1 //===- BlockExtractor.cpp - Extracts blocks into their own functions ------===//
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 pass extracts the specified basic blocks from the module into their
10 // own functions.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "llvm/Transforms/IPO/BlockExtractor.h"
15 #include "llvm/ADT/STLExtras.h"
16 #include "llvm/ADT/Statistic.h"
17 #include "llvm/IR/Instructions.h"
18 #include "llvm/IR/Module.h"
19 #include "llvm/IR/PassManager.h"
20 #include "llvm/InitializePasses.h"
21 #include "llvm/Pass.h"
22 #include "llvm/Support/CommandLine.h"
23 #include "llvm/Support/Debug.h"
24 #include "llvm/Support/MemoryBuffer.h"
25 #include "llvm/Transforms/IPO.h"
26 #include "llvm/Transforms/Utils/BasicBlockUtils.h"
27 #include "llvm/Transforms/Utils/CodeExtractor.h"
28 
29 using namespace llvm;
30 
31 #define DEBUG_TYPE "block-extractor"
32 
33 STATISTIC(NumExtracted, "Number of basic blocks extracted");
34 
35 static cl::opt<std::string> BlockExtractorFile(
36     "extract-blocks-file", cl::value_desc("filename"),
37     cl::desc("A file containing list of basic blocks to extract"), cl::Hidden);
38 
39 static cl::opt<bool>
40     BlockExtractorEraseFuncs("extract-blocks-erase-funcs",
41                              cl::desc("Erase the existing functions"),
42                              cl::Hidden);
43 namespace {
44 class BlockExtractor {
45 public:
46   BlockExtractor(bool EraseFunctions) : EraseFunctions(EraseFunctions) {}
47   bool runOnModule(Module &M);
48   void
49   init(const std::vector<std::vector<BasicBlock *>> &GroupsOfBlocksToExtract) {
50     GroupsOfBlocks = GroupsOfBlocksToExtract;
51     if (!BlockExtractorFile.empty())
52       loadFile();
53   }
54 
55 private:
56   std::vector<std::vector<BasicBlock *>> GroupsOfBlocks;
57   bool EraseFunctions;
58   /// Map a function name to groups of blocks.
59   SmallVector<std::pair<std::string, SmallVector<std::string, 4>>, 4>
60       BlocksByName;
61 
62   void loadFile();
63   void splitLandingPadPreds(Function &F);
64 };
65 
66 } // end anonymous namespace
67 
68 /// Gets all of the blocks specified in the input file.
69 void BlockExtractor::loadFile() {
70   auto ErrOrBuf = MemoryBuffer::getFile(BlockExtractorFile);
71   if (ErrOrBuf.getError())
72     report_fatal_error("BlockExtractor couldn't load the file.");
73   // Read the file.
74   auto &Buf = *ErrOrBuf;
75   SmallVector<StringRef, 16> Lines;
76   Buf->getBuffer().split(Lines, '\n', /*MaxSplit=*/-1,
77                          /*KeepEmpty=*/false);
78   for (const auto &Line : Lines) {
79     SmallVector<StringRef, 4> LineSplit;
80     Line.split(LineSplit, ' ', /*MaxSplit=*/-1,
81                /*KeepEmpty=*/false);
82     if (LineSplit.empty())
83       continue;
84     if (LineSplit.size()!=2)
85       report_fatal_error("Invalid line format, expecting lines like: 'funcname bb1[;bb2..]'",
86                          /*GenCrashDiag=*/false);
87     SmallVector<StringRef, 4> BBNames;
88     LineSplit[1].split(BBNames, ';', /*MaxSplit=*/-1,
89                        /*KeepEmpty=*/false);
90     if (BBNames.empty())
91       report_fatal_error("Missing bbs name");
92     BlocksByName.push_back(
93         {std::string(LineSplit[0]), {BBNames.begin(), BBNames.end()}});
94   }
95 }
96 
97 /// Extracts the landing pads to make sure all of them have only one
98 /// predecessor.
99 void BlockExtractor::splitLandingPadPreds(Function &F) {
100   for (BasicBlock &BB : F) {
101     for (Instruction &I : BB) {
102       if (!isa<InvokeInst>(&I))
103         continue;
104       InvokeInst *II = cast<InvokeInst>(&I);
105       BasicBlock *Parent = II->getParent();
106       BasicBlock *LPad = II->getUnwindDest();
107 
108       // Look through the landing pad's predecessors. If one of them ends in an
109       // 'invoke', then we want to split the landing pad.
110       bool Split = false;
111       for (auto *PredBB : predecessors(LPad)) {
112         if (PredBB->isLandingPad() && PredBB != Parent &&
113             isa<InvokeInst>(Parent->getTerminator())) {
114           Split = true;
115           break;
116         }
117       }
118 
119       if (!Split)
120         continue;
121 
122       SmallVector<BasicBlock *, 2> NewBBs;
123       SplitLandingPadPredecessors(LPad, Parent, ".1", ".2", NewBBs);
124     }
125   }
126 }
127 
128 bool BlockExtractor::runOnModule(Module &M) {
129   bool Changed = false;
130 
131   // Get all the functions.
132   SmallVector<Function *, 4> Functions;
133   for (Function &F : M) {
134     splitLandingPadPreds(F);
135     Functions.push_back(&F);
136   }
137 
138   // Get all the blocks specified in the input file.
139   unsigned NextGroupIdx = GroupsOfBlocks.size();
140   GroupsOfBlocks.resize(NextGroupIdx + BlocksByName.size());
141   for (const auto &BInfo : BlocksByName) {
142     Function *F = M.getFunction(BInfo.first);
143     if (!F)
144       report_fatal_error("Invalid function name specified in the input file",
145                          /*GenCrashDiag=*/false);
146     for (const auto &BBInfo : BInfo.second) {
147       auto Res = llvm::find_if(*F, [&](const BasicBlock &BB) {
148         return BB.getName().equals(BBInfo);
149       });
150       if (Res == F->end())
151         report_fatal_error("Invalid block name specified in the input file",
152                            /*GenCrashDiag=*/false);
153       GroupsOfBlocks[NextGroupIdx].push_back(&*Res);
154     }
155     ++NextGroupIdx;
156   }
157 
158   // Extract each group of basic blocks.
159   for (auto &BBs : GroupsOfBlocks) {
160     SmallVector<BasicBlock *, 32> BlocksToExtractVec;
161     for (BasicBlock *BB : BBs) {
162       // Check if the module contains BB.
163       if (BB->getParent()->getParent() != &M)
164         report_fatal_error("Invalid basic block", /*GenCrashDiag=*/false);
165       LLVM_DEBUG(dbgs() << "BlockExtractor: Extracting "
166                         << BB->getParent()->getName() << ":" << BB->getName()
167                         << "\n");
168       BlocksToExtractVec.push_back(BB);
169       if (const InvokeInst *II = dyn_cast<InvokeInst>(BB->getTerminator()))
170         BlocksToExtractVec.push_back(II->getUnwindDest());
171       ++NumExtracted;
172       Changed = true;
173     }
174     CodeExtractorAnalysisCache CEAC(*BBs[0]->getParent());
175     Function *F = CodeExtractor(BlocksToExtractVec).extractCodeRegion(CEAC);
176     if (F)
177       LLVM_DEBUG(dbgs() << "Extracted group '" << (*BBs.begin())->getName()
178                         << "' in: " << F->getName() << '\n');
179     else
180       LLVM_DEBUG(dbgs() << "Failed to extract for group '"
181                         << (*BBs.begin())->getName() << "'\n");
182   }
183 
184   // Erase the functions.
185   if (EraseFunctions || BlockExtractorEraseFuncs) {
186     for (Function *F : Functions) {
187       LLVM_DEBUG(dbgs() << "BlockExtractor: Trying to delete " << F->getName()
188                         << "\n");
189       F->deleteBody();
190     }
191     // Set linkage as ExternalLinkage to avoid erasing unreachable functions.
192     for (Function &F : M)
193       F.setLinkage(GlobalValue::ExternalLinkage);
194     Changed = true;
195   }
196 
197   return Changed;
198 }
199 
200 BlockExtractorPass::BlockExtractorPass(
201     std::vector<std::vector<BasicBlock *>> &&GroupsOfBlocks,
202     bool EraseFunctions)
203     : GroupsOfBlocks(GroupsOfBlocks), EraseFunctions(EraseFunctions) {}
204 
205 PreservedAnalyses BlockExtractorPass::run(Module &M,
206                                           ModuleAnalysisManager &AM) {
207   BlockExtractor BE(EraseFunctions);
208   BE.init(GroupsOfBlocks);
209   return BE.runOnModule(M) ? PreservedAnalyses::none()
210                            : PreservedAnalyses::all();
211 }
212