/*===-- Lexer.l - Scanner for llvm assembly files --------------*- C++ -*--===// // // This file implements the flex scanner for LLVM assembly languages files. // //===----------------------------------------------------------------------===*/ %option prefix="llvmAsm" %option yylineno %option nostdinit %option never-interactive %option batch %option noyywrap %option nodefault %option 8bit %option outfile="Lexer.cpp" %option ecs %option noreject %option noyymore %{ #include "ParserInternals.h" #include #include "llvmAsmParser.h" #include #include #define RET_TOK(type, Enum, sym) \ llvmAsmlval.type = Instruction::Enum; return sym // TODO: All of the static identifiers are figured out by the lexer, // these should be hashed to reduce the lexer size // atoull - Convert an ascii string of decimal digits into the unsigned long // long representation... this does not have to do input error checking, // because we know that the input will be matched by a suitable regex... // static uint64_t atoull(const char *Buffer) { uint64_t Result = 0; for (; *Buffer; Buffer++) { uint64_t OldRes = Result; Result *= 10; Result += *Buffer-'0'; if (Result < OldRes) // Uh, oh, overflow detected!!! ThrowException("constant bigger than 64 bits detected!"); } return Result; } // HexToFP - Convert the ascii string in hexidecimal format to the floating // point representation of it. // static double HexToFP(const char *Buffer) { uint64_t Result = 0; for (; *Buffer; ++Buffer) { uint64_t OldRes = Result; Result *= 16; char C = *Buffer; if (C >= '0' && C <= '9') Result += C-'0'; else if (C >= 'A' && C <= 'F') Result += C-'A'+10; else if (C >= 'a' && C <= 'f') Result += C-'a'+10; if (Result < OldRes) // Uh, oh, overflow detected!!! ThrowException("constant bigger than 64 bits detected!"); } assert(sizeof(double) == sizeof(Result) && "Data sizes incompatible on this target!"); // Behave nicely in the face of C TBAA rules... see: // http://www.nullstone.com/htmls/category/aliastyp.htm // char *ProxyPointer = (char*)&Result; return *(double*)ProxyPointer; // Cast Hex constant to double } // UnEscapeLexed - Run through the specified buffer and change \xx codes to the // appropriate character. If AllowNull is set to false, a \00 value will cause // an exception to be thrown. // // If AllowNull is set to true, the return value of the function points to the // last character of the string in memory. // char *UnEscapeLexed(char *Buffer, bool AllowNull) { char *BOut = Buffer; for (char *BIn = Buffer; *BIn; ) { if (BIn[0] == '\\' && isxdigit(BIn[1]) && isxdigit(BIn[2])) { char Tmp = BIn[3]; BIn[3] = 0; // Terminate string *BOut = strtol(BIn+1, 0, 16); // Convert to number if (!AllowNull && !*BOut) ThrowException("String literal cannot accept \\00 escape!"); BIn[3] = Tmp; // Restore character BIn += 3; // Skip over handled chars ++BOut; } else { *BOut++ = *BIn++; } } return BOut; } #define YY_NEVER_INTERACTIVE 1 %} /* Comments start with a ; and go till end of line */ Comment ;.* /* Variable(Value) identifiers start with a % sign */ VarID %[-a-zA-Z$._][-a-zA-Z$._0-9]* /* Label identifiers end with a colon */ Label [-a-zA-Z$._0-9]+: /* Quoted names can contain any character except " and \ */ StringConstant \"[^\"]+\" /* [PN]Integer: match positive and negative literal integer values that * are preceeded by a '%' character. These represent unnamed variable slots. */ EPInteger %[0-9]+ ENInteger %-[0-9]+ /* E[PN]Integer: match positive and negative literal integer values */ PInteger [0-9]+ NInteger -[0-9]+ /* FPConstant - A Floating point constant. */ FPConstant [-+]?[0-9]+[.][0-9]*([eE][-+]?[0-9]+)? /* HexFPConstant - Floating point constant represented in IEEE format as a * hexadecimal number for when exponential notation is not precise enough. */ HexFPConstant 0x[0-9A-Fa-f]+ %% {Comment} { /* Ignore comments for now */ } begin { return BEGINTOK; } end { return ENDTOK; } true { return TRUE; } false { return FALSE; } declare { return DECLARE; } global { return GLOBAL; } constant { return CONSTANT; } const { return CONST; } internal { return INTERNAL; } linkonce { return LINKONCE; } appending { return APPENDING; } uninitialized { return EXTERNAL; } /* Deprecated, turn into external */ external { return EXTERNAL; } implementation { return IMPLEMENTATION; } \.\.\. { return DOTDOTDOT; } null { return NULL_TOK; } to { return TO; } except { return EXCEPT; } not { return NOT; } /* Deprecated, turned into XOR */ void { llvmAsmlval.PrimType = Type::VoidTy ; return VOID; } bool { llvmAsmlval.PrimType = Type::BoolTy ; return BOOL; } sbyte { llvmAsmlval.PrimType = Type::SByteTy ; return SBYTE; } ubyte { llvmAsmlval.PrimType = Type::UByteTy ; return UBYTE; } short { llvmAsmlval.PrimType = Type::ShortTy ; return SHORT; } ushort { llvmAsmlval.PrimType = Type::UShortTy; return USHORT; } int { llvmAsmlval.PrimType = Type::IntTy ; return INT; } uint { llvmAsmlval.PrimType = Type::UIntTy ; return UINT; } long { llvmAsmlval.PrimType = Type::LongTy ; return LONG; } ulong { llvmAsmlval.PrimType = Type::ULongTy ; return ULONG; } float { llvmAsmlval.PrimType = Type::FloatTy ; return FLOAT; } double { llvmAsmlval.PrimType = Type::DoubleTy; return DOUBLE; } type { llvmAsmlval.PrimType = Type::TypeTy ; return TYPE; } label { llvmAsmlval.PrimType = Type::LabelTy ; return LABEL; } opaque { return OPAQUE; } add { RET_TOK(BinaryOpVal, Add, ADD); } sub { RET_TOK(BinaryOpVal, Sub, SUB); } mul { RET_TOK(BinaryOpVal, Mul, MUL); } div { RET_TOK(BinaryOpVal, Div, DIV); } rem { RET_TOK(BinaryOpVal, Rem, REM); } and { RET_TOK(BinaryOpVal, And, AND); } or { RET_TOK(BinaryOpVal, Or , OR ); } xor { RET_TOK(BinaryOpVal, Xor, XOR); } setne { RET_TOK(BinaryOpVal, SetNE, SETNE); } seteq { RET_TOK(BinaryOpVal, SetEQ, SETEQ); } setlt { RET_TOK(BinaryOpVal, SetLT, SETLT); } setgt { RET_TOK(BinaryOpVal, SetGT, SETGT); } setle { RET_TOK(BinaryOpVal, SetLE, SETLE); } setge { RET_TOK(BinaryOpVal, SetGE, SETGE); } phi { RET_TOK(OtherOpVal, PHINode, PHI); } call { RET_TOK(OtherOpVal, Call, CALL); } cast { RET_TOK(OtherOpVal, Cast, CAST); } shl { RET_TOK(OtherOpVal, Shl, SHL); } shr { RET_TOK(OtherOpVal, Shr, SHR); } ret { RET_TOK(TermOpVal, Ret, RET); } br { RET_TOK(TermOpVal, Br, BR); } switch { RET_TOK(TermOpVal, Switch, SWITCH); } invoke { RET_TOK(TermOpVal, Invoke, INVOKE); } malloc { RET_TOK(MemOpVal, Malloc, MALLOC); } alloca { RET_TOK(MemOpVal, Alloca, ALLOCA); } free { RET_TOK(MemOpVal, Free, FREE); } load { RET_TOK(MemOpVal, Load, LOAD); } store { RET_TOK(MemOpVal, Store, STORE); } getelementptr { RET_TOK(MemOpVal, GetElementPtr, GETELEMENTPTR); } {VarID} { UnEscapeLexed(yytext+1); llvmAsmlval.StrVal = strdup(yytext+1); // Skip % return VAR_ID; } {Label} { yytext[strlen(yytext)-1] = 0; // nuke colon UnEscapeLexed(yytext); llvmAsmlval.StrVal = strdup(yytext); return LABELSTR; } {StringConstant} { // Note that we cannot unescape a string constant here! The // string constant might contain a \00 which would not be // understood by the string stuff. It is valid to make a // [sbyte] c"Hello World\00" constant, for example. // yytext[strlen(yytext)-1] = 0; // nuke end quote llvmAsmlval.StrVal = strdup(yytext+1); // Nuke start quote return STRINGCONSTANT; } {PInteger} { llvmAsmlval.UInt64Val = atoull(yytext); return EUINT64VAL; } {NInteger} { uint64_t Val = atoull(yytext+1); // +1: we have bigger negative range if (Val > (uint64_t)INT64_MAX+1) ThrowException("Constant too large for signed 64 bits!"); llvmAsmlval.SInt64Val = -Val; return ESINT64VAL; } {EPInteger} { llvmAsmlval.UIntVal = atoull(yytext+1); return UINTVAL; } {ENInteger} { uint64_t Val = atoull(yytext+2); // +1: we have bigger negative range if (Val > (uint64_t)INT32_MAX+1) ThrowException("Constant too large for signed 32 bits!"); llvmAsmlval.SIntVal = -Val; return SINTVAL; } {FPConstant} { llvmAsmlval.FPVal = atof(yytext); return FPVAL; } {HexFPConstant} { llvmAsmlval.FPVal = HexToFP(yytext); return FPVAL; } [ \t\n] { /* Ignore whitespace */ } . { return yytext[0]; } %%