1 //===-- RISCVAsmParser.cpp - Parse RISCV assembly to MCInst instructions --===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 
9 #include "MCTargetDesc/RISCVAsmBackend.h"
10 #include "MCTargetDesc/RISCVMCExpr.h"
11 #include "MCTargetDesc/RISCVMCTargetDesc.h"
12 #include "MCTargetDesc/RISCVTargetStreamer.h"
13 #include "TargetInfo/RISCVTargetInfo.h"
14 #include "Utils/RISCVBaseInfo.h"
15 #include "Utils/RISCVMatInt.h"
16 #include "llvm/ADT/STLExtras.h"
17 #include "llvm/ADT/SmallVector.h"
18 #include "llvm/ADT/StringSwitch.h"
19 #include "llvm/CodeGen/Register.h"
20 #include "llvm/MC/MCAssembler.h"
21 #include "llvm/MC/MCContext.h"
22 #include "llvm/MC/MCExpr.h"
23 #include "llvm/MC/MCInst.h"
24 #include "llvm/MC/MCInstBuilder.h"
25 #include "llvm/MC/MCObjectFileInfo.h"
26 #include "llvm/MC/MCParser/MCAsmLexer.h"
27 #include "llvm/MC/MCParser/MCParsedAsmOperand.h"
28 #include "llvm/MC/MCParser/MCTargetAsmParser.h"
29 #include "llvm/MC/MCRegisterInfo.h"
30 #include "llvm/MC/MCStreamer.h"
31 #include "llvm/MC/MCSubtargetInfo.h"
32 #include "llvm/Support/Casting.h"
33 #include "llvm/Support/MathExtras.h"
34 #include "llvm/Support/TargetRegistry.h"
35 
36 #include <limits>
37 
38 using namespace llvm;
39 
40 // Include the auto-generated portion of the compress emitter.
41 #define GEN_COMPRESS_INSTR
42 #include "RISCVGenCompressInstEmitter.inc"
43 
44 namespace {
45 struct RISCVOperand;
46 
47 class RISCVAsmParser : public MCTargetAsmParser {
48   SmallVector<FeatureBitset, 4> FeatureBitStack;
49 
50   SMLoc getLoc() const { return getParser().getTok().getLoc(); }
51   bool isRV64() const { return getSTI().hasFeature(RISCV::Feature64Bit); }
52   bool isRV32E() const { return getSTI().hasFeature(RISCV::FeatureRV32E); }
53 
54   RISCVTargetStreamer &getTargetStreamer() {
55     MCTargetStreamer &TS = *getParser().getStreamer().getTargetStreamer();
56     return static_cast<RISCVTargetStreamer &>(TS);
57   }
58 
59   unsigned validateTargetOperandClass(MCParsedAsmOperand &Op,
60                                       unsigned Kind) override;
61 
62   bool generateImmOutOfRangeError(OperandVector &Operands, uint64_t ErrorInfo,
63                                   int64_t Lower, int64_t Upper, Twine Msg);
64 
65   bool MatchAndEmitInstruction(SMLoc IDLoc, unsigned &Opcode,
66                                OperandVector &Operands, MCStreamer &Out,
67                                uint64_t &ErrorInfo,
68                                bool MatchingInlineAsm) override;
69 
70   bool ParseRegister(unsigned &RegNo, SMLoc &StartLoc, SMLoc &EndLoc) override;
71 
72   bool ParseInstruction(ParseInstructionInfo &Info, StringRef Name,
73                         SMLoc NameLoc, OperandVector &Operands) override;
74 
75   bool ParseDirective(AsmToken DirectiveID) override;
76 
77   // Helper to actually emit an instruction to the MCStreamer. Also, when
78   // possible, compression of the instruction is performed.
79   void emitToStreamer(MCStreamer &S, const MCInst &Inst);
80 
81   // Helper to emit a combination of LUI, ADDI(W), and SLLI instructions that
82   // synthesize the desired immedate value into the destination register.
83   void emitLoadImm(Register DestReg, int64_t Value, MCStreamer &Out);
84 
85   // Helper to emit a combination of AUIPC and SecondOpcode. Used to implement
86   // helpers such as emitLoadLocalAddress and emitLoadAddress.
87   void emitAuipcInstPair(MCOperand DestReg, MCOperand TmpReg,
88                          const MCExpr *Symbol, RISCVMCExpr::VariantKind VKHi,
89                          unsigned SecondOpcode, SMLoc IDLoc, MCStreamer &Out);
90 
91   // Helper to emit pseudo instruction "lla" used in PC-rel addressing.
92   void emitLoadLocalAddress(MCInst &Inst, SMLoc IDLoc, MCStreamer &Out);
93 
94   // Helper to emit pseudo instruction "la" used in GOT/PC-rel addressing.
95   void emitLoadAddress(MCInst &Inst, SMLoc IDLoc, MCStreamer &Out);
96 
97   // Helper to emit pseudo instruction "la.tls.ie" used in initial-exec TLS
98   // addressing.
99   void emitLoadTLSIEAddress(MCInst &Inst, SMLoc IDLoc, MCStreamer &Out);
100 
101   // Helper to emit pseudo instruction "la.tls.gd" used in global-dynamic TLS
102   // addressing.
103   void emitLoadTLSGDAddress(MCInst &Inst, SMLoc IDLoc, MCStreamer &Out);
104 
105   // Helper to emit pseudo load/store instruction with a symbol.
106   void emitLoadStoreSymbol(MCInst &Inst, unsigned Opcode, SMLoc IDLoc,
107                            MCStreamer &Out, bool HasTmpReg);
108 
109   // Checks that a PseudoAddTPRel is using x4/tp in its second input operand.
110   // Enforcing this using a restricted register class for the second input
111   // operand of PseudoAddTPRel results in a poor diagnostic due to the fact
112   // 'add' is an overloaded mnemonic.
113   bool checkPseudoAddTPRel(MCInst &Inst, OperandVector &Operands);
114 
115   /// Helper for processing MC instructions that have been successfully matched
116   /// by MatchAndEmitInstruction. Modifications to the emitted instructions,
117   /// like the expansion of pseudo instructions (e.g., "li"), can be performed
118   /// in this method.
119   bool processInstruction(MCInst &Inst, SMLoc IDLoc, OperandVector &Operands,
120                           MCStreamer &Out);
121 
122 // Auto-generated instruction matching functions
123 #define GET_ASSEMBLER_HEADER
124 #include "RISCVGenAsmMatcher.inc"
125 
126   OperandMatchResultTy parseCSRSystemRegister(OperandVector &Operands);
127   OperandMatchResultTy parseImmediate(OperandVector &Operands);
128   OperandMatchResultTy parseRegister(OperandVector &Operands,
129                                      bool AllowParens = false);
130   OperandMatchResultTy parseMemOpBaseReg(OperandVector &Operands);
131   OperandMatchResultTy parseAtomicMemOp(OperandVector &Operands);
132   OperandMatchResultTy parseOperandWithModifier(OperandVector &Operands);
133   OperandMatchResultTy parseBareSymbol(OperandVector &Operands);
134   OperandMatchResultTy parseCallSymbol(OperandVector &Operands);
135   OperandMatchResultTy parseJALOffset(OperandVector &Operands);
136 
137   bool parseOperand(OperandVector &Operands, StringRef Mnemonic);
138 
139   bool parseDirectiveOption();
140 
141   void setFeatureBits(uint64_t Feature, StringRef FeatureString) {
142     if (!(getSTI().getFeatureBits()[Feature])) {
143       MCSubtargetInfo &STI = copySTI();
144       setAvailableFeatures(
145           ComputeAvailableFeatures(STI.ToggleFeature(FeatureString)));
146     }
147   }
148 
149   void clearFeatureBits(uint64_t Feature, StringRef FeatureString) {
150     if (getSTI().getFeatureBits()[Feature]) {
151       MCSubtargetInfo &STI = copySTI();
152       setAvailableFeatures(
153           ComputeAvailableFeatures(STI.ToggleFeature(FeatureString)));
154     }
155   }
156 
157   void pushFeatureBits() {
158     FeatureBitStack.push_back(getSTI().getFeatureBits());
159   }
160 
161   bool popFeatureBits() {
162     if (FeatureBitStack.empty())
163       return true;
164 
165     FeatureBitset FeatureBits = FeatureBitStack.pop_back_val();
166     copySTI().setFeatureBits(FeatureBits);
167     setAvailableFeatures(ComputeAvailableFeatures(FeatureBits));
168 
169     return false;
170   }
171 public:
172   enum RISCVMatchResultTy {
173     Match_Dummy = FIRST_TARGET_MATCH_RESULT_TY,
174 #define GET_OPERAND_DIAGNOSTIC_TYPES
175 #include "RISCVGenAsmMatcher.inc"
176 #undef GET_OPERAND_DIAGNOSTIC_TYPES
177   };
178 
179   static bool classifySymbolRef(const MCExpr *Expr,
180                                 RISCVMCExpr::VariantKind &Kind,
181                                 int64_t &Addend);
182 
183   RISCVAsmParser(const MCSubtargetInfo &STI, MCAsmParser &Parser,
184                  const MCInstrInfo &MII, const MCTargetOptions &Options)
185       : MCTargetAsmParser(Options, STI, MII) {
186     Parser.addAliasForDirective(".half", ".2byte");
187     Parser.addAliasForDirective(".hword", ".2byte");
188     Parser.addAliasForDirective(".word", ".4byte");
189     Parser.addAliasForDirective(".dword", ".8byte");
190     setAvailableFeatures(ComputeAvailableFeatures(STI.getFeatureBits()));
191   }
192 };
193 
194 /// RISCVOperand - Instances of this class represent a parsed machine
195 /// instruction
196 struct RISCVOperand : public MCParsedAsmOperand {
197 
198   enum class KindTy {
199     Token,
200     Register,
201     Immediate,
202     SystemRegister
203   } Kind;
204 
205   bool IsRV64;
206 
207   struct RegOp {
208     Register RegNum;
209   };
210 
211   struct ImmOp {
212     const MCExpr *Val;
213   };
214 
215   struct SysRegOp {
216     const char *Data;
217     unsigned Length;
218     unsigned Encoding;
219     // FIXME: Add the Encoding parsed fields as needed for checks,
220     // e.g.: read/write or user/supervisor/machine privileges.
221   };
222 
223   SMLoc StartLoc, EndLoc;
224   union {
225     StringRef Tok;
226     RegOp Reg;
227     ImmOp Imm;
228     struct SysRegOp SysReg;
229   };
230 
231   RISCVOperand(KindTy K) : MCParsedAsmOperand(), Kind(K) {}
232 
233 public:
234   RISCVOperand(const RISCVOperand &o) : MCParsedAsmOperand() {
235     Kind = o.Kind;
236     IsRV64 = o.IsRV64;
237     StartLoc = o.StartLoc;
238     EndLoc = o.EndLoc;
239     switch (Kind) {
240     case KindTy::Register:
241       Reg = o.Reg;
242       break;
243     case KindTy::Immediate:
244       Imm = o.Imm;
245       break;
246     case KindTy::Token:
247       Tok = o.Tok;
248       break;
249     case KindTy::SystemRegister:
250       SysReg = o.SysReg;
251       break;
252     }
253   }
254 
255   bool isToken() const override { return Kind == KindTy::Token; }
256   bool isReg() const override { return Kind == KindTy::Register; }
257   bool isImm() const override { return Kind == KindTy::Immediate; }
258   bool isMem() const override { return false; }
259   bool isSystemRegister() const { return Kind == KindTy::SystemRegister; }
260 
261   static bool evaluateConstantImm(const MCExpr *Expr, int64_t &Imm,
262                                   RISCVMCExpr::VariantKind &VK) {
263     if (auto *RE = dyn_cast<RISCVMCExpr>(Expr)) {
264       VK = RE->getKind();
265       return RE->evaluateAsConstant(Imm);
266     }
267 
268     if (auto CE = dyn_cast<MCConstantExpr>(Expr)) {
269       VK = RISCVMCExpr::VK_RISCV_None;
270       Imm = CE->getValue();
271       return true;
272     }
273 
274     return false;
275   }
276 
277   // True if operand is a symbol with no modifiers, or a constant with no
278   // modifiers and isShiftedInt<N-1, 1>(Op).
279   template <int N> bool isBareSimmNLsb0() const {
280     int64_t Imm;
281     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
282     if (!isImm())
283       return false;
284     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
285     bool IsValid;
286     if (!IsConstantImm)
287       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
288     else
289       IsValid = isShiftedInt<N - 1, 1>(Imm);
290     return IsValid && VK == RISCVMCExpr::VK_RISCV_None;
291   }
292 
293   // Predicate methods for AsmOperands defined in RISCVInstrInfo.td
294 
295   bool isBareSymbol() const {
296     int64_t Imm;
297     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
298     // Must be of 'immediate' type but not a constant.
299     if (!isImm() || evaluateConstantImm(getImm(), Imm, VK))
300       return false;
301     return RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm) &&
302            VK == RISCVMCExpr::VK_RISCV_None;
303   }
304 
305   bool isCallSymbol() const {
306     int64_t Imm;
307     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
308     // Must be of 'immediate' type but not a constant.
309     if (!isImm() || evaluateConstantImm(getImm(), Imm, VK))
310       return false;
311     return RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm) &&
312            (VK == RISCVMCExpr::VK_RISCV_CALL ||
313             VK == RISCVMCExpr::VK_RISCV_CALL_PLT);
314   }
315 
316   bool isTPRelAddSymbol() const {
317     int64_t Imm;
318     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
319     // Must be of 'immediate' type but not a constant.
320     if (!isImm() || evaluateConstantImm(getImm(), Imm, VK))
321       return false;
322     return RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm) &&
323            VK == RISCVMCExpr::VK_RISCV_TPREL_ADD;
324   }
325 
326   bool isCSRSystemRegister() const { return isSystemRegister(); }
327 
328   /// Return true if the operand is a valid for the fence instruction e.g.
329   /// ('iorw').
330   bool isFenceArg() const {
331     if (!isImm())
332       return false;
333     const MCExpr *Val = getImm();
334     auto *SVal = dyn_cast<MCSymbolRefExpr>(Val);
335     if (!SVal || SVal->getKind() != MCSymbolRefExpr::VK_None)
336       return false;
337 
338     StringRef Str = SVal->getSymbol().getName();
339     // Letters must be unique, taken from 'iorw', and in ascending order. This
340     // holds as long as each individual character is one of 'iorw' and is
341     // greater than the previous character.
342     char Prev = '\0';
343     for (char c : Str) {
344       if (c != 'i' && c != 'o' && c != 'r' && c != 'w')
345         return false;
346       if (c <= Prev)
347         return false;
348       Prev = c;
349     }
350     return true;
351   }
352 
353   /// Return true if the operand is a valid floating point rounding mode.
354   bool isFRMArg() const {
355     if (!isImm())
356       return false;
357     const MCExpr *Val = getImm();
358     auto *SVal = dyn_cast<MCSymbolRefExpr>(Val);
359     if (!SVal || SVal->getKind() != MCSymbolRefExpr::VK_None)
360       return false;
361 
362     StringRef Str = SVal->getSymbol().getName();
363 
364     return RISCVFPRndMode::stringToRoundingMode(Str) != RISCVFPRndMode::Invalid;
365   }
366 
367   bool isImmXLenLI() const {
368     int64_t Imm;
369     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
370     if (!isImm())
371       return false;
372     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
373     if (VK == RISCVMCExpr::VK_RISCV_LO || VK == RISCVMCExpr::VK_RISCV_PCREL_LO)
374       return true;
375     // Given only Imm, ensuring that the actually specified constant is either
376     // a signed or unsigned 64-bit number is unfortunately impossible.
377     return IsConstantImm && VK == RISCVMCExpr::VK_RISCV_None &&
378            (isRV64() || (isInt<32>(Imm) || isUInt<32>(Imm)));
379   }
380 
381   bool isUImmLog2XLen() const {
382     int64_t Imm;
383     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
384     if (!isImm())
385       return false;
386     if (!evaluateConstantImm(getImm(), Imm, VK) ||
387         VK != RISCVMCExpr::VK_RISCV_None)
388       return false;
389     return (isRV64() && isUInt<6>(Imm)) || isUInt<5>(Imm);
390   }
391 
392   bool isUImmLog2XLenNonZero() const {
393     int64_t Imm;
394     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
395     if (!isImm())
396       return false;
397     if (!evaluateConstantImm(getImm(), Imm, VK) ||
398         VK != RISCVMCExpr::VK_RISCV_None)
399       return false;
400     if (Imm == 0)
401       return false;
402     return (isRV64() && isUInt<6>(Imm)) || isUInt<5>(Imm);
403   }
404 
405   bool isUImm5() const {
406     int64_t Imm;
407     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
408     if (!isImm())
409       return false;
410     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
411     return IsConstantImm && isUInt<5>(Imm) && VK == RISCVMCExpr::VK_RISCV_None;
412   }
413 
414   bool isUImm5NonZero() const {
415     int64_t Imm;
416     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
417     if (!isImm())
418       return false;
419     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
420     return IsConstantImm && isUInt<5>(Imm) && (Imm != 0) &&
421            VK == RISCVMCExpr::VK_RISCV_None;
422   }
423 
424   bool isSImm6() const {
425     if (!isImm())
426       return false;
427     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
428     int64_t Imm;
429     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
430     return IsConstantImm && isInt<6>(Imm) &&
431            VK == RISCVMCExpr::VK_RISCV_None;
432   }
433 
434   bool isSImm6NonZero() const {
435     if (!isImm())
436       return false;
437     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
438     int64_t Imm;
439     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
440     return IsConstantImm && isInt<6>(Imm) && (Imm != 0) &&
441            VK == RISCVMCExpr::VK_RISCV_None;
442   }
443 
444   bool isCLUIImm() const {
445     if (!isImm())
446       return false;
447     int64_t Imm;
448     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
449     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
450     return IsConstantImm && (Imm != 0) &&
451            (isUInt<5>(Imm) || (Imm >= 0xfffe0 && Imm <= 0xfffff)) &&
452            VK == RISCVMCExpr::VK_RISCV_None;
453   }
454 
455   bool isUImm7Lsb00() const {
456     if (!isImm())
457       return false;
458     int64_t Imm;
459     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
460     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
461     return IsConstantImm && isShiftedUInt<5, 2>(Imm) &&
462            VK == RISCVMCExpr::VK_RISCV_None;
463   }
464 
465   bool isUImm8Lsb00() const {
466     if (!isImm())
467       return false;
468     int64_t Imm;
469     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
470     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
471     return IsConstantImm && isShiftedUInt<6, 2>(Imm) &&
472            VK == RISCVMCExpr::VK_RISCV_None;
473   }
474 
475   bool isUImm8Lsb000() const {
476     if (!isImm())
477       return false;
478     int64_t Imm;
479     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
480     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
481     return IsConstantImm && isShiftedUInt<5, 3>(Imm) &&
482            VK == RISCVMCExpr::VK_RISCV_None;
483   }
484 
485   bool isSImm9Lsb0() const { return isBareSimmNLsb0<9>(); }
486 
487   bool isUImm9Lsb000() const {
488     if (!isImm())
489       return false;
490     int64_t Imm;
491     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
492     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
493     return IsConstantImm && isShiftedUInt<6, 3>(Imm) &&
494            VK == RISCVMCExpr::VK_RISCV_None;
495   }
496 
497   bool isUImm10Lsb00NonZero() const {
498     if (!isImm())
499       return false;
500     int64_t Imm;
501     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
502     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
503     return IsConstantImm && isShiftedUInt<8, 2>(Imm) && (Imm != 0) &&
504            VK == RISCVMCExpr::VK_RISCV_None;
505   }
506 
507   bool isSImm12() const {
508     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
509     int64_t Imm;
510     bool IsValid;
511     if (!isImm())
512       return false;
513     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
514     if (!IsConstantImm)
515       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
516     else
517       IsValid = isInt<12>(Imm);
518     return IsValid && ((IsConstantImm && VK == RISCVMCExpr::VK_RISCV_None) ||
519                        VK == RISCVMCExpr::VK_RISCV_LO ||
520                        VK == RISCVMCExpr::VK_RISCV_PCREL_LO ||
521                        VK == RISCVMCExpr::VK_RISCV_TPREL_LO);
522   }
523 
524   bool isSImm12Lsb0() const { return isBareSimmNLsb0<12>(); }
525 
526   bool isSImm13Lsb0() const { return isBareSimmNLsb0<13>(); }
527 
528   bool isSImm10Lsb0000NonZero() const {
529     if (!isImm())
530       return false;
531     int64_t Imm;
532     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
533     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
534     return IsConstantImm && (Imm != 0) && isShiftedInt<6, 4>(Imm) &&
535            VK == RISCVMCExpr::VK_RISCV_None;
536   }
537 
538   bool isUImm20LUI() const {
539     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
540     int64_t Imm;
541     bool IsValid;
542     if (!isImm())
543       return false;
544     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
545     if (!IsConstantImm) {
546       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
547       return IsValid && (VK == RISCVMCExpr::VK_RISCV_HI ||
548                          VK == RISCVMCExpr::VK_RISCV_TPREL_HI);
549     } else {
550       return isUInt<20>(Imm) && (VK == RISCVMCExpr::VK_RISCV_None ||
551                                  VK == RISCVMCExpr::VK_RISCV_HI ||
552                                  VK == RISCVMCExpr::VK_RISCV_TPREL_HI);
553     }
554   }
555 
556   bool isUImm20AUIPC() const {
557     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
558     int64_t Imm;
559     bool IsValid;
560     if (!isImm())
561       return false;
562     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
563     if (!IsConstantImm) {
564       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
565       return IsValid && (VK == RISCVMCExpr::VK_RISCV_PCREL_HI ||
566                          VK == RISCVMCExpr::VK_RISCV_GOT_HI ||
567                          VK == RISCVMCExpr::VK_RISCV_TLS_GOT_HI ||
568                          VK == RISCVMCExpr::VK_RISCV_TLS_GD_HI);
569     } else {
570       return isUInt<20>(Imm) && (VK == RISCVMCExpr::VK_RISCV_None ||
571                                  VK == RISCVMCExpr::VK_RISCV_PCREL_HI ||
572                                  VK == RISCVMCExpr::VK_RISCV_GOT_HI ||
573                                  VK == RISCVMCExpr::VK_RISCV_TLS_GOT_HI ||
574                                  VK == RISCVMCExpr::VK_RISCV_TLS_GD_HI);
575     }
576   }
577 
578   bool isSImm21Lsb0JAL() const { return isBareSimmNLsb0<21>(); }
579 
580   bool isImmZero() const {
581     if (!isImm())
582       return false;
583     int64_t Imm;
584     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
585     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
586     return IsConstantImm && (Imm == 0) && VK == RISCVMCExpr::VK_RISCV_None;
587   }
588 
589   /// getStartLoc - Gets location of the first token of this operand
590   SMLoc getStartLoc() const override { return StartLoc; }
591   /// getEndLoc - Gets location of the last token of this operand
592   SMLoc getEndLoc() const override { return EndLoc; }
593   /// True if this operand is for an RV64 instruction
594   bool isRV64() const { return IsRV64; }
595 
596   unsigned getReg() const override {
597     assert(Kind == KindTy::Register && "Invalid type access!");
598     return Reg.RegNum.id();
599   }
600 
601   StringRef getSysReg() const {
602     assert(Kind == KindTy::SystemRegister && "Invalid access!");
603     return StringRef(SysReg.Data, SysReg.Length);
604   }
605 
606   const MCExpr *getImm() const {
607     assert(Kind == KindTy::Immediate && "Invalid type access!");
608     return Imm.Val;
609   }
610 
611   StringRef getToken() const {
612     assert(Kind == KindTy::Token && "Invalid type access!");
613     return Tok;
614   }
615 
616   void print(raw_ostream &OS) const override {
617     switch (Kind) {
618     case KindTy::Immediate:
619       OS << *getImm();
620       break;
621     case KindTy::Register:
622       OS << "<register x";
623       OS << getReg() << ">";
624       break;
625     case KindTy::Token:
626       OS << "'" << getToken() << "'";
627       break;
628     case KindTy::SystemRegister:
629       OS << "<sysreg: " << getSysReg() << '>';
630       break;
631     }
632   }
633 
634   static std::unique_ptr<RISCVOperand> createToken(StringRef Str, SMLoc S,
635                                                    bool IsRV64) {
636     auto Op = std::make_unique<RISCVOperand>(KindTy::Token);
637     Op->Tok = Str;
638     Op->StartLoc = S;
639     Op->EndLoc = S;
640     Op->IsRV64 = IsRV64;
641     return Op;
642   }
643 
644   static std::unique_ptr<RISCVOperand> createReg(unsigned RegNo, SMLoc S,
645                                                  SMLoc E, bool IsRV64) {
646     auto Op = std::make_unique<RISCVOperand>(KindTy::Register);
647     Op->Reg.RegNum = RegNo;
648     Op->StartLoc = S;
649     Op->EndLoc = E;
650     Op->IsRV64 = IsRV64;
651     return Op;
652   }
653 
654   static std::unique_ptr<RISCVOperand> createImm(const MCExpr *Val, SMLoc S,
655                                                  SMLoc E, bool IsRV64) {
656     auto Op = std::make_unique<RISCVOperand>(KindTy::Immediate);
657     Op->Imm.Val = Val;
658     Op->StartLoc = S;
659     Op->EndLoc = E;
660     Op->IsRV64 = IsRV64;
661     return Op;
662   }
663 
664   static std::unique_ptr<RISCVOperand>
665   createSysReg(StringRef Str, SMLoc S, unsigned Encoding, bool IsRV64) {
666     auto Op = std::make_unique<RISCVOperand>(KindTy::SystemRegister);
667     Op->SysReg.Data = Str.data();
668     Op->SysReg.Length = Str.size();
669     Op->SysReg.Encoding = Encoding;
670     Op->StartLoc = S;
671     Op->IsRV64 = IsRV64;
672     return Op;
673   }
674 
675   void addExpr(MCInst &Inst, const MCExpr *Expr) const {
676     assert(Expr && "Expr shouldn't be null!");
677     int64_t Imm = 0;
678     RISCVMCExpr::VariantKind VK = RISCVMCExpr::VK_RISCV_None;
679     bool IsConstant = evaluateConstantImm(Expr, Imm, VK);
680 
681     if (IsConstant)
682       Inst.addOperand(MCOperand::createImm(Imm));
683     else
684       Inst.addOperand(MCOperand::createExpr(Expr));
685   }
686 
687   // Used by the TableGen Code
688   void addRegOperands(MCInst &Inst, unsigned N) const {
689     assert(N == 1 && "Invalid number of operands!");
690     Inst.addOperand(MCOperand::createReg(getReg()));
691   }
692 
693   void addImmOperands(MCInst &Inst, unsigned N) const {
694     assert(N == 1 && "Invalid number of operands!");
695     addExpr(Inst, getImm());
696   }
697 
698   void addFenceArgOperands(MCInst &Inst, unsigned N) const {
699     assert(N == 1 && "Invalid number of operands!");
700     // isFenceArg has validated the operand, meaning this cast is safe
701     auto SE = cast<MCSymbolRefExpr>(getImm());
702 
703     unsigned Imm = 0;
704     for (char c : SE->getSymbol().getName()) {
705       switch (c) {
706       default:
707         llvm_unreachable("FenceArg must contain only [iorw]");
708       case 'i': Imm |= RISCVFenceField::I; break;
709       case 'o': Imm |= RISCVFenceField::O; break;
710       case 'r': Imm |= RISCVFenceField::R; break;
711       case 'w': Imm |= RISCVFenceField::W; break;
712       }
713     }
714     Inst.addOperand(MCOperand::createImm(Imm));
715   }
716 
717   void addCSRSystemRegisterOperands(MCInst &Inst, unsigned N) const {
718     assert(N == 1 && "Invalid number of operands!");
719     Inst.addOperand(MCOperand::createImm(SysReg.Encoding));
720   }
721 
722   // Returns the rounding mode represented by this RISCVOperand. Should only
723   // be called after checking isFRMArg.
724   RISCVFPRndMode::RoundingMode getRoundingMode() const {
725     // isFRMArg has validated the operand, meaning this cast is safe.
726     auto SE = cast<MCSymbolRefExpr>(getImm());
727     RISCVFPRndMode::RoundingMode FRM =
728         RISCVFPRndMode::stringToRoundingMode(SE->getSymbol().getName());
729     assert(FRM != RISCVFPRndMode::Invalid && "Invalid rounding mode");
730     return FRM;
731   }
732 
733   void addFRMArgOperands(MCInst &Inst, unsigned N) const {
734     assert(N == 1 && "Invalid number of operands!");
735     Inst.addOperand(MCOperand::createImm(getRoundingMode()));
736   }
737 };
738 } // end anonymous namespace.
739 
740 #define GET_REGISTER_MATCHER
741 #define GET_SUBTARGET_FEATURE_NAME
742 #define GET_MATCHER_IMPLEMENTATION
743 #define GET_MNEMONIC_SPELL_CHECKER
744 #include "RISCVGenAsmMatcher.inc"
745 
746 static Register convertFPR64ToFPR32(Register Reg) {
747   assert(Reg >= RISCV::F0_D && Reg <= RISCV::F31_D && "Invalid register");
748   return Reg - RISCV::F0_D + RISCV::F0_F;
749 }
750 
751 unsigned RISCVAsmParser::validateTargetOperandClass(MCParsedAsmOperand &AsmOp,
752                                                     unsigned Kind) {
753   RISCVOperand &Op = static_cast<RISCVOperand &>(AsmOp);
754   if (!Op.isReg())
755     return Match_InvalidOperand;
756 
757   Register Reg = Op.getReg();
758   bool IsRegFPR64 =
759       RISCVMCRegisterClasses[RISCV::FPR64RegClassID].contains(Reg);
760   bool IsRegFPR64C =
761       RISCVMCRegisterClasses[RISCV::FPR64CRegClassID].contains(Reg);
762 
763   // As the parser couldn't differentiate an FPR32 from an FPR64, coerce the
764   // register from FPR64 to FPR32 or FPR64C to FPR32C if necessary.
765   if ((IsRegFPR64 && Kind == MCK_FPR32) ||
766       (IsRegFPR64C && Kind == MCK_FPR32C)) {
767     Op.Reg.RegNum = convertFPR64ToFPR32(Reg);
768     return Match_Success;
769   }
770   return Match_InvalidOperand;
771 }
772 
773 bool RISCVAsmParser::generateImmOutOfRangeError(
774     OperandVector &Operands, uint64_t ErrorInfo, int64_t Lower, int64_t Upper,
775     Twine Msg = "immediate must be an integer in the range") {
776   SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
777   return Error(ErrorLoc, Msg + " [" + Twine(Lower) + ", " + Twine(Upper) + "]");
778 }
779 
780 static std::string RISCVMnemonicSpellCheck(StringRef S,
781                                           const FeatureBitset &FBS,
782                                           unsigned VariantID = 0);
783 
784 bool RISCVAsmParser::MatchAndEmitInstruction(SMLoc IDLoc, unsigned &Opcode,
785                                              OperandVector &Operands,
786                                              MCStreamer &Out,
787                                              uint64_t &ErrorInfo,
788                                              bool MatchingInlineAsm) {
789   MCInst Inst;
790   FeatureBitset MissingFeatures;
791 
792   auto Result =
793     MatchInstructionImpl(Operands, Inst, ErrorInfo, MissingFeatures,
794                          MatchingInlineAsm);
795   switch (Result) {
796   default:
797     break;
798   case Match_Success:
799     return processInstruction(Inst, IDLoc, Operands, Out);
800   case Match_MissingFeature: {
801     assert(MissingFeatures.any() && "Unknown missing features!");
802     bool FirstFeature = true;
803     std::string Msg = "instruction requires the following:";
804     for (unsigned i = 0, e = MissingFeatures.size(); i != e; ++i) {
805       if (MissingFeatures[i]) {
806         Msg += FirstFeature ? " " : ", ";
807         Msg += getSubtargetFeatureName(i);
808         FirstFeature = false;
809       }
810     }
811     return Error(IDLoc, Msg);
812   }
813   case Match_MnemonicFail: {
814     FeatureBitset FBS = ComputeAvailableFeatures(getSTI().getFeatureBits());
815     std::string Suggestion = RISCVMnemonicSpellCheck(
816       ((RISCVOperand &)*Operands[0]).getToken(), FBS);
817     return Error(IDLoc, "unrecognized instruction mnemonic" + Suggestion);
818   }
819   case Match_InvalidOperand: {
820     SMLoc ErrorLoc = IDLoc;
821     if (ErrorInfo != ~0U) {
822       if (ErrorInfo >= Operands.size())
823         return Error(ErrorLoc, "too few operands for instruction");
824 
825       ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
826       if (ErrorLoc == SMLoc())
827         ErrorLoc = IDLoc;
828     }
829     return Error(ErrorLoc, "invalid operand for instruction");
830   }
831   }
832 
833   // Handle the case when the error message is of specific type
834   // other than the generic Match_InvalidOperand, and the
835   // corresponding operand is missing.
836   if (Result > FIRST_TARGET_MATCH_RESULT_TY) {
837     SMLoc ErrorLoc = IDLoc;
838     if (ErrorInfo != ~0U && ErrorInfo >= Operands.size())
839         return Error(ErrorLoc, "too few operands for instruction");
840   }
841 
842   switch(Result) {
843   default:
844     break;
845   case Match_InvalidImmXLenLI:
846     if (isRV64()) {
847       SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
848       return Error(ErrorLoc, "operand must be a constant 64-bit integer");
849     }
850     return generateImmOutOfRangeError(Operands, ErrorInfo,
851                                       std::numeric_limits<int32_t>::min(),
852                                       std::numeric_limits<uint32_t>::max());
853   case Match_InvalidImmZero: {
854     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
855     return Error(ErrorLoc, "immediate must be zero");
856   }
857   case Match_InvalidUImmLog2XLen:
858     if (isRV64())
859       return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 6) - 1);
860     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 5) - 1);
861   case Match_InvalidUImmLog2XLenNonZero:
862     if (isRV64())
863       return generateImmOutOfRangeError(Operands, ErrorInfo, 1, (1 << 6) - 1);
864     return generateImmOutOfRangeError(Operands, ErrorInfo, 1, (1 << 5) - 1);
865   case Match_InvalidUImm5:
866     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 5) - 1);
867   case Match_InvalidSImm6:
868     return generateImmOutOfRangeError(Operands, ErrorInfo, -(1 << 5),
869                                       (1 << 5) - 1);
870   case Match_InvalidSImm6NonZero:
871     return generateImmOutOfRangeError(
872         Operands, ErrorInfo, -(1 << 5), (1 << 5) - 1,
873         "immediate must be non-zero in the range");
874   case Match_InvalidCLUIImm:
875     return generateImmOutOfRangeError(
876         Operands, ErrorInfo, 1, (1 << 5) - 1,
877         "immediate must be in [0xfffe0, 0xfffff] or");
878   case Match_InvalidUImm7Lsb00:
879     return generateImmOutOfRangeError(
880         Operands, ErrorInfo, 0, (1 << 7) - 4,
881         "immediate must be a multiple of 4 bytes in the range");
882   case Match_InvalidUImm8Lsb00:
883     return generateImmOutOfRangeError(
884         Operands, ErrorInfo, 0, (1 << 8) - 4,
885         "immediate must be a multiple of 4 bytes in the range");
886   case Match_InvalidUImm8Lsb000:
887     return generateImmOutOfRangeError(
888         Operands, ErrorInfo, 0, (1 << 8) - 8,
889         "immediate must be a multiple of 8 bytes in the range");
890   case Match_InvalidSImm9Lsb0:
891     return generateImmOutOfRangeError(
892         Operands, ErrorInfo, -(1 << 8), (1 << 8) - 2,
893         "immediate must be a multiple of 2 bytes in the range");
894   case Match_InvalidUImm9Lsb000:
895     return generateImmOutOfRangeError(
896         Operands, ErrorInfo, 0, (1 << 9) - 8,
897         "immediate must be a multiple of 8 bytes in the range");
898   case Match_InvalidUImm10Lsb00NonZero:
899     return generateImmOutOfRangeError(
900         Operands, ErrorInfo, 4, (1 << 10) - 4,
901         "immediate must be a multiple of 4 bytes in the range");
902   case Match_InvalidSImm10Lsb0000NonZero:
903     return generateImmOutOfRangeError(
904         Operands, ErrorInfo, -(1 << 9), (1 << 9) - 16,
905         "immediate must be a multiple of 16 bytes and non-zero in the range");
906   case Match_InvalidSImm12:
907     return generateImmOutOfRangeError(
908         Operands, ErrorInfo, -(1 << 11), (1 << 11) - 1,
909         "operand must be a symbol with %lo/%pcrel_lo/%tprel_lo modifier or an "
910         "integer in the range");
911   case Match_InvalidSImm12Lsb0:
912     return generateImmOutOfRangeError(
913         Operands, ErrorInfo, -(1 << 11), (1 << 11) - 2,
914         "immediate must be a multiple of 2 bytes in the range");
915   case Match_InvalidSImm13Lsb0:
916     return generateImmOutOfRangeError(
917         Operands, ErrorInfo, -(1 << 12), (1 << 12) - 2,
918         "immediate must be a multiple of 2 bytes in the range");
919   case Match_InvalidUImm20LUI:
920     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 20) - 1,
921                                       "operand must be a symbol with "
922                                       "%hi/%tprel_hi modifier or an integer in "
923                                       "the range");
924   case Match_InvalidUImm20AUIPC:
925     return generateImmOutOfRangeError(
926         Operands, ErrorInfo, 0, (1 << 20) - 1,
927         "operand must be a symbol with a "
928         "%pcrel_hi/%got_pcrel_hi/%tls_ie_pcrel_hi/%tls_gd_pcrel_hi modifier or "
929         "an integer in the range");
930   case Match_InvalidSImm21Lsb0JAL:
931     return generateImmOutOfRangeError(
932         Operands, ErrorInfo, -(1 << 20), (1 << 20) - 2,
933         "immediate must be a multiple of 2 bytes in the range");
934   case Match_InvalidCSRSystemRegister: {
935     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 12) - 1,
936                                       "operand must be a valid system register "
937                                       "name or an integer in the range");
938   }
939   case Match_InvalidFenceArg: {
940     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
941     return Error(
942         ErrorLoc,
943         "operand must be formed of letters selected in-order from 'iorw'");
944   }
945   case Match_InvalidFRMArg: {
946     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
947     return Error(
948         ErrorLoc,
949         "operand must be a valid floating point rounding mode mnemonic");
950   }
951   case Match_InvalidBareSymbol: {
952     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
953     return Error(ErrorLoc, "operand must be a bare symbol name");
954   }
955   case Match_InvalidCallSymbol: {
956     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
957     return Error(ErrorLoc, "operand must be a bare symbol name");
958   }
959   case Match_InvalidTPRelAddSymbol: {
960     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
961     return Error(ErrorLoc, "operand must be a symbol with %tprel_add modifier");
962   }
963   }
964 
965   llvm_unreachable("Unknown match type detected!");
966 }
967 
968 // Attempts to match Name as a register (either using the default name or
969 // alternative ABI names), setting RegNo to the matching register. Upon
970 // failure, returns true and sets RegNo to 0. If IsRV32E then registers
971 // x16-x31 will be rejected.
972 static bool matchRegisterNameHelper(bool IsRV32E, Register &RegNo,
973                                     StringRef Name) {
974   RegNo = MatchRegisterName(Name);
975   // The 32- and 64-bit FPRs have the same asm name. Check that the initial
976   // match always matches the 64-bit variant, and not the 32-bit one.
977   assert(!(RegNo >= RISCV::F0_F && RegNo <= RISCV::F31_F));
978   // The default FPR register class is based on the tablegen enum ordering.
979   static_assert(RISCV::F0_D < RISCV::F0_F, "FPR matching must be updated");
980   if (RegNo == RISCV::NoRegister)
981     RegNo = MatchRegisterAltName(Name);
982   if (IsRV32E && RegNo >= RISCV::X16 && RegNo <= RISCV::X31)
983     RegNo = RISCV::NoRegister;
984   return RegNo == RISCV::NoRegister;
985 }
986 
987 bool RISCVAsmParser::ParseRegister(unsigned &RegNo, SMLoc &StartLoc,
988                                    SMLoc &EndLoc) {
989   const AsmToken &Tok = getParser().getTok();
990   StartLoc = Tok.getLoc();
991   EndLoc = Tok.getEndLoc();
992   RegNo = 0;
993   StringRef Name = getLexer().getTok().getIdentifier();
994 
995   if (matchRegisterNameHelper(isRV32E(), (Register&)RegNo, Name))
996     return Error(StartLoc, "invalid register name");
997 
998   getParser().Lex(); // Eat identifier token.
999   return false;
1000 }
1001 
1002 OperandMatchResultTy RISCVAsmParser::parseRegister(OperandVector &Operands,
1003                                                    bool AllowParens) {
1004   SMLoc FirstS = getLoc();
1005   bool HadParens = false;
1006   AsmToken LParen;
1007 
1008   // If this is an LParen and a parenthesised register name is allowed, parse it
1009   // atomically.
1010   if (AllowParens && getLexer().is(AsmToken::LParen)) {
1011     AsmToken Buf[2];
1012     size_t ReadCount = getLexer().peekTokens(Buf);
1013     if (ReadCount == 2 && Buf[1].getKind() == AsmToken::RParen) {
1014       HadParens = true;
1015       LParen = getParser().getTok();
1016       getParser().Lex(); // Eat '('
1017     }
1018   }
1019 
1020   switch (getLexer().getKind()) {
1021   default:
1022     if (HadParens)
1023       getLexer().UnLex(LParen);
1024     return MatchOperand_NoMatch;
1025   case AsmToken::Identifier:
1026     StringRef Name = getLexer().getTok().getIdentifier();
1027     Register RegNo;
1028     matchRegisterNameHelper(isRV32E(), RegNo, Name);
1029 
1030     if (RegNo == RISCV::NoRegister) {
1031       if (HadParens)
1032         getLexer().UnLex(LParen);
1033       return MatchOperand_NoMatch;
1034     }
1035     if (HadParens)
1036       Operands.push_back(RISCVOperand::createToken("(", FirstS, isRV64()));
1037     SMLoc S = getLoc();
1038     SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1039     getLexer().Lex();
1040     Operands.push_back(RISCVOperand::createReg(RegNo, S, E, isRV64()));
1041   }
1042 
1043   if (HadParens) {
1044     getParser().Lex(); // Eat ')'
1045     Operands.push_back(RISCVOperand::createToken(")", getLoc(), isRV64()));
1046   }
1047 
1048   return MatchOperand_Success;
1049 }
1050 
1051 OperandMatchResultTy
1052 RISCVAsmParser::parseCSRSystemRegister(OperandVector &Operands) {
1053   SMLoc S = getLoc();
1054   const MCExpr *Res;
1055 
1056   switch (getLexer().getKind()) {
1057   default:
1058     return MatchOperand_NoMatch;
1059   case AsmToken::LParen:
1060   case AsmToken::Minus:
1061   case AsmToken::Plus:
1062   case AsmToken::Exclaim:
1063   case AsmToken::Tilde:
1064   case AsmToken::Integer:
1065   case AsmToken::String: {
1066     if (getParser().parseExpression(Res))
1067       return MatchOperand_ParseFail;
1068 
1069     auto *CE = dyn_cast<MCConstantExpr>(Res);
1070     if (CE) {
1071       int64_t Imm = CE->getValue();
1072       if (isUInt<12>(Imm)) {
1073         auto SysReg = RISCVSysReg::lookupSysRegByEncoding(Imm);
1074         // Accept an immediate representing a named or un-named Sys Reg
1075         // if the range is valid, regardless of the required features.
1076         Operands.push_back(RISCVOperand::createSysReg(
1077             SysReg ? SysReg->Name : "", S, Imm, isRV64()));
1078         return MatchOperand_Success;
1079       }
1080     }
1081 
1082     Twine Msg = "immediate must be an integer in the range";
1083     Error(S, Msg + " [" + Twine(0) + ", " + Twine((1 << 12) - 1) + "]");
1084     return MatchOperand_ParseFail;
1085   }
1086   case AsmToken::Identifier: {
1087     StringRef Identifier;
1088     if (getParser().parseIdentifier(Identifier))
1089       return MatchOperand_ParseFail;
1090 
1091     auto SysReg = RISCVSysReg::lookupSysRegByName(Identifier);
1092     // Accept a named Sys Reg if the required features are present.
1093     if (SysReg) {
1094       if (!SysReg->haveRequiredFeatures(getSTI().getFeatureBits())) {
1095         Error(S, "system register use requires an option to be enabled");
1096         return MatchOperand_ParseFail;
1097       }
1098       Operands.push_back(RISCVOperand::createSysReg(
1099           Identifier, S, SysReg->Encoding, isRV64()));
1100       return MatchOperand_Success;
1101     }
1102 
1103     Twine Msg = "operand must be a valid system register name "
1104                 "or an integer in the range";
1105     Error(S, Msg + " [" + Twine(0) + ", " + Twine((1 << 12) - 1) + "]");
1106     return MatchOperand_ParseFail;
1107   }
1108   case AsmToken::Percent: {
1109     // Discard operand with modifier.
1110     Twine Msg = "immediate must be an integer in the range";
1111     Error(S, Msg + " [" + Twine(0) + ", " + Twine((1 << 12) - 1) + "]");
1112     return MatchOperand_ParseFail;
1113   }
1114   }
1115 
1116   return MatchOperand_NoMatch;
1117 }
1118 
1119 OperandMatchResultTy RISCVAsmParser::parseImmediate(OperandVector &Operands) {
1120   SMLoc S = getLoc();
1121   SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1122   const MCExpr *Res;
1123 
1124   switch (getLexer().getKind()) {
1125   default:
1126     return MatchOperand_NoMatch;
1127   case AsmToken::LParen:
1128   case AsmToken::Dot:
1129   case AsmToken::Minus:
1130   case AsmToken::Plus:
1131   case AsmToken::Exclaim:
1132   case AsmToken::Tilde:
1133   case AsmToken::Integer:
1134   case AsmToken::String:
1135   case AsmToken::Identifier:
1136     if (getParser().parseExpression(Res))
1137       return MatchOperand_ParseFail;
1138     break;
1139   case AsmToken::Percent:
1140     return parseOperandWithModifier(Operands);
1141   }
1142 
1143   Operands.push_back(RISCVOperand::createImm(Res, S, E, isRV64()));
1144   return MatchOperand_Success;
1145 }
1146 
1147 OperandMatchResultTy
1148 RISCVAsmParser::parseOperandWithModifier(OperandVector &Operands) {
1149   SMLoc S = getLoc();
1150   SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1151 
1152   if (getLexer().getKind() != AsmToken::Percent) {
1153     Error(getLoc(), "expected '%' for operand modifier");
1154     return MatchOperand_ParseFail;
1155   }
1156 
1157   getParser().Lex(); // Eat '%'
1158 
1159   if (getLexer().getKind() != AsmToken::Identifier) {
1160     Error(getLoc(), "expected valid identifier for operand modifier");
1161     return MatchOperand_ParseFail;
1162   }
1163   StringRef Identifier = getParser().getTok().getIdentifier();
1164   RISCVMCExpr::VariantKind VK = RISCVMCExpr::getVariantKindForName(Identifier);
1165   if (VK == RISCVMCExpr::VK_RISCV_Invalid) {
1166     Error(getLoc(), "unrecognized operand modifier");
1167     return MatchOperand_ParseFail;
1168   }
1169 
1170   getParser().Lex(); // Eat the identifier
1171   if (getLexer().getKind() != AsmToken::LParen) {
1172     Error(getLoc(), "expected '('");
1173     return MatchOperand_ParseFail;
1174   }
1175   getParser().Lex(); // Eat '('
1176 
1177   const MCExpr *SubExpr;
1178   if (getParser().parseParenExpression(SubExpr, E)) {
1179     return MatchOperand_ParseFail;
1180   }
1181 
1182   const MCExpr *ModExpr = RISCVMCExpr::create(SubExpr, VK, getContext());
1183   Operands.push_back(RISCVOperand::createImm(ModExpr, S, E, isRV64()));
1184   return MatchOperand_Success;
1185 }
1186 
1187 OperandMatchResultTy RISCVAsmParser::parseBareSymbol(OperandVector &Operands) {
1188   SMLoc S = getLoc();
1189   SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1190   const MCExpr *Res;
1191 
1192   if (getLexer().getKind() != AsmToken::Identifier)
1193     return MatchOperand_NoMatch;
1194 
1195   StringRef Identifier;
1196   AsmToken Tok = getLexer().getTok();
1197 
1198   if (getParser().parseIdentifier(Identifier))
1199     return MatchOperand_ParseFail;
1200 
1201   if (Identifier.consume_back("@plt")) {
1202     Error(getLoc(), "'@plt' operand not valid for instruction");
1203     return MatchOperand_ParseFail;
1204   }
1205 
1206   MCSymbol *Sym = getContext().getOrCreateSymbol(Identifier);
1207 
1208   if (Sym->isVariable()) {
1209     const MCExpr *V = Sym->getVariableValue(/*SetUsed=*/false);
1210     if (!isa<MCSymbolRefExpr>(V)) {
1211       getLexer().UnLex(Tok); // Put back if it's not a bare symbol.
1212       return MatchOperand_NoMatch;
1213     }
1214     Res = V;
1215   } else
1216     Res = MCSymbolRefExpr::create(Sym, MCSymbolRefExpr::VK_None, getContext());
1217 
1218   MCBinaryExpr::Opcode Opcode;
1219   switch (getLexer().getKind()) {
1220   default:
1221     Operands.push_back(RISCVOperand::createImm(Res, S, E, isRV64()));
1222     return MatchOperand_Success;
1223   case AsmToken::Plus:
1224     Opcode = MCBinaryExpr::Add;
1225     break;
1226   case AsmToken::Minus:
1227     Opcode = MCBinaryExpr::Sub;
1228     break;
1229   }
1230 
1231   const MCExpr *Expr;
1232   if (getParser().parseExpression(Expr))
1233     return MatchOperand_ParseFail;
1234   Res = MCBinaryExpr::create(Opcode, Res, Expr, getContext());
1235   Operands.push_back(RISCVOperand::createImm(Res, S, E, isRV64()));
1236   return MatchOperand_Success;
1237 }
1238 
1239 OperandMatchResultTy RISCVAsmParser::parseCallSymbol(OperandVector &Operands) {
1240   SMLoc S = getLoc();
1241   SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1242   const MCExpr *Res;
1243 
1244   if (getLexer().getKind() != AsmToken::Identifier)
1245     return MatchOperand_NoMatch;
1246 
1247   // Avoid parsing the register in `call rd, foo` as a call symbol.
1248   if (getLexer().peekTok().getKind() != AsmToken::EndOfStatement)
1249     return MatchOperand_NoMatch;
1250 
1251   StringRef Identifier;
1252   if (getParser().parseIdentifier(Identifier))
1253     return MatchOperand_ParseFail;
1254 
1255   RISCVMCExpr::VariantKind Kind = RISCVMCExpr::VK_RISCV_CALL;
1256   if (Identifier.consume_back("@plt"))
1257     Kind = RISCVMCExpr::VK_RISCV_CALL_PLT;
1258 
1259   MCSymbol *Sym = getContext().getOrCreateSymbol(Identifier);
1260   Res = MCSymbolRefExpr::create(Sym, MCSymbolRefExpr::VK_None, getContext());
1261   Res = RISCVMCExpr::create(Res, Kind, getContext());
1262   Operands.push_back(RISCVOperand::createImm(Res, S, E, isRV64()));
1263   return MatchOperand_Success;
1264 }
1265 
1266 OperandMatchResultTy RISCVAsmParser::parseJALOffset(OperandVector &Operands) {
1267   // Parsing jal operands is fiddly due to the `jal foo` and `jal ra, foo`
1268   // both being acceptable forms. When parsing `jal ra, foo` this function
1269   // will be called for the `ra` register operand in an attempt to match the
1270   // single-operand alias. parseJALOffset must fail for this case. It would
1271   // seem logical to try parse the operand using parseImmediate and return
1272   // NoMatch if the next token is a comma (meaning we must be parsing a jal in
1273   // the second form rather than the first). We can't do this as there's no
1274   // way of rewinding the lexer state. Instead, return NoMatch if this operand
1275   // is an identifier and is followed by a comma.
1276   if (getLexer().is(AsmToken::Identifier) &&
1277       getLexer().peekTok().is(AsmToken::Comma))
1278     return MatchOperand_NoMatch;
1279 
1280   return parseImmediate(Operands);
1281 }
1282 
1283 OperandMatchResultTy
1284 RISCVAsmParser::parseMemOpBaseReg(OperandVector &Operands) {
1285   if (getLexer().isNot(AsmToken::LParen)) {
1286     Error(getLoc(), "expected '('");
1287     return MatchOperand_ParseFail;
1288   }
1289 
1290   getParser().Lex(); // Eat '('
1291   Operands.push_back(RISCVOperand::createToken("(", getLoc(), isRV64()));
1292 
1293   if (parseRegister(Operands) != MatchOperand_Success) {
1294     Error(getLoc(), "expected register");
1295     return MatchOperand_ParseFail;
1296   }
1297 
1298   if (getLexer().isNot(AsmToken::RParen)) {
1299     Error(getLoc(), "expected ')'");
1300     return MatchOperand_ParseFail;
1301   }
1302 
1303   getParser().Lex(); // Eat ')'
1304   Operands.push_back(RISCVOperand::createToken(")", getLoc(), isRV64()));
1305 
1306   return MatchOperand_Success;
1307 }
1308 
1309 OperandMatchResultTy RISCVAsmParser::parseAtomicMemOp(OperandVector &Operands) {
1310   // Atomic operations such as lr.w, sc.w, and amo*.w accept a "memory operand"
1311   // as one of their register operands, such as `(a0)`. This just denotes that
1312   // the register (in this case `a0`) contains a memory address.
1313   //
1314   // Normally, we would be able to parse these by putting the parens into the
1315   // instruction string. However, GNU as also accepts a zero-offset memory
1316   // operand (such as `0(a0)`), and ignores the 0. Normally this would be parsed
1317   // with parseImmediate followed by parseMemOpBaseReg, but these instructions
1318   // do not accept an immediate operand, and we do not want to add a "dummy"
1319   // operand that is silently dropped.
1320   //
1321   // Instead, we use this custom parser. This will: allow (and discard) an
1322   // offset if it is zero; require (and discard) parentheses; and add only the
1323   // parsed register operand to `Operands`.
1324   //
1325   // These operands are printed with RISCVInstPrinter::printAtomicMemOp, which
1326   // will only print the register surrounded by parentheses (which GNU as also
1327   // uses as its canonical representation for these operands).
1328   std::unique_ptr<RISCVOperand> OptionalImmOp;
1329 
1330   if (getLexer().isNot(AsmToken::LParen)) {
1331     // Parse an Integer token. We do not accept arbritrary constant expressions
1332     // in the offset field (because they may include parens, which complicates
1333     // parsing a lot).
1334     int64_t ImmVal;
1335     SMLoc ImmStart = getLoc();
1336     if (getParser().parseIntToken(ImmVal,
1337                                   "expected '(' or optional integer offset"))
1338       return MatchOperand_ParseFail;
1339 
1340     // Create a RISCVOperand for checking later (so the error messages are
1341     // nicer), but we don't add it to Operands.
1342     SMLoc ImmEnd = getLoc();
1343     OptionalImmOp =
1344         RISCVOperand::createImm(MCConstantExpr::create(ImmVal, getContext()),
1345                                 ImmStart, ImmEnd, isRV64());
1346   }
1347 
1348   if (getLexer().isNot(AsmToken::LParen)) {
1349     Error(getLoc(), OptionalImmOp ? "expected '(' after optional integer offset"
1350                                   : "expected '(' or optional integer offset");
1351     return MatchOperand_ParseFail;
1352   }
1353   getParser().Lex(); // Eat '('
1354 
1355   if (parseRegister(Operands) != MatchOperand_Success) {
1356     Error(getLoc(), "expected register");
1357     return MatchOperand_ParseFail;
1358   }
1359 
1360   if (getLexer().isNot(AsmToken::RParen)) {
1361     Error(getLoc(), "expected ')'");
1362     return MatchOperand_ParseFail;
1363   }
1364   getParser().Lex(); // Eat ')'
1365 
1366   // Deferred Handling of non-zero offsets. This makes the error messages nicer.
1367   if (OptionalImmOp && !OptionalImmOp->isImmZero()) {
1368     Error(OptionalImmOp->getStartLoc(), "optional integer offset must be 0",
1369           SMRange(OptionalImmOp->getStartLoc(), OptionalImmOp->getEndLoc()));
1370     return MatchOperand_ParseFail;
1371   }
1372 
1373   return MatchOperand_Success;
1374 }
1375 
1376 /// Looks at a token type and creates the relevant operand from this
1377 /// information, adding to Operands. If operand was parsed, returns false, else
1378 /// true.
1379 bool RISCVAsmParser::parseOperand(OperandVector &Operands, StringRef Mnemonic) {
1380   // Check if the current operand has a custom associated parser, if so, try to
1381   // custom parse the operand, or fallback to the general approach.
1382   OperandMatchResultTy Result =
1383       MatchOperandParserImpl(Operands, Mnemonic, /*ParseForAllFeatures=*/true);
1384   if (Result == MatchOperand_Success)
1385     return false;
1386   if (Result == MatchOperand_ParseFail)
1387     return true;
1388 
1389   // Attempt to parse token as a register.
1390   if (parseRegister(Operands, true) == MatchOperand_Success)
1391     return false;
1392 
1393   // Attempt to parse token as an immediate
1394   if (parseImmediate(Operands) == MatchOperand_Success) {
1395     // Parse memory base register if present
1396     if (getLexer().is(AsmToken::LParen))
1397       return parseMemOpBaseReg(Operands) != MatchOperand_Success;
1398     return false;
1399   }
1400 
1401   // Finally we have exhausted all options and must declare defeat.
1402   Error(getLoc(), "unknown operand");
1403   return true;
1404 }
1405 
1406 bool RISCVAsmParser::ParseInstruction(ParseInstructionInfo &Info,
1407                                       StringRef Name, SMLoc NameLoc,
1408                                       OperandVector &Operands) {
1409   // Ensure that if the instruction occurs when relaxation is enabled,
1410   // relocations are forced for the file. Ideally this would be done when there
1411   // is enough information to reliably determine if the instruction itself may
1412   // cause relaxations. Unfortunately instruction processing stage occurs in the
1413   // same pass as relocation emission, so it's too late to set a 'sticky bit'
1414   // for the entire file.
1415   if (getSTI().getFeatureBits()[RISCV::FeatureRelax]) {
1416     auto *Assembler = getTargetStreamer().getStreamer().getAssemblerPtr();
1417     if (Assembler != nullptr) {
1418       RISCVAsmBackend &MAB =
1419           static_cast<RISCVAsmBackend &>(Assembler->getBackend());
1420       MAB.setForceRelocs();
1421     }
1422   }
1423 
1424   // First operand is token for instruction
1425   Operands.push_back(RISCVOperand::createToken(Name, NameLoc, isRV64()));
1426 
1427   // If there are no more operands, then finish
1428   if (getLexer().is(AsmToken::EndOfStatement))
1429     return false;
1430 
1431   // Parse first operand
1432   if (parseOperand(Operands, Name))
1433     return true;
1434 
1435   // Parse until end of statement, consuming commas between operands
1436   unsigned OperandIdx = 1;
1437   while (getLexer().is(AsmToken::Comma)) {
1438     // Consume comma token
1439     getLexer().Lex();
1440 
1441     // Parse next operand
1442     if (parseOperand(Operands, Name))
1443       return true;
1444 
1445     ++OperandIdx;
1446   }
1447 
1448   if (getLexer().isNot(AsmToken::EndOfStatement)) {
1449     SMLoc Loc = getLexer().getLoc();
1450     getParser().eatToEndOfStatement();
1451     return Error(Loc, "unexpected token");
1452   }
1453 
1454   getParser().Lex(); // Consume the EndOfStatement.
1455   return false;
1456 }
1457 
1458 bool RISCVAsmParser::classifySymbolRef(const MCExpr *Expr,
1459                                        RISCVMCExpr::VariantKind &Kind,
1460                                        int64_t &Addend) {
1461   Kind = RISCVMCExpr::VK_RISCV_None;
1462   Addend = 0;
1463 
1464   if (const RISCVMCExpr *RE = dyn_cast<RISCVMCExpr>(Expr)) {
1465     Kind = RE->getKind();
1466     Expr = RE->getSubExpr();
1467   }
1468 
1469   // It's a simple symbol reference or constant with no addend.
1470   if (isa<MCConstantExpr>(Expr) || isa<MCSymbolRefExpr>(Expr))
1471     return true;
1472 
1473   const MCBinaryExpr *BE = dyn_cast<MCBinaryExpr>(Expr);
1474   if (!BE)
1475     return false;
1476 
1477   if (!isa<MCSymbolRefExpr>(BE->getLHS()))
1478     return false;
1479 
1480   if (BE->getOpcode() != MCBinaryExpr::Add &&
1481       BE->getOpcode() != MCBinaryExpr::Sub)
1482     return false;
1483 
1484   // We are able to support the subtraction of two symbol references
1485   if (BE->getOpcode() == MCBinaryExpr::Sub &&
1486       isa<MCSymbolRefExpr>(BE->getRHS()))
1487     return true;
1488 
1489   // See if the addend is a constant, otherwise there's more going
1490   // on here than we can deal with.
1491   auto AddendExpr = dyn_cast<MCConstantExpr>(BE->getRHS());
1492   if (!AddendExpr)
1493     return false;
1494 
1495   Addend = AddendExpr->getValue();
1496   if (BE->getOpcode() == MCBinaryExpr::Sub)
1497     Addend = -Addend;
1498 
1499   // It's some symbol reference + a constant addend
1500   return Kind != RISCVMCExpr::VK_RISCV_Invalid;
1501 }
1502 
1503 bool RISCVAsmParser::ParseDirective(AsmToken DirectiveID) {
1504   // This returns false if this function recognizes the directive
1505   // regardless of whether it is successfully handles or reports an
1506   // error. Otherwise it returns true to give the generic parser a
1507   // chance at recognizing it.
1508   StringRef IDVal = DirectiveID.getString();
1509 
1510   if (IDVal == ".option")
1511     return parseDirectiveOption();
1512 
1513   return true;
1514 }
1515 
1516 bool RISCVAsmParser::parseDirectiveOption() {
1517   MCAsmParser &Parser = getParser();
1518   // Get the option token.
1519   AsmToken Tok = Parser.getTok();
1520   // At the moment only identifiers are supported.
1521   if (Tok.isNot(AsmToken::Identifier))
1522     return Error(Parser.getTok().getLoc(),
1523                  "unexpected token, expected identifier");
1524 
1525   StringRef Option = Tok.getIdentifier();
1526 
1527   if (Option == "push") {
1528     getTargetStreamer().emitDirectiveOptionPush();
1529 
1530     Parser.Lex();
1531     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1532       return Error(Parser.getTok().getLoc(),
1533                    "unexpected token, expected end of statement");
1534 
1535     pushFeatureBits();
1536     return false;
1537   }
1538 
1539   if (Option == "pop") {
1540     SMLoc StartLoc = Parser.getTok().getLoc();
1541     getTargetStreamer().emitDirectiveOptionPop();
1542 
1543     Parser.Lex();
1544     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1545       return Error(Parser.getTok().getLoc(),
1546                    "unexpected token, expected end of statement");
1547 
1548     if (popFeatureBits())
1549       return Error(StartLoc, ".option pop with no .option push");
1550 
1551     return false;
1552   }
1553 
1554   if (Option == "rvc") {
1555     getTargetStreamer().emitDirectiveOptionRVC();
1556 
1557     Parser.Lex();
1558     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1559       return Error(Parser.getTok().getLoc(),
1560                    "unexpected token, expected end of statement");
1561 
1562     setFeatureBits(RISCV::FeatureStdExtC, "c");
1563     return false;
1564   }
1565 
1566   if (Option == "norvc") {
1567     getTargetStreamer().emitDirectiveOptionNoRVC();
1568 
1569     Parser.Lex();
1570     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1571       return Error(Parser.getTok().getLoc(),
1572                    "unexpected token, expected end of statement");
1573 
1574     clearFeatureBits(RISCV::FeatureStdExtC, "c");
1575     return false;
1576   }
1577 
1578   if (Option == "relax") {
1579     getTargetStreamer().emitDirectiveOptionRelax();
1580 
1581     Parser.Lex();
1582     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1583       return Error(Parser.getTok().getLoc(),
1584                    "unexpected token, expected end of statement");
1585 
1586     setFeatureBits(RISCV::FeatureRelax, "relax");
1587     return false;
1588   }
1589 
1590   if (Option == "norelax") {
1591     getTargetStreamer().emitDirectiveOptionNoRelax();
1592 
1593     Parser.Lex();
1594     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1595       return Error(Parser.getTok().getLoc(),
1596                    "unexpected token, expected end of statement");
1597 
1598     clearFeatureBits(RISCV::FeatureRelax, "relax");
1599     return false;
1600   }
1601 
1602   // Unknown option.
1603   Warning(Parser.getTok().getLoc(),
1604           "unknown option, expected 'push', 'pop', 'rvc', 'norvc', 'relax' or "
1605           "'norelax'");
1606   Parser.eatToEndOfStatement();
1607   return false;
1608 }
1609 
1610 void RISCVAsmParser::emitToStreamer(MCStreamer &S, const MCInst &Inst) {
1611   MCInst CInst;
1612   bool Res = compressInst(CInst, Inst, getSTI(), S.getContext());
1613   CInst.setLoc(Inst.getLoc());
1614   S.EmitInstruction((Res ? CInst : Inst), getSTI());
1615 }
1616 
1617 void RISCVAsmParser::emitLoadImm(Register DestReg, int64_t Value,
1618                                  MCStreamer &Out) {
1619   RISCVMatInt::InstSeq Seq;
1620   RISCVMatInt::generateInstSeq(Value, isRV64(), Seq);
1621 
1622   Register SrcReg = RISCV::X0;
1623   for (RISCVMatInt::Inst &Inst : Seq) {
1624     if (Inst.Opc == RISCV::LUI) {
1625       emitToStreamer(
1626           Out, MCInstBuilder(RISCV::LUI).addReg(DestReg).addImm(Inst.Imm));
1627     } else {
1628       emitToStreamer(
1629           Out, MCInstBuilder(Inst.Opc).addReg(DestReg).addReg(SrcReg).addImm(
1630                    Inst.Imm));
1631     }
1632 
1633     // Only the first instruction has X0 as its source.
1634     SrcReg = DestReg;
1635   }
1636 }
1637 
1638 void RISCVAsmParser::emitAuipcInstPair(MCOperand DestReg, MCOperand TmpReg,
1639                                        const MCExpr *Symbol,
1640                                        RISCVMCExpr::VariantKind VKHi,
1641                                        unsigned SecondOpcode, SMLoc IDLoc,
1642                                        MCStreamer &Out) {
1643   // A pair of instructions for PC-relative addressing; expands to
1644   //   TmpLabel: AUIPC TmpReg, VKHi(symbol)
1645   //             OP DestReg, TmpReg, %pcrel_lo(TmpLabel)
1646   MCContext &Ctx = getContext();
1647 
1648   MCSymbol *TmpLabel = Ctx.createTempSymbol(
1649       "pcrel_hi", /* AlwaysAddSuffix */ true, /* CanBeUnnamed */ false);
1650   Out.EmitLabel(TmpLabel);
1651 
1652   const RISCVMCExpr *SymbolHi = RISCVMCExpr::create(Symbol, VKHi, Ctx);
1653   emitToStreamer(
1654       Out, MCInstBuilder(RISCV::AUIPC).addOperand(TmpReg).addExpr(SymbolHi));
1655 
1656   const MCExpr *RefToLinkTmpLabel =
1657       RISCVMCExpr::create(MCSymbolRefExpr::create(TmpLabel, Ctx),
1658                           RISCVMCExpr::VK_RISCV_PCREL_LO, Ctx);
1659 
1660   emitToStreamer(Out, MCInstBuilder(SecondOpcode)
1661                           .addOperand(DestReg)
1662                           .addOperand(TmpReg)
1663                           .addExpr(RefToLinkTmpLabel));
1664 }
1665 
1666 void RISCVAsmParser::emitLoadLocalAddress(MCInst &Inst, SMLoc IDLoc,
1667                                           MCStreamer &Out) {
1668   // The load local address pseudo-instruction "lla" is used in PC-relative
1669   // addressing of local symbols:
1670   //   lla rdest, symbol
1671   // expands to
1672   //   TmpLabel: AUIPC rdest, %pcrel_hi(symbol)
1673   //             ADDI rdest, rdest, %pcrel_lo(TmpLabel)
1674   MCOperand DestReg = Inst.getOperand(0);
1675   const MCExpr *Symbol = Inst.getOperand(1).getExpr();
1676   emitAuipcInstPair(DestReg, DestReg, Symbol, RISCVMCExpr::VK_RISCV_PCREL_HI,
1677                     RISCV::ADDI, IDLoc, Out);
1678 }
1679 
1680 void RISCVAsmParser::emitLoadAddress(MCInst &Inst, SMLoc IDLoc,
1681                                      MCStreamer &Out) {
1682   // The load address pseudo-instruction "la" is used in PC-relative and
1683   // GOT-indirect addressing of global symbols:
1684   //   la rdest, symbol
1685   // expands to either (for non-PIC)
1686   //   TmpLabel: AUIPC rdest, %pcrel_hi(symbol)
1687   //             ADDI rdest, rdest, %pcrel_lo(TmpLabel)
1688   // or (for PIC)
1689   //   TmpLabel: AUIPC rdest, %got_pcrel_hi(symbol)
1690   //             Lx rdest, %pcrel_lo(TmpLabel)(rdest)
1691   MCOperand DestReg = Inst.getOperand(0);
1692   const MCExpr *Symbol = Inst.getOperand(1).getExpr();
1693   unsigned SecondOpcode;
1694   RISCVMCExpr::VariantKind VKHi;
1695   // FIXME: Should check .option (no)pic when implemented
1696   if (getContext().getObjectFileInfo()->isPositionIndependent()) {
1697     SecondOpcode = isRV64() ? RISCV::LD : RISCV::LW;
1698     VKHi = RISCVMCExpr::VK_RISCV_GOT_HI;
1699   } else {
1700     SecondOpcode = RISCV::ADDI;
1701     VKHi = RISCVMCExpr::VK_RISCV_PCREL_HI;
1702   }
1703   emitAuipcInstPair(DestReg, DestReg, Symbol, VKHi, SecondOpcode, IDLoc, Out);
1704 }
1705 
1706 void RISCVAsmParser::emitLoadTLSIEAddress(MCInst &Inst, SMLoc IDLoc,
1707                                           MCStreamer &Out) {
1708   // The load TLS IE address pseudo-instruction "la.tls.ie" is used in
1709   // initial-exec TLS model addressing of global symbols:
1710   //   la.tls.ie rdest, symbol
1711   // expands to
1712   //   TmpLabel: AUIPC rdest, %tls_ie_pcrel_hi(symbol)
1713   //             Lx rdest, %pcrel_lo(TmpLabel)(rdest)
1714   MCOperand DestReg = Inst.getOperand(0);
1715   const MCExpr *Symbol = Inst.getOperand(1).getExpr();
1716   unsigned SecondOpcode = isRV64() ? RISCV::LD : RISCV::LW;
1717   emitAuipcInstPair(DestReg, DestReg, Symbol, RISCVMCExpr::VK_RISCV_TLS_GOT_HI,
1718                     SecondOpcode, IDLoc, Out);
1719 }
1720 
1721 void RISCVAsmParser::emitLoadTLSGDAddress(MCInst &Inst, SMLoc IDLoc,
1722                                           MCStreamer &Out) {
1723   // The load TLS GD address pseudo-instruction "la.tls.gd" is used in
1724   // global-dynamic TLS model addressing of global symbols:
1725   //   la.tls.gd rdest, symbol
1726   // expands to
1727   //   TmpLabel: AUIPC rdest, %tls_gd_pcrel_hi(symbol)
1728   //             ADDI rdest, rdest, %pcrel_lo(TmpLabel)
1729   MCOperand DestReg = Inst.getOperand(0);
1730   const MCExpr *Symbol = Inst.getOperand(1).getExpr();
1731   emitAuipcInstPair(DestReg, DestReg, Symbol, RISCVMCExpr::VK_RISCV_TLS_GD_HI,
1732                     RISCV::ADDI, IDLoc, Out);
1733 }
1734 
1735 void RISCVAsmParser::emitLoadStoreSymbol(MCInst &Inst, unsigned Opcode,
1736                                          SMLoc IDLoc, MCStreamer &Out,
1737                                          bool HasTmpReg) {
1738   // The load/store pseudo-instruction does a pc-relative load with
1739   // a symbol.
1740   //
1741   // The expansion looks like this
1742   //
1743   //   TmpLabel: AUIPC tmp, %pcrel_hi(symbol)
1744   //             [S|L]X    rd, %pcrel_lo(TmpLabel)(tmp)
1745   MCOperand DestReg = Inst.getOperand(0);
1746   unsigned SymbolOpIdx = HasTmpReg ? 2 : 1;
1747   unsigned TmpRegOpIdx = HasTmpReg ? 1 : 0;
1748   MCOperand TmpReg = Inst.getOperand(TmpRegOpIdx);
1749   const MCExpr *Symbol = Inst.getOperand(SymbolOpIdx).getExpr();
1750   emitAuipcInstPair(DestReg, TmpReg, Symbol, RISCVMCExpr::VK_RISCV_PCREL_HI,
1751                     Opcode, IDLoc, Out);
1752 }
1753 
1754 bool RISCVAsmParser::checkPseudoAddTPRel(MCInst &Inst,
1755                                          OperandVector &Operands) {
1756   assert(Inst.getOpcode() == RISCV::PseudoAddTPRel && "Invalid instruction");
1757   assert(Inst.getOperand(2).isReg() && "Unexpected second operand kind");
1758   if (Inst.getOperand(2).getReg() != RISCV::X4) {
1759     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[3]).getStartLoc();
1760     return Error(ErrorLoc, "the second input operand must be tp/x4 when using "
1761                            "%tprel_add modifier");
1762   }
1763 
1764   return false;
1765 }
1766 
1767 bool RISCVAsmParser::processInstruction(MCInst &Inst, SMLoc IDLoc,
1768                                         OperandVector &Operands,
1769                                         MCStreamer &Out) {
1770   Inst.setLoc(IDLoc);
1771 
1772   switch (Inst.getOpcode()) {
1773   default:
1774     break;
1775   case RISCV::PseudoLI: {
1776     Register Reg = Inst.getOperand(0).getReg();
1777     const MCOperand &Op1 = Inst.getOperand(1);
1778     if (Op1.isExpr()) {
1779       // We must have li reg, %lo(sym) or li reg, %pcrel_lo(sym) or similar.
1780       // Just convert to an addi. This allows compatibility with gas.
1781       emitToStreamer(Out, MCInstBuilder(RISCV::ADDI)
1782                               .addReg(Reg)
1783                               .addReg(RISCV::X0)
1784                               .addExpr(Op1.getExpr()));
1785       return false;
1786     }
1787     int64_t Imm = Inst.getOperand(1).getImm();
1788     // On RV32 the immediate here can either be a signed or an unsigned
1789     // 32-bit number. Sign extension has to be performed to ensure that Imm
1790     // represents the expected signed 64-bit number.
1791     if (!isRV64())
1792       Imm = SignExtend64<32>(Imm);
1793     emitLoadImm(Reg, Imm, Out);
1794     return false;
1795   }
1796   case RISCV::PseudoLLA:
1797     emitLoadLocalAddress(Inst, IDLoc, Out);
1798     return false;
1799   case RISCV::PseudoLA:
1800     emitLoadAddress(Inst, IDLoc, Out);
1801     return false;
1802   case RISCV::PseudoLA_TLS_IE:
1803     emitLoadTLSIEAddress(Inst, IDLoc, Out);
1804     return false;
1805   case RISCV::PseudoLA_TLS_GD:
1806     emitLoadTLSGDAddress(Inst, IDLoc, Out);
1807     return false;
1808   case RISCV::PseudoLB:
1809     emitLoadStoreSymbol(Inst, RISCV::LB, IDLoc, Out, /*HasTmpReg=*/false);
1810     return false;
1811   case RISCV::PseudoLBU:
1812     emitLoadStoreSymbol(Inst, RISCV::LBU, IDLoc, Out, /*HasTmpReg=*/false);
1813     return false;
1814   case RISCV::PseudoLH:
1815     emitLoadStoreSymbol(Inst, RISCV::LH, IDLoc, Out, /*HasTmpReg=*/false);
1816     return false;
1817   case RISCV::PseudoLHU:
1818     emitLoadStoreSymbol(Inst, RISCV::LHU, IDLoc, Out, /*HasTmpReg=*/false);
1819     return false;
1820   case RISCV::PseudoLW:
1821     emitLoadStoreSymbol(Inst, RISCV::LW, IDLoc, Out, /*HasTmpReg=*/false);
1822     return false;
1823   case RISCV::PseudoLWU:
1824     emitLoadStoreSymbol(Inst, RISCV::LWU, IDLoc, Out, /*HasTmpReg=*/false);
1825     return false;
1826   case RISCV::PseudoLD:
1827     emitLoadStoreSymbol(Inst, RISCV::LD, IDLoc, Out, /*HasTmpReg=*/false);
1828     return false;
1829   case RISCV::PseudoFLW:
1830     emitLoadStoreSymbol(Inst, RISCV::FLW, IDLoc, Out, /*HasTmpReg=*/true);
1831     return false;
1832   case RISCV::PseudoFLD:
1833     emitLoadStoreSymbol(Inst, RISCV::FLD, IDLoc, Out, /*HasTmpReg=*/true);
1834     return false;
1835   case RISCV::PseudoSB:
1836     emitLoadStoreSymbol(Inst, RISCV::SB, IDLoc, Out, /*HasTmpReg=*/true);
1837     return false;
1838   case RISCV::PseudoSH:
1839     emitLoadStoreSymbol(Inst, RISCV::SH, IDLoc, Out, /*HasTmpReg=*/true);
1840     return false;
1841   case RISCV::PseudoSW:
1842     emitLoadStoreSymbol(Inst, RISCV::SW, IDLoc, Out, /*HasTmpReg=*/true);
1843     return false;
1844   case RISCV::PseudoSD:
1845     emitLoadStoreSymbol(Inst, RISCV::SD, IDLoc, Out, /*HasTmpReg=*/true);
1846     return false;
1847   case RISCV::PseudoFSW:
1848     emitLoadStoreSymbol(Inst, RISCV::FSW, IDLoc, Out, /*HasTmpReg=*/true);
1849     return false;
1850   case RISCV::PseudoFSD:
1851     emitLoadStoreSymbol(Inst, RISCV::FSD, IDLoc, Out, /*HasTmpReg=*/true);
1852     return false;
1853   case RISCV::PseudoAddTPRel:
1854     if (checkPseudoAddTPRel(Inst, Operands))
1855       return true;
1856     break;
1857   }
1858 
1859   emitToStreamer(Out, Inst);
1860   return false;
1861 }
1862 
1863 extern "C" void LLVMInitializeRISCVAsmParser() {
1864   RegisterMCAsmParser<RISCVAsmParser> X(getTheRISCV32Target());
1865   RegisterMCAsmParser<RISCVAsmParser> Y(getTheRISCV64Target());
1866 }
1867