1 //===-- llvm-lto2: test harness for the resolution-based LTO interface ----===//
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 program takes in a list of bitcode files, links them and performs
10 // link-time optimization according to the provided symbol resolutions using the
11 // resolution-based LTO interface, and outputs one or more object files.
12 //
13 // This program is intended to eventually replace llvm-lto which uses the legacy
14 // LTO interface.
15 //
16 //===----------------------------------------------------------------------===//
17 
18 #include "llvm/Bitcode/BitcodeReader.h"
19 #include "llvm/CodeGen/CommandFlags.h"
20 #include "llvm/Config/llvm-config.h"
21 #include "llvm/IR/DiagnosticPrinter.h"
22 #include "llvm/LTO/Caching.h"
23 #include "llvm/LTO/LTO.h"
24 #include "llvm/Passes/PassPlugin.h"
25 #include "llvm/Remarks/HotnessThresholdParser.h"
26 #include "llvm/Support/CommandLine.h"
27 #include "llvm/Support/FileSystem.h"
28 #include "llvm/Support/InitLLVM.h"
29 #include "llvm/Support/PluginLoader.h"
30 #include "llvm/Support/TargetSelect.h"
31 #include "llvm/Support/Threading.h"
32 
33 using namespace llvm;
34 using namespace lto;
35 
36 static codegen::RegisterCodeGenFlags CGF;
37 
38 static cl::opt<char>
39     OptLevel("O", cl::desc("Optimization level. [-O0, -O1, -O2, or -O3] "
40                            "(default = '-O2')"),
41              cl::Prefix, cl::ZeroOrMore, cl::init('2'));
42 
43 static cl::opt<char> CGOptLevel(
44     "cg-opt-level",
45     cl::desc("Codegen optimization level (0, 1, 2 or 3, default = '2')"),
46     cl::init('2'));
47 
48 static cl::list<std::string> InputFilenames(cl::Positional, cl::OneOrMore,
49                                             cl::desc("<input bitcode files>"));
50 
51 static cl::opt<std::string> OutputFilename("o", cl::Required,
52                                            cl::desc("Output filename"),
53                                            cl::value_desc("filename"));
54 
55 static cl::opt<std::string> CacheDir("cache-dir", cl::desc("Cache Directory"),
56                                      cl::value_desc("directory"));
57 
58 static cl::opt<std::string> OptPipeline("opt-pipeline",
59                                         cl::desc("Optimizer Pipeline"),
60                                         cl::value_desc("pipeline"));
61 
62 static cl::opt<std::string> AAPipeline("aa-pipeline",
63                                        cl::desc("Alias Analysis Pipeline"),
64                                        cl::value_desc("aapipeline"));
65 
66 static cl::opt<bool> SaveTemps("save-temps", cl::desc("Save temporary files"));
67 
68 static cl::opt<bool>
69     ThinLTODistributedIndexes("thinlto-distributed-indexes", cl::init(false),
70                               cl::desc("Write out individual index and "
71                                        "import files for the "
72                                        "distributed backend case"));
73 
74 // Default to using all available threads in the system, but using only one
75 // thread per core (no SMT).
76 // Use -thinlto-threads=all to use hardware_concurrency() instead, which means
77 // to use all hardware threads or cores in the system.
78 static cl::opt<std::string> Threads("thinlto-threads");
79 
80 static cl::list<std::string> SymbolResolutions(
81     "r",
82     cl::desc("Specify a symbol resolution: filename,symbolname,resolution\n"
83              "where \"resolution\" is a sequence (which may be empty) of the\n"
84              "following characters:\n"
85              " p - prevailing: the linker has chosen this definition of the\n"
86              "     symbol\n"
87              " l - local: the definition of this symbol is unpreemptable at\n"
88              "     runtime and is known to be in this linkage unit\n"
89              " x - externally visible: the definition of this symbol is\n"
90              "     visible outside of the LTO unit\n"
91              "A resolution for each symbol must be specified."),
92     cl::ZeroOrMore);
93 
94 static cl::opt<std::string> OverrideTriple(
95     "override-triple",
96     cl::desc("Replace target triples in input files with this triple"));
97 
98 static cl::opt<std::string> DefaultTriple(
99     "default-triple",
100     cl::desc(
101         "Replace unspecified target triples in input files with this triple"));
102 
103 static cl::opt<bool> RemarksWithHotness(
104     "pass-remarks-with-hotness",
105     cl::desc("With PGO, include profile count in optimization remarks"),
106     cl::Hidden);
107 
108 cl::opt<Optional<uint64_t>, false, remarks::HotnessThresholdParser>
109     RemarksHotnessThreshold(
110         "pass-remarks-hotness-threshold",
111         cl::desc("Minimum profile count required for an "
112                  "optimization remark to be output."
113                  " Use 'auto' to apply the threshold from profile summary."),
114         cl::value_desc("uint or 'auto'"), cl::init(0), cl::Hidden);
115 
116 static cl::opt<std::string>
117     RemarksFilename("pass-remarks-output",
118                     cl::desc("Output filename for pass remarks"),
119                     cl::value_desc("filename"));
120 
121 static cl::opt<std::string>
122     RemarksPasses("pass-remarks-filter",
123                   cl::desc("Only record optimization remarks from passes whose "
124                            "names match the given regular expression"),
125                   cl::value_desc("regex"));
126 
127 static cl::opt<std::string> RemarksFormat(
128     "pass-remarks-format",
129     cl::desc("The format used for serializing remarks (default: YAML)"),
130     cl::value_desc("format"), cl::init("yaml"));
131 
132 static cl::opt<std::string>
133     SamplePGOFile("lto-sample-profile-file",
134                   cl::desc("Specify a SamplePGO profile file"));
135 
136 static cl::opt<std::string>
137     CSPGOFile("lto-cspgo-profile-file",
138               cl::desc("Specify a context sensitive PGO profile file"));
139 
140 static cl::opt<bool>
141     RunCSIRInstr("lto-cspgo-gen",
142                  cl::desc("Run PGO context sensitive IR instrumentation"),
143                  cl::init(false), cl::Hidden);
144 
145 static cl::opt<bool>
146     UseNewPM("use-new-pm",
147              cl::desc("Run LTO passes using the new pass manager"),
148              cl::init(LLVM_ENABLE_NEW_PASS_MANAGER), cl::Hidden);
149 
150 static cl::opt<bool>
151     DebugPassManager("debug-pass-manager", cl::init(false), cl::Hidden,
152                      cl::desc("Print pass management debugging information"));
153 
154 static cl::opt<std::string>
155     StatsFile("stats-file", cl::desc("Filename to write statistics to"));
156 
157 static cl::list<std::string>
158     PassPlugins("load-pass-plugin",
159                 cl::desc("Load passes from plugin library"));
160 
161 static cl::opt<bool> EnableFreestanding(
162     "lto-freestanding",
163     cl::desc("Enable Freestanding (disable builtins / TLI) during LTO"),
164     cl::init(false), cl::Hidden);
165 
166 static void check(Error E, std::string Msg) {
167   if (!E)
168     return;
169   handleAllErrors(std::move(E), [&](ErrorInfoBase &EIB) {
170     errs() << "llvm-lto2: " << Msg << ": " << EIB.message().c_str() << '\n';
171   });
172   exit(1);
173 }
174 
175 template <typename T> static T check(Expected<T> E, std::string Msg) {
176   if (E)
177     return std::move(*E);
178   check(E.takeError(), Msg);
179   return T();
180 }
181 
182 static void check(std::error_code EC, std::string Msg) {
183   check(errorCodeToError(EC), Msg);
184 }
185 
186 template <typename T> static T check(ErrorOr<T> E, std::string Msg) {
187   if (E)
188     return std::move(*E);
189   check(E.getError(), Msg);
190   return T();
191 }
192 
193 static int usage() {
194   errs() << "Available subcommands: dump-symtab run\n";
195   return 1;
196 }
197 
198 static int run(int argc, char **argv) {
199   cl::ParseCommandLineOptions(argc, argv, "Resolution-based LTO test harness");
200 
201   // FIXME: Workaround PR30396 which means that a symbol can appear
202   // more than once if it is defined in module-level assembly and
203   // has a GV declaration. We allow (file, symbol) pairs to have multiple
204   // resolutions and apply them in the order observed.
205   std::map<std::pair<std::string, std::string>, std::list<SymbolResolution>>
206       CommandLineResolutions;
207   for (std::string R : SymbolResolutions) {
208     StringRef Rest = R;
209     StringRef FileName, SymbolName;
210     std::tie(FileName, Rest) = Rest.split(',');
211     if (Rest.empty()) {
212       llvm::errs() << "invalid resolution: " << R << '\n';
213       return 1;
214     }
215     std::tie(SymbolName, Rest) = Rest.split(',');
216     SymbolResolution Res;
217     for (char C : Rest) {
218       if (C == 'p')
219         Res.Prevailing = true;
220       else if (C == 'l')
221         Res.FinalDefinitionInLinkageUnit = true;
222       else if (C == 'x')
223         Res.VisibleToRegularObj = true;
224       else if (C == 'r')
225         Res.LinkerRedefined = true;
226       else {
227         llvm::errs() << "invalid character " << C << " in resolution: " << R
228                      << '\n';
229         return 1;
230       }
231     }
232     CommandLineResolutions[{std::string(FileName), std::string(SymbolName)}]
233         .push_back(Res);
234   }
235 
236   std::vector<std::unique_ptr<MemoryBuffer>> MBs;
237 
238   Config Conf;
239   Conf.DiagHandler = [](const DiagnosticInfo &DI) {
240     DiagnosticPrinterRawOStream DP(errs());
241     DI.print(DP);
242     errs() << '\n';
243     if (DI.getSeverity() == DS_Error)
244       exit(1);
245   };
246 
247   Conf.CPU = codegen::getMCPU();
248   Conf.Options = codegen::InitTargetOptionsFromCodeGenFlags(Triple());
249   Conf.MAttrs = codegen::getMAttrs();
250   if (auto RM = codegen::getExplicitRelocModel())
251     Conf.RelocModel = RM.getValue();
252   Conf.CodeModel = codegen::getExplicitCodeModel();
253 
254   Conf.DebugPassManager = DebugPassManager;
255 
256   if (SaveTemps)
257     check(Conf.addSaveTemps(OutputFilename + "."),
258           "Config::addSaveTemps failed");
259 
260   // Optimization remarks.
261   Conf.RemarksFilename = RemarksFilename;
262   Conf.RemarksPasses = RemarksPasses;
263   Conf.RemarksWithHotness = RemarksWithHotness;
264   Conf.RemarksHotnessThreshold = RemarksHotnessThreshold;
265   Conf.RemarksFormat = RemarksFormat;
266 
267   Conf.SampleProfile = SamplePGOFile;
268   Conf.CSIRProfile = CSPGOFile;
269   Conf.RunCSIRInstr = RunCSIRInstr;
270 
271   // Run a custom pipeline, if asked for.
272   Conf.OptPipeline = OptPipeline;
273   Conf.AAPipeline = AAPipeline;
274 
275   Conf.OptLevel = OptLevel - '0';
276   Conf.UseNewPM = UseNewPM;
277   Conf.Freestanding = EnableFreestanding;
278   for (auto &PluginFN : PassPlugins)
279     Conf.PassPlugins.push_back(PluginFN);
280   switch (CGOptLevel) {
281   case '0':
282     Conf.CGOptLevel = CodeGenOpt::None;
283     break;
284   case '1':
285     Conf.CGOptLevel = CodeGenOpt::Less;
286     break;
287   case '2':
288     Conf.CGOptLevel = CodeGenOpt::Default;
289     break;
290   case '3':
291     Conf.CGOptLevel = CodeGenOpt::Aggressive;
292     break;
293   default:
294     llvm::errs() << "invalid cg optimization level: " << CGOptLevel << '\n';
295     return 1;
296   }
297 
298   if (auto FT = codegen::getExplicitFileType())
299     Conf.CGFileType = FT.getValue();
300 
301   Conf.OverrideTriple = OverrideTriple;
302   Conf.DefaultTriple = DefaultTriple;
303   Conf.StatsFile = StatsFile;
304   Conf.PTO.LoopVectorization = Conf.OptLevel > 1;
305   Conf.PTO.SLPVectorization = Conf.OptLevel > 1;
306 
307   ThinBackend Backend;
308   if (ThinLTODistributedIndexes)
309     Backend = createWriteIndexesThinBackend(/* OldPrefix */ "",
310                                             /* NewPrefix */ "",
311                                             /* ShouldEmitImportsFiles */ true,
312                                             /* LinkedObjectsFile */ nullptr,
313                                             /* OnWrite */ {});
314   else
315     Backend = createInProcessThinBackend(
316         llvm::heavyweight_hardware_concurrency(Threads));
317   LTO Lto(std::move(Conf), std::move(Backend));
318 
319   bool HasErrors = false;
320   for (std::string F : InputFilenames) {
321     std::unique_ptr<MemoryBuffer> MB = check(MemoryBuffer::getFile(F), F);
322     std::unique_ptr<InputFile> Input =
323         check(InputFile::create(MB->getMemBufferRef()), F);
324 
325     std::vector<SymbolResolution> Res;
326     for (const InputFile::Symbol &Sym : Input->symbols()) {
327       auto I = CommandLineResolutions.find({F, std::string(Sym.getName())});
328       // If it isn't found, look for "$", which would have been added
329       // (followed by a hash) when the symbol was promoted during module
330       // splitting if it was defined in one part and used in the other.
331       // Try looking up the symbol name before the "$".
332       if (I == CommandLineResolutions.end()) {
333         auto SplitName = Sym.getName().rsplit("$");
334         I = CommandLineResolutions.find({F, std::string(SplitName.first)});
335       }
336       if (I == CommandLineResolutions.end()) {
337         llvm::errs() << argv[0] << ": missing symbol resolution for " << F
338                      << ',' << Sym.getName() << '\n';
339         HasErrors = true;
340       } else {
341         Res.push_back(I->second.front());
342         I->second.pop_front();
343         if (I->second.empty())
344           CommandLineResolutions.erase(I);
345       }
346     }
347 
348     if (HasErrors)
349       continue;
350 
351     MBs.push_back(std::move(MB));
352     check(Lto.add(std::move(Input), Res), F);
353   }
354 
355   if (!CommandLineResolutions.empty()) {
356     HasErrors = true;
357     for (auto UnusedRes : CommandLineResolutions)
358       llvm::errs() << argv[0] << ": unused symbol resolution for "
359                    << UnusedRes.first.first << ',' << UnusedRes.first.second
360                    << '\n';
361   }
362   if (HasErrors)
363     return 1;
364 
365   auto AddStream =
366       [&](size_t Task) -> std::unique_ptr<lto::NativeObjectStream> {
367     std::string Path = OutputFilename + "." + utostr(Task);
368 
369     std::error_code EC;
370     auto S = std::make_unique<raw_fd_ostream>(Path, EC, sys::fs::OF_None);
371     check(EC, Path);
372     return std::make_unique<lto::NativeObjectStream>(std::move(S));
373   };
374 
375   auto AddBuffer = [&](size_t Task, std::unique_ptr<MemoryBuffer> MB) {
376     *AddStream(Task)->OS << MB->getBuffer();
377   };
378 
379   NativeObjectCache Cache;
380   if (!CacheDir.empty())
381     Cache = check(localCache(CacheDir, AddBuffer), "failed to create cache");
382 
383   check(Lto.run(AddStream, Cache), "LTO::run failed");
384   return 0;
385 }
386 
387 static int dumpSymtab(int argc, char **argv) {
388   for (StringRef F : make_range(argv + 1, argv + argc)) {
389     std::unique_ptr<MemoryBuffer> MB =
390         check(MemoryBuffer::getFile(F), std::string(F));
391     BitcodeFileContents BFC =
392         check(getBitcodeFileContents(*MB), std::string(F));
393 
394     if (BFC.Symtab.size() >= sizeof(irsymtab::storage::Header)) {
395       auto *Hdr = reinterpret_cast<const irsymtab::storage::Header *>(
396           BFC.Symtab.data());
397       outs() << "version: " << Hdr->Version << '\n';
398       if (Hdr->Version == irsymtab::storage::Header::kCurrentVersion)
399         outs() << "producer: " << Hdr->Producer.get(BFC.StrtabForSymtab)
400                << '\n';
401     }
402 
403     std::unique_ptr<InputFile> Input =
404         check(InputFile::create(MB->getMemBufferRef()), std::string(F));
405 
406     outs() << "target triple: " << Input->getTargetTriple() << '\n';
407     Triple TT(Input->getTargetTriple());
408 
409     outs() << "source filename: " << Input->getSourceFileName() << '\n';
410 
411     if (TT.isOSBinFormatCOFF())
412       outs() << "linker opts: " << Input->getCOFFLinkerOpts() << '\n';
413 
414     if (TT.isOSBinFormatELF()) {
415       outs() << "dependent libraries:";
416       for (auto L : Input->getDependentLibraries())
417         outs() << " \"" << L << "\"";
418       outs() << '\n';
419     }
420 
421     std::vector<StringRef> ComdatTable = Input->getComdatTable();
422     for (const InputFile::Symbol &Sym : Input->symbols()) {
423       switch (Sym.getVisibility()) {
424       case GlobalValue::HiddenVisibility:
425         outs() << 'H';
426         break;
427       case GlobalValue::ProtectedVisibility:
428         outs() << 'P';
429         break;
430       case GlobalValue::DefaultVisibility:
431         outs() << 'D';
432         break;
433       }
434 
435       auto PrintBool = [&](char C, bool B) { outs() << (B ? C : '-'); };
436       PrintBool('U', Sym.isUndefined());
437       PrintBool('C', Sym.isCommon());
438       PrintBool('W', Sym.isWeak());
439       PrintBool('I', Sym.isIndirect());
440       PrintBool('O', Sym.canBeOmittedFromSymbolTable());
441       PrintBool('T', Sym.isTLS());
442       PrintBool('X', Sym.isExecutable());
443       outs() << ' ' << Sym.getName() << '\n';
444 
445       if (Sym.isCommon())
446         outs() << "         size " << Sym.getCommonSize() << " align "
447                << Sym.getCommonAlignment() << '\n';
448 
449       int Comdat = Sym.getComdatIndex();
450       if (Comdat != -1)
451         outs() << "         comdat " << ComdatTable[Comdat] << '\n';
452 
453       if (TT.isOSBinFormatCOFF() && Sym.isWeak() && Sym.isIndirect())
454         outs() << "         fallback " << Sym.getCOFFWeakExternalFallback() << '\n';
455 
456       if (!Sym.getSectionName().empty())
457         outs() << "         section " << Sym.getSectionName() << "\n";
458     }
459 
460     outs() << '\n';
461   }
462 
463   return 0;
464 }
465 
466 int main(int argc, char **argv) {
467   InitLLVM X(argc, argv);
468   InitializeAllTargets();
469   InitializeAllTargetMCs();
470   InitializeAllAsmPrinters();
471   InitializeAllAsmParsers();
472 
473   // FIXME: This should use llvm::cl subcommands, but it isn't currently
474   // possible to pass an argument not associated with a subcommand to a
475   // subcommand (e.g. -use-new-pm).
476   if (argc < 2)
477     return usage();
478 
479   StringRef Subcommand = argv[1];
480   // Ensure that argv[0] is correct after adjusting argv/argc.
481   argv[1] = argv[0];
482   if (Subcommand == "dump-symtab")
483     return dumpSymtab(argc - 1, argv + 1);
484   if (Subcommand == "run")
485     return run(argc - 1, argv + 1);
486   return usage();
487 }
488