SemaChecking.cpp revision bd2e27e8484c1d378c9c9d089f61ff496627f391 
1//===--- SemaChecking.cpp - Extra Semantic Checking -----------------------===//
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 file implements extra semantic analysis beyond what is enforced
11//  by the C type system.
12//
13//===----------------------------------------------------------------------===//
14
15#include "clang/Sema/Initialization.h"
16#include "clang/Sema/Sema.h"
17#include "clang/Sema/SemaInternal.h"
18#include "clang/Sema/Initialization.h"
19#include "clang/Sema/Lookup.h"
20#include "clang/Sema/ScopeInfo.h"
21#include "clang/Analysis/Analyses/FormatString.h"
22#include "clang/AST/ASTContext.h"
23#include "clang/AST/CharUnits.h"
24#include "clang/AST/DeclCXX.h"
25#include "clang/AST/DeclObjC.h"
26#include "clang/AST/Expr.h"
27#include "clang/AST/ExprCXX.h"
28#include "clang/AST/ExprObjC.h"
29#include "clang/AST/EvaluatedExprVisitor.h"
30#include "clang/AST/DeclObjC.h"
31#include "clang/AST/StmtCXX.h"
32#include "clang/AST/StmtObjC.h"
33#include "clang/Lex/Preprocessor.h"
34#include "llvm/ADT/BitVector.h"
35#include "llvm/ADT/SmallString.h"
36#include "llvm/ADT/STLExtras.h"
37#include "llvm/Support/raw_ostream.h"
38#include "clang/Basic/TargetBuiltins.h"
39#include "clang/Basic/TargetInfo.h"
40#include "clang/Basic/ConvertUTF.h"
41#include <limits>
42using namespace clang;
43using namespace sema;
44
45SourceLocation Sema::getLocationOfStringLiteralByte(const StringLiteral *SL,
46                                                    unsigned ByteNo) const {
47  return SL->getLocationOfByte(ByteNo, PP.getSourceManager(),
48                               PP.getLangOpts(), PP.getTargetInfo());
49}
50
51/// Checks that a call expression's argument count is the desired number.
52/// This is useful when doing custom type-checking.  Returns true on error.
53static bool checkArgCount(Sema &S, CallExpr *call, unsigned desiredArgCount) {
54  unsigned argCount = call->getNumArgs();
55  if (argCount == desiredArgCount) return false;
56
57  if (argCount < desiredArgCount)
58    return S.Diag(call->getLocEnd(), diag::err_typecheck_call_too_few_args)
59        << 0 /*function call*/ << desiredArgCount << argCount
60        << call->getSourceRange();
61
62  // Highlight all the excess arguments.
63  SourceRange range(call->getArg(desiredArgCount)->getLocStart(),
64                    call->getArg(argCount - 1)->getLocEnd());
65
66  return S.Diag(range.getBegin(), diag::err_typecheck_call_too_many_args)
67    << 0 /*function call*/ << desiredArgCount << argCount
68    << call->getArg(1)->getSourceRange();
69}
70
71/// Check that the first argument to __builtin_annotation is an integer
72/// and the second argument is a non-wide string literal.
73static bool SemaBuiltinAnnotation(Sema &S, CallExpr *TheCall) {
74  if (checkArgCount(S, TheCall, 2))
75    return true;
76
77  // First argument should be an integer.
78  Expr *ValArg = TheCall->getArg(0);
79  QualType Ty = ValArg->getType();
80  if (!Ty->isIntegerType()) {
81    S.Diag(ValArg->getLocStart(), diag::err_builtin_annotation_first_arg)
82      << ValArg->getSourceRange();
83    return true;
84  }
85
86  // Second argument should be a constant string.
87  Expr *StrArg = TheCall->getArg(1)->IgnoreParenCasts();
88  StringLiteral *Literal = dyn_cast<StringLiteral>(StrArg);
89  if (!Literal || !Literal->isAscii()) {
90    S.Diag(StrArg->getLocStart(), diag::err_builtin_annotation_second_arg)
91      << StrArg->getSourceRange();
92    return true;
93  }
94
95  TheCall->setType(Ty);
96  return false;
97}
98
99ExprResult
100Sema::CheckBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
101  ExprResult TheCallResult(Owned(TheCall));
102
103  // Find out if any arguments are required to be integer constant expressions.
104  unsigned ICEArguments = 0;
105  ASTContext::GetBuiltinTypeError Error;
106  Context.GetBuiltinType(BuiltinID, Error, &ICEArguments);
107  if (Error != ASTContext::GE_None)
108    ICEArguments = 0;  // Don't diagnose previously diagnosed errors.
109
110  // If any arguments are required to be ICE's, check and diagnose.
111  for (unsigned ArgNo = 0; ICEArguments != 0; ++ArgNo) {
112    // Skip arguments not required to be ICE's.
113    if ((ICEArguments & (1 << ArgNo)) == 0) continue;
114
115    llvm::APSInt Result;
116    if (SemaBuiltinConstantArg(TheCall, ArgNo, Result))
117      return true;
118    ICEArguments &= ~(1 << ArgNo);
119  }
120
121  switch (BuiltinID) {
122  case Builtin::BI__builtin___CFStringMakeConstantString:
123    assert(TheCall->getNumArgs() == 1 &&
124           "Wrong # arguments to builtin CFStringMakeConstantString");
125    if (CheckObjCString(TheCall->getArg(0)))
126      return ExprError();
127    break;
128  case Builtin::BI__builtin_stdarg_start:
129  case Builtin::BI__builtin_va_start:
130    if (SemaBuiltinVAStart(TheCall))
131      return ExprError();
132    break;
133  case Builtin::BI__builtin_isgreater:
134  case Builtin::BI__builtin_isgreaterequal:
135  case Builtin::BI__builtin_isless:
136  case Builtin::BI__builtin_islessequal:
137  case Builtin::BI__builtin_islessgreater:
138  case Builtin::BI__builtin_isunordered:
139    if (SemaBuiltinUnorderedCompare(TheCall))
140      return ExprError();
141    break;
142  case Builtin::BI__builtin_fpclassify:
143    if (SemaBuiltinFPClassification(TheCall, 6))
144      return ExprError();
145    break;
146  case Builtin::BI__builtin_isfinite:
147  case Builtin::BI__builtin_isinf:
148  case Builtin::BI__builtin_isinf_sign:
149  case Builtin::BI__builtin_isnan:
150  case Builtin::BI__builtin_isnormal:
151    if (SemaBuiltinFPClassification(TheCall, 1))
152      return ExprError();
153    break;
154  case Builtin::BI__builtin_shufflevector:
155    return SemaBuiltinShuffleVector(TheCall);
156    // TheCall will be freed by the smart pointer here, but that's fine, since
157    // SemaBuiltinShuffleVector guts it, but then doesn't release it.
158  case Builtin::BI__builtin_prefetch:
159    if (SemaBuiltinPrefetch(TheCall))
160      return ExprError();
161    break;
162  case Builtin::BI__builtin_object_size:
163    if (SemaBuiltinObjectSize(TheCall))
164      return ExprError();
165    break;
166  case Builtin::BI__builtin_longjmp:
167    if (SemaBuiltinLongjmp(TheCall))
168      return ExprError();
169    break;
170
171  case Builtin::BI__builtin_classify_type:
172    if (checkArgCount(*this, TheCall, 1)) return true;
173    TheCall->setType(Context.IntTy);
174    break;
175  case Builtin::BI__builtin_constant_p:
176    if (checkArgCount(*this, TheCall, 1)) return true;
177    TheCall->setType(Context.IntTy);
178    break;
179  case Builtin::BI__sync_fetch_and_add:
180  case Builtin::BI__sync_fetch_and_add_1:
181  case Builtin::BI__sync_fetch_and_add_2:
182  case Builtin::BI__sync_fetch_and_add_4:
183  case Builtin::BI__sync_fetch_and_add_8:
184  case Builtin::BI__sync_fetch_and_add_16:
185  case Builtin::BI__sync_fetch_and_sub:
186  case Builtin::BI__sync_fetch_and_sub_1:
187  case Builtin::BI__sync_fetch_and_sub_2:
188  case Builtin::BI__sync_fetch_and_sub_4:
189  case Builtin::BI__sync_fetch_and_sub_8:
190  case Builtin::BI__sync_fetch_and_sub_16:
191  case Builtin::BI__sync_fetch_and_or:
192  case Builtin::BI__sync_fetch_and_or_1:
193  case Builtin::BI__sync_fetch_and_or_2:
194  case Builtin::BI__sync_fetch_and_or_4:
195  case Builtin::BI__sync_fetch_and_or_8:
196  case Builtin::BI__sync_fetch_and_or_16:
197  case Builtin::BI__sync_fetch_and_and:
198  case Builtin::BI__sync_fetch_and_and_1:
199  case Builtin::BI__sync_fetch_and_and_2:
200  case Builtin::BI__sync_fetch_and_and_4:
201  case Builtin::BI__sync_fetch_and_and_8:
202  case Builtin::BI__sync_fetch_and_and_16:
203  case Builtin::BI__sync_fetch_and_xor:
204  case Builtin::BI__sync_fetch_and_xor_1:
205  case Builtin::BI__sync_fetch_and_xor_2:
206  case Builtin::BI__sync_fetch_and_xor_4:
207  case Builtin::BI__sync_fetch_and_xor_8:
208  case Builtin::BI__sync_fetch_and_xor_16:
209  case Builtin::BI__sync_add_and_fetch:
210  case Builtin::BI__sync_add_and_fetch_1:
211  case Builtin::BI__sync_add_and_fetch_2:
212  case Builtin::BI__sync_add_and_fetch_4:
213  case Builtin::BI__sync_add_and_fetch_8:
214  case Builtin::BI__sync_add_and_fetch_16:
215  case Builtin::BI__sync_sub_and_fetch:
216  case Builtin::BI__sync_sub_and_fetch_1:
217  case Builtin::BI__sync_sub_and_fetch_2:
218  case Builtin::BI__sync_sub_and_fetch_4:
219  case Builtin::BI__sync_sub_and_fetch_8:
220  case Builtin::BI__sync_sub_and_fetch_16:
221  case Builtin::BI__sync_and_and_fetch:
222  case Builtin::BI__sync_and_and_fetch_1:
223  case Builtin::BI__sync_and_and_fetch_2:
224  case Builtin::BI__sync_and_and_fetch_4:
225  case Builtin::BI__sync_and_and_fetch_8:
226  case Builtin::BI__sync_and_and_fetch_16:
227  case Builtin::BI__sync_or_and_fetch:
228  case Builtin::BI__sync_or_and_fetch_1:
229  case Builtin::BI__sync_or_and_fetch_2:
230  case Builtin::BI__sync_or_and_fetch_4:
231  case Builtin::BI__sync_or_and_fetch_8:
232  case Builtin::BI__sync_or_and_fetch_16:
233  case Builtin::BI__sync_xor_and_fetch:
234  case Builtin::BI__sync_xor_and_fetch_1:
235  case Builtin::BI__sync_xor_and_fetch_2:
236  case Builtin::BI__sync_xor_and_fetch_4:
237  case Builtin::BI__sync_xor_and_fetch_8:
238  case Builtin::BI__sync_xor_and_fetch_16:
239  case Builtin::BI__sync_val_compare_and_swap:
240  case Builtin::BI__sync_val_compare_and_swap_1:
241  case Builtin::BI__sync_val_compare_and_swap_2:
242  case Builtin::BI__sync_val_compare_and_swap_4:
243  case Builtin::BI__sync_val_compare_and_swap_8:
244  case Builtin::BI__sync_val_compare_and_swap_16:
245  case Builtin::BI__sync_bool_compare_and_swap:
246  case Builtin::BI__sync_bool_compare_and_swap_1:
247  case Builtin::BI__sync_bool_compare_and_swap_2:
248  case Builtin::BI__sync_bool_compare_and_swap_4:
249  case Builtin::BI__sync_bool_compare_and_swap_8:
250  case Builtin::BI__sync_bool_compare_and_swap_16:
251  case Builtin::BI__sync_lock_test_and_set:
252  case Builtin::BI__sync_lock_test_and_set_1:
253  case Builtin::BI__sync_lock_test_and_set_2:
254  case Builtin::BI__sync_lock_test_and_set_4:
255  case Builtin::BI__sync_lock_test_and_set_8:
256  case Builtin::BI__sync_lock_test_and_set_16:
257  case Builtin::BI__sync_lock_release:
258  case Builtin::BI__sync_lock_release_1:
259  case Builtin::BI__sync_lock_release_2:
260  case Builtin::BI__sync_lock_release_4:
261  case Builtin::BI__sync_lock_release_8:
262  case Builtin::BI__sync_lock_release_16:
263  case Builtin::BI__sync_swap:
264  case Builtin::BI__sync_swap_1:
265  case Builtin::BI__sync_swap_2:
266  case Builtin::BI__sync_swap_4:
267  case Builtin::BI__sync_swap_8:
268  case Builtin::BI__sync_swap_16:
269    return SemaBuiltinAtomicOverloaded(move(TheCallResult));
270#define BUILTIN(ID, TYPE, ATTRS)
271#define ATOMIC_BUILTIN(ID, TYPE, ATTRS) \
272  case Builtin::BI##ID: \
273    return SemaAtomicOpsOverloaded(move(TheCallResult), AtomicExpr::AO##ID);
274#include "clang/Basic/Builtins.def"
275  case Builtin::BI__builtin_annotation:
276    if (SemaBuiltinAnnotation(*this, TheCall))
277      return ExprError();
278    break;
279  }
280
281  // Since the target specific builtins for each arch overlap, only check those
282  // of the arch we are compiling for.
283  if (BuiltinID >= Builtin::FirstTSBuiltin) {
284    switch (Context.getTargetInfo().getTriple().getArch()) {
285      case llvm::Triple::arm:
286      case llvm::Triple::thumb:
287        if (CheckARMBuiltinFunctionCall(BuiltinID, TheCall))
288          return ExprError();
289        break;
290      default:
291        break;
292    }
293  }
294
295  return move(TheCallResult);
296}
297
298// Get the valid immediate range for the specified NEON type code.
299static unsigned RFT(unsigned t, bool shift = false) {
300  NeonTypeFlags Type(t);
301  int IsQuad = Type.isQuad();
302  switch (Type.getEltType()) {
303  case NeonTypeFlags::Int8:
304  case NeonTypeFlags::Poly8:
305    return shift ? 7 : (8 << IsQuad) - 1;
306  case NeonTypeFlags::Int16:
307  case NeonTypeFlags::Poly16:
308    return shift ? 15 : (4 << IsQuad) - 1;
309  case NeonTypeFlags::Int32:
310    return shift ? 31 : (2 << IsQuad) - 1;
311  case NeonTypeFlags::Int64:
312    return shift ? 63 : (1 << IsQuad) - 1;
313  case NeonTypeFlags::Float16:
314    assert(!shift && "cannot shift float types!");
315    return (4 << IsQuad) - 1;
316  case NeonTypeFlags::Float32:
317    assert(!shift && "cannot shift float types!");
318    return (2 << IsQuad) - 1;
319  }
320  llvm_unreachable("Invalid NeonTypeFlag!");
321}
322
323/// getNeonEltType - Return the QualType corresponding to the elements of
324/// the vector type specified by the NeonTypeFlags.  This is used to check
325/// the pointer arguments for Neon load/store intrinsics.
326static QualType getNeonEltType(NeonTypeFlags Flags, ASTContext &Context) {
327  switch (Flags.getEltType()) {
328  case NeonTypeFlags::Int8:
329    return Flags.isUnsigned() ? Context.UnsignedCharTy : Context.SignedCharTy;
330  case NeonTypeFlags::Int16:
331    return Flags.isUnsigned() ? Context.UnsignedShortTy : Context.ShortTy;
332  case NeonTypeFlags::Int32:
333    return Flags.isUnsigned() ? Context.UnsignedIntTy : Context.IntTy;
334  case NeonTypeFlags::Int64:
335    return Flags.isUnsigned() ? Context.UnsignedLongLongTy : Context.LongLongTy;
336  case NeonTypeFlags::Poly8:
337    return Context.SignedCharTy;
338  case NeonTypeFlags::Poly16:
339    return Context.ShortTy;
340  case NeonTypeFlags::Float16:
341    return Context.UnsignedShortTy;
342  case NeonTypeFlags::Float32:
343    return Context.FloatTy;
344  }
345  llvm_unreachable("Invalid NeonTypeFlag!");
346}
347
348bool Sema::CheckARMBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
349  llvm::APSInt Result;
350
351  unsigned mask = 0;
352  unsigned TV = 0;
353  int PtrArgNum = -1;
354  bool HasConstPtr = false;
355  switch (BuiltinID) {
356#define GET_NEON_OVERLOAD_CHECK
357#include "clang/Basic/arm_neon.inc"
358#undef GET_NEON_OVERLOAD_CHECK
359  }
360
361  // For NEON intrinsics which are overloaded on vector element type, validate
362  // the immediate which specifies which variant to emit.
363  unsigned ImmArg = TheCall->getNumArgs()-1;
364  if (mask) {
365    if (SemaBuiltinConstantArg(TheCall, ImmArg, Result))
366      return true;
367
368    TV = Result.getLimitedValue(64);
369    if ((TV > 63) || (mask & (1 << TV)) == 0)
370      return Diag(TheCall->getLocStart(), diag::err_invalid_neon_type_code)
371        << TheCall->getArg(ImmArg)->getSourceRange();
372  }
373
374  if (PtrArgNum >= 0) {
375    // Check that pointer arguments have the specified type.
376    Expr *Arg = TheCall->getArg(PtrArgNum);
377    if (ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(Arg))
378      Arg = ICE->getSubExpr();
379    ExprResult RHS = DefaultFunctionArrayLvalueConversion(Arg);
380    QualType RHSTy = RHS.get()->getType();
381    QualType EltTy = getNeonEltType(NeonTypeFlags(TV), Context);
382    if (HasConstPtr)
383      EltTy = EltTy.withConst();
384    QualType LHSTy = Context.getPointerType(EltTy);
385    AssignConvertType ConvTy;
386    ConvTy = CheckSingleAssignmentConstraints(LHSTy, RHS);
387    if (RHS.isInvalid())
388      return true;
389    if (DiagnoseAssignmentResult(ConvTy, Arg->getLocStart(), LHSTy, RHSTy,
390                                 RHS.get(), AA_Assigning))
391      return true;
392  }
393
394  // For NEON intrinsics which take an immediate value as part of the
395  // instruction, range check them here.
396  unsigned i = 0, l = 0, u = 0;
397  switch (BuiltinID) {
398  default: return false;
399  case ARM::BI__builtin_arm_ssat: i = 1; l = 1; u = 31; break;
400  case ARM::BI__builtin_arm_usat: i = 1; u = 31; break;
401  case ARM::BI__builtin_arm_vcvtr_f:
402  case ARM::BI__builtin_arm_vcvtr_d: i = 1; u = 1; break;
403#define GET_NEON_IMMEDIATE_CHECK
404#include "clang/Basic/arm_neon.inc"
405#undef GET_NEON_IMMEDIATE_CHECK
406  };
407
408  // We can't check the value of a dependent argument.
409  if (TheCall->getArg(i)->isTypeDependent() ||
410      TheCall->getArg(i)->isValueDependent())
411    return false;
412
413  // Check that the immediate argument is actually a constant.
414  if (SemaBuiltinConstantArg(TheCall, i, Result))
415    return true;
416
417  // Range check against the upper/lower values for this isntruction.
418  unsigned Val = Result.getZExtValue();
419  if (Val < l || Val > (u + l))
420    return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
421      << l << u+l << TheCall->getArg(i)->getSourceRange();
422
423  // FIXME: VFP Intrinsics should error if VFP not present.
424  return false;
425}
426
427/// Given a FunctionDecl's FormatAttr, attempts to populate the FomatStringInfo
428/// parameter with the FormatAttr's correct format_idx and firstDataArg.
429/// Returns true when the format fits the function and the FormatStringInfo has
430/// been populated.
431bool Sema::getFormatStringInfo(const FormatAttr *Format, bool IsCXXMember,
432                               FormatStringInfo *FSI) {
433  FSI->HasVAListArg = Format->getFirstArg() == 0;
434  FSI->FormatIdx = Format->getFormatIdx() - 1;
435  FSI->FirstDataArg = FSI->HasVAListArg ? 0 : Format->getFirstArg() - 1;
436
437  // The way the format attribute works in GCC, the implicit this argument
438  // of member functions is counted. However, it doesn't appear in our own
439  // lists, so decrement format_idx in that case.
440  if (IsCXXMember) {
441    if(FSI->FormatIdx == 0)
442      return false;
443    --FSI->FormatIdx;
444    if (FSI->FirstDataArg != 0)
445      --FSI->FirstDataArg;
446  }
447  return true;
448}
449
450/// Handles the checks for format strings, non-POD arguments to vararg
451/// functions, and NULL arguments passed to non-NULL parameters.
452void Sema::checkCall(NamedDecl *FDecl, Expr **Args,
453                     unsigned NumArgs,
454                     unsigned NumProtoArgs,
455                     bool IsMemberFunction,
456                     SourceLocation Loc,
457                     SourceRange Range,
458                     VariadicCallType CallType) {
459  // FIXME: This mechanism should be abstracted to be less fragile and
460  // more efficient. For example, just map function ids to custom
461  // handlers.
462
463  // Printf and scanf checking.
464  bool HandledFormatString = false;
465  for (specific_attr_iterator<FormatAttr>
466         I = FDecl->specific_attr_begin<FormatAttr>(),
467         E = FDecl->specific_attr_end<FormatAttr>(); I != E ; ++I)
468    if (CheckFormatArguments(*I, Args, NumArgs, IsMemberFunction, Loc, Range))
469        HandledFormatString = true;
470
471  // Refuse POD arguments that weren't caught by the format string
472  // checks above.
473  if (!HandledFormatString && CallType != VariadicDoesNotApply)
474    for (unsigned ArgIdx = NumProtoArgs; ArgIdx < NumArgs; ++ArgIdx)
475      variadicArgumentPODCheck(Args[ArgIdx], CallType);
476
477  for (specific_attr_iterator<NonNullAttr>
478         I = FDecl->specific_attr_begin<NonNullAttr>(),
479         E = FDecl->specific_attr_end<NonNullAttr>(); I != E; ++I)
480    CheckNonNullArguments(*I, Args, Loc);
481}
482
483/// CheckConstructorCall - Check a constructor call for correctness and safety
484/// properties not enforced by the C type system.
485void Sema::CheckConstructorCall(FunctionDecl *FDecl, Expr **Args,
486                                unsigned NumArgs,
487                                const FunctionProtoType *Proto,
488                                SourceLocation Loc) {
489  VariadicCallType CallType =
490    Proto->isVariadic() ? VariadicConstructor : VariadicDoesNotApply;
491  checkCall(FDecl, Args, NumArgs, Proto->getNumArgs(),
492            /*IsMemberFunction=*/true, Loc, SourceRange(), CallType);
493}
494
495/// CheckFunctionCall - Check a direct function call for various correctness
496/// and safety properties not strictly enforced by the C type system.
497bool Sema::CheckFunctionCall(FunctionDecl *FDecl, CallExpr *TheCall,
498                             const FunctionProtoType *Proto) {
499  bool IsMemberFunction = isa<CXXMemberCallExpr>(TheCall);
500  VariadicCallType CallType = getVariadicCallType(FDecl, Proto,
501                                                  TheCall->getCallee());
502  unsigned NumProtoArgs = Proto ? Proto->getNumArgs() : 0;
503  checkCall(FDecl, TheCall->getArgs(), TheCall->getNumArgs(), NumProtoArgs,
504            IsMemberFunction, TheCall->getRParenLoc(),
505            TheCall->getCallee()->getSourceRange(), CallType);
506
507  IdentifierInfo *FnInfo = FDecl->getIdentifier();
508  // None of the checks below are needed for functions that don't have
509  // simple names (e.g., C++ conversion functions).
510  if (!FnInfo)
511    return false;
512
513  unsigned CMId = FDecl->getMemoryFunctionKind();
514  if (CMId == 0)
515    return false;
516
517  // Handle memory setting and copying functions.
518  if (CMId == Builtin::BIstrlcpy || CMId == Builtin::BIstrlcat)
519    CheckStrlcpycatArguments(TheCall, FnInfo);
520  else if (CMId == Builtin::BIstrncat)
521    CheckStrncatArguments(TheCall, FnInfo);
522  else
523    CheckMemaccessArguments(TheCall, CMId, FnInfo);
524
525  return false;
526}
527
528bool Sema::CheckObjCMethodCall(ObjCMethodDecl *Method, SourceLocation lbrac,
529                               Expr **Args, unsigned NumArgs) {
530  VariadicCallType CallType =
531      Method->isVariadic() ? VariadicMethod : VariadicDoesNotApply;
532
533  checkCall(Method, Args, NumArgs, Method->param_size(),
534            /*IsMemberFunction=*/false,
535            lbrac, Method->getSourceRange(), CallType);
536
537  return false;
538}
539
540bool Sema::CheckBlockCall(NamedDecl *NDecl, CallExpr *TheCall,
541                          const FunctionProtoType *Proto) {
542  const VarDecl *V = dyn_cast<VarDecl>(NDecl);
543  if (!V)
544    return false;
545
546  QualType Ty = V->getType();
547  if (!Ty->isBlockPointerType())
548    return false;
549
550  VariadicCallType CallType =
551      Proto && Proto->isVariadic() ? VariadicBlock : VariadicDoesNotApply ;
552  unsigned NumProtoArgs = Proto ? Proto->getNumArgs() : 0;
553
554  checkCall(NDecl, TheCall->getArgs(), TheCall->getNumArgs(),
555            NumProtoArgs, /*IsMemberFunction=*/false,
556            TheCall->getRParenLoc(),
557            TheCall->getCallee()->getSourceRange(), CallType);
558
559  return false;
560}
561
562ExprResult Sema::SemaAtomicOpsOverloaded(ExprResult TheCallResult,
563                                         AtomicExpr::AtomicOp Op) {
564  CallExpr *TheCall = cast<CallExpr>(TheCallResult.get());
565  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
566
567  // All these operations take one of the following forms:
568  enum {
569    // C    __c11_atomic_init(A *, C)
570    Init,
571    // C    __c11_atomic_load(A *, int)
572    Load,
573    // void __atomic_load(A *, CP, int)
574    Copy,
575    // C    __c11_atomic_add(A *, M, int)
576    Arithmetic,
577    // C    __atomic_exchange_n(A *, CP, int)
578    Xchg,
579    // void __atomic_exchange(A *, C *, CP, int)
580    GNUXchg,
581    // bool __c11_atomic_compare_exchange_strong(A *, C *, CP, int, int)
582    C11CmpXchg,
583    // bool __atomic_compare_exchange(A *, C *, CP, bool, int, int)
584    GNUCmpXchg
585  } Form = Init;
586  const unsigned NumArgs[] = { 2, 2, 3, 3, 3, 4, 5, 6 };
587  const unsigned NumVals[] = { 1, 0, 1, 1, 1, 2, 2, 3 };
588  // where:
589  //   C is an appropriate type,
590  //   A is volatile _Atomic(C) for __c11 builtins and is C for GNU builtins,
591  //   CP is C for __c11 builtins and GNU _n builtins and is C * otherwise,
592  //   M is C if C is an integer, and ptrdiff_t if C is a pointer, and
593  //   the int parameters are for orderings.
594
595  assert(AtomicExpr::AO__c11_atomic_init == 0 &&
596         AtomicExpr::AO__c11_atomic_fetch_xor + 1 == AtomicExpr::AO__atomic_load
597         && "need to update code for modified C11 atomics");
598  bool IsC11 = Op >= AtomicExpr::AO__c11_atomic_init &&
599               Op <= AtomicExpr::AO__c11_atomic_fetch_xor;
600  bool IsN = Op == AtomicExpr::AO__atomic_load_n ||
601             Op == AtomicExpr::AO__atomic_store_n ||
602             Op == AtomicExpr::AO__atomic_exchange_n ||
603             Op == AtomicExpr::AO__atomic_compare_exchange_n;
604  bool IsAddSub = false;
605
606  switch (Op) {
607  case AtomicExpr::AO__c11_atomic_init:
608    Form = Init;
609    break;
610
611  case AtomicExpr::AO__c11_atomic_load:
612  case AtomicExpr::AO__atomic_load_n:
613    Form = Load;
614    break;
615
616  case AtomicExpr::AO__c11_atomic_store:
617  case AtomicExpr::AO__atomic_load:
618  case AtomicExpr::AO__atomic_store:
619  case AtomicExpr::AO__atomic_store_n:
620    Form = Copy;
621    break;
622
623  case AtomicExpr::AO__c11_atomic_fetch_add:
624  case AtomicExpr::AO__c11_atomic_fetch_sub:
625  case AtomicExpr::AO__atomic_fetch_add:
626  case AtomicExpr::AO__atomic_fetch_sub:
627  case AtomicExpr::AO__atomic_add_fetch:
628  case AtomicExpr::AO__atomic_sub_fetch:
629    IsAddSub = true;
630    // Fall through.
631  case AtomicExpr::AO__c11_atomic_fetch_and:
632  case AtomicExpr::AO__c11_atomic_fetch_or:
633  case AtomicExpr::AO__c11_atomic_fetch_xor:
634  case AtomicExpr::AO__atomic_fetch_and:
635  case AtomicExpr::AO__atomic_fetch_or:
636  case AtomicExpr::AO__atomic_fetch_xor:
637  case AtomicExpr::AO__atomic_fetch_nand:
638  case AtomicExpr::AO__atomic_and_fetch:
639  case AtomicExpr::AO__atomic_or_fetch:
640  case AtomicExpr::AO__atomic_xor_fetch:
641  case AtomicExpr::AO__atomic_nand_fetch:
642    Form = Arithmetic;
643    break;
644
645  case AtomicExpr::AO__c11_atomic_exchange:
646  case AtomicExpr::AO__atomic_exchange_n:
647    Form = Xchg;
648    break;
649
650  case AtomicExpr::AO__atomic_exchange:
651    Form = GNUXchg;
652    break;
653
654  case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
655  case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
656    Form = C11CmpXchg;
657    break;
658
659  case AtomicExpr::AO__atomic_compare_exchange:
660  case AtomicExpr::AO__atomic_compare_exchange_n:
661    Form = GNUCmpXchg;
662    break;
663  }
664
665  // Check we have the right number of arguments.
666  if (TheCall->getNumArgs() < NumArgs[Form]) {
667    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
668      << 0 << NumArgs[Form] << TheCall->getNumArgs()
669      << TheCall->getCallee()->getSourceRange();
670    return ExprError();
671  } else if (TheCall->getNumArgs() > NumArgs[Form]) {
672    Diag(TheCall->getArg(NumArgs[Form])->getLocStart(),
673         diag::err_typecheck_call_too_many_args)
674      << 0 << NumArgs[Form] << TheCall->getNumArgs()
675      << TheCall->getCallee()->getSourceRange();
676    return ExprError();
677  }
678
679  // Inspect the first argument of the atomic operation.
680  Expr *Ptr = TheCall->getArg(0);
681  Ptr = DefaultFunctionArrayLvalueConversion(Ptr).get();
682  const PointerType *pointerType = Ptr->getType()->getAs<PointerType>();
683  if (!pointerType) {
684    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer)
685      << Ptr->getType() << Ptr->getSourceRange();
686    return ExprError();
687  }
688
689  // For a __c11 builtin, this should be a pointer to an _Atomic type.
690  QualType AtomTy = pointerType->getPointeeType(); // 'A'
691  QualType ValType = AtomTy; // 'C'
692  if (IsC11) {
693    if (!AtomTy->isAtomicType()) {
694      Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic)
695        << Ptr->getType() << Ptr->getSourceRange();
696      return ExprError();
697    }
698    ValType = AtomTy->getAs<AtomicType>()->getValueType();
699  }
700
701  // For an arithmetic operation, the implied arithmetic must be well-formed.
702  if (Form == Arithmetic) {
703    // gcc does not enforce these rules for GNU atomics, but we do so for sanity.
704    if (IsAddSub && !ValType->isIntegerType() && !ValType->isPointerType()) {
705      Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic_int_or_ptr)
706        << IsC11 << Ptr->getType() << Ptr->getSourceRange();
707      return ExprError();
708    }
709    if (!IsAddSub && !ValType->isIntegerType()) {
710      Diag(DRE->getLocStart(), diag::err_atomic_op_bitwise_needs_atomic_int)
711        << IsC11 << Ptr->getType() << Ptr->getSourceRange();
712      return ExprError();
713    }
714  } else if (IsN && !ValType->isIntegerType() && !ValType->isPointerType()) {
715    // For __atomic_*_n operations, the value type must be a scalar integral or
716    // pointer type which is 1, 2, 4, 8 or 16 bytes in length.
717    Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic_int_or_ptr)
718      << IsC11 << Ptr->getType() << Ptr->getSourceRange();
719    return ExprError();
720  }
721
722  if (!IsC11 && !AtomTy.isTriviallyCopyableType(Context)) {
723    // For GNU atomics, require a trivially-copyable type. This is not part of
724    // the GNU atomics specification, but we enforce it for sanity.
725    Diag(DRE->getLocStart(), diag::err_atomic_op_needs_trivial_copy)
726      << Ptr->getType() << Ptr->getSourceRange();
727    return ExprError();
728  }
729
730  // FIXME: For any builtin other than a load, the ValType must not be
731  // const-qualified.
732
733  switch (ValType.getObjCLifetime()) {
734  case Qualifiers::OCL_None:
735  case Qualifiers::OCL_ExplicitNone:
736    // okay
737    break;
738
739  case Qualifiers::OCL_Weak:
740  case Qualifiers::OCL_Strong:
741  case Qualifiers::OCL_Autoreleasing:
742    // FIXME: Can this happen? By this point, ValType should be known
743    // to be trivially copyable.
744    Diag(DRE->getLocStart(), diag::err_arc_atomic_ownership)
745      << ValType << Ptr->getSourceRange();
746    return ExprError();
747  }
748
749  QualType ResultType = ValType;
750  if (Form == Copy || Form == GNUXchg || Form == Init)
751    ResultType = Context.VoidTy;
752  else if (Form == C11CmpXchg || Form == GNUCmpXchg)
753    ResultType = Context.BoolTy;
754
755  // The type of a parameter passed 'by value'. In the GNU atomics, such
756  // arguments are actually passed as pointers.
757  QualType ByValType = ValType; // 'CP'
758  if (!IsC11 && !IsN)
759    ByValType = Ptr->getType();
760
761  // The first argument --- the pointer --- has a fixed type; we
762  // deduce the types of the rest of the arguments accordingly.  Walk
763  // the remaining arguments, converting them to the deduced value type.
764  for (unsigned i = 1; i != NumArgs[Form]; ++i) {
765    QualType Ty;
766    if (i < NumVals[Form] + 1) {
767      switch (i) {
768      case 1:
769        // The second argument is the non-atomic operand. For arithmetic, this
770        // is always passed by value, and for a compare_exchange it is always
771        // passed by address. For the rest, GNU uses by-address and C11 uses
772        // by-value.
773        assert(Form != Load);
774        if (Form == Init || (Form == Arithmetic && ValType->isIntegerType()))
775          Ty = ValType;
776        else if (Form == Copy || Form == Xchg)
777          Ty = ByValType;
778        else if (Form == Arithmetic)
779          Ty = Context.getPointerDiffType();
780        else
781          Ty = Context.getPointerType(ValType.getUnqualifiedType());
782        break;
783      case 2:
784        // The third argument to compare_exchange / GNU exchange is a
785        // (pointer to a) desired value.
786        Ty = ByValType;
787        break;
788      case 3:
789        // The fourth argument to GNU compare_exchange is a 'weak' flag.
790        Ty = Context.BoolTy;
791        break;
792      }
793    } else {
794      // The order(s) are always converted to int.
795      Ty = Context.IntTy;
796    }
797
798    InitializedEntity Entity =
799        InitializedEntity::InitializeParameter(Context, Ty, false);
800    ExprResult Arg = TheCall->getArg(i);
801    Arg = PerformCopyInitialization(Entity, SourceLocation(), Arg);
802    if (Arg.isInvalid())
803      return true;
804    TheCall->setArg(i, Arg.get());
805  }
806
807  // Permute the arguments into a 'consistent' order.
808  SmallVector<Expr*, 5> SubExprs;
809  SubExprs.push_back(Ptr);
810  switch (Form) {
811  case Init:
812    // Note, AtomicExpr::getVal1() has a special case for this atomic.
813    SubExprs.push_back(TheCall->getArg(1)); // Val1
814    break;
815  case Load:
816    SubExprs.push_back(TheCall->getArg(1)); // Order
817    break;
818  case Copy:
819  case Arithmetic:
820  case Xchg:
821    SubExprs.push_back(TheCall->getArg(2)); // Order
822    SubExprs.push_back(TheCall->getArg(1)); // Val1
823    break;
824  case GNUXchg:
825    // Note, AtomicExpr::getVal2() has a special case for this atomic.
826    SubExprs.push_back(TheCall->getArg(3)); // Order
827    SubExprs.push_back(TheCall->getArg(1)); // Val1
828    SubExprs.push_back(TheCall->getArg(2)); // Val2
829    break;
830  case C11CmpXchg:
831    SubExprs.push_back(TheCall->getArg(3)); // Order
832    SubExprs.push_back(TheCall->getArg(1)); // Val1
833    SubExprs.push_back(TheCall->getArg(4)); // OrderFail
834    SubExprs.push_back(TheCall->getArg(2)); // Val2
835    break;
836  case GNUCmpXchg:
837    SubExprs.push_back(TheCall->getArg(4)); // Order
838    SubExprs.push_back(TheCall->getArg(1)); // Val1
839    SubExprs.push_back(TheCall->getArg(5)); // OrderFail
840    SubExprs.push_back(TheCall->getArg(2)); // Val2
841    SubExprs.push_back(TheCall->getArg(3)); // Weak
842    break;
843  }
844
845  return Owned(new (Context) AtomicExpr(TheCall->getCallee()->getLocStart(),
846                                        SubExprs.data(), SubExprs.size(),
847                                        ResultType, Op,
848                                        TheCall->getRParenLoc()));
849}
850
851
852/// checkBuiltinArgument - Given a call to a builtin function, perform
853/// normal type-checking on the given argument, updating the call in
854/// place.  This is useful when a builtin function requires custom
855/// type-checking for some of its arguments but not necessarily all of
856/// them.
857///
858/// Returns true on error.
859static bool checkBuiltinArgument(Sema &S, CallExpr *E, unsigned ArgIndex) {
860  FunctionDecl *Fn = E->getDirectCallee();
861  assert(Fn && "builtin call without direct callee!");
862
863  ParmVarDecl *Param = Fn->getParamDecl(ArgIndex);
864  InitializedEntity Entity =
865    InitializedEntity::InitializeParameter(S.Context, Param);
866
867  ExprResult Arg = E->getArg(0);
868  Arg = S.PerformCopyInitialization(Entity, SourceLocation(), Arg);
869  if (Arg.isInvalid())
870    return true;
871
872  E->setArg(ArgIndex, Arg.take());
873  return false;
874}
875
876/// SemaBuiltinAtomicOverloaded - We have a call to a function like
877/// __sync_fetch_and_add, which is an overloaded function based on the pointer
878/// type of its first argument.  The main ActOnCallExpr routines have already
879/// promoted the types of arguments because all of these calls are prototyped as
880/// void(...).
881///
882/// This function goes through and does final semantic checking for these
883/// builtins,
884ExprResult
885Sema::SemaBuiltinAtomicOverloaded(ExprResult TheCallResult) {
886  CallExpr *TheCall = (CallExpr *)TheCallResult.get();
887  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
888  FunctionDecl *FDecl = cast<FunctionDecl>(DRE->getDecl());
889
890  // Ensure that we have at least one argument to do type inference from.
891  if (TheCall->getNumArgs() < 1) {
892    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args_at_least)
893      << 0 << 1 << TheCall->getNumArgs()
894      << TheCall->getCallee()->getSourceRange();
895    return ExprError();
896  }
897
898  // Inspect the first argument of the atomic builtin.  This should always be
899  // a pointer type, whose element is an integral scalar or pointer type.
900  // Because it is a pointer type, we don't have to worry about any implicit
901  // casts here.
902  // FIXME: We don't allow floating point scalars as input.
903  Expr *FirstArg = TheCall->getArg(0);
904  ExprResult FirstArgResult = DefaultFunctionArrayLvalueConversion(FirstArg);
905  if (FirstArgResult.isInvalid())
906    return ExprError();
907  FirstArg = FirstArgResult.take();
908  TheCall->setArg(0, FirstArg);
909
910  const PointerType *pointerType = FirstArg->getType()->getAs<PointerType>();
911  if (!pointerType) {
912    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer)
913      << FirstArg->getType() << FirstArg->getSourceRange();
914    return ExprError();
915  }
916
917  QualType ValType = pointerType->getPointeeType();
918  if (!ValType->isIntegerType() && !ValType->isAnyPointerType() &&
919      !ValType->isBlockPointerType()) {
920    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer_intptr)
921      << FirstArg->getType() << FirstArg->getSourceRange();
922    return ExprError();
923  }
924
925  switch (ValType.getObjCLifetime()) {
926  case Qualifiers::OCL_None:
927  case Qualifiers::OCL_ExplicitNone:
928    // okay
929    break;
930
931  case Qualifiers::OCL_Weak:
932  case Qualifiers::OCL_Strong:
933  case Qualifiers::OCL_Autoreleasing:
934    Diag(DRE->getLocStart(), diag::err_arc_atomic_ownership)
935      << ValType << FirstArg->getSourceRange();
936    return ExprError();
937  }
938
939  // Strip any qualifiers off ValType.
940  ValType = ValType.getUnqualifiedType();
941
942  // The majority of builtins return a value, but a few have special return
943  // types, so allow them to override appropriately below.
944  QualType ResultType = ValType;
945
946  // We need to figure out which concrete builtin this maps onto.  For example,
947  // __sync_fetch_and_add with a 2 byte object turns into
948  // __sync_fetch_and_add_2.
949#define BUILTIN_ROW(x) \
950  { Builtin::BI##x##_1, Builtin::BI##x##_2, Builtin::BI##x##_4, \
951    Builtin::BI##x##_8, Builtin::BI##x##_16 }
952
953  static const unsigned BuiltinIndices[][5] = {
954    BUILTIN_ROW(__sync_fetch_and_add),
955    BUILTIN_ROW(__sync_fetch_and_sub),
956    BUILTIN_ROW(__sync_fetch_and_or),
957    BUILTIN_ROW(__sync_fetch_and_and),
958    BUILTIN_ROW(__sync_fetch_and_xor),
959
960    BUILTIN_ROW(__sync_add_and_fetch),
961    BUILTIN_ROW(__sync_sub_and_fetch),
962    BUILTIN_ROW(__sync_and_and_fetch),
963    BUILTIN_ROW(__sync_or_and_fetch),
964    BUILTIN_ROW(__sync_xor_and_fetch),
965
966    BUILTIN_ROW(__sync_val_compare_and_swap),
967    BUILTIN_ROW(__sync_bool_compare_and_swap),
968    BUILTIN_ROW(__sync_lock_test_and_set),
969    BUILTIN_ROW(__sync_lock_release),
970    BUILTIN_ROW(__sync_swap)
971  };
972#undef BUILTIN_ROW
973
974  // Determine the index of the size.
975  unsigned SizeIndex;
976  switch (Context.getTypeSizeInChars(ValType).getQuantity()) {
977  case 1: SizeIndex = 0; break;
978  case 2: SizeIndex = 1; break;
979  case 4: SizeIndex = 2; break;
980  case 8: SizeIndex = 3; break;
981  case 16: SizeIndex = 4; break;
982  default:
983    Diag(DRE->getLocStart(), diag::err_atomic_builtin_pointer_size)
984      << FirstArg->getType() << FirstArg->getSourceRange();
985    return ExprError();
986  }
987
988  // Each of these builtins has one pointer argument, followed by some number of
989  // values (0, 1 or 2) followed by a potentially empty varags list of stuff
990  // that we ignore.  Find out which row of BuiltinIndices to read from as well
991  // as the number of fixed args.
992  unsigned BuiltinID = FDecl->getBuiltinID();
993  unsigned BuiltinIndex, NumFixed = 1;
994  switch (BuiltinID) {
995  default: llvm_unreachable("Unknown overloaded atomic builtin!");
996  case Builtin::BI__sync_fetch_and_add:
997  case Builtin::BI__sync_fetch_and_add_1:
998  case Builtin::BI__sync_fetch_and_add_2:
999  case Builtin::BI__sync_fetch_and_add_4:
1000  case Builtin::BI__sync_fetch_and_add_8:
1001  case Builtin::BI__sync_fetch_and_add_16:
1002    BuiltinIndex = 0;
1003    break;
1004
1005  case Builtin::BI__sync_fetch_and_sub:
1006  case Builtin::BI__sync_fetch_and_sub_1:
1007  case Builtin::BI__sync_fetch_and_sub_2:
1008  case Builtin::BI__sync_fetch_and_sub_4:
1009  case Builtin::BI__sync_fetch_and_sub_8:
1010  case Builtin::BI__sync_fetch_and_sub_16:
1011    BuiltinIndex = 1;
1012    break;
1013
1014  case Builtin::BI__sync_fetch_and_or:
1015  case Builtin::BI__sync_fetch_and_or_1:
1016  case Builtin::BI__sync_fetch_and_or_2:
1017  case Builtin::BI__sync_fetch_and_or_4:
1018  case Builtin::BI__sync_fetch_and_or_8:
1019  case Builtin::BI__sync_fetch_and_or_16:
1020    BuiltinIndex = 2;
1021    break;
1022
1023  case Builtin::BI__sync_fetch_and_and:
1024  case Builtin::BI__sync_fetch_and_and_1:
1025  case Builtin::BI__sync_fetch_and_and_2:
1026  case Builtin::BI__sync_fetch_and_and_4:
1027  case Builtin::BI__sync_fetch_and_and_8:
1028  case Builtin::BI__sync_fetch_and_and_16:
1029    BuiltinIndex = 3;
1030    break;
1031
1032  case Builtin::BI__sync_fetch_and_xor:
1033  case Builtin::BI__sync_fetch_and_xor_1:
1034  case Builtin::BI__sync_fetch_and_xor_2:
1035  case Builtin::BI__sync_fetch_and_xor_4:
1036  case Builtin::BI__sync_fetch_and_xor_8:
1037  case Builtin::BI__sync_fetch_and_xor_16:
1038    BuiltinIndex = 4;
1039    break;
1040
1041  case Builtin::BI__sync_add_and_fetch:
1042  case Builtin::BI__sync_add_and_fetch_1:
1043  case Builtin::BI__sync_add_and_fetch_2:
1044  case Builtin::BI__sync_add_and_fetch_4:
1045  case Builtin::BI__sync_add_and_fetch_8:
1046  case Builtin::BI__sync_add_and_fetch_16:
1047    BuiltinIndex = 5;
1048    break;
1049
1050  case Builtin::BI__sync_sub_and_fetch:
1051  case Builtin::BI__sync_sub_and_fetch_1:
1052  case Builtin::BI__sync_sub_and_fetch_2:
1053  case Builtin::BI__sync_sub_and_fetch_4:
1054  case Builtin::BI__sync_sub_and_fetch_8:
1055  case Builtin::BI__sync_sub_and_fetch_16:
1056    BuiltinIndex = 6;
1057    break;
1058
1059  case Builtin::BI__sync_and_and_fetch:
1060  case Builtin::BI__sync_and_and_fetch_1:
1061  case Builtin::BI__sync_and_and_fetch_2:
1062  case Builtin::BI__sync_and_and_fetch_4:
1063  case Builtin::BI__sync_and_and_fetch_8:
1064  case Builtin::BI__sync_and_and_fetch_16:
1065    BuiltinIndex = 7;
1066    break;
1067
1068  case Builtin::BI__sync_or_and_fetch:
1069  case Builtin::BI__sync_or_and_fetch_1:
1070  case Builtin::BI__sync_or_and_fetch_2:
1071  case Builtin::BI__sync_or_and_fetch_4:
1072  case Builtin::BI__sync_or_and_fetch_8:
1073  case Builtin::BI__sync_or_and_fetch_16:
1074    BuiltinIndex = 8;
1075    break;
1076
1077  case Builtin::BI__sync_xor_and_fetch:
1078  case Builtin::BI__sync_xor_and_fetch_1:
1079  case Builtin::BI__sync_xor_and_fetch_2:
1080  case Builtin::BI__sync_xor_and_fetch_4:
1081  case Builtin::BI__sync_xor_and_fetch_8:
1082  case Builtin::BI__sync_xor_and_fetch_16:
1083    BuiltinIndex = 9;
1084    break;
1085
1086  case Builtin::BI__sync_val_compare_and_swap:
1087  case Builtin::BI__sync_val_compare_and_swap_1:
1088  case Builtin::BI__sync_val_compare_and_swap_2:
1089  case Builtin::BI__sync_val_compare_and_swap_4:
1090  case Builtin::BI__sync_val_compare_and_swap_8:
1091  case Builtin::BI__sync_val_compare_and_swap_16:
1092    BuiltinIndex = 10;
1093    NumFixed = 2;
1094    break;
1095
1096  case Builtin::BI__sync_bool_compare_and_swap:
1097  case Builtin::BI__sync_bool_compare_and_swap_1:
1098  case Builtin::BI__sync_bool_compare_and_swap_2:
1099  case Builtin::BI__sync_bool_compare_and_swap_4:
1100  case Builtin::BI__sync_bool_compare_and_swap_8:
1101  case Builtin::BI__sync_bool_compare_and_swap_16:
1102    BuiltinIndex = 11;
1103    NumFixed = 2;
1104    ResultType = Context.BoolTy;
1105    break;
1106
1107  case Builtin::BI__sync_lock_test_and_set:
1108  case Builtin::BI__sync_lock_test_and_set_1:
1109  case Builtin::BI__sync_lock_test_and_set_2:
1110  case Builtin::BI__sync_lock_test_and_set_4:
1111  case Builtin::BI__sync_lock_test_and_set_8:
1112  case Builtin::BI__sync_lock_test_and_set_16:
1113    BuiltinIndex = 12;
1114    break;
1115
1116  case Builtin::BI__sync_lock_release:
1117  case Builtin::BI__sync_lock_release_1:
1118  case Builtin::BI__sync_lock_release_2:
1119  case Builtin::BI__sync_lock_release_4:
1120  case Builtin::BI__sync_lock_release_8:
1121  case Builtin::BI__sync_lock_release_16:
1122    BuiltinIndex = 13;
1123    NumFixed = 0;
1124    ResultType = Context.VoidTy;
1125    break;
1126
1127  case Builtin::BI__sync_swap:
1128  case Builtin::BI__sync_swap_1:
1129  case Builtin::BI__sync_swap_2:
1130  case Builtin::BI__sync_swap_4:
1131  case Builtin::BI__sync_swap_8:
1132  case Builtin::BI__sync_swap_16:
1133    BuiltinIndex = 14;
1134    break;
1135  }
1136
1137  // Now that we know how many fixed arguments we expect, first check that we
1138  // have at least that many.
1139  if (TheCall->getNumArgs() < 1+NumFixed) {
1140    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args_at_least)
1141      << 0 << 1+NumFixed << TheCall->getNumArgs()
1142      << TheCall->getCallee()->getSourceRange();
1143    return ExprError();
1144  }
1145
1146  // Get the decl for the concrete builtin from this, we can tell what the
1147  // concrete integer type we should convert to is.
1148  unsigned NewBuiltinID = BuiltinIndices[BuiltinIndex][SizeIndex];
1149  const char *NewBuiltinName = Context.BuiltinInfo.GetName(NewBuiltinID);
1150  IdentifierInfo *NewBuiltinII = PP.getIdentifierInfo(NewBuiltinName);
1151  FunctionDecl *NewBuiltinDecl =
1152    cast<FunctionDecl>(LazilyCreateBuiltin(NewBuiltinII, NewBuiltinID,
1153                                           TUScope, false, DRE->getLocStart()));
1154
1155  // The first argument --- the pointer --- has a fixed type; we
1156  // deduce the types of the rest of the arguments accordingly.  Walk
1157  // the remaining arguments, converting them to the deduced value type.
1158  for (unsigned i = 0; i != NumFixed; ++i) {
1159    ExprResult Arg = TheCall->getArg(i+1);
1160
1161    // GCC does an implicit conversion to the pointer or integer ValType.  This
1162    // can fail in some cases (1i -> int**), check for this error case now.
1163    // Initialize the argument.
1164    InitializedEntity Entity = InitializedEntity::InitializeParameter(Context,
1165                                                   ValType, /*consume*/ false);
1166    Arg = PerformCopyInitialization(Entity, SourceLocation(), Arg);
1167    if (Arg.isInvalid())
1168      return ExprError();
1169
1170    // Okay, we have something that *can* be converted to the right type.  Check
1171    // to see if there is a potentially weird extension going on here.  This can
1172    // happen when you do an atomic operation on something like an char* and
1173    // pass in 42.  The 42 gets converted to char.  This is even more strange
1174    // for things like 45.123 -> char, etc.
1175    // FIXME: Do this check.
1176    TheCall->setArg(i+1, Arg.take());
1177  }
1178
1179  ASTContext& Context = this->getASTContext();
1180
1181  // Create a new DeclRefExpr to refer to the new decl.
1182  DeclRefExpr* NewDRE = DeclRefExpr::Create(
1183      Context,
1184      DRE->getQualifierLoc(),
1185      SourceLocation(),
1186      NewBuiltinDecl,
1187      /*enclosing*/ false,
1188      DRE->getLocation(),
1189      NewBuiltinDecl->getType(),
1190      DRE->getValueKind());
1191
1192  // Set the callee in the CallExpr.
1193  // FIXME: This leaks the original parens and implicit casts.
1194  ExprResult PromotedCall = UsualUnaryConversions(NewDRE);
1195  if (PromotedCall.isInvalid())
1196    return ExprError();
1197  TheCall->setCallee(PromotedCall.take());
1198
1199  // Change the result type of the call to match the original value type. This
1200  // is arbitrary, but the codegen for these builtins ins design to handle it
1201  // gracefully.
1202  TheCall->setType(ResultType);
1203
1204  return move(TheCallResult);
1205}
1206
1207/// CheckObjCString - Checks that the argument to the builtin
1208/// CFString constructor is correct
1209/// Note: It might also make sense to do the UTF-16 conversion here (would
1210/// simplify the backend).
1211bool Sema::CheckObjCString(Expr *Arg) {
1212  Arg = Arg->IgnoreParenCasts();
1213  StringLiteral *Literal = dyn_cast<StringLiteral>(Arg);
1214
1215  if (!Literal || !Literal->isAscii()) {
1216    Diag(Arg->getLocStart(), diag::err_cfstring_literal_not_string_constant)
1217      << Arg->getSourceRange();
1218    return true;
1219  }
1220
1221  if (Literal->containsNonAsciiOrNull()) {
1222    StringRef String = Literal->getString();
1223    unsigned NumBytes = String.size();
1224    SmallVector<UTF16, 128> ToBuf(NumBytes);
1225    const UTF8 *FromPtr = (UTF8 *)String.data();
1226    UTF16 *ToPtr = &ToBuf[0];
1227
1228    ConversionResult Result = ConvertUTF8toUTF16(&FromPtr, FromPtr + NumBytes,
1229                                                 &ToPtr, ToPtr + NumBytes,
1230                                                 strictConversion);
1231    // Check for conversion failure.
1232    if (Result != conversionOK)
1233      Diag(Arg->getLocStart(),
1234           diag::warn_cfstring_truncated) << Arg->getSourceRange();
1235  }
1236  return false;
1237}
1238
1239/// SemaBuiltinVAStart - Check the arguments to __builtin_va_start for validity.
1240/// Emit an error and return true on failure, return false on success.
1241bool Sema::SemaBuiltinVAStart(CallExpr *TheCall) {
1242  Expr *Fn = TheCall->getCallee();
1243  if (TheCall->getNumArgs() > 2) {
1244    Diag(TheCall->getArg(2)->getLocStart(),
1245         diag::err_typecheck_call_too_many_args)
1246      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1247      << Fn->getSourceRange()
1248      << SourceRange(TheCall->getArg(2)->getLocStart(),
1249                     (*(TheCall->arg_end()-1))->getLocEnd());
1250    return true;
1251  }
1252
1253  if (TheCall->getNumArgs() < 2) {
1254    return Diag(TheCall->getLocEnd(),
1255      diag::err_typecheck_call_too_few_args_at_least)
1256      << 0 /*function call*/ << 2 << TheCall->getNumArgs();
1257  }
1258
1259  // Type-check the first argument normally.
1260  if (checkBuiltinArgument(*this, TheCall, 0))
1261    return true;
1262
1263  // Determine whether the current function is variadic or not.
1264  BlockScopeInfo *CurBlock = getCurBlock();
1265  bool isVariadic;
1266  if (CurBlock)
1267    isVariadic = CurBlock->TheDecl->isVariadic();
1268  else if (FunctionDecl *FD = getCurFunctionDecl())
1269    isVariadic = FD->isVariadic();
1270  else
1271    isVariadic = getCurMethodDecl()->isVariadic();
1272
1273  if (!isVariadic) {
1274    Diag(Fn->getLocStart(), diag::err_va_start_used_in_non_variadic_function);
1275    return true;
1276  }
1277
1278  // Verify that the second argument to the builtin is the last argument of the
1279  // current function or method.
1280  bool SecondArgIsLastNamedArgument = false;
1281  const Expr *Arg = TheCall->getArg(1)->IgnoreParenCasts();
1282
1283  if (const DeclRefExpr *DR = dyn_cast<DeclRefExpr>(Arg)) {
1284    if (const ParmVarDecl *PV = dyn_cast<ParmVarDecl>(DR->getDecl())) {
1285      // FIXME: This isn't correct for methods (results in bogus warning).
1286      // Get the last formal in the current function.
1287      const ParmVarDecl *LastArg;
1288      if (CurBlock)
1289        LastArg = *(CurBlock->TheDecl->param_end()-1);
1290      else if (FunctionDecl *FD = getCurFunctionDecl())
1291        LastArg = *(FD->param_end()-1);
1292      else
1293        LastArg = *(getCurMethodDecl()->param_end()-1);
1294      SecondArgIsLastNamedArgument = PV == LastArg;
1295    }
1296  }
1297
1298  if (!SecondArgIsLastNamedArgument)
1299    Diag(TheCall->getArg(1)->getLocStart(),
1300         diag::warn_second_parameter_of_va_start_not_last_named_argument);
1301  return false;
1302}
1303
1304/// SemaBuiltinUnorderedCompare - Handle functions like __builtin_isgreater and
1305/// friends.  This is declared to take (...), so we have to check everything.
1306bool Sema::SemaBuiltinUnorderedCompare(CallExpr *TheCall) {
1307  if (TheCall->getNumArgs() < 2)
1308    return Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
1309      << 0 << 2 << TheCall->getNumArgs()/*function call*/;
1310  if (TheCall->getNumArgs() > 2)
1311    return Diag(TheCall->getArg(2)->getLocStart(),
1312                diag::err_typecheck_call_too_many_args)
1313      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1314      << SourceRange(TheCall->getArg(2)->getLocStart(),
1315                     (*(TheCall->arg_end()-1))->getLocEnd());
1316
1317  ExprResult OrigArg0 = TheCall->getArg(0);
1318  ExprResult OrigArg1 = TheCall->getArg(1);
1319
1320  // Do standard promotions between the two arguments, returning their common
1321  // type.
1322  QualType Res = UsualArithmeticConversions(OrigArg0, OrigArg1, false);
1323  if (OrigArg0.isInvalid() || OrigArg1.isInvalid())
1324    return true;
1325
1326  // Make sure any conversions are pushed back into the call; this is
1327  // type safe since unordered compare builtins are declared as "_Bool
1328  // foo(...)".
1329  TheCall->setArg(0, OrigArg0.get());
1330  TheCall->setArg(1, OrigArg1.get());
1331
1332  if (OrigArg0.get()->isTypeDependent() || OrigArg1.get()->isTypeDependent())
1333    return false;
1334
1335  // If the common type isn't a real floating type, then the arguments were
1336  // invalid for this operation.
1337  if (Res.isNull() || !Res->isRealFloatingType())
1338    return Diag(OrigArg0.get()->getLocStart(),
1339                diag::err_typecheck_call_invalid_ordered_compare)
1340      << OrigArg0.get()->getType() << OrigArg1.get()->getType()
1341      << SourceRange(OrigArg0.get()->getLocStart(), OrigArg1.get()->getLocEnd());
1342
1343  return false;
1344}
1345
1346/// SemaBuiltinSemaBuiltinFPClassification - Handle functions like
1347/// __builtin_isnan and friends.  This is declared to take (...), so we have
1348/// to check everything. We expect the last argument to be a floating point
1349/// value.
1350bool Sema::SemaBuiltinFPClassification(CallExpr *TheCall, unsigned NumArgs) {
1351  if (TheCall->getNumArgs() < NumArgs)
1352    return Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
1353      << 0 << NumArgs << TheCall->getNumArgs()/*function call*/;
1354  if (TheCall->getNumArgs() > NumArgs)
1355    return Diag(TheCall->getArg(NumArgs)->getLocStart(),
1356                diag::err_typecheck_call_too_many_args)
1357      << 0 /*function call*/ << NumArgs << TheCall->getNumArgs()
1358      << SourceRange(TheCall->getArg(NumArgs)->getLocStart(),
1359                     (*(TheCall->arg_end()-1))->getLocEnd());
1360
1361  Expr *OrigArg = TheCall->getArg(NumArgs-1);
1362
1363  if (OrigArg->isTypeDependent())
1364    return false;
1365
1366  // This operation requires a non-_Complex floating-point number.
1367  if (!OrigArg->getType()->isRealFloatingType())
1368    return Diag(OrigArg->getLocStart(),
1369                diag::err_typecheck_call_invalid_unary_fp)
1370      << OrigArg->getType() << OrigArg->getSourceRange();
1371
1372  // If this is an implicit conversion from float -> double, remove it.
1373  if (ImplicitCastExpr *Cast = dyn_cast<ImplicitCastExpr>(OrigArg)) {
1374    Expr *CastArg = Cast->getSubExpr();
1375    if (CastArg->getType()->isSpecificBuiltinType(BuiltinType::Float)) {
1376      assert(Cast->getType()->isSpecificBuiltinType(BuiltinType::Double) &&
1377             "promotion from float to double is the only expected cast here");
1378      Cast->setSubExpr(0);
1379      TheCall->setArg(NumArgs-1, CastArg);
1380    }
1381  }
1382
1383  return false;
1384}
1385
1386/// SemaBuiltinShuffleVector - Handle __builtin_shufflevector.
1387// This is declared to take (...), so we have to check everything.
1388ExprResult Sema::SemaBuiltinShuffleVector(CallExpr *TheCall) {
1389  if (TheCall->getNumArgs() < 2)
1390    return ExprError(Diag(TheCall->getLocEnd(),
1391                          diag::err_typecheck_call_too_few_args_at_least)
1392      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1393      << TheCall->getSourceRange());
1394
1395  // Determine which of the following types of shufflevector we're checking:
1396  // 1) unary, vector mask: (lhs, mask)
1397  // 2) binary, vector mask: (lhs, rhs, mask)
1398  // 3) binary, scalar mask: (lhs, rhs, index, ..., index)
1399  QualType resType = TheCall->getArg(0)->getType();
1400  unsigned numElements = 0;
1401
1402  if (!TheCall->getArg(0)->isTypeDependent() &&
1403      !TheCall->getArg(1)->isTypeDependent()) {
1404    QualType LHSType = TheCall->getArg(0)->getType();
1405    QualType RHSType = TheCall->getArg(1)->getType();
1406
1407    if (!LHSType->isVectorType() || !RHSType->isVectorType()) {
1408      Diag(TheCall->getLocStart(), diag::err_shufflevector_non_vector)
1409        << SourceRange(TheCall->getArg(0)->getLocStart(),
1410                       TheCall->getArg(1)->getLocEnd());
1411      return ExprError();
1412    }
1413
1414    numElements = LHSType->getAs<VectorType>()->getNumElements();
1415    unsigned numResElements = TheCall->getNumArgs() - 2;
1416
1417    // Check to see if we have a call with 2 vector arguments, the unary shuffle
1418    // with mask.  If so, verify that RHS is an integer vector type with the
1419    // same number of elts as lhs.
1420    if (TheCall->getNumArgs() == 2) {
1421      if (!RHSType->hasIntegerRepresentation() ||
1422          RHSType->getAs<VectorType>()->getNumElements() != numElements)
1423        Diag(TheCall->getLocStart(), diag::err_shufflevector_incompatible_vector)
1424          << SourceRange(TheCall->getArg(1)->getLocStart(),
1425                         TheCall->getArg(1)->getLocEnd());
1426      numResElements = numElements;
1427    }
1428    else if (!Context.hasSameUnqualifiedType(LHSType, RHSType)) {
1429      Diag(TheCall->getLocStart(), diag::err_shufflevector_incompatible_vector)
1430        << SourceRange(TheCall->getArg(0)->getLocStart(),
1431                       TheCall->getArg(1)->getLocEnd());
1432      return ExprError();
1433    } else if (numElements != numResElements) {
1434      QualType eltType = LHSType->getAs<VectorType>()->getElementType();
1435      resType = Context.getVectorType(eltType, numResElements,
1436                                      VectorType::GenericVector);
1437    }
1438  }
1439
1440  for (unsigned i = 2; i < TheCall->getNumArgs(); i++) {
1441    if (TheCall->getArg(i)->isTypeDependent() ||
1442        TheCall->getArg(i)->isValueDependent())
1443      continue;
1444
1445    llvm::APSInt Result(32);
1446    if (!TheCall->getArg(i)->isIntegerConstantExpr(Result, Context))
1447      return ExprError(Diag(TheCall->getLocStart(),
1448                  diag::err_shufflevector_nonconstant_argument)
1449                << TheCall->getArg(i)->getSourceRange());
1450
1451    if (Result.getActiveBits() > 64 || Result.getZExtValue() >= numElements*2)
1452      return ExprError(Diag(TheCall->getLocStart(),
1453                  diag::err_shufflevector_argument_too_large)
1454               << TheCall->getArg(i)->getSourceRange());
1455  }
1456
1457  SmallVector<Expr*, 32> exprs;
1458
1459  for (unsigned i = 0, e = TheCall->getNumArgs(); i != e; i++) {
1460    exprs.push_back(TheCall->getArg(i));
1461    TheCall->setArg(i, 0);
1462  }
1463
1464  return Owned(new (Context) ShuffleVectorExpr(Context, exprs.begin(),
1465                                            exprs.size(), resType,
1466                                            TheCall->getCallee()->getLocStart(),
1467                                            TheCall->getRParenLoc()));
1468}
1469
1470/// SemaBuiltinPrefetch - Handle __builtin_prefetch.
1471// This is declared to take (const void*, ...) and can take two
1472// optional constant int args.
1473bool Sema::SemaBuiltinPrefetch(CallExpr *TheCall) {
1474  unsigned NumArgs = TheCall->getNumArgs();
1475
1476  if (NumArgs > 3)
1477    return Diag(TheCall->getLocEnd(),
1478             diag::err_typecheck_call_too_many_args_at_most)
1479             << 0 /*function call*/ << 3 << NumArgs
1480             << TheCall->getSourceRange();
1481
1482  // Argument 0 is checked for us and the remaining arguments must be
1483  // constant integers.
1484  for (unsigned i = 1; i != NumArgs; ++i) {
1485    Expr *Arg = TheCall->getArg(i);
1486
1487    // We can't check the value of a dependent argument.
1488    if (Arg->isTypeDependent() || Arg->isValueDependent())
1489      continue;
1490
1491    llvm::APSInt Result;
1492    if (SemaBuiltinConstantArg(TheCall, i, Result))
1493      return true;
1494
1495    // FIXME: gcc issues a warning and rewrites these to 0. These
1496    // seems especially odd for the third argument since the default
1497    // is 3.
1498    if (i == 1) {
1499      if (Result.getLimitedValue() > 1)
1500        return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1501             << "0" << "1" << Arg->getSourceRange();
1502    } else {
1503      if (Result.getLimitedValue() > 3)
1504        return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1505            << "0" << "3" << Arg->getSourceRange();
1506    }
1507  }
1508
1509  return false;
1510}
1511
1512/// SemaBuiltinConstantArg - Handle a check if argument ArgNum of CallExpr
1513/// TheCall is a constant expression.
1514bool Sema::SemaBuiltinConstantArg(CallExpr *TheCall, int ArgNum,
1515                                  llvm::APSInt &Result) {
1516  Expr *Arg = TheCall->getArg(ArgNum);
1517  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
1518  FunctionDecl *FDecl = cast<FunctionDecl>(DRE->getDecl());
1519
1520  if (Arg->isTypeDependent() || Arg->isValueDependent()) return false;
1521
1522  if (!Arg->isIntegerConstantExpr(Result, Context))
1523    return Diag(TheCall->getLocStart(), diag::err_constant_integer_arg_type)
1524                << FDecl->getDeclName() <<  Arg->getSourceRange();
1525
1526  return false;
1527}
1528
1529/// SemaBuiltinObjectSize - Handle __builtin_object_size(void *ptr,
1530/// int type). This simply type checks that type is one of the defined
1531/// constants (0-3).
1532// For compatibility check 0-3, llvm only handles 0 and 2.
1533bool Sema::SemaBuiltinObjectSize(CallExpr *TheCall) {
1534  llvm::APSInt Result;
1535
1536  // We can't check the value of a dependent argument.
1537  if (TheCall->getArg(1)->isTypeDependent() ||
1538      TheCall->getArg(1)->isValueDependent())
1539    return false;
1540
1541  // Check constant-ness first.
1542  if (SemaBuiltinConstantArg(TheCall, 1, Result))
1543    return true;
1544
1545  Expr *Arg = TheCall->getArg(1);
1546  if (Result.getSExtValue() < 0 || Result.getSExtValue() > 3) {
1547    return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1548             << "0" << "3" << SourceRange(Arg->getLocStart(), Arg->getLocEnd());
1549  }
1550
1551  return false;
1552}
1553
1554/// SemaBuiltinLongjmp - Handle __builtin_longjmp(void *env[5], int val).
1555/// This checks that val is a constant 1.
1556bool Sema::SemaBuiltinLongjmp(CallExpr *TheCall) {
1557  Expr *Arg = TheCall->getArg(1);
1558  llvm::APSInt Result;
1559
1560  // TODO: This is less than ideal. Overload this to take a value.
1561  if (SemaBuiltinConstantArg(TheCall, 1, Result))
1562    return true;
1563
1564  if (Result != 1)
1565    return Diag(TheCall->getLocStart(), diag::err_builtin_longjmp_invalid_val)
1566             << SourceRange(Arg->getLocStart(), Arg->getLocEnd());
1567
1568  return false;
1569}
1570
1571// Determine if an expression is a string literal or constant string.
1572// If this function returns false on the arguments to a function expecting a
1573// format string, we will usually need to emit a warning.
1574// True string literals are then checked by CheckFormatString.
1575Sema::StringLiteralCheckType
1576Sema::checkFormatStringExpr(const Expr *E, Expr **Args,
1577                            unsigned NumArgs, bool HasVAListArg,
1578                            unsigned format_idx, unsigned firstDataArg,
1579                            FormatStringType Type, bool inFunctionCall) {
1580 tryAgain:
1581  if (E->isTypeDependent() || E->isValueDependent())
1582    return SLCT_NotALiteral;
1583
1584  E = E->IgnoreParenCasts();
1585
1586  if (E->isNullPointerConstant(Context, Expr::NPC_ValueDependentIsNotNull))
1587    // Technically -Wformat-nonliteral does not warn about this case.
1588    // The behavior of printf and friends in this case is implementation
1589    // dependent.  Ideally if the format string cannot be null then
1590    // it should have a 'nonnull' attribute in the function prototype.
1591    return SLCT_CheckedLiteral;
1592
1593  switch (E->getStmtClass()) {
1594  case Stmt::BinaryConditionalOperatorClass:
1595  case Stmt::ConditionalOperatorClass: {
1596    // The expression is a literal if both sub-expressions were, and it was
1597    // completely checked only if both sub-expressions were checked.
1598    const AbstractConditionalOperator *C =
1599        cast<AbstractConditionalOperator>(E);
1600    StringLiteralCheckType Left =
1601        checkFormatStringExpr(C->getTrueExpr(), Args, NumArgs,
1602                              HasVAListArg, format_idx, firstDataArg,
1603                              Type, inFunctionCall);
1604    if (Left == SLCT_NotALiteral)
1605      return SLCT_NotALiteral;
1606    StringLiteralCheckType Right =
1607        checkFormatStringExpr(C->getFalseExpr(), Args, NumArgs,
1608                              HasVAListArg, format_idx, firstDataArg,
1609                              Type, inFunctionCall);
1610    return Left < Right ? Left : Right;
1611  }
1612
1613  case Stmt::ImplicitCastExprClass: {
1614    E = cast<ImplicitCastExpr>(E)->getSubExpr();
1615    goto tryAgain;
1616  }
1617
1618  case Stmt::OpaqueValueExprClass:
1619    if (const Expr *src = cast<OpaqueValueExpr>(E)->getSourceExpr()) {
1620      E = src;
1621      goto tryAgain;
1622    }
1623    return SLCT_NotALiteral;
1624
1625  case Stmt::PredefinedExprClass:
1626    // While __func__, etc., are technically not string literals, they
1627    // cannot contain format specifiers and thus are not a security
1628    // liability.
1629    return SLCT_UncheckedLiteral;
1630
1631  case Stmt::DeclRefExprClass: {
1632    const DeclRefExpr *DR = cast<DeclRefExpr>(E);
1633
1634    // As an exception, do not flag errors for variables binding to
1635    // const string literals.
1636    if (const VarDecl *VD = dyn_cast<VarDecl>(DR->getDecl())) {
1637      bool isConstant = false;
1638      QualType T = DR->getType();
1639
1640      if (const ArrayType *AT = Context.getAsArrayType(T)) {
1641        isConstant = AT->getElementType().isConstant(Context);
1642      } else if (const PointerType *PT = T->getAs<PointerType>()) {
1643        isConstant = T.isConstant(Context) &&
1644                     PT->getPointeeType().isConstant(Context);
1645      } else if (T->isObjCObjectPointerType()) {
1646        // In ObjC, there is usually no "const ObjectPointer" type,
1647        // so don't check if the pointee type is constant.
1648        isConstant = T.isConstant(Context);
1649      }
1650
1651      if (isConstant) {
1652        if (const Expr *Init = VD->getAnyInitializer()) {
1653          // Look through initializers like const char c[] = { "foo" }
1654          if (const InitListExpr *InitList = dyn_cast<InitListExpr>(Init)) {
1655            if (InitList->isStringLiteralInit())
1656              Init = InitList->getInit(0)->IgnoreParenImpCasts();
1657          }
1658          return checkFormatStringExpr(Init, Args, NumArgs,
1659                                       HasVAListArg, format_idx,
1660                                       firstDataArg, Type,
1661                                       /*inFunctionCall*/false);
1662        }
1663      }
1664
1665      // For vprintf* functions (i.e., HasVAListArg==true), we add a
1666      // special check to see if the format string is a function parameter
1667      // of the function calling the printf function.  If the function
1668      // has an attribute indicating it is a printf-like function, then we
1669      // should suppress warnings concerning non-literals being used in a call
1670      // to a vprintf function.  For example:
1671      //
1672      // void
1673      // logmessage(char const *fmt __attribute__ (format (printf, 1, 2)), ...){
1674      //      va_list ap;
1675      //      va_start(ap, fmt);
1676      //      vprintf(fmt, ap);  // Do NOT emit a warning about "fmt".
1677      //      ...
1678      //
1679      if (HasVAListArg) {
1680        if (const ParmVarDecl *PV = dyn_cast<ParmVarDecl>(VD)) {
1681          if (const NamedDecl *ND = dyn_cast<NamedDecl>(PV->getDeclContext())) {
1682            int PVIndex = PV->getFunctionScopeIndex() + 1;
1683            for (specific_attr_iterator<FormatAttr>
1684                 i = ND->specific_attr_begin<FormatAttr>(),
1685                 e = ND->specific_attr_end<FormatAttr>(); i != e ; ++i) {
1686              FormatAttr *PVFormat = *i;
1687              // adjust for implicit parameter
1688              if (const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(ND))
1689                if (MD->isInstance())
1690                  ++PVIndex;
1691              // We also check if the formats are compatible.
1692              // We can't pass a 'scanf' string to a 'printf' function.
1693              if (PVIndex == PVFormat->getFormatIdx() &&
1694                  Type == GetFormatStringType(PVFormat))
1695                return SLCT_UncheckedLiteral;
1696            }
1697          }
1698        }
1699      }
1700    }
1701
1702    return SLCT_NotALiteral;
1703  }
1704
1705  case Stmt::CallExprClass:
1706  case Stmt::CXXMemberCallExprClass: {
1707    const CallExpr *CE = cast<CallExpr>(E);
1708    if (const NamedDecl *ND = dyn_cast_or_null<NamedDecl>(CE->getCalleeDecl())) {
1709      if (const FormatArgAttr *FA = ND->getAttr<FormatArgAttr>()) {
1710        unsigned ArgIndex = FA->getFormatIdx();
1711        if (const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(ND))
1712          if (MD->isInstance())
1713            --ArgIndex;
1714        const Expr *Arg = CE->getArg(ArgIndex - 1);
1715
1716        return checkFormatStringExpr(Arg, Args, NumArgs,
1717                                     HasVAListArg, format_idx, firstDataArg,
1718                                     Type, inFunctionCall);
1719      } else if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(ND)) {
1720        unsigned BuiltinID = FD->getBuiltinID();
1721        if (BuiltinID == Builtin::BI__builtin___CFStringMakeConstantString ||
1722            BuiltinID == Builtin::BI__builtin___NSStringMakeConstantString) {
1723          const Expr *Arg = CE->getArg(0);
1724          return checkFormatStringExpr(Arg, Args, NumArgs,
1725                                       HasVAListArg, format_idx,
1726                                       firstDataArg, Type, inFunctionCall);
1727        }
1728      }
1729    }
1730
1731    return SLCT_NotALiteral;
1732  }
1733  case Stmt::ObjCStringLiteralClass:
1734  case Stmt::StringLiteralClass: {
1735    const StringLiteral *StrE = NULL;
1736
1737    if (const ObjCStringLiteral *ObjCFExpr = dyn_cast<ObjCStringLiteral>(E))
1738      StrE = ObjCFExpr->getString();
1739    else
1740      StrE = cast<StringLiteral>(E);
1741
1742    if (StrE) {
1743      CheckFormatString(StrE, E, Args, NumArgs, HasVAListArg, format_idx,
1744                        firstDataArg, Type, inFunctionCall);
1745      return SLCT_CheckedLiteral;
1746    }
1747
1748    return SLCT_NotALiteral;
1749  }
1750
1751  default:
1752    return SLCT_NotALiteral;
1753  }
1754}
1755
1756void
1757Sema::CheckNonNullArguments(const NonNullAttr *NonNull,
1758                            const Expr * const *ExprArgs,
1759                            SourceLocation CallSiteLoc) {
1760  for (NonNullAttr::args_iterator i = NonNull->args_begin(),
1761                                  e = NonNull->args_end();
1762       i != e; ++i) {
1763    const Expr *ArgExpr = ExprArgs[*i];
1764    if (ArgExpr->isNullPointerConstant(Context,
1765                                       Expr::NPC_ValueDependentIsNotNull))
1766      Diag(CallSiteLoc, diag::warn_null_arg) << ArgExpr->getSourceRange();
1767  }
1768}
1769
1770Sema::FormatStringType Sema::GetFormatStringType(const FormatAttr *Format) {
1771  return llvm::StringSwitch<FormatStringType>(Format->getType())
1772  .Case("scanf", FST_Scanf)
1773  .Cases("printf", "printf0", FST_Printf)
1774  .Cases("NSString", "CFString", FST_NSString)
1775  .Case("strftime", FST_Strftime)
1776  .Case("strfmon", FST_Strfmon)
1777  .Cases("kprintf", "cmn_err", "vcmn_err", "zcmn_err", FST_Kprintf)
1778  .Default(FST_Unknown);
1779}
1780
1781/// CheckPrintfScanfArguments - Check calls to printf and scanf (and similar
1782/// functions) for correct use of format strings.
1783/// Returns true if a format string has been fully checked.
1784bool Sema::CheckFormatArguments(const FormatAttr *Format, CallExpr *TheCall) {
1785  bool IsCXXMember = isa<CXXMemberCallExpr>(TheCall);
1786  return CheckFormatArguments(Format, TheCall->getArgs(),
1787                              TheCall->getNumArgs(),
1788                              IsCXXMember, TheCall->getRParenLoc(),
1789                              TheCall->getCallee()->getSourceRange());
1790}
1791
1792bool Sema::CheckFormatArguments(const FormatAttr *Format, Expr **Args,
1793                                unsigned NumArgs, bool IsCXXMember,
1794                                SourceLocation Loc, SourceRange Range) {
1795  FormatStringInfo FSI;
1796  if (getFormatStringInfo(Format, IsCXXMember, &FSI))
1797    return CheckFormatArguments(Args, NumArgs, FSI.HasVAListArg, FSI.FormatIdx,
1798                                FSI.FirstDataArg, GetFormatStringType(Format),
1799                                Loc, Range);
1800  return false;
1801}
1802
1803bool Sema::CheckFormatArguments(Expr **Args, unsigned NumArgs,
1804                                bool HasVAListArg, unsigned format_idx,
1805                                unsigned firstDataArg, FormatStringType Type,
1806                                SourceLocation Loc, SourceRange Range) {
1807  // CHECK: printf/scanf-like function is called with no format string.
1808  if (format_idx >= NumArgs) {
1809    Diag(Loc, diag::warn_missing_format_string) << Range;
1810    return false;
1811  }
1812
1813  const Expr *OrigFormatExpr = Args[format_idx]->IgnoreParenCasts();
1814
1815  // CHECK: format string is not a string literal.
1816  //
1817  // Dynamically generated format strings are difficult to
1818  // automatically vet at compile time.  Requiring that format strings
1819  // are string literals: (1) permits the checking of format strings by
1820  // the compiler and thereby (2) can practically remove the source of
1821  // many format string exploits.
1822
1823  // Format string can be either ObjC string (e.g. @"%d") or
1824  // C string (e.g. "%d")
1825  // ObjC string uses the same format specifiers as C string, so we can use
1826  // the same format string checking logic for both ObjC and C strings.
1827  StringLiteralCheckType CT =
1828      checkFormatStringExpr(OrigFormatExpr, Args, NumArgs, HasVAListArg,
1829                            format_idx, firstDataArg, Type);
1830  if (CT != SLCT_NotALiteral)
1831    // Literal format string found, check done!
1832    return CT == SLCT_CheckedLiteral;
1833
1834  // Strftime is particular as it always uses a single 'time' argument,
1835  // so it is safe to pass a non-literal string.
1836  if (Type == FST_Strftime)
1837    return false;
1838
1839  // Do not emit diag when the string param is a macro expansion and the
1840  // format is either NSString or CFString. This is a hack to prevent
1841  // diag when using the NSLocalizedString and CFCopyLocalizedString macros
1842  // which are usually used in place of NS and CF string literals.
1843  if (Type == FST_NSString &&
1844      SourceMgr.isInSystemMacro(Args[format_idx]->getLocStart()))
1845    return false;
1846
1847  // If there are no arguments specified, warn with -Wformat-security, otherwise
1848  // warn only with -Wformat-nonliteral.
1849  if (NumArgs == format_idx+1)
1850    Diag(Args[format_idx]->getLocStart(),
1851         diag::warn_format_nonliteral_noargs)
1852      << OrigFormatExpr->getSourceRange();
1853  else
1854    Diag(Args[format_idx]->getLocStart(),
1855         diag::warn_format_nonliteral)
1856           << OrigFormatExpr->getSourceRange();
1857  return false;
1858}
1859
1860namespace {
1861class CheckFormatHandler : public analyze_format_string::FormatStringHandler {
1862protected:
1863  Sema &S;
1864  const StringLiteral *FExpr;
1865  const Expr *OrigFormatExpr;
1866  const unsigned FirstDataArg;
1867  const unsigned NumDataArgs;
1868  const char *Beg; // Start of format string.
1869  const bool HasVAListArg;
1870  const Expr * const *Args;
1871  const unsigned NumArgs;
1872  unsigned FormatIdx;
1873  llvm::BitVector CoveredArgs;
1874  bool usesPositionalArgs;
1875  bool atFirstArg;
1876  bool inFunctionCall;
1877public:
1878  CheckFormatHandler(Sema &s, const StringLiteral *fexpr,
1879                     const Expr *origFormatExpr, unsigned firstDataArg,
1880                     unsigned numDataArgs, const char *beg, bool hasVAListArg,
1881                     Expr **args, unsigned numArgs,
1882                     unsigned formatIdx, bool inFunctionCall)
1883    : S(s), FExpr(fexpr), OrigFormatExpr(origFormatExpr),
1884      FirstDataArg(firstDataArg), NumDataArgs(numDataArgs),
1885      Beg(beg), HasVAListArg(hasVAListArg),
1886      Args(args), NumArgs(numArgs), FormatIdx(formatIdx),
1887      usesPositionalArgs(false), atFirstArg(true),
1888      inFunctionCall(inFunctionCall) {
1889        CoveredArgs.resize(numDataArgs);
1890        CoveredArgs.reset();
1891      }
1892
1893  void DoneProcessing();
1894
1895  void HandleIncompleteSpecifier(const char *startSpecifier,
1896                                 unsigned specifierLen);
1897
1898  void HandleNonStandardLengthModifier(
1899      const analyze_format_string::LengthModifier &LM,
1900      const char *startSpecifier, unsigned specifierLen);
1901
1902  void HandleNonStandardConversionSpecifier(
1903      const analyze_format_string::ConversionSpecifier &CS,
1904      const char *startSpecifier, unsigned specifierLen);
1905
1906  void HandleNonStandardConversionSpecification(
1907      const analyze_format_string::LengthModifier &LM,
1908      const analyze_format_string::ConversionSpecifier &CS,
1909      const char *startSpecifier, unsigned specifierLen);
1910
1911  virtual void HandlePosition(const char *startPos, unsigned posLen);
1912
1913  virtual void HandleInvalidPosition(const char *startSpecifier,
1914                                     unsigned specifierLen,
1915                                     analyze_format_string::PositionContext p);
1916
1917  virtual void HandleZeroPosition(const char *startPos, unsigned posLen);
1918
1919  void HandleNullChar(const char *nullCharacter);
1920
1921  template <typename Range>
1922  static void EmitFormatDiagnostic(Sema &S, bool inFunctionCall,
1923                                   const Expr *ArgumentExpr,
1924                                   PartialDiagnostic PDiag,
1925                                   SourceLocation StringLoc,
1926                                   bool IsStringLocation, Range StringRange,
1927                                   FixItHint Fixit = FixItHint());
1928
1929protected:
1930  bool HandleInvalidConversionSpecifier(unsigned argIndex, SourceLocation Loc,
1931                                        const char *startSpec,
1932                                        unsigned specifierLen,
1933                                        const char *csStart, unsigned csLen);
1934
1935  void HandlePositionalNonpositionalArgs(SourceLocation Loc,
1936                                         const char *startSpec,
1937                                         unsigned specifierLen);
1938
1939  SourceRange getFormatStringRange();
1940  CharSourceRange getSpecifierRange(const char *startSpecifier,
1941                                    unsigned specifierLen);
1942  SourceLocation getLocationOfByte(const char *x);
1943
1944  const Expr *getDataArg(unsigned i) const;
1945
1946  bool CheckNumArgs(const analyze_format_string::FormatSpecifier &FS,
1947                    const analyze_format_string::ConversionSpecifier &CS,
1948                    const char *startSpecifier, unsigned specifierLen,
1949                    unsigned argIndex);
1950
1951  template <typename Range>
1952  void EmitFormatDiagnostic(PartialDiagnostic PDiag, SourceLocation StringLoc,
1953                            bool IsStringLocation, Range StringRange,
1954                            FixItHint Fixit = FixItHint());
1955
1956  void CheckPositionalAndNonpositionalArgs(
1957      const analyze_format_string::FormatSpecifier *FS);
1958};
1959}
1960
1961SourceRange CheckFormatHandler::getFormatStringRange() {
1962  return OrigFormatExpr->getSourceRange();
1963}
1964
1965CharSourceRange CheckFormatHandler::
1966getSpecifierRange(const char *startSpecifier, unsigned specifierLen) {
1967  SourceLocation Start = getLocationOfByte(startSpecifier);
1968  SourceLocation End   = getLocationOfByte(startSpecifier + specifierLen - 1);
1969
1970  // Advance the end SourceLocation by one due to half-open ranges.
1971  End = End.getLocWithOffset(1);
1972
1973  return CharSourceRange::getCharRange(Start, End);
1974}
1975
1976SourceLocation CheckFormatHandler::getLocationOfByte(const char *x) {
1977  return S.getLocationOfStringLiteralByte(FExpr, x - Beg);
1978}
1979
1980void CheckFormatHandler::HandleIncompleteSpecifier(const char *startSpecifier,
1981                                                   unsigned specifierLen){
1982  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_incomplete_specifier),
1983                       getLocationOfByte(startSpecifier),
1984                       /*IsStringLocation*/true,
1985                       getSpecifierRange(startSpecifier, specifierLen));
1986}
1987
1988void CheckFormatHandler::HandleNonStandardLengthModifier(
1989    const analyze_format_string::LengthModifier &LM,
1990    const char *startSpecifier, unsigned specifierLen) {
1991  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard) << LM.toString()
1992                       << 0,
1993                       getLocationOfByte(LM.getStart()),
1994                       /*IsStringLocation*/true,
1995                       getSpecifierRange(startSpecifier, specifierLen));
1996}
1997
1998void CheckFormatHandler::HandleNonStandardConversionSpecifier(
1999    const analyze_format_string::ConversionSpecifier &CS,
2000    const char *startSpecifier, unsigned specifierLen) {
2001  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard) << CS.toString()
2002                       << 1,
2003                       getLocationOfByte(CS.getStart()),
2004                       /*IsStringLocation*/true,
2005                       getSpecifierRange(startSpecifier, specifierLen));
2006}
2007
2008void CheckFormatHandler::HandleNonStandardConversionSpecification(
2009    const analyze_format_string::LengthModifier &LM,
2010    const analyze_format_string::ConversionSpecifier &CS,
2011    const char *startSpecifier, unsigned specifierLen) {
2012  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard_conversion_spec)
2013                       << LM.toString() << CS.toString(),
2014                       getLocationOfByte(LM.getStart()),
2015                       /*IsStringLocation*/true,
2016                       getSpecifierRange(startSpecifier, specifierLen));
2017}
2018
2019void CheckFormatHandler::HandlePosition(const char *startPos,
2020                                        unsigned posLen) {
2021  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard_positional_arg),
2022                               getLocationOfByte(startPos),
2023                               /*IsStringLocation*/true,
2024                               getSpecifierRange(startPos, posLen));
2025}
2026
2027void
2028CheckFormatHandler::HandleInvalidPosition(const char *startPos, unsigned posLen,
2029                                     analyze_format_string::PositionContext p) {
2030  EmitFormatDiagnostic(S.PDiag(diag::warn_format_invalid_positional_specifier)
2031                         << (unsigned) p,
2032                       getLocationOfByte(startPos), /*IsStringLocation*/true,
2033                       getSpecifierRange(startPos, posLen));
2034}
2035
2036void CheckFormatHandler::HandleZeroPosition(const char *startPos,
2037                                            unsigned posLen) {
2038  EmitFormatDiagnostic(S.PDiag(diag::warn_format_zero_positional_specifier),
2039                               getLocationOfByte(startPos),
2040                               /*IsStringLocation*/true,
2041                               getSpecifierRange(startPos, posLen));
2042}
2043
2044void CheckFormatHandler::HandleNullChar(const char *nullCharacter) {
2045  if (!isa<ObjCStringLiteral>(OrigFormatExpr)) {
2046    // The presence of a null character is likely an error.
2047    EmitFormatDiagnostic(
2048      S.PDiag(diag::warn_printf_format_string_contains_null_char),
2049      getLocationOfByte(nullCharacter), /*IsStringLocation*/true,
2050      getFormatStringRange());
2051  }
2052}
2053
2054const Expr *CheckFormatHandler::getDataArg(unsigned i) const {
2055  return Args[FirstDataArg + i];
2056}
2057
2058void CheckFormatHandler::DoneProcessing() {
2059    // Does the number of data arguments exceed the number of
2060    // format conversions in the format string?
2061  if (!HasVAListArg) {
2062      // Find any arguments that weren't covered.
2063    CoveredArgs.flip();
2064    signed notCoveredArg = CoveredArgs.find_first();
2065    if (notCoveredArg >= 0) {
2066      assert((unsigned)notCoveredArg < NumDataArgs);
2067      SourceLocation Loc = getDataArg((unsigned) notCoveredArg)->getLocStart();
2068      if (!S.getSourceManager().isInSystemMacro(Loc)) {
2069        EmitFormatDiagnostic(S.PDiag(diag::warn_printf_data_arg_not_used),
2070                             Loc,
2071                             /*IsStringLocation*/false, getFormatStringRange());
2072      }
2073    }
2074  }
2075}
2076
2077bool
2078CheckFormatHandler::HandleInvalidConversionSpecifier(unsigned argIndex,
2079                                                     SourceLocation Loc,
2080                                                     const char *startSpec,
2081                                                     unsigned specifierLen,
2082                                                     const char *csStart,
2083                                                     unsigned csLen) {
2084
2085  bool keepGoing = true;
2086  if (argIndex < NumDataArgs) {
2087    // Consider the argument coverered, even though the specifier doesn't
2088    // make sense.
2089    CoveredArgs.set(argIndex);
2090  }
2091  else {
2092    // If argIndex exceeds the number of data arguments we
2093    // don't issue a warning because that is just a cascade of warnings (and
2094    // they may have intended '%%' anyway). We don't want to continue processing
2095    // the format string after this point, however, as we will like just get
2096    // gibberish when trying to match arguments.
2097    keepGoing = false;
2098  }
2099
2100  EmitFormatDiagnostic(S.PDiag(diag::warn_format_invalid_conversion)
2101                         << StringRef(csStart, csLen),
2102                       Loc, /*IsStringLocation*/true,
2103                       getSpecifierRange(startSpec, specifierLen));
2104
2105  return keepGoing;
2106}
2107
2108void
2109CheckFormatHandler::HandlePositionalNonpositionalArgs(SourceLocation Loc,
2110                                                      const char *startSpec,
2111                                                      unsigned specifierLen) {
2112  EmitFormatDiagnostic(
2113    S.PDiag(diag::warn_format_mix_positional_nonpositional_args),
2114    Loc, /*isStringLoc*/true, getSpecifierRange(startSpec, specifierLen));
2115}
2116
2117bool
2118CheckFormatHandler::CheckNumArgs(
2119  const analyze_format_string::FormatSpecifier &FS,
2120  const analyze_format_string::ConversionSpecifier &CS,
2121  const char *startSpecifier, unsigned specifierLen, unsigned argIndex) {
2122
2123  if (argIndex >= NumDataArgs) {
2124    PartialDiagnostic PDiag = FS.usesPositionalArg()
2125      ? (S.PDiag(diag::warn_printf_positional_arg_exceeds_data_args)
2126           << (argIndex+1) << NumDataArgs)
2127      : S.PDiag(diag::warn_printf_insufficient_data_args);
2128    EmitFormatDiagnostic(
2129      PDiag, getLocationOfByte(CS.getStart()), /*IsStringLocation*/true,
2130      getSpecifierRange(startSpecifier, specifierLen));
2131    return false;
2132  }
2133  return true;
2134}
2135
2136template<typename Range>
2137void CheckFormatHandler::EmitFormatDiagnostic(PartialDiagnostic PDiag,
2138                                              SourceLocation Loc,
2139                                              bool IsStringLocation,
2140                                              Range StringRange,
2141                                              FixItHint FixIt) {
2142  EmitFormatDiagnostic(S, inFunctionCall, Args[FormatIdx], PDiag,
2143                       Loc, IsStringLocation, StringRange, FixIt);
2144}
2145
2146/// \brief If the format string is not within the funcion call, emit a note
2147/// so that the function call and string are in diagnostic messages.
2148///
2149/// \param inFunctionCall if true, the format string is within the function
2150/// call and only one diagnostic message will be produced.  Otherwise, an
2151/// extra note will be emitted pointing to location of the format string.
2152///
2153/// \param ArgumentExpr the expression that is passed as the format string
2154/// argument in the function call.  Used for getting locations when two
2155/// diagnostics are emitted.
2156///
2157/// \param PDiag the callee should already have provided any strings for the
2158/// diagnostic message.  This function only adds locations and fixits
2159/// to diagnostics.
2160///
2161/// \param Loc primary location for diagnostic.  If two diagnostics are
2162/// required, one will be at Loc and a new SourceLocation will be created for
2163/// the other one.
2164///
2165/// \param IsStringLocation if true, Loc points to the format string should be
2166/// used for the note.  Otherwise, Loc points to the argument list and will
2167/// be used with PDiag.
2168///
2169/// \param StringRange some or all of the string to highlight.  This is
2170/// templated so it can accept either a CharSourceRange or a SourceRange.
2171///
2172/// \param Fixit optional fix it hint for the format string.
2173template<typename Range>
2174void CheckFormatHandler::EmitFormatDiagnostic(Sema &S, bool InFunctionCall,
2175                                              const Expr *ArgumentExpr,
2176                                              PartialDiagnostic PDiag,
2177                                              SourceLocation Loc,
2178                                              bool IsStringLocation,
2179                                              Range StringRange,
2180                                              FixItHint FixIt) {
2181  if (InFunctionCall)
2182    S.Diag(Loc, PDiag) << StringRange << FixIt;
2183  else {
2184    S.Diag(IsStringLocation ? ArgumentExpr->getExprLoc() : Loc, PDiag)
2185      << ArgumentExpr->getSourceRange();
2186    S.Diag(IsStringLocation ? Loc : StringRange.getBegin(),
2187           diag::note_format_string_defined)
2188      << StringRange << FixIt;
2189  }
2190}
2191
2192//===--- CHECK: Printf format string checking ------------------------------===//
2193
2194namespace {
2195class CheckPrintfHandler : public CheckFormatHandler {
2196  bool ObjCContext;
2197public:
2198  CheckPrintfHandler(Sema &s, const StringLiteral *fexpr,
2199                     const Expr *origFormatExpr, unsigned firstDataArg,
2200                     unsigned numDataArgs, bool isObjC,
2201                     const char *beg, bool hasVAListArg,
2202                     Expr **Args, unsigned NumArgs,
2203                     unsigned formatIdx, bool inFunctionCall)
2204  : CheckFormatHandler(s, fexpr, origFormatExpr, firstDataArg,
2205                       numDataArgs, beg, hasVAListArg, Args, NumArgs,
2206                       formatIdx, inFunctionCall), ObjCContext(isObjC) {}
2207
2208
2209  bool HandleInvalidPrintfConversionSpecifier(
2210                                      const analyze_printf::PrintfSpecifier &FS,
2211                                      const char *startSpecifier,
2212                                      unsigned specifierLen);
2213
2214  bool HandlePrintfSpecifier(const analyze_printf::PrintfSpecifier &FS,
2215                             const char *startSpecifier,
2216                             unsigned specifierLen);
2217  bool checkFormatExpr(const analyze_printf::PrintfSpecifier &FS,
2218                       const char *StartSpecifier,
2219                       unsigned SpecifierLen,
2220                       const Expr *E);
2221
2222  bool HandleAmount(const analyze_format_string::OptionalAmount &Amt, unsigned k,
2223                    const char *startSpecifier, unsigned specifierLen);
2224  void HandleInvalidAmount(const analyze_printf::PrintfSpecifier &FS,
2225                           const analyze_printf::OptionalAmount &Amt,
2226                           unsigned type,
2227                           const char *startSpecifier, unsigned specifierLen);
2228  void HandleFlag(const analyze_printf::PrintfSpecifier &FS,
2229                  const analyze_printf::OptionalFlag &flag,
2230                  const char *startSpecifier, unsigned specifierLen);
2231  void HandleIgnoredFlag(const analyze_printf::PrintfSpecifier &FS,
2232                         const analyze_printf::OptionalFlag &ignoredFlag,
2233                         const analyze_printf::OptionalFlag &flag,
2234                         const char *startSpecifier, unsigned specifierLen);
2235  bool checkForCStrMembers(const analyze_printf::ArgTypeResult &ATR,
2236                           const Expr *E, const CharSourceRange &CSR);
2237
2238};
2239}
2240
2241bool CheckPrintfHandler::HandleInvalidPrintfConversionSpecifier(
2242                                      const analyze_printf::PrintfSpecifier &FS,
2243                                      const char *startSpecifier,
2244                                      unsigned specifierLen) {
2245  const analyze_printf::PrintfConversionSpecifier &CS =
2246    FS.getConversionSpecifier();
2247
2248  return HandleInvalidConversionSpecifier(FS.getArgIndex(),
2249                                          getLocationOfByte(CS.getStart()),
2250                                          startSpecifier, specifierLen,
2251                                          CS.getStart(), CS.getLength());
2252}
2253
2254bool CheckPrintfHandler::HandleAmount(
2255                               const analyze_format_string::OptionalAmount &Amt,
2256                               unsigned k, const char *startSpecifier,
2257                               unsigned specifierLen) {
2258
2259  if (Amt.hasDataArgument()) {
2260    if (!HasVAListArg) {
2261      unsigned argIndex = Amt.getArgIndex();
2262      if (argIndex >= NumDataArgs) {
2263        EmitFormatDiagnostic(S.PDiag(diag::warn_printf_asterisk_missing_arg)
2264                               << k,
2265                             getLocationOfByte(Amt.getStart()),
2266                             /*IsStringLocation*/true,
2267                             getSpecifierRange(startSpecifier, specifierLen));
2268        // Don't do any more checking.  We will just emit
2269        // spurious errors.
2270        return false;
2271      }
2272
2273      // Type check the data argument.  It should be an 'int'.
2274      // Although not in conformance with C99, we also allow the argument to be
2275      // an 'unsigned int' as that is a reasonably safe case.  GCC also
2276      // doesn't emit a warning for that case.
2277      CoveredArgs.set(argIndex);
2278      const Expr *Arg = getDataArg(argIndex);
2279      QualType T = Arg->getType();
2280
2281      const analyze_printf::ArgTypeResult &ATR = Amt.getArgType(S.Context);
2282      assert(ATR.isValid());
2283
2284      if (!ATR.matchesType(S.Context, T)) {
2285        EmitFormatDiagnostic(S.PDiag(diag::warn_printf_asterisk_wrong_type)
2286                               << k << ATR.getRepresentativeTypeName(S.Context)
2287                               << T << Arg->getSourceRange(),
2288                             getLocationOfByte(Amt.getStart()),
2289                             /*IsStringLocation*/true,
2290                             getSpecifierRange(startSpecifier, specifierLen));
2291        // Don't do any more checking.  We will just emit
2292        // spurious errors.
2293        return false;
2294      }
2295    }
2296  }
2297  return true;
2298}
2299
2300void CheckPrintfHandler::HandleInvalidAmount(
2301                                      const analyze_printf::PrintfSpecifier &FS,
2302                                      const analyze_printf::OptionalAmount &Amt,
2303                                      unsigned type,
2304                                      const char *startSpecifier,
2305                                      unsigned specifierLen) {
2306  const analyze_printf::PrintfConversionSpecifier &CS =
2307    FS.getConversionSpecifier();
2308
2309  FixItHint fixit =
2310    Amt.getHowSpecified() == analyze_printf::OptionalAmount::Constant
2311      ? FixItHint::CreateRemoval(getSpecifierRange(Amt.getStart(),
2312                                 Amt.getConstantLength()))
2313      : FixItHint();
2314
2315  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_nonsensical_optional_amount)
2316                         << type << CS.toString(),
2317                       getLocationOfByte(Amt.getStart()),
2318                       /*IsStringLocation*/true,
2319                       getSpecifierRange(startSpecifier, specifierLen),
2320                       fixit);
2321}
2322
2323void CheckPrintfHandler::HandleFlag(const analyze_printf::PrintfSpecifier &FS,
2324                                    const analyze_printf::OptionalFlag &flag,
2325                                    const char *startSpecifier,
2326                                    unsigned specifierLen) {
2327  // Warn about pointless flag with a fixit removal.
2328  const analyze_printf::PrintfConversionSpecifier &CS =
2329    FS.getConversionSpecifier();
2330  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_nonsensical_flag)
2331                         << flag.toString() << CS.toString(),
2332                       getLocationOfByte(flag.getPosition()),
2333                       /*IsStringLocation*/true,
2334                       getSpecifierRange(startSpecifier, specifierLen),
2335                       FixItHint::CreateRemoval(
2336                         getSpecifierRange(flag.getPosition(), 1)));
2337}
2338
2339void CheckPrintfHandler::HandleIgnoredFlag(
2340                                const analyze_printf::PrintfSpecifier &FS,
2341                                const analyze_printf::OptionalFlag &ignoredFlag,
2342                                const analyze_printf::OptionalFlag &flag,
2343                                const char *startSpecifier,
2344                                unsigned specifierLen) {
2345  // Warn about ignored flag with a fixit removal.
2346  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_ignored_flag)
2347                         << ignoredFlag.toString() << flag.toString(),
2348                       getLocationOfByte(ignoredFlag.getPosition()),
2349                       /*IsStringLocation*/true,
2350                       getSpecifierRange(startSpecifier, specifierLen),
2351                       FixItHint::CreateRemoval(
2352                         getSpecifierRange(ignoredFlag.getPosition(), 1)));
2353}
2354
2355// Determines if the specified is a C++ class or struct containing
2356// a member with the specified name and kind (e.g. a CXXMethodDecl named
2357// "c_str()").
2358template<typename MemberKind>
2359static llvm::SmallPtrSet<MemberKind*, 1>
2360CXXRecordMembersNamed(StringRef Name, Sema &S, QualType Ty) {
2361  const RecordType *RT = Ty->getAs<RecordType>();
2362  llvm::SmallPtrSet<MemberKind*, 1> Results;
2363
2364  if (!RT)
2365    return Results;
2366  const CXXRecordDecl *RD = dyn_cast<CXXRecordDecl>(RT->getDecl());
2367  if (!RD)
2368    return Results;
2369
2370  LookupResult R(S, &S.PP.getIdentifierTable().get(Name), SourceLocation(),
2371                 Sema::LookupMemberName);
2372
2373  // We just need to include all members of the right kind turned up by the
2374  // filter, at this point.
2375  if (S.LookupQualifiedName(R, RT->getDecl()))
2376    for (LookupResult::iterator I = R.begin(), E = R.end(); I != E; ++I) {
2377      NamedDecl *decl = (*I)->getUnderlyingDecl();
2378      if (MemberKind *FK = dyn_cast<MemberKind>(decl))
2379        Results.insert(FK);
2380    }
2381  return Results;
2382}
2383
2384// Check if a (w)string was passed when a (w)char* was needed, and offer a
2385// better diagnostic if so. ATR is assumed to be valid.
2386// Returns true when a c_str() conversion method is found.
2387bool CheckPrintfHandler::checkForCStrMembers(
2388    const analyze_printf::ArgTypeResult &ATR, const Expr *E,
2389    const CharSourceRange &CSR) {
2390  typedef llvm::SmallPtrSet<CXXMethodDecl*, 1> MethodSet;
2391
2392  MethodSet Results =
2393      CXXRecordMembersNamed<CXXMethodDecl>("c_str", S, E->getType());
2394
2395  for (MethodSet::iterator MI = Results.begin(), ME = Results.end();
2396       MI != ME; ++MI) {
2397    const CXXMethodDecl *Method = *MI;
2398    if (Method->getNumParams() == 0 &&
2399          ATR.matchesType(S.Context, Method->getResultType())) {
2400      // FIXME: Suggest parens if the expression needs them.
2401      SourceLocation EndLoc =
2402          S.getPreprocessor().getLocForEndOfToken(E->getLocEnd());
2403      S.Diag(E->getLocStart(), diag::note_printf_c_str)
2404          << "c_str()"
2405          << FixItHint::CreateInsertion(EndLoc, ".c_str()");
2406      return true;
2407    }
2408  }
2409
2410  return false;
2411}
2412
2413bool
2414CheckPrintfHandler::HandlePrintfSpecifier(const analyze_printf::PrintfSpecifier
2415                                            &FS,
2416                                          const char *startSpecifier,
2417                                          unsigned specifierLen) {
2418
2419  using namespace analyze_format_string;
2420  using namespace analyze_printf;
2421  const PrintfConversionSpecifier &CS = FS.getConversionSpecifier();
2422
2423  if (FS.consumesDataArgument()) {
2424    if (atFirstArg) {
2425        atFirstArg = false;
2426        usesPositionalArgs = FS.usesPositionalArg();
2427    }
2428    else if (usesPositionalArgs != FS.usesPositionalArg()) {
2429      HandlePositionalNonpositionalArgs(getLocationOfByte(CS.getStart()),
2430                                        startSpecifier, specifierLen);
2431      return false;
2432    }
2433  }
2434
2435  // First check if the field width, precision, and conversion specifier
2436  // have matching data arguments.
2437  if (!HandleAmount(FS.getFieldWidth(), /* field width */ 0,
2438                    startSpecifier, specifierLen)) {
2439    return false;
2440  }
2441
2442  if (!HandleAmount(FS.getPrecision(), /* precision */ 1,
2443                    startSpecifier, specifierLen)) {
2444    return false;
2445  }
2446
2447  if (!CS.consumesDataArgument()) {
2448    // FIXME: Technically specifying a precision or field width here
2449    // makes no sense.  Worth issuing a warning at some point.
2450    return true;
2451  }
2452
2453  // Consume the argument.
2454  unsigned argIndex = FS.getArgIndex();
2455  if (argIndex < NumDataArgs) {
2456    // The check to see if the argIndex is valid will come later.
2457    // We set the bit here because we may exit early from this
2458    // function if we encounter some other error.
2459    CoveredArgs.set(argIndex);
2460  }
2461
2462  // Check for using an Objective-C specific conversion specifier
2463  // in a non-ObjC literal.
2464  if (!ObjCContext && CS.isObjCArg()) {
2465    return HandleInvalidPrintfConversionSpecifier(FS, startSpecifier,
2466                                                  specifierLen);
2467  }
2468
2469  // Check for invalid use of field width
2470  if (!FS.hasValidFieldWidth()) {
2471    HandleInvalidAmount(FS, FS.getFieldWidth(), /* field width */ 0,
2472        startSpecifier, specifierLen);
2473  }
2474
2475  // Check for invalid use of precision
2476  if (!FS.hasValidPrecision()) {
2477    HandleInvalidAmount(FS, FS.getPrecision(), /* precision */ 1,
2478        startSpecifier, specifierLen);
2479  }
2480
2481  // Check each flag does not conflict with any other component.
2482  if (!FS.hasValidThousandsGroupingPrefix())
2483    HandleFlag(FS, FS.hasThousandsGrouping(), startSpecifier, specifierLen);
2484  if (!FS.hasValidLeadingZeros())
2485    HandleFlag(FS, FS.hasLeadingZeros(), startSpecifier, specifierLen);
2486  if (!FS.hasValidPlusPrefix())
2487    HandleFlag(FS, FS.hasPlusPrefix(), startSpecifier, specifierLen);
2488  if (!FS.hasValidSpacePrefix())
2489    HandleFlag(FS, FS.hasSpacePrefix(), startSpecifier, specifierLen);
2490  if (!FS.hasValidAlternativeForm())
2491    HandleFlag(FS, FS.hasAlternativeForm(), startSpecifier, specifierLen);
2492  if (!FS.hasValidLeftJustified())
2493    HandleFlag(FS, FS.isLeftJustified(), startSpecifier, specifierLen);
2494
2495  // Check that flags are not ignored by another flag
2496  if (FS.hasSpacePrefix() && FS.hasPlusPrefix()) // ' ' ignored by '+'
2497    HandleIgnoredFlag(FS, FS.hasSpacePrefix(), FS.hasPlusPrefix(),
2498        startSpecifier, specifierLen);
2499  if (FS.hasLeadingZeros() && FS.isLeftJustified()) // '0' ignored by '-'
2500    HandleIgnoredFlag(FS, FS.hasLeadingZeros(), FS.isLeftJustified(),
2501            startSpecifier, specifierLen);
2502
2503  // Check the length modifier is valid with the given conversion specifier.
2504  const LengthModifier &LM = FS.getLengthModifier();
2505  if (!FS.hasValidLengthModifier())
2506    EmitFormatDiagnostic(S.PDiag(diag::warn_format_nonsensical_length)
2507                           << LM.toString() << CS.toString(),
2508                         getLocationOfByte(LM.getStart()),
2509                         /*IsStringLocation*/true,
2510                         getSpecifierRange(startSpecifier, specifierLen),
2511                         FixItHint::CreateRemoval(
2512                           getSpecifierRange(LM.getStart(),
2513                                             LM.getLength())));
2514  if (!FS.hasStandardLengthModifier())
2515    HandleNonStandardLengthModifier(LM, startSpecifier, specifierLen);
2516  if (!FS.hasStandardConversionSpecifier(S.getLangOpts()))
2517    HandleNonStandardConversionSpecifier(CS, startSpecifier, specifierLen);
2518  if (!FS.hasStandardLengthConversionCombination())
2519    HandleNonStandardConversionSpecification(LM, CS, startSpecifier,
2520                                             specifierLen);
2521
2522  // Are we using '%n'?
2523  if (CS.getKind() == ConversionSpecifier::nArg) {
2524    // Issue a warning about this being a possible security issue.
2525    EmitFormatDiagnostic(S.PDiag(diag::warn_printf_write_back),
2526                         getLocationOfByte(CS.getStart()),
2527                         /*IsStringLocation*/true,
2528                         getSpecifierRange(startSpecifier, specifierLen));
2529    // Continue checking the other format specifiers.
2530    return true;
2531  }
2532
2533  // The remaining checks depend on the data arguments.
2534  if (HasVAListArg)
2535    return true;
2536
2537  if (!CheckNumArgs(FS, CS, startSpecifier, specifierLen, argIndex))
2538    return false;
2539
2540  return checkFormatExpr(FS, startSpecifier, specifierLen,
2541                         getDataArg(argIndex));
2542}
2543
2544bool
2545CheckPrintfHandler::checkFormatExpr(const analyze_printf::PrintfSpecifier &FS,
2546                                    const char *StartSpecifier,
2547                                    unsigned SpecifierLen,
2548                                    const Expr *E) {
2549  using namespace analyze_format_string;
2550  using namespace analyze_printf;
2551  // Now type check the data expression that matches the
2552  // format specifier.
2553  const analyze_printf::ArgTypeResult &ATR = FS.getArgType(S.Context,
2554                                                           ObjCContext);
2555  if (ATR.isValid() && !ATR.matchesType(S.Context, E->getType())) {
2556    // Look through argument promotions for our error message's reported type.
2557    // This includes the integral and floating promotions, but excludes array
2558    // and function pointer decay; seeing that an argument intended to be a
2559    // string has type 'char [6]' is probably more confusing than 'char *'.
2560    if (const ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(E)) {
2561      if (ICE->getCastKind() == CK_IntegralCast ||
2562          ICE->getCastKind() == CK_FloatingCast) {
2563        E = ICE->getSubExpr();
2564
2565        // Check if we didn't match because of an implicit cast from a 'char'
2566        // or 'short' to an 'int'.  This is done because printf is a varargs
2567        // function.
2568        if (ICE->getType() == S.Context.IntTy ||
2569            ICE->getType() == S.Context.UnsignedIntTy) {
2570          // All further checking is done on the subexpression.
2571          if (ATR.matchesType(S.Context, E->getType()))
2572            return true;
2573        }
2574      }
2575    }
2576
2577    // We may be able to offer a FixItHint if it is a supported type.
2578    PrintfSpecifier fixedFS = FS;
2579    bool success = fixedFS.fixType(E->getType(), S.getLangOpts(),
2580                                   S.Context, ObjCContext);
2581
2582    if (success) {
2583      // Get the fix string from the fixed format specifier
2584      SmallString<16> buf;
2585      llvm::raw_svector_ostream os(buf);
2586      fixedFS.toString(os);
2587
2588      EmitFormatDiagnostic(
2589        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2590          << ATR.getRepresentativeTypeName(S.Context) << E->getType()
2591          << E->getSourceRange(),
2592        E->getLocStart(),
2593        /*IsStringLocation*/false,
2594        getSpecifierRange(StartSpecifier, SpecifierLen),
2595        FixItHint::CreateReplacement(
2596          getSpecifierRange(StartSpecifier, SpecifierLen),
2597          os.str()));
2598    } else {
2599      const CharSourceRange &CSR = getSpecifierRange(StartSpecifier,
2600                                                     SpecifierLen);
2601      // Since the warning for passing non-POD types to variadic functions
2602      // was deferred until now, we emit a warning for non-POD
2603      // arguments here.
2604      if (S.isValidVarArgType(E->getType()) == Sema::VAK_Invalid) {
2605        EmitFormatDiagnostic(
2606          S.PDiag(diag::warn_non_pod_vararg_with_format_string)
2607            << S.getLangOpts().CPlusPlus0x
2608            << E->getType()
2609            << ATR.getRepresentativeTypeName(S.Context)
2610            << CSR
2611            << E->getSourceRange(),
2612          E->getLocStart(), /*IsStringLocation*/false, CSR);
2613
2614        checkForCStrMembers(ATR, E, CSR);
2615      } else
2616        EmitFormatDiagnostic(
2617          S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2618            << ATR.getRepresentativeTypeName(S.Context) << E->getType()
2619            << CSR
2620            << E->getSourceRange(),
2621          E->getLocStart(), /*IsStringLocation*/false, CSR);
2622    }
2623  }
2624
2625  return true;
2626}
2627
2628//===--- CHECK: Scanf format string checking ------------------------------===//
2629
2630namespace {
2631class CheckScanfHandler : public CheckFormatHandler {
2632public:
2633  CheckScanfHandler(Sema &s, const StringLiteral *fexpr,
2634                    const Expr *origFormatExpr, unsigned firstDataArg,
2635                    unsigned numDataArgs, const char *beg, bool hasVAListArg,
2636                    Expr **Args, unsigned NumArgs,
2637                    unsigned formatIdx, bool inFunctionCall)
2638  : CheckFormatHandler(s, fexpr, origFormatExpr, firstDataArg,
2639                       numDataArgs, beg, hasVAListArg,
2640                       Args, NumArgs, formatIdx, inFunctionCall) {}
2641
2642  bool HandleScanfSpecifier(const analyze_scanf::ScanfSpecifier &FS,
2643                            const char *startSpecifier,
2644                            unsigned specifierLen);
2645
2646  bool HandleInvalidScanfConversionSpecifier(
2647          const analyze_scanf::ScanfSpecifier &FS,
2648          const char *startSpecifier,
2649          unsigned specifierLen);
2650
2651  void HandleIncompleteScanList(const char *start, const char *end);
2652};
2653}
2654
2655void CheckScanfHandler::HandleIncompleteScanList(const char *start,
2656                                                 const char *end) {
2657  EmitFormatDiagnostic(S.PDiag(diag::warn_scanf_scanlist_incomplete),
2658                       getLocationOfByte(end), /*IsStringLocation*/true,
2659                       getSpecifierRange(start, end - start));
2660}
2661
2662bool CheckScanfHandler::HandleInvalidScanfConversionSpecifier(
2663                                        const analyze_scanf::ScanfSpecifier &FS,
2664                                        const char *startSpecifier,
2665                                        unsigned specifierLen) {
2666
2667  const analyze_scanf::ScanfConversionSpecifier &CS =
2668    FS.getConversionSpecifier();
2669
2670  return HandleInvalidConversionSpecifier(FS.getArgIndex(),
2671                                          getLocationOfByte(CS.getStart()),
2672                                          startSpecifier, specifierLen,
2673                                          CS.getStart(), CS.getLength());
2674}
2675
2676bool CheckScanfHandler::HandleScanfSpecifier(
2677                                       const analyze_scanf::ScanfSpecifier &FS,
2678                                       const char *startSpecifier,
2679                                       unsigned specifierLen) {
2680
2681  using namespace analyze_scanf;
2682  using namespace analyze_format_string;
2683
2684  const ScanfConversionSpecifier &CS = FS.getConversionSpecifier();
2685
2686  // Handle case where '%' and '*' don't consume an argument.  These shouldn't
2687  // be used to decide if we are using positional arguments consistently.
2688  if (FS.consumesDataArgument()) {
2689    if (atFirstArg) {
2690      atFirstArg = false;
2691      usesPositionalArgs = FS.usesPositionalArg();
2692    }
2693    else if (usesPositionalArgs != FS.usesPositionalArg()) {
2694      HandlePositionalNonpositionalArgs(getLocationOfByte(CS.getStart()),
2695                                        startSpecifier, specifierLen);
2696      return false;
2697    }
2698  }
2699
2700  // Check if the field with is non-zero.
2701  const OptionalAmount &Amt = FS.getFieldWidth();
2702  if (Amt.getHowSpecified() == OptionalAmount::Constant) {
2703    if (Amt.getConstantAmount() == 0) {
2704      const CharSourceRange &R = getSpecifierRange(Amt.getStart(),
2705                                                   Amt.getConstantLength());
2706      EmitFormatDiagnostic(S.PDiag(diag::warn_scanf_nonzero_width),
2707                           getLocationOfByte(Amt.getStart()),
2708                           /*IsStringLocation*/true, R,
2709                           FixItHint::CreateRemoval(R));
2710    }
2711  }
2712
2713  if (!FS.consumesDataArgument()) {
2714    // FIXME: Technically specifying a precision or field width here
2715    // makes no sense.  Worth issuing a warning at some point.
2716    return true;
2717  }
2718
2719  // Consume the argument.
2720  unsigned argIndex = FS.getArgIndex();
2721  if (argIndex < NumDataArgs) {
2722      // The check to see if the argIndex is valid will come later.
2723      // We set the bit here because we may exit early from this
2724      // function if we encounter some other error.
2725    CoveredArgs.set(argIndex);
2726  }
2727
2728  // Check the length modifier is valid with the given conversion specifier.
2729  const LengthModifier &LM = FS.getLengthModifier();
2730  if (!FS.hasValidLengthModifier()) {
2731    const CharSourceRange &R = getSpecifierRange(LM.getStart(), LM.getLength());
2732    EmitFormatDiagnostic(S.PDiag(diag::warn_format_nonsensical_length)
2733                         << LM.toString() << CS.toString()
2734                         << getSpecifierRange(startSpecifier, specifierLen),
2735                         getLocationOfByte(LM.getStart()),
2736                         /*IsStringLocation*/true, R,
2737                         FixItHint::CreateRemoval(R));
2738  }
2739
2740  if (!FS.hasStandardLengthModifier())
2741    HandleNonStandardLengthModifier(LM, startSpecifier, specifierLen);
2742  if (!FS.hasStandardConversionSpecifier(S.getLangOpts()))
2743    HandleNonStandardConversionSpecifier(CS, startSpecifier, specifierLen);
2744  if (!FS.hasStandardLengthConversionCombination())
2745    HandleNonStandardConversionSpecification(LM, CS, startSpecifier,
2746                                             specifierLen);
2747
2748  // The remaining checks depend on the data arguments.
2749  if (HasVAListArg)
2750    return true;
2751
2752  if (!CheckNumArgs(FS, CS, startSpecifier, specifierLen, argIndex))
2753    return false;
2754
2755  // Check that the argument type matches the format specifier.
2756  const Expr *Ex = getDataArg(argIndex);
2757  const analyze_scanf::ScanfArgTypeResult &ATR = FS.getArgType(S.Context);
2758  if (ATR.isValid() && !ATR.matchesType(S.Context, Ex->getType())) {
2759    ScanfSpecifier fixedFS = FS;
2760    bool success = fixedFS.fixType(Ex->getType(), S.getLangOpts(),
2761                                   S.Context);
2762
2763    if (success) {
2764      // Get the fix string from the fixed format specifier.
2765      SmallString<128> buf;
2766      llvm::raw_svector_ostream os(buf);
2767      fixedFS.toString(os);
2768
2769      EmitFormatDiagnostic(
2770        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2771          << ATR.getRepresentativeTypeName(S.Context) << Ex->getType()
2772          << Ex->getSourceRange(),
2773        Ex->getLocStart(),
2774        /*IsStringLocation*/false,
2775        getSpecifierRange(startSpecifier, specifierLen),
2776        FixItHint::CreateReplacement(
2777          getSpecifierRange(startSpecifier, specifierLen),
2778          os.str()));
2779    } else {
2780      EmitFormatDiagnostic(
2781        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2782          << ATR.getRepresentativeTypeName(S.Context) << Ex->getType()
2783          << Ex->getSourceRange(),
2784        Ex->getLocStart(),
2785        /*IsStringLocation*/false,
2786        getSpecifierRange(startSpecifier, specifierLen));
2787    }
2788  }
2789
2790  return true;
2791}
2792
2793void Sema::CheckFormatString(const StringLiteral *FExpr,
2794                             const Expr *OrigFormatExpr,
2795                             Expr **Args, unsigned NumArgs,
2796                             bool HasVAListArg, unsigned format_idx,
2797                             unsigned firstDataArg, FormatStringType Type,
2798                             bool inFunctionCall) {
2799
2800  // CHECK: is the format string a wide literal?
2801  if (!FExpr->isAscii() && !FExpr->isUTF8()) {
2802    CheckFormatHandler::EmitFormatDiagnostic(
2803      *this, inFunctionCall, Args[format_idx],
2804      PDiag(diag::warn_format_string_is_wide_literal), FExpr->getLocStart(),
2805      /*IsStringLocation*/true, OrigFormatExpr->getSourceRange());
2806    return;
2807  }
2808
2809  // Str - The format string.  NOTE: this is NOT null-terminated!
2810  StringRef StrRef = FExpr->getString();
2811  const char *Str = StrRef.data();
2812  unsigned StrLen = StrRef.size();
2813  const unsigned numDataArgs = NumArgs - firstDataArg;
2814
2815  // CHECK: empty format string?
2816  if (StrLen == 0 && numDataArgs > 0) {
2817    CheckFormatHandler::EmitFormatDiagnostic(
2818      *this, inFunctionCall, Args[format_idx],
2819      PDiag(diag::warn_empty_format_string), FExpr->getLocStart(),
2820      /*IsStringLocation*/true, OrigFormatExpr->getSourceRange());
2821    return;
2822  }
2823
2824  if (Type == FST_Printf || Type == FST_NSString) {
2825    CheckPrintfHandler H(*this, FExpr, OrigFormatExpr, firstDataArg,
2826                         numDataArgs, (Type == FST_NSString),
2827                         Str, HasVAListArg, Args, NumArgs, format_idx,
2828                         inFunctionCall);
2829
2830    if (!analyze_format_string::ParsePrintfString(H, Str, Str + StrLen,
2831                                                  getLangOpts()))
2832      H.DoneProcessing();
2833  } else if (Type == FST_Scanf) {
2834    CheckScanfHandler H(*this, FExpr, OrigFormatExpr, firstDataArg, numDataArgs,
2835                        Str, HasVAListArg, Args, NumArgs, format_idx,
2836                        inFunctionCall);
2837
2838    if (!analyze_format_string::ParseScanfString(H, Str, Str + StrLen,
2839                                                 getLangOpts()))
2840      H.DoneProcessing();
2841  } // TODO: handle other formats
2842}
2843
2844//===--- CHECK: Standard memory functions ---------------------------------===//
2845
2846/// \brief Determine whether the given type is a dynamic class type (e.g.,
2847/// whether it has a vtable).
2848static bool isDynamicClassType(QualType T) {
2849  if (CXXRecordDecl *Record = T->getAsCXXRecordDecl())
2850    if (CXXRecordDecl *Definition = Record->getDefinition())
2851      if (Definition->isDynamicClass())
2852        return true;
2853
2854  return false;
2855}
2856
2857/// \brief If E is a sizeof expression, returns its argument expression,
2858/// otherwise returns NULL.
2859static const Expr *getSizeOfExprArg(const Expr* E) {
2860  if (const UnaryExprOrTypeTraitExpr *SizeOf =
2861      dyn_cast<UnaryExprOrTypeTraitExpr>(E))
2862    if (SizeOf->getKind() == clang::UETT_SizeOf && !SizeOf->isArgumentType())
2863      return SizeOf->getArgumentExpr()->IgnoreParenImpCasts();
2864
2865  return 0;
2866}
2867
2868/// \brief If E is a sizeof expression, returns its argument type.
2869static QualType getSizeOfArgType(const Expr* E) {
2870  if (const UnaryExprOrTypeTraitExpr *SizeOf =
2871      dyn_cast<UnaryExprOrTypeTraitExpr>(E))
2872    if (SizeOf->getKind() == clang::UETT_SizeOf)
2873      return SizeOf->getTypeOfArgument();
2874
2875  return QualType();
2876}
2877
2878/// \brief Check for dangerous or invalid arguments to memset().
2879///
2880/// This issues warnings on known problematic, dangerous or unspecified
2881/// arguments to the standard 'memset', 'memcpy', 'memmove', and 'memcmp'
2882/// function calls.
2883///
2884/// \param Call The call expression to diagnose.
2885void Sema::CheckMemaccessArguments(const CallExpr *Call,
2886                                   unsigned BId,
2887                                   IdentifierInfo *FnName) {
2888  assert(BId != 0);
2889
2890  // It is possible to have a non-standard definition of memset.  Validate
2891  // we have enough arguments, and if not, abort further checking.
2892  unsigned ExpectedNumArgs = (BId == Builtin::BIstrndup ? 2 : 3);
2893  if (Call->getNumArgs() < ExpectedNumArgs)
2894    return;
2895
2896  unsigned LastArg = (BId == Builtin::BImemset ||
2897                      BId == Builtin::BIstrndup ? 1 : 2);
2898  unsigned LenArg = (BId == Builtin::BIstrndup ? 1 : 2);
2899  const Expr *LenExpr = Call->getArg(LenArg)->IgnoreParenImpCasts();
2900
2901  // We have special checking when the length is a sizeof expression.
2902  QualType SizeOfArgTy = getSizeOfArgType(LenExpr);
2903  const Expr *SizeOfArg = getSizeOfExprArg(LenExpr);
2904  llvm::FoldingSetNodeID SizeOfArgID;
2905
2906  for (unsigned ArgIdx = 0; ArgIdx != LastArg; ++ArgIdx) {
2907    const Expr *Dest = Call->getArg(ArgIdx)->IgnoreParenImpCasts();
2908    SourceRange ArgRange = Call->getArg(ArgIdx)->getSourceRange();
2909
2910    QualType DestTy = Dest->getType();
2911    if (const PointerType *DestPtrTy = DestTy->getAs<PointerType>()) {
2912      QualType PointeeTy = DestPtrTy->getPointeeType();
2913
2914      // Never warn about void type pointers. This can be used to suppress
2915      // false positives.
2916      if (PointeeTy->isVoidType())
2917        continue;
2918
2919      // Catch "memset(p, 0, sizeof(p))" -- needs to be sizeof(*p). Do this by
2920      // actually comparing the expressions for equality. Because computing the
2921      // expression IDs can be expensive, we only do this if the diagnostic is
2922      // enabled.
2923      if (SizeOfArg &&
2924          Diags.getDiagnosticLevel(diag::warn_sizeof_pointer_expr_memaccess,
2925                                   SizeOfArg->getExprLoc())) {
2926        // We only compute IDs for expressions if the warning is enabled, and
2927        // cache the sizeof arg's ID.
2928        if (SizeOfArgID == llvm::FoldingSetNodeID())
2929          SizeOfArg->Profile(SizeOfArgID, Context, true);
2930        llvm::FoldingSetNodeID DestID;
2931        Dest->Profile(DestID, Context, true);
2932        if (DestID == SizeOfArgID) {
2933          // TODO: For strncpy() and friends, this could suggest sizeof(dst)
2934          //       over sizeof(src) as well.
2935          unsigned ActionIdx = 0; // Default is to suggest dereferencing.
2936          StringRef ReadableName = FnName->getName();
2937
2938          if (const UnaryOperator *UnaryOp = dyn_cast<UnaryOperator>(Dest))
2939            if (UnaryOp->getOpcode() == UO_AddrOf)
2940              ActionIdx = 1; // If its an address-of operator, just remove it.
2941          if (Context.getTypeSize(PointeeTy) == Context.getCharWidth())
2942            ActionIdx = 2; // If the pointee's size is sizeof(char),
2943                           // suggest an explicit length.
2944
2945          // If the function is defined as a builtin macro, do not show macro
2946          // expansion.
2947          SourceLocation SL = SizeOfArg->getExprLoc();
2948          SourceRange DSR = Dest->getSourceRange();
2949          SourceRange SSR = SizeOfArg->getSourceRange();
2950          SourceManager &SM  = PP.getSourceManager();
2951
2952          if (SM.isMacroArgExpansion(SL)) {
2953            ReadableName = Lexer::getImmediateMacroName(SL, SM, LangOpts);
2954            SL = SM.getSpellingLoc(SL);
2955            DSR = SourceRange(SM.getSpellingLoc(DSR.getBegin()),
2956                             SM.getSpellingLoc(DSR.getEnd()));
2957            SSR = SourceRange(SM.getSpellingLoc(SSR.getBegin()),
2958                             SM.getSpellingLoc(SSR.getEnd()));
2959          }
2960
2961          DiagRuntimeBehavior(SL, SizeOfArg,
2962                              PDiag(diag::warn_sizeof_pointer_expr_memaccess)
2963                                << ReadableName
2964                                << PointeeTy
2965                                << DestTy
2966                                << DSR
2967                                << SSR);
2968          DiagRuntimeBehavior(SL, SizeOfArg,
2969                         PDiag(diag::warn_sizeof_pointer_expr_memaccess_note)
2970                                << ActionIdx
2971                                << SSR);
2972
2973          break;
2974        }
2975      }
2976
2977      // Also check for cases where the sizeof argument is the exact same
2978      // type as the memory argument, and where it points to a user-defined
2979      // record type.
2980      if (SizeOfArgTy != QualType()) {
2981        if (PointeeTy->isRecordType() &&
2982            Context.typesAreCompatible(SizeOfArgTy, DestTy)) {
2983          DiagRuntimeBehavior(LenExpr->getExprLoc(), Dest,
2984                              PDiag(diag::warn_sizeof_pointer_type_memaccess)
2985                                << FnName << SizeOfArgTy << ArgIdx
2986                                << PointeeTy << Dest->getSourceRange()
2987                                << LenExpr->getSourceRange());
2988          break;
2989        }
2990      }
2991
2992      // Always complain about dynamic classes.
2993      if (isDynamicClassType(PointeeTy)) {
2994
2995        unsigned OperationType = 0;
2996        // "overwritten" if we're warning about the destination for any call
2997        // but memcmp; otherwise a verb appropriate to the call.
2998        if (ArgIdx != 0 || BId == Builtin::BImemcmp) {
2999          if (BId == Builtin::BImemcpy)
3000            OperationType = 1;
3001          else if(BId == Builtin::BImemmove)
3002            OperationType = 2;
3003          else if (BId == Builtin::BImemcmp)
3004            OperationType = 3;
3005        }
3006
3007        DiagRuntimeBehavior(
3008          Dest->getExprLoc(), Dest,
3009          PDiag(diag::warn_dyn_class_memaccess)
3010            << (BId == Builtin::BImemcmp ? ArgIdx + 2 : ArgIdx)
3011            << FnName << PointeeTy
3012            << OperationType
3013            << Call->getCallee()->getSourceRange());
3014      } else if (PointeeTy.hasNonTrivialObjCLifetime() &&
3015               BId != Builtin::BImemset)
3016        DiagRuntimeBehavior(
3017          Dest->getExprLoc(), Dest,
3018          PDiag(diag::warn_arc_object_memaccess)
3019            << ArgIdx << FnName << PointeeTy
3020            << Call->getCallee()->getSourceRange());
3021      else
3022        continue;
3023
3024      DiagRuntimeBehavior(
3025        Dest->getExprLoc(), Dest,
3026        PDiag(diag::note_bad_memaccess_silence)
3027          << FixItHint::CreateInsertion(ArgRange.getBegin(), "(void*)"));
3028      break;
3029    }
3030  }
3031}
3032
3033// A little helper routine: ignore addition and subtraction of integer literals.
3034// This intentionally does not ignore all integer constant expressions because
3035// we don't want to remove sizeof().
3036static const Expr *ignoreLiteralAdditions(const Expr *Ex, ASTContext &Ctx) {
3037  Ex = Ex->IgnoreParenCasts();
3038
3039  for (;;) {
3040    const BinaryOperator * BO = dyn_cast<BinaryOperator>(Ex);
3041    if (!BO || !BO->isAdditiveOp())
3042      break;
3043
3044    const Expr *RHS = BO->getRHS()->IgnoreParenCasts();
3045    const Expr *LHS = BO->getLHS()->IgnoreParenCasts();
3046
3047    if (isa<IntegerLiteral>(RHS))
3048      Ex = LHS;
3049    else if (isa<IntegerLiteral>(LHS))
3050      Ex = RHS;
3051    else
3052      break;
3053  }
3054
3055  return Ex;
3056}
3057
3058// Warn if the user has made the 'size' argument to strlcpy or strlcat
3059// be the size of the source, instead of the destination.
3060void Sema::CheckStrlcpycatArguments(const CallExpr *Call,
3061                                    IdentifierInfo *FnName) {
3062
3063  // Don't crash if the user has the wrong number of arguments
3064  if (Call->getNumArgs() != 3)
3065    return;
3066
3067  const Expr *SrcArg = ignoreLiteralAdditions(Call->getArg(1), Context);
3068  const Expr *SizeArg = ignoreLiteralAdditions(Call->getArg(2), Context);
3069  const Expr *CompareWithSrc = NULL;
3070
3071  // Look for 'strlcpy(dst, x, sizeof(x))'
3072  if (const Expr *Ex = getSizeOfExprArg(SizeArg))
3073    CompareWithSrc = Ex;
3074  else {
3075    // Look for 'strlcpy(dst, x, strlen(x))'
3076    if (const CallExpr *SizeCall = dyn_cast<CallExpr>(SizeArg)) {
3077      if (SizeCall->isBuiltinCall() == Builtin::BIstrlen
3078          && SizeCall->getNumArgs() == 1)
3079        CompareWithSrc = ignoreLiteralAdditions(SizeCall->getArg(0), Context);
3080    }
3081  }
3082
3083  if (!CompareWithSrc)
3084    return;
3085
3086  // Determine if the argument to sizeof/strlen is equal to the source
3087  // argument.  In principle there's all kinds of things you could do
3088  // here, for instance creating an == expression and evaluating it with
3089  // EvaluateAsBooleanCondition, but this uses a more direct technique:
3090  const DeclRefExpr *SrcArgDRE = dyn_cast<DeclRefExpr>(SrcArg);
3091  if (!SrcArgDRE)
3092    return;
3093
3094  const DeclRefExpr *CompareWithSrcDRE = dyn_cast<DeclRefExpr>(CompareWithSrc);
3095  if (!CompareWithSrcDRE ||
3096      SrcArgDRE->getDecl() != CompareWithSrcDRE->getDecl())
3097    return;
3098
3099  const Expr *OriginalSizeArg = Call->getArg(2);
3100  Diag(CompareWithSrcDRE->getLocStart(), diag::warn_strlcpycat_wrong_size)
3101    << OriginalSizeArg->getSourceRange() << FnName;
3102
3103  // Output a FIXIT hint if the destination is an array (rather than a
3104  // pointer to an array).  This could be enhanced to handle some
3105  // pointers if we know the actual size, like if DstArg is 'array+2'
3106  // we could say 'sizeof(array)-2'.
3107  const Expr *DstArg = Call->getArg(0)->IgnoreParenImpCasts();
3108  QualType DstArgTy = DstArg->getType();
3109
3110  // Only handle constant-sized or VLAs, but not flexible members.
3111  if (const ConstantArrayType *CAT = Context.getAsConstantArrayType(DstArgTy)) {
3112    // Only issue the FIXIT for arrays of size > 1.
3113    if (CAT->getSize().getSExtValue() <= 1)
3114      return;
3115  } else if (!DstArgTy->isVariableArrayType()) {
3116    return;
3117  }
3118
3119  SmallString<128> sizeString;
3120  llvm::raw_svector_ostream OS(sizeString);
3121  OS << "sizeof(";
3122  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3123  OS << ")";
3124
3125  Diag(OriginalSizeArg->getLocStart(), diag::note_strlcpycat_wrong_size)
3126    << FixItHint::CreateReplacement(OriginalSizeArg->getSourceRange(),
3127                                    OS.str());
3128}
3129
3130/// Check if two expressions refer to the same declaration.
3131static bool referToTheSameDecl(const Expr *E1, const Expr *E2) {
3132  if (const DeclRefExpr *D1 = dyn_cast_or_null<DeclRefExpr>(E1))
3133    if (const DeclRefExpr *D2 = dyn_cast_or_null<DeclRefExpr>(E2))
3134      return D1->getDecl() == D2->getDecl();
3135  return false;
3136}
3137
3138static const Expr *getStrlenExprArg(const Expr *E) {
3139  if (const CallExpr *CE = dyn_cast<CallExpr>(E)) {
3140    const FunctionDecl *FD = CE->getDirectCallee();
3141    if (!FD || FD->getMemoryFunctionKind() != Builtin::BIstrlen)
3142      return 0;
3143    return CE->getArg(0)->IgnoreParenCasts();
3144  }
3145  return 0;
3146}
3147
3148// Warn on anti-patterns as the 'size' argument to strncat.
3149// The correct size argument should look like following:
3150//   strncat(dst, src, sizeof(dst) - strlen(dest) - 1);
3151void Sema::CheckStrncatArguments(const CallExpr *CE,
3152                                 IdentifierInfo *FnName) {
3153  // Don't crash if the user has the wrong number of arguments.
3154  if (CE->getNumArgs() < 3)
3155    return;
3156  const Expr *DstArg = CE->getArg(0)->IgnoreParenCasts();
3157  const Expr *SrcArg = CE->getArg(1)->IgnoreParenCasts();
3158  const Expr *LenArg = CE->getArg(2)->IgnoreParenCasts();
3159
3160  // Identify common expressions, which are wrongly used as the size argument
3161  // to strncat and may lead to buffer overflows.
3162  unsigned PatternType = 0;
3163  if (const Expr *SizeOfArg = getSizeOfExprArg(LenArg)) {
3164    // - sizeof(dst)
3165    if (referToTheSameDecl(SizeOfArg, DstArg))
3166      PatternType = 1;
3167    // - sizeof(src)
3168    else if (referToTheSameDecl(SizeOfArg, SrcArg))
3169      PatternType = 2;
3170  } else if (const BinaryOperator *BE = dyn_cast<BinaryOperator>(LenArg)) {
3171    if (BE->getOpcode() == BO_Sub) {
3172      const Expr *L = BE->getLHS()->IgnoreParenCasts();
3173      const Expr *R = BE->getRHS()->IgnoreParenCasts();
3174      // - sizeof(dst) - strlen(dst)
3175      if (referToTheSameDecl(DstArg, getSizeOfExprArg(L)) &&
3176          referToTheSameDecl(DstArg, getStrlenExprArg(R)))
3177        PatternType = 1;
3178      // - sizeof(src) - (anything)
3179      else if (referToTheSameDecl(SrcArg, getSizeOfExprArg(L)))
3180        PatternType = 2;
3181    }
3182  }
3183
3184  if (PatternType == 0)
3185    return;
3186
3187  // Generate the diagnostic.
3188  SourceLocation SL = LenArg->getLocStart();
3189  SourceRange SR = LenArg->getSourceRange();
3190  SourceManager &SM  = PP.getSourceManager();
3191
3192  // If the function is defined as a builtin macro, do not show macro expansion.
3193  if (SM.isMacroArgExpansion(SL)) {
3194    SL = SM.getSpellingLoc(SL);
3195    SR = SourceRange(SM.getSpellingLoc(SR.getBegin()),
3196                     SM.getSpellingLoc(SR.getEnd()));
3197  }
3198
3199  if (PatternType == 1)
3200    Diag(SL, diag::warn_strncat_large_size) << SR;
3201  else
3202    Diag(SL, diag::warn_strncat_src_size) << SR;
3203
3204  // Output a FIXIT hint if the destination is an array (rather than a
3205  // pointer to an array).  This could be enhanced to handle some
3206  // pointers if we know the actual size, like if DstArg is 'array+2'
3207  // we could say 'sizeof(array)-2'.
3208  QualType DstArgTy = DstArg->getType();
3209
3210  // Only handle constant-sized or VLAs, but not flexible members.
3211  if (const ConstantArrayType *CAT = Context.getAsConstantArrayType(DstArgTy)) {
3212    // Only issue the FIXIT for arrays of size > 1.
3213    if (CAT->getSize().getSExtValue() <= 1)
3214      return;
3215  } else if (!DstArgTy->isVariableArrayType()) {
3216    return;
3217  }
3218
3219  SmallString<128> sizeString;
3220  llvm::raw_svector_ostream OS(sizeString);
3221  OS << "sizeof(";
3222  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3223  OS << ") - ";
3224  OS << "strlen(";
3225  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3226  OS << ") - 1";
3227
3228  Diag(SL, diag::note_strncat_wrong_size)
3229    << FixItHint::CreateReplacement(SR, OS.str());
3230}
3231
3232//===--- CHECK: Return Address of Stack Variable --------------------------===//
3233
3234static Expr *EvalVal(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars,
3235                     Decl *ParentDecl);
3236static Expr *EvalAddr(Expr* E, SmallVectorImpl<DeclRefExpr *> &refVars,
3237                      Decl *ParentDecl);
3238
3239/// CheckReturnStackAddr - Check if a return statement returns the address
3240///   of a stack variable.
3241void
3242Sema::CheckReturnStackAddr(Expr *RetValExp, QualType lhsType,
3243                           SourceLocation ReturnLoc) {
3244
3245  Expr *stackE = 0;
3246  SmallVector<DeclRefExpr *, 8> refVars;
3247
3248  // Perform checking for returned stack addresses, local blocks,
3249  // label addresses or references to temporaries.
3250  if (lhsType->isPointerType() ||
3251      (!getLangOpts().ObjCAutoRefCount && lhsType->isBlockPointerType())) {
3252    stackE = EvalAddr(RetValExp, refVars, /*ParentDecl=*/0);
3253  } else if (lhsType->isReferenceType()) {
3254    stackE = EvalVal(RetValExp, refVars, /*ParentDecl=*/0);
3255  }
3256
3257  if (stackE == 0)
3258    return; // Nothing suspicious was found.
3259
3260  SourceLocation diagLoc;
3261  SourceRange diagRange;
3262  if (refVars.empty()) {
3263    diagLoc = stackE->getLocStart();
3264    diagRange = stackE->getSourceRange();
3265  } else {
3266    // We followed through a reference variable. 'stackE' contains the
3267    // problematic expression but we will warn at the return statement pointing
3268    // at the reference variable. We will later display the "trail" of
3269    // reference variables using notes.
3270    diagLoc = refVars[0]->getLocStart();
3271    diagRange = refVars[0]->getSourceRange();
3272  }
3273
3274  if (DeclRefExpr *DR = dyn_cast<DeclRefExpr>(stackE)) { //address of local var.
3275    Diag(diagLoc, lhsType->isReferenceType() ? diag::warn_ret_stack_ref
3276                                             : diag::warn_ret_stack_addr)
3277     << DR->getDecl()->getDeclName() << diagRange;
3278  } else if (isa<BlockExpr>(stackE)) { // local block.
3279    Diag(diagLoc, diag::err_ret_local_block) << diagRange;
3280  } else if (isa<AddrLabelExpr>(stackE)) { // address of label.
3281    Diag(diagLoc, diag::warn_ret_addr_label) << diagRange;
3282  } else { // local temporary.
3283    Diag(diagLoc, lhsType->isReferenceType() ? diag::warn_ret_local_temp_ref
3284                                             : diag::warn_ret_local_temp_addr)
3285     << diagRange;
3286  }
3287
3288  // Display the "trail" of reference variables that we followed until we
3289  // found the problematic expression using notes.
3290  for (unsigned i = 0, e = refVars.size(); i != e; ++i) {
3291    VarDecl *VD = cast<VarDecl>(refVars[i]->getDecl());
3292    // If this var binds to another reference var, show the range of the next
3293    // var, otherwise the var binds to the problematic expression, in which case
3294    // show the range of the expression.
3295    SourceRange range = (i < e-1) ? refVars[i+1]->getSourceRange()
3296                                  : stackE->getSourceRange();
3297    Diag(VD->getLocation(), diag::note_ref_var_local_bind)
3298      << VD->getDeclName() << range;
3299  }
3300}
3301
3302/// EvalAddr - EvalAddr and EvalVal are mutually recursive functions that
3303///  check if the expression in a return statement evaluates to an address
3304///  to a location on the stack, a local block, an address of a label, or a
3305///  reference to local temporary. The recursion is used to traverse the
3306///  AST of the return expression, with recursion backtracking when we
3307///  encounter a subexpression that (1) clearly does not lead to one of the
3308///  above problematic expressions (2) is something we cannot determine leads to
3309///  a problematic expression based on such local checking.
3310///
3311///  Both EvalAddr and EvalVal follow through reference variables to evaluate
3312///  the expression that they point to. Such variables are added to the
3313///  'refVars' vector so that we know what the reference variable "trail" was.
3314///
3315///  EvalAddr processes expressions that are pointers that are used as
3316///  references (and not L-values).  EvalVal handles all other values.
3317///  At the base case of the recursion is a check for the above problematic
3318///  expressions.
3319///
3320///  This implementation handles:
3321///
3322///   * pointer-to-pointer casts
3323///   * implicit conversions from array references to pointers
3324///   * taking the address of fields
3325///   * arbitrary interplay between "&" and "*" operators
3326///   * pointer arithmetic from an address of a stack variable
3327///   * taking the address of an array element where the array is on the stack
3328static Expr *EvalAddr(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars,
3329                      Decl *ParentDecl) {
3330  if (E->isTypeDependent())
3331      return NULL;
3332
3333  // We should only be called for evaluating pointer expressions.
3334  assert((E->getType()->isAnyPointerType() ||
3335          E->getType()->isBlockPointerType() ||
3336          E->getType()->isObjCQualifiedIdType()) &&
3337         "EvalAddr only works on pointers");
3338
3339  E = E->IgnoreParens();
3340
3341  // Our "symbolic interpreter" is just a dispatch off the currently
3342  // viewed AST node.  We then recursively traverse the AST by calling
3343  // EvalAddr and EvalVal appropriately.
3344  switch (E->getStmtClass()) {
3345  case Stmt::DeclRefExprClass: {
3346    DeclRefExpr *DR = cast<DeclRefExpr>(E);
3347
3348    if (VarDecl *V = dyn_cast<VarDecl>(DR->getDecl()))
3349      // If this is a reference variable, follow through to the expression that
3350      // it points to.
3351      if (V->hasLocalStorage() &&
3352          V->getType()->isReferenceType() && V->hasInit()) {
3353        // Add the reference variable to the "trail".
3354        refVars.push_back(DR);
3355        return EvalAddr(V->getInit(), refVars, ParentDecl);
3356      }
3357
3358    return NULL;
3359  }
3360
3361  case Stmt::UnaryOperatorClass: {
3362    // The only unary operator that make sense to handle here
3363    // is AddrOf.  All others don't make sense as pointers.
3364    UnaryOperator *U = cast<UnaryOperator>(E);
3365
3366    if (U->getOpcode() == UO_AddrOf)
3367      return EvalVal(U->getSubExpr(), refVars, ParentDecl);
3368    else
3369      return NULL;
3370  }
3371
3372  case Stmt::BinaryOperatorClass: {
3373    // Handle pointer arithmetic.  All other binary operators are not valid
3374    // in this context.
3375    BinaryOperator *B = cast<BinaryOperator>(E);
3376    BinaryOperatorKind op = B->getOpcode();
3377
3378    if (op != BO_Add && op != BO_Sub)
3379      return NULL;
3380
3381    Expr *Base = B->getLHS();
3382
3383    // Determine which argument is the real pointer base.  It could be
3384    // the RHS argument instead of the LHS.
3385    if (!Base->getType()->isPointerType()) Base = B->getRHS();
3386
3387    assert (Base->getType()->isPointerType());
3388    return EvalAddr(Base, refVars, ParentDecl);
3389  }
3390
3391  // For conditional operators we need to see if either the LHS or RHS are
3392  // valid DeclRefExpr*s.  If one of them is valid, we return it.
3393  case Stmt::ConditionalOperatorClass: {
3394    ConditionalOperator *C = cast<ConditionalOperator>(E);
3395
3396    // Handle the GNU extension for missing LHS.
3397    if (Expr *lhsExpr = C->getLHS()) {
3398    // In C++, we can have a throw-expression, which has 'void' type.
3399      if (!lhsExpr->getType()->isVoidType())
3400        if (Expr* LHS = EvalAddr(lhsExpr, refVars, ParentDecl))
3401          return LHS;
3402    }
3403
3404    // In C++, we can have a throw-expression, which has 'void' type.
3405    if (C->getRHS()->getType()->isVoidType())
3406      return NULL;
3407
3408    return EvalAddr(C->getRHS(), refVars, ParentDecl);
3409  }
3410
3411  case Stmt::BlockExprClass:
3412    if (cast<BlockExpr>(E)->getBlockDecl()->hasCaptures())
3413      return E; // local block.
3414    return NULL;
3415
3416  case Stmt::AddrLabelExprClass:
3417    return E; // address of label.
3418
3419  case Stmt::ExprWithCleanupsClass:
3420    return EvalAddr(cast<ExprWithCleanups>(E)->getSubExpr(), refVars,
3421                    ParentDecl);
3422
3423  // For casts, we need to handle conversions from arrays to
3424  // pointer values, and pointer-to-pointer conversions.
3425  case Stmt::ImplicitCastExprClass:
3426  case Stmt::CStyleCastExprClass:
3427  case Stmt::CXXFunctionalCastExprClass:
3428  case Stmt::ObjCBridgedCastExprClass:
3429  case Stmt::CXXStaticCastExprClass:
3430  case Stmt::CXXDynamicCastExprClass:
3431  case Stmt::CXXConstCastExprClass:
3432  case Stmt::CXXReinterpretCastExprClass: {
3433    Expr* SubExpr = cast<CastExpr>(E)->getSubExpr();
3434    switch (cast<CastExpr>(E)->getCastKind()) {
3435    case CK_BitCast:
3436    case CK_LValueToRValue:
3437    case CK_NoOp:
3438    case CK_BaseToDerived:
3439    case CK_DerivedToBase:
3440    case CK_UncheckedDerivedToBase:
3441    case CK_Dynamic:
3442    case CK_CPointerToObjCPointerCast:
3443    case CK_BlockPointerToObjCPointerCast:
3444    case CK_AnyPointerToBlockPointerCast:
3445      return EvalAddr(SubExpr, refVars, ParentDecl);
3446
3447    case CK_ArrayToPointerDecay:
3448      return EvalVal(SubExpr, refVars, ParentDecl);
3449
3450    default:
3451      return 0;
3452    }
3453  }
3454
3455  case Stmt::MaterializeTemporaryExprClass:
3456    if (Expr *Result = EvalAddr(
3457                         cast<MaterializeTemporaryExpr>(E)->GetTemporaryExpr(),
3458                                refVars, ParentDecl))
3459      return Result;
3460
3461    return E;
3462
3463  // Everything else: we simply don't reason about them.
3464  default:
3465    return NULL;
3466  }
3467}
3468
3469
3470///  EvalVal - This function is complements EvalAddr in the mutual recursion.
3471///   See the comments for EvalAddr for more details.
3472static Expr *EvalVal(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars,
3473                     Decl *ParentDecl) {
3474do {
3475  // We should only be called for evaluating non-pointer expressions, or
3476  // expressions with a pointer type that are not used as references but instead
3477  // are l-values (e.g., DeclRefExpr with a pointer type).
3478
3479  // Our "symbolic interpreter" is just a dispatch off the currently
3480  // viewed AST node.  We then recursively traverse the AST by calling
3481  // EvalAddr and EvalVal appropriately.
3482
3483  E = E->IgnoreParens();
3484  switch (E->getStmtClass()) {
3485  case Stmt::ImplicitCastExprClass: {
3486    ImplicitCastExpr *IE = cast<ImplicitCastExpr>(E);
3487    if (IE->getValueKind() == VK_LValue) {
3488      E = IE->getSubExpr();
3489      continue;
3490    }
3491    return NULL;
3492  }
3493
3494  case Stmt::ExprWithCleanupsClass:
3495    return EvalVal(cast<ExprWithCleanups>(E)->getSubExpr(), refVars,ParentDecl);
3496
3497  case Stmt::DeclRefExprClass: {
3498    // When we hit a DeclRefExpr we are looking at code that refers to a
3499    // variable's name. If it's not a reference variable we check if it has
3500    // local storage within the function, and if so, return the expression.
3501    DeclRefExpr *DR = cast<DeclRefExpr>(E);
3502
3503    if (VarDecl *V = dyn_cast<VarDecl>(DR->getDecl())) {
3504      // Check if it refers to itself, e.g. "int& i = i;".
3505      if (V == ParentDecl)
3506        return DR;
3507
3508      if (V->hasLocalStorage()) {
3509        if (!V->getType()->isReferenceType())
3510          return DR;
3511
3512        // Reference variable, follow through to the expression that
3513        // it points to.
3514        if (V->hasInit()) {
3515          // Add the reference variable to the "trail".
3516          refVars.push_back(DR);
3517          return EvalVal(V->getInit(), refVars, V);
3518        }
3519      }
3520    }
3521
3522    return NULL;
3523  }
3524
3525  case Stmt::UnaryOperatorClass: {
3526    // The only unary operator that make sense to handle here
3527    // is Deref.  All others don't resolve to a "name."  This includes
3528    // handling all sorts of rvalues passed to a unary operator.
3529    UnaryOperator *U = cast<UnaryOperator>(E);
3530
3531    if (U->getOpcode() == UO_Deref)
3532      return EvalAddr(U->getSubExpr(), refVars, ParentDecl);
3533
3534    return NULL;
3535  }
3536
3537  case Stmt::ArraySubscriptExprClass: {
3538    // Array subscripts are potential references to data on the stack.  We
3539    // retrieve the DeclRefExpr* for the array variable if it indeed
3540    // has local storage.
3541    return EvalAddr(cast<ArraySubscriptExpr>(E)->getBase(), refVars,ParentDecl);
3542  }
3543
3544  case Stmt::ConditionalOperatorClass: {
3545    // For conditional operators we need to see if either the LHS or RHS are
3546    // non-NULL Expr's.  If one is non-NULL, we return it.
3547    ConditionalOperator *C = cast<ConditionalOperator>(E);
3548
3549    // Handle the GNU extension for missing LHS.
3550    if (Expr *lhsExpr = C->getLHS())
3551      if (Expr *LHS = EvalVal(lhsExpr, refVars, ParentDecl))
3552        return LHS;
3553
3554    return EvalVal(C->getRHS(), refVars, ParentDecl);
3555  }
3556
3557  // Accesses to members are potential references to data on the stack.
3558  case Stmt::MemberExprClass: {
3559    MemberExpr *M = cast<MemberExpr>(E);
3560
3561    // Check for indirect access.  We only want direct field accesses.
3562    if (M->isArrow())
3563      return NULL;
3564
3565    // Check whether the member type is itself a reference, in which case
3566    // we're not going to refer to the member, but to what the member refers to.
3567    if (M->getMemberDecl()->getType()->isReferenceType())
3568      return NULL;
3569
3570    return EvalVal(M->getBase(), refVars, ParentDecl);
3571  }
3572
3573  case Stmt::MaterializeTemporaryExprClass:
3574    if (Expr *Result = EvalVal(
3575                          cast<MaterializeTemporaryExpr>(E)->GetTemporaryExpr(),
3576                               refVars, ParentDecl))
3577      return Result;
3578
3579    return E;
3580
3581  default:
3582    // Check that we don't return or take the address of a reference to a
3583    // temporary. This is only useful in C++.
3584    if (!E->isTypeDependent() && E->isRValue())
3585      return E;
3586
3587    // Everything else: we simply don't reason about them.
3588    return NULL;
3589  }
3590} while (true);
3591}
3592
3593//===--- CHECK: Floating-Point comparisons (-Wfloat-equal) ---------------===//
3594
3595/// Check for comparisons of floating point operands using != and ==.
3596/// Issue a warning if these are no self-comparisons, as they are not likely
3597/// to do what the programmer intended.
3598void Sema::CheckFloatComparison(SourceLocation Loc, Expr* LHS, Expr *RHS) {
3599  bool EmitWarning = true;
3600
3601  Expr* LeftExprSansParen = LHS->IgnoreParenImpCasts();
3602  Expr* RightExprSansParen = RHS->IgnoreParenImpCasts();
3603
3604  // Special case: check for x == x (which is OK).
3605  // Do not emit warnings for such cases.
3606  if (DeclRefExpr* DRL = dyn_cast<DeclRefExpr>(LeftExprSansParen))
3607    if (DeclRefExpr* DRR = dyn_cast<DeclRefExpr>(RightExprSansParen))
3608      if (DRL->getDecl() == DRR->getDecl())
3609        EmitWarning = false;
3610
3611
3612  // Special case: check for comparisons against literals that can be exactly
3613  //  represented by APFloat.  In such cases, do not emit a warning.  This
3614  //  is a heuristic: often comparison against such literals are used to
3615  //  detect if a value in a variable has not changed.  This clearly can
3616  //  lead to false negatives.
3617  if (EmitWarning) {
3618    if (FloatingLiteral* FLL = dyn_cast<FloatingLiteral>(LeftExprSansParen)) {
3619      if (FLL->isExact())
3620        EmitWarning = false;
3621    } else
3622      if (FloatingLiteral* FLR = dyn_cast<FloatingLiteral>(RightExprSansParen)){
3623        if (FLR->isExact())
3624          EmitWarning = false;
3625    }
3626  }
3627
3628  // Check for comparisons with builtin types.
3629  if (EmitWarning)
3630    if (CallExpr* CL = dyn_cast<CallExpr>(LeftExprSansParen))
3631      if (CL->isBuiltinCall())
3632        EmitWarning = false;
3633
3634  if (EmitWarning)
3635    if (CallExpr* CR = dyn_cast<CallExpr>(RightExprSansParen))
3636      if (CR->isBuiltinCall())
3637        EmitWarning = false;
3638
3639  // Emit the diagnostic.
3640  if (EmitWarning)
3641    Diag(Loc, diag::warn_floatingpoint_eq)
3642      << LHS->getSourceRange() << RHS->getSourceRange();
3643}
3644
3645//===--- CHECK: Integer mixed-sign comparisons (-Wsign-compare) --------===//
3646//===--- CHECK: Lossy implicit conversions (-Wconversion) --------------===//
3647
3648namespace {
3649
3650/// Structure recording the 'active' range of an integer-valued
3651/// expression.
3652struct IntRange {
3653  /// The number of bits active in the int.
3654  unsigned Width;
3655
3656  /// True if the int is known not to have negative values.
3657  bool NonNegative;
3658
3659  IntRange(unsigned Width, bool NonNegative)
3660    : Width(Width), NonNegative(NonNegative)
3661  {}
3662
3663  /// Returns the range of the bool type.
3664  static IntRange forBoolType() {
3665    return IntRange(1, true);
3666  }
3667
3668  /// Returns the range of an opaque value of the given integral type.
3669  static IntRange forValueOfType(ASTContext &C, QualType T) {
3670    return forValueOfCanonicalType(C,
3671                          T->getCanonicalTypeInternal().getTypePtr());
3672  }
3673
3674  /// Returns the range of an opaque value of a canonical integral type.
3675  static IntRange forValueOfCanonicalType(ASTContext &C, const Type *T) {
3676    assert(T->isCanonicalUnqualified());
3677
3678    if (const VectorType *VT = dyn_cast<VectorType>(T))
3679      T = VT->getElementType().getTypePtr();
3680    if (const ComplexType *CT = dyn_cast<ComplexType>(T))
3681      T = CT->getElementType().getTypePtr();
3682
3683    // For enum types, use the known bit width of the enumerators.
3684    if (const EnumType *ET = dyn_cast<EnumType>(T)) {
3685      EnumDecl *Enum = ET->getDecl();
3686      if (!Enum->isCompleteDefinition())
3687        return IntRange(C.getIntWidth(QualType(T, 0)), false);
3688
3689      unsigned NumPositive = Enum->getNumPositiveBits();
3690      unsigned NumNegative = Enum->getNumNegativeBits();
3691
3692      return IntRange(std::max(NumPositive, NumNegative), NumNegative == 0);
3693    }
3694
3695    const BuiltinType *BT = cast<BuiltinType>(T);
3696    assert(BT->isInteger());
3697
3698    return IntRange(C.getIntWidth(QualType(T, 0)), BT->isUnsignedInteger());
3699  }
3700
3701  /// Returns the "target" range of a canonical integral type, i.e.
3702  /// the range of values expressible in the type.
3703  ///
3704  /// This matches forValueOfCanonicalType except that enums have the
3705  /// full range of their type, not the range of their enumerators.
3706  static IntRange forTargetOfCanonicalType(ASTContext &C, const Type *T) {
3707    assert(T->isCanonicalUnqualified());
3708
3709    if (const VectorType *VT = dyn_cast<VectorType>(T))
3710      T = VT->getElementType().getTypePtr();
3711    if (const ComplexType *CT = dyn_cast<ComplexType>(T))
3712      T = CT->getElementType().getTypePtr();
3713    if (const EnumType *ET = dyn_cast<EnumType>(T))
3714      T = C.getCanonicalType(ET->getDecl()->getIntegerType()).getTypePtr();
3715
3716    const BuiltinType *BT = cast<BuiltinType>(T);
3717    assert(BT->isInteger());
3718
3719    return IntRange(C.getIntWidth(QualType(T, 0)), BT->isUnsignedInteger());
3720  }
3721
3722  /// Returns the supremum of two ranges: i.e. their conservative merge.
3723  static IntRange join(IntRange L, IntRange R) {
3724    return IntRange(std::max(L.Width, R.Width),
3725                    L.NonNegative && R.NonNegative);
3726  }
3727
3728  /// Returns the infinum of two ranges: i.e. their aggressive merge.
3729  static IntRange meet(IntRange L, IntRange R) {
3730    return IntRange(std::min(L.Width, R.Width),
3731                    L.NonNegative || R.NonNegative);
3732  }
3733};
3734
3735static IntRange GetValueRange(ASTContext &C, llvm::APSInt &value,
3736                              unsigned MaxWidth) {
3737  if (value.isSigned() && value.isNegative())
3738    return IntRange(value.getMinSignedBits(), false);
3739
3740  if (value.getBitWidth() > MaxWidth)
3741    value = value.trunc(MaxWidth);
3742
3743  // isNonNegative() just checks the sign bit without considering
3744  // signedness.
3745  return IntRange(value.getActiveBits(), true);
3746}
3747
3748static IntRange GetValueRange(ASTContext &C, APValue &result, QualType Ty,
3749                              unsigned MaxWidth) {
3750  if (result.isInt())
3751    return GetValueRange(C, result.getInt(), MaxWidth);
3752
3753  if (result.isVector()) {
3754    IntRange R = GetValueRange(C, result.getVectorElt(0), Ty, MaxWidth);
3755    for (unsigned i = 1, e = result.getVectorLength(); i != e; ++i) {
3756      IntRange El = GetValueRange(C, result.getVectorElt(i), Ty, MaxWidth);
3757      R = IntRange::join(R, El);
3758    }
3759    return R;
3760  }
3761
3762  if (result.isComplexInt()) {
3763    IntRange R = GetValueRange(C, result.getComplexIntReal(), MaxWidth);
3764    IntRange I = GetValueRange(C, result.getComplexIntImag(), MaxWidth);
3765    return IntRange::join(R, I);
3766  }
3767
3768  // This can happen with lossless casts to intptr_t of "based" lvalues.
3769  // Assume it might use arbitrary bits.
3770  // FIXME: The only reason we need to pass the type in here is to get
3771  // the sign right on this one case.  It would be nice if APValue
3772  // preserved this.
3773  assert(result.isLValue() || result.isAddrLabelDiff());
3774  return IntRange(MaxWidth, Ty->isUnsignedIntegerOrEnumerationType());
3775}
3776
3777/// Pseudo-evaluate the given integer expression, estimating the
3778/// range of values it might take.
3779///
3780/// \param MaxWidth - the width to which the value will be truncated
3781static IntRange GetExprRange(ASTContext &C, Expr *E, unsigned MaxWidth) {
3782  E = E->IgnoreParens();
3783
3784  // Try a full evaluation first.
3785  Expr::EvalResult result;
3786  if (E->EvaluateAsRValue(result, C))
3787    return GetValueRange(C, result.Val, E->getType(), MaxWidth);
3788
3789  // I think we only want to look through implicit casts here; if the
3790  // user has an explicit widening cast, we should treat the value as
3791  // being of the new, wider type.
3792  if (ImplicitCastExpr *CE = dyn_cast<ImplicitCastExpr>(E)) {
3793    if (CE->getCastKind() == CK_NoOp || CE->getCastKind() == CK_LValueToRValue)
3794      return GetExprRange(C, CE->getSubExpr(), MaxWidth);
3795
3796    IntRange OutputTypeRange = IntRange::forValueOfType(C, CE->getType());
3797
3798    bool isIntegerCast = (CE->getCastKind() == CK_IntegralCast);
3799
3800    // Assume that non-integer casts can span the full range of the type.
3801    if (!isIntegerCast)
3802      return OutputTypeRange;
3803
3804    IntRange SubRange
3805      = GetExprRange(C, CE->getSubExpr(),
3806                     std::min(MaxWidth, OutputTypeRange.Width));
3807
3808    // Bail out if the subexpr's range is as wide as the cast type.
3809    if (SubRange.Width >= OutputTypeRange.Width)
3810      return OutputTypeRange;
3811
3812    // Otherwise, we take the smaller width, and we're non-negative if
3813    // either the output type or the subexpr is.
3814    return IntRange(SubRange.Width,
3815                    SubRange.NonNegative || OutputTypeRange.NonNegative);
3816  }
3817
3818  if (ConditionalOperator *CO = dyn_cast<ConditionalOperator>(E)) {
3819    // If we can fold the condition, just take that operand.
3820    bool CondResult;
3821    if (CO->getCond()->EvaluateAsBooleanCondition(CondResult, C))
3822      return GetExprRange(C, CondResult ? CO->getTrueExpr()
3823                                        : CO->getFalseExpr(),
3824                          MaxWidth);
3825
3826    // Otherwise, conservatively merge.
3827    IntRange L = GetExprRange(C, CO->getTrueExpr(), MaxWidth);
3828    IntRange R = GetExprRange(C, CO->getFalseExpr(), MaxWidth);
3829    return IntRange::join(L, R);
3830  }
3831
3832  if (BinaryOperator *BO = dyn_cast<BinaryOperator>(E)) {
3833    switch (BO->getOpcode()) {
3834
3835    // Boolean-valued operations are single-bit and positive.
3836    case BO_LAnd:
3837    case BO_LOr:
3838    case BO_LT:
3839    case BO_GT:
3840    case BO_LE:
3841    case BO_GE:
3842    case BO_EQ:
3843    case BO_NE:
3844      return IntRange::forBoolType();
3845
3846    // The type of the assignments is the type of the LHS, so the RHS
3847    // is not necessarily the same type.
3848    case BO_MulAssign:
3849    case BO_DivAssign:
3850    case BO_RemAssign:
3851    case BO_AddAssign:
3852    case BO_SubAssign:
3853    case BO_XorAssign:
3854    case BO_OrAssign:
3855      // TODO: bitfields?
3856      return IntRange::forValueOfType(C, E->getType());
3857
3858    // Simple assignments just pass through the RHS, which will have
3859    // been coerced to the LHS type.
3860    case BO_Assign:
3861      // TODO: bitfields?
3862      return GetExprRange(C, BO->getRHS(), MaxWidth);
3863
3864    // Operations with opaque sources are black-listed.
3865    case BO_PtrMemD:
3866    case BO_PtrMemI:
3867      return IntRange::forValueOfType(C, E->getType());
3868
3869    // Bitwise-and uses the *infinum* of the two source ranges.
3870    case BO_And:
3871    case BO_AndAssign:
3872      return IntRange::meet(GetExprRange(C, BO->getLHS(), MaxWidth),
3873                            GetExprRange(C, BO->getRHS(), MaxWidth));
3874
3875    // Left shift gets black-listed based on a judgement call.
3876    case BO_Shl:
3877      // ...except that we want to treat '1 << (blah)' as logically
3878      // positive.  It's an important idiom.
3879      if (IntegerLiteral *I
3880            = dyn_cast<IntegerLiteral>(BO->getLHS()->IgnoreParenCasts())) {
3881        if (I->getValue() == 1) {
3882          IntRange R = IntRange::forValueOfType(C, E->getType());
3883          return IntRange(R.Width, /*NonNegative*/ true);
3884        }
3885      }
3886      // fallthrough
3887
3888    case BO_ShlAssign:
3889      return IntRange::forValueOfType(C, E->getType());
3890
3891    // Right shift by a constant can narrow its left argument.
3892    case BO_Shr:
3893    case BO_ShrAssign: {
3894      IntRange L = GetExprRange(C, BO->getLHS(), MaxWidth);
3895
3896      // If the shift amount is a positive constant, drop the width by
3897      // that much.
3898      llvm::APSInt shift;
3899      if (BO->getRHS()->isIntegerConstantExpr(shift, C) &&
3900          shift.isNonNegative()) {
3901        unsigned zext = shift.getZExtValue();
3902        if (zext >= L.Width)
3903          L.Width = (L.NonNegative ? 0 : 1);
3904        else
3905          L.Width -= zext;
3906      }
3907
3908      return L;
3909    }
3910
3911    // Comma acts as its right operand.
3912    case BO_Comma:
3913      return GetExprRange(C, BO->getRHS(), MaxWidth);
3914
3915    // Black-list pointer subtractions.
3916    case BO_Sub:
3917      if (BO->getLHS()->getType()->isPointerType())
3918        return IntRange::forValueOfType(C, E->getType());
3919      break;
3920
3921    // The width of a division result is mostly determined by the size
3922    // of the LHS.
3923    case BO_Div: {
3924      // Don't 'pre-truncate' the operands.
3925      unsigned opWidth = C.getIntWidth(E->getType());
3926      IntRange L = GetExprRange(C, BO->getLHS(), opWidth);
3927
3928      // If the divisor is constant, use that.
3929      llvm::APSInt divisor;
3930      if (BO->getRHS()->isIntegerConstantExpr(divisor, C)) {
3931        unsigned log2 = divisor.logBase2(); // floor(log_2(divisor))
3932        if (log2 >= L.Width)
3933          L.Width = (L.NonNegative ? 0 : 1);
3934        else
3935          L.Width = std::min(L.Width - log2, MaxWidth);
3936        return L;
3937      }
3938
3939      // Otherwise, just use the LHS's width.
3940      IntRange R = GetExprRange(C, BO->getRHS(), opWidth);
3941      return IntRange(L.Width, L.NonNegative && R.NonNegative);
3942    }
3943
3944    // The result of a remainder can't be larger than the result of
3945    // either side.
3946    case BO_Rem: {
3947      // Don't 'pre-truncate' the operands.
3948      unsigned opWidth = C.getIntWidth(E->getType());
3949      IntRange L = GetExprRange(C, BO->getLHS(), opWidth);
3950      IntRange R = GetExprRange(C, BO->getRHS(), opWidth);
3951
3952      IntRange meet = IntRange::meet(L, R);
3953      meet.Width = std::min(meet.Width, MaxWidth);
3954      return meet;
3955    }
3956
3957    // The default behavior is okay for these.
3958    case BO_Mul:
3959    case BO_Add:
3960    case BO_Xor:
3961    case BO_Or:
3962      break;
3963    }
3964
3965    // The default case is to treat the operation as if it were closed
3966    // on the narrowest type that encompasses both operands.
3967    IntRange L = GetExprRange(C, BO->getLHS(), MaxWidth);
3968    IntRange R = GetExprRange(C, BO->getRHS(), MaxWidth);
3969    return IntRange::join(L, R);
3970  }
3971
3972  if (UnaryOperator *UO = dyn_cast<UnaryOperator>(E)) {
3973    switch (UO->getOpcode()) {
3974    // Boolean-valued operations are white-listed.
3975    case UO_LNot:
3976      return IntRange::forBoolType();
3977
3978    // Operations with opaque sources are black-listed.
3979    case UO_Deref:
3980    case UO_AddrOf: // should be impossible
3981      return IntRange::forValueOfType(C, E->getType());
3982
3983    default:
3984      return GetExprRange(C, UO->getSubExpr(), MaxWidth);
3985    }
3986  }
3987
3988  if (dyn_cast<OffsetOfExpr>(E)) {
3989    IntRange::forValueOfType(C, E->getType());
3990  }
3991
3992  if (FieldDecl *BitField = E->getBitField())
3993    return IntRange(BitField->getBitWidthValue(C),
3994                    BitField->getType()->isUnsignedIntegerOrEnumerationType());
3995
3996  return IntRange::forValueOfType(C, E->getType());
3997}
3998
3999static IntRange GetExprRange(ASTContext &C, Expr *E) {
4000  return GetExprRange(C, E, C.getIntWidth(E->getType()));
4001}
4002
4003/// Checks whether the given value, which currently has the given
4004/// source semantics, has the same value when coerced through the
4005/// target semantics.
4006static bool IsSameFloatAfterCast(const llvm::APFloat &value,
4007                                 const llvm::fltSemantics &Src,
4008                                 const llvm::fltSemantics &Tgt) {
4009  llvm::APFloat truncated = value;
4010
4011  bool ignored;
4012  truncated.convert(Src, llvm::APFloat::rmNearestTiesToEven, &ignored);
4013  truncated.convert(Tgt, llvm::APFloat::rmNearestTiesToEven, &ignored);
4014
4015  return truncated.bitwiseIsEqual(value);
4016}
4017
4018/// Checks whether the given value, which currently has the given
4019/// source semantics, has the same value when coerced through the
4020/// target semantics.
4021///
4022/// The value might be a vector of floats (or a complex number).
4023static bool IsSameFloatAfterCast(const APValue &value,
4024                                 const llvm::fltSemantics &Src,
4025                                 const llvm::fltSemantics &Tgt) {
4026  if (value.isFloat())
4027    return IsSameFloatAfterCast(value.getFloat(), Src, Tgt);
4028
4029  if (value.isVector()) {
4030    for (unsigned i = 0, e = value.getVectorLength(); i != e; ++i)
4031      if (!IsSameFloatAfterCast(value.getVectorElt(i), Src, Tgt))
4032        return false;
4033    return true;
4034  }
4035
4036  assert(value.isComplexFloat());
4037  return (IsSameFloatAfterCast(value.getComplexFloatReal(), Src, Tgt) &&
4038          IsSameFloatAfterCast(value.getComplexFloatImag(), Src, Tgt));
4039}
4040
4041static void AnalyzeImplicitConversions(Sema &S, Expr *E, SourceLocation CC);
4042
4043static bool IsZero(Sema &S, Expr *E) {
4044  // Suppress cases where we are comparing against an enum constant.
4045  if (const DeclRefExpr *DR =
4046      dyn_cast<DeclRefExpr>(E->IgnoreParenImpCasts()))
4047    if (isa<EnumConstantDecl>(DR->getDecl()))
4048      return false;
4049
4050  // Suppress cases where the '0' value is expanded from a macro.
4051  if (E->getLocStart().isMacroID())
4052    return false;
4053
4054  llvm::APSInt Value;
4055  return E->isIntegerConstantExpr(Value, S.Context) && Value == 0;
4056}
4057
4058static bool HasEnumType(Expr *E) {
4059  // Strip off implicit integral promotions.
4060  while (ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(E)) {
4061    if (ICE->getCastKind() != CK_IntegralCast &&
4062        ICE->getCastKind() != CK_NoOp)
4063      break;
4064    E = ICE->getSubExpr();
4065  }
4066
4067  return E->getType()->isEnumeralType();
4068}
4069
4070static void CheckTrivialUnsignedComparison(Sema &S, BinaryOperator *E) {
4071  BinaryOperatorKind op = E->getOpcode();
4072  if (E->isValueDependent())
4073    return;
4074
4075  if (op == BO_LT && IsZero(S, E->getRHS())) {
4076    S.Diag(E->getOperatorLoc(), diag::warn_lunsigned_always_true_comparison)
4077      << "< 0" << "false" << HasEnumType(E->getLHS())
4078      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4079  } else if (op == BO_GE && IsZero(S, E->getRHS())) {
4080    S.Diag(E->getOperatorLoc(), diag::warn_lunsigned_always_true_comparison)
4081      << ">= 0" << "true" << HasEnumType(E->getLHS())
4082      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4083  } else if (op == BO_GT && IsZero(S, E->getLHS())) {
4084    S.Diag(E->getOperatorLoc(), diag::warn_runsigned_always_true_comparison)
4085      << "0 >" << "false" << HasEnumType(E->getRHS())
4086      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4087  } else if (op == BO_LE && IsZero(S, E->getLHS())) {
4088    S.Diag(E->getOperatorLoc(), diag::warn_runsigned_always_true_comparison)
4089      << "0 <=" << "true" << HasEnumType(E->getRHS())
4090      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4091  }
4092}
4093
4094/// Analyze the operands of the given comparison.  Implements the
4095/// fallback case from AnalyzeComparison.
4096static void AnalyzeImpConvsInComparison(Sema &S, BinaryOperator *E) {
4097  AnalyzeImplicitConversions(S, E->getLHS(), E->getOperatorLoc());
4098  AnalyzeImplicitConversions(S, E->getRHS(), E->getOperatorLoc());
4099}
4100
4101/// \brief Implements -Wsign-compare.
4102///
4103/// \param E the binary operator to check for warnings
4104static void AnalyzeComparison(Sema &S, BinaryOperator *E) {
4105  // The type the comparison is being performed in.
4106  QualType T = E->getLHS()->getType();
4107  assert(S.Context.hasSameUnqualifiedType(T, E->getRHS()->getType())
4108         && "comparison with mismatched types");
4109
4110  // We don't do anything special if this isn't an unsigned integral
4111  // comparison:  we're only interested in integral comparisons, and
4112  // signed comparisons only happen in cases we don't care to warn about.
4113  //
4114  // We also don't care about value-dependent expressions or expressions
4115  // whose result is a constant.
4116  if (!T->hasUnsignedIntegerRepresentation()
4117      || E->isValueDependent() || E->isIntegerConstantExpr(S.Context))
4118    return AnalyzeImpConvsInComparison(S, E);
4119
4120  Expr *LHS = E->getLHS()->IgnoreParenImpCasts();
4121  Expr *RHS = E->getRHS()->IgnoreParenImpCasts();
4122
4123  // Check to see if one of the (unmodified) operands is of different
4124  // signedness.
4125  Expr *signedOperand, *unsignedOperand;
4126  if (LHS->getType()->hasSignedIntegerRepresentation()) {
4127    assert(!RHS->getType()->hasSignedIntegerRepresentation() &&
4128           "unsigned comparison between two signed integer expressions?");
4129    signedOperand = LHS;
4130    unsignedOperand = RHS;
4131  } else if (RHS->getType()->hasSignedIntegerRepresentation()) {
4132    signedOperand = RHS;
4133    unsignedOperand = LHS;
4134  } else {
4135    CheckTrivialUnsignedComparison(S, E);
4136    return AnalyzeImpConvsInComparison(S, E);
4137  }
4138
4139  // Otherwise, calculate the effective range of the signed operand.
4140  IntRange signedRange = GetExprRange(S.Context, signedOperand);
4141
4142  // Go ahead and analyze implicit conversions in the operands.  Note
4143  // that we skip the implicit conversions on both sides.
4144  AnalyzeImplicitConversions(S, LHS, E->getOperatorLoc());
4145  AnalyzeImplicitConversions(S, RHS, E->getOperatorLoc());
4146
4147  // If the signed range is non-negative, -Wsign-compare won't fire,
4148  // but we should still check for comparisons which are always true
4149  // or false.
4150  if (signedRange.NonNegative)
4151    return CheckTrivialUnsignedComparison(S, E);
4152
4153  // For (in)equality comparisons, if the unsigned operand is a
4154  // constant which cannot collide with a overflowed signed operand,
4155  // then reinterpreting the signed operand as unsigned will not
4156  // change the result of the comparison.
4157  if (E->isEqualityOp()) {
4158    unsigned comparisonWidth = S.Context.getIntWidth(T);
4159    IntRange unsignedRange = GetExprRange(S.Context, unsignedOperand);
4160
4161    // We should never be unable to prove that the unsigned operand is
4162    // non-negative.
4163    assert(unsignedRange.NonNegative && "unsigned range includes negative?");
4164
4165    if (unsignedRange.Width < comparisonWidth)
4166      return;
4167  }
4168
4169  S.DiagRuntimeBehavior(E->getOperatorLoc(), E,
4170    S.PDiag(diag::warn_mixed_sign_comparison)
4171      << LHS->getType() << RHS->getType()
4172      << LHS->getSourceRange() << RHS->getSourceRange());
4173}
4174
4175/// Analyzes an attempt to assign the given value to a bitfield.
4176///
4177/// Returns true if there was something fishy about the attempt.
4178static bool AnalyzeBitFieldAssignment(Sema &S, FieldDecl *Bitfield, Expr *Init,
4179                                      SourceLocation InitLoc) {
4180  assert(Bitfield->isBitField());
4181  if (Bitfield->isInvalidDecl())
4182    return false;
4183
4184  // White-list bool bitfields.
4185  if (Bitfield->getType()->isBooleanType())
4186    return false;
4187
4188  // Ignore value- or type-dependent expressions.
4189  if (Bitfield->getBitWidth()->isValueDependent() ||
4190      Bitfield->getBitWidth()->isTypeDependent() ||
4191      Init->isValueDependent() ||
4192      Init->isTypeDependent())
4193    return false;
4194
4195  Expr *OriginalInit = Init->IgnoreParenImpCasts();
4196
4197  llvm::APSInt Value;
4198  if (!OriginalInit->EvaluateAsInt(Value, S.Context, Expr::SE_AllowSideEffects))
4199    return false;
4200
4201  unsigned OriginalWidth = Value.getBitWidth();
4202  unsigned FieldWidth = Bitfield->getBitWidthValue(S.Context);
4203
4204  if (OriginalWidth <= FieldWidth)
4205    return false;
4206
4207  // Compute the value which the bitfield will contain.
4208  llvm::APSInt TruncatedValue = Value.trunc(FieldWidth);
4209  TruncatedValue.setIsSigned(Bitfield->getType()->isSignedIntegerType());
4210
4211  // Check whether the stored value is equal to the original value.
4212  TruncatedValue = TruncatedValue.extend(OriginalWidth);
4213  if (Value == TruncatedValue)
4214    return false;
4215
4216  // Special-case bitfields of width 1: booleans are naturally 0/1, and
4217  // therefore don't strictly fit into a signed bitfield of width 1.
4218  if (FieldWidth == 1 && Value == 1)
4219    return false;
4220
4221  std::string PrettyValue = Value.toString(10);
4222  std::string PrettyTrunc = TruncatedValue.toString(10);
4223
4224  S.Diag(InitLoc, diag::warn_impcast_bitfield_precision_constant)
4225    << PrettyValue << PrettyTrunc << OriginalInit->getType()
4226    << Init->getSourceRange();
4227
4228  return true;
4229}
4230
4231/// Analyze the given simple or compound assignment for warning-worthy
4232/// operations.
4233static void AnalyzeAssignment(Sema &S, BinaryOperator *E) {
4234  // Just recurse on the LHS.
4235  AnalyzeImplicitConversions(S, E->getLHS(), E->getOperatorLoc());
4236
4237  // We want to recurse on the RHS as normal unless we're assigning to
4238  // a bitfield.
4239  if (FieldDecl *Bitfield = E->getLHS()->getBitField()) {
4240    if (AnalyzeBitFieldAssignment(S, Bitfield, E->getRHS(),
4241                                  E->getOperatorLoc())) {
4242      // Recurse, ignoring any implicit conversions on the RHS.
4243      return AnalyzeImplicitConversions(S, E->getRHS()->IgnoreParenImpCasts(),
4244                                        E->getOperatorLoc());
4245    }
4246  }
4247
4248  AnalyzeImplicitConversions(S, E->getRHS(), E->getOperatorLoc());
4249}
4250
4251/// Diagnose an implicit cast;  purely a helper for CheckImplicitConversion.
4252static void DiagnoseImpCast(Sema &S, Expr *E, QualType SourceType, QualType T,
4253                            SourceLocation CContext, unsigned diag,
4254                            bool pruneControlFlow = false) {
4255  if (pruneControlFlow) {
4256    S.DiagRuntimeBehavior(E->getExprLoc(), E,
4257                          S.PDiag(diag)
4258                            << SourceType << T << E->getSourceRange()
4259                            << SourceRange(CContext));
4260    return;
4261  }
4262  S.Diag(E->getExprLoc(), diag)
4263    << SourceType << T << E->getSourceRange() << SourceRange(CContext);
4264}
4265
4266/// Diagnose an implicit cast;  purely a helper for CheckImplicitConversion.
4267static void DiagnoseImpCast(Sema &S, Expr *E, QualType T,
4268                            SourceLocation CContext, unsigned diag,
4269                            bool pruneControlFlow = false) {
4270  DiagnoseImpCast(S, E, E->getType(), T, CContext, diag, pruneControlFlow);
4271}
4272
4273/// Diagnose an implicit cast from a literal expression. Does not warn when the
4274/// cast wouldn't lose information.
4275void DiagnoseFloatingLiteralImpCast(Sema &S, FloatingLiteral *FL, QualType T,
4276                                    SourceLocation CContext) {
4277  // Try to convert the literal exactly to an integer. If we can, don't warn.
4278  bool isExact = false;
4279  const llvm::APFloat &Value = FL->getValue();
4280  llvm::APSInt IntegerValue(S.Context.getIntWidth(T),
4281                            T->hasUnsignedIntegerRepresentation());
4282  if (Value.convertToInteger(IntegerValue,
4283                             llvm::APFloat::rmTowardZero, &isExact)
4284      == llvm::APFloat::opOK && isExact)
4285    return;
4286
4287  SmallString<16> PrettySourceValue;
4288  Value.toString(PrettySourceValue);
4289  SmallString<16> PrettyTargetValue;
4290  if (T->isSpecificBuiltinType(BuiltinType::Bool))
4291    PrettyTargetValue = IntegerValue == 0 ? "false" : "true";
4292  else
4293    IntegerValue.toString(PrettyTargetValue);
4294
4295  S.Diag(FL->getExprLoc(), diag::warn_impcast_literal_float_to_integer)
4296    << FL->getType() << T.getUnqualifiedType() << PrettySourceValue
4297    << PrettyTargetValue << FL->getSourceRange() << SourceRange(CContext);
4298}
4299
4300std::string PrettyPrintInRange(const llvm::APSInt &Value, IntRange Range) {
4301  if (!Range.Width) return "0";
4302
4303  llvm::APSInt ValueInRange = Value;
4304  ValueInRange.setIsSigned(!Range.NonNegative);
4305  ValueInRange = ValueInRange.trunc(Range.Width);
4306  return ValueInRange.toString(10);
4307}
4308
4309void CheckImplicitConversion(Sema &S, Expr *E, QualType T,
4310                             SourceLocation CC, bool *ICContext = 0) {
4311  if (E->isTypeDependent() || E->isValueDependent()) return;
4312
4313  const Type *Source = S.Context.getCanonicalType(E->getType()).getTypePtr();
4314  const Type *Target = S.Context.getCanonicalType(T).getTypePtr();
4315  if (Source == Target) return;
4316  if (Target->isDependentType()) return;
4317
4318  // If the conversion context location is invalid don't complain. We also
4319  // don't want to emit a warning if the issue occurs from the expansion of
4320  // a system macro. The problem is that 'getSpellingLoc()' is slow, so we
4321  // delay this check as long as possible. Once we detect we are in that
4322  // scenario, we just return.
4323  if (CC.isInvalid())
4324    return;
4325
4326  // Diagnose implicit casts to bool.
4327  if (Target->isSpecificBuiltinType(BuiltinType::Bool)) {
4328    if (isa<StringLiteral>(E))
4329      // Warn on string literal to bool.  Checks for string literals in logical
4330      // expressions, for instances, assert(0 && "error here"), is prevented
4331      // by a check in AnalyzeImplicitConversions().
4332      return DiagnoseImpCast(S, E, T, CC,
4333                             diag::warn_impcast_string_literal_to_bool);
4334    if (Source->isFunctionType()) {
4335      // Warn on function to bool. Checks free functions and static member
4336      // functions. Weakly imported functions are excluded from the check,
4337      // since it's common to test their value to check whether the linker
4338      // found a definition for them.
4339      ValueDecl *D = 0;
4340      if (DeclRefExpr* R = dyn_cast<DeclRefExpr>(E)) {
4341        D = R->getDecl();
4342      } else if (MemberExpr *M = dyn_cast<MemberExpr>(E)) {
4343        D = M->getMemberDecl();
4344      }
4345
4346      if (D && !D->isWeak()) {
4347        if (FunctionDecl* F = dyn_cast<FunctionDecl>(D)) {
4348          S.Diag(E->getExprLoc(), diag::warn_impcast_function_to_bool)
4349            << F << E->getSourceRange() << SourceRange(CC);
4350          S.Diag(E->getExprLoc(), diag::note_function_to_bool_silence)
4351            << FixItHint::CreateInsertion(E->getExprLoc(), "&");
4352          QualType ReturnType;
4353          UnresolvedSet<4> NonTemplateOverloads;
4354          S.isExprCallable(*E, ReturnType, NonTemplateOverloads);
4355          if (!ReturnType.isNull()
4356              && ReturnType->isSpecificBuiltinType(BuiltinType::Bool))
4357            S.Diag(E->getExprLoc(), diag::note_function_to_bool_call)
4358              << FixItHint::CreateInsertion(
4359                 S.getPreprocessor().getLocForEndOfToken(E->getLocEnd()), "()");
4360          return;
4361        }
4362      }
4363    }
4364  }
4365
4366  // Strip vector types.
4367  if (isa<VectorType>(Source)) {
4368    if (!isa<VectorType>(Target)) {
4369      if (S.SourceMgr.isInSystemMacro(CC))
4370        return;
4371      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_vector_scalar);
4372    }
4373
4374    // If the vector cast is cast between two vectors of the same size, it is
4375    // a bitcast, not a conversion.
4376    if (S.Context.getTypeSize(Source) == S.Context.getTypeSize(Target))
4377      return;
4378
4379    Source = cast<VectorType>(Source)->getElementType().getTypePtr();
4380    Target = cast<VectorType>(Target)->getElementType().getTypePtr();
4381  }
4382
4383  // Strip complex types.
4384  if (isa<ComplexType>(Source)) {
4385    if (!isa<ComplexType>(Target)) {
4386      if (S.SourceMgr.isInSystemMacro(CC))
4387        return;
4388
4389      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_complex_scalar);
4390    }
4391
4392    Source = cast<ComplexType>(Source)->getElementType().getTypePtr();
4393    Target = cast<ComplexType>(Target)->getElementType().getTypePtr();
4394  }
4395
4396  const BuiltinType *SourceBT = dyn_cast<BuiltinType>(Source);
4397  const BuiltinType *TargetBT = dyn_cast<BuiltinType>(Target);
4398
4399  // If the source is floating point...
4400  if (SourceBT && SourceBT->isFloatingPoint()) {
4401    // ...and the target is floating point...
4402    if (TargetBT && TargetBT->isFloatingPoint()) {
4403      // ...then warn if we're dropping FP rank.
4404
4405      // Builtin FP kinds are ordered by increasing FP rank.
4406      if (SourceBT->getKind() > TargetBT->getKind()) {
4407        // Don't warn about float constants that are precisely
4408        // representable in the target type.
4409        Expr::EvalResult result;
4410        if (E->EvaluateAsRValue(result, S.Context)) {
4411          // Value might be a float, a float vector, or a float complex.
4412          if (IsSameFloatAfterCast(result.Val,
4413                   S.Context.getFloatTypeSemantics(QualType(TargetBT, 0)),
4414                   S.Context.getFloatTypeSemantics(QualType(SourceBT, 0))))
4415            return;
4416        }
4417
4418        if (S.SourceMgr.isInSystemMacro(CC))
4419          return;
4420
4421        DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_float_precision);
4422      }
4423      return;
4424    }
4425
4426    // If the target is integral, always warn.
4427    if (TargetBT && TargetBT->isInteger()) {
4428      if (S.SourceMgr.isInSystemMacro(CC))
4429        return;
4430
4431      Expr *InnerE = E->IgnoreParenImpCasts();
4432      // We also want to warn on, e.g., "int i = -1.234"
4433      if (UnaryOperator *UOp = dyn_cast<UnaryOperator>(InnerE))
4434        if (UOp->getOpcode() == UO_Minus || UOp->getOpcode() == UO_Plus)
4435          InnerE = UOp->getSubExpr()->IgnoreParenImpCasts();
4436
4437      if (FloatingLiteral *FL = dyn_cast<FloatingLiteral>(InnerE)) {
4438        DiagnoseFloatingLiteralImpCast(S, FL, T, CC);
4439      } else {
4440        DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_float_integer);
4441      }
4442    }
4443
4444    return;
4445  }
4446
4447  if ((E->isNullPointerConstant(S.Context, Expr::NPC_ValueDependentIsNotNull)
4448           == Expr::NPCK_GNUNull) && !Target->isAnyPointerType()
4449      && !Target->isBlockPointerType() && !Target->isMemberPointerType()) {
4450    SourceLocation Loc = E->getSourceRange().getBegin();
4451    if (Loc.isMacroID())
4452      Loc = S.SourceMgr.getImmediateExpansionRange(Loc).first;
4453    if (!Loc.isMacroID() || CC.isMacroID())
4454      S.Diag(Loc, diag::warn_impcast_null_pointer_to_integer)
4455          << T << clang::SourceRange(CC)
4456          << FixItHint::CreateReplacement(Loc, S.getFixItZeroLiteralForType(T));
4457  }
4458
4459  if (!Source->isIntegerType() || !Target->isIntegerType())
4460    return;
4461
4462  // TODO: remove this early return once the false positives for constant->bool
4463  // in templates, macros, etc, are reduced or removed.
4464  if (Target->isSpecificBuiltinType(BuiltinType::Bool))
4465    return;
4466
4467  IntRange SourceRange = GetExprRange(S.Context, E);
4468  IntRange TargetRange = IntRange::forTargetOfCanonicalType(S.Context, Target);
4469
4470  if (SourceRange.Width > TargetRange.Width) {
4471    // If the source is a constant, use a default-on diagnostic.
4472    // TODO: this should happen for bitfield stores, too.
4473    llvm::APSInt Value(32);
4474    if (E->isIntegerConstantExpr(Value, S.Context)) {
4475      if (S.SourceMgr.isInSystemMacro(CC))
4476        return;
4477
4478      std::string PrettySourceValue = Value.toString(10);
4479      std::string PrettyTargetValue = PrettyPrintInRange(Value, TargetRange);
4480
4481      S.DiagRuntimeBehavior(E->getExprLoc(), E,
4482        S.PDiag(diag::warn_impcast_integer_precision_constant)
4483            << PrettySourceValue << PrettyTargetValue
4484            << E->getType() << T << E->getSourceRange()
4485            << clang::SourceRange(CC));
4486      return;
4487    }
4488
4489    // People want to build with -Wshorten-64-to-32 and not -Wconversion.
4490    if (S.SourceMgr.isInSystemMacro(CC))
4491      return;
4492
4493    if (TargetRange.Width == 32 && S.Context.getIntWidth(E->getType()) == 64)
4494      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_integer_64_32,
4495                             /* pruneControlFlow */ true);
4496    return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_integer_precision);
4497  }
4498
4499  if ((TargetRange.NonNegative && !SourceRange.NonNegative) ||
4500      (!TargetRange.NonNegative && SourceRange.NonNegative &&
4501       SourceRange.Width == TargetRange.Width)) {
4502
4503    if (S.SourceMgr.isInSystemMacro(CC))
4504      return;
4505
4506    unsigned DiagID = diag::warn_impcast_integer_sign;
4507
4508    // Traditionally, gcc has warned about this under -Wsign-compare.
4509    // We also want to warn about it in -Wconversion.
4510    // So if -Wconversion is off, use a completely identical diagnostic
4511    // in the sign-compare group.
4512    // The conditional-checking code will
4513    if (ICContext) {
4514      DiagID = diag::warn_impcast_integer_sign_conditional;
4515      *ICContext = true;
4516    }
4517
4518    return DiagnoseImpCast(S, E, T, CC, DiagID);
4519  }
4520
4521  // Diagnose conversions between different enumeration types.
4522  // In C, we pretend that the type of an EnumConstantDecl is its enumeration
4523  // type, to give us better diagnostics.
4524  QualType SourceType = E->getType();
4525  if (!S.getLangOpts().CPlusPlus) {
4526    if (DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(E))
4527      if (EnumConstantDecl *ECD = dyn_cast<EnumConstantDecl>(DRE->getDecl())) {
4528        EnumDecl *Enum = cast<EnumDecl>(ECD->getDeclContext());
4529        SourceType = S.Context.getTypeDeclType(Enum);
4530        Source = S.Context.getCanonicalType(SourceType).getTypePtr();
4531      }
4532  }
4533
4534  if (const EnumType *SourceEnum = Source->getAs<EnumType>())
4535    if (const EnumType *TargetEnum = Target->getAs<EnumType>())
4536      if ((SourceEnum->getDecl()->getIdentifier() ||
4537           SourceEnum->getDecl()->getTypedefNameForAnonDecl()) &&
4538          (TargetEnum->getDecl()->getIdentifier() ||
4539           TargetEnum->getDecl()->getTypedefNameForAnonDecl()) &&
4540          SourceEnum != TargetEnum) {
4541        if (S.SourceMgr.isInSystemMacro(CC))
4542          return;
4543
4544        return DiagnoseImpCast(S, E, SourceType, T, CC,
4545                               diag::warn_impcast_different_enum_types);
4546      }
4547
4548  return;
4549}
4550
4551void CheckConditionalOperator(Sema &S, ConditionalOperator *E,
4552                              SourceLocation CC, QualType T);
4553
4554void CheckConditionalOperand(Sema &S, Expr *E, QualType T,
4555                             SourceLocation CC, bool &ICContext) {
4556  E = E->IgnoreParenImpCasts();
4557
4558  if (isa<ConditionalOperator>(E))
4559    return CheckConditionalOperator(S, cast<ConditionalOperator>(E), CC, T);
4560
4561  AnalyzeImplicitConversions(S, E, CC);
4562  if (E->getType() != T)
4563    return CheckImplicitConversion(S, E, T, CC, &ICContext);
4564  return;
4565}
4566
4567void CheckConditionalOperator(Sema &S, ConditionalOperator *E,
4568                              SourceLocation CC, QualType T) {
4569  AnalyzeImplicitConversions(S, E->getCond(), CC);
4570
4571  bool Suspicious = false;
4572  CheckConditionalOperand(S, E->getTrueExpr(), T, CC, Suspicious);
4573  CheckConditionalOperand(S, E->getFalseExpr(), T, CC, Suspicious);
4574
4575  // If -Wconversion would have warned about either of the candidates
4576  // for a signedness conversion to the context type...
4577  if (!Suspicious) return;
4578
4579  // ...but it's currently ignored...
4580  if (S.Diags.getDiagnosticLevel(diag::warn_impcast_integer_sign_conditional,
4581                                 CC))
4582    return;
4583
4584  // ...then check whether it would have warned about either of the
4585  // candidates for a signedness conversion to the condition type.
4586  if (E->getType() == T) return;
4587
4588  Suspicious = false;
4589  CheckImplicitConversion(S, E->getTrueExpr()->IgnoreParenImpCasts(),
4590                          E->getType(), CC, &Suspicious);
4591  if (!Suspicious)
4592    CheckImplicitConversion(S, E->getFalseExpr()->IgnoreParenImpCasts(),
4593                            E->getType(), CC, &Suspicious);
4594}
4595
4596/// AnalyzeImplicitConversions - Find and report any interesting
4597/// implicit conversions in the given expression.  There are a couple
4598/// of competing diagnostics here, -Wconversion and -Wsign-compare.
4599void AnalyzeImplicitConversions(Sema &S, Expr *OrigE, SourceLocation CC) {
4600  QualType T = OrigE->getType();
4601  Expr *E = OrigE->IgnoreParenImpCasts();
4602
4603  if (E->isTypeDependent() || E->isValueDependent())
4604    return;
4605
4606  // For conditional operators, we analyze the arguments as if they
4607  // were being fed directly into the output.
4608  if (isa<ConditionalOperator>(E)) {
4609    ConditionalOperator *CO = cast<ConditionalOperator>(E);
4610    CheckConditionalOperator(S, CO, CC, T);
4611    return;
4612  }
4613
4614  // Go ahead and check any implicit conversions we might have skipped.
4615  // The non-canonical typecheck is just an optimization;
4616  // CheckImplicitConversion will filter out dead implicit conversions.
4617  if (E->getType() != T)
4618    CheckImplicitConversion(S, E, T, CC);
4619
4620  // Now continue drilling into this expression.
4621
4622  // Skip past explicit casts.
4623  if (isa<ExplicitCastExpr>(E)) {
4624    E = cast<ExplicitCastExpr>(E)->getSubExpr()->IgnoreParenImpCasts();
4625    return AnalyzeImplicitConversions(S, E, CC);
4626  }
4627
4628  if (BinaryOperator *BO = dyn_cast<BinaryOperator>(E)) {
4629    // Do a somewhat different check with comparison operators.
4630    if (BO->isComparisonOp())
4631      return AnalyzeComparison(S, BO);
4632
4633    // And with simple assignments.
4634    if (BO->getOpcode() == BO_Assign)
4635      return AnalyzeAssignment(S, BO);
4636  }
4637
4638  // These break the otherwise-useful invariant below.  Fortunately,
4639  // we don't really need to recurse into them, because any internal
4640  // expressions should have been analyzed already when they were
4641  // built into statements.
4642  if (isa<StmtExpr>(E)) return;
4643
4644  // Don't descend into unevaluated contexts.
4645  if (isa<UnaryExprOrTypeTraitExpr>(E)) return;
4646
4647  // Now just recurse over the expression's children.
4648  CC = E->getExprLoc();
4649  BinaryOperator *BO = dyn_cast<BinaryOperator>(E);
4650  bool IsLogicalOperator = BO && BO->isLogicalOp();
4651  for (Stmt::child_range I = E->children(); I; ++I) {
4652    Expr *ChildExpr = dyn_cast_or_null<Expr>(*I);
4653    if (!ChildExpr)
4654      continue;
4655
4656    if (IsLogicalOperator &&
4657        isa<StringLiteral>(ChildExpr->IgnoreParenImpCasts()))
4658      // Ignore checking string literals that are in logical operators.
4659      continue;
4660    AnalyzeImplicitConversions(S, ChildExpr, CC);
4661  }
4662}
4663
4664} // end anonymous namespace
4665
4666/// Diagnoses "dangerous" implicit conversions within the given
4667/// expression (which is a full expression).  Implements -Wconversion
4668/// and -Wsign-compare.
4669///
4670/// \param CC the "context" location of the implicit conversion, i.e.
4671///   the most location of the syntactic entity requiring the implicit
4672///   conversion
4673void Sema::CheckImplicitConversions(Expr *E, SourceLocation CC) {
4674  // Don't diagnose in unevaluated contexts.
4675  if (ExprEvalContexts.back().Context == Sema::Unevaluated)
4676    return;
4677
4678  // Don't diagnose for value- or type-dependent expressions.
4679  if (E->isTypeDependent() || E->isValueDependent())
4680    return;
4681
4682  // Check for array bounds violations in cases where the check isn't triggered
4683  // elsewhere for other Expr types (like BinaryOperators), e.g. when an
4684  // ArraySubscriptExpr is on the RHS of a variable initialization.
4685  CheckArrayAccess(E);
4686
4687  // This is not the right CC for (e.g.) a variable initialization.
4688  AnalyzeImplicitConversions(*this, E, CC);
4689}
4690
4691void Sema::CheckBitFieldInitialization(SourceLocation InitLoc,
4692                                       FieldDecl *BitField,
4693                                       Expr *Init) {
4694  (void) AnalyzeBitFieldAssignment(*this, BitField, Init, InitLoc);
4695}
4696
4697/// CheckParmsForFunctionDef - Check that the parameters of the given
4698/// function are appropriate for the definition of a function. This
4699/// takes care of any checks that cannot be performed on the
4700/// declaration itself, e.g., that the types of each of the function
4701/// parameters are complete.
4702bool Sema::CheckParmsForFunctionDef(ParmVarDecl **P, ParmVarDecl **PEnd,
4703                                    bool CheckParameterNames) {
4704  bool HasInvalidParm = false;
4705  for (; P != PEnd; ++P) {
4706    ParmVarDecl *Param = *P;
4707
4708    // C99 6.7.5.3p4: the parameters in a parameter type list in a
4709    // function declarator that is part of a function definition of
4710    // that function shall not have incomplete type.
4711    //
4712    // This is also C++ [dcl.fct]p6.
4713    if (!Param->isInvalidDecl() &&
4714        RequireCompleteType(Param->getLocation(), Param->getType(),
4715                            diag::err_typecheck_decl_incomplete_type)) {
4716      Param->setInvalidDecl();
4717      HasInvalidParm = true;
4718    }
4719
4720    // C99 6.9.1p5: If the declarator includes a parameter type list, the
4721    // declaration of each parameter shall include an identifier.
4722    if (CheckParameterNames &&
4723        Param->getIdentifier() == 0 &&
4724        !Param->isImplicit() &&
4725        !getLangOpts().CPlusPlus)
4726      Diag(Param->getLocation(), diag::err_parameter_name_omitted);
4727
4728    // C99 6.7.5.3p12:
4729    //   If the function declarator is not part of a definition of that
4730    //   function, parameters may have incomplete type and may use the [*]
4731    //   notation in their sequences of declarator specifiers to specify
4732    //   variable length array types.
4733    QualType PType = Param->getOriginalType();
4734    if (const ArrayType *AT = Context.getAsArrayType(PType)) {
4735      if (AT->getSizeModifier() == ArrayType::Star) {
4736        // FIXME: This diagnosic should point the the '[*]' if source-location
4737        // information is added for it.
4738        Diag(Param->getLocation(), diag::err_array_star_in_function_definition);
4739      }
4740    }
4741  }
4742
4743  return HasInvalidParm;
4744}
4745
4746/// CheckCastAlign - Implements -Wcast-align, which warns when a
4747/// pointer cast increases the alignment requirements.
4748void Sema::CheckCastAlign(Expr *Op, QualType T, SourceRange TRange) {
4749  // This is actually a lot of work to potentially be doing on every
4750  // cast; don't do it if we're ignoring -Wcast_align (as is the default).
4751  if (getDiagnostics().getDiagnosticLevel(diag::warn_cast_align,
4752                                          TRange.getBegin())
4753        == DiagnosticsEngine::Ignored)
4754    return;
4755
4756  // Ignore dependent types.
4757  if (T->isDependentType() || Op->getType()->isDependentType())
4758    return;
4759
4760  // Require that the destination be a pointer type.
4761  const PointerType *DestPtr = T->getAs<PointerType>();
4762  if (!DestPtr) return;
4763
4764  // If the destination has alignment 1, we're done.
4765  QualType DestPointee = DestPtr->getPointeeType();
4766  if (DestPointee->isIncompleteType()) return;
4767  CharUnits DestAlign = Context.getTypeAlignInChars(DestPointee);
4768  if (DestAlign.isOne()) return;
4769
4770  // Require that the source be a pointer type.
4771  const PointerType *SrcPtr = Op->getType()->getAs<PointerType>();
4772  if (!SrcPtr) return;
4773  QualType SrcPointee = SrcPtr->getPointeeType();
4774
4775  // Whitelist casts from cv void*.  We already implicitly
4776  // whitelisted casts to cv void*, since they have alignment 1.
4777  // Also whitelist casts involving incomplete types, which implicitly
4778  // includes 'void'.
4779  if (SrcPointee->isIncompleteType()) return;
4780
4781  CharUnits SrcAlign = Context.getTypeAlignInChars(SrcPointee);
4782  if (SrcAlign >= DestAlign) return;
4783
4784  Diag(TRange.getBegin(), diag::warn_cast_align)
4785    << Op->getType() << T
4786    << static_cast<unsigned>(SrcAlign.getQuantity())
4787    << static_cast<unsigned>(DestAlign.getQuantity())
4788    << TRange << Op->getSourceRange();
4789}
4790
4791static const Type* getElementType(const Expr *BaseExpr) {
4792  const Type* EltType = BaseExpr->getType().getTypePtr();
4793  if (EltType->isAnyPointerType())
4794    return EltType->getPointeeType().getTypePtr();
4795  else if (EltType->isArrayType())
4796    return EltType->getBaseElementTypeUnsafe();
4797  return EltType;
4798}
4799
4800/// \brief Check whether this array fits the idiom of a size-one tail padded
4801/// array member of a struct.
4802///
4803/// We avoid emitting out-of-bounds access warnings for such arrays as they are
4804/// commonly used to emulate flexible arrays in C89 code.
4805static bool IsTailPaddedMemberArray(Sema &S, llvm::APInt Size,
4806                                    const NamedDecl *ND) {
4807  if (Size != 1 || !ND) return false;
4808
4809  const FieldDecl *FD = dyn_cast<FieldDecl>(ND);
4810  if (!FD) return false;
4811
4812  // Don't consider sizes resulting from macro expansions or template argument
4813  // substitution to form C89 tail-padded arrays.
4814
4815  TypeSourceInfo *TInfo = FD->getTypeSourceInfo();
4816  while (TInfo) {
4817    TypeLoc TL = TInfo->getTypeLoc();
4818    // Look through typedefs.
4819    const TypedefTypeLoc *TTL = dyn_cast<TypedefTypeLoc>(&TL);
4820    if (TTL) {
4821      const TypedefNameDecl *TDL = TTL->getTypedefNameDecl();
4822      TInfo = TDL->getTypeSourceInfo();
4823      continue;
4824    }
4825    ConstantArrayTypeLoc CTL = cast<ConstantArrayTypeLoc>(TL);
4826    const Expr *SizeExpr = dyn_cast<IntegerLiteral>(CTL.getSizeExpr());
4827    if (!SizeExpr || SizeExpr->getExprLoc().isMacroID())
4828      return false;
4829    break;
4830  }
4831
4832  const RecordDecl *RD = dyn_cast<RecordDecl>(FD->getDeclContext());
4833  if (!RD) return false;
4834  if (RD->isUnion()) return false;
4835  if (const CXXRecordDecl *CRD = dyn_cast<CXXRecordDecl>(RD)) {
4836    if (!CRD->isStandardLayout()) return false;
4837  }
4838
4839  // See if this is the last field decl in the record.
4840  const Decl *D = FD;
4841  while ((D = D->getNextDeclInContext()))
4842    if (isa<FieldDecl>(D))
4843      return false;
4844  return true;
4845}
4846
4847void Sema::CheckArrayAccess(const Expr *BaseExpr, const Expr *IndexExpr,
4848                            const ArraySubscriptExpr *ASE,
4849                            bool AllowOnePastEnd, bool IndexNegated) {
4850  IndexExpr = IndexExpr->IgnoreParenImpCasts();
4851  if (IndexExpr->isValueDependent())
4852    return;
4853
4854  const Type *EffectiveType = getElementType(BaseExpr);
4855  BaseExpr = BaseExpr->IgnoreParenCasts();
4856  const ConstantArrayType *ArrayTy =
4857    Context.getAsConstantArrayType(BaseExpr->getType());
4858  if (!ArrayTy)
4859    return;
4860
4861  llvm::APSInt index;
4862  if (!IndexExpr->EvaluateAsInt(index, Context))
4863    return;
4864  if (IndexNegated)
4865    index = -index;
4866
4867  const NamedDecl *ND = NULL;
4868  if (const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(BaseExpr))
4869    ND = dyn_cast<NamedDecl>(DRE->getDecl());
4870  if (const MemberExpr *ME = dyn_cast<MemberExpr>(BaseExpr))
4871    ND = dyn_cast<NamedDecl>(ME->getMemberDecl());
4872
4873  if (index.isUnsigned() || !index.isNegative()) {
4874    llvm::APInt size = ArrayTy->getSize();
4875    if (!size.isStrictlyPositive())
4876      return;
4877
4878    const Type* BaseType = getElementType(BaseExpr);
4879    if (BaseType != EffectiveType) {
4880      // Make sure we're comparing apples to apples when comparing index to size
4881      uint64_t ptrarith_typesize = Context.getTypeSize(EffectiveType);
4882      uint64_t array_typesize = Context.getTypeSize(BaseType);
4883      // Handle ptrarith_typesize being zero, such as when casting to void*
4884      if (!ptrarith_typesize) ptrarith_typesize = 1;
4885      if (ptrarith_typesize != array_typesize) {
4886        // There's a cast to a different size type involved
4887        uint64_t ratio = array_typesize / ptrarith_typesize;
4888        // TODO: Be smarter about handling cases where array_typesize is not a
4889        // multiple of ptrarith_typesize
4890        if (ptrarith_typesize * ratio == array_typesize)
4891          size *= llvm::APInt(size.getBitWidth(), ratio);
4892      }
4893    }
4894
4895    if (size.getBitWidth() > index.getBitWidth())
4896      index = index.zext(size.getBitWidth());
4897    else if (size.getBitWidth() < index.getBitWidth())
4898      size = size.zext(index.getBitWidth());
4899
4900    // For array subscripting the index must be less than size, but for pointer
4901    // arithmetic also allow the index (offset) to be equal to size since
4902    // computing the next address after the end of the array is legal and
4903    // commonly done e.g. in C++ iterators and range-based for loops.
4904    if (AllowOnePastEnd ? index.ule(size) : index.ult(size))
4905      return;
4906
4907    // Also don't warn for arrays of size 1 which are members of some
4908    // structure. These are often used to approximate flexible arrays in C89
4909    // code.
4910    if (IsTailPaddedMemberArray(*this, size, ND))
4911      return;
4912
4913    // Suppress the warning if the subscript expression (as identified by the
4914    // ']' location) and the index expression are both from macro expansions
4915    // within a system header.
4916    if (ASE) {
4917      SourceLocation RBracketLoc = SourceMgr.getSpellingLoc(
4918          ASE->getRBracketLoc());
4919      if (SourceMgr.isInSystemHeader(RBracketLoc)) {
4920        SourceLocation IndexLoc = SourceMgr.getSpellingLoc(
4921            IndexExpr->getLocStart());
4922        if (SourceMgr.isFromSameFile(RBracketLoc, IndexLoc))
4923          return;
4924      }
4925    }
4926
4927    unsigned DiagID = diag::warn_ptr_arith_exceeds_bounds;
4928    if (ASE)
4929      DiagID = diag::warn_array_index_exceeds_bounds;
4930
4931    DiagRuntimeBehavior(BaseExpr->getLocStart(), BaseExpr,
4932                        PDiag(DiagID) << index.toString(10, true)
4933                          << size.toString(10, true)
4934                          << (unsigned)size.getLimitedValue(~0U)
4935                          << IndexExpr->getSourceRange());
4936  } else {
4937    unsigned DiagID = diag::warn_array_index_precedes_bounds;
4938    if (!ASE) {
4939      DiagID = diag::warn_ptr_arith_precedes_bounds;
4940      if (index.isNegative()) index = -index;
4941    }
4942
4943    DiagRuntimeBehavior(BaseExpr->getLocStart(), BaseExpr,
4944                        PDiag(DiagID) << index.toString(10, true)
4945                          << IndexExpr->getSourceRange());
4946  }
4947
4948  if (!ND) {
4949    // Try harder to find a NamedDecl to point at in the note.
4950    while (const ArraySubscriptExpr *ASE =
4951           dyn_cast<ArraySubscriptExpr>(BaseExpr))
4952      BaseExpr = ASE->getBase()->IgnoreParenCasts();
4953    if (const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(BaseExpr))
4954      ND = dyn_cast<NamedDecl>(DRE->getDecl());
4955    if (const MemberExpr *ME = dyn_cast<MemberExpr>(BaseExpr))
4956      ND = dyn_cast<NamedDecl>(ME->getMemberDecl());
4957  }
4958
4959  if (ND)
4960    DiagRuntimeBehavior(ND->getLocStart(), BaseExpr,
4961                        PDiag(diag::note_array_index_out_of_bounds)
4962                          << ND->getDeclName());
4963}
4964
4965void Sema::CheckArrayAccess(const Expr *expr) {
4966  int AllowOnePastEnd = 0;
4967  while (expr) {
4968    expr = expr->IgnoreParenImpCasts();
4969    switch (expr->getStmtClass()) {
4970      case Stmt::ArraySubscriptExprClass: {
4971        const ArraySubscriptExpr *ASE = cast<ArraySubscriptExpr>(expr);
4972        CheckArrayAccess(ASE->getBase(), ASE->getIdx(), ASE,
4973                         AllowOnePastEnd > 0);
4974        return;
4975      }
4976      case Stmt::UnaryOperatorClass: {
4977        // Only unwrap the * and & unary operators
4978        const UnaryOperator *UO = cast<UnaryOperator>(expr);
4979        expr = UO->getSubExpr();
4980        switch (UO->getOpcode()) {
4981          case UO_AddrOf:
4982            AllowOnePastEnd++;
4983            break;
4984          case UO_Deref:
4985            AllowOnePastEnd--;
4986            break;
4987          default:
4988            return;
4989        }
4990        break;
4991      }
4992      case Stmt::ConditionalOperatorClass: {
4993        const ConditionalOperator *cond = cast<ConditionalOperator>(expr);
4994        if (const Expr *lhs = cond->getLHS())
4995          CheckArrayAccess(lhs);
4996        if (const Expr *rhs = cond->getRHS())
4997          CheckArrayAccess(rhs);
4998        return;
4999      }
5000      default:
5001        return;
5002    }
5003  }
5004}
5005
5006//===--- CHECK: Objective-C retain cycles ----------------------------------//
5007
5008namespace {
5009  struct RetainCycleOwner {
5010    RetainCycleOwner() : Variable(0), Indirect(false) {}
5011    VarDecl *Variable;
5012    SourceRange Range;
5013    SourceLocation Loc;
5014    bool Indirect;
5015
5016    void setLocsFrom(Expr *e) {
5017      Loc = e->getExprLoc();
5018      Range = e->getSourceRange();
5019    }
5020  };
5021}
5022
5023/// Consider whether capturing the given variable can possibly lead to
5024/// a retain cycle.
5025static bool considerVariable(VarDecl *var, Expr *ref, RetainCycleOwner &owner) {
5026  // In ARC, it's captured strongly iff the variable has __strong
5027  // lifetime.  In MRR, it's captured strongly if the variable is
5028  // __block and has an appropriate type.
5029  if (var->getType().getObjCLifetime() != Qualifiers::OCL_Strong)
5030    return false;
5031
5032  owner.Variable = var;
5033  owner.setLocsFrom(ref);
5034  return true;
5035}
5036
5037static bool findRetainCycleOwner(Sema &S, Expr *e, RetainCycleOwner &owner) {
5038  while (true) {
5039    e = e->IgnoreParens();
5040    if (CastExpr *cast = dyn_cast<CastExpr>(e)) {
5041      switch (cast->getCastKind()) {
5042      case CK_BitCast:
5043      case CK_LValueBitCast:
5044      case CK_LValueToRValue:
5045      case CK_ARCReclaimReturnedObject:
5046        e = cast->getSubExpr();
5047        continue;
5048
5049      default:
5050        return false;
5051      }
5052    }
5053
5054    if (ObjCIvarRefExpr *ref = dyn_cast<ObjCIvarRefExpr>(e)) {
5055      ObjCIvarDecl *ivar = ref->getDecl();
5056      if (ivar->getType().getObjCLifetime() != Qualifiers::OCL_Strong)
5057        return false;
5058
5059      // Try to find a retain cycle in the base.
5060      if (!findRetainCycleOwner(S, ref->getBase(), owner))
5061        return false;
5062
5063      if (ref->isFreeIvar()) owner.setLocsFrom(ref);
5064      owner.Indirect = true;
5065      return true;
5066    }
5067
5068    if (DeclRefExpr *ref = dyn_cast<DeclRefExpr>(e)) {
5069      VarDecl *var = dyn_cast<VarDecl>(ref->getDecl());
5070      if (!var) return false;
5071      return considerVariable(var, ref, owner);
5072    }
5073
5074    if (MemberExpr *member = dyn_cast<MemberExpr>(e)) {
5075      if (member->isArrow()) return false;
5076
5077      // Don't count this as an indirect ownership.
5078      e = member->getBase();
5079      continue;
5080    }
5081
5082    if (PseudoObjectExpr *pseudo = dyn_cast<PseudoObjectExpr>(e)) {
5083      // Only pay attention to pseudo-objects on property references.
5084      ObjCPropertyRefExpr *pre
5085        = dyn_cast<ObjCPropertyRefExpr>(pseudo->getSyntacticForm()
5086                                              ->IgnoreParens());
5087      if (!pre) return false;
5088      if (pre->isImplicitProperty()) return false;
5089      ObjCPropertyDecl *property = pre->getExplicitProperty();
5090      if (!property->isRetaining() &&
5091          !(property->getPropertyIvarDecl() &&
5092            property->getPropertyIvarDecl()->getType()
5093              .getObjCLifetime() == Qualifiers::OCL_Strong))
5094          return false;
5095
5096      owner.Indirect = true;
5097      if (pre->isSuperReceiver()) {
5098        owner.Variable = S.getCurMethodDecl()->getSelfDecl();
5099        if (!owner.Variable)
5100          return false;
5101        owner.Loc = pre->getLocation();
5102        owner.Range = pre->getSourceRange();
5103        return true;
5104      }
5105      e = const_cast<Expr*>(cast<OpaqueValueExpr>(pre->getBase())
5106                              ->getSourceExpr());
5107      continue;
5108    }
5109
5110    // Array ivars?
5111
5112    return false;
5113  }
5114}
5115
5116namespace {
5117  struct FindCaptureVisitor : EvaluatedExprVisitor<FindCaptureVisitor> {
5118    FindCaptureVisitor(ASTContext &Context, VarDecl *variable)
5119      : EvaluatedExprVisitor<FindCaptureVisitor>(Context),
5120        Variable(variable), Capturer(0) {}
5121
5122    VarDecl *Variable;
5123    Expr *Capturer;
5124
5125    void VisitDeclRefExpr(DeclRefExpr *ref) {
5126      if (ref->getDecl() == Variable && !Capturer)
5127        Capturer = ref;
5128    }
5129
5130    void VisitObjCIvarRefExpr(ObjCIvarRefExpr *ref) {
5131      if (Capturer) return;
5132      Visit(ref->getBase());
5133      if (Capturer && ref->isFreeIvar())
5134        Capturer = ref;
5135    }
5136
5137    void VisitBlockExpr(BlockExpr *block) {
5138      // Look inside nested blocks
5139      if (block->getBlockDecl()->capturesVariable(Variable))
5140        Visit(block->getBlockDecl()->getBody());
5141    }
5142  };
5143}
5144
5145/// Check whether the given argument is a block which captures a
5146/// variable.
5147static Expr *findCapturingExpr(Sema &S, Expr *e, RetainCycleOwner &owner) {
5148  assert(owner.Variable && owner.Loc.isValid());
5149
5150  e = e->IgnoreParenCasts();
5151  BlockExpr *block = dyn_cast<BlockExpr>(e);
5152  if (!block || !block->getBlockDecl()->capturesVariable(owner.Variable))
5153    return 0;
5154
5155  FindCaptureVisitor visitor(S.Context, owner.Variable);
5156  visitor.Visit(block->getBlockDecl()->getBody());
5157  return visitor.Capturer;
5158}
5159
5160static void diagnoseRetainCycle(Sema &S, Expr *capturer,
5161                                RetainCycleOwner &owner) {
5162  assert(capturer);
5163  assert(owner.Variable && owner.Loc.isValid());
5164
5165  S.Diag(capturer->getExprLoc(), diag::warn_arc_retain_cycle)
5166    << owner.Variable << capturer->getSourceRange();
5167  S.Diag(owner.Loc, diag::note_arc_retain_cycle_owner)
5168    << owner.Indirect << owner.Range;
5169}
5170
5171/// Check for a keyword selector that starts with the word 'add' or
5172/// 'set'.
5173static bool isSetterLikeSelector(Selector sel) {
5174  if (sel.isUnarySelector()) return false;
5175
5176  StringRef str = sel.getNameForSlot(0);
5177  while (!str.empty() && str.front() == '_') str = str.substr(1);
5178  if (str.startswith("set"))
5179    str = str.substr(3);
5180  else if (str.startswith("add")) {
5181    // Specially whitelist 'addOperationWithBlock:'.
5182    if (sel.getNumArgs() == 1 && str.startswith("addOperationWithBlock"))
5183      return false;
5184    str = str.substr(3);
5185  }
5186  else
5187    return false;
5188
5189  if (str.empty()) return true;
5190  return !islower(str.front());
5191}
5192
5193/// Check a message send to see if it's likely to cause a retain cycle.
5194void Sema::checkRetainCycles(ObjCMessageExpr *msg) {
5195  // Only check instance methods whose selector looks like a setter.
5196  if (!msg->isInstanceMessage() || !isSetterLikeSelector(msg->getSelector()))
5197    return;
5198
5199  // Try to find a variable that the receiver is strongly owned by.
5200  RetainCycleOwner owner;
5201  if (msg->getReceiverKind() == ObjCMessageExpr::Instance) {
5202    if (!findRetainCycleOwner(*this, msg->getInstanceReceiver(), owner))
5203      return;
5204  } else {
5205    assert(msg->getReceiverKind() == ObjCMessageExpr::SuperInstance);
5206    owner.Variable = getCurMethodDecl()->getSelfDecl();
5207    owner.Loc = msg->getSuperLoc();
5208    owner.Range = msg->getSuperLoc();
5209  }
5210
5211  // Check whether the receiver is captured by any of the arguments.
5212  for (unsigned i = 0, e = msg->getNumArgs(); i != e; ++i)
5213    if (Expr *capturer = findCapturingExpr(*this, msg->getArg(i), owner))
5214      return diagnoseRetainCycle(*this, capturer, owner);
5215}
5216
5217/// Check a property assign to see if it's likely to cause a retain cycle.
5218void Sema::checkRetainCycles(Expr *receiver, Expr *argument) {
5219  RetainCycleOwner owner;
5220  if (!findRetainCycleOwner(*this, receiver, owner))
5221    return;
5222
5223  if (Expr *capturer = findCapturingExpr(*this, argument, owner))
5224    diagnoseRetainCycle(*this, capturer, owner);
5225}
5226
5227bool Sema::checkUnsafeAssigns(SourceLocation Loc,
5228                              QualType LHS, Expr *RHS) {
5229  Qualifiers::ObjCLifetime LT = LHS.getObjCLifetime();
5230  if (LT != Qualifiers::OCL_Weak && LT != Qualifiers::OCL_ExplicitNone)
5231    return false;
5232  // strip off any implicit cast added to get to the one arc-specific
5233  while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
5234    if (cast->getCastKind() == CK_ARCConsumeObject) {
5235      Diag(Loc, diag::warn_arc_retained_assign)
5236        << (LT == Qualifiers::OCL_ExplicitNone) << 1
5237        << RHS->getSourceRange();
5238      return true;
5239    }
5240    RHS = cast->getSubExpr();
5241  }
5242  return false;
5243}
5244
5245void Sema::checkUnsafeExprAssigns(SourceLocation Loc,
5246                              Expr *LHS, Expr *RHS) {
5247  QualType LHSType;
5248  // PropertyRef on LHS type need be directly obtained from
5249  // its declaration as it has a PsuedoType.
5250  ObjCPropertyRefExpr *PRE
5251    = dyn_cast<ObjCPropertyRefExpr>(LHS->IgnoreParens());
5252  if (PRE && !PRE->isImplicitProperty()) {
5253    const ObjCPropertyDecl *PD = PRE->getExplicitProperty();
5254    if (PD)
5255      LHSType = PD->getType();
5256  }
5257
5258  if (LHSType.isNull())
5259    LHSType = LHS->getType();
5260  if (checkUnsafeAssigns(Loc, LHSType, RHS))
5261    return;
5262  Qualifiers::ObjCLifetime LT = LHSType.getObjCLifetime();
5263  // FIXME. Check for other life times.
5264  if (LT != Qualifiers::OCL_None)
5265    return;
5266
5267  if (PRE) {
5268    if (PRE->isImplicitProperty())
5269      return;
5270    const ObjCPropertyDecl *PD = PRE->getExplicitProperty();
5271    if (!PD)
5272      return;
5273
5274    unsigned Attributes = PD->getPropertyAttributes();
5275    if (Attributes & ObjCPropertyDecl::OBJC_PR_assign) {
5276      // when 'assign' attribute was not explicitly specified
5277      // by user, ignore it and rely on property type itself
5278      // for lifetime info.
5279      unsigned AsWrittenAttr = PD->getPropertyAttributesAsWritten();
5280      if (!(AsWrittenAttr & ObjCPropertyDecl::OBJC_PR_assign) &&
5281          LHSType->isObjCRetainableType())
5282        return;
5283
5284      while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
5285        if (cast->getCastKind() == CK_ARCConsumeObject) {
5286          Diag(Loc, diag::warn_arc_retained_property_assign)
5287          << RHS->getSourceRange();
5288          return;
5289        }
5290        RHS = cast->getSubExpr();
5291      }
5292    }
5293    else if (Attributes & ObjCPropertyDecl::OBJC_PR_weak) {
5294      while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
5295        if (cast->getCastKind() == CK_ARCConsumeObject) {
5296          Diag(Loc, diag::warn_arc_retained_assign)
5297          << 0 << 0<< RHS->getSourceRange();
5298          return;
5299        }
5300        RHS = cast->getSubExpr();
5301      }
5302    }
5303  }
5304}
5305
5306//===--- CHECK: Empty statement body (-Wempty-body) ---------------------===//
5307
5308namespace {
5309bool ShouldDiagnoseEmptyStmtBody(const SourceManager &SourceMgr,
5310                                 SourceLocation StmtLoc,
5311                                 const NullStmt *Body) {
5312  // Do not warn if the body is a macro that expands to nothing, e.g:
5313  //
5314  // #define CALL(x)
5315  // if (condition)
5316  //   CALL(0);
5317  //
5318  if (Body->hasLeadingEmptyMacro())
5319    return false;
5320
5321  // Get line numbers of statement and body.
5322  bool StmtLineInvalid;
5323  unsigned StmtLine = SourceMgr.getSpellingLineNumber(StmtLoc,
5324                                                      &StmtLineInvalid);
5325  if (StmtLineInvalid)
5326    return false;
5327
5328  bool BodyLineInvalid;
5329  unsigned BodyLine = SourceMgr.getSpellingLineNumber(Body->getSemiLoc(),
5330                                                      &BodyLineInvalid);
5331  if (BodyLineInvalid)
5332    return false;
5333
5334  // Warn if null statement and body are on the same line.
5335  if (StmtLine != BodyLine)
5336    return false;
5337
5338  return true;
5339}
5340} // Unnamed namespace
5341
5342void Sema::DiagnoseEmptyStmtBody(SourceLocation StmtLoc,
5343                                 const Stmt *Body,
5344                                 unsigned DiagID) {
5345  // Since this is a syntactic check, don't emit diagnostic for template
5346  // instantiations, this just adds noise.
5347  if (CurrentInstantiationScope)
5348    return;
5349
5350  // The body should be a null statement.
5351  const NullStmt *NBody = dyn_cast<NullStmt>(Body);
5352  if (!NBody)
5353    return;
5354
5355  // Do the usual checks.
5356  if (!ShouldDiagnoseEmptyStmtBody(SourceMgr, StmtLoc, NBody))
5357    return;
5358
5359  Diag(NBody->getSemiLoc(), DiagID);
5360  Diag(NBody->getSemiLoc(), diag::note_empty_body_on_separate_line);
5361}
5362
5363void Sema::DiagnoseEmptyLoopBody(const Stmt *S,
5364                                 const Stmt *PossibleBody) {
5365  assert(!CurrentInstantiationScope); // Ensured by caller
5366
5367  SourceLocation StmtLoc;
5368  const Stmt *Body;
5369  unsigned DiagID;
5370  if (const ForStmt *FS = dyn_cast<ForStmt>(S)) {
5371    StmtLoc = FS->getRParenLoc();
5372    Body = FS->getBody();
5373    DiagID = diag::warn_empty_for_body;
5374  } else if (const WhileStmt *WS = dyn_cast<WhileStmt>(S)) {
5375    StmtLoc = WS->getCond()->getSourceRange().getEnd();
5376    Body = WS->getBody();
5377    DiagID = diag::warn_empty_while_body;
5378  } else
5379    return; // Neither `for' nor `while'.
5380
5381  // The body should be a null statement.
5382  const NullStmt *NBody = dyn_cast<NullStmt>(Body);
5383  if (!NBody)
5384    return;
5385
5386  // Skip expensive checks if diagnostic is disabled.
5387  if (Diags.getDiagnosticLevel(DiagID, NBody->getSemiLoc()) ==
5388          DiagnosticsEngine::Ignored)
5389    return;
5390
5391  // Do the usual checks.
5392  if (!ShouldDiagnoseEmptyStmtBody(SourceMgr, StmtLoc, NBody))
5393    return;
5394
5395  // `for(...);' and `while(...);' are popular idioms, so in order to keep
5396  // noise level low, emit diagnostics only if for/while is followed by a
5397  // CompoundStmt, e.g.:
5398  //    for (int i = 0; i < n; i++);
5399  //    {
5400  //      a(i);
5401  //    }
5402  // or if for/while is followed by a statement with more indentation
5403  // than for/while itself:
5404  //    for (int i = 0; i < n; i++);
5405  //      a(i);
5406  bool ProbableTypo = isa<CompoundStmt>(PossibleBody);
5407  if (!ProbableTypo) {
5408    bool BodyColInvalid;
5409    unsigned BodyCol = SourceMgr.getPresumedColumnNumber(
5410                             PossibleBody->getLocStart(),
5411                             &BodyColInvalid);
5412    if (BodyColInvalid)
5413      return;
5414
5415    bool StmtColInvalid;
5416    unsigned StmtCol = SourceMgr.getPresumedColumnNumber(
5417                             S->getLocStart(),
5418                             &StmtColInvalid);
5419    if (StmtColInvalid)
5420      return;
5421
5422    if (BodyCol > StmtCol)
5423      ProbableTypo = true;
5424  }
5425
5426  if (ProbableTypo) {
5427    Diag(NBody->getSemiLoc(), DiagID);
5428    Diag(NBody->getSemiLoc(), diag::note_empty_body_on_separate_line);
5429  }
5430}
5431