#include "X86AsmInstrumentation.h"
#include "X86AsmParserCommon.h"
#include "X86Operand.h"
+#include "X86ISelLowering.h"
#include "llvm/ADT/APFloat.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallString.h"
static const char OpPrecedence[] = {
0, // IC_OR
- 1, // IC_AND
- 2, // IC_LSHIFT
- 2, // IC_RSHIFT
- 3, // IC_PLUS
- 3, // IC_MINUS
- 4, // IC_MULTIPLY
- 4, // IC_DIVIDE
- 5, // IC_RPAREN
- 6, // IC_LPAREN
+ 1, // IC_XOR
+ 2, // IC_AND
+ 3, // IC_LSHIFT
+ 3, // IC_RSHIFT
+ 4, // IC_PLUS
+ 4, // IC_MINUS
+ 5, // IC_MULTIPLY
+ 5, // IC_DIVIDE
+ 6, // IC_RPAREN
+ 7, // IC_LPAREN
0, // IC_IMM
0 // IC_REGISTER
};
class X86AsmParser : public MCTargetAsmParser {
MCSubtargetInfo &STI;
- MCAsmParser &Parser;
const MCInstrInfo &MII;
ParseInstructionInfo *InstInfo;
std::unique_ptr<X86AsmInstrumentation> Instrumentation;
private:
SMLoc consumeToken() {
+ MCAsmParser &Parser = getParser();
SMLoc Result = Parser.getTok().getLoc();
Parser.Lex();
return Result;
enum InfixCalculatorTok {
IC_OR = 0,
+ IC_XOR,
IC_AND,
IC_LSHIFT,
IC_RSHIFT,
typedef std::pair< InfixCalculatorTok, int64_t > ICToken;
SmallVector<InfixCalculatorTok, 4> InfixOperatorStack;
SmallVector<ICToken, 4> PostfixStack;
-
+
public:
int64_t popOperand() {
assert (!PostfixStack.empty() && "Poped an empty stack!");
"Unexpected operand!");
PostfixStack.push_back(std::make_pair(Op, Val));
}
-
+
void popOperator() { InfixOperatorStack.pop_back(); }
void pushOperator(InfixCalculatorTok Op) {
// Push the new operator if the stack is empty.
InfixOperatorStack.push_back(Op);
return;
}
-
+
// Push the new operator if it has a higher precedence than the operator
// on the top of the stack or the operator on the top of the stack is a
// left parentheses.
InfixOperatorStack.push_back(Op);
return;
}
-
+
// The operator on the top of the stack has higher precedence than the
// new operator.
unsigned ParenCount = 0;
// Nothing to process.
if (InfixOperatorStack.empty())
break;
-
+
Idx = InfixOperatorStack.size() - 1;
StackOp = InfixOperatorStack[Idx];
if (!(OpPrecedence[StackOp] >= OpPrecedence[Op] || ParenCount))
break;
-
+
// If we have an even parentheses count and we see a left parentheses,
// then stop processing.
if (!ParenCount && StackOp == IC_LPAREN)
break;
-
+
if (StackOp == IC_RPAREN) {
++ParenCount;
InfixOperatorStack.pop_back();
// Push the new operator.
InfixOperatorStack.push_back(Op);
}
+
int64_t execute() {
// Push any remaining operators onto the postfix stack.
while (!InfixOperatorStack.empty()) {
if (StackOp != IC_LPAREN && StackOp != IC_RPAREN)
PostfixStack.push_back(std::make_pair(StackOp, 0));
}
-
+
if (PostfixStack.empty())
return 0;
-
+
SmallVector<ICToken, 16> OperandStack;
for (unsigned i = 0, e = PostfixStack.size(); i != e; ++i) {
ICToken Op = PostfixStack[i];
Val = Op1.second | Op2.second;
OperandStack.push_back(std::make_pair(IC_IMM, Val));
break;
+ case IC_XOR:
+ assert(Op1.first == IC_IMM && Op2.first == IC_IMM &&
+ "Xor operation with an immediate and a register!");
+ Val = Op1.second ^ Op2.second;
+ OperandStack.push_back(std::make_pair(IC_IMM, Val));
+ break;
case IC_AND:
assert (Op1.first == IC_IMM && Op2.first == IC_IMM &&
"And operation with an immediate and a register!");
enum IntelExprState {
IES_OR,
+ IES_XOR,
IES_AND,
IES_LSHIFT,
IES_RSHIFT,
State(IES_PLUS), PrevState(IES_ERROR), BaseReg(0), IndexReg(0), TmpReg(0),
Scale(1), Imm(imm), Sym(nullptr), StopOnLBrac(stoponlbrac),
AddImmPrefix(addimmprefix) { Info.clear(); }
-
+
unsigned getBaseReg() { return BaseReg; }
unsigned getIndexReg() { return IndexReg; }
unsigned getScale() { return Scale; }
}
PrevState = CurrState;
}
+ void onXor() {
+ IntelExprState CurrState = State;
+ switch (State) {
+ default:
+ State = IES_ERROR;
+ break;
+ case IES_INTEGER:
+ case IES_RPAREN:
+ case IES_REGISTER:
+ State = IES_XOR;
+ IC.pushOperator(IC_XOR);
+ break;
+ }
+ PrevState = CurrState;
+ }
void onAnd() {
IntelExprState CurrState = State;
switch (State) {
case IES_MINUS:
case IES_NOT:
case IES_OR:
+ case IES_XOR:
case IES_AND:
case IES_LSHIFT:
case IES_RSHIFT:
PrevState == IES_LSHIFT || PrevState == IES_RSHIFT ||
PrevState == IES_MULTIPLY || PrevState == IES_DIVIDE ||
PrevState == IES_LPAREN || PrevState == IES_LBRAC ||
- PrevState == IES_NOT) &&
+ PrevState == IES_NOT || PrevState == IES_XOR) &&
CurrState == IES_MINUS) {
// Unary minus. No need to pop the minus operand because it was never
// pushed.
PrevState == IES_LSHIFT || PrevState == IES_RSHIFT ||
PrevState == IES_MULTIPLY || PrevState == IES_DIVIDE ||
PrevState == IES_LPAREN || PrevState == IES_LBRAC ||
- PrevState == IES_NOT) &&
+ PrevState == IES_NOT || PrevState == IES_XOR) &&
CurrState == IES_NOT) {
// Unary not. No need to pop the not operand because it was never
// pushed.
case IES_MINUS:
case IES_NOT:
case IES_OR:
+ case IES_XOR:
case IES_AND:
case IES_LSHIFT:
case IES_RSHIFT:
PrevState == IES_LSHIFT || PrevState == IES_RSHIFT ||
PrevState == IES_MULTIPLY || PrevState == IES_DIVIDE ||
PrevState == IES_LPAREN || PrevState == IES_LBRAC ||
- PrevState == IES_NOT) &&
+ PrevState == IES_NOT || PrevState == IES_XOR) &&
(CurrState == IES_MINUS || CurrState == IES_NOT)) {
State = IES_ERROR;
break;
}
};
- MCAsmParser &getParser() const { return Parser; }
-
- MCAsmLexer &getLexer() const { return Parser.getLexer(); }
-
bool Error(SMLoc L, const Twine &Msg,
ArrayRef<SMRange> Ranges = None,
bool MatchingInlineAsm = false) {
+ MCAsmParser &Parser = getParser();
if (MatchingInlineAsm) return true;
return Parser.Error(L, Msg, Ranges);
}
bool ErrorAndEatStatement(SMLoc L, const Twine &Msg,
ArrayRef<SMRange> Ranges = None,
bool MatchingInlineAsm = false) {
- Parser.eatToEndOfStatement();
- return Error(L, Msg, Ranges, MatchingInlineAsm);
+ MCAsmParser &Parser = getParser();
+ Parser.eatToEndOfStatement();
+ return Error(L, Msg, Ranges, MatchingInlineAsm);
}
std::nullptr_t ErrorOperand(SMLoc Loc, StringRef Msg) {
std::unique_ptr<X86Operand> DefaultMemSIOperand(SMLoc Loc);
std::unique_ptr<X86Operand> DefaultMemDIOperand(SMLoc Loc);
+ void AddDefaultSrcDestOperands(
+ OperandVector& Operands, std::unique_ptr<llvm::MCParsedAsmOperand> &&Src,
+ std::unique_ptr<llvm::MCParsedAsmOperand> &&Dst);
std::unique_ptr<X86Operand> ParseOperand();
std::unique_ptr<X86Operand> ParseATTOperand();
std::unique_ptr<X86Operand> ParseIntelOperand();
ParseIntelSegmentOverride(unsigned SegReg, SMLoc Start, unsigned Size);
std::unique_ptr<X86Operand>
ParseIntelMemOperand(int64_t ImmDisp, SMLoc StartLoc, unsigned Size);
+ std::unique_ptr<X86Operand> ParseRoundingModeOp(SMLoc Start, SMLoc End);
bool ParseIntelExpression(IntelExprStateMachine &SM, SMLoc &End);
std::unique_ptr<X86Operand> ParseIntelBracExpression(unsigned SegReg,
SMLoc Start,
bool ParseDirectiveWord(unsigned Size, SMLoc L);
bool ParseDirectiveCode(StringRef IDVal, SMLoc L);
+ bool validateInstruction(MCInst &Inst, const OperandVector &Ops);
bool processInstruction(MCInst &Inst, const OperandVector &Ops);
/// Wrapper around MCStreamer::EmitInstruction(). Possibly adds
uint64_t &ErrorInfo,
bool MatchingInlineAsm);
- unsigned getPointerSize() {
- if (is16BitMode()) return 16;
- if (is32BitMode()) return 32;
- if (is64BitMode()) return 64;
- llvm_unreachable("invalid mode");
- }
-
bool OmitRegisterFromClobberLists(unsigned RegNo) override;
/// doSrcDstMatch - Returns true if operands are matching in their
bool is64BitMode() const {
// FIXME: Can tablegen auto-generate this?
- return (STI.getFeatureBits() & X86::Mode64Bit) != 0;
+ return STI.getFeatureBits()[X86::Mode64Bit];
}
bool is32BitMode() const {
// FIXME: Can tablegen auto-generate this?
- return (STI.getFeatureBits() & X86::Mode32Bit) != 0;
+ return STI.getFeatureBits()[X86::Mode32Bit];
}
bool is16BitMode() const {
// FIXME: Can tablegen auto-generate this?
- return (STI.getFeatureBits() & X86::Mode16Bit) != 0;
+ return STI.getFeatureBits()[X86::Mode16Bit];
}
- void SwitchMode(uint64_t mode) {
- uint64_t oldMode = STI.getFeatureBits() &
- (X86::Mode64Bit | X86::Mode32Bit | X86::Mode16Bit);
- unsigned FB = ComputeAvailableFeatures(STI.ToggleFeature(oldMode | mode));
+ void SwitchMode(unsigned mode) {
+ FeatureBitset AllModes({X86::Mode64Bit, X86::Mode32Bit, X86::Mode16Bit});
+ FeatureBitset OldMode = STI.getFeatureBits() & AllModes;
+ unsigned FB = ComputeAvailableFeatures(
+ STI.ToggleFeature(OldMode.flip(mode)));
setAvailableFeatures(FB);
- assert(mode == (STI.getFeatureBits() &
- (X86::Mode64Bit | X86::Mode32Bit | X86::Mode16Bit)));
+
+ assert(FeatureBitset({mode}) == (STI.getFeatureBits() & AllModes));
}
unsigned getPointerWidth() {
/// }
public:
- X86AsmParser(MCSubtargetInfo &sti, MCAsmParser &parser,
- const MCInstrInfo &mii,
- const MCTargetOptions &Options)
- : MCTargetAsmParser(), STI(sti), Parser(parser), MII(mii),
- InstInfo(nullptr) {
+ X86AsmParser(MCSubtargetInfo &sti, MCAsmParser &Parser,
+ const MCInstrInfo &mii, const MCTargetOptions &Options)
+ : MCTargetAsmParser(Options), STI(sti), MII(mii), InstInfo(nullptr) {
// Initialize the set of available features.
setAvailableFeatures(ComputeAvailableFeatures(STI.getFeatureBits()));
bool X86AsmParser::ParseRegister(unsigned &RegNo,
SMLoc &StartLoc, SMLoc &EndLoc) {
+ MCAsmParser &Parser = getParser();
RegNo = 0;
const AsmToken &PercentTok = Parser.getTok();
StartLoc = PercentTok.getLoc();
if (RegNo == 0)
RegNo = MatchRegisterName(Tok.getString().lower());
+ // The "flags" register cannot be referenced directly.
+ // Treat it as an identifier instead.
+ if (isParsingInlineAsm() && isParsingIntelSyntax() && RegNo == X86::EFLAGS)
+ RegNo = 0;
+
if (!is64BitMode()) {
// FIXME: This should be done using Requires<Not64BitMode> and
// Requires<In64BitMode> so "eiz" usage in 64-bit instructions can be also
std::unique_ptr<X86Operand> X86AsmParser::DefaultMemSIOperand(SMLoc Loc) {
unsigned basereg =
is64BitMode() ? X86::RSI : (is32BitMode() ? X86::ESI : X86::SI);
- const MCExpr *Disp = MCConstantExpr::Create(0, getContext());
- return X86Operand::CreateMem(/*SegReg=*/0, Disp, /*BaseReg=*/basereg,
- /*IndexReg=*/0, /*Scale=*/1, Loc, Loc, 0);
+ const MCExpr *Disp = MCConstantExpr::create(0, getContext());
+ return X86Operand::CreateMem(getPointerWidth(), /*SegReg=*/0, Disp,
+ /*BaseReg=*/basereg, /*IndexReg=*/0, /*Scale=*/1,
+ Loc, Loc, 0);
}
std::unique_ptr<X86Operand> X86AsmParser::DefaultMemDIOperand(SMLoc Loc) {
unsigned basereg =
is64BitMode() ? X86::RDI : (is32BitMode() ? X86::EDI : X86::DI);
- const MCExpr *Disp = MCConstantExpr::Create(0, getContext());
- return X86Operand::CreateMem(/*SegReg=*/0, Disp, /*BaseReg=*/basereg,
- /*IndexReg=*/0, /*Scale=*/1, Loc, Loc, 0);
+ const MCExpr *Disp = MCConstantExpr::create(0, getContext());
+ return X86Operand::CreateMem(getPointerWidth(), /*SegReg=*/0, Disp,
+ /*BaseReg=*/basereg, /*IndexReg=*/0, /*Scale=*/1,
+ Loc, Loc, 0);
+}
+
+void X86AsmParser::AddDefaultSrcDestOperands(
+ OperandVector& Operands, std::unique_ptr<llvm::MCParsedAsmOperand> &&Src,
+ std::unique_ptr<llvm::MCParsedAsmOperand> &&Dst) {
+ if (isParsingIntelSyntax()) {
+ Operands.push_back(std::move(Dst));
+ Operands.push_back(std::move(Src));
+ }
+ else {
+ Operands.push_back(std::move(Src));
+ Operands.push_back(std::move(Dst));
+ }
}
std::unique_ptr<X86Operand> X86AsmParser::ParseOperand() {
.Cases("DWORD", "dword", 32)
.Cases("QWORD", "qword", 64)
.Cases("XWORD", "xword", 80)
+ //
+ .Cases("TBYTE", "tbyte", 80)
.Cases("XMMWORD", "xmmword", 128)
.Cases("YMMWORD", "ymmword", 256)
.Cases("ZMMWORD", "zmmword", 512)
// Create an absolute memory reference in order to match against
// instructions taking a PC relative operand.
- return X86Operand::CreateMem(Disp, Start, End, Size, Identifier,
- Info.OpDecl);
+ return X86Operand::CreateMem(getPointerWidth(), Disp, Start, End, Size,
+ Identifier, Info.OpDecl);
}
// We either have a direct symbol reference, or an offset from a symbol. The
// if we don't know the actual value at this time. This is necessary to
// get the matching correct in some cases.
BaseReg = BaseReg ? BaseReg : 1;
- return X86Operand::CreateMem(SegReg, Disp, BaseReg, IndexReg, Scale, Start,
- End, Size, Identifier, Info.OpDecl);
+ return X86Operand::CreateMem(getPointerWidth(), SegReg, Disp, BaseReg,
+ IndexReg, Scale, Start, End, Size, Identifier,
+ Info.OpDecl);
}
static void
(*I).Kind = AOK_Delete;
}
const char *SymLocPtr = SymName.data();
- // Skip everything before the symbol.
+ // Skip everything before the symbol.
if (unsigned Len = SymLocPtr - StartInBrac.getPointer()) {
assert(Len > 0 && "Expected a non-negative length.");
AsmRewrites->push_back(AsmRewrite(AOK_Skip, StartInBrac, Len));
}
bool X86AsmParser::ParseIntelExpression(IntelExprStateMachine &SM, SMLoc &End) {
+ MCAsmParser &Parser = getParser();
const AsmToken &Tok = Parser.getTok();
bool Done = false;
// identifier. Don't try an parse it as a register.
if (Tok.getString().startswith("."))
break;
-
+
// If we're parsing an immediate expression, we don't expect a '['.
if (SM.getStopOnLBrac() && getLexer().getKind() == AsmToken::LBrac)
break;
StringRef IDVal = getTok().getString();
if (IDVal == "f" || IDVal == "b") {
MCSymbol *Sym =
- getContext().GetDirectionalLocalSymbol(IntVal, IDVal == "b");
+ getContext().getDirectionalLocalSymbol(IntVal, IDVal == "b");
MCSymbolRefExpr::VariantKind Variant = MCSymbolRefExpr::VK_None;
- const MCExpr *Val =
- MCSymbolRefExpr::Create(Sym, Variant, getContext());
+ const MCExpr *Val =
+ MCSymbolRefExpr::create(Sym, Variant, getContext());
if (IDVal == "b" && Sym->isUndefined())
return Error(Loc, "invalid reference to undefined symbol");
StringRef Identifier = Sym->getName();
case AsmToken::Star: SM.onStar(); break;
case AsmToken::Slash: SM.onDivide(); break;
case AsmToken::Pipe: SM.onOr(); break;
+ case AsmToken::Caret: SM.onXor(); break;
case AsmToken::Amp: SM.onAnd(); break;
case AsmToken::LessLess:
SM.onLShift(); break;
std::unique_ptr<X86Operand>
X86AsmParser::ParseIntelBracExpression(unsigned SegReg, SMLoc Start,
int64_t ImmDisp, unsigned Size) {
+ MCAsmParser &Parser = getParser();
const AsmToken &Tok = Parser.getTok();
SMLoc BracLoc = Tok.getLoc(), End = Tok.getEndLoc();
if (getLexer().isNot(AsmToken::LBrac))
}
if (SM.getImm() || !Disp) {
- const MCExpr *Imm = MCConstantExpr::Create(SM.getImm(), getContext());
+ const MCExpr *Imm = MCConstantExpr::create(SM.getImm(), getContext());
if (Disp)
- Disp = MCBinaryExpr::CreateAdd(Disp, Imm, getContext());
+ Disp = MCBinaryExpr::createAdd(Disp, Imm, getContext());
else
Disp = Imm; // An immediate displacement only.
}
const MCExpr *NewDisp;
if (ParseIntelDotOperator(Disp, NewDisp))
return nullptr;
-
+
End = Tok.getEndLoc();
Parser.Lex(); // Eat the field.
Disp = NewDisp;
// handle [-42]
if (!BaseReg && !IndexReg) {
if (!SegReg)
- return X86Operand::CreateMem(Disp, Start, End, Size);
- else
- return X86Operand::CreateMem(SegReg, Disp, 0, 0, 1, Start, End, Size);
+ return X86Operand::CreateMem(getPointerWidth(), Disp, Start, End, Size);
+ return X86Operand::CreateMem(getPointerWidth(), SegReg, Disp, 0, 0, 1,
+ Start, End, Size);
}
StringRef ErrMsg;
if (CheckBaseRegAndIndexReg(BaseReg, IndexReg, ErrMsg)) {
Error(StartInBrac, ErrMsg);
return nullptr;
}
- return X86Operand::CreateMem(SegReg, Disp, BaseReg, IndexReg, Scale, Start,
- End, Size);
+ return X86Operand::CreateMem(getPointerWidth(), SegReg, Disp, BaseReg,
+ IndexReg, Scale, Start, End, Size);
}
InlineAsmIdentifierInfo &Info = SM.getIdentifierInfo();
StringRef &Identifier,
InlineAsmIdentifierInfo &Info,
bool IsUnevaluatedOperand, SMLoc &End) {
+ MCAsmParser &Parser = getParser();
assert (isParsingInlineAsm() && "Expected to be parsing inline assembly.");
Val = nullptr;
}
// Create the symbol reference.
- MCSymbol *Sym = getContext().GetOrCreateSymbol(Identifier);
+ MCSymbol *Sym = getContext().getOrCreateSymbol(Identifier);
MCSymbolRefExpr::VariantKind Variant = MCSymbolRefExpr::VK_None;
- Val = MCSymbolRefExpr::Create(Sym, Variant, getParser().getContext());
+ Val = MCSymbolRefExpr::create(Sym, Variant, getParser().getContext());
return false;
}
std::unique_ptr<X86Operand>
X86AsmParser::ParseIntelSegmentOverride(unsigned SegReg, SMLoc Start,
unsigned Size) {
+ MCAsmParser &Parser = getParser();
assert(SegReg != 0 && "Tried to parse a segment override without a segment!");
const AsmToken &Tok = Parser.getTok(); // Eat colon.
if (Tok.isNot(AsmToken::Colon))
// An immediate following a 'segment register', 'colon' token sequence can
// be followed by a bracketed expression. If it isn't we know we have our
// final segment override.
- const MCExpr *Disp = MCConstantExpr::Create(ImmDisp, getContext());
- return X86Operand::CreateMem(SegReg, Disp, /*BaseReg=*/0, /*IndexReg=*/0,
- /*Scale=*/1, Start, ImmDispToken.getEndLoc(),
- Size);
+ const MCExpr *Disp = MCConstantExpr::create(ImmDisp, getContext());
+ return X86Operand::CreateMem(getPointerWidth(), SegReg, Disp,
+ /*BaseReg=*/0, /*IndexReg=*/0, /*Scale=*/1,
+ Start, ImmDispToken.getEndLoc(), Size);
}
}
if (getParser().parsePrimaryExpr(Val, End))
return ErrorOperand(Tok.getLoc(), "unknown token in expression");
- return X86Operand::CreateMem(Val, Start, End, Size);
+ return X86Operand::CreateMem(getPointerWidth(), Val, Start, End, Size);
}
InlineAsmIdentifierInfo Info;
/*Scale=*/1, Start, End, Size, Identifier, Info);
}
+//ParseRoundingModeOp - Parse AVX-512 rounding mode operand
+std::unique_ptr<X86Operand>
+X86AsmParser::ParseRoundingModeOp(SMLoc Start, SMLoc End) {
+ MCAsmParser &Parser = getParser();
+ const AsmToken &Tok = Parser.getTok();
+ // Eat "{" and mark the current place.
+ const SMLoc consumedToken = consumeToken();
+ if (Tok.getIdentifier().startswith("r")){
+ int rndMode = StringSwitch<int>(Tok.getIdentifier())
+ .Case("rn", X86::STATIC_ROUNDING::TO_NEAREST_INT)
+ .Case("rd", X86::STATIC_ROUNDING::TO_NEG_INF)
+ .Case("ru", X86::STATIC_ROUNDING::TO_POS_INF)
+ .Case("rz", X86::STATIC_ROUNDING::TO_ZERO)
+ .Default(-1);
+ if (-1 == rndMode)
+ return ErrorOperand(Tok.getLoc(), "Invalid rounding mode.");
+ Parser.Lex(); // Eat "r*" of r*-sae
+ if (!getLexer().is(AsmToken::Minus))
+ return ErrorOperand(Tok.getLoc(), "Expected - at this point");
+ Parser.Lex(); // Eat "-"
+ Parser.Lex(); // Eat the sae
+ if (!getLexer().is(AsmToken::RCurly))
+ return ErrorOperand(Tok.getLoc(), "Expected } at this point");
+ Parser.Lex(); // Eat "}"
+ const MCExpr *RndModeOp =
+ MCConstantExpr::create(rndMode, Parser.getContext());
+ return X86Operand::CreateImm(RndModeOp, Start, End);
+ }
+ if(Tok.getIdentifier().equals("sae")){
+ Parser.Lex(); // Eat the sae
+ if (!getLexer().is(AsmToken::RCurly))
+ return ErrorOperand(Tok.getLoc(), "Expected } at this point");
+ Parser.Lex(); // Eat "}"
+ return X86Operand::CreateToken("{sae}", consumedToken);
+ }
+ return ErrorOperand(Tok.getLoc(), "unknown token in expression");
+}
/// ParseIntelMemOperand - Parse intel style memory operand.
std::unique_ptr<X86Operand> X86AsmParser::ParseIntelMemOperand(int64_t ImmDisp,
SMLoc Start,
unsigned Size) {
+ MCAsmParser &Parser = getParser();
const AsmToken &Tok = Parser.getTok();
SMLoc End;
if (getParser().parsePrimaryExpr(Val, End))
return ErrorOperand(Tok.getLoc(), "unknown token in expression");
- return X86Operand::CreateMem(Val, Start, End, Size);
+ return X86Operand::CreateMem(getPointerWidth(), Val, Start, End, Size);
}
InlineAsmIdentifierInfo Info;
return nullptr;
}
- const MCExpr *Disp = MCConstantExpr::Create(SM.getImm(), getContext());
+ const MCExpr *Disp = MCConstantExpr::create(SM.getImm(), getContext());
// BaseReg is non-zero to avoid assertions. In the context of inline asm,
// we're pointing to a local variable in memory, so the base register is
// really the frame or stack pointer.
- return X86Operand::CreateMem(/*SegReg=*/0, Disp, /*BaseReg=*/1, /*IndexReg=*/0,
- /*Scale=*/1, Start, End, Size, Identifier,
- Info.OpDecl);
+ return X86Operand::CreateMem(getPointerWidth(), /*SegReg=*/0, Disp,
+ /*BaseReg=*/1, /*IndexReg=*/0, /*Scale=*/1,
+ Start, End, Size, Identifier, Info.OpDecl);
}
/// Parse the '.' operator.
bool X86AsmParser::ParseIntelDotOperator(const MCExpr *Disp,
const MCExpr *&NewDisp) {
+ MCAsmParser &Parser = getParser();
const AsmToken &Tok = Parser.getTok();
int64_t OrigDispVal, DotDispVal;
Val));
}
- NewDisp = MCConstantExpr::Create(OrigDispVal + DotDispVal, getContext());
+ NewDisp = MCConstantExpr::create(OrigDispVal + DotDispVal, getContext());
return false;
}
/// Parse the 'offset' operator. This operator is used to specify the
/// location rather then the content of a variable.
std::unique_ptr<X86Operand> X86AsmParser::ParseIntelOffsetOfOperator() {
+ MCAsmParser &Parser = getParser();
const AsmToken &Tok = Parser.getTok();
SMLoc OffsetOfLoc = Tok.getLoc();
Parser.Lex(); // Eat offset.
/// TYPE operator returns the size of a C or C++ type or variable. If the
/// variable is an array, TYPE returns the size of a single element.
std::unique_ptr<X86Operand> X86AsmParser::ParseIntelOperator(unsigned OpKind) {
+ MCAsmParser &Parser = getParser();
const AsmToken &Tok = Parser.getTok();
SMLoc TypeLoc = Tok.getLoc();
Parser.Lex(); // Eat operator.
unsigned Len = End.getPointer() - TypeLoc.getPointer();
InstInfo->AsmRewrites->push_back(AsmRewrite(AOK_Imm, TypeLoc, Len, CVal));
- const MCExpr *Imm = MCConstantExpr::Create(CVal, getContext());
+ const MCExpr *Imm = MCConstantExpr::create(CVal, getContext());
return X86Operand::CreateImm(Imm, Start, End);
}
std::unique_ptr<X86Operand> X86AsmParser::ParseIntelOperand() {
+ MCAsmParser &Parser = getParser();
const AsmToken &Tok = Parser.getTok();
SMLoc Start, End;
// to the MCExpr with the directional local symbol and this is a
// memory operand not an immediate operand.
if (SM.getSym())
- return X86Operand::CreateMem(SM.getSym(), Start, End, Size);
+ return X86Operand::CreateMem(getPointerWidth(), SM.getSym(), Start, End,
+ Size);
- const MCExpr *ImmExpr = MCConstantExpr::Create(Imm, getContext());
+ const MCExpr *ImmExpr = MCConstantExpr::create(Imm, getContext());
return X86Operand::CreateImm(ImmExpr, Start, End);
}
return ParseIntelMemOperand(Imm, Start, Size);
}
+ // rounding mode token
+ if (STI.getFeatureBits()[X86::FeatureAVX512] &&
+ getLexer().is(AsmToken::LCurly))
+ return ParseRoundingModeOp(Start, End);
+
// Register.
unsigned RegNo = 0;
if (!ParseRegister(RegNo, Start, End)) {
}
std::unique_ptr<X86Operand> X86AsmParser::ParseATTOperand() {
+ MCAsmParser &Parser = getParser();
switch (getLexer().getKind()) {
default:
// Parse a memory operand with no segment register.
return nullptr;
return X86Operand::CreateImm(Val, Start, End);
}
+ case AsmToken::LCurly:{
+ SMLoc Start = Parser.getTok().getLoc(), End;
+ if (STI.getFeatureBits()[X86::FeatureAVX512])
+ return ParseRoundingModeOp(Start, End);
+ return ErrorOperand(Start, "unknown token in expression");
+ }
}
}
bool X86AsmParser::HandleAVX512Operand(OperandVector &Operands,
const MCParsedAsmOperand &Op) {
- if(STI.getFeatureBits() & X86::FeatureAVX512) {
+ MCAsmParser &Parser = getParser();
+ if(STI.getFeatureBits()[X86::FeatureAVX512]) {
if (getLexer().is(AsmToken::LCurly)) {
// Eat "{" and mark the current place.
const SMLoc consumedToken = consumeToken();
std::unique_ptr<X86Operand> X86AsmParser::ParseMemOperand(unsigned SegReg,
SMLoc MemStart) {
+ MCAsmParser &Parser = getParser();
// We have to disambiguate a parenthesized expression "(4+5)" from the start
// of a memory operand with a missing displacement "(%ebx)" or "(,%eax)". The
// only way to do this without lookahead is to eat the '(' and see what is
// after it.
- const MCExpr *Disp = MCConstantExpr::Create(0, getParser().getContext());
+ const MCExpr *Disp = MCConstantExpr::create(0, getParser().getContext());
if (getLexer().isNot(AsmToken::LParen)) {
SMLoc ExprEnd;
if (getParser().parseExpression(Disp, ExprEnd)) return nullptr;
if (getLexer().isNot(AsmToken::LParen)) {
// Unless we have a segment register, treat this as an immediate.
if (SegReg == 0)
- return X86Operand::CreateMem(Disp, MemStart, ExprEnd);
- return X86Operand::CreateMem(SegReg, Disp, 0, 0, 1, MemStart, ExprEnd);
+ return X86Operand::CreateMem(getPointerWidth(), Disp, MemStart, ExprEnd);
+ return X86Operand::CreateMem(getPointerWidth(), SegReg, Disp, 0, 0, 1,
+ MemStart, ExprEnd);
}
// Eat the '('.
if (getLexer().isNot(AsmToken::LParen)) {
// Unless we have a segment register, treat this as an immediate.
if (SegReg == 0)
- return X86Operand::CreateMem(Disp, LParenLoc, ExprEnd);
- return X86Operand::CreateMem(SegReg, Disp, 0, 0, 1, MemStart, ExprEnd);
+ return X86Operand::CreateMem(getPointerWidth(), Disp, LParenLoc,
+ ExprEnd);
+ return X86Operand::CreateMem(getPointerWidth(), SegReg, Disp, 0, 0, 1,
+ MemStart, ExprEnd);
}
// Eat the '('.
}
if (SegReg || BaseReg || IndexReg)
- return X86Operand::CreateMem(SegReg, Disp, BaseReg, IndexReg, Scale,
- MemStart, MemEnd);
- return X86Operand::CreateMem(Disp, MemStart, MemEnd);
+ return X86Operand::CreateMem(getPointerWidth(), SegReg, Disp, BaseReg,
+ IndexReg, Scale, MemStart, MemEnd);
+ return X86Operand::CreateMem(getPointerWidth(), Disp, MemStart, MemEnd);
}
bool X86AsmParser::ParseInstruction(ParseInstructionInfo &Info, StringRef Name,
SMLoc NameLoc, OperandVector &Operands) {
+ MCAsmParser &Parser = getParser();
InstInfo = &Info;
StringRef PatchedName = Name;
PatchedName = PatchedName.substr(0, Name.size()-1);
// FIXME: Hack to recognize cmp<comparison code>{ss,sd,ps,pd}.
- const MCExpr *ExtraImmOp = nullptr;
if ((PatchedName.startswith("cmp") || PatchedName.startswith("vcmp")) &&
(PatchedName.endswith("ss") || PatchedName.endswith("sd") ||
PatchedName.endswith("ps") || PatchedName.endswith("pd"))) {
bool IsVCMP = PatchedName[0] == 'v';
- unsigned SSECCIdx = IsVCMP ? 4 : 3;
- unsigned SSEComparisonCode = StringSwitch<unsigned>(
- PatchedName.slice(SSECCIdx, PatchedName.size() - 2))
+ unsigned CCIdx = IsVCMP ? 4 : 3;
+ unsigned ComparisonCode = StringSwitch<unsigned>(
+ PatchedName.slice(CCIdx, PatchedName.size() - 2))
.Case("eq", 0x00)
.Case("lt", 0x01)
.Case("le", 0x02)
.Case("gt_oq", 0x1E)
.Case("true_us", 0x1F)
.Default(~0U);
- if (SSEComparisonCode != ~0U && (IsVCMP || SSEComparisonCode < 8)) {
- ExtraImmOp = MCConstantExpr::Create(SSEComparisonCode,
- getParser().getContext());
- if (PatchedName.endswith("ss")) {
- PatchedName = IsVCMP ? "vcmpss" : "cmpss";
- } else if (PatchedName.endswith("sd")) {
- PatchedName = IsVCMP ? "vcmpsd" : "cmpsd";
- } else if (PatchedName.endswith("ps")) {
- PatchedName = IsVCMP ? "vcmpps" : "cmpps";
- } else {
- assert(PatchedName.endswith("pd") && "Unexpected mnemonic!");
- PatchedName = IsVCMP ? "vcmppd" : "cmppd";
- }
+ if (ComparisonCode != ~0U && (IsVCMP || ComparisonCode < 8)) {
+
+ Operands.push_back(X86Operand::CreateToken(PatchedName.slice(0, CCIdx),
+ NameLoc));
+
+ const MCExpr *ImmOp = MCConstantExpr::create(ComparisonCode,
+ getParser().getContext());
+ Operands.push_back(X86Operand::CreateImm(ImmOp, NameLoc, NameLoc));
+
+ PatchedName = PatchedName.substr(PatchedName.size() - 2);
+ }
+ }
+
+ // FIXME: Hack to recognize vpcmp<comparison code>{ub,uw,ud,uq,b,w,d,q}.
+ if (PatchedName.startswith("vpcmp") &&
+ (PatchedName.endswith("b") || PatchedName.endswith("w") ||
+ PatchedName.endswith("d") || PatchedName.endswith("q"))) {
+ unsigned CCIdx = PatchedName.drop_back().back() == 'u' ? 2 : 1;
+ unsigned ComparisonCode = StringSwitch<unsigned>(
+ PatchedName.slice(5, PatchedName.size() - CCIdx))
+ .Case("eq", 0x0) // Only allowed on unsigned. Checked below.
+ .Case("lt", 0x1)
+ .Case("le", 0x2)
+ //.Case("false", 0x3) // Not a documented alias.
+ .Case("neq", 0x4)
+ .Case("nlt", 0x5)
+ .Case("nle", 0x6)
+ //.Case("true", 0x7) // Not a documented alias.
+ .Default(~0U);
+ if (ComparisonCode != ~0U && (ComparisonCode != 0 || CCIdx == 2)) {
+ Operands.push_back(X86Operand::CreateToken("vpcmp", NameLoc));
+
+ const MCExpr *ImmOp = MCConstantExpr::create(ComparisonCode,
+ getParser().getContext());
+ Operands.push_back(X86Operand::CreateImm(ImmOp, NameLoc, NameLoc));
+
+ PatchedName = PatchedName.substr(PatchedName.size() - CCIdx);
+ }
+ }
+
+ // FIXME: Hack to recognize vpcom<comparison code>{ub,uw,ud,uq,b,w,d,q}.
+ if (PatchedName.startswith("vpcom") &&
+ (PatchedName.endswith("b") || PatchedName.endswith("w") ||
+ PatchedName.endswith("d") || PatchedName.endswith("q"))) {
+ unsigned CCIdx = PatchedName.drop_back().back() == 'u' ? 2 : 1;
+ unsigned ComparisonCode = StringSwitch<unsigned>(
+ PatchedName.slice(5, PatchedName.size() - CCIdx))
+ .Case("lt", 0x0)
+ .Case("le", 0x1)
+ .Case("gt", 0x2)
+ .Case("ge", 0x3)
+ .Case("eq", 0x4)
+ .Case("neq", 0x5)
+ .Case("false", 0x6)
+ .Case("true", 0x7)
+ .Default(~0U);
+ if (ComparisonCode != ~0U) {
+ Operands.push_back(X86Operand::CreateToken("vpcom", NameLoc));
+
+ const MCExpr *ImmOp = MCConstantExpr::create(ComparisonCode,
+ getParser().getContext());
+ Operands.push_back(X86Operand::CreateImm(ImmOp, NameLoc, NameLoc));
+
+ PatchedName = PatchedName.substr(PatchedName.size() - CCIdx);
}
}
Operands.push_back(X86Operand::CreateToken(PatchedName, NameLoc));
- if (ExtraImmOp && !isParsingIntelSyntax())
- Operands.push_back(X86Operand::CreateImm(ExtraImmOp, NameLoc, NameLoc));
-
// Determine whether this is an instruction prefix.
bool isPrefix =
Name == "lock" || Name == "rep" ||
(isPrefix && getLexer().is(AsmToken::Slash)))
Parser.Lex();
- if (ExtraImmOp && isParsingIntelSyntax())
- Operands.push_back(X86Operand::CreateImm(ExtraImmOp, NameLoc, NameLoc));
-
// This is a terrible hack to handle "out[bwl]? %al, (%dx)" ->
// "outb %al, %dx". Out doesn't take a memory form, but this is a widely
// documented form in various unofficial manuals, so a lot of code uses it.
if (Name.startswith("ins") && Operands.size() == 1 &&
(Name == "insb" || Name == "insw" || Name == "insl" ||
Name == "insd" )) {
- if (isParsingIntelSyntax()) {
- Operands.push_back(X86Operand::CreateReg(X86::DX, NameLoc, NameLoc));
- Operands.push_back(DefaultMemDIOperand(NameLoc));
- } else {
- Operands.push_back(X86Operand::CreateReg(X86::DX, NameLoc, NameLoc));
- Operands.push_back(DefaultMemDIOperand(NameLoc));
- }
+ AddDefaultSrcDestOperands(Operands,
+ X86Operand::CreateReg(X86::DX, NameLoc, NameLoc),
+ DefaultMemDIOperand(NameLoc));
}
// Append default arguments to "outs[bwld]"
if (Name.startswith("outs") && Operands.size() == 1 &&
(Name == "outsb" || Name == "outsw" || Name == "outsl" ||
Name == "outsd" )) {
- if (isParsingIntelSyntax()) {
- Operands.push_back(DefaultMemSIOperand(NameLoc));
- Operands.push_back(X86Operand::CreateReg(X86::DX, NameLoc, NameLoc));
- } else {
- Operands.push_back(DefaultMemSIOperand(NameLoc));
- Operands.push_back(X86Operand::CreateReg(X86::DX, NameLoc, NameLoc));
- }
+ AddDefaultSrcDestOperands(Operands,
+ DefaultMemSIOperand(NameLoc),
+ X86Operand::CreateReg(X86::DX, NameLoc, NameLoc));
}
// Transform "lods[bwlq]" into "lods[bwlq] ($SIREG)" for appropriate
(Name == "cmps" || Name == "cmpsb" || Name == "cmpsw" ||
Name == "cmpsl" || Name == "cmpsd" || Name == "cmpsq")) {
if (Operands.size() == 1) {
- if (isParsingIntelSyntax()) {
- Operands.push_back(DefaultMemSIOperand(NameLoc));
- Operands.push_back(DefaultMemDIOperand(NameLoc));
- } else {
- Operands.push_back(DefaultMemDIOperand(NameLoc));
- Operands.push_back(DefaultMemSIOperand(NameLoc));
- }
+ AddDefaultSrcDestOperands(Operands,
+ DefaultMemDIOperand(NameLoc),
+ DefaultMemSIOperand(NameLoc));
} else if (Operands.size() == 3) {
X86Operand &Op = (X86Operand &)*Operands[1];
X86Operand &Op2 = (X86Operand &)*Operands[2];
if (Operands.size() == 1) {
if (Name == "movsd")
Operands.back() = X86Operand::CreateToken("movsl", NameLoc);
- if (isParsingIntelSyntax()) {
- Operands.push_back(DefaultMemDIOperand(NameLoc));
- Operands.push_back(DefaultMemSIOperand(NameLoc));
- } else {
- Operands.push_back(DefaultMemSIOperand(NameLoc));
- Operands.push_back(DefaultMemDIOperand(NameLoc));
- }
+ AddDefaultSrcDestOperands(Operands,
+ DefaultMemSIOperand(NameLoc),
+ DefaultMemDIOperand(NameLoc));
} else if (Operands.size() == 3) {
X86Operand &Op = (X86Operand &)*Operands[1];
X86Operand &Op2 = (X86Operand &)*Operands[2];
// instalias with an immediate operand yet.
if (Name == "int" && Operands.size() == 2) {
X86Operand &Op1 = static_cast<X86Operand &>(*Operands[1]);
- if (Op1.isImm() && isa<MCConstantExpr>(Op1.getImm()) &&
- cast<MCConstantExpr>(Op1.getImm())->getValue() == 3) {
- Operands.erase(Operands.begin() + 1);
- static_cast<X86Operand &>(*Operands[0]).setTokenValue("int3");
- }
+ if (Op1.isImm())
+ if (auto *CE = dyn_cast<MCConstantExpr>(Op1.getImm()))
+ if (CE->getValue() == 3) {
+ Operands.erase(Operands.begin() + 1);
+ static_cast<X86Operand &>(*Operands[0]).setTokenValue("int3");
+ }
}
return false;
MCInst TmpInst;
TmpInst.setOpcode(Opcode);
if (!isCmp)
- TmpInst.addOperand(MCOperand::CreateReg(Reg));
- TmpInst.addOperand(MCOperand::CreateReg(Reg));
+ TmpInst.addOperand(MCOperand::createReg(Reg));
+ TmpInst.addOperand(MCOperand::createReg(Reg));
TmpInst.addOperand(Inst.getOperand(0));
Inst = TmpInst;
return true;
return convertToSExti8(Inst, Opcode, X86::RAX, isCmp);
}
+bool X86AsmParser::validateInstruction(MCInst &Inst, const OperandVector &Ops) {
+ switch (Inst.getOpcode()) {
+ default: return true;
+ case X86::INT:
+ X86Operand &Op = static_cast<X86Operand &>(*Ops[1]);
+ assert(Op.isImm() && "expected immediate");
+ int64_t Res;
+ if (!Op.getImm()->evaluateAsAbsolute(Res) || Res > 255) {
+ Error(Op.getStartLoc(), "interrupt vector must be in range [0-255]");
+ return false;
+ }
+ return true;
+ }
+ llvm_unreachable("handle the instruction appropriately");
+}
+
bool X86AsmParser::processInstruction(MCInst &Inst, const OperandVector &Ops) {
switch (Inst.getOpcode()) {
default: return false;
switch (MatchInstructionImpl(Operands, Inst,
ErrorInfo, MatchingInlineAsm,
isParsingIntelSyntax())) {
- default: break;
+ default: llvm_unreachable("Unexpected match result!");
case Match_Success:
+ if (!validateInstruction(Inst, Operands))
+ return true;
+
// Some instructions need post-processing to, for example, tweak which
// encoding is selected. Loop on it while changes happen so the
// individual transformations can chain off each other.
SmallString<16> Tmp;
Tmp += Base;
Tmp += ' ';
- Op.setTokenValue(Tmp.str());
+ Op.setTokenValue(Tmp);
// If this instruction starts with an 'f', then it is a floating point stack
// instruction. These come in up to three forms for 32-bit, 64-bit, and
static const char *const PtrSizedInstrs[] = {"call", "jmp", "push"};
for (const char *Instr : PtrSizedInstrs) {
if (Mnemonic == Instr) {
- UnsizedMemOp->Mem.Size = getPointerSize();
+ UnsizedMemOp->Mem.Size = getPointerWidth();
break;
}
}
SmallVector<unsigned, 8> Match;
uint64_t ErrorInfoMissingFeature = 0;
if (UnsizedMemOp && UnsizedMemOp->isMemUnsized()) {
- static const unsigned MopSizes[] = {8, 16, 32, 64, 80};
+ static const unsigned MopSizes[] = {8, 16, 32, 64, 80, 128, 256, 512};
for (unsigned Size : MopSizes) {
UnsizedMemOp->Mem.Size = Size;
uint64_t ErrorInfoIgnore;
}
// If we haven't matched anything yet, this is not a basic integer or FPU
- // operation. There shouldn't be any ambiguity in our mneumonic table, so try
+ // operation. There shouldn't be any ambiguity in our mnemonic table, so try
// matching with the unsized operand.
if (Match.empty()) {
Match.push_back(MatchInstructionImpl(Operands, Inst, ErrorInfo,
unsigned NumSuccessfulMatches =
std::count(std::begin(Match), std::end(Match), Match_Success);
if (NumSuccessfulMatches == 1) {
+ if (!validateInstruction(Inst, Operands))
+ return true;
+
// Some instructions need post-processing to, for example, tweak which
// encoding is selected. Loop on it while changes happen so the individual
// transformations can chain off each other.
}
bool X86AsmParser::ParseDirective(AsmToken DirectiveID) {
+ MCAsmParser &Parser = getParser();
StringRef IDVal = DirectiveID.getIdentifier();
if (IDVal == ".word")
return ParseDirectiveWord(2, DirectiveID.getLoc());
/// ParseDirectiveWord
/// ::= .word [ expression (, expression)* ]
bool X86AsmParser::ParseDirectiveWord(unsigned Size, SMLoc L) {
+ MCAsmParser &Parser = getParser();
if (getLexer().isNot(AsmToken::EndOfStatement)) {
for (;;) {
const MCExpr *Value;
/// ParseDirectiveCode
/// ::= .code16 | .code32 | .code64
bool X86AsmParser::ParseDirectiveCode(StringRef IDVal, SMLoc L) {
+ MCAsmParser &Parser = getParser();
if (IDVal == ".code16") {
Parser.Lex();
if (!is16BitMode()) {