1 //===-- RISCVAsmParser.cpp - Parse RISCV assembly to MCInst instructions --===//
2 //
3 //                     The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 
10 #include "MCTargetDesc/RISCVAsmBackend.h"
11 #include "MCTargetDesc/RISCVMCExpr.h"
12 #include "MCTargetDesc/RISCVMCTargetDesc.h"
13 #include "MCTargetDesc/RISCVTargetStreamer.h"
14 #include "Utils/RISCVBaseInfo.h"
15 #include "llvm/ADT/STLExtras.h"
16 #include "llvm/ADT/StringSwitch.h"
17 #include "llvm/MC/MCAssembler.h"
18 #include "llvm/MC/MCContext.h"
19 #include "llvm/MC/MCExpr.h"
20 #include "llvm/MC/MCInst.h"
21 #include "llvm/MC/MCInstBuilder.h"
22 #include "llvm/MC/MCParser/MCAsmLexer.h"
23 #include "llvm/MC/MCParser/MCParsedAsmOperand.h"
24 #include "llvm/MC/MCParser/MCTargetAsmParser.h"
25 #include "llvm/MC/MCRegisterInfo.h"
26 #include "llvm/MC/MCStreamer.h"
27 #include "llvm/MC/MCSubtargetInfo.h"
28 #include "llvm/Support/Casting.h"
29 #include "llvm/Support/MathExtras.h"
30 #include "llvm/Support/TargetRegistry.h"
31 
32 #include <limits>
33 
34 using namespace llvm;
35 
36 // Include the auto-generated portion of the compress emitter.
37 #define GEN_COMPRESS_INSTR
38 #include "RISCVGenCompressInstEmitter.inc"
39 
40 namespace {
41 struct RISCVOperand;
42 
43 class RISCVAsmParser : public MCTargetAsmParser {
44   SMLoc getLoc() const { return getParser().getTok().getLoc(); }
45   bool isRV64() const { return getSTI().hasFeature(RISCV::Feature64Bit); }
46 
47   RISCVTargetStreamer &getTargetStreamer() {
48     MCTargetStreamer &TS = *getParser().getStreamer().getTargetStreamer();
49     return static_cast<RISCVTargetStreamer &>(TS);
50   }
51 
52   unsigned validateTargetOperandClass(MCParsedAsmOperand &Op,
53                                       unsigned Kind) override;
54 
55   bool generateImmOutOfRangeError(OperandVector &Operands, uint64_t ErrorInfo,
56                                   int64_t Lower, int64_t Upper, Twine Msg);
57 
58   bool MatchAndEmitInstruction(SMLoc IDLoc, unsigned &Opcode,
59                                OperandVector &Operands, MCStreamer &Out,
60                                uint64_t &ErrorInfo,
61                                bool MatchingInlineAsm) override;
62 
63   bool ParseRegister(unsigned &RegNo, SMLoc &StartLoc, SMLoc &EndLoc) override;
64 
65   bool ParseInstruction(ParseInstructionInfo &Info, StringRef Name,
66                         SMLoc NameLoc, OperandVector &Operands) override;
67 
68   bool ParseDirective(AsmToken DirectiveID) override;
69 
70   // Helper to actually emit an instruction to the MCStreamer. Also, when
71   // possible, compression of the instruction is performed.
72   void emitToStreamer(MCStreamer &S, const MCInst &Inst);
73 
74   // Helper to emit a combination of LUI, ADDI(W), and SLLI instructions that
75   // synthesize the desired immedate value into the destination register.
76   void emitLoadImm(unsigned DestReg, int64_t Value, MCStreamer &Out);
77 
78   // Helper to emit pseudo instruction "lla" used in PC-rel addressing.
79   void emitLoadLocalAddress(MCInst &Inst, SMLoc IDLoc, MCStreamer &Out);
80 
81   /// Helper for processing MC instructions that have been successfully matched
82   /// by MatchAndEmitInstruction. Modifications to the emitted instructions,
83   /// like the expansion of pseudo instructions (e.g., "li"), can be performed
84   /// in this method.
85   bool processInstruction(MCInst &Inst, SMLoc IDLoc, MCStreamer &Out);
86 
87 // Auto-generated instruction matching functions
88 #define GET_ASSEMBLER_HEADER
89 #include "RISCVGenAsmMatcher.inc"
90 
91   OperandMatchResultTy parseCSRSystemRegister(OperandVector &Operands);
92   OperandMatchResultTy parseImmediate(OperandVector &Operands);
93   OperandMatchResultTy parseRegister(OperandVector &Operands,
94                                      bool AllowParens = false);
95   OperandMatchResultTy parseMemOpBaseReg(OperandVector &Operands);
96   OperandMatchResultTy parseOperandWithModifier(OperandVector &Operands);
97   OperandMatchResultTy parseBareSymbol(OperandVector &Operands);
98   OperandMatchResultTy parseJALOffset(OperandVector &Operands);
99 
100   bool parseOperand(OperandVector &Operands, StringRef Mnemonic);
101 
102   bool parseDirectiveOption();
103 
104   void setFeatureBits(uint64_t Feature, StringRef FeatureString) {
105     if (!(getSTI().getFeatureBits()[Feature])) {
106       MCSubtargetInfo &STI = copySTI();
107       setAvailableFeatures(
108           ComputeAvailableFeatures(STI.ToggleFeature(FeatureString)));
109     }
110   }
111 
112   void clearFeatureBits(uint64_t Feature, StringRef FeatureString) {
113     if (getSTI().getFeatureBits()[Feature]) {
114       MCSubtargetInfo &STI = copySTI();
115       setAvailableFeatures(
116           ComputeAvailableFeatures(STI.ToggleFeature(FeatureString)));
117     }
118   }
119 
120 public:
121   enum RISCVMatchResultTy {
122     Match_Dummy = FIRST_TARGET_MATCH_RESULT_TY,
123 #define GET_OPERAND_DIAGNOSTIC_TYPES
124 #include "RISCVGenAsmMatcher.inc"
125 #undef GET_OPERAND_DIAGNOSTIC_TYPES
126   };
127 
128   static bool classifySymbolRef(const MCExpr *Expr,
129                                 RISCVMCExpr::VariantKind &Kind,
130                                 int64_t &Addend);
131 
132   RISCVAsmParser(const MCSubtargetInfo &STI, MCAsmParser &Parser,
133                  const MCInstrInfo &MII, const MCTargetOptions &Options)
134       : MCTargetAsmParser(Options, STI, MII) {
135     Parser.addAliasForDirective(".half", ".2byte");
136     Parser.addAliasForDirective(".hword", ".2byte");
137     Parser.addAliasForDirective(".word", ".4byte");
138     Parser.addAliasForDirective(".dword", ".8byte");
139     setAvailableFeatures(ComputeAvailableFeatures(STI.getFeatureBits()));
140   }
141 };
142 
143 /// RISCVOperand - Instances of this class represent a parsed machine
144 /// instruction
145 struct RISCVOperand : public MCParsedAsmOperand {
146 
147   enum KindTy {
148     Token,
149     Register,
150     Immediate,
151     SystemRegister
152   } Kind;
153 
154   bool IsRV64;
155 
156   struct RegOp {
157     unsigned RegNum;
158   };
159 
160   struct ImmOp {
161     const MCExpr *Val;
162   };
163 
164   struct SysRegOp {
165     const char *Data;
166     unsigned Length;
167     unsigned Encoding;
168     // FIXME: Add the Encoding parsed fields as needed for checks,
169     // e.g.: read/write or user/supervisor/machine privileges.
170   };
171 
172   SMLoc StartLoc, EndLoc;
173   union {
174     StringRef Tok;
175     RegOp Reg;
176     ImmOp Imm;
177     struct SysRegOp SysReg;
178   };
179 
180   RISCVOperand(KindTy K) : MCParsedAsmOperand(), Kind(K) {}
181 
182 public:
183   RISCVOperand(const RISCVOperand &o) : MCParsedAsmOperand() {
184     Kind = o.Kind;
185     IsRV64 = o.IsRV64;
186     StartLoc = o.StartLoc;
187     EndLoc = o.EndLoc;
188     switch (Kind) {
189     case Register:
190       Reg = o.Reg;
191       break;
192     case Immediate:
193       Imm = o.Imm;
194       break;
195     case Token:
196       Tok = o.Tok;
197       break;
198     case SystemRegister:
199       SysReg = o.SysReg;
200       break;
201     }
202   }
203 
204   bool isToken() const override { return Kind == Token; }
205   bool isReg() const override { return Kind == Register; }
206   bool isImm() const override { return Kind == Immediate; }
207   bool isMem() const override { return false; }
208   bool isSystemRegister() const { return Kind == SystemRegister; }
209 
210   static bool evaluateConstantImm(const MCExpr *Expr, int64_t &Imm,
211                                   RISCVMCExpr::VariantKind &VK) {
212     if (auto *RE = dyn_cast<RISCVMCExpr>(Expr)) {
213       VK = RE->getKind();
214       return RE->evaluateAsConstant(Imm);
215     }
216 
217     if (auto CE = dyn_cast<MCConstantExpr>(Expr)) {
218       VK = RISCVMCExpr::VK_RISCV_None;
219       Imm = CE->getValue();
220       return true;
221     }
222 
223     return false;
224   }
225 
226   // True if operand is a symbol with no modifiers, or a constant with no
227   // modifiers and isShiftedInt<N-1, 1>(Op).
228   template <int N> bool isBareSimmNLsb0() const {
229     int64_t Imm;
230     RISCVMCExpr::VariantKind VK;
231     if (!isImm())
232       return false;
233     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
234     bool IsValid;
235     if (!IsConstantImm)
236       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
237     else
238       IsValid = isShiftedInt<N - 1, 1>(Imm);
239     return IsValid && VK == RISCVMCExpr::VK_RISCV_None;
240   }
241 
242   // Predicate methods for AsmOperands defined in RISCVInstrInfo.td
243 
244   bool isBareSymbol() const {
245     int64_t Imm;
246     RISCVMCExpr::VariantKind VK;
247     // Must be of 'immediate' type but not a constant.
248     if (!isImm() || evaluateConstantImm(getImm(), Imm, VK))
249       return false;
250     return RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm) &&
251            VK == RISCVMCExpr::VK_RISCV_None;
252   }
253 
254   bool isCSRSystemRegister() const { return isSystemRegister(); }
255 
256   /// Return true if the operand is a valid for the fence instruction e.g.
257   /// ('iorw').
258   bool isFenceArg() const {
259     if (!isImm())
260       return false;
261     const MCExpr *Val = getImm();
262     auto *SVal = dyn_cast<MCSymbolRefExpr>(Val);
263     if (!SVal || SVal->getKind() != MCSymbolRefExpr::VK_None)
264       return false;
265 
266     StringRef Str = SVal->getSymbol().getName();
267     // Letters must be unique, taken from 'iorw', and in ascending order. This
268     // holds as long as each individual character is one of 'iorw' and is
269     // greater than the previous character.
270     char Prev = '\0';
271     for (char c : Str) {
272       if (c != 'i' && c != 'o' && c != 'r' && c != 'w')
273         return false;
274       if (c <= Prev)
275         return false;
276       Prev = c;
277     }
278     return true;
279   }
280 
281   /// Return true if the operand is a valid floating point rounding mode.
282   bool isFRMArg() const {
283     if (!isImm())
284       return false;
285     const MCExpr *Val = getImm();
286     auto *SVal = dyn_cast<MCSymbolRefExpr>(Val);
287     if (!SVal || SVal->getKind() != MCSymbolRefExpr::VK_None)
288       return false;
289 
290     StringRef Str = SVal->getSymbol().getName();
291 
292     return RISCVFPRndMode::stringToRoundingMode(Str) != RISCVFPRndMode::Invalid;
293   }
294 
295   bool isImmXLen() const {
296     int64_t Imm;
297     RISCVMCExpr::VariantKind VK;
298     if (!isImm())
299       return false;
300     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
301     // Given only Imm, ensuring that the actually specified constant is either
302     // a signed or unsigned 64-bit number is unfortunately impossible.
303     bool IsInRange = isRV64() ? true : isInt<32>(Imm) || isUInt<32>(Imm);
304     return IsConstantImm && IsInRange && VK == RISCVMCExpr::VK_RISCV_None;
305   }
306 
307   bool isUImmLog2XLen() const {
308     int64_t Imm;
309     RISCVMCExpr::VariantKind VK;
310     if (!isImm())
311       return false;
312     if (!evaluateConstantImm(getImm(), Imm, VK) ||
313         VK != RISCVMCExpr::VK_RISCV_None)
314       return false;
315     return (isRV64() && isUInt<6>(Imm)) || isUInt<5>(Imm);
316   }
317 
318   bool isUImmLog2XLenNonZero() const {
319     int64_t Imm;
320     RISCVMCExpr::VariantKind VK;
321     if (!isImm())
322       return false;
323     if (!evaluateConstantImm(getImm(), Imm, VK) ||
324         VK != RISCVMCExpr::VK_RISCV_None)
325       return false;
326     if (Imm == 0)
327       return false;
328     return (isRV64() && isUInt<6>(Imm)) || isUInt<5>(Imm);
329   }
330 
331   bool isUImm5() const {
332     int64_t Imm;
333     RISCVMCExpr::VariantKind VK;
334     if (!isImm())
335       return false;
336     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
337     return IsConstantImm && isUInt<5>(Imm) && VK == RISCVMCExpr::VK_RISCV_None;
338   }
339 
340   bool isUImm5NonZero() const {
341     int64_t Imm;
342     RISCVMCExpr::VariantKind VK;
343     if (!isImm())
344       return false;
345     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
346     return IsConstantImm && isUInt<5>(Imm) && (Imm != 0) &&
347            VK == RISCVMCExpr::VK_RISCV_None;
348   }
349 
350   bool isSImm6() const {
351     if (!isImm())
352       return false;
353     RISCVMCExpr::VariantKind VK;
354     int64_t Imm;
355     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
356     return IsConstantImm && isInt<6>(Imm) &&
357            VK == RISCVMCExpr::VK_RISCV_None;
358   }
359 
360   bool isSImm6NonZero() const {
361     if (!isImm())
362       return false;
363     RISCVMCExpr::VariantKind VK;
364     int64_t Imm;
365     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
366     return IsConstantImm && isInt<6>(Imm) && (Imm != 0) &&
367            VK == RISCVMCExpr::VK_RISCV_None;
368   }
369 
370   bool isCLUIImm() const {
371     if (!isImm())
372       return false;
373     int64_t Imm;
374     RISCVMCExpr::VariantKind VK;
375     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
376     return IsConstantImm && (Imm != 0) &&
377            (isUInt<5>(Imm) || (Imm >= 0xfffe0 && Imm <= 0xfffff)) &&
378            VK == RISCVMCExpr::VK_RISCV_None;
379   }
380 
381   bool isUImm7Lsb00() const {
382     if (!isImm())
383       return false;
384     int64_t Imm;
385     RISCVMCExpr::VariantKind VK;
386     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
387     return IsConstantImm && isShiftedUInt<5, 2>(Imm) &&
388            VK == RISCVMCExpr::VK_RISCV_None;
389   }
390 
391   bool isUImm8Lsb00() const {
392     if (!isImm())
393       return false;
394     int64_t Imm;
395     RISCVMCExpr::VariantKind VK;
396     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
397     return IsConstantImm && isShiftedUInt<6, 2>(Imm) &&
398            VK == RISCVMCExpr::VK_RISCV_None;
399   }
400 
401   bool isUImm8Lsb000() const {
402     if (!isImm())
403       return false;
404     int64_t Imm;
405     RISCVMCExpr::VariantKind VK;
406     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
407     return IsConstantImm && isShiftedUInt<5, 3>(Imm) &&
408            VK == RISCVMCExpr::VK_RISCV_None;
409   }
410 
411   bool isSImm9Lsb0() const { return isBareSimmNLsb0<9>(); }
412 
413   bool isUImm9Lsb000() const {
414     if (!isImm())
415       return false;
416     int64_t Imm;
417     RISCVMCExpr::VariantKind VK;
418     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
419     return IsConstantImm && isShiftedUInt<6, 3>(Imm) &&
420            VK == RISCVMCExpr::VK_RISCV_None;
421   }
422 
423   bool isUImm10Lsb00NonZero() const {
424     if (!isImm())
425       return false;
426     int64_t Imm;
427     RISCVMCExpr::VariantKind VK;
428     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
429     return IsConstantImm && isShiftedUInt<8, 2>(Imm) && (Imm != 0) &&
430            VK == RISCVMCExpr::VK_RISCV_None;
431   }
432 
433   bool isSImm12() const {
434     RISCVMCExpr::VariantKind VK;
435     int64_t Imm;
436     bool IsValid;
437     if (!isImm())
438       return false;
439     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
440     if (!IsConstantImm)
441       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
442     else
443       IsValid = isInt<12>(Imm);
444     return IsValid && ((IsConstantImm && VK == RISCVMCExpr::VK_RISCV_None) ||
445                        VK == RISCVMCExpr::VK_RISCV_LO ||
446                        VK == RISCVMCExpr::VK_RISCV_PCREL_LO);
447   }
448 
449   bool isSImm12Lsb0() const { return isBareSimmNLsb0<12>(); }
450 
451   bool isSImm13Lsb0() const { return isBareSimmNLsb0<13>(); }
452 
453   bool isSImm10Lsb0000NonZero() const {
454     if (!isImm())
455       return false;
456     int64_t Imm;
457     RISCVMCExpr::VariantKind VK;
458     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
459     return IsConstantImm && (Imm != 0) && isShiftedInt<6, 4>(Imm) &&
460            VK == RISCVMCExpr::VK_RISCV_None;
461   }
462 
463   bool isUImm20LUI() const {
464     RISCVMCExpr::VariantKind VK;
465     int64_t Imm;
466     bool IsValid;
467     if (!isImm())
468       return false;
469     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
470     if (!IsConstantImm) {
471       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
472       return IsValid && VK == RISCVMCExpr::VK_RISCV_HI;
473     } else {
474       return isUInt<20>(Imm) && (VK == RISCVMCExpr::VK_RISCV_None ||
475                                  VK == RISCVMCExpr::VK_RISCV_HI);
476     }
477   }
478 
479   bool isUImm20AUIPC() const {
480     RISCVMCExpr::VariantKind VK;
481     int64_t Imm;
482     bool IsValid;
483     if (!isImm())
484       return false;
485     bool IsConstantImm = evaluateConstantImm(getImm(), Imm, VK);
486     if (!IsConstantImm) {
487       IsValid = RISCVAsmParser::classifySymbolRef(getImm(), VK, Imm);
488       return IsValid && VK == RISCVMCExpr::VK_RISCV_PCREL_HI;
489     } else {
490       return isUInt<20>(Imm) && (VK == RISCVMCExpr::VK_RISCV_None ||
491                                  VK == RISCVMCExpr::VK_RISCV_PCREL_HI);
492     }
493   }
494 
495   bool isSImm21Lsb0JAL() const { return isBareSimmNLsb0<21>(); }
496 
497   /// getStartLoc - Gets location of the first token of this operand
498   SMLoc getStartLoc() const override { return StartLoc; }
499   /// getEndLoc - Gets location of the last token of this operand
500   SMLoc getEndLoc() const override { return EndLoc; }
501   /// True if this operand is for an RV64 instruction
502   bool isRV64() const { return IsRV64; }
503 
504   unsigned getReg() const override {
505     assert(Kind == Register && "Invalid type access!");
506     return Reg.RegNum;
507   }
508 
509   StringRef getSysReg() const {
510     assert(Kind == SystemRegister && "Invalid access!");
511     return StringRef(SysReg.Data, SysReg.Length);
512   }
513 
514   const MCExpr *getImm() const {
515     assert(Kind == Immediate && "Invalid type access!");
516     return Imm.Val;
517   }
518 
519   StringRef getToken() const {
520     assert(Kind == Token && "Invalid type access!");
521     return Tok;
522   }
523 
524   void print(raw_ostream &OS) const override {
525     switch (Kind) {
526     case Immediate:
527       OS << *getImm();
528       break;
529     case Register:
530       OS << "<register x";
531       OS << getReg() << ">";
532       break;
533     case Token:
534       OS << "'" << getToken() << "'";
535       break;
536     case SystemRegister:
537       OS << "<sysreg: " << getSysReg() << '>';
538       break;
539     }
540   }
541 
542   static std::unique_ptr<RISCVOperand> createToken(StringRef Str, SMLoc S,
543                                                    bool IsRV64) {
544     auto Op = make_unique<RISCVOperand>(Token);
545     Op->Tok = Str;
546     Op->StartLoc = S;
547     Op->EndLoc = S;
548     Op->IsRV64 = IsRV64;
549     return Op;
550   }
551 
552   static std::unique_ptr<RISCVOperand> createReg(unsigned RegNo, SMLoc S,
553                                                  SMLoc E, bool IsRV64) {
554     auto Op = make_unique<RISCVOperand>(Register);
555     Op->Reg.RegNum = RegNo;
556     Op->StartLoc = S;
557     Op->EndLoc = E;
558     Op->IsRV64 = IsRV64;
559     return Op;
560   }
561 
562   static std::unique_ptr<RISCVOperand> createImm(const MCExpr *Val, SMLoc S,
563                                                  SMLoc E, bool IsRV64) {
564     auto Op = make_unique<RISCVOperand>(Immediate);
565     Op->Imm.Val = Val;
566     Op->StartLoc = S;
567     Op->EndLoc = E;
568     Op->IsRV64 = IsRV64;
569     return Op;
570   }
571 
572   static std::unique_ptr<RISCVOperand>
573   createSysReg(StringRef Str, SMLoc S, unsigned Encoding, bool IsRV64) {
574     auto Op = make_unique<RISCVOperand>(SystemRegister);
575     Op->SysReg.Data = Str.data();
576     Op->SysReg.Length = Str.size();
577     Op->SysReg.Encoding = Encoding;
578     Op->StartLoc = S;
579     Op->IsRV64 = IsRV64;
580     return Op;
581   }
582 
583   void addExpr(MCInst &Inst, const MCExpr *Expr) const {
584     assert(Expr && "Expr shouldn't be null!");
585     int64_t Imm = 0;
586     RISCVMCExpr::VariantKind VK;
587     bool IsConstant = evaluateConstantImm(Expr, Imm, VK);
588 
589     if (IsConstant)
590       Inst.addOperand(MCOperand::createImm(Imm));
591     else
592       Inst.addOperand(MCOperand::createExpr(Expr));
593   }
594 
595   // Used by the TableGen Code
596   void addRegOperands(MCInst &Inst, unsigned N) const {
597     assert(N == 1 && "Invalid number of operands!");
598     Inst.addOperand(MCOperand::createReg(getReg()));
599   }
600 
601   void addImmOperands(MCInst &Inst, unsigned N) const {
602     assert(N == 1 && "Invalid number of operands!");
603     addExpr(Inst, getImm());
604   }
605 
606   void addFenceArgOperands(MCInst &Inst, unsigned N) const {
607     assert(N == 1 && "Invalid number of operands!");
608     // isFenceArg has validated the operand, meaning this cast is safe
609     auto SE = cast<MCSymbolRefExpr>(getImm());
610 
611     unsigned Imm = 0;
612     for (char c : SE->getSymbol().getName()) {
613       switch (c) {
614       default:
615         llvm_unreachable("FenceArg must contain only [iorw]");
616       case 'i': Imm |= RISCVFenceField::I; break;
617       case 'o': Imm |= RISCVFenceField::O; break;
618       case 'r': Imm |= RISCVFenceField::R; break;
619       case 'w': Imm |= RISCVFenceField::W; break;
620       }
621     }
622     Inst.addOperand(MCOperand::createImm(Imm));
623   }
624 
625   void addCSRSystemRegisterOperands(MCInst &Inst, unsigned N) const {
626     assert(N == 1 && "Invalid number of operands!");
627     Inst.addOperand(MCOperand::createImm(SysReg.Encoding));
628   }
629 
630   // Returns the rounding mode represented by this RISCVOperand. Should only
631   // be called after checking isFRMArg.
632   RISCVFPRndMode::RoundingMode getRoundingMode() const {
633     // isFRMArg has validated the operand, meaning this cast is safe.
634     auto SE = cast<MCSymbolRefExpr>(getImm());
635     RISCVFPRndMode::RoundingMode FRM =
636         RISCVFPRndMode::stringToRoundingMode(SE->getSymbol().getName());
637     assert(FRM != RISCVFPRndMode::Invalid && "Invalid rounding mode");
638     return FRM;
639   }
640 
641   void addFRMArgOperands(MCInst &Inst, unsigned N) const {
642     assert(N == 1 && "Invalid number of operands!");
643     Inst.addOperand(MCOperand::createImm(getRoundingMode()));
644   }
645 };
646 } // end anonymous namespace.
647 
648 #define GET_REGISTER_MATCHER
649 #define GET_MATCHER_IMPLEMENTATION
650 #include "RISCVGenAsmMatcher.inc"
651 
652 // Return the matching FPR64 register for the given FPR32.
653 // FIXME: Ideally this function could be removed in favour of using
654 // information from TableGen.
655 unsigned convertFPR32ToFPR64(unsigned Reg) {
656   switch (Reg) {
657   default:
658     llvm_unreachable("Not a recognised FPR32 register");
659   case RISCV::F0_32: return RISCV::F0_64;
660   case RISCV::F1_32: return RISCV::F1_64;
661   case RISCV::F2_32: return RISCV::F2_64;
662   case RISCV::F3_32: return RISCV::F3_64;
663   case RISCV::F4_32: return RISCV::F4_64;
664   case RISCV::F5_32: return RISCV::F5_64;
665   case RISCV::F6_32: return RISCV::F6_64;
666   case RISCV::F7_32: return RISCV::F7_64;
667   case RISCV::F8_32: return RISCV::F8_64;
668   case RISCV::F9_32: return RISCV::F9_64;
669   case RISCV::F10_32: return RISCV::F10_64;
670   case RISCV::F11_32: return RISCV::F11_64;
671   case RISCV::F12_32: return RISCV::F12_64;
672   case RISCV::F13_32: return RISCV::F13_64;
673   case RISCV::F14_32: return RISCV::F14_64;
674   case RISCV::F15_32: return RISCV::F15_64;
675   case RISCV::F16_32: return RISCV::F16_64;
676   case RISCV::F17_32: return RISCV::F17_64;
677   case RISCV::F18_32: return RISCV::F18_64;
678   case RISCV::F19_32: return RISCV::F19_64;
679   case RISCV::F20_32: return RISCV::F20_64;
680   case RISCV::F21_32: return RISCV::F21_64;
681   case RISCV::F22_32: return RISCV::F22_64;
682   case RISCV::F23_32: return RISCV::F23_64;
683   case RISCV::F24_32: return RISCV::F24_64;
684   case RISCV::F25_32: return RISCV::F25_64;
685   case RISCV::F26_32: return RISCV::F26_64;
686   case RISCV::F27_32: return RISCV::F27_64;
687   case RISCV::F28_32: return RISCV::F28_64;
688   case RISCV::F29_32: return RISCV::F29_64;
689   case RISCV::F30_32: return RISCV::F30_64;
690   case RISCV::F31_32: return RISCV::F31_64;
691   }
692 }
693 
694 unsigned RISCVAsmParser::validateTargetOperandClass(MCParsedAsmOperand &AsmOp,
695                                                     unsigned Kind) {
696   RISCVOperand &Op = static_cast<RISCVOperand &>(AsmOp);
697   if (!Op.isReg())
698     return Match_InvalidOperand;
699 
700   unsigned Reg = Op.getReg();
701   bool IsRegFPR32 =
702       RISCVMCRegisterClasses[RISCV::FPR32RegClassID].contains(Reg);
703   bool IsRegFPR32C =
704       RISCVMCRegisterClasses[RISCV::FPR32CRegClassID].contains(Reg);
705 
706   // As the parser couldn't differentiate an FPR32 from an FPR64, coerce the
707   // register from FPR32 to FPR64 or FPR32C to FPR64C if necessary.
708   if ((IsRegFPR32 && Kind == MCK_FPR64) ||
709       (IsRegFPR32C && Kind == MCK_FPR64C)) {
710     Op.Reg.RegNum = convertFPR32ToFPR64(Reg);
711     return Match_Success;
712   }
713   return Match_InvalidOperand;
714 }
715 
716 bool RISCVAsmParser::generateImmOutOfRangeError(
717     OperandVector &Operands, uint64_t ErrorInfo, int64_t Lower, int64_t Upper,
718     Twine Msg = "immediate must be an integer in the range") {
719   SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
720   return Error(ErrorLoc, Msg + " [" + Twine(Lower) + ", " + Twine(Upper) + "]");
721 }
722 
723 bool RISCVAsmParser::MatchAndEmitInstruction(SMLoc IDLoc, unsigned &Opcode,
724                                              OperandVector &Operands,
725                                              MCStreamer &Out,
726                                              uint64_t &ErrorInfo,
727                                              bool MatchingInlineAsm) {
728   MCInst Inst;
729 
730   auto Result =
731     MatchInstructionImpl(Operands, Inst, ErrorInfo, MatchingInlineAsm);
732   switch (Result) {
733   default:
734     break;
735   case Match_Success:
736     return processInstruction(Inst, IDLoc, Out);
737   case Match_MissingFeature:
738     return Error(IDLoc, "instruction use requires an option to be enabled");
739   case Match_MnemonicFail:
740     return Error(IDLoc, "unrecognized instruction mnemonic");
741   case Match_InvalidOperand: {
742     SMLoc ErrorLoc = IDLoc;
743     if (ErrorInfo != ~0U) {
744       if (ErrorInfo >= Operands.size())
745         return Error(ErrorLoc, "too few operands for instruction");
746 
747       ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
748       if (ErrorLoc == SMLoc())
749         ErrorLoc = IDLoc;
750     }
751     return Error(ErrorLoc, "invalid operand for instruction");
752   }
753   }
754 
755   // Handle the case when the error message is of specific type
756   // other than the generic Match_InvalidOperand, and the
757   // corresponding operand is missing.
758   if (Result > FIRST_TARGET_MATCH_RESULT_TY) {
759     SMLoc ErrorLoc = IDLoc;
760     if (ErrorInfo != ~0U && ErrorInfo >= Operands.size())
761         return Error(ErrorLoc, "too few operands for instruction");
762   }
763 
764   switch(Result) {
765   default:
766     break;
767   case Match_InvalidImmXLen:
768     if (isRV64()) {
769       SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
770       return Error(ErrorLoc, "operand must be a constant 64-bit integer");
771     }
772     return generateImmOutOfRangeError(Operands, ErrorInfo,
773                                       std::numeric_limits<int32_t>::min(),
774                                       std::numeric_limits<uint32_t>::max());
775   case Match_InvalidUImmLog2XLen:
776     if (isRV64())
777       return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 6) - 1);
778     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 5) - 1);
779   case Match_InvalidUImmLog2XLenNonZero:
780     if (isRV64())
781       return generateImmOutOfRangeError(Operands, ErrorInfo, 1, (1 << 6) - 1);
782     return generateImmOutOfRangeError(Operands, ErrorInfo, 1, (1 << 5) - 1);
783   case Match_InvalidUImm5:
784     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 5) - 1);
785   case Match_InvalidSImm6:
786     return generateImmOutOfRangeError(Operands, ErrorInfo, -(1 << 5),
787                                       (1 << 5) - 1);
788   case Match_InvalidSImm6NonZero:
789     return generateImmOutOfRangeError(
790         Operands, ErrorInfo, -(1 << 5), (1 << 5) - 1,
791         "immediate must be non-zero in the range");
792   case Match_InvalidCLUIImm:
793     return generateImmOutOfRangeError(
794         Operands, ErrorInfo, 1, (1 << 5) - 1,
795         "immediate must be in [0xfffe0, 0xfffff] or");
796   case Match_InvalidUImm7Lsb00:
797     return generateImmOutOfRangeError(
798         Operands, ErrorInfo, 0, (1 << 7) - 4,
799         "immediate must be a multiple of 4 bytes in the range");
800   case Match_InvalidUImm8Lsb00:
801     return generateImmOutOfRangeError(
802         Operands, ErrorInfo, 0, (1 << 8) - 4,
803         "immediate must be a multiple of 4 bytes in the range");
804   case Match_InvalidUImm8Lsb000:
805     return generateImmOutOfRangeError(
806         Operands, ErrorInfo, 0, (1 << 8) - 8,
807         "immediate must be a multiple of 8 bytes in the range");
808   case Match_InvalidSImm9Lsb0:
809     return generateImmOutOfRangeError(
810         Operands, ErrorInfo, -(1 << 8), (1 << 8) - 2,
811         "immediate must be a multiple of 2 bytes in the range");
812   case Match_InvalidUImm9Lsb000:
813     return generateImmOutOfRangeError(
814         Operands, ErrorInfo, 0, (1 << 9) - 8,
815         "immediate must be a multiple of 8 bytes in the range");
816   case Match_InvalidUImm10Lsb00NonZero:
817     return generateImmOutOfRangeError(
818         Operands, ErrorInfo, 4, (1 << 10) - 4,
819         "immediate must be a multiple of 4 bytes in the range");
820   case Match_InvalidSImm10Lsb0000NonZero:
821     return generateImmOutOfRangeError(
822         Operands, ErrorInfo, -(1 << 9), (1 << 9) - 16,
823         "immediate must be a multiple of 16 bytes and non-zero in the range");
824   case Match_InvalidSImm12:
825     return generateImmOutOfRangeError(
826         Operands, ErrorInfo, -(1 << 11), (1 << 11) - 1,
827         "operand must be a symbol with %lo/%pcrel_lo modifier or an integer in "
828         "the range");
829   case Match_InvalidSImm12Lsb0:
830     return generateImmOutOfRangeError(
831         Operands, ErrorInfo, -(1 << 11), (1 << 11) - 2,
832         "immediate must be a multiple of 2 bytes in the range");
833   case Match_InvalidSImm13Lsb0:
834     return generateImmOutOfRangeError(
835         Operands, ErrorInfo, -(1 << 12), (1 << 12) - 2,
836         "immediate must be a multiple of 2 bytes in the range");
837   case Match_InvalidUImm20LUI:
838     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 20) - 1,
839                                       "operand must be a symbol with %hi() "
840                                       "modifier or an integer in the range");
841   case Match_InvalidUImm20AUIPC:
842     return generateImmOutOfRangeError(
843         Operands, ErrorInfo, 0, (1 << 20) - 1,
844         "operand must be a symbol with %pcrel_hi() modifier or an integer in "
845         "the range");
846   case Match_InvalidSImm21Lsb0JAL:
847     return generateImmOutOfRangeError(
848         Operands, ErrorInfo, -(1 << 20), (1 << 20) - 2,
849         "immediate must be a multiple of 2 bytes in the range");
850   case Match_InvalidCSRSystemRegister: {
851     return generateImmOutOfRangeError(Operands, ErrorInfo, 0, (1 << 12) - 1,
852                                       "operand must be a valid system register "
853                                       "name or an integer in the range");
854   }
855   case Match_InvalidFenceArg: {
856     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
857     return Error(
858         ErrorLoc,
859         "operand must be formed of letters selected in-order from 'iorw'");
860   }
861   case Match_InvalidFRMArg: {
862     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
863     return Error(
864         ErrorLoc,
865         "operand must be a valid floating point rounding mode mnemonic");
866   }
867   case Match_InvalidBareSymbol: {
868     SMLoc ErrorLoc = ((RISCVOperand &)*Operands[ErrorInfo]).getStartLoc();
869     return Error(ErrorLoc, "operand must be a bare symbol name");
870   }
871   }
872 
873   llvm_unreachable("Unknown match type detected!");
874 }
875 
876 bool RISCVAsmParser::ParseRegister(unsigned &RegNo, SMLoc &StartLoc,
877                                    SMLoc &EndLoc) {
878   const AsmToken &Tok = getParser().getTok();
879   StartLoc = Tok.getLoc();
880   EndLoc = Tok.getEndLoc();
881   RegNo = 0;
882   StringRef Name = getLexer().getTok().getIdentifier();
883 
884   if (!MatchRegisterName(Name) || !MatchRegisterAltName(Name)) {
885     getParser().Lex(); // Eat identifier token.
886     return false;
887   }
888 
889   return Error(StartLoc, "invalid register name");
890 }
891 
892 OperandMatchResultTy RISCVAsmParser::parseRegister(OperandVector &Operands,
893                                                    bool AllowParens) {
894   SMLoc FirstS = getLoc();
895   bool HadParens = false;
896   AsmToken Buf[2];
897 
898   // If this a parenthesised register name is allowed, parse it atomically
899   if (AllowParens && getLexer().is(AsmToken::LParen)) {
900     size_t ReadCount = getLexer().peekTokens(Buf);
901     if (ReadCount == 2 && Buf[1].getKind() == AsmToken::RParen) {
902       HadParens = true;
903       getParser().Lex(); // Eat '('
904     }
905   }
906 
907   switch (getLexer().getKind()) {
908   default:
909     return MatchOperand_NoMatch;
910   case AsmToken::Identifier:
911     StringRef Name = getLexer().getTok().getIdentifier();
912     unsigned RegNo = MatchRegisterName(Name);
913     if (RegNo == 0) {
914       RegNo = MatchRegisterAltName(Name);
915       if (RegNo == 0) {
916         if (HadParens)
917           getLexer().UnLex(Buf[0]);
918         return MatchOperand_NoMatch;
919       }
920     }
921     if (HadParens)
922       Operands.push_back(RISCVOperand::createToken("(", FirstS, isRV64()));
923     SMLoc S = getLoc();
924     SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
925     getLexer().Lex();
926     Operands.push_back(RISCVOperand::createReg(RegNo, S, E, isRV64()));
927   }
928 
929   if (HadParens) {
930     getParser().Lex(); // Eat ')'
931     Operands.push_back(RISCVOperand::createToken(")", getLoc(), isRV64()));
932   }
933 
934   return MatchOperand_Success;
935 }
936 
937 OperandMatchResultTy
938 RISCVAsmParser::parseCSRSystemRegister(OperandVector &Operands) {
939   SMLoc S = getLoc();
940   const MCExpr *Res;
941 
942   switch (getLexer().getKind()) {
943   default:
944     return MatchOperand_NoMatch;
945   case AsmToken::LParen:
946   case AsmToken::Minus:
947   case AsmToken::Plus:
948   case AsmToken::Integer:
949   case AsmToken::String: {
950     if (getParser().parseExpression(Res))
951       return MatchOperand_ParseFail;
952 
953     auto *CE = dyn_cast<MCConstantExpr>(Res);
954     if (CE) {
955       int64_t Imm = CE->getValue();
956       if (isUInt<12>(Imm)) {
957         auto SysReg = RISCVSysReg::lookupSysRegByEncoding(Imm);
958         // Accept an immediate representing a named or un-named Sys Reg
959         // if the range is valid, regardless of the required features.
960         Operands.push_back(RISCVOperand::createSysReg(
961             SysReg ? SysReg->Name : "", S, Imm, isRV64()));
962         return MatchOperand_Success;
963       }
964     }
965 
966     Twine Msg = "immediate must be an integer in the range";
967     Error(S, Msg + " [" + Twine(0) + ", " + Twine((1 << 12) - 1) + "]");
968     return MatchOperand_ParseFail;
969   }
970   case AsmToken::Identifier: {
971     StringRef Identifier;
972     if (getParser().parseIdentifier(Identifier))
973       return MatchOperand_ParseFail;
974 
975     auto SysReg = RISCVSysReg::lookupSysRegByName(Identifier);
976     // Accept a named Sys Reg if the required features are present.
977     if (SysReg) {
978       if (!SysReg->haveRequiredFeatures(getSTI().getFeatureBits())) {
979         Error(S, "system register use requires an option to be enabled");
980         return MatchOperand_ParseFail;
981       }
982       Operands.push_back(RISCVOperand::createSysReg(
983           Identifier, S, SysReg->Encoding, isRV64()));
984       return MatchOperand_Success;
985     }
986 
987     Twine Msg = "operand must be a valid system register name "
988                 "or an integer in the range";
989     Error(S, Msg + " [" + Twine(0) + ", " + Twine((1 << 12) - 1) + "]");
990     return MatchOperand_ParseFail;
991   }
992   case AsmToken::Percent: {
993     // Discard operand with modifier.
994     Twine Msg = "immediate must be an integer in the range";
995     Error(S, Msg + " [" + Twine(0) + ", " + Twine((1 << 12) - 1) + "]");
996     return MatchOperand_ParseFail;
997   }
998   }
999 
1000   return MatchOperand_NoMatch;
1001 }
1002 
1003 OperandMatchResultTy RISCVAsmParser::parseImmediate(OperandVector &Operands) {
1004   SMLoc S = getLoc();
1005   SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1006   const MCExpr *Res;
1007 
1008   switch (getLexer().getKind()) {
1009   default:
1010     return MatchOperand_NoMatch;
1011   case AsmToken::LParen:
1012   case AsmToken::Minus:
1013   case AsmToken::Plus:
1014   case AsmToken::Integer:
1015   case AsmToken::String:
1016     if (getParser().parseExpression(Res))
1017       return MatchOperand_ParseFail;
1018     break;
1019   case AsmToken::Identifier: {
1020     StringRef Identifier;
1021     if (getParser().parseIdentifier(Identifier))
1022       return MatchOperand_ParseFail;
1023     MCSymbol *Sym = getContext().getOrCreateSymbol(Identifier);
1024     Res = MCSymbolRefExpr::create(Sym, MCSymbolRefExpr::VK_None, getContext());
1025     break;
1026   }
1027   case AsmToken::Percent:
1028     return parseOperandWithModifier(Operands);
1029   }
1030 
1031   Operands.push_back(RISCVOperand::createImm(Res, S, E, isRV64()));
1032   return MatchOperand_Success;
1033 }
1034 
1035 OperandMatchResultTy
1036 RISCVAsmParser::parseOperandWithModifier(OperandVector &Operands) {
1037   SMLoc S = getLoc();
1038   SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1039 
1040   if (getLexer().getKind() != AsmToken::Percent) {
1041     Error(getLoc(), "expected '%' for operand modifier");
1042     return MatchOperand_ParseFail;
1043   }
1044 
1045   getParser().Lex(); // Eat '%'
1046 
1047   if (getLexer().getKind() != AsmToken::Identifier) {
1048     Error(getLoc(), "expected valid identifier for operand modifier");
1049     return MatchOperand_ParseFail;
1050   }
1051   StringRef Identifier = getParser().getTok().getIdentifier();
1052   RISCVMCExpr::VariantKind VK = RISCVMCExpr::getVariantKindForName(Identifier);
1053   if (VK == RISCVMCExpr::VK_RISCV_Invalid) {
1054     Error(getLoc(), "unrecognized operand modifier");
1055     return MatchOperand_ParseFail;
1056   }
1057 
1058   getParser().Lex(); // Eat the identifier
1059   if (getLexer().getKind() != AsmToken::LParen) {
1060     Error(getLoc(), "expected '('");
1061     return MatchOperand_ParseFail;
1062   }
1063   getParser().Lex(); // Eat '('
1064 
1065   const MCExpr *SubExpr;
1066   if (getParser().parseParenExpression(SubExpr, E)) {
1067     return MatchOperand_ParseFail;
1068   }
1069 
1070   const MCExpr *ModExpr = RISCVMCExpr::create(SubExpr, VK, getContext());
1071   Operands.push_back(RISCVOperand::createImm(ModExpr, S, E, isRV64()));
1072   return MatchOperand_Success;
1073 }
1074 
1075 OperandMatchResultTy RISCVAsmParser::parseBareSymbol(OperandVector &Operands) {
1076   SMLoc S = getLoc();
1077   SMLoc E = SMLoc::getFromPointer(S.getPointer() - 1);
1078   const MCExpr *Res;
1079 
1080   if (getLexer().getKind() != AsmToken::Identifier)
1081     return MatchOperand_NoMatch;
1082 
1083   StringRef Identifier;
1084   if (getParser().parseIdentifier(Identifier))
1085     return MatchOperand_ParseFail;
1086 
1087   MCSymbol *Sym = getContext().getOrCreateSymbol(Identifier);
1088   Res = MCSymbolRefExpr::create(Sym, MCSymbolRefExpr::VK_None, getContext());
1089   Operands.push_back(RISCVOperand::createImm(Res, S, E, isRV64()));
1090   return MatchOperand_Success;
1091 }
1092 
1093 OperandMatchResultTy RISCVAsmParser::parseJALOffset(OperandVector &Operands) {
1094   // Parsing jal operands is fiddly due to the `jal foo` and `jal ra, foo`
1095   // both being acceptable forms. When parsing `jal ra, foo` this function
1096   // will be called for the `ra` register operand in an attempt to match the
1097   // single-operand alias. parseJALOffset must fail for this case. It would
1098   // seem logical to try parse the operand using parseImmediate and return
1099   // NoMatch if the next token is a comma (meaning we must be parsing a jal in
1100   // the second form rather than the first). We can't do this as there's no
1101   // way of rewinding the lexer state. Instead, return NoMatch if this operand
1102   // is an identifier and is followed by a comma.
1103   if (getLexer().is(AsmToken::Identifier) &&
1104       getLexer().peekTok().is(AsmToken::Comma))
1105     return MatchOperand_NoMatch;
1106 
1107   return parseImmediate(Operands);
1108 }
1109 
1110 OperandMatchResultTy
1111 RISCVAsmParser::parseMemOpBaseReg(OperandVector &Operands) {
1112   if (getLexer().isNot(AsmToken::LParen)) {
1113     Error(getLoc(), "expected '('");
1114     return MatchOperand_ParseFail;
1115   }
1116 
1117   getParser().Lex(); // Eat '('
1118   Operands.push_back(RISCVOperand::createToken("(", getLoc(), isRV64()));
1119 
1120   if (parseRegister(Operands) != MatchOperand_Success) {
1121     Error(getLoc(), "expected register");
1122     return MatchOperand_ParseFail;
1123   }
1124 
1125   if (getLexer().isNot(AsmToken::RParen)) {
1126     Error(getLoc(), "expected ')'");
1127     return MatchOperand_ParseFail;
1128   }
1129 
1130   getParser().Lex(); // Eat ')'
1131   Operands.push_back(RISCVOperand::createToken(")", getLoc(), isRV64()));
1132 
1133   return MatchOperand_Success;
1134 }
1135 
1136 /// Looks at a token type and creates the relevant operand from this
1137 /// information, adding to Operands. If operand was parsed, returns false, else
1138 /// true.
1139 bool RISCVAsmParser::parseOperand(OperandVector &Operands, StringRef Mnemonic) {
1140   // Check if the current operand has a custom associated parser, if so, try to
1141   // custom parse the operand, or fallback to the general approach.
1142   OperandMatchResultTy Result =
1143       MatchOperandParserImpl(Operands, Mnemonic, /*ParseForAllFeatures=*/true);
1144   if (Result == MatchOperand_Success)
1145     return false;
1146   if (Result == MatchOperand_ParseFail)
1147     return true;
1148 
1149   // Attempt to parse token as a register.
1150   if (parseRegister(Operands, true) == MatchOperand_Success)
1151     return false;
1152 
1153   // Attempt to parse token as an immediate
1154   if (parseImmediate(Operands) == MatchOperand_Success) {
1155     // Parse memory base register if present
1156     if (getLexer().is(AsmToken::LParen))
1157       return parseMemOpBaseReg(Operands) != MatchOperand_Success;
1158     return false;
1159   }
1160 
1161   // Finally we have exhausted all options and must declare defeat.
1162   Error(getLoc(), "unknown operand");
1163   return true;
1164 }
1165 
1166 bool RISCVAsmParser::ParseInstruction(ParseInstructionInfo &Info,
1167                                       StringRef Name, SMLoc NameLoc,
1168                                       OperandVector &Operands) {
1169   // Ensure that if the instruction occurs when relaxation is enabled,
1170   // relocations are forced for the file. Ideally this would be done when there
1171   // is enough information to reliably determine if the instruction itself may
1172   // cause relaxations. Unfortunately instruction processing stage occurs in the
1173   // same pass as relocation emission, so it's too late to set a 'sticky bit'
1174   // for the entire file.
1175   if (getSTI().getFeatureBits()[RISCV::FeatureRelax]) {
1176     auto *Assembler = getTargetStreamer().getStreamer().getAssemblerPtr();
1177     if (Assembler != nullptr) {
1178       RISCVAsmBackend &MAB =
1179           static_cast<RISCVAsmBackend &>(Assembler->getBackend());
1180       MAB.setForceRelocs();
1181     }
1182   }
1183 
1184   // First operand is token for instruction
1185   Operands.push_back(RISCVOperand::createToken(Name, NameLoc, isRV64()));
1186 
1187   // If there are no more operands, then finish
1188   if (getLexer().is(AsmToken::EndOfStatement))
1189     return false;
1190 
1191   // Parse first operand
1192   if (parseOperand(Operands, Name))
1193     return true;
1194 
1195   // Parse until end of statement, consuming commas between operands
1196   unsigned OperandIdx = 1;
1197   while (getLexer().is(AsmToken::Comma)) {
1198     // Consume comma token
1199     getLexer().Lex();
1200 
1201     // Parse next operand
1202     if (parseOperand(Operands, Name))
1203       return true;
1204 
1205     ++OperandIdx;
1206   }
1207 
1208   if (getLexer().isNot(AsmToken::EndOfStatement)) {
1209     SMLoc Loc = getLexer().getLoc();
1210     getParser().eatToEndOfStatement();
1211     return Error(Loc, "unexpected token");
1212   }
1213 
1214   getParser().Lex(); // Consume the EndOfStatement.
1215   return false;
1216 }
1217 
1218 bool RISCVAsmParser::classifySymbolRef(const MCExpr *Expr,
1219                                        RISCVMCExpr::VariantKind &Kind,
1220                                        int64_t &Addend) {
1221   Kind = RISCVMCExpr::VK_RISCV_None;
1222   Addend = 0;
1223 
1224   if (const RISCVMCExpr *RE = dyn_cast<RISCVMCExpr>(Expr)) {
1225     Kind = RE->getKind();
1226     Expr = RE->getSubExpr();
1227   }
1228 
1229   // It's a simple symbol reference or constant with no addend.
1230   if (isa<MCConstantExpr>(Expr) || isa<MCSymbolRefExpr>(Expr))
1231     return true;
1232 
1233   const MCBinaryExpr *BE = dyn_cast<MCBinaryExpr>(Expr);
1234   if (!BE)
1235     return false;
1236 
1237   if (!isa<MCSymbolRefExpr>(BE->getLHS()))
1238     return false;
1239 
1240   if (BE->getOpcode() != MCBinaryExpr::Add &&
1241       BE->getOpcode() != MCBinaryExpr::Sub)
1242     return false;
1243 
1244   // We are able to support the subtraction of two symbol references
1245   if (BE->getOpcode() == MCBinaryExpr::Sub &&
1246       isa<MCSymbolRefExpr>(BE->getRHS()))
1247     return true;
1248 
1249   // See if the addend is a constant, otherwise there's more going
1250   // on here than we can deal with.
1251   auto AddendExpr = dyn_cast<MCConstantExpr>(BE->getRHS());
1252   if (!AddendExpr)
1253     return false;
1254 
1255   Addend = AddendExpr->getValue();
1256   if (BE->getOpcode() == MCBinaryExpr::Sub)
1257     Addend = -Addend;
1258 
1259   // It's some symbol reference + a constant addend
1260   return Kind != RISCVMCExpr::VK_RISCV_Invalid;
1261 }
1262 
1263 bool RISCVAsmParser::ParseDirective(AsmToken DirectiveID) {
1264   // This returns false if this function recognizes the directive
1265   // regardless of whether it is successfully handles or reports an
1266   // error. Otherwise it returns true to give the generic parser a
1267   // chance at recognizing it.
1268   StringRef IDVal = DirectiveID.getString();
1269 
1270   if (IDVal == ".option")
1271     return parseDirectiveOption();
1272 
1273   return true;
1274 }
1275 
1276 bool RISCVAsmParser::parseDirectiveOption() {
1277   MCAsmParser &Parser = getParser();
1278   // Get the option token.
1279   AsmToken Tok = Parser.getTok();
1280   // At the moment only identifiers are supported.
1281   if (Tok.isNot(AsmToken::Identifier))
1282     return Error(Parser.getTok().getLoc(),
1283                  "unexpected token, expected identifier");
1284 
1285   StringRef Option = Tok.getIdentifier();
1286 
1287   if (Option == "rvc") {
1288     getTargetStreamer().emitDirectiveOptionRVC();
1289 
1290     Parser.Lex();
1291     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1292       return Error(Parser.getTok().getLoc(),
1293                    "unexpected token, expected end of statement");
1294 
1295     setFeatureBits(RISCV::FeatureStdExtC, "c");
1296     return false;
1297   }
1298 
1299   if (Option == "norvc") {
1300     getTargetStreamer().emitDirectiveOptionNoRVC();
1301 
1302     Parser.Lex();
1303     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1304       return Error(Parser.getTok().getLoc(),
1305                    "unexpected token, expected end of statement");
1306 
1307     clearFeatureBits(RISCV::FeatureStdExtC, "c");
1308     return false;
1309   }
1310 
1311   if (Option == "relax") {
1312     getTargetStreamer().emitDirectiveOptionRelax();
1313 
1314     Parser.Lex();
1315     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1316       return Error(Parser.getTok().getLoc(),
1317                    "unexpected token, expected end of statement");
1318 
1319     setFeatureBits(RISCV::FeatureRelax, "relax");
1320     return false;
1321   }
1322 
1323   if (Option == "norelax") {
1324     getTargetStreamer().emitDirectiveOptionNoRelax();
1325 
1326     Parser.Lex();
1327     if (Parser.getTok().isNot(AsmToken::EndOfStatement))
1328       return Error(Parser.getTok().getLoc(),
1329                    "unexpected token, expected end of statement");
1330 
1331     clearFeatureBits(RISCV::FeatureRelax, "relax");
1332     return false;
1333   }
1334 
1335   // Unknown option.
1336   Warning(Parser.getTok().getLoc(),
1337           "unknown option, expected 'rvc', 'norvc', 'relax' or 'norelax'");
1338   Parser.eatToEndOfStatement();
1339   return false;
1340 }
1341 
1342 void RISCVAsmParser::emitToStreamer(MCStreamer &S, const MCInst &Inst) {
1343   MCInst CInst;
1344   bool Res = compressInst(CInst, Inst, getSTI(), S.getContext());
1345   CInst.setLoc(Inst.getLoc());
1346   S.EmitInstruction((Res ? CInst : Inst), getSTI());
1347 }
1348 
1349 void RISCVAsmParser::emitLoadImm(unsigned DestReg, int64_t Value,
1350                                  MCStreamer &Out) {
1351   if (isInt<32>(Value)) {
1352     // Emits the MC instructions for loading a 32-bit constant into a register.
1353     //
1354     // Depending on the active bits in the immediate Value v, the following
1355     // instruction sequences are emitted:
1356     //
1357     // v == 0                        : ADDI(W)
1358     // v[0,12) != 0 && v[12,32) == 0 : ADDI(W)
1359     // v[0,12) == 0 && v[12,32) != 0 : LUI
1360     // v[0,32) != 0                  : LUI+ADDI(W)
1361     //
1362     int64_t Hi20 = ((Value + 0x800) >> 12) & 0xFFFFF;
1363     int64_t Lo12 = SignExtend64<12>(Value);
1364     unsigned SrcReg = RISCV::X0;
1365 
1366     if (Hi20) {
1367       emitToStreamer(Out,
1368                      MCInstBuilder(RISCV::LUI).addReg(DestReg).addImm(Hi20));
1369       SrcReg = DestReg;
1370     }
1371 
1372     if (Lo12 || Hi20 == 0) {
1373       unsigned AddiOpcode =
1374           STI->hasFeature(RISCV::Feature64Bit) ? RISCV::ADDIW : RISCV::ADDI;
1375       emitToStreamer(Out, MCInstBuilder(AddiOpcode)
1376                               .addReg(DestReg)
1377                               .addReg(SrcReg)
1378                               .addImm(Lo12));
1379     }
1380     return;
1381   }
1382   assert(STI->hasFeature(RISCV::Feature64Bit) &&
1383          "Target must be 64-bit to support a >32-bit constant");
1384 
1385   // In the worst case, for a full 64-bit constant, a sequence of 8 instructions
1386   // (i.e., LUI+ADDIW+SLLI+ADDI+SLLI+ADDI+SLLI+ADDI) has to be emmitted. Note
1387   // that the first two instructions (LUI+ADDIW) can contribute up to 32 bits
1388   // while the following ADDI instructions contribute up to 12 bits each.
1389   //
1390   // On the first glance, implementing this seems to be possible by simply
1391   // emitting the most significant 32 bits (LUI+ADDIW) followed by as many left
1392   // shift (SLLI) and immediate additions (ADDI) as needed. However, due to the
1393   // fact that ADDI performs a sign extended addition, doing it like that would
1394   // only be possible when at most 11 bits of the ADDI instructions are used.
1395   // Using all 12 bits of the ADDI instructions, like done by GAS, actually
1396   // requires that the constant is processed starting with the least significant
1397   // bit.
1398   //
1399   // In the following, constants are processed from LSB to MSB but instruction
1400   // emission is performed from MSB to LSB by recursively calling
1401   // emitLoadImm. In each recursion, first the lowest 12 bits are removed
1402   // from the constant and the optimal shift amount, which can be greater than
1403   // 12 bits if the constant is sparse, is determined. Then, the shifted
1404   // remaining constant is processed recursively and gets emitted as soon as it
1405   // fits into 32 bits. The emission of the shifts and additions is subsequently
1406   // performed when the recursion returns.
1407   //
1408   int64_t Lo12 = SignExtend64<12>(Value);
1409   int64_t Hi52 = (Value + 0x800) >> 12;
1410   int ShiftAmount = 12 + findFirstSet((uint64_t)Hi52);
1411   Hi52 = SignExtend64(Hi52 >> (ShiftAmount - 12), 64 - ShiftAmount);
1412 
1413   emitLoadImm(DestReg, Hi52, Out);
1414 
1415   emitToStreamer(Out, MCInstBuilder(RISCV::SLLI)
1416                           .addReg(DestReg)
1417                           .addReg(DestReg)
1418                           .addImm(ShiftAmount));
1419 
1420   if (Lo12)
1421     emitToStreamer(Out, MCInstBuilder(RISCV::ADDI)
1422                             .addReg(DestReg)
1423                             .addReg(DestReg)
1424                             .addImm(Lo12));
1425 }
1426 
1427 void RISCVAsmParser::emitLoadLocalAddress(MCInst &Inst, SMLoc IDLoc,
1428                                           MCStreamer &Out) {
1429   // The local load address pseudo-instruction "lla" is used in PC-relative
1430   // addressing of symbols:
1431   //   lla rdest, symbol
1432   // expands to
1433   //   TmpLabel: AUIPC rdest, %pcrel_hi(symbol)
1434   //             ADDI rdest, %pcrel_lo(TmpLabel)
1435   MCContext &Ctx = getContext();
1436 
1437   MCSymbol *TmpLabel = Ctx.createTempSymbol(
1438       "pcrel_hi", /* AlwaysAddSuffix */ true, /* CanBeUnnamed */ false);
1439   Out.EmitLabel(TmpLabel);
1440 
1441   MCOperand DestReg = Inst.getOperand(0);
1442   const RISCVMCExpr *Symbol = RISCVMCExpr::create(
1443       Inst.getOperand(1).getExpr(), RISCVMCExpr::VK_RISCV_PCREL_HI, Ctx);
1444 
1445   emitToStreamer(
1446       Out, MCInstBuilder(RISCV::AUIPC).addOperand(DestReg).addExpr(Symbol));
1447 
1448   const MCExpr *RefToLinkTmpLabel =
1449       RISCVMCExpr::create(MCSymbolRefExpr::create(TmpLabel, Ctx),
1450                           RISCVMCExpr::VK_RISCV_PCREL_LO, Ctx);
1451 
1452   emitToStreamer(Out, MCInstBuilder(RISCV::ADDI)
1453                           .addOperand(DestReg)
1454                           .addOperand(DestReg)
1455                           .addExpr(RefToLinkTmpLabel));
1456 }
1457 
1458 bool RISCVAsmParser::processInstruction(MCInst &Inst, SMLoc IDLoc,
1459                                         MCStreamer &Out) {
1460   Inst.setLoc(IDLoc);
1461 
1462   if (Inst.getOpcode() == RISCV::PseudoLI) {
1463     auto Reg = Inst.getOperand(0).getReg();
1464     int64_t Imm = Inst.getOperand(1).getImm();
1465     // On RV32 the immediate here can either be a signed or an unsigned
1466     // 32-bit number. Sign extension has to be performed to ensure that Imm
1467     // represents the expected signed 64-bit number.
1468     if (!isRV64())
1469       Imm = SignExtend64<32>(Imm);
1470     emitLoadImm(Reg, Imm, Out);
1471     return false;
1472   } else if (Inst.getOpcode() == RISCV::PseudoLLA) {
1473     emitLoadLocalAddress(Inst, IDLoc, Out);
1474     return false;
1475   }
1476 
1477   emitToStreamer(Out, Inst);
1478   return false;
1479 }
1480 
1481 extern "C" void LLVMInitializeRISCVAsmParser() {
1482   RegisterMCAsmParser<RISCVAsmParser> X(getTheRISCV32Target());
1483   RegisterMCAsmParser<RISCVAsmParser> Y(getTheRISCV64Target());
1484 }
1485