xref: /llvm-project/llvm/tools/llvm-diff/llvm-diff.cpp (revision 58c5088d484700692f3447d27e536e3ac405a6dc)
1 #include <string>
2 #include <utility>
3 
4 #include <llvm/ADT/StringRef.h>
5 #include <llvm/ADT/SmallVector.h>
6 #include <llvm/ADT/DenseMap.h>
7 
8 // Required to parse .ll files.
9 #include <llvm/Support/SourceMgr.h>
10 #include <llvm/Assembly/Parser.h>
11 
12 // Required to parse .bc files.
13 #include <llvm/Support/MemoryBuffer.h>
14 #include <llvm/Bitcode/ReaderWriter.h>
15 
16 #include <llvm/Support/CommandLine.h>
17 #include <llvm/Support/raw_ostream.h>
18 #include <llvm/Support/ErrorHandling.h>
19 #include <llvm/LLVMContext.h>
20 #include <llvm/Module.h>
21 #include <llvm/Type.h>
22 #include <llvm/Instructions.h>
23 
24 #include "DifferenceEngine.h"
25 
26 using namespace llvm;
27 
28 /// Reads a module from a file.  If the filename ends in .ll, it is
29 /// interpreted as an assembly file;  otherwise, it is interpreted as
30 /// bitcode.  On error, messages are written to stderr and null is
31 /// returned.
32 static Module *ReadModule(LLVMContext &Context, StringRef Name) {
33   // LLVM assembly path.
34   if (Name.endswith(".ll")) {
35     SMDiagnostic Diag;
36     Module *M = ParseAssemblyFile(Name, Diag, Context);
37     if (M) return M;
38 
39     Diag.Print("llvmdiff", errs());
40     return 0;
41   }
42 
43   // Bitcode path.
44   MemoryBuffer *Buffer = MemoryBuffer::getFile(Name);
45 
46   // ParseBitcodeFile takes ownership of the buffer if it succeeds.
47   std::string Error;
48   Module *M = ParseBitcodeFile(Buffer, Context, &Error);
49   if (M) return M;
50 
51   errs() << "error parsing " << Name << ": " << Error;
52   delete Buffer;
53   return 0;
54 }
55 
56 namespace {
57 struct DiffContext {
58   DiffContext(Value *L, Value *R)
59     : L(L), R(R), Differences(false), IsFunction(isa<Function>(L)) {}
60   Value *L;
61   Value *R;
62   bool Differences;
63   bool IsFunction;
64   DenseMap<Value*,unsigned> LNumbering;
65   DenseMap<Value*,unsigned> RNumbering;
66 };
67 
68 void ComputeNumbering(Function *F, DenseMap<Value*,unsigned> &Numbering) {
69   unsigned BBN = 0;
70   unsigned IN = 0;
71 
72   // Arguments get the first numbers.
73   for (Function::arg_iterator
74          AI = F->arg_begin(), AE = F->arg_end(); AI != AE; ++AI)
75     if (!AI->hasName())
76       Numbering[&*AI] = IN++;
77 
78   // Walk the basic blocks in order.
79   for (Function::iterator FI = F->begin(), FE = F->end(); FI != FE; ++FI) {
80     // Basic blocks have their own 'namespace'.
81     if (!FI->hasName())
82       Numbering[&*FI] = BBN++;
83 
84     // Walk the instructions in order.
85     for (BasicBlock::iterator BI = FI->begin(), BE = FI->end(); BI != BE; ++BI)
86       // void instructions don't get numbers.
87       if (!BI->hasName() && !BI->getType()->isVoidTy())
88         Numbering[&*BI] = IN++;
89   }
90 
91   assert(!Numbering.empty() && "asked for numbering but numbering was no-op");
92 }
93 
94 class DiffConsumer : public DifferenceEngine::Consumer {
95 private:
96   raw_ostream &out;
97   Module *LModule;
98   Module *RModule;
99   SmallVector<DiffContext, 5> contexts;
100   bool Differences;
101   unsigned Indent;
102 
103   void printValue(Value *V, bool isL) {
104     if (V->hasName()) {
105       out << (isa<GlobalValue>(V) ? '@' : '%') << V->getName();
106       return;
107     }
108     if (V->getType()->isVoidTy()) {
109       if (isa<StoreInst>(V)) {
110         out << "store to ";
111         printValue(cast<StoreInst>(V)->getPointerOperand(), isL);
112       } else if (isa<CallInst>(V)) {
113         out << "call to ";
114         printValue(cast<CallInst>(V)->getCalledValue(), isL);
115       } else if (isa<InvokeInst>(V)) {
116         out << "invoke to ";
117         printValue(cast<InvokeInst>(V)->getCalledValue(), isL);
118       } else {
119         out << *V;
120       }
121       return;
122     }
123 
124     unsigned N = contexts.size();
125     while (N > 0) {
126       --N;
127       DiffContext &ctxt = contexts[N];
128       if (!ctxt.IsFunction) continue;
129       if (isL) {
130         if (ctxt.LNumbering.empty())
131           ComputeNumbering(cast<Function>(ctxt.L), ctxt.LNumbering);
132         out << '%' << ctxt.LNumbering[V];
133         return;
134       } else {
135         if (ctxt.RNumbering.empty())
136           ComputeNumbering(cast<Function>(ctxt.R), ctxt.RNumbering);
137         out << '%' << ctxt.RNumbering[V];
138         return;
139       }
140     }
141 
142     out << "<anonymous>";
143   }
144 
145   void header() {
146     if (contexts.empty()) return;
147     for (SmallVectorImpl<DiffContext>::iterator
148            I = contexts.begin(), E = contexts.end(); I != E; ++I) {
149       if (I->Differences) continue;
150       if (isa<Function>(I->L)) {
151         // Extra newline between functions.
152         if (Differences) out << "\n";
153 
154         Function *L = cast<Function>(I->L);
155         Function *R = cast<Function>(I->R);
156         if (L->getName() != R->getName())
157           out << "in function " << L->getName() << " / " << R->getName() << ":\n";
158         else
159           out << "in function " << L->getName() << ":\n";
160       } else if (isa<BasicBlock>(I->L)) {
161         BasicBlock *L = cast<BasicBlock>(I->L);
162         BasicBlock *R = cast<BasicBlock>(I->R);
163         out << "  in block ";
164         printValue(L, true);
165         out << " / ";
166         printValue(R, false);
167         out << ":\n";
168       } else if (isa<Instruction>(I->L)) {
169         out << "    in instruction ";
170         printValue(I->L, true);
171         out << " / ";
172         printValue(I->R, false);
173         out << ":\n";
174       }
175 
176       I->Differences = true;
177     }
178   }
179 
180   void indent() {
181     unsigned N = Indent;
182     while (N--) out << ' ';
183   }
184 
185 public:
186   DiffConsumer(Module *L, Module *R)
187     : out(errs()), LModule(L), RModule(R), Differences(false), Indent(0) {}
188 
189   bool hadDifferences() const { return Differences; }
190 
191   void enterContext(Value *L, Value *R) {
192     contexts.push_back(DiffContext(L, R));
193     Indent += 2;
194   }
195   void exitContext() {
196     Differences |= contexts.back().Differences;
197     contexts.pop_back();
198     Indent -= 2;
199   }
200 
201   void log(StringRef text) {
202     header();
203     indent();
204     out << text << '\n';
205   }
206 
207   void logf(const DifferenceEngine::LogBuilder &Log) {
208     header();
209     indent();
210 
211     unsigned arg = 0;
212 
213     StringRef format = Log.getFormat();
214     while (true) {
215       size_t percent = format.find('%');
216       if (percent == StringRef::npos) {
217         out << format;
218         break;
219       }
220       assert(format[percent] == '%');
221 
222       if (percent > 0) out << format.substr(0, percent);
223 
224       switch (format[percent+1]) {
225       case '%': out << '%'; break;
226       case 'l': printValue(Log.getArgument(arg++), true); break;
227       case 'r': printValue(Log.getArgument(arg++), false); break;
228       default: llvm_unreachable("unknown format character");
229       }
230 
231       format = format.substr(percent+2);
232     }
233 
234     out << '\n';
235   }
236 
237   void logd(const DifferenceEngine::DiffLogBuilder &Log) {
238     header();
239 
240     for (unsigned I = 0, E = Log.getNumLines(); I != E; ++I) {
241       indent();
242       switch (Log.getLineKind(I)) {
243       case DifferenceEngine::DC_match:
244         out << "  ";
245         Log.getLeft(I)->dump();
246         //printValue(Log.getLeft(I), true);
247         break;
248       case DifferenceEngine::DC_left:
249         out << "< ";
250         Log.getLeft(I)->dump();
251         //printValue(Log.getLeft(I), true);
252         break;
253       case DifferenceEngine::DC_right:
254         out << "> ";
255         Log.getRight(I)->dump();
256         //printValue(Log.getRight(I), false);
257         break;
258       }
259       //out << "\n";
260     }
261   }
262 
263 };
264 }
265 
266 static void diffGlobal(DifferenceEngine &Engine, Module *L, Module *R,
267                        StringRef Name) {
268   // Drop leading sigils from the global name.
269   if (Name.startswith("@")) Name = Name.substr(1);
270 
271   Function *LFn = L->getFunction(Name);
272   Function *RFn = R->getFunction(Name);
273   if (LFn && RFn)
274     Engine.diff(LFn, RFn);
275   else if (!LFn && !RFn)
276     errs() << "No function named @" << Name << " in either module\n";
277   else if (!LFn)
278     errs() << "No function named @" << Name << " in left module\n";
279   else
280     errs() << "No function named @" << Name << " in right module\n";
281 }
282 
283 cl::opt<std::string> LeftFilename(cl::Positional, cl::desc("<first file>"), cl::Required);
284 cl::opt<std::string> RightFilename(cl::Positional, cl::desc("<second file>"), cl::Required);
285 cl::list<std::string> GlobalsToCompare(cl::Positional, cl::desc("<globals to compare>"));
286 
287 int main(int argc, char **argv) {
288   cl::ParseCommandLineOptions(argc, argv);
289 
290   LLVMContext Context;
291 
292   // Load both modules.  Die if that fails.
293   Module *LModule = ReadModule(Context, LeftFilename);
294   Module *RModule = ReadModule(Context, RightFilename);
295   if (!LModule || !RModule) return 1;
296 
297   DiffConsumer Consumer(LModule, RModule);
298   DifferenceEngine Engine(Context, Consumer);
299 
300   // If any global names were given, just diff those.
301   if (!GlobalsToCompare.empty()) {
302     for (unsigned I = 0, E = GlobalsToCompare.size(); I != E; ++I)
303       diffGlobal(Engine, LModule, RModule, GlobalsToCompare[I]);
304 
305   // Otherwise, diff everything in the module.
306   } else {
307     Engine.diff(LModule, RModule);
308   }
309 
310   delete LModule;
311   delete RModule;
312 
313   return Consumer.hadDifferences();
314 }
315