1 //===- MIRParser.cpp - MIR serialization format parser implementation -----===//
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 // This file implements the class that parses the optional LLVM IR and machine
11 // functions that are stored in MIR files.
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #include "llvm/CodeGen/MIRParser/MIRParser.h"
16 #include "MIParser.h"
17 #include "llvm/ADT/DenseMap.h"
18 #include "llvm/ADT/StringRef.h"
19 #include "llvm/ADT/StringMap.h"
20 #include "llvm/ADT/STLExtras.h"
21 #include "llvm/AsmParser/Parser.h"
22 #include "llvm/AsmParser/SlotMapping.h"
23 #include "llvm/CodeGen/MachineFunction.h"
24 #include "llvm/CodeGen/MachineFrameInfo.h"
25 #include "llvm/CodeGen/MachineRegisterInfo.h"
26 #include "llvm/CodeGen/MIRYamlMapping.h"
27 #include "llvm/IR/BasicBlock.h"
28 #include "llvm/IR/DiagnosticInfo.h"
29 #include "llvm/IR/Instructions.h"
30 #include "llvm/IR/LLVMContext.h"
31 #include "llvm/IR/Module.h"
32 #include "llvm/IR/ValueSymbolTable.h"
33 #include "llvm/Support/LineIterator.h"
34 #include "llvm/Support/SMLoc.h"
35 #include "llvm/Support/SourceMgr.h"
36 #include "llvm/Support/MemoryBuffer.h"
37 #include "llvm/Support/YAMLTraits.h"
38 #include <memory>
39 
40 using namespace llvm;
41 
42 namespace llvm {
43 
44 /// This class implements the parsing of LLVM IR that's embedded inside a MIR
45 /// file.
46 class MIRParserImpl {
47   SourceMgr SM;
48   StringRef Filename;
49   LLVMContext &Context;
50   StringMap<std::unique_ptr<yaml::MachineFunction>> Functions;
51   SlotMapping IRSlots;
52   /// Maps from register class names to register classes.
53   StringMap<const TargetRegisterClass *> Names2RegClasses;
54 
55 public:
56   MIRParserImpl(std::unique_ptr<MemoryBuffer> Contents, StringRef Filename,
57                 LLVMContext &Context);
58 
59   void reportDiagnostic(const SMDiagnostic &Diag);
60 
61   /// Report an error with the given message at unknown location.
62   ///
63   /// Always returns true.
64   bool error(const Twine &Message);
65 
66   /// Report an error with the given message at the given location.
67   ///
68   /// Always returns true.
69   bool error(SMLoc Loc, const Twine &Message);
70 
71   /// Report a given error with the location translated from the location in an
72   /// embedded string literal to a location in the MIR file.
73   ///
74   /// Always returns true.
75   bool error(const SMDiagnostic &Error, SMRange SourceRange);
76 
77   /// Try to parse the optional LLVM module and the machine functions in the MIR
78   /// file.
79   ///
80   /// Return null if an error occurred.
81   std::unique_ptr<Module> parse();
82 
83   /// Parse the machine function in the current YAML document.
84   ///
85   /// \param NoLLVMIR - set to true when the MIR file doesn't have LLVM IR.
86   /// A dummy IR function is created and inserted into the given module when
87   /// this parameter is true.
88   ///
89   /// Return true if an error occurred.
90   bool parseMachineFunction(yaml::Input &In, Module &M, bool NoLLVMIR);
91 
92   /// Initialize the machine function to the state that's described in the MIR
93   /// file.
94   ///
95   /// Return true if error occurred.
96   bool initializeMachineFunction(MachineFunction &MF);
97 
98   /// Initialize the machine basic block using it's YAML representation.
99   ///
100   /// Return true if an error occurred.
101   bool initializeMachineBasicBlock(MachineFunction &MF, MachineBasicBlock &MBB,
102                                    const yaml::MachineBasicBlock &YamlMBB,
103                                    const PerFunctionMIParsingState &PFS);
104 
105   bool
106   initializeRegisterInfo(const MachineFunction &MF,
107                          MachineRegisterInfo &RegInfo,
108                          const yaml::MachineFunction &YamlMF,
109                          DenseMap<unsigned, unsigned> &VirtualRegisterSlots);
110 
111   bool initializeFrameInfo(MachineFrameInfo &MFI,
112                            const yaml::MachineFunction &YamlMF);
113 
114 private:
115   /// Return a MIR diagnostic converted from an MI string diagnostic.
116   SMDiagnostic diagFromMIStringDiag(const SMDiagnostic &Error,
117                                     SMRange SourceRange);
118 
119   /// Return a MIR diagnostic converted from an LLVM assembly diagnostic.
120   SMDiagnostic diagFromLLVMAssemblyDiag(const SMDiagnostic &Error,
121                                         SMRange SourceRange);
122 
123   /// Create an empty function with the given name.
124   void createDummyFunction(StringRef Name, Module &M);
125 
126   void initNames2RegClasses(const MachineFunction &MF);
127 
128   /// Check if the given identifier is a name of a register class.
129   ///
130   /// Return null if the name isn't a register class.
131   const TargetRegisterClass *getRegClass(const MachineFunction &MF,
132                                          StringRef Name);
133 };
134 
135 } // end namespace llvm
136 
137 MIRParserImpl::MIRParserImpl(std::unique_ptr<MemoryBuffer> Contents,
138                              StringRef Filename, LLVMContext &Context)
139     : SM(), Filename(Filename), Context(Context) {
140   SM.AddNewSourceBuffer(std::move(Contents), SMLoc());
141 }
142 
143 bool MIRParserImpl::error(const Twine &Message) {
144   Context.diagnose(DiagnosticInfoMIRParser(
145       DS_Error, SMDiagnostic(Filename, SourceMgr::DK_Error, Message.str())));
146   return true;
147 }
148 
149 bool MIRParserImpl::error(SMLoc Loc, const Twine &Message) {
150   Context.diagnose(DiagnosticInfoMIRParser(
151       DS_Error, SM.GetMessage(Loc, SourceMgr::DK_Error, Message)));
152   return true;
153 }
154 
155 bool MIRParserImpl::error(const SMDiagnostic &Error, SMRange SourceRange) {
156   assert(Error.getKind() == SourceMgr::DK_Error && "Expected an error");
157   reportDiagnostic(diagFromMIStringDiag(Error, SourceRange));
158   return true;
159 }
160 
161 void MIRParserImpl::reportDiagnostic(const SMDiagnostic &Diag) {
162   DiagnosticSeverity Kind;
163   switch (Diag.getKind()) {
164   case SourceMgr::DK_Error:
165     Kind = DS_Error;
166     break;
167   case SourceMgr::DK_Warning:
168     Kind = DS_Warning;
169     break;
170   case SourceMgr::DK_Note:
171     Kind = DS_Note;
172     break;
173   }
174   Context.diagnose(DiagnosticInfoMIRParser(Kind, Diag));
175 }
176 
177 static void handleYAMLDiag(const SMDiagnostic &Diag, void *Context) {
178   reinterpret_cast<MIRParserImpl *>(Context)->reportDiagnostic(Diag);
179 }
180 
181 std::unique_ptr<Module> MIRParserImpl::parse() {
182   yaml::Input In(SM.getMemoryBuffer(SM.getMainFileID())->getBuffer(),
183                  /*Ctxt=*/nullptr, handleYAMLDiag, this);
184   In.setContext(&In);
185 
186   if (!In.setCurrentDocument()) {
187     if (In.error())
188       return nullptr;
189     // Create an empty module when the MIR file is empty.
190     return llvm::make_unique<Module>(Filename, Context);
191   }
192 
193   std::unique_ptr<Module> M;
194   bool NoLLVMIR = false;
195   // Parse the block scalar manually so that we can return unique pointer
196   // without having to go trough YAML traits.
197   if (const auto *BSN =
198           dyn_cast_or_null<yaml::BlockScalarNode>(In.getCurrentNode())) {
199     SMDiagnostic Error;
200     M = parseAssembly(MemoryBufferRef(BSN->getValue(), Filename), Error,
201                       Context, &IRSlots);
202     if (!M) {
203       reportDiagnostic(diagFromLLVMAssemblyDiag(Error, BSN->getSourceRange()));
204       return M;
205     }
206     In.nextDocument();
207     if (!In.setCurrentDocument())
208       return M;
209   } else {
210     // Create an new, empty module.
211     M = llvm::make_unique<Module>(Filename, Context);
212     NoLLVMIR = true;
213   }
214 
215   // Parse the machine functions.
216   do {
217     if (parseMachineFunction(In, *M, NoLLVMIR))
218       return nullptr;
219     In.nextDocument();
220   } while (In.setCurrentDocument());
221 
222   return M;
223 }
224 
225 bool MIRParserImpl::parseMachineFunction(yaml::Input &In, Module &M,
226                                          bool NoLLVMIR) {
227   auto MF = llvm::make_unique<yaml::MachineFunction>();
228   yaml::yamlize(In, *MF, false);
229   if (In.error())
230     return true;
231   auto FunctionName = MF->Name;
232   if (Functions.find(FunctionName) != Functions.end())
233     return error(Twine("redefinition of machine function '") + FunctionName +
234                  "'");
235   Functions.insert(std::make_pair(FunctionName, std::move(MF)));
236   if (NoLLVMIR)
237     createDummyFunction(FunctionName, M);
238   else if (!M.getFunction(FunctionName))
239     return error(Twine("function '") + FunctionName +
240                  "' isn't defined in the provided LLVM IR");
241   return false;
242 }
243 
244 void MIRParserImpl::createDummyFunction(StringRef Name, Module &M) {
245   auto &Context = M.getContext();
246   Function *F = cast<Function>(M.getOrInsertFunction(
247       Name, FunctionType::get(Type::getVoidTy(Context), false)));
248   BasicBlock *BB = BasicBlock::Create(Context, "entry", F);
249   new UnreachableInst(Context, BB);
250 }
251 
252 bool MIRParserImpl::initializeMachineFunction(MachineFunction &MF) {
253   auto It = Functions.find(MF.getName());
254   if (It == Functions.end())
255     return error(Twine("no machine function information for function '") +
256                  MF.getName() + "' in the MIR file");
257   // TODO: Recreate the machine function.
258   const yaml::MachineFunction &YamlMF = *It->getValue();
259   if (YamlMF.Alignment)
260     MF.setAlignment(YamlMF.Alignment);
261   MF.setExposesReturnsTwice(YamlMF.ExposesReturnsTwice);
262   MF.setHasInlineAsm(YamlMF.HasInlineAsm);
263   PerFunctionMIParsingState PFS;
264   if (initializeRegisterInfo(MF, MF.getRegInfo(), YamlMF,
265                              PFS.VirtualRegisterSlots))
266     return true;
267   if (initializeFrameInfo(*MF.getFrameInfo(), YamlMF))
268     return true;
269 
270   const auto &F = *MF.getFunction();
271   for (const auto &YamlMBB : YamlMF.BasicBlocks) {
272     const BasicBlock *BB = nullptr;
273     const yaml::StringValue &Name = YamlMBB.Name;
274     if (!Name.Value.empty()) {
275       BB = dyn_cast_or_null<BasicBlock>(
276           F.getValueSymbolTable().lookup(Name.Value));
277       if (!BB)
278         return error(Name.SourceRange.Start,
279                      Twine("basic block '") + Name.Value +
280                          "' is not defined in the function '" + MF.getName() +
281                          "'");
282     }
283     auto *MBB = MF.CreateMachineBasicBlock(BB);
284     MF.insert(MF.end(), MBB);
285     bool WasInserted =
286         PFS.MBBSlots.insert(std::make_pair(YamlMBB.ID, MBB)).second;
287     if (!WasInserted)
288       return error(Twine("redefinition of machine basic block with id #") +
289                    Twine(YamlMBB.ID));
290   }
291 
292   if (YamlMF.BasicBlocks.empty())
293     return error(Twine("machine function '") + Twine(MF.getName()) +
294                  "' requires at least one machine basic block in its body");
295   // Initialize the machine basic blocks after creating them all so that the
296   // machine instructions parser can resolve the MBB references.
297   unsigned I = 0;
298   for (const auto &YamlMBB : YamlMF.BasicBlocks) {
299     if (initializeMachineBasicBlock(MF, *MF.getBlockNumbered(I++), YamlMBB,
300                                     PFS))
301       return true;
302   }
303   return false;
304 }
305 
306 bool MIRParserImpl::initializeMachineBasicBlock(
307     MachineFunction &MF, MachineBasicBlock &MBB,
308     const yaml::MachineBasicBlock &YamlMBB,
309     const PerFunctionMIParsingState &PFS) {
310   MBB.setAlignment(YamlMBB.Alignment);
311   if (YamlMBB.AddressTaken)
312     MBB.setHasAddressTaken();
313   MBB.setIsLandingPad(YamlMBB.IsLandingPad);
314   SMDiagnostic Error;
315   // Parse the successors.
316   for (const auto &MBBSource : YamlMBB.Successors) {
317     MachineBasicBlock *SuccMBB = nullptr;
318     if (parseMBBReference(SuccMBB, SM, MF, MBBSource.Value, PFS, IRSlots,
319                           Error))
320       return error(Error, MBBSource.SourceRange);
321     // TODO: Report an error when adding the same successor more than once.
322     MBB.addSuccessor(SuccMBB);
323   }
324   // Parse the liveins.
325   for (const auto &LiveInSource : YamlMBB.LiveIns) {
326     unsigned Reg = 0;
327     if (parseNamedRegisterReference(Reg, SM, MF, LiveInSource.Value, PFS,
328                                     IRSlots, Error))
329       return error(Error, LiveInSource.SourceRange);
330     MBB.addLiveIn(Reg);
331   }
332   // Parse the instructions.
333   for (const auto &MISource : YamlMBB.Instructions) {
334     MachineInstr *MI = nullptr;
335     if (parseMachineInstr(MI, SM, MF, MISource.Value, PFS, IRSlots, Error))
336       return error(Error, MISource.SourceRange);
337     MBB.insert(MBB.end(), MI);
338   }
339   return false;
340 }
341 
342 bool MIRParserImpl::initializeRegisterInfo(
343     const MachineFunction &MF, MachineRegisterInfo &RegInfo,
344     const yaml::MachineFunction &YamlMF,
345     DenseMap<unsigned, unsigned> &VirtualRegisterSlots) {
346   assert(RegInfo.isSSA());
347   if (!YamlMF.IsSSA)
348     RegInfo.leaveSSA();
349   assert(RegInfo.tracksLiveness());
350   if (!YamlMF.TracksRegLiveness)
351     RegInfo.invalidateLiveness();
352   RegInfo.enableSubRegLiveness(YamlMF.TracksSubRegLiveness);
353 
354   // Parse the virtual register information.
355   for (const auto &VReg : YamlMF.VirtualRegisters) {
356     const auto *RC = getRegClass(MF, VReg.Class.Value);
357     if (!RC)
358       return error(VReg.Class.SourceRange.Start,
359                    Twine("use of undefined register class '") +
360                        VReg.Class.Value + "'");
361     unsigned Reg = RegInfo.createVirtualRegister(RC);
362     // TODO: Report an error when the same virtual register with the same ID is
363     // redefined.
364     VirtualRegisterSlots.insert(std::make_pair(VReg.ID, Reg));
365   }
366   return false;
367 }
368 
369 bool MIRParserImpl::initializeFrameInfo(MachineFrameInfo &MFI,
370                                         const yaml::MachineFunction &YamlMF) {
371   const yaml::MachineFrameInfo &YamlMFI = YamlMF.FrameInfo;
372   MFI.setFrameAddressIsTaken(YamlMFI.IsFrameAddressTaken);
373   MFI.setReturnAddressIsTaken(YamlMFI.IsReturnAddressTaken);
374   MFI.setHasStackMap(YamlMFI.HasStackMap);
375   MFI.setHasPatchPoint(YamlMFI.HasPatchPoint);
376   MFI.setStackSize(YamlMFI.StackSize);
377   MFI.setOffsetAdjustment(YamlMFI.OffsetAdjustment);
378   if (YamlMFI.MaxAlignment)
379     MFI.ensureMaxAlignment(YamlMFI.MaxAlignment);
380   MFI.setAdjustsStack(YamlMFI.AdjustsStack);
381   MFI.setHasCalls(YamlMFI.HasCalls);
382   MFI.setMaxCallFrameSize(YamlMFI.MaxCallFrameSize);
383   MFI.setHasOpaqueSPAdjustment(YamlMFI.HasOpaqueSPAdjustment);
384   MFI.setHasVAStart(YamlMFI.HasVAStart);
385   MFI.setHasMustTailInVarArgFunc(YamlMFI.HasMustTailInVarArgFunc);
386 
387   // Initialize the fixed frame objects.
388   for (const auto &Object : YamlMF.FixedStackObjects) {
389     int ObjectIdx;
390     if (Object.Type != yaml::FixedMachineStackObject::SpillSlot)
391       ObjectIdx = MFI.CreateFixedObject(Object.Size, Object.Offset,
392                                         Object.IsImmutable, Object.IsAliased);
393     else
394       ObjectIdx = MFI.CreateFixedSpillStackObject(Object.Size, Object.Offset);
395     MFI.setObjectAlignment(ObjectIdx, Object.Alignment);
396     // TODO: Store the mapping between fixed object IDs and object indices to
397     // parse fixed stack object references correctly.
398   }
399 
400   // Initialize the ordinary frame objects.
401   for (const auto &Object : YamlMF.StackObjects) {
402     int ObjectIdx;
403     if (Object.Type == yaml::MachineStackObject::VariableSized)
404       ObjectIdx =
405           MFI.CreateVariableSizedObject(Object.Alignment, /*Alloca=*/nullptr);
406     else
407       ObjectIdx = MFI.CreateStackObject(
408           Object.Size, Object.Alignment,
409           Object.Type == yaml::MachineStackObject::SpillSlot);
410     MFI.setObjectOffset(ObjectIdx, Object.Offset);
411     // TODO: Store the mapping between object IDs and object indices to parse
412     // stack object references correctly.
413   }
414   return false;
415 }
416 
417 SMDiagnostic MIRParserImpl::diagFromMIStringDiag(const SMDiagnostic &Error,
418                                                  SMRange SourceRange) {
419   assert(SourceRange.isValid() && "Invalid source range");
420   SMLoc Loc = SourceRange.Start;
421   bool HasQuote = Loc.getPointer() < SourceRange.End.getPointer() &&
422                   *Loc.getPointer() == '\'';
423   // Translate the location of the error from the location in the MI string to
424   // the corresponding location in the MIR file.
425   Loc = Loc.getFromPointer(Loc.getPointer() + Error.getColumnNo() +
426                            (HasQuote ? 1 : 0));
427 
428   // TODO: Translate any source ranges as well.
429   return SM.GetMessage(Loc, Error.getKind(), Error.getMessage(), None,
430                        Error.getFixIts());
431 }
432 
433 SMDiagnostic MIRParserImpl::diagFromLLVMAssemblyDiag(const SMDiagnostic &Error,
434                                                      SMRange SourceRange) {
435   assert(SourceRange.isValid());
436 
437   // Translate the location of the error from the location in the llvm IR string
438   // to the corresponding location in the MIR file.
439   auto LineAndColumn = SM.getLineAndColumn(SourceRange.Start);
440   unsigned Line = LineAndColumn.first + Error.getLineNo() - 1;
441   unsigned Column = Error.getColumnNo();
442   StringRef LineStr = Error.getLineContents();
443   SMLoc Loc = Error.getLoc();
444 
445   // Get the full line and adjust the column number by taking the indentation of
446   // LLVM IR into account.
447   for (line_iterator L(*SM.getMemoryBuffer(SM.getMainFileID()), false), E;
448        L != E; ++L) {
449     if (L.line_number() == Line) {
450       LineStr = *L;
451       Loc = SMLoc::getFromPointer(LineStr.data());
452       auto Indent = LineStr.find(Error.getLineContents());
453       if (Indent != StringRef::npos)
454         Column += Indent;
455       break;
456     }
457   }
458 
459   return SMDiagnostic(SM, Loc, Filename, Line, Column, Error.getKind(),
460                       Error.getMessage(), LineStr, Error.getRanges(),
461                       Error.getFixIts());
462 }
463 
464 void MIRParserImpl::initNames2RegClasses(const MachineFunction &MF) {
465   if (!Names2RegClasses.empty())
466     return;
467   const TargetRegisterInfo *TRI = MF.getSubtarget().getRegisterInfo();
468   for (unsigned I = 0, E = TRI->getNumRegClasses(); I < E; ++I) {
469     const auto *RC = TRI->getRegClass(I);
470     Names2RegClasses.insert(
471         std::make_pair(StringRef(TRI->getRegClassName(RC)).lower(), RC));
472   }
473 }
474 
475 const TargetRegisterClass *MIRParserImpl::getRegClass(const MachineFunction &MF,
476                                                       StringRef Name) {
477   initNames2RegClasses(MF);
478   auto RegClassInfo = Names2RegClasses.find(Name);
479   if (RegClassInfo == Names2RegClasses.end())
480     return nullptr;
481   return RegClassInfo->getValue();
482 }
483 
484 MIRParser::MIRParser(std::unique_ptr<MIRParserImpl> Impl)
485     : Impl(std::move(Impl)) {}
486 
487 MIRParser::~MIRParser() {}
488 
489 std::unique_ptr<Module> MIRParser::parseLLVMModule() { return Impl->parse(); }
490 
491 bool MIRParser::initializeMachineFunction(MachineFunction &MF) {
492   return Impl->initializeMachineFunction(MF);
493 }
494 
495 std::unique_ptr<MIRParser> llvm::createMIRParserFromFile(StringRef Filename,
496                                                          SMDiagnostic &Error,
497                                                          LLVMContext &Context) {
498   auto FileOrErr = MemoryBuffer::getFile(Filename);
499   if (std::error_code EC = FileOrErr.getError()) {
500     Error = SMDiagnostic(Filename, SourceMgr::DK_Error,
501                          "Could not open input file: " + EC.message());
502     return nullptr;
503   }
504   return createMIRParser(std::move(FileOrErr.get()), Context);
505 }
506 
507 std::unique_ptr<MIRParser>
508 llvm::createMIRParser(std::unique_ptr<MemoryBuffer> Contents,
509                       LLVMContext &Context) {
510   auto Filename = Contents->getBufferIdentifier();
511   return llvm::make_unique<MIRParser>(
512       llvm::make_unique<MIRParserImpl>(std::move(Contents), Filename, Context));
513 }
514