SemaStmtAsm.cpp revision cf81cd270e31f97bb88b51e24708e75cc604980e 
1//===--- SemaStmtAsm.cpp - Semantic Analysis for Asm Statements -----------===//
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 semantic analysis for inline asm statements.
11//
12//===----------------------------------------------------------------------===//
13
14#include "clang/Sema/SemaInternal.h"
15#include "clang/Sema/Scope.h"
16#include "clang/Sema/ScopeInfo.h"
17#include "clang/Sema/Initialization.h"
18#include "clang/Sema/Lookup.h"
19#include "clang/AST/TypeLoc.h"
20#include "clang/Lex/Preprocessor.h"
21#include "clang/Basic/TargetInfo.h"
22#include "llvm/ADT/ArrayRef.h"
23#include "llvm/ADT/BitVector.h"
24#include "llvm/ADT/SmallString.h"
25#include "llvm/MC/MCAsmInfo.h"
26#include "llvm/MC/MCContext.h"
27#include "llvm/MC/MCObjectFileInfo.h"
28#include "llvm/MC/MCRegisterInfo.h"
29#include "llvm/MC/MCStreamer.h"
30#include "llvm/MC/MCSubtargetInfo.h"
31#include "llvm/MC/MCTargetAsmParser.h"
32#include "llvm/MC/MCParser/MCAsmParser.h"
33#include "llvm/Support/SourceMgr.h"
34#include "llvm/Support/TargetRegistry.h"
35#include "llvm/Support/TargetSelect.h"
36using namespace clang;
37using namespace sema;
38
39/// CheckAsmLValue - GNU C has an extremely ugly extension whereby they silently
40/// ignore "noop" casts in places where an lvalue is required by an inline asm.
41/// We emulate this behavior when -fheinous-gnu-extensions is specified, but
42/// provide a strong guidance to not use it.
43///
44/// This method checks to see if the argument is an acceptable l-value and
45/// returns false if it is a case we can handle.
46static bool CheckAsmLValue(const Expr *E, Sema &S) {
47  // Type dependent expressions will be checked during instantiation.
48  if (E->isTypeDependent())
49    return false;
50
51  if (E->isLValue())
52    return false;  // Cool, this is an lvalue.
53
54  // Okay, this is not an lvalue, but perhaps it is the result of a cast that we
55  // are supposed to allow.
56  const Expr *E2 = E->IgnoreParenNoopCasts(S.Context);
57  if (E != E2 && E2->isLValue()) {
58    if (!S.getLangOpts().HeinousExtensions)
59      S.Diag(E2->getLocStart(), diag::err_invalid_asm_cast_lvalue)
60        << E->getSourceRange();
61    else
62      S.Diag(E2->getLocStart(), diag::warn_invalid_asm_cast_lvalue)
63        << E->getSourceRange();
64    // Accept, even if we emitted an error diagnostic.
65    return false;
66  }
67
68  // None of the above, just randomly invalid non-lvalue.
69  return true;
70}
71
72/// isOperandMentioned - Return true if the specified operand # is mentioned
73/// anywhere in the decomposed asm string.
74static bool isOperandMentioned(unsigned OpNo,
75                         ArrayRef<GCCAsmStmt::AsmStringPiece> AsmStrPieces) {
76  for (unsigned p = 0, e = AsmStrPieces.size(); p != e; ++p) {
77    const GCCAsmStmt::AsmStringPiece &Piece = AsmStrPieces[p];
78    if (!Piece.isOperand()) continue;
79
80    // If this is a reference to the input and if the input was the smaller
81    // one, then we have to reject this asm.
82    if (Piece.getOperandNo() == OpNo)
83      return true;
84  }
85  return false;
86}
87
88StmtResult Sema::ActOnGCCAsmStmt(SourceLocation AsmLoc, bool IsSimple,
89                                 bool IsVolatile, unsigned NumOutputs,
90                                 unsigned NumInputs, IdentifierInfo **Names,
91                                 MultiExprArg constraints, MultiExprArg exprs,
92                                 Expr *asmString, MultiExprArg clobbers,
93                                 SourceLocation RParenLoc) {
94  unsigned NumClobbers = clobbers.size();
95  StringLiteral **Constraints =
96    reinterpret_cast<StringLiteral**>(constraints.data());
97  Expr **Exprs = exprs.data();
98  StringLiteral *AsmString = cast<StringLiteral>(asmString);
99  StringLiteral **Clobbers = reinterpret_cast<StringLiteral**>(clobbers.data());
100
101  SmallVector<TargetInfo::ConstraintInfo, 4> OutputConstraintInfos;
102
103  // The parser verifies that there is a string literal here.
104  if (!AsmString->isAscii())
105    return StmtError(Diag(AsmString->getLocStart(),diag::err_asm_wide_character)
106      << AsmString->getSourceRange());
107
108  for (unsigned i = 0; i != NumOutputs; i++) {
109    StringLiteral *Literal = Constraints[i];
110    if (!Literal->isAscii())
111      return StmtError(Diag(Literal->getLocStart(),diag::err_asm_wide_character)
112        << Literal->getSourceRange());
113
114    StringRef OutputName;
115    if (Names[i])
116      OutputName = Names[i]->getName();
117
118    TargetInfo::ConstraintInfo Info(Literal->getString(), OutputName);
119    if (!Context.getTargetInfo().validateOutputConstraint(Info))
120      return StmtError(Diag(Literal->getLocStart(),
121                            diag::err_asm_invalid_output_constraint)
122                       << Info.getConstraintStr());
123
124    // Check that the output exprs are valid lvalues.
125    Expr *OutputExpr = Exprs[i];
126    if (CheckAsmLValue(OutputExpr, *this)) {
127      return StmtError(Diag(OutputExpr->getLocStart(),
128                  diag::err_asm_invalid_lvalue_in_output)
129        << OutputExpr->getSourceRange());
130    }
131
132    OutputConstraintInfos.push_back(Info);
133  }
134
135  SmallVector<TargetInfo::ConstraintInfo, 4> InputConstraintInfos;
136
137  for (unsigned i = NumOutputs, e = NumOutputs + NumInputs; i != e; i++) {
138    StringLiteral *Literal = Constraints[i];
139    if (!Literal->isAscii())
140      return StmtError(Diag(Literal->getLocStart(),diag::err_asm_wide_character)
141        << Literal->getSourceRange());
142
143    StringRef InputName;
144    if (Names[i])
145      InputName = Names[i]->getName();
146
147    TargetInfo::ConstraintInfo Info(Literal->getString(), InputName);
148    if (!Context.getTargetInfo().validateInputConstraint(OutputConstraintInfos.data(),
149                                                NumOutputs, Info)) {
150      return StmtError(Diag(Literal->getLocStart(),
151                            diag::err_asm_invalid_input_constraint)
152                       << Info.getConstraintStr());
153    }
154
155    Expr *InputExpr = Exprs[i];
156
157    // Only allow void types for memory constraints.
158    if (Info.allowsMemory() && !Info.allowsRegister()) {
159      if (CheckAsmLValue(InputExpr, *this))
160        return StmtError(Diag(InputExpr->getLocStart(),
161                              diag::err_asm_invalid_lvalue_in_input)
162                         << Info.getConstraintStr()
163                         << InputExpr->getSourceRange());
164    }
165
166    if (Info.allowsRegister()) {
167      if (InputExpr->getType()->isVoidType()) {
168        return StmtError(Diag(InputExpr->getLocStart(),
169                              diag::err_asm_invalid_type_in_input)
170          << InputExpr->getType() << Info.getConstraintStr()
171          << InputExpr->getSourceRange());
172      }
173    }
174
175    ExprResult Result = DefaultFunctionArrayLvalueConversion(Exprs[i]);
176    if (Result.isInvalid())
177      return StmtError();
178
179    Exprs[i] = Result.take();
180    InputConstraintInfos.push_back(Info);
181  }
182
183  // Check that the clobbers are valid.
184  for (unsigned i = 0; i != NumClobbers; i++) {
185    StringLiteral *Literal = Clobbers[i];
186    if (!Literal->isAscii())
187      return StmtError(Diag(Literal->getLocStart(),diag::err_asm_wide_character)
188        << Literal->getSourceRange());
189
190    StringRef Clobber = Literal->getString();
191
192    if (!Context.getTargetInfo().isValidClobber(Clobber))
193      return StmtError(Diag(Literal->getLocStart(),
194                  diag::err_asm_unknown_register_name) << Clobber);
195  }
196
197  GCCAsmStmt *NS =
198    new (Context) GCCAsmStmt(Context, AsmLoc, IsSimple, IsVolatile, NumOutputs,
199                             NumInputs, Names, Constraints, Exprs, AsmString,
200                             NumClobbers, Clobbers, RParenLoc);
201  // Validate the asm string, ensuring it makes sense given the operands we
202  // have.
203  SmallVector<GCCAsmStmt::AsmStringPiece, 8> Pieces;
204  unsigned DiagOffs;
205  if (unsigned DiagID = NS->AnalyzeAsmString(Pieces, Context, DiagOffs)) {
206    Diag(getLocationOfStringLiteralByte(AsmString, DiagOffs), DiagID)
207           << AsmString->getSourceRange();
208    return StmtError();
209  }
210
211  // Validate tied input operands for type mismatches.
212  for (unsigned i = 0, e = InputConstraintInfos.size(); i != e; ++i) {
213    TargetInfo::ConstraintInfo &Info = InputConstraintInfos[i];
214
215    // If this is a tied constraint, verify that the output and input have
216    // either exactly the same type, or that they are int/ptr operands with the
217    // same size (int/long, int*/long, are ok etc).
218    if (!Info.hasTiedOperand()) continue;
219
220    unsigned TiedTo = Info.getTiedOperand();
221    unsigned InputOpNo = i+NumOutputs;
222    Expr *OutputExpr = Exprs[TiedTo];
223    Expr *InputExpr = Exprs[InputOpNo];
224
225    if (OutputExpr->isTypeDependent() || InputExpr->isTypeDependent())
226      continue;
227
228    QualType InTy = InputExpr->getType();
229    QualType OutTy = OutputExpr->getType();
230    if (Context.hasSameType(InTy, OutTy))
231      continue;  // All types can be tied to themselves.
232
233    // Decide if the input and output are in the same domain (integer/ptr or
234    // floating point.
235    enum AsmDomain {
236      AD_Int, AD_FP, AD_Other
237    } InputDomain, OutputDomain;
238
239    if (InTy->isIntegerType() || InTy->isPointerType())
240      InputDomain = AD_Int;
241    else if (InTy->isRealFloatingType())
242      InputDomain = AD_FP;
243    else
244      InputDomain = AD_Other;
245
246    if (OutTy->isIntegerType() || OutTy->isPointerType())
247      OutputDomain = AD_Int;
248    else if (OutTy->isRealFloatingType())
249      OutputDomain = AD_FP;
250    else
251      OutputDomain = AD_Other;
252
253    // They are ok if they are the same size and in the same domain.  This
254    // allows tying things like:
255    //   void* to int*
256    //   void* to int            if they are the same size.
257    //   double to long double   if they are the same size.
258    //
259    uint64_t OutSize = Context.getTypeSize(OutTy);
260    uint64_t InSize = Context.getTypeSize(InTy);
261    if (OutSize == InSize && InputDomain == OutputDomain &&
262        InputDomain != AD_Other)
263      continue;
264
265    // If the smaller input/output operand is not mentioned in the asm string,
266    // then we can promote the smaller one to a larger input and the asm string
267    // won't notice.
268    bool SmallerValueMentioned = false;
269
270    // If this is a reference to the input and if the input was the smaller
271    // one, then we have to reject this asm.
272    if (isOperandMentioned(InputOpNo, Pieces)) {
273      // This is a use in the asm string of the smaller operand.  Since we
274      // codegen this by promoting to a wider value, the asm will get printed
275      // "wrong".
276      SmallerValueMentioned |= InSize < OutSize;
277    }
278    if (isOperandMentioned(TiedTo, Pieces)) {
279      // If this is a reference to the output, and if the output is the larger
280      // value, then it's ok because we'll promote the input to the larger type.
281      SmallerValueMentioned |= OutSize < InSize;
282    }
283
284    // If the smaller value wasn't mentioned in the asm string, and if the
285    // output was a register, just extend the shorter one to the size of the
286    // larger one.
287    if (!SmallerValueMentioned && InputDomain != AD_Other &&
288        OutputConstraintInfos[TiedTo].allowsRegister())
289      continue;
290
291    // Either both of the operands were mentioned or the smaller one was
292    // mentioned.  One more special case that we'll allow: if the tied input is
293    // integer, unmentioned, and is a constant, then we'll allow truncating it
294    // down to the size of the destination.
295    if (InputDomain == AD_Int && OutputDomain == AD_Int &&
296        !isOperandMentioned(InputOpNo, Pieces) &&
297        InputExpr->isEvaluatable(Context)) {
298      CastKind castKind =
299        (OutTy->isBooleanType() ? CK_IntegralToBoolean : CK_IntegralCast);
300      InputExpr = ImpCastExprToType(InputExpr, OutTy, castKind).take();
301      Exprs[InputOpNo] = InputExpr;
302      NS->setInputExpr(i, InputExpr);
303      continue;
304    }
305
306    Diag(InputExpr->getLocStart(),
307         diag::err_asm_tying_incompatible_types)
308      << InTy << OutTy << OutputExpr->getSourceRange()
309      << InputExpr->getSourceRange();
310    return StmtError();
311  }
312
313  return Owned(NS);
314}
315
316// getSpelling - Get the spelling of the AsmTok token.
317static StringRef getSpelling(Sema &SemaRef, Token AsmTok) {
318  StringRef Asm;
319  SmallString<512> TokenBuf;
320  TokenBuf.resize(512);
321  bool StringInvalid = false;
322  Asm = SemaRef.PP.getSpelling(AsmTok, TokenBuf, &StringInvalid);
323  assert (!StringInvalid && "Expected valid string!");
324  return Asm;
325}
326
327// Build the inline assembly string.  Returns true on error.
328static bool buildMSAsmString(Sema &SemaRef,
329                             SourceLocation AsmLoc,
330                             ArrayRef<Token> AsmToks,
331                             std::string &AsmString) {
332  assert (!AsmToks.empty() && "Didn't expect an empty AsmToks!");
333
334  SmallString<512> Asm;
335  for (unsigned i = 0, e = AsmToks.size(); i < e; ++i) {
336    bool isNewAsm = ((i == 0) ||
337                     AsmToks[i].isAtStartOfLine() ||
338                     AsmToks[i].is(tok::kw_asm));
339    if (isNewAsm) {
340      if (i != 0)
341        Asm += "\n\t";
342
343      if (AsmToks[i].is(tok::kw_asm)) {
344        i++; // Skip __asm
345        if (i == e) {
346          SemaRef.Diag(AsmLoc, diag::err_asm_empty);
347          return true;
348        }
349
350      }
351    }
352
353    if (i && AsmToks[i].hasLeadingSpace() && !isNewAsm)
354      Asm += ' ';
355
356    StringRef Spelling = getSpelling(SemaRef, AsmToks[i]);
357    Asm += Spelling;
358  }
359  AsmString = Asm.str();
360  return false;
361}
362
363class MCAsmParserSemaCallbackImpl : public llvm::MCAsmParserSemaCallback {
364  Sema *SemaRef;
365
366public:
367  MCAsmParserSemaCallbackImpl(class Sema *Ref) { SemaRef = Ref; }
368  ~MCAsmParserSemaCallbackImpl() {}
369
370  void *LookupInlineAsmIdentifier(StringRef Name, void *SrcLoc, unsigned &Size){
371    SourceLocation Loc = SourceLocation::getFromPtrEncoding(SrcLoc);
372    NamedDecl *OpDecl = SemaRef->LookupInlineAsmIdentifier(Name, Loc, Size);
373    return static_cast<void *>(OpDecl);
374  }
375};
376
377NamedDecl *Sema::LookupInlineAsmIdentifier(StringRef Name, SourceLocation Loc,
378                                           unsigned &Size) {
379  Size = 0;
380  LookupResult Result(*this, &Context.Idents.get(Name), Loc,
381                      Sema::LookupOrdinaryName);
382
383  if (!LookupName(Result, getCurScope())) {
384    // If we don't find anything, return null; the AsmParser will assume
385    // it is a label of some sort.
386    return 0;
387  }
388
389  if (!Result.isSingleResult()) {
390    // FIXME: Diagnose result.
391    return 0;
392  }
393
394  NamedDecl *ND = Result.getFoundDecl();
395  if (isa<VarDecl>(ND) || isa<FunctionDecl>(ND)) {
396    if (VarDecl *Var = dyn_cast<VarDecl>(ND))
397      Size = Context.getTypeInfo(Var->getType()).first;
398
399    return ND;
400  }
401
402  // FIXME: Handle other kinds of results? (FieldDecl, etc.)
403  // FIXME: Diagnose if we find something we can't handle, like a typedef.
404  return 0;
405}
406
407StmtResult Sema::ActOnMSAsmStmt(SourceLocation AsmLoc, SourceLocation LBraceLoc,
408                                ArrayRef<Token> AsmToks,SourceLocation EndLoc) {
409  SmallVector<IdentifierInfo*, 4> Names;
410  SmallVector<StringRef, 4> ConstraintRefs;
411  SmallVector<Expr*, 4> Exprs;
412  SmallVector<StringRef, 4> ClobberRefs;
413
414  // Empty asm statements don't need to instantiate the AsmParser, etc.
415  if (AsmToks.empty()) {
416    StringRef EmptyAsmStr;
417    MSAsmStmt *NS =
418      new (Context) MSAsmStmt(Context, AsmLoc, LBraceLoc, /*IsSimple*/ true,
419                              /*IsVolatile*/ true, AsmToks, /*NumOutputs*/ 0,
420                              /*NumInputs*/ 0, Names, ConstraintRefs, Exprs,
421                              EmptyAsmStr, ClobberRefs, EndLoc);
422    return Owned(NS);
423  }
424
425  std::string AsmString;
426  if (buildMSAsmString(*this, AsmLoc, AsmToks, AsmString))
427    return StmtError();
428
429  // Get the target specific parser.
430  std::string Error;
431  const std::string &TT = Context.getTargetInfo().getTriple().getTriple();
432  const llvm::Target *TheTarget(llvm::TargetRegistry::lookupTarget(TT, Error));
433
434  OwningPtr<llvm::MCAsmInfo> MAI(TheTarget->createMCAsmInfo(TT));
435  OwningPtr<llvm::MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TT));
436  OwningPtr<llvm::MCObjectFileInfo> MOFI(new llvm::MCObjectFileInfo());
437  OwningPtr<llvm::MCSubtargetInfo>
438    STI(TheTarget->createMCSubtargetInfo(TT, "", ""));
439
440  llvm::SourceMgr SrcMgr;
441  llvm::MCContext Ctx(*MAI, *MRI, MOFI.get(), &SrcMgr);
442  llvm::MemoryBuffer *Buffer =
443    llvm::MemoryBuffer::getMemBuffer(AsmString, "<inline asm>");
444
445  // Tell SrcMgr about this buffer, which is what the parser will pick up.
446  SrcMgr.AddNewSourceBuffer(Buffer, llvm::SMLoc());
447
448  OwningPtr<llvm::MCStreamer> Str(createNullStreamer(Ctx));
449  OwningPtr<llvm::MCAsmParser>
450    Parser(createMCAsmParser(SrcMgr, Ctx, *Str.get(), *MAI));
451  OwningPtr<llvm::MCTargetAsmParser>
452    TargetParser(TheTarget->createMCAsmParser(*STI, *Parser));
453
454  // Get the instruction descriptor.
455  const llvm::MCInstrInfo *MII = TheTarget->createMCInstrInfo();
456  llvm::MCInstPrinter *IP =
457    TheTarget->createMCInstPrinter(1, *MAI, *MII, *MRI, *STI);
458
459  // Change to the Intel dialect.
460  Parser->setAssemblerDialect(1);
461  Parser->setTargetParser(*TargetParser.get());
462  Parser->setParsingInlineAsm(true);
463  TargetParser->setParsingInlineAsm(true);
464
465  unsigned NumOutputs;
466  unsigned NumInputs;
467  std::string AsmStringIR;
468  SmallVector<void *, 4> OpDecls;
469  SmallVector<std::string, 4> Constraints;
470  SmallVector<std::string, 4> Clobbers;
471  MCAsmParserSemaCallbackImpl MCAPSI(this);
472  if (Parser->ParseMSInlineAsm(AsmLoc.getPtrEncoding(), AsmStringIR,
473                               NumOutputs, NumInputs, OpDecls, Constraints,
474                               Clobbers, MII, IP, MCAPSI))
475    return StmtError();
476
477  // Build the vector of clobber StringRefs.
478  unsigned NumClobbers = Clobbers.size();
479  ClobberRefs.resize(NumClobbers);
480  for (unsigned i = 0; i != NumClobbers; ++i)
481    ClobberRefs[i] = StringRef(Clobbers[i]);
482
483  // Recast the void pointers and build the vector of constraint StringRefs.
484  unsigned NumExprs = NumOutputs + NumInputs;
485  Names.resize(NumExprs);
486  ConstraintRefs.resize(NumExprs);
487  Exprs.resize(NumExprs);
488  for (unsigned i = 0, e = NumExprs; i != e; ++i) {
489    NamedDecl *OpDecl = static_cast<NamedDecl *>(OpDecls[i]);
490    if (!OpDecl)
491      return StmtError();
492
493    DeclarationNameInfo NameInfo(OpDecl->getDeclName(), AsmLoc);
494    ExprResult OpExpr = BuildDeclarationNameExpr(CXXScopeSpec(), NameInfo,
495                                                 OpDecl);
496    if (OpExpr.isInvalid())
497      return StmtError();
498
499    Names[i] = OpDecl->getIdentifier();
500    ConstraintRefs[i] = StringRef(Constraints[i]);
501    Exprs[i] = OpExpr.take();
502  }
503
504  bool IsSimple = NumExprs > 0;
505  MSAsmStmt *NS =
506    new (Context) MSAsmStmt(Context, AsmLoc, LBraceLoc, IsSimple,
507                            /*IsVolatile*/ true, AsmToks, NumOutputs, NumInputs,
508                            Names, ConstraintRefs, Exprs, AsmStringIR,
509                            ClobberRefs, EndLoc);
510  return Owned(NS);
511}
512