1 //===- ChainedIncludesSource.cpp - Chained PCHs in Memory -------*- 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 //  This file defines the ChainedIncludesSource class, which converts headers
10 //  to chained PCHs in memory, mainly used for testing.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "clang/Basic/Builtins.h"
15 #include "clang/Basic/TargetInfo.h"
16 #include "clang/Frontend/ASTUnit.h"
17 #include "clang/Frontend/CompilerInstance.h"
18 #include "clang/Frontend/TextDiagnosticPrinter.h"
19 #include "clang/Lex/Preprocessor.h"
20 #include "clang/Lex/PreprocessorOptions.h"
21 #include "clang/Parse/ParseAST.h"
22 #include "clang/Sema/MultiplexExternalSemaSource.h"
23 #include "clang/Serialization/ASTReader.h"
24 #include "clang/Serialization/ASTWriter.h"
25 #include "llvm/Support/MemoryBuffer.h"
26 
27 using namespace clang;
28 
29 namespace {
30 class ChainedIncludesSourceImpl : public ExternalSemaSource {
31 public:
32   ChainedIncludesSourceImpl(std::vector<std::unique_ptr<CompilerInstance>> CIs)
33       : CIs(std::move(CIs)) {}
34 
35 protected:
36   //===----------------------------------------------------------------------===//
37   // ExternalASTSource interface.
38   //===----------------------------------------------------------------------===//
39 
40   /// Return the amount of memory used by memory buffers, breaking down
41   /// by heap-backed versus mmap'ed memory.
42   void getMemoryBufferSizes(MemoryBufferSizes &sizes) const override {
43     for (unsigned i = 0, e = CIs.size(); i != e; ++i) {
44       if (const ExternalASTSource *eSrc =
45           CIs[i]->getASTContext().getExternalSource()) {
46         eSrc->getMemoryBufferSizes(sizes);
47       }
48     }
49   }
50 
51 private:
52   std::vector<std::unique_ptr<CompilerInstance>> CIs;
53 };
54 
55 /// Members of ChainedIncludesSource, factored out so we can initialize
56 /// them before we initialize the ExternalSemaSource base class.
57 struct ChainedIncludesSourceMembers {
58   ChainedIncludesSourceMembers(
59       std::vector<std::unique_ptr<CompilerInstance>> CIs,
60       IntrusiveRefCntPtr<ExternalSemaSource> FinalReader)
61       : Impl(std::move(CIs)), FinalReader(std::move(FinalReader)) {}
62   ChainedIncludesSourceImpl Impl;
63   IntrusiveRefCntPtr<ExternalSemaSource> FinalReader;
64 };
65 
66 /// Use MultiplexExternalSemaSource to dispatch all ExternalSemaSource
67 /// calls to the final reader.
68 class ChainedIncludesSource
69     : private ChainedIncludesSourceMembers,
70       public MultiplexExternalSemaSource {
71 public:
72   ChainedIncludesSource(std::vector<std::unique_ptr<CompilerInstance>> CIs,
73                         IntrusiveRefCntPtr<ExternalSemaSource> FinalReader)
74       : ChainedIncludesSourceMembers(std::move(CIs), std::move(FinalReader)),
75         MultiplexExternalSemaSource(Impl, *this->FinalReader) {}
76 };
77 }
78 
79 static ASTReader *
80 createASTReader(CompilerInstance &CI, StringRef pchFile,
81                 SmallVectorImpl<std::unique_ptr<llvm::MemoryBuffer>> &MemBufs,
82                 SmallVectorImpl<std::string> &bufNames,
83                 ASTDeserializationListener *deserialListener = nullptr) {
84   Preprocessor &PP = CI.getPreprocessor();
85   std::unique_ptr<ASTReader> Reader;
86   Reader.reset(new ASTReader(PP, CI.getModuleCache(), &CI.getASTContext(),
87                              CI.getPCHContainerReader(),
88                              /*Extensions=*/{},
89                              /*isysroot=*/"", /*DisableValidation=*/true));
90   for (unsigned ti = 0; ti < bufNames.size(); ++ti) {
91     StringRef sr(bufNames[ti]);
92     Reader->addInMemoryBuffer(sr, std::move(MemBufs[ti]));
93   }
94   Reader->setDeserializationListener(deserialListener);
95   switch (Reader->ReadAST(pchFile, serialization::MK_PCH, SourceLocation(),
96                           ASTReader::ARR_None)) {
97   case ASTReader::Success:
98     // Set the predefines buffer as suggested by the PCH reader.
99     PP.setPredefines(Reader->getSuggestedPredefines());
100     return Reader.release();
101 
102   case ASTReader::Failure:
103   case ASTReader::Missing:
104   case ASTReader::OutOfDate:
105   case ASTReader::VersionMismatch:
106   case ASTReader::ConfigurationMismatch:
107   case ASTReader::HadErrors:
108     break;
109   }
110   return nullptr;
111 }
112 
113 IntrusiveRefCntPtr<ExternalSemaSource> clang::createChainedIncludesSource(
114     CompilerInstance &CI, IntrusiveRefCntPtr<ExternalSemaSource> &Reader) {
115 
116   std::vector<std::string> &includes = CI.getPreprocessorOpts().ChainedIncludes;
117   assert(!includes.empty() && "No '-chain-include' in options!");
118 
119   std::vector<std::unique_ptr<CompilerInstance>> CIs;
120   InputKind IK = CI.getFrontendOpts().Inputs[0].getKind();
121 
122   SmallVector<std::unique_ptr<llvm::MemoryBuffer>, 4> SerialBufs;
123   SmallVector<std::string, 4> serialBufNames;
124 
125   for (unsigned i = 0, e = includes.size(); i != e; ++i) {
126     bool firstInclude = (i == 0);
127     std::unique_ptr<CompilerInvocation> CInvok;
128     CInvok.reset(new CompilerInvocation(CI.getInvocation()));
129 
130     CInvok->getPreprocessorOpts().ChainedIncludes.clear();
131     CInvok->getPreprocessorOpts().ImplicitPCHInclude.clear();
132     CInvok->getPreprocessorOpts().DisablePCHValidation = true;
133     CInvok->getPreprocessorOpts().Includes.clear();
134     CInvok->getPreprocessorOpts().MacroIncludes.clear();
135     CInvok->getPreprocessorOpts().Macros.clear();
136 
137     CInvok->getFrontendOpts().Inputs.clear();
138     FrontendInputFile InputFile(includes[i], IK);
139     CInvok->getFrontendOpts().Inputs.push_back(InputFile);
140 
141     TextDiagnosticPrinter *DiagClient =
142       new TextDiagnosticPrinter(llvm::errs(), new DiagnosticOptions());
143     IntrusiveRefCntPtr<DiagnosticIDs> DiagID(new DiagnosticIDs());
144     IntrusiveRefCntPtr<DiagnosticsEngine> Diags(
145         new DiagnosticsEngine(DiagID, &CI.getDiagnosticOpts(), DiagClient));
146 
147     std::unique_ptr<CompilerInstance> Clang(
148         new CompilerInstance(CI.getPCHContainerOperations()));
149     Clang->setInvocation(std::move(CInvok));
150     Clang->setDiagnostics(Diags.get());
151     Clang->setTarget(TargetInfo::CreateTargetInfo(
152         Clang->getDiagnostics(), Clang->getInvocation().TargetOpts));
153     Clang->createFileManager();
154     Clang->createSourceManager(Clang->getFileManager());
155     Clang->createPreprocessor(TU_Prefix);
156     Clang->getDiagnosticClient().BeginSourceFile(Clang->getLangOpts(),
157                                                  &Clang->getPreprocessor());
158     Clang->createASTContext();
159 
160     auto Buffer = std::make_shared<PCHBuffer>();
161     ArrayRef<std::shared_ptr<ModuleFileExtension>> Extensions;
162     auto consumer = std::make_unique<PCHGenerator>(
163         Clang->getPreprocessor(), Clang->getModuleCache(), "-", /*isysroot=*/"",
164         Buffer, Extensions, /*AllowASTWithErrors=*/true);
165     Clang->getASTContext().setASTMutationListener(
166                                             consumer->GetASTMutationListener());
167     Clang->setASTConsumer(std::move(consumer));
168     Clang->createSema(TU_Prefix, nullptr);
169 
170     if (firstInclude) {
171       Preprocessor &PP = Clang->getPreprocessor();
172       PP.getBuiltinInfo().initializeBuiltins(PP.getIdentifierTable(),
173                                              PP.getLangOpts());
174     } else {
175       assert(!SerialBufs.empty());
176       SmallVector<std::unique_ptr<llvm::MemoryBuffer>, 4> Bufs;
177       // TODO: Pass through the existing MemoryBuffer instances instead of
178       // allocating new ones.
179       for (auto &SB : SerialBufs)
180         Bufs.push_back(llvm::MemoryBuffer::getMemBuffer(SB->getBuffer()));
181       std::string pchName = includes[i-1];
182       llvm::raw_string_ostream os(pchName);
183       os << ".pch" << i-1;
184       serialBufNames.push_back(os.str());
185 
186       IntrusiveRefCntPtr<ASTReader> Reader;
187       Reader = createASTReader(
188           *Clang, pchName, Bufs, serialBufNames,
189           Clang->getASTConsumer().GetASTDeserializationListener());
190       if (!Reader)
191         return nullptr;
192       Clang->setASTReader(Reader);
193       Clang->getASTContext().setExternalSource(Reader);
194     }
195 
196     if (!Clang->InitializeSourceManager(InputFile))
197       return nullptr;
198 
199     ParseAST(Clang->getSema());
200     Clang->getDiagnosticClient().EndSourceFile();
201     assert(Buffer->IsComplete && "serialization did not complete");
202     auto &serialAST = Buffer->Data;
203     SerialBufs.push_back(llvm::MemoryBuffer::getMemBufferCopy(
204         StringRef(serialAST.data(), serialAST.size())));
205     serialAST.clear();
206     CIs.push_back(std::move(Clang));
207   }
208 
209   assert(!SerialBufs.empty());
210   std::string pchName = includes.back() + ".pch-final";
211   serialBufNames.push_back(pchName);
212   Reader = createASTReader(CI, pchName, SerialBufs, serialBufNames);
213   if (!Reader)
214     return nullptr;
215 
216   return IntrusiveRefCntPtr<ChainedIncludesSource>(
217       new ChainedIncludesSource(std::move(CIs), Reader));
218 }
219