%{
#include "ParserInternals.h"
#include "llvm/Module.h"
+#include "llvm/Support/MathExtras.h"
#include <list>
#include "llvmAsmParser.h"
#include <cctype>
yy_scan_string (str);
}
+// Construct a token value for a non-obsolete token
#define RET_TOK(type, Enum, sym) \
- llvmAsmlval.type = Instruction::Enum; return sym
+ llvmAsmlval.type = Instruction::Enum; \
+ return sym
+
+// Construct a token value for an obsolete token
+#define RET_TY(CTYPE, SYM) \
+ llvmAsmlval.PrimType = CTYPE;\
+ return SYM
namespace llvm {
Result *= 10;
Result += *Buffer-'0';
if (Result < OldRes) // Uh, oh, overflow detected!!!
- ThrowException("constant bigger than 64 bits detected!");
+ GenerateError("constant bigger than 64 bits detected!");
}
return Result;
}
Result += C-'a'+10;
if (Result < OldRes) // Uh, oh, overflow detected!!!
- ThrowException("constant bigger than 64 bits detected!");
+ GenerateError("constant bigger than 64 bits detected!");
}
return Result;
}
-
-// HexToFP - Convert the ascii string in hexidecimal format to the floating
+// HexToFP - Convert the ascii string in hexadecimal format to the floating
// point representation of it.
//
static double HexToFP(const char *Buffer) {
- // Behave nicely in the face of C TBAA rules... see:
- // http://www.nullstone.com/htmls/category/aliastyp.htm
- union {
- uint64_t UI;
- double FP;
- } UIntToFP;
- UIntToFP.UI = HexIntToVal(Buffer);
-
- assert(sizeof(double) == sizeof(uint64_t) &&
- "Data sizes incompatible on this target!");
- return UIntToFP.FP; // Cast Hex constant to double
+ return BitsToDouble(HexIntToVal(Buffer)); // Cast Hex constant to double
}
+static void HexToIntPair(const char *Buffer, uint64_t Pair[2]) {
+ Pair[0] = 0;
+ for (int i=0; i<16; i++, Buffer++) {
+ assert(*Buffer);
+ Pair[0] *= 16;
+ char C = *Buffer;
+ if (C >= '0' && C <= '9')
+ Pair[0] += C-'0';
+ else if (C >= 'A' && C <= 'F')
+ Pair[0] += C-'A'+10;
+ else if (C >= 'a' && C <= 'f')
+ Pair[0] += C-'a'+10;
+ }
+ Pair[1] = 0;
+ for (int i=0; i<16 && *Buffer; i++, Buffer++) {
+ Pair[1] *= 16;
+ char C = *Buffer;
+ if (C >= '0' && C <= '9')
+ Pair[1] += C-'0';
+ else if (C >= 'A' && C <= 'F')
+ Pair[1] += C-'A'+10;
+ else if (C >= 'a' && C <= 'f')
+ Pair[1] += C-'a'+10;
+ }
+ if (*Buffer)
+ GenerateError("constant bigger than 128 bits detected!");
+}
// 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) {
+// appropriate character.
+char *UnEscapeLexed(char *Buffer, char* EndBuffer) {
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 = (char)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;
+ if (BIn[0] == '\\') {
+ if (BIn < EndBuffer-1 && BIn[1] == '\\') {
+ *BOut++ = '\\'; // Two \ becomes one
+ BIn += 2;
+ } else if (BIn < EndBuffer-2 && isxdigit(BIn[1]) && isxdigit(BIn[2])) {
+ char Tmp = BIn[3]; BIn[3] = 0; // Terminate string
+ *BOut = (char)strtol(BIn+1, 0, 16); // Convert to number
+ BIn[3] = Tmp; // Restore character
+ BIn += 3; // Skip over handled chars
+ ++BOut;
+ } else {
+ *BOut++ = *BIn++;
+ }
} else {
*BOut++ = *BIn++;
}
}
-
return BOut;
}
/* 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]*
+/* Local Values and Type identifiers start with a % sign */
+LocalVarName %[-a-zA-Z$._][-a-zA-Z$._0-9]*
+
+/* Global Value identifiers start with an @ sign */
+GlobalVarName @[-a-zA-Z$._][-a-zA-Z$._0-9]*
/* Label identifiers end with a colon */
Label [-a-zA-Z$._0-9]+:
/* Quoted names can contain any character except " and \ */
StringConstant \"[^\"]*\"
+AtStringConstant @\"[^\"]*\"
+PctStringConstant %\"[^\"]*\"
+
+/* LocalVarID/GlobalVarID: match an unnamed local variable slot ID. */
+LocalVarID %[0-9]+
+GlobalVarID @[0-9]+
+/* Integer types are specified with i and a bitwidth */
+IntegerType i[0-9]+
-/* [PN]Integer: match positive and negative literal integer values that
- * are preceeded by a '%' character. These represent unnamed variable slots.
- */
-EPInteger %[0-9]+
-ENInteger %-[0-9]+
-
-
-/* E[PN]Integer: match positive and negative literal integer values */
+/* E[PN]Integer: match positive and negative literal integer values. */
PInteger [0-9]+
NInteger -[0-9]+
-/* FPConstant - A Floating point constant.
+/* FPConstant - A Floating point constant. Float and double only.
*/
FPConstant [-+]?[0-9]+[.][0-9]*([eE][-+]?[0-9]+)?
/* HexFPConstant - Floating point constant represented in IEEE format as a
* hexadecimal number for when exponential notation is not precise enough.
+ * Float and double only.
*/
HexFPConstant 0x[0-9A-Fa-f]+
+/* F80HexFPConstant - x87 long double in hexadecimal format (10 bytes)
+ */
+HexFP80Constant 0xK[0-9A-Fa-f]+
+
+/* F128HexFPConstant - IEEE 128-bit in hexadecimal format (16 bytes)
+ */
+HexFP128Constant 0xL[0-9A-Fa-f]+
+
+/* PPC128HexFPConstant - PowerPC 128-bit in hexadecimal format (16 bytes)
+ */
+HexPPC128Constant 0xM[0-9A-Fa-f]+
+
/* HexIntConstant - Hexadecimal constant generated by the CFE to avoid forcing
* it to deal with 64 bit numbers.
*/
HexIntConstant [us]0x[0-9A-Fa-f]+
+
+/* WSNL - shorthand for whitespace followed by newline */
+WSNL [ \r\t]*$
%%
{Comment} { /* Ignore comments for now */ }
true { return TRUETOK; }
false { return FALSETOK; }
declare { return DECLARE; }
+define { return DEFINE; }
global { return GLOBAL; }
constant { return CONSTANT; }
internal { return INTERNAL; }
linkonce { return LINKONCE; }
weak { return WEAK; }
appending { return APPENDING; }
-uninitialized { return EXTERNAL; } /* Deprecated, turn into external */
+dllimport { return DLLIMPORT; }
+dllexport { return DLLEXPORT; }
+hidden { return HIDDEN; }
+protected { return PROTECTED; }
+extern_weak { return EXTERN_WEAK; }
external { return EXTERNAL; }
-implementation { return IMPLEMENTATION; }
+thread_local { return THREAD_LOCAL; }
zeroinitializer { return ZEROINITIALIZER; }
\.\.\. { return DOTDOTDOT; }
undef { return UNDEF; }
null { return NULL_TOK; }
to { return TO; }
-except { RET_TOK(TermOpVal, Unwind, UNWIND); }
-not { return NOT; } /* Deprecated, turned into XOR */
tail { return TAIL; }
target { return TARGET; }
triple { return TRIPLE; }
deplibs { return DEPLIBS; }
-endian { return ENDIAN; }
-pointersize { return POINTERSIZE; }
-little { return LITTLE; }
-big { return BIG; }
+datalayout { return DATALAYOUT; }
volatile { return VOLATILE; }
align { return ALIGN; }
section { return SECTION; }
+alias { return ALIAS; }
module { return MODULE; }
asm { return ASM_TOK; }
sideeffect { return SIDEEFFECT; }
ccc { return CCC_TOK; }
fastcc { return FASTCC_TOK; }
coldcc { return COLDCC_TOK; }
-
-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; }
-label { llvmAsmlval.PrimType = Type::LabelTy ; return LABEL; }
+x86_stdcallcc { return X86_STDCALLCC_TOK; }
+x86_fastcallcc { return X86_FASTCALLCC_TOK; }
+
+signext { return SIGNEXT; }
+zeroext { return ZEROEXT; }
+inreg { return INREG; }
+sret { return SRET; }
+nounwind { return NOUNWIND; }
+noreturn { return NORETURN; }
+noalias { return NOALIAS; }
+byval { return BYVAL; }
+nest { return NEST; }
+sext{WSNL} { // For auto-upgrade only, drop in LLVM 3.0
+ return SIGNEXT; }
+zext{WSNL} { // For auto-upgrade only, drop in LLVM 3.0
+ return ZEROEXT; }
+
+void { RET_TY(Type::VoidTy, VOID); }
+float { RET_TY(Type::FloatTy, FLOAT); }
+double { RET_TY(Type::DoubleTy,DOUBLE);}
+x86_fp80 { RET_TY(Type::X86_FP80Ty, X86_FP80);}
+fp128 { RET_TY(Type::FP128Ty, FP128);}
+ppc_fp128 { RET_TY(Type::PPC_FP128Ty, PPC_FP128);}
+label { RET_TY(Type::LabelTy, LABEL); }
type { return TYPE; }
opaque { return OPAQUE; }
+{IntegerType} { uint64_t NumBits = atoull(yytext+1);
+ if (NumBits < IntegerType::MIN_INT_BITS ||
+ NumBits > IntegerType::MAX_INT_BITS)
+ GenerateError("Bitwidth for integer type out of range!");
+ const Type* Ty = IntegerType::get(NumBits);
+ RET_TY(Ty, INTTYPE);
+ }
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); }
+udiv { RET_TOK(BinaryOpVal, UDiv, UDIV); }
+sdiv { RET_TOK(BinaryOpVal, SDiv, SDIV); }
+fdiv { RET_TOK(BinaryOpVal, FDiv, FDIV); }
+urem { RET_TOK(BinaryOpVal, URem, UREM); }
+srem { RET_TOK(BinaryOpVal, SRem, SREM); }
+frem { RET_TOK(BinaryOpVal, FRem, FREM); }
+shl { RET_TOK(BinaryOpVal, Shl, SHL); }
+lshr { RET_TOK(BinaryOpVal, LShr, LSHR); }
+ashr { RET_TOK(BinaryOpVal, AShr, ASHR); }
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); }
+icmp { RET_TOK(OtherOpVal, ICmp, ICMP); }
+fcmp { RET_TOK(OtherOpVal, FCmp, FCMP); }
+
+eq { return EQ; }
+ne { return NE; }
+slt { return SLT; }
+sgt { return SGT; }
+sle { return SLE; }
+sge { return SGE; }
+ult { return ULT; }
+ugt { return UGT; }
+ule { return ULE; }
+uge { return UGE; }
+oeq { return OEQ; }
+one { return ONE; }
+olt { return OLT; }
+ogt { return OGT; }
+ole { return OLE; }
+oge { return OGE; }
+ord { return ORD; }
+uno { return UNO; }
+ueq { return UEQ; }
+une { return UNE; }
phi { RET_TOK(OtherOpVal, PHI, PHI_TOK); }
call { RET_TOK(OtherOpVal, Call, CALL); }
-cast { RET_TOK(OtherOpVal, Cast, CAST); }
+trunc { RET_TOK(CastOpVal, Trunc, TRUNC); }
+zext { RET_TOK(CastOpVal, ZExt, ZEXT); }
+sext { RET_TOK(CastOpVal, SExt, SEXT); }
+fptrunc { RET_TOK(CastOpVal, FPTrunc, FPTRUNC); }
+fpext { RET_TOK(CastOpVal, FPExt, FPEXT); }
+uitofp { RET_TOK(CastOpVal, UIToFP, UITOFP); }
+sitofp { RET_TOK(CastOpVal, SIToFP, SITOFP); }
+fptoui { RET_TOK(CastOpVal, FPToUI, FPTOUI); }
+fptosi { RET_TOK(CastOpVal, FPToSI, FPTOSI); }
+inttoptr { RET_TOK(CastOpVal, IntToPtr, INTTOPTR); }
+ptrtoint { RET_TOK(CastOpVal, PtrToInt, PTRTOINT); }
+bitcast { RET_TOK(CastOpVal, BitCast, BITCAST); }
select { RET_TOK(OtherOpVal, Select, SELECT); }
-shl { RET_TOK(OtherOpVal, Shl, SHL); }
-shr { RET_TOK(OtherOpVal, Shr, SHR); }
-vanext { return VANEXT_old; }
-vaarg { return VAARG_old; }
va_arg { RET_TOK(OtherOpVal, VAArg , VAARG); }
ret { RET_TOK(TermOpVal, Ret, RET); }
br { RET_TOK(TermOpVal, Br, BR); }
extractelement { RET_TOK(OtherOpVal, ExtractElement, EXTRACTELEMENT); }
insertelement { RET_TOK(OtherOpVal, InsertElement, INSERTELEMENT); }
+shufflevector { RET_TOK(OtherOpVal, ShuffleVector, SHUFFLEVECTOR); }
-{VarID} {
- UnEscapeLexed(yytext+1);
- llvmAsmlval.StrVal = strdup(yytext+1); // Skip %
- return VAR_ID;
+{LocalVarName} {
+ llvmAsmlval.StrVal = new std::string(yytext+1); // Skip %
+ return LOCALVAR;
+ }
+{GlobalVarName} {
+ llvmAsmlval.StrVal = new std::string(yytext+1); // Skip @
+ return GLOBALVAR;
}
{Label} {
- yytext[strlen(yytext)-1] = 0; // nuke colon
- UnEscapeLexed(yytext);
- llvmAsmlval.StrVal = strdup(yytext);
+ yytext[yyleng-1] = 0; // nuke colon
+ llvmAsmlval.StrVal = new std::string(yytext);
return LABELSTR;
}
{QuoteLabel} {
- yytext[strlen(yytext)-2] = 0; // nuke colon, end quote
- UnEscapeLexed(yytext+1);
- llvmAsmlval.StrVal = strdup(yytext+1);
+ yytext[yyleng-2] = 0; // nuke colon, end quote
+ const char* EndChar = UnEscapeLexed(yytext+1, yytext+yyleng);
+ llvmAsmlval.StrVal =
+ new std::string(yytext+1, EndChar - yytext - 1);
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
+{StringConstant} { yytext[yyleng-1] = 0; // nuke end quote
+ const char* EndChar = UnEscapeLexed(yytext+1, yytext+yyleng);
+ llvmAsmlval.StrVal =
+ new std::string(yytext+1, EndChar - yytext - 1);
return STRINGCONSTANT;
}
-
-
-{PInteger} { llvmAsmlval.UInt64Val = atoull(yytext); return EUINT64VAL; }
+{AtStringConstant} {
+ yytext[yyleng-1] = 0; // nuke end quote
+ const char* EndChar =
+ UnEscapeLexed(yytext+2, yytext+yyleng);
+ llvmAsmlval.StrVal =
+ new std::string(yytext+2, EndChar - yytext - 2);
+ return ATSTRINGCONSTANT;
+ }
+{PctStringConstant} {
+ yytext[yyleng-1] = 0; // nuke end quote
+ const char* EndChar =
+ UnEscapeLexed(yytext+2, yytext+yyleng);
+ llvmAsmlval.StrVal =
+ new std::string(yytext+2, EndChar - yytext - 2);
+ return PCTSTRINGCONSTANT;
+ }
+{PInteger} {
+ uint32_t numBits = ((yyleng * 64) / 19) + 1;
+ APInt Tmp(numBits, yytext, yyleng, 10);
+ uint32_t activeBits = Tmp.getActiveBits();
+ if (activeBits > 0 && activeBits < numBits)
+ Tmp.trunc(activeBits);
+ if (Tmp.getBitWidth() > 64) {
+ llvmAsmlval.APIntVal = new APInt(Tmp);
+ return EUAPINTVAL;
+ } else {
+ llvmAsmlval.UInt64Val = Tmp.getZExtValue();
+ 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;
+ uint32_t numBits = (((yyleng-1) * 64) / 19) + 2;
+ APInt Tmp(numBits, yytext, yyleng, 10);
+ uint32_t minBits = Tmp.getMinSignedBits();
+ if (minBits > 0 && minBits < numBits)
+ Tmp.trunc(minBits);
+ if (Tmp.getBitWidth() > 64) {
+ llvmAsmlval.APIntVal = new APInt(Tmp);
+ return ESAPINTVAL;
+ } else {
+ llvmAsmlval.SInt64Val = Tmp.getSExtValue();
+ return ESINT64VAL;
+ }
}
-{HexIntConstant} {
- llvmAsmlval.UInt64Val = HexIntToVal(yytext+3);
- return yytext[0] == 's' ? ESINT64VAL : EUINT64VAL;
+
+{HexIntConstant} { int len = yyleng - 3;
+ uint32_t bits = len * 4;
+ APInt Tmp(bits, yytext+3, len, 16);
+ uint32_t activeBits = Tmp.getActiveBits();
+ if (activeBits > 0 && activeBits < bits)
+ Tmp.trunc(activeBits);
+ if (Tmp.getBitWidth() > 64) {
+ llvmAsmlval.APIntVal = new APInt(Tmp);
+ return yytext[0] == 's' ? ESAPINTVAL : EUAPINTVAL;
+ } else if (yytext[0] == 's') {
+ llvmAsmlval.SInt64Val = Tmp.getSExtValue();
+ return ESINT64VAL;
+ } else {
+ llvmAsmlval.UInt64Val = Tmp.getZExtValue();
+ return EUINT64VAL;
+ }
}
-{EPInteger} {
+{LocalVarID} {
uint64_t Val = atoull(yytext+1);
if ((unsigned)Val != Val)
- ThrowException("Invalid value number (too large)!");
+ GenerateError("Invalid value number (too large)!");
llvmAsmlval.UIntVal = unsigned(Val);
- return UINTVAL;
+ return LOCALVAL_ID;
}
-{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 = (int)-Val;
- return SINTVAL;
+{GlobalVarID} {
+ uint64_t Val = atoull(yytext+1);
+ if ((unsigned)Val != Val)
+ GenerateError("Invalid value number (too large)!");
+ llvmAsmlval.UIntVal = unsigned(Val);
+ return GLOBALVAL_ID;
}
-{FPConstant} { llvmAsmlval.FPVal = atof(yytext); return FPVAL; }
-{HexFPConstant} { llvmAsmlval.FPVal = HexToFP(yytext); return FPVAL; }
+{FPConstant} { llvmAsmlval.FPVal = new APFloat(atof(yytext)); return FPVAL; }
+{HexFPConstant} { llvmAsmlval.FPVal = new APFloat(HexToFP(yytext+2));
+ return FPVAL;
+ }
+{HexFP80Constant} { uint64_t Pair[2];
+ HexToIntPair(yytext+3, Pair);
+ llvmAsmlval.FPVal = new APFloat(APInt(80, 2, Pair));
+ return FPVAL;
+ }
+{HexFP128Constant} { uint64_t Pair[2];
+ HexToIntPair(yytext+3, Pair);
+ llvmAsmlval.FPVal = new APFloat(APInt(128, 2, Pair), true);
+ return FPVAL;
+ }
+{HexPPC128Constant} { uint64_t Pair[2];
+ HexToIntPair(yytext+3, Pair);
+ llvmAsmlval.FPVal = new APFloat(APInt(128, 2, Pair));
+ return FPVAL;
+ }
<<EOF>> {
/* Make sure to free the internal buffers for flex when we are
. { return yytext[0]; }
%%
-