1//== ObjCSelfInitChecker.cpp - Checker for 'self' initialization -*- C++ -*--=// 2// 3// The LLVM Compiler Infrastructure 4// 5// This file is distributed under the University of Illinois Open Source 6// License. See LICENSE.TXT for details. 7// 8//===----------------------------------------------------------------------===// 9// 10// This defines ObjCSelfInitChecker, a builtin check that checks for uses of 11// 'self' before proper initialization. 12// 13//===----------------------------------------------------------------------===// 14 15// This checks initialization methods to verify that they assign 'self' to the 16// result of an initialization call (e.g. [super init], or [self initWith..]) 17// before using 'self' or any instance variable. 18// 19// To perform the required checking, values are tagged with flags that indicate 20// 1) if the object is the one pointed to by 'self', and 2) if the object 21// is the result of an initializer (e.g. [super init]). 22// 23// Uses of an object that is true for 1) but not 2) trigger a diagnostic. 24// The uses that are currently checked are: 25// - Using instance variables. 26// - Returning the object. 27// 28// Note that we don't check for an invalid 'self' that is the receiver of an 29// obj-c message expression to cut down false positives where logging functions 30// get information from self (like its class) or doing "invalidation" on self 31// when the initialization fails. 32// 33// Because the object that 'self' points to gets invalidated when a call 34// receives a reference to 'self', the checker keeps track and passes the flags 35// for 1) and 2) to the new object that 'self' points to after the call. 36// 37//===----------------------------------------------------------------------===// 38 39#include "ClangSACheckers.h" 40#include "clang/AST/ParentMap.h" 41#include "clang/StaticAnalyzer/Core/BugReporter/BugType.h" 42#include "clang/StaticAnalyzer/Core/Checker.h" 43#include "clang/StaticAnalyzer/Core/CheckerManager.h" 44#include "clang/StaticAnalyzer/Core/PathSensitive/CallEvent.h" 45#include "clang/StaticAnalyzer/Core/PathSensitive/CheckerContext.h" 46#include "clang/StaticAnalyzer/Core/PathSensitive/ProgramStateTrait.h" 47#include "llvm/Support/raw_ostream.h" 48 49using namespace clang; 50using namespace ento; 51 52static bool shouldRunOnFunctionOrMethod(const NamedDecl *ND); 53static bool isInitializationMethod(const ObjCMethodDecl *MD); 54static bool isInitMessage(const ObjCMethodCall &Msg); 55static bool isSelfVar(SVal location, CheckerContext &C); 56 57namespace { 58class ObjCSelfInitChecker : public Checker< check::PostObjCMessage, 59 check::PostStmt<ObjCIvarRefExpr>, 60 check::PreStmt<ReturnStmt>, 61 check::PreCall, 62 check::PostCall, 63 check::Location, 64 check::Bind > { 65 mutable std::unique_ptr<BugType> BT; 66 67 void checkForInvalidSelf(const Expr *E, CheckerContext &C, 68 const char *errorStr) const; 69 70public: 71 ObjCSelfInitChecker() {} 72 void checkPostObjCMessage(const ObjCMethodCall &Msg, CheckerContext &C) const; 73 void checkPostStmt(const ObjCIvarRefExpr *E, CheckerContext &C) const; 74 void checkPreStmt(const ReturnStmt *S, CheckerContext &C) const; 75 void checkLocation(SVal location, bool isLoad, const Stmt *S, 76 CheckerContext &C) const; 77 void checkBind(SVal loc, SVal val, const Stmt *S, CheckerContext &C) const; 78 79 void checkPreCall(const CallEvent &CE, CheckerContext &C) const; 80 void checkPostCall(const CallEvent &CE, CheckerContext &C) const; 81 82 void printState(raw_ostream &Out, ProgramStateRef State, 83 const char *NL, const char *Sep) const override; 84}; 85} // end anonymous namespace 86 87namespace { 88enum SelfFlagEnum { 89 /// \brief No flag set. 90 SelfFlag_None = 0x0, 91 /// \brief Value came from 'self'. 92 SelfFlag_Self = 0x1, 93 /// \brief Value came from the result of an initializer (e.g. [super init]). 94 SelfFlag_InitRes = 0x2 95}; 96} 97 98REGISTER_MAP_WITH_PROGRAMSTATE(SelfFlag, SymbolRef, unsigned) 99REGISTER_TRAIT_WITH_PROGRAMSTATE(CalledInit, bool) 100 101/// \brief A call receiving a reference to 'self' invalidates the object that 102/// 'self' contains. This keeps the "self flags" assigned to the 'self' 103/// object before the call so we can assign them to the new object that 'self' 104/// points to after the call. 105REGISTER_TRAIT_WITH_PROGRAMSTATE(PreCallSelfFlags, unsigned) 106 107static SelfFlagEnum getSelfFlags(SVal val, ProgramStateRef state) { 108 if (SymbolRef sym = val.getAsSymbol()) 109 if (const unsigned *attachedFlags = state->get<SelfFlag>(sym)) 110 return (SelfFlagEnum)*attachedFlags; 111 return SelfFlag_None; 112} 113 114static SelfFlagEnum getSelfFlags(SVal val, CheckerContext &C) { 115 return getSelfFlags(val, C.getState()); 116} 117 118static void addSelfFlag(ProgramStateRef state, SVal val, 119 SelfFlagEnum flag, CheckerContext &C) { 120 // We tag the symbol that the SVal wraps. 121 if (SymbolRef sym = val.getAsSymbol()) { 122 state = state->set<SelfFlag>(sym, getSelfFlags(val, state) | flag); 123 C.addTransition(state); 124 } 125} 126 127static bool hasSelfFlag(SVal val, SelfFlagEnum flag, CheckerContext &C) { 128 return getSelfFlags(val, C) & flag; 129} 130 131/// \brief Returns true of the value of the expression is the object that 'self' 132/// points to and is an object that did not come from the result of calling 133/// an initializer. 134static bool isInvalidSelf(const Expr *E, CheckerContext &C) { 135 SVal exprVal = C.getState()->getSVal(E, C.getLocationContext()); 136 if (!hasSelfFlag(exprVal, SelfFlag_Self, C)) 137 return false; // value did not come from 'self'. 138 if (hasSelfFlag(exprVal, SelfFlag_InitRes, C)) 139 return false; // 'self' is properly initialized. 140 141 return true; 142} 143 144void ObjCSelfInitChecker::checkForInvalidSelf(const Expr *E, CheckerContext &C, 145 const char *errorStr) const { 146 if (!E) 147 return; 148 149 if (!C.getState()->get<CalledInit>()) 150 return; 151 152 if (!isInvalidSelf(E, C)) 153 return; 154 155 // Generate an error node. 156 ExplodedNode *N = C.generateSink(); 157 if (!N) 158 return; 159 160 if (!BT) 161 BT.reset(new BugType(this, "Missing \"self = [(super or self) init...]\"", 162 categories::CoreFoundationObjectiveC)); 163 BugReport *report = new BugReport(*BT, errorStr, N); 164 C.emitReport(report); 165} 166 167void ObjCSelfInitChecker::checkPostObjCMessage(const ObjCMethodCall &Msg, 168 CheckerContext &C) const { 169 // When encountering a message that does initialization (init rule), 170 // tag the return value so that we know later on that if self has this value 171 // then it is properly initialized. 172 173 // FIXME: A callback should disable checkers at the start of functions. 174 if (!shouldRunOnFunctionOrMethod(dyn_cast<NamedDecl>( 175 C.getCurrentAnalysisDeclContext()->getDecl()))) 176 return; 177 178 if (isInitMessage(Msg)) { 179 // Tag the return value as the result of an initializer. 180 ProgramStateRef state = C.getState(); 181 182 // FIXME this really should be context sensitive, where we record 183 // the current stack frame (for IPA). Also, we need to clean this 184 // value out when we return from this method. 185 state = state->set<CalledInit>(true); 186 187 SVal V = state->getSVal(Msg.getOriginExpr(), C.getLocationContext()); 188 addSelfFlag(state, V, SelfFlag_InitRes, C); 189 return; 190 } 191 192 // We don't check for an invalid 'self' in an obj-c message expression to cut 193 // down false positives where logging functions get information from self 194 // (like its class) or doing "invalidation" on self when the initialization 195 // fails. 196} 197 198void ObjCSelfInitChecker::checkPostStmt(const ObjCIvarRefExpr *E, 199 CheckerContext &C) const { 200 // FIXME: A callback should disable checkers at the start of functions. 201 if (!shouldRunOnFunctionOrMethod(dyn_cast<NamedDecl>( 202 C.getCurrentAnalysisDeclContext()->getDecl()))) 203 return; 204 205 checkForInvalidSelf( 206 E->getBase(), C, 207 "Instance variable used while 'self' is not set to the result of " 208 "'[(super or self) init...]'"); 209} 210 211void ObjCSelfInitChecker::checkPreStmt(const ReturnStmt *S, 212 CheckerContext &C) const { 213 // FIXME: A callback should disable checkers at the start of functions. 214 if (!shouldRunOnFunctionOrMethod(dyn_cast<NamedDecl>( 215 C.getCurrentAnalysisDeclContext()->getDecl()))) 216 return; 217 218 checkForInvalidSelf(S->getRetValue(), C, 219 "Returning 'self' while it is not set to the result of " 220 "'[(super or self) init...]'"); 221} 222 223// When a call receives a reference to 'self', [Pre/Post]Call pass 224// the SelfFlags from the object 'self' points to before the call to the new 225// object after the call. This is to avoid invalidation of 'self' by logging 226// functions. 227// Another common pattern in classes with multiple initializers is to put the 228// subclass's common initialization bits into a static function that receives 229// the value of 'self', e.g: 230// @code 231// if (!(self = [super init])) 232// return nil; 233// if (!(self = _commonInit(self))) 234// return nil; 235// @endcode 236// Until we can use inter-procedural analysis, in such a call, transfer the 237// SelfFlags to the result of the call. 238 239void ObjCSelfInitChecker::checkPreCall(const CallEvent &CE, 240 CheckerContext &C) const { 241 // FIXME: A callback should disable checkers at the start of functions. 242 if (!shouldRunOnFunctionOrMethod(dyn_cast<NamedDecl>( 243 C.getCurrentAnalysisDeclContext()->getDecl()))) 244 return; 245 246 ProgramStateRef state = C.getState(); 247 unsigned NumArgs = CE.getNumArgs(); 248 // If we passed 'self' as and argument to the call, record it in the state 249 // to be propagated after the call. 250 // Note, we could have just given up, but try to be more optimistic here and 251 // assume that the functions are going to continue initialization or will not 252 // modify self. 253 for (unsigned i = 0; i < NumArgs; ++i) { 254 SVal argV = CE.getArgSVal(i); 255 if (isSelfVar(argV, C)) { 256 unsigned selfFlags = getSelfFlags(state->getSVal(argV.castAs<Loc>()), C); 257 C.addTransition(state->set<PreCallSelfFlags>(selfFlags)); 258 return; 259 } else if (hasSelfFlag(argV, SelfFlag_Self, C)) { 260 unsigned selfFlags = getSelfFlags(argV, C); 261 C.addTransition(state->set<PreCallSelfFlags>(selfFlags)); 262 return; 263 } 264 } 265} 266 267void ObjCSelfInitChecker::checkPostCall(const CallEvent &CE, 268 CheckerContext &C) const { 269 // FIXME: A callback should disable checkers at the start of functions. 270 if (!shouldRunOnFunctionOrMethod(dyn_cast<NamedDecl>( 271 C.getCurrentAnalysisDeclContext()->getDecl()))) 272 return; 273 274 ProgramStateRef state = C.getState(); 275 SelfFlagEnum prevFlags = (SelfFlagEnum)state->get<PreCallSelfFlags>(); 276 if (!prevFlags) 277 return; 278 state = state->remove<PreCallSelfFlags>(); 279 280 unsigned NumArgs = CE.getNumArgs(); 281 for (unsigned i = 0; i < NumArgs; ++i) { 282 SVal argV = CE.getArgSVal(i); 283 if (isSelfVar(argV, C)) { 284 // If the address of 'self' is being passed to the call, assume that the 285 // 'self' after the call will have the same flags. 286 // EX: log(&self) 287 addSelfFlag(state, state->getSVal(argV.castAs<Loc>()), prevFlags, C); 288 return; 289 } else if (hasSelfFlag(argV, SelfFlag_Self, C)) { 290 // If 'self' is passed to the call by value, assume that the function 291 // returns 'self'. So assign the flags, which were set on 'self' to the 292 // return value. 293 // EX: self = performMoreInitialization(self) 294 addSelfFlag(state, CE.getReturnValue(), prevFlags, C); 295 return; 296 } 297 } 298 299 C.addTransition(state); 300} 301 302void ObjCSelfInitChecker::checkLocation(SVal location, bool isLoad, 303 const Stmt *S, 304 CheckerContext &C) const { 305 if (!shouldRunOnFunctionOrMethod(dyn_cast<NamedDecl>( 306 C.getCurrentAnalysisDeclContext()->getDecl()))) 307 return; 308 309 // Tag the result of a load from 'self' so that we can easily know that the 310 // value is the object that 'self' points to. 311 ProgramStateRef state = C.getState(); 312 if (isSelfVar(location, C)) 313 addSelfFlag(state, state->getSVal(location.castAs<Loc>()), SelfFlag_Self, 314 C); 315} 316 317 318void ObjCSelfInitChecker::checkBind(SVal loc, SVal val, const Stmt *S, 319 CheckerContext &C) const { 320 // Allow assignment of anything to self. Self is a local variable in the 321 // initializer, so it is legal to assign anything to it, like results of 322 // static functions/method calls. After self is assigned something we cannot 323 // reason about, stop enforcing the rules. 324 // (Only continue checking if the assigned value should be treated as self.) 325 if ((isSelfVar(loc, C)) && 326 !hasSelfFlag(val, SelfFlag_InitRes, C) && 327 !hasSelfFlag(val, SelfFlag_Self, C) && 328 !isSelfVar(val, C)) { 329 330 // Stop tracking the checker-specific state in the state. 331 ProgramStateRef State = C.getState(); 332 State = State->remove<CalledInit>(); 333 if (SymbolRef sym = loc.getAsSymbol()) 334 State = State->remove<SelfFlag>(sym); 335 C.addTransition(State); 336 } 337} 338 339void ObjCSelfInitChecker::printState(raw_ostream &Out, ProgramStateRef State, 340 const char *NL, const char *Sep) const { 341 SelfFlagTy FlagMap = State->get<SelfFlag>(); 342 bool DidCallInit = State->get<CalledInit>(); 343 SelfFlagEnum PreCallFlags = (SelfFlagEnum)State->get<PreCallSelfFlags>(); 344 345 if (FlagMap.isEmpty() && !DidCallInit && !PreCallFlags) 346 return; 347 348 Out << Sep << NL << *this << " :" << NL; 349 350 if (DidCallInit) 351 Out << " An init method has been called." << NL; 352 353 if (PreCallFlags != SelfFlag_None) { 354 if (PreCallFlags & SelfFlag_Self) { 355 Out << " An argument of the current call came from the 'self' variable." 356 << NL; 357 } 358 if (PreCallFlags & SelfFlag_InitRes) { 359 Out << " An argument of the current call came from an init method." 360 << NL; 361 } 362 } 363 364 Out << NL; 365 for (SelfFlagTy::iterator I = FlagMap.begin(), E = FlagMap.end(); 366 I != E; ++I) { 367 Out << I->first << " : "; 368 369 if (I->second == SelfFlag_None) 370 Out << "none"; 371 372 if (I->second & SelfFlag_Self) 373 Out << "self variable"; 374 375 if (I->second & SelfFlag_InitRes) { 376 if (I->second != SelfFlag_InitRes) 377 Out << " | "; 378 Out << "result of init method"; 379 } 380 381 Out << NL; 382 } 383} 384 385 386// FIXME: A callback should disable checkers at the start of functions. 387static bool shouldRunOnFunctionOrMethod(const NamedDecl *ND) { 388 if (!ND) 389 return false; 390 391 const ObjCMethodDecl *MD = dyn_cast<ObjCMethodDecl>(ND); 392 if (!MD) 393 return false; 394 if (!isInitializationMethod(MD)) 395 return false; 396 397 // self = [super init] applies only to NSObject subclasses. 398 // For instance, NSProxy doesn't implement -init. 399 ASTContext &Ctx = MD->getASTContext(); 400 IdentifierInfo* NSObjectII = &Ctx.Idents.get("NSObject"); 401 ObjCInterfaceDecl *ID = MD->getClassInterface()->getSuperClass(); 402 for ( ; ID ; ID = ID->getSuperClass()) { 403 IdentifierInfo *II = ID->getIdentifier(); 404 405 if (II == NSObjectII) 406 break; 407 } 408 if (!ID) 409 return false; 410 411 return true; 412} 413 414/// \brief Returns true if the location is 'self'. 415static bool isSelfVar(SVal location, CheckerContext &C) { 416 AnalysisDeclContext *analCtx = C.getCurrentAnalysisDeclContext(); 417 if (!analCtx->getSelfDecl()) 418 return false; 419 if (!location.getAs<loc::MemRegionVal>()) 420 return false; 421 422 loc::MemRegionVal MRV = location.castAs<loc::MemRegionVal>(); 423 if (const DeclRegion *DR = dyn_cast<DeclRegion>(MRV.stripCasts())) 424 return (DR->getDecl() == analCtx->getSelfDecl()); 425 426 return false; 427} 428 429static bool isInitializationMethod(const ObjCMethodDecl *MD) { 430 return MD->getMethodFamily() == OMF_init; 431} 432 433static bool isInitMessage(const ObjCMethodCall &Call) { 434 return Call.getMethodFamily() == OMF_init; 435} 436 437//===----------------------------------------------------------------------===// 438// Registration. 439//===----------------------------------------------------------------------===// 440 441void ento::registerObjCSelfInitChecker(CheckerManager &mgr) { 442 mgr.registerChecker<ObjCSelfInitChecker>(); 443} 444