xref: /freebsd/contrib/llvm-project/clang/tools/driver/driver.cpp (revision 85868e8a1daeaae7a0e48effb2ea2310ae3b02c6)
1 //===-- driver.cpp - Clang GCC-Compatible Driver --------------------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This is the entry point to the clang driver; it is a thin wrapper
10 // for functionality in the Driver clang library.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "clang/Driver/Driver.h"
15 #include "clang/Basic/DiagnosticOptions.h"
16 #include "clang/Basic/Stack.h"
17 #include "clang/Driver/Compilation.h"
18 #include "clang/Driver/DriverDiagnostic.h"
19 #include "clang/Driver/Options.h"
20 #include "clang/Driver/ToolChain.h"
21 #include "clang/Frontend/ChainedDiagnosticConsumer.h"
22 #include "clang/Frontend/CompilerInvocation.h"
23 #include "clang/Frontend/SerializedDiagnosticPrinter.h"
24 #include "clang/Frontend/TextDiagnosticPrinter.h"
25 #include "clang/Frontend/Utils.h"
26 #include "llvm/ADT/ArrayRef.h"
27 #include "llvm/ADT/SmallString.h"
28 #include "llvm/ADT/SmallVector.h"
29 #include "llvm/Option/ArgList.h"
30 #include "llvm/Option/OptTable.h"
31 #include "llvm/Option/Option.h"
32 #include "llvm/Support/CommandLine.h"
33 #include "llvm/Support/ErrorHandling.h"
34 #include "llvm/Support/FileSystem.h"
35 #include "llvm/Support/Host.h"
36 #include "llvm/Support/InitLLVM.h"
37 #include "llvm/Support/Path.h"
38 #include "llvm/Support/Process.h"
39 #include "llvm/Support/Program.h"
40 #include "llvm/Support/Regex.h"
41 #include "llvm/Support/Signals.h"
42 #include "llvm/Support/StringSaver.h"
43 #include "llvm/Support/TargetSelect.h"
44 #include "llvm/Support/Timer.h"
45 #include "llvm/Support/raw_ostream.h"
46 #include <memory>
47 #include <set>
48 #include <system_error>
49 using namespace clang;
50 using namespace clang::driver;
51 using namespace llvm::opt;
52 
53 std::string GetExecutablePath(const char *Argv0, bool CanonicalPrefixes) {
54   if (!CanonicalPrefixes) {
55     SmallString<128> ExecutablePath(Argv0);
56     // Do a PATH lookup if Argv0 isn't a valid path.
57     if (!llvm::sys::fs::exists(ExecutablePath))
58       if (llvm::ErrorOr<std::string> P =
59               llvm::sys::findProgramByName(ExecutablePath))
60         ExecutablePath = *P;
61     return ExecutablePath.str();
62   }
63 
64   // This just needs to be some symbol in the binary; C++ doesn't
65   // allow taking the address of ::main however.
66   void *P = (void*) (intptr_t) GetExecutablePath;
67   return llvm::sys::fs::getMainExecutable(Argv0, P);
68 }
69 
70 static const char *GetStableCStr(std::set<std::string> &SavedStrings,
71                                  StringRef S) {
72   return SavedStrings.insert(S).first->c_str();
73 }
74 
75 /// ApplyQAOverride - Apply a list of edits to the input argument lists.
76 ///
77 /// The input string is a space separate list of edits to perform,
78 /// they are applied in order to the input argument lists. Edits
79 /// should be one of the following forms:
80 ///
81 ///  '#': Silence information about the changes to the command line arguments.
82 ///
83 ///  '^': Add FOO as a new argument at the beginning of the command line.
84 ///
85 ///  '+': Add FOO as a new argument at the end of the command line.
86 ///
87 ///  's/XXX/YYY/': Substitute the regular expression XXX with YYY in the command
88 ///  line.
89 ///
90 ///  'xOPTION': Removes all instances of the literal argument OPTION.
91 ///
92 ///  'XOPTION': Removes all instances of the literal argument OPTION,
93 ///  and the following argument.
94 ///
95 ///  'Ox': Removes all flags matching 'O' or 'O[sz0-9]' and adds 'Ox'
96 ///  at the end of the command line.
97 ///
98 /// \param OS - The stream to write edit information to.
99 /// \param Args - The vector of command line arguments.
100 /// \param Edit - The override command to perform.
101 /// \param SavedStrings - Set to use for storing string representations.
102 static void ApplyOneQAOverride(raw_ostream &OS,
103                                SmallVectorImpl<const char*> &Args,
104                                StringRef Edit,
105                                std::set<std::string> &SavedStrings) {
106   // This does not need to be efficient.
107 
108   if (Edit[0] == '^') {
109     const char *Str =
110       GetStableCStr(SavedStrings, Edit.substr(1));
111     OS << "### Adding argument " << Str << " at beginning\n";
112     Args.insert(Args.begin() + 1, Str);
113   } else if (Edit[0] == '+') {
114     const char *Str =
115       GetStableCStr(SavedStrings, Edit.substr(1));
116     OS << "### Adding argument " << Str << " at end\n";
117     Args.push_back(Str);
118   } else if (Edit[0] == 's' && Edit[1] == '/' && Edit.endswith("/") &&
119              Edit.slice(2, Edit.size()-1).find('/') != StringRef::npos) {
120     StringRef MatchPattern = Edit.substr(2).split('/').first;
121     StringRef ReplPattern = Edit.substr(2).split('/').second;
122     ReplPattern = ReplPattern.slice(0, ReplPattern.size()-1);
123 
124     for (unsigned i = 1, e = Args.size(); i != e; ++i) {
125       // Ignore end-of-line response file markers
126       if (Args[i] == nullptr)
127         continue;
128       std::string Repl = llvm::Regex(MatchPattern).sub(ReplPattern, Args[i]);
129 
130       if (Repl != Args[i]) {
131         OS << "### Replacing '" << Args[i] << "' with '" << Repl << "'\n";
132         Args[i] = GetStableCStr(SavedStrings, Repl);
133       }
134     }
135   } else if (Edit[0] == 'x' || Edit[0] == 'X') {
136     auto Option = Edit.substr(1);
137     for (unsigned i = 1; i < Args.size();) {
138       if (Option == Args[i]) {
139         OS << "### Deleting argument " << Args[i] << '\n';
140         Args.erase(Args.begin() + i);
141         if (Edit[0] == 'X') {
142           if (i < Args.size()) {
143             OS << "### Deleting argument " << Args[i] << '\n';
144             Args.erase(Args.begin() + i);
145           } else
146             OS << "### Invalid X edit, end of command line!\n";
147         }
148       } else
149         ++i;
150     }
151   } else if (Edit[0] == 'O') {
152     for (unsigned i = 1; i < Args.size();) {
153       const char *A = Args[i];
154       // Ignore end-of-line response file markers
155       if (A == nullptr)
156         continue;
157       if (A[0] == '-' && A[1] == 'O' &&
158           (A[2] == '\0' ||
159            (A[3] == '\0' && (A[2] == 's' || A[2] == 'z' ||
160                              ('0' <= A[2] && A[2] <= '9'))))) {
161         OS << "### Deleting argument " << Args[i] << '\n';
162         Args.erase(Args.begin() + i);
163       } else
164         ++i;
165     }
166     OS << "### Adding argument " << Edit << " at end\n";
167     Args.push_back(GetStableCStr(SavedStrings, '-' + Edit.str()));
168   } else {
169     OS << "### Unrecognized edit: " << Edit << "\n";
170   }
171 }
172 
173 /// ApplyQAOverride - Apply a comma separate list of edits to the
174 /// input argument lists. See ApplyOneQAOverride.
175 static void ApplyQAOverride(SmallVectorImpl<const char*> &Args,
176                             const char *OverrideStr,
177                             std::set<std::string> &SavedStrings) {
178   raw_ostream *OS = &llvm::errs();
179 
180   if (OverrideStr[0] == '#') {
181     ++OverrideStr;
182     OS = &llvm::nulls();
183   }
184 
185   *OS << "### CCC_OVERRIDE_OPTIONS: " << OverrideStr << "\n";
186 
187   // This does not need to be efficient.
188 
189   const char *S = OverrideStr;
190   while (*S) {
191     const char *End = ::strchr(S, ' ');
192     if (!End)
193       End = S + strlen(S);
194     if (End != S)
195       ApplyOneQAOverride(*OS, Args, std::string(S, End), SavedStrings);
196     S = End;
197     if (*S != '\0')
198       ++S;
199   }
200 }
201 
202 extern int cc1_main(ArrayRef<const char *> Argv, const char *Argv0,
203                     void *MainAddr);
204 extern int cc1as_main(ArrayRef<const char *> Argv, const char *Argv0,
205                       void *MainAddr);
206 extern int cc1gen_reproducer_main(ArrayRef<const char *> Argv,
207                                   const char *Argv0, void *MainAddr);
208 
209 static void insertTargetAndModeArgs(const ParsedClangName &NameParts,
210                                     SmallVectorImpl<const char *> &ArgVector,
211                                     std::set<std::string> &SavedStrings) {
212   // Put target and mode arguments at the start of argument list so that
213   // arguments specified in command line could override them. Avoid putting
214   // them at index 0, as an option like '-cc1' must remain the first.
215   int InsertionPoint = 0;
216   if (ArgVector.size() > 0)
217     ++InsertionPoint;
218 
219   if (NameParts.DriverMode) {
220     // Add the mode flag to the arguments.
221     ArgVector.insert(ArgVector.begin() + InsertionPoint,
222                      GetStableCStr(SavedStrings, NameParts.DriverMode));
223   }
224 
225   if (NameParts.TargetIsValid) {
226     const char *arr[] = {"-target", GetStableCStr(SavedStrings,
227                                                   NameParts.TargetPrefix)};
228     ArgVector.insert(ArgVector.begin() + InsertionPoint,
229                      std::begin(arr), std::end(arr));
230   }
231 }
232 
233 static void getCLEnvVarOptions(std::string &EnvValue, llvm::StringSaver &Saver,
234                                SmallVectorImpl<const char *> &Opts) {
235   llvm::cl::TokenizeWindowsCommandLine(EnvValue, Saver, Opts);
236   // The first instance of '#' should be replaced with '=' in each option.
237   for (const char *Opt : Opts)
238     if (char *NumberSignPtr = const_cast<char *>(::strchr(Opt, '#')))
239       *NumberSignPtr = '=';
240 }
241 
242 static void SetBackdoorDriverOutputsFromEnvVars(Driver &TheDriver) {
243   // Handle CC_PRINT_OPTIONS and CC_PRINT_OPTIONS_FILE.
244   TheDriver.CCPrintOptions = !!::getenv("CC_PRINT_OPTIONS");
245   if (TheDriver.CCPrintOptions)
246     TheDriver.CCPrintOptionsFilename = ::getenv("CC_PRINT_OPTIONS_FILE");
247 
248   // Handle CC_PRINT_HEADERS and CC_PRINT_HEADERS_FILE.
249   TheDriver.CCPrintHeaders = !!::getenv("CC_PRINT_HEADERS");
250   if (TheDriver.CCPrintHeaders)
251     TheDriver.CCPrintHeadersFilename = ::getenv("CC_PRINT_HEADERS_FILE");
252 
253   // Handle CC_LOG_DIAGNOSTICS and CC_LOG_DIAGNOSTICS_FILE.
254   TheDriver.CCLogDiagnostics = !!::getenv("CC_LOG_DIAGNOSTICS");
255   if (TheDriver.CCLogDiagnostics)
256     TheDriver.CCLogDiagnosticsFilename = ::getenv("CC_LOG_DIAGNOSTICS_FILE");
257 }
258 
259 static void FixupDiagPrefixExeName(TextDiagnosticPrinter *DiagClient,
260                                    const std::string &Path) {
261   // If the clang binary happens to be named cl.exe for compatibility reasons,
262   // use clang-cl.exe as the prefix to avoid confusion between clang and MSVC.
263   StringRef ExeBasename(llvm::sys::path::stem(Path));
264   if (ExeBasename.equals_lower("cl"))
265     ExeBasename = "clang-cl";
266   DiagClient->setPrefix(ExeBasename);
267 }
268 
269 // This lets us create the DiagnosticsEngine with a properly-filled-out
270 // DiagnosticOptions instance.
271 static DiagnosticOptions *
272 CreateAndPopulateDiagOpts(ArrayRef<const char *> argv) {
273   auto *DiagOpts = new DiagnosticOptions;
274   unsigned MissingArgIndex, MissingArgCount;
275   InputArgList Args = getDriverOptTable().ParseArgs(
276       argv.slice(1), MissingArgIndex, MissingArgCount);
277   // We ignore MissingArgCount and the return value of ParseDiagnosticArgs.
278   // Any errors that would be diagnosed here will also be diagnosed later,
279   // when the DiagnosticsEngine actually exists.
280   (void)ParseDiagnosticArgs(*DiagOpts, Args);
281   return DiagOpts;
282 }
283 
284 static void SetInstallDir(SmallVectorImpl<const char *> &argv,
285                           Driver &TheDriver, bool CanonicalPrefixes) {
286   // Attempt to find the original path used to invoke the driver, to determine
287   // the installed path. We do this manually, because we want to support that
288   // path being a symlink.
289   SmallString<128> InstalledPath(argv[0]);
290 
291   // Do a PATH lookup, if there are no directory components.
292   if (llvm::sys::path::filename(InstalledPath) == InstalledPath)
293     if (llvm::ErrorOr<std::string> Tmp = llvm::sys::findProgramByName(
294             llvm::sys::path::filename(InstalledPath.str())))
295       InstalledPath = *Tmp;
296 
297   // FIXME: We don't actually canonicalize this, we just make it absolute.
298   if (CanonicalPrefixes)
299     llvm::sys::fs::make_absolute(InstalledPath);
300 
301   StringRef InstalledPathParent(llvm::sys::path::parent_path(InstalledPath));
302   if (llvm::sys::fs::exists(InstalledPathParent))
303     TheDriver.setInstalledDir(InstalledPathParent);
304 }
305 
306 static int ExecuteCC1Tool(ArrayRef<const char *> argv, StringRef Tool) {
307   void *GetExecutablePathVP = (void *)(intptr_t) GetExecutablePath;
308   if (Tool == "")
309     return cc1_main(argv.slice(2), argv[0], GetExecutablePathVP);
310   if (Tool == "as")
311     return cc1as_main(argv.slice(2), argv[0], GetExecutablePathVP);
312   if (Tool == "gen-reproducer")
313     return cc1gen_reproducer_main(argv.slice(2), argv[0], GetExecutablePathVP);
314 
315   // Reject unknown tools.
316   llvm::errs() << "error: unknown integrated tool '" << Tool << "'. "
317                << "Valid tools include '-cc1' and '-cc1as'.\n";
318   return 1;
319 }
320 
321 int main(int argc_, const char **argv_) {
322   noteBottomOfStack();
323   llvm::InitLLVM X(argc_, argv_);
324   SmallVector<const char *, 256> argv(argv_, argv_ + argc_);
325 
326   if (llvm::sys::Process::FixupStandardFileDescriptors())
327     return 1;
328 
329   llvm::InitializeAllTargets();
330   auto TargetAndMode = ToolChain::getTargetAndModeFromProgramName(argv[0]);
331 
332   llvm::BumpPtrAllocator A;
333   llvm::StringSaver Saver(A);
334 
335   // Parse response files using the GNU syntax, unless we're in CL mode. There
336   // are two ways to put clang in CL compatibility mode: argv[0] is either
337   // clang-cl or cl, or --driver-mode=cl is on the command line. The normal
338   // command line parsing can't happen until after response file parsing, so we
339   // have to manually search for a --driver-mode=cl argument the hard way.
340   // Finally, our -cc1 tools don't care which tokenization mode we use because
341   // response files written by clang will tokenize the same way in either mode.
342   bool ClangCLMode = false;
343   if (StringRef(TargetAndMode.DriverMode).equals("--driver-mode=cl") ||
344       llvm::find_if(argv, [](const char *F) {
345         return F && strcmp(F, "--driver-mode=cl") == 0;
346       }) != argv.end()) {
347     ClangCLMode = true;
348   }
349   enum { Default, POSIX, Windows } RSPQuoting = Default;
350   for (const char *F : argv) {
351     if (strcmp(F, "--rsp-quoting=posix") == 0)
352       RSPQuoting = POSIX;
353     else if (strcmp(F, "--rsp-quoting=windows") == 0)
354       RSPQuoting = Windows;
355   }
356 
357   // Determines whether we want nullptr markers in argv to indicate response
358   // files end-of-lines. We only use this for the /LINK driver argument with
359   // clang-cl.exe on Windows.
360   bool MarkEOLs = ClangCLMode;
361 
362   llvm::cl::TokenizerCallback Tokenizer;
363   if (RSPQuoting == Windows || (RSPQuoting == Default && ClangCLMode))
364     Tokenizer = &llvm::cl::TokenizeWindowsCommandLine;
365   else
366     Tokenizer = &llvm::cl::TokenizeGNUCommandLine;
367 
368   if (MarkEOLs && argv.size() > 1 && StringRef(argv[1]).startswith("-cc1"))
369     MarkEOLs = false;
370   llvm::cl::ExpandResponseFiles(Saver, Tokenizer, argv, MarkEOLs);
371 
372   // Handle -cc1 integrated tools, even if -cc1 was expanded from a response
373   // file.
374   auto FirstArg = std::find_if(argv.begin() + 1, argv.end(),
375                                [](const char *A) { return A != nullptr; });
376   if (FirstArg != argv.end() && StringRef(*FirstArg).startswith("-cc1")) {
377     // If -cc1 came from a response file, remove the EOL sentinels.
378     if (MarkEOLs) {
379       auto newEnd = std::remove(argv.begin(), argv.end(), nullptr);
380       argv.resize(newEnd - argv.begin());
381     }
382     return ExecuteCC1Tool(argv, argv[1] + 4);
383   }
384 
385   bool CanonicalPrefixes = true;
386   for (int i = 1, size = argv.size(); i < size; ++i) {
387     // Skip end-of-line response file markers
388     if (argv[i] == nullptr)
389       continue;
390     if (StringRef(argv[i]) == "-no-canonical-prefixes") {
391       CanonicalPrefixes = false;
392       break;
393     }
394   }
395 
396   // Handle CL and _CL_ which permits additional command line options to be
397   // prepended or appended.
398   if (ClangCLMode) {
399     // Arguments in "CL" are prepended.
400     llvm::Optional<std::string> OptCL = llvm::sys::Process::GetEnv("CL");
401     if (OptCL.hasValue()) {
402       SmallVector<const char *, 8> PrependedOpts;
403       getCLEnvVarOptions(OptCL.getValue(), Saver, PrependedOpts);
404 
405       // Insert right after the program name to prepend to the argument list.
406       argv.insert(argv.begin() + 1, PrependedOpts.begin(), PrependedOpts.end());
407     }
408     // Arguments in "_CL_" are appended.
409     llvm::Optional<std::string> Opt_CL_ = llvm::sys::Process::GetEnv("_CL_");
410     if (Opt_CL_.hasValue()) {
411       SmallVector<const char *, 8> AppendedOpts;
412       getCLEnvVarOptions(Opt_CL_.getValue(), Saver, AppendedOpts);
413 
414       // Insert at the end of the argument list to append.
415       argv.append(AppendedOpts.begin(), AppendedOpts.end());
416     }
417   }
418 
419   std::set<std::string> SavedStrings;
420   // Handle CCC_OVERRIDE_OPTIONS, used for editing a command line behind the
421   // scenes.
422   if (const char *OverrideStr = ::getenv("CCC_OVERRIDE_OPTIONS")) {
423     // FIXME: Driver shouldn't take extra initial argument.
424     ApplyQAOverride(argv, OverrideStr, SavedStrings);
425   }
426 
427   std::string Path = GetExecutablePath(argv[0], CanonicalPrefixes);
428 
429   IntrusiveRefCntPtr<DiagnosticOptions> DiagOpts =
430       CreateAndPopulateDiagOpts(argv);
431 
432   TextDiagnosticPrinter *DiagClient
433     = new TextDiagnosticPrinter(llvm::errs(), &*DiagOpts);
434   FixupDiagPrefixExeName(DiagClient, Path);
435 
436   IntrusiveRefCntPtr<DiagnosticIDs> DiagID(new DiagnosticIDs());
437 
438   DiagnosticsEngine Diags(DiagID, &*DiagOpts, DiagClient);
439 
440   if (!DiagOpts->DiagnosticSerializationFile.empty()) {
441     auto SerializedConsumer =
442         clang::serialized_diags::create(DiagOpts->DiagnosticSerializationFile,
443                                         &*DiagOpts, /*MergeChildRecords=*/true);
444     Diags.setClient(new ChainedDiagnosticConsumer(
445         Diags.takeClient(), std::move(SerializedConsumer)));
446   }
447 
448   ProcessWarningOptions(Diags, *DiagOpts, /*ReportDiags=*/false);
449 
450   Driver TheDriver(Path, llvm::sys::getDefaultTargetTriple(), Diags);
451   SetInstallDir(argv, TheDriver, CanonicalPrefixes);
452   TheDriver.setTargetAndMode(TargetAndMode);
453 
454   insertTargetAndModeArgs(TargetAndMode, argv, SavedStrings);
455 
456   SetBackdoorDriverOutputsFromEnvVars(TheDriver);
457 
458   std::unique_ptr<Compilation> C(TheDriver.BuildCompilation(argv));
459   int Res = 1;
460   if (C && !C->containsError()) {
461     SmallVector<std::pair<int, const Command *>, 4> FailingCommands;
462     Res = TheDriver.ExecuteCompilation(*C, FailingCommands);
463 
464     // Force a crash to test the diagnostics.
465     if (TheDriver.GenReproducer) {
466       Diags.Report(diag::err_drv_force_crash)
467         << !::getenv("FORCE_CLANG_DIAGNOSTICS_CRASH");
468 
469       // Pretend that every command failed.
470       FailingCommands.clear();
471       for (const auto &J : C->getJobs())
472         if (const Command *C = dyn_cast<Command>(&J))
473           FailingCommands.push_back(std::make_pair(-1, C));
474     }
475 
476     for (const auto &P : FailingCommands) {
477       int CommandRes = P.first;
478       const Command *FailingCommand = P.second;
479       if (!Res)
480         Res = CommandRes;
481 
482       // If result status is < 0, then the driver command signalled an error.
483       // If result status is 70, then the driver command reported a fatal error.
484       // On Windows, abort will return an exit code of 3.  In these cases,
485       // generate additional diagnostic information if possible.
486       bool DiagnoseCrash = CommandRes < 0 || CommandRes == 70;
487 #ifdef _WIN32
488       DiagnoseCrash |= CommandRes == 3;
489 #endif
490       if (DiagnoseCrash) {
491         TheDriver.generateCompilationDiagnostics(*C, *FailingCommand);
492         break;
493       }
494     }
495   }
496 
497   Diags.getClient()->finish();
498 
499   // If any timers were active but haven't been destroyed yet, print their
500   // results now.  This happens in -disable-free mode.
501   llvm::TimerGroup::printAll(llvm::errs());
502   llvm::TimerGroup::clearAll();
503 
504 #ifdef _WIN32
505   // Exit status should not be negative on Win32, unless abnormal termination.
506   // Once abnormal termiation was caught, negative status should not be
507   // propagated.
508   if (Res < 0)
509     Res = 1;
510 #endif
511 
512   // If we have multiple failing commands, we return the result of the first
513   // failing command.
514   return Res;
515 }
516