1 //===- DriverUtils.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 "Config.h"
10 #include "Driver.h"
11 #include "InputFiles.h"
12 #include "ObjC.h"
13 #include "Target.h"
14
15 #include "lld/Common/Args.h"
16 #include "lld/Common/CommonLinkerContext.h"
17 #include "lld/Common/Reproduce.h"
18 #include "llvm/ADT/CachedHashString.h"
19 #include "llvm/ADT/DenseMap.h"
20 #include "llvm/LTO/LTO.h"
21 #include "llvm/Option/Arg.h"
22 #include "llvm/Option/ArgList.h"
23 #include "llvm/Option/Option.h"
24 #include "llvm/Support/CommandLine.h"
25 #include "llvm/Support/FileSystem.h"
26 #include "llvm/Support/Path.h"
27 #include "llvm/TextAPI/InterfaceFile.h"
28 #include "llvm/TextAPI/TextAPIReader.h"
29
30 using namespace llvm;
31 using namespace llvm::MachO;
32 using namespace llvm::opt;
33 using namespace llvm::sys;
34 using namespace lld;
35 using namespace lld::macho;
36
37 // Create prefix string literals used in Options.td
38 #define PREFIX(NAME, VALUE) \
39 static constexpr StringLiteral NAME##_init[] = VALUE; \
40 static constexpr ArrayRef<StringLiteral> NAME(NAME##_init, \
41 std::size(NAME##_init) - 1);
42 #include "Options.inc"
43 #undef PREFIX
44
45 // Create table mapping all options defined in Options.td
46 static constexpr OptTable::Info optInfo[] = {
47 #define OPTION(PREFIX, NAME, ID, KIND, GROUP, ALIAS, ALIASARGS, FLAGS, \
48 VISIBILITY, PARAM, HELPTEXT, HELPTEXTSFORVARIANTS, METAVAR, \
49 VALUES) \
50 {PREFIX, \
51 NAME, \
52 HELPTEXT, \
53 HELPTEXTSFORVARIANTS, \
54 METAVAR, \
55 OPT_##ID, \
56 opt::Option::KIND##Class, \
57 PARAM, \
58 FLAGS, \
59 VISIBILITY, \
60 OPT_##GROUP, \
61 OPT_##ALIAS, \
62 ALIASARGS, \
63 VALUES},
64 #include "Options.inc"
65 #undef OPTION
66 };
67
MachOOptTable()68 MachOOptTable::MachOOptTable() : GenericOptTable(optInfo) {}
69
70 // Set color diagnostics according to --color-diagnostics={auto,always,never}
71 // or --no-color-diagnostics flags.
handleColorDiagnostics(InputArgList & args)72 static void handleColorDiagnostics(InputArgList &args) {
73 const Arg *arg =
74 args.getLastArg(OPT_color_diagnostics, OPT_color_diagnostics_eq,
75 OPT_no_color_diagnostics);
76 if (!arg)
77 return;
78 if (arg->getOption().getID() == OPT_color_diagnostics) {
79 lld::errs().enable_colors(true);
80 } else if (arg->getOption().getID() == OPT_no_color_diagnostics) {
81 lld::errs().enable_colors(false);
82 } else {
83 StringRef s = arg->getValue();
84 if (s == "always")
85 lld::errs().enable_colors(true);
86 else if (s == "never")
87 lld::errs().enable_colors(false);
88 else if (s != "auto")
89 error("unknown option: --color-diagnostics=" + s);
90 }
91 }
92
parse(ArrayRef<const char * > argv)93 InputArgList MachOOptTable::parse(ArrayRef<const char *> argv) {
94 // Make InputArgList from string vectors.
95 unsigned missingIndex;
96 unsigned missingCount;
97 SmallVector<const char *, 256> vec(argv.data(), argv.data() + argv.size());
98
99 // Expand response files (arguments in the form of @<filename>)
100 // and then parse the argument again.
101 cl::ExpandResponseFiles(saver(), cl::TokenizeGNUCommandLine, vec);
102 InputArgList args = ParseArgs(vec, missingIndex, missingCount);
103
104 // Handle -fatal_warnings early since it converts missing argument warnings
105 // to errors.
106 errorHandler().fatalWarnings = args.hasArg(OPT_fatal_warnings);
107 errorHandler().suppressWarnings = args.hasArg(OPT_w);
108
109 if (missingCount)
110 error(Twine(args.getArgString(missingIndex)) + ": missing argument");
111
112 handleColorDiagnostics(args);
113
114 for (const Arg *arg : args.filtered(OPT_UNKNOWN)) {
115 std::string nearest;
116 if (findNearest(arg->getAsString(args), nearest) > 1)
117 error("unknown argument '" + arg->getAsString(args) + "'");
118 else
119 error("unknown argument '" + arg->getAsString(args) +
120 "', did you mean '" + nearest + "'");
121 }
122 return args;
123 }
124
printHelp(const char * argv0,bool showHidden) const125 void MachOOptTable::printHelp(const char *argv0, bool showHidden) const {
126 OptTable::printHelp(lld::outs(),
127 (std::string(argv0) + " [options] file...").c_str(),
128 "LLVM Linker", showHidden);
129 lld::outs() << "\n";
130 }
131
rewritePath(StringRef s)132 static std::string rewritePath(StringRef s) {
133 if (fs::exists(s))
134 return relativeToRoot(s);
135 return std::string(s);
136 }
137
rewriteInputPath(StringRef s)138 static std::string rewriteInputPath(StringRef s) {
139 // Don't bother rewriting "absolute" paths that are actually under the
140 // syslibroot; simply rewriting the syslibroot is sufficient.
141 if (rerootPath(s) == s && fs::exists(s))
142 return relativeToRoot(s);
143 return std::string(s);
144 }
145
146 // Reconstructs command line arguments so that so that you can re-run
147 // the same command with the same inputs. This is for --reproduce.
createResponseFile(const InputArgList & args)148 std::string macho::createResponseFile(const InputArgList &args) {
149 SmallString<0> data;
150 raw_svector_ostream os(data);
151
152 // Copy the command line to the output while rewriting paths.
153 for (const Arg *arg : args) {
154 switch (arg->getOption().getID()) {
155 case OPT_reproduce:
156 break;
157 case OPT_INPUT:
158 os << quote(rewriteInputPath(arg->getValue())) << "\n";
159 break;
160 case OPT_o:
161 os << "-o " << quote(path::filename(arg->getValue())) << "\n";
162 break;
163 case OPT_filelist:
164 if (std::optional<MemoryBufferRef> buffer = readFile(arg->getValue()))
165 for (StringRef path : args::getLines(*buffer))
166 os << quote(rewriteInputPath(path)) << "\n";
167 break;
168 case OPT_force_load:
169 case OPT_weak_library:
170 case OPT_load_hidden:
171 os << arg->getSpelling() << " "
172 << quote(rewriteInputPath(arg->getValue())) << "\n";
173 break;
174 case OPT_F:
175 case OPT_L:
176 case OPT_bundle_loader:
177 case OPT_exported_symbols_list:
178 case OPT_order_file:
179 case OPT_syslibroot:
180 case OPT_unexported_symbols_list:
181 os << arg->getSpelling() << " " << quote(rewritePath(arg->getValue()))
182 << "\n";
183 break;
184 case OPT_sectcreate:
185 os << arg->getSpelling() << " " << quote(arg->getValue(0)) << " "
186 << quote(arg->getValue(1)) << " "
187 << quote(rewritePath(arg->getValue(2))) << "\n";
188 break;
189 default:
190 os << toString(*arg) << "\n";
191 }
192 }
193 return std::string(data);
194 }
195
searchedDylib(const Twine & path,bool found)196 static void searchedDylib(const Twine &path, bool found) {
197 if (config->printDylibSearch)
198 message("searched " + path + (found ? ", found " : ", not found"));
199 if (!found)
200 depTracker->logFileNotFound(path);
201 }
202
resolveDylibPath(StringRef dylibPath)203 std::optional<StringRef> macho::resolveDylibPath(StringRef dylibPath) {
204 // TODO: if a tbd and dylib are both present, we should check to make sure
205 // they are consistent.
206 SmallString<261> tbdPath = dylibPath;
207 path::replace_extension(tbdPath, ".tbd");
208 bool tbdExists = fs::exists(tbdPath);
209 searchedDylib(tbdPath, tbdExists);
210 if (tbdExists)
211 return saver().save(tbdPath.str());
212
213 bool dylibExists = fs::exists(dylibPath);
214 searchedDylib(dylibPath, dylibExists);
215 if (dylibExists)
216 return saver().save(dylibPath);
217 return {};
218 }
219
220 // It's not uncommon to have multiple attempts to load a single dylib,
221 // especially if it's a commonly re-exported core library.
222 static DenseMap<CachedHashStringRef, DylibFile *> loadedDylibs;
223
loadDylib(MemoryBufferRef mbref,DylibFile * umbrella,bool isBundleLoader,bool explicitlyLinked)224 DylibFile *macho::loadDylib(MemoryBufferRef mbref, DylibFile *umbrella,
225 bool isBundleLoader, bool explicitlyLinked) {
226 CachedHashStringRef path(mbref.getBufferIdentifier());
227 DylibFile *&file = loadedDylibs[path];
228 if (file) {
229 if (explicitlyLinked)
230 file->setExplicitlyLinked();
231 return file;
232 }
233
234 DylibFile *newFile;
235 file_magic magic = identify_magic(mbref.getBuffer());
236 if (magic == file_magic::tapi_file) {
237 Expected<std::unique_ptr<InterfaceFile>> result = TextAPIReader::get(mbref);
238 if (!result) {
239 error("could not load TAPI file at " + mbref.getBufferIdentifier() +
240 ": " + toString(result.takeError()));
241 return nullptr;
242 }
243 file =
244 make<DylibFile>(**result, umbrella, isBundleLoader, explicitlyLinked);
245
246 // parseReexports() can recursively call loadDylib(). That's fine since
247 // we wrote the DylibFile we just loaded to the loadDylib cache via the
248 // `file` reference. But the recursive load can grow loadDylibs, so the
249 // `file` reference might become invalid after parseReexports() -- so copy
250 // the pointer it refers to before continuing.
251 newFile = file;
252 if (newFile->exportingFile)
253 newFile->parseReexports(**result);
254 } else {
255 assert(magic == file_magic::macho_dynamically_linked_shared_lib ||
256 magic == file_magic::macho_dynamically_linked_shared_lib_stub ||
257 magic == file_magic::macho_executable ||
258 magic == file_magic::macho_bundle);
259 file = make<DylibFile>(mbref, umbrella, isBundleLoader, explicitlyLinked);
260
261 // parseLoadCommands() can also recursively call loadDylib(). See comment
262 // in previous block for why this means we must copy `file` here.
263 newFile = file;
264 if (newFile->exportingFile)
265 newFile->parseLoadCommands(mbref);
266 }
267 return newFile;
268 }
269
resetLoadedDylibs()270 void macho::resetLoadedDylibs() { loadedDylibs.clear(); }
271
272 std::optional<StringRef>
findPathCombination(const Twine & name,const std::vector<StringRef> & roots,ArrayRef<StringRef> extensions)273 macho::findPathCombination(const Twine &name,
274 const std::vector<StringRef> &roots,
275 ArrayRef<StringRef> extensions) {
276 SmallString<261> base;
277 for (StringRef dir : roots) {
278 base = dir;
279 path::append(base, name);
280 for (StringRef ext : extensions) {
281 Twine location = base + ext;
282 bool exists = fs::exists(location);
283 searchedDylib(location, exists);
284 if (exists)
285 return saver().save(location.str());
286 }
287 }
288 return {};
289 }
290
rerootPath(StringRef path)291 StringRef macho::rerootPath(StringRef path) {
292 if (!path::is_absolute(path, path::Style::posix) || path.ends_with(".o"))
293 return path;
294
295 if (std::optional<StringRef> rerootedPath =
296 findPathCombination(path, config->systemLibraryRoots))
297 return *rerootedPath;
298
299 return path;
300 }
301
getModTime(StringRef path)302 uint32_t macho::getModTime(StringRef path) {
303 if (config->zeroModTime)
304 return 0;
305
306 fs::file_status stat;
307 if (!fs::status(path, stat))
308 if (fs::exists(stat))
309 return toTimeT(stat.getLastModificationTime());
310
311 warn("failed to get modification time of " + path);
312 return 0;
313 }
314
printArchiveMemberLoad(StringRef reason,const InputFile * f)315 void macho::printArchiveMemberLoad(StringRef reason, const InputFile *f) {
316 if (config->printEachFile)
317 message(toString(f));
318 if (config->printWhyLoad)
319 message(reason + " forced load of " + toString(f));
320 }
321
DependencyTracker(StringRef path)322 macho::DependencyTracker::DependencyTracker(StringRef path)
323 : path(path), active(!path.empty()) {
324 if (active && fs::exists(path) && !fs::can_write(path)) {
325 warn("Ignoring dependency_info option since specified path is not "
326 "writeable.");
327 active = false;
328 }
329 }
330
write(StringRef version,const SetVector<InputFile * > & inputs,StringRef output)331 void macho::DependencyTracker::write(StringRef version,
332 const SetVector<InputFile *> &inputs,
333 StringRef output) {
334 if (!active)
335 return;
336
337 std::error_code ec;
338 raw_fd_ostream os(path, ec, fs::OF_None);
339 if (ec) {
340 warn("Error writing dependency info to file");
341 return;
342 }
343
344 auto addDep = [&os](DepOpCode opcode, const StringRef &path) {
345 // XXX: Even though DepOpCode's underlying type is uint8_t,
346 // this cast is still needed because Clang older than 10.x has a bug,
347 // where it doesn't know to cast the enum to its underlying type.
348 // Hence `<< DepOpCode` is ambiguous to it.
349 os << static_cast<uint8_t>(opcode);
350 os << path;
351 os << '\0';
352 };
353
354 addDep(DepOpCode::Version, version);
355
356 // Sort the input by its names.
357 std::vector<StringRef> inputNames;
358 inputNames.reserve(inputs.size());
359 for (InputFile *f : inputs)
360 inputNames.push_back(f->getName());
361 llvm::sort(inputNames);
362
363 for (const StringRef &in : inputNames)
364 addDep(DepOpCode::Input, in);
365
366 for (const std::string &f : notFounds)
367 addDep(DepOpCode::NotFound, f);
368
369 addDep(DepOpCode::Output, output);
370 }
371