xref: /freebsd/contrib/llvm-project/lld/COFF/MinGW.cpp (revision 5f757f3ff9144b609b3c433dfd370cc6bdc191ad)
10b57cec5SDimitry Andric //===- MinGW.cpp ----------------------------------------------------------===//
20b57cec5SDimitry Andric //
30b57cec5SDimitry Andric // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
40b57cec5SDimitry Andric // See https://llvm.org/LICENSE.txt for license information.
50b57cec5SDimitry Andric // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
60b57cec5SDimitry Andric //
70b57cec5SDimitry Andric //===----------------------------------------------------------------------===//
80b57cec5SDimitry Andric 
90b57cec5SDimitry Andric #include "MinGW.h"
10349cc55cSDimitry Andric #include "COFFLinkerContext.h"
11e8d8bef9SDimitry Andric #include "Driver.h"
12e8d8bef9SDimitry Andric #include "InputFiles.h"
130b57cec5SDimitry Andric #include "SymbolTable.h"
14e8d8bef9SDimitry Andric #include "llvm/ADT/DenseMap.h"
15e8d8bef9SDimitry Andric #include "llvm/ADT/DenseSet.h"
160b57cec5SDimitry Andric #include "llvm/Object/COFF.h"
17e8d8bef9SDimitry Andric #include "llvm/Support/Parallel.h"
180b57cec5SDimitry Andric #include "llvm/Support/Path.h"
19*5f757f3fSDimitry Andric #include "llvm/Support/TimeProfiler.h"
200b57cec5SDimitry Andric #include "llvm/Support/raw_ostream.h"
210b57cec5SDimitry Andric 
220b57cec5SDimitry Andric using namespace llvm;
230b57cec5SDimitry Andric using namespace llvm::COFF;
245ffd83dbSDimitry Andric using namespace lld;
255ffd83dbSDimitry Andric using namespace lld::coff;
2685868e8aSDimitry Andric 
2761cfbce3SDimitry Andric AutoExporter::AutoExporter(
28bdd1243dSDimitry Andric     COFFLinkerContext &ctx,
2961cfbce3SDimitry Andric     const llvm::DenseSet<StringRef> &manualExcludeSymbols)
30bdd1243dSDimitry Andric     : manualExcludeSymbols(manualExcludeSymbols), ctx(ctx) {
310b57cec5SDimitry Andric   excludeLibs = {
320b57cec5SDimitry Andric       "libgcc",
330b57cec5SDimitry Andric       "libgcc_s",
340b57cec5SDimitry Andric       "libstdc++",
350b57cec5SDimitry Andric       "libmingw32",
360b57cec5SDimitry Andric       "libmingwex",
370b57cec5SDimitry Andric       "libg2c",
380b57cec5SDimitry Andric       "libsupc++",
390b57cec5SDimitry Andric       "libobjc",
400b57cec5SDimitry Andric       "libgcj",
410b57cec5SDimitry Andric       "libclang_rt.builtins",
420b57cec5SDimitry Andric       "libclang_rt.builtins-aarch64",
430b57cec5SDimitry Andric       "libclang_rt.builtins-arm",
440b57cec5SDimitry Andric       "libclang_rt.builtins-i386",
450b57cec5SDimitry Andric       "libclang_rt.builtins-x86_64",
46979e22ffSDimitry Andric       "libclang_rt.profile",
47979e22ffSDimitry Andric       "libclang_rt.profile-aarch64",
48979e22ffSDimitry Andric       "libclang_rt.profile-arm",
49979e22ffSDimitry Andric       "libclang_rt.profile-i386",
50979e22ffSDimitry Andric       "libclang_rt.profile-x86_64",
510b57cec5SDimitry Andric       "libc++",
520b57cec5SDimitry Andric       "libc++abi",
531ac55f4cSDimitry Andric       "libFortran_main",
541ac55f4cSDimitry Andric       "libFortranRuntime",
551ac55f4cSDimitry Andric       "libFortranDecimal",
560b57cec5SDimitry Andric       "libunwind",
570b57cec5SDimitry Andric       "libmsvcrt",
580b57cec5SDimitry Andric       "libucrtbase",
590b57cec5SDimitry Andric   };
600b57cec5SDimitry Andric 
610b57cec5SDimitry Andric   excludeObjects = {
620b57cec5SDimitry Andric       "crt0.o",    "crt1.o",  "crt1u.o", "crt2.o",  "crt2u.o",    "dllcrt1.o",
630b57cec5SDimitry Andric       "dllcrt2.o", "gcrt0.o", "gcrt1.o", "gcrt2.o", "crtbegin.o", "crtend.o",
640b57cec5SDimitry Andric   };
650b57cec5SDimitry Andric 
660b57cec5SDimitry Andric   excludeSymbolPrefixes = {
670b57cec5SDimitry Andric       // Import symbols
680b57cec5SDimitry Andric       "__imp_",
690b57cec5SDimitry Andric       "__IMPORT_DESCRIPTOR_",
700b57cec5SDimitry Andric       // Extra import symbols from GNU import libraries
710b57cec5SDimitry Andric       "__nm_",
720b57cec5SDimitry Andric       // C++ symbols
730b57cec5SDimitry Andric       "__rtti_",
740b57cec5SDimitry Andric       "__builtin_",
7585868e8aSDimitry Andric       // Artificial symbols such as .refptr
760b57cec5SDimitry Andric       ".",
77979e22ffSDimitry Andric       // profile generate symbols
78979e22ffSDimitry Andric       "__profc_",
79979e22ffSDimitry Andric       "__profd_",
80979e22ffSDimitry Andric       "__profvp_",
810b57cec5SDimitry Andric   };
820b57cec5SDimitry Andric 
830b57cec5SDimitry Andric   excludeSymbolSuffixes = {
840b57cec5SDimitry Andric       "_iname",
850b57cec5SDimitry Andric       "_NULL_THUNK_DATA",
860b57cec5SDimitry Andric   };
870b57cec5SDimitry Andric 
88bdd1243dSDimitry Andric   if (ctx.config.machine == I386) {
890b57cec5SDimitry Andric     excludeSymbols = {
900b57cec5SDimitry Andric         "__NULL_IMPORT_DESCRIPTOR",
910b57cec5SDimitry Andric         "__pei386_runtime_relocator",
920b57cec5SDimitry Andric         "_do_pseudo_reloc",
930b57cec5SDimitry Andric         "_impure_ptr",
940b57cec5SDimitry Andric         "__impure_ptr",
950b57cec5SDimitry Andric         "__fmode",
960b57cec5SDimitry Andric         "_environ",
970b57cec5SDimitry Andric         "___dso_handle",
980b57cec5SDimitry Andric         // These are the MinGW names that differ from the standard
990b57cec5SDimitry Andric         // ones (lacking an extra underscore).
1000b57cec5SDimitry Andric         "_DllMain@12",
1010b57cec5SDimitry Andric         "_DllEntryPoint@12",
1020b57cec5SDimitry Andric         "_DllMainCRTStartup@12",
1030b57cec5SDimitry Andric     };
1040b57cec5SDimitry Andric     excludeSymbolPrefixes.insert("__head_");
1050b57cec5SDimitry Andric   } else {
1060b57cec5SDimitry Andric     excludeSymbols = {
1070b57cec5SDimitry Andric         "__NULL_IMPORT_DESCRIPTOR",
1080b57cec5SDimitry Andric         "_pei386_runtime_relocator",
1090b57cec5SDimitry Andric         "do_pseudo_reloc",
1100b57cec5SDimitry Andric         "impure_ptr",
1110b57cec5SDimitry Andric         "_impure_ptr",
1120b57cec5SDimitry Andric         "_fmode",
1130b57cec5SDimitry Andric         "environ",
1140b57cec5SDimitry Andric         "__dso_handle",
1150b57cec5SDimitry Andric         // These are the MinGW names that differ from the standard
1160b57cec5SDimitry Andric         // ones (lacking an extra underscore).
1170b57cec5SDimitry Andric         "DllMain",
1180b57cec5SDimitry Andric         "DllEntryPoint",
1190b57cec5SDimitry Andric         "DllMainCRTStartup",
1200b57cec5SDimitry Andric     };
1210b57cec5SDimitry Andric     excludeSymbolPrefixes.insert("_head_");
1220b57cec5SDimitry Andric   }
1230b57cec5SDimitry Andric }
1240b57cec5SDimitry Andric 
1250b57cec5SDimitry Andric void AutoExporter::addWholeArchive(StringRef path) {
1260b57cec5SDimitry Andric   StringRef libName = sys::path::filename(path);
1270b57cec5SDimitry Andric   // Drop the file extension, to match the processing below.
1280b57cec5SDimitry Andric   libName = libName.substr(0, libName.rfind('.'));
1290b57cec5SDimitry Andric   excludeLibs.erase(libName);
1300b57cec5SDimitry Andric }
1310b57cec5SDimitry Andric 
13261cfbce3SDimitry Andric void AutoExporter::addExcludedSymbol(StringRef symbol) {
13361cfbce3SDimitry Andric   excludeSymbols.insert(symbol);
13461cfbce3SDimitry Andric }
13561cfbce3SDimitry Andric 
136bdd1243dSDimitry Andric bool AutoExporter::shouldExport(Defined *sym) const {
137fe6060f1SDimitry Andric   if (!sym || !sym->getChunk())
1380b57cec5SDimitry Andric     return false;
1390b57cec5SDimitry Andric 
1400b57cec5SDimitry Andric   // Only allow the symbol kinds that make sense to export; in particular,
1410b57cec5SDimitry Andric   // disallow import symbols.
1420b57cec5SDimitry Andric   if (!isa<DefinedRegular>(sym) && !isa<DefinedCommon>(sym))
1430b57cec5SDimitry Andric     return false;
14461cfbce3SDimitry Andric   if (excludeSymbols.count(sym->getName()) || manualExcludeSymbols.count(sym->getName()))
1450b57cec5SDimitry Andric     return false;
1460b57cec5SDimitry Andric 
1470b57cec5SDimitry Andric   for (StringRef prefix : excludeSymbolPrefixes.keys())
14806c3fb27SDimitry Andric     if (sym->getName().starts_with(prefix))
1490b57cec5SDimitry Andric       return false;
1500b57cec5SDimitry Andric   for (StringRef suffix : excludeSymbolSuffixes.keys())
15106c3fb27SDimitry Andric     if (sym->getName().ends_with(suffix))
1520b57cec5SDimitry Andric       return false;
1530b57cec5SDimitry Andric 
1540b57cec5SDimitry Andric   // If a corresponding __imp_ symbol exists and is defined, don't export it.
155349cc55cSDimitry Andric   if (ctx.symtab.find(("__imp_" + sym->getName()).str()))
1560b57cec5SDimitry Andric     return false;
1570b57cec5SDimitry Andric 
1580b57cec5SDimitry Andric   // Check that file is non-null before dereferencing it, symbols not
1590b57cec5SDimitry Andric   // originating in regular object files probably shouldn't be exported.
1600b57cec5SDimitry Andric   if (!sym->getFile())
1610b57cec5SDimitry Andric     return false;
1620b57cec5SDimitry Andric 
1630b57cec5SDimitry Andric   StringRef libName = sys::path::filename(sym->getFile()->parentName);
1640b57cec5SDimitry Andric 
1650b57cec5SDimitry Andric   // Drop the file extension.
1660b57cec5SDimitry Andric   libName = libName.substr(0, libName.rfind('.'));
1670b57cec5SDimitry Andric   if (!libName.empty())
1680b57cec5SDimitry Andric     return !excludeLibs.count(libName);
1690b57cec5SDimitry Andric 
1700b57cec5SDimitry Andric   StringRef fileName = sys::path::filename(sym->getFile()->getName());
1710b57cec5SDimitry Andric   return !excludeObjects.count(fileName);
1720b57cec5SDimitry Andric }
1730b57cec5SDimitry Andric 
174bdd1243dSDimitry Andric void lld::coff::writeDefFile(StringRef name,
175bdd1243dSDimitry Andric                              const std::vector<Export> &exports) {
176*5f757f3fSDimitry Andric   llvm::TimeTraceScope timeScope("Write .def file");
1770b57cec5SDimitry Andric   std::error_code ec;
17885868e8aSDimitry Andric   raw_fd_ostream os(name, ec, sys::fs::OF_None);
1790b57cec5SDimitry Andric   if (ec)
1800b57cec5SDimitry Andric     fatal("cannot open " + name + ": " + ec.message());
1810b57cec5SDimitry Andric 
1820b57cec5SDimitry Andric   os << "EXPORTS\n";
183bdd1243dSDimitry Andric   for (const Export &e : exports) {
1840b57cec5SDimitry Andric     os << "    " << e.exportName << " "
1850b57cec5SDimitry Andric        << "@" << e.ordinal;
1860b57cec5SDimitry Andric     if (auto *def = dyn_cast_or_null<Defined>(e.sym)) {
1870b57cec5SDimitry Andric       if (def && def->getChunk() &&
1880b57cec5SDimitry Andric           !(def->getChunk()->getOutputCharacteristics() & IMAGE_SCN_MEM_EXECUTE))
1890b57cec5SDimitry Andric         os << " DATA";
1900b57cec5SDimitry Andric     }
1910b57cec5SDimitry Andric     os << "\n";
1920b57cec5SDimitry Andric   }
1930b57cec5SDimitry Andric }
194e8d8bef9SDimitry Andric 
195bdd1243dSDimitry Andric static StringRef mangle(Twine sym, MachineTypes machine) {
196bdd1243dSDimitry Andric   assert(machine != IMAGE_FILE_MACHINE_UNKNOWN);
197bdd1243dSDimitry Andric   if (machine == I386)
19804eeddc0SDimitry Andric     return saver().save("_" + sym);
19904eeddc0SDimitry Andric   return saver().save(sym);
200e8d8bef9SDimitry Andric }
201e8d8bef9SDimitry Andric 
202e8d8bef9SDimitry Andric // Handles -wrap option.
203e8d8bef9SDimitry Andric //
204e8d8bef9SDimitry Andric // This function instantiates wrapper symbols. At this point, they seem
205e8d8bef9SDimitry Andric // like they are not being used at all, so we explicitly set some flags so
206e8d8bef9SDimitry Andric // that LTO won't eliminate them.
207e8d8bef9SDimitry Andric std::vector<WrappedSymbol>
208349cc55cSDimitry Andric lld::coff::addWrappedSymbols(COFFLinkerContext &ctx, opt::InputArgList &args) {
209e8d8bef9SDimitry Andric   std::vector<WrappedSymbol> v;
210e8d8bef9SDimitry Andric   DenseSet<StringRef> seen;
211e8d8bef9SDimitry Andric 
212e8d8bef9SDimitry Andric   for (auto *arg : args.filtered(OPT_wrap)) {
213e8d8bef9SDimitry Andric     StringRef name = arg->getValue();
214e8d8bef9SDimitry Andric     if (!seen.insert(name).second)
215e8d8bef9SDimitry Andric       continue;
216e8d8bef9SDimitry Andric 
217349cc55cSDimitry Andric     Symbol *sym = ctx.symtab.findUnderscore(name);
218e8d8bef9SDimitry Andric     if (!sym)
219e8d8bef9SDimitry Andric       continue;
220e8d8bef9SDimitry Andric 
221bdd1243dSDimitry Andric     Symbol *real =
222bdd1243dSDimitry Andric         ctx.symtab.addUndefined(mangle("__real_" + name, ctx.config.machine));
223bdd1243dSDimitry Andric     Symbol *wrap =
224bdd1243dSDimitry Andric         ctx.symtab.addUndefined(mangle("__wrap_" + name, ctx.config.machine));
225e8d8bef9SDimitry Andric     v.push_back({sym, real, wrap});
226e8d8bef9SDimitry Andric 
227e8d8bef9SDimitry Andric     // These symbols may seem undefined initially, but don't bail out
228349cc55cSDimitry Andric     // at symtab.reportUnresolvable() due to them, but let wrapSymbols
229e8d8bef9SDimitry Andric     // below sort things out before checking finally with
230349cc55cSDimitry Andric     // symtab.resolveRemainingUndefines().
231e8d8bef9SDimitry Andric     sym->deferUndefined = true;
232e8d8bef9SDimitry Andric     real->deferUndefined = true;
233e8d8bef9SDimitry Andric     // We want to tell LTO not to inline symbols to be overwritten
234e8d8bef9SDimitry Andric     // because LTO doesn't know the final symbol contents after renaming.
235e8d8bef9SDimitry Andric     real->canInline = false;
236e8d8bef9SDimitry Andric     sym->canInline = false;
237e8d8bef9SDimitry Andric 
238e8d8bef9SDimitry Andric     // Tell LTO not to eliminate these symbols.
239e8d8bef9SDimitry Andric     sym->isUsedInRegularObj = true;
240e8d8bef9SDimitry Andric     if (!isa<Undefined>(wrap))
241e8d8bef9SDimitry Andric       wrap->isUsedInRegularObj = true;
242e8d8bef9SDimitry Andric   }
243e8d8bef9SDimitry Andric   return v;
244e8d8bef9SDimitry Andric }
245e8d8bef9SDimitry Andric 
246e8d8bef9SDimitry Andric // Do renaming for -wrap by updating pointers to symbols.
247e8d8bef9SDimitry Andric //
248e8d8bef9SDimitry Andric // When this function is executed, only InputFiles and symbol table
249e8d8bef9SDimitry Andric // contain pointers to symbol objects. We visit them to replace pointers,
250e8d8bef9SDimitry Andric // so that wrapped symbols are swapped as instructed by the command line.
251349cc55cSDimitry Andric void lld::coff::wrapSymbols(COFFLinkerContext &ctx,
252349cc55cSDimitry Andric                             ArrayRef<WrappedSymbol> wrapped) {
253e8d8bef9SDimitry Andric   DenseMap<Symbol *, Symbol *> map;
254e8d8bef9SDimitry Andric   for (const WrappedSymbol &w : wrapped) {
255e8d8bef9SDimitry Andric     map[w.sym] = w.wrap;
256e8d8bef9SDimitry Andric     map[w.real] = w.sym;
257e8d8bef9SDimitry Andric     if (Defined *d = dyn_cast<Defined>(w.wrap)) {
258349cc55cSDimitry Andric       Symbol *imp = ctx.symtab.find(("__imp_" + w.sym->getName()).str());
259e8d8bef9SDimitry Andric       // Create a new defined local import for the wrap symbol. If
260e8d8bef9SDimitry Andric       // no imp prefixed symbol existed, there's no need for it.
261e8d8bef9SDimitry Andric       // (We can't easily distinguish whether any object file actually
262e8d8bef9SDimitry Andric       // referenced it or not, though.)
263e8d8bef9SDimitry Andric       if (imp) {
264e8d8bef9SDimitry Andric         DefinedLocalImport *wrapimp = make<DefinedLocalImport>(
265bdd1243dSDimitry Andric             ctx, saver().save("__imp_" + w.wrap->getName()), d);
266349cc55cSDimitry Andric         ctx.symtab.localImportChunks.push_back(wrapimp->getChunk());
267e8d8bef9SDimitry Andric         map[imp] = wrapimp;
268e8d8bef9SDimitry Andric       }
269e8d8bef9SDimitry Andric     }
270e8d8bef9SDimitry Andric   }
271e8d8bef9SDimitry Andric 
272e8d8bef9SDimitry Andric   // Update pointers in input files.
273349cc55cSDimitry Andric   parallelForEach(ctx.objFileInstances, [&](ObjFile *file) {
274e8d8bef9SDimitry Andric     MutableArrayRef<Symbol *> syms = file->getMutableSymbols();
27506c3fb27SDimitry Andric     for (auto &sym : syms)
27606c3fb27SDimitry Andric       if (Symbol *s = map.lookup(sym))
27706c3fb27SDimitry Andric         sym = s;
278e8d8bef9SDimitry Andric   });
279e8d8bef9SDimitry Andric }
280