SemaChecking.cpp revision 0f38acee92014cb15af980808f87144e5564031d 
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      case llvm::Triple::mips:
291      case llvm::Triple::mipsel:
292      case llvm::Triple::mips64:
293      case llvm::Triple::mips64el:
294        if (CheckMipsBuiltinFunctionCall(BuiltinID, TheCall))
295          return ExprError();
296        break;
297      default:
298        break;
299    }
300  }
301
302  return move(TheCallResult);
303}
304
305// Get the valid immediate range for the specified NEON type code.
306static unsigned RFT(unsigned t, bool shift = false) {
307  NeonTypeFlags Type(t);
308  int IsQuad = Type.isQuad();
309  switch (Type.getEltType()) {
310  case NeonTypeFlags::Int8:
311  case NeonTypeFlags::Poly8:
312    return shift ? 7 : (8 << IsQuad) - 1;
313  case NeonTypeFlags::Int16:
314  case NeonTypeFlags::Poly16:
315    return shift ? 15 : (4 << IsQuad) - 1;
316  case NeonTypeFlags::Int32:
317    return shift ? 31 : (2 << IsQuad) - 1;
318  case NeonTypeFlags::Int64:
319    return shift ? 63 : (1 << IsQuad) - 1;
320  case NeonTypeFlags::Float16:
321    assert(!shift && "cannot shift float types!");
322    return (4 << IsQuad) - 1;
323  case NeonTypeFlags::Float32:
324    assert(!shift && "cannot shift float types!");
325    return (2 << IsQuad) - 1;
326  }
327  llvm_unreachable("Invalid NeonTypeFlag!");
328}
329
330/// getNeonEltType - Return the QualType corresponding to the elements of
331/// the vector type specified by the NeonTypeFlags.  This is used to check
332/// the pointer arguments for Neon load/store intrinsics.
333static QualType getNeonEltType(NeonTypeFlags Flags, ASTContext &Context) {
334  switch (Flags.getEltType()) {
335  case NeonTypeFlags::Int8:
336    return Flags.isUnsigned() ? Context.UnsignedCharTy : Context.SignedCharTy;
337  case NeonTypeFlags::Int16:
338    return Flags.isUnsigned() ? Context.UnsignedShortTy : Context.ShortTy;
339  case NeonTypeFlags::Int32:
340    return Flags.isUnsigned() ? Context.UnsignedIntTy : Context.IntTy;
341  case NeonTypeFlags::Int64:
342    return Flags.isUnsigned() ? Context.UnsignedLongLongTy : Context.LongLongTy;
343  case NeonTypeFlags::Poly8:
344    return Context.SignedCharTy;
345  case NeonTypeFlags::Poly16:
346    return Context.ShortTy;
347  case NeonTypeFlags::Float16:
348    return Context.UnsignedShortTy;
349  case NeonTypeFlags::Float32:
350    return Context.FloatTy;
351  }
352  llvm_unreachable("Invalid NeonTypeFlag!");
353}
354
355bool Sema::CheckARMBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
356  llvm::APSInt Result;
357
358  unsigned mask = 0;
359  unsigned TV = 0;
360  int PtrArgNum = -1;
361  bool HasConstPtr = false;
362  switch (BuiltinID) {
363#define GET_NEON_OVERLOAD_CHECK
364#include "clang/Basic/arm_neon.inc"
365#undef GET_NEON_OVERLOAD_CHECK
366  }
367
368  // For NEON intrinsics which are overloaded on vector element type, validate
369  // the immediate which specifies which variant to emit.
370  unsigned ImmArg = TheCall->getNumArgs()-1;
371  if (mask) {
372    if (SemaBuiltinConstantArg(TheCall, ImmArg, Result))
373      return true;
374
375    TV = Result.getLimitedValue(64);
376    if ((TV > 63) || (mask & (1 << TV)) == 0)
377      return Diag(TheCall->getLocStart(), diag::err_invalid_neon_type_code)
378        << TheCall->getArg(ImmArg)->getSourceRange();
379  }
380
381  if (PtrArgNum >= 0) {
382    // Check that pointer arguments have the specified type.
383    Expr *Arg = TheCall->getArg(PtrArgNum);
384    if (ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(Arg))
385      Arg = ICE->getSubExpr();
386    ExprResult RHS = DefaultFunctionArrayLvalueConversion(Arg);
387    QualType RHSTy = RHS.get()->getType();
388    QualType EltTy = getNeonEltType(NeonTypeFlags(TV), Context);
389    if (HasConstPtr)
390      EltTy = EltTy.withConst();
391    QualType LHSTy = Context.getPointerType(EltTy);
392    AssignConvertType ConvTy;
393    ConvTy = CheckSingleAssignmentConstraints(LHSTy, RHS);
394    if (RHS.isInvalid())
395      return true;
396    if (DiagnoseAssignmentResult(ConvTy, Arg->getLocStart(), LHSTy, RHSTy,
397                                 RHS.get(), AA_Assigning))
398      return true;
399  }
400
401  // For NEON intrinsics which take an immediate value as part of the
402  // instruction, range check them here.
403  unsigned i = 0, l = 0, u = 0;
404  switch (BuiltinID) {
405  default: return false;
406  case ARM::BI__builtin_arm_ssat: i = 1; l = 1; u = 31; break;
407  case ARM::BI__builtin_arm_usat: i = 1; u = 31; break;
408  case ARM::BI__builtin_arm_vcvtr_f:
409  case ARM::BI__builtin_arm_vcvtr_d: i = 1; u = 1; break;
410#define GET_NEON_IMMEDIATE_CHECK
411#include "clang/Basic/arm_neon.inc"
412#undef GET_NEON_IMMEDIATE_CHECK
413  };
414
415  // We can't check the value of a dependent argument.
416  if (TheCall->getArg(i)->isTypeDependent() ||
417      TheCall->getArg(i)->isValueDependent())
418    return false;
419
420  // Check that the immediate argument is actually a constant.
421  if (SemaBuiltinConstantArg(TheCall, i, Result))
422    return true;
423
424  // Range check against the upper/lower values for this isntruction.
425  unsigned Val = Result.getZExtValue();
426  if (Val < l || Val > (u + l))
427    return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
428      << l << u+l << TheCall->getArg(i)->getSourceRange();
429
430  // FIXME: VFP Intrinsics should error if VFP not present.
431  return false;
432}
433
434bool Sema::CheckMipsBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
435  unsigned i = 0, l = 0, u = 0;
436  switch (BuiltinID) {
437  default: return false;
438  case Mips::BI__builtin_mips_wrdsp: i = 1; l = 0; u = 63; break;
439  case Mips::BI__builtin_mips_rddsp: i = 0; l = 0; u = 63; break;
440  };
441
442  // We can't check the value of a dependent argument.
443  if (TheCall->getArg(i)->isTypeDependent() ||
444      TheCall->getArg(i)->isValueDependent())
445    return false;
446
447  // Check that the immediate argument is actually a constant.
448  llvm::APSInt Result;
449  if (SemaBuiltinConstantArg(TheCall, i, Result))
450    return true;
451
452  // Range check against the upper/lower values for this instruction.
453  unsigned Val = Result.getZExtValue();
454  if (Val < l || Val > u)
455    return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
456      << l << u << TheCall->getArg(i)->getSourceRange();
457
458  return false;
459}
460
461/// Given a FunctionDecl's FormatAttr, attempts to populate the FomatStringInfo
462/// parameter with the FormatAttr's correct format_idx and firstDataArg.
463/// Returns true when the format fits the function and the FormatStringInfo has
464/// been populated.
465bool Sema::getFormatStringInfo(const FormatAttr *Format, bool IsCXXMember,
466                               FormatStringInfo *FSI) {
467  FSI->HasVAListArg = Format->getFirstArg() == 0;
468  FSI->FormatIdx = Format->getFormatIdx() - 1;
469  FSI->FirstDataArg = FSI->HasVAListArg ? 0 : Format->getFirstArg() - 1;
470
471  // The way the format attribute works in GCC, the implicit this argument
472  // of member functions is counted. However, it doesn't appear in our own
473  // lists, so decrement format_idx in that case.
474  if (IsCXXMember) {
475    if(FSI->FormatIdx == 0)
476      return false;
477    --FSI->FormatIdx;
478    if (FSI->FirstDataArg != 0)
479      --FSI->FirstDataArg;
480  }
481  return true;
482}
483
484/// Handles the checks for format strings, non-POD arguments to vararg
485/// functions, and NULL arguments passed to non-NULL parameters.
486void Sema::checkCall(NamedDecl *FDecl, Expr **Args,
487                     unsigned NumArgs,
488                     unsigned NumProtoArgs,
489                     bool IsMemberFunction,
490                     SourceLocation Loc,
491                     SourceRange Range,
492                     VariadicCallType CallType) {
493  // FIXME: This mechanism should be abstracted to be less fragile and
494  // more efficient. For example, just map function ids to custom
495  // handlers.
496
497  // Printf and scanf checking.
498  bool HandledFormatString = false;
499  for (specific_attr_iterator<FormatAttr>
500         I = FDecl->specific_attr_begin<FormatAttr>(),
501         E = FDecl->specific_attr_end<FormatAttr>(); I != E ; ++I)
502    if (CheckFormatArguments(*I, Args, NumArgs, IsMemberFunction, CallType,
503                             Loc, Range))
504        HandledFormatString = true;
505
506  // Refuse POD arguments that weren't caught by the format string
507  // checks above.
508  if (!HandledFormatString && CallType != VariadicDoesNotApply)
509    for (unsigned ArgIdx = NumProtoArgs; ArgIdx < NumArgs; ++ArgIdx)
510      variadicArgumentPODCheck(Args[ArgIdx], CallType);
511
512  for (specific_attr_iterator<NonNullAttr>
513         I = FDecl->specific_attr_begin<NonNullAttr>(),
514         E = FDecl->specific_attr_end<NonNullAttr>(); I != E; ++I)
515    CheckNonNullArguments(*I, Args, Loc);
516}
517
518/// CheckConstructorCall - Check a constructor call for correctness and safety
519/// properties not enforced by the C type system.
520void Sema::CheckConstructorCall(FunctionDecl *FDecl, Expr **Args,
521                                unsigned NumArgs,
522                                const FunctionProtoType *Proto,
523                                SourceLocation Loc) {
524  VariadicCallType CallType =
525    Proto->isVariadic() ? VariadicConstructor : VariadicDoesNotApply;
526  checkCall(FDecl, Args, NumArgs, Proto->getNumArgs(),
527            /*IsMemberFunction=*/true, Loc, SourceRange(), CallType);
528}
529
530/// CheckFunctionCall - Check a direct function call for various correctness
531/// and safety properties not strictly enforced by the C type system.
532bool Sema::CheckFunctionCall(FunctionDecl *FDecl, CallExpr *TheCall,
533                             const FunctionProtoType *Proto) {
534  bool IsMemberFunction = isa<CXXMemberCallExpr>(TheCall);
535  VariadicCallType CallType = getVariadicCallType(FDecl, Proto,
536                                                  TheCall->getCallee());
537  unsigned NumProtoArgs = Proto ? Proto->getNumArgs() : 0;
538  checkCall(FDecl, TheCall->getArgs(), TheCall->getNumArgs(), NumProtoArgs,
539            IsMemberFunction, TheCall->getRParenLoc(),
540            TheCall->getCallee()->getSourceRange(), CallType);
541
542  IdentifierInfo *FnInfo = FDecl->getIdentifier();
543  // None of the checks below are needed for functions that don't have
544  // simple names (e.g., C++ conversion functions).
545  if (!FnInfo)
546    return false;
547
548  unsigned CMId = FDecl->getMemoryFunctionKind();
549  if (CMId == 0)
550    return false;
551
552  // Handle memory setting and copying functions.
553  if (CMId == Builtin::BIstrlcpy || CMId == Builtin::BIstrlcat)
554    CheckStrlcpycatArguments(TheCall, FnInfo);
555  else if (CMId == Builtin::BIstrncat)
556    CheckStrncatArguments(TheCall, FnInfo);
557  else
558    CheckMemaccessArguments(TheCall, CMId, FnInfo);
559
560  return false;
561}
562
563bool Sema::CheckObjCMethodCall(ObjCMethodDecl *Method, SourceLocation lbrac,
564                               Expr **Args, unsigned NumArgs) {
565  VariadicCallType CallType =
566      Method->isVariadic() ? VariadicMethod : VariadicDoesNotApply;
567
568  checkCall(Method, Args, NumArgs, Method->param_size(),
569            /*IsMemberFunction=*/false,
570            lbrac, Method->getSourceRange(), CallType);
571
572  return false;
573}
574
575bool Sema::CheckBlockCall(NamedDecl *NDecl, CallExpr *TheCall,
576                          const FunctionProtoType *Proto) {
577  const VarDecl *V = dyn_cast<VarDecl>(NDecl);
578  if (!V)
579    return false;
580
581  QualType Ty = V->getType();
582  if (!Ty->isBlockPointerType())
583    return false;
584
585  VariadicCallType CallType =
586      Proto && Proto->isVariadic() ? VariadicBlock : VariadicDoesNotApply ;
587  unsigned NumProtoArgs = Proto ? Proto->getNumArgs() : 0;
588
589  checkCall(NDecl, TheCall->getArgs(), TheCall->getNumArgs(),
590            NumProtoArgs, /*IsMemberFunction=*/false,
591            TheCall->getRParenLoc(),
592            TheCall->getCallee()->getSourceRange(), CallType);
593
594  return false;
595}
596
597ExprResult Sema::SemaAtomicOpsOverloaded(ExprResult TheCallResult,
598                                         AtomicExpr::AtomicOp Op) {
599  CallExpr *TheCall = cast<CallExpr>(TheCallResult.get());
600  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
601
602  // All these operations take one of the following forms:
603  enum {
604    // C    __c11_atomic_init(A *, C)
605    Init,
606    // C    __c11_atomic_load(A *, int)
607    Load,
608    // void __atomic_load(A *, CP, int)
609    Copy,
610    // C    __c11_atomic_add(A *, M, int)
611    Arithmetic,
612    // C    __atomic_exchange_n(A *, CP, int)
613    Xchg,
614    // void __atomic_exchange(A *, C *, CP, int)
615    GNUXchg,
616    // bool __c11_atomic_compare_exchange_strong(A *, C *, CP, int, int)
617    C11CmpXchg,
618    // bool __atomic_compare_exchange(A *, C *, CP, bool, int, int)
619    GNUCmpXchg
620  } Form = Init;
621  const unsigned NumArgs[] = { 2, 2, 3, 3, 3, 4, 5, 6 };
622  const unsigned NumVals[] = { 1, 0, 1, 1, 1, 2, 2, 3 };
623  // where:
624  //   C is an appropriate type,
625  //   A is volatile _Atomic(C) for __c11 builtins and is C for GNU builtins,
626  //   CP is C for __c11 builtins and GNU _n builtins and is C * otherwise,
627  //   M is C if C is an integer, and ptrdiff_t if C is a pointer, and
628  //   the int parameters are for orderings.
629
630  assert(AtomicExpr::AO__c11_atomic_init == 0 &&
631         AtomicExpr::AO__c11_atomic_fetch_xor + 1 == AtomicExpr::AO__atomic_load
632         && "need to update code for modified C11 atomics");
633  bool IsC11 = Op >= AtomicExpr::AO__c11_atomic_init &&
634               Op <= AtomicExpr::AO__c11_atomic_fetch_xor;
635  bool IsN = Op == AtomicExpr::AO__atomic_load_n ||
636             Op == AtomicExpr::AO__atomic_store_n ||
637             Op == AtomicExpr::AO__atomic_exchange_n ||
638             Op == AtomicExpr::AO__atomic_compare_exchange_n;
639  bool IsAddSub = false;
640
641  switch (Op) {
642  case AtomicExpr::AO__c11_atomic_init:
643    Form = Init;
644    break;
645
646  case AtomicExpr::AO__c11_atomic_load:
647  case AtomicExpr::AO__atomic_load_n:
648    Form = Load;
649    break;
650
651  case AtomicExpr::AO__c11_atomic_store:
652  case AtomicExpr::AO__atomic_load:
653  case AtomicExpr::AO__atomic_store:
654  case AtomicExpr::AO__atomic_store_n:
655    Form = Copy;
656    break;
657
658  case AtomicExpr::AO__c11_atomic_fetch_add:
659  case AtomicExpr::AO__c11_atomic_fetch_sub:
660  case AtomicExpr::AO__atomic_fetch_add:
661  case AtomicExpr::AO__atomic_fetch_sub:
662  case AtomicExpr::AO__atomic_add_fetch:
663  case AtomicExpr::AO__atomic_sub_fetch:
664    IsAddSub = true;
665    // Fall through.
666  case AtomicExpr::AO__c11_atomic_fetch_and:
667  case AtomicExpr::AO__c11_atomic_fetch_or:
668  case AtomicExpr::AO__c11_atomic_fetch_xor:
669  case AtomicExpr::AO__atomic_fetch_and:
670  case AtomicExpr::AO__atomic_fetch_or:
671  case AtomicExpr::AO__atomic_fetch_xor:
672  case AtomicExpr::AO__atomic_fetch_nand:
673  case AtomicExpr::AO__atomic_and_fetch:
674  case AtomicExpr::AO__atomic_or_fetch:
675  case AtomicExpr::AO__atomic_xor_fetch:
676  case AtomicExpr::AO__atomic_nand_fetch:
677    Form = Arithmetic;
678    break;
679
680  case AtomicExpr::AO__c11_atomic_exchange:
681  case AtomicExpr::AO__atomic_exchange_n:
682    Form = Xchg;
683    break;
684
685  case AtomicExpr::AO__atomic_exchange:
686    Form = GNUXchg;
687    break;
688
689  case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
690  case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
691    Form = C11CmpXchg;
692    break;
693
694  case AtomicExpr::AO__atomic_compare_exchange:
695  case AtomicExpr::AO__atomic_compare_exchange_n:
696    Form = GNUCmpXchg;
697    break;
698  }
699
700  // Check we have the right number of arguments.
701  if (TheCall->getNumArgs() < NumArgs[Form]) {
702    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
703      << 0 << NumArgs[Form] << TheCall->getNumArgs()
704      << TheCall->getCallee()->getSourceRange();
705    return ExprError();
706  } else if (TheCall->getNumArgs() > NumArgs[Form]) {
707    Diag(TheCall->getArg(NumArgs[Form])->getLocStart(),
708         diag::err_typecheck_call_too_many_args)
709      << 0 << NumArgs[Form] << TheCall->getNumArgs()
710      << TheCall->getCallee()->getSourceRange();
711    return ExprError();
712  }
713
714  // Inspect the first argument of the atomic operation.
715  Expr *Ptr = TheCall->getArg(0);
716  Ptr = DefaultFunctionArrayLvalueConversion(Ptr).get();
717  const PointerType *pointerType = Ptr->getType()->getAs<PointerType>();
718  if (!pointerType) {
719    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer)
720      << Ptr->getType() << Ptr->getSourceRange();
721    return ExprError();
722  }
723
724  // For a __c11 builtin, this should be a pointer to an _Atomic type.
725  QualType AtomTy = pointerType->getPointeeType(); // 'A'
726  QualType ValType = AtomTy; // 'C'
727  if (IsC11) {
728    if (!AtomTy->isAtomicType()) {
729      Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic)
730        << Ptr->getType() << Ptr->getSourceRange();
731      return ExprError();
732    }
733    ValType = AtomTy->getAs<AtomicType>()->getValueType();
734  }
735
736  // For an arithmetic operation, the implied arithmetic must be well-formed.
737  if (Form == Arithmetic) {
738    // gcc does not enforce these rules for GNU atomics, but we do so for sanity.
739    if (IsAddSub && !ValType->isIntegerType() && !ValType->isPointerType()) {
740      Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic_int_or_ptr)
741        << IsC11 << Ptr->getType() << Ptr->getSourceRange();
742      return ExprError();
743    }
744    if (!IsAddSub && !ValType->isIntegerType()) {
745      Diag(DRE->getLocStart(), diag::err_atomic_op_bitwise_needs_atomic_int)
746        << IsC11 << Ptr->getType() << Ptr->getSourceRange();
747      return ExprError();
748    }
749  } else if (IsN && !ValType->isIntegerType() && !ValType->isPointerType()) {
750    // For __atomic_*_n operations, the value type must be a scalar integral or
751    // pointer type which is 1, 2, 4, 8 or 16 bytes in length.
752    Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic_int_or_ptr)
753      << IsC11 << Ptr->getType() << Ptr->getSourceRange();
754    return ExprError();
755  }
756
757  if (!IsC11 && !AtomTy.isTriviallyCopyableType(Context)) {
758    // For GNU atomics, require a trivially-copyable type. This is not part of
759    // the GNU atomics specification, but we enforce it for sanity.
760    Diag(DRE->getLocStart(), diag::err_atomic_op_needs_trivial_copy)
761      << Ptr->getType() << Ptr->getSourceRange();
762    return ExprError();
763  }
764
765  // FIXME: For any builtin other than a load, the ValType must not be
766  // const-qualified.
767
768  switch (ValType.getObjCLifetime()) {
769  case Qualifiers::OCL_None:
770  case Qualifiers::OCL_ExplicitNone:
771    // okay
772    break;
773
774  case Qualifiers::OCL_Weak:
775  case Qualifiers::OCL_Strong:
776  case Qualifiers::OCL_Autoreleasing:
777    // FIXME: Can this happen? By this point, ValType should be known
778    // to be trivially copyable.
779    Diag(DRE->getLocStart(), diag::err_arc_atomic_ownership)
780      << ValType << Ptr->getSourceRange();
781    return ExprError();
782  }
783
784  QualType ResultType = ValType;
785  if (Form == Copy || Form == GNUXchg || Form == Init)
786    ResultType = Context.VoidTy;
787  else if (Form == C11CmpXchg || Form == GNUCmpXchg)
788    ResultType = Context.BoolTy;
789
790  // The type of a parameter passed 'by value'. In the GNU atomics, such
791  // arguments are actually passed as pointers.
792  QualType ByValType = ValType; // 'CP'
793  if (!IsC11 && !IsN)
794    ByValType = Ptr->getType();
795
796  // The first argument --- the pointer --- has a fixed type; we
797  // deduce the types of the rest of the arguments accordingly.  Walk
798  // the remaining arguments, converting them to the deduced value type.
799  for (unsigned i = 1; i != NumArgs[Form]; ++i) {
800    QualType Ty;
801    if (i < NumVals[Form] + 1) {
802      switch (i) {
803      case 1:
804        // The second argument is the non-atomic operand. For arithmetic, this
805        // is always passed by value, and for a compare_exchange it is always
806        // passed by address. For the rest, GNU uses by-address and C11 uses
807        // by-value.
808        assert(Form != Load);
809        if (Form == Init || (Form == Arithmetic && ValType->isIntegerType()))
810          Ty = ValType;
811        else if (Form == Copy || Form == Xchg)
812          Ty = ByValType;
813        else if (Form == Arithmetic)
814          Ty = Context.getPointerDiffType();
815        else
816          Ty = Context.getPointerType(ValType.getUnqualifiedType());
817        break;
818      case 2:
819        // The third argument to compare_exchange / GNU exchange is a
820        // (pointer to a) desired value.
821        Ty = ByValType;
822        break;
823      case 3:
824        // The fourth argument to GNU compare_exchange is a 'weak' flag.
825        Ty = Context.BoolTy;
826        break;
827      }
828    } else {
829      // The order(s) are always converted to int.
830      Ty = Context.IntTy;
831    }
832
833    InitializedEntity Entity =
834        InitializedEntity::InitializeParameter(Context, Ty, false);
835    ExprResult Arg = TheCall->getArg(i);
836    Arg = PerformCopyInitialization(Entity, SourceLocation(), Arg);
837    if (Arg.isInvalid())
838      return true;
839    TheCall->setArg(i, Arg.get());
840  }
841
842  // Permute the arguments into a 'consistent' order.
843  SmallVector<Expr*, 5> SubExprs;
844  SubExprs.push_back(Ptr);
845  switch (Form) {
846  case Init:
847    // Note, AtomicExpr::getVal1() has a special case for this atomic.
848    SubExprs.push_back(TheCall->getArg(1)); // Val1
849    break;
850  case Load:
851    SubExprs.push_back(TheCall->getArg(1)); // Order
852    break;
853  case Copy:
854  case Arithmetic:
855  case Xchg:
856    SubExprs.push_back(TheCall->getArg(2)); // Order
857    SubExprs.push_back(TheCall->getArg(1)); // Val1
858    break;
859  case GNUXchg:
860    // Note, AtomicExpr::getVal2() has a special case for this atomic.
861    SubExprs.push_back(TheCall->getArg(3)); // Order
862    SubExprs.push_back(TheCall->getArg(1)); // Val1
863    SubExprs.push_back(TheCall->getArg(2)); // Val2
864    break;
865  case C11CmpXchg:
866    SubExprs.push_back(TheCall->getArg(3)); // Order
867    SubExprs.push_back(TheCall->getArg(1)); // Val1
868    SubExprs.push_back(TheCall->getArg(4)); // OrderFail
869    SubExprs.push_back(TheCall->getArg(2)); // Val2
870    break;
871  case GNUCmpXchg:
872    SubExprs.push_back(TheCall->getArg(4)); // Order
873    SubExprs.push_back(TheCall->getArg(1)); // Val1
874    SubExprs.push_back(TheCall->getArg(5)); // OrderFail
875    SubExprs.push_back(TheCall->getArg(2)); // Val2
876    SubExprs.push_back(TheCall->getArg(3)); // Weak
877    break;
878  }
879
880  return Owned(new (Context) AtomicExpr(TheCall->getCallee()->getLocStart(),
881                                        SubExprs.data(), SubExprs.size(),
882                                        ResultType, Op,
883                                        TheCall->getRParenLoc()));
884}
885
886
887/// checkBuiltinArgument - Given a call to a builtin function, perform
888/// normal type-checking on the given argument, updating the call in
889/// place.  This is useful when a builtin function requires custom
890/// type-checking for some of its arguments but not necessarily all of
891/// them.
892///
893/// Returns true on error.
894static bool checkBuiltinArgument(Sema &S, CallExpr *E, unsigned ArgIndex) {
895  FunctionDecl *Fn = E->getDirectCallee();
896  assert(Fn && "builtin call without direct callee!");
897
898  ParmVarDecl *Param = Fn->getParamDecl(ArgIndex);
899  InitializedEntity Entity =
900    InitializedEntity::InitializeParameter(S.Context, Param);
901
902  ExprResult Arg = E->getArg(0);
903  Arg = S.PerformCopyInitialization(Entity, SourceLocation(), Arg);
904  if (Arg.isInvalid())
905    return true;
906
907  E->setArg(ArgIndex, Arg.take());
908  return false;
909}
910
911/// SemaBuiltinAtomicOverloaded - We have a call to a function like
912/// __sync_fetch_and_add, which is an overloaded function based on the pointer
913/// type of its first argument.  The main ActOnCallExpr routines have already
914/// promoted the types of arguments because all of these calls are prototyped as
915/// void(...).
916///
917/// This function goes through and does final semantic checking for these
918/// builtins,
919ExprResult
920Sema::SemaBuiltinAtomicOverloaded(ExprResult TheCallResult) {
921  CallExpr *TheCall = (CallExpr *)TheCallResult.get();
922  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
923  FunctionDecl *FDecl = cast<FunctionDecl>(DRE->getDecl());
924
925  // Ensure that we have at least one argument to do type inference from.
926  if (TheCall->getNumArgs() < 1) {
927    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args_at_least)
928      << 0 << 1 << TheCall->getNumArgs()
929      << TheCall->getCallee()->getSourceRange();
930    return ExprError();
931  }
932
933  // Inspect the first argument of the atomic builtin.  This should always be
934  // a pointer type, whose element is an integral scalar or pointer type.
935  // Because it is a pointer type, we don't have to worry about any implicit
936  // casts here.
937  // FIXME: We don't allow floating point scalars as input.
938  Expr *FirstArg = TheCall->getArg(0);
939  ExprResult FirstArgResult = DefaultFunctionArrayLvalueConversion(FirstArg);
940  if (FirstArgResult.isInvalid())
941    return ExprError();
942  FirstArg = FirstArgResult.take();
943  TheCall->setArg(0, FirstArg);
944
945  const PointerType *pointerType = FirstArg->getType()->getAs<PointerType>();
946  if (!pointerType) {
947    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer)
948      << FirstArg->getType() << FirstArg->getSourceRange();
949    return ExprError();
950  }
951
952  QualType ValType = pointerType->getPointeeType();
953  if (!ValType->isIntegerType() && !ValType->isAnyPointerType() &&
954      !ValType->isBlockPointerType()) {
955    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer_intptr)
956      << FirstArg->getType() << FirstArg->getSourceRange();
957    return ExprError();
958  }
959
960  switch (ValType.getObjCLifetime()) {
961  case Qualifiers::OCL_None:
962  case Qualifiers::OCL_ExplicitNone:
963    // okay
964    break;
965
966  case Qualifiers::OCL_Weak:
967  case Qualifiers::OCL_Strong:
968  case Qualifiers::OCL_Autoreleasing:
969    Diag(DRE->getLocStart(), diag::err_arc_atomic_ownership)
970      << ValType << FirstArg->getSourceRange();
971    return ExprError();
972  }
973
974  // Strip any qualifiers off ValType.
975  ValType = ValType.getUnqualifiedType();
976
977  // The majority of builtins return a value, but a few have special return
978  // types, so allow them to override appropriately below.
979  QualType ResultType = ValType;
980
981  // We need to figure out which concrete builtin this maps onto.  For example,
982  // __sync_fetch_and_add with a 2 byte object turns into
983  // __sync_fetch_and_add_2.
984#define BUILTIN_ROW(x) \
985  { Builtin::BI##x##_1, Builtin::BI##x##_2, Builtin::BI##x##_4, \
986    Builtin::BI##x##_8, Builtin::BI##x##_16 }
987
988  static const unsigned BuiltinIndices[][5] = {
989    BUILTIN_ROW(__sync_fetch_and_add),
990    BUILTIN_ROW(__sync_fetch_and_sub),
991    BUILTIN_ROW(__sync_fetch_and_or),
992    BUILTIN_ROW(__sync_fetch_and_and),
993    BUILTIN_ROW(__sync_fetch_and_xor),
994
995    BUILTIN_ROW(__sync_add_and_fetch),
996    BUILTIN_ROW(__sync_sub_and_fetch),
997    BUILTIN_ROW(__sync_and_and_fetch),
998    BUILTIN_ROW(__sync_or_and_fetch),
999    BUILTIN_ROW(__sync_xor_and_fetch),
1000
1001    BUILTIN_ROW(__sync_val_compare_and_swap),
1002    BUILTIN_ROW(__sync_bool_compare_and_swap),
1003    BUILTIN_ROW(__sync_lock_test_and_set),
1004    BUILTIN_ROW(__sync_lock_release),
1005    BUILTIN_ROW(__sync_swap)
1006  };
1007#undef BUILTIN_ROW
1008
1009  // Determine the index of the size.
1010  unsigned SizeIndex;
1011  switch (Context.getTypeSizeInChars(ValType).getQuantity()) {
1012  case 1: SizeIndex = 0; break;
1013  case 2: SizeIndex = 1; break;
1014  case 4: SizeIndex = 2; break;
1015  case 8: SizeIndex = 3; break;
1016  case 16: SizeIndex = 4; break;
1017  default:
1018    Diag(DRE->getLocStart(), diag::err_atomic_builtin_pointer_size)
1019      << FirstArg->getType() << FirstArg->getSourceRange();
1020    return ExprError();
1021  }
1022
1023  // Each of these builtins has one pointer argument, followed by some number of
1024  // values (0, 1 or 2) followed by a potentially empty varags list of stuff
1025  // that we ignore.  Find out which row of BuiltinIndices to read from as well
1026  // as the number of fixed args.
1027  unsigned BuiltinID = FDecl->getBuiltinID();
1028  unsigned BuiltinIndex, NumFixed = 1;
1029  switch (BuiltinID) {
1030  default: llvm_unreachable("Unknown overloaded atomic builtin!");
1031  case Builtin::BI__sync_fetch_and_add:
1032  case Builtin::BI__sync_fetch_and_add_1:
1033  case Builtin::BI__sync_fetch_and_add_2:
1034  case Builtin::BI__sync_fetch_and_add_4:
1035  case Builtin::BI__sync_fetch_and_add_8:
1036  case Builtin::BI__sync_fetch_and_add_16:
1037    BuiltinIndex = 0;
1038    break;
1039
1040  case Builtin::BI__sync_fetch_and_sub:
1041  case Builtin::BI__sync_fetch_and_sub_1:
1042  case Builtin::BI__sync_fetch_and_sub_2:
1043  case Builtin::BI__sync_fetch_and_sub_4:
1044  case Builtin::BI__sync_fetch_and_sub_8:
1045  case Builtin::BI__sync_fetch_and_sub_16:
1046    BuiltinIndex = 1;
1047    break;
1048
1049  case Builtin::BI__sync_fetch_and_or:
1050  case Builtin::BI__sync_fetch_and_or_1:
1051  case Builtin::BI__sync_fetch_and_or_2:
1052  case Builtin::BI__sync_fetch_and_or_4:
1053  case Builtin::BI__sync_fetch_and_or_8:
1054  case Builtin::BI__sync_fetch_and_or_16:
1055    BuiltinIndex = 2;
1056    break;
1057
1058  case Builtin::BI__sync_fetch_and_and:
1059  case Builtin::BI__sync_fetch_and_and_1:
1060  case Builtin::BI__sync_fetch_and_and_2:
1061  case Builtin::BI__sync_fetch_and_and_4:
1062  case Builtin::BI__sync_fetch_and_and_8:
1063  case Builtin::BI__sync_fetch_and_and_16:
1064    BuiltinIndex = 3;
1065    break;
1066
1067  case Builtin::BI__sync_fetch_and_xor:
1068  case Builtin::BI__sync_fetch_and_xor_1:
1069  case Builtin::BI__sync_fetch_and_xor_2:
1070  case Builtin::BI__sync_fetch_and_xor_4:
1071  case Builtin::BI__sync_fetch_and_xor_8:
1072  case Builtin::BI__sync_fetch_and_xor_16:
1073    BuiltinIndex = 4;
1074    break;
1075
1076  case Builtin::BI__sync_add_and_fetch:
1077  case Builtin::BI__sync_add_and_fetch_1:
1078  case Builtin::BI__sync_add_and_fetch_2:
1079  case Builtin::BI__sync_add_and_fetch_4:
1080  case Builtin::BI__sync_add_and_fetch_8:
1081  case Builtin::BI__sync_add_and_fetch_16:
1082    BuiltinIndex = 5;
1083    break;
1084
1085  case Builtin::BI__sync_sub_and_fetch:
1086  case Builtin::BI__sync_sub_and_fetch_1:
1087  case Builtin::BI__sync_sub_and_fetch_2:
1088  case Builtin::BI__sync_sub_and_fetch_4:
1089  case Builtin::BI__sync_sub_and_fetch_8:
1090  case Builtin::BI__sync_sub_and_fetch_16:
1091    BuiltinIndex = 6;
1092    break;
1093
1094  case Builtin::BI__sync_and_and_fetch:
1095  case Builtin::BI__sync_and_and_fetch_1:
1096  case Builtin::BI__sync_and_and_fetch_2:
1097  case Builtin::BI__sync_and_and_fetch_4:
1098  case Builtin::BI__sync_and_and_fetch_8:
1099  case Builtin::BI__sync_and_and_fetch_16:
1100    BuiltinIndex = 7;
1101    break;
1102
1103  case Builtin::BI__sync_or_and_fetch:
1104  case Builtin::BI__sync_or_and_fetch_1:
1105  case Builtin::BI__sync_or_and_fetch_2:
1106  case Builtin::BI__sync_or_and_fetch_4:
1107  case Builtin::BI__sync_or_and_fetch_8:
1108  case Builtin::BI__sync_or_and_fetch_16:
1109    BuiltinIndex = 8;
1110    break;
1111
1112  case Builtin::BI__sync_xor_and_fetch:
1113  case Builtin::BI__sync_xor_and_fetch_1:
1114  case Builtin::BI__sync_xor_and_fetch_2:
1115  case Builtin::BI__sync_xor_and_fetch_4:
1116  case Builtin::BI__sync_xor_and_fetch_8:
1117  case Builtin::BI__sync_xor_and_fetch_16:
1118    BuiltinIndex = 9;
1119    break;
1120
1121  case Builtin::BI__sync_val_compare_and_swap:
1122  case Builtin::BI__sync_val_compare_and_swap_1:
1123  case Builtin::BI__sync_val_compare_and_swap_2:
1124  case Builtin::BI__sync_val_compare_and_swap_4:
1125  case Builtin::BI__sync_val_compare_and_swap_8:
1126  case Builtin::BI__sync_val_compare_and_swap_16:
1127    BuiltinIndex = 10;
1128    NumFixed = 2;
1129    break;
1130
1131  case Builtin::BI__sync_bool_compare_and_swap:
1132  case Builtin::BI__sync_bool_compare_and_swap_1:
1133  case Builtin::BI__sync_bool_compare_and_swap_2:
1134  case Builtin::BI__sync_bool_compare_and_swap_4:
1135  case Builtin::BI__sync_bool_compare_and_swap_8:
1136  case Builtin::BI__sync_bool_compare_and_swap_16:
1137    BuiltinIndex = 11;
1138    NumFixed = 2;
1139    ResultType = Context.BoolTy;
1140    break;
1141
1142  case Builtin::BI__sync_lock_test_and_set:
1143  case Builtin::BI__sync_lock_test_and_set_1:
1144  case Builtin::BI__sync_lock_test_and_set_2:
1145  case Builtin::BI__sync_lock_test_and_set_4:
1146  case Builtin::BI__sync_lock_test_and_set_8:
1147  case Builtin::BI__sync_lock_test_and_set_16:
1148    BuiltinIndex = 12;
1149    break;
1150
1151  case Builtin::BI__sync_lock_release:
1152  case Builtin::BI__sync_lock_release_1:
1153  case Builtin::BI__sync_lock_release_2:
1154  case Builtin::BI__sync_lock_release_4:
1155  case Builtin::BI__sync_lock_release_8:
1156  case Builtin::BI__sync_lock_release_16:
1157    BuiltinIndex = 13;
1158    NumFixed = 0;
1159    ResultType = Context.VoidTy;
1160    break;
1161
1162  case Builtin::BI__sync_swap:
1163  case Builtin::BI__sync_swap_1:
1164  case Builtin::BI__sync_swap_2:
1165  case Builtin::BI__sync_swap_4:
1166  case Builtin::BI__sync_swap_8:
1167  case Builtin::BI__sync_swap_16:
1168    BuiltinIndex = 14;
1169    break;
1170  }
1171
1172  // Now that we know how many fixed arguments we expect, first check that we
1173  // have at least that many.
1174  if (TheCall->getNumArgs() < 1+NumFixed) {
1175    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args_at_least)
1176      << 0 << 1+NumFixed << TheCall->getNumArgs()
1177      << TheCall->getCallee()->getSourceRange();
1178    return ExprError();
1179  }
1180
1181  // Get the decl for the concrete builtin from this, we can tell what the
1182  // concrete integer type we should convert to is.
1183  unsigned NewBuiltinID = BuiltinIndices[BuiltinIndex][SizeIndex];
1184  const char *NewBuiltinName = Context.BuiltinInfo.GetName(NewBuiltinID);
1185  IdentifierInfo *NewBuiltinII = PP.getIdentifierInfo(NewBuiltinName);
1186  FunctionDecl *NewBuiltinDecl =
1187    cast<FunctionDecl>(LazilyCreateBuiltin(NewBuiltinII, NewBuiltinID,
1188                                           TUScope, false, DRE->getLocStart()));
1189
1190  // The first argument --- the pointer --- has a fixed type; we
1191  // deduce the types of the rest of the arguments accordingly.  Walk
1192  // the remaining arguments, converting them to the deduced value type.
1193  for (unsigned i = 0; i != NumFixed; ++i) {
1194    ExprResult Arg = TheCall->getArg(i+1);
1195
1196    // GCC does an implicit conversion to the pointer or integer ValType.  This
1197    // can fail in some cases (1i -> int**), check for this error case now.
1198    // Initialize the argument.
1199    InitializedEntity Entity = InitializedEntity::InitializeParameter(Context,
1200                                                   ValType, /*consume*/ false);
1201    Arg = PerformCopyInitialization(Entity, SourceLocation(), Arg);
1202    if (Arg.isInvalid())
1203      return ExprError();
1204
1205    // Okay, we have something that *can* be converted to the right type.  Check
1206    // to see if there is a potentially weird extension going on here.  This can
1207    // happen when you do an atomic operation on something like an char* and
1208    // pass in 42.  The 42 gets converted to char.  This is even more strange
1209    // for things like 45.123 -> char, etc.
1210    // FIXME: Do this check.
1211    TheCall->setArg(i+1, Arg.take());
1212  }
1213
1214  ASTContext& Context = this->getASTContext();
1215
1216  // Create a new DeclRefExpr to refer to the new decl.
1217  DeclRefExpr* NewDRE = DeclRefExpr::Create(
1218      Context,
1219      DRE->getQualifierLoc(),
1220      SourceLocation(),
1221      NewBuiltinDecl,
1222      /*enclosing*/ false,
1223      DRE->getLocation(),
1224      NewBuiltinDecl->getType(),
1225      DRE->getValueKind());
1226
1227  // Set the callee in the CallExpr.
1228  // FIXME: This leaks the original parens and implicit casts.
1229  ExprResult PromotedCall = UsualUnaryConversions(NewDRE);
1230  if (PromotedCall.isInvalid())
1231    return ExprError();
1232  TheCall->setCallee(PromotedCall.take());
1233
1234  // Change the result type of the call to match the original value type. This
1235  // is arbitrary, but the codegen for these builtins ins design to handle it
1236  // gracefully.
1237  TheCall->setType(ResultType);
1238
1239  return move(TheCallResult);
1240}
1241
1242/// CheckObjCString - Checks that the argument to the builtin
1243/// CFString constructor is correct
1244/// Note: It might also make sense to do the UTF-16 conversion here (would
1245/// simplify the backend).
1246bool Sema::CheckObjCString(Expr *Arg) {
1247  Arg = Arg->IgnoreParenCasts();
1248  StringLiteral *Literal = dyn_cast<StringLiteral>(Arg);
1249
1250  if (!Literal || !Literal->isAscii()) {
1251    Diag(Arg->getLocStart(), diag::err_cfstring_literal_not_string_constant)
1252      << Arg->getSourceRange();
1253    return true;
1254  }
1255
1256  if (Literal->containsNonAsciiOrNull()) {
1257    StringRef String = Literal->getString();
1258    unsigned NumBytes = String.size();
1259    SmallVector<UTF16, 128> ToBuf(NumBytes);
1260    const UTF8 *FromPtr = (UTF8 *)String.data();
1261    UTF16 *ToPtr = &ToBuf[0];
1262
1263    ConversionResult Result = ConvertUTF8toUTF16(&FromPtr, FromPtr + NumBytes,
1264                                                 &ToPtr, ToPtr + NumBytes,
1265                                                 strictConversion);
1266    // Check for conversion failure.
1267    if (Result != conversionOK)
1268      Diag(Arg->getLocStart(),
1269           diag::warn_cfstring_truncated) << Arg->getSourceRange();
1270  }
1271  return false;
1272}
1273
1274/// SemaBuiltinVAStart - Check the arguments to __builtin_va_start for validity.
1275/// Emit an error and return true on failure, return false on success.
1276bool Sema::SemaBuiltinVAStart(CallExpr *TheCall) {
1277  Expr *Fn = TheCall->getCallee();
1278  if (TheCall->getNumArgs() > 2) {
1279    Diag(TheCall->getArg(2)->getLocStart(),
1280         diag::err_typecheck_call_too_many_args)
1281      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1282      << Fn->getSourceRange()
1283      << SourceRange(TheCall->getArg(2)->getLocStart(),
1284                     (*(TheCall->arg_end()-1))->getLocEnd());
1285    return true;
1286  }
1287
1288  if (TheCall->getNumArgs() < 2) {
1289    return Diag(TheCall->getLocEnd(),
1290      diag::err_typecheck_call_too_few_args_at_least)
1291      << 0 /*function call*/ << 2 << TheCall->getNumArgs();
1292  }
1293
1294  // Type-check the first argument normally.
1295  if (checkBuiltinArgument(*this, TheCall, 0))
1296    return true;
1297
1298  // Determine whether the current function is variadic or not.
1299  BlockScopeInfo *CurBlock = getCurBlock();
1300  bool isVariadic;
1301  if (CurBlock)
1302    isVariadic = CurBlock->TheDecl->isVariadic();
1303  else if (FunctionDecl *FD = getCurFunctionDecl())
1304    isVariadic = FD->isVariadic();
1305  else
1306    isVariadic = getCurMethodDecl()->isVariadic();
1307
1308  if (!isVariadic) {
1309    Diag(Fn->getLocStart(), diag::err_va_start_used_in_non_variadic_function);
1310    return true;
1311  }
1312
1313  // Verify that the second argument to the builtin is the last argument of the
1314  // current function or method.
1315  bool SecondArgIsLastNamedArgument = false;
1316  const Expr *Arg = TheCall->getArg(1)->IgnoreParenCasts();
1317
1318  if (const DeclRefExpr *DR = dyn_cast<DeclRefExpr>(Arg)) {
1319    if (const ParmVarDecl *PV = dyn_cast<ParmVarDecl>(DR->getDecl())) {
1320      // FIXME: This isn't correct for methods (results in bogus warning).
1321      // Get the last formal in the current function.
1322      const ParmVarDecl *LastArg;
1323      if (CurBlock)
1324        LastArg = *(CurBlock->TheDecl->param_end()-1);
1325      else if (FunctionDecl *FD = getCurFunctionDecl())
1326        LastArg = *(FD->param_end()-1);
1327      else
1328        LastArg = *(getCurMethodDecl()->param_end()-1);
1329      SecondArgIsLastNamedArgument = PV == LastArg;
1330    }
1331  }
1332
1333  if (!SecondArgIsLastNamedArgument)
1334    Diag(TheCall->getArg(1)->getLocStart(),
1335         diag::warn_second_parameter_of_va_start_not_last_named_argument);
1336  return false;
1337}
1338
1339/// SemaBuiltinUnorderedCompare - Handle functions like __builtin_isgreater and
1340/// friends.  This is declared to take (...), so we have to check everything.
1341bool Sema::SemaBuiltinUnorderedCompare(CallExpr *TheCall) {
1342  if (TheCall->getNumArgs() < 2)
1343    return Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
1344      << 0 << 2 << TheCall->getNumArgs()/*function call*/;
1345  if (TheCall->getNumArgs() > 2)
1346    return Diag(TheCall->getArg(2)->getLocStart(),
1347                diag::err_typecheck_call_too_many_args)
1348      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1349      << SourceRange(TheCall->getArg(2)->getLocStart(),
1350                     (*(TheCall->arg_end()-1))->getLocEnd());
1351
1352  ExprResult OrigArg0 = TheCall->getArg(0);
1353  ExprResult OrigArg1 = TheCall->getArg(1);
1354
1355  // Do standard promotions between the two arguments, returning their common
1356  // type.
1357  QualType Res = UsualArithmeticConversions(OrigArg0, OrigArg1, false);
1358  if (OrigArg0.isInvalid() || OrigArg1.isInvalid())
1359    return true;
1360
1361  // Make sure any conversions are pushed back into the call; this is
1362  // type safe since unordered compare builtins are declared as "_Bool
1363  // foo(...)".
1364  TheCall->setArg(0, OrigArg0.get());
1365  TheCall->setArg(1, OrigArg1.get());
1366
1367  if (OrigArg0.get()->isTypeDependent() || OrigArg1.get()->isTypeDependent())
1368    return false;
1369
1370  // If the common type isn't a real floating type, then the arguments were
1371  // invalid for this operation.
1372  if (Res.isNull() || !Res->isRealFloatingType())
1373    return Diag(OrigArg0.get()->getLocStart(),
1374                diag::err_typecheck_call_invalid_ordered_compare)
1375      << OrigArg0.get()->getType() << OrigArg1.get()->getType()
1376      << SourceRange(OrigArg0.get()->getLocStart(), OrigArg1.get()->getLocEnd());
1377
1378  return false;
1379}
1380
1381/// SemaBuiltinSemaBuiltinFPClassification - Handle functions like
1382/// __builtin_isnan and friends.  This is declared to take (...), so we have
1383/// to check everything. We expect the last argument to be a floating point
1384/// value.
1385bool Sema::SemaBuiltinFPClassification(CallExpr *TheCall, unsigned NumArgs) {
1386  if (TheCall->getNumArgs() < NumArgs)
1387    return Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
1388      << 0 << NumArgs << TheCall->getNumArgs()/*function call*/;
1389  if (TheCall->getNumArgs() > NumArgs)
1390    return Diag(TheCall->getArg(NumArgs)->getLocStart(),
1391                diag::err_typecheck_call_too_many_args)
1392      << 0 /*function call*/ << NumArgs << TheCall->getNumArgs()
1393      << SourceRange(TheCall->getArg(NumArgs)->getLocStart(),
1394                     (*(TheCall->arg_end()-1))->getLocEnd());
1395
1396  Expr *OrigArg = TheCall->getArg(NumArgs-1);
1397
1398  if (OrigArg->isTypeDependent())
1399    return false;
1400
1401  // This operation requires a non-_Complex floating-point number.
1402  if (!OrigArg->getType()->isRealFloatingType())
1403    return Diag(OrigArg->getLocStart(),
1404                diag::err_typecheck_call_invalid_unary_fp)
1405      << OrigArg->getType() << OrigArg->getSourceRange();
1406
1407  // If this is an implicit conversion from float -> double, remove it.
1408  if (ImplicitCastExpr *Cast = dyn_cast<ImplicitCastExpr>(OrigArg)) {
1409    Expr *CastArg = Cast->getSubExpr();
1410    if (CastArg->getType()->isSpecificBuiltinType(BuiltinType::Float)) {
1411      assert(Cast->getType()->isSpecificBuiltinType(BuiltinType::Double) &&
1412             "promotion from float to double is the only expected cast here");
1413      Cast->setSubExpr(0);
1414      TheCall->setArg(NumArgs-1, CastArg);
1415    }
1416  }
1417
1418  return false;
1419}
1420
1421/// SemaBuiltinShuffleVector - Handle __builtin_shufflevector.
1422// This is declared to take (...), so we have to check everything.
1423ExprResult Sema::SemaBuiltinShuffleVector(CallExpr *TheCall) {
1424  if (TheCall->getNumArgs() < 2)
1425    return ExprError(Diag(TheCall->getLocEnd(),
1426                          diag::err_typecheck_call_too_few_args_at_least)
1427      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1428      << TheCall->getSourceRange());
1429
1430  // Determine which of the following types of shufflevector we're checking:
1431  // 1) unary, vector mask: (lhs, mask)
1432  // 2) binary, vector mask: (lhs, rhs, mask)
1433  // 3) binary, scalar mask: (lhs, rhs, index, ..., index)
1434  QualType resType = TheCall->getArg(0)->getType();
1435  unsigned numElements = 0;
1436
1437  if (!TheCall->getArg(0)->isTypeDependent() &&
1438      !TheCall->getArg(1)->isTypeDependent()) {
1439    QualType LHSType = TheCall->getArg(0)->getType();
1440    QualType RHSType = TheCall->getArg(1)->getType();
1441
1442    if (!LHSType->isVectorType() || !RHSType->isVectorType()) {
1443      Diag(TheCall->getLocStart(), diag::err_shufflevector_non_vector)
1444        << SourceRange(TheCall->getArg(0)->getLocStart(),
1445                       TheCall->getArg(1)->getLocEnd());
1446      return ExprError();
1447    }
1448
1449    numElements = LHSType->getAs<VectorType>()->getNumElements();
1450    unsigned numResElements = TheCall->getNumArgs() - 2;
1451
1452    // Check to see if we have a call with 2 vector arguments, the unary shuffle
1453    // with mask.  If so, verify that RHS is an integer vector type with the
1454    // same number of elts as lhs.
1455    if (TheCall->getNumArgs() == 2) {
1456      if (!RHSType->hasIntegerRepresentation() ||
1457          RHSType->getAs<VectorType>()->getNumElements() != numElements)
1458        Diag(TheCall->getLocStart(), diag::err_shufflevector_incompatible_vector)
1459          << SourceRange(TheCall->getArg(1)->getLocStart(),
1460                         TheCall->getArg(1)->getLocEnd());
1461      numResElements = numElements;
1462    }
1463    else if (!Context.hasSameUnqualifiedType(LHSType, RHSType)) {
1464      Diag(TheCall->getLocStart(), diag::err_shufflevector_incompatible_vector)
1465        << SourceRange(TheCall->getArg(0)->getLocStart(),
1466                       TheCall->getArg(1)->getLocEnd());
1467      return ExprError();
1468    } else if (numElements != numResElements) {
1469      QualType eltType = LHSType->getAs<VectorType>()->getElementType();
1470      resType = Context.getVectorType(eltType, numResElements,
1471                                      VectorType::GenericVector);
1472    }
1473  }
1474
1475  for (unsigned i = 2; i < TheCall->getNumArgs(); i++) {
1476    if (TheCall->getArg(i)->isTypeDependent() ||
1477        TheCall->getArg(i)->isValueDependent())
1478      continue;
1479
1480    llvm::APSInt Result(32);
1481    if (!TheCall->getArg(i)->isIntegerConstantExpr(Result, Context))
1482      return ExprError(Diag(TheCall->getLocStart(),
1483                  diag::err_shufflevector_nonconstant_argument)
1484                << TheCall->getArg(i)->getSourceRange());
1485
1486    if (Result.getActiveBits() > 64 || Result.getZExtValue() >= numElements*2)
1487      return ExprError(Diag(TheCall->getLocStart(),
1488                  diag::err_shufflevector_argument_too_large)
1489               << TheCall->getArg(i)->getSourceRange());
1490  }
1491
1492  SmallVector<Expr*, 32> exprs;
1493
1494  for (unsigned i = 0, e = TheCall->getNumArgs(); i != e; i++) {
1495    exprs.push_back(TheCall->getArg(i));
1496    TheCall->setArg(i, 0);
1497  }
1498
1499  return Owned(new (Context) ShuffleVectorExpr(Context, exprs.begin(),
1500                                            exprs.size(), resType,
1501                                            TheCall->getCallee()->getLocStart(),
1502                                            TheCall->getRParenLoc()));
1503}
1504
1505/// SemaBuiltinPrefetch - Handle __builtin_prefetch.
1506// This is declared to take (const void*, ...) and can take two
1507// optional constant int args.
1508bool Sema::SemaBuiltinPrefetch(CallExpr *TheCall) {
1509  unsigned NumArgs = TheCall->getNumArgs();
1510
1511  if (NumArgs > 3)
1512    return Diag(TheCall->getLocEnd(),
1513             diag::err_typecheck_call_too_many_args_at_most)
1514             << 0 /*function call*/ << 3 << NumArgs
1515             << TheCall->getSourceRange();
1516
1517  // Argument 0 is checked for us and the remaining arguments must be
1518  // constant integers.
1519  for (unsigned i = 1; i != NumArgs; ++i) {
1520    Expr *Arg = TheCall->getArg(i);
1521
1522    // We can't check the value of a dependent argument.
1523    if (Arg->isTypeDependent() || Arg->isValueDependent())
1524      continue;
1525
1526    llvm::APSInt Result;
1527    if (SemaBuiltinConstantArg(TheCall, i, Result))
1528      return true;
1529
1530    // FIXME: gcc issues a warning and rewrites these to 0. These
1531    // seems especially odd for the third argument since the default
1532    // is 3.
1533    if (i == 1) {
1534      if (Result.getLimitedValue() > 1)
1535        return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1536             << "0" << "1" << Arg->getSourceRange();
1537    } else {
1538      if (Result.getLimitedValue() > 3)
1539        return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1540            << "0" << "3" << Arg->getSourceRange();
1541    }
1542  }
1543
1544  return false;
1545}
1546
1547/// SemaBuiltinConstantArg - Handle a check if argument ArgNum of CallExpr
1548/// TheCall is a constant expression.
1549bool Sema::SemaBuiltinConstantArg(CallExpr *TheCall, int ArgNum,
1550                                  llvm::APSInt &Result) {
1551  Expr *Arg = TheCall->getArg(ArgNum);
1552  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
1553  FunctionDecl *FDecl = cast<FunctionDecl>(DRE->getDecl());
1554
1555  if (Arg->isTypeDependent() || Arg->isValueDependent()) return false;
1556
1557  if (!Arg->isIntegerConstantExpr(Result, Context))
1558    return Diag(TheCall->getLocStart(), diag::err_constant_integer_arg_type)
1559                << FDecl->getDeclName() <<  Arg->getSourceRange();
1560
1561  return false;
1562}
1563
1564/// SemaBuiltinObjectSize - Handle __builtin_object_size(void *ptr,
1565/// int type). This simply type checks that type is one of the defined
1566/// constants (0-3).
1567// For compatibility check 0-3, llvm only handles 0 and 2.
1568bool Sema::SemaBuiltinObjectSize(CallExpr *TheCall) {
1569  llvm::APSInt Result;
1570
1571  // We can't check the value of a dependent argument.
1572  if (TheCall->getArg(1)->isTypeDependent() ||
1573      TheCall->getArg(1)->isValueDependent())
1574    return false;
1575
1576  // Check constant-ness first.
1577  if (SemaBuiltinConstantArg(TheCall, 1, Result))
1578    return true;
1579
1580  Expr *Arg = TheCall->getArg(1);
1581  if (Result.getSExtValue() < 0 || Result.getSExtValue() > 3) {
1582    return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1583             << "0" << "3" << SourceRange(Arg->getLocStart(), Arg->getLocEnd());
1584  }
1585
1586  return false;
1587}
1588
1589/// SemaBuiltinLongjmp - Handle __builtin_longjmp(void *env[5], int val).
1590/// This checks that val is a constant 1.
1591bool Sema::SemaBuiltinLongjmp(CallExpr *TheCall) {
1592  Expr *Arg = TheCall->getArg(1);
1593  llvm::APSInt Result;
1594
1595  // TODO: This is less than ideal. Overload this to take a value.
1596  if (SemaBuiltinConstantArg(TheCall, 1, Result))
1597    return true;
1598
1599  if (Result != 1)
1600    return Diag(TheCall->getLocStart(), diag::err_builtin_longjmp_invalid_val)
1601             << SourceRange(Arg->getLocStart(), Arg->getLocEnd());
1602
1603  return false;
1604}
1605
1606// Determine if an expression is a string literal or constant string.
1607// If this function returns false on the arguments to a function expecting a
1608// format string, we will usually need to emit a warning.
1609// True string literals are then checked by CheckFormatString.
1610Sema::StringLiteralCheckType
1611Sema::checkFormatStringExpr(const Expr *E, Expr **Args,
1612                            unsigned NumArgs, bool HasVAListArg,
1613                            unsigned format_idx, unsigned firstDataArg,
1614                            FormatStringType Type, VariadicCallType CallType,
1615                            bool inFunctionCall) {
1616 tryAgain:
1617  if (E->isTypeDependent() || E->isValueDependent())
1618    return SLCT_NotALiteral;
1619
1620  E = E->IgnoreParenCasts();
1621
1622  if (E->isNullPointerConstant(Context, Expr::NPC_ValueDependentIsNotNull))
1623    // Technically -Wformat-nonliteral does not warn about this case.
1624    // The behavior of printf and friends in this case is implementation
1625    // dependent.  Ideally if the format string cannot be null then
1626    // it should have a 'nonnull' attribute in the function prototype.
1627    return SLCT_CheckedLiteral;
1628
1629  switch (E->getStmtClass()) {
1630  case Stmt::BinaryConditionalOperatorClass:
1631  case Stmt::ConditionalOperatorClass: {
1632    // The expression is a literal if both sub-expressions were, and it was
1633    // completely checked only if both sub-expressions were checked.
1634    const AbstractConditionalOperator *C =
1635        cast<AbstractConditionalOperator>(E);
1636    StringLiteralCheckType Left =
1637        checkFormatStringExpr(C->getTrueExpr(), Args, NumArgs,
1638                              HasVAListArg, format_idx, firstDataArg,
1639                              Type, CallType, inFunctionCall);
1640    if (Left == SLCT_NotALiteral)
1641      return SLCT_NotALiteral;
1642    StringLiteralCheckType Right =
1643        checkFormatStringExpr(C->getFalseExpr(), Args, NumArgs,
1644                              HasVAListArg, format_idx, firstDataArg,
1645                              Type, CallType, inFunctionCall);
1646    return Left < Right ? Left : Right;
1647  }
1648
1649  case Stmt::ImplicitCastExprClass: {
1650    E = cast<ImplicitCastExpr>(E)->getSubExpr();
1651    goto tryAgain;
1652  }
1653
1654  case Stmt::OpaqueValueExprClass:
1655    if (const Expr *src = cast<OpaqueValueExpr>(E)->getSourceExpr()) {
1656      E = src;
1657      goto tryAgain;
1658    }
1659    return SLCT_NotALiteral;
1660
1661  case Stmt::PredefinedExprClass:
1662    // While __func__, etc., are technically not string literals, they
1663    // cannot contain format specifiers and thus are not a security
1664    // liability.
1665    return SLCT_UncheckedLiteral;
1666
1667  case Stmt::DeclRefExprClass: {
1668    const DeclRefExpr *DR = cast<DeclRefExpr>(E);
1669
1670    // As an exception, do not flag errors for variables binding to
1671    // const string literals.
1672    if (const VarDecl *VD = dyn_cast<VarDecl>(DR->getDecl())) {
1673      bool isConstant = false;
1674      QualType T = DR->getType();
1675
1676      if (const ArrayType *AT = Context.getAsArrayType(T)) {
1677        isConstant = AT->getElementType().isConstant(Context);
1678      } else if (const PointerType *PT = T->getAs<PointerType>()) {
1679        isConstant = T.isConstant(Context) &&
1680                     PT->getPointeeType().isConstant(Context);
1681      } else if (T->isObjCObjectPointerType()) {
1682        // In ObjC, there is usually no "const ObjectPointer" type,
1683        // so don't check if the pointee type is constant.
1684        isConstant = T.isConstant(Context);
1685      }
1686
1687      if (isConstant) {
1688        if (const Expr *Init = VD->getAnyInitializer()) {
1689          // Look through initializers like const char c[] = { "foo" }
1690          if (const InitListExpr *InitList = dyn_cast<InitListExpr>(Init)) {
1691            if (InitList->isStringLiteralInit())
1692              Init = InitList->getInit(0)->IgnoreParenImpCasts();
1693          }
1694          return checkFormatStringExpr(Init, Args, NumArgs,
1695                                       HasVAListArg, format_idx,
1696                                       firstDataArg, Type, CallType,
1697                                       /*inFunctionCall*/false);
1698        }
1699      }
1700
1701      // For vprintf* functions (i.e., HasVAListArg==true), we add a
1702      // special check to see if the format string is a function parameter
1703      // of the function calling the printf function.  If the function
1704      // has an attribute indicating it is a printf-like function, then we
1705      // should suppress warnings concerning non-literals being used in a call
1706      // to a vprintf function.  For example:
1707      //
1708      // void
1709      // logmessage(char const *fmt __attribute__ (format (printf, 1, 2)), ...){
1710      //      va_list ap;
1711      //      va_start(ap, fmt);
1712      //      vprintf(fmt, ap);  // Do NOT emit a warning about "fmt".
1713      //      ...
1714      //
1715      if (HasVAListArg) {
1716        if (const ParmVarDecl *PV = dyn_cast<ParmVarDecl>(VD)) {
1717          if (const NamedDecl *ND = dyn_cast<NamedDecl>(PV->getDeclContext())) {
1718            int PVIndex = PV->getFunctionScopeIndex() + 1;
1719            for (specific_attr_iterator<FormatAttr>
1720                 i = ND->specific_attr_begin<FormatAttr>(),
1721                 e = ND->specific_attr_end<FormatAttr>(); i != e ; ++i) {
1722              FormatAttr *PVFormat = *i;
1723              // adjust for implicit parameter
1724              if (const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(ND))
1725                if (MD->isInstance())
1726                  ++PVIndex;
1727              // We also check if the formats are compatible.
1728              // We can't pass a 'scanf' string to a 'printf' function.
1729              if (PVIndex == PVFormat->getFormatIdx() &&
1730                  Type == GetFormatStringType(PVFormat))
1731                return SLCT_UncheckedLiteral;
1732            }
1733          }
1734        }
1735      }
1736    }
1737
1738    return SLCT_NotALiteral;
1739  }
1740
1741  case Stmt::CallExprClass:
1742  case Stmt::CXXMemberCallExprClass: {
1743    const CallExpr *CE = cast<CallExpr>(E);
1744    if (const NamedDecl *ND = dyn_cast_or_null<NamedDecl>(CE->getCalleeDecl())) {
1745      if (const FormatArgAttr *FA = ND->getAttr<FormatArgAttr>()) {
1746        unsigned ArgIndex = FA->getFormatIdx();
1747        if (const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(ND))
1748          if (MD->isInstance())
1749            --ArgIndex;
1750        const Expr *Arg = CE->getArg(ArgIndex - 1);
1751
1752        return checkFormatStringExpr(Arg, Args, NumArgs,
1753                                     HasVAListArg, format_idx, firstDataArg,
1754                                     Type, CallType, inFunctionCall);
1755      } else if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(ND)) {
1756        unsigned BuiltinID = FD->getBuiltinID();
1757        if (BuiltinID == Builtin::BI__builtin___CFStringMakeConstantString ||
1758            BuiltinID == Builtin::BI__builtin___NSStringMakeConstantString) {
1759          const Expr *Arg = CE->getArg(0);
1760          return checkFormatStringExpr(Arg, Args, NumArgs,
1761                                       HasVAListArg, format_idx,
1762                                       firstDataArg, Type, CallType,
1763                                       inFunctionCall);
1764        }
1765      }
1766    }
1767
1768    return SLCT_NotALiteral;
1769  }
1770  case Stmt::ObjCStringLiteralClass:
1771  case Stmt::StringLiteralClass: {
1772    const StringLiteral *StrE = NULL;
1773
1774    if (const ObjCStringLiteral *ObjCFExpr = dyn_cast<ObjCStringLiteral>(E))
1775      StrE = ObjCFExpr->getString();
1776    else
1777      StrE = cast<StringLiteral>(E);
1778
1779    if (StrE) {
1780      CheckFormatString(StrE, E, Args, NumArgs, HasVAListArg, format_idx,
1781                        firstDataArg, Type, inFunctionCall, CallType);
1782      return SLCT_CheckedLiteral;
1783    }
1784
1785    return SLCT_NotALiteral;
1786  }
1787
1788  default:
1789    return SLCT_NotALiteral;
1790  }
1791}
1792
1793void
1794Sema::CheckNonNullArguments(const NonNullAttr *NonNull,
1795                            const Expr * const *ExprArgs,
1796                            SourceLocation CallSiteLoc) {
1797  for (NonNullAttr::args_iterator i = NonNull->args_begin(),
1798                                  e = NonNull->args_end();
1799       i != e; ++i) {
1800    const Expr *ArgExpr = ExprArgs[*i];
1801    if (ArgExpr->isNullPointerConstant(Context,
1802                                       Expr::NPC_ValueDependentIsNotNull))
1803      Diag(CallSiteLoc, diag::warn_null_arg) << ArgExpr->getSourceRange();
1804  }
1805}
1806
1807Sema::FormatStringType Sema::GetFormatStringType(const FormatAttr *Format) {
1808  return llvm::StringSwitch<FormatStringType>(Format->getType())
1809  .Case("scanf", FST_Scanf)
1810  .Cases("printf", "printf0", FST_Printf)
1811  .Cases("NSString", "CFString", FST_NSString)
1812  .Case("strftime", FST_Strftime)
1813  .Case("strfmon", FST_Strfmon)
1814  .Cases("kprintf", "cmn_err", "vcmn_err", "zcmn_err", FST_Kprintf)
1815  .Default(FST_Unknown);
1816}
1817
1818/// CheckFormatArguments - Check calls to printf and scanf (and similar
1819/// functions) for correct use of format strings.
1820/// Returns true if a format string has been fully checked.
1821bool Sema::CheckFormatArguments(const FormatAttr *Format, Expr **Args,
1822                                unsigned NumArgs, bool IsCXXMember,
1823                                VariadicCallType CallType,
1824                                SourceLocation Loc, SourceRange Range) {
1825  FormatStringInfo FSI;
1826  if (getFormatStringInfo(Format, IsCXXMember, &FSI))
1827    return CheckFormatArguments(Args, NumArgs, FSI.HasVAListArg, FSI.FormatIdx,
1828                                FSI.FirstDataArg, GetFormatStringType(Format),
1829                                CallType, Loc, Range);
1830  return false;
1831}
1832
1833bool Sema::CheckFormatArguments(Expr **Args, unsigned NumArgs,
1834                                bool HasVAListArg, unsigned format_idx,
1835                                unsigned firstDataArg, FormatStringType Type,
1836                                VariadicCallType CallType,
1837                                SourceLocation Loc, SourceRange Range) {
1838  // CHECK: printf/scanf-like function is called with no format string.
1839  if (format_idx >= NumArgs) {
1840    Diag(Loc, diag::warn_missing_format_string) << Range;
1841    return false;
1842  }
1843
1844  const Expr *OrigFormatExpr = Args[format_idx]->IgnoreParenCasts();
1845
1846  // CHECK: format string is not a string literal.
1847  //
1848  // Dynamically generated format strings are difficult to
1849  // automatically vet at compile time.  Requiring that format strings
1850  // are string literals: (1) permits the checking of format strings by
1851  // the compiler and thereby (2) can practically remove the source of
1852  // many format string exploits.
1853
1854  // Format string can be either ObjC string (e.g. @"%d") or
1855  // C string (e.g. "%d")
1856  // ObjC string uses the same format specifiers as C string, so we can use
1857  // the same format string checking logic for both ObjC and C strings.
1858  StringLiteralCheckType CT =
1859      checkFormatStringExpr(OrigFormatExpr, Args, NumArgs, HasVAListArg,
1860                            format_idx, firstDataArg, Type, CallType);
1861  if (CT != SLCT_NotALiteral)
1862    // Literal format string found, check done!
1863    return CT == SLCT_CheckedLiteral;
1864
1865  // Strftime is particular as it always uses a single 'time' argument,
1866  // so it is safe to pass a non-literal string.
1867  if (Type == FST_Strftime)
1868    return false;
1869
1870  // Do not emit diag when the string param is a macro expansion and the
1871  // format is either NSString or CFString. This is a hack to prevent
1872  // diag when using the NSLocalizedString and CFCopyLocalizedString macros
1873  // which are usually used in place of NS and CF string literals.
1874  if (Type == FST_NSString &&
1875      SourceMgr.isInSystemMacro(Args[format_idx]->getLocStart()))
1876    return false;
1877
1878  // If there are no arguments specified, warn with -Wformat-security, otherwise
1879  // warn only with -Wformat-nonliteral.
1880  if (NumArgs == format_idx+1)
1881    Diag(Args[format_idx]->getLocStart(),
1882         diag::warn_format_nonliteral_noargs)
1883      << OrigFormatExpr->getSourceRange();
1884  else
1885    Diag(Args[format_idx]->getLocStart(),
1886         diag::warn_format_nonliteral)
1887           << OrigFormatExpr->getSourceRange();
1888  return false;
1889}
1890
1891namespace {
1892class CheckFormatHandler : public analyze_format_string::FormatStringHandler {
1893protected:
1894  Sema &S;
1895  const StringLiteral *FExpr;
1896  const Expr *OrigFormatExpr;
1897  const unsigned FirstDataArg;
1898  const unsigned NumDataArgs;
1899  const char *Beg; // Start of format string.
1900  const bool HasVAListArg;
1901  const Expr * const *Args;
1902  const unsigned NumArgs;
1903  unsigned FormatIdx;
1904  llvm::BitVector CoveredArgs;
1905  bool usesPositionalArgs;
1906  bool atFirstArg;
1907  bool inFunctionCall;
1908  Sema::VariadicCallType CallType;
1909public:
1910  CheckFormatHandler(Sema &s, const StringLiteral *fexpr,
1911                     const Expr *origFormatExpr, unsigned firstDataArg,
1912                     unsigned numDataArgs, const char *beg, bool hasVAListArg,
1913                     Expr **args, unsigned numArgs,
1914                     unsigned formatIdx, bool inFunctionCall,
1915                     Sema::VariadicCallType callType)
1916    : S(s), FExpr(fexpr), OrigFormatExpr(origFormatExpr),
1917      FirstDataArg(firstDataArg), NumDataArgs(numDataArgs),
1918      Beg(beg), HasVAListArg(hasVAListArg),
1919      Args(args), NumArgs(numArgs), FormatIdx(formatIdx),
1920      usesPositionalArgs(false), atFirstArg(true),
1921      inFunctionCall(inFunctionCall), CallType(callType) {
1922        CoveredArgs.resize(numDataArgs);
1923        CoveredArgs.reset();
1924      }
1925
1926  void DoneProcessing();
1927
1928  void HandleIncompleteSpecifier(const char *startSpecifier,
1929                                 unsigned specifierLen);
1930
1931  void HandleNonStandardLengthModifier(
1932      const analyze_format_string::LengthModifier &LM,
1933      const char *startSpecifier, unsigned specifierLen);
1934
1935  void HandleNonStandardConversionSpecifier(
1936      const analyze_format_string::ConversionSpecifier &CS,
1937      const char *startSpecifier, unsigned specifierLen);
1938
1939  void HandleNonStandardConversionSpecification(
1940      const analyze_format_string::LengthModifier &LM,
1941      const analyze_format_string::ConversionSpecifier &CS,
1942      const char *startSpecifier, unsigned specifierLen);
1943
1944  virtual void HandlePosition(const char *startPos, unsigned posLen);
1945
1946  virtual void HandleInvalidPosition(const char *startSpecifier,
1947                                     unsigned specifierLen,
1948                                     analyze_format_string::PositionContext p);
1949
1950  virtual void HandleZeroPosition(const char *startPos, unsigned posLen);
1951
1952  void HandleNullChar(const char *nullCharacter);
1953
1954  template <typename Range>
1955  static void EmitFormatDiagnostic(Sema &S, bool inFunctionCall,
1956                                   const Expr *ArgumentExpr,
1957                                   PartialDiagnostic PDiag,
1958                                   SourceLocation StringLoc,
1959                                   bool IsStringLocation, Range StringRange,
1960                                   FixItHint Fixit = FixItHint());
1961
1962protected:
1963  bool HandleInvalidConversionSpecifier(unsigned argIndex, SourceLocation Loc,
1964                                        const char *startSpec,
1965                                        unsigned specifierLen,
1966                                        const char *csStart, unsigned csLen);
1967
1968  void HandlePositionalNonpositionalArgs(SourceLocation Loc,
1969                                         const char *startSpec,
1970                                         unsigned specifierLen);
1971
1972  SourceRange getFormatStringRange();
1973  CharSourceRange getSpecifierRange(const char *startSpecifier,
1974                                    unsigned specifierLen);
1975  SourceLocation getLocationOfByte(const char *x);
1976
1977  const Expr *getDataArg(unsigned i) const;
1978
1979  bool CheckNumArgs(const analyze_format_string::FormatSpecifier &FS,
1980                    const analyze_format_string::ConversionSpecifier &CS,
1981                    const char *startSpecifier, unsigned specifierLen,
1982                    unsigned argIndex);
1983
1984  template <typename Range>
1985  void EmitFormatDiagnostic(PartialDiagnostic PDiag, SourceLocation StringLoc,
1986                            bool IsStringLocation, Range StringRange,
1987                            FixItHint Fixit = FixItHint());
1988
1989  void CheckPositionalAndNonpositionalArgs(
1990      const analyze_format_string::FormatSpecifier *FS);
1991};
1992}
1993
1994SourceRange CheckFormatHandler::getFormatStringRange() {
1995  return OrigFormatExpr->getSourceRange();
1996}
1997
1998CharSourceRange CheckFormatHandler::
1999getSpecifierRange(const char *startSpecifier, unsigned specifierLen) {
2000  SourceLocation Start = getLocationOfByte(startSpecifier);
2001  SourceLocation End   = getLocationOfByte(startSpecifier + specifierLen - 1);
2002
2003  // Advance the end SourceLocation by one due to half-open ranges.
2004  End = End.getLocWithOffset(1);
2005
2006  return CharSourceRange::getCharRange(Start, End);
2007}
2008
2009SourceLocation CheckFormatHandler::getLocationOfByte(const char *x) {
2010  return S.getLocationOfStringLiteralByte(FExpr, x - Beg);
2011}
2012
2013void CheckFormatHandler::HandleIncompleteSpecifier(const char *startSpecifier,
2014                                                   unsigned specifierLen){
2015  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_incomplete_specifier),
2016                       getLocationOfByte(startSpecifier),
2017                       /*IsStringLocation*/true,
2018                       getSpecifierRange(startSpecifier, specifierLen));
2019}
2020
2021void CheckFormatHandler::HandleNonStandardLengthModifier(
2022    const analyze_format_string::LengthModifier &LM,
2023    const char *startSpecifier, unsigned specifierLen) {
2024  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard) << LM.toString()
2025                       << 0,
2026                       getLocationOfByte(LM.getStart()),
2027                       /*IsStringLocation*/true,
2028                       getSpecifierRange(startSpecifier, specifierLen));
2029}
2030
2031void CheckFormatHandler::HandleNonStandardConversionSpecifier(
2032    const analyze_format_string::ConversionSpecifier &CS,
2033    const char *startSpecifier, unsigned specifierLen) {
2034  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard) << CS.toString()
2035                       << 1,
2036                       getLocationOfByte(CS.getStart()),
2037                       /*IsStringLocation*/true,
2038                       getSpecifierRange(startSpecifier, specifierLen));
2039}
2040
2041void CheckFormatHandler::HandleNonStandardConversionSpecification(
2042    const analyze_format_string::LengthModifier &LM,
2043    const analyze_format_string::ConversionSpecifier &CS,
2044    const char *startSpecifier, unsigned specifierLen) {
2045  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard_conversion_spec)
2046                       << LM.toString() << CS.toString(),
2047                       getLocationOfByte(LM.getStart()),
2048                       /*IsStringLocation*/true,
2049                       getSpecifierRange(startSpecifier, specifierLen));
2050}
2051
2052void CheckFormatHandler::HandlePosition(const char *startPos,
2053                                        unsigned posLen) {
2054  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard_positional_arg),
2055                               getLocationOfByte(startPos),
2056                               /*IsStringLocation*/true,
2057                               getSpecifierRange(startPos, posLen));
2058}
2059
2060void
2061CheckFormatHandler::HandleInvalidPosition(const char *startPos, unsigned posLen,
2062                                     analyze_format_string::PositionContext p) {
2063  EmitFormatDiagnostic(S.PDiag(diag::warn_format_invalid_positional_specifier)
2064                         << (unsigned) p,
2065                       getLocationOfByte(startPos), /*IsStringLocation*/true,
2066                       getSpecifierRange(startPos, posLen));
2067}
2068
2069void CheckFormatHandler::HandleZeroPosition(const char *startPos,
2070                                            unsigned posLen) {
2071  EmitFormatDiagnostic(S.PDiag(diag::warn_format_zero_positional_specifier),
2072                               getLocationOfByte(startPos),
2073                               /*IsStringLocation*/true,
2074                               getSpecifierRange(startPos, posLen));
2075}
2076
2077void CheckFormatHandler::HandleNullChar(const char *nullCharacter) {
2078  if (!isa<ObjCStringLiteral>(OrigFormatExpr)) {
2079    // The presence of a null character is likely an error.
2080    EmitFormatDiagnostic(
2081      S.PDiag(diag::warn_printf_format_string_contains_null_char),
2082      getLocationOfByte(nullCharacter), /*IsStringLocation*/true,
2083      getFormatStringRange());
2084  }
2085}
2086
2087// Note that this may return NULL if there was an error parsing or building
2088// one of the argument expressions.
2089const Expr *CheckFormatHandler::getDataArg(unsigned i) const {
2090  return Args[FirstDataArg + i];
2091}
2092
2093void CheckFormatHandler::DoneProcessing() {
2094    // Does the number of data arguments exceed the number of
2095    // format conversions in the format string?
2096  if (!HasVAListArg) {
2097      // Find any arguments that weren't covered.
2098    CoveredArgs.flip();
2099    signed notCoveredArg = CoveredArgs.find_first();
2100    if (notCoveredArg >= 0) {
2101      assert((unsigned)notCoveredArg < NumDataArgs);
2102      if (const Expr *E = getDataArg((unsigned) notCoveredArg)) {
2103        SourceLocation Loc = E->getLocStart();
2104        if (!S.getSourceManager().isInSystemMacro(Loc)) {
2105          EmitFormatDiagnostic(S.PDiag(diag::warn_printf_data_arg_not_used),
2106                               Loc, /*IsStringLocation*/false,
2107                               getFormatStringRange());
2108        }
2109      }
2110    }
2111  }
2112}
2113
2114bool
2115CheckFormatHandler::HandleInvalidConversionSpecifier(unsigned argIndex,
2116                                                     SourceLocation Loc,
2117                                                     const char *startSpec,
2118                                                     unsigned specifierLen,
2119                                                     const char *csStart,
2120                                                     unsigned csLen) {
2121
2122  bool keepGoing = true;
2123  if (argIndex < NumDataArgs) {
2124    // Consider the argument coverered, even though the specifier doesn't
2125    // make sense.
2126    CoveredArgs.set(argIndex);
2127  }
2128  else {
2129    // If argIndex exceeds the number of data arguments we
2130    // don't issue a warning because that is just a cascade of warnings (and
2131    // they may have intended '%%' anyway). We don't want to continue processing
2132    // the format string after this point, however, as we will like just get
2133    // gibberish when trying to match arguments.
2134    keepGoing = false;
2135  }
2136
2137  EmitFormatDiagnostic(S.PDiag(diag::warn_format_invalid_conversion)
2138                         << StringRef(csStart, csLen),
2139                       Loc, /*IsStringLocation*/true,
2140                       getSpecifierRange(startSpec, specifierLen));
2141
2142  return keepGoing;
2143}
2144
2145void
2146CheckFormatHandler::HandlePositionalNonpositionalArgs(SourceLocation Loc,
2147                                                      const char *startSpec,
2148                                                      unsigned specifierLen) {
2149  EmitFormatDiagnostic(
2150    S.PDiag(diag::warn_format_mix_positional_nonpositional_args),
2151    Loc, /*isStringLoc*/true, getSpecifierRange(startSpec, specifierLen));
2152}
2153
2154bool
2155CheckFormatHandler::CheckNumArgs(
2156  const analyze_format_string::FormatSpecifier &FS,
2157  const analyze_format_string::ConversionSpecifier &CS,
2158  const char *startSpecifier, unsigned specifierLen, unsigned argIndex) {
2159
2160  if (argIndex >= NumDataArgs) {
2161    PartialDiagnostic PDiag = FS.usesPositionalArg()
2162      ? (S.PDiag(diag::warn_printf_positional_arg_exceeds_data_args)
2163           << (argIndex+1) << NumDataArgs)
2164      : S.PDiag(diag::warn_printf_insufficient_data_args);
2165    EmitFormatDiagnostic(
2166      PDiag, getLocationOfByte(CS.getStart()), /*IsStringLocation*/true,
2167      getSpecifierRange(startSpecifier, specifierLen));
2168    return false;
2169  }
2170  return true;
2171}
2172
2173template<typename Range>
2174void CheckFormatHandler::EmitFormatDiagnostic(PartialDiagnostic PDiag,
2175                                              SourceLocation Loc,
2176                                              bool IsStringLocation,
2177                                              Range StringRange,
2178                                              FixItHint FixIt) {
2179  EmitFormatDiagnostic(S, inFunctionCall, Args[FormatIdx], PDiag,
2180                       Loc, IsStringLocation, StringRange, FixIt);
2181}
2182
2183/// \brief If the format string is not within the funcion call, emit a note
2184/// so that the function call and string are in diagnostic messages.
2185///
2186/// \param inFunctionCall if true, the format string is within the function
2187/// call and only one diagnostic message will be produced.  Otherwise, an
2188/// extra note will be emitted pointing to location of the format string.
2189///
2190/// \param ArgumentExpr the expression that is passed as the format string
2191/// argument in the function call.  Used for getting locations when two
2192/// diagnostics are emitted.
2193///
2194/// \param PDiag the callee should already have provided any strings for the
2195/// diagnostic message.  This function only adds locations and fixits
2196/// to diagnostics.
2197///
2198/// \param Loc primary location for diagnostic.  If two diagnostics are
2199/// required, one will be at Loc and a new SourceLocation will be created for
2200/// the other one.
2201///
2202/// \param IsStringLocation if true, Loc points to the format string should be
2203/// used for the note.  Otherwise, Loc points to the argument list and will
2204/// be used with PDiag.
2205///
2206/// \param StringRange some or all of the string to highlight.  This is
2207/// templated so it can accept either a CharSourceRange or a SourceRange.
2208///
2209/// \param Fixit optional fix it hint for the format string.
2210template<typename Range>
2211void CheckFormatHandler::EmitFormatDiagnostic(Sema &S, bool InFunctionCall,
2212                                              const Expr *ArgumentExpr,
2213                                              PartialDiagnostic PDiag,
2214                                              SourceLocation Loc,
2215                                              bool IsStringLocation,
2216                                              Range StringRange,
2217                                              FixItHint FixIt) {
2218  if (InFunctionCall)
2219    S.Diag(Loc, PDiag) << StringRange << FixIt;
2220  else {
2221    S.Diag(IsStringLocation ? ArgumentExpr->getExprLoc() : Loc, PDiag)
2222      << ArgumentExpr->getSourceRange();
2223    S.Diag(IsStringLocation ? Loc : StringRange.getBegin(),
2224           diag::note_format_string_defined)
2225      << StringRange << FixIt;
2226  }
2227}
2228
2229//===--- CHECK: Printf format string checking ------------------------------===//
2230
2231namespace {
2232class CheckPrintfHandler : public CheckFormatHandler {
2233  bool ObjCContext;
2234public:
2235  CheckPrintfHandler(Sema &s, const StringLiteral *fexpr,
2236                     const Expr *origFormatExpr, unsigned firstDataArg,
2237                     unsigned numDataArgs, bool isObjC,
2238                     const char *beg, bool hasVAListArg,
2239                     Expr **Args, unsigned NumArgs,
2240                     unsigned formatIdx, bool inFunctionCall,
2241                     Sema::VariadicCallType CallType)
2242  : CheckFormatHandler(s, fexpr, origFormatExpr, firstDataArg,
2243                       numDataArgs, beg, hasVAListArg, Args, NumArgs,
2244                       formatIdx, inFunctionCall, CallType), ObjCContext(isObjC)
2245  {}
2246
2247
2248  bool HandleInvalidPrintfConversionSpecifier(
2249                                      const analyze_printf::PrintfSpecifier &FS,
2250                                      const char *startSpecifier,
2251                                      unsigned specifierLen);
2252
2253  bool HandlePrintfSpecifier(const analyze_printf::PrintfSpecifier &FS,
2254                             const char *startSpecifier,
2255                             unsigned specifierLen);
2256  bool checkFormatExpr(const analyze_printf::PrintfSpecifier &FS,
2257                       const char *StartSpecifier,
2258                       unsigned SpecifierLen,
2259                       const Expr *E);
2260
2261  bool HandleAmount(const analyze_format_string::OptionalAmount &Amt, unsigned k,
2262                    const char *startSpecifier, unsigned specifierLen);
2263  void HandleInvalidAmount(const analyze_printf::PrintfSpecifier &FS,
2264                           const analyze_printf::OptionalAmount &Amt,
2265                           unsigned type,
2266                           const char *startSpecifier, unsigned specifierLen);
2267  void HandleFlag(const analyze_printf::PrintfSpecifier &FS,
2268                  const analyze_printf::OptionalFlag &flag,
2269                  const char *startSpecifier, unsigned specifierLen);
2270  void HandleIgnoredFlag(const analyze_printf::PrintfSpecifier &FS,
2271                         const analyze_printf::OptionalFlag &ignoredFlag,
2272                         const analyze_printf::OptionalFlag &flag,
2273                         const char *startSpecifier, unsigned specifierLen);
2274  bool checkForCStrMembers(const analyze_printf::ArgType &AT,
2275                           const Expr *E, const CharSourceRange &CSR);
2276
2277};
2278}
2279
2280bool CheckPrintfHandler::HandleInvalidPrintfConversionSpecifier(
2281                                      const analyze_printf::PrintfSpecifier &FS,
2282                                      const char *startSpecifier,
2283                                      unsigned specifierLen) {
2284  const analyze_printf::PrintfConversionSpecifier &CS =
2285    FS.getConversionSpecifier();
2286
2287  return HandleInvalidConversionSpecifier(FS.getArgIndex(),
2288                                          getLocationOfByte(CS.getStart()),
2289                                          startSpecifier, specifierLen,
2290                                          CS.getStart(), CS.getLength());
2291}
2292
2293bool CheckPrintfHandler::HandleAmount(
2294                               const analyze_format_string::OptionalAmount &Amt,
2295                               unsigned k, const char *startSpecifier,
2296                               unsigned specifierLen) {
2297
2298  if (Amt.hasDataArgument()) {
2299    if (!HasVAListArg) {
2300      unsigned argIndex = Amt.getArgIndex();
2301      if (argIndex >= NumDataArgs) {
2302        EmitFormatDiagnostic(S.PDiag(diag::warn_printf_asterisk_missing_arg)
2303                               << k,
2304                             getLocationOfByte(Amt.getStart()),
2305                             /*IsStringLocation*/true,
2306                             getSpecifierRange(startSpecifier, specifierLen));
2307        // Don't do any more checking.  We will just emit
2308        // spurious errors.
2309        return false;
2310      }
2311
2312      // Type check the data argument.  It should be an 'int'.
2313      // Although not in conformance with C99, we also allow the argument to be
2314      // an 'unsigned int' as that is a reasonably safe case.  GCC also
2315      // doesn't emit a warning for that case.
2316      CoveredArgs.set(argIndex);
2317      const Expr *Arg = getDataArg(argIndex);
2318      if (!Arg)
2319        return false;
2320
2321      QualType T = Arg->getType();
2322
2323      const analyze_printf::ArgType &AT = Amt.getArgType(S.Context);
2324      assert(AT.isValid());
2325
2326      if (!AT.matchesType(S.Context, T)) {
2327        EmitFormatDiagnostic(S.PDiag(diag::warn_printf_asterisk_wrong_type)
2328                               << k << AT.getRepresentativeTypeName(S.Context)
2329                               << T << Arg->getSourceRange(),
2330                             getLocationOfByte(Amt.getStart()),
2331                             /*IsStringLocation*/true,
2332                             getSpecifierRange(startSpecifier, specifierLen));
2333        // Don't do any more checking.  We will just emit
2334        // spurious errors.
2335        return false;
2336      }
2337    }
2338  }
2339  return true;
2340}
2341
2342void CheckPrintfHandler::HandleInvalidAmount(
2343                                      const analyze_printf::PrintfSpecifier &FS,
2344                                      const analyze_printf::OptionalAmount &Amt,
2345                                      unsigned type,
2346                                      const char *startSpecifier,
2347                                      unsigned specifierLen) {
2348  const analyze_printf::PrintfConversionSpecifier &CS =
2349    FS.getConversionSpecifier();
2350
2351  FixItHint fixit =
2352    Amt.getHowSpecified() == analyze_printf::OptionalAmount::Constant
2353      ? FixItHint::CreateRemoval(getSpecifierRange(Amt.getStart(),
2354                                 Amt.getConstantLength()))
2355      : FixItHint();
2356
2357  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_nonsensical_optional_amount)
2358                         << type << CS.toString(),
2359                       getLocationOfByte(Amt.getStart()),
2360                       /*IsStringLocation*/true,
2361                       getSpecifierRange(startSpecifier, specifierLen),
2362                       fixit);
2363}
2364
2365void CheckPrintfHandler::HandleFlag(const analyze_printf::PrintfSpecifier &FS,
2366                                    const analyze_printf::OptionalFlag &flag,
2367                                    const char *startSpecifier,
2368                                    unsigned specifierLen) {
2369  // Warn about pointless flag with a fixit removal.
2370  const analyze_printf::PrintfConversionSpecifier &CS =
2371    FS.getConversionSpecifier();
2372  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_nonsensical_flag)
2373                         << flag.toString() << CS.toString(),
2374                       getLocationOfByte(flag.getPosition()),
2375                       /*IsStringLocation*/true,
2376                       getSpecifierRange(startSpecifier, specifierLen),
2377                       FixItHint::CreateRemoval(
2378                         getSpecifierRange(flag.getPosition(), 1)));
2379}
2380
2381void CheckPrintfHandler::HandleIgnoredFlag(
2382                                const analyze_printf::PrintfSpecifier &FS,
2383                                const analyze_printf::OptionalFlag &ignoredFlag,
2384                                const analyze_printf::OptionalFlag &flag,
2385                                const char *startSpecifier,
2386                                unsigned specifierLen) {
2387  // Warn about ignored flag with a fixit removal.
2388  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_ignored_flag)
2389                         << ignoredFlag.toString() << flag.toString(),
2390                       getLocationOfByte(ignoredFlag.getPosition()),
2391                       /*IsStringLocation*/true,
2392                       getSpecifierRange(startSpecifier, specifierLen),
2393                       FixItHint::CreateRemoval(
2394                         getSpecifierRange(ignoredFlag.getPosition(), 1)));
2395}
2396
2397// Determines if the specified is a C++ class or struct containing
2398// a member with the specified name and kind (e.g. a CXXMethodDecl named
2399// "c_str()").
2400template<typename MemberKind>
2401static llvm::SmallPtrSet<MemberKind*, 1>
2402CXXRecordMembersNamed(StringRef Name, Sema &S, QualType Ty) {
2403  const RecordType *RT = Ty->getAs<RecordType>();
2404  llvm::SmallPtrSet<MemberKind*, 1> Results;
2405
2406  if (!RT)
2407    return Results;
2408  const CXXRecordDecl *RD = dyn_cast<CXXRecordDecl>(RT->getDecl());
2409  if (!RD)
2410    return Results;
2411
2412  LookupResult R(S, &S.PP.getIdentifierTable().get(Name), SourceLocation(),
2413                 Sema::LookupMemberName);
2414
2415  // We just need to include all members of the right kind turned up by the
2416  // filter, at this point.
2417  if (S.LookupQualifiedName(R, RT->getDecl()))
2418    for (LookupResult::iterator I = R.begin(), E = R.end(); I != E; ++I) {
2419      NamedDecl *decl = (*I)->getUnderlyingDecl();
2420      if (MemberKind *FK = dyn_cast<MemberKind>(decl))
2421        Results.insert(FK);
2422    }
2423  return Results;
2424}
2425
2426// Check if a (w)string was passed when a (w)char* was needed, and offer a
2427// better diagnostic if so. AT is assumed to be valid.
2428// Returns true when a c_str() conversion method is found.
2429bool CheckPrintfHandler::checkForCStrMembers(
2430    const analyze_printf::ArgType &AT, const Expr *E,
2431    const CharSourceRange &CSR) {
2432  typedef llvm::SmallPtrSet<CXXMethodDecl*, 1> MethodSet;
2433
2434  MethodSet Results =
2435      CXXRecordMembersNamed<CXXMethodDecl>("c_str", S, E->getType());
2436
2437  for (MethodSet::iterator MI = Results.begin(), ME = Results.end();
2438       MI != ME; ++MI) {
2439    const CXXMethodDecl *Method = *MI;
2440    if (Method->getNumParams() == 0 &&
2441          AT.matchesType(S.Context, Method->getResultType())) {
2442      // FIXME: Suggest parens if the expression needs them.
2443      SourceLocation EndLoc =
2444          S.getPreprocessor().getLocForEndOfToken(E->getLocEnd());
2445      S.Diag(E->getLocStart(), diag::note_printf_c_str)
2446          << "c_str()"
2447          << FixItHint::CreateInsertion(EndLoc, ".c_str()");
2448      return true;
2449    }
2450  }
2451
2452  return false;
2453}
2454
2455bool
2456CheckPrintfHandler::HandlePrintfSpecifier(const analyze_printf::PrintfSpecifier
2457                                            &FS,
2458                                          const char *startSpecifier,
2459                                          unsigned specifierLen) {
2460
2461  using namespace analyze_format_string;
2462  using namespace analyze_printf;
2463  const PrintfConversionSpecifier &CS = FS.getConversionSpecifier();
2464
2465  if (FS.consumesDataArgument()) {
2466    if (atFirstArg) {
2467        atFirstArg = false;
2468        usesPositionalArgs = FS.usesPositionalArg();
2469    }
2470    else if (usesPositionalArgs != FS.usesPositionalArg()) {
2471      HandlePositionalNonpositionalArgs(getLocationOfByte(CS.getStart()),
2472                                        startSpecifier, specifierLen);
2473      return false;
2474    }
2475  }
2476
2477  // First check if the field width, precision, and conversion specifier
2478  // have matching data arguments.
2479  if (!HandleAmount(FS.getFieldWidth(), /* field width */ 0,
2480                    startSpecifier, specifierLen)) {
2481    return false;
2482  }
2483
2484  if (!HandleAmount(FS.getPrecision(), /* precision */ 1,
2485                    startSpecifier, specifierLen)) {
2486    return false;
2487  }
2488
2489  if (!CS.consumesDataArgument()) {
2490    // FIXME: Technically specifying a precision or field width here
2491    // makes no sense.  Worth issuing a warning at some point.
2492    return true;
2493  }
2494
2495  // Consume the argument.
2496  unsigned argIndex = FS.getArgIndex();
2497  if (argIndex < NumDataArgs) {
2498    // The check to see if the argIndex is valid will come later.
2499    // We set the bit here because we may exit early from this
2500    // function if we encounter some other error.
2501    CoveredArgs.set(argIndex);
2502  }
2503
2504  // Check for using an Objective-C specific conversion specifier
2505  // in a non-ObjC literal.
2506  if (!ObjCContext && CS.isObjCArg()) {
2507    return HandleInvalidPrintfConversionSpecifier(FS, startSpecifier,
2508                                                  specifierLen);
2509  }
2510
2511  // Check for invalid use of field width
2512  if (!FS.hasValidFieldWidth()) {
2513    HandleInvalidAmount(FS, FS.getFieldWidth(), /* field width */ 0,
2514        startSpecifier, specifierLen);
2515  }
2516
2517  // Check for invalid use of precision
2518  if (!FS.hasValidPrecision()) {
2519    HandleInvalidAmount(FS, FS.getPrecision(), /* precision */ 1,
2520        startSpecifier, specifierLen);
2521  }
2522
2523  // Check each flag does not conflict with any other component.
2524  if (!FS.hasValidThousandsGroupingPrefix())
2525    HandleFlag(FS, FS.hasThousandsGrouping(), startSpecifier, specifierLen);
2526  if (!FS.hasValidLeadingZeros())
2527    HandleFlag(FS, FS.hasLeadingZeros(), startSpecifier, specifierLen);
2528  if (!FS.hasValidPlusPrefix())
2529    HandleFlag(FS, FS.hasPlusPrefix(), startSpecifier, specifierLen);
2530  if (!FS.hasValidSpacePrefix())
2531    HandleFlag(FS, FS.hasSpacePrefix(), startSpecifier, specifierLen);
2532  if (!FS.hasValidAlternativeForm())
2533    HandleFlag(FS, FS.hasAlternativeForm(), startSpecifier, specifierLen);
2534  if (!FS.hasValidLeftJustified())
2535    HandleFlag(FS, FS.isLeftJustified(), startSpecifier, specifierLen);
2536
2537  // Check that flags are not ignored by another flag
2538  if (FS.hasSpacePrefix() && FS.hasPlusPrefix()) // ' ' ignored by '+'
2539    HandleIgnoredFlag(FS, FS.hasSpacePrefix(), FS.hasPlusPrefix(),
2540        startSpecifier, specifierLen);
2541  if (FS.hasLeadingZeros() && FS.isLeftJustified()) // '0' ignored by '-'
2542    HandleIgnoredFlag(FS, FS.hasLeadingZeros(), FS.isLeftJustified(),
2543            startSpecifier, specifierLen);
2544
2545  // Check the length modifier is valid with the given conversion specifier.
2546  const LengthModifier &LM = FS.getLengthModifier();
2547  if (!FS.hasValidLengthModifier())
2548    EmitFormatDiagnostic(S.PDiag(diag::warn_format_nonsensical_length)
2549                           << LM.toString() << CS.toString(),
2550                         getLocationOfByte(LM.getStart()),
2551                         /*IsStringLocation*/true,
2552                         getSpecifierRange(startSpecifier, specifierLen),
2553                         FixItHint::CreateRemoval(
2554                           getSpecifierRange(LM.getStart(),
2555                                             LM.getLength())));
2556  if (!FS.hasStandardLengthModifier())
2557    HandleNonStandardLengthModifier(LM, startSpecifier, specifierLen);
2558  if (!FS.hasStandardConversionSpecifier(S.getLangOpts()))
2559    HandleNonStandardConversionSpecifier(CS, startSpecifier, specifierLen);
2560  if (!FS.hasStandardLengthConversionCombination())
2561    HandleNonStandardConversionSpecification(LM, CS, startSpecifier,
2562                                             specifierLen);
2563
2564  // The remaining checks depend on the data arguments.
2565  if (HasVAListArg)
2566    return true;
2567
2568  if (!CheckNumArgs(FS, CS, startSpecifier, specifierLen, argIndex))
2569    return false;
2570
2571  const Expr *Arg = getDataArg(argIndex);
2572  if (!Arg)
2573    return true;
2574
2575  return checkFormatExpr(FS, startSpecifier, specifierLen, Arg);
2576}
2577
2578bool
2579CheckPrintfHandler::checkFormatExpr(const analyze_printf::PrintfSpecifier &FS,
2580                                    const char *StartSpecifier,
2581                                    unsigned SpecifierLen,
2582                                    const Expr *E) {
2583  using namespace analyze_format_string;
2584  using namespace analyze_printf;
2585  // Now type check the data expression that matches the
2586  // format specifier.
2587  const analyze_printf::ArgType &AT = FS.getArgType(S.Context,
2588                                                    ObjCContext);
2589  if (AT.isValid() && !AT.matchesType(S.Context, E->getType())) {
2590    // Look through argument promotions for our error message's reported type.
2591    // This includes the integral and floating promotions, but excludes array
2592    // and function pointer decay; seeing that an argument intended to be a
2593    // string has type 'char [6]' is probably more confusing than 'char *'.
2594    if (const ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(E)) {
2595      if (ICE->getCastKind() == CK_IntegralCast ||
2596          ICE->getCastKind() == CK_FloatingCast) {
2597        E = ICE->getSubExpr();
2598
2599        // Check if we didn't match because of an implicit cast from a 'char'
2600        // or 'short' to an 'int'.  This is done because printf is a varargs
2601        // function.
2602        if (ICE->getType() == S.Context.IntTy ||
2603            ICE->getType() == S.Context.UnsignedIntTy) {
2604          // All further checking is done on the subexpression.
2605          if (AT.matchesType(S.Context, E->getType()))
2606            return true;
2607        }
2608      }
2609    }
2610
2611    // We may be able to offer a FixItHint if it is a supported type.
2612    PrintfSpecifier fixedFS = FS;
2613    bool success = fixedFS.fixType(E->getType(), S.getLangOpts(),
2614                                   S.Context, ObjCContext);
2615
2616    if (success) {
2617      // Get the fix string from the fixed format specifier
2618      SmallString<16> buf;
2619      llvm::raw_svector_ostream os(buf);
2620      fixedFS.toString(os);
2621
2622      EmitFormatDiagnostic(
2623        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2624          << AT.getRepresentativeTypeName(S.Context) << E->getType()
2625          << E->getSourceRange(),
2626        E->getLocStart(),
2627        /*IsStringLocation*/false,
2628        getSpecifierRange(StartSpecifier, SpecifierLen),
2629        FixItHint::CreateReplacement(
2630          getSpecifierRange(StartSpecifier, SpecifierLen),
2631          os.str()));
2632    } else {
2633      const CharSourceRange &CSR = getSpecifierRange(StartSpecifier,
2634                                                     SpecifierLen);
2635      // Since the warning for passing non-POD types to variadic functions
2636      // was deferred until now, we emit a warning for non-POD
2637      // arguments here.
2638      if (S.isValidVarArgType(E->getType()) == Sema::VAK_Invalid) {
2639        unsigned DiagKind;
2640        if (E->getType()->isObjCObjectType())
2641          DiagKind = diag::err_cannot_pass_objc_interface_to_vararg_format;
2642        else
2643          DiagKind = diag::warn_non_pod_vararg_with_format_string;
2644
2645        EmitFormatDiagnostic(
2646          S.PDiag(DiagKind)
2647            << S.getLangOpts().CPlusPlus0x
2648            << E->getType()
2649            << CallType
2650            << AT.getRepresentativeTypeName(S.Context)
2651            << CSR
2652            << E->getSourceRange(),
2653          E->getLocStart(), /*IsStringLocation*/false, CSR);
2654
2655        checkForCStrMembers(AT, E, CSR);
2656      } else
2657        EmitFormatDiagnostic(
2658          S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2659            << AT.getRepresentativeTypeName(S.Context) << E->getType()
2660            << CSR
2661            << E->getSourceRange(),
2662          E->getLocStart(), /*IsStringLocation*/false, CSR);
2663    }
2664  }
2665
2666  return true;
2667}
2668
2669//===--- CHECK: Scanf format string checking ------------------------------===//
2670
2671namespace {
2672class CheckScanfHandler : public CheckFormatHandler {
2673public:
2674  CheckScanfHandler(Sema &s, const StringLiteral *fexpr,
2675                    const Expr *origFormatExpr, unsigned firstDataArg,
2676                    unsigned numDataArgs, const char *beg, bool hasVAListArg,
2677                    Expr **Args, unsigned NumArgs,
2678                    unsigned formatIdx, bool inFunctionCall,
2679                    Sema::VariadicCallType CallType)
2680  : CheckFormatHandler(s, fexpr, origFormatExpr, firstDataArg,
2681                       numDataArgs, beg, hasVAListArg,
2682                       Args, NumArgs, formatIdx, inFunctionCall, CallType)
2683  {}
2684
2685  bool HandleScanfSpecifier(const analyze_scanf::ScanfSpecifier &FS,
2686                            const char *startSpecifier,
2687                            unsigned specifierLen);
2688
2689  bool HandleInvalidScanfConversionSpecifier(
2690          const analyze_scanf::ScanfSpecifier &FS,
2691          const char *startSpecifier,
2692          unsigned specifierLen);
2693
2694  void HandleIncompleteScanList(const char *start, const char *end);
2695};
2696}
2697
2698void CheckScanfHandler::HandleIncompleteScanList(const char *start,
2699                                                 const char *end) {
2700  EmitFormatDiagnostic(S.PDiag(diag::warn_scanf_scanlist_incomplete),
2701                       getLocationOfByte(end), /*IsStringLocation*/true,
2702                       getSpecifierRange(start, end - start));
2703}
2704
2705bool CheckScanfHandler::HandleInvalidScanfConversionSpecifier(
2706                                        const analyze_scanf::ScanfSpecifier &FS,
2707                                        const char *startSpecifier,
2708                                        unsigned specifierLen) {
2709
2710  const analyze_scanf::ScanfConversionSpecifier &CS =
2711    FS.getConversionSpecifier();
2712
2713  return HandleInvalidConversionSpecifier(FS.getArgIndex(),
2714                                          getLocationOfByte(CS.getStart()),
2715                                          startSpecifier, specifierLen,
2716                                          CS.getStart(), CS.getLength());
2717}
2718
2719bool CheckScanfHandler::HandleScanfSpecifier(
2720                                       const analyze_scanf::ScanfSpecifier &FS,
2721                                       const char *startSpecifier,
2722                                       unsigned specifierLen) {
2723
2724  using namespace analyze_scanf;
2725  using namespace analyze_format_string;
2726
2727  const ScanfConversionSpecifier &CS = FS.getConversionSpecifier();
2728
2729  // Handle case where '%' and '*' don't consume an argument.  These shouldn't
2730  // be used to decide if we are using positional arguments consistently.
2731  if (FS.consumesDataArgument()) {
2732    if (atFirstArg) {
2733      atFirstArg = false;
2734      usesPositionalArgs = FS.usesPositionalArg();
2735    }
2736    else if (usesPositionalArgs != FS.usesPositionalArg()) {
2737      HandlePositionalNonpositionalArgs(getLocationOfByte(CS.getStart()),
2738                                        startSpecifier, specifierLen);
2739      return false;
2740    }
2741  }
2742
2743  // Check if the field with is non-zero.
2744  const OptionalAmount &Amt = FS.getFieldWidth();
2745  if (Amt.getHowSpecified() == OptionalAmount::Constant) {
2746    if (Amt.getConstantAmount() == 0) {
2747      const CharSourceRange &R = getSpecifierRange(Amt.getStart(),
2748                                                   Amt.getConstantLength());
2749      EmitFormatDiagnostic(S.PDiag(diag::warn_scanf_nonzero_width),
2750                           getLocationOfByte(Amt.getStart()),
2751                           /*IsStringLocation*/true, R,
2752                           FixItHint::CreateRemoval(R));
2753    }
2754  }
2755
2756  if (!FS.consumesDataArgument()) {
2757    // FIXME: Technically specifying a precision or field width here
2758    // makes no sense.  Worth issuing a warning at some point.
2759    return true;
2760  }
2761
2762  // Consume the argument.
2763  unsigned argIndex = FS.getArgIndex();
2764  if (argIndex < NumDataArgs) {
2765      // The check to see if the argIndex is valid will come later.
2766      // We set the bit here because we may exit early from this
2767      // function if we encounter some other error.
2768    CoveredArgs.set(argIndex);
2769  }
2770
2771  // Check the length modifier is valid with the given conversion specifier.
2772  const LengthModifier &LM = FS.getLengthModifier();
2773  if (!FS.hasValidLengthModifier()) {
2774    const CharSourceRange &R = getSpecifierRange(LM.getStart(), LM.getLength());
2775    EmitFormatDiagnostic(S.PDiag(diag::warn_format_nonsensical_length)
2776                         << LM.toString() << CS.toString()
2777                         << getSpecifierRange(startSpecifier, specifierLen),
2778                         getLocationOfByte(LM.getStart()),
2779                         /*IsStringLocation*/true, R,
2780                         FixItHint::CreateRemoval(R));
2781  }
2782
2783  if (!FS.hasStandardLengthModifier())
2784    HandleNonStandardLengthModifier(LM, startSpecifier, specifierLen);
2785  if (!FS.hasStandardConversionSpecifier(S.getLangOpts()))
2786    HandleNonStandardConversionSpecifier(CS, startSpecifier, specifierLen);
2787  if (!FS.hasStandardLengthConversionCombination())
2788    HandleNonStandardConversionSpecification(LM, CS, startSpecifier,
2789                                             specifierLen);
2790
2791  // The remaining checks depend on the data arguments.
2792  if (HasVAListArg)
2793    return true;
2794
2795  if (!CheckNumArgs(FS, CS, startSpecifier, specifierLen, argIndex))
2796    return false;
2797
2798  // Check that the argument type matches the format specifier.
2799  const Expr *Ex = getDataArg(argIndex);
2800  if (!Ex)
2801    return true;
2802
2803  const analyze_format_string::ArgType &AT = FS.getArgType(S.Context);
2804  if (AT.isValid() && !AT.matchesType(S.Context, Ex->getType())) {
2805    ScanfSpecifier fixedFS = FS;
2806    bool success = fixedFS.fixType(Ex->getType(), S.getLangOpts(),
2807                                   S.Context);
2808
2809    if (success) {
2810      // Get the fix string from the fixed format specifier.
2811      SmallString<128> buf;
2812      llvm::raw_svector_ostream os(buf);
2813      fixedFS.toString(os);
2814
2815      EmitFormatDiagnostic(
2816        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2817          << AT.getRepresentativeTypeName(S.Context) << Ex->getType()
2818          << Ex->getSourceRange(),
2819        Ex->getLocStart(),
2820        /*IsStringLocation*/false,
2821        getSpecifierRange(startSpecifier, specifierLen),
2822        FixItHint::CreateReplacement(
2823          getSpecifierRange(startSpecifier, specifierLen),
2824          os.str()));
2825    } else {
2826      EmitFormatDiagnostic(
2827        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2828          << AT.getRepresentativeTypeName(S.Context) << Ex->getType()
2829          << Ex->getSourceRange(),
2830        Ex->getLocStart(),
2831        /*IsStringLocation*/false,
2832        getSpecifierRange(startSpecifier, specifierLen));
2833    }
2834  }
2835
2836  return true;
2837}
2838
2839void Sema::CheckFormatString(const StringLiteral *FExpr,
2840                             const Expr *OrigFormatExpr,
2841                             Expr **Args, unsigned NumArgs,
2842                             bool HasVAListArg, unsigned format_idx,
2843                             unsigned firstDataArg, FormatStringType Type,
2844                             bool inFunctionCall, VariadicCallType CallType) {
2845
2846  // CHECK: is the format string a wide literal?
2847  if (!FExpr->isAscii() && !FExpr->isUTF8()) {
2848    CheckFormatHandler::EmitFormatDiagnostic(
2849      *this, inFunctionCall, Args[format_idx],
2850      PDiag(diag::warn_format_string_is_wide_literal), FExpr->getLocStart(),
2851      /*IsStringLocation*/true, OrigFormatExpr->getSourceRange());
2852    return;
2853  }
2854
2855  // Str - The format string.  NOTE: this is NOT null-terminated!
2856  StringRef StrRef = FExpr->getString();
2857  const char *Str = StrRef.data();
2858  unsigned StrLen = StrRef.size();
2859  const unsigned numDataArgs = NumArgs - firstDataArg;
2860
2861  // CHECK: empty format string?
2862  if (StrLen == 0 && numDataArgs > 0) {
2863    CheckFormatHandler::EmitFormatDiagnostic(
2864      *this, inFunctionCall, Args[format_idx],
2865      PDiag(diag::warn_empty_format_string), FExpr->getLocStart(),
2866      /*IsStringLocation*/true, OrigFormatExpr->getSourceRange());
2867    return;
2868  }
2869
2870  if (Type == FST_Printf || Type == FST_NSString) {
2871    CheckPrintfHandler H(*this, FExpr, OrigFormatExpr, firstDataArg,
2872                         numDataArgs, (Type == FST_NSString),
2873                         Str, HasVAListArg, Args, NumArgs, format_idx,
2874                         inFunctionCall, CallType);
2875
2876    if (!analyze_format_string::ParsePrintfString(H, Str, Str + StrLen,
2877                                                  getLangOpts()))
2878      H.DoneProcessing();
2879  } else if (Type == FST_Scanf) {
2880    CheckScanfHandler H(*this, FExpr, OrigFormatExpr, firstDataArg, numDataArgs,
2881                        Str, HasVAListArg, Args, NumArgs, format_idx,
2882                        inFunctionCall, CallType);
2883
2884    if (!analyze_format_string::ParseScanfString(H, Str, Str + StrLen,
2885                                                 getLangOpts()))
2886      H.DoneProcessing();
2887  } // TODO: handle other formats
2888}
2889
2890//===--- CHECK: Standard memory functions ---------------------------------===//
2891
2892/// \brief Determine whether the given type is a dynamic class type (e.g.,
2893/// whether it has a vtable).
2894static bool isDynamicClassType(QualType T) {
2895  if (CXXRecordDecl *Record = T->getAsCXXRecordDecl())
2896    if (CXXRecordDecl *Definition = Record->getDefinition())
2897      if (Definition->isDynamicClass())
2898        return true;
2899
2900  return false;
2901}
2902
2903/// \brief If E is a sizeof expression, returns its argument expression,
2904/// otherwise returns NULL.
2905static const Expr *getSizeOfExprArg(const Expr* E) {
2906  if (const UnaryExprOrTypeTraitExpr *SizeOf =
2907      dyn_cast<UnaryExprOrTypeTraitExpr>(E))
2908    if (SizeOf->getKind() == clang::UETT_SizeOf && !SizeOf->isArgumentType())
2909      return SizeOf->getArgumentExpr()->IgnoreParenImpCasts();
2910
2911  return 0;
2912}
2913
2914/// \brief If E is a sizeof expression, returns its argument type.
2915static QualType getSizeOfArgType(const Expr* E) {
2916  if (const UnaryExprOrTypeTraitExpr *SizeOf =
2917      dyn_cast<UnaryExprOrTypeTraitExpr>(E))
2918    if (SizeOf->getKind() == clang::UETT_SizeOf)
2919      return SizeOf->getTypeOfArgument();
2920
2921  return QualType();
2922}
2923
2924/// \brief Check for dangerous or invalid arguments to memset().
2925///
2926/// This issues warnings on known problematic, dangerous or unspecified
2927/// arguments to the standard 'memset', 'memcpy', 'memmove', and 'memcmp'
2928/// function calls.
2929///
2930/// \param Call The call expression to diagnose.
2931void Sema::CheckMemaccessArguments(const CallExpr *Call,
2932                                   unsigned BId,
2933                                   IdentifierInfo *FnName) {
2934  assert(BId != 0);
2935
2936  // It is possible to have a non-standard definition of memset.  Validate
2937  // we have enough arguments, and if not, abort further checking.
2938  unsigned ExpectedNumArgs = (BId == Builtin::BIstrndup ? 2 : 3);
2939  if (Call->getNumArgs() < ExpectedNumArgs)
2940    return;
2941
2942  unsigned LastArg = (BId == Builtin::BImemset ||
2943                      BId == Builtin::BIstrndup ? 1 : 2);
2944  unsigned LenArg = (BId == Builtin::BIstrndup ? 1 : 2);
2945  const Expr *LenExpr = Call->getArg(LenArg)->IgnoreParenImpCasts();
2946
2947  // We have special checking when the length is a sizeof expression.
2948  QualType SizeOfArgTy = getSizeOfArgType(LenExpr);
2949  const Expr *SizeOfArg = getSizeOfExprArg(LenExpr);
2950  llvm::FoldingSetNodeID SizeOfArgID;
2951
2952  for (unsigned ArgIdx = 0; ArgIdx != LastArg; ++ArgIdx) {
2953    const Expr *Dest = Call->getArg(ArgIdx)->IgnoreParenImpCasts();
2954    SourceRange ArgRange = Call->getArg(ArgIdx)->getSourceRange();
2955
2956    QualType DestTy = Dest->getType();
2957    if (const PointerType *DestPtrTy = DestTy->getAs<PointerType>()) {
2958      QualType PointeeTy = DestPtrTy->getPointeeType();
2959
2960      // Never warn about void type pointers. This can be used to suppress
2961      // false positives.
2962      if (PointeeTy->isVoidType())
2963        continue;
2964
2965      // Catch "memset(p, 0, sizeof(p))" -- needs to be sizeof(*p). Do this by
2966      // actually comparing the expressions for equality. Because computing the
2967      // expression IDs can be expensive, we only do this if the diagnostic is
2968      // enabled.
2969      if (SizeOfArg &&
2970          Diags.getDiagnosticLevel(diag::warn_sizeof_pointer_expr_memaccess,
2971                                   SizeOfArg->getExprLoc())) {
2972        // We only compute IDs for expressions if the warning is enabled, and
2973        // cache the sizeof arg's ID.
2974        if (SizeOfArgID == llvm::FoldingSetNodeID())
2975          SizeOfArg->Profile(SizeOfArgID, Context, true);
2976        llvm::FoldingSetNodeID DestID;
2977        Dest->Profile(DestID, Context, true);
2978        if (DestID == SizeOfArgID) {
2979          // TODO: For strncpy() and friends, this could suggest sizeof(dst)
2980          //       over sizeof(src) as well.
2981          unsigned ActionIdx = 0; // Default is to suggest dereferencing.
2982          StringRef ReadableName = FnName->getName();
2983
2984          if (const UnaryOperator *UnaryOp = dyn_cast<UnaryOperator>(Dest))
2985            if (UnaryOp->getOpcode() == UO_AddrOf)
2986              ActionIdx = 1; // If its an address-of operator, just remove it.
2987          if (Context.getTypeSize(PointeeTy) == Context.getCharWidth())
2988            ActionIdx = 2; // If the pointee's size is sizeof(char),
2989                           // suggest an explicit length.
2990
2991          // If the function is defined as a builtin macro, do not show macro
2992          // expansion.
2993          SourceLocation SL = SizeOfArg->getExprLoc();
2994          SourceRange DSR = Dest->getSourceRange();
2995          SourceRange SSR = SizeOfArg->getSourceRange();
2996          SourceManager &SM  = PP.getSourceManager();
2997
2998          if (SM.isMacroArgExpansion(SL)) {
2999            ReadableName = Lexer::getImmediateMacroName(SL, SM, LangOpts);
3000            SL = SM.getSpellingLoc(SL);
3001            DSR = SourceRange(SM.getSpellingLoc(DSR.getBegin()),
3002                             SM.getSpellingLoc(DSR.getEnd()));
3003            SSR = SourceRange(SM.getSpellingLoc(SSR.getBegin()),
3004                             SM.getSpellingLoc(SSR.getEnd()));
3005          }
3006
3007          DiagRuntimeBehavior(SL, SizeOfArg,
3008                              PDiag(diag::warn_sizeof_pointer_expr_memaccess)
3009                                << ReadableName
3010                                << PointeeTy
3011                                << DestTy
3012                                << DSR
3013                                << SSR);
3014          DiagRuntimeBehavior(SL, SizeOfArg,
3015                         PDiag(diag::warn_sizeof_pointer_expr_memaccess_note)
3016                                << ActionIdx
3017                                << SSR);
3018
3019          break;
3020        }
3021      }
3022
3023      // Also check for cases where the sizeof argument is the exact same
3024      // type as the memory argument, and where it points to a user-defined
3025      // record type.
3026      if (SizeOfArgTy != QualType()) {
3027        if (PointeeTy->isRecordType() &&
3028            Context.typesAreCompatible(SizeOfArgTy, DestTy)) {
3029          DiagRuntimeBehavior(LenExpr->getExprLoc(), Dest,
3030                              PDiag(diag::warn_sizeof_pointer_type_memaccess)
3031                                << FnName << SizeOfArgTy << ArgIdx
3032                                << PointeeTy << Dest->getSourceRange()
3033                                << LenExpr->getSourceRange());
3034          break;
3035        }
3036      }
3037
3038      // Always complain about dynamic classes.
3039      if (isDynamicClassType(PointeeTy)) {
3040
3041        unsigned OperationType = 0;
3042        // "overwritten" if we're warning about the destination for any call
3043        // but memcmp; otherwise a verb appropriate to the call.
3044        if (ArgIdx != 0 || BId == Builtin::BImemcmp) {
3045          if (BId == Builtin::BImemcpy)
3046            OperationType = 1;
3047          else if(BId == Builtin::BImemmove)
3048            OperationType = 2;
3049          else if (BId == Builtin::BImemcmp)
3050            OperationType = 3;
3051        }
3052
3053        DiagRuntimeBehavior(
3054          Dest->getExprLoc(), Dest,
3055          PDiag(diag::warn_dyn_class_memaccess)
3056            << (BId == Builtin::BImemcmp ? ArgIdx + 2 : ArgIdx)
3057            << FnName << PointeeTy
3058            << OperationType
3059            << Call->getCallee()->getSourceRange());
3060      } else if (PointeeTy.hasNonTrivialObjCLifetime() &&
3061               BId != Builtin::BImemset)
3062        DiagRuntimeBehavior(
3063          Dest->getExprLoc(), Dest,
3064          PDiag(diag::warn_arc_object_memaccess)
3065            << ArgIdx << FnName << PointeeTy
3066            << Call->getCallee()->getSourceRange());
3067      else
3068        continue;
3069
3070      DiagRuntimeBehavior(
3071        Dest->getExprLoc(), Dest,
3072        PDiag(diag::note_bad_memaccess_silence)
3073          << FixItHint::CreateInsertion(ArgRange.getBegin(), "(void*)"));
3074      break;
3075    }
3076  }
3077}
3078
3079// A little helper routine: ignore addition and subtraction of integer literals.
3080// This intentionally does not ignore all integer constant expressions because
3081// we don't want to remove sizeof().
3082static const Expr *ignoreLiteralAdditions(const Expr *Ex, ASTContext &Ctx) {
3083  Ex = Ex->IgnoreParenCasts();
3084
3085  for (;;) {
3086    const BinaryOperator * BO = dyn_cast<BinaryOperator>(Ex);
3087    if (!BO || !BO->isAdditiveOp())
3088      break;
3089
3090    const Expr *RHS = BO->getRHS()->IgnoreParenCasts();
3091    const Expr *LHS = BO->getLHS()->IgnoreParenCasts();
3092
3093    if (isa<IntegerLiteral>(RHS))
3094      Ex = LHS;
3095    else if (isa<IntegerLiteral>(LHS))
3096      Ex = RHS;
3097    else
3098      break;
3099  }
3100
3101  return Ex;
3102}
3103
3104static bool isConstantSizeArrayWithMoreThanOneElement(QualType Ty,
3105                                                      ASTContext &Context) {
3106  // Only handle constant-sized or VLAs, but not flexible members.
3107  if (const ConstantArrayType *CAT = Context.getAsConstantArrayType(Ty)) {
3108    // Only issue the FIXIT for arrays of size > 1.
3109    if (CAT->getSize().getSExtValue() <= 1)
3110      return false;
3111  } else if (!Ty->isVariableArrayType()) {
3112    return false;
3113  }
3114  return true;
3115}
3116
3117// Warn if the user has made the 'size' argument to strlcpy or strlcat
3118// be the size of the source, instead of the destination.
3119void Sema::CheckStrlcpycatArguments(const CallExpr *Call,
3120                                    IdentifierInfo *FnName) {
3121
3122  // Don't crash if the user has the wrong number of arguments
3123  if (Call->getNumArgs() != 3)
3124    return;
3125
3126  const Expr *SrcArg = ignoreLiteralAdditions(Call->getArg(1), Context);
3127  const Expr *SizeArg = ignoreLiteralAdditions(Call->getArg(2), Context);
3128  const Expr *CompareWithSrc = NULL;
3129
3130  // Look for 'strlcpy(dst, x, sizeof(x))'
3131  if (const Expr *Ex = getSizeOfExprArg(SizeArg))
3132    CompareWithSrc = Ex;
3133  else {
3134    // Look for 'strlcpy(dst, x, strlen(x))'
3135    if (const CallExpr *SizeCall = dyn_cast<CallExpr>(SizeArg)) {
3136      if (SizeCall->isBuiltinCall() == Builtin::BIstrlen
3137          && SizeCall->getNumArgs() == 1)
3138        CompareWithSrc = ignoreLiteralAdditions(SizeCall->getArg(0), Context);
3139    }
3140  }
3141
3142  if (!CompareWithSrc)
3143    return;
3144
3145  // Determine if the argument to sizeof/strlen is equal to the source
3146  // argument.  In principle there's all kinds of things you could do
3147  // here, for instance creating an == expression and evaluating it with
3148  // EvaluateAsBooleanCondition, but this uses a more direct technique:
3149  const DeclRefExpr *SrcArgDRE = dyn_cast<DeclRefExpr>(SrcArg);
3150  if (!SrcArgDRE)
3151    return;
3152
3153  const DeclRefExpr *CompareWithSrcDRE = dyn_cast<DeclRefExpr>(CompareWithSrc);
3154  if (!CompareWithSrcDRE ||
3155      SrcArgDRE->getDecl() != CompareWithSrcDRE->getDecl())
3156    return;
3157
3158  const Expr *OriginalSizeArg = Call->getArg(2);
3159  Diag(CompareWithSrcDRE->getLocStart(), diag::warn_strlcpycat_wrong_size)
3160    << OriginalSizeArg->getSourceRange() << FnName;
3161
3162  // Output a FIXIT hint if the destination is an array (rather than a
3163  // pointer to an array).  This could be enhanced to handle some
3164  // pointers if we know the actual size, like if DstArg is 'array+2'
3165  // we could say 'sizeof(array)-2'.
3166  const Expr *DstArg = Call->getArg(0)->IgnoreParenImpCasts();
3167  if (!isConstantSizeArrayWithMoreThanOneElement(DstArg->getType(), Context))
3168    return;
3169
3170  SmallString<128> sizeString;
3171  llvm::raw_svector_ostream OS(sizeString);
3172  OS << "sizeof(";
3173  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3174  OS << ")";
3175
3176  Diag(OriginalSizeArg->getLocStart(), diag::note_strlcpycat_wrong_size)
3177    << FixItHint::CreateReplacement(OriginalSizeArg->getSourceRange(),
3178                                    OS.str());
3179}
3180
3181/// Check if two expressions refer to the same declaration.
3182static bool referToTheSameDecl(const Expr *E1, const Expr *E2) {
3183  if (const DeclRefExpr *D1 = dyn_cast_or_null<DeclRefExpr>(E1))
3184    if (const DeclRefExpr *D2 = dyn_cast_or_null<DeclRefExpr>(E2))
3185      return D1->getDecl() == D2->getDecl();
3186  return false;
3187}
3188
3189static const Expr *getStrlenExprArg(const Expr *E) {
3190  if (const CallExpr *CE = dyn_cast<CallExpr>(E)) {
3191    const FunctionDecl *FD = CE->getDirectCallee();
3192    if (!FD || FD->getMemoryFunctionKind() != Builtin::BIstrlen)
3193      return 0;
3194    return CE->getArg(0)->IgnoreParenCasts();
3195  }
3196  return 0;
3197}
3198
3199// Warn on anti-patterns as the 'size' argument to strncat.
3200// The correct size argument should look like following:
3201//   strncat(dst, src, sizeof(dst) - strlen(dest) - 1);
3202void Sema::CheckStrncatArguments(const CallExpr *CE,
3203                                 IdentifierInfo *FnName) {
3204  // Don't crash if the user has the wrong number of arguments.
3205  if (CE->getNumArgs() < 3)
3206    return;
3207  const Expr *DstArg = CE->getArg(0)->IgnoreParenCasts();
3208  const Expr *SrcArg = CE->getArg(1)->IgnoreParenCasts();
3209  const Expr *LenArg = CE->getArg(2)->IgnoreParenCasts();
3210
3211  // Identify common expressions, which are wrongly used as the size argument
3212  // to strncat and may lead to buffer overflows.
3213  unsigned PatternType = 0;
3214  if (const Expr *SizeOfArg = getSizeOfExprArg(LenArg)) {
3215    // - sizeof(dst)
3216    if (referToTheSameDecl(SizeOfArg, DstArg))
3217      PatternType = 1;
3218    // - sizeof(src)
3219    else if (referToTheSameDecl(SizeOfArg, SrcArg))
3220      PatternType = 2;
3221  } else if (const BinaryOperator *BE = dyn_cast<BinaryOperator>(LenArg)) {
3222    if (BE->getOpcode() == BO_Sub) {
3223      const Expr *L = BE->getLHS()->IgnoreParenCasts();
3224      const Expr *R = BE->getRHS()->IgnoreParenCasts();
3225      // - sizeof(dst) - strlen(dst)
3226      if (referToTheSameDecl(DstArg, getSizeOfExprArg(L)) &&
3227          referToTheSameDecl(DstArg, getStrlenExprArg(R)))
3228        PatternType = 1;
3229      // - sizeof(src) - (anything)
3230      else if (referToTheSameDecl(SrcArg, getSizeOfExprArg(L)))
3231        PatternType = 2;
3232    }
3233  }
3234
3235  if (PatternType == 0)
3236    return;
3237
3238  // Generate the diagnostic.
3239  SourceLocation SL = LenArg->getLocStart();
3240  SourceRange SR = LenArg->getSourceRange();
3241  SourceManager &SM  = PP.getSourceManager();
3242
3243  // If the function is defined as a builtin macro, do not show macro expansion.
3244  if (SM.isMacroArgExpansion(SL)) {
3245    SL = SM.getSpellingLoc(SL);
3246    SR = SourceRange(SM.getSpellingLoc(SR.getBegin()),
3247                     SM.getSpellingLoc(SR.getEnd()));
3248  }
3249
3250  // Check if the destination is an array (rather than a pointer to an array).
3251  QualType DstTy = DstArg->getType();
3252  bool isKnownSizeArray = isConstantSizeArrayWithMoreThanOneElement(DstTy,
3253                                                                    Context);
3254  if (!isKnownSizeArray) {
3255    if (PatternType == 1)
3256      Diag(SL, diag::warn_strncat_wrong_size) << SR;
3257    else
3258      Diag(SL, diag::warn_strncat_src_size) << SR;
3259    return;
3260  }
3261
3262  if (PatternType == 1)
3263    Diag(SL, diag::warn_strncat_large_size) << SR;
3264  else
3265    Diag(SL, diag::warn_strncat_src_size) << SR;
3266
3267  SmallString<128> sizeString;
3268  llvm::raw_svector_ostream OS(sizeString);
3269  OS << "sizeof(";
3270  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3271  OS << ") - ";
3272  OS << "strlen(";
3273  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3274  OS << ") - 1";
3275
3276  Diag(SL, diag::note_strncat_wrong_size)
3277    << FixItHint::CreateReplacement(SR, OS.str());
3278}
3279
3280//===--- CHECK: Return Address of Stack Variable --------------------------===//
3281
3282static Expr *EvalVal(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars,
3283                     Decl *ParentDecl);
3284static Expr *EvalAddr(Expr* E, SmallVectorImpl<DeclRefExpr *> &refVars,
3285                      Decl *ParentDecl);
3286
3287/// CheckReturnStackAddr - Check if a return statement returns the address
3288///   of a stack variable.
3289void
3290Sema::CheckReturnStackAddr(Expr *RetValExp, QualType lhsType,
3291                           SourceLocation ReturnLoc) {
3292
3293  Expr *stackE = 0;
3294  SmallVector<DeclRefExpr *, 8> refVars;
3295
3296  // Perform checking for returned stack addresses, local blocks,
3297  // label addresses or references to temporaries.
3298  if (lhsType->isPointerType() ||
3299      (!getLangOpts().ObjCAutoRefCount && lhsType->isBlockPointerType())) {
3300    stackE = EvalAddr(RetValExp, refVars, /*ParentDecl=*/0);
3301  } else if (lhsType->isReferenceType()) {
3302    stackE = EvalVal(RetValExp, refVars, /*ParentDecl=*/0);
3303  }
3304
3305  if (stackE == 0)
3306    return; // Nothing suspicious was found.
3307
3308  SourceLocation diagLoc;
3309  SourceRange diagRange;
3310  if (refVars.empty()) {
3311    diagLoc = stackE->getLocStart();
3312    diagRange = stackE->getSourceRange();
3313  } else {
3314    // We followed through a reference variable. 'stackE' contains the
3315    // problematic expression but we will warn at the return statement pointing
3316    // at the reference variable. We will later display the "trail" of
3317    // reference variables using notes.
3318    diagLoc = refVars[0]->getLocStart();
3319    diagRange = refVars[0]->getSourceRange();
3320  }
3321
3322  if (DeclRefExpr *DR = dyn_cast<DeclRefExpr>(stackE)) { //address of local var.
3323    Diag(diagLoc, lhsType->isReferenceType() ? diag::warn_ret_stack_ref
3324                                             : diag::warn_ret_stack_addr)
3325     << DR->getDecl()->getDeclName() << diagRange;
3326  } else if (isa<BlockExpr>(stackE)) { // local block.
3327    Diag(diagLoc, diag::err_ret_local_block) << diagRange;
3328  } else if (isa<AddrLabelExpr>(stackE)) { // address of label.
3329    Diag(diagLoc, diag::warn_ret_addr_label) << diagRange;
3330  } else { // local temporary.
3331    Diag(diagLoc, lhsType->isReferenceType() ? diag::warn_ret_local_temp_ref
3332                                             : diag::warn_ret_local_temp_addr)
3333     << diagRange;
3334  }
3335
3336  // Display the "trail" of reference variables that we followed until we
3337  // found the problematic expression using notes.
3338  for (unsigned i = 0, e = refVars.size(); i != e; ++i) {
3339    VarDecl *VD = cast<VarDecl>(refVars[i]->getDecl());
3340    // If this var binds to another reference var, show the range of the next
3341    // var, otherwise the var binds to the problematic expression, in which case
3342    // show the range of the expression.
3343    SourceRange range = (i < e-1) ? refVars[i+1]->getSourceRange()
3344                                  : stackE->getSourceRange();
3345    Diag(VD->getLocation(), diag::note_ref_var_local_bind)
3346      << VD->getDeclName() << range;
3347  }
3348}
3349
3350/// EvalAddr - EvalAddr and EvalVal are mutually recursive functions that
3351///  check if the expression in a return statement evaluates to an address
3352///  to a location on the stack, a local block, an address of a label, or a
3353///  reference to local temporary. The recursion is used to traverse the
3354///  AST of the return expression, with recursion backtracking when we
3355///  encounter a subexpression that (1) clearly does not lead to one of the
3356///  above problematic expressions (2) is something we cannot determine leads to
3357///  a problematic expression based on such local checking.
3358///
3359///  Both EvalAddr and EvalVal follow through reference variables to evaluate
3360///  the expression that they point to. Such variables are added to the
3361///  'refVars' vector so that we know what the reference variable "trail" was.
3362///
3363///  EvalAddr processes expressions that are pointers that are used as
3364///  references (and not L-values).  EvalVal handles all other values.
3365///  At the base case of the recursion is a check for the above problematic
3366///  expressions.
3367///
3368///  This implementation handles:
3369///
3370///   * pointer-to-pointer casts
3371///   * implicit conversions from array references to pointers
3372///   * taking the address of fields
3373///   * arbitrary interplay between "&" and "*" operators
3374///   * pointer arithmetic from an address of a stack variable
3375///   * taking the address of an array element where the array is on the stack
3376static Expr *EvalAddr(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars,
3377                      Decl *ParentDecl) {
3378  if (E->isTypeDependent())
3379      return NULL;
3380
3381  // We should only be called for evaluating pointer expressions.
3382  assert((E->getType()->isAnyPointerType() ||
3383          E->getType()->isBlockPointerType() ||
3384          E->getType()->isObjCQualifiedIdType()) &&
3385         "EvalAddr only works on pointers");
3386
3387  E = E->IgnoreParens();
3388
3389  // Our "symbolic interpreter" is just a dispatch off the currently
3390  // viewed AST node.  We then recursively traverse the AST by calling
3391  // EvalAddr and EvalVal appropriately.
3392  switch (E->getStmtClass()) {
3393  case Stmt::DeclRefExprClass: {
3394    DeclRefExpr *DR = cast<DeclRefExpr>(E);
3395
3396    if (VarDecl *V = dyn_cast<VarDecl>(DR->getDecl()))
3397      // If this is a reference variable, follow through to the expression that
3398      // it points to.
3399      if (V->hasLocalStorage() &&
3400          V->getType()->isReferenceType() && V->hasInit()) {
3401        // Add the reference variable to the "trail".
3402        refVars.push_back(DR);
3403        return EvalAddr(V->getInit(), refVars, ParentDecl);
3404      }
3405
3406    return NULL;
3407  }
3408
3409  case Stmt::UnaryOperatorClass: {
3410    // The only unary operator that make sense to handle here
3411    // is AddrOf.  All others don't make sense as pointers.
3412    UnaryOperator *U = cast<UnaryOperator>(E);
3413
3414    if (U->getOpcode() == UO_AddrOf)
3415      return EvalVal(U->getSubExpr(), refVars, ParentDecl);
3416    else
3417      return NULL;
3418  }
3419
3420  case Stmt::BinaryOperatorClass: {
3421    // Handle pointer arithmetic.  All other binary operators are not valid
3422    // in this context.
3423    BinaryOperator *B = cast<BinaryOperator>(E);
3424    BinaryOperatorKind op = B->getOpcode();
3425
3426    if (op != BO_Add && op != BO_Sub)
3427      return NULL;
3428
3429    Expr *Base = B->getLHS();
3430
3431    // Determine which argument is the real pointer base.  It could be
3432    // the RHS argument instead of the LHS.
3433    if (!Base->getType()->isPointerType()) Base = B->getRHS();
3434
3435    assert (Base->getType()->isPointerType());
3436    return EvalAddr(Base, refVars, ParentDecl);
3437  }
3438
3439  // For conditional operators we need to see if either the LHS or RHS are
3440  // valid DeclRefExpr*s.  If one of them is valid, we return it.
3441  case Stmt::ConditionalOperatorClass: {
3442    ConditionalOperator *C = cast<ConditionalOperator>(E);
3443
3444    // Handle the GNU extension for missing LHS.
3445    if (Expr *lhsExpr = C->getLHS()) {
3446    // In C++, we can have a throw-expression, which has 'void' type.
3447      if (!lhsExpr->getType()->isVoidType())
3448        if (Expr* LHS = EvalAddr(lhsExpr, refVars, ParentDecl))
3449          return LHS;
3450    }
3451
3452    // In C++, we can have a throw-expression, which has 'void' type.
3453    if (C->getRHS()->getType()->isVoidType())
3454      return NULL;
3455
3456    return EvalAddr(C->getRHS(), refVars, ParentDecl);
3457  }
3458
3459  case Stmt::BlockExprClass:
3460    if (cast<BlockExpr>(E)->getBlockDecl()->hasCaptures())
3461      return E; // local block.
3462    return NULL;
3463
3464  case Stmt::AddrLabelExprClass:
3465    return E; // address of label.
3466
3467  case Stmt::ExprWithCleanupsClass:
3468    return EvalAddr(cast<ExprWithCleanups>(E)->getSubExpr(), refVars,
3469                    ParentDecl);
3470
3471  // For casts, we need to handle conversions from arrays to
3472  // pointer values, and pointer-to-pointer conversions.
3473  case Stmt::ImplicitCastExprClass:
3474  case Stmt::CStyleCastExprClass:
3475  case Stmt::CXXFunctionalCastExprClass:
3476  case Stmt::ObjCBridgedCastExprClass:
3477  case Stmt::CXXStaticCastExprClass:
3478  case Stmt::CXXDynamicCastExprClass:
3479  case Stmt::CXXConstCastExprClass:
3480  case Stmt::CXXReinterpretCastExprClass: {
3481    Expr* SubExpr = cast<CastExpr>(E)->getSubExpr();
3482    switch (cast<CastExpr>(E)->getCastKind()) {
3483    case CK_BitCast:
3484    case CK_LValueToRValue:
3485    case CK_NoOp:
3486    case CK_BaseToDerived:
3487    case CK_DerivedToBase:
3488    case CK_UncheckedDerivedToBase:
3489    case CK_Dynamic:
3490    case CK_CPointerToObjCPointerCast:
3491    case CK_BlockPointerToObjCPointerCast:
3492    case CK_AnyPointerToBlockPointerCast:
3493      return EvalAddr(SubExpr, refVars, ParentDecl);
3494
3495    case CK_ArrayToPointerDecay:
3496      return EvalVal(SubExpr, refVars, ParentDecl);
3497
3498    default:
3499      return 0;
3500    }
3501  }
3502
3503  case Stmt::MaterializeTemporaryExprClass:
3504    if (Expr *Result = EvalAddr(
3505                         cast<MaterializeTemporaryExpr>(E)->GetTemporaryExpr(),
3506                                refVars, ParentDecl))
3507      return Result;
3508
3509    return E;
3510
3511  // Everything else: we simply don't reason about them.
3512  default:
3513    return NULL;
3514  }
3515}
3516
3517
3518///  EvalVal - This function is complements EvalAddr in the mutual recursion.
3519///   See the comments for EvalAddr for more details.
3520static Expr *EvalVal(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars,
3521                     Decl *ParentDecl) {
3522do {
3523  // We should only be called for evaluating non-pointer expressions, or
3524  // expressions with a pointer type that are not used as references but instead
3525  // are l-values (e.g., DeclRefExpr with a pointer type).
3526
3527  // Our "symbolic interpreter" is just a dispatch off the currently
3528  // viewed AST node.  We then recursively traverse the AST by calling
3529  // EvalAddr and EvalVal appropriately.
3530
3531  E = E->IgnoreParens();
3532  switch (E->getStmtClass()) {
3533  case Stmt::ImplicitCastExprClass: {
3534    ImplicitCastExpr *IE = cast<ImplicitCastExpr>(E);
3535    if (IE->getValueKind() == VK_LValue) {
3536      E = IE->getSubExpr();
3537      continue;
3538    }
3539    return NULL;
3540  }
3541
3542  case Stmt::ExprWithCleanupsClass:
3543    return EvalVal(cast<ExprWithCleanups>(E)->getSubExpr(), refVars,ParentDecl);
3544
3545  case Stmt::DeclRefExprClass: {
3546    // When we hit a DeclRefExpr we are looking at code that refers to a
3547    // variable's name. If it's not a reference variable we check if it has
3548    // local storage within the function, and if so, return the expression.
3549    DeclRefExpr *DR = cast<DeclRefExpr>(E);
3550
3551    if (VarDecl *V = dyn_cast<VarDecl>(DR->getDecl())) {
3552      // Check if it refers to itself, e.g. "int& i = i;".
3553      if (V == ParentDecl)
3554        return DR;
3555
3556      if (V->hasLocalStorage()) {
3557        if (!V->getType()->isReferenceType())
3558          return DR;
3559
3560        // Reference variable, follow through to the expression that
3561        // it points to.
3562        if (V->hasInit()) {
3563          // Add the reference variable to the "trail".
3564          refVars.push_back(DR);
3565          return EvalVal(V->getInit(), refVars, V);
3566        }
3567      }
3568    }
3569
3570    return NULL;
3571  }
3572
3573  case Stmt::UnaryOperatorClass: {
3574    // The only unary operator that make sense to handle here
3575    // is Deref.  All others don't resolve to a "name."  This includes
3576    // handling all sorts of rvalues passed to a unary operator.
3577    UnaryOperator *U = cast<UnaryOperator>(E);
3578
3579    if (U->getOpcode() == UO_Deref)
3580      return EvalAddr(U->getSubExpr(), refVars, ParentDecl);
3581
3582    return NULL;
3583  }
3584
3585  case Stmt::ArraySubscriptExprClass: {
3586    // Array subscripts are potential references to data on the stack.  We
3587    // retrieve the DeclRefExpr* for the array variable if it indeed
3588    // has local storage.
3589    return EvalAddr(cast<ArraySubscriptExpr>(E)->getBase(), refVars,ParentDecl);
3590  }
3591
3592  case Stmt::ConditionalOperatorClass: {
3593    // For conditional operators we need to see if either the LHS or RHS are
3594    // non-NULL Expr's.  If one is non-NULL, we return it.
3595    ConditionalOperator *C = cast<ConditionalOperator>(E);
3596
3597    // Handle the GNU extension for missing LHS.
3598    if (Expr *lhsExpr = C->getLHS())
3599      if (Expr *LHS = EvalVal(lhsExpr, refVars, ParentDecl))
3600        return LHS;
3601
3602    return EvalVal(C->getRHS(), refVars, ParentDecl);
3603  }
3604
3605  // Accesses to members are potential references to data on the stack.
3606  case Stmt::MemberExprClass: {
3607    MemberExpr *M = cast<MemberExpr>(E);
3608
3609    // Check for indirect access.  We only want direct field accesses.
3610    if (M->isArrow())
3611      return NULL;
3612
3613    // Check whether the member type is itself a reference, in which case
3614    // we're not going to refer to the member, but to what the member refers to.
3615    if (M->getMemberDecl()->getType()->isReferenceType())
3616      return NULL;
3617
3618    return EvalVal(M->getBase(), refVars, ParentDecl);
3619  }
3620
3621  case Stmt::MaterializeTemporaryExprClass:
3622    if (Expr *Result = EvalVal(
3623                          cast<MaterializeTemporaryExpr>(E)->GetTemporaryExpr(),
3624                               refVars, ParentDecl))
3625      return Result;
3626
3627    return E;
3628
3629  default:
3630    // Check that we don't return or take the address of a reference to a
3631    // temporary. This is only useful in C++.
3632    if (!E->isTypeDependent() && E->isRValue())
3633      return E;
3634
3635    // Everything else: we simply don't reason about them.
3636    return NULL;
3637  }
3638} while (true);
3639}
3640
3641//===--- CHECK: Floating-Point comparisons (-Wfloat-equal) ---------------===//
3642
3643/// Check for comparisons of floating point operands using != and ==.
3644/// Issue a warning if these are no self-comparisons, as they are not likely
3645/// to do what the programmer intended.
3646void Sema::CheckFloatComparison(SourceLocation Loc, Expr* LHS, Expr *RHS) {
3647  Expr* LeftExprSansParen = LHS->IgnoreParenImpCasts();
3648  Expr* RightExprSansParen = RHS->IgnoreParenImpCasts();
3649
3650  // Special case: check for x == x (which is OK).
3651  // Do not emit warnings for such cases.
3652  if (DeclRefExpr* DRL = dyn_cast<DeclRefExpr>(LeftExprSansParen))
3653    if (DeclRefExpr* DRR = dyn_cast<DeclRefExpr>(RightExprSansParen))
3654      if (DRL->getDecl() == DRR->getDecl())
3655        return;
3656
3657
3658  // Special case: check for comparisons against literals that can be exactly
3659  //  represented by APFloat.  In such cases, do not emit a warning.  This
3660  //  is a heuristic: often comparison against such literals are used to
3661  //  detect if a value in a variable has not changed.  This clearly can
3662  //  lead to false negatives.
3663  if (FloatingLiteral* FLL = dyn_cast<FloatingLiteral>(LeftExprSansParen)) {
3664    if (FLL->isExact())
3665      return;
3666  } else
3667    if (FloatingLiteral* FLR = dyn_cast<FloatingLiteral>(RightExprSansParen))
3668      if (FLR->isExact())
3669        return;
3670
3671  // Check for comparisons with builtin types.
3672  if (CallExpr* CL = dyn_cast<CallExpr>(LeftExprSansParen))
3673    if (CL->isBuiltinCall())
3674      return;
3675
3676  if (CallExpr* CR = dyn_cast<CallExpr>(RightExprSansParen))
3677    if (CR->isBuiltinCall())
3678      return;
3679
3680  // Emit the diagnostic.
3681  Diag(Loc, diag::warn_floatingpoint_eq)
3682    << LHS->getSourceRange() << RHS->getSourceRange();
3683}
3684
3685//===--- CHECK: Integer mixed-sign comparisons (-Wsign-compare) --------===//
3686//===--- CHECK: Lossy implicit conversions (-Wconversion) --------------===//
3687
3688namespace {
3689
3690/// Structure recording the 'active' range of an integer-valued
3691/// expression.
3692struct IntRange {
3693  /// The number of bits active in the int.
3694  unsigned Width;
3695
3696  /// True if the int is known not to have negative values.
3697  bool NonNegative;
3698
3699  IntRange(unsigned Width, bool NonNegative)
3700    : Width(Width), NonNegative(NonNegative)
3701  {}
3702
3703  /// Returns the range of the bool type.
3704  static IntRange forBoolType() {
3705    return IntRange(1, true);
3706  }
3707
3708  /// Returns the range of an opaque value of the given integral type.
3709  static IntRange forValueOfType(ASTContext &C, QualType T) {
3710    return forValueOfCanonicalType(C,
3711                          T->getCanonicalTypeInternal().getTypePtr());
3712  }
3713
3714  /// Returns the range of an opaque value of a canonical integral type.
3715  static IntRange forValueOfCanonicalType(ASTContext &C, const Type *T) {
3716    assert(T->isCanonicalUnqualified());
3717
3718    if (const VectorType *VT = dyn_cast<VectorType>(T))
3719      T = VT->getElementType().getTypePtr();
3720    if (const ComplexType *CT = dyn_cast<ComplexType>(T))
3721      T = CT->getElementType().getTypePtr();
3722
3723    // For enum types, use the known bit width of the enumerators.
3724    if (const EnumType *ET = dyn_cast<EnumType>(T)) {
3725      EnumDecl *Enum = ET->getDecl();
3726      if (!Enum->isCompleteDefinition())
3727        return IntRange(C.getIntWidth(QualType(T, 0)), false);
3728
3729      unsigned NumPositive = Enum->getNumPositiveBits();
3730      unsigned NumNegative = Enum->getNumNegativeBits();
3731
3732      return IntRange(std::max(NumPositive, NumNegative), NumNegative == 0);
3733    }
3734
3735    const BuiltinType *BT = cast<BuiltinType>(T);
3736    assert(BT->isInteger());
3737
3738    return IntRange(C.getIntWidth(QualType(T, 0)), BT->isUnsignedInteger());
3739  }
3740
3741  /// Returns the "target" range of a canonical integral type, i.e.
3742  /// the range of values expressible in the type.
3743  ///
3744  /// This matches forValueOfCanonicalType except that enums have the
3745  /// full range of their type, not the range of their enumerators.
3746  static IntRange forTargetOfCanonicalType(ASTContext &C, const Type *T) {
3747    assert(T->isCanonicalUnqualified());
3748
3749    if (const VectorType *VT = dyn_cast<VectorType>(T))
3750      T = VT->getElementType().getTypePtr();
3751    if (const ComplexType *CT = dyn_cast<ComplexType>(T))
3752      T = CT->getElementType().getTypePtr();
3753    if (const EnumType *ET = dyn_cast<EnumType>(T))
3754      T = C.getCanonicalType(ET->getDecl()->getIntegerType()).getTypePtr();
3755
3756    const BuiltinType *BT = cast<BuiltinType>(T);
3757    assert(BT->isInteger());
3758
3759    return IntRange(C.getIntWidth(QualType(T, 0)), BT->isUnsignedInteger());
3760  }
3761
3762  /// Returns the supremum of two ranges: i.e. their conservative merge.
3763  static IntRange join(IntRange L, IntRange R) {
3764    return IntRange(std::max(L.Width, R.Width),
3765                    L.NonNegative && R.NonNegative);
3766  }
3767
3768  /// Returns the infinum of two ranges: i.e. their aggressive merge.
3769  static IntRange meet(IntRange L, IntRange R) {
3770    return IntRange(std::min(L.Width, R.Width),
3771                    L.NonNegative || R.NonNegative);
3772  }
3773};
3774
3775static IntRange GetValueRange(ASTContext &C, llvm::APSInt &value,
3776                              unsigned MaxWidth) {
3777  if (value.isSigned() && value.isNegative())
3778    return IntRange(value.getMinSignedBits(), false);
3779
3780  if (value.getBitWidth() > MaxWidth)
3781    value = value.trunc(MaxWidth);
3782
3783  // isNonNegative() just checks the sign bit without considering
3784  // signedness.
3785  return IntRange(value.getActiveBits(), true);
3786}
3787
3788static IntRange GetValueRange(ASTContext &C, APValue &result, QualType Ty,
3789                              unsigned MaxWidth) {
3790  if (result.isInt())
3791    return GetValueRange(C, result.getInt(), MaxWidth);
3792
3793  if (result.isVector()) {
3794    IntRange R = GetValueRange(C, result.getVectorElt(0), Ty, MaxWidth);
3795    for (unsigned i = 1, e = result.getVectorLength(); i != e; ++i) {
3796      IntRange El = GetValueRange(C, result.getVectorElt(i), Ty, MaxWidth);
3797      R = IntRange::join(R, El);
3798    }
3799    return R;
3800  }
3801
3802  if (result.isComplexInt()) {
3803    IntRange R = GetValueRange(C, result.getComplexIntReal(), MaxWidth);
3804    IntRange I = GetValueRange(C, result.getComplexIntImag(), MaxWidth);
3805    return IntRange::join(R, I);
3806  }
3807
3808  // This can happen with lossless casts to intptr_t of "based" lvalues.
3809  // Assume it might use arbitrary bits.
3810  // FIXME: The only reason we need to pass the type in here is to get
3811  // the sign right on this one case.  It would be nice if APValue
3812  // preserved this.
3813  assert(result.isLValue() || result.isAddrLabelDiff());
3814  return IntRange(MaxWidth, Ty->isUnsignedIntegerOrEnumerationType());
3815}
3816
3817/// Pseudo-evaluate the given integer expression, estimating the
3818/// range of values it might take.
3819///
3820/// \param MaxWidth - the width to which the value will be truncated
3821static IntRange GetExprRange(ASTContext &C, Expr *E, unsigned MaxWidth) {
3822  E = E->IgnoreParens();
3823
3824  // Try a full evaluation first.
3825  Expr::EvalResult result;
3826  if (E->EvaluateAsRValue(result, C))
3827    return GetValueRange(C, result.Val, E->getType(), MaxWidth);
3828
3829  // I think we only want to look through implicit casts here; if the
3830  // user has an explicit widening cast, we should treat the value as
3831  // being of the new, wider type.
3832  if (ImplicitCastExpr *CE = dyn_cast<ImplicitCastExpr>(E)) {
3833    if (CE->getCastKind() == CK_NoOp || CE->getCastKind() == CK_LValueToRValue)
3834      return GetExprRange(C, CE->getSubExpr(), MaxWidth);
3835
3836    IntRange OutputTypeRange = IntRange::forValueOfType(C, CE->getType());
3837
3838    bool isIntegerCast = (CE->getCastKind() == CK_IntegralCast);
3839
3840    // Assume that non-integer casts can span the full range of the type.
3841    if (!isIntegerCast)
3842      return OutputTypeRange;
3843
3844    IntRange SubRange
3845      = GetExprRange(C, CE->getSubExpr(),
3846                     std::min(MaxWidth, OutputTypeRange.Width));
3847
3848    // Bail out if the subexpr's range is as wide as the cast type.
3849    if (SubRange.Width >= OutputTypeRange.Width)
3850      return OutputTypeRange;
3851
3852    // Otherwise, we take the smaller width, and we're non-negative if
3853    // either the output type or the subexpr is.
3854    return IntRange(SubRange.Width,
3855                    SubRange.NonNegative || OutputTypeRange.NonNegative);
3856  }
3857
3858  if (ConditionalOperator *CO = dyn_cast<ConditionalOperator>(E)) {
3859    // If we can fold the condition, just take that operand.
3860    bool CondResult;
3861    if (CO->getCond()->EvaluateAsBooleanCondition(CondResult, C))
3862      return GetExprRange(C, CondResult ? CO->getTrueExpr()
3863                                        : CO->getFalseExpr(),
3864                          MaxWidth);
3865
3866    // Otherwise, conservatively merge.
3867    IntRange L = GetExprRange(C, CO->getTrueExpr(), MaxWidth);
3868    IntRange R = GetExprRange(C, CO->getFalseExpr(), MaxWidth);
3869    return IntRange::join(L, R);
3870  }
3871
3872  if (BinaryOperator *BO = dyn_cast<BinaryOperator>(E)) {
3873    switch (BO->getOpcode()) {
3874
3875    // Boolean-valued operations are single-bit and positive.
3876    case BO_LAnd:
3877    case BO_LOr:
3878    case BO_LT:
3879    case BO_GT:
3880    case BO_LE:
3881    case BO_GE:
3882    case BO_EQ:
3883    case BO_NE:
3884      return IntRange::forBoolType();
3885
3886    // The type of the assignments is the type of the LHS, so the RHS
3887    // is not necessarily the same type.
3888    case BO_MulAssign:
3889    case BO_DivAssign:
3890    case BO_RemAssign:
3891    case BO_AddAssign:
3892    case BO_SubAssign:
3893    case BO_XorAssign:
3894    case BO_OrAssign:
3895      // TODO: bitfields?
3896      return IntRange::forValueOfType(C, E->getType());
3897
3898    // Simple assignments just pass through the RHS, which will have
3899    // been coerced to the LHS type.
3900    case BO_Assign:
3901      // TODO: bitfields?
3902      return GetExprRange(C, BO->getRHS(), MaxWidth);
3903
3904    // Operations with opaque sources are black-listed.
3905    case BO_PtrMemD:
3906    case BO_PtrMemI:
3907      return IntRange::forValueOfType(C, E->getType());
3908
3909    // Bitwise-and uses the *infinum* of the two source ranges.
3910    case BO_And:
3911    case BO_AndAssign:
3912      return IntRange::meet(GetExprRange(C, BO->getLHS(), MaxWidth),
3913                            GetExprRange(C, BO->getRHS(), MaxWidth));
3914
3915    // Left shift gets black-listed based on a judgement call.
3916    case BO_Shl:
3917      // ...except that we want to treat '1 << (blah)' as logically
3918      // positive.  It's an important idiom.
3919      if (IntegerLiteral *I
3920            = dyn_cast<IntegerLiteral>(BO->getLHS()->IgnoreParenCasts())) {
3921        if (I->getValue() == 1) {
3922          IntRange R = IntRange::forValueOfType(C, E->getType());
3923          return IntRange(R.Width, /*NonNegative*/ true);
3924        }
3925      }
3926      // fallthrough
3927
3928    case BO_ShlAssign:
3929      return IntRange::forValueOfType(C, E->getType());
3930
3931    // Right shift by a constant can narrow its left argument.
3932    case BO_Shr:
3933    case BO_ShrAssign: {
3934      IntRange L = GetExprRange(C, BO->getLHS(), MaxWidth);
3935
3936      // If the shift amount is a positive constant, drop the width by
3937      // that much.
3938      llvm::APSInt shift;
3939      if (BO->getRHS()->isIntegerConstantExpr(shift, C) &&
3940          shift.isNonNegative()) {
3941        unsigned zext = shift.getZExtValue();
3942        if (zext >= L.Width)
3943          L.Width = (L.NonNegative ? 0 : 1);
3944        else
3945          L.Width -= zext;
3946      }
3947
3948      return L;
3949    }
3950
3951    // Comma acts as its right operand.
3952    case BO_Comma:
3953      return GetExprRange(C, BO->getRHS(), MaxWidth);
3954
3955    // Black-list pointer subtractions.
3956    case BO_Sub:
3957      if (BO->getLHS()->getType()->isPointerType())
3958        return IntRange::forValueOfType(C, E->getType());
3959      break;
3960
3961    // The width of a division result is mostly determined by the size
3962    // of the LHS.
3963    case BO_Div: {
3964      // Don't 'pre-truncate' the operands.
3965      unsigned opWidth = C.getIntWidth(E->getType());
3966      IntRange L = GetExprRange(C, BO->getLHS(), opWidth);
3967
3968      // If the divisor is constant, use that.
3969      llvm::APSInt divisor;
3970      if (BO->getRHS()->isIntegerConstantExpr(divisor, C)) {
3971        unsigned log2 = divisor.logBase2(); // floor(log_2(divisor))
3972        if (log2 >= L.Width)
3973          L.Width = (L.NonNegative ? 0 : 1);
3974        else
3975          L.Width = std::min(L.Width - log2, MaxWidth);
3976        return L;
3977      }
3978
3979      // Otherwise, just use the LHS's width.
3980      IntRange R = GetExprRange(C, BO->getRHS(), opWidth);
3981      return IntRange(L.Width, L.NonNegative && R.NonNegative);
3982    }
3983
3984    // The result of a remainder can't be larger than the result of
3985    // either side.
3986    case BO_Rem: {
3987      // Don't 'pre-truncate' the operands.
3988      unsigned opWidth = C.getIntWidth(E->getType());
3989      IntRange L = GetExprRange(C, BO->getLHS(), opWidth);
3990      IntRange R = GetExprRange(C, BO->getRHS(), opWidth);
3991
3992      IntRange meet = IntRange::meet(L, R);
3993      meet.Width = std::min(meet.Width, MaxWidth);
3994      return meet;
3995    }
3996
3997    // The default behavior is okay for these.
3998    case BO_Mul:
3999    case BO_Add:
4000    case BO_Xor:
4001    case BO_Or:
4002      break;
4003    }
4004
4005    // The default case is to treat the operation as if it were closed
4006    // on the narrowest type that encompasses both operands.
4007    IntRange L = GetExprRange(C, BO->getLHS(), MaxWidth);
4008    IntRange R = GetExprRange(C, BO->getRHS(), MaxWidth);
4009    return IntRange::join(L, R);
4010  }
4011
4012  if (UnaryOperator *UO = dyn_cast<UnaryOperator>(E)) {
4013    switch (UO->getOpcode()) {
4014    // Boolean-valued operations are white-listed.
4015    case UO_LNot:
4016      return IntRange::forBoolType();
4017
4018    // Operations with opaque sources are black-listed.
4019    case UO_Deref:
4020    case UO_AddrOf: // should be impossible
4021      return IntRange::forValueOfType(C, E->getType());
4022
4023    default:
4024      return GetExprRange(C, UO->getSubExpr(), MaxWidth);
4025    }
4026  }
4027
4028  if (dyn_cast<OffsetOfExpr>(E)) {
4029    IntRange::forValueOfType(C, E->getType());
4030  }
4031
4032  if (FieldDecl *BitField = E->getBitField())
4033    return IntRange(BitField->getBitWidthValue(C),
4034                    BitField->getType()->isUnsignedIntegerOrEnumerationType());
4035
4036  return IntRange::forValueOfType(C, E->getType());
4037}
4038
4039static IntRange GetExprRange(ASTContext &C, Expr *E) {
4040  return GetExprRange(C, E, C.getIntWidth(E->getType()));
4041}
4042
4043/// Checks whether the given value, which currently has the given
4044/// source semantics, has the same value when coerced through the
4045/// target semantics.
4046static bool IsSameFloatAfterCast(const llvm::APFloat &value,
4047                                 const llvm::fltSemantics &Src,
4048                                 const llvm::fltSemantics &Tgt) {
4049  llvm::APFloat truncated = value;
4050
4051  bool ignored;
4052  truncated.convert(Src, llvm::APFloat::rmNearestTiesToEven, &ignored);
4053  truncated.convert(Tgt, llvm::APFloat::rmNearestTiesToEven, &ignored);
4054
4055  return truncated.bitwiseIsEqual(value);
4056}
4057
4058/// Checks whether the given value, which currently has the given
4059/// source semantics, has the same value when coerced through the
4060/// target semantics.
4061///
4062/// The value might be a vector of floats (or a complex number).
4063static bool IsSameFloatAfterCast(const APValue &value,
4064                                 const llvm::fltSemantics &Src,
4065                                 const llvm::fltSemantics &Tgt) {
4066  if (value.isFloat())
4067    return IsSameFloatAfterCast(value.getFloat(), Src, Tgt);
4068
4069  if (value.isVector()) {
4070    for (unsigned i = 0, e = value.getVectorLength(); i != e; ++i)
4071      if (!IsSameFloatAfterCast(value.getVectorElt(i), Src, Tgt))
4072        return false;
4073    return true;
4074  }
4075
4076  assert(value.isComplexFloat());
4077  return (IsSameFloatAfterCast(value.getComplexFloatReal(), Src, Tgt) &&
4078          IsSameFloatAfterCast(value.getComplexFloatImag(), Src, Tgt));
4079}
4080
4081static void AnalyzeImplicitConversions(Sema &S, Expr *E, SourceLocation CC);
4082
4083static bool IsZero(Sema &S, Expr *E) {
4084  // Suppress cases where we are comparing against an enum constant.
4085  if (const DeclRefExpr *DR =
4086      dyn_cast<DeclRefExpr>(E->IgnoreParenImpCasts()))
4087    if (isa<EnumConstantDecl>(DR->getDecl()))
4088      return false;
4089
4090  // Suppress cases where the '0' value is expanded from a macro.
4091  if (E->getLocStart().isMacroID())
4092    return false;
4093
4094  llvm::APSInt Value;
4095  return E->isIntegerConstantExpr(Value, S.Context) && Value == 0;
4096}
4097
4098static bool HasEnumType(Expr *E) {
4099  // Strip off implicit integral promotions.
4100  while (ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(E)) {
4101    if (ICE->getCastKind() != CK_IntegralCast &&
4102        ICE->getCastKind() != CK_NoOp)
4103      break;
4104    E = ICE->getSubExpr();
4105  }
4106
4107  return E->getType()->isEnumeralType();
4108}
4109
4110static void CheckTrivialUnsignedComparison(Sema &S, BinaryOperator *E) {
4111  BinaryOperatorKind op = E->getOpcode();
4112  if (E->isValueDependent())
4113    return;
4114
4115  if (op == BO_LT && IsZero(S, E->getRHS())) {
4116    S.Diag(E->getOperatorLoc(), diag::warn_lunsigned_always_true_comparison)
4117      << "< 0" << "false" << HasEnumType(E->getLHS())
4118      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4119  } else if (op == BO_GE && IsZero(S, E->getRHS())) {
4120    S.Diag(E->getOperatorLoc(), diag::warn_lunsigned_always_true_comparison)
4121      << ">= 0" << "true" << HasEnumType(E->getLHS())
4122      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4123  } else if (op == BO_GT && IsZero(S, E->getLHS())) {
4124    S.Diag(E->getOperatorLoc(), diag::warn_runsigned_always_true_comparison)
4125      << "0 >" << "false" << HasEnumType(E->getRHS())
4126      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4127  } else if (op == BO_LE && IsZero(S, E->getLHS())) {
4128    S.Diag(E->getOperatorLoc(), diag::warn_runsigned_always_true_comparison)
4129      << "0 <=" << "true" << HasEnumType(E->getRHS())
4130      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
4131  }
4132}
4133
4134/// Analyze the operands of the given comparison.  Implements the
4135/// fallback case from AnalyzeComparison.
4136static void AnalyzeImpConvsInComparison(Sema &S, BinaryOperator *E) {
4137  AnalyzeImplicitConversions(S, E->getLHS(), E->getOperatorLoc());
4138  AnalyzeImplicitConversions(S, E->getRHS(), E->getOperatorLoc());
4139}
4140
4141/// \brief Implements -Wsign-compare.
4142///
4143/// \param E the binary operator to check for warnings
4144static void AnalyzeComparison(Sema &S, BinaryOperator *E) {
4145  // The type the comparison is being performed in.
4146  QualType T = E->getLHS()->getType();
4147  assert(S.Context.hasSameUnqualifiedType(T, E->getRHS()->getType())
4148         && "comparison with mismatched types");
4149
4150  // We don't do anything special if this isn't an unsigned integral
4151  // comparison:  we're only interested in integral comparisons, and
4152  // signed comparisons only happen in cases we don't care to warn about.
4153  //
4154  // We also don't care about value-dependent expressions or expressions
4155  // whose result is a constant.
4156  if (!T->hasUnsignedIntegerRepresentation()
4157      || E->isValueDependent() || E->isIntegerConstantExpr(S.Context))
4158    return AnalyzeImpConvsInComparison(S, E);
4159
4160  Expr *LHS = E->getLHS()->IgnoreParenImpCasts();
4161  Expr *RHS = E->getRHS()->IgnoreParenImpCasts();
4162
4163  // Check to see if one of the (unmodified) operands is of different
4164  // signedness.
4165  Expr *signedOperand, *unsignedOperand;
4166  if (LHS->getType()->hasSignedIntegerRepresentation()) {
4167    assert(!RHS->getType()->hasSignedIntegerRepresentation() &&
4168           "unsigned comparison between two signed integer expressions?");
4169    signedOperand = LHS;
4170    unsignedOperand = RHS;
4171  } else if (RHS->getType()->hasSignedIntegerRepresentation()) {
4172    signedOperand = RHS;
4173    unsignedOperand = LHS;
4174  } else {
4175    CheckTrivialUnsignedComparison(S, E);
4176    return AnalyzeImpConvsInComparison(S, E);
4177  }
4178
4179  // Otherwise, calculate the effective range of the signed operand.
4180  IntRange signedRange = GetExprRange(S.Context, signedOperand);
4181
4182  // Go ahead and analyze implicit conversions in the operands.  Note
4183  // that we skip the implicit conversions on both sides.
4184  AnalyzeImplicitConversions(S, LHS, E->getOperatorLoc());
4185  AnalyzeImplicitConversions(S, RHS, E->getOperatorLoc());
4186
4187  // If the signed range is non-negative, -Wsign-compare won't fire,
4188  // but we should still check for comparisons which are always true
4189  // or false.
4190  if (signedRange.NonNegative)
4191    return CheckTrivialUnsignedComparison(S, E);
4192
4193  // For (in)equality comparisons, if the unsigned operand is a
4194  // constant which cannot collide with a overflowed signed operand,
4195  // then reinterpreting the signed operand as unsigned will not
4196  // change the result of the comparison.
4197  if (E->isEqualityOp()) {
4198    unsigned comparisonWidth = S.Context.getIntWidth(T);
4199    IntRange unsignedRange = GetExprRange(S.Context, unsignedOperand);
4200
4201    // We should never be unable to prove that the unsigned operand is
4202    // non-negative.
4203    assert(unsignedRange.NonNegative && "unsigned range includes negative?");
4204
4205    if (unsignedRange.Width < comparisonWidth)
4206      return;
4207  }
4208
4209  S.DiagRuntimeBehavior(E->getOperatorLoc(), E,
4210    S.PDiag(diag::warn_mixed_sign_comparison)
4211      << LHS->getType() << RHS->getType()
4212      << LHS->getSourceRange() << RHS->getSourceRange());
4213}
4214
4215/// Analyzes an attempt to assign the given value to a bitfield.
4216///
4217/// Returns true if there was something fishy about the attempt.
4218static bool AnalyzeBitFieldAssignment(Sema &S, FieldDecl *Bitfield, Expr *Init,
4219                                      SourceLocation InitLoc) {
4220  assert(Bitfield->isBitField());
4221  if (Bitfield->isInvalidDecl())
4222    return false;
4223
4224  // White-list bool bitfields.
4225  if (Bitfield->getType()->isBooleanType())
4226    return false;
4227
4228  // Ignore value- or type-dependent expressions.
4229  if (Bitfield->getBitWidth()->isValueDependent() ||
4230      Bitfield->getBitWidth()->isTypeDependent() ||
4231      Init->isValueDependent() ||
4232      Init->isTypeDependent())
4233    return false;
4234
4235  Expr *OriginalInit = Init->IgnoreParenImpCasts();
4236
4237  llvm::APSInt Value;
4238  if (!OriginalInit->EvaluateAsInt(Value, S.Context, Expr::SE_AllowSideEffects))
4239    return false;
4240
4241  unsigned OriginalWidth = Value.getBitWidth();
4242  unsigned FieldWidth = Bitfield->getBitWidthValue(S.Context);
4243
4244  if (OriginalWidth <= FieldWidth)
4245    return false;
4246
4247  // Compute the value which the bitfield will contain.
4248  llvm::APSInt TruncatedValue = Value.trunc(FieldWidth);
4249  TruncatedValue.setIsSigned(Bitfield->getType()->isSignedIntegerType());
4250
4251  // Check whether the stored value is equal to the original value.
4252  TruncatedValue = TruncatedValue.extend(OriginalWidth);
4253  if (llvm::APSInt::isSameValue(Value, TruncatedValue))
4254    return false;
4255
4256  // Special-case bitfields of width 1: booleans are naturally 0/1, and
4257  // therefore don't strictly fit into a signed bitfield of width 1.
4258  if (FieldWidth == 1 && Value == 1)
4259    return false;
4260
4261  std::string PrettyValue = Value.toString(10);
4262  std::string PrettyTrunc = TruncatedValue.toString(10);
4263
4264  S.Diag(InitLoc, diag::warn_impcast_bitfield_precision_constant)
4265    << PrettyValue << PrettyTrunc << OriginalInit->getType()
4266    << Init->getSourceRange();
4267
4268  return true;
4269}
4270
4271/// Analyze the given simple or compound assignment for warning-worthy
4272/// operations.
4273static void AnalyzeAssignment(Sema &S, BinaryOperator *E) {
4274  // Just recurse on the LHS.
4275  AnalyzeImplicitConversions(S, E->getLHS(), E->getOperatorLoc());
4276
4277  // We want to recurse on the RHS as normal unless we're assigning to
4278  // a bitfield.
4279  if (FieldDecl *Bitfield = E->getLHS()->getBitField()) {
4280    if (AnalyzeBitFieldAssignment(S, Bitfield, E->getRHS(),
4281                                  E->getOperatorLoc())) {
4282      // Recurse, ignoring any implicit conversions on the RHS.
4283      return AnalyzeImplicitConversions(S, E->getRHS()->IgnoreParenImpCasts(),
4284                                        E->getOperatorLoc());
4285    }
4286  }
4287
4288  AnalyzeImplicitConversions(S, E->getRHS(), E->getOperatorLoc());
4289}
4290
4291/// Diagnose an implicit cast;  purely a helper for CheckImplicitConversion.
4292static void DiagnoseImpCast(Sema &S, Expr *E, QualType SourceType, QualType T,
4293                            SourceLocation CContext, unsigned diag,
4294                            bool pruneControlFlow = false) {
4295  if (pruneControlFlow) {
4296    S.DiagRuntimeBehavior(E->getExprLoc(), E,
4297                          S.PDiag(diag)
4298                            << SourceType << T << E->getSourceRange()
4299                            << SourceRange(CContext));
4300    return;
4301  }
4302  S.Diag(E->getExprLoc(), diag)
4303    << SourceType << T << E->getSourceRange() << SourceRange(CContext);
4304}
4305
4306/// Diagnose an implicit cast;  purely a helper for CheckImplicitConversion.
4307static void DiagnoseImpCast(Sema &S, Expr *E, QualType T,
4308                            SourceLocation CContext, unsigned diag,
4309                            bool pruneControlFlow = false) {
4310  DiagnoseImpCast(S, E, E->getType(), T, CContext, diag, pruneControlFlow);
4311}
4312
4313/// Diagnose an implicit cast from a literal expression. Does not warn when the
4314/// cast wouldn't lose information.
4315void DiagnoseFloatingLiteralImpCast(Sema &S, FloatingLiteral *FL, QualType T,
4316                                    SourceLocation CContext) {
4317  // Try to convert the literal exactly to an integer. If we can, don't warn.
4318  bool isExact = false;
4319  const llvm::APFloat &Value = FL->getValue();
4320  llvm::APSInt IntegerValue(S.Context.getIntWidth(T),
4321                            T->hasUnsignedIntegerRepresentation());
4322  if (Value.convertToInteger(IntegerValue,
4323                             llvm::APFloat::rmTowardZero, &isExact)
4324      == llvm::APFloat::opOK && isExact)
4325    return;
4326
4327  SmallString<16> PrettySourceValue;
4328  Value.toString(PrettySourceValue);
4329  SmallString<16> PrettyTargetValue;
4330  if (T->isSpecificBuiltinType(BuiltinType::Bool))
4331    PrettyTargetValue = IntegerValue == 0 ? "false" : "true";
4332  else
4333    IntegerValue.toString(PrettyTargetValue);
4334
4335  S.Diag(FL->getExprLoc(), diag::warn_impcast_literal_float_to_integer)
4336    << FL->getType() << T.getUnqualifiedType() << PrettySourceValue
4337    << PrettyTargetValue << FL->getSourceRange() << SourceRange(CContext);
4338}
4339
4340std::string PrettyPrintInRange(const llvm::APSInt &Value, IntRange Range) {
4341  if (!Range.Width) return "0";
4342
4343  llvm::APSInt ValueInRange = Value;
4344  ValueInRange.setIsSigned(!Range.NonNegative);
4345  ValueInRange = ValueInRange.trunc(Range.Width);
4346  return ValueInRange.toString(10);
4347}
4348
4349void CheckImplicitConversion(Sema &S, Expr *E, QualType T,
4350                             SourceLocation CC, bool *ICContext = 0) {
4351  if (E->isTypeDependent() || E->isValueDependent()) return;
4352
4353  const Type *Source = S.Context.getCanonicalType(E->getType()).getTypePtr();
4354  const Type *Target = S.Context.getCanonicalType(T).getTypePtr();
4355  if (Source == Target) return;
4356  if (Target->isDependentType()) return;
4357
4358  // If the conversion context location is invalid don't complain. We also
4359  // don't want to emit a warning if the issue occurs from the expansion of
4360  // a system macro. The problem is that 'getSpellingLoc()' is slow, so we
4361  // delay this check as long as possible. Once we detect we are in that
4362  // scenario, we just return.
4363  if (CC.isInvalid())
4364    return;
4365
4366  // Diagnose implicit casts to bool.
4367  if (Target->isSpecificBuiltinType(BuiltinType::Bool)) {
4368    if (isa<StringLiteral>(E))
4369      // Warn on string literal to bool.  Checks for string literals in logical
4370      // expressions, for instances, assert(0 && "error here"), is prevented
4371      // by a check in AnalyzeImplicitConversions().
4372      return DiagnoseImpCast(S, E, T, CC,
4373                             diag::warn_impcast_string_literal_to_bool);
4374    if (Source->isFunctionType()) {
4375      // Warn on function to bool. Checks free functions and static member
4376      // functions. Weakly imported functions are excluded from the check,
4377      // since it's common to test their value to check whether the linker
4378      // found a definition for them.
4379      ValueDecl *D = 0;
4380      if (DeclRefExpr* R = dyn_cast<DeclRefExpr>(E)) {
4381        D = R->getDecl();
4382      } else if (MemberExpr *M = dyn_cast<MemberExpr>(E)) {
4383        D = M->getMemberDecl();
4384      }
4385
4386      if (D && !D->isWeak()) {
4387        if (FunctionDecl* F = dyn_cast<FunctionDecl>(D)) {
4388          S.Diag(E->getExprLoc(), diag::warn_impcast_function_to_bool)
4389            << F << E->getSourceRange() << SourceRange(CC);
4390          S.Diag(E->getExprLoc(), diag::note_function_to_bool_silence)
4391            << FixItHint::CreateInsertion(E->getExprLoc(), "&");
4392          QualType ReturnType;
4393          UnresolvedSet<4> NonTemplateOverloads;
4394          S.isExprCallable(*E, ReturnType, NonTemplateOverloads);
4395          if (!ReturnType.isNull()
4396              && ReturnType->isSpecificBuiltinType(BuiltinType::Bool))
4397            S.Diag(E->getExprLoc(), diag::note_function_to_bool_call)
4398              << FixItHint::CreateInsertion(
4399                 S.getPreprocessor().getLocForEndOfToken(E->getLocEnd()), "()");
4400          return;
4401        }
4402      }
4403    }
4404  }
4405
4406  // Strip vector types.
4407  if (isa<VectorType>(Source)) {
4408    if (!isa<VectorType>(Target)) {
4409      if (S.SourceMgr.isInSystemMacro(CC))
4410        return;
4411      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_vector_scalar);
4412    }
4413
4414    // If the vector cast is cast between two vectors of the same size, it is
4415    // a bitcast, not a conversion.
4416    if (S.Context.getTypeSize(Source) == S.Context.getTypeSize(Target))
4417      return;
4418
4419    Source = cast<VectorType>(Source)->getElementType().getTypePtr();
4420    Target = cast<VectorType>(Target)->getElementType().getTypePtr();
4421  }
4422
4423  // Strip complex types.
4424  if (isa<ComplexType>(Source)) {
4425    if (!isa<ComplexType>(Target)) {
4426      if (S.SourceMgr.isInSystemMacro(CC))
4427        return;
4428
4429      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_complex_scalar);
4430    }
4431
4432    Source = cast<ComplexType>(Source)->getElementType().getTypePtr();
4433    Target = cast<ComplexType>(Target)->getElementType().getTypePtr();
4434  }
4435
4436  const BuiltinType *SourceBT = dyn_cast<BuiltinType>(Source);
4437  const BuiltinType *TargetBT = dyn_cast<BuiltinType>(Target);
4438
4439  // If the source is floating point...
4440  if (SourceBT && SourceBT->isFloatingPoint()) {
4441    // ...and the target is floating point...
4442    if (TargetBT && TargetBT->isFloatingPoint()) {
4443      // ...then warn if we're dropping FP rank.
4444
4445      // Builtin FP kinds are ordered by increasing FP rank.
4446      if (SourceBT->getKind() > TargetBT->getKind()) {
4447        // Don't warn about float constants that are precisely
4448        // representable in the target type.
4449        Expr::EvalResult result;
4450        if (E->EvaluateAsRValue(result, S.Context)) {
4451          // Value might be a float, a float vector, or a float complex.
4452          if (IsSameFloatAfterCast(result.Val,
4453                   S.Context.getFloatTypeSemantics(QualType(TargetBT, 0)),
4454                   S.Context.getFloatTypeSemantics(QualType(SourceBT, 0))))
4455            return;
4456        }
4457
4458        if (S.SourceMgr.isInSystemMacro(CC))
4459          return;
4460
4461        DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_float_precision);
4462      }
4463      return;
4464    }
4465
4466    // If the target is integral, always warn.
4467    if (TargetBT && TargetBT->isInteger()) {
4468      if (S.SourceMgr.isInSystemMacro(CC))
4469        return;
4470
4471      Expr *InnerE = E->IgnoreParenImpCasts();
4472      // We also want to warn on, e.g., "int i = -1.234"
4473      if (UnaryOperator *UOp = dyn_cast<UnaryOperator>(InnerE))
4474        if (UOp->getOpcode() == UO_Minus || UOp->getOpcode() == UO_Plus)
4475          InnerE = UOp->getSubExpr()->IgnoreParenImpCasts();
4476
4477      if (FloatingLiteral *FL = dyn_cast<FloatingLiteral>(InnerE)) {
4478        DiagnoseFloatingLiteralImpCast(S, FL, T, CC);
4479      } else {
4480        DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_float_integer);
4481      }
4482    }
4483
4484    return;
4485  }
4486
4487  if ((E->isNullPointerConstant(S.Context, Expr::NPC_ValueDependentIsNotNull)
4488           == Expr::NPCK_GNUNull) && !Target->isAnyPointerType()
4489      && !Target->isBlockPointerType() && !Target->isMemberPointerType()) {
4490    SourceLocation Loc = E->getSourceRange().getBegin();
4491    if (Loc.isMacroID())
4492      Loc = S.SourceMgr.getImmediateExpansionRange(Loc).first;
4493    if (!Loc.isMacroID() || CC.isMacroID())
4494      S.Diag(Loc, diag::warn_impcast_null_pointer_to_integer)
4495          << T << clang::SourceRange(CC)
4496          << FixItHint::CreateReplacement(Loc, S.getFixItZeroLiteralForType(T));
4497  }
4498
4499  if (!Source->isIntegerType() || !Target->isIntegerType())
4500    return;
4501
4502  // TODO: remove this early return once the false positives for constant->bool
4503  // in templates, macros, etc, are reduced or removed.
4504  if (Target->isSpecificBuiltinType(BuiltinType::Bool))
4505    return;
4506
4507  IntRange SourceRange = GetExprRange(S.Context, E);
4508  IntRange TargetRange = IntRange::forTargetOfCanonicalType(S.Context, Target);
4509
4510  if (SourceRange.Width > TargetRange.Width) {
4511    // If the source is a constant, use a default-on diagnostic.
4512    // TODO: this should happen for bitfield stores, too.
4513    llvm::APSInt Value(32);
4514    if (E->isIntegerConstantExpr(Value, S.Context)) {
4515      if (S.SourceMgr.isInSystemMacro(CC))
4516        return;
4517
4518      std::string PrettySourceValue = Value.toString(10);
4519      std::string PrettyTargetValue = PrettyPrintInRange(Value, TargetRange);
4520
4521      S.DiagRuntimeBehavior(E->getExprLoc(), E,
4522        S.PDiag(diag::warn_impcast_integer_precision_constant)
4523            << PrettySourceValue << PrettyTargetValue
4524            << E->getType() << T << E->getSourceRange()
4525            << clang::SourceRange(CC));
4526      return;
4527    }
4528
4529    // People want to build with -Wshorten-64-to-32 and not -Wconversion.
4530    if (S.SourceMgr.isInSystemMacro(CC))
4531      return;
4532
4533    if (TargetRange.Width == 32 && S.Context.getIntWidth(E->getType()) == 64)
4534      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_integer_64_32,
4535                             /* pruneControlFlow */ true);
4536    return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_integer_precision);
4537  }
4538
4539  if ((TargetRange.NonNegative && !SourceRange.NonNegative) ||
4540      (!TargetRange.NonNegative && SourceRange.NonNegative &&
4541       SourceRange.Width == TargetRange.Width)) {
4542
4543    if (S.SourceMgr.isInSystemMacro(CC))
4544      return;
4545
4546    unsigned DiagID = diag::warn_impcast_integer_sign;
4547
4548    // Traditionally, gcc has warned about this under -Wsign-compare.
4549    // We also want to warn about it in -Wconversion.
4550    // So if -Wconversion is off, use a completely identical diagnostic
4551    // in the sign-compare group.
4552    // The conditional-checking code will
4553    if (ICContext) {
4554      DiagID = diag::warn_impcast_integer_sign_conditional;
4555      *ICContext = true;
4556    }
4557
4558    return DiagnoseImpCast(S, E, T, CC, DiagID);
4559  }
4560
4561  // Diagnose conversions between different enumeration types.
4562  // In C, we pretend that the type of an EnumConstantDecl is its enumeration
4563  // type, to give us better diagnostics.
4564  QualType SourceType = E->getType();
4565  if (!S.getLangOpts().CPlusPlus) {
4566    if (DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(E))
4567      if (EnumConstantDecl *ECD = dyn_cast<EnumConstantDecl>(DRE->getDecl())) {
4568        EnumDecl *Enum = cast<EnumDecl>(ECD->getDeclContext());
4569        SourceType = S.Context.getTypeDeclType(Enum);
4570        Source = S.Context.getCanonicalType(SourceType).getTypePtr();
4571      }
4572  }
4573
4574  if (const EnumType *SourceEnum = Source->getAs<EnumType>())
4575    if (const EnumType *TargetEnum = Target->getAs<EnumType>())
4576      if ((SourceEnum->getDecl()->getIdentifier() ||
4577           SourceEnum->getDecl()->getTypedefNameForAnonDecl()) &&
4578          (TargetEnum->getDecl()->getIdentifier() ||
4579           TargetEnum->getDecl()->getTypedefNameForAnonDecl()) &&
4580          SourceEnum != TargetEnum) {
4581        if (S.SourceMgr.isInSystemMacro(CC))
4582          return;
4583
4584        return DiagnoseImpCast(S, E, SourceType, T, CC,
4585                               diag::warn_impcast_different_enum_types);
4586      }
4587
4588  return;
4589}
4590
4591void CheckConditionalOperator(Sema &S, ConditionalOperator *E,
4592                              SourceLocation CC, QualType T);
4593
4594void CheckConditionalOperand(Sema &S, Expr *E, QualType T,
4595                             SourceLocation CC, bool &ICContext) {
4596  E = E->IgnoreParenImpCasts();
4597
4598  if (isa<ConditionalOperator>(E))
4599    return CheckConditionalOperator(S, cast<ConditionalOperator>(E), CC, T);
4600
4601  AnalyzeImplicitConversions(S, E, CC);
4602  if (E->getType() != T)
4603    return CheckImplicitConversion(S, E, T, CC, &ICContext);
4604  return;
4605}
4606
4607void CheckConditionalOperator(Sema &S, ConditionalOperator *E,
4608                              SourceLocation CC, QualType T) {
4609  AnalyzeImplicitConversions(S, E->getCond(), CC);
4610
4611  bool Suspicious = false;
4612  CheckConditionalOperand(S, E->getTrueExpr(), T, CC, Suspicious);
4613  CheckConditionalOperand(S, E->getFalseExpr(), T, CC, Suspicious);
4614
4615  // If -Wconversion would have warned about either of the candidates
4616  // for a signedness conversion to the context type...
4617  if (!Suspicious) return;
4618
4619  // ...but it's currently ignored...
4620  if (S.Diags.getDiagnosticLevel(diag::warn_impcast_integer_sign_conditional,
4621                                 CC))
4622    return;
4623
4624  // ...then check whether it would have warned about either of the
4625  // candidates for a signedness conversion to the condition type.
4626  if (E->getType() == T) return;
4627
4628  Suspicious = false;
4629  CheckImplicitConversion(S, E->getTrueExpr()->IgnoreParenImpCasts(),
4630                          E->getType(), CC, &Suspicious);
4631  if (!Suspicious)
4632    CheckImplicitConversion(S, E->getFalseExpr()->IgnoreParenImpCasts(),
4633                            E->getType(), CC, &Suspicious);
4634}
4635
4636/// AnalyzeImplicitConversions - Find and report any interesting
4637/// implicit conversions in the given expression.  There are a couple
4638/// of competing diagnostics here, -Wconversion and -Wsign-compare.
4639void AnalyzeImplicitConversions(Sema &S, Expr *OrigE, SourceLocation CC) {
4640  QualType T = OrigE->getType();
4641  Expr *E = OrigE->IgnoreParenImpCasts();
4642
4643  if (E->isTypeDependent() || E->isValueDependent())
4644    return;
4645
4646  // For conditional operators, we analyze the arguments as if they
4647  // were being fed directly into the output.
4648  if (isa<ConditionalOperator>(E)) {
4649    ConditionalOperator *CO = cast<ConditionalOperator>(E);
4650    CheckConditionalOperator(S, CO, CC, T);
4651    return;
4652  }
4653
4654  // Go ahead and check any implicit conversions we might have skipped.
4655  // The non-canonical typecheck is just an optimization;
4656  // CheckImplicitConversion will filter out dead implicit conversions.
4657  if (E->getType() != T)
4658    CheckImplicitConversion(S, E, T, CC);
4659
4660  // Now continue drilling into this expression.
4661
4662  // Skip past explicit casts.
4663  if (isa<ExplicitCastExpr>(E)) {
4664    E = cast<ExplicitCastExpr>(E)->getSubExpr()->IgnoreParenImpCasts();
4665    return AnalyzeImplicitConversions(S, E, CC);
4666  }
4667
4668  if (BinaryOperator *BO = dyn_cast<BinaryOperator>(E)) {
4669    // Do a somewhat different check with comparison operators.
4670    if (BO->isComparisonOp())
4671      return AnalyzeComparison(S, BO);
4672
4673    // And with simple assignments.
4674    if (BO->getOpcode() == BO_Assign)
4675      return AnalyzeAssignment(S, BO);
4676  }
4677
4678  // These break the otherwise-useful invariant below.  Fortunately,
4679  // we don't really need to recurse into them, because any internal
4680  // expressions should have been analyzed already when they were
4681  // built into statements.
4682  if (isa<StmtExpr>(E)) return;
4683
4684  // Don't descend into unevaluated contexts.
4685  if (isa<UnaryExprOrTypeTraitExpr>(E)) return;
4686
4687  // Now just recurse over the expression's children.
4688  CC = E->getExprLoc();
4689  BinaryOperator *BO = dyn_cast<BinaryOperator>(E);
4690  bool IsLogicalOperator = BO && BO->isLogicalOp();
4691  for (Stmt::child_range I = E->children(); I; ++I) {
4692    Expr *ChildExpr = dyn_cast_or_null<Expr>(*I);
4693    if (!ChildExpr)
4694      continue;
4695
4696    if (IsLogicalOperator &&
4697        isa<StringLiteral>(ChildExpr->IgnoreParenImpCasts()))
4698      // Ignore checking string literals that are in logical operators.
4699      continue;
4700    AnalyzeImplicitConversions(S, ChildExpr, CC);
4701  }
4702}
4703
4704} // end anonymous namespace
4705
4706/// Diagnoses "dangerous" implicit conversions within the given
4707/// expression (which is a full expression).  Implements -Wconversion
4708/// and -Wsign-compare.
4709///
4710/// \param CC the "context" location of the implicit conversion, i.e.
4711///   the most location of the syntactic entity requiring the implicit
4712///   conversion
4713void Sema::CheckImplicitConversions(Expr *E, SourceLocation CC) {
4714  // Don't diagnose in unevaluated contexts.
4715  if (isUnevaluatedContext())
4716    return;
4717
4718  // Don't diagnose for value- or type-dependent expressions.
4719  if (E->isTypeDependent() || E->isValueDependent())
4720    return;
4721
4722  // Check for array bounds violations in cases where the check isn't triggered
4723  // elsewhere for other Expr types (like BinaryOperators), e.g. when an
4724  // ArraySubscriptExpr is on the RHS of a variable initialization.
4725  CheckArrayAccess(E);
4726
4727  // This is not the right CC for (e.g.) a variable initialization.
4728  AnalyzeImplicitConversions(*this, E, CC);
4729}
4730
4731void Sema::CheckBitFieldInitialization(SourceLocation InitLoc,
4732                                       FieldDecl *BitField,
4733                                       Expr *Init) {
4734  (void) AnalyzeBitFieldAssignment(*this, BitField, Init, InitLoc);
4735}
4736
4737/// CheckParmsForFunctionDef - Check that the parameters of the given
4738/// function are appropriate for the definition of a function. This
4739/// takes care of any checks that cannot be performed on the
4740/// declaration itself, e.g., that the types of each of the function
4741/// parameters are complete.
4742bool Sema::CheckParmsForFunctionDef(ParmVarDecl **P, ParmVarDecl **PEnd,
4743                                    bool CheckParameterNames) {
4744  bool HasInvalidParm = false;
4745  for (; P != PEnd; ++P) {
4746    ParmVarDecl *Param = *P;
4747
4748    // C99 6.7.5.3p4: the parameters in a parameter type list in a
4749    // function declarator that is part of a function definition of
4750    // that function shall not have incomplete type.
4751    //
4752    // This is also C++ [dcl.fct]p6.
4753    if (!Param->isInvalidDecl() &&
4754        RequireCompleteType(Param->getLocation(), Param->getType(),
4755                            diag::err_typecheck_decl_incomplete_type)) {
4756      Param->setInvalidDecl();
4757      HasInvalidParm = true;
4758    }
4759
4760    // C99 6.9.1p5: If the declarator includes a parameter type list, the
4761    // declaration of each parameter shall include an identifier.
4762    if (CheckParameterNames &&
4763        Param->getIdentifier() == 0 &&
4764        !Param->isImplicit() &&
4765        !getLangOpts().CPlusPlus)
4766      Diag(Param->getLocation(), diag::err_parameter_name_omitted);
4767
4768    // C99 6.7.5.3p12:
4769    //   If the function declarator is not part of a definition of that
4770    //   function, parameters may have incomplete type and may use the [*]
4771    //   notation in their sequences of declarator specifiers to specify
4772    //   variable length array types.
4773    QualType PType = Param->getOriginalType();
4774    if (const ArrayType *AT = Context.getAsArrayType(PType)) {
4775      if (AT->getSizeModifier() == ArrayType::Star) {
4776        // FIXME: This diagnosic should point the '[*]' if source-location
4777        // information is added for it.
4778        Diag(Param->getLocation(), diag::err_array_star_in_function_definition);
4779      }
4780    }
4781  }
4782
4783  return HasInvalidParm;
4784}
4785
4786/// CheckCastAlign - Implements -Wcast-align, which warns when a
4787/// pointer cast increases the alignment requirements.
4788void Sema::CheckCastAlign(Expr *Op, QualType T, SourceRange TRange) {
4789  // This is actually a lot of work to potentially be doing on every
4790  // cast; don't do it if we're ignoring -Wcast_align (as is the default).
4791  if (getDiagnostics().getDiagnosticLevel(diag::warn_cast_align,
4792                                          TRange.getBegin())
4793        == DiagnosticsEngine::Ignored)
4794    return;
4795
4796  // Ignore dependent types.
4797  if (T->isDependentType() || Op->getType()->isDependentType())
4798    return;
4799
4800  // Require that the destination be a pointer type.
4801  const PointerType *DestPtr = T->getAs<PointerType>();
4802  if (!DestPtr) return;
4803
4804  // If the destination has alignment 1, we're done.
4805  QualType DestPointee = DestPtr->getPointeeType();
4806  if (DestPointee->isIncompleteType()) return;
4807  CharUnits DestAlign = Context.getTypeAlignInChars(DestPointee);
4808  if (DestAlign.isOne()) return;
4809
4810  // Require that the source be a pointer type.
4811  const PointerType *SrcPtr = Op->getType()->getAs<PointerType>();
4812  if (!SrcPtr) return;
4813  QualType SrcPointee = SrcPtr->getPointeeType();
4814
4815  // Whitelist casts from cv void*.  We already implicitly
4816  // whitelisted casts to cv void*, since they have alignment 1.
4817  // Also whitelist casts involving incomplete types, which implicitly
4818  // includes 'void'.
4819  if (SrcPointee->isIncompleteType()) return;
4820
4821  CharUnits SrcAlign = Context.getTypeAlignInChars(SrcPointee);
4822  if (SrcAlign >= DestAlign) return;
4823
4824  Diag(TRange.getBegin(), diag::warn_cast_align)
4825    << Op->getType() << T
4826    << static_cast<unsigned>(SrcAlign.getQuantity())
4827    << static_cast<unsigned>(DestAlign.getQuantity())
4828    << TRange << Op->getSourceRange();
4829}
4830
4831static const Type* getElementType(const Expr *BaseExpr) {
4832  const Type* EltType = BaseExpr->getType().getTypePtr();
4833  if (EltType->isAnyPointerType())
4834    return EltType->getPointeeType().getTypePtr();
4835  else if (EltType->isArrayType())
4836    return EltType->getBaseElementTypeUnsafe();
4837  return EltType;
4838}
4839
4840/// \brief Check whether this array fits the idiom of a size-one tail padded
4841/// array member of a struct.
4842///
4843/// We avoid emitting out-of-bounds access warnings for such arrays as they are
4844/// commonly used to emulate flexible arrays in C89 code.
4845static bool IsTailPaddedMemberArray(Sema &S, llvm::APInt Size,
4846                                    const NamedDecl *ND) {
4847  if (Size != 1 || !ND) return false;
4848
4849  const FieldDecl *FD = dyn_cast<FieldDecl>(ND);
4850  if (!FD) return false;
4851
4852  // Don't consider sizes resulting from macro expansions or template argument
4853  // substitution to form C89 tail-padded arrays.
4854
4855  TypeSourceInfo *TInfo = FD->getTypeSourceInfo();
4856  while (TInfo) {
4857    TypeLoc TL = TInfo->getTypeLoc();
4858    // Look through typedefs.
4859    const TypedefTypeLoc *TTL = dyn_cast<TypedefTypeLoc>(&TL);
4860    if (TTL) {
4861      const TypedefNameDecl *TDL = TTL->getTypedefNameDecl();
4862      TInfo = TDL->getTypeSourceInfo();
4863      continue;
4864    }
4865    ConstantArrayTypeLoc CTL = cast<ConstantArrayTypeLoc>(TL);
4866    const Expr *SizeExpr = dyn_cast<IntegerLiteral>(CTL.getSizeExpr());
4867    if (!SizeExpr || SizeExpr->getExprLoc().isMacroID())
4868      return false;
4869    break;
4870  }
4871
4872  const RecordDecl *RD = dyn_cast<RecordDecl>(FD->getDeclContext());
4873  if (!RD) return false;
4874  if (RD->isUnion()) return false;
4875  if (const CXXRecordDecl *CRD = dyn_cast<CXXRecordDecl>(RD)) {
4876    if (!CRD->isStandardLayout()) return false;
4877  }
4878
4879  // See if this is the last field decl in the record.
4880  const Decl *D = FD;
4881  while ((D = D->getNextDeclInContext()))
4882    if (isa<FieldDecl>(D))
4883      return false;
4884  return true;
4885}
4886
4887void Sema::CheckArrayAccess(const Expr *BaseExpr, const Expr *IndexExpr,
4888                            const ArraySubscriptExpr *ASE,
4889                            bool AllowOnePastEnd, bool IndexNegated) {
4890  IndexExpr = IndexExpr->IgnoreParenImpCasts();
4891  if (IndexExpr->isValueDependent())
4892    return;
4893
4894  const Type *EffectiveType = getElementType(BaseExpr);
4895  BaseExpr = BaseExpr->IgnoreParenCasts();
4896  const ConstantArrayType *ArrayTy =
4897    Context.getAsConstantArrayType(BaseExpr->getType());
4898  if (!ArrayTy)
4899    return;
4900
4901  llvm::APSInt index;
4902  if (!IndexExpr->EvaluateAsInt(index, Context))
4903    return;
4904  if (IndexNegated)
4905    index = -index;
4906
4907  const NamedDecl *ND = NULL;
4908  if (const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(BaseExpr))
4909    ND = dyn_cast<NamedDecl>(DRE->getDecl());
4910  if (const MemberExpr *ME = dyn_cast<MemberExpr>(BaseExpr))
4911    ND = dyn_cast<NamedDecl>(ME->getMemberDecl());
4912
4913  if (index.isUnsigned() || !index.isNegative()) {
4914    llvm::APInt size = ArrayTy->getSize();
4915    if (!size.isStrictlyPositive())
4916      return;
4917
4918    const Type* BaseType = getElementType(BaseExpr);
4919    if (BaseType != EffectiveType) {
4920      // Make sure we're comparing apples to apples when comparing index to size
4921      uint64_t ptrarith_typesize = Context.getTypeSize(EffectiveType);
4922      uint64_t array_typesize = Context.getTypeSize(BaseType);
4923      // Handle ptrarith_typesize being zero, such as when casting to void*
4924      if (!ptrarith_typesize) ptrarith_typesize = 1;
4925      if (ptrarith_typesize != array_typesize) {
4926        // There's a cast to a different size type involved
4927        uint64_t ratio = array_typesize / ptrarith_typesize;
4928        // TODO: Be smarter about handling cases where array_typesize is not a
4929        // multiple of ptrarith_typesize
4930        if (ptrarith_typesize * ratio == array_typesize)
4931          size *= llvm::APInt(size.getBitWidth(), ratio);
4932      }
4933    }
4934
4935    if (size.getBitWidth() > index.getBitWidth())
4936      index = index.zext(size.getBitWidth());
4937    else if (size.getBitWidth() < index.getBitWidth())
4938      size = size.zext(index.getBitWidth());
4939
4940    // For array subscripting the index must be less than size, but for pointer
4941    // arithmetic also allow the index (offset) to be equal to size since
4942    // computing the next address after the end of the array is legal and
4943    // commonly done e.g. in C++ iterators and range-based for loops.
4944    if (AllowOnePastEnd ? index.ule(size) : index.ult(size))
4945      return;
4946
4947    // Also don't warn for arrays of size 1 which are members of some
4948    // structure. These are often used to approximate flexible arrays in C89
4949    // code.
4950    if (IsTailPaddedMemberArray(*this, size, ND))
4951      return;
4952
4953    // Suppress the warning if the subscript expression (as identified by the
4954    // ']' location) and the index expression are both from macro expansions
4955    // within a system header.
4956    if (ASE) {
4957      SourceLocation RBracketLoc = SourceMgr.getSpellingLoc(
4958          ASE->getRBracketLoc());
4959      if (SourceMgr.isInSystemHeader(RBracketLoc)) {
4960        SourceLocation IndexLoc = SourceMgr.getSpellingLoc(
4961            IndexExpr->getLocStart());
4962        if (SourceMgr.isFromSameFile(RBracketLoc, IndexLoc))
4963          return;
4964      }
4965    }
4966
4967    unsigned DiagID = diag::warn_ptr_arith_exceeds_bounds;
4968    if (ASE)
4969      DiagID = diag::warn_array_index_exceeds_bounds;
4970
4971    DiagRuntimeBehavior(BaseExpr->getLocStart(), BaseExpr,
4972                        PDiag(DiagID) << index.toString(10, true)
4973                          << size.toString(10, true)
4974                          << (unsigned)size.getLimitedValue(~0U)
4975                          << IndexExpr->getSourceRange());
4976  } else {
4977    unsigned DiagID = diag::warn_array_index_precedes_bounds;
4978    if (!ASE) {
4979      DiagID = diag::warn_ptr_arith_precedes_bounds;
4980      if (index.isNegative()) index = -index;
4981    }
4982
4983    DiagRuntimeBehavior(BaseExpr->getLocStart(), BaseExpr,
4984                        PDiag(DiagID) << index.toString(10, true)
4985                          << IndexExpr->getSourceRange());
4986  }
4987
4988  if (!ND) {
4989    // Try harder to find a NamedDecl to point at in the note.
4990    while (const ArraySubscriptExpr *ASE =
4991           dyn_cast<ArraySubscriptExpr>(BaseExpr))
4992      BaseExpr = ASE->getBase()->IgnoreParenCasts();
4993    if (const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(BaseExpr))
4994      ND = dyn_cast<NamedDecl>(DRE->getDecl());
4995    if (const MemberExpr *ME = dyn_cast<MemberExpr>(BaseExpr))
4996      ND = dyn_cast<NamedDecl>(ME->getMemberDecl());
4997  }
4998
4999  if (ND)
5000    DiagRuntimeBehavior(ND->getLocStart(), BaseExpr,
5001                        PDiag(diag::note_array_index_out_of_bounds)
5002                          << ND->getDeclName());
5003}
5004
5005void Sema::CheckArrayAccess(const Expr *expr) {
5006  int AllowOnePastEnd = 0;
5007  while (expr) {
5008    expr = expr->IgnoreParenImpCasts();
5009    switch (expr->getStmtClass()) {
5010      case Stmt::ArraySubscriptExprClass: {
5011        const ArraySubscriptExpr *ASE = cast<ArraySubscriptExpr>(expr);
5012        CheckArrayAccess(ASE->getBase(), ASE->getIdx(), ASE,
5013                         AllowOnePastEnd > 0);
5014        return;
5015      }
5016      case Stmt::UnaryOperatorClass: {
5017        // Only unwrap the * and & unary operators
5018        const UnaryOperator *UO = cast<UnaryOperator>(expr);
5019        expr = UO->getSubExpr();
5020        switch (UO->getOpcode()) {
5021          case UO_AddrOf:
5022            AllowOnePastEnd++;
5023            break;
5024          case UO_Deref:
5025            AllowOnePastEnd--;
5026            break;
5027          default:
5028            return;
5029        }
5030        break;
5031      }
5032      case Stmt::ConditionalOperatorClass: {
5033        const ConditionalOperator *cond = cast<ConditionalOperator>(expr);
5034        if (const Expr *lhs = cond->getLHS())
5035          CheckArrayAccess(lhs);
5036        if (const Expr *rhs = cond->getRHS())
5037          CheckArrayAccess(rhs);
5038        return;
5039      }
5040      default:
5041        return;
5042    }
5043  }
5044}
5045
5046//===--- CHECK: Objective-C retain cycles ----------------------------------//
5047
5048namespace {
5049  struct RetainCycleOwner {
5050    RetainCycleOwner() : Variable(0), Indirect(false) {}
5051    VarDecl *Variable;
5052    SourceRange Range;
5053    SourceLocation Loc;
5054    bool Indirect;
5055
5056    void setLocsFrom(Expr *e) {
5057      Loc = e->getExprLoc();
5058      Range = e->getSourceRange();
5059    }
5060  };
5061}
5062
5063/// Consider whether capturing the given variable can possibly lead to
5064/// a retain cycle.
5065static bool considerVariable(VarDecl *var, Expr *ref, RetainCycleOwner &owner) {
5066  // In ARC, it's captured strongly iff the variable has __strong
5067  // lifetime.  In MRR, it's captured strongly if the variable is
5068  // __block and has an appropriate type.
5069  if (var->getType().getObjCLifetime() != Qualifiers::OCL_Strong)
5070    return false;
5071
5072  owner.Variable = var;
5073  owner.setLocsFrom(ref);
5074  return true;
5075}
5076
5077static bool findRetainCycleOwner(Sema &S, Expr *e, RetainCycleOwner &owner) {
5078  while (true) {
5079    e = e->IgnoreParens();
5080    if (CastExpr *cast = dyn_cast<CastExpr>(e)) {
5081      switch (cast->getCastKind()) {
5082      case CK_BitCast:
5083      case CK_LValueBitCast:
5084      case CK_LValueToRValue:
5085      case CK_ARCReclaimReturnedObject:
5086        e = cast->getSubExpr();
5087        continue;
5088
5089      default:
5090        return false;
5091      }
5092    }
5093
5094    if (ObjCIvarRefExpr *ref = dyn_cast<ObjCIvarRefExpr>(e)) {
5095      ObjCIvarDecl *ivar = ref->getDecl();
5096      if (ivar->getType().getObjCLifetime() != Qualifiers::OCL_Strong)
5097        return false;
5098
5099      // Try to find a retain cycle in the base.
5100      if (!findRetainCycleOwner(S, ref->getBase(), owner))
5101        return false;
5102
5103      if (ref->isFreeIvar()) owner.setLocsFrom(ref);
5104      owner.Indirect = true;
5105      return true;
5106    }
5107
5108    if (DeclRefExpr *ref = dyn_cast<DeclRefExpr>(e)) {
5109      VarDecl *var = dyn_cast<VarDecl>(ref->getDecl());
5110      if (!var) return false;
5111      return considerVariable(var, ref, owner);
5112    }
5113
5114    if (MemberExpr *member = dyn_cast<MemberExpr>(e)) {
5115      if (member->isArrow()) return false;
5116
5117      // Don't count this as an indirect ownership.
5118      e = member->getBase();
5119      continue;
5120    }
5121
5122    if (PseudoObjectExpr *pseudo = dyn_cast<PseudoObjectExpr>(e)) {
5123      // Only pay attention to pseudo-objects on property references.
5124      ObjCPropertyRefExpr *pre
5125        = dyn_cast<ObjCPropertyRefExpr>(pseudo->getSyntacticForm()
5126                                              ->IgnoreParens());
5127      if (!pre) return false;
5128      if (pre->isImplicitProperty()) return false;
5129      ObjCPropertyDecl *property = pre->getExplicitProperty();
5130      if (!property->isRetaining() &&
5131          !(property->getPropertyIvarDecl() &&
5132            property->getPropertyIvarDecl()->getType()
5133              .getObjCLifetime() == Qualifiers::OCL_Strong))
5134          return false;
5135
5136      owner.Indirect = true;
5137      if (pre->isSuperReceiver()) {
5138        owner.Variable = S.getCurMethodDecl()->getSelfDecl();
5139        if (!owner.Variable)
5140          return false;
5141        owner.Loc = pre->getLocation();
5142        owner.Range = pre->getSourceRange();
5143        return true;
5144      }
5145      e = const_cast<Expr*>(cast<OpaqueValueExpr>(pre->getBase())
5146                              ->getSourceExpr());
5147      continue;
5148    }
5149
5150    // Array ivars?
5151
5152    return false;
5153  }
5154}
5155
5156namespace {
5157  struct FindCaptureVisitor : EvaluatedExprVisitor<FindCaptureVisitor> {
5158    FindCaptureVisitor(ASTContext &Context, VarDecl *variable)
5159      : EvaluatedExprVisitor<FindCaptureVisitor>(Context),
5160        Variable(variable), Capturer(0) {}
5161
5162    VarDecl *Variable;
5163    Expr *Capturer;
5164
5165    void VisitDeclRefExpr(DeclRefExpr *ref) {
5166      if (ref->getDecl() == Variable && !Capturer)
5167        Capturer = ref;
5168    }
5169
5170    void VisitObjCIvarRefExpr(ObjCIvarRefExpr *ref) {
5171      if (Capturer) return;
5172      Visit(ref->getBase());
5173      if (Capturer && ref->isFreeIvar())
5174        Capturer = ref;
5175    }
5176
5177    void VisitBlockExpr(BlockExpr *block) {
5178      // Look inside nested blocks
5179      if (block->getBlockDecl()->capturesVariable(Variable))
5180        Visit(block->getBlockDecl()->getBody());
5181    }
5182  };
5183}
5184
5185/// Check whether the given argument is a block which captures a
5186/// variable.
5187static Expr *findCapturingExpr(Sema &S, Expr *e, RetainCycleOwner &owner) {
5188  assert(owner.Variable && owner.Loc.isValid());
5189
5190  e = e->IgnoreParenCasts();
5191  BlockExpr *block = dyn_cast<BlockExpr>(e);
5192  if (!block || !block->getBlockDecl()->capturesVariable(owner.Variable))
5193    return 0;
5194
5195  FindCaptureVisitor visitor(S.Context, owner.Variable);
5196  visitor.Visit(block->getBlockDecl()->getBody());
5197  return visitor.Capturer;
5198}
5199
5200static void diagnoseRetainCycle(Sema &S, Expr *capturer,
5201                                RetainCycleOwner &owner) {
5202  assert(capturer);
5203  assert(owner.Variable && owner.Loc.isValid());
5204
5205  S.Diag(capturer->getExprLoc(), diag::warn_arc_retain_cycle)
5206    << owner.Variable << capturer->getSourceRange();
5207  S.Diag(owner.Loc, diag::note_arc_retain_cycle_owner)
5208    << owner.Indirect << owner.Range;
5209}
5210
5211/// Check for a keyword selector that starts with the word 'add' or
5212/// 'set'.
5213static bool isSetterLikeSelector(Selector sel) {
5214  if (sel.isUnarySelector()) return false;
5215
5216  StringRef str = sel.getNameForSlot(0);
5217  while (!str.empty() && str.front() == '_') str = str.substr(1);
5218  if (str.startswith("set"))
5219    str = str.substr(3);
5220  else if (str.startswith("add")) {
5221    // Specially whitelist 'addOperationWithBlock:'.
5222    if (sel.getNumArgs() == 1 && str.startswith("addOperationWithBlock"))
5223      return false;
5224    str = str.substr(3);
5225  }
5226  else
5227    return false;
5228
5229  if (str.empty()) return true;
5230  return !islower(str.front());
5231}
5232
5233/// Check a message send to see if it's likely to cause a retain cycle.
5234void Sema::checkRetainCycles(ObjCMessageExpr *msg) {
5235  // Only check instance methods whose selector looks like a setter.
5236  if (!msg->isInstanceMessage() || !isSetterLikeSelector(msg->getSelector()))
5237    return;
5238
5239  // Try to find a variable that the receiver is strongly owned by.
5240  RetainCycleOwner owner;
5241  if (msg->getReceiverKind() == ObjCMessageExpr::Instance) {
5242    if (!findRetainCycleOwner(*this, msg->getInstanceReceiver(), owner))
5243      return;
5244  } else {
5245    assert(msg->getReceiverKind() == ObjCMessageExpr::SuperInstance);
5246    owner.Variable = getCurMethodDecl()->getSelfDecl();
5247    owner.Loc = msg->getSuperLoc();
5248    owner.Range = msg->getSuperLoc();
5249  }
5250
5251  // Check whether the receiver is captured by any of the arguments.
5252  for (unsigned i = 0, e = msg->getNumArgs(); i != e; ++i)
5253    if (Expr *capturer = findCapturingExpr(*this, msg->getArg(i), owner))
5254      return diagnoseRetainCycle(*this, capturer, owner);
5255}
5256
5257/// Check a property assign to see if it's likely to cause a retain cycle.
5258void Sema::checkRetainCycles(Expr *receiver, Expr *argument) {
5259  RetainCycleOwner owner;
5260  if (!findRetainCycleOwner(*this, receiver, owner))
5261    return;
5262
5263  if (Expr *capturer = findCapturingExpr(*this, argument, owner))
5264    diagnoseRetainCycle(*this, capturer, owner);
5265}
5266
5267bool Sema::checkUnsafeAssigns(SourceLocation Loc,
5268                              QualType LHS, Expr *RHS) {
5269  Qualifiers::ObjCLifetime LT = LHS.getObjCLifetime();
5270  if (LT != Qualifiers::OCL_Weak && LT != Qualifiers::OCL_ExplicitNone)
5271    return false;
5272  // strip off any implicit cast added to get to the one arc-specific
5273  while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
5274    if (cast->getCastKind() == CK_ARCConsumeObject) {
5275      Diag(Loc, diag::warn_arc_retained_assign)
5276        << (LT == Qualifiers::OCL_ExplicitNone) << 1
5277        << RHS->getSourceRange();
5278      return true;
5279    }
5280    RHS = cast->getSubExpr();
5281  }
5282  return false;
5283}
5284
5285void Sema::checkUnsafeExprAssigns(SourceLocation Loc,
5286                              Expr *LHS, Expr *RHS) {
5287  QualType LHSType;
5288  // PropertyRef on LHS type need be directly obtained from
5289  // its declaration as it has a PsuedoType.
5290  ObjCPropertyRefExpr *PRE
5291    = dyn_cast<ObjCPropertyRefExpr>(LHS->IgnoreParens());
5292  if (PRE && !PRE->isImplicitProperty()) {
5293    const ObjCPropertyDecl *PD = PRE->getExplicitProperty();
5294    if (PD)
5295      LHSType = PD->getType();
5296  }
5297
5298  if (LHSType.isNull())
5299    LHSType = LHS->getType();
5300  if (checkUnsafeAssigns(Loc, LHSType, RHS))
5301    return;
5302  Qualifiers::ObjCLifetime LT = LHSType.getObjCLifetime();
5303  // FIXME. Check for other life times.
5304  if (LT != Qualifiers::OCL_None)
5305    return;
5306
5307  if (PRE) {
5308    if (PRE->isImplicitProperty())
5309      return;
5310    const ObjCPropertyDecl *PD = PRE->getExplicitProperty();
5311    if (!PD)
5312      return;
5313
5314    unsigned Attributes = PD->getPropertyAttributes();
5315    if (Attributes & ObjCPropertyDecl::OBJC_PR_assign) {
5316      // when 'assign' attribute was not explicitly specified
5317      // by user, ignore it and rely on property type itself
5318      // for lifetime info.
5319      unsigned AsWrittenAttr = PD->getPropertyAttributesAsWritten();
5320      if (!(AsWrittenAttr & ObjCPropertyDecl::OBJC_PR_assign) &&
5321          LHSType->isObjCRetainableType())
5322        return;
5323
5324      while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
5325        if (cast->getCastKind() == CK_ARCConsumeObject) {
5326          Diag(Loc, diag::warn_arc_retained_property_assign)
5327          << RHS->getSourceRange();
5328          return;
5329        }
5330        RHS = cast->getSubExpr();
5331      }
5332    }
5333    else if (Attributes & ObjCPropertyDecl::OBJC_PR_weak) {
5334      while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
5335        if (cast->getCastKind() == CK_ARCConsumeObject) {
5336          Diag(Loc, diag::warn_arc_retained_assign)
5337          << 0 << 0<< RHS->getSourceRange();
5338          return;
5339        }
5340        RHS = cast->getSubExpr();
5341      }
5342    }
5343  }
5344}
5345
5346//===--- CHECK: Empty statement body (-Wempty-body) ---------------------===//
5347
5348namespace {
5349bool ShouldDiagnoseEmptyStmtBody(const SourceManager &SourceMgr,
5350                                 SourceLocation StmtLoc,
5351                                 const NullStmt *Body) {
5352  // Do not warn if the body is a macro that expands to nothing, e.g:
5353  //
5354  // #define CALL(x)
5355  // if (condition)
5356  //   CALL(0);
5357  //
5358  if (Body->hasLeadingEmptyMacro())
5359    return false;
5360
5361  // Get line numbers of statement and body.
5362  bool StmtLineInvalid;
5363  unsigned StmtLine = SourceMgr.getSpellingLineNumber(StmtLoc,
5364                                                      &StmtLineInvalid);
5365  if (StmtLineInvalid)
5366    return false;
5367
5368  bool BodyLineInvalid;
5369  unsigned BodyLine = SourceMgr.getSpellingLineNumber(Body->getSemiLoc(),
5370                                                      &BodyLineInvalid);
5371  if (BodyLineInvalid)
5372    return false;
5373
5374  // Warn if null statement and body are on the same line.
5375  if (StmtLine != BodyLine)
5376    return false;
5377
5378  return true;
5379}
5380} // Unnamed namespace
5381
5382void Sema::DiagnoseEmptyStmtBody(SourceLocation StmtLoc,
5383                                 const Stmt *Body,
5384                                 unsigned DiagID) {
5385  // Since this is a syntactic check, don't emit diagnostic for template
5386  // instantiations, this just adds noise.
5387  if (CurrentInstantiationScope)
5388    return;
5389
5390  // The body should be a null statement.
5391  const NullStmt *NBody = dyn_cast<NullStmt>(Body);
5392  if (!NBody)
5393    return;
5394
5395  // Do the usual checks.
5396  if (!ShouldDiagnoseEmptyStmtBody(SourceMgr, StmtLoc, NBody))
5397    return;
5398
5399  Diag(NBody->getSemiLoc(), DiagID);
5400  Diag(NBody->getSemiLoc(), diag::note_empty_body_on_separate_line);
5401}
5402
5403void Sema::DiagnoseEmptyLoopBody(const Stmt *S,
5404                                 const Stmt *PossibleBody) {
5405  assert(!CurrentInstantiationScope); // Ensured by caller
5406
5407  SourceLocation StmtLoc;
5408  const Stmt *Body;
5409  unsigned DiagID;
5410  if (const ForStmt *FS = dyn_cast<ForStmt>(S)) {
5411    StmtLoc = FS->getRParenLoc();
5412    Body = FS->getBody();
5413    DiagID = diag::warn_empty_for_body;
5414  } else if (const WhileStmt *WS = dyn_cast<WhileStmt>(S)) {
5415    StmtLoc = WS->getCond()->getSourceRange().getEnd();
5416    Body = WS->getBody();
5417    DiagID = diag::warn_empty_while_body;
5418  } else
5419    return; // Neither `for' nor `while'.
5420
5421  // The body should be a null statement.
5422  const NullStmt *NBody = dyn_cast<NullStmt>(Body);
5423  if (!NBody)
5424    return;
5425
5426  // Skip expensive checks if diagnostic is disabled.
5427  if (Diags.getDiagnosticLevel(DiagID, NBody->getSemiLoc()) ==
5428          DiagnosticsEngine::Ignored)
5429    return;
5430
5431  // Do the usual checks.
5432  if (!ShouldDiagnoseEmptyStmtBody(SourceMgr, StmtLoc, NBody))
5433    return;
5434
5435  // `for(...);' and `while(...);' are popular idioms, so in order to keep
5436  // noise level low, emit diagnostics only if for/while is followed by a
5437  // CompoundStmt, e.g.:
5438  //    for (int i = 0; i < n; i++);
5439  //    {
5440  //      a(i);
5441  //    }
5442  // or if for/while is followed by a statement with more indentation
5443  // than for/while itself:
5444  //    for (int i = 0; i < n; i++);
5445  //      a(i);
5446  bool ProbableTypo = isa<CompoundStmt>(PossibleBody);
5447  if (!ProbableTypo) {
5448    bool BodyColInvalid;
5449    unsigned BodyCol = SourceMgr.getPresumedColumnNumber(
5450                             PossibleBody->getLocStart(),
5451                             &BodyColInvalid);
5452    if (BodyColInvalid)
5453      return;
5454
5455    bool StmtColInvalid;
5456    unsigned StmtCol = SourceMgr.getPresumedColumnNumber(
5457                             S->getLocStart(),
5458                             &StmtColInvalid);
5459    if (StmtColInvalid)
5460      return;
5461
5462    if (BodyCol > StmtCol)
5463      ProbableTypo = true;
5464  }
5465
5466  if (ProbableTypo) {
5467    Diag(NBody->getSemiLoc(), DiagID);
5468    Diag(NBody->getSemiLoc(), diag::note_empty_body_on_separate_line);
5469  }
5470}
5471