xref: /llvm-project/clang/lib/StaticAnalyzer/Checkers/ExprInspectionChecker.cpp (revision a1580d7b59b65b17f2ce7fdb95f46379e7df4089) 
1 //==- ExprInspectionChecker.cpp - Used for regression tests ------*- 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 #include "clang/Analysis/IssueHash.h"
10 #include "clang/StaticAnalyzer/Checkers/BuiltinCheckerRegistration.h"
11 #include "clang/StaticAnalyzer/Checkers/SValExplainer.h"
12 #include "clang/StaticAnalyzer/Checkers/Taint.h"
13 #include "clang/StaticAnalyzer/Core/BugReporter/BugType.h"
14 #include "clang/StaticAnalyzer/Core/Checker.h"
15 #include "clang/StaticAnalyzer/Core/PathSensitive/CallEvent.h"
16 #include "clang/StaticAnalyzer/Core/PathSensitive/CheckerContext.h"
17 #include "clang/StaticAnalyzer/Core/PathSensitive/DynamicExtent.h"
18 #include "llvm/ADT/StringSwitch.h"
19 #include "llvm/Support/ScopedPrinter.h"
20 #include <optional>
21 
22 using namespace clang;
23 using namespace ento;
24 
25 namespace {
26 class ExprInspectionChecker
27     : public Checker<eval::Call, check::DeadSymbols, check::EndAnalysis> {
28   mutable std::unique_ptr<BugType> BT;
29 
30   // These stats are per-analysis, not per-branch, hence they shouldn't
31   // stay inside the program state.
32   struct ReachedStat {
33     ExplodedNode *ExampleNode;
34     unsigned NumTimesReached;
35   };
36   mutable llvm::DenseMap<const CallExpr *, ReachedStat> ReachedStats;
37 
38   void analyzerEval(const CallExpr *CE, CheckerContext &C) const;
39   void analyzerCheckInlined(const CallExpr *CE, CheckerContext &C) const;
40   void analyzerWarnIfReached(const CallExpr *CE, CheckerContext &C) const;
41   void analyzerNumTimesReached(const CallExpr *CE, CheckerContext &C) const;
42   void analyzerCrash(const CallExpr *CE, CheckerContext &C) const;
43   void analyzerWarnOnDeadSymbol(const CallExpr *CE, CheckerContext &C) const;
44   void analyzerValue(const CallExpr *CE, CheckerContext &C) const;
45   void analyzerDumpSValType(const CallExpr *CE, CheckerContext &C) const;
46   void analyzerDump(const CallExpr *CE, CheckerContext &C) const;
47   void analyzerExplain(const CallExpr *CE, CheckerContext &C) const;
48   void analyzerPrintState(const CallExpr *CE, CheckerContext &C) const;
49   void analyzerGetExtent(const CallExpr *CE, CheckerContext &C) const;
50   void analyzerDumpExtent(const CallExpr *CE, CheckerContext &C) const;
51   void analyzerDumpElementCount(const CallExpr *CE, CheckerContext &C) const;
52   void analyzerHashDump(const CallExpr *CE, CheckerContext &C) const;
53   void analyzerDenote(const CallExpr *CE, CheckerContext &C) const;
54   void analyzerExpress(const CallExpr *CE, CheckerContext &C) const;
55   void analyzerIsTainted(const CallExpr *CE, CheckerContext &C) const;
56 
57   typedef void (ExprInspectionChecker::*FnCheck)(const CallExpr *,
58                                                  CheckerContext &C) const;
59 
60   // Optional parameter `ExprVal` for expression value to be marked interesting.
61   ExplodedNode *reportBug(llvm::StringRef Msg, CheckerContext &C,
62                           Optional<SVal> ExprVal = std::nullopt) const;
63   ExplodedNode *reportBug(llvm::StringRef Msg, BugReporter &BR, ExplodedNode *N,
64                           Optional<SVal> ExprVal = std::nullopt) const;
65   template <typename T> void printAndReport(CheckerContext &C, T What) const;
66 
67   const Expr *getArgExpr(const CallExpr *CE, CheckerContext &C) const;
68   const MemRegion *getArgRegion(const CallExpr *CE, CheckerContext &C) const;
69 
70 public:
71   bool evalCall(const CallEvent &Call, CheckerContext &C) const;
72   void checkDeadSymbols(SymbolReaper &SymReaper, CheckerContext &C) const;
73   void checkEndAnalysis(ExplodedGraph &G, BugReporter &BR,
74                         ExprEngine &Eng) const;
75 };
76 } // namespace
77 
78 REGISTER_SET_WITH_PROGRAMSTATE(MarkedSymbols, SymbolRef)
79 REGISTER_MAP_WITH_PROGRAMSTATE(DenotedSymbols, SymbolRef, const StringLiteral *)
80 
81 bool ExprInspectionChecker::evalCall(const CallEvent &Call,
82                                      CheckerContext &C) const {
83   const auto *CE = dyn_cast_or_null<CallExpr>(Call.getOriginExpr());
84   if (!CE)
85     return false;
86 
87   // These checks should have no effect on the surrounding environment
88   // (globals should not be invalidated, etc), hence the use of evalCall.
89   FnCheck Handler =
90       llvm::StringSwitch<FnCheck>(C.getCalleeName(CE))
91           .Case("clang_analyzer_eval", &ExprInspectionChecker::analyzerEval)
92           .Case("clang_analyzer_checkInlined",
93                 &ExprInspectionChecker::analyzerCheckInlined)
94           .Case("clang_analyzer_crash", &ExprInspectionChecker::analyzerCrash)
95           .Case("clang_analyzer_warnIfReached",
96                 &ExprInspectionChecker::analyzerWarnIfReached)
97           .Case("clang_analyzer_warnOnDeadSymbol",
98                 &ExprInspectionChecker::analyzerWarnOnDeadSymbol)
99           .StartsWith("clang_analyzer_explain",
100                       &ExprInspectionChecker::analyzerExplain)
101           .Case("clang_analyzer_dumpExtent",
102                 &ExprInspectionChecker::analyzerDumpExtent)
103           .Case("clang_analyzer_dumpElementCount",
104                 &ExprInspectionChecker::analyzerDumpElementCount)
105           .Case("clang_analyzer_value", &ExprInspectionChecker::analyzerValue)
106           .StartsWith("clang_analyzer_dumpSvalType",
107                       &ExprInspectionChecker::analyzerDumpSValType)
108           .StartsWith("clang_analyzer_dump",
109                       &ExprInspectionChecker::analyzerDump)
110           .Case("clang_analyzer_getExtent",
111                 &ExprInspectionChecker::analyzerGetExtent)
112           .Case("clang_analyzer_printState",
113                 &ExprInspectionChecker::analyzerPrintState)
114           .Case("clang_analyzer_numTimesReached",
115                 &ExprInspectionChecker::analyzerNumTimesReached)
116           .Case("clang_analyzer_hashDump",
117                 &ExprInspectionChecker::analyzerHashDump)
118           .Case("clang_analyzer_denote", &ExprInspectionChecker::analyzerDenote)
119           .Case("clang_analyzer_express", // This also marks the argument as
120                                           // interesting.
121                 &ExprInspectionChecker::analyzerExpress)
122           .StartsWith("clang_analyzer_isTainted",
123                       &ExprInspectionChecker::analyzerIsTainted)
124           .Default(nullptr);
125 
126   if (!Handler)
127     return false;
128 
129   (this->*Handler)(CE, C);
130   return true;
131 }
132 
133 static const char *getArgumentValueString(const CallExpr *CE,
134                                           CheckerContext &C) {
135   if (CE->getNumArgs() == 0)
136     return "Missing assertion argument";
137 
138   ExplodedNode *N = C.getPredecessor();
139   const LocationContext *LC = N->getLocationContext();
140   ProgramStateRef State = N->getState();
141 
142   const Expr *Assertion = CE->getArg(0);
143   SVal AssertionVal = State->getSVal(Assertion, LC);
144 
145   if (AssertionVal.isUndef())
146     return "UNDEFINED";
147 
148   ProgramStateRef StTrue, StFalse;
149   std::tie(StTrue, StFalse) =
150       State->assume(AssertionVal.castAs<DefinedOrUnknownSVal>());
151 
152   if (StTrue) {
153     if (StFalse)
154       return "UNKNOWN";
155     else
156       return "TRUE";
157   } else {
158     if (StFalse)
159       return "FALSE";
160     else
161       llvm_unreachable("Invalid constraint; neither true or false.");
162   }
163 }
164 
165 ExplodedNode *ExprInspectionChecker::reportBug(llvm::StringRef Msg,
166                                                CheckerContext &C,
167                                                Optional<SVal> ExprVal) const {
168   ExplodedNode *N = C.generateNonFatalErrorNode();
169   reportBug(Msg, C.getBugReporter(), N, ExprVal);
170   return N;
171 }
172 
173 ExplodedNode *ExprInspectionChecker::reportBug(llvm::StringRef Msg,
174                                                BugReporter &BR, ExplodedNode *N,
175                                                Optional<SVal> ExprVal) const {
176   if (!N)
177     return nullptr;
178 
179   if (!BT)
180     BT.reset(new BugType(this, "Checking analyzer assumptions", "debug"));
181 
182   auto R = std::make_unique<PathSensitiveBugReport>(*BT, Msg, N);
183   if (ExprVal) {
184     R->markInteresting(*ExprVal);
185   }
186   BR.emitReport(std::move(R));
187   return N;
188 }
189 
190 const Expr *ExprInspectionChecker::getArgExpr(const CallExpr *CE,
191                                               CheckerContext &C) const {
192   if (CE->getNumArgs() == 0) {
193     reportBug("Missing argument", C);
194     return nullptr;
195   }
196   return CE->getArg(0);
197 }
198 
199 const MemRegion *ExprInspectionChecker::getArgRegion(const CallExpr *CE,
200                                                      CheckerContext &C) const {
201   const Expr *Arg = getArgExpr(CE, C);
202   if (!Arg)
203     return nullptr;
204 
205   const MemRegion *MR = C.getSVal(Arg).getAsRegion();
206   if (!MR) {
207     reportBug("Cannot obtain the region", C);
208     return nullptr;
209   }
210 
211   return MR;
212 }
213 
214 void ExprInspectionChecker::analyzerEval(const CallExpr *CE,
215                                          CheckerContext &C) const {
216   const LocationContext *LC = C.getPredecessor()->getLocationContext();
217 
218   // A specific instantiation of an inlined function may have more constrained
219   // values than can generally be assumed. Skip the check.
220   if (LC->getStackFrame()->getParent() != nullptr)
221     return;
222 
223   reportBug(getArgumentValueString(CE, C), C);
224 }
225 
226 void ExprInspectionChecker::analyzerWarnIfReached(const CallExpr *CE,
227                                                   CheckerContext &C) const {
228   reportBug("REACHABLE", C);
229 }
230 
231 void ExprInspectionChecker::analyzerNumTimesReached(const CallExpr *CE,
232                                                     CheckerContext &C) const {
233   ++ReachedStats[CE].NumTimesReached;
234   if (!ReachedStats[CE].ExampleNode) {
235     // Later, in checkEndAnalysis, we'd throw a report against it.
236     ReachedStats[CE].ExampleNode = C.generateNonFatalErrorNode();
237   }
238 }
239 
240 void ExprInspectionChecker::analyzerCheckInlined(const CallExpr *CE,
241                                                  CheckerContext &C) const {
242   const LocationContext *LC = C.getPredecessor()->getLocationContext();
243 
244   // An inlined function could conceivably also be analyzed as a top-level
245   // function. We ignore this case and only emit a message (TRUE or FALSE)
246   // when we are analyzing it as an inlined function. This means that
247   // clang_analyzer_checkInlined(true) should always print TRUE, but
248   // clang_analyzer_checkInlined(false) should never actually print anything.
249   if (LC->getStackFrame()->getParent() == nullptr)
250     return;
251 
252   reportBug(getArgumentValueString(CE, C), C);
253 }
254 
255 void ExprInspectionChecker::analyzerExplain(const CallExpr *CE,
256                                             CheckerContext &C) const {
257   const Expr *Arg = getArgExpr(CE, C);
258   if (!Arg)
259     return;
260 
261   SVal V = C.getSVal(Arg);
262   SValExplainer Ex(C.getASTContext());
263   reportBug(Ex.Visit(V), C);
264 }
265 
266 static void printHelper(llvm::raw_svector_ostream &Out, CheckerContext &C,
267                         const llvm::APSInt &I) {
268   Out << I.getBitWidth() << (I.isUnsigned() ? "u:" : "s:");
269   Out << I;
270 }
271 
272 static void printHelper(llvm::raw_svector_ostream &Out, CheckerContext &C,
273                         SymbolRef Sym) {
274   C.getConstraintManager().printValue(Out, C.getState(), Sym);
275 }
276 
277 static void printHelper(llvm::raw_svector_ostream &Out, CheckerContext &C,
278                         SVal V) {
279   Out << V;
280 }
281 
282 template <typename T>
283 void ExprInspectionChecker::printAndReport(CheckerContext &C, T What) const {
284   llvm::SmallString<64> Str;
285   llvm::raw_svector_ostream OS(Str);
286   printHelper(OS, C, What);
287   reportBug(OS.str(), C);
288 }
289 
290 void ExprInspectionChecker::analyzerValue(const CallExpr *CE,
291                                           CheckerContext &C) const {
292   const Expr *Arg = getArgExpr(CE, C);
293   if (!Arg)
294     return;
295 
296   SVal V = C.getSVal(Arg);
297   if (const SymbolRef Sym = V.getAsSymbol())
298     printAndReport(C, Sym);
299   else if (const llvm::APSInt *I = V.getAsInteger())
300     printAndReport(C, *I);
301   else
302     reportBug("n/a", C);
303 }
304 
305 void ExprInspectionChecker::analyzerDumpSValType(const CallExpr *CE,
306                                                  CheckerContext &C) const {
307   const Expr *Arg = getArgExpr(CE, C);
308   if (!Arg)
309     return;
310 
311   QualType Ty = C.getSVal(Arg).getType(C.getASTContext());
312   reportBug(Ty.getAsString(), C);
313 }
314 
315 void ExprInspectionChecker::analyzerDump(const CallExpr *CE,
316                                          CheckerContext &C) const {
317   const Expr *Arg = getArgExpr(CE, C);
318   if (!Arg)
319     return;
320 
321   SVal V = C.getSVal(Arg);
322   printAndReport(C, V);
323 }
324 
325 void ExprInspectionChecker::analyzerGetExtent(const CallExpr *CE,
326                                               CheckerContext &C) const {
327   const MemRegion *MR = getArgRegion(CE, C);
328   if (!MR)
329     return;
330 
331   ProgramStateRef State = C.getState();
332   DefinedOrUnknownSVal Size = getDynamicExtent(State, MR, C.getSValBuilder());
333 
334   State = State->BindExpr(CE, C.getLocationContext(), Size);
335   C.addTransition(State);
336 }
337 
338 void ExprInspectionChecker::analyzerDumpExtent(const CallExpr *CE,
339                                                CheckerContext &C) const {
340   const MemRegion *MR = getArgRegion(CE, C);
341   if (!MR)
342     return;
343 
344   DefinedOrUnknownSVal Size =
345       getDynamicExtent(C.getState(), MR, C.getSValBuilder());
346   printAndReport(C, Size);
347 }
348 
349 void ExprInspectionChecker::analyzerDumpElementCount(const CallExpr *CE,
350                                                      CheckerContext &C) const {
351   const MemRegion *MR = getArgRegion(CE, C);
352   if (!MR)
353     return;
354 
355   QualType ElementTy;
356   if (const auto *TVR = MR->getAs<TypedValueRegion>()) {
357     ElementTy = TVR->getValueType();
358   } else {
359     ElementTy = MR->castAs<SymbolicRegion>()->getPointeeStaticType();
360   }
361 
362   assert(!ElementTy->isPointerType());
363 
364   DefinedOrUnknownSVal ElementCount =
365       getDynamicElementCount(C.getState(), MR, C.getSValBuilder(), ElementTy);
366   printAndReport(C, ElementCount);
367 }
368 
369 void ExprInspectionChecker::analyzerPrintState(const CallExpr *CE,
370                                                CheckerContext &C) const {
371   C.getState()->dump();
372 }
373 
374 void ExprInspectionChecker::analyzerWarnOnDeadSymbol(const CallExpr *CE,
375                                                      CheckerContext &C) const {
376   const Expr *Arg = getArgExpr(CE, C);
377   if (!Arg)
378     return;
379 
380   SVal Val = C.getSVal(Arg);
381   SymbolRef Sym = Val.getAsSymbol();
382   if (!Sym)
383     return;
384 
385   ProgramStateRef State = C.getState();
386   State = State->add<MarkedSymbols>(Sym);
387   C.addTransition(State);
388 }
389 
390 void ExprInspectionChecker::checkDeadSymbols(SymbolReaper &SymReaper,
391                                              CheckerContext &C) const {
392   ProgramStateRef State = C.getState();
393   const MarkedSymbolsTy &Syms = State->get<MarkedSymbols>();
394   ExplodedNode *N = C.getPredecessor();
395   for (auto I = Syms.begin(), E = Syms.end(); I != E; ++I) {
396     SymbolRef Sym = *I;
397     if (!SymReaper.isDead(Sym))
398       continue;
399 
400     // The non-fatal error node should be the same for all reports.
401     if (ExplodedNode *BugNode = reportBug("SYMBOL DEAD", C))
402       N = BugNode;
403     State = State->remove<MarkedSymbols>(Sym);
404   }
405 
406   for (auto I : State->get<DenotedSymbols>()) {
407     SymbolRef Sym = I.first;
408     if (!SymReaper.isLive(Sym))
409       State = State->remove<DenotedSymbols>(Sym);
410   }
411 
412   C.addTransition(State, N);
413 }
414 
415 void ExprInspectionChecker::checkEndAnalysis(ExplodedGraph &G, BugReporter &BR,
416                                              ExprEngine &Eng) const {
417   for (auto Item : ReachedStats) {
418     unsigned NumTimesReached = Item.second.NumTimesReached;
419     ExplodedNode *N = Item.second.ExampleNode;
420 
421     reportBug(llvm::to_string(NumTimesReached), BR, N);
422   }
423   ReachedStats.clear();
424 }
425 
426 void ExprInspectionChecker::analyzerCrash(const CallExpr *CE,
427                                           CheckerContext &C) const {
428   LLVM_BUILTIN_TRAP;
429 }
430 
431 void ExprInspectionChecker::analyzerHashDump(const CallExpr *CE,
432                                              CheckerContext &C) const {
433   const LangOptions &Opts = C.getLangOpts();
434   const SourceManager &SM = C.getSourceManager();
435   FullSourceLoc FL(CE->getArg(0)->getBeginLoc(), SM);
436   std::string HashContent =
437       getIssueString(FL, getCheckerName().getName(), "Category",
438                      C.getLocationContext()->getDecl(), Opts);
439 
440   reportBug(HashContent, C);
441 }
442 
443 void ExprInspectionChecker::analyzerDenote(const CallExpr *CE,
444                                            CheckerContext &C) const {
445   if (CE->getNumArgs() < 2) {
446     reportBug("clang_analyzer_denote() requires a symbol and a string literal",
447               C);
448     return;
449   }
450 
451   SymbolRef Sym = C.getSVal(CE->getArg(0)).getAsSymbol();
452   if (!Sym) {
453     reportBug("Not a symbol", C);
454     return;
455   }
456 
457   const auto *E = dyn_cast<StringLiteral>(CE->getArg(1)->IgnoreParenCasts());
458   if (!E) {
459     reportBug("Not a string literal", C);
460     return;
461   }
462 
463   ProgramStateRef State = C.getState();
464 
465   C.addTransition(C.getState()->set<DenotedSymbols>(Sym, E));
466 }
467 
468 namespace {
469 class SymbolExpressor
470     : public SymExprVisitor<SymbolExpressor, Optional<std::string>> {
471   ProgramStateRef State;
472 
473 public:
474   SymbolExpressor(ProgramStateRef State) : State(State) {}
475 
476   Optional<std::string> lookup(const SymExpr *S) {
477     if (const StringLiteral *const *SLPtr = State->get<DenotedSymbols>(S)) {
478       const StringLiteral *SL = *SLPtr;
479       return std::string(SL->getBytes());
480     }
481     return std::nullopt;
482   }
483 
484   Optional<std::string> VisitSymExpr(const SymExpr *S) { return lookup(S); }
485 
486   Optional<std::string> VisitSymIntExpr(const SymIntExpr *S) {
487     if (Optional<std::string> Str = lookup(S))
488       return Str;
489     if (Optional<std::string> Str = Visit(S->getLHS()))
490       return (*Str + " " + BinaryOperator::getOpcodeStr(S->getOpcode()) + " " +
491               std::to_string(S->getRHS().getLimitedValue()) +
492               (S->getRHS().isUnsigned() ? "U" : ""))
493           .str();
494     return std::nullopt;
495   }
496 
497   Optional<std::string> VisitSymSymExpr(const SymSymExpr *S) {
498     if (Optional<std::string> Str = lookup(S))
499       return Str;
500     if (Optional<std::string> Str1 = Visit(S->getLHS()))
501       if (Optional<std::string> Str2 = Visit(S->getRHS()))
502         return (*Str1 + " " + BinaryOperator::getOpcodeStr(S->getOpcode()) +
503                 " " + *Str2)
504             .str();
505     return std::nullopt;
506   }
507 
508   Optional<std::string> VisitUnarySymExpr(const UnarySymExpr *S) {
509     if (Optional<std::string> Str = lookup(S))
510       return Str;
511     if (Optional<std::string> Str = Visit(S->getOperand()))
512       return (UnaryOperator::getOpcodeStr(S->getOpcode()) + *Str).str();
513     return std::nullopt;
514   }
515 
516   Optional<std::string> VisitSymbolCast(const SymbolCast *S) {
517     if (Optional<std::string> Str = lookup(S))
518       return Str;
519     if (Optional<std::string> Str = Visit(S->getOperand()))
520       return (Twine("(") + S->getType().getAsString() + ")" + *Str).str();
521     return std::nullopt;
522   }
523 };
524 } // namespace
525 
526 void ExprInspectionChecker::analyzerExpress(const CallExpr *CE,
527                                             CheckerContext &C) const {
528   const Expr *Arg = getArgExpr(CE, C);
529   if (!Arg)
530     return;
531 
532   SVal ArgVal = C.getSVal(CE->getArg(0));
533   SymbolRef Sym = ArgVal.getAsSymbol();
534   if (!Sym) {
535     reportBug("Not a symbol", C, ArgVal);
536     return;
537   }
538 
539   SymbolExpressor V(C.getState());
540   auto Str = V.Visit(Sym);
541   if (!Str) {
542     reportBug("Unable to express", C, ArgVal);
543     return;
544   }
545 
546   reportBug(*Str, C, ArgVal);
547 }
548 
549 void ExprInspectionChecker::analyzerIsTainted(const CallExpr *CE,
550                                               CheckerContext &C) const {
551   if (CE->getNumArgs() != 1) {
552     reportBug("clang_analyzer_isTainted() requires exactly one argument", C);
553     return;
554   }
555   const bool IsTainted =
556       taint::isTainted(C.getState(), CE->getArg(0), C.getLocationContext());
557   reportBug(IsTainted ? "YES" : "NO", C);
558 }
559 
560 void ento::registerExprInspectionChecker(CheckerManager &Mgr) {
561   Mgr.registerChecker<ExprInspectionChecker>();
562 }
563 
564 bool ento::shouldRegisterExprInspectionChecker(const CheckerManager &mgr) {
565   return true;
566 }
567