MFC 261991:

Upgrade our copy of llvm/clang to 3.4 release.  This version supports
all of the features in the current working draft of the upcoming C++
standard, provisionally named C++1y.

The code generator's performance is greatly increased, and the loop
auto-vectorizer is now enabled at -Os and -O2 in addition to -O3.  The
PowerPC backend has made several major improvements to code generation
quality and compile time, and the X86, SPARC, ARM32, Aarch64 and SystemZ
backends have all seen major feature work.

Release notes for llvm and clang can be found here:
<http://llvm.org/releases/3.4/docs/ReleaseNotes.html>
<http://llvm.org/releases/3.4/tools/clang/docs/ReleaseNotes.html>

MFC 262121 (by emaste):

Update lldb for clang/llvm 3.4 import

This commit largely restores the lldb source to the upstream r196259
snapshot with the addition of threaded inferior support and a few bug
fixes.

Specific upstream lldb revisions restored include:
   SVN      git
  181387  779e6ac
  181703  7bef4e2
  182099  b31044e
  182650  f2dcf35
  182683  0d91b80
  183862  15c1774
  183929  99447a6
  184177  0b2934b
  184948  4dc3761
  184954  007e7bc
  186990  eebd175

Sponsored by:	DARPA, AFRL

MFC 262186 (by emaste):

Fix mismerge in r262121

A break statement was lost in the merge.  The error had no functional
impact, but restore it to reduce the diff against upstream.

MFC 262303:

Pull in r197521 from upstream clang trunk (by rdivacky):

  Use the integrated assembler by default on FreeBSD/ppc and ppc64.

Requested by:	jhibbits

MFC 262611:

Pull in r196874 from upstream llvm trunk:

  Fix a crash that occurs when PWD is invalid.

  MCJIT needs to be able to run in hostile environments, even when PWD
  is invalid. There's no need to crash MCJIT in this case.

  The obvious fix is to simply leave MCContext's CompilationDir empty
  when PWD can't be determined. This way, MCJIT clients,
  and other clients that link with LLVM don't need a valid working directory.

  If we do want to guarantee valid CompilationDir, that should be done
  only for clients of getCompilationDir(). This is as simple as checking
  for an empty string.

  The only current use of getCompilationDir is EmitGenDwarfInfo, which
  won't conceivably run with an invalid working dir. However, in the
  purely hypothetically and untestable case that this happens, the
  AT_comp_dir will be omitted from the compilation_unit DIE.

This should help fix assertions occurring with ports-mgmt/tinderbox,
when it is using jails, and sometimes invalidates clang's current
working directory.

Reported by:	decke

MFC 262809:

Pull in r203007 from upstream clang trunk:

  Don't produce an alias between destructors with different calling conventions.

  Fixes pr19007.

(Please note that is an LLVM PR identifier, not a FreeBSD one.)

This should fix Firefox and/or libxul crashes (due to problems with
regparm/stdcall calling conventions) on i386.

Reported by:	multiple users on freebsd-current
PR:		bin/187103

MFC 263048:

Repair recognition of "CC" as an alias for the C++ compiler, since it
was silently broken by upstream for a Windows-specific use-case.

Apparently some versions of CMake still rely on this archaic feature...

Reported by:	rakuco

MFC 263049:

Garbage collect the old way of adding the libstdc++ include directories
in clang's InitHeaderSearch.cpp.  This has been superseded by David
Chisnall's commit in r255321.

Moreover, if libc++ is used, the libstdc++ include directories should
not be in the search path at all.  These directories are now only used
if you pass -stdlib=libstdc++.

1 files changed, 461 insertions, 0 deletions

diff --git a/contrib/llvm/lib/MC/MCModuleYAML.cpp b/contrib/llvm/lib/MC/MCModuleYAML.cpp
new file mode 100644
index 0000000..e2de578
--- /dev/null
+++ b/contrib/llvm/lib/MC/MCModuleYAML.cpp
@@ -0,0 +1,461 @@
+//===- MCModuleYAML.cpp - MCModule YAMLIO implementation ------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file defines classes for handling the YAML representation of MCModule.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/MC/MCModuleYAML.h"
+#include "llvm/ADT/StringMap.h"
+#include "llvm/MC/MCAtom.h"
+#include "llvm/MC/MCFunction.h"
+#include "llvm/MC/MCInstrInfo.h"
+#include "llvm/MC/MCRegisterInfo.h"
+#include "llvm/Object/YAML.h"
+#include "llvm/Support/Allocator.h"
+#include "llvm/Support/MathExtras.h"
+#include "llvm/Support/YAMLTraits.h"
+#include <vector>
+
+namespace llvm {
+
+namespace {
+
+// This class is used to map opcode and register names to enum values.
+//
+// There are at least 3 obvious ways to do this:
+// 1- Generate an MII/MRI method using a tablegen StringMatcher
+// 2- Write an MII/MRI method using std::lower_bound and the assumption that
+//    the enums are sorted (starting at a fixed value).
+// 3- Do the matching manually as is done here.
+//
+// Why 3?
+// 1- A StringMatcher function for thousands of entries would incur
+//    a non-negligible binary size overhead.
+// 2- The lower_bound comparators would be somewhat involved and aren't
+//    obviously reusable (see LessRecordRegister in llvm/TableGen/Record.h)
+// 3- This isn't actually something useful outside tests (but the same argument
+//    can be made against having {MII,MRI}::getName).
+//
+// If this becomes useful outside this specific situation, feel free to do
+// the Right Thing (tm) and move the functionality to MII/MRI.
+//
+class InstrRegInfoHolder {
+  typedef StringMap<unsigned, BumpPtrAllocator> EnumValByNameTy;
+  EnumValByNameTy InstEnumValueByName;
+  EnumValByNameTy RegEnumValueByName;
+
+public:
+  const MCInstrInfo &MII;
+  const MCRegisterInfo &MRI;
+  InstrRegInfoHolder(const MCInstrInfo &MII, const MCRegisterInfo &MRI)
+      : InstEnumValueByName(NextPowerOf2(MII.getNumOpcodes())),
+        RegEnumValueByName(NextPowerOf2(MRI.getNumRegs())), MII(MII), MRI(MRI) {
+    for (int i = 0, e = MII.getNumOpcodes(); i != e; ++i)
+      InstEnumValueByName[MII.getName(i)] = i;
+    for (int i = 0, e = MRI.getNumRegs(); i != e; ++i)
+      RegEnumValueByName[MRI.getName(i)] = i;
+  }
+
+  bool matchRegister(StringRef Name, unsigned &Reg) {
+    EnumValByNameTy::const_iterator It = RegEnumValueByName.find(Name);
+    if (It == RegEnumValueByName.end())
+      return false;
+    Reg = It->getValue();
+    return true;
+  }
+  bool matchOpcode(StringRef Name, unsigned &Opc) {
+    EnumValByNameTy::const_iterator It = InstEnumValueByName.find(Name);
+    if (It == InstEnumValueByName.end())
+      return false;
+    Opc = It->getValue();
+    return true;
+  }
+};
+
+} // end unnamed namespace
+
+namespace MCModuleYAML {
+
+LLVM_YAML_STRONG_TYPEDEF(unsigned, OpcodeEnum)
+
+struct Operand {
+  MCOperand MCOp;
+};
+
+struct Inst {
+  OpcodeEnum Opcode;
+  std::vector<Operand> Operands;
+  uint64_t Size;
+};
+
+struct Atom {
+  MCAtom::AtomKind Type;
+  yaml::Hex64 StartAddress;
+  uint64_t Size;
+
+  std::vector<Inst> Insts;
+  object::yaml::BinaryRef Data;
+};
+
+struct BasicBlock {
+  yaml::Hex64 Address;
+  std::vector<yaml::Hex64> Preds;
+  std::vector<yaml::Hex64> Succs;
+};
+
+struct Function {
+  StringRef Name;
+  std::vector<BasicBlock> BasicBlocks;
+};
+
+struct Module {
+  std::vector<Atom> Atoms;
+  std::vector<Function> Functions;
+};
+
+} // end namespace MCModuleYAML
+} // end namespace llvm
+
+LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(llvm::yaml::Hex64)
+LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(llvm::MCModuleYAML::Operand)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MCModuleYAML::Inst)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MCModuleYAML::Atom)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MCModuleYAML::BasicBlock)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MCModuleYAML::Function)
+
+namespace llvm {
+
+namespace yaml {
+
+template <> struct ScalarEnumerationTraits<MCAtom::AtomKind> {
+  static void enumeration(IO &IO, MCAtom::AtomKind &Kind);
+};
+
+template <> struct MappingTraits<MCModuleYAML::Atom> {
+  static void mapping(IO &IO, MCModuleYAML::Atom &A);
+};
+
+template <> struct MappingTraits<MCModuleYAML::Inst> {
+  static void mapping(IO &IO, MCModuleYAML::Inst &I);
+};
+
+template <> struct MappingTraits<MCModuleYAML::BasicBlock> {
+  static void mapping(IO &IO, MCModuleYAML::BasicBlock &BB);
+};
+
+template <> struct MappingTraits<MCModuleYAML::Function> {
+  static void mapping(IO &IO, MCModuleYAML::Function &Fn);
+};
+
+template <> struct MappingTraits<MCModuleYAML::Module> {
+  static void mapping(IO &IO, MCModuleYAML::Module &M);
+};
+
+template <> struct ScalarTraits<MCModuleYAML::Operand> {
+  static void output(const MCModuleYAML::Operand &, void *,
+                     llvm::raw_ostream &);
+  static StringRef input(StringRef, void *, MCModuleYAML::Operand &);
+};
+
+template <> struct ScalarTraits<MCModuleYAML::OpcodeEnum> {
+  static void output(const MCModuleYAML::OpcodeEnum &, void *,
+                     llvm::raw_ostream &);
+  static StringRef input(StringRef, void *, MCModuleYAML::OpcodeEnum &);
+};
+
+void ScalarEnumerationTraits<MCAtom::AtomKind>::enumeration(
+    IO &IO, MCAtom::AtomKind &Value) {
+  IO.enumCase(Value, "Text", MCAtom::TextAtom);
+  IO.enumCase(Value, "Data", MCAtom::DataAtom);
+}
+
+void MappingTraits<MCModuleYAML::Atom>::mapping(IO &IO, MCModuleYAML::Atom &A) {
+  IO.mapRequired("StartAddress", A.StartAddress);
+  IO.mapRequired("Size", A.Size);
+  IO.mapRequired("Type", A.Type);
+  if (A.Type == MCAtom::TextAtom)
+    IO.mapRequired("Content", A.Insts);
+  else if (A.Type == MCAtom::DataAtom)
+    IO.mapRequired("Content", A.Data);
+}
+
+void MappingTraits<MCModuleYAML::Inst>::mapping(IO &IO, MCModuleYAML::Inst &I) {
+  IO.mapRequired("Inst", I.Opcode);
+  IO.mapRequired("Size", I.Size);
+  IO.mapRequired("Ops", I.Operands);
+}
+
+void
+MappingTraits<MCModuleYAML::BasicBlock>::mapping(IO &IO,
+                                                 MCModuleYAML::BasicBlock &BB) {
+  IO.mapRequired("Address", BB.Address);
+  IO.mapRequired("Preds", BB.Preds);
+  IO.mapRequired("Succs", BB.Succs);
+}
+
+void MappingTraits<MCModuleYAML::Function>::mapping(IO &IO,
+                                                    MCModuleYAML::Function &F) {
+  IO.mapRequired("Name", F.Name);
+  IO.mapRequired("BasicBlocks", F.BasicBlocks);
+}
+
+void MappingTraits<MCModuleYAML::Module>::mapping(IO &IO,
+                                                  MCModuleYAML::Module &M) {
+  IO.mapRequired("Atoms", M.Atoms);
+  IO.mapOptional("Functions", M.Functions);
+}
+
+void
+ScalarTraits<MCModuleYAML::Operand>::output(const MCModuleYAML::Operand &Val,
+                                            void *Ctx, raw_ostream &Out) {
+  InstrRegInfoHolder *IRI = (InstrRegInfoHolder *)Ctx;
+
+  // FIXME: Doesn't support FPImm and expr/inst, but do these make sense?
+  if (Val.MCOp.isImm())
+    Out << "I" << Val.MCOp.getImm();
+  else if (Val.MCOp.isReg())
+    Out << "R" << IRI->MRI.getName(Val.MCOp.getReg());
+  else
+    llvm_unreachable("Trying to output invalid MCOperand!");
+}
+
+StringRef
+ScalarTraits<MCModuleYAML::Operand>::input(StringRef Scalar, void *Ctx,
+                                           MCModuleYAML::Operand &Val) {
+  InstrRegInfoHolder *IRI = (InstrRegInfoHolder *)Ctx;
+  char Type = 0;
+  if (Scalar.size() >= 1)
+    Type = Scalar.front();
+  if (Type != 'R' && Type != 'I')
+    return "Operand must start with 'R' (register) or 'I' (immediate).";
+  if (Type == 'R') {
+    unsigned Reg;
+    if (!IRI->matchRegister(Scalar.substr(1), Reg))
+      return "Invalid register name.";
+    Val.MCOp = MCOperand::CreateReg(Reg);
+  } else if (Type == 'I') {
+    int64_t RIVal;
+    if (Scalar.substr(1).getAsInteger(10, RIVal))
+      return "Invalid immediate value.";
+    Val.MCOp = MCOperand::CreateImm(RIVal);
+  } else {
+    Val.MCOp = MCOperand();
+  }
+  return StringRef();
+}
+
+void ScalarTraits<MCModuleYAML::OpcodeEnum>::output(
+    const MCModuleYAML::OpcodeEnum &Val, void *Ctx, raw_ostream &Out) {
+  InstrRegInfoHolder *IRI = (InstrRegInfoHolder *)Ctx;
+  Out << IRI->MII.getName(Val);
+}
+
+StringRef
+ScalarTraits<MCModuleYAML::OpcodeEnum>::input(StringRef Scalar, void *Ctx,
+                                              MCModuleYAML::OpcodeEnum &Val) {
+  InstrRegInfoHolder *IRI = (InstrRegInfoHolder *)Ctx;
+  unsigned Opc;
+  if (!IRI->matchOpcode(Scalar, Opc))
+    return "Invalid instruction opcode.";
+  Val = Opc;
+  return "";
+}
+
+} // end namespace yaml
+
+namespace {
+
+class MCModule2YAML {
+  const MCModule &MCM;
+  MCModuleYAML::Module YAMLModule;
+  void dumpAtom(const MCAtom *MCA);
+  void dumpFunction(const MCFunction *MCF);
+  void dumpBasicBlock(const MCBasicBlock *MCBB);
+
+public:
+  MCModule2YAML(const MCModule &MCM);
+  MCModuleYAML::Module &getYAMLModule();
+};
+
+class YAML2MCModule {
+  MCModule &MCM;
+
+public:
+  YAML2MCModule(MCModule &MCM);
+  StringRef parse(const MCModuleYAML::Module &YAMLModule);
+};
+
+} // end unnamed namespace
+
+MCModule2YAML::MCModule2YAML(const MCModule &MCM) : MCM(MCM), YAMLModule() {
+  for (MCModule::const_atom_iterator AI = MCM.atom_begin(), AE = MCM.atom_end();
+       AI != AE; ++AI)
+    dumpAtom(*AI);
+  for (MCModule::const_func_iterator FI = MCM.func_begin(), FE = MCM.func_end();
+       FI != FE; ++FI)
+    dumpFunction(*FI);
+}
+
+void MCModule2YAML::dumpAtom(const MCAtom *MCA) {
+  YAMLModule.Atoms.resize(YAMLModule.Atoms.size() + 1);
+  MCModuleYAML::Atom &A = YAMLModule.Atoms.back();
+  A.Type = MCA->getKind();
+  A.StartAddress = MCA->getBeginAddr();
+  A.Size = MCA->getEndAddr() - MCA->getBeginAddr() + 1;
+  if (const MCTextAtom *TA = dyn_cast<MCTextAtom>(MCA)) {
+    const size_t InstCount = TA->size();
+    A.Insts.resize(InstCount);
+    for (size_t i = 0; i != InstCount; ++i) {
+      const MCDecodedInst &MCDI = TA->at(i);
+      A.Insts[i].Opcode = MCDI.Inst.getOpcode();
+      A.Insts[i].Size = MCDI.Size;
+      const unsigned OpCount = MCDI.Inst.getNumOperands();
+      A.Insts[i].Operands.resize(OpCount);
+      for (unsigned oi = 0; oi != OpCount; ++oi)
+        A.Insts[i].Operands[oi].MCOp = MCDI.Inst.getOperand(oi);
+    }
+  } else if (const MCDataAtom *DA = dyn_cast<MCDataAtom>(MCA)) {
+    A.Data = DA->getData();
+  } else {
+    llvm_unreachable("Unknown atom type.");
+  }
+}
+
+void MCModule2YAML::dumpFunction(const MCFunction *MCF) {
+  YAMLModule.Functions.resize(YAMLModule.Functions.size() + 1);
+  MCModuleYAML::Function &F = YAMLModule.Functions.back();
+  F.Name = MCF->getName();
+  for (MCFunction::const_iterator BBI = MCF->begin(), BBE = MCF->end();
+       BBI != BBE; ++BBI) {
+    const MCBasicBlock *MCBB = *BBI;
+    F.BasicBlocks.resize(F.BasicBlocks.size() + 1);
+    MCModuleYAML::BasicBlock &BB = F.BasicBlocks.back();
+    BB.Address = MCBB->getInsts()->getBeginAddr();
+    for (MCBasicBlock::pred_const_iterator PI = MCBB->pred_begin(),
+                                           PE = MCBB->pred_end();
+         PI != PE; ++PI)
+      BB.Preds.push_back((*PI)->getInsts()->getBeginAddr());
+    for (MCBasicBlock::succ_const_iterator SI = MCBB->succ_begin(),
+                                           SE = MCBB->succ_end();
+         SI != SE; ++SI)
+      BB.Succs.push_back((*SI)->getInsts()->getBeginAddr());
+  }
+}
+
+MCModuleYAML::Module &MCModule2YAML::getYAMLModule() { return YAMLModule; }
+
+YAML2MCModule::YAML2MCModule(MCModule &MCM) : MCM(MCM) {}
+
+StringRef YAML2MCModule::parse(const MCModuleYAML::Module &YAMLModule) {
+  typedef std::vector<MCModuleYAML::Atom>::const_iterator AtomIt;
+  typedef std::vector<MCModuleYAML::Inst>::const_iterator InstIt;
+  typedef std::vector<MCModuleYAML::Operand>::const_iterator OpIt;
+
+  typedef DenseMap<uint64_t, MCTextAtom *> AddrToTextAtomTy;
+  AddrToTextAtomTy TAByAddr;
+
+  for (AtomIt AI = YAMLModule.Atoms.begin(), AE = YAMLModule.Atoms.end();
+       AI != AE; ++AI) {
+    uint64_t StartAddress = AI->StartAddress;
+    if (AI->Size == 0)
+      return "Atoms can't be empty!";
+    uint64_t EndAddress = StartAddress + AI->Size - 1;
+    switch (AI->Type) {
+    case MCAtom::TextAtom: {
+      MCTextAtom *TA = MCM.createTextAtom(StartAddress, EndAddress);
+      TAByAddr[StartAddress] = TA;
+      for (InstIt II = AI->Insts.begin(), IE = AI->Insts.end(); II != IE;
+           ++II) {
+        MCInst MI;
+        MI.setOpcode(II->Opcode);
+        for (OpIt OI = II->Operands.begin(), OE = II->Operands.end(); OI != OE;
+             ++OI)
+          MI.addOperand(OI->MCOp);
+        TA->addInst(MI, II->Size);
+      }
+      break;
+    }
+    case MCAtom::DataAtom: {
+      MCDataAtom *DA = MCM.createDataAtom(StartAddress, EndAddress);
+      SmallVector<char, 64> Data;
+      raw_svector_ostream OS(Data);
+      AI->Data.writeAsBinary(OS);
+      OS.flush();
+      for (size_t i = 0, e = Data.size(); i != e; ++i)
+        DA->addData((uint8_t)Data[i]);
+      break;
+    }
+    }
+  }
+
+  typedef std::vector<MCModuleYAML::Function>::const_iterator FuncIt;
+  typedef std::vector<MCModuleYAML::BasicBlock>::const_iterator BBIt;
+  typedef std::vector<yaml::Hex64>::const_iterator AddrIt;
+  for (FuncIt FI = YAMLModule.Functions.begin(),
+              FE = YAMLModule.Functions.end();
+       FI != FE; ++FI) {
+    MCFunction *MCFN = MCM.createFunction(FI->Name);
+    for (BBIt BBI = FI->BasicBlocks.begin(), BBE = FI->BasicBlocks.end();
+         BBI != BBE; ++BBI) {
+      AddrToTextAtomTy::const_iterator It = TAByAddr.find(BBI->Address);
+      if (It == TAByAddr.end())
+        return "Basic block start address doesn't match any text atom!";
+      MCFN->createBlock(*It->second);
+    }
+    for (BBIt BBI = FI->BasicBlocks.begin(), BBE = FI->BasicBlocks.end();
+         BBI != BBE; ++BBI) {
+      MCBasicBlock *MCBB = MCFN->find(BBI->Address);
+      if (!MCBB)
+        return "Couldn't find matching basic block in function.";
+      for (AddrIt PI = BBI->Preds.begin(), PE = BBI->Preds.end(); PI != PE;
+           ++PI) {
+        MCBasicBlock *Pred = MCFN->find(*PI);
+        if (!Pred)
+          return "Couldn't find predecessor basic block.";
+        MCBB->addPredecessor(Pred);
+      }
+      for (AddrIt SI = BBI->Succs.begin(), SE = BBI->Succs.end(); SI != SE;
+           ++SI) {
+        MCBasicBlock *Succ = MCFN->find(*SI);
+        if (!Succ)
+          return "Couldn't find predecessor basic block.";
+        MCBB->addSuccessor(Succ);
+      }
+    }
+  }
+  return "";
+}
+
+StringRef mcmodule2yaml(raw_ostream &OS, const MCModule &MCM,
+                        const MCInstrInfo &MII, const MCRegisterInfo &MRI) {
+  MCModule2YAML Dumper(MCM);
+  InstrRegInfoHolder IRI(MII, MRI);
+  yaml::Output YOut(OS, (void *)&IRI);
+  YOut << Dumper.getYAMLModule();
+  return "";
+}
+
+StringRef yaml2mcmodule(OwningPtr<MCModule> &MCM, StringRef YamlContent,
+                        const MCInstrInfo &MII, const MCRegisterInfo &MRI) {
+  MCM.reset(new MCModule);
+  YAML2MCModule Parser(*MCM);
+  MCModuleYAML::Module YAMLModule;
+  InstrRegInfoHolder IRI(MII, MRI);
+  yaml::Input YIn(YamlContent, (void *)&IRI);
+  YIn >> YAMLModule;
+  if (error_code ec = YIn.error())
+    return ec.message();
+  StringRef err = Parser.parse(YAMLModule);
+  if (!err.empty())
+    return err;
+  return "";
+}
+
+} // end namespace llvm