023d 188 { return (_compiledAtom ? _compiledAtom->endUnwind() : NULL); }
0a8dc3df 189 ld::Atom::LineInfo::iterator beginLineInfo() const override
a645023d 190 { return (_compiledAtom ? _compiledAtom->beginLineInfo() : NULL); }
0a8dc3df 191 ld::Atom::LineInfo::iterator endLineInfo() const override
a645023d
A
192 { return (_compiledAtom ? _compiledAtom->endLineInfo() : NULL); }
193
194 const ld::Atom* compiledAtom() { return _compiledAtom; }
195 void setCompiledAtom(const ld::Atom& atom);
196
197private:
198
199 File& _file;
200 const char* _name;
201 const ld::Atom* _compiledAtom;
202};
203
204
205
206
207
208
209
210class Parser
211{
212public:
213 static bool validFile(const uint8_t* fileContent, uint64_t fileLength, cpu_type_t architecture, cpu_subtype_t subarch);
214 static const char* fileKind(const uint8_t* fileContent, uint64_t fileLength);
215 static File* parse(const uint8_t* fileContent, uint64_t fileLength, const char* path,
9543cb2f
A
216 time_t modTime, ld::File::Ordinal ordinal, cpu_type_t architecture, cpu_subtype_t subarch,
217 bool logAllFiles, bool verboseOptimizationHints);
a645023d
A
218 static bool libLTOisLoaded() { return (::lto_get_version() != NULL); }
219 static bool optimize( const std::vector<const ld::Atom*>& allAtoms,
220 ld::Internal& state,
a645023d
A
221 const OptimizeOptions& options,
222 ld::File::AtomHandler& handler,
223 std::vector<const ld::Atom*>& newAtoms,
224 std::vector<const char*>& additionalUndefines);
225
0a8dc3df 226 static const char* ltoVersion() { return ::lto_get_version(); }
a645023d
A
227
228private:
0a8dc3df 229
a645023d 230 static const char* tripletPrefixForArch(cpu_type_t arch);
0a8dc3df
A
231 static ld::relocatable::File* parseMachOFile(const uint8_t* p, size_t len, const std::string &path, const OptimizeOptions& options,
232 ld::File::Ordinal ordinal);
9543cb2f
A
233#if LTO_API_VERSION >= 7
234 static void ltoDiagnosticHandler(lto_codegen_diagnostic_severity_t, const char*, void*);
235#endif
a645023d 236
d425e388
A
237 typedef std::unordered_set<const char*, ld::CStringHash, ld::CStringEquals> CStringSet;
238 typedef std::unordered_map<const char*, Atom*, ld::CStringHash, ld::CStringEquals> CStringToAtom;
a645023d
A
239
240 class AtomSyncer : public ld::File::AtomHandler {
241 public:
242 AtomSyncer(std::vector<const char*>& a, std::vector<const ld::Atom*>&na,
0a8dc3df
A
243 const CStringToAtom &la, const CStringToAtom &dla, const OptimizeOptions& options) :
244 _options(options), _additionalUndefines(a), _newAtoms(na), _llvmAtoms(la), _deadllvmAtoms(dla), _lastProxiedAtom(NULL), _lastProxiedFile(NULL) {}
245 void doAtom(const class ld::Atom&) override;
246 void doFile(const class ld::File&) override { }
a645023d
A
247
248 const OptimizeOptions& _options;
249 std::vector<const char*>& _additionalUndefines;
250 std::vector<const ld::Atom*>& _newAtoms;
0a8dc3df
A
251 const CStringToAtom &_llvmAtoms;
252 const CStringToAtom &_deadllvmAtoms;
253 const ld::Atom* _lastProxiedAtom;
254 const ld::File* _lastProxiedFile;
a645023d
A
255 };
256
0a8dc3df
A
257 static void setPreservedSymbols(const std::vector<const ld::Atom*>& allAtoms,
258 ld::Internal& state,
259 const OptimizeOptions& options,
260 CStringToAtom &deadllvmAtoms,
261 CStringToAtom &llvmAtoms,
262 lto_code_gen_t generator);
263
264 static std::tuple<uint8_t *, size_t> codegen(const OptimizeOptions& options,
265 ld::Internal& state,
266 lto_code_gen_t generator,
267 std::string& object_path);
268
269
270 static void loadMachO(ld::relocatable::File* machoFile,
271 const OptimizeOptions& options,
272 ld::File::AtomHandler& handler,
273 std::vector<const ld::Atom*>& newAtoms,
274 std::vector<const char*>& additionalUndefines,
275 CStringToAtom &llvmAtoms,
276 CStringToAtom &deadllvmAtoms);
277
278 static bool optimizeLTO(const std::vector<File*> files,
279 const std::vector<const ld::Atom*>& allAtoms,
280 ld::Internal& state,
281 const OptimizeOptions& options,
282 ld::File::AtomHandler& handler,
283 std::vector<const ld::Atom*>& newAtoms,
284 std::vector<const char*>& additionalUndefines);
285
286 static bool optimizeThinLTO(const std::vector<File*>& Files,
287 const std::vector<const ld::Atom*>& allAtoms,
288 ld::Internal& state,
289 const OptimizeOptions& options,
290 ld::File::AtomHandler& handler,
291 std::vector<const ld::Atom*>& newAtoms,
292 std::vector<const char*>& additionalUndefines);
293
82b4b32b 294#if LTO_API_VERSION >= 18
0a8dc3df
A
295 static thinlto_code_gen_t init_thinlto_codegen(const std::vector<File*>& files,
296 const std::vector<const ld::Atom*>& allAtoms,
297 ld::Internal& state,
298 const OptimizeOptions& options,
299 CStringToAtom& deadllvmAtoms,
300 CStringToAtom& llvmAtoms);
82b4b32b 301#endif
0a8dc3df 302
a645023d 303 static std::vector<File*> _s_files;
0a8dc3df 304 static bool _s_llvmOptionsProcessed;
a645023d
A
305};
306
307std::vector<File*> Parser::_s_files;
0a8dc3df 308bool Parser::_s_llvmOptionsProcessed = false;
a645023d
A
309
310
311bool Parser::validFile(const uint8_t* fileContent, uint64_t fileLength, cpu_type_t architecture, cpu_subtype_t subarch)
312{
ebf6f434
A
313 for (const ArchInfo* t=archInfoArray; t->archName != NULL; ++t) {
314 if ( (architecture == t->cpuType) && (!(t->isSubType) || (subarch == t->cpuSubType)) ) {
315 bool result = ::lto_module_is_object_file_in_memory_for_target(fileContent, fileLength, t->llvmTriplePrefix);
316 if ( !result ) {
317 // <rdar://problem/8434487> LTO only supports thumbv7 not armv7
318 if ( t->llvmTriplePrefixAlt[0] != '\0' ) {
319 result = ::lto_module_is_object_file_in_memory_for_target(fileContent, fileLength, t->llvmTriplePrefixAlt);
320 }
a645023d 321 }
ebf6f434
A
322 return result;
323 }
a645023d
A
324 }
325 return false;
326}
327
328const char* Parser::fileKind(const uint8_t* p, uint64_t fileLength)
329{
330 if ( (p[0] == 0xDE) && (p[1] == 0xC0) && (p[2] == 0x17) && (p[3] == 0x0B) ) {
ebf6f434
A
331 cpu_type_t arch = LittleEndian::get32(*((uint32_t*)(&p[16])));
332 for (const ArchInfo* t=archInfoArray; t->archName != NULL; ++t) {
333 if ( arch == t->cpuType ) {
334 if ( t->isSubType ) {
afe874b1 335 if ( ::lto_module_is_object_file_in_memory_for_target(p, fileLength, t->llvmTriplePrefix) )
ebf6f434
A
336 return t->archName;
337 }
338 else {
339 return t->archName;
afe874b1 340 }
ebf6f434 341 }
a645023d
A
342 }
343 return "unknown bitcode architecture";
344 }
345 return NULL;
346}
347
b1f7435d 348File* Parser::parse(const uint8_t* fileContent, uint64_t fileLength, const char* path, time_t modTime, ld::File::Ordinal ordinal,
9543cb2f 349 cpu_type_t architecture, cpu_subtype_t subarch, bool logAllFiles, bool verboseOptimizationHints)
a645023d 350{
b1f7435d 351 File* f = new File(path, modTime, ordinal, fileContent, fileLength, architecture);
a645023d
A
352 _s_files.push_back(f);
353 if ( logAllFiles )
354 printf("%s\n", path);
355 return f;
356}
357
358
0a8dc3df
A
359ld::relocatable::File* Parser::parseMachOFile(const uint8_t* p, size_t len, const std::string &path, const OptimizeOptions& options,
360 ld::File::Ordinal ordinal)
a645023d
A
361{
362 mach_o::relocatable::ParserOptions objOpts;
363 objOpts.architecture = options.arch;
364 objOpts.objSubtypeMustMatch = false;
365 objOpts.logAllFiles = false;
f80fe69f
A
366 objOpts.warnUnwindConversionProblems = options.needsUnwindInfoSection;
367 objOpts.keepDwarfUnwind = options.keepDwarfUnwind;
368 objOpts.forceDwarfConversion = false;
9543cb2f
A
369 objOpts.neverConvertDwarf = false;
370 objOpts.verboseOptimizationHints = options.verboseOptimizationHints;
ba348e21 371 objOpts.armUsesZeroCostExceptions = options.armUsesZeroCostExceptions;
eaf282aa
A
372 objOpts.simulator = options.simulator;
373 objOpts.ignoreMismatchPlatform = options.ignoreMismatchPlatform;
e456bf10
A
374#if SUPPORT_ARCH_arm64e
375 objOpts.supportsAuthenticatedPointers = options.supportsAuthenticatedPointers;
376#endif
377 objOpts.platforms = options.platforms;
a645023d 378 objOpts.subType = 0;
eaf282aa 379 objOpts.srcKind = ld::relocatable::File::kSourceLTO;
dd9e569f
A
380 objOpts.treateBitcodeAsData = false;
381 objOpts.usingBitcode = options.bitcodeBundle;
ec29ba20
A
382 objOpts.maxDefaultCommonAlignment = options.maxDefaultCommonAlignment;
383
0a8dc3df
A
384 const char *object_path = path.c_str();
385 if (path.empty())
386 object_path = "/tmp/lto.o";
387
a645023d 388 time_t modTime = 0;
0a8dc3df
A
389 struct stat statBuffer;
390 if ( stat(object_path, &statBuffer) == 0 )
391 modTime = statBuffer.st_mtime;
392
393 ld::relocatable::File* result = mach_o::relocatable::parse(p, len, strdup(object_path), modTime, ordinal, objOpts);
a645023d
A
394 if ( result != NULL )
395 return result;
396 throw "LLVM LTO, file is not of required architecture";
397}
398
399
400
b1f7435d 401File::File(const char* pth, time_t mTime, ld::File::Ordinal ordinal, const uint8_t* content, uint32_t contentLength, cpu_type_t arch)
0a8dc3df 402 : ld::relocatable::File(pth,mTime,ordinal), _isThinLTO(false), _architecture(arch), _internalAtom(*this),
eaf282aa
A
403 _atomArray(NULL), _atomArrayCount(0), _module(NULL), _path(pth),
404 _content(content), _contentLength(contentLength), _debugInfoPath(pth),
a645023d
A
405 _section("__TEXT_", "__tmp_lto", ld::Section::typeTempLTO),
406 _fixupToInternal(0, ld::Fixup::k1of1, ld::Fixup::kindNone, &_internalAtom),
407 _debugInfo(ld::relocatable::File::kDebugInfoNone), _cpuSubType(0)
408{
409 const bool log = false;
410
411 // create llvm module
eaf282aa
A
412#if LTO_API_VERSION >= 11
413 if ( sSupportsLocalContext || !sHasTriedLocalContext ) {
414 _module = ::lto_module_create_in_local_context(content, contentLength, pth);
415 }
416 if ( !sHasTriedLocalContext ) {
417 sHasTriedLocalContext = true;
418 sSupportsLocalContext = (_module != NULL);
419 }
420 if ( (_module == NULL) && !sSupportsLocalContext )
421#endif
599556ff
A
422#if LTO_API_VERSION >= 9
423 _module = ::lto_module_create_from_memory_with_path(content, contentLength, pth);
eaf282aa 424 if ( _module == NULL && !sSupportsLocalContext )
599556ff 425#endif
a645023d
A
426 _module = ::lto_module_create_from_memory(content, contentLength);
427 if ( _module == NULL )
d425e388 428 throwf("could not parse object file %s: '%s', using libLTO version '%s'", pth, ::lto_get_error_message(), ::lto_get_version());
a645023d
A
429
430 if ( log ) fprintf(stderr, "bitcode file: %s\n", pth);
0a8dc3df
A
431
432#if LTO_API_VERSION >= 18
433 _isThinLTO = ::lto_module_is_thinlto(_module);
434#endif
435
a645023d
A
436 // create atom for each global symbol in module
437 uint32_t count = ::lto_module_get_num_symbols(_module);
438 _atomArray = (Atom*)malloc(sizeof(Atom)*count);
439 for (uint32_t i=0; i < count; ++i) {
440 const char* name = ::lto_module_get_symbol_name(_module, i);
441 lto_symbol_attributes attr = lto_module_get_symbol_attribute(_module, i);
442
443 // <rdar://problem/6378110> LTO doesn't like dtrace symbols
444 // ignore dtrace static probes for now
445 // later when codegen is done and a mach-o file is produces the probes will be processed
446 if ( (strncmp(name, "___dtrace_probe$", 16) == 0) || (strncmp(name, "___dtrace_isenabled$", 20) == 0) )
447 continue;
448
449 ld::Atom::Definition def;
450 ld::Atom::Combine combine = ld::Atom::combineNever;
451 switch ( attr & LTO_SYMBOL_DEFINITION_MASK ) {
452 case LTO_SYMBOL_DEFINITION_REGULAR:
453 def = ld::Atom::definitionRegular;
454 break;
455 case LTO_SYMBOL_DEFINITION_TENTATIVE:
456 def = ld::Atom::definitionTentative;
457 break;
458 case LTO_SYMBOL_DEFINITION_WEAK:
459 def = ld::Atom::definitionRegular;
460 combine = ld::Atom::combineByName;
461 break;
462 case LTO_SYMBOL_DEFINITION_UNDEFINED:
463 case LTO_SYMBOL_DEFINITION_WEAKUNDEF:
464 def = ld::Atom::definitionProxy;
465 break;
466 default:
467 throwf("unknown definition kind for symbol %s in bitcode file %s", name, pth);
468 }
469
470 // make LLVM atoms for definitions and a reference for undefines
471 if ( def != ld::Atom::definitionProxy ) {
472 ld::Atom::Scope scope;
b2fa67a8 473 bool autohide = false;
a645023d
A
474 switch ( attr & LTO_SYMBOL_SCOPE_MASK) {
475 case LTO_SYMBOL_SCOPE_INTERNAL:
476 scope = ld::Atom::scopeTranslationUnit;
477 break;
478 case LTO_SYMBOL_SCOPE_HIDDEN:
479 scope = ld::Atom::scopeLinkageUnit;
480 break;
481 case LTO_SYMBOL_SCOPE_DEFAULT:
482 scope = ld::Atom::scopeGlobal;
483 break;
b2fa67a8
A
484#if LTO_API_VERSION >= 4
485 case LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN:
486 scope = ld::Atom::scopeGlobal;
487 autohide = true;
488 break;
489#endif
a645023d
A
490 default:
491 throwf("unknown scope for symbol %s in bitcode file %s", name, pth);
492 }
493 // only make atoms for non-internal symbols
494 if ( scope == ld::Atom::scopeTranslationUnit )
495 continue;
496 uint8_t alignment = (attr & LTO_SYMBOL_ALIGNMENT_MASK);
497 // make Atom using placement new operator
b2fa67a8 498 new (&_atomArray[_atomArrayCount++]) Atom(*this, name, scope, def, combine, alignment, autohide);
ebf6f434 499 if ( scope != ld::Atom::scopeTranslationUnit )
a645023d
A
500 _internalAtom.addReference(name);
501 if ( log ) fprintf(stderr, "\t0x%08X %s\n", attr, name);
502 }
503 else {
504 // add to list of external references
505 _internalAtom.addReference(name);
506 if ( log ) fprintf(stderr, "\t%s (undefined)\n", name);
507 }
508 }
eaf282aa
A
509
510#if LTO_API_VERSION >= 11
511 if ( sSupportsLocalContext )
512 this->release();
513#endif
a645023d
A
514}
515
516File::~File()
599556ff
A
517{
518 this->release();
519}
520
eaf282aa
A
521bool File::mergeIntoGenerator(lto_code_gen_t generator, bool useSetModule) {
522#if LTO_API_VERSION >= 11
523 if ( sSupportsLocalContext ) {
524 assert(!_module && "Expected module to be disposed");
525 _module = ::lto_module_create_in_codegen_context(_content, _contentLength,
526 _path, generator);
527 if ( _module == NULL )
528 throwf("could not reparse object file %s: '%s', using libLTO version '%s'",
529 _path, ::lto_get_error_message(), ::lto_get_version());
530 }
531#endif
532 assert(_module && "Expected module to stick around");
533#if LTO_API_VERSION >= 13
534 if (useSetModule) {
535 // lto_codegen_set_module will transfer ownership of the module to LTO code generator,
536 // so we don't need to release the module here.
537 ::lto_codegen_set_module(generator, _module);
538 return false;
539 }
540#endif
541 if ( ::lto_codegen_add_module(generator, _module) )
542 return true;
543
544 // <rdar://problem/15471128> linker should release module as soon as possible
545 this->release();
546 return false;
547}
548
0a8dc3df 549#if LTO_API_VERSION >= 18
82b4b32b 550void File::addToThinGenerator(thinlto_code_gen_t generator, int id) {
0a8dc3df 551 assert(!_module && "Expected module to be disposed");
82b4b32b
A
552 std::string pathWithID = _path;
553 pathWithID += std::to_string(id);
bee7e226 554 ::thinlto_codegen_add_module(generator, strdup(pathWithID.c_str()), (const char *)_content, _contentLength);
0a8dc3df
A
555}
556#endif
557
599556ff 558void File::release()
a645023d
A
559{
560 if ( _module != NULL )
561 ::lto_module_dispose(_module);
599556ff 562 _module = NULL;
a645023d
A
563}
564
565bool File::forEachAtom(ld::File::AtomHandler& handler) const
566{
567 handler.doAtom(_internalAtom);
568 for(uint32_t i=0; i < _atomArrayCount; ++i) {
569 handler.doAtom(_atomArray[i]);
570 }
571 return true;
572}
573
574InternalAtom::InternalAtom(File& f)
575 : ld::Atom(f._section, ld::Atom::definitionRegular, ld::Atom::combineNever, ld::Atom::scopeTranslationUnit,
576 ld::Atom::typeLTOtemporary, ld::Atom::symbolTableNotIn, true, false, false, ld::Atom::Alignment(0)),
577 _file(f)
578{
579}
580
b2fa67a8
A
581Atom::Atom(File& f, const char* nm, ld::Atom::Scope s, ld::Atom::Definition d, ld::Atom::Combine c,
582 ld::Atom::Alignment a, bool ah)
a645023d
A
583 : ld::Atom(f._section, d, c, s, ld::Atom::typeLTOtemporary,
584 ld::Atom::symbolTableIn, false, false, false, a),
599556ff 585 _file(f), _name(strdup(nm)), _compiledAtom(NULL)
a645023d 586{
b2fa67a8
A
587 if ( ah )
588 this->setAutoHide();
a645023d
A
589}
590
591void Atom::setCompiledAtom(const ld::Atom& atom)
592{
593 // set delegate so virtual methods go to it
594 _compiledAtom = &atom;
595
596 //fprintf(stderr, "setting lto atom %p to delegate to mach-o atom %p (%s)\n", this, &atom, atom.name());
597
598 // update fields in ld::Atom to match newly constructed mach-o atom
599 (const_cast<Atom*>(this))->setAttributesFromAtom(atom);
600}
601
602
603
f80fe69f
A
604// <rdar://problem/12379604> The order that files are merged must match command line order
605struct CommandLineOrderFileSorter
606{
607 bool operator()(File* left, File* right)
608 {
609 return ( left->ordinal() < right->ordinal() );
610 }
611};
612
613
9543cb2f
A
614#if LTO_API_VERSION >= 7
615void Parser::ltoDiagnosticHandler(lto_codegen_diagnostic_severity_t severity, const char* message, void*)
616{
617 switch ( severity ) {
599556ff
A
618#if LTO_API_VERSION >= 10
619 case LTO_DS_REMARK:
eaf282aa
A
620 fprintf(stderr, "ld: LTO remark: %s\n", message);
621 break;
599556ff 622#endif
9543cb2f
A
623 case LTO_DS_NOTE:
624 case LTO_DS_WARNING:
625 warning("%s", message);
626 break;
627 case LTO_DS_ERROR:
628 throwf("%s", message);
629 }
630}
631#endif
632
a645023d 633
0a8dc3df
A
634/// Instruct libLTO about the list of symbols to preserve, compute deadllvmAtoms and llvmAtoms
635void Parser::setPreservedSymbols( const std::vector<const ld::Atom*>& allAtoms,
636 ld::Internal& state,
637 const OptimizeOptions& options,
638 CStringToAtom &deadllvmAtoms,
639 CStringToAtom &llvmAtoms,
640 lto_code_gen_t generator) {
641 const bool logMustPreserve = false;
f80fe69f 642
0a8dc3df
A
643 // The atom graph uses directed edges (references). Collect all references where
644 // originating atom is not part of any LTO Reader. This allows optimizer to optimize an
645 // external (i.e. not originated from same .o file) reference if all originating atoms are also
a645023d
A
646 // defined in llvm bitcode file.
647 CStringSet nonLLVMRefs;
0a8dc3df 648 bool hasNonllvmAtoms = false;
a645023d
A
649 for (std::vector<const ld::Atom*>::const_iterator it = allAtoms.begin(); it != allAtoms.end(); ++it) {
650 const ld::Atom* atom = *it;
0a8dc3df
A
651 // only look at references that come from an atom that is not an LTO atom
652 if (atom->contentType() != ld::Atom::typeLTOtemporary ||
653 ((lto::File *)atom->file())->isThinLTO()) {
a645023d
A
654 if ( (atom->section().type() != ld::Section::typeMachHeader) && (atom->definition() != ld::Atom::definitionProxy) ) {
655 hasNonllvmAtoms = true;
656 }
657 const ld::Atom* target;
658 for (ld::Fixup::iterator fit=atom->fixupsBegin(); fit != atom->fixupsEnd(); ++fit) {
659 switch ( fit->binding ) {
660 case ld::Fixup::bindingDirectlyBound:
661 // that reference an llvm atom
0a8dc3df 662 if ( fit->u.target->contentType() == ld::Atom::typeLTOtemporary )
a645023d
A
663 nonLLVMRefs.insert(fit->u.target->name());
664 break;
665 case ld::Fixup::bindingsIndirectlyBound:
666 target = state.indirectBindingTable[fit->u.bindingIndex];
9543cb2f 667 if ( (target != NULL) && (target->contentType() == ld::Atom::typeLTOtemporary) )
a645023d
A
668 nonLLVMRefs.insert(target->name());
669 default:
670 break;
671 }
672 }
673 }
599556ff 674 else if ( atom->scope() >= ld::Atom::scopeLinkageUnit ) {
a645023d
A
675 llvmAtoms[atom->name()] = (Atom*)atom;
676 }
677 }
678 // if entry point is in a llvm bitcode file, it must be preserved by LTO
679 if ( state.entryPoint!= NULL ) {
0a8dc3df 680 if ( state.entryPoint->contentType() == ld::Atom::typeLTOtemporary )
a645023d
A
681 nonLLVMRefs.insert(state.entryPoint->name());
682 }
0a8dc3df 683
a645023d
A
684 // deadAtoms are the atoms that the linker coalesced. For instance weak or tentative definitions
685 // overriden by another atom. If any of these deadAtoms are llvm atoms and they were replaced
0a8dc3df 686 // with a mach-o atom, we need to tell the lto engine to preserve (not optimize away) its dead
a645023d 687 // atom so that the linker can replace it with the mach-o one later.
a645023d
A
688 for (std::vector<const ld::Atom*>::const_iterator it = allAtoms.begin(); it != allAtoms.end(); ++it) {
689 const ld::Atom* atom = *it;
690 if ( atom->coalescedAway() && (atom->contentType() == ld::Atom::typeLTOtemporary) ) {
691 const char* name = atom->name();
692 if ( logMustPreserve ) fprintf(stderr, "lto_codegen_add_must_preserve_symbol(%s) because linker coalesce away and replace with a mach-o atom\n", name);
693 ::lto_codegen_add_must_preserve_symbol(generator, name);
694 deadllvmAtoms[name] = (Atom*)atom;
695 }
696 }
697 for (std::vector<File*>::iterator it=_s_files.begin(); it != _s_files.end(); ++it) {
698 File* file = *it;
699 for(uint32_t i=0; i < file->_atomArrayCount; ++i) {
700 Atom* llvmAtom = &file->_atomArray[i];
701 if ( llvmAtom->coalescedAway() ) {
702 const char* name = llvmAtom->name();
703 if ( deadllvmAtoms.find(name) == deadllvmAtoms.end() ) {
0a8dc3df 704 if ( logMustPreserve )
a645023d
A
705 fprintf(stderr, "lto_codegen_add_must_preserve_symbol(%s) because linker coalesce away and replace with a mach-o atom\n", name);
706 ::lto_codegen_add_must_preserve_symbol(generator, name);
707 deadllvmAtoms[name] = (Atom*)llvmAtom;
708 }
709 }
710 else if ( options.linkerDeadStripping && !llvmAtom->live() ) {
711 const char* name = llvmAtom->name();
712 deadllvmAtoms[name] = (Atom*)llvmAtom;
713 }
714 }
715 }
0a8dc3df 716
a645023d
A
717 // tell code generator about symbols that must be preserved
718 for (CStringToAtom::iterator it = llvmAtoms.begin(); it != llvmAtoms.end(); ++it) {
719 const char* name = it->first;
720 Atom* atom = it->second;
721 // Include llvm Symbol in export list if it meets one of following two conditions
722 // 1 - atom scope is global (and not linkage unit).
723 // 2 - included in nonLLVMRefs set.
724 // If a symbol is not listed in exportList then LTO is free to optimize it away.
0a8dc3df 725 if ( (atom->scope() == ld::Atom::scopeGlobal) && options.preserveAllGlobals ) {
a645023d
A
726 if ( logMustPreserve ) fprintf(stderr, "lto_codegen_add_must_preserve_symbol(%s) because global symbol\n", name);
727 ::lto_codegen_add_must_preserve_symbol(generator, name);
728 }
729 else if ( nonLLVMRefs.find(name) != nonLLVMRefs.end() ) {
730 if ( logMustPreserve ) fprintf(stderr, "lto_codegen_add_must_preserve_symbol(%s) because referenced by a mach-o atom\n", name);
731 ::lto_codegen_add_must_preserve_symbol(generator, name);
732 }
599556ff
A
733 else if ( options.relocatable && hasNonllvmAtoms ) {
734 // <rdar://problem/14334895> ld -r mode but merging in some mach-o files, so need to keep libLTO from optimizing away anything
735 if ( logMustPreserve ) fprintf(stderr, "lto_codegen_add_must_preserve_symbol(%s) because -r mode disable LTO dead stripping\n", name);
736 ::lto_codegen_add_must_preserve_symbol(generator, name);
737 }
a645023d 738 }
0a8dc3df 739
599556ff
A
740 // <rdar://problem/16165191> tell code generator to preserve initial undefines
741 for( std::vector<const char*>::const_iterator it=options.initialUndefines->begin(); it != options.initialUndefines->end(); ++it) {
742 if ( logMustPreserve ) fprintf(stderr, "lto_codegen_add_must_preserve_symbol(%s) because it is an initial undefine\n", *it);
743 ::lto_codegen_add_must_preserve_symbol(generator, *it);
744 }
745
0a8dc3df
A
746 // special case running ld -r on all bitcode files to produce another bitcode file (instead of mach-o)
747 if ( options.relocatable && !hasNonllvmAtoms ) {
eaf282aa
A
748#if LTO_API_VERSION >= 15
749 ::lto_codegen_set_should_embed_uselists(generator, false);
750#endif
a645023d
A
751 if ( ! ::lto_codegen_write_merged_modules(generator, options.outputFilePath) ) {
752 // HACK, no good way to tell linker we are all done, so just quit
753 exit(0);
754 }
755 warning("could not produce merged bitcode file");
0a8dc3df
A
756 }
757
758}
759
760// Retrieve the codegen model from the options
761static lto_codegen_model getCodeModel(const OptimizeOptions& options) {
a645023d
A
762 if ( options.mainExecutable ) {
763 if ( options.staticExecutable ) {
d425e388
A
764 // x86_64 "static" or any "-static -pie" is really dynamic code model
765 if ( (options.arch == CPU_TYPE_X86_64) || options.pie )
0a8dc3df 766 return LTO_CODEGEN_PIC_MODEL_DYNAMIC;
a645023d 767 else
0a8dc3df 768 return LTO_CODEGEN_PIC_MODEL_STATIC;
a645023d 769 }
82b4b32b
A
770 else if ( options.preload ) {
771 if ( options.pie )
772 return LTO_CODEGEN_PIC_MODEL_DYNAMIC;
773 else
774 return LTO_CODEGEN_PIC_MODEL_STATIC;
775 }
a645023d
A
776 else {
777 if ( options.pie )
0a8dc3df 778 return LTO_CODEGEN_PIC_MODEL_DYNAMIC;
a645023d 779 else
0a8dc3df 780 return LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
a645023d
A
781 }
782 }
783 else {
784 if ( options.allowTextRelocs )
0a8dc3df 785 return LTO_CODEGEN_PIC_MODEL_DYNAMIC_NO_PIC;
a645023d 786 else
0a8dc3df 787 return LTO_CODEGEN_PIC_MODEL_DYNAMIC;
a645023d 788 }
0a8dc3df
A
789
790}
791
792std::tuple<uint8_t *, size_t> Parser::codegen(const OptimizeOptions& options,
793 ld::Internal& state,
794 lto_code_gen_t generator,
795 std::string& object_path) {
796 uint8_t *machOFile;
797 size_t machOFileLen;
798
799 if ( ::lto_codegen_set_pic_model(generator, getCodeModel(options)) )
a645023d
A
800 throwf("could not create set codegen model: %s", lto_get_error_message());
801
0a8dc3df
A
802 // if requested, save off merged bitcode file
803 if ( options.saveTemps ) {
804 char tempBitcodePath[MAXPATHLEN];
805 strcpy(tempBitcodePath, options.outputFilePath);
806 strcat(tempBitcodePath, ".lto.bc");
eaf282aa 807#if LTO_API_VERSION >= 15
0a8dc3df 808 ::lto_codegen_set_should_embed_uselists(generator, true);
eaf282aa 809#endif
0a8dc3df
A
810 ::lto_codegen_write_merged_modules(generator, tempBitcodePath);
811 }
a645023d
A
812
813#if LTO_API_VERSION >= 3
814 // find assembler next to linker
815 char path[PATH_MAX];
816 uint32_t bufSize = PATH_MAX;
817 if ( _NSGetExecutablePath(path, &bufSize) != -1 ) {
818 char* lastSlash = strrchr(path, '/');
819 if ( lastSlash != NULL ) {
820 strcpy(lastSlash+1, "as");
821 struct stat statInfo;
822 if ( stat(path, &statInfo) == 0 )
823 ::lto_codegen_set_assembler_path(generator, path);
824 }
825 }
826#endif
eaf282aa
A
827
828 // When lto API version is greater than or equal to 12, we use lto_codegen_optimize and lto_codegen_compile_optimized
829 // instead of lto_codegen_compile, and we save the merged bitcode file in between.
830 bool useSplitAPI = false;
831#if LTO_API_VERSION >= 12
832 if ( ::lto_api_version() >= 12)
833 useSplitAPI = true;
834#endif
835
eaf282aa
A
836 if ( useSplitAPI) {
837#if LTO_API_VERSION >= 12
838#if LTO_API_VERSION >= 14
0a8dc3df
A
839 if ( ::lto_api_version() >= 14 && options.ltoCodegenOnly)
840 lto_codegen_set_should_internalize(generator, false);
eaf282aa
A
841#endif
842 // run optimizer
843 if ( !options.ltoCodegenOnly && ::lto_codegen_optimize(generator) )
844 throwf("could not do LTO optimization: '%s', using libLTO version '%s'", ::lto_get_error_message(), ::lto_get_version());
845
846 if ( options.saveTemps || options.bitcodeBundle ) {
847 // save off merged bitcode file
848 char tempOptBitcodePath[MAXPATHLEN];
849 strcpy(tempOptBitcodePath, options.outputFilePath);
850 strcat(tempOptBitcodePath, ".lto.opt.bc");
851#if LTO_API_VERSION >= 15
852 ::lto_codegen_set_should_embed_uselists(generator, true);
853#endif
854 ::lto_codegen_write_merged_modules(generator, tempOptBitcodePath);
855 if ( options.bitcodeBundle )
0a8dc3df 856 state.ltoBitcodePath.push_back(tempOptBitcodePath);
eaf282aa
A
857 }
858
859 // run code generator
860 machOFile = (uint8_t*)::lto_codegen_compile_optimized(generator, &machOFileLen);
861#endif
862 if ( machOFile == NULL )
863 throwf("could not do LTO codegen: '%s', using libLTO version '%s'", ::lto_get_error_message(), ::lto_get_version());
864 }
865 else {
866 // run optimizer and code generator
867 machOFile = (uint8_t*)::lto_codegen_compile(generator, &machOFileLen);
868 if ( machOFile == NULL )
869 throwf("could not do LTO codegen: '%s', using libLTO version '%s'", ::lto_get_error_message(), ::lto_get_version());
870 if ( options.saveTemps ) {
871 // save off merged bitcode file
872 char tempOptBitcodePath[MAXPATHLEN];
873 strcpy(tempOptBitcodePath, options.outputFilePath);
874 strcat(tempOptBitcodePath, ".lto.opt.bc");
875#if LTO_API_VERSION >= 15
876 ::lto_codegen_set_should_embed_uselists(generator, true);
877#endif
878 ::lto_codegen_write_merged_modules(generator, tempOptBitcodePath);
879 }
880 }
0a8dc3df
A
881
882 // if requested, save off temp mach-o file
883 if ( options.saveTemps ) {
884 char tempMachoPath[MAXPATHLEN];
885 strcpy(tempMachoPath, options.outputFilePath);
886 strcat(tempMachoPath, ".lto.o");
887 int fd = ::open(tempMachoPath, O_CREAT | O_WRONLY | O_TRUNC, 0666);
a645023d
A
888 if ( fd != -1) {
889 ::write(fd, machOFile, machOFileLen);
890 ::close(fd);
891 }
a645023d
A
892 }
893
894 // if needed, save temp mach-o file to specific location
0a8dc3df
A
895 if ( !object_path.empty() ) {
896 int fd = ::open(object_path.c_str(), O_CREAT | O_WRONLY | O_TRUNC, 0666);
a645023d
A
897 if ( fd != -1) {
898 ::write(fd, machOFile, machOFileLen);
899 ::close(fd);
900 }
901 else {
0a8dc3df 902 warning("could not write LTO temp file '%s', errno=%d", object_path.c_str(), errno);
a645023d
A
903 }
904 }
0a8dc3df
A
905 return std::make_tuple(machOFile, machOFileLen);
906}
907
908/// Load the MachO located in buffer \p machOFile with size \p machOFileLen.
909/// The loaded atoms are sync'ed using all the supplied lists.
910void Parser::loadMachO(ld::relocatable::File* machoFile,
911 const OptimizeOptions& options,
912 ld::File::AtomHandler& handler,
913 std::vector<const ld::Atom*>& newAtoms,
914 std::vector<const char*>& additionalUndefines,
915 CStringToAtom &llvmAtoms,
916 CStringToAtom &deadllvmAtoms) {
917 const bool logAtomsBeforeSync = false;
918
a645023d
A
919 // sync generated mach-o atoms with existing atoms ld knows about
920 if ( logAtomsBeforeSync ) {
921 fprintf(stderr, "llvmAtoms:\n");
922 for (CStringToAtom::iterator it = llvmAtoms.begin(); it != llvmAtoms.end(); ++it) {
923 const char* name = it->first;
599556ff
A
924 Atom* atom = it->second;
925 fprintf(stderr, "\t%p\t%s\n", atom, name);
a645023d
A
926 }
927 fprintf(stderr, "deadllvmAtoms:\n");
928 for (CStringToAtom::iterator it = deadllvmAtoms.begin(); it != deadllvmAtoms.end(); ++it) {
929 const char* name = it->first;
599556ff
A
930 Atom* atom = it->second;
931 fprintf(stderr, "\t%p\t%s\n", atom, name);
a645023d
A
932 }
933 }
934 AtomSyncer syncer(additionalUndefines, newAtoms, llvmAtoms, deadllvmAtoms, options);
935 machoFile->forEachAtom(syncer);
0a8dc3df
A
936
937 // notify about file level attributes
938 handler.doFile(*machoFile);
939}
940
941// Full LTO processing
942bool Parser::optimizeLTO(const std::vector<File*> files,
943 const std::vector<const ld::Atom*>& allAtoms,
944 ld::Internal& state,
945 const OptimizeOptions& options,
946 ld::File::AtomHandler& handler,
947 std::vector<const ld::Atom*>& newAtoms,
948 std::vector<const char*>& additionalUndefines) {
949 const bool logExtraOptions = false;
950 const bool logBitcodeFiles = false;
951
952 if (files.empty())
953 return true;
954
955 // create optimizer and add each Reader
956 lto_code_gen_t generator = NULL;
957#if LTO_API_VERSION >= 11
958 if ( File::sSupportsLocalContext )
959 generator = ::lto_codegen_create_in_local_context();
960 else
961#endif
962 generator = ::lto_codegen_create();
963#if LTO_API_VERSION >= 7
964 lto_codegen_set_diagnostic_handler(generator, ltoDiagnosticHandler, NULL);
965#endif
966
967 ld::File::Ordinal lastOrdinal;
968
969 // When flto_codegen_only is on and we have a single .bc file, use lto_codegen_set_module instead of
970 // lto_codegen_add_module, to make sure the the destination module will be the same as the input .bc file.
971 bool useSetModule = false;
972#if LTO_API_VERSION >= 13
973 useSetModule = (files.size() == 1) && options.ltoCodegenOnly && (::lto_api_version() >= 13);
974#endif
975 for (auto *f : files) {
976 assert(f->ordinal() > lastOrdinal);
977 if ( logBitcodeFiles && !useSetModule ) fprintf(stderr, "lto_codegen_add_module(%s)\n", f->path());
978 if ( logBitcodeFiles && useSetModule ) fprintf(stderr, "lto_codegen_set_module(%s)\n", f->path());
979 if ( f->mergeIntoGenerator(generator, useSetModule) )
980 throwf("lto: could not merge in %s because '%s', using libLTO version '%s'", f->path(), ::lto_get_error_message(), ::lto_get_version());
981 lastOrdinal = f->ordinal();
982 }
983
984 // add any -mllvm command line options
985 if ( !_s_llvmOptionsProcessed ) {
986 for (const char* opt : *options.llvmOptions) {
987 if ( logExtraOptions ) fprintf(stderr, "passing option to llvm: %s\n", opt);
988 ::lto_codegen_debug_options(generator, opt);
989 }
990 _s_llvmOptionsProcessed = true;
a645023d 991 }
0a8dc3df
A
992
993 // <rdar://problem/13687397> Need a way for LTO to get cpu variants (until that info is in bitcode)
994 if ( options.mcpu != NULL )
995 ::lto_codegen_set_cpu(generator, options.mcpu);
996
997 // Compute the preserved symbols
998 CStringToAtom deadllvmAtoms, llvmAtoms;
999 setPreservedSymbols(allAtoms, state, options, deadllvmAtoms, llvmAtoms, generator);
1000
1001 size_t machOFileLen = 0;
1002 const uint8_t* machOFile = NULL;
1003
1004 // mach-o parsing is done in-memory, but need path for debug notes
1005 std::string object_path;
1006 if ( options.tmpObjectFilePath != NULL ) {
1007 object_path = options.tmpObjectFilePath;
1008 // If the path exists and is a directory (for instance if some files
1009 // were processed with ThinLTO before), we create the LTO file inside
1010 // the directory.
1011 struct stat statBuffer;
1012 if( stat(object_path.c_str(), &statBuffer) == 0 && S_ISDIR(statBuffer.st_mode) ) {
1013 object_path += "/lto.o";
1014 }
1015 }
1016
1017 // Codegen Now
1018 std::tie(machOFile, machOFileLen) = codegen(options, state, generator, object_path);
1019
1020 // parse generated mach-o file into a MachOReader
1021 ld::relocatable::File* machoFile = parseMachOFile(machOFile, machOFileLen, object_path, options, ld::File::Ordinal::LTOOrdinal());
1022
1023 // Load the generated MachO file
1024 loadMachO(machoFile, options, handler, newAtoms, additionalUndefines, llvmAtoms, deadllvmAtoms);
1025
a645023d
A
1026 // Remove Atoms from ld if code generator optimized them away
1027 for (CStringToAtom::iterator li = llvmAtoms.begin(), le = llvmAtoms.end(); li != le; ++li) {
1028 // check if setRealAtom() called on this Atom
1029 if ( li->second->compiledAtom() == NULL ) {
1030 //fprintf(stderr, "llvm optimized away %p %s\n", li->second, li->second->name());
1031 li->second->setCoalescedAway();
1032 }
1033 }
0a8dc3df 1034
a645023d 1035 // if final mach-o file has debug info, update original bitcode files to match
0a8dc3df
A
1036 for (auto *f : files) {
1037 f->setDebugInfo(machoFile->debugInfo(), machoFile->path(), machoFile->modificationTime(), machoFile->cpuSubType());
a645023d 1038 }
0a8dc3df 1039
a645023d
A
1040 return true;
1041}
1042
82b4b32b 1043#if LTO_API_VERSION >= 18
0a8dc3df
A
1044// Create the ThinLTO codegenerator
1045thinlto_code_gen_t Parser::init_thinlto_codegen(const std::vector<File*>& files,
1046 const std::vector<const ld::Atom*>& allAtoms,
1047 ld::Internal& state,
1048 const OptimizeOptions& options,
1049 CStringToAtom& deadllvmAtoms,
1050 CStringToAtom& llvmAtoms) {
1051 const bool logMustPreserve = false;
1052
1053 thinlto_code_gen_t thingenerator = ::thinlto_create_codegen();
1054
1055 // Caching control
1056 if (options.ltoCachePath && !options.bitcodeBundle) {
1057 struct stat statBuffer;
1058 if( stat(options.ltoCachePath, &statBuffer) != 0 || !S_ISDIR(statBuffer.st_mode) ) {
1059 if ( mkdir(options.ltoCachePath, 0700) !=0 ) {
1060 warning("unable to create ThinLTO cache directory: %s", options.ltoCachePath);
1061 }
1062 }
1063 thinlto_codegen_set_cache_dir(thingenerator, options.ltoCachePath);
e456bf10
A
1064 if (options.ltoPruneIntervalOverwrite)
1065 thinlto_codegen_set_cache_pruning_interval(thingenerator, options.ltoPruneInterval);
0a8dc3df
A
1066 thinlto_codegen_set_cache_entry_expiration(thingenerator, options.ltoPruneAfter);
1067 thinlto_codegen_set_final_cache_size_relative_to_available_space(thingenerator, options.ltoMaxCacheSize);
1068 }
1069
1070 // if requested, ask the code generator to save off intermediate bitcode files
1071 if ( options.saveTemps ) {
1072 std::string tempPath = options.outputFilePath;
1073 tempPath += ".thinlto.bcs/";
1074 struct stat statBuffer;
1075 if( stat(tempPath.c_str(), &statBuffer) != 0 || !S_ISDIR(statBuffer.st_mode) ) {
1076 if ( mkdir(tempPath.c_str(), 0700) !=0 ) {
1077 warning("unable to create ThinLTO output directory for temporary bitcode files: %s", tempPath.c_str());
1078 }
1079 }
1080 thinlto_codegen_set_savetemps_dir(thingenerator, tempPath.c_str());
1081 }
1082
1083 // Set some codegen options
1084 if ( thinlto_codegen_set_pic_model(thingenerator, getCodeModel(options)) )
1085 throwf("could not create set codegen model: %s", lto_get_error_message());
1086
1087 // Expose reachability informations for internalization in LTO
1088
1089 // The atom graph uses directed edges (references). Collect all references where
1090 // originating atom is not part of any LTO Reader. This allows optimizer to optimize an
1091 // external (i.e. not originated from same .o file) reference if all originating atoms are also
1092 // defined in llvm bitcode file.
1093 CStringSet nonLLVMRefs;
1094 CStringSet LLVMRefs;
1095 for (std::vector<const ld::Atom*>::const_iterator it = allAtoms.begin(); it != allAtoms.end(); ++it) {
1096 const ld::Atom* atom = *it;
1097 const ld::Atom* target;
1098 for (ld::Fixup::iterator fit=atom->fixupsBegin(); fit != atom->fixupsEnd(); ++fit) {
1099 switch ( fit->binding ) {
1100 case ld::Fixup::bindingDirectlyBound:
1101 // that reference a ThinLTO llvm atom
1102 target = fit->u.target;
1103 if ( target->contentType() == ld::Atom::typeLTOtemporary &&
1104 ((lto::File *)target->file())->isThinLTO() &&
1105 atom->file() != target->file()
1106 ) {
1107 if (atom->contentType() != ld::Atom::typeLTOtemporary ||
1108 !((lto::File *)atom->file())->isThinLTO())
1109 nonLLVMRefs.insert(target->name());
1110 else
1111 LLVMRefs.insert(target->name());
1112 if ( logMustPreserve )
1113 fprintf(stderr, "Found a reference from %s -> %s\n", atom->name(), target->name());
1114 }
1115 break;
1116 case ld::Fixup::bindingsIndirectlyBound:
1117 target = state.indirectBindingTable[fit->u.bindingIndex];
1118 if ( (target != NULL) && (target->contentType() == ld::Atom::typeLTOtemporary) &&
1119 ((lto::File *)target->file())->isThinLTO() &&
1120 atom->file() != target->file()
1121 ) {
1122 if (atom->contentType() != ld::Atom::typeLTOtemporary ||
1123 !((lto::File *)atom->file())->isThinLTO())
1124 nonLLVMRefs.insert(target->name());
1125 else
1126 LLVMRefs.insert(target->name());
1127 if ( logMustPreserve )
1128 fprintf(stderr, "Found a reference from %s -> %s\n", atom->name(), target->name());
1129 }
1130 default:
1131 break;
1132 }
1133 }
1134 if (atom->contentType() == ld::Atom::typeLTOtemporary &&
e456bf10
A
1135 ((lto::File *)atom->file())->isThinLTO() &&
1136 atom != &((lto::File *)atom->file())->internalAtom()) {
1137 assert(atom->scope() != ld::Atom::scopeTranslationUnit && "LTO should not expose static atoms");
1138 assert(llvmAtoms.find(atom->name()) == llvmAtoms.end() && "Unexpected llvmAtom with duplicate name");
0a8dc3df
A
1139 llvmAtoms[atom->name()] = (Atom*)atom;
1140 }
1141 }
1142 // if entry point is in a llvm bitcode file, it must be preserved by LTO
1143 if ( state.entryPoint != NULL ) {
1144 if ( state.entryPoint->contentType() == ld::Atom::typeLTOtemporary )
1145 nonLLVMRefs.insert(state.entryPoint->name());
1146 }
1147 for (auto file : files) {
1148 for(uint32_t i=0; i < file->_atomArrayCount; ++i) {
1149 Atom* llvmAtom = &file->_atomArray[i];
1150 if ( llvmAtom->coalescedAway() ) {
1151 const char* name = llvmAtom->name();
1152 if ( deadllvmAtoms.find(name) == deadllvmAtoms.end() ) {
1153 if ( logMustPreserve )
bee7e226 1154 fprintf(stderr, "thinlto_codegen_add_must_preserve_symbol(%s) because linker coalesce away and replace with a mach-o atom\n", name);
0a8dc3df
A
1155 ::thinlto_codegen_add_must_preserve_symbol(thingenerator, name, strlen(name));
1156 deadllvmAtoms[name] = (Atom*)llvmAtom;
1157 }
1158 }
1159 else if ( options.linkerDeadStripping && !llvmAtom->live() ) {
1160 const char* name = llvmAtom->name();
1161 deadllvmAtoms[name] = (Atom*)llvmAtom;
1162 }
1163 }
1164 }
1165
1166 // tell code generator about symbols that must be preserved
1167 for (CStringToAtom::iterator it = llvmAtoms.begin(); it != llvmAtoms.end(); ++it) {
1168 const char* name = it->first;
1169 Atom* atom = it->second;
1170 // Include llvm Symbol in export list if it meets one of following two conditions
1171 // 1 - atom scope is global (and not linkage unit).
1172 // 2 - included in nonLLVMRefs set.
1173 // If a symbol is not listed in exportList then LTO is free to optimize it away.
1174 if ( (atom->scope() == ld::Atom::scopeGlobal) && options.preserveAllGlobals ) {
bee7e226 1175 if ( logMustPreserve ) fprintf(stderr, "thinlto_codegen_add_must_preserve_symbol(%s) because global symbol\n", name);
0a8dc3df
A
1176 ::thinlto_codegen_add_must_preserve_symbol(thingenerator, name, strlen(name));
1177 }
1178 else if ( nonLLVMRefs.find(name) != nonLLVMRefs.end() ) {
bee7e226 1179 if ( logMustPreserve ) fprintf(stderr, "thinlto_codegen_add_must_preserve_symbol(%s) because referenced from outside of ThinLTO\n", name);
0a8dc3df
A
1180 ::thinlto_codegen_add_must_preserve_symbol(thingenerator, name, strlen(name));
1181 }
1182 else if ( LLVMRefs.find(name) != LLVMRefs.end() ) {
bee7e226 1183 if ( logMustPreserve ) fprintf(stderr, "thinlto_codegen_add_cross_referenced_symbol(%s) because referenced from another file\n", name);
0a8dc3df
A
1184 ::thinlto_codegen_add_cross_referenced_symbol(thingenerator, name, strlen(name));
1185 } else {
1186 if ( logMustPreserve ) fprintf(stderr, "NOT preserving(%s)\n", name);
1187 }
e456bf10
A
1188
1189 // <rdar://problem/16165191> tell code generator to preserve initial undefines
1190 for (const char* undefName : *options.initialUndefines) {
1191 if ( logMustPreserve ) fprintf(stderr, "thinlto_codegen_add_cross_referenced_symbol(%s) because it is an initial undefine\n", undefName);
1192 ::thinlto_codegen_add_cross_referenced_symbol(thingenerator, undefName, strlen(undefName));
1193 }
1194
0a8dc3df
A
1195// FIXME: to be implemented
1196// else if ( options.relocatable && hasNonllvmAtoms ) {
1197// // <rdar://problem/14334895> ld -r mode but merging in some mach-o files, so need to keep libLTO from optimizing away anything
1198// if ( logMustPreserve ) fprintf(stderr, "lto_codegen_add_must_preserve_symbol(%s) because -r mode disable LTO dead stripping\n", name);
1199// ::thinlto_codegen_add_must_preserve_symbol(thingenerator, name, strlen(name));
1200// }
1201 }
1202
1203 return thingenerator;
1204}
82b4b32b 1205#endif
0a8dc3df
A
1206
1207// Full LTO processing
1208bool Parser::optimizeThinLTO(const std::vector<File*>& files,
1209 const std::vector<const ld::Atom*>& allAtoms,
1210 ld::Internal& state,
1211 const OptimizeOptions& options,
1212 ld::File::AtomHandler& handler,
1213 std::vector<const ld::Atom*>& newAtoms,
1214 std::vector<const char*>& additionalUndefines) {
1215 const bool logBitcodeFiles = false;
1216
1217 if (files.empty())
1218 return true;
1219
1220#if LTO_API_VERSION >= 18
1221
1222 if (::lto_api_version() < 18)
1223 throwf("lto: could not use -thinlto because libLTO is too old (version '%d', >=18 is required)", ::lto_api_version());
1224
1225 // Handle -mllvm options
1226 if ( !_s_llvmOptionsProcessed ) {
1227 thinlto_debug_options(options.llvmOptions->data(), options.llvmOptions->size());
1228 _s_llvmOptionsProcessed = true;
1229 }
1230
1231 // Create the ThinLTO codegenerator
1232 CStringToAtom deadllvmAtoms;
1233 CStringToAtom llvmAtoms;
1234 thinlto_code_gen_t thingenerator = init_thinlto_codegen(files, allAtoms, state, options, deadllvmAtoms, llvmAtoms);
1235
1236
1237 ld::File::Ordinal lastOrdinal;
82b4b32b 1238 int FileId = 0;
0a8dc3df
A
1239 for (auto *f : files) {
1240 if ( logBitcodeFiles) fprintf(stderr, "thinlto_codegen_add_module(%s)\n", f->path());
82b4b32b 1241 f->addToThinGenerator(thingenerator, FileId++);
0a8dc3df
A
1242 lastOrdinal = f->ordinal();
1243 }
1244
1245#if LTO_API_VERSION >= 19
1246 // In the bitcode bundle case, we first run the generator with codegen disabled
1247 // and get the bitcode output. These files are added for later bundling, and a
1248 // new codegenerator is setup with these as input, and the optimizer disabled.
1249 if (options.bitcodeBundle) {
1250 // Bitcode Bundle case
1251 thinlto_codegen_disable_codegen(thingenerator, true);
1252 // Process the optimizer only
1253 thinlto_codegen_process(thingenerator);
1254 auto numObjects = thinlto_module_get_num_objects(thingenerator);
1255 // Save the codegenerator
1256 thinlto_code_gen_t bitcode_generator = thingenerator;
1257 // Create a new codegen generator for the codegen part.
e456bf10
A
1258 // Clear out the stored atoms so we can recompute them.
1259 deadllvmAtoms.clear();
1260 llvmAtoms.clear();
0a8dc3df
A
1261 thingenerator = init_thinlto_codegen(files, allAtoms, state, options, deadllvmAtoms, llvmAtoms);
1262 // Disable the optimizer
1263 thinlto_codegen_set_codegen_only(thingenerator, true);
1264
1265 // Save bitcode files for later, and add them to the codegen generator.
1266 for (unsigned bufID = 0; bufID < numObjects; ++bufID) {
1267 auto machOFile = thinlto_module_get_object(bitcode_generator, bufID);
1268 std::string tempMachoPath = options.outputFilePath;
1269 tempMachoPath += ".";
1270 tempMachoPath += std::to_string(bufID);
1271 tempMachoPath += ".thinlto.o.bc";
1272 state.ltoBitcodePath.push_back(tempMachoPath);
1273 int fd = ::open(tempMachoPath.c_str(), O_CREAT | O_WRONLY | O_TRUNC, 0666);
1274 if ( fd != -1 ) {
1275 ::write(fd, machOFile.Buffer, machOFile.Size);
1276 ::close(fd);
1277 } else {
1278 throwf("unable to write temporary ThinLTO output: %s", tempMachoPath.c_str());
1279 }
1280
1281 // Add the optimized bitcode to the codegen generator now.
bee7e226 1282 ::thinlto_codegen_add_module(thingenerator, strdup(tempMachoPath.c_str()), (const char *)machOFile.Buffer, machOFile.Size);
0a8dc3df
A
1283 }
1284 }
1285
1286 if (options.ltoCodegenOnly)
1287 // Disable the optimizer
1288 thinlto_codegen_set_codegen_only(thingenerator, true);
1289#endif
1290
82b4b32b
A
1291 // If object_path_lto is used, we switch to a file-based API: libLTO will
1292 // generate the files on disk and we'll map them on-demand.
1293
1294#if LTO_API_VERSION >= 21
1295 bool useFileBasedAPI = (options.tmpObjectFilePath && ::lto_api_version() >= 21);
1296 if ( useFileBasedAPI )
1297 thinlto_set_generated_objects_dir(thingenerator, options.tmpObjectFilePath);
1298#endif
1299
0a8dc3df
A
1300 // run code generator
1301 thinlto_codegen_process(thingenerator);
82b4b32b
A
1302
1303 unsigned numObjects;
1304#if LTO_API_VERSION >= 21
1305 if ( useFileBasedAPI )
1306 numObjects = thinlto_module_get_num_object_files(thingenerator);
1307 else
1308#endif
1309 numObjects = thinlto_module_get_num_objects(thingenerator);
1310 if ( numObjects == 0 )
0a8dc3df
A
1311 throwf("could not do ThinLTO codegen (thinlto_codegen_process didn't produce any object): '%s', using libLTO version '%s'", ::lto_get_error_message(), ::lto_get_version());
1312
bee7e226
A
1313 auto get_thinlto_buffer_or_load_file = [&] (unsigned ID) {
1314#if LTO_API_VERSION >= 21
1315 if ( useFileBasedAPI ) {
1316 const char* path = thinlto_module_get_object_file(thingenerator, ID);
1317 // map in whole file
1318 struct stat stat_buf;
1319 int fd = ::open(path, O_RDONLY, 0);
1320 if ( fd == -1 )
1321 throwf("can't open thinlto file '%s', errno=%d", path, errno);
1322 if ( ::fstat(fd, &stat_buf) != 0 )
1323 throwf("fstat thinlto file '%s' failed, errno=%d\n", path, errno);
1324 size_t len = stat_buf.st_size;
1325 if ( len < 20 )
1326 throwf("ThinLTO file '%s' too small (length=%zu)", path, len);
1327 const char* p = (const char*)::mmap(NULL, len, PROT_READ, MAP_FILE | MAP_PRIVATE, fd, 0);
1328 if ( p == (const char*)(-1) )
1329 throwf("can't map file, errno=%d", errno);
1330 ::close(fd);
1331 return LTOObjectBuffer{ p, len };
1332 }
1333#endif
1334 return thinlto_module_get_object(thingenerator, ID);
1335 };
1336
0a8dc3df
A
1337 // if requested, save off objects files
1338 if ( options.saveTemps ) {
1339 for (unsigned bufID = 0; bufID < numObjects; ++bufID) {
bee7e226 1340 auto machOFile = get_thinlto_buffer_or_load_file(bufID);
0a8dc3df
A
1341 std::string tempMachoPath = options.outputFilePath;
1342 tempMachoPath += ".";
1343 tempMachoPath += std::to_string(bufID);
1344 tempMachoPath += ".thinlto.o";
1345 int fd = ::open(tempMachoPath.c_str(), O_CREAT | O_WRONLY | O_TRUNC, 0666);
1346 if ( fd != -1 ) {
1347 ::write(fd, machOFile.Buffer, machOFile.Size);
1348 ::close(fd);
82b4b32b
A
1349 }
1350 else {
0a8dc3df
A
1351 warning("unable to write temporary ThinLTO output: %s", tempMachoPath.c_str());
1352 }
1353 }
1354 }
1355
0a8dc3df
A
1356 // mach-o parsing is done in-memory, but need path for debug notes
1357 std::string macho_dirpath = "/tmp/thinlto.o";
1358 if ( options.tmpObjectFilePath != NULL ) {
1359 macho_dirpath = options.tmpObjectFilePath;
1360 struct stat statBuffer;
1361 if( stat(macho_dirpath.c_str(), &statBuffer) != 0 || !S_ISDIR(statBuffer.st_mode) ) {
82b4b32b 1362 unlink(macho_dirpath.c_str());
0a8dc3df
A
1363 if ( mkdir(macho_dirpath.c_str(), 0700) !=0 ) {
1364 warning("unable to create ThinLTO output directory for temporary object files: %s", macho_dirpath.c_str());
1365 }
1366 }
1367 }
1368
1369 auto ordinal = ld::File::Ordinal::LTOOrdinal().nextFileListOrdinal();
1370 for (unsigned bufID = 0; bufID < numObjects; ++bufID) {
82b4b32b 1371 auto machOFile = get_thinlto_buffer_or_load_file(bufID);
0a8dc3df
A
1372 if (!machOFile.Size) {
1373 warning("Ignoring empty buffer generated by ThinLTO");
1374 continue;
1375 }
1376
1377 // mach-o parsing is done in-memory, but need path for debug notes
82b4b32b
A
1378 std::string tmp_path;
1379#if LTO_API_VERSION >= 21
1380 if ( useFileBasedAPI ) {
1381 tmp_path = thinlto_module_get_object_file(thingenerator, bufID);
1382 }
1383 else
1384#endif
1385 if ( options.tmpObjectFilePath != NULL) {
1386 tmp_path = macho_dirpath + "/" + std::to_string(bufID) + ".o";
1387 // if needed, save temp mach-o file to specific location
0a8dc3df
A
1388 int fd = ::open(tmp_path.c_str(), O_CREAT | O_WRONLY | O_TRUNC, 0666);
1389 if ( fd != -1) {
1390 ::write(fd, (const uint8_t *)machOFile.Buffer, machOFile.Size);
1391 ::close(fd);
1392 }
1393 else {
1394 warning("could not write ThinLTO temp file '%s', errno=%d", tmp_path.c_str(), errno);
1395 }
1396 }
1397
7f09b935
A
1398 // parse generated mach-o file into a MachOReader
1399 ld::relocatable::File* machoFile = parseMachOFile((const uint8_t *)machOFile.Buffer, machOFile.Size, tmp_path, options, ordinal);
1400 ordinal = ordinal.nextFileListOrdinal();
1401
0a8dc3df
A
1402 // Load the generated MachO file
1403 loadMachO(machoFile, options, handler, newAtoms, additionalUndefines, llvmAtoms, deadllvmAtoms);
1404 }
1405
1406 // Remove Atoms from ld if code generator optimized them away
1407 for (CStringToAtom::iterator li = llvmAtoms.begin(), le = llvmAtoms.end(); li != le; ++li) {
1408 // check if setRealAtom() called on this Atom
1409 if ( li->second->compiledAtom() == NULL ) {
1410 //fprintf(stderr, "llvm optimized away %p %s\n", li->second, li->second->name());
1411 li->second->setCoalescedAway();
1412 }
1413 }
1414
1415 return true;
1416#else // ! (LTO_API_VERSION >= 18)
1417 throwf("lto: could not use -thinlto because ld was built against a version of libLTO too old (version '%d', >=18 is required)", LTO_API_VERSION);
1418#endif
1419}
1420
1421bool Parser::optimize( const std::vector<const ld::Atom*>& allAtoms,
1422 ld::Internal& state,
1423 const OptimizeOptions& options,
1424 ld::File::AtomHandler& handler,
1425 std::vector<const ld::Atom*>& newAtoms,
1426 std::vector<const char*>& additionalUndefines)
1427{
1428
1429 // exit quickly if nothing to do
1430 if ( _s_files.size() == 0 )
1431 return false;
1432
1433 // print out LTO version string if -v was used
1434 if ( options.verbose )
1435 fprintf(stderr, "%s\n", ::lto_get_version());
1436
1437 // <rdar://problem/12379604> The order that files are merged must match command line order
1438 std::sort(_s_files.begin(), _s_files.end(), CommandLineOrderFileSorter());
1439
1440#if LTO_API_VERSION >= 19
1441 // If ltoCodegenOnly is set, we don't want to merge any bitcode files and perform FullLTO
1442 // we just take the ThinLTO path (optimization will be disabled anyway).
1443 if (options.ltoCodegenOnly) {
1444 for (auto *file : _s_files) {
1445 file->setIsThinLTO(true);
1446 }
1447 }
1448#endif
1449
1450 std::vector<File *> theLTOFiles;
1451 std::vector<File *> theThinLTOFiles;
1452 for (auto *file : _s_files) {
1453 if (file->isThinLTO()) {
1454 theThinLTOFiles.push_back(file);
1455 } else {
1456 theLTOFiles.push_back(file);
1457 }
1458 }
1459
1460 auto result = optimizeThinLTO(theThinLTOFiles, allAtoms, state, options, handler, newAtoms, additionalUndefines) &&
1461 optimizeLTO(theLTOFiles, allAtoms, state, options, handler, newAtoms, additionalUndefines);
1462
1463 // Remove InternalAtoms from ld
1464 for (std::vector<File*>::iterator it=_s_files.begin(); it != _s_files.end(); ++it) {
1465 (*it)->internalAtom().setCoalescedAway();
1466 }
1467
1468 return result;
1469}
1470
a645023d
A
1471
1472void Parser::AtomSyncer::doAtom(const ld::Atom& machoAtom)
1473{
599556ff 1474 static const bool log = false;
a645023d
A
1475 // update proxy atoms to point to real atoms and find new atoms
1476 const char* name = machoAtom.name();
0a8dc3df 1477 CStringToAtom::const_iterator pos = _llvmAtoms.find(name);
f410558f 1478 if ( (pos != _llvmAtoms.end()) && (machoAtom.scope() != ld::Atom::scopeTranslationUnit) ) {
599556ff 1479 // turn Atom into a proxy for this mach-o atom
82b4b32b
A
1480 if (pos->second->scope() == ld::Atom::scopeLinkageUnit) {
1481 if (log) fprintf(stderr, "demote %s to hidden after LTO\n", name);
1482 (const_cast<ld::Atom*>(&machoAtom))->setScope(ld::Atom::scopeLinkageUnit);
1483 }
e456bf10
A
1484 // If both llvmAtom and machoAtom has the same scope and combine, but machoAtom loses auto hide, add it back.
1485 // rdar://problem/38646854
1486 if (pos->second->scope() == machoAtom.scope() &&
1487 pos->second->combine() == machoAtom.combine() &&
1488 pos->second->autoHide() && !machoAtom.autoHide()) {
1489 if (log) fprintf(stderr, "set %s to auto hide after LTO\n", name);
1490 (const_cast<ld::Atom*>(&machoAtom))->setAutoHide();
1491 }
599556ff 1492 pos->second->setCompiledAtom(machoAtom);
0a8dc3df
A
1493 _lastProxiedAtom = &machoAtom;
1494 _lastProxiedFile = pos->second->file();
599556ff
A
1495 if (log) fprintf(stderr, "AtomSyncer, mach-o atom %p synced to lto atom %p (name=%s)\n", &machoAtom, pos->second, machoAtom.name());
1496 }
1497 else {
1498 // an atom of this name was not in the allAtoms list the linker gave us
0a8dc3df
A
1499 auto llvmAtom = _deadllvmAtoms.find(name);
1500 if ( llvmAtom != _deadllvmAtoms.end() ) {
599556ff
A
1501 // this corresponding to an atom that the linker coalesced away or marked not-live
1502 if ( _options.linkerDeadStripping ) {
1503 // llvm seems to want this atom and -dead_strip is enabled, so it will be deleted if not needed, so add back
0a8dc3df 1504 llvmAtom->second->setCompiledAtom(machoAtom);
a645023d 1505 _newAtoms.push_back(&machoAtom);
0a8dc3df 1506 if (log) fprintf(stderr, "AtomSyncer, mach-o atom %p matches dead lto atom %p but adding back (name=%s)\n", &machoAtom, llvmAtom->second, machoAtom.name());
a645023d 1507 }
599556ff
A
1508 else {
1509 // Don't pass it back as a new atom
0a8dc3df 1510 if (log) fprintf(stderr, "AtomSyncer, mach-o atom %p matches dead lto atom %p (name=%s)\n", &machoAtom, llvmAtom->second, machoAtom.name());
82b4b32b
A
1511 if ( llvmAtom->second->coalescedAway() ) {
1512 if (log) fprintf(stderr, "AtomSyncer: dead coalesced atom %s\n", machoAtom.name());
1513 // <rdar://problem/28269547>
1514 // We told libLTO to keep a weak atom that will replaced by an native mach-o atom.
1515 // We also need to remove any atoms directly dependent on this (FDE, LSDA).
1516 for (ld::Fixup::iterator fit=machoAtom.fixupsBegin(), fend=machoAtom.fixupsEnd(); fit != fend; ++fit) {
1517 switch ( fit->kind ) {
1518 case ld::Fixup::kindNoneGroupSubordinate:
1519 case ld::Fixup::kindNoneGroupSubordinateFDE:
1520 case ld::Fixup::kindNoneGroupSubordinateLSDA:
1521 assert(fit->binding == ld::Fixup::bindingDirectlyBound);
1522 (const_cast<ld::Atom*>(fit->u.target))->setCoalescedAway();
1523 if (log) fprintf(stderr, "AtomSyncer: mark coalesced-away subordinate atom %s\n", fit->u.target->name());
1524 break;
1525 default:
1526 break;
1527 }
1528 }
1529 }
599556ff 1530 }
a645023d 1531 }
599556ff
A
1532 else
1533 {
1534 // this is something new that lto conjured up, tell ld its new
1535 _newAtoms.push_back(&machoAtom);
1536 // <rdar://problem/15469363> if new static atom in same section as previous non-static atom, assign to same file as previous
0a8dc3df 1537 if ( (_lastProxiedAtom != NULL) && (_lastProxiedAtom->section() == machoAtom.section()) ) {
599556ff 1538 ld::Atom* ma = const_cast<ld::Atom*>(&machoAtom);
0a8dc3df 1539 ma->setFile(_lastProxiedFile);
e456bf10 1540 if (log) fprintf(stderr, "AtomSyncer, mach-o atom %s is static and being assigned to %s\n", machoAtom.name(), _lastProxiedFile->path());
599556ff
A
1541 }
1542 if (log) fprintf(stderr, "AtomSyncer, mach-o atom %p is totally new (name=%s)\n", &machoAtom, machoAtom.name());
9543cb2f 1543 }
a645023d
A
1544 }
1545
1546 // adjust fixups to go through proxy atoms
599556ff 1547 if (log) fprintf(stderr, " adjusting fixups in atom: %s\n", machoAtom.name());
a645023d
A
1548 for (ld::Fixup::iterator fit=machoAtom.fixupsBegin(); fit != machoAtom.fixupsEnd(); ++fit) {
1549 switch ( fit->binding ) {
1550 case ld::Fixup::bindingNone:
1551 break;
1552 case ld::Fixup::bindingByNameUnbound:
1553 // don't know if this target has been seen by linker before or if it is new
1554 // be conservative and tell linker it is new
1555 _additionalUndefines.push_back(fit->u.name);
599556ff 1556 if (log) fprintf(stderr, " adding by-name symbol %s\n", fit->u.name);
a645023d
A
1557 break;
1558 case ld::Fixup::bindingDirectlyBound:
1559 // If mach-o atom is referencing another mach-o atom then
1560 // reference is not going through Atom proxy. Fix it here to ensure that all
1561 // llvm symbol references always go through Atom proxy.
f410558f 1562 if ( fit->u.target->scope() != ld::Atom::scopeTranslationUnit )
599556ff 1563 {
a645023d 1564 const char* targetName = fit->u.target->name();
0a8dc3df 1565 CStringToAtom::const_iterator post = _llvmAtoms.find(targetName);
599556ff
A
1566 if ( post != _llvmAtoms.end() ) {
1567 const ld::Atom* t = post->second;
1568 if (log) fprintf(stderr, " updating direct reference to %p to be ref to %p: %s\n", fit->u.target, t, targetName);
1569 fit->u.target = t;
a645023d
A
1570 }
1571 else {
f80fe69f 1572 // <rdar://problem/12859831> Don't unbind follow-on reference into by-name reference
eaf282aa 1573 if ( (_deadllvmAtoms.find(targetName) != _deadllvmAtoms.end()) && (fit->kind != ld::Fixup::kindNoneFollowOn) && (fit->u.target->scope() != ld::Atom::scopeTranslationUnit) ) {
a645023d
A
1574 // target was coalesed away and replace by mach-o atom from a non llvm .o file
1575 fit->binding = ld::Fixup::bindingByNameUnbound;
1576 fit->u.name = targetName;
1577 }
1578 }
1579 }
1580 //fprintf(stderr, " direct ref to: %s (scope=%d)\n", fit->u.target->name(), fit->u.target->scope());
1581 break;
1582 case ld::Fixup::bindingByContentBound:
1583 //fprintf(stderr, " direct by content to: %s\n", fit->u.target->name());
1584 break;
1585 case ld::Fixup::bindingsIndirectlyBound:
1586 assert(0 && "indirect binding found in initial mach-o file?");
1587 //fprintf(stderr, " indirect by content to: %u\n", fit->u.bindingIndex);
1588 break;
1589 }
1590 }
1591
1592}
1593
ebf6f434
A
1594class Mutex {
1595 static pthread_mutex_t lto_lock;
1596public:
1597 Mutex() { pthread_mutex_lock(&lto_lock); }
1598 ~Mutex() { pthread_mutex_unlock(&lto_lock); }
1599};
1600pthread_mutex_t Mutex::lto_lock = PTHREAD_MUTEX_INITIALIZER;
eaf282aa
A
1601bool File::sSupportsLocalContext = false;
1602bool File::sHasTriedLocalContext = false;
a645023d
A
1603
1604//
1605// Used by archive reader to see if member is an llvm bitcode file
1606//
1607bool isObjectFile(const uint8_t* fileContent, uint64_t fileLength, cpu_type_t architecture, cpu_subtype_t subarch)
1608{
ebf6f434 1609 Mutex lock;
a645023d
A
1610 return Parser::validFile(fileContent, fileLength, architecture, subarch);
1611}
1612
0a8dc3df
A
1613//
1614// Used by archive reader to see if member defines a Category (for -ObjC semantics)
1615//
1616bool hasObjCCategory(const uint8_t* fileContent, uint64_t fileLength)
1617{
1618#if LTO_API_VERSION >= 20
1619 // note: if run with older libLTO.dylib that does not implement
1620 // lto_module_has_objc_category, the call will return 0 which is "false"
1621 return lto_module_has_objc_category(fileContent, fileLength);
1622#else
1623 return false;
1624#endif
1625}
1626
1627
eaf282aa
A
1628static ld::relocatable::File *parseImpl(
1629 const uint8_t *fileContent, uint64_t fileLength, const char *path,
1630 time_t modTime, ld::File::Ordinal ordinal, cpu_type_t architecture,
1631 cpu_subtype_t subarch, bool logAllFiles,
1632 bool verboseOptimizationHints)
1633{
1634 if ( Parser::validFile(fileContent, fileLength, architecture, subarch) )
1635 return Parser::parse(fileContent, fileLength, path, modTime, ordinal, architecture, subarch, logAllFiles, verboseOptimizationHints);
1636 else
1637 return NULL;
1638}
1639
a645023d
A
1640//
1641// main function used by linker to instantiate ld::Files
1642//
1643ld::relocatable::File* parse(const uint8_t* fileContent, uint64_t fileLength,
b1f7435d 1644 const char* path, time_t modTime, ld::File::Ordinal ordinal,
9543cb2f
A
1645 cpu_type_t architecture, cpu_subtype_t subarch, bool logAllFiles,
1646 bool verboseOptimizationHints)
a645023d 1647{
ec29ba20
A
1648 // do light weight check before acquiring lock
1649 if ( fileLength < 4 )
1650 return NULL;
1651 if ( (fileContent[0] != 0xDE) || (fileContent[1] != 0xC0) || (fileContent[2] != 0x17) || (fileContent[3] != 0x0B) )
1652 return NULL;
1653
eaf282aa
A
1654 // Note: Once lto_module_create_in_local_context() and friends are thread safe
1655 // this lock can be removed.
ebf6f434 1656 Mutex lock;
eaf282aa
A
1657 return parseImpl(fileContent, fileLength, path, modTime, ordinal,
1658 architecture, subarch, logAllFiles,
1659 verboseOptimizationHints);
a645023d
A
1660}
1661
1662//
1663// used by "ld -v" to report version of libLTO.dylib being used
1664//
1665const char* version()
1666{
ebf6f434 1667 Mutex lock;
a645023d
A
1668 return ::lto_get_version();
1669}
1670
bee7e226
A
1671//
1672// used by "ld -v" to report static version of libLTO.dylib API being compiled
1673//
1674unsigned int static_api_version()
1675{
1676 return LTO_API_VERSION;
1677}
1678
1679//
1680// used by "ld -v" to report version of libLTO.dylib being used
1681//
1682unsigned int runtime_api_version()
1683{
1684 return ::lto_api_version();
1685}
1686
a645023d
A
1687
1688//
1689// used by ld for error reporting
1690//
1691bool libLTOisLoaded()
1692{
ebf6f434 1693 Mutex lock;
a645023d
A
1694 return (::lto_get_version() != NULL);
1695}
1696
1697//
1698// used by ld for error reporting
1699//
1700const char* archName(const uint8_t* fileContent, uint64_t fileLength)
1701{
ebf6f434 1702 Mutex lock;
a645023d
A
1703 return Parser::fileKind(fileContent, fileLength);
1704}
1705
1706//
1707// used by ld for doing link time optimization
1708//
1709bool optimize( const std::vector<const ld::Atom*>& allAtoms,
1710 ld::Internal& state,
a645023d
A
1711 const OptimizeOptions& options,
1712 ld::File::AtomHandler& handler,
1713 std::vector<const ld::Atom*>& newAtoms,
1714 std::vector<const char*>& additionalUndefines)
1715{
ebf6f434
A
1716 Mutex lock;
1717 return Parser::optimize(allAtoms, state, options, handler, newAtoms, additionalUndefines);
a645023d
A
1718}
1719
1720
1721
1722}; // namespace lto
1723
1724
1725#endif
1726