SemaChecking.cpp revision 2c0abf4ae33ab2ba690ccae724b8d6f196e7cfda 
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/ScopeInfo.h"
20#include "clang/Analysis/Analyses/FormatString.h"
21#include "clang/AST/ASTContext.h"
22#include "clang/AST/CharUnits.h"
23#include "clang/AST/DeclCXX.h"
24#include "clang/AST/DeclObjC.h"
25#include "clang/AST/ExprCXX.h"
26#include "clang/AST/ExprObjC.h"
27#include "clang/AST/EvaluatedExprVisitor.h"
28#include "clang/AST/DeclObjC.h"
29#include "clang/AST/StmtCXX.h"
30#include "clang/AST/StmtObjC.h"
31#include "clang/Lex/Preprocessor.h"
32#include "llvm/ADT/BitVector.h"
33#include "llvm/ADT/SmallString.h"
34#include "llvm/ADT/STLExtras.h"
35#include "llvm/Support/raw_ostream.h"
36#include "clang/Basic/TargetBuiltins.h"
37#include "clang/Basic/TargetInfo.h"
38#include "clang/Basic/ConvertUTF.h"
39#include <limits>
40using namespace clang;
41using namespace sema;
42
43SourceLocation Sema::getLocationOfStringLiteralByte(const StringLiteral *SL,
44                                                    unsigned ByteNo) const {
45  return SL->getLocationOfByte(ByteNo, PP.getSourceManager(),
46                               PP.getLangOpts(), PP.getTargetInfo());
47}
48
49/// Checks that a call expression's argument count is the desired number.
50/// This is useful when doing custom type-checking.  Returns true on error.
51static bool checkArgCount(Sema &S, CallExpr *call, unsigned desiredArgCount) {
52  unsigned argCount = call->getNumArgs();
53  if (argCount == desiredArgCount) return false;
54
55  if (argCount < desiredArgCount)
56    return S.Diag(call->getLocEnd(), diag::err_typecheck_call_too_few_args)
57        << 0 /*function call*/ << desiredArgCount << argCount
58        << call->getSourceRange();
59
60  // Highlight all the excess arguments.
61  SourceRange range(call->getArg(desiredArgCount)->getLocStart(),
62                    call->getArg(argCount - 1)->getLocEnd());
63
64  return S.Diag(range.getBegin(), diag::err_typecheck_call_too_many_args)
65    << 0 /*function call*/ << desiredArgCount << argCount
66    << call->getArg(1)->getSourceRange();
67}
68
69/// Check that the first argument to __builtin_annotation is an integer
70/// and the second argument is a non-wide string literal.
71static bool SemaBuiltinAnnotation(Sema &S, CallExpr *TheCall) {
72  if (checkArgCount(S, TheCall, 2))
73    return true;
74
75  // First argument should be an integer.
76  Expr *ValArg = TheCall->getArg(0);
77  QualType Ty = ValArg->getType();
78  if (!Ty->isIntegerType()) {
79    S.Diag(ValArg->getLocStart(), diag::err_builtin_annotation_first_arg)
80      << ValArg->getSourceRange();
81    return true;
82  }
83
84  // Second argument should be a constant string.
85  Expr *StrArg = TheCall->getArg(1)->IgnoreParenCasts();
86  StringLiteral *Literal = dyn_cast<StringLiteral>(StrArg);
87  if (!Literal || !Literal->isAscii()) {
88    S.Diag(StrArg->getLocStart(), diag::err_builtin_annotation_second_arg)
89      << StrArg->getSourceRange();
90    return true;
91  }
92
93  TheCall->setType(Ty);
94  return false;
95}
96
97ExprResult
98Sema::CheckBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
99  ExprResult TheCallResult(Owned(TheCall));
100
101  // Find out if any arguments are required to be integer constant expressions.
102  unsigned ICEArguments = 0;
103  ASTContext::GetBuiltinTypeError Error;
104  Context.GetBuiltinType(BuiltinID, Error, &ICEArguments);
105  if (Error != ASTContext::GE_None)
106    ICEArguments = 0;  // Don't diagnose previously diagnosed errors.
107
108  // If any arguments are required to be ICE's, check and diagnose.
109  for (unsigned ArgNo = 0; ICEArguments != 0; ++ArgNo) {
110    // Skip arguments not required to be ICE's.
111    if ((ICEArguments & (1 << ArgNo)) == 0) continue;
112
113    llvm::APSInt Result;
114    if (SemaBuiltinConstantArg(TheCall, ArgNo, Result))
115      return true;
116    ICEArguments &= ~(1 << ArgNo);
117  }
118
119  switch (BuiltinID) {
120  case Builtin::BI__builtin___CFStringMakeConstantString:
121    assert(TheCall->getNumArgs() == 1 &&
122           "Wrong # arguments to builtin CFStringMakeConstantString");
123    if (CheckObjCString(TheCall->getArg(0)))
124      return ExprError();
125    break;
126  case Builtin::BI__builtin_stdarg_start:
127  case Builtin::BI__builtin_va_start:
128    if (SemaBuiltinVAStart(TheCall))
129      return ExprError();
130    break;
131  case Builtin::BI__builtin_isgreater:
132  case Builtin::BI__builtin_isgreaterequal:
133  case Builtin::BI__builtin_isless:
134  case Builtin::BI__builtin_islessequal:
135  case Builtin::BI__builtin_islessgreater:
136  case Builtin::BI__builtin_isunordered:
137    if (SemaBuiltinUnorderedCompare(TheCall))
138      return ExprError();
139    break;
140  case Builtin::BI__builtin_fpclassify:
141    if (SemaBuiltinFPClassification(TheCall, 6))
142      return ExprError();
143    break;
144  case Builtin::BI__builtin_isfinite:
145  case Builtin::BI__builtin_isinf:
146  case Builtin::BI__builtin_isinf_sign:
147  case Builtin::BI__builtin_isnan:
148  case Builtin::BI__builtin_isnormal:
149    if (SemaBuiltinFPClassification(TheCall, 1))
150      return ExprError();
151    break;
152  case Builtin::BI__builtin_shufflevector:
153    return SemaBuiltinShuffleVector(TheCall);
154    // TheCall will be freed by the smart pointer here, but that's fine, since
155    // SemaBuiltinShuffleVector guts it, but then doesn't release it.
156  case Builtin::BI__builtin_prefetch:
157    if (SemaBuiltinPrefetch(TheCall))
158      return ExprError();
159    break;
160  case Builtin::BI__builtin_object_size:
161    if (SemaBuiltinObjectSize(TheCall))
162      return ExprError();
163    break;
164  case Builtin::BI__builtin_longjmp:
165    if (SemaBuiltinLongjmp(TheCall))
166      return ExprError();
167    break;
168
169  case Builtin::BI__builtin_classify_type:
170    if (checkArgCount(*this, TheCall, 1)) return true;
171    TheCall->setType(Context.IntTy);
172    break;
173  case Builtin::BI__builtin_constant_p:
174    if (checkArgCount(*this, TheCall, 1)) return true;
175    TheCall->setType(Context.IntTy);
176    break;
177  case Builtin::BI__sync_fetch_and_add:
178  case Builtin::BI__sync_fetch_and_add_1:
179  case Builtin::BI__sync_fetch_and_add_2:
180  case Builtin::BI__sync_fetch_and_add_4:
181  case Builtin::BI__sync_fetch_and_add_8:
182  case Builtin::BI__sync_fetch_and_add_16:
183  case Builtin::BI__sync_fetch_and_sub:
184  case Builtin::BI__sync_fetch_and_sub_1:
185  case Builtin::BI__sync_fetch_and_sub_2:
186  case Builtin::BI__sync_fetch_and_sub_4:
187  case Builtin::BI__sync_fetch_and_sub_8:
188  case Builtin::BI__sync_fetch_and_sub_16:
189  case Builtin::BI__sync_fetch_and_or:
190  case Builtin::BI__sync_fetch_and_or_1:
191  case Builtin::BI__sync_fetch_and_or_2:
192  case Builtin::BI__sync_fetch_and_or_4:
193  case Builtin::BI__sync_fetch_and_or_8:
194  case Builtin::BI__sync_fetch_and_or_16:
195  case Builtin::BI__sync_fetch_and_and:
196  case Builtin::BI__sync_fetch_and_and_1:
197  case Builtin::BI__sync_fetch_and_and_2:
198  case Builtin::BI__sync_fetch_and_and_4:
199  case Builtin::BI__sync_fetch_and_and_8:
200  case Builtin::BI__sync_fetch_and_and_16:
201  case Builtin::BI__sync_fetch_and_xor:
202  case Builtin::BI__sync_fetch_and_xor_1:
203  case Builtin::BI__sync_fetch_and_xor_2:
204  case Builtin::BI__sync_fetch_and_xor_4:
205  case Builtin::BI__sync_fetch_and_xor_8:
206  case Builtin::BI__sync_fetch_and_xor_16:
207  case Builtin::BI__sync_add_and_fetch:
208  case Builtin::BI__sync_add_and_fetch_1:
209  case Builtin::BI__sync_add_and_fetch_2:
210  case Builtin::BI__sync_add_and_fetch_4:
211  case Builtin::BI__sync_add_and_fetch_8:
212  case Builtin::BI__sync_add_and_fetch_16:
213  case Builtin::BI__sync_sub_and_fetch:
214  case Builtin::BI__sync_sub_and_fetch_1:
215  case Builtin::BI__sync_sub_and_fetch_2:
216  case Builtin::BI__sync_sub_and_fetch_4:
217  case Builtin::BI__sync_sub_and_fetch_8:
218  case Builtin::BI__sync_sub_and_fetch_16:
219  case Builtin::BI__sync_and_and_fetch:
220  case Builtin::BI__sync_and_and_fetch_1:
221  case Builtin::BI__sync_and_and_fetch_2:
222  case Builtin::BI__sync_and_and_fetch_4:
223  case Builtin::BI__sync_and_and_fetch_8:
224  case Builtin::BI__sync_and_and_fetch_16:
225  case Builtin::BI__sync_or_and_fetch:
226  case Builtin::BI__sync_or_and_fetch_1:
227  case Builtin::BI__sync_or_and_fetch_2:
228  case Builtin::BI__sync_or_and_fetch_4:
229  case Builtin::BI__sync_or_and_fetch_8:
230  case Builtin::BI__sync_or_and_fetch_16:
231  case Builtin::BI__sync_xor_and_fetch:
232  case Builtin::BI__sync_xor_and_fetch_1:
233  case Builtin::BI__sync_xor_and_fetch_2:
234  case Builtin::BI__sync_xor_and_fetch_4:
235  case Builtin::BI__sync_xor_and_fetch_8:
236  case Builtin::BI__sync_xor_and_fetch_16:
237  case Builtin::BI__sync_val_compare_and_swap:
238  case Builtin::BI__sync_val_compare_and_swap_1:
239  case Builtin::BI__sync_val_compare_and_swap_2:
240  case Builtin::BI__sync_val_compare_and_swap_4:
241  case Builtin::BI__sync_val_compare_and_swap_8:
242  case Builtin::BI__sync_val_compare_and_swap_16:
243  case Builtin::BI__sync_bool_compare_and_swap:
244  case Builtin::BI__sync_bool_compare_and_swap_1:
245  case Builtin::BI__sync_bool_compare_and_swap_2:
246  case Builtin::BI__sync_bool_compare_and_swap_4:
247  case Builtin::BI__sync_bool_compare_and_swap_8:
248  case Builtin::BI__sync_bool_compare_and_swap_16:
249  case Builtin::BI__sync_lock_test_and_set:
250  case Builtin::BI__sync_lock_test_and_set_1:
251  case Builtin::BI__sync_lock_test_and_set_2:
252  case Builtin::BI__sync_lock_test_and_set_4:
253  case Builtin::BI__sync_lock_test_and_set_8:
254  case Builtin::BI__sync_lock_test_and_set_16:
255  case Builtin::BI__sync_lock_release:
256  case Builtin::BI__sync_lock_release_1:
257  case Builtin::BI__sync_lock_release_2:
258  case Builtin::BI__sync_lock_release_4:
259  case Builtin::BI__sync_lock_release_8:
260  case Builtin::BI__sync_lock_release_16:
261  case Builtin::BI__sync_swap:
262  case Builtin::BI__sync_swap_1:
263  case Builtin::BI__sync_swap_2:
264  case Builtin::BI__sync_swap_4:
265  case Builtin::BI__sync_swap_8:
266  case Builtin::BI__sync_swap_16:
267    return SemaBuiltinAtomicOverloaded(move(TheCallResult));
268#define BUILTIN(ID, TYPE, ATTRS)
269#define ATOMIC_BUILTIN(ID, TYPE, ATTRS) \
270  case Builtin::BI##ID: \
271    return SemaAtomicOpsOverloaded(move(TheCallResult), AtomicExpr::AO##ID);
272#include "clang/Basic/Builtins.def"
273  case Builtin::BI__builtin_annotation:
274    if (SemaBuiltinAnnotation(*this, TheCall))
275      return ExprError();
276    break;
277  }
278
279  // Since the target specific builtins for each arch overlap, only check those
280  // of the arch we are compiling for.
281  if (BuiltinID >= Builtin::FirstTSBuiltin) {
282    switch (Context.getTargetInfo().getTriple().getArch()) {
283      case llvm::Triple::arm:
284      case llvm::Triple::thumb:
285        if (CheckARMBuiltinFunctionCall(BuiltinID, TheCall))
286          return ExprError();
287        break;
288      default:
289        break;
290    }
291  }
292
293  return move(TheCallResult);
294}
295
296// Get the valid immediate range for the specified NEON type code.
297static unsigned RFT(unsigned t, bool shift = false) {
298  NeonTypeFlags Type(t);
299  int IsQuad = Type.isQuad();
300  switch (Type.getEltType()) {
301  case NeonTypeFlags::Int8:
302  case NeonTypeFlags::Poly8:
303    return shift ? 7 : (8 << IsQuad) - 1;
304  case NeonTypeFlags::Int16:
305  case NeonTypeFlags::Poly16:
306    return shift ? 15 : (4 << IsQuad) - 1;
307  case NeonTypeFlags::Int32:
308    return shift ? 31 : (2 << IsQuad) - 1;
309  case NeonTypeFlags::Int64:
310    return shift ? 63 : (1 << IsQuad) - 1;
311  case NeonTypeFlags::Float16:
312    assert(!shift && "cannot shift float types!");
313    return (4 << IsQuad) - 1;
314  case NeonTypeFlags::Float32:
315    assert(!shift && "cannot shift float types!");
316    return (2 << IsQuad) - 1;
317  }
318  llvm_unreachable("Invalid NeonTypeFlag!");
319}
320
321/// getNeonEltType - Return the QualType corresponding to the elements of
322/// the vector type specified by the NeonTypeFlags.  This is used to check
323/// the pointer arguments for Neon load/store intrinsics.
324static QualType getNeonEltType(NeonTypeFlags Flags, ASTContext &Context) {
325  switch (Flags.getEltType()) {
326  case NeonTypeFlags::Int8:
327    return Flags.isUnsigned() ? Context.UnsignedCharTy : Context.SignedCharTy;
328  case NeonTypeFlags::Int16:
329    return Flags.isUnsigned() ? Context.UnsignedShortTy : Context.ShortTy;
330  case NeonTypeFlags::Int32:
331    return Flags.isUnsigned() ? Context.UnsignedIntTy : Context.IntTy;
332  case NeonTypeFlags::Int64:
333    return Flags.isUnsigned() ? Context.UnsignedLongLongTy : Context.LongLongTy;
334  case NeonTypeFlags::Poly8:
335    return Context.SignedCharTy;
336  case NeonTypeFlags::Poly16:
337    return Context.ShortTy;
338  case NeonTypeFlags::Float16:
339    return Context.UnsignedShortTy;
340  case NeonTypeFlags::Float32:
341    return Context.FloatTy;
342  }
343  llvm_unreachable("Invalid NeonTypeFlag!");
344}
345
346bool Sema::CheckARMBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
347  llvm::APSInt Result;
348
349  unsigned mask = 0;
350  unsigned TV = 0;
351  int PtrArgNum = -1;
352  bool HasConstPtr = false;
353  switch (BuiltinID) {
354#define GET_NEON_OVERLOAD_CHECK
355#include "clang/Basic/arm_neon.inc"
356#undef GET_NEON_OVERLOAD_CHECK
357  }
358
359  // For NEON intrinsics which are overloaded on vector element type, validate
360  // the immediate which specifies which variant to emit.
361  unsigned ImmArg = TheCall->getNumArgs()-1;
362  if (mask) {
363    if (SemaBuiltinConstantArg(TheCall, ImmArg, Result))
364      return true;
365
366    TV = Result.getLimitedValue(64);
367    if ((TV > 63) || (mask & (1 << TV)) == 0)
368      return Diag(TheCall->getLocStart(), diag::err_invalid_neon_type_code)
369        << TheCall->getArg(ImmArg)->getSourceRange();
370  }
371
372  if (PtrArgNum >= 0) {
373    // Check that pointer arguments have the specified type.
374    Expr *Arg = TheCall->getArg(PtrArgNum);
375    if (ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(Arg))
376      Arg = ICE->getSubExpr();
377    ExprResult RHS = DefaultFunctionArrayLvalueConversion(Arg);
378    QualType RHSTy = RHS.get()->getType();
379    QualType EltTy = getNeonEltType(NeonTypeFlags(TV), Context);
380    if (HasConstPtr)
381      EltTy = EltTy.withConst();
382    QualType LHSTy = Context.getPointerType(EltTy);
383    AssignConvertType ConvTy;
384    ConvTy = CheckSingleAssignmentConstraints(LHSTy, RHS);
385    if (RHS.isInvalid())
386      return true;
387    if (DiagnoseAssignmentResult(ConvTy, Arg->getLocStart(), LHSTy, RHSTy,
388                                 RHS.get(), AA_Assigning))
389      return true;
390  }
391
392  // For NEON intrinsics which take an immediate value as part of the
393  // instruction, range check them here.
394  unsigned i = 0, l = 0, u = 0;
395  switch (BuiltinID) {
396  default: return false;
397  case ARM::BI__builtin_arm_ssat: i = 1; l = 1; u = 31; break;
398  case ARM::BI__builtin_arm_usat: i = 1; u = 31; break;
399  case ARM::BI__builtin_arm_vcvtr_f:
400  case ARM::BI__builtin_arm_vcvtr_d: i = 1; u = 1; break;
401#define GET_NEON_IMMEDIATE_CHECK
402#include "clang/Basic/arm_neon.inc"
403#undef GET_NEON_IMMEDIATE_CHECK
404  };
405
406  // Check that the immediate argument is actually a constant.
407  if (SemaBuiltinConstantArg(TheCall, i, Result))
408    return true;
409
410  // Range check against the upper/lower values for this isntruction.
411  unsigned Val = Result.getZExtValue();
412  if (Val < l || Val > (u + l))
413    return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
414      << l << u+l << TheCall->getArg(i)->getSourceRange();
415
416  // FIXME: VFP Intrinsics should error if VFP not present.
417  return false;
418}
419
420/// CheckFunctionCall - Check a direct function call for various correctness
421/// and safety properties not strictly enforced by the C type system.
422bool Sema::CheckFunctionCall(FunctionDecl *FDecl, CallExpr *TheCall) {
423  // Get the IdentifierInfo* for the called function.
424  IdentifierInfo *FnInfo = FDecl->getIdentifier();
425
426  // None of the checks below are needed for functions that don't have
427  // simple names (e.g., C++ conversion functions).
428  if (!FnInfo)
429    return false;
430
431  // FIXME: This mechanism should be abstracted to be less fragile and
432  // more efficient. For example, just map function ids to custom
433  // handlers.
434
435  // Printf and scanf checking.
436  for (specific_attr_iterator<FormatAttr>
437         i = FDecl->specific_attr_begin<FormatAttr>(),
438         e = FDecl->specific_attr_end<FormatAttr>(); i != e ; ++i) {
439    CheckFormatArguments(*i, TheCall);
440  }
441
442  for (specific_attr_iterator<NonNullAttr>
443         i = FDecl->specific_attr_begin<NonNullAttr>(),
444         e = FDecl->specific_attr_end<NonNullAttr>(); i != e; ++i) {
445    CheckNonNullArguments(*i, TheCall->getArgs(),
446                          TheCall->getCallee()->getLocStart());
447  }
448
449  unsigned CMId = FDecl->getMemoryFunctionKind();
450  if (CMId == 0)
451    return false;
452
453  // Handle memory setting and copying functions.
454  if (CMId == Builtin::BIstrlcpy || CMId == Builtin::BIstrlcat)
455    CheckStrlcpycatArguments(TheCall, FnInfo);
456  else if (CMId == Builtin::BIstrncat)
457    CheckStrncatArguments(TheCall, FnInfo);
458  else
459    CheckMemaccessArguments(TheCall, CMId, FnInfo);
460
461  return false;
462}
463
464bool Sema::CheckObjCMethodCall(ObjCMethodDecl *Method, SourceLocation lbrac,
465                               Expr **Args, unsigned NumArgs) {
466  for (specific_attr_iterator<FormatAttr>
467       i = Method->specific_attr_begin<FormatAttr>(),
468       e = Method->specific_attr_end<FormatAttr>(); i != e ; ++i) {
469
470    CheckFormatArguments(*i, Args, NumArgs, false, lbrac,
471                         Method->getSourceRange());
472  }
473
474  // diagnose nonnull arguments.
475  for (specific_attr_iterator<NonNullAttr>
476       i = Method->specific_attr_begin<NonNullAttr>(),
477       e = Method->specific_attr_end<NonNullAttr>(); i != e; ++i) {
478    CheckNonNullArguments(*i, Args, lbrac);
479  }
480
481  return false;
482}
483
484bool Sema::CheckBlockCall(NamedDecl *NDecl, CallExpr *TheCall) {
485  const VarDecl *V = dyn_cast<VarDecl>(NDecl);
486  if (!V)
487    return false;
488
489  QualType Ty = V->getType();
490  if (!Ty->isBlockPointerType())
491    return false;
492
493  // format string checking.
494  for (specific_attr_iterator<FormatAttr>
495       i = NDecl->specific_attr_begin<FormatAttr>(),
496       e = NDecl->specific_attr_end<FormatAttr>(); i != e ; ++i) {
497    CheckFormatArguments(*i, TheCall);
498  }
499
500  return false;
501}
502
503ExprResult Sema::SemaAtomicOpsOverloaded(ExprResult TheCallResult,
504                                         AtomicExpr::AtomicOp Op) {
505  CallExpr *TheCall = cast<CallExpr>(TheCallResult.get());
506  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
507
508  // All these operations take one of the following forms:
509  enum {
510    // C    __c11_atomic_init(A *, C)
511    Init,
512    // C    __c11_atomic_load(A *, int)
513    Load,
514    // void __atomic_load(A *, CP, int)
515    Copy,
516    // C    __c11_atomic_add(A *, M, int)
517    Arithmetic,
518    // C    __atomic_exchange_n(A *, CP, int)
519    Xchg,
520    // void __atomic_exchange(A *, C *, CP, int)
521    GNUXchg,
522    // bool __c11_atomic_compare_exchange_strong(A *, C *, CP, int, int)
523    C11CmpXchg,
524    // bool __atomic_compare_exchange(A *, C *, CP, bool, int, int)
525    GNUCmpXchg
526  } Form = Init;
527  const unsigned NumArgs[] = { 2, 2, 3, 3, 3, 4, 5, 6 };
528  const unsigned NumVals[] = { 1, 0, 1, 1, 1, 2, 2, 3 };
529  // where:
530  //   C is an appropriate type,
531  //   A is volatile _Atomic(C) for __c11 builtins and is C for GNU builtins,
532  //   CP is C for __c11 builtins and GNU _n builtins and is C * otherwise,
533  //   M is C if C is an integer, and ptrdiff_t if C is a pointer, and
534  //   the int parameters are for orderings.
535
536  assert(AtomicExpr::AO__c11_atomic_init == 0 &&
537         AtomicExpr::AO__c11_atomic_fetch_xor + 1 == AtomicExpr::AO__atomic_load
538         && "need to update code for modified C11 atomics");
539  bool IsC11 = Op >= AtomicExpr::AO__c11_atomic_init &&
540               Op <= AtomicExpr::AO__c11_atomic_fetch_xor;
541  bool IsN = Op == AtomicExpr::AO__atomic_load_n ||
542             Op == AtomicExpr::AO__atomic_store_n ||
543             Op == AtomicExpr::AO__atomic_exchange_n ||
544             Op == AtomicExpr::AO__atomic_compare_exchange_n;
545  bool IsAddSub = false;
546
547  switch (Op) {
548  case AtomicExpr::AO__c11_atomic_init:
549    Form = Init;
550    break;
551
552  case AtomicExpr::AO__c11_atomic_load:
553  case AtomicExpr::AO__atomic_load_n:
554    Form = Load;
555    break;
556
557  case AtomicExpr::AO__c11_atomic_store:
558  case AtomicExpr::AO__atomic_load:
559  case AtomicExpr::AO__atomic_store:
560  case AtomicExpr::AO__atomic_store_n:
561    Form = Copy;
562    break;
563
564  case AtomicExpr::AO__c11_atomic_fetch_add:
565  case AtomicExpr::AO__c11_atomic_fetch_sub:
566  case AtomicExpr::AO__atomic_fetch_add:
567  case AtomicExpr::AO__atomic_fetch_sub:
568  case AtomicExpr::AO__atomic_add_fetch:
569  case AtomicExpr::AO__atomic_sub_fetch:
570    IsAddSub = true;
571    // Fall through.
572  case AtomicExpr::AO__c11_atomic_fetch_and:
573  case AtomicExpr::AO__c11_atomic_fetch_or:
574  case AtomicExpr::AO__c11_atomic_fetch_xor:
575  case AtomicExpr::AO__atomic_fetch_and:
576  case AtomicExpr::AO__atomic_fetch_or:
577  case AtomicExpr::AO__atomic_fetch_xor:
578  case AtomicExpr::AO__atomic_fetch_nand:
579  case AtomicExpr::AO__atomic_and_fetch:
580  case AtomicExpr::AO__atomic_or_fetch:
581  case AtomicExpr::AO__atomic_xor_fetch:
582  case AtomicExpr::AO__atomic_nand_fetch:
583    Form = Arithmetic;
584    break;
585
586  case AtomicExpr::AO__c11_atomic_exchange:
587  case AtomicExpr::AO__atomic_exchange_n:
588    Form = Xchg;
589    break;
590
591  case AtomicExpr::AO__atomic_exchange:
592    Form = GNUXchg;
593    break;
594
595  case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
596  case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
597    Form = C11CmpXchg;
598    break;
599
600  case AtomicExpr::AO__atomic_compare_exchange:
601  case AtomicExpr::AO__atomic_compare_exchange_n:
602    Form = GNUCmpXchg;
603    break;
604  }
605
606  // Check we have the right number of arguments.
607  if (TheCall->getNumArgs() < NumArgs[Form]) {
608    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
609      << 0 << NumArgs[Form] << TheCall->getNumArgs()
610      << TheCall->getCallee()->getSourceRange();
611    return ExprError();
612  } else if (TheCall->getNumArgs() > NumArgs[Form]) {
613    Diag(TheCall->getArg(NumArgs[Form])->getLocStart(),
614         diag::err_typecheck_call_too_many_args)
615      << 0 << NumArgs[Form] << TheCall->getNumArgs()
616      << TheCall->getCallee()->getSourceRange();
617    return ExprError();
618  }
619
620  // Inspect the first argument of the atomic operation.
621  Expr *Ptr = TheCall->getArg(0);
622  Ptr = DefaultFunctionArrayLvalueConversion(Ptr).get();
623  const PointerType *pointerType = Ptr->getType()->getAs<PointerType>();
624  if (!pointerType) {
625    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer)
626      << Ptr->getType() << Ptr->getSourceRange();
627    return ExprError();
628  }
629
630  // For a __c11 builtin, this should be a pointer to an _Atomic type.
631  QualType AtomTy = pointerType->getPointeeType(); // 'A'
632  QualType ValType = AtomTy; // 'C'
633  if (IsC11) {
634    if (!AtomTy->isAtomicType()) {
635      Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic)
636        << Ptr->getType() << Ptr->getSourceRange();
637      return ExprError();
638    }
639    ValType = AtomTy->getAs<AtomicType>()->getValueType();
640  }
641
642  // For an arithmetic operation, the implied arithmetic must be well-formed.
643  if (Form == Arithmetic) {
644    // gcc does not enforce these rules for GNU atomics, but we do so for sanity.
645    if (IsAddSub && !ValType->isIntegerType() && !ValType->isPointerType()) {
646      Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic_int_or_ptr)
647        << IsC11 << Ptr->getType() << Ptr->getSourceRange();
648      return ExprError();
649    }
650    if (!IsAddSub && !ValType->isIntegerType()) {
651      Diag(DRE->getLocStart(), diag::err_atomic_op_bitwise_needs_atomic_int)
652        << IsC11 << Ptr->getType() << Ptr->getSourceRange();
653      return ExprError();
654    }
655  } else if (IsN && !ValType->isIntegerType() && !ValType->isPointerType()) {
656    // For __atomic_*_n operations, the value type must be a scalar integral or
657    // pointer type which is 1, 2, 4, 8 or 16 bytes in length.
658    Diag(DRE->getLocStart(), diag::err_atomic_op_needs_atomic_int_or_ptr)
659      << IsC11 << Ptr->getType() << Ptr->getSourceRange();
660    return ExprError();
661  }
662
663  if (!IsC11 && !AtomTy.isTriviallyCopyableType(Context)) {
664    // For GNU atomics, require a trivially-copyable type. This is not part of
665    // the GNU atomics specification, but we enforce it for sanity.
666    Diag(DRE->getLocStart(), diag::err_atomic_op_needs_trivial_copy)
667      << Ptr->getType() << Ptr->getSourceRange();
668    return ExprError();
669  }
670
671  // FIXME: For any builtin other than a load, the ValType must not be
672  // const-qualified.
673
674  switch (ValType.getObjCLifetime()) {
675  case Qualifiers::OCL_None:
676  case Qualifiers::OCL_ExplicitNone:
677    // okay
678    break;
679
680  case Qualifiers::OCL_Weak:
681  case Qualifiers::OCL_Strong:
682  case Qualifiers::OCL_Autoreleasing:
683    // FIXME: Can this happen? By this point, ValType should be known
684    // to be trivially copyable.
685    Diag(DRE->getLocStart(), diag::err_arc_atomic_ownership)
686      << ValType << Ptr->getSourceRange();
687    return ExprError();
688  }
689
690  QualType ResultType = ValType;
691  if (Form == Copy || Form == GNUXchg || Form == Init)
692    ResultType = Context.VoidTy;
693  else if (Form == C11CmpXchg || Form == GNUCmpXchg)
694    ResultType = Context.BoolTy;
695
696  // The type of a parameter passed 'by value'. In the GNU atomics, such
697  // arguments are actually passed as pointers.
698  QualType ByValType = ValType; // 'CP'
699  if (!IsC11 && !IsN)
700    ByValType = Ptr->getType();
701
702  // The first argument --- the pointer --- has a fixed type; we
703  // deduce the types of the rest of the arguments accordingly.  Walk
704  // the remaining arguments, converting them to the deduced value type.
705  for (unsigned i = 1; i != NumArgs[Form]; ++i) {
706    QualType Ty;
707    if (i < NumVals[Form] + 1) {
708      switch (i) {
709      case 1:
710        // The second argument is the non-atomic operand. For arithmetic, this
711        // is always passed by value, and for a compare_exchange it is always
712        // passed by address. For the rest, GNU uses by-address and C11 uses
713        // by-value.
714        assert(Form != Load);
715        if (Form == Init || (Form == Arithmetic && ValType->isIntegerType()))
716          Ty = ValType;
717        else if (Form == Copy || Form == Xchg)
718          Ty = ByValType;
719        else if (Form == Arithmetic)
720          Ty = Context.getPointerDiffType();
721        else
722          Ty = Context.getPointerType(ValType.getUnqualifiedType());
723        break;
724      case 2:
725        // The third argument to compare_exchange / GNU exchange is a
726        // (pointer to a) desired value.
727        Ty = ByValType;
728        break;
729      case 3:
730        // The fourth argument to GNU compare_exchange is a 'weak' flag.
731        Ty = Context.BoolTy;
732        break;
733      }
734    } else {
735      // The order(s) are always converted to int.
736      Ty = Context.IntTy;
737    }
738
739    InitializedEntity Entity =
740        InitializedEntity::InitializeParameter(Context, Ty, false);
741    ExprResult Arg = TheCall->getArg(i);
742    Arg = PerformCopyInitialization(Entity, SourceLocation(), Arg);
743    if (Arg.isInvalid())
744      return true;
745    TheCall->setArg(i, Arg.get());
746  }
747
748  // Permute the arguments into a 'consistent' order.
749  SmallVector<Expr*, 5> SubExprs;
750  SubExprs.push_back(Ptr);
751  switch (Form) {
752  case Init:
753    // Note, AtomicExpr::getVal1() has a special case for this atomic.
754    SubExprs.push_back(TheCall->getArg(1)); // Val1
755    break;
756  case Load:
757    SubExprs.push_back(TheCall->getArg(1)); // Order
758    break;
759  case Copy:
760  case Arithmetic:
761  case Xchg:
762    SubExprs.push_back(TheCall->getArg(2)); // Order
763    SubExprs.push_back(TheCall->getArg(1)); // Val1
764    break;
765  case GNUXchg:
766    // Note, AtomicExpr::getVal2() has a special case for this atomic.
767    SubExprs.push_back(TheCall->getArg(3)); // Order
768    SubExprs.push_back(TheCall->getArg(1)); // Val1
769    SubExprs.push_back(TheCall->getArg(2)); // Val2
770    break;
771  case C11CmpXchg:
772    SubExprs.push_back(TheCall->getArg(3)); // Order
773    SubExprs.push_back(TheCall->getArg(1)); // Val1
774    SubExprs.push_back(TheCall->getArg(4)); // OrderFail
775    SubExprs.push_back(TheCall->getArg(2)); // Val2
776    break;
777  case GNUCmpXchg:
778    SubExprs.push_back(TheCall->getArg(4)); // Order
779    SubExprs.push_back(TheCall->getArg(1)); // Val1
780    SubExprs.push_back(TheCall->getArg(5)); // OrderFail
781    SubExprs.push_back(TheCall->getArg(2)); // Val2
782    SubExprs.push_back(TheCall->getArg(3)); // Weak
783    break;
784  }
785
786  return Owned(new (Context) AtomicExpr(TheCall->getCallee()->getLocStart(),
787                                        SubExprs.data(), SubExprs.size(),
788                                        ResultType, Op,
789                                        TheCall->getRParenLoc()));
790}
791
792
793/// checkBuiltinArgument - Given a call to a builtin function, perform
794/// normal type-checking on the given argument, updating the call in
795/// place.  This is useful when a builtin function requires custom
796/// type-checking for some of its arguments but not necessarily all of
797/// them.
798///
799/// Returns true on error.
800static bool checkBuiltinArgument(Sema &S, CallExpr *E, unsigned ArgIndex) {
801  FunctionDecl *Fn = E->getDirectCallee();
802  assert(Fn && "builtin call without direct callee!");
803
804  ParmVarDecl *Param = Fn->getParamDecl(ArgIndex);
805  InitializedEntity Entity =
806    InitializedEntity::InitializeParameter(S.Context, Param);
807
808  ExprResult Arg = E->getArg(0);
809  Arg = S.PerformCopyInitialization(Entity, SourceLocation(), Arg);
810  if (Arg.isInvalid())
811    return true;
812
813  E->setArg(ArgIndex, Arg.take());
814  return false;
815}
816
817/// SemaBuiltinAtomicOverloaded - We have a call to a function like
818/// __sync_fetch_and_add, which is an overloaded function based on the pointer
819/// type of its first argument.  The main ActOnCallExpr routines have already
820/// promoted the types of arguments because all of these calls are prototyped as
821/// void(...).
822///
823/// This function goes through and does final semantic checking for these
824/// builtins,
825ExprResult
826Sema::SemaBuiltinAtomicOverloaded(ExprResult TheCallResult) {
827  CallExpr *TheCall = (CallExpr *)TheCallResult.get();
828  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
829  FunctionDecl *FDecl = cast<FunctionDecl>(DRE->getDecl());
830
831  // Ensure that we have at least one argument to do type inference from.
832  if (TheCall->getNumArgs() < 1) {
833    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args_at_least)
834      << 0 << 1 << TheCall->getNumArgs()
835      << TheCall->getCallee()->getSourceRange();
836    return ExprError();
837  }
838
839  // Inspect the first argument of the atomic builtin.  This should always be
840  // a pointer type, whose element is an integral scalar or pointer type.
841  // Because it is a pointer type, we don't have to worry about any implicit
842  // casts here.
843  // FIXME: We don't allow floating point scalars as input.
844  Expr *FirstArg = TheCall->getArg(0);
845  ExprResult FirstArgResult = DefaultFunctionArrayLvalueConversion(FirstArg);
846  if (FirstArgResult.isInvalid())
847    return ExprError();
848  FirstArg = FirstArgResult.take();
849  TheCall->setArg(0, FirstArg);
850
851  const PointerType *pointerType = FirstArg->getType()->getAs<PointerType>();
852  if (!pointerType) {
853    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer)
854      << FirstArg->getType() << FirstArg->getSourceRange();
855    return ExprError();
856  }
857
858  QualType ValType = pointerType->getPointeeType();
859  if (!ValType->isIntegerType() && !ValType->isAnyPointerType() &&
860      !ValType->isBlockPointerType()) {
861    Diag(DRE->getLocStart(), diag::err_atomic_builtin_must_be_pointer_intptr)
862      << FirstArg->getType() << FirstArg->getSourceRange();
863    return ExprError();
864  }
865
866  switch (ValType.getObjCLifetime()) {
867  case Qualifiers::OCL_None:
868  case Qualifiers::OCL_ExplicitNone:
869    // okay
870    break;
871
872  case Qualifiers::OCL_Weak:
873  case Qualifiers::OCL_Strong:
874  case Qualifiers::OCL_Autoreleasing:
875    Diag(DRE->getLocStart(), diag::err_arc_atomic_ownership)
876      << ValType << FirstArg->getSourceRange();
877    return ExprError();
878  }
879
880  // Strip any qualifiers off ValType.
881  ValType = ValType.getUnqualifiedType();
882
883  // The majority of builtins return a value, but a few have special return
884  // types, so allow them to override appropriately below.
885  QualType ResultType = ValType;
886
887  // We need to figure out which concrete builtin this maps onto.  For example,
888  // __sync_fetch_and_add with a 2 byte object turns into
889  // __sync_fetch_and_add_2.
890#define BUILTIN_ROW(x) \
891  { Builtin::BI##x##_1, Builtin::BI##x##_2, Builtin::BI##x##_4, \
892    Builtin::BI##x##_8, Builtin::BI##x##_16 }
893
894  static const unsigned BuiltinIndices[][5] = {
895    BUILTIN_ROW(__sync_fetch_and_add),
896    BUILTIN_ROW(__sync_fetch_and_sub),
897    BUILTIN_ROW(__sync_fetch_and_or),
898    BUILTIN_ROW(__sync_fetch_and_and),
899    BUILTIN_ROW(__sync_fetch_and_xor),
900
901    BUILTIN_ROW(__sync_add_and_fetch),
902    BUILTIN_ROW(__sync_sub_and_fetch),
903    BUILTIN_ROW(__sync_and_and_fetch),
904    BUILTIN_ROW(__sync_or_and_fetch),
905    BUILTIN_ROW(__sync_xor_and_fetch),
906
907    BUILTIN_ROW(__sync_val_compare_and_swap),
908    BUILTIN_ROW(__sync_bool_compare_and_swap),
909    BUILTIN_ROW(__sync_lock_test_and_set),
910    BUILTIN_ROW(__sync_lock_release),
911    BUILTIN_ROW(__sync_swap)
912  };
913#undef BUILTIN_ROW
914
915  // Determine the index of the size.
916  unsigned SizeIndex;
917  switch (Context.getTypeSizeInChars(ValType).getQuantity()) {
918  case 1: SizeIndex = 0; break;
919  case 2: SizeIndex = 1; break;
920  case 4: SizeIndex = 2; break;
921  case 8: SizeIndex = 3; break;
922  case 16: SizeIndex = 4; break;
923  default:
924    Diag(DRE->getLocStart(), diag::err_atomic_builtin_pointer_size)
925      << FirstArg->getType() << FirstArg->getSourceRange();
926    return ExprError();
927  }
928
929  // Each of these builtins has one pointer argument, followed by some number of
930  // values (0, 1 or 2) followed by a potentially empty varags list of stuff
931  // that we ignore.  Find out which row of BuiltinIndices to read from as well
932  // as the number of fixed args.
933  unsigned BuiltinID = FDecl->getBuiltinID();
934  unsigned BuiltinIndex, NumFixed = 1;
935  switch (BuiltinID) {
936  default: llvm_unreachable("Unknown overloaded atomic builtin!");
937  case Builtin::BI__sync_fetch_and_add:
938  case Builtin::BI__sync_fetch_and_add_1:
939  case Builtin::BI__sync_fetch_and_add_2:
940  case Builtin::BI__sync_fetch_and_add_4:
941  case Builtin::BI__sync_fetch_and_add_8:
942  case Builtin::BI__sync_fetch_and_add_16:
943    BuiltinIndex = 0;
944    break;
945
946  case Builtin::BI__sync_fetch_and_sub:
947  case Builtin::BI__sync_fetch_and_sub_1:
948  case Builtin::BI__sync_fetch_and_sub_2:
949  case Builtin::BI__sync_fetch_and_sub_4:
950  case Builtin::BI__sync_fetch_and_sub_8:
951  case Builtin::BI__sync_fetch_and_sub_16:
952    BuiltinIndex = 1;
953    break;
954
955  case Builtin::BI__sync_fetch_and_or:
956  case Builtin::BI__sync_fetch_and_or_1:
957  case Builtin::BI__sync_fetch_and_or_2:
958  case Builtin::BI__sync_fetch_and_or_4:
959  case Builtin::BI__sync_fetch_and_or_8:
960  case Builtin::BI__sync_fetch_and_or_16:
961    BuiltinIndex = 2;
962    break;
963
964  case Builtin::BI__sync_fetch_and_and:
965  case Builtin::BI__sync_fetch_and_and_1:
966  case Builtin::BI__sync_fetch_and_and_2:
967  case Builtin::BI__sync_fetch_and_and_4:
968  case Builtin::BI__sync_fetch_and_and_8:
969  case Builtin::BI__sync_fetch_and_and_16:
970    BuiltinIndex = 3;
971    break;
972
973  case Builtin::BI__sync_fetch_and_xor:
974  case Builtin::BI__sync_fetch_and_xor_1:
975  case Builtin::BI__sync_fetch_and_xor_2:
976  case Builtin::BI__sync_fetch_and_xor_4:
977  case Builtin::BI__sync_fetch_and_xor_8:
978  case Builtin::BI__sync_fetch_and_xor_16:
979    BuiltinIndex = 4;
980    break;
981
982  case Builtin::BI__sync_add_and_fetch:
983  case Builtin::BI__sync_add_and_fetch_1:
984  case Builtin::BI__sync_add_and_fetch_2:
985  case Builtin::BI__sync_add_and_fetch_4:
986  case Builtin::BI__sync_add_and_fetch_8:
987  case Builtin::BI__sync_add_and_fetch_16:
988    BuiltinIndex = 5;
989    break;
990
991  case Builtin::BI__sync_sub_and_fetch:
992  case Builtin::BI__sync_sub_and_fetch_1:
993  case Builtin::BI__sync_sub_and_fetch_2:
994  case Builtin::BI__sync_sub_and_fetch_4:
995  case Builtin::BI__sync_sub_and_fetch_8:
996  case Builtin::BI__sync_sub_and_fetch_16:
997    BuiltinIndex = 6;
998    break;
999
1000  case Builtin::BI__sync_and_and_fetch:
1001  case Builtin::BI__sync_and_and_fetch_1:
1002  case Builtin::BI__sync_and_and_fetch_2:
1003  case Builtin::BI__sync_and_and_fetch_4:
1004  case Builtin::BI__sync_and_and_fetch_8:
1005  case Builtin::BI__sync_and_and_fetch_16:
1006    BuiltinIndex = 7;
1007    break;
1008
1009  case Builtin::BI__sync_or_and_fetch:
1010  case Builtin::BI__sync_or_and_fetch_1:
1011  case Builtin::BI__sync_or_and_fetch_2:
1012  case Builtin::BI__sync_or_and_fetch_4:
1013  case Builtin::BI__sync_or_and_fetch_8:
1014  case Builtin::BI__sync_or_and_fetch_16:
1015    BuiltinIndex = 8;
1016    break;
1017
1018  case Builtin::BI__sync_xor_and_fetch:
1019  case Builtin::BI__sync_xor_and_fetch_1:
1020  case Builtin::BI__sync_xor_and_fetch_2:
1021  case Builtin::BI__sync_xor_and_fetch_4:
1022  case Builtin::BI__sync_xor_and_fetch_8:
1023  case Builtin::BI__sync_xor_and_fetch_16:
1024    BuiltinIndex = 9;
1025    break;
1026
1027  case Builtin::BI__sync_val_compare_and_swap:
1028  case Builtin::BI__sync_val_compare_and_swap_1:
1029  case Builtin::BI__sync_val_compare_and_swap_2:
1030  case Builtin::BI__sync_val_compare_and_swap_4:
1031  case Builtin::BI__sync_val_compare_and_swap_8:
1032  case Builtin::BI__sync_val_compare_and_swap_16:
1033    BuiltinIndex = 10;
1034    NumFixed = 2;
1035    break;
1036
1037  case Builtin::BI__sync_bool_compare_and_swap:
1038  case Builtin::BI__sync_bool_compare_and_swap_1:
1039  case Builtin::BI__sync_bool_compare_and_swap_2:
1040  case Builtin::BI__sync_bool_compare_and_swap_4:
1041  case Builtin::BI__sync_bool_compare_and_swap_8:
1042  case Builtin::BI__sync_bool_compare_and_swap_16:
1043    BuiltinIndex = 11;
1044    NumFixed = 2;
1045    ResultType = Context.BoolTy;
1046    break;
1047
1048  case Builtin::BI__sync_lock_test_and_set:
1049  case Builtin::BI__sync_lock_test_and_set_1:
1050  case Builtin::BI__sync_lock_test_and_set_2:
1051  case Builtin::BI__sync_lock_test_and_set_4:
1052  case Builtin::BI__sync_lock_test_and_set_8:
1053  case Builtin::BI__sync_lock_test_and_set_16:
1054    BuiltinIndex = 12;
1055    break;
1056
1057  case Builtin::BI__sync_lock_release:
1058  case Builtin::BI__sync_lock_release_1:
1059  case Builtin::BI__sync_lock_release_2:
1060  case Builtin::BI__sync_lock_release_4:
1061  case Builtin::BI__sync_lock_release_8:
1062  case Builtin::BI__sync_lock_release_16:
1063    BuiltinIndex = 13;
1064    NumFixed = 0;
1065    ResultType = Context.VoidTy;
1066    break;
1067
1068  case Builtin::BI__sync_swap:
1069  case Builtin::BI__sync_swap_1:
1070  case Builtin::BI__sync_swap_2:
1071  case Builtin::BI__sync_swap_4:
1072  case Builtin::BI__sync_swap_8:
1073  case Builtin::BI__sync_swap_16:
1074    BuiltinIndex = 14;
1075    break;
1076  }
1077
1078  // Now that we know how many fixed arguments we expect, first check that we
1079  // have at least that many.
1080  if (TheCall->getNumArgs() < 1+NumFixed) {
1081    Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args_at_least)
1082      << 0 << 1+NumFixed << TheCall->getNumArgs()
1083      << TheCall->getCallee()->getSourceRange();
1084    return ExprError();
1085  }
1086
1087  // Get the decl for the concrete builtin from this, we can tell what the
1088  // concrete integer type we should convert to is.
1089  unsigned NewBuiltinID = BuiltinIndices[BuiltinIndex][SizeIndex];
1090  const char *NewBuiltinName = Context.BuiltinInfo.GetName(NewBuiltinID);
1091  IdentifierInfo *NewBuiltinII = PP.getIdentifierInfo(NewBuiltinName);
1092  FunctionDecl *NewBuiltinDecl =
1093    cast<FunctionDecl>(LazilyCreateBuiltin(NewBuiltinII, NewBuiltinID,
1094                                           TUScope, false, DRE->getLocStart()));
1095
1096  // The first argument --- the pointer --- has a fixed type; we
1097  // deduce the types of the rest of the arguments accordingly.  Walk
1098  // the remaining arguments, converting them to the deduced value type.
1099  for (unsigned i = 0; i != NumFixed; ++i) {
1100    ExprResult Arg = TheCall->getArg(i+1);
1101
1102    // GCC does an implicit conversion to the pointer or integer ValType.  This
1103    // can fail in some cases (1i -> int**), check for this error case now.
1104    // Initialize the argument.
1105    InitializedEntity Entity = InitializedEntity::InitializeParameter(Context,
1106                                                   ValType, /*consume*/ false);
1107    Arg = PerformCopyInitialization(Entity, SourceLocation(), Arg);
1108    if (Arg.isInvalid())
1109      return ExprError();
1110
1111    // Okay, we have something that *can* be converted to the right type.  Check
1112    // to see if there is a potentially weird extension going on here.  This can
1113    // happen when you do an atomic operation on something like an char* and
1114    // pass in 42.  The 42 gets converted to char.  This is even more strange
1115    // for things like 45.123 -> char, etc.
1116    // FIXME: Do this check.
1117    TheCall->setArg(i+1, Arg.take());
1118  }
1119
1120  ASTContext& Context = this->getASTContext();
1121
1122  // Create a new DeclRefExpr to refer to the new decl.
1123  DeclRefExpr* NewDRE = DeclRefExpr::Create(
1124      Context,
1125      DRE->getQualifierLoc(),
1126      SourceLocation(),
1127      NewBuiltinDecl,
1128      /*enclosing*/ false,
1129      DRE->getLocation(),
1130      NewBuiltinDecl->getType(),
1131      DRE->getValueKind());
1132
1133  // Set the callee in the CallExpr.
1134  // FIXME: This leaks the original parens and implicit casts.
1135  ExprResult PromotedCall = UsualUnaryConversions(NewDRE);
1136  if (PromotedCall.isInvalid())
1137    return ExprError();
1138  TheCall->setCallee(PromotedCall.take());
1139
1140  // Change the result type of the call to match the original value type. This
1141  // is arbitrary, but the codegen for these builtins ins design to handle it
1142  // gracefully.
1143  TheCall->setType(ResultType);
1144
1145  return move(TheCallResult);
1146}
1147
1148/// CheckObjCString - Checks that the argument to the builtin
1149/// CFString constructor is correct
1150/// Note: It might also make sense to do the UTF-16 conversion here (would
1151/// simplify the backend).
1152bool Sema::CheckObjCString(Expr *Arg) {
1153  Arg = Arg->IgnoreParenCasts();
1154  StringLiteral *Literal = dyn_cast<StringLiteral>(Arg);
1155
1156  if (!Literal || !Literal->isAscii()) {
1157    Diag(Arg->getLocStart(), diag::err_cfstring_literal_not_string_constant)
1158      << Arg->getSourceRange();
1159    return true;
1160  }
1161
1162  if (Literal->containsNonAsciiOrNull()) {
1163    StringRef String = Literal->getString();
1164    unsigned NumBytes = String.size();
1165    SmallVector<UTF16, 128> ToBuf(NumBytes);
1166    const UTF8 *FromPtr = (UTF8 *)String.data();
1167    UTF16 *ToPtr = &ToBuf[0];
1168
1169    ConversionResult Result = ConvertUTF8toUTF16(&FromPtr, FromPtr + NumBytes,
1170                                                 &ToPtr, ToPtr + NumBytes,
1171                                                 strictConversion);
1172    // Check for conversion failure.
1173    if (Result != conversionOK)
1174      Diag(Arg->getLocStart(),
1175           diag::warn_cfstring_truncated) << Arg->getSourceRange();
1176  }
1177  return false;
1178}
1179
1180/// SemaBuiltinVAStart - Check the arguments to __builtin_va_start for validity.
1181/// Emit an error and return true on failure, return false on success.
1182bool Sema::SemaBuiltinVAStart(CallExpr *TheCall) {
1183  Expr *Fn = TheCall->getCallee();
1184  if (TheCall->getNumArgs() > 2) {
1185    Diag(TheCall->getArg(2)->getLocStart(),
1186         diag::err_typecheck_call_too_many_args)
1187      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1188      << Fn->getSourceRange()
1189      << SourceRange(TheCall->getArg(2)->getLocStart(),
1190                     (*(TheCall->arg_end()-1))->getLocEnd());
1191    return true;
1192  }
1193
1194  if (TheCall->getNumArgs() < 2) {
1195    return Diag(TheCall->getLocEnd(),
1196      diag::err_typecheck_call_too_few_args_at_least)
1197      << 0 /*function call*/ << 2 << TheCall->getNumArgs();
1198  }
1199
1200  // Type-check the first argument normally.
1201  if (checkBuiltinArgument(*this, TheCall, 0))
1202    return true;
1203
1204  // Determine whether the current function is variadic or not.
1205  BlockScopeInfo *CurBlock = getCurBlock();
1206  bool isVariadic;
1207  if (CurBlock)
1208    isVariadic = CurBlock->TheDecl->isVariadic();
1209  else if (FunctionDecl *FD = getCurFunctionDecl())
1210    isVariadic = FD->isVariadic();
1211  else
1212    isVariadic = getCurMethodDecl()->isVariadic();
1213
1214  if (!isVariadic) {
1215    Diag(Fn->getLocStart(), diag::err_va_start_used_in_non_variadic_function);
1216    return true;
1217  }
1218
1219  // Verify that the second argument to the builtin is the last argument of the
1220  // current function or method.
1221  bool SecondArgIsLastNamedArgument = false;
1222  const Expr *Arg = TheCall->getArg(1)->IgnoreParenCasts();
1223
1224  if (const DeclRefExpr *DR = dyn_cast<DeclRefExpr>(Arg)) {
1225    if (const ParmVarDecl *PV = dyn_cast<ParmVarDecl>(DR->getDecl())) {
1226      // FIXME: This isn't correct for methods (results in bogus warning).
1227      // Get the last formal in the current function.
1228      const ParmVarDecl *LastArg;
1229      if (CurBlock)
1230        LastArg = *(CurBlock->TheDecl->param_end()-1);
1231      else if (FunctionDecl *FD = getCurFunctionDecl())
1232        LastArg = *(FD->param_end()-1);
1233      else
1234        LastArg = *(getCurMethodDecl()->param_end()-1);
1235      SecondArgIsLastNamedArgument = PV == LastArg;
1236    }
1237  }
1238
1239  if (!SecondArgIsLastNamedArgument)
1240    Diag(TheCall->getArg(1)->getLocStart(),
1241         diag::warn_second_parameter_of_va_start_not_last_named_argument);
1242  return false;
1243}
1244
1245/// SemaBuiltinUnorderedCompare - Handle functions like __builtin_isgreater and
1246/// friends.  This is declared to take (...), so we have to check everything.
1247bool Sema::SemaBuiltinUnorderedCompare(CallExpr *TheCall) {
1248  if (TheCall->getNumArgs() < 2)
1249    return Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
1250      << 0 << 2 << TheCall->getNumArgs()/*function call*/;
1251  if (TheCall->getNumArgs() > 2)
1252    return Diag(TheCall->getArg(2)->getLocStart(),
1253                diag::err_typecheck_call_too_many_args)
1254      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1255      << SourceRange(TheCall->getArg(2)->getLocStart(),
1256                     (*(TheCall->arg_end()-1))->getLocEnd());
1257
1258  ExprResult OrigArg0 = TheCall->getArg(0);
1259  ExprResult OrigArg1 = TheCall->getArg(1);
1260
1261  // Do standard promotions between the two arguments, returning their common
1262  // type.
1263  QualType Res = UsualArithmeticConversions(OrigArg0, OrigArg1, false);
1264  if (OrigArg0.isInvalid() || OrigArg1.isInvalid())
1265    return true;
1266
1267  // Make sure any conversions are pushed back into the call; this is
1268  // type safe since unordered compare builtins are declared as "_Bool
1269  // foo(...)".
1270  TheCall->setArg(0, OrigArg0.get());
1271  TheCall->setArg(1, OrigArg1.get());
1272
1273  if (OrigArg0.get()->isTypeDependent() || OrigArg1.get()->isTypeDependent())
1274    return false;
1275
1276  // If the common type isn't a real floating type, then the arguments were
1277  // invalid for this operation.
1278  if (!Res->isRealFloatingType())
1279    return Diag(OrigArg0.get()->getLocStart(),
1280                diag::err_typecheck_call_invalid_ordered_compare)
1281      << OrigArg0.get()->getType() << OrigArg1.get()->getType()
1282      << SourceRange(OrigArg0.get()->getLocStart(), OrigArg1.get()->getLocEnd());
1283
1284  return false;
1285}
1286
1287/// SemaBuiltinSemaBuiltinFPClassification - Handle functions like
1288/// __builtin_isnan and friends.  This is declared to take (...), so we have
1289/// to check everything. We expect the last argument to be a floating point
1290/// value.
1291bool Sema::SemaBuiltinFPClassification(CallExpr *TheCall, unsigned NumArgs) {
1292  if (TheCall->getNumArgs() < NumArgs)
1293    return Diag(TheCall->getLocEnd(), diag::err_typecheck_call_too_few_args)
1294      << 0 << NumArgs << TheCall->getNumArgs()/*function call*/;
1295  if (TheCall->getNumArgs() > NumArgs)
1296    return Diag(TheCall->getArg(NumArgs)->getLocStart(),
1297                diag::err_typecheck_call_too_many_args)
1298      << 0 /*function call*/ << NumArgs << TheCall->getNumArgs()
1299      << SourceRange(TheCall->getArg(NumArgs)->getLocStart(),
1300                     (*(TheCall->arg_end()-1))->getLocEnd());
1301
1302  Expr *OrigArg = TheCall->getArg(NumArgs-1);
1303
1304  if (OrigArg->isTypeDependent())
1305    return false;
1306
1307  // This operation requires a non-_Complex floating-point number.
1308  if (!OrigArg->getType()->isRealFloatingType())
1309    return Diag(OrigArg->getLocStart(),
1310                diag::err_typecheck_call_invalid_unary_fp)
1311      << OrigArg->getType() << OrigArg->getSourceRange();
1312
1313  // If this is an implicit conversion from float -> double, remove it.
1314  if (ImplicitCastExpr *Cast = dyn_cast<ImplicitCastExpr>(OrigArg)) {
1315    Expr *CastArg = Cast->getSubExpr();
1316    if (CastArg->getType()->isSpecificBuiltinType(BuiltinType::Float)) {
1317      assert(Cast->getType()->isSpecificBuiltinType(BuiltinType::Double) &&
1318             "promotion from float to double is the only expected cast here");
1319      Cast->setSubExpr(0);
1320      TheCall->setArg(NumArgs-1, CastArg);
1321    }
1322  }
1323
1324  return false;
1325}
1326
1327/// SemaBuiltinShuffleVector - Handle __builtin_shufflevector.
1328// This is declared to take (...), so we have to check everything.
1329ExprResult Sema::SemaBuiltinShuffleVector(CallExpr *TheCall) {
1330  if (TheCall->getNumArgs() < 2)
1331    return ExprError(Diag(TheCall->getLocEnd(),
1332                          diag::err_typecheck_call_too_few_args_at_least)
1333      << 0 /*function call*/ << 2 << TheCall->getNumArgs()
1334      << TheCall->getSourceRange());
1335
1336  // Determine which of the following types of shufflevector we're checking:
1337  // 1) unary, vector mask: (lhs, mask)
1338  // 2) binary, vector mask: (lhs, rhs, mask)
1339  // 3) binary, scalar mask: (lhs, rhs, index, ..., index)
1340  QualType resType = TheCall->getArg(0)->getType();
1341  unsigned numElements = 0;
1342
1343  if (!TheCall->getArg(0)->isTypeDependent() &&
1344      !TheCall->getArg(1)->isTypeDependent()) {
1345    QualType LHSType = TheCall->getArg(0)->getType();
1346    QualType RHSType = TheCall->getArg(1)->getType();
1347
1348    if (!LHSType->isVectorType() || !RHSType->isVectorType()) {
1349      Diag(TheCall->getLocStart(), diag::err_shufflevector_non_vector)
1350        << SourceRange(TheCall->getArg(0)->getLocStart(),
1351                       TheCall->getArg(1)->getLocEnd());
1352      return ExprError();
1353    }
1354
1355    numElements = LHSType->getAs<VectorType>()->getNumElements();
1356    unsigned numResElements = TheCall->getNumArgs() - 2;
1357
1358    // Check to see if we have a call with 2 vector arguments, the unary shuffle
1359    // with mask.  If so, verify that RHS is an integer vector type with the
1360    // same number of elts as lhs.
1361    if (TheCall->getNumArgs() == 2) {
1362      if (!RHSType->hasIntegerRepresentation() ||
1363          RHSType->getAs<VectorType>()->getNumElements() != numElements)
1364        Diag(TheCall->getLocStart(), diag::err_shufflevector_incompatible_vector)
1365          << SourceRange(TheCall->getArg(1)->getLocStart(),
1366                         TheCall->getArg(1)->getLocEnd());
1367      numResElements = numElements;
1368    }
1369    else if (!Context.hasSameUnqualifiedType(LHSType, RHSType)) {
1370      Diag(TheCall->getLocStart(), diag::err_shufflevector_incompatible_vector)
1371        << SourceRange(TheCall->getArg(0)->getLocStart(),
1372                       TheCall->getArg(1)->getLocEnd());
1373      return ExprError();
1374    } else if (numElements != numResElements) {
1375      QualType eltType = LHSType->getAs<VectorType>()->getElementType();
1376      resType = Context.getVectorType(eltType, numResElements,
1377                                      VectorType::GenericVector);
1378    }
1379  }
1380
1381  for (unsigned i = 2; i < TheCall->getNumArgs(); i++) {
1382    if (TheCall->getArg(i)->isTypeDependent() ||
1383        TheCall->getArg(i)->isValueDependent())
1384      continue;
1385
1386    llvm::APSInt Result(32);
1387    if (!TheCall->getArg(i)->isIntegerConstantExpr(Result, Context))
1388      return ExprError(Diag(TheCall->getLocStart(),
1389                  diag::err_shufflevector_nonconstant_argument)
1390                << TheCall->getArg(i)->getSourceRange());
1391
1392    if (Result.getActiveBits() > 64 || Result.getZExtValue() >= numElements*2)
1393      return ExprError(Diag(TheCall->getLocStart(),
1394                  diag::err_shufflevector_argument_too_large)
1395               << TheCall->getArg(i)->getSourceRange());
1396  }
1397
1398  SmallVector<Expr*, 32> exprs;
1399
1400  for (unsigned i = 0, e = TheCall->getNumArgs(); i != e; i++) {
1401    exprs.push_back(TheCall->getArg(i));
1402    TheCall->setArg(i, 0);
1403  }
1404
1405  return Owned(new (Context) ShuffleVectorExpr(Context, exprs.begin(),
1406                                            exprs.size(), resType,
1407                                            TheCall->getCallee()->getLocStart(),
1408                                            TheCall->getRParenLoc()));
1409}
1410
1411/// SemaBuiltinPrefetch - Handle __builtin_prefetch.
1412// This is declared to take (const void*, ...) and can take two
1413// optional constant int args.
1414bool Sema::SemaBuiltinPrefetch(CallExpr *TheCall) {
1415  unsigned NumArgs = TheCall->getNumArgs();
1416
1417  if (NumArgs > 3)
1418    return Diag(TheCall->getLocEnd(),
1419             diag::err_typecheck_call_too_many_args_at_most)
1420             << 0 /*function call*/ << 3 << NumArgs
1421             << TheCall->getSourceRange();
1422
1423  // Argument 0 is checked for us and the remaining arguments must be
1424  // constant integers.
1425  for (unsigned i = 1; i != NumArgs; ++i) {
1426    Expr *Arg = TheCall->getArg(i);
1427
1428    llvm::APSInt Result;
1429    if (SemaBuiltinConstantArg(TheCall, i, Result))
1430      return true;
1431
1432    // FIXME: gcc issues a warning and rewrites these to 0. These
1433    // seems especially odd for the third argument since the default
1434    // is 3.
1435    if (i == 1) {
1436      if (Result.getLimitedValue() > 1)
1437        return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1438             << "0" << "1" << Arg->getSourceRange();
1439    } else {
1440      if (Result.getLimitedValue() > 3)
1441        return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1442            << "0" << "3" << Arg->getSourceRange();
1443    }
1444  }
1445
1446  return false;
1447}
1448
1449/// SemaBuiltinConstantArg - Handle a check if argument ArgNum of CallExpr
1450/// TheCall is a constant expression.
1451bool Sema::SemaBuiltinConstantArg(CallExpr *TheCall, int ArgNum,
1452                                  llvm::APSInt &Result) {
1453  Expr *Arg = TheCall->getArg(ArgNum);
1454  DeclRefExpr *DRE =cast<DeclRefExpr>(TheCall->getCallee()->IgnoreParenCasts());
1455  FunctionDecl *FDecl = cast<FunctionDecl>(DRE->getDecl());
1456
1457  if (Arg->isTypeDependent() || Arg->isValueDependent()) return false;
1458
1459  if (!Arg->isIntegerConstantExpr(Result, Context))
1460    return Diag(TheCall->getLocStart(), diag::err_constant_integer_arg_type)
1461                << FDecl->getDeclName() <<  Arg->getSourceRange();
1462
1463  return false;
1464}
1465
1466/// SemaBuiltinObjectSize - Handle __builtin_object_size(void *ptr,
1467/// int type). This simply type checks that type is one of the defined
1468/// constants (0-3).
1469// For compatibility check 0-3, llvm only handles 0 and 2.
1470bool Sema::SemaBuiltinObjectSize(CallExpr *TheCall) {
1471  llvm::APSInt Result;
1472
1473  // Check constant-ness first.
1474  if (SemaBuiltinConstantArg(TheCall, 1, Result))
1475    return true;
1476
1477  Expr *Arg = TheCall->getArg(1);
1478  if (Result.getSExtValue() < 0 || Result.getSExtValue() > 3) {
1479    return Diag(TheCall->getLocStart(), diag::err_argument_invalid_range)
1480             << "0" << "3" << SourceRange(Arg->getLocStart(), Arg->getLocEnd());
1481  }
1482
1483  return false;
1484}
1485
1486/// SemaBuiltinLongjmp - Handle __builtin_longjmp(void *env[5], int val).
1487/// This checks that val is a constant 1.
1488bool Sema::SemaBuiltinLongjmp(CallExpr *TheCall) {
1489  Expr *Arg = TheCall->getArg(1);
1490  llvm::APSInt Result;
1491
1492  // TODO: This is less than ideal. Overload this to take a value.
1493  if (SemaBuiltinConstantArg(TheCall, 1, Result))
1494    return true;
1495
1496  if (Result != 1)
1497    return Diag(TheCall->getLocStart(), diag::err_builtin_longjmp_invalid_val)
1498             << SourceRange(Arg->getLocStart(), Arg->getLocEnd());
1499
1500  return false;
1501}
1502
1503// Handle i > 1 ? "x" : "y", recursively.
1504bool Sema::SemaCheckStringLiteral(const Expr *E, Expr **Args,
1505                                  unsigned NumArgs, bool HasVAListArg,
1506                                  unsigned format_idx, unsigned firstDataArg,
1507                                  FormatStringType Type, bool inFunctionCall) {
1508 tryAgain:
1509  if (E->isTypeDependent() || E->isValueDependent())
1510    return false;
1511
1512  E = E->IgnoreParenCasts();
1513
1514  if (E->isNullPointerConstant(Context, Expr::NPC_ValueDependentIsNotNull))
1515    // Technically -Wformat-nonliteral does not warn about this case.
1516    // The behavior of printf and friends in this case is implementation
1517    // dependent.  Ideally if the format string cannot be null then
1518    // it should have a 'nonnull' attribute in the function prototype.
1519    return true;
1520
1521  switch (E->getStmtClass()) {
1522  case Stmt::BinaryConditionalOperatorClass:
1523  case Stmt::ConditionalOperatorClass: {
1524    const AbstractConditionalOperator *C = cast<AbstractConditionalOperator>(E);
1525    return SemaCheckStringLiteral(C->getTrueExpr(), Args, NumArgs, HasVAListArg,
1526                                  format_idx, firstDataArg, Type,
1527                                  inFunctionCall)
1528       && SemaCheckStringLiteral(C->getFalseExpr(), Args, NumArgs, HasVAListArg,
1529                                 format_idx, firstDataArg, Type,
1530                                 inFunctionCall);
1531  }
1532
1533  case Stmt::ImplicitCastExprClass: {
1534    E = cast<ImplicitCastExpr>(E)->getSubExpr();
1535    goto tryAgain;
1536  }
1537
1538  case Stmt::OpaqueValueExprClass:
1539    if (const Expr *src = cast<OpaqueValueExpr>(E)->getSourceExpr()) {
1540      E = src;
1541      goto tryAgain;
1542    }
1543    return false;
1544
1545  case Stmt::PredefinedExprClass:
1546    // While __func__, etc., are technically not string literals, they
1547    // cannot contain format specifiers and thus are not a security
1548    // liability.
1549    return true;
1550
1551  case Stmt::DeclRefExprClass: {
1552    const DeclRefExpr *DR = cast<DeclRefExpr>(E);
1553
1554    // As an exception, do not flag errors for variables binding to
1555    // const string literals.
1556    if (const VarDecl *VD = dyn_cast<VarDecl>(DR->getDecl())) {
1557      bool isConstant = false;
1558      QualType T = DR->getType();
1559
1560      if (const ArrayType *AT = Context.getAsArrayType(T)) {
1561        isConstant = AT->getElementType().isConstant(Context);
1562      } else if (const PointerType *PT = T->getAs<PointerType>()) {
1563        isConstant = T.isConstant(Context) &&
1564                     PT->getPointeeType().isConstant(Context);
1565      } else if (T->isObjCObjectPointerType()) {
1566        // In ObjC, there is usually no "const ObjectPointer" type,
1567        // so don't check if the pointee type is constant.
1568        isConstant = T.isConstant(Context);
1569      }
1570
1571      if (isConstant) {
1572        if (const Expr *Init = VD->getAnyInitializer())
1573          return SemaCheckStringLiteral(Init, Args, NumArgs,
1574                                        HasVAListArg, format_idx, firstDataArg,
1575                                        Type, /*inFunctionCall*/false);
1576      }
1577
1578      // For vprintf* functions (i.e., HasVAListArg==true), we add a
1579      // special check to see if the format string is a function parameter
1580      // of the function calling the printf function.  If the function
1581      // has an attribute indicating it is a printf-like function, then we
1582      // should suppress warnings concerning non-literals being used in a call
1583      // to a vprintf function.  For example:
1584      //
1585      // void
1586      // logmessage(char const *fmt __attribute__ (format (printf, 1, 2)), ...){
1587      //      va_list ap;
1588      //      va_start(ap, fmt);
1589      //      vprintf(fmt, ap);  // Do NOT emit a warning about "fmt".
1590      //      ...
1591      //
1592      if (HasVAListArg) {
1593        if (const ParmVarDecl *PV = dyn_cast<ParmVarDecl>(VD)) {
1594          if (const NamedDecl *ND = dyn_cast<NamedDecl>(PV->getDeclContext())) {
1595            int PVIndex = PV->getFunctionScopeIndex() + 1;
1596            for (specific_attr_iterator<FormatAttr>
1597                 i = ND->specific_attr_begin<FormatAttr>(),
1598                 e = ND->specific_attr_end<FormatAttr>(); i != e ; ++i) {
1599              FormatAttr *PVFormat = *i;
1600              // adjust for implicit parameter
1601              if (const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(ND))
1602                if (MD->isInstance())
1603                  ++PVIndex;
1604              // We also check if the formats are compatible.
1605              // We can't pass a 'scanf' string to a 'printf' function.
1606              if (PVIndex == PVFormat->getFormatIdx() &&
1607                  Type == GetFormatStringType(PVFormat))
1608                return true;
1609            }
1610          }
1611        }
1612      }
1613    }
1614
1615    return false;
1616  }
1617
1618  case Stmt::CallExprClass:
1619  case Stmt::CXXMemberCallExprClass: {
1620    const CallExpr *CE = cast<CallExpr>(E);
1621    if (const NamedDecl *ND = dyn_cast_or_null<NamedDecl>(CE->getCalleeDecl())) {
1622      if (const FormatArgAttr *FA = ND->getAttr<FormatArgAttr>()) {
1623        unsigned ArgIndex = FA->getFormatIdx();
1624        if (const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(ND))
1625          if (MD->isInstance())
1626            --ArgIndex;
1627        const Expr *Arg = CE->getArg(ArgIndex - 1);
1628
1629        return SemaCheckStringLiteral(Arg, Args, NumArgs, HasVAListArg,
1630                                      format_idx, firstDataArg, Type,
1631                                      inFunctionCall);
1632      }
1633    }
1634
1635    return false;
1636  }
1637  case Stmt::ObjCStringLiteralClass:
1638  case Stmt::StringLiteralClass: {
1639    const StringLiteral *StrE = NULL;
1640
1641    if (const ObjCStringLiteral *ObjCFExpr = dyn_cast<ObjCStringLiteral>(E))
1642      StrE = ObjCFExpr->getString();
1643    else
1644      StrE = cast<StringLiteral>(E);
1645
1646    if (StrE) {
1647      CheckFormatString(StrE, E, Args, NumArgs, HasVAListArg, format_idx,
1648                        firstDataArg, Type, inFunctionCall);
1649      return true;
1650    }
1651
1652    return false;
1653  }
1654
1655  default:
1656    return false;
1657  }
1658}
1659
1660void
1661Sema::CheckNonNullArguments(const NonNullAttr *NonNull,
1662                            const Expr * const *ExprArgs,
1663                            SourceLocation CallSiteLoc) {
1664  for (NonNullAttr::args_iterator i = NonNull->args_begin(),
1665                                  e = NonNull->args_end();
1666       i != e; ++i) {
1667    const Expr *ArgExpr = ExprArgs[*i];
1668    if (ArgExpr->isNullPointerConstant(Context,
1669                                       Expr::NPC_ValueDependentIsNotNull))
1670      Diag(CallSiteLoc, diag::warn_null_arg) << ArgExpr->getSourceRange();
1671  }
1672}
1673
1674Sema::FormatStringType Sema::GetFormatStringType(const FormatAttr *Format) {
1675  return llvm::StringSwitch<FormatStringType>(Format->getType())
1676  .Case("scanf", FST_Scanf)
1677  .Cases("printf", "printf0", FST_Printf)
1678  .Cases("NSString", "CFString", FST_NSString)
1679  .Case("strftime", FST_Strftime)
1680  .Case("strfmon", FST_Strfmon)
1681  .Cases("kprintf", "cmn_err", "vcmn_err", "zcmn_err", FST_Kprintf)
1682  .Default(FST_Unknown);
1683}
1684
1685/// CheckPrintfScanfArguments - Check calls to printf and scanf (and similar
1686/// functions) for correct use of format strings.
1687void Sema::CheckFormatArguments(const FormatAttr *Format, CallExpr *TheCall) {
1688  bool IsCXXMember = false;
1689  // The way the format attribute works in GCC, the implicit this argument
1690  // of member functions is counted. However, it doesn't appear in our own
1691  // lists, so decrement format_idx in that case.
1692  IsCXXMember = isa<CXXMemberCallExpr>(TheCall);
1693  CheckFormatArguments(Format, TheCall->getArgs(), TheCall->getNumArgs(),
1694                       IsCXXMember, TheCall->getRParenLoc(),
1695                       TheCall->getCallee()->getSourceRange());
1696}
1697
1698void Sema::CheckFormatArguments(const FormatAttr *Format, Expr **Args,
1699                                unsigned NumArgs, bool IsCXXMember,
1700                                SourceLocation Loc, SourceRange Range) {
1701  bool HasVAListArg = Format->getFirstArg() == 0;
1702  unsigned format_idx = Format->getFormatIdx() - 1;
1703  unsigned firstDataArg = HasVAListArg ? 0 : Format->getFirstArg() - 1;
1704  if (IsCXXMember) {
1705    if (format_idx == 0)
1706      return;
1707    --format_idx;
1708    if(firstDataArg != 0)
1709      --firstDataArg;
1710  }
1711  CheckFormatArguments(Args, NumArgs, HasVAListArg, format_idx,
1712                       firstDataArg, GetFormatStringType(Format), Loc, Range);
1713}
1714
1715void Sema::CheckFormatArguments(Expr **Args, unsigned NumArgs,
1716                                bool HasVAListArg, unsigned format_idx,
1717                                unsigned firstDataArg, FormatStringType Type,
1718                                SourceLocation Loc, SourceRange Range) {
1719  // CHECK: printf/scanf-like function is called with no format string.
1720  if (format_idx >= NumArgs) {
1721    Diag(Loc, diag::warn_missing_format_string) << Range;
1722    return;
1723  }
1724
1725  const Expr *OrigFormatExpr = Args[format_idx]->IgnoreParenCasts();
1726
1727  // CHECK: format string is not a string literal.
1728  //
1729  // Dynamically generated format strings are difficult to
1730  // automatically vet at compile time.  Requiring that format strings
1731  // are string literals: (1) permits the checking of format strings by
1732  // the compiler and thereby (2) can practically remove the source of
1733  // many format string exploits.
1734
1735  // Format string can be either ObjC string (e.g. @"%d") or
1736  // C string (e.g. "%d")
1737  // ObjC string uses the same format specifiers as C string, so we can use
1738  // the same format string checking logic for both ObjC and C strings.
1739  if (SemaCheckStringLiteral(OrigFormatExpr, Args, NumArgs, HasVAListArg,
1740                             format_idx, firstDataArg, Type))
1741    return;  // Literal format string found, check done!
1742
1743  // Strftime is particular as it always uses a single 'time' argument,
1744  // so it is safe to pass a non-literal string.
1745  if (Type == FST_Strftime)
1746    return;
1747
1748  // Do not emit diag when the string param is a macro expansion and the
1749  // format is either NSString or CFString. This is a hack to prevent
1750  // diag when using the NSLocalizedString and CFCopyLocalizedString macros
1751  // which are usually used in place of NS and CF string literals.
1752  if (Type == FST_NSString && Args[format_idx]->getLocStart().isMacroID())
1753    return;
1754
1755  // If there are no arguments specified, warn with -Wformat-security, otherwise
1756  // warn only with -Wformat-nonliteral.
1757  if (NumArgs == format_idx+1)
1758    Diag(Args[format_idx]->getLocStart(),
1759         diag::warn_format_nonliteral_noargs)
1760      << OrigFormatExpr->getSourceRange();
1761  else
1762    Diag(Args[format_idx]->getLocStart(),
1763         diag::warn_format_nonliteral)
1764           << OrigFormatExpr->getSourceRange();
1765}
1766
1767namespace {
1768class CheckFormatHandler : public analyze_format_string::FormatStringHandler {
1769protected:
1770  Sema &S;
1771  const StringLiteral *FExpr;
1772  const Expr *OrigFormatExpr;
1773  const unsigned FirstDataArg;
1774  const unsigned NumDataArgs;
1775  const bool IsObjCLiteral;
1776  const char *Beg; // Start of format string.
1777  const bool HasVAListArg;
1778  const Expr * const *Args;
1779  const unsigned NumArgs;
1780  unsigned FormatIdx;
1781  llvm::BitVector CoveredArgs;
1782  bool usesPositionalArgs;
1783  bool atFirstArg;
1784  bool inFunctionCall;
1785public:
1786  CheckFormatHandler(Sema &s, const StringLiteral *fexpr,
1787                     const Expr *origFormatExpr, unsigned firstDataArg,
1788                     unsigned numDataArgs, bool isObjCLiteral,
1789                     const char *beg, bool hasVAListArg,
1790                     Expr **args, unsigned numArgs,
1791                     unsigned formatIdx, bool inFunctionCall)
1792    : S(s), FExpr(fexpr), OrigFormatExpr(origFormatExpr),
1793      FirstDataArg(firstDataArg),
1794      NumDataArgs(numDataArgs),
1795      IsObjCLiteral(isObjCLiteral), Beg(beg),
1796      HasVAListArg(hasVAListArg),
1797      Args(args), NumArgs(numArgs), FormatIdx(formatIdx),
1798      usesPositionalArgs(false), atFirstArg(true),
1799      inFunctionCall(inFunctionCall) {
1800        CoveredArgs.resize(numDataArgs);
1801        CoveredArgs.reset();
1802      }
1803
1804  void DoneProcessing();
1805
1806  void HandleIncompleteSpecifier(const char *startSpecifier,
1807                                 unsigned specifierLen);
1808
1809  void HandleNonStandardLengthModifier(
1810      const analyze_format_string::LengthModifier &LM,
1811      const char *startSpecifier, unsigned specifierLen);
1812
1813  void HandleNonStandardConversionSpecifier(
1814      const analyze_format_string::ConversionSpecifier &CS,
1815      const char *startSpecifier, unsigned specifierLen);
1816
1817  void HandleNonStandardConversionSpecification(
1818      const analyze_format_string::LengthModifier &LM,
1819      const analyze_format_string::ConversionSpecifier &CS,
1820      const char *startSpecifier, unsigned specifierLen);
1821
1822  virtual void HandlePosition(const char *startPos, unsigned posLen);
1823
1824  virtual void HandleInvalidPosition(const char *startSpecifier,
1825                                     unsigned specifierLen,
1826                                     analyze_format_string::PositionContext p);
1827
1828  virtual void HandleZeroPosition(const char *startPos, unsigned posLen);
1829
1830  void HandleNullChar(const char *nullCharacter);
1831
1832  template <typename Range>
1833  static void EmitFormatDiagnostic(Sema &S, bool inFunctionCall,
1834                                   const Expr *ArgumentExpr,
1835                                   PartialDiagnostic PDiag,
1836                                   SourceLocation StringLoc,
1837                                   bool IsStringLocation, Range StringRange,
1838                                   FixItHint Fixit = FixItHint());
1839
1840protected:
1841  bool HandleInvalidConversionSpecifier(unsigned argIndex, SourceLocation Loc,
1842                                        const char *startSpec,
1843                                        unsigned specifierLen,
1844                                        const char *csStart, unsigned csLen);
1845
1846  void HandlePositionalNonpositionalArgs(SourceLocation Loc,
1847                                         const char *startSpec,
1848                                         unsigned specifierLen);
1849
1850  SourceRange getFormatStringRange();
1851  CharSourceRange getSpecifierRange(const char *startSpecifier,
1852                                    unsigned specifierLen);
1853  SourceLocation getLocationOfByte(const char *x);
1854
1855  const Expr *getDataArg(unsigned i) const;
1856
1857  bool CheckNumArgs(const analyze_format_string::FormatSpecifier &FS,
1858                    const analyze_format_string::ConversionSpecifier &CS,
1859                    const char *startSpecifier, unsigned specifierLen,
1860                    unsigned argIndex);
1861
1862  template <typename Range>
1863  void EmitFormatDiagnostic(PartialDiagnostic PDiag, SourceLocation StringLoc,
1864                            bool IsStringLocation, Range StringRange,
1865                            FixItHint Fixit = FixItHint());
1866
1867  void CheckPositionalAndNonpositionalArgs(
1868      const analyze_format_string::FormatSpecifier *FS);
1869};
1870}
1871
1872SourceRange CheckFormatHandler::getFormatStringRange() {
1873  return OrigFormatExpr->getSourceRange();
1874}
1875
1876CharSourceRange CheckFormatHandler::
1877getSpecifierRange(const char *startSpecifier, unsigned specifierLen) {
1878  SourceLocation Start = getLocationOfByte(startSpecifier);
1879  SourceLocation End   = getLocationOfByte(startSpecifier + specifierLen - 1);
1880
1881  // Advance the end SourceLocation by one due to half-open ranges.
1882  End = End.getLocWithOffset(1);
1883
1884  return CharSourceRange::getCharRange(Start, End);
1885}
1886
1887SourceLocation CheckFormatHandler::getLocationOfByte(const char *x) {
1888  return S.getLocationOfStringLiteralByte(FExpr, x - Beg);
1889}
1890
1891void CheckFormatHandler::HandleIncompleteSpecifier(const char *startSpecifier,
1892                                                   unsigned specifierLen){
1893  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_incomplete_specifier),
1894                       getLocationOfByte(startSpecifier),
1895                       /*IsStringLocation*/true,
1896                       getSpecifierRange(startSpecifier, specifierLen));
1897}
1898
1899void CheckFormatHandler::HandleNonStandardLengthModifier(
1900    const analyze_format_string::LengthModifier &LM,
1901    const char *startSpecifier, unsigned specifierLen) {
1902  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard) << LM.toString()
1903                       << 0,
1904                       getLocationOfByte(LM.getStart()),
1905                       /*IsStringLocation*/true,
1906                       getSpecifierRange(startSpecifier, specifierLen));
1907}
1908
1909void CheckFormatHandler::HandleNonStandardConversionSpecifier(
1910    const analyze_format_string::ConversionSpecifier &CS,
1911    const char *startSpecifier, unsigned specifierLen) {
1912  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard) << CS.toString()
1913                       << 1,
1914                       getLocationOfByte(CS.getStart()),
1915                       /*IsStringLocation*/true,
1916                       getSpecifierRange(startSpecifier, specifierLen));
1917}
1918
1919void CheckFormatHandler::HandleNonStandardConversionSpecification(
1920    const analyze_format_string::LengthModifier &LM,
1921    const analyze_format_string::ConversionSpecifier &CS,
1922    const char *startSpecifier, unsigned specifierLen) {
1923  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard_conversion_spec)
1924                       << LM.toString() << CS.toString(),
1925                       getLocationOfByte(LM.getStart()),
1926                       /*IsStringLocation*/true,
1927                       getSpecifierRange(startSpecifier, specifierLen));
1928}
1929
1930void CheckFormatHandler::HandlePosition(const char *startPos,
1931                                        unsigned posLen) {
1932  EmitFormatDiagnostic(S.PDiag(diag::warn_format_non_standard_positional_arg),
1933                               getLocationOfByte(startPos),
1934                               /*IsStringLocation*/true,
1935                               getSpecifierRange(startPos, posLen));
1936}
1937
1938void
1939CheckFormatHandler::HandleInvalidPosition(const char *startPos, unsigned posLen,
1940                                     analyze_format_string::PositionContext p) {
1941  EmitFormatDiagnostic(S.PDiag(diag::warn_format_invalid_positional_specifier)
1942                         << (unsigned) p,
1943                       getLocationOfByte(startPos), /*IsStringLocation*/true,
1944                       getSpecifierRange(startPos, posLen));
1945}
1946
1947void CheckFormatHandler::HandleZeroPosition(const char *startPos,
1948                                            unsigned posLen) {
1949  EmitFormatDiagnostic(S.PDiag(diag::warn_format_zero_positional_specifier),
1950                               getLocationOfByte(startPos),
1951                               /*IsStringLocation*/true,
1952                               getSpecifierRange(startPos, posLen));
1953}
1954
1955void CheckFormatHandler::HandleNullChar(const char *nullCharacter) {
1956  if (!IsObjCLiteral) {
1957    // The presence of a null character is likely an error.
1958    EmitFormatDiagnostic(
1959      S.PDiag(diag::warn_printf_format_string_contains_null_char),
1960      getLocationOfByte(nullCharacter), /*IsStringLocation*/true,
1961      getFormatStringRange());
1962  }
1963}
1964
1965const Expr *CheckFormatHandler::getDataArg(unsigned i) const {
1966  return Args[FirstDataArg + i];
1967}
1968
1969void CheckFormatHandler::DoneProcessing() {
1970    // Does the number of data arguments exceed the number of
1971    // format conversions in the format string?
1972  if (!HasVAListArg) {
1973      // Find any arguments that weren't covered.
1974    CoveredArgs.flip();
1975    signed notCoveredArg = CoveredArgs.find_first();
1976    if (notCoveredArg >= 0) {
1977      assert((unsigned)notCoveredArg < NumDataArgs);
1978      EmitFormatDiagnostic(S.PDiag(diag::warn_printf_data_arg_not_used),
1979                           getDataArg((unsigned) notCoveredArg)->getLocStart(),
1980                           /*IsStringLocation*/false, getFormatStringRange());
1981    }
1982  }
1983}
1984
1985bool
1986CheckFormatHandler::HandleInvalidConversionSpecifier(unsigned argIndex,
1987                                                     SourceLocation Loc,
1988                                                     const char *startSpec,
1989                                                     unsigned specifierLen,
1990                                                     const char *csStart,
1991                                                     unsigned csLen) {
1992
1993  bool keepGoing = true;
1994  if (argIndex < NumDataArgs) {
1995    // Consider the argument coverered, even though the specifier doesn't
1996    // make sense.
1997    CoveredArgs.set(argIndex);
1998  }
1999  else {
2000    // If argIndex exceeds the number of data arguments we
2001    // don't issue a warning because that is just a cascade of warnings (and
2002    // they may have intended '%%' anyway). We don't want to continue processing
2003    // the format string after this point, however, as we will like just get
2004    // gibberish when trying to match arguments.
2005    keepGoing = false;
2006  }
2007
2008  EmitFormatDiagnostic(S.PDiag(diag::warn_format_invalid_conversion)
2009                         << StringRef(csStart, csLen),
2010                       Loc, /*IsStringLocation*/true,
2011                       getSpecifierRange(startSpec, specifierLen));
2012
2013  return keepGoing;
2014}
2015
2016void
2017CheckFormatHandler::HandlePositionalNonpositionalArgs(SourceLocation Loc,
2018                                                      const char *startSpec,
2019                                                      unsigned specifierLen) {
2020  EmitFormatDiagnostic(
2021    S.PDiag(diag::warn_format_mix_positional_nonpositional_args),
2022    Loc, /*isStringLoc*/true, getSpecifierRange(startSpec, specifierLen));
2023}
2024
2025bool
2026CheckFormatHandler::CheckNumArgs(
2027  const analyze_format_string::FormatSpecifier &FS,
2028  const analyze_format_string::ConversionSpecifier &CS,
2029  const char *startSpecifier, unsigned specifierLen, unsigned argIndex) {
2030
2031  if (argIndex >= NumDataArgs) {
2032    PartialDiagnostic PDiag = FS.usesPositionalArg()
2033      ? (S.PDiag(diag::warn_printf_positional_arg_exceeds_data_args)
2034           << (argIndex+1) << NumDataArgs)
2035      : S.PDiag(diag::warn_printf_insufficient_data_args);
2036    EmitFormatDiagnostic(
2037      PDiag, getLocationOfByte(CS.getStart()), /*IsStringLocation*/true,
2038      getSpecifierRange(startSpecifier, specifierLen));
2039    return false;
2040  }
2041  return true;
2042}
2043
2044template<typename Range>
2045void CheckFormatHandler::EmitFormatDiagnostic(PartialDiagnostic PDiag,
2046                                              SourceLocation Loc,
2047                                              bool IsStringLocation,
2048                                              Range StringRange,
2049                                              FixItHint FixIt) {
2050  EmitFormatDiagnostic(S, inFunctionCall, Args[FormatIdx], PDiag,
2051                       Loc, IsStringLocation, StringRange, FixIt);
2052}
2053
2054/// \brief If the format string is not within the funcion call, emit a note
2055/// so that the function call and string are in diagnostic messages.
2056///
2057/// \param inFunctionCall if true, the format string is within the function
2058/// call and only one diagnostic message will be produced.  Otherwise, an
2059/// extra note will be emitted pointing to location of the format string.
2060///
2061/// \param ArgumentExpr the expression that is passed as the format string
2062/// argument in the function call.  Used for getting locations when two
2063/// diagnostics are emitted.
2064///
2065/// \param PDiag the callee should already have provided any strings for the
2066/// diagnostic message.  This function only adds locations and fixits
2067/// to diagnostics.
2068///
2069/// \param Loc primary location for diagnostic.  If two diagnostics are
2070/// required, one will be at Loc and a new SourceLocation will be created for
2071/// the other one.
2072///
2073/// \param IsStringLocation if true, Loc points to the format string should be
2074/// used for the note.  Otherwise, Loc points to the argument list and will
2075/// be used with PDiag.
2076///
2077/// \param StringRange some or all of the string to highlight.  This is
2078/// templated so it can accept either a CharSourceRange or a SourceRange.
2079///
2080/// \param Fixit optional fix it hint for the format string.
2081template<typename Range>
2082void CheckFormatHandler::EmitFormatDiagnostic(Sema &S, bool InFunctionCall,
2083                                              const Expr *ArgumentExpr,
2084                                              PartialDiagnostic PDiag,
2085                                              SourceLocation Loc,
2086                                              bool IsStringLocation,
2087                                              Range StringRange,
2088                                              FixItHint FixIt) {
2089  if (InFunctionCall)
2090    S.Diag(Loc, PDiag) << StringRange << FixIt;
2091  else {
2092    S.Diag(IsStringLocation ? ArgumentExpr->getExprLoc() : Loc, PDiag)
2093      << ArgumentExpr->getSourceRange();
2094    S.Diag(IsStringLocation ? Loc : StringRange.getBegin(),
2095           diag::note_format_string_defined)
2096      << StringRange << FixIt;
2097  }
2098}
2099
2100//===--- CHECK: Printf format string checking ------------------------------===//
2101
2102namespace {
2103class CheckPrintfHandler : public CheckFormatHandler {
2104public:
2105  CheckPrintfHandler(Sema &s, const StringLiteral *fexpr,
2106                     const Expr *origFormatExpr, unsigned firstDataArg,
2107                     unsigned numDataArgs, bool isObjCLiteral,
2108                     const char *beg, bool hasVAListArg,
2109                     Expr **Args, unsigned NumArgs,
2110                     unsigned formatIdx, bool inFunctionCall)
2111  : CheckFormatHandler(s, fexpr, origFormatExpr, firstDataArg,
2112                       numDataArgs, isObjCLiteral, beg, hasVAListArg,
2113                       Args, NumArgs, formatIdx, inFunctionCall) {}
2114
2115
2116  bool HandleInvalidPrintfConversionSpecifier(
2117                                      const analyze_printf::PrintfSpecifier &FS,
2118                                      const char *startSpecifier,
2119                                      unsigned specifierLen);
2120
2121  bool HandlePrintfSpecifier(const analyze_printf::PrintfSpecifier &FS,
2122                             const char *startSpecifier,
2123                             unsigned specifierLen);
2124
2125  bool HandleAmount(const analyze_format_string::OptionalAmount &Amt, unsigned k,
2126                    const char *startSpecifier, unsigned specifierLen);
2127  void HandleInvalidAmount(const analyze_printf::PrintfSpecifier &FS,
2128                           const analyze_printf::OptionalAmount &Amt,
2129                           unsigned type,
2130                           const char *startSpecifier, unsigned specifierLen);
2131  void HandleFlag(const analyze_printf::PrintfSpecifier &FS,
2132                  const analyze_printf::OptionalFlag &flag,
2133                  const char *startSpecifier, unsigned specifierLen);
2134  void HandleIgnoredFlag(const analyze_printf::PrintfSpecifier &FS,
2135                         const analyze_printf::OptionalFlag &ignoredFlag,
2136                         const analyze_printf::OptionalFlag &flag,
2137                         const char *startSpecifier, unsigned specifierLen);
2138};
2139}
2140
2141bool CheckPrintfHandler::HandleInvalidPrintfConversionSpecifier(
2142                                      const analyze_printf::PrintfSpecifier &FS,
2143                                      const char *startSpecifier,
2144                                      unsigned specifierLen) {
2145  const analyze_printf::PrintfConversionSpecifier &CS =
2146    FS.getConversionSpecifier();
2147
2148  return HandleInvalidConversionSpecifier(FS.getArgIndex(),
2149                                          getLocationOfByte(CS.getStart()),
2150                                          startSpecifier, specifierLen,
2151                                          CS.getStart(), CS.getLength());
2152}
2153
2154bool CheckPrintfHandler::HandleAmount(
2155                               const analyze_format_string::OptionalAmount &Amt,
2156                               unsigned k, const char *startSpecifier,
2157                               unsigned specifierLen) {
2158
2159  if (Amt.hasDataArgument()) {
2160    if (!HasVAListArg) {
2161      unsigned argIndex = Amt.getArgIndex();
2162      if (argIndex >= NumDataArgs) {
2163        EmitFormatDiagnostic(S.PDiag(diag::warn_printf_asterisk_missing_arg)
2164                               << k,
2165                             getLocationOfByte(Amt.getStart()),
2166                             /*IsStringLocation*/true,
2167                             getSpecifierRange(startSpecifier, specifierLen));
2168        // Don't do any more checking.  We will just emit
2169        // spurious errors.
2170        return false;
2171      }
2172
2173      // Type check the data argument.  It should be an 'int'.
2174      // Although not in conformance with C99, we also allow the argument to be
2175      // an 'unsigned int' as that is a reasonably safe case.  GCC also
2176      // doesn't emit a warning for that case.
2177      CoveredArgs.set(argIndex);
2178      const Expr *Arg = getDataArg(argIndex);
2179      QualType T = Arg->getType();
2180
2181      const analyze_printf::ArgTypeResult &ATR = Amt.getArgType(S.Context);
2182      assert(ATR.isValid());
2183
2184      if (!ATR.matchesType(S.Context, T)) {
2185        EmitFormatDiagnostic(S.PDiag(diag::warn_printf_asterisk_wrong_type)
2186                               << k << ATR.getRepresentativeTypeName(S.Context)
2187                               << T << Arg->getSourceRange(),
2188                             getLocationOfByte(Amt.getStart()),
2189                             /*IsStringLocation*/true,
2190                             getSpecifierRange(startSpecifier, specifierLen));
2191        // Don't do any more checking.  We will just emit
2192        // spurious errors.
2193        return false;
2194      }
2195    }
2196  }
2197  return true;
2198}
2199
2200void CheckPrintfHandler::HandleInvalidAmount(
2201                                      const analyze_printf::PrintfSpecifier &FS,
2202                                      const analyze_printf::OptionalAmount &Amt,
2203                                      unsigned type,
2204                                      const char *startSpecifier,
2205                                      unsigned specifierLen) {
2206  const analyze_printf::PrintfConversionSpecifier &CS =
2207    FS.getConversionSpecifier();
2208
2209  FixItHint fixit =
2210    Amt.getHowSpecified() == analyze_printf::OptionalAmount::Constant
2211      ? FixItHint::CreateRemoval(getSpecifierRange(Amt.getStart(),
2212                                 Amt.getConstantLength()))
2213      : FixItHint();
2214
2215  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_nonsensical_optional_amount)
2216                         << type << CS.toString(),
2217                       getLocationOfByte(Amt.getStart()),
2218                       /*IsStringLocation*/true,
2219                       getSpecifierRange(startSpecifier, specifierLen),
2220                       fixit);
2221}
2222
2223void CheckPrintfHandler::HandleFlag(const analyze_printf::PrintfSpecifier &FS,
2224                                    const analyze_printf::OptionalFlag &flag,
2225                                    const char *startSpecifier,
2226                                    unsigned specifierLen) {
2227  // Warn about pointless flag with a fixit removal.
2228  const analyze_printf::PrintfConversionSpecifier &CS =
2229    FS.getConversionSpecifier();
2230  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_nonsensical_flag)
2231                         << flag.toString() << CS.toString(),
2232                       getLocationOfByte(flag.getPosition()),
2233                       /*IsStringLocation*/true,
2234                       getSpecifierRange(startSpecifier, specifierLen),
2235                       FixItHint::CreateRemoval(
2236                         getSpecifierRange(flag.getPosition(), 1)));
2237}
2238
2239void CheckPrintfHandler::HandleIgnoredFlag(
2240                                const analyze_printf::PrintfSpecifier &FS,
2241                                const analyze_printf::OptionalFlag &ignoredFlag,
2242                                const analyze_printf::OptionalFlag &flag,
2243                                const char *startSpecifier,
2244                                unsigned specifierLen) {
2245  // Warn about ignored flag with a fixit removal.
2246  EmitFormatDiagnostic(S.PDiag(diag::warn_printf_ignored_flag)
2247                         << ignoredFlag.toString() << flag.toString(),
2248                       getLocationOfByte(ignoredFlag.getPosition()),
2249                       /*IsStringLocation*/true,
2250                       getSpecifierRange(startSpecifier, specifierLen),
2251                       FixItHint::CreateRemoval(
2252                         getSpecifierRange(ignoredFlag.getPosition(), 1)));
2253}
2254
2255bool
2256CheckPrintfHandler::HandlePrintfSpecifier(const analyze_printf::PrintfSpecifier
2257                                            &FS,
2258                                          const char *startSpecifier,
2259                                          unsigned specifierLen) {
2260
2261  using namespace analyze_format_string;
2262  using namespace analyze_printf;
2263  const PrintfConversionSpecifier &CS = FS.getConversionSpecifier();
2264
2265  if (FS.consumesDataArgument()) {
2266    if (atFirstArg) {
2267        atFirstArg = false;
2268        usesPositionalArgs = FS.usesPositionalArg();
2269    }
2270    else if (usesPositionalArgs != FS.usesPositionalArg()) {
2271      HandlePositionalNonpositionalArgs(getLocationOfByte(CS.getStart()),
2272                                        startSpecifier, specifierLen);
2273      return false;
2274    }
2275  }
2276
2277  // First check if the field width, precision, and conversion specifier
2278  // have matching data arguments.
2279  if (!HandleAmount(FS.getFieldWidth(), /* field width */ 0,
2280                    startSpecifier, specifierLen)) {
2281    return false;
2282  }
2283
2284  if (!HandleAmount(FS.getPrecision(), /* precision */ 1,
2285                    startSpecifier, specifierLen)) {
2286    return false;
2287  }
2288
2289  if (!CS.consumesDataArgument()) {
2290    // FIXME: Technically specifying a precision or field width here
2291    // makes no sense.  Worth issuing a warning at some point.
2292    return true;
2293  }
2294
2295  // Consume the argument.
2296  unsigned argIndex = FS.getArgIndex();
2297  if (argIndex < NumDataArgs) {
2298    // The check to see if the argIndex is valid will come later.
2299    // We set the bit here because we may exit early from this
2300    // function if we encounter some other error.
2301    CoveredArgs.set(argIndex);
2302  }
2303
2304  // Check for using an Objective-C specific conversion specifier
2305  // in a non-ObjC literal.
2306  if (!IsObjCLiteral && CS.isObjCArg()) {
2307    return HandleInvalidPrintfConversionSpecifier(FS, startSpecifier,
2308                                                  specifierLen);
2309  }
2310
2311  // Check for invalid use of field width
2312  if (!FS.hasValidFieldWidth()) {
2313    HandleInvalidAmount(FS, FS.getFieldWidth(), /* field width */ 0,
2314        startSpecifier, specifierLen);
2315  }
2316
2317  // Check for invalid use of precision
2318  if (!FS.hasValidPrecision()) {
2319    HandleInvalidAmount(FS, FS.getPrecision(), /* precision */ 1,
2320        startSpecifier, specifierLen);
2321  }
2322
2323  // Check each flag does not conflict with any other component.
2324  if (!FS.hasValidThousandsGroupingPrefix())
2325    HandleFlag(FS, FS.hasThousandsGrouping(), startSpecifier, specifierLen);
2326  if (!FS.hasValidLeadingZeros())
2327    HandleFlag(FS, FS.hasLeadingZeros(), startSpecifier, specifierLen);
2328  if (!FS.hasValidPlusPrefix())
2329    HandleFlag(FS, FS.hasPlusPrefix(), startSpecifier, specifierLen);
2330  if (!FS.hasValidSpacePrefix())
2331    HandleFlag(FS, FS.hasSpacePrefix(), startSpecifier, specifierLen);
2332  if (!FS.hasValidAlternativeForm())
2333    HandleFlag(FS, FS.hasAlternativeForm(), startSpecifier, specifierLen);
2334  if (!FS.hasValidLeftJustified())
2335    HandleFlag(FS, FS.isLeftJustified(), startSpecifier, specifierLen);
2336
2337  // Check that flags are not ignored by another flag
2338  if (FS.hasSpacePrefix() && FS.hasPlusPrefix()) // ' ' ignored by '+'
2339    HandleIgnoredFlag(FS, FS.hasSpacePrefix(), FS.hasPlusPrefix(),
2340        startSpecifier, specifierLen);
2341  if (FS.hasLeadingZeros() && FS.isLeftJustified()) // '0' ignored by '-'
2342    HandleIgnoredFlag(FS, FS.hasLeadingZeros(), FS.isLeftJustified(),
2343            startSpecifier, specifierLen);
2344
2345  // Check the length modifier is valid with the given conversion specifier.
2346  const LengthModifier &LM = FS.getLengthModifier();
2347  if (!FS.hasValidLengthModifier())
2348    EmitFormatDiagnostic(S.PDiag(diag::warn_format_nonsensical_length)
2349                           << LM.toString() << CS.toString(),
2350                         getLocationOfByte(LM.getStart()),
2351                         /*IsStringLocation*/true,
2352                         getSpecifierRange(startSpecifier, specifierLen),
2353                         FixItHint::CreateRemoval(
2354                           getSpecifierRange(LM.getStart(),
2355                                             LM.getLength())));
2356  if (!FS.hasStandardLengthModifier())
2357    HandleNonStandardLengthModifier(LM, startSpecifier, specifierLen);
2358  if (!FS.hasStandardConversionSpecifier(S.getLangOpts()))
2359    HandleNonStandardConversionSpecifier(CS, startSpecifier, specifierLen);
2360  if (!FS.hasStandardLengthConversionCombination())
2361    HandleNonStandardConversionSpecification(LM, CS, startSpecifier,
2362                                             specifierLen);
2363
2364  // Are we using '%n'?
2365  if (CS.getKind() == ConversionSpecifier::nArg) {
2366    // Issue a warning about this being a possible security issue.
2367    EmitFormatDiagnostic(S.PDiag(diag::warn_printf_write_back),
2368                         getLocationOfByte(CS.getStart()),
2369                         /*IsStringLocation*/true,
2370                         getSpecifierRange(startSpecifier, specifierLen));
2371    // Continue checking the other format specifiers.
2372    return true;
2373  }
2374
2375  // The remaining checks depend on the data arguments.
2376  if (HasVAListArg)
2377    return true;
2378
2379  if (!CheckNumArgs(FS, CS, startSpecifier, specifierLen, argIndex))
2380    return false;
2381
2382  // Now type check the data expression that matches the
2383  // format specifier.
2384  const Expr *Ex = getDataArg(argIndex);
2385  const analyze_printf::ArgTypeResult &ATR = FS.getArgType(S.Context,
2386                                                           IsObjCLiteral);
2387  if (ATR.isValid() && !ATR.matchesType(S.Context, Ex->getType())) {
2388    // Check if we didn't match because of an implicit cast from a 'char'
2389    // or 'short' to an 'int'.  This is done because printf is a varargs
2390    // function.
2391    if (const ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(Ex))
2392      if (ICE->getType() == S.Context.IntTy) {
2393        // All further checking is done on the subexpression.
2394        Ex = ICE->getSubExpr();
2395        if (ATR.matchesType(S.Context, Ex->getType()))
2396          return true;
2397      }
2398
2399    // We may be able to offer a FixItHint if it is a supported type.
2400    PrintfSpecifier fixedFS = FS;
2401    bool success = fixedFS.fixType(Ex->getType(), S.getLangOpts(),
2402                                   S.Context, IsObjCLiteral);
2403
2404    if (success) {
2405      // Get the fix string from the fixed format specifier
2406      SmallString<128> buf;
2407      llvm::raw_svector_ostream os(buf);
2408      fixedFS.toString(os);
2409
2410      EmitFormatDiagnostic(
2411        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2412          << ATR.getRepresentativeTypeName(S.Context) << Ex->getType()
2413          << Ex->getSourceRange(),
2414        getLocationOfByte(CS.getStart()),
2415        /*IsStringLocation*/true,
2416        getSpecifierRange(startSpecifier, specifierLen),
2417        FixItHint::CreateReplacement(
2418          getSpecifierRange(startSpecifier, specifierLen),
2419          os.str()));
2420    }
2421    else {
2422      EmitFormatDiagnostic(
2423        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2424          << ATR.getRepresentativeTypeName(S.Context) << Ex->getType()
2425          << getSpecifierRange(startSpecifier, specifierLen)
2426          << Ex->getSourceRange(),
2427        getLocationOfByte(CS.getStart()),
2428        true,
2429        getSpecifierRange(startSpecifier, specifierLen));
2430    }
2431  }
2432
2433  return true;
2434}
2435
2436//===--- CHECK: Scanf format string checking ------------------------------===//
2437
2438namespace {
2439class CheckScanfHandler : public CheckFormatHandler {
2440public:
2441  CheckScanfHandler(Sema &s, const StringLiteral *fexpr,
2442                    const Expr *origFormatExpr, unsigned firstDataArg,
2443                    unsigned numDataArgs, bool isObjCLiteral,
2444                    const char *beg, bool hasVAListArg,
2445                    Expr **Args, unsigned NumArgs,
2446                    unsigned formatIdx, bool inFunctionCall)
2447  : CheckFormatHandler(s, fexpr, origFormatExpr, firstDataArg,
2448                       numDataArgs, isObjCLiteral, beg, hasVAListArg,
2449                       Args, NumArgs, formatIdx, inFunctionCall) {}
2450
2451  bool HandleScanfSpecifier(const analyze_scanf::ScanfSpecifier &FS,
2452                            const char *startSpecifier,
2453                            unsigned specifierLen);
2454
2455  bool HandleInvalidScanfConversionSpecifier(
2456          const analyze_scanf::ScanfSpecifier &FS,
2457          const char *startSpecifier,
2458          unsigned specifierLen);
2459
2460  void HandleIncompleteScanList(const char *start, const char *end);
2461};
2462}
2463
2464void CheckScanfHandler::HandleIncompleteScanList(const char *start,
2465                                                 const char *end) {
2466  EmitFormatDiagnostic(S.PDiag(diag::warn_scanf_scanlist_incomplete),
2467                       getLocationOfByte(end), /*IsStringLocation*/true,
2468                       getSpecifierRange(start, end - start));
2469}
2470
2471bool CheckScanfHandler::HandleInvalidScanfConversionSpecifier(
2472                                        const analyze_scanf::ScanfSpecifier &FS,
2473                                        const char *startSpecifier,
2474                                        unsigned specifierLen) {
2475
2476  const analyze_scanf::ScanfConversionSpecifier &CS =
2477    FS.getConversionSpecifier();
2478
2479  return HandleInvalidConversionSpecifier(FS.getArgIndex(),
2480                                          getLocationOfByte(CS.getStart()),
2481                                          startSpecifier, specifierLen,
2482                                          CS.getStart(), CS.getLength());
2483}
2484
2485bool CheckScanfHandler::HandleScanfSpecifier(
2486                                       const analyze_scanf::ScanfSpecifier &FS,
2487                                       const char *startSpecifier,
2488                                       unsigned specifierLen) {
2489
2490  using namespace analyze_scanf;
2491  using namespace analyze_format_string;
2492
2493  const ScanfConversionSpecifier &CS = FS.getConversionSpecifier();
2494
2495  // Handle case where '%' and '*' don't consume an argument.  These shouldn't
2496  // be used to decide if we are using positional arguments consistently.
2497  if (FS.consumesDataArgument()) {
2498    if (atFirstArg) {
2499      atFirstArg = false;
2500      usesPositionalArgs = FS.usesPositionalArg();
2501    }
2502    else if (usesPositionalArgs != FS.usesPositionalArg()) {
2503      HandlePositionalNonpositionalArgs(getLocationOfByte(CS.getStart()),
2504                                        startSpecifier, specifierLen);
2505      return false;
2506    }
2507  }
2508
2509  // Check if the field with is non-zero.
2510  const OptionalAmount &Amt = FS.getFieldWidth();
2511  if (Amt.getHowSpecified() == OptionalAmount::Constant) {
2512    if (Amt.getConstantAmount() == 0) {
2513      const CharSourceRange &R = getSpecifierRange(Amt.getStart(),
2514                                                   Amt.getConstantLength());
2515      EmitFormatDiagnostic(S.PDiag(diag::warn_scanf_nonzero_width),
2516                           getLocationOfByte(Amt.getStart()),
2517                           /*IsStringLocation*/true, R,
2518                           FixItHint::CreateRemoval(R));
2519    }
2520  }
2521
2522  if (!FS.consumesDataArgument()) {
2523    // FIXME: Technically specifying a precision or field width here
2524    // makes no sense.  Worth issuing a warning at some point.
2525    return true;
2526  }
2527
2528  // Consume the argument.
2529  unsigned argIndex = FS.getArgIndex();
2530  if (argIndex < NumDataArgs) {
2531      // The check to see if the argIndex is valid will come later.
2532      // We set the bit here because we may exit early from this
2533      // function if we encounter some other error.
2534    CoveredArgs.set(argIndex);
2535  }
2536
2537  // Check the length modifier is valid with the given conversion specifier.
2538  const LengthModifier &LM = FS.getLengthModifier();
2539  if (!FS.hasValidLengthModifier()) {
2540    const CharSourceRange &R = getSpecifierRange(LM.getStart(), LM.getLength());
2541    EmitFormatDiagnostic(S.PDiag(diag::warn_format_nonsensical_length)
2542                         << LM.toString() << CS.toString()
2543                         << getSpecifierRange(startSpecifier, specifierLen),
2544                         getLocationOfByte(LM.getStart()),
2545                         /*IsStringLocation*/true, R,
2546                         FixItHint::CreateRemoval(R));
2547  }
2548
2549  if (!FS.hasStandardLengthModifier())
2550    HandleNonStandardLengthModifier(LM, startSpecifier, specifierLen);
2551  if (!FS.hasStandardConversionSpecifier(S.getLangOpts()))
2552    HandleNonStandardConversionSpecifier(CS, startSpecifier, specifierLen);
2553  if (!FS.hasStandardLengthConversionCombination())
2554    HandleNonStandardConversionSpecification(LM, CS, startSpecifier,
2555                                             specifierLen);
2556
2557  // The remaining checks depend on the data arguments.
2558  if (HasVAListArg)
2559    return true;
2560
2561  if (!CheckNumArgs(FS, CS, startSpecifier, specifierLen, argIndex))
2562    return false;
2563
2564  // Check that the argument type matches the format specifier.
2565  const Expr *Ex = getDataArg(argIndex);
2566  const analyze_scanf::ScanfArgTypeResult &ATR = FS.getArgType(S.Context);
2567  if (ATR.isValid() && !ATR.matchesType(S.Context, Ex->getType())) {
2568    ScanfSpecifier fixedFS = FS;
2569    bool success = fixedFS.fixType(Ex->getType(), S.getLangOpts(),
2570                                   S.Context);
2571
2572    if (success) {
2573      // Get the fix string from the fixed format specifier.
2574      SmallString<128> buf;
2575      llvm::raw_svector_ostream os(buf);
2576      fixedFS.toString(os);
2577
2578      EmitFormatDiagnostic(
2579        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2580          << ATR.getRepresentativeTypeName(S.Context) << Ex->getType()
2581          << Ex->getSourceRange(),
2582        getLocationOfByte(CS.getStart()),
2583        /*IsStringLocation*/true,
2584        getSpecifierRange(startSpecifier, specifierLen),
2585        FixItHint::CreateReplacement(
2586          getSpecifierRange(startSpecifier, specifierLen),
2587          os.str()));
2588    } else {
2589      EmitFormatDiagnostic(
2590        S.PDiag(diag::warn_printf_conversion_argument_type_mismatch)
2591          << ATR.getRepresentativeTypeName(S.Context) << Ex->getType()
2592          << Ex->getSourceRange(),
2593        getLocationOfByte(CS.getStart()),
2594        /*IsStringLocation*/true,
2595        getSpecifierRange(startSpecifier, specifierLen));
2596    }
2597  }
2598
2599  return true;
2600}
2601
2602void Sema::CheckFormatString(const StringLiteral *FExpr,
2603                             const Expr *OrigFormatExpr,
2604                             Expr **Args, unsigned NumArgs,
2605                             bool HasVAListArg, unsigned format_idx,
2606                             unsigned firstDataArg, FormatStringType Type,
2607                             bool inFunctionCall) {
2608
2609  // CHECK: is the format string a wide literal?
2610  if (!FExpr->isAscii()) {
2611    CheckFormatHandler::EmitFormatDiagnostic(
2612      *this, inFunctionCall, Args[format_idx],
2613      PDiag(diag::warn_format_string_is_wide_literal), FExpr->getLocStart(),
2614      /*IsStringLocation*/true, OrigFormatExpr->getSourceRange());
2615    return;
2616  }
2617
2618  // Str - The format string.  NOTE: this is NOT null-terminated!
2619  StringRef StrRef = FExpr->getString();
2620  const char *Str = StrRef.data();
2621  unsigned StrLen = StrRef.size();
2622  const unsigned numDataArgs = NumArgs - firstDataArg;
2623
2624  // CHECK: empty format string?
2625  if (StrLen == 0 && numDataArgs > 0) {
2626    CheckFormatHandler::EmitFormatDiagnostic(
2627      *this, inFunctionCall, Args[format_idx],
2628      PDiag(diag::warn_empty_format_string), FExpr->getLocStart(),
2629      /*IsStringLocation*/true, OrigFormatExpr->getSourceRange());
2630    return;
2631  }
2632
2633  if (Type == FST_Printf || Type == FST_NSString) {
2634    CheckPrintfHandler H(*this, FExpr, OrigFormatExpr, firstDataArg,
2635                         numDataArgs, isa<ObjCStringLiteral>(OrigFormatExpr),
2636                         Str, HasVAListArg, Args, NumArgs, format_idx,
2637                         inFunctionCall);
2638
2639    if (!analyze_format_string::ParsePrintfString(H, Str, Str + StrLen,
2640                                                  getLangOpts()))
2641      H.DoneProcessing();
2642  } else if (Type == FST_Scanf) {
2643    CheckScanfHandler H(*this, FExpr, OrigFormatExpr, firstDataArg,
2644                        numDataArgs, isa<ObjCStringLiteral>(OrigFormatExpr),
2645                        Str, HasVAListArg, Args, NumArgs, format_idx,
2646                        inFunctionCall);
2647
2648    if (!analyze_format_string::ParseScanfString(H, Str, Str + StrLen,
2649                                                 getLangOpts()))
2650      H.DoneProcessing();
2651  } // TODO: handle other formats
2652}
2653
2654//===--- CHECK: Standard memory functions ---------------------------------===//
2655
2656/// \brief Determine whether the given type is a dynamic class type (e.g.,
2657/// whether it has a vtable).
2658static bool isDynamicClassType(QualType T) {
2659  if (CXXRecordDecl *Record = T->getAsCXXRecordDecl())
2660    if (CXXRecordDecl *Definition = Record->getDefinition())
2661      if (Definition->isDynamicClass())
2662        return true;
2663
2664  return false;
2665}
2666
2667/// \brief If E is a sizeof expression, returns its argument expression,
2668/// otherwise returns NULL.
2669static const Expr *getSizeOfExprArg(const Expr* E) {
2670  if (const UnaryExprOrTypeTraitExpr *SizeOf =
2671      dyn_cast<UnaryExprOrTypeTraitExpr>(E))
2672    if (SizeOf->getKind() == clang::UETT_SizeOf && !SizeOf->isArgumentType())
2673      return SizeOf->getArgumentExpr()->IgnoreParenImpCasts();
2674
2675  return 0;
2676}
2677
2678/// \brief If E is a sizeof expression, returns its argument type.
2679static QualType getSizeOfArgType(const Expr* E) {
2680  if (const UnaryExprOrTypeTraitExpr *SizeOf =
2681      dyn_cast<UnaryExprOrTypeTraitExpr>(E))
2682    if (SizeOf->getKind() == clang::UETT_SizeOf)
2683      return SizeOf->getTypeOfArgument();
2684
2685  return QualType();
2686}
2687
2688/// \brief Check for dangerous or invalid arguments to memset().
2689///
2690/// This issues warnings on known problematic, dangerous or unspecified
2691/// arguments to the standard 'memset', 'memcpy', 'memmove', and 'memcmp'
2692/// function calls.
2693///
2694/// \param Call The call expression to diagnose.
2695void Sema::CheckMemaccessArguments(const CallExpr *Call,
2696                                   unsigned BId,
2697                                   IdentifierInfo *FnName) {
2698  assert(BId != 0);
2699
2700  // It is possible to have a non-standard definition of memset.  Validate
2701  // we have enough arguments, and if not, abort further checking.
2702  unsigned ExpectedNumArgs = (BId == Builtin::BIstrndup ? 2 : 3);
2703  if (Call->getNumArgs() < ExpectedNumArgs)
2704    return;
2705
2706  unsigned LastArg = (BId == Builtin::BImemset ||
2707                      BId == Builtin::BIstrndup ? 1 : 2);
2708  unsigned LenArg = (BId == Builtin::BIstrndup ? 1 : 2);
2709  const Expr *LenExpr = Call->getArg(LenArg)->IgnoreParenImpCasts();
2710
2711  // We have special checking when the length is a sizeof expression.
2712  QualType SizeOfArgTy = getSizeOfArgType(LenExpr);
2713  const Expr *SizeOfArg = getSizeOfExprArg(LenExpr);
2714  llvm::FoldingSetNodeID SizeOfArgID;
2715
2716  for (unsigned ArgIdx = 0; ArgIdx != LastArg; ++ArgIdx) {
2717    const Expr *Dest = Call->getArg(ArgIdx)->IgnoreParenImpCasts();
2718    SourceRange ArgRange = Call->getArg(ArgIdx)->getSourceRange();
2719
2720    QualType DestTy = Dest->getType();
2721    if (const PointerType *DestPtrTy = DestTy->getAs<PointerType>()) {
2722      QualType PointeeTy = DestPtrTy->getPointeeType();
2723
2724      // Never warn about void type pointers. This can be used to suppress
2725      // false positives.
2726      if (PointeeTy->isVoidType())
2727        continue;
2728
2729      // Catch "memset(p, 0, sizeof(p))" -- needs to be sizeof(*p). Do this by
2730      // actually comparing the expressions for equality. Because computing the
2731      // expression IDs can be expensive, we only do this if the diagnostic is
2732      // enabled.
2733      if (SizeOfArg &&
2734          Diags.getDiagnosticLevel(diag::warn_sizeof_pointer_expr_memaccess,
2735                                   SizeOfArg->getExprLoc())) {
2736        // We only compute IDs for expressions if the warning is enabled, and
2737        // cache the sizeof arg's ID.
2738        if (SizeOfArgID == llvm::FoldingSetNodeID())
2739          SizeOfArg->Profile(SizeOfArgID, Context, true);
2740        llvm::FoldingSetNodeID DestID;
2741        Dest->Profile(DestID, Context, true);
2742        if (DestID == SizeOfArgID) {
2743          // TODO: For strncpy() and friends, this could suggest sizeof(dst)
2744          //       over sizeof(src) as well.
2745          unsigned ActionIdx = 0; // Default is to suggest dereferencing.
2746          if (const UnaryOperator *UnaryOp = dyn_cast<UnaryOperator>(Dest))
2747            if (UnaryOp->getOpcode() == UO_AddrOf)
2748              ActionIdx = 1; // If its an address-of operator, just remove it.
2749          if (Context.getTypeSize(PointeeTy) == Context.getCharWidth())
2750            ActionIdx = 2; // If the pointee's size is sizeof(char),
2751                           // suggest an explicit length.
2752          unsigned DestSrcSelect =
2753            (BId == Builtin::BIstrndup ? 1 : ArgIdx);
2754          DiagRuntimeBehavior(SizeOfArg->getExprLoc(), Dest,
2755                              PDiag(diag::warn_sizeof_pointer_expr_memaccess)
2756                                << FnName << DestSrcSelect << ActionIdx
2757                                << Dest->getSourceRange()
2758                                << SizeOfArg->getSourceRange());
2759          break;
2760        }
2761      }
2762
2763      // Also check for cases where the sizeof argument is the exact same
2764      // type as the memory argument, and where it points to a user-defined
2765      // record type.
2766      if (SizeOfArgTy != QualType()) {
2767        if (PointeeTy->isRecordType() &&
2768            Context.typesAreCompatible(SizeOfArgTy, DestTy)) {
2769          DiagRuntimeBehavior(LenExpr->getExprLoc(), Dest,
2770                              PDiag(diag::warn_sizeof_pointer_type_memaccess)
2771                                << FnName << SizeOfArgTy << ArgIdx
2772                                << PointeeTy << Dest->getSourceRange()
2773                                << LenExpr->getSourceRange());
2774          break;
2775        }
2776      }
2777
2778      // Always complain about dynamic classes.
2779      if (isDynamicClassType(PointeeTy)) {
2780
2781        unsigned OperationType = 0;
2782        // "overwritten" if we're warning about the destination for any call
2783        // but memcmp; otherwise a verb appropriate to the call.
2784        if (ArgIdx != 0 || BId == Builtin::BImemcmp) {
2785          if (BId == Builtin::BImemcpy)
2786            OperationType = 1;
2787          else if(BId == Builtin::BImemmove)
2788            OperationType = 2;
2789          else if (BId == Builtin::BImemcmp)
2790            OperationType = 3;
2791        }
2792
2793        DiagRuntimeBehavior(
2794          Dest->getExprLoc(), Dest,
2795          PDiag(diag::warn_dyn_class_memaccess)
2796            << (BId == Builtin::BImemcmp ? ArgIdx + 2 : ArgIdx)
2797            << FnName << PointeeTy
2798            << OperationType
2799            << Call->getCallee()->getSourceRange());
2800      } else if (PointeeTy.hasNonTrivialObjCLifetime() &&
2801               BId != Builtin::BImemset)
2802        DiagRuntimeBehavior(
2803          Dest->getExprLoc(), Dest,
2804          PDiag(diag::warn_arc_object_memaccess)
2805            << ArgIdx << FnName << PointeeTy
2806            << Call->getCallee()->getSourceRange());
2807      else
2808        continue;
2809
2810      DiagRuntimeBehavior(
2811        Dest->getExprLoc(), Dest,
2812        PDiag(diag::note_bad_memaccess_silence)
2813          << FixItHint::CreateInsertion(ArgRange.getBegin(), "(void*)"));
2814      break;
2815    }
2816  }
2817}
2818
2819// A little helper routine: ignore addition and subtraction of integer literals.
2820// This intentionally does not ignore all integer constant expressions because
2821// we don't want to remove sizeof().
2822static const Expr *ignoreLiteralAdditions(const Expr *Ex, ASTContext &Ctx) {
2823  Ex = Ex->IgnoreParenCasts();
2824
2825  for (;;) {
2826    const BinaryOperator * BO = dyn_cast<BinaryOperator>(Ex);
2827    if (!BO || !BO->isAdditiveOp())
2828      break;
2829
2830    const Expr *RHS = BO->getRHS()->IgnoreParenCasts();
2831    const Expr *LHS = BO->getLHS()->IgnoreParenCasts();
2832
2833    if (isa<IntegerLiteral>(RHS))
2834      Ex = LHS;
2835    else if (isa<IntegerLiteral>(LHS))
2836      Ex = RHS;
2837    else
2838      break;
2839  }
2840
2841  return Ex;
2842}
2843
2844// Warn if the user has made the 'size' argument to strlcpy or strlcat
2845// be the size of the source, instead of the destination.
2846void Sema::CheckStrlcpycatArguments(const CallExpr *Call,
2847                                    IdentifierInfo *FnName) {
2848
2849  // Don't crash if the user has the wrong number of arguments
2850  if (Call->getNumArgs() != 3)
2851    return;
2852
2853  const Expr *SrcArg = ignoreLiteralAdditions(Call->getArg(1), Context);
2854  const Expr *SizeArg = ignoreLiteralAdditions(Call->getArg(2), Context);
2855  const Expr *CompareWithSrc = NULL;
2856
2857  // Look for 'strlcpy(dst, x, sizeof(x))'
2858  if (const Expr *Ex = getSizeOfExprArg(SizeArg))
2859    CompareWithSrc = Ex;
2860  else {
2861    // Look for 'strlcpy(dst, x, strlen(x))'
2862    if (const CallExpr *SizeCall = dyn_cast<CallExpr>(SizeArg)) {
2863      if (SizeCall->isBuiltinCall() == Builtin::BIstrlen
2864          && SizeCall->getNumArgs() == 1)
2865        CompareWithSrc = ignoreLiteralAdditions(SizeCall->getArg(0), Context);
2866    }
2867  }
2868
2869  if (!CompareWithSrc)
2870    return;
2871
2872  // Determine if the argument to sizeof/strlen is equal to the source
2873  // argument.  In principle there's all kinds of things you could do
2874  // here, for instance creating an == expression and evaluating it with
2875  // EvaluateAsBooleanCondition, but this uses a more direct technique:
2876  const DeclRefExpr *SrcArgDRE = dyn_cast<DeclRefExpr>(SrcArg);
2877  if (!SrcArgDRE)
2878    return;
2879
2880  const DeclRefExpr *CompareWithSrcDRE = dyn_cast<DeclRefExpr>(CompareWithSrc);
2881  if (!CompareWithSrcDRE ||
2882      SrcArgDRE->getDecl() != CompareWithSrcDRE->getDecl())
2883    return;
2884
2885  const Expr *OriginalSizeArg = Call->getArg(2);
2886  Diag(CompareWithSrcDRE->getLocStart(), diag::warn_strlcpycat_wrong_size)
2887    << OriginalSizeArg->getSourceRange() << FnName;
2888
2889  // Output a FIXIT hint if the destination is an array (rather than a
2890  // pointer to an array).  This could be enhanced to handle some
2891  // pointers if we know the actual size, like if DstArg is 'array+2'
2892  // we could say 'sizeof(array)-2'.
2893  const Expr *DstArg = Call->getArg(0)->IgnoreParenImpCasts();
2894  QualType DstArgTy = DstArg->getType();
2895
2896  // Only handle constant-sized or VLAs, but not flexible members.
2897  if (const ConstantArrayType *CAT = Context.getAsConstantArrayType(DstArgTy)) {
2898    // Only issue the FIXIT for arrays of size > 1.
2899    if (CAT->getSize().getSExtValue() <= 1)
2900      return;
2901  } else if (!DstArgTy->isVariableArrayType()) {
2902    return;
2903  }
2904
2905  SmallString<128> sizeString;
2906  llvm::raw_svector_ostream OS(sizeString);
2907  OS << "sizeof(";
2908  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
2909  OS << ")";
2910
2911  Diag(OriginalSizeArg->getLocStart(), diag::note_strlcpycat_wrong_size)
2912    << FixItHint::CreateReplacement(OriginalSizeArg->getSourceRange(),
2913                                    OS.str());
2914}
2915
2916/// Check if two expressions refer to the same declaration.
2917static bool referToTheSameDecl(const Expr *E1, const Expr *E2) {
2918  if (const DeclRefExpr *D1 = dyn_cast_or_null<DeclRefExpr>(E1))
2919    if (const DeclRefExpr *D2 = dyn_cast_or_null<DeclRefExpr>(E2))
2920      return D1->getDecl() == D2->getDecl();
2921  return false;
2922}
2923
2924static const Expr *getStrlenExprArg(const Expr *E) {
2925  if (const CallExpr *CE = dyn_cast<CallExpr>(E)) {
2926    const FunctionDecl *FD = CE->getDirectCallee();
2927    if (!FD || FD->getMemoryFunctionKind() != Builtin::BIstrlen)
2928      return 0;
2929    return CE->getArg(0)->IgnoreParenCasts();
2930  }
2931  return 0;
2932}
2933
2934// Warn on anti-patterns as the 'size' argument to strncat.
2935// The correct size argument should look like following:
2936//   strncat(dst, src, sizeof(dst) - strlen(dest) - 1);
2937void Sema::CheckStrncatArguments(const CallExpr *CE,
2938                                 IdentifierInfo *FnName) {
2939  // Don't crash if the user has the wrong number of arguments.
2940  if (CE->getNumArgs() < 3)
2941    return;
2942  const Expr *DstArg = CE->getArg(0)->IgnoreParenCasts();
2943  const Expr *SrcArg = CE->getArg(1)->IgnoreParenCasts();
2944  const Expr *LenArg = CE->getArg(2)->IgnoreParenCasts();
2945
2946  // Identify common expressions, which are wrongly used as the size argument
2947  // to strncat and may lead to buffer overflows.
2948  unsigned PatternType = 0;
2949  if (const Expr *SizeOfArg = getSizeOfExprArg(LenArg)) {
2950    // - sizeof(dst)
2951    if (referToTheSameDecl(SizeOfArg, DstArg))
2952      PatternType = 1;
2953    // - sizeof(src)
2954    else if (referToTheSameDecl(SizeOfArg, SrcArg))
2955      PatternType = 2;
2956  } else if (const BinaryOperator *BE = dyn_cast<BinaryOperator>(LenArg)) {
2957    if (BE->getOpcode() == BO_Sub) {
2958      const Expr *L = BE->getLHS()->IgnoreParenCasts();
2959      const Expr *R = BE->getRHS()->IgnoreParenCasts();
2960      // - sizeof(dst) - strlen(dst)
2961      if (referToTheSameDecl(DstArg, getSizeOfExprArg(L)) &&
2962          referToTheSameDecl(DstArg, getStrlenExprArg(R)))
2963        PatternType = 1;
2964      // - sizeof(src) - (anything)
2965      else if (referToTheSameDecl(SrcArg, getSizeOfExprArg(L)))
2966        PatternType = 2;
2967    }
2968  }
2969
2970  if (PatternType == 0)
2971    return;
2972
2973  // Generate the diagnostic.
2974  SourceLocation SL = LenArg->getLocStart();
2975  SourceRange SR = LenArg->getSourceRange();
2976  SourceManager &SM  = PP.getSourceManager();
2977
2978  // If the function is defined as a builtin macro, do not show macro expansion.
2979  if (SM.isMacroArgExpansion(SL)) {
2980    SL = SM.getSpellingLoc(SL);
2981    SR = SourceRange(SM.getSpellingLoc(SR.getBegin()),
2982                     SM.getSpellingLoc(SR.getEnd()));
2983  }
2984
2985  if (PatternType == 1)
2986    Diag(SL, diag::warn_strncat_large_size) << SR;
2987  else
2988    Diag(SL, diag::warn_strncat_src_size) << SR;
2989
2990  // Output a FIXIT hint if the destination is an array (rather than a
2991  // pointer to an array).  This could be enhanced to handle some
2992  // pointers if we know the actual size, like if DstArg is 'array+2'
2993  // we could say 'sizeof(array)-2'.
2994  QualType DstArgTy = DstArg->getType();
2995
2996  // Only handle constant-sized or VLAs, but not flexible members.
2997  if (const ConstantArrayType *CAT = Context.getAsConstantArrayType(DstArgTy)) {
2998    // Only issue the FIXIT for arrays of size > 1.
2999    if (CAT->getSize().getSExtValue() <= 1)
3000      return;
3001  } else if (!DstArgTy->isVariableArrayType()) {
3002    return;
3003  }
3004
3005  SmallString<128> sizeString;
3006  llvm::raw_svector_ostream OS(sizeString);
3007  OS << "sizeof(";
3008  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3009  OS << ") - ";
3010  OS << "strlen(";
3011  DstArg->printPretty(OS, Context, 0, getPrintingPolicy());
3012  OS << ") - 1";
3013
3014  Diag(SL, diag::note_strncat_wrong_size)
3015    << FixItHint::CreateReplacement(SR, OS.str());
3016}
3017
3018//===--- CHECK: Return Address of Stack Variable --------------------------===//
3019
3020static Expr *EvalVal(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars);
3021static Expr *EvalAddr(Expr* E, SmallVectorImpl<DeclRefExpr *> &refVars);
3022
3023/// CheckReturnStackAddr - Check if a return statement returns the address
3024///   of a stack variable.
3025void
3026Sema::CheckReturnStackAddr(Expr *RetValExp, QualType lhsType,
3027                           SourceLocation ReturnLoc) {
3028
3029  Expr *stackE = 0;
3030  SmallVector<DeclRefExpr *, 8> refVars;
3031
3032  // Perform checking for returned stack addresses, local blocks,
3033  // label addresses or references to temporaries.
3034  if (lhsType->isPointerType() ||
3035      (!getLangOpts().ObjCAutoRefCount && lhsType->isBlockPointerType())) {
3036    stackE = EvalAddr(RetValExp, refVars);
3037  } else if (lhsType->isReferenceType()) {
3038    stackE = EvalVal(RetValExp, refVars);
3039  }
3040
3041  if (stackE == 0)
3042    return; // Nothing suspicious was found.
3043
3044  SourceLocation diagLoc;
3045  SourceRange diagRange;
3046  if (refVars.empty()) {
3047    diagLoc = stackE->getLocStart();
3048    diagRange = stackE->getSourceRange();
3049  } else {
3050    // We followed through a reference variable. 'stackE' contains the
3051    // problematic expression but we will warn at the return statement pointing
3052    // at the reference variable. We will later display the "trail" of
3053    // reference variables using notes.
3054    diagLoc = refVars[0]->getLocStart();
3055    diagRange = refVars[0]->getSourceRange();
3056  }
3057
3058  if (DeclRefExpr *DR = dyn_cast<DeclRefExpr>(stackE)) { //address of local var.
3059    Diag(diagLoc, lhsType->isReferenceType() ? diag::warn_ret_stack_ref
3060                                             : diag::warn_ret_stack_addr)
3061     << DR->getDecl()->getDeclName() << diagRange;
3062  } else if (isa<BlockExpr>(stackE)) { // local block.
3063    Diag(diagLoc, diag::err_ret_local_block) << diagRange;
3064  } else if (isa<AddrLabelExpr>(stackE)) { // address of label.
3065    Diag(diagLoc, diag::warn_ret_addr_label) << diagRange;
3066  } else { // local temporary.
3067    Diag(diagLoc, lhsType->isReferenceType() ? diag::warn_ret_local_temp_ref
3068                                             : diag::warn_ret_local_temp_addr)
3069     << diagRange;
3070  }
3071
3072  // Display the "trail" of reference variables that we followed until we
3073  // found the problematic expression using notes.
3074  for (unsigned i = 0, e = refVars.size(); i != e; ++i) {
3075    VarDecl *VD = cast<VarDecl>(refVars[i]->getDecl());
3076    // If this var binds to another reference var, show the range of the next
3077    // var, otherwise the var binds to the problematic expression, in which case
3078    // show the range of the expression.
3079    SourceRange range = (i < e-1) ? refVars[i+1]->getSourceRange()
3080                                  : stackE->getSourceRange();
3081    Diag(VD->getLocation(), diag::note_ref_var_local_bind)
3082      << VD->getDeclName() << range;
3083  }
3084}
3085
3086/// EvalAddr - EvalAddr and EvalVal are mutually recursive functions that
3087///  check if the expression in a return statement evaluates to an address
3088///  to a location on the stack, a local block, an address of a label, or a
3089///  reference to local temporary. The recursion is used to traverse the
3090///  AST of the return expression, with recursion backtracking when we
3091///  encounter a subexpression that (1) clearly does not lead to one of the
3092///  above problematic expressions (2) is something we cannot determine leads to
3093///  a problematic expression based on such local checking.
3094///
3095///  Both EvalAddr and EvalVal follow through reference variables to evaluate
3096///  the expression that they point to. Such variables are added to the
3097///  'refVars' vector so that we know what the reference variable "trail" was.
3098///
3099///  EvalAddr processes expressions that are pointers that are used as
3100///  references (and not L-values).  EvalVal handles all other values.
3101///  At the base case of the recursion is a check for the above problematic
3102///  expressions.
3103///
3104///  This implementation handles:
3105///
3106///   * pointer-to-pointer casts
3107///   * implicit conversions from array references to pointers
3108///   * taking the address of fields
3109///   * arbitrary interplay between "&" and "*" operators
3110///   * pointer arithmetic from an address of a stack variable
3111///   * taking the address of an array element where the array is on the stack
3112static Expr *EvalAddr(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars) {
3113  if (E->isTypeDependent())
3114      return NULL;
3115
3116  // We should only be called for evaluating pointer expressions.
3117  assert((E->getType()->isAnyPointerType() ||
3118          E->getType()->isBlockPointerType() ||
3119          E->getType()->isObjCQualifiedIdType()) &&
3120         "EvalAddr only works on pointers");
3121
3122  E = E->IgnoreParens();
3123
3124  // Our "symbolic interpreter" is just a dispatch off the currently
3125  // viewed AST node.  We then recursively traverse the AST by calling
3126  // EvalAddr and EvalVal appropriately.
3127  switch (E->getStmtClass()) {
3128  case Stmt::DeclRefExprClass: {
3129    DeclRefExpr *DR = cast<DeclRefExpr>(E);
3130
3131    if (VarDecl *V = dyn_cast<VarDecl>(DR->getDecl()))
3132      // If this is a reference variable, follow through to the expression that
3133      // it points to.
3134      if (V->hasLocalStorage() &&
3135          V->getType()->isReferenceType() && V->hasInit()) {
3136        // Add the reference variable to the "trail".
3137        refVars.push_back(DR);
3138        return EvalAddr(V->getInit(), refVars);
3139      }
3140
3141    return NULL;
3142  }
3143
3144  case Stmt::UnaryOperatorClass: {
3145    // The only unary operator that make sense to handle here
3146    // is AddrOf.  All others don't make sense as pointers.
3147    UnaryOperator *U = cast<UnaryOperator>(E);
3148
3149    if (U->getOpcode() == UO_AddrOf)
3150      return EvalVal(U->getSubExpr(), refVars);
3151    else
3152      return NULL;
3153  }
3154
3155  case Stmt::BinaryOperatorClass: {
3156    // Handle pointer arithmetic.  All other binary operators are not valid
3157    // in this context.
3158    BinaryOperator *B = cast<BinaryOperator>(E);
3159    BinaryOperatorKind op = B->getOpcode();
3160
3161    if (op != BO_Add && op != BO_Sub)
3162      return NULL;
3163
3164    Expr *Base = B->getLHS();
3165
3166    // Determine which argument is the real pointer base.  It could be
3167    // the RHS argument instead of the LHS.
3168    if (!Base->getType()->isPointerType()) Base = B->getRHS();
3169
3170    assert (Base->getType()->isPointerType());
3171    return EvalAddr(Base, refVars);
3172  }
3173
3174  // For conditional operators we need to see if either the LHS or RHS are
3175  // valid DeclRefExpr*s.  If one of them is valid, we return it.
3176  case Stmt::ConditionalOperatorClass: {
3177    ConditionalOperator *C = cast<ConditionalOperator>(E);
3178
3179    // Handle the GNU extension for missing LHS.
3180    if (Expr *lhsExpr = C->getLHS()) {
3181    // In C++, we can have a throw-expression, which has 'void' type.
3182      if (!lhsExpr->getType()->isVoidType())
3183        if (Expr* LHS = EvalAddr(lhsExpr, refVars))
3184          return LHS;
3185    }
3186
3187    // In C++, we can have a throw-expression, which has 'void' type.
3188    if (C->getRHS()->getType()->isVoidType())
3189      return NULL;
3190
3191    return EvalAddr(C->getRHS(), refVars);
3192  }
3193
3194  case Stmt::BlockExprClass:
3195    if (cast<BlockExpr>(E)->getBlockDecl()->hasCaptures())
3196      return E; // local block.
3197    return NULL;
3198
3199  case Stmt::AddrLabelExprClass:
3200    return E; // address of label.
3201
3202  case Stmt::ExprWithCleanupsClass:
3203    return EvalAddr(cast<ExprWithCleanups>(E)->getSubExpr(), refVars);
3204
3205  // For casts, we need to handle conversions from arrays to
3206  // pointer values, and pointer-to-pointer conversions.
3207  case Stmt::ImplicitCastExprClass:
3208  case Stmt::CStyleCastExprClass:
3209  case Stmt::CXXFunctionalCastExprClass:
3210  case Stmt::ObjCBridgedCastExprClass:
3211  case Stmt::CXXStaticCastExprClass:
3212  case Stmt::CXXDynamicCastExprClass:
3213  case Stmt::CXXConstCastExprClass:
3214  case Stmt::CXXReinterpretCastExprClass: {
3215    Expr* SubExpr = cast<CastExpr>(E)->getSubExpr();
3216    switch (cast<CastExpr>(E)->getCastKind()) {
3217    case CK_BitCast:
3218    case CK_LValueToRValue:
3219    case CK_NoOp:
3220    case CK_BaseToDerived:
3221    case CK_DerivedToBase:
3222    case CK_UncheckedDerivedToBase:
3223    case CK_Dynamic:
3224    case CK_CPointerToObjCPointerCast:
3225    case CK_BlockPointerToObjCPointerCast:
3226    case CK_AnyPointerToBlockPointerCast:
3227      return EvalAddr(SubExpr, refVars);
3228
3229    case CK_ArrayToPointerDecay:
3230      return EvalVal(SubExpr, refVars);
3231
3232    default:
3233      return 0;
3234    }
3235  }
3236
3237  case Stmt::MaterializeTemporaryExprClass:
3238    if (Expr *Result = EvalAddr(
3239                         cast<MaterializeTemporaryExpr>(E)->GetTemporaryExpr(),
3240                                refVars))
3241      return Result;
3242
3243    return E;
3244
3245  // Everything else: we simply don't reason about them.
3246  default:
3247    return NULL;
3248  }
3249}
3250
3251
3252///  EvalVal - This function is complements EvalAddr in the mutual recursion.
3253///   See the comments for EvalAddr for more details.
3254static Expr *EvalVal(Expr *E, SmallVectorImpl<DeclRefExpr *> &refVars) {
3255do {
3256  // We should only be called for evaluating non-pointer expressions, or
3257  // expressions with a pointer type that are not used as references but instead
3258  // are l-values (e.g., DeclRefExpr with a pointer type).
3259
3260  // Our "symbolic interpreter" is just a dispatch off the currently
3261  // viewed AST node.  We then recursively traverse the AST by calling
3262  // EvalAddr and EvalVal appropriately.
3263
3264  E = E->IgnoreParens();
3265  switch (E->getStmtClass()) {
3266  case Stmt::ImplicitCastExprClass: {
3267    ImplicitCastExpr *IE = cast<ImplicitCastExpr>(E);
3268    if (IE->getValueKind() == VK_LValue) {
3269      E = IE->getSubExpr();
3270      continue;
3271    }
3272    return NULL;
3273  }
3274
3275  case Stmt::ExprWithCleanupsClass:
3276    return EvalVal(cast<ExprWithCleanups>(E)->getSubExpr(), refVars);
3277
3278  case Stmt::DeclRefExprClass: {
3279    // When we hit a DeclRefExpr we are looking at code that refers to a
3280    // variable's name. If it's not a reference variable we check if it has
3281    // local storage within the function, and if so, return the expression.
3282    DeclRefExpr *DR = cast<DeclRefExpr>(E);
3283
3284    if (VarDecl *V = dyn_cast<VarDecl>(DR->getDecl()))
3285      if (V->hasLocalStorage()) {
3286        if (!V->getType()->isReferenceType())
3287          return DR;
3288
3289        // Reference variable, follow through to the expression that
3290        // it points to.
3291        if (V->hasInit()) {
3292          // Add the reference variable to the "trail".
3293          refVars.push_back(DR);
3294          return EvalVal(V->getInit(), refVars);
3295        }
3296      }
3297
3298    return NULL;
3299  }
3300
3301  case Stmt::UnaryOperatorClass: {
3302    // The only unary operator that make sense to handle here
3303    // is Deref.  All others don't resolve to a "name."  This includes
3304    // handling all sorts of rvalues passed to a unary operator.
3305    UnaryOperator *U = cast<UnaryOperator>(E);
3306
3307    if (U->getOpcode() == UO_Deref)
3308      return EvalAddr(U->getSubExpr(), refVars);
3309
3310    return NULL;
3311  }
3312
3313  case Stmt::ArraySubscriptExprClass: {
3314    // Array subscripts are potential references to data on the stack.  We
3315    // retrieve the DeclRefExpr* for the array variable if it indeed
3316    // has local storage.
3317    return EvalAddr(cast<ArraySubscriptExpr>(E)->getBase(), refVars);
3318  }
3319
3320  case Stmt::ConditionalOperatorClass: {
3321    // For conditional operators we need to see if either the LHS or RHS are
3322    // non-NULL Expr's.  If one is non-NULL, we return it.
3323    ConditionalOperator *C = cast<ConditionalOperator>(E);
3324
3325    // Handle the GNU extension for missing LHS.
3326    if (Expr *lhsExpr = C->getLHS())
3327      if (Expr *LHS = EvalVal(lhsExpr, refVars))
3328        return LHS;
3329
3330    return EvalVal(C->getRHS(), refVars);
3331  }
3332
3333  // Accesses to members are potential references to data on the stack.
3334  case Stmt::MemberExprClass: {
3335    MemberExpr *M = cast<MemberExpr>(E);
3336
3337    // Check for indirect access.  We only want direct field accesses.
3338    if (M->isArrow())
3339      return NULL;
3340
3341    // Check whether the member type is itself a reference, in which case
3342    // we're not going to refer to the member, but to what the member refers to.
3343    if (M->getMemberDecl()->getType()->isReferenceType())
3344      return NULL;
3345
3346    return EvalVal(M->getBase(), refVars);
3347  }
3348
3349  case Stmt::MaterializeTemporaryExprClass:
3350    if (Expr *Result = EvalVal(
3351                          cast<MaterializeTemporaryExpr>(E)->GetTemporaryExpr(),
3352                               refVars))
3353      return Result;
3354
3355    return E;
3356
3357  default:
3358    // Check that we don't return or take the address of a reference to a
3359    // temporary. This is only useful in C++.
3360    if (!E->isTypeDependent() && E->isRValue())
3361      return E;
3362
3363    // Everything else: we simply don't reason about them.
3364    return NULL;
3365  }
3366} while (true);
3367}
3368
3369//===--- CHECK: Floating-Point comparisons (-Wfloat-equal) ---------------===//
3370
3371/// Check for comparisons of floating point operands using != and ==.
3372/// Issue a warning if these are no self-comparisons, as they are not likely
3373/// to do what the programmer intended.
3374void Sema::CheckFloatComparison(SourceLocation Loc, Expr* LHS, Expr *RHS) {
3375  bool EmitWarning = true;
3376
3377  Expr* LeftExprSansParen = LHS->IgnoreParenImpCasts();
3378  Expr* RightExprSansParen = RHS->IgnoreParenImpCasts();
3379
3380  // Special case: check for x == x (which is OK).
3381  // Do not emit warnings for such cases.
3382  if (DeclRefExpr* DRL = dyn_cast<DeclRefExpr>(LeftExprSansParen))
3383    if (DeclRefExpr* DRR = dyn_cast<DeclRefExpr>(RightExprSansParen))
3384      if (DRL->getDecl() == DRR->getDecl())
3385        EmitWarning = false;
3386
3387
3388  // Special case: check for comparisons against literals that can be exactly
3389  //  represented by APFloat.  In such cases, do not emit a warning.  This
3390  //  is a heuristic: often comparison against such literals are used to
3391  //  detect if a value in a variable has not changed.  This clearly can
3392  //  lead to false negatives.
3393  if (EmitWarning) {
3394    if (FloatingLiteral* FLL = dyn_cast<FloatingLiteral>(LeftExprSansParen)) {
3395      if (FLL->isExact())
3396        EmitWarning = false;
3397    } else
3398      if (FloatingLiteral* FLR = dyn_cast<FloatingLiteral>(RightExprSansParen)){
3399        if (FLR->isExact())
3400          EmitWarning = false;
3401    }
3402  }
3403
3404  // Check for comparisons with builtin types.
3405  if (EmitWarning)
3406    if (CallExpr* CL = dyn_cast<CallExpr>(LeftExprSansParen))
3407      if (CL->isBuiltinCall())
3408        EmitWarning = false;
3409
3410  if (EmitWarning)
3411    if (CallExpr* CR = dyn_cast<CallExpr>(RightExprSansParen))
3412      if (CR->isBuiltinCall())
3413        EmitWarning = false;
3414
3415  // Emit the diagnostic.
3416  if (EmitWarning)
3417    Diag(Loc, diag::warn_floatingpoint_eq)
3418      << LHS->getSourceRange() << RHS->getSourceRange();
3419}
3420
3421//===--- CHECK: Integer mixed-sign comparisons (-Wsign-compare) --------===//
3422//===--- CHECK: Lossy implicit conversions (-Wconversion) --------------===//
3423
3424namespace {
3425
3426/// Structure recording the 'active' range of an integer-valued
3427/// expression.
3428struct IntRange {
3429  /// The number of bits active in the int.
3430  unsigned Width;
3431
3432  /// True if the int is known not to have negative values.
3433  bool NonNegative;
3434
3435  IntRange(unsigned Width, bool NonNegative)
3436    : Width(Width), NonNegative(NonNegative)
3437  {}
3438
3439  /// Returns the range of the bool type.
3440  static IntRange forBoolType() {
3441    return IntRange(1, true);
3442  }
3443
3444  /// Returns the range of an opaque value of the given integral type.
3445  static IntRange forValueOfType(ASTContext &C, QualType T) {
3446    return forValueOfCanonicalType(C,
3447                          T->getCanonicalTypeInternal().getTypePtr());
3448  }
3449
3450  /// Returns the range of an opaque value of a canonical integral type.
3451  static IntRange forValueOfCanonicalType(ASTContext &C, const Type *T) {
3452    assert(T->isCanonicalUnqualified());
3453
3454    if (const VectorType *VT = dyn_cast<VectorType>(T))
3455      T = VT->getElementType().getTypePtr();
3456    if (const ComplexType *CT = dyn_cast<ComplexType>(T))
3457      T = CT->getElementType().getTypePtr();
3458
3459    // For enum types, use the known bit width of the enumerators.
3460    if (const EnumType *ET = dyn_cast<EnumType>(T)) {
3461      EnumDecl *Enum = ET->getDecl();
3462      if (!Enum->isCompleteDefinition())
3463        return IntRange(C.getIntWidth(QualType(T, 0)), false);
3464
3465      unsigned NumPositive = Enum->getNumPositiveBits();
3466      unsigned NumNegative = Enum->getNumNegativeBits();
3467
3468      return IntRange(std::max(NumPositive, NumNegative), NumNegative == 0);
3469    }
3470
3471    const BuiltinType *BT = cast<BuiltinType>(T);
3472    assert(BT->isInteger());
3473
3474    return IntRange(C.getIntWidth(QualType(T, 0)), BT->isUnsignedInteger());
3475  }
3476
3477  /// Returns the "target" range of a canonical integral type, i.e.
3478  /// the range of values expressible in the type.
3479  ///
3480  /// This matches forValueOfCanonicalType except that enums have the
3481  /// full range of their type, not the range of their enumerators.
3482  static IntRange forTargetOfCanonicalType(ASTContext &C, const Type *T) {
3483    assert(T->isCanonicalUnqualified());
3484
3485    if (const VectorType *VT = dyn_cast<VectorType>(T))
3486      T = VT->getElementType().getTypePtr();
3487    if (const ComplexType *CT = dyn_cast<ComplexType>(T))
3488      T = CT->getElementType().getTypePtr();
3489    if (const EnumType *ET = dyn_cast<EnumType>(T))
3490      T = C.getCanonicalType(ET->getDecl()->getIntegerType()).getTypePtr();
3491
3492    const BuiltinType *BT = cast<BuiltinType>(T);
3493    assert(BT->isInteger());
3494
3495    return IntRange(C.getIntWidth(QualType(T, 0)), BT->isUnsignedInteger());
3496  }
3497
3498  /// Returns the supremum of two ranges: i.e. their conservative merge.
3499  static IntRange join(IntRange L, IntRange R) {
3500    return IntRange(std::max(L.Width, R.Width),
3501                    L.NonNegative && R.NonNegative);
3502  }
3503
3504  /// Returns the infinum of two ranges: i.e. their aggressive merge.
3505  static IntRange meet(IntRange L, IntRange R) {
3506    return IntRange(std::min(L.Width, R.Width),
3507                    L.NonNegative || R.NonNegative);
3508  }
3509};
3510
3511static IntRange GetValueRange(ASTContext &C, llvm::APSInt &value,
3512                              unsigned MaxWidth) {
3513  if (value.isSigned() && value.isNegative())
3514    return IntRange(value.getMinSignedBits(), false);
3515
3516  if (value.getBitWidth() > MaxWidth)
3517    value = value.trunc(MaxWidth);
3518
3519  // isNonNegative() just checks the sign bit without considering
3520  // signedness.
3521  return IntRange(value.getActiveBits(), true);
3522}
3523
3524static IntRange GetValueRange(ASTContext &C, APValue &result, QualType Ty,
3525                              unsigned MaxWidth) {
3526  if (result.isInt())
3527    return GetValueRange(C, result.getInt(), MaxWidth);
3528
3529  if (result.isVector()) {
3530    IntRange R = GetValueRange(C, result.getVectorElt(0), Ty, MaxWidth);
3531    for (unsigned i = 1, e = result.getVectorLength(); i != e; ++i) {
3532      IntRange El = GetValueRange(C, result.getVectorElt(i), Ty, MaxWidth);
3533      R = IntRange::join(R, El);
3534    }
3535    return R;
3536  }
3537
3538  if (result.isComplexInt()) {
3539    IntRange R = GetValueRange(C, result.getComplexIntReal(), MaxWidth);
3540    IntRange I = GetValueRange(C, result.getComplexIntImag(), MaxWidth);
3541    return IntRange::join(R, I);
3542  }
3543
3544  // This can happen with lossless casts to intptr_t of "based" lvalues.
3545  // Assume it might use arbitrary bits.
3546  // FIXME: The only reason we need to pass the type in here is to get
3547  // the sign right on this one case.  It would be nice if APValue
3548  // preserved this.
3549  assert(result.isLValue() || result.isAddrLabelDiff());
3550  return IntRange(MaxWidth, Ty->isUnsignedIntegerOrEnumerationType());
3551}
3552
3553/// Pseudo-evaluate the given integer expression, estimating the
3554/// range of values it might take.
3555///
3556/// \param MaxWidth - the width to which the value will be truncated
3557static IntRange GetExprRange(ASTContext &C, Expr *E, unsigned MaxWidth) {
3558  E = E->IgnoreParens();
3559
3560  // Try a full evaluation first.
3561  Expr::EvalResult result;
3562  if (E->EvaluateAsRValue(result, C))
3563    return GetValueRange(C, result.Val, E->getType(), MaxWidth);
3564
3565  // I think we only want to look through implicit casts here; if the
3566  // user has an explicit widening cast, we should treat the value as
3567  // being of the new, wider type.
3568  if (ImplicitCastExpr *CE = dyn_cast<ImplicitCastExpr>(E)) {
3569    if (CE->getCastKind() == CK_NoOp || CE->getCastKind() == CK_LValueToRValue)
3570      return GetExprRange(C, CE->getSubExpr(), MaxWidth);
3571
3572    IntRange OutputTypeRange = IntRange::forValueOfType(C, CE->getType());
3573
3574    bool isIntegerCast = (CE->getCastKind() == CK_IntegralCast);
3575
3576    // Assume that non-integer casts can span the full range of the type.
3577    if (!isIntegerCast)
3578      return OutputTypeRange;
3579
3580    IntRange SubRange
3581      = GetExprRange(C, CE->getSubExpr(),
3582                     std::min(MaxWidth, OutputTypeRange.Width));
3583
3584    // Bail out if the subexpr's range is as wide as the cast type.
3585    if (SubRange.Width >= OutputTypeRange.Width)
3586      return OutputTypeRange;
3587
3588    // Otherwise, we take the smaller width, and we're non-negative if
3589    // either the output type or the subexpr is.
3590    return IntRange(SubRange.Width,
3591                    SubRange.NonNegative || OutputTypeRange.NonNegative);
3592  }
3593
3594  if (ConditionalOperator *CO = dyn_cast<ConditionalOperator>(E)) {
3595    // If we can fold the condition, just take that operand.
3596    bool CondResult;
3597    if (CO->getCond()->EvaluateAsBooleanCondition(CondResult, C))
3598      return GetExprRange(C, CondResult ? CO->getTrueExpr()
3599                                        : CO->getFalseExpr(),
3600                          MaxWidth);
3601
3602    // Otherwise, conservatively merge.
3603    IntRange L = GetExprRange(C, CO->getTrueExpr(), MaxWidth);
3604    IntRange R = GetExprRange(C, CO->getFalseExpr(), MaxWidth);
3605    return IntRange::join(L, R);
3606  }
3607
3608  if (BinaryOperator *BO = dyn_cast<BinaryOperator>(E)) {
3609    switch (BO->getOpcode()) {
3610
3611    // Boolean-valued operations are single-bit and positive.
3612    case BO_LAnd:
3613    case BO_LOr:
3614    case BO_LT:
3615    case BO_GT:
3616    case BO_LE:
3617    case BO_GE:
3618    case BO_EQ:
3619    case BO_NE:
3620      return IntRange::forBoolType();
3621
3622    // The type of the assignments is the type of the LHS, so the RHS
3623    // is not necessarily the same type.
3624    case BO_MulAssign:
3625    case BO_DivAssign:
3626    case BO_RemAssign:
3627    case BO_AddAssign:
3628    case BO_SubAssign:
3629    case BO_XorAssign:
3630    case BO_OrAssign:
3631      // TODO: bitfields?
3632      return IntRange::forValueOfType(C, E->getType());
3633
3634    // Simple assignments just pass through the RHS, which will have
3635    // been coerced to the LHS type.
3636    case BO_Assign:
3637      // TODO: bitfields?
3638      return GetExprRange(C, BO->getRHS(), MaxWidth);
3639
3640    // Operations with opaque sources are black-listed.
3641    case BO_PtrMemD:
3642    case BO_PtrMemI:
3643      return IntRange::forValueOfType(C, E->getType());
3644
3645    // Bitwise-and uses the *infinum* of the two source ranges.
3646    case BO_And:
3647    case BO_AndAssign:
3648      return IntRange::meet(GetExprRange(C, BO->getLHS(), MaxWidth),
3649                            GetExprRange(C, BO->getRHS(), MaxWidth));
3650
3651    // Left shift gets black-listed based on a judgement call.
3652    case BO_Shl:
3653      // ...except that we want to treat '1 << (blah)' as logically
3654      // positive.  It's an important idiom.
3655      if (IntegerLiteral *I
3656            = dyn_cast<IntegerLiteral>(BO->getLHS()->IgnoreParenCasts())) {
3657        if (I->getValue() == 1) {
3658          IntRange R = IntRange::forValueOfType(C, E->getType());
3659          return IntRange(R.Width, /*NonNegative*/ true);
3660        }
3661      }
3662      // fallthrough
3663
3664    case BO_ShlAssign:
3665      return IntRange::forValueOfType(C, E->getType());
3666
3667    // Right shift by a constant can narrow its left argument.
3668    case BO_Shr:
3669    case BO_ShrAssign: {
3670      IntRange L = GetExprRange(C, BO->getLHS(), MaxWidth);
3671
3672      // If the shift amount is a positive constant, drop the width by
3673      // that much.
3674      llvm::APSInt shift;
3675      if (BO->getRHS()->isIntegerConstantExpr(shift, C) &&
3676          shift.isNonNegative()) {
3677        unsigned zext = shift.getZExtValue();
3678        if (zext >= L.Width)
3679          L.Width = (L.NonNegative ? 0 : 1);
3680        else
3681          L.Width -= zext;
3682      }
3683
3684      return L;
3685    }
3686
3687    // Comma acts as its right operand.
3688    case BO_Comma:
3689      return GetExprRange(C, BO->getRHS(), MaxWidth);
3690
3691    // Black-list pointer subtractions.
3692    case BO_Sub:
3693      if (BO->getLHS()->getType()->isPointerType())
3694        return IntRange::forValueOfType(C, E->getType());
3695      break;
3696
3697    // The width of a division result is mostly determined by the size
3698    // of the LHS.
3699    case BO_Div: {
3700      // Don't 'pre-truncate' the operands.
3701      unsigned opWidth = C.getIntWidth(E->getType());
3702      IntRange L = GetExprRange(C, BO->getLHS(), opWidth);
3703
3704      // If the divisor is constant, use that.
3705      llvm::APSInt divisor;
3706      if (BO->getRHS()->isIntegerConstantExpr(divisor, C)) {
3707        unsigned log2 = divisor.logBase2(); // floor(log_2(divisor))
3708        if (log2 >= L.Width)
3709          L.Width = (L.NonNegative ? 0 : 1);
3710        else
3711          L.Width = std::min(L.Width - log2, MaxWidth);
3712        return L;
3713      }
3714
3715      // Otherwise, just use the LHS's width.
3716      IntRange R = GetExprRange(C, BO->getRHS(), opWidth);
3717      return IntRange(L.Width, L.NonNegative && R.NonNegative);
3718    }
3719
3720    // The result of a remainder can't be larger than the result of
3721    // either side.
3722    case BO_Rem: {
3723      // Don't 'pre-truncate' the operands.
3724      unsigned opWidth = C.getIntWidth(E->getType());
3725      IntRange L = GetExprRange(C, BO->getLHS(), opWidth);
3726      IntRange R = GetExprRange(C, BO->getRHS(), opWidth);
3727
3728      IntRange meet = IntRange::meet(L, R);
3729      meet.Width = std::min(meet.Width, MaxWidth);
3730      return meet;
3731    }
3732
3733    // The default behavior is okay for these.
3734    case BO_Mul:
3735    case BO_Add:
3736    case BO_Xor:
3737    case BO_Or:
3738      break;
3739    }
3740
3741    // The default case is to treat the operation as if it were closed
3742    // on the narrowest type that encompasses both operands.
3743    IntRange L = GetExprRange(C, BO->getLHS(), MaxWidth);
3744    IntRange R = GetExprRange(C, BO->getRHS(), MaxWidth);
3745    return IntRange::join(L, R);
3746  }
3747
3748  if (UnaryOperator *UO = dyn_cast<UnaryOperator>(E)) {
3749    switch (UO->getOpcode()) {
3750    // Boolean-valued operations are white-listed.
3751    case UO_LNot:
3752      return IntRange::forBoolType();
3753
3754    // Operations with opaque sources are black-listed.
3755    case UO_Deref:
3756    case UO_AddrOf: // should be impossible
3757      return IntRange::forValueOfType(C, E->getType());
3758
3759    default:
3760      return GetExprRange(C, UO->getSubExpr(), MaxWidth);
3761    }
3762  }
3763
3764  if (dyn_cast<OffsetOfExpr>(E)) {
3765    IntRange::forValueOfType(C, E->getType());
3766  }
3767
3768  if (FieldDecl *BitField = E->getBitField())
3769    return IntRange(BitField->getBitWidthValue(C),
3770                    BitField->getType()->isUnsignedIntegerOrEnumerationType());
3771
3772  return IntRange::forValueOfType(C, E->getType());
3773}
3774
3775static IntRange GetExprRange(ASTContext &C, Expr *E) {
3776  return GetExprRange(C, E, C.getIntWidth(E->getType()));
3777}
3778
3779/// Checks whether the given value, which currently has the given
3780/// source semantics, has the same value when coerced through the
3781/// target semantics.
3782static bool IsSameFloatAfterCast(const llvm::APFloat &value,
3783                                 const llvm::fltSemantics &Src,
3784                                 const llvm::fltSemantics &Tgt) {
3785  llvm::APFloat truncated = value;
3786
3787  bool ignored;
3788  truncated.convert(Src, llvm::APFloat::rmNearestTiesToEven, &ignored);
3789  truncated.convert(Tgt, llvm::APFloat::rmNearestTiesToEven, &ignored);
3790
3791  return truncated.bitwiseIsEqual(value);
3792}
3793
3794/// Checks whether the given value, which currently has the given
3795/// source semantics, has the same value when coerced through the
3796/// target semantics.
3797///
3798/// The value might be a vector of floats (or a complex number).
3799static bool IsSameFloatAfterCast(const APValue &value,
3800                                 const llvm::fltSemantics &Src,
3801                                 const llvm::fltSemantics &Tgt) {
3802  if (value.isFloat())
3803    return IsSameFloatAfterCast(value.getFloat(), Src, Tgt);
3804
3805  if (value.isVector()) {
3806    for (unsigned i = 0, e = value.getVectorLength(); i != e; ++i)
3807      if (!IsSameFloatAfterCast(value.getVectorElt(i), Src, Tgt))
3808        return false;
3809    return true;
3810  }
3811
3812  assert(value.isComplexFloat());
3813  return (IsSameFloatAfterCast(value.getComplexFloatReal(), Src, Tgt) &&
3814          IsSameFloatAfterCast(value.getComplexFloatImag(), Src, Tgt));
3815}
3816
3817static void AnalyzeImplicitConversions(Sema &S, Expr *E, SourceLocation CC);
3818
3819static bool IsZero(Sema &S, Expr *E) {
3820  // Suppress cases where we are comparing against an enum constant.
3821  if (const DeclRefExpr *DR =
3822      dyn_cast<DeclRefExpr>(E->IgnoreParenImpCasts()))
3823    if (isa<EnumConstantDecl>(DR->getDecl()))
3824      return false;
3825
3826  // Suppress cases where the '0' value is expanded from a macro.
3827  if (E->getLocStart().isMacroID())
3828    return false;
3829
3830  llvm::APSInt Value;
3831  return E->isIntegerConstantExpr(Value, S.Context) && Value == 0;
3832}
3833
3834static bool HasEnumType(Expr *E) {
3835  // Strip off implicit integral promotions.
3836  while (ImplicitCastExpr *ICE = dyn_cast<ImplicitCastExpr>(E)) {
3837    if (ICE->getCastKind() != CK_IntegralCast &&
3838        ICE->getCastKind() != CK_NoOp)
3839      break;
3840    E = ICE->getSubExpr();
3841  }
3842
3843  return E->getType()->isEnumeralType();
3844}
3845
3846static void CheckTrivialUnsignedComparison(Sema &S, BinaryOperator *E) {
3847  BinaryOperatorKind op = E->getOpcode();
3848  if (E->isValueDependent())
3849    return;
3850
3851  if (op == BO_LT && IsZero(S, E->getRHS())) {
3852    S.Diag(E->getOperatorLoc(), diag::warn_lunsigned_always_true_comparison)
3853      << "< 0" << "false" << HasEnumType(E->getLHS())
3854      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
3855  } else if (op == BO_GE && IsZero(S, E->getRHS())) {
3856    S.Diag(E->getOperatorLoc(), diag::warn_lunsigned_always_true_comparison)
3857      << ">= 0" << "true" << HasEnumType(E->getLHS())
3858      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
3859  } else if (op == BO_GT && IsZero(S, E->getLHS())) {
3860    S.Diag(E->getOperatorLoc(), diag::warn_runsigned_always_true_comparison)
3861      << "0 >" << "false" << HasEnumType(E->getRHS())
3862      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
3863  } else if (op == BO_LE && IsZero(S, E->getLHS())) {
3864    S.Diag(E->getOperatorLoc(), diag::warn_runsigned_always_true_comparison)
3865      << "0 <=" << "true" << HasEnumType(E->getRHS())
3866      << E->getLHS()->getSourceRange() << E->getRHS()->getSourceRange();
3867  }
3868}
3869
3870/// Analyze the operands of the given comparison.  Implements the
3871/// fallback case from AnalyzeComparison.
3872static void AnalyzeImpConvsInComparison(Sema &S, BinaryOperator *E) {
3873  AnalyzeImplicitConversions(S, E->getLHS(), E->getOperatorLoc());
3874  AnalyzeImplicitConversions(S, E->getRHS(), E->getOperatorLoc());
3875}
3876
3877/// \brief Implements -Wsign-compare.
3878///
3879/// \param E the binary operator to check for warnings
3880static void AnalyzeComparison(Sema &S, BinaryOperator *E) {
3881  // The type the comparison is being performed in.
3882  QualType T = E->getLHS()->getType();
3883  assert(S.Context.hasSameUnqualifiedType(T, E->getRHS()->getType())
3884         && "comparison with mismatched types");
3885
3886  // We don't do anything special if this isn't an unsigned integral
3887  // comparison:  we're only interested in integral comparisons, and
3888  // signed comparisons only happen in cases we don't care to warn about.
3889  //
3890  // We also don't care about value-dependent expressions or expressions
3891  // whose result is a constant.
3892  if (!T->hasUnsignedIntegerRepresentation()
3893      || E->isValueDependent() || E->isIntegerConstantExpr(S.Context))
3894    return AnalyzeImpConvsInComparison(S, E);
3895
3896  Expr *LHS = E->getLHS()->IgnoreParenImpCasts();
3897  Expr *RHS = E->getRHS()->IgnoreParenImpCasts();
3898
3899  // Check to see if one of the (unmodified) operands is of different
3900  // signedness.
3901  Expr *signedOperand, *unsignedOperand;
3902  if (LHS->getType()->hasSignedIntegerRepresentation()) {
3903    assert(!RHS->getType()->hasSignedIntegerRepresentation() &&
3904           "unsigned comparison between two signed integer expressions?");
3905    signedOperand = LHS;
3906    unsignedOperand = RHS;
3907  } else if (RHS->getType()->hasSignedIntegerRepresentation()) {
3908    signedOperand = RHS;
3909    unsignedOperand = LHS;
3910  } else {
3911    CheckTrivialUnsignedComparison(S, E);
3912    return AnalyzeImpConvsInComparison(S, E);
3913  }
3914
3915  // Otherwise, calculate the effective range of the signed operand.
3916  IntRange signedRange = GetExprRange(S.Context, signedOperand);
3917
3918  // Go ahead and analyze implicit conversions in the operands.  Note
3919  // that we skip the implicit conversions on both sides.
3920  AnalyzeImplicitConversions(S, LHS, E->getOperatorLoc());
3921  AnalyzeImplicitConversions(S, RHS, E->getOperatorLoc());
3922
3923  // If the signed range is non-negative, -Wsign-compare won't fire,
3924  // but we should still check for comparisons which are always true
3925  // or false.
3926  if (signedRange.NonNegative)
3927    return CheckTrivialUnsignedComparison(S, E);
3928
3929  // For (in)equality comparisons, if the unsigned operand is a
3930  // constant which cannot collide with a overflowed signed operand,
3931  // then reinterpreting the signed operand as unsigned will not
3932  // change the result of the comparison.
3933  if (E->isEqualityOp()) {
3934    unsigned comparisonWidth = S.Context.getIntWidth(T);
3935    IntRange unsignedRange = GetExprRange(S.Context, unsignedOperand);
3936
3937    // We should never be unable to prove that the unsigned operand is
3938    // non-negative.
3939    assert(unsignedRange.NonNegative && "unsigned range includes negative?");
3940
3941    if (unsignedRange.Width < comparisonWidth)
3942      return;
3943  }
3944
3945  S.Diag(E->getOperatorLoc(), diag::warn_mixed_sign_comparison)
3946    << LHS->getType() << RHS->getType()
3947    << LHS->getSourceRange() << RHS->getSourceRange();
3948}
3949
3950/// Analyzes an attempt to assign the given value to a bitfield.
3951///
3952/// Returns true if there was something fishy about the attempt.
3953static bool AnalyzeBitFieldAssignment(Sema &S, FieldDecl *Bitfield, Expr *Init,
3954                                      SourceLocation InitLoc) {
3955  assert(Bitfield->isBitField());
3956  if (Bitfield->isInvalidDecl())
3957    return false;
3958
3959  // White-list bool bitfields.
3960  if (Bitfield->getType()->isBooleanType())
3961    return false;
3962
3963  // Ignore value- or type-dependent expressions.
3964  if (Bitfield->getBitWidth()->isValueDependent() ||
3965      Bitfield->getBitWidth()->isTypeDependent() ||
3966      Init->isValueDependent() ||
3967      Init->isTypeDependent())
3968    return false;
3969
3970  Expr *OriginalInit = Init->IgnoreParenImpCasts();
3971
3972  llvm::APSInt Value;
3973  if (!OriginalInit->EvaluateAsInt(Value, S.Context, Expr::SE_AllowSideEffects))
3974    return false;
3975
3976  unsigned OriginalWidth = Value.getBitWidth();
3977  unsigned FieldWidth = Bitfield->getBitWidthValue(S.Context);
3978
3979  if (OriginalWidth <= FieldWidth)
3980    return false;
3981
3982  // Compute the value which the bitfield will contain.
3983  llvm::APSInt TruncatedValue = Value.trunc(FieldWidth);
3984  TruncatedValue.setIsSigned(Bitfield->getType()->isSignedIntegerType());
3985
3986  // Check whether the stored value is equal to the original value.
3987  TruncatedValue = TruncatedValue.extend(OriginalWidth);
3988  if (Value == TruncatedValue)
3989    return false;
3990
3991  // Special-case bitfields of width 1: booleans are naturally 0/1, and
3992  // therefore don't strictly fit into a signed bitfield of width 1.
3993  if (FieldWidth == 1 && Value == 1)
3994    return false;
3995
3996  std::string PrettyValue = Value.toString(10);
3997  std::string PrettyTrunc = TruncatedValue.toString(10);
3998
3999  S.Diag(InitLoc, diag::warn_impcast_bitfield_precision_constant)
4000    << PrettyValue << PrettyTrunc << OriginalInit->getType()
4001    << Init->getSourceRange();
4002
4003  return true;
4004}
4005
4006/// Analyze the given simple or compound assignment for warning-worthy
4007/// operations.
4008static void AnalyzeAssignment(Sema &S, BinaryOperator *E) {
4009  // Just recurse on the LHS.
4010  AnalyzeImplicitConversions(S, E->getLHS(), E->getOperatorLoc());
4011
4012  // We want to recurse on the RHS as normal unless we're assigning to
4013  // a bitfield.
4014  if (FieldDecl *Bitfield = E->getLHS()->getBitField()) {
4015    if (AnalyzeBitFieldAssignment(S, Bitfield, E->getRHS(),
4016                                  E->getOperatorLoc())) {
4017      // Recurse, ignoring any implicit conversions on the RHS.
4018      return AnalyzeImplicitConversions(S, E->getRHS()->IgnoreParenImpCasts(),
4019                                        E->getOperatorLoc());
4020    }
4021  }
4022
4023  AnalyzeImplicitConversions(S, E->getRHS(), E->getOperatorLoc());
4024}
4025
4026/// Diagnose an implicit cast;  purely a helper for CheckImplicitConversion.
4027static void DiagnoseImpCast(Sema &S, Expr *E, QualType SourceType, QualType T,
4028                            SourceLocation CContext, unsigned diag,
4029                            bool pruneControlFlow = false) {
4030  if (pruneControlFlow) {
4031    S.DiagRuntimeBehavior(E->getExprLoc(), E,
4032                          S.PDiag(diag)
4033                            << SourceType << T << E->getSourceRange()
4034                            << SourceRange(CContext));
4035    return;
4036  }
4037  S.Diag(E->getExprLoc(), diag)
4038    << SourceType << T << E->getSourceRange() << SourceRange(CContext);
4039}
4040
4041/// Diagnose an implicit cast;  purely a helper for CheckImplicitConversion.
4042static void DiagnoseImpCast(Sema &S, Expr *E, QualType T,
4043                            SourceLocation CContext, unsigned diag,
4044                            bool pruneControlFlow = false) {
4045  DiagnoseImpCast(S, E, E->getType(), T, CContext, diag, pruneControlFlow);
4046}
4047
4048/// Diagnose an implicit cast from a literal expression. Does not warn when the
4049/// cast wouldn't lose information.
4050void DiagnoseFloatingLiteralImpCast(Sema &S, FloatingLiteral *FL, QualType T,
4051                                    SourceLocation CContext) {
4052  // Try to convert the literal exactly to an integer. If we can, don't warn.
4053  bool isExact = false;
4054  const llvm::APFloat &Value = FL->getValue();
4055  llvm::APSInt IntegerValue(S.Context.getIntWidth(T),
4056                            T->hasUnsignedIntegerRepresentation());
4057  if (Value.convertToInteger(IntegerValue,
4058                             llvm::APFloat::rmTowardZero, &isExact)
4059      == llvm::APFloat::opOK && isExact)
4060    return;
4061
4062  S.Diag(FL->getExprLoc(), diag::warn_impcast_literal_float_to_integer)
4063    << FL->getType() << T << FL->getSourceRange() << SourceRange(CContext);
4064}
4065
4066std::string PrettyPrintInRange(const llvm::APSInt &Value, IntRange Range) {
4067  if (!Range.Width) return "0";
4068
4069  llvm::APSInt ValueInRange = Value;
4070  ValueInRange.setIsSigned(!Range.NonNegative);
4071  ValueInRange = ValueInRange.trunc(Range.Width);
4072  return ValueInRange.toString(10);
4073}
4074
4075void CheckImplicitConversion(Sema &S, Expr *E, QualType T,
4076                             SourceLocation CC, bool *ICContext = 0) {
4077  if (E->isTypeDependent() || E->isValueDependent()) return;
4078
4079  const Type *Source = S.Context.getCanonicalType(E->getType()).getTypePtr();
4080  const Type *Target = S.Context.getCanonicalType(T).getTypePtr();
4081  if (Source == Target) return;
4082  if (Target->isDependentType()) return;
4083
4084  // If the conversion context location is invalid don't complain. We also
4085  // don't want to emit a warning if the issue occurs from the expansion of
4086  // a system macro. The problem is that 'getSpellingLoc()' is slow, so we
4087  // delay this check as long as possible. Once we detect we are in that
4088  // scenario, we just return.
4089  if (CC.isInvalid())
4090    return;
4091
4092  // Diagnose implicit casts to bool.
4093  if (Target->isSpecificBuiltinType(BuiltinType::Bool)) {
4094    if (isa<StringLiteral>(E))
4095      // Warn on string literal to bool.  Checks for string literals in logical
4096      // expressions, for instances, assert(0 && "error here"), is prevented
4097      // by a check in AnalyzeImplicitConversions().
4098      return DiagnoseImpCast(S, E, T, CC,
4099                             diag::warn_impcast_string_literal_to_bool);
4100    if (Source->isFunctionType()) {
4101      // Warn on function to bool. Checks free functions and static member
4102      // functions. Weakly imported functions are excluded from the check,
4103      // since it's common to test their value to check whether the linker
4104      // found a definition for them.
4105      ValueDecl *D = 0;
4106      if (DeclRefExpr* R = dyn_cast<DeclRefExpr>(E)) {
4107        D = R->getDecl();
4108      } else if (MemberExpr *M = dyn_cast<MemberExpr>(E)) {
4109        D = M->getMemberDecl();
4110      }
4111
4112      if (D && !D->isWeak()) {
4113        if (FunctionDecl* F = dyn_cast<FunctionDecl>(D)) {
4114          S.Diag(E->getExprLoc(), diag::warn_impcast_function_to_bool)
4115            << F << E->getSourceRange() << SourceRange(CC);
4116          S.Diag(E->getExprLoc(), diag::note_function_to_bool_silence)
4117            << FixItHint::CreateInsertion(E->getExprLoc(), "&");
4118          QualType ReturnType;
4119          UnresolvedSet<4> NonTemplateOverloads;
4120          S.isExprCallable(*E, ReturnType, NonTemplateOverloads);
4121          if (!ReturnType.isNull()
4122              && ReturnType->isSpecificBuiltinType(BuiltinType::Bool))
4123            S.Diag(E->getExprLoc(), diag::note_function_to_bool_call)
4124              << FixItHint::CreateInsertion(
4125                 S.getPreprocessor().getLocForEndOfToken(E->getLocEnd()), "()");
4126          return;
4127        }
4128      }
4129    }
4130    return; // Other casts to bool are not checked.
4131  }
4132
4133  // Strip vector types.
4134  if (isa<VectorType>(Source)) {
4135    if (!isa<VectorType>(Target)) {
4136      if (S.SourceMgr.isInSystemMacro(CC))
4137        return;
4138      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_vector_scalar);
4139    }
4140
4141    // If the vector cast is cast between two vectors of the same size, it is
4142    // a bitcast, not a conversion.
4143    if (S.Context.getTypeSize(Source) == S.Context.getTypeSize(Target))
4144      return;
4145
4146    Source = cast<VectorType>(Source)->getElementType().getTypePtr();
4147    Target = cast<VectorType>(Target)->getElementType().getTypePtr();
4148  }
4149
4150  // Strip complex types.
4151  if (isa<ComplexType>(Source)) {
4152    if (!isa<ComplexType>(Target)) {
4153      if (S.SourceMgr.isInSystemMacro(CC))
4154        return;
4155
4156      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_complex_scalar);
4157    }
4158
4159    Source = cast<ComplexType>(Source)->getElementType().getTypePtr();
4160    Target = cast<ComplexType>(Target)->getElementType().getTypePtr();
4161  }
4162
4163  const BuiltinType *SourceBT = dyn_cast<BuiltinType>(Source);
4164  const BuiltinType *TargetBT = dyn_cast<BuiltinType>(Target);
4165
4166  // If the source is floating point...
4167  if (SourceBT && SourceBT->isFloatingPoint()) {
4168    // ...and the target is floating point...
4169    if (TargetBT && TargetBT->isFloatingPoint()) {
4170      // ...then warn if we're dropping FP rank.
4171
4172      // Builtin FP kinds are ordered by increasing FP rank.
4173      if (SourceBT->getKind() > TargetBT->getKind()) {
4174        // Don't warn about float constants that are precisely
4175        // representable in the target type.
4176        Expr::EvalResult result;
4177        if (E->EvaluateAsRValue(result, S.Context)) {
4178          // Value might be a float, a float vector, or a float complex.
4179          if (IsSameFloatAfterCast(result.Val,
4180                   S.Context.getFloatTypeSemantics(QualType(TargetBT, 0)),
4181                   S.Context.getFloatTypeSemantics(QualType(SourceBT, 0))))
4182            return;
4183        }
4184
4185        if (S.SourceMgr.isInSystemMacro(CC))
4186          return;
4187
4188        DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_float_precision);
4189      }
4190      return;
4191    }
4192
4193    // If the target is integral, always warn.
4194    if ((TargetBT && TargetBT->isInteger())) {
4195      if (S.SourceMgr.isInSystemMacro(CC))
4196        return;
4197
4198      Expr *InnerE = E->IgnoreParenImpCasts();
4199      // We also want to warn on, e.g., "int i = -1.234"
4200      if (UnaryOperator *UOp = dyn_cast<UnaryOperator>(InnerE))
4201        if (UOp->getOpcode() == UO_Minus || UOp->getOpcode() == UO_Plus)
4202          InnerE = UOp->getSubExpr()->IgnoreParenImpCasts();
4203
4204      if (FloatingLiteral *FL = dyn_cast<FloatingLiteral>(InnerE)) {
4205        DiagnoseFloatingLiteralImpCast(S, FL, T, CC);
4206      } else {
4207        DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_float_integer);
4208      }
4209    }
4210
4211    return;
4212  }
4213
4214  if (!Source->isIntegerType() || !Target->isIntegerType())
4215    return;
4216
4217  if ((E->isNullPointerConstant(S.Context, Expr::NPC_ValueDependentIsNotNull)
4218           == Expr::NPCK_GNUNull) && Target->isIntegerType()) {
4219    SourceLocation Loc = E->getSourceRange().getBegin();
4220    if (Loc.isMacroID())
4221      Loc = S.SourceMgr.getImmediateExpansionRange(Loc).first;
4222    S.Diag(Loc, diag::warn_impcast_null_pointer_to_integer)
4223        << T << clang::SourceRange(CC)
4224        << FixItHint::CreateReplacement(Loc, S.getFixItZeroLiteralForType(T));
4225    return;
4226  }
4227
4228  IntRange SourceRange = GetExprRange(S.Context, E);
4229  IntRange TargetRange = IntRange::forTargetOfCanonicalType(S.Context, Target);
4230
4231  if (SourceRange.Width > TargetRange.Width) {
4232    // If the source is a constant, use a default-on diagnostic.
4233    // TODO: this should happen for bitfield stores, too.
4234    llvm::APSInt Value(32);
4235    if (E->isIntegerConstantExpr(Value, S.Context)) {
4236      if (S.SourceMgr.isInSystemMacro(CC))
4237        return;
4238
4239      std::string PrettySourceValue = Value.toString(10);
4240      std::string PrettyTargetValue = PrettyPrintInRange(Value, TargetRange);
4241
4242      S.DiagRuntimeBehavior(E->getExprLoc(), E,
4243        S.PDiag(diag::warn_impcast_integer_precision_constant)
4244            << PrettySourceValue << PrettyTargetValue
4245            << E->getType() << T << E->getSourceRange()
4246            << clang::SourceRange(CC));
4247      return;
4248    }
4249
4250    // People want to build with -Wshorten-64-to-32 and not -Wconversion.
4251    if (S.SourceMgr.isInSystemMacro(CC))
4252      return;
4253
4254    if (TargetRange.Width == 32 && S.Context.getIntWidth(E->getType()) == 64)
4255      return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_integer_64_32,
4256                             /* pruneControlFlow */ true);
4257    return DiagnoseImpCast(S, E, T, CC, diag::warn_impcast_integer_precision);
4258  }
4259
4260  if ((TargetRange.NonNegative && !SourceRange.NonNegative) ||
4261      (!TargetRange.NonNegative && SourceRange.NonNegative &&
4262       SourceRange.Width == TargetRange.Width)) {
4263
4264    if (S.SourceMgr.isInSystemMacro(CC))
4265      return;
4266
4267    unsigned DiagID = diag::warn_impcast_integer_sign;
4268
4269    // Traditionally, gcc has warned about this under -Wsign-compare.
4270    // We also want to warn about it in -Wconversion.
4271    // So if -Wconversion is off, use a completely identical diagnostic
4272    // in the sign-compare group.
4273    // The conditional-checking code will
4274    if (ICContext) {
4275      DiagID = diag::warn_impcast_integer_sign_conditional;
4276      *ICContext = true;
4277    }
4278
4279    return DiagnoseImpCast(S, E, T, CC, DiagID);
4280  }
4281
4282  // Diagnose conversions between different enumeration types.
4283  // In C, we pretend that the type of an EnumConstantDecl is its enumeration
4284  // type, to give us better diagnostics.
4285  QualType SourceType = E->getType();
4286  if (!S.getLangOpts().CPlusPlus) {
4287    if (DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(E))
4288      if (EnumConstantDecl *ECD = dyn_cast<EnumConstantDecl>(DRE->getDecl())) {
4289        EnumDecl *Enum = cast<EnumDecl>(ECD->getDeclContext());
4290        SourceType = S.Context.getTypeDeclType(Enum);
4291        Source = S.Context.getCanonicalType(SourceType).getTypePtr();
4292      }
4293  }
4294
4295  if (const EnumType *SourceEnum = Source->getAs<EnumType>())
4296    if (const EnumType *TargetEnum = Target->getAs<EnumType>())
4297      if ((SourceEnum->getDecl()->getIdentifier() ||
4298           SourceEnum->getDecl()->getTypedefNameForAnonDecl()) &&
4299          (TargetEnum->getDecl()->getIdentifier() ||
4300           TargetEnum->getDecl()->getTypedefNameForAnonDecl()) &&
4301          SourceEnum != TargetEnum) {
4302        if (S.SourceMgr.isInSystemMacro(CC))
4303          return;
4304
4305        return DiagnoseImpCast(S, E, SourceType, T, CC,
4306                               diag::warn_impcast_different_enum_types);
4307      }
4308
4309  return;
4310}
4311
4312void CheckConditionalOperator(Sema &S, ConditionalOperator *E, QualType T);
4313
4314void CheckConditionalOperand(Sema &S, Expr *E, QualType T,
4315                             SourceLocation CC, bool &ICContext) {
4316  E = E->IgnoreParenImpCasts();
4317
4318  if (isa<ConditionalOperator>(E))
4319    return CheckConditionalOperator(S, cast<ConditionalOperator>(E), T);
4320
4321  AnalyzeImplicitConversions(S, E, CC);
4322  if (E->getType() != T)
4323    return CheckImplicitConversion(S, E, T, CC, &ICContext);
4324  return;
4325}
4326
4327void CheckConditionalOperator(Sema &S, ConditionalOperator *E, QualType T) {
4328  SourceLocation CC = E->getQuestionLoc();
4329
4330  AnalyzeImplicitConversions(S, E->getCond(), CC);
4331
4332  bool Suspicious = false;
4333  CheckConditionalOperand(S, E->getTrueExpr(), T, CC, Suspicious);
4334  CheckConditionalOperand(S, E->getFalseExpr(), T, CC, Suspicious);
4335
4336  // If -Wconversion would have warned about either of the candidates
4337  // for a signedness conversion to the context type...
4338  if (!Suspicious) return;
4339
4340  // ...but it's currently ignored...
4341  if (S.Diags.getDiagnosticLevel(diag::warn_impcast_integer_sign_conditional,
4342                                 CC))
4343    return;
4344
4345  // ...then check whether it would have warned about either of the
4346  // candidates for a signedness conversion to the condition type.
4347  if (E->getType() == T) return;
4348
4349  Suspicious = false;
4350  CheckImplicitConversion(S, E->getTrueExpr()->IgnoreParenImpCasts(),
4351                          E->getType(), CC, &Suspicious);
4352  if (!Suspicious)
4353    CheckImplicitConversion(S, E->getFalseExpr()->IgnoreParenImpCasts(),
4354                            E->getType(), CC, &Suspicious);
4355}
4356
4357/// AnalyzeImplicitConversions - Find and report any interesting
4358/// implicit conversions in the given expression.  There are a couple
4359/// of competing diagnostics here, -Wconversion and -Wsign-compare.
4360void AnalyzeImplicitConversions(Sema &S, Expr *OrigE, SourceLocation CC) {
4361  QualType T = OrigE->getType();
4362  Expr *E = OrigE->IgnoreParenImpCasts();
4363
4364  if (E->isTypeDependent() || E->isValueDependent())
4365    return;
4366
4367  // For conditional operators, we analyze the arguments as if they
4368  // were being fed directly into the output.
4369  if (isa<ConditionalOperator>(E)) {
4370    ConditionalOperator *CO = cast<ConditionalOperator>(E);
4371    CheckConditionalOperator(S, CO, T);
4372    return;
4373  }
4374
4375  // Go ahead and check any implicit conversions we might have skipped.
4376  // The non-canonical typecheck is just an optimization;
4377  // CheckImplicitConversion will filter out dead implicit conversions.
4378  if (E->getType() != T)
4379    CheckImplicitConversion(S, E, T, CC);
4380
4381  // Now continue drilling into this expression.
4382
4383  // Skip past explicit casts.
4384  if (isa<ExplicitCastExpr>(E)) {
4385    E = cast<ExplicitCastExpr>(E)->getSubExpr()->IgnoreParenImpCasts();
4386    return AnalyzeImplicitConversions(S, E, CC);
4387  }
4388
4389  if (BinaryOperator *BO = dyn_cast<BinaryOperator>(E)) {
4390    // Do a somewhat different check with comparison operators.
4391    if (BO->isComparisonOp())
4392      return AnalyzeComparison(S, BO);
4393
4394    // And with simple assignments.
4395    if (BO->getOpcode() == BO_Assign)
4396      return AnalyzeAssignment(S, BO);
4397  }
4398
4399  // These break the otherwise-useful invariant below.  Fortunately,
4400  // we don't really need to recurse into them, because any internal
4401  // expressions should have been analyzed already when they were
4402  // built into statements.
4403  if (isa<StmtExpr>(E)) return;
4404
4405  // Don't descend into unevaluated contexts.
4406  if (isa<UnaryExprOrTypeTraitExpr>(E)) return;
4407
4408  // Now just recurse over the expression's children.
4409  CC = E->getExprLoc();
4410  BinaryOperator *BO = dyn_cast<BinaryOperator>(E);
4411  bool IsLogicalOperator = BO && BO->isLogicalOp();
4412  for (Stmt::child_range I = E->children(); I; ++I) {
4413    Expr *ChildExpr = dyn_cast_or_null<Expr>(*I);
4414    if (!ChildExpr)
4415      continue;
4416
4417    if (IsLogicalOperator &&
4418        isa<StringLiteral>(ChildExpr->IgnoreParenImpCasts()))
4419      // Ignore checking string literals that are in logical operators.
4420      continue;
4421    AnalyzeImplicitConversions(S, ChildExpr, CC);
4422  }
4423}
4424
4425} // end anonymous namespace
4426
4427/// Diagnoses "dangerous" implicit conversions within the given
4428/// expression (which is a full expression).  Implements -Wconversion
4429/// and -Wsign-compare.
4430///
4431/// \param CC the "context" location of the implicit conversion, i.e.
4432///   the most location of the syntactic entity requiring the implicit
4433///   conversion
4434void Sema::CheckImplicitConversions(Expr *E, SourceLocation CC) {
4435  // Don't diagnose in unevaluated contexts.
4436  if (ExprEvalContexts.back().Context == Sema::Unevaluated)
4437    return;
4438
4439  // Don't diagnose for value- or type-dependent expressions.
4440  if (E->isTypeDependent() || E->isValueDependent())
4441    return;
4442
4443  // Check for array bounds violations in cases where the check isn't triggered
4444  // elsewhere for other Expr types (like BinaryOperators), e.g. when an
4445  // ArraySubscriptExpr is on the RHS of a variable initialization.
4446  CheckArrayAccess(E);
4447
4448  // This is not the right CC for (e.g.) a variable initialization.
4449  AnalyzeImplicitConversions(*this, E, CC);
4450}
4451
4452void Sema::CheckBitFieldInitialization(SourceLocation InitLoc,
4453                                       FieldDecl *BitField,
4454                                       Expr *Init) {
4455  (void) AnalyzeBitFieldAssignment(*this, BitField, Init, InitLoc);
4456}
4457
4458/// CheckParmsForFunctionDef - Check that the parameters of the given
4459/// function are appropriate for the definition of a function. This
4460/// takes care of any checks that cannot be performed on the
4461/// declaration itself, e.g., that the types of each of the function
4462/// parameters are complete.
4463bool Sema::CheckParmsForFunctionDef(ParmVarDecl **P, ParmVarDecl **PEnd,
4464                                    bool CheckParameterNames) {
4465  bool HasInvalidParm = false;
4466  for (; P != PEnd; ++P) {
4467    ParmVarDecl *Param = *P;
4468
4469    // C99 6.7.5.3p4: the parameters in a parameter type list in a
4470    // function declarator that is part of a function definition of
4471    // that function shall not have incomplete type.
4472    //
4473    // This is also C++ [dcl.fct]p6.
4474    if (!Param->isInvalidDecl() &&
4475        RequireCompleteType(Param->getLocation(), Param->getType(),
4476                               diag::err_typecheck_decl_incomplete_type)) {
4477      Param->setInvalidDecl();
4478      HasInvalidParm = true;
4479    }
4480
4481    // C99 6.9.1p5: If the declarator includes a parameter type list, the
4482    // declaration of each parameter shall include an identifier.
4483    if (CheckParameterNames &&
4484        Param->getIdentifier() == 0 &&
4485        !Param->isImplicit() &&
4486        !getLangOpts().CPlusPlus)
4487      Diag(Param->getLocation(), diag::err_parameter_name_omitted);
4488
4489    // C99 6.7.5.3p12:
4490    //   If the function declarator is not part of a definition of that
4491    //   function, parameters may have incomplete type and may use the [*]
4492    //   notation in their sequences of declarator specifiers to specify
4493    //   variable length array types.
4494    QualType PType = Param->getOriginalType();
4495    if (const ArrayType *AT = Context.getAsArrayType(PType)) {
4496      if (AT->getSizeModifier() == ArrayType::Star) {
4497        // FIXME: This diagnosic should point the the '[*]' if source-location
4498        // information is added for it.
4499        Diag(Param->getLocation(), diag::err_array_star_in_function_definition);
4500      }
4501    }
4502  }
4503
4504  return HasInvalidParm;
4505}
4506
4507/// CheckCastAlign - Implements -Wcast-align, which warns when a
4508/// pointer cast increases the alignment requirements.
4509void Sema::CheckCastAlign(Expr *Op, QualType T, SourceRange TRange) {
4510  // This is actually a lot of work to potentially be doing on every
4511  // cast; don't do it if we're ignoring -Wcast_align (as is the default).
4512  if (getDiagnostics().getDiagnosticLevel(diag::warn_cast_align,
4513                                          TRange.getBegin())
4514        == DiagnosticsEngine::Ignored)
4515    return;
4516
4517  // Ignore dependent types.
4518  if (T->isDependentType() || Op->getType()->isDependentType())
4519    return;
4520
4521  // Require that the destination be a pointer type.
4522  const PointerType *DestPtr = T->getAs<PointerType>();
4523  if (!DestPtr) return;
4524
4525  // If the destination has alignment 1, we're done.
4526  QualType DestPointee = DestPtr->getPointeeType();
4527  if (DestPointee->isIncompleteType()) return;
4528  CharUnits DestAlign = Context.getTypeAlignInChars(DestPointee);
4529  if (DestAlign.isOne()) return;
4530
4531  // Require that the source be a pointer type.
4532  const PointerType *SrcPtr = Op->getType()->getAs<PointerType>();
4533  if (!SrcPtr) return;
4534  QualType SrcPointee = SrcPtr->getPointeeType();
4535
4536  // Whitelist casts from cv void*.  We already implicitly
4537  // whitelisted casts to cv void*, since they have alignment 1.
4538  // Also whitelist casts involving incomplete types, which implicitly
4539  // includes 'void'.
4540  if (SrcPointee->isIncompleteType()) return;
4541
4542  CharUnits SrcAlign = Context.getTypeAlignInChars(SrcPointee);
4543  if (SrcAlign >= DestAlign) return;
4544
4545  Diag(TRange.getBegin(), diag::warn_cast_align)
4546    << Op->getType() << T
4547    << static_cast<unsigned>(SrcAlign.getQuantity())
4548    << static_cast<unsigned>(DestAlign.getQuantity())
4549    << TRange << Op->getSourceRange();
4550}
4551
4552static const Type* getElementType(const Expr *BaseExpr) {
4553  const Type* EltType = BaseExpr->getType().getTypePtr();
4554  if (EltType->isAnyPointerType())
4555    return EltType->getPointeeType().getTypePtr();
4556  else if (EltType->isArrayType())
4557    return EltType->getBaseElementTypeUnsafe();
4558  return EltType;
4559}
4560
4561/// \brief Check whether this array fits the idiom of a size-one tail padded
4562/// array member of a struct.
4563///
4564/// We avoid emitting out-of-bounds access warnings for such arrays as they are
4565/// commonly used to emulate flexible arrays in C89 code.
4566static bool IsTailPaddedMemberArray(Sema &S, llvm::APInt Size,
4567                                    const NamedDecl *ND) {
4568  if (Size != 1 || !ND) return false;
4569
4570  const FieldDecl *FD = dyn_cast<FieldDecl>(ND);
4571  if (!FD) return false;
4572
4573  // Don't consider sizes resulting from macro expansions or template argument
4574  // substitution to form C89 tail-padded arrays.
4575  ConstantArrayTypeLoc TL =
4576    cast<ConstantArrayTypeLoc>(FD->getTypeSourceInfo()->getTypeLoc());
4577  const Expr *SizeExpr = dyn_cast<IntegerLiteral>(TL.getSizeExpr());
4578  if (!SizeExpr || SizeExpr->getExprLoc().isMacroID())
4579    return false;
4580
4581  const RecordDecl *RD = dyn_cast<RecordDecl>(FD->getDeclContext());
4582  if (!RD) return false;
4583  if (RD->isUnion()) return false;
4584  if (const CXXRecordDecl *CRD = dyn_cast<CXXRecordDecl>(RD)) {
4585    if (!CRD->isStandardLayout()) return false;
4586  }
4587
4588  // See if this is the last field decl in the record.
4589  const Decl *D = FD;
4590  while ((D = D->getNextDeclInContext()))
4591    if (isa<FieldDecl>(D))
4592      return false;
4593  return true;
4594}
4595
4596void Sema::CheckArrayAccess(const Expr *BaseExpr, const Expr *IndexExpr,
4597                            const ArraySubscriptExpr *ASE,
4598                            bool AllowOnePastEnd, bool IndexNegated) {
4599  IndexExpr = IndexExpr->IgnoreParenImpCasts();
4600  if (IndexExpr->isValueDependent())
4601    return;
4602
4603  const Type *EffectiveType = getElementType(BaseExpr);
4604  BaseExpr = BaseExpr->IgnoreParenCasts();
4605  const ConstantArrayType *ArrayTy =
4606    Context.getAsConstantArrayType(BaseExpr->getType());
4607  if (!ArrayTy)
4608    return;
4609
4610  llvm::APSInt index;
4611  if (!IndexExpr->EvaluateAsInt(index, Context))
4612    return;
4613  if (IndexNegated)
4614    index = -index;
4615
4616  const NamedDecl *ND = NULL;
4617  if (const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(BaseExpr))
4618    ND = dyn_cast<NamedDecl>(DRE->getDecl());
4619  if (const MemberExpr *ME = dyn_cast<MemberExpr>(BaseExpr))
4620    ND = dyn_cast<NamedDecl>(ME->getMemberDecl());
4621
4622  if (index.isUnsigned() || !index.isNegative()) {
4623    llvm::APInt size = ArrayTy->getSize();
4624    if (!size.isStrictlyPositive())
4625      return;
4626
4627    const Type* BaseType = getElementType(BaseExpr);
4628    if (BaseType != EffectiveType) {
4629      // Make sure we're comparing apples to apples when comparing index to size
4630      uint64_t ptrarith_typesize = Context.getTypeSize(EffectiveType);
4631      uint64_t array_typesize = Context.getTypeSize(BaseType);
4632      // Handle ptrarith_typesize being zero, such as when casting to void*
4633      if (!ptrarith_typesize) ptrarith_typesize = 1;
4634      if (ptrarith_typesize != array_typesize) {
4635        // There's a cast to a different size type involved
4636        uint64_t ratio = array_typesize / ptrarith_typesize;
4637        // TODO: Be smarter about handling cases where array_typesize is not a
4638        // multiple of ptrarith_typesize
4639        if (ptrarith_typesize * ratio == array_typesize)
4640          size *= llvm::APInt(size.getBitWidth(), ratio);
4641      }
4642    }
4643
4644    if (size.getBitWidth() > index.getBitWidth())
4645      index = index.zext(size.getBitWidth());
4646    else if (size.getBitWidth() < index.getBitWidth())
4647      size = size.zext(index.getBitWidth());
4648
4649    // For array subscripting the index must be less than size, but for pointer
4650    // arithmetic also allow the index (offset) to be equal to size since
4651    // computing the next address after the end of the array is legal and
4652    // commonly done e.g. in C++ iterators and range-based for loops.
4653    if (AllowOnePastEnd ? index.ule(size) : index.ult(size))
4654      return;
4655
4656    // Also don't warn for arrays of size 1 which are members of some
4657    // structure. These are often used to approximate flexible arrays in C89
4658    // code.
4659    if (IsTailPaddedMemberArray(*this, size, ND))
4660      return;
4661
4662    // Suppress the warning if the subscript expression (as identified by the
4663    // ']' location) and the index expression are both from macro expansions
4664    // within a system header.
4665    if (ASE) {
4666      SourceLocation RBracketLoc = SourceMgr.getSpellingLoc(
4667          ASE->getRBracketLoc());
4668      if (SourceMgr.isInSystemHeader(RBracketLoc)) {
4669        SourceLocation IndexLoc = SourceMgr.getSpellingLoc(
4670            IndexExpr->getLocStart());
4671        if (SourceMgr.isFromSameFile(RBracketLoc, IndexLoc))
4672          return;
4673      }
4674    }
4675
4676    unsigned DiagID = diag::warn_ptr_arith_exceeds_bounds;
4677    if (ASE)
4678      DiagID = diag::warn_array_index_exceeds_bounds;
4679
4680    DiagRuntimeBehavior(BaseExpr->getLocStart(), BaseExpr,
4681                        PDiag(DiagID) << index.toString(10, true)
4682                          << size.toString(10, true)
4683                          << (unsigned)size.getLimitedValue(~0U)
4684                          << IndexExpr->getSourceRange());
4685  } else {
4686    unsigned DiagID = diag::warn_array_index_precedes_bounds;
4687    if (!ASE) {
4688      DiagID = diag::warn_ptr_arith_precedes_bounds;
4689      if (index.isNegative()) index = -index;
4690    }
4691
4692    DiagRuntimeBehavior(BaseExpr->getLocStart(), BaseExpr,
4693                        PDiag(DiagID) << index.toString(10, true)
4694                          << IndexExpr->getSourceRange());
4695  }
4696
4697  if (!ND) {
4698    // Try harder to find a NamedDecl to point at in the note.
4699    while (const ArraySubscriptExpr *ASE =
4700           dyn_cast<ArraySubscriptExpr>(BaseExpr))
4701      BaseExpr = ASE->getBase()->IgnoreParenCasts();
4702    if (const DeclRefExpr *DRE = dyn_cast<DeclRefExpr>(BaseExpr))
4703      ND = dyn_cast<NamedDecl>(DRE->getDecl());
4704    if (const MemberExpr *ME = dyn_cast<MemberExpr>(BaseExpr))
4705      ND = dyn_cast<NamedDecl>(ME->getMemberDecl());
4706  }
4707
4708  if (ND)
4709    DiagRuntimeBehavior(ND->getLocStart(), BaseExpr,
4710                        PDiag(diag::note_array_index_out_of_bounds)
4711                          << ND->getDeclName());
4712}
4713
4714void Sema::CheckArrayAccess(const Expr *expr) {
4715  int AllowOnePastEnd = 0;
4716  while (expr) {
4717    expr = expr->IgnoreParenImpCasts();
4718    switch (expr->getStmtClass()) {
4719      case Stmt::ArraySubscriptExprClass: {
4720        const ArraySubscriptExpr *ASE = cast<ArraySubscriptExpr>(expr);
4721        CheckArrayAccess(ASE->getBase(), ASE->getIdx(), ASE,
4722                         AllowOnePastEnd > 0);
4723        return;
4724      }
4725      case Stmt::UnaryOperatorClass: {
4726        // Only unwrap the * and & unary operators
4727        const UnaryOperator *UO = cast<UnaryOperator>(expr);
4728        expr = UO->getSubExpr();
4729        switch (UO->getOpcode()) {
4730          case UO_AddrOf:
4731            AllowOnePastEnd++;
4732            break;
4733          case UO_Deref:
4734            AllowOnePastEnd--;
4735            break;
4736          default:
4737            return;
4738        }
4739        break;
4740      }
4741      case Stmt::ConditionalOperatorClass: {
4742        const ConditionalOperator *cond = cast<ConditionalOperator>(expr);
4743        if (const Expr *lhs = cond->getLHS())
4744          CheckArrayAccess(lhs);
4745        if (const Expr *rhs = cond->getRHS())
4746          CheckArrayAccess(rhs);
4747        return;
4748      }
4749      default:
4750        return;
4751    }
4752  }
4753}
4754
4755//===--- CHECK: Objective-C retain cycles ----------------------------------//
4756
4757namespace {
4758  struct RetainCycleOwner {
4759    RetainCycleOwner() : Variable(0), Indirect(false) {}
4760    VarDecl *Variable;
4761    SourceRange Range;
4762    SourceLocation Loc;
4763    bool Indirect;
4764
4765    void setLocsFrom(Expr *e) {
4766      Loc = e->getExprLoc();
4767      Range = e->getSourceRange();
4768    }
4769  };
4770}
4771
4772/// Consider whether capturing the given variable can possibly lead to
4773/// a retain cycle.
4774static bool considerVariable(VarDecl *var, Expr *ref, RetainCycleOwner &owner) {
4775  // In ARC, it's captured strongly iff the variable has __strong
4776  // lifetime.  In MRR, it's captured strongly if the variable is
4777  // __block and has an appropriate type.
4778  if (var->getType().getObjCLifetime() != Qualifiers::OCL_Strong)
4779    return false;
4780
4781  owner.Variable = var;
4782  owner.setLocsFrom(ref);
4783  return true;
4784}
4785
4786static bool findRetainCycleOwner(Sema &S, Expr *e, RetainCycleOwner &owner) {
4787  while (true) {
4788    e = e->IgnoreParens();
4789    if (CastExpr *cast = dyn_cast<CastExpr>(e)) {
4790      switch (cast->getCastKind()) {
4791      case CK_BitCast:
4792      case CK_LValueBitCast:
4793      case CK_LValueToRValue:
4794      case CK_ARCReclaimReturnedObject:
4795        e = cast->getSubExpr();
4796        continue;
4797
4798      default:
4799        return false;
4800      }
4801    }
4802
4803    if (ObjCIvarRefExpr *ref = dyn_cast<ObjCIvarRefExpr>(e)) {
4804      ObjCIvarDecl *ivar = ref->getDecl();
4805      if (ivar->getType().getObjCLifetime() != Qualifiers::OCL_Strong)
4806        return false;
4807
4808      // Try to find a retain cycle in the base.
4809      if (!findRetainCycleOwner(S, ref->getBase(), owner))
4810        return false;
4811
4812      if (ref->isFreeIvar()) owner.setLocsFrom(ref);
4813      owner.Indirect = true;
4814      return true;
4815    }
4816
4817    if (DeclRefExpr *ref = dyn_cast<DeclRefExpr>(e)) {
4818      VarDecl *var = dyn_cast<VarDecl>(ref->getDecl());
4819      if (!var) return false;
4820      return considerVariable(var, ref, owner);
4821    }
4822
4823    if (MemberExpr *member = dyn_cast<MemberExpr>(e)) {
4824      if (member->isArrow()) return false;
4825
4826      // Don't count this as an indirect ownership.
4827      e = member->getBase();
4828      continue;
4829    }
4830
4831    if (PseudoObjectExpr *pseudo = dyn_cast<PseudoObjectExpr>(e)) {
4832      // Only pay attention to pseudo-objects on property references.
4833      ObjCPropertyRefExpr *pre
4834        = dyn_cast<ObjCPropertyRefExpr>(pseudo->getSyntacticForm()
4835                                              ->IgnoreParens());
4836      if (!pre) return false;
4837      if (pre->isImplicitProperty()) return false;
4838      ObjCPropertyDecl *property = pre->getExplicitProperty();
4839      if (!property->isRetaining() &&
4840          !(property->getPropertyIvarDecl() &&
4841            property->getPropertyIvarDecl()->getType()
4842              .getObjCLifetime() == Qualifiers::OCL_Strong))
4843          return false;
4844
4845      owner.Indirect = true;
4846      if (pre->isSuperReceiver()) {
4847        owner.Variable = S.getCurMethodDecl()->getSelfDecl();
4848        if (!owner.Variable)
4849          return false;
4850        owner.Loc = pre->getLocation();
4851        owner.Range = pre->getSourceRange();
4852        return true;
4853      }
4854      e = const_cast<Expr*>(cast<OpaqueValueExpr>(pre->getBase())
4855                              ->getSourceExpr());
4856      continue;
4857    }
4858
4859    // Array ivars?
4860
4861    return false;
4862  }
4863}
4864
4865namespace {
4866  struct FindCaptureVisitor : EvaluatedExprVisitor<FindCaptureVisitor> {
4867    FindCaptureVisitor(ASTContext &Context, VarDecl *variable)
4868      : EvaluatedExprVisitor<FindCaptureVisitor>(Context),
4869        Variable(variable), Capturer(0) {}
4870
4871    VarDecl *Variable;
4872    Expr *Capturer;
4873
4874    void VisitDeclRefExpr(DeclRefExpr *ref) {
4875      if (ref->getDecl() == Variable && !Capturer)
4876        Capturer = ref;
4877    }
4878
4879    void VisitObjCIvarRefExpr(ObjCIvarRefExpr *ref) {
4880      if (Capturer) return;
4881      Visit(ref->getBase());
4882      if (Capturer && ref->isFreeIvar())
4883        Capturer = ref;
4884    }
4885
4886    void VisitBlockExpr(BlockExpr *block) {
4887      // Look inside nested blocks
4888      if (block->getBlockDecl()->capturesVariable(Variable))
4889        Visit(block->getBlockDecl()->getBody());
4890    }
4891  };
4892}
4893
4894/// Check whether the given argument is a block which captures a
4895/// variable.
4896static Expr *findCapturingExpr(Sema &S, Expr *e, RetainCycleOwner &owner) {
4897  assert(owner.Variable && owner.Loc.isValid());
4898
4899  e = e->IgnoreParenCasts();
4900  BlockExpr *block = dyn_cast<BlockExpr>(e);
4901  if (!block || !block->getBlockDecl()->capturesVariable(owner.Variable))
4902    return 0;
4903
4904  FindCaptureVisitor visitor(S.Context, owner.Variable);
4905  visitor.Visit(block->getBlockDecl()->getBody());
4906  return visitor.Capturer;
4907}
4908
4909static void diagnoseRetainCycle(Sema &S, Expr *capturer,
4910                                RetainCycleOwner &owner) {
4911  assert(capturer);
4912  assert(owner.Variable && owner.Loc.isValid());
4913
4914  S.Diag(capturer->getExprLoc(), diag::warn_arc_retain_cycle)
4915    << owner.Variable << capturer->getSourceRange();
4916  S.Diag(owner.Loc, diag::note_arc_retain_cycle_owner)
4917    << owner.Indirect << owner.Range;
4918}
4919
4920/// Check for a keyword selector that starts with the word 'add' or
4921/// 'set'.
4922static bool isSetterLikeSelector(Selector sel) {
4923  if (sel.isUnarySelector()) return false;
4924
4925  StringRef str = sel.getNameForSlot(0);
4926  while (!str.empty() && str.front() == '_') str = str.substr(1);
4927  if (str.startswith("set"))
4928    str = str.substr(3);
4929  else if (str.startswith("add")) {
4930    // Specially whitelist 'addOperationWithBlock:'.
4931    if (sel.getNumArgs() == 1 && str.startswith("addOperationWithBlock"))
4932      return false;
4933    str = str.substr(3);
4934  }
4935  else
4936    return false;
4937
4938  if (str.empty()) return true;
4939  return !islower(str.front());
4940}
4941
4942/// Check a message send to see if it's likely to cause a retain cycle.
4943void Sema::checkRetainCycles(ObjCMessageExpr *msg) {
4944  // Only check instance methods whose selector looks like a setter.
4945  if (!msg->isInstanceMessage() || !isSetterLikeSelector(msg->getSelector()))
4946    return;
4947
4948  // Try to find a variable that the receiver is strongly owned by.
4949  RetainCycleOwner owner;
4950  if (msg->getReceiverKind() == ObjCMessageExpr::Instance) {
4951    if (!findRetainCycleOwner(*this, msg->getInstanceReceiver(), owner))
4952      return;
4953  } else {
4954    assert(msg->getReceiverKind() == ObjCMessageExpr::SuperInstance);
4955    owner.Variable = getCurMethodDecl()->getSelfDecl();
4956    owner.Loc = msg->getSuperLoc();
4957    owner.Range = msg->getSuperLoc();
4958  }
4959
4960  // Check whether the receiver is captured by any of the arguments.
4961  for (unsigned i = 0, e = msg->getNumArgs(); i != e; ++i)
4962    if (Expr *capturer = findCapturingExpr(*this, msg->getArg(i), owner))
4963      return diagnoseRetainCycle(*this, capturer, owner);
4964}
4965
4966/// Check a property assign to see if it's likely to cause a retain cycle.
4967void Sema::checkRetainCycles(Expr *receiver, Expr *argument) {
4968  RetainCycleOwner owner;
4969  if (!findRetainCycleOwner(*this, receiver, owner))
4970    return;
4971
4972  if (Expr *capturer = findCapturingExpr(*this, argument, owner))
4973    diagnoseRetainCycle(*this, capturer, owner);
4974}
4975
4976bool Sema::checkUnsafeAssigns(SourceLocation Loc,
4977                              QualType LHS, Expr *RHS) {
4978  Qualifiers::ObjCLifetime LT = LHS.getObjCLifetime();
4979  if (LT != Qualifiers::OCL_Weak && LT != Qualifiers::OCL_ExplicitNone)
4980    return false;
4981  // strip off any implicit cast added to get to the one arc-specific
4982  while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
4983    if (cast->getCastKind() == CK_ARCConsumeObject) {
4984      Diag(Loc, diag::warn_arc_retained_assign)
4985        << (LT == Qualifiers::OCL_ExplicitNone)
4986        << RHS->getSourceRange();
4987      return true;
4988    }
4989    RHS = cast->getSubExpr();
4990  }
4991  return false;
4992}
4993
4994void Sema::checkUnsafeExprAssigns(SourceLocation Loc,
4995                              Expr *LHS, Expr *RHS) {
4996  QualType LHSType;
4997  // PropertyRef on LHS type need be directly obtained from
4998  // its declaration as it has a PsuedoType.
4999  ObjCPropertyRefExpr *PRE
5000    = dyn_cast<ObjCPropertyRefExpr>(LHS->IgnoreParens());
5001  if (PRE && !PRE->isImplicitProperty()) {
5002    const ObjCPropertyDecl *PD = PRE->getExplicitProperty();
5003    if (PD)
5004      LHSType = PD->getType();
5005  }
5006
5007  if (LHSType.isNull())
5008    LHSType = LHS->getType();
5009  if (checkUnsafeAssigns(Loc, LHSType, RHS))
5010    return;
5011  Qualifiers::ObjCLifetime LT = LHSType.getObjCLifetime();
5012  // FIXME. Check for other life times.
5013  if (LT != Qualifiers::OCL_None)
5014    return;
5015
5016  if (PRE) {
5017    if (PRE->isImplicitProperty())
5018      return;
5019    const ObjCPropertyDecl *PD = PRE->getExplicitProperty();
5020    if (!PD)
5021      return;
5022
5023    unsigned Attributes = PD->getPropertyAttributes();
5024    if (Attributes & ObjCPropertyDecl::OBJC_PR_assign) {
5025      // when 'assign' attribute was not explicitly specified
5026      // by user, ignore it and rely on property type itself
5027      // for lifetime info.
5028      unsigned AsWrittenAttr = PD->getPropertyAttributesAsWritten();
5029      if (!(AsWrittenAttr & ObjCPropertyDecl::OBJC_PR_assign) &&
5030          LHSType->isObjCRetainableType())
5031        return;
5032
5033      while (ImplicitCastExpr *cast = dyn_cast<ImplicitCastExpr>(RHS)) {
5034        if (cast->getCastKind() == CK_ARCConsumeObject) {
5035          Diag(Loc, diag::warn_arc_retained_property_assign)
5036          << RHS->getSourceRange();
5037          return;
5038        }
5039        RHS = cast->getSubExpr();
5040      }
5041    }
5042  }
5043}
5044
5045//===--- CHECK: Empty statement body (-Wempty-body) ---------------------===//
5046
5047namespace {
5048bool ShouldDiagnoseEmptyStmtBody(const SourceManager &SourceMgr,
5049                                 SourceLocation StmtLoc,
5050                                 const NullStmt *Body) {
5051  // Do not warn if the body is a macro that expands to nothing, e.g:
5052  //
5053  // #define CALL(x)
5054  // if (condition)
5055  //   CALL(0);
5056  //
5057  if (Body->hasLeadingEmptyMacro())
5058    return false;
5059
5060  // Get line numbers of statement and body.
5061  bool StmtLineInvalid;
5062  unsigned StmtLine = SourceMgr.getSpellingLineNumber(StmtLoc,
5063                                                      &StmtLineInvalid);
5064  if (StmtLineInvalid)
5065    return false;
5066
5067  bool BodyLineInvalid;
5068  unsigned BodyLine = SourceMgr.getSpellingLineNumber(Body->getSemiLoc(),
5069                                                      &BodyLineInvalid);
5070  if (BodyLineInvalid)
5071    return false;
5072
5073  // Warn if null statement and body are on the same line.
5074  if (StmtLine != BodyLine)
5075    return false;
5076
5077  return true;
5078}
5079} // Unnamed namespace
5080
5081void Sema::DiagnoseEmptyStmtBody(SourceLocation StmtLoc,
5082                                 const Stmt *Body,
5083                                 unsigned DiagID) {
5084  // Since this is a syntactic check, don't emit diagnostic for template
5085  // instantiations, this just adds noise.
5086  if (CurrentInstantiationScope)
5087    return;
5088
5089  // The body should be a null statement.
5090  const NullStmt *NBody = dyn_cast<NullStmt>(Body);
5091  if (!NBody)
5092    return;
5093
5094  // Do the usual checks.
5095  if (!ShouldDiagnoseEmptyStmtBody(SourceMgr, StmtLoc, NBody))
5096    return;
5097
5098  Diag(NBody->getSemiLoc(), DiagID);
5099  Diag(NBody->getSemiLoc(), diag::note_empty_body_on_separate_line);
5100}
5101
5102void Sema::DiagnoseEmptyLoopBody(const Stmt *S,
5103                                 const Stmt *PossibleBody) {
5104  assert(!CurrentInstantiationScope); // Ensured by caller
5105
5106  SourceLocation StmtLoc;
5107  const Stmt *Body;
5108  unsigned DiagID;
5109  if (const ForStmt *FS = dyn_cast<ForStmt>(S)) {
5110    StmtLoc = FS->getRParenLoc();
5111    Body = FS->getBody();
5112    DiagID = diag::warn_empty_for_body;
5113  } else if (const WhileStmt *WS = dyn_cast<WhileStmt>(S)) {
5114    StmtLoc = WS->getCond()->getSourceRange().getEnd();
5115    Body = WS->getBody();
5116    DiagID = diag::warn_empty_while_body;
5117  } else
5118    return; // Neither `for' nor `while'.
5119
5120  // The body should be a null statement.
5121  const NullStmt *NBody = dyn_cast<NullStmt>(Body);
5122  if (!NBody)
5123    return;
5124
5125  // Skip expensive checks if diagnostic is disabled.
5126  if (Diags.getDiagnosticLevel(DiagID, NBody->getSemiLoc()) ==
5127          DiagnosticsEngine::Ignored)
5128    return;
5129
5130  // Do the usual checks.
5131  if (!ShouldDiagnoseEmptyStmtBody(SourceMgr, StmtLoc, NBody))
5132    return;
5133
5134  // `for(...);' and `while(...);' are popular idioms, so in order to keep
5135  // noise level low, emit diagnostics only if for/while is followed by a
5136  // CompoundStmt, e.g.:
5137  //    for (int i = 0; i < n; i++);
5138  //    {
5139  //      a(i);
5140  //    }
5141  // or if for/while is followed by a statement with more indentation
5142  // than for/while itself:
5143  //    for (int i = 0; i < n; i++);
5144  //      a(i);
5145  bool ProbableTypo = isa<CompoundStmt>(PossibleBody);
5146  if (!ProbableTypo) {
5147    bool BodyColInvalid;
5148    unsigned BodyCol = SourceMgr.getPresumedColumnNumber(
5149                             PossibleBody->getLocStart(),
5150                             &BodyColInvalid);
5151    if (BodyColInvalid)
5152      return;
5153
5154    bool StmtColInvalid;
5155    unsigned StmtCol = SourceMgr.getPresumedColumnNumber(
5156                             S->getLocStart(),
5157                             &StmtColInvalid);
5158    if (StmtColInvalid)
5159      return;
5160
5161    if (BodyCol > StmtCol)
5162      ProbableTypo = true;
5163  }
5164
5165  if (ProbableTypo) {
5166    Diag(NBody->getSemiLoc(), DiagID);
5167    Diag(NBody->getSemiLoc(), diag::note_empty_body_on_separate_line);
5168  }
5169}
5170