1 //===--- Mangle.cpp - Mangle C++ Names --------------------------*- C++ -*-===//
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 // Implements generic name mangling support for blocks and Objective-C.
10 //
11 //===----------------------------------------------------------------------===//
12 #include "clang/AST/Attr.h"
13 #include "clang/AST/ASTContext.h"
14 #include "clang/AST/Decl.h"
15 #include "clang/AST/DeclCXX.h"
16 #include "clang/AST/DeclObjC.h"
17 #include "clang/AST/DeclTemplate.h"
18 #include "clang/AST/ExprCXX.h"
19 #include "clang/AST/Mangle.h"
20 #include "clang/AST/VTableBuilder.h"
21 #include "clang/Basic/ABI.h"
22 #include "clang/Basic/SourceManager.h"
23 #include "clang/Basic/TargetInfo.h"
24 #include "llvm/ADT/StringExtras.h"
25 #include "llvm/IR/DataLayout.h"
26 #include "llvm/IR/Mangler.h"
27 #include "llvm/Support/ErrorHandling.h"
28 #include "llvm/Support/Format.h"
29 #include "llvm/Support/raw_ostream.h"
30 
31 using namespace clang;
32 
33 // FIXME: For blocks we currently mimic GCC's mangling scheme, which leaves
34 // much to be desired. Come up with a better mangling scheme.
35 
36 static void mangleFunctionBlock(MangleContext &Context,
37                                 StringRef Outer,
38                                 const BlockDecl *BD,
39                                 raw_ostream &Out) {
40   unsigned discriminator = Context.getBlockId(BD, true);
41   if (discriminator == 0)
42     Out << "__" << Outer << "_block_invoke";
43   else
44     Out << "__" << Outer << "_block_invoke_" << discriminator+1;
45 }
46 
47 void MangleContext::anchor() { }
48 
49 enum CCMangling {
50   CCM_Other,
51   CCM_Fast,
52   CCM_RegCall,
53   CCM_Vector,
54   CCM_Std,
55   CCM_WasmMainArgcArgv
56 };
57 
58 static bool isExternC(const NamedDecl *ND) {
59   if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(ND))
60     return FD->isExternC();
61   if (const VarDecl *VD = dyn_cast<VarDecl>(ND))
62     return VD->isExternC();
63   return false;
64 }
65 
66 static CCMangling getCallingConvMangling(const ASTContext &Context,
67                                          const NamedDecl *ND) {
68   const TargetInfo &TI = Context.getTargetInfo();
69   const llvm::Triple &Triple = TI.getTriple();
70 
71   // On wasm, the argc/argv form of "main" is renamed so that the startup code
72   // can call it with the correct function signature.
73   // On Emscripten, users may be exporting "main" and expecting to call it
74   // themselves, so we can't mangle it.
75   if (Triple.isWasm() && !Triple.isOSEmscripten())
76     if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(ND))
77       if (FD->isMain() && FD->hasPrototype() && FD->param_size() == 2)
78         return CCM_WasmMainArgcArgv;
79 
80   if (!Triple.isOSWindows() || !Triple.isX86())
81     return CCM_Other;
82 
83   if (Context.getLangOpts().CPlusPlus && !isExternC(ND) &&
84       TI.getCXXABI() == TargetCXXABI::Microsoft)
85     return CCM_Other;
86 
87   const FunctionDecl *FD = dyn_cast<FunctionDecl>(ND);
88   if (!FD)
89     return CCM_Other;
90   QualType T = FD->getType();
91 
92   const FunctionType *FT = T->castAs<FunctionType>();
93 
94   CallingConv CC = FT->getCallConv();
95   switch (CC) {
96   default:
97     return CCM_Other;
98   case CC_X86FastCall:
99     return CCM_Fast;
100   case CC_X86StdCall:
101     return CCM_Std;
102   case CC_X86VectorCall:
103     return CCM_Vector;
104   }
105 }
106 
107 bool MangleContext::shouldMangleDeclName(const NamedDecl *D) {
108   const ASTContext &ASTContext = getASTContext();
109 
110   CCMangling CC = getCallingConvMangling(ASTContext, D);
111   if (CC != CCM_Other)
112     return true;
113 
114   // If the declaration has an owning module for linkage purposes that needs to
115   // be mangled, we must mangle its name.
116   if (!D->hasExternalFormalLinkage() && D->getOwningModuleForLinkage())
117     return true;
118 
119   // C functions with internal linkage have to be mangled with option
120   // -funique-internal-linkage-names.
121   if (!getASTContext().getLangOpts().CPlusPlus &&
122       isUniqueInternalLinkageDecl(D))
123     return true;
124 
125   // In C, functions with no attributes never need to be mangled. Fastpath them.
126   if (!getASTContext().getLangOpts().CPlusPlus && !D->hasAttrs())
127     return false;
128 
129   // Any decl can be declared with __asm("foo") on it, and this takes precedence
130   // over all other naming in the .o file.
131   if (D->hasAttr<AsmLabelAttr>())
132     return true;
133 
134   // Declarations that don't have identifier names always need to be mangled.
135   if (isa<MSGuidDecl>(D))
136     return true;
137 
138   return shouldMangleCXXName(D);
139 }
140 
141 void MangleContext::mangleName(GlobalDecl GD, raw_ostream &Out) {
142   const NamedDecl *D = cast<NamedDecl>(GD.getDecl());
143   // Any decl can be declared with __asm("foo") on it, and this takes precedence
144   // over all other naming in the .o file.
145   if (const AsmLabelAttr *ALA = D->getAttr<AsmLabelAttr>()) {
146     // If we have an asm name, then we use it as the mangling.
147 
148     // If the label isn't literal, or if this is an alias for an LLVM intrinsic,
149     // do not add a "\01" prefix.
150     if (!ALA->getIsLiteralLabel() || ALA->getLabel().startswith("llvm.")) {
151       Out << ALA->getLabel();
152       return;
153     }
154 
155     // Adding the prefix can cause problems when one file has a "foo" and
156     // another has a "\01foo". That is known to happen on ELF with the
157     // tricks normally used for producing aliases (PR9177). Fortunately the
158     // llvm mangler on ELF is a nop, so we can just avoid adding the \01
159     // marker.
160     char GlobalPrefix =
161         getASTContext().getTargetInfo().getDataLayout().getGlobalPrefix();
162     if (GlobalPrefix)
163       Out << '\01'; // LLVM IR Marker for __asm("foo")
164 
165     Out << ALA->getLabel();
166     return;
167   }
168 
169   if (auto *GD = dyn_cast<MSGuidDecl>(D))
170     return mangleMSGuidDecl(GD, Out);
171 
172   const ASTContext &ASTContext = getASTContext();
173   CCMangling CC = getCallingConvMangling(ASTContext, D);
174 
175   if (CC == CCM_WasmMainArgcArgv) {
176     Out << "__main_argc_argv";
177     return;
178   }
179 
180   bool MCXX = shouldMangleCXXName(D);
181   const TargetInfo &TI = Context.getTargetInfo();
182   if (CC == CCM_Other || (MCXX && TI.getCXXABI() == TargetCXXABI::Microsoft)) {
183     if (const ObjCMethodDecl *OMD = dyn_cast<ObjCMethodDecl>(D))
184       mangleObjCMethodNameAsSourceName(OMD, Out);
185     else
186       mangleCXXName(GD, Out);
187     return;
188   }
189 
190   Out << '\01';
191   if (CC == CCM_Std)
192     Out << '_';
193   else if (CC == CCM_Fast)
194     Out << '@';
195   else if (CC == CCM_RegCall)
196     Out << "__regcall3__";
197 
198   if (!MCXX)
199     Out << D->getIdentifier()->getName();
200   else if (const ObjCMethodDecl *OMD = dyn_cast<ObjCMethodDecl>(D))
201     mangleObjCMethodNameAsSourceName(OMD, Out);
202   else
203     mangleCXXName(GD, Out);
204 
205   const FunctionDecl *FD = cast<FunctionDecl>(D);
206   const FunctionType *FT = FD->getType()->castAs<FunctionType>();
207   const FunctionProtoType *Proto = dyn_cast<FunctionProtoType>(FT);
208   if (CC == CCM_Vector)
209     Out << '@';
210   Out << '@';
211   if (!Proto) {
212     Out << '0';
213     return;
214   }
215   assert(!Proto->isVariadic());
216   unsigned ArgWords = 0;
217   if (const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(FD))
218     if (!MD->isStatic())
219       ++ArgWords;
220   for (const auto &AT : Proto->param_types())
221     // Size should be aligned to pointer size.
222     ArgWords +=
223         llvm::alignTo(ASTContext.getTypeSize(AT), TI.getPointerWidth(0)) /
224         TI.getPointerWidth(0);
225   Out << ((TI.getPointerWidth(0) / 8) * ArgWords);
226 }
227 
228 void MangleContext::mangleMSGuidDecl(const MSGuidDecl *GD, raw_ostream &Out) {
229   // For now, follow the MSVC naming convention for GUID objects on all
230   // targets.
231   MSGuidDecl::Parts P = GD->getParts();
232   Out << llvm::format("_GUID_%08" PRIx32 "_%04" PRIx32 "_%04" PRIx32 "_",
233                       P.Part1, P.Part2, P.Part3);
234   unsigned I = 0;
235   for (uint8_t C : P.Part4And5) {
236     Out << llvm::format("%02" PRIx8, C);
237     if (++I == 2)
238       Out << "_";
239   }
240 }
241 
242 void MangleContext::mangleGlobalBlock(const BlockDecl *BD,
243                                       const NamedDecl *ID,
244                                       raw_ostream &Out) {
245   unsigned discriminator = getBlockId(BD, false);
246   if (ID) {
247     if (shouldMangleDeclName(ID))
248       mangleName(ID, Out);
249     else {
250       Out << ID->getIdentifier()->getName();
251     }
252   }
253   if (discriminator == 0)
254     Out << "_block_invoke";
255   else
256     Out << "_block_invoke_" << discriminator+1;
257 }
258 
259 void MangleContext::mangleCtorBlock(const CXXConstructorDecl *CD,
260                                     CXXCtorType CT, const BlockDecl *BD,
261                                     raw_ostream &ResStream) {
262   SmallString<64> Buffer;
263   llvm::raw_svector_ostream Out(Buffer);
264   mangleName(GlobalDecl(CD, CT), Out);
265   mangleFunctionBlock(*this, Buffer, BD, ResStream);
266 }
267 
268 void MangleContext::mangleDtorBlock(const CXXDestructorDecl *DD,
269                                     CXXDtorType DT, const BlockDecl *BD,
270                                     raw_ostream &ResStream) {
271   SmallString<64> Buffer;
272   llvm::raw_svector_ostream Out(Buffer);
273   mangleName(GlobalDecl(DD, DT), Out);
274   mangleFunctionBlock(*this, Buffer, BD, ResStream);
275 }
276 
277 void MangleContext::mangleBlock(const DeclContext *DC, const BlockDecl *BD,
278                                 raw_ostream &Out) {
279   assert(!isa<CXXConstructorDecl>(DC) && !isa<CXXDestructorDecl>(DC));
280 
281   SmallString<64> Buffer;
282   llvm::raw_svector_ostream Stream(Buffer);
283   if (const ObjCMethodDecl *Method = dyn_cast<ObjCMethodDecl>(DC)) {
284     mangleObjCMethodNameAsSourceName(Method, Stream);
285   } else {
286     assert((isa<NamedDecl>(DC) || isa<BlockDecl>(DC)) &&
287            "expected a NamedDecl or BlockDecl");
288     if (isa<BlockDecl>(DC))
289       for (; DC && isa<BlockDecl>(DC); DC = DC->getParent())
290         (void) getBlockId(cast<BlockDecl>(DC), true);
291     assert((isa<TranslationUnitDecl>(DC) || isa<NamedDecl>(DC)) &&
292            "expected a TranslationUnitDecl or a NamedDecl");
293     if (const auto *CD = dyn_cast<CXXConstructorDecl>(DC))
294       mangleCtorBlock(CD, /*CT*/ Ctor_Complete, BD, Out);
295     else if (const auto *DD = dyn_cast<CXXDestructorDecl>(DC))
296       mangleDtorBlock(DD, /*DT*/ Dtor_Complete, BD, Out);
297     else if (auto ND = dyn_cast<NamedDecl>(DC)) {
298       if (!shouldMangleDeclName(ND) && ND->getIdentifier())
299         Stream << ND->getIdentifier()->getName();
300       else {
301         // FIXME: We were doing a mangleUnqualifiedName() before, but that's
302         // a private member of a class that will soon itself be private to the
303         // Itanium C++ ABI object. What should we do now? Right now, I'm just
304         // calling the mangleName() method on the MangleContext; is there a
305         // better way?
306         mangleName(ND, Stream);
307       }
308     }
309   }
310   mangleFunctionBlock(*this, Buffer, BD, Out);
311 }
312 
313 void MangleContext::mangleObjCMethodName(const ObjCMethodDecl *MD,
314                                          raw_ostream &OS,
315                                          bool includePrefixByte,
316                                          bool includeCategoryNamespace) {
317   if (getASTContext().getLangOpts().ObjCRuntime.isGNUFamily()) {
318     // This is the mangling we've always used on the GNU runtimes, but it
319     // has obvious collisions in the face of underscores within class
320     // names, category names, and selectors; maybe we should improve it.
321 
322     OS << (MD->isClassMethod() ? "_c_" : "_i_")
323        << MD->getClassInterface()->getName() << '_';
324 
325     if (includeCategoryNamespace) {
326       if (auto category = MD->getCategory())
327         OS << category->getName();
328     }
329     OS << '_';
330 
331     auto selector = MD->getSelector();
332     for (unsigned slotIndex = 0,
333                   numArgs = selector.getNumArgs(),
334                   slotEnd = std::max(numArgs, 1U);
335            slotIndex != slotEnd; ++slotIndex) {
336       if (auto name = selector.getIdentifierInfoForSlot(slotIndex))
337         OS << name->getName();
338 
339       // Replace all the positions that would've been ':' with '_'.
340       // That's after each slot except that a unary selector doesn't
341       // end in ':'.
342       if (numArgs)
343         OS << '_';
344     }
345 
346     return;
347   }
348 
349   // \01+[ContainerName(CategoryName) SelectorName]
350   if (includePrefixByte) {
351     OS << '\01';
352   }
353   OS << (MD->isInstanceMethod() ? '-' : '+') << '[';
354   if (const auto *CID = MD->getCategory()) {
355     OS << CID->getClassInterface()->getName();
356     if (includeCategoryNamespace) {
357       OS << '(' << *CID << ')';
358     }
359   } else if (const auto *CD =
360                  dyn_cast<ObjCContainerDecl>(MD->getDeclContext())) {
361     OS << CD->getName();
362   } else {
363     llvm_unreachable("Unexpected ObjC method decl context");
364   }
365   OS << ' ';
366   MD->getSelector().print(OS);
367   OS << ']';
368 }
369 
370 void MangleContext::mangleObjCMethodNameAsSourceName(const ObjCMethodDecl *MD,
371                                                      raw_ostream &Out) {
372   SmallString<64> Name;
373   llvm::raw_svector_ostream OS(Name);
374 
375   mangleObjCMethodName(MD, OS, /*includePrefixByte=*/false,
376                        /*includeCategoryNamespace=*/true);
377   Out << OS.str().size() << OS.str();
378 }
379 
380 class ASTNameGenerator::Implementation {
381   std::unique_ptr<MangleContext> MC;
382   llvm::DataLayout DL;
383 
384 public:
385   explicit Implementation(ASTContext &Ctx)
386       : MC(Ctx.createMangleContext()), DL(Ctx.getTargetInfo().getDataLayout()) {
387   }
388 
389   bool writeName(const Decl *D, raw_ostream &OS) {
390     // First apply frontend mangling.
391     SmallString<128> FrontendBuf;
392     llvm::raw_svector_ostream FrontendBufOS(FrontendBuf);
393     if (auto *FD = dyn_cast<FunctionDecl>(D)) {
394       if (FD->isDependentContext())
395         return true;
396       if (writeFuncOrVarName(FD, FrontendBufOS))
397         return true;
398     } else if (auto *VD = dyn_cast<VarDecl>(D)) {
399       if (writeFuncOrVarName(VD, FrontendBufOS))
400         return true;
401     } else if (auto *MD = dyn_cast<ObjCMethodDecl>(D)) {
402       MC->mangleObjCMethodName(MD, OS, /*includePrefixByte=*/false,
403                                /*includeCategoryNamespace=*/true);
404       return false;
405     } else if (auto *ID = dyn_cast<ObjCInterfaceDecl>(D)) {
406       writeObjCClassName(ID, FrontendBufOS);
407     } else {
408       return true;
409     }
410 
411     // Now apply backend mangling.
412     llvm::Mangler::getNameWithPrefix(OS, FrontendBufOS.str(), DL);
413     return false;
414   }
415 
416   std::string getName(const Decl *D) {
417     std::string Name;
418     {
419       llvm::raw_string_ostream OS(Name);
420       writeName(D, OS);
421     }
422     return Name;
423   }
424 
425   enum ObjCKind {
426     ObjCClass,
427     ObjCMetaclass,
428   };
429 
430   static StringRef getClassSymbolPrefix(ObjCKind Kind,
431                                         const ASTContext &Context) {
432     if (Context.getLangOpts().ObjCRuntime.isGNUFamily())
433       return Kind == ObjCMetaclass ? "_OBJC_METACLASS_" : "_OBJC_CLASS_";
434     return Kind == ObjCMetaclass ? "OBJC_METACLASS_$_" : "OBJC_CLASS_$_";
435   }
436 
437   std::vector<std::string> getAllManglings(const ObjCContainerDecl *OCD) {
438     StringRef ClassName;
439     if (const auto *OID = dyn_cast<ObjCInterfaceDecl>(OCD))
440       ClassName = OID->getObjCRuntimeNameAsString();
441     else if (const auto *OID = dyn_cast<ObjCImplementationDecl>(OCD))
442       ClassName = OID->getObjCRuntimeNameAsString();
443 
444     if (ClassName.empty())
445       return {};
446 
447     auto Mangle = [&](ObjCKind Kind, StringRef ClassName) -> std::string {
448       SmallString<40> Mangled;
449       auto Prefix = getClassSymbolPrefix(Kind, OCD->getASTContext());
450       llvm::Mangler::getNameWithPrefix(Mangled, Prefix + ClassName, DL);
451       return std::string(Mangled.str());
452     };
453 
454     return {
455         Mangle(ObjCClass, ClassName),
456         Mangle(ObjCMetaclass, ClassName),
457     };
458   }
459 
460   std::vector<std::string> getAllManglings(const Decl *D) {
461     if (const auto *OCD = dyn_cast<ObjCContainerDecl>(D))
462       return getAllManglings(OCD);
463 
464     if (!(isa<CXXRecordDecl>(D) || isa<CXXMethodDecl>(D)))
465       return {};
466 
467     const NamedDecl *ND = cast<NamedDecl>(D);
468 
469     ASTContext &Ctx = ND->getASTContext();
470     std::unique_ptr<MangleContext> M(Ctx.createMangleContext());
471 
472     std::vector<std::string> Manglings;
473 
474     auto hasDefaultCXXMethodCC = [](ASTContext &C, const CXXMethodDecl *MD) {
475       auto DefaultCC = C.getDefaultCallingConvention(/*IsVariadic=*/false,
476                                                      /*IsCXXMethod=*/true);
477       auto CC = MD->getType()->castAs<FunctionProtoType>()->getCallConv();
478       return CC == DefaultCC;
479     };
480 
481     if (const auto *CD = dyn_cast_or_null<CXXConstructorDecl>(ND)) {
482       Manglings.emplace_back(getMangledStructor(CD, Ctor_Base));
483 
484       if (Ctx.getTargetInfo().getCXXABI().isItaniumFamily())
485         if (!CD->getParent()->isAbstract())
486           Manglings.emplace_back(getMangledStructor(CD, Ctor_Complete));
487 
488       if (Ctx.getTargetInfo().getCXXABI().isMicrosoft())
489         if (CD->hasAttr<DLLExportAttr>() && CD->isDefaultConstructor())
490           if (!(hasDefaultCXXMethodCC(Ctx, CD) && CD->getNumParams() == 0))
491             Manglings.emplace_back(getMangledStructor(CD, Ctor_DefaultClosure));
492     } else if (const auto *DD = dyn_cast_or_null<CXXDestructorDecl>(ND)) {
493       Manglings.emplace_back(getMangledStructor(DD, Dtor_Base));
494       if (Ctx.getTargetInfo().getCXXABI().isItaniumFamily()) {
495         Manglings.emplace_back(getMangledStructor(DD, Dtor_Complete));
496         if (DD->isVirtual())
497           Manglings.emplace_back(getMangledStructor(DD, Dtor_Deleting));
498       }
499     } else if (const auto *MD = dyn_cast_or_null<CXXMethodDecl>(ND)) {
500       Manglings.emplace_back(getName(ND));
501       if (MD->isVirtual())
502         if (const auto *TIV = Ctx.getVTableContext()->getThunkInfo(MD))
503           for (const auto &T : *TIV)
504             Manglings.emplace_back(getMangledThunk(MD, T));
505     }
506 
507     return Manglings;
508   }
509 
510 private:
511   bool writeFuncOrVarName(const NamedDecl *D, raw_ostream &OS) {
512     if (MC->shouldMangleDeclName(D)) {
513       GlobalDecl GD;
514       if (const auto *CtorD = dyn_cast<CXXConstructorDecl>(D))
515         GD = GlobalDecl(CtorD, Ctor_Complete);
516       else if (const auto *DtorD = dyn_cast<CXXDestructorDecl>(D))
517         GD = GlobalDecl(DtorD, Dtor_Complete);
518       else if (D->hasAttr<CUDAGlobalAttr>())
519         GD = GlobalDecl(cast<FunctionDecl>(D));
520       else
521         GD = GlobalDecl(D);
522       MC->mangleName(GD, OS);
523       return false;
524     } else {
525       IdentifierInfo *II = D->getIdentifier();
526       if (!II)
527         return true;
528       OS << II->getName();
529       return false;
530     }
531   }
532 
533   void writeObjCClassName(const ObjCInterfaceDecl *D, raw_ostream &OS) {
534     OS << getClassSymbolPrefix(ObjCClass, D->getASTContext());
535     OS << D->getObjCRuntimeNameAsString();
536   }
537 
538   std::string getMangledStructor(const NamedDecl *ND, unsigned StructorType) {
539     std::string FrontendBuf;
540     llvm::raw_string_ostream FOS(FrontendBuf);
541 
542     GlobalDecl GD;
543     if (const auto *CD = dyn_cast_or_null<CXXConstructorDecl>(ND))
544       GD = GlobalDecl(CD, static_cast<CXXCtorType>(StructorType));
545     else if (const auto *DD = dyn_cast_or_null<CXXDestructorDecl>(ND))
546       GD = GlobalDecl(DD, static_cast<CXXDtorType>(StructorType));
547     MC->mangleName(GD, FOS);
548 
549     std::string BackendBuf;
550     llvm::raw_string_ostream BOS(BackendBuf);
551 
552     llvm::Mangler::getNameWithPrefix(BOS, FOS.str(), DL);
553 
554     return BOS.str();
555   }
556 
557   std::string getMangledThunk(const CXXMethodDecl *MD, const ThunkInfo &T) {
558     std::string FrontendBuf;
559     llvm::raw_string_ostream FOS(FrontendBuf);
560 
561     MC->mangleThunk(MD, T, FOS);
562 
563     std::string BackendBuf;
564     llvm::raw_string_ostream BOS(BackendBuf);
565 
566     llvm::Mangler::getNameWithPrefix(BOS, FOS.str(), DL);
567 
568     return BOS.str();
569   }
570 };
571 
572 ASTNameGenerator::ASTNameGenerator(ASTContext &Ctx)
573     : Impl(std::make_unique<Implementation>(Ctx)) {}
574 
575 ASTNameGenerator::~ASTNameGenerator() {}
576 
577 bool ASTNameGenerator::writeName(const Decl *D, raw_ostream &OS) {
578   return Impl->writeName(D, OS);
579 }
580 
581 std::string ASTNameGenerator::getName(const Decl *D) {
582   return Impl->getName(D);
583 }
584 
585 std::vector<std::string> ASTNameGenerator::getAllManglings(const Decl *D) {
586   return Impl->getAllManglings(D);
587 }
588