llvm/lib/CodeGen/AsmPrinter/AsmPrinterInlineAsm.cpp

   1 //===-- AsmPrinterInlineAsm.cpp - AsmPrinter Inline Asm Handling ----------===//
   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 // This file implements the inline assembler pieces of the AsmPrinter class.
  10 //
  11 //===----------------------------------------------------------------------===//
  12
  13 #include "llvm/ADT/SmallString.h"
  14 #include "llvm/ADT/SmallVector.h"
  15 #include "llvm/ADT/StringExtras.h"
  16 #include "llvm/ADT/Twine.h"
  17 #include "llvm/CodeGen/AsmPrinter.h"
  18 #include "llvm/CodeGen/MachineBasicBlock.h"
  19 #include "llvm/CodeGen/MachineFunction.h"
  20 #include "llvm/CodeGen/MachineModuleInfo.h"
  21 #include "llvm/CodeGen/TargetRegisterInfo.h"
  22 #include "llvm/CodeGen/TargetSubtargetInfo.h"
  23 #include "llvm/IR/Constants.h"
  24 #include "llvm/IR/DataLayout.h"
  25 #include "llvm/IR/DiagnosticInfo.h"
  26 #include "llvm/IR/InlineAsm.h"
  27 #include "llvm/IR/LLVMContext.h"
  28 #include "llvm/IR/Module.h"
  29 #include "llvm/MC/MCAsmInfo.h"
  30 #include "llvm/MC/MCInstrInfo.h"
  31 #include "llvm/MC/MCParser/MCAsmLexer.h"
  32 #include "llvm/MC/MCParser/MCTargetAsmParser.h"
  33 #include "llvm/MC/MCStreamer.h"
  34 #include "llvm/MC/MCSymbol.h"
  35 #include "llvm/MC/TargetRegistry.h"
  36 #include "llvm/Support/ErrorHandling.h"
  37 #include "llvm/Support/MemoryBuffer.h"
  38 #include "llvm/Support/SourceMgr.h"
  39 #include "llvm/Support/raw_ostream.h"
  40 #include "llvm/Target/TargetMachine.h"
  41 using namespace llvm;
  42
  43 #define DEBUG_TYPE "asm-printer"
  44
  45 unsigned AsmPrinter::addInlineAsmDiagBuffer(StringRef AsmStr,
  46                                             const MDNode *LocMDNode) const {
  47   MCContext &Context = MMI->getContext();
  48   Context.initInlineSourceManager();
  49   SourceMgr &SrcMgr = *Context.getInlineSourceManager();
  50   std::vector<const MDNode *> &LocInfos = Context.getLocInfos();
  51
  52   std::unique_ptr<MemoryBuffer> Buffer;
  53   // The inline asm source manager will outlive AsmStr, so make a copy of the
  54   // string for SourceMgr to own.
  55   Buffer = MemoryBuffer::getMemBufferCopy(AsmStr, "<inline asm>");
  56
  57   // Tell SrcMgr about this buffer, it takes ownership of the buffer.
  58   unsigned BufNum = SrcMgr.AddNewSourceBuffer(std::move(Buffer), SMLoc());
  59
  60   // Store LocMDNode in DiagInfo, using BufNum as an identifier.
  61   if (LocMDNode) {
  62     LocInfos.resize(BufNum);
  63     LocInfos[BufNum - 1] = LocMDNode;
  64   }
  65
  66   return BufNum;
  67 }
  68
  69
  70 /// EmitInlineAsm - Emit a blob of inline asm to the output streamer.
  71 void AsmPrinter::emitInlineAsm(StringRef Str, const MCSubtargetInfo &STI,
  72                                const MCTargetOptions &MCOptions,
  73                                const MDNode *LocMDNode,
  74                                InlineAsm::AsmDialect Dialect) const {
  75   assert(!Str.empty() && "Can't emit empty inline asm block");
  76
  77   // Remember if the buffer is nul terminated or not so we can avoid a copy.
  78   bool isNullTerminated = Str.back() == 0;
  79   if (isNullTerminated)
  80     Str = Str.substr(0, Str.size()-1);
  81
  82   // If the output streamer does not have mature MC support or the integrated
  83   // assembler has been disabled or not required, just emit the blob textually.
  84   // Otherwise parse the asm and emit it via MC support.
  85   // This is useful in case the asm parser doesn't handle something but the
  86   // system assembler does.
  87   const MCAsmInfo *MCAI = TM.getMCAsmInfo();
  88   assert(MCAI && "No MCAsmInfo");
  89   if (!MCAI->useIntegratedAssembler() &&
  90       !MCAI->parseInlineAsmUsingAsmParser() &&
  91       !OutStreamer->isIntegratedAssemblerRequired()) {
  92     emitInlineAsmStart();
  93     OutStreamer->emitRawText(Str);
  94     emitInlineAsmEnd(STI, nullptr);
  95     return;
  96   }
  97
  98   unsigned BufNum = addInlineAsmDiagBuffer(Str, LocMDNode);
  99   SourceMgr &SrcMgr = *MMI->getContext().getInlineSourceManager();
 100   SrcMgr.setIncludeDirs(MCOptions.IASSearchPaths);
 101
 102   std::unique_ptr<MCAsmParser> Parser(
 103       createMCAsmParser(SrcMgr, OutContext, *OutStreamer, *MAI, BufNum));
 104
 105   // Do not use assembler-level information for parsing inline assembly.
 106   OutStreamer->setUseAssemblerInfoForParsing(false);
 107
 108   // We create a new MCInstrInfo here since we might be at the module level
 109   // and not have a MachineFunction to initialize the TargetInstrInfo from and
 110   // we only need MCInstrInfo for asm parsing. We create one unconditionally
 111   // because it's not subtarget dependent.
 112   std::unique_ptr<MCInstrInfo> MII(TM.getTarget().createMCInstrInfo());
 113   assert(MII && "Failed to create instruction info");
 114   std::unique_ptr<MCTargetAsmParser> TAP(TM.getTarget().createMCAsmParser(
 115       STI, *Parser, *MII, MCOptions));
 116   if (!TAP)
 117     report_fatal_error("Inline asm not supported by this streamer because"
 118                        " we don't have an asm parser for this target\n");
 119   Parser->setAssemblerDialect(Dialect);
 120   Parser->setTargetParser(*TAP);
 121   // Enable lexing Masm binary and hex integer literals in intel inline
 122   // assembly.
 123   if (Dialect == InlineAsm::AD_Intel)
 124     Parser->getLexer().setLexMasmIntegers(true);
 125
 126   emitInlineAsmStart();
 127   // Don't implicitly switch to the text section before the asm.
 128   (void)Parser->Run(/*NoInitialTextSection*/ true,
 129                     /*NoFinalize*/ true);
 130   emitInlineAsmEnd(STI, &TAP->getSTI());
 131 }
 132
 133 static void EmitInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
 134                              MachineModuleInfo *MMI, const MCAsmInfo *MAI,
 135                              AsmPrinter *AP, uint64_t LocCookie,
 136                              raw_ostream &OS) {
 137   bool InputIsIntelDialect = MI->getInlineAsmDialect() == InlineAsm::AD_Intel;
 138
 139   if (InputIsIntelDialect) {
 140     // Switch to the inline assembly variant.
 141     OS << "\t.intel_syntax\n\t";
 142   }
 143
 144   int CurVariant = -1; // The number of the {.|.|.} region we are in.
 145   const char *LastEmitted = AsmStr; // One past the last character emitted.
 146   unsigned NumOperands = MI->getNumOperands();
 147
 148   int AsmPrinterVariant;
 149   if (InputIsIntelDialect)
 150     AsmPrinterVariant = 1; // X86MCAsmInfo.cpp's AsmWriterFlavorTy::Intel.
 151   else
 152     AsmPrinterVariant = MMI->getTarget().unqualifiedInlineAsmVariant();
 153
 154   // FIXME: Should this happen for `asm inteldialect` as well?
 155   if (!InputIsIntelDialect && MAI->getEmitGNUAsmStartIndentationMarker())
 156     OS << '\t';
 157
 158   while (*LastEmitted) {
 159     switch (*LastEmitted) {
 160     default: {
 161       // Not a special case, emit the string section literally.
 162       const char *LiteralEnd = LastEmitted+1;
 163       while (*LiteralEnd && *LiteralEnd != '{' && *LiteralEnd != '|' &&
 164              *LiteralEnd != '}' && *LiteralEnd != '$' && *LiteralEnd != '\n')
 165         ++LiteralEnd;
 166       if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
 167         OS.write(LastEmitted, LiteralEnd - LastEmitted);
 168       LastEmitted = LiteralEnd;
 169       break;
 170     }
 171     case '\n':
 172       ++LastEmitted;   // Consume newline character.
 173       OS << '\n';      // Indent code with newline.
 174       break;
 175     case '$': {
 176       ++LastEmitted;   // Consume '$' character.
 177       bool Done = true;
 178
 179       // Handle escapes.
 180       switch (*LastEmitted) {
 181       default: Done = false; break;
 182       case '$':     // $$ -> $
 183         if (!InputIsIntelDialect)
 184           if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
 185             OS << '$';
 186         ++LastEmitted;  // Consume second '$' character.
 187         break;
 188       case '(':        // $( -> same as GCC's { character.
 189         ++LastEmitted; // Consume '(' character.
 190         if (CurVariant != -1)
 191           report_fatal_error("Nested variants found in inline asm string: '" +
 192                              Twine(AsmStr) + "'");
 193         CurVariant = 0; // We're in the first variant now.
 194         break;
 195       case '|':
 196         ++LastEmitted; // Consume '|' character.
 197         if (CurVariant == -1)
 198           OS << '|'; // This is gcc's behavior for | outside a variant.
 199         else
 200           ++CurVariant; // We're in the next variant.
 201         break;
 202       case ')':        // $) -> same as GCC's } char.
 203         ++LastEmitted; // Consume ')' character.
 204         if (CurVariant == -1)
 205           OS << '}'; // This is gcc's behavior for } outside a variant.
 206         else
 207           CurVariant = -1;
 208         break;
 209       }
 210       if (Done) break;
 211
 212       bool HasCurlyBraces = false;
 213       if (*LastEmitted == '{') {     // ${variable}
 214         ++LastEmitted;               // Consume '{' character.
 215         HasCurlyBraces = true;
 216       }
 217
 218       // If we have ${:foo}, then this is not a real operand reference, it is a
 219       // "magic" string reference, just like in .td files.  Arrange to call
 220       // PrintSpecial.
 221       if (HasCurlyBraces && *LastEmitted == ':') {
 222         ++LastEmitted;
 223         const char *StrStart = LastEmitted;
 224         const char *StrEnd = strchr(StrStart, '}');
 225         if (!StrEnd)
 226           report_fatal_error("Unterminated ${:foo} operand in inline asm"
 227                              " string: '" + Twine(AsmStr) + "'");
 228         if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
 229           AP->PrintSpecial(MI, OS, StringRef(StrStart, StrEnd - StrStart));
 230         LastEmitted = StrEnd+1;
 231         break;
 232       }
 233
 234       const char *IDStart = LastEmitted;
 235       const char *IDEnd = IDStart;
 236       while (isDigit(*IDEnd))
 237         ++IDEnd;
 238
 239       unsigned Val;
 240       if (StringRef(IDStart, IDEnd-IDStart).getAsInteger(10, Val))
 241         report_fatal_error("Bad $ operand number in inline asm string: '" +
 242                            Twine(AsmStr) + "'");
 243       LastEmitted = IDEnd;
 244
 245       if (Val >= NumOperands - 1)
 246         report_fatal_error("Invalid $ operand number in inline asm string: '" +
 247                            Twine(AsmStr) + "'");
 248
 249       char Modifier[2] = { 0, 0 };
 250
 251       if (HasCurlyBraces) {
 252         // If we have curly braces, check for a modifier character.  This
 253         // supports syntax like ${0:u}, which correspond to "%u0" in GCC asm.
 254         if (*LastEmitted == ':') {
 255           ++LastEmitted;    // Consume ':' character.
 256           if (*LastEmitted == 0)
 257             report_fatal_error("Bad ${:} expression in inline asm string: '" +
 258                                Twine(AsmStr) + "'");
 259
 260           Modifier[0] = *LastEmitted;
 261           ++LastEmitted;    // Consume modifier character.
 262         }
 263
 264         if (*LastEmitted != '}')
 265           report_fatal_error("Bad ${} expression in inline asm string: '" +
 266                              Twine(AsmStr) + "'");
 267         ++LastEmitted;    // Consume '}' character.
 268       }
 269
 270       // Okay, we finally have a value number.  Ask the target to print this
 271       // operand!
 272       if (CurVariant == -1 || CurVariant == AsmPrinterVariant) {
 273         unsigned OpNo = InlineAsm::MIOp_FirstOperand;
 274
 275         bool Error = false;
 276
 277         // Scan to find the machine operand number for the operand.
 278         for (; Val; --Val) {
 279           if (OpNo >= MI->getNumOperands())
 280             break;
 281           const InlineAsm::Flag F(MI->getOperand(OpNo).getImm());
 282           OpNo += F.getNumOperandRegisters() + 1;
 283         }
 284
 285         // We may have a location metadata attached to the end of the
 286         // instruction, and at no point should see metadata at any
 287         // other point while processing. It's an error if so.
 288         if (OpNo >= MI->getNumOperands() || MI->getOperand(OpNo).isMetadata()) {
 289           Error = true;
 290         } else {
 291           const InlineAsm::Flag F(MI->getOperand(OpNo).getImm());
 292           ++OpNo; // Skip over the ID number.
 293
 294           // FIXME: Shouldn't arch-independent output template handling go into
 295           // PrintAsmOperand?
 296           // Labels are target independent.
 297           if (MI->getOperand(OpNo).isBlockAddress()) {
 298             const BlockAddress *BA = MI->getOperand(OpNo).getBlockAddress();
 299             MCSymbol *Sym = AP->GetBlockAddressSymbol(BA);
 300             Sym->print(OS, AP->MAI);
 301             MMI->getContext().registerInlineAsmLabel(Sym);
 302           } else if (MI->getOperand(OpNo).isMBB()) {
 303             const MCSymbol *Sym = MI->getOperand(OpNo).getMBB()->getSymbol();
 304             Sym->print(OS, AP->MAI);
 305           } else if (F.isMemKind()) {
 306             Error = AP->PrintAsmMemoryOperand(
 307                 MI, OpNo, Modifier[0] ? Modifier : nullptr, OS);
 308           } else {
 309             Error = AP->PrintAsmOperand(MI, OpNo,
 310                                         Modifier[0] ? Modifier : nullptr, OS);
 311           }
 312         }
 313         if (Error) {
 314           std::string msg;
 315           raw_string_ostream Msg(msg);
 316           Msg << "invalid operand in inline asm: '" << AsmStr << "'";
 317           MMI->getModule()->getContext().emitError(LocCookie, Msg.str());
 318         }
 319       }
 320       break;
 321     }
 322     }
 323   }
 324   if (InputIsIntelDialect)
 325     OS << "\n\t.att_syntax";
 326   OS << '\n' << (char)0;  // null terminate string.
 327 }
 328
 329 /// This method formats and emits the specified machine instruction that is an
 330 /// inline asm.
 331 void AsmPrinter::emitInlineAsm(const MachineInstr *MI) const {
 332   assert(MI->isInlineAsm() && "printInlineAsm only works on inline asms");
 333
 334   // Disassemble the AsmStr, printing out the literal pieces, the operands, etc.
 335   const char *AsmStr = MI->getOperand(0).getSymbolName();
 336
 337   // If this asmstr is empty, just print the #APP/#NOAPP markers.
 338   // These are useful to see where empty asm's wound up.
 339   if (AsmStr[0] == 0) {
 340     OutStreamer->emitRawComment(MAI->getInlineAsmStart());
 341     OutStreamer->emitRawComment(MAI->getInlineAsmEnd());
 342     return;
 343   }
 344
 345   // Emit the #APP start marker.  This has to happen even if verbose-asm isn't
 346   // enabled, so we use emitRawComment.
 347   OutStreamer->emitRawComment(MAI->getInlineAsmStart());
 348
 349   // Get the !srcloc metadata node if we have it, and decode the loc cookie from
 350   // it.
 351   uint64_t LocCookie = 0;
 352   const MDNode *LocMD = nullptr;
 353   for (const MachineOperand &MO : llvm::reverse(MI->operands())) {
 354     if (MO.isMetadata() && (LocMD = MO.getMetadata()) &&
 355         LocMD->getNumOperands() != 0) {
 356       if (const ConstantInt *CI =
 357               mdconst::dyn_extract<ConstantInt>(LocMD->getOperand(0))) {
 358         LocCookie = CI->getZExtValue();
 359         break;
 360       }
 361     }
 362   }
 363
 364   // Emit the inline asm to a temporary string so we can emit it through
 365   // EmitInlineAsm.
 366   SmallString<256> StringData;
 367   raw_svector_ostream OS(StringData);
 368
 369   AsmPrinter *AP = const_cast<AsmPrinter*>(this);
 370   EmitInlineAsmStr(AsmStr, MI, MMI, MAI, AP, LocCookie, OS);
 371
 372   // Emit warnings if we use reserved registers on the clobber list, as
 373   // that might lead to undefined behaviour.
 374   SmallVector<Register, 8> RestrRegs;
 375   const TargetRegisterInfo *TRI = MF->getSubtarget().getRegisterInfo();
 376   // Start with the first operand descriptor, and iterate over them.
 377   for (unsigned I = InlineAsm::MIOp_FirstOperand, NumOps = MI->getNumOperands();
 378        I < NumOps; ++I) {
 379     const MachineOperand &MO = MI->getOperand(I);
 380     if (!MO.isImm())
 381       continue;
 382     const InlineAsm::Flag F(MO.getImm());
 383     if (F.isClobberKind()) {
 384       Register Reg = MI->getOperand(I + 1).getReg();
 385       if (!TRI->isAsmClobberable(*MF, Reg))
 386         RestrRegs.push_back(Reg);
 387     }
 388     // Skip to one before the next operand descriptor, if it exists.
 389     I += F.getNumOperandRegisters();
 390   }
 391
 392   if (!RestrRegs.empty()) {
 393     std::string Msg = "inline asm clobber list contains reserved registers: ";
 394     ListSeparator LS;
 395     for (const Register RR : RestrRegs) {
 396       Msg += LS;
 397       Msg += TRI->getRegAsmName(RR);
 398     }
 399     const char *Note =
 400         "Reserved registers on the clobber list may not be "
 401         "preserved across the asm statement, and clobbering them may "
 402         "lead to undefined behaviour.";
 403     MMI->getModule()->getContext().diagnose(DiagnosticInfoInlineAsm(
 404         LocCookie, Msg, DiagnosticSeverity::DS_Warning));
 405     MMI->getModule()->getContext().diagnose(
 406         DiagnosticInfoInlineAsm(LocCookie, Note, DiagnosticSeverity::DS_Note));
 407
 408     for (const Register RR : RestrRegs) {
 409       if (std::optional<std::string> reason =
 410               TRI->explainReservedReg(*MF, RR)) {
 411         MMI->getModule()->getContext().diagnose(DiagnosticInfoInlineAsm(
 412             LocCookie, *reason, DiagnosticSeverity::DS_Note));
 413       }
 414     }
 415   }
 416
 417   emitInlineAsm(OS.str(), getSubtargetInfo(), TM.Options.MCOptions, LocMD,
 418                 MI->getInlineAsmDialect());
 419
 420   // Emit the #NOAPP end marker.  This has to happen even if verbose-asm isn't
 421   // enabled, so we use emitRawComment.
 422   OutStreamer->emitRawComment(MAI->getInlineAsmEnd());
 423 }
 424
 425 /// PrintSpecial - Print information related to the specified machine instr
 426 /// that is independent of the operand, and may be independent of the instr
 427 /// itself.  This can be useful for portably encoding the comment character
 428 /// or other bits of target-specific knowledge into the asmstrings.  The
 429 /// syntax used is ${:comment}.  Targets can override this to add support
 430 /// for their own strange codes.
 431 void AsmPrinter::PrintSpecial(const MachineInstr *MI, raw_ostream &OS,
 432                               StringRef Code) const {
 433   if (Code == "private") {
 434     const DataLayout &DL = MF->getDataLayout();
 435     OS << DL.getPrivateGlobalPrefix();
 436   } else if (Code == "comment") {
 437     OS << MAI->getCommentString();
 438   } else if (Code == "uid") {
 439     // Comparing the address of MI isn't sufficient, because machineinstrs may
 440     // be allocated to the same address across functions.
 441
 442     // If this is a new LastFn instruction, bump the counter.
 443     if (LastMI != MI || LastFn != getFunctionNumber()) {
 444       ++Counter;
 445       LastMI = MI;
 446       LastFn = getFunctionNumber();
 447     }
 448     OS << Counter;
 449   } else {
 450     std::string msg;
 451     raw_string_ostream Msg(msg);
 452     Msg << "Unknown special formatter '" << Code
 453          << "' for machine instr: " << *MI;
 454     report_fatal_error(Twine(Msg.str()));
 455   }
 456 }
 457
 458 void AsmPrinter::PrintSymbolOperand(const MachineOperand &MO, raw_ostream &OS) {
 459   assert(MO.isGlobal() && "caller should check MO.isGlobal");
 460   getSymbolPreferLocal(*MO.getGlobal())->print(OS, MAI);
 461   printOffset(MO.getOffset(), OS);
 462 }
 463
 464 /// PrintAsmOperand - Print the specified operand of MI, an INLINEASM
 465 /// instruction, using the specified assembler variant.  Targets should
 466 /// override this to format as appropriate for machine specific ExtraCodes
 467 /// or when the arch-independent handling would be too complex otherwise.
 468 bool AsmPrinter::PrintAsmOperand(const MachineInstr *MI, unsigned OpNo,
 469                                  const char *ExtraCode, raw_ostream &O) {
 470   // Does this asm operand have a single letter operand modifier?
 471   if (ExtraCode && ExtraCode[0]) {
 472     if (ExtraCode[1] != 0) return true; // Unknown modifier.
 473
 474     // https://gcc.gnu.org/onlinedocs/gccint/Output-Template.html
 475     const MachineOperand &MO = MI->getOperand(OpNo);
 476     switch (ExtraCode[0]) {
 477     default:
 478       return true;  // Unknown modifier.
 479     case 'a': // Print as memory address.
 480       if (MO.isReg()) {
 481         PrintAsmMemoryOperand(MI, OpNo, nullptr, O);
 482         return false;
 483       }
 484       [[fallthrough]]; // GCC allows '%a' to behave like '%c' with immediates.
 485     case 'c': // Substitute immediate value without immediate syntax
 486       if (MO.isImm()) {
 487         O << MO.getImm();
 488         return false;
 489       }
 490       if (MO.isGlobal()) {
 491         PrintSymbolOperand(MO, O);
 492         return false;
 493       }
 494       return true;
 495     case 'n':  // Negate the immediate constant.
 496       if (!MO.isImm())
 497         return true;
 498       O << -MO.getImm();
 499       return false;
 500     case 's':  // The GCC deprecated s modifier
 501       if (!MO.isImm())
 502         return true;
 503       O << ((32 - MO.getImm()) & 31);
 504       return false;
 505     }
 506   }
 507   return true;
 508 }
 509
 510 bool AsmPrinter::PrintAsmMemoryOperand(const MachineInstr *MI, unsigned OpNo,
 511                                        const char *ExtraCode, raw_ostream &O) {
 512   // Target doesn't support this yet!
 513   return true;
 514 }
 515
 516 void AsmPrinter::emitInlineAsmStart() const {}
 517
 518 void AsmPrinter::emitInlineAsmEnd(const MCSubtargetInfo &StartInfo,
 519                                   const MCSubtargetInfo *EndInfo) const {}