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, 199 insertions, 71 deletions

diff --git a/contrib/llvm/lib/Target/X86/X86MCInstLower.cpp b/contrib/llvm/lib/Target/X86/X86MCInstLower.cpp
index a8a9fd8..6649c82 100644
--- a/contrib/llvm/lib/Target/X86/X86MCInstLower.cpp
+++ b/contrib/llvm/lib/Target/X86/X86MCInstLower.cpp
@@ -17,6 +17,7 @@
 #include "X86COFFMachineModuleInfo.h"
 #include "llvm/ADT/SmallString.h"
 #include "llvm/CodeGen/MachineModuleInfoImpls.h"
+#include "llvm/CodeGen/StackMaps.h"
 #include "llvm/IR/Type.h"
 #include "llvm/MC/MCAsmInfo.h"
 #include "llvm/MC/MCContext.h"
@@ -34,14 +35,12 @@ namespace {
 /// X86MCInstLower - This class is used to lower an MachineInstr into an MCInst.
 class X86MCInstLower {
   MCContext &Ctx;
-  Mangler *Mang;
   const MachineFunction &MF;
   const TargetMachine &TM;
   const MCAsmInfo &MAI;
   X86AsmPrinter &AsmPrinter;
 public:
-  X86MCInstLower(Mangler *mang, const MachineFunction &MF,
-                 X86AsmPrinter &asmprinter);
+  X86MCInstLower(const MachineFunction &MF, X86AsmPrinter &asmprinter);
 
   void Lower(const MachineInstr *MI, MCInst &OutMI) const;
 
@@ -50,13 +49,16 @@ public:
 
 private:
   MachineModuleInfoMachO &getMachOMMI() const;
+  Mangler *getMang() const {
+    return AsmPrinter.Mang;
+  }
 };
 
 } // end anonymous namespace
 
-X86MCInstLower::X86MCInstLower(Mangler *mang, const MachineFunction &mf,
+X86MCInstLower::X86MCInstLower(const MachineFunction &mf,
                                X86AsmPrinter &asmprinter)
-: Ctx(mf.getContext()), Mang(mang), MF(mf), TM(mf.getTarget()),
+: Ctx(mf.getContext()), MF(mf), TM(mf.getTarget()),
   MAI(*TM.getMCAsmInfo()), AsmPrinter(asmprinter) {}
 
 MachineModuleInfoMachO &X86MCInstLower::getMachOMMI() const {
@@ -81,7 +83,7 @@ GetSymbolFromOperand(const MachineOperand &MO) const {
         MO.getTargetFlags() == X86II::MO_DARWIN_HIDDEN_NONLAZY_PIC_BASE)
       isImplicitlyPrivate = true;
 
-    Mang->getNameWithPrefix(Name, GV, isImplicitlyPrivate);
+    getMang()->getNameWithPrefix(Name, GV, isImplicitlyPrivate);
   } else if (MO.isSymbol()) {
     Name += MAI.getGlobalPrefix();
     Name += MO.getSymbolName();
@@ -110,7 +112,7 @@ GetSymbolFromOperand(const MachineOperand &MO) const {
       assert(MO.isGlobal() && "Extern symbol not handled yet");
       StubSym =
         MachineModuleInfoImpl::
-        StubValueTy(Mang->getSymbol(MO.getGlobal()),
+        StubValueTy(AsmPrinter.getSymbol(MO.getGlobal()),
                     !MO.getGlobal()->hasInternalLinkage());
     }
     return Sym;
@@ -124,7 +126,7 @@ GetSymbolFromOperand(const MachineOperand &MO) const {
       assert(MO.isGlobal() && "Extern symbol not handled yet");
       StubSym =
         MachineModuleInfoImpl::
-        StubValueTy(Mang->getSymbol(MO.getGlobal()),
+        StubValueTy(AsmPrinter.getSymbol(MO.getGlobal()),
                     !MO.getGlobal()->hasInternalLinkage());
     }
     return Sym;
@@ -140,7 +142,7 @@ GetSymbolFromOperand(const MachineOperand &MO) const {
     if (MO.isGlobal()) {
       StubSym =
         MachineModuleInfoImpl::
-        StubValueTy(Mang->getSymbol(MO.getGlobal()),
+        StubValueTy(AsmPrinter.getSymbol(MO.getGlobal()),
                     !MO.getGlobal()->hasInternalLinkage());
     } else {
       Name.erase(Name.end()-5, Name.end());
@@ -225,32 +227,6 @@ MCOperand X86MCInstLower::LowerSymbolOperand(const MachineOperand &MO,
 }
 
 
-
-static void lower_subreg32(MCInst *MI, unsigned OpNo) {
-  // Convert registers in the addr mode according to subreg32.
-  unsigned Reg = MI->getOperand(OpNo).getReg();
-  if (Reg != 0)
-    MI->getOperand(OpNo).setReg(getX86SubSuperRegister(Reg, MVT::i32));
-}
-
-static void lower_lea64_32mem(MCInst *MI, unsigned OpNo) {
-  // Convert registers in the addr mode according to subreg64.
-  for (unsigned i = 0; i != 4; ++i) {
-    if (!MI->getOperand(OpNo+i).isReg()) continue;
-
-    unsigned Reg = MI->getOperand(OpNo+i).getReg();
-    // LEAs can use RIP-relative addressing, and RIP has no sub/super register.
-    if (Reg == 0 || Reg == X86::RIP) continue;
-
-    MI->getOperand(OpNo+i).setReg(getX86SubSuperRegister(Reg, MVT::i64));
-  }
-}
-
-/// LowerSubReg32_Op0 - Things like MOVZX16rr8 -> MOVZX32rr8.
-static void LowerSubReg32_Op0(MCInst &OutMI, unsigned NewOpc) {
-  OutMI.setOpcode(NewOpc);
-  lower_subreg32(&OutMI, 0);
-}
 /// LowerUnaryToTwoAddr - R = setb   -> R = sbb R, R
 static void LowerUnaryToTwoAddr(MCInst &OutMI, unsigned NewOpc) {
   OutMI.setOpcode(NewOpc);
@@ -280,6 +256,34 @@ static void SimplifyShortImmForm(MCInst &Inst, unsigned Opcode) {
   Inst.addOperand(Saved);
 }
 
+/// \brief If a movsx instruction has a shorter encoding for the used register
+/// simplify the instruction to use it instead.
+static void SimplifyMOVSX(MCInst &Inst) {
+  unsigned NewOpcode = 0;
+  unsigned Op0 = Inst.getOperand(0).getReg(), Op1 = Inst.getOperand(1).getReg();
+  switch (Inst.getOpcode()) {
+  default:
+    llvm_unreachable("Unexpected instruction!");
+  case X86::MOVSX16rr8:  // movsbw %al, %ax   --> cbtw
+    if (Op0 == X86::AX && Op1 == X86::AL)
+      NewOpcode = X86::CBW;
+    break;
+  case X86::MOVSX32rr16: // movswl %ax, %eax  --> cwtl
+    if (Op0 == X86::EAX && Op1 == X86::AX)
+      NewOpcode = X86::CWDE;
+    break;
+  case X86::MOVSX64rr32: // movslq %eax, %rax --> cltq
+    if (Op0 == X86::RAX && Op1 == X86::EAX)
+      NewOpcode = X86::CDQE;
+    break;
+  }
+
+  if (NewOpcode != 0) {
+    Inst = MCInst();
+    Inst.setOpcode(NewOpcode);
+  }
+}
+
 /// \brief Simplify things like MOV32rm to MOV32o32a.
 static void SimplifyShortMoveForm(X86AsmPrinter &Printer, MCInst &Inst,
                                   unsigned Opcode) {
@@ -376,9 +380,7 @@ void X86MCInstLower::Lower(const MachineInstr *MI, MCInst &OutMI) const {
   // Handle a few special cases to eliminate operand modifiers.
 ReSimplify:
   switch (OutMI.getOpcode()) {
-  case X86::LEA64_32r: // Handle 'subreg rewriting' for the lea64_32mem operand.
-    lower_lea64_32mem(&OutMI, 1);
-    // FALL THROUGH.
+  case X86::LEA64_32r:
   case X86::LEA64r:
   case X86::LEA16r:
   case X86::LEA32r:
@@ -388,23 +390,10 @@ ReSimplify:
     assert(OutMI.getOperand(1+X86::AddrSegmentReg).getReg() == 0 &&
            "LEA has segment specified!");
     break;
-  case X86::MOVZX64rr32:  LowerSubReg32_Op0(OutMI, X86::MOV32rr); break;
-  case X86::MOVZX64rm32:  LowerSubReg32_Op0(OutMI, X86::MOV32rm); break;
-  case X86::MOV64ri64i32: LowerSubReg32_Op0(OutMI, X86::MOV32ri); break;
-  case X86::MOVZX64rr8:   LowerSubReg32_Op0(OutMI, X86::MOVZX32rr8); break;
-  case X86::MOVZX64rm8:   LowerSubReg32_Op0(OutMI, X86::MOVZX32rm8); break;
-  case X86::MOVZX64rr16:  LowerSubReg32_Op0(OutMI, X86::MOVZX32rr16); break;
-  case X86::MOVZX64rm16:  LowerSubReg32_Op0(OutMI, X86::MOVZX32rm16); break;
-  case X86::MOV8r0:       LowerUnaryToTwoAddr(OutMI, X86::XOR8rr); break;
   case X86::MOV32r0:      LowerUnaryToTwoAddr(OutMI, X86::XOR32rr); break;
 
-  case X86::MOV16r0:
-    LowerSubReg32_Op0(OutMI, X86::MOV32r0);   // MOV16r0 -> MOV32r0
-    LowerUnaryToTwoAddr(OutMI, X86::XOR32rr); // MOV32r0 -> XOR32rr
-    break;
-  case X86::MOV64r0:
-    LowerSubReg32_Op0(OutMI, X86::MOV32r0);   // MOV64r0 -> MOV32r0
-    LowerUnaryToTwoAddr(OutMI, X86::XOR32rr); // MOV32r0 -> XOR32rr
+  case X86::MOV32ri64:
+    OutMI.setOpcode(X86::MOV32ri);
     break;
 
   // Commute operands to get a smaller encoding by using VEX.R instead of VEX.B
@@ -598,16 +587,11 @@ ReSimplify:
   case X86::XOR32ri:    SimplifyShortImmForm(OutMI, X86::XOR32i32);  break;
   case X86::XOR64ri32:  SimplifyShortImmForm(OutMI, X86::XOR64i32);  break;
 
-  case X86::MORESTACK_RET:
-    OutMI.setOpcode(X86::RET);
-    break;
-
-  case X86::MORESTACK_RET_RESTORE_R10:
-    OutMI.setOpcode(X86::MOV64rr);
-    OutMI.addOperand(MCOperand::CreateReg(X86::R10));
-    OutMI.addOperand(MCOperand::CreateReg(X86::RAX));
-
-    AsmPrinter.OutStreamer.EmitInstruction(MCInstBuilder(X86::RET));
+  // Try to shrink some forms of movsx.
+  case X86::MOVSX16rr8:
+  case X86::MOVSX32rr16:
+  case X86::MOVSX64rr32:
+    SimplifyMOVSX(OutMI);
     break;
   }
 }
@@ -691,17 +675,143 @@ static void LowerTlsAddr(MCStreamer &OutStreamer,
     .addExpr(tlsRef));
 }
 
+static std::pair<StackMaps::Location, MachineInstr::const_mop_iterator>
+parseMemoryOperand(StackMaps::Location::LocationType LocTy, unsigned Size,
+                   MachineInstr::const_mop_iterator MOI,
+                   MachineInstr::const_mop_iterator MOE) {
+
+  typedef StackMaps::Location Location;
+
+  assert(std::distance(MOI, MOE) >= 5 && "Too few operands to encode mem op.");
+
+  const MachineOperand &Base = *MOI;
+  const MachineOperand &Scale = *(++MOI);
+  const MachineOperand &Index = *(++MOI);
+  const MachineOperand &Disp = *(++MOI);
+  const MachineOperand &ZeroReg = *(++MOI);
+
+  // Sanity check for supported operand format.
+  assert(Base.isReg() &&
+         Scale.isImm() && Scale.getImm() == 1 &&
+         Index.isReg() && Index.getReg() == 0 &&
+         Disp.isImm() && ZeroReg.isReg() && (ZeroReg.getReg() == 0) &&
+         "Unsupported x86 memory operand sequence.");
+  (void)Scale;
+  (void)Index;
+  (void)ZeroReg;
+
+  return std::make_pair(
+    Location(LocTy, Size, Base.getReg(), Disp.getImm()), ++MOI);
+}
+
+std::pair<StackMaps::Location, MachineInstr::const_mop_iterator>
+X86AsmPrinter::stackmapOperandParser(MachineInstr::const_mop_iterator MOI,
+                                     MachineInstr::const_mop_iterator MOE,
+                                     const TargetMachine &TM) {
+
+  typedef StackMaps::Location Location;
+
+  const MachineOperand &MOP = *MOI;
+  assert(!MOP.isRegMask() && (!MOP.isReg() || !MOP.isImplicit()) &&
+         "Register mask and implicit operands should not be processed.");
+
+  if (MOP.isImm()) {
+    // Verify anyregcc
+    // [<def>], <id>, <numBytes>, <target>, <numArgs>, <cc>, ...
+
+    switch (MOP.getImm()) {
+    default: llvm_unreachable("Unrecognized operand type.");
+    case StackMaps::DirectMemRefOp: {
+      unsigned Size = TM.getDataLayout()->getPointerSizeInBits();
+      assert((Size % 8) == 0 && "Need pointer size in bytes.");
+      Size /= 8;
+      return parseMemoryOperand(StackMaps::Location::Direct, Size,
+                                llvm::next(MOI), MOE);
+    }
+    case StackMaps::IndirectMemRefOp: {
+      ++MOI;
+      int64_t Size = MOI->getImm();
+      assert(Size > 0 && "Need a valid size for indirect memory locations.");
+      return parseMemoryOperand(StackMaps::Location::Indirect, Size,
+                                llvm::next(MOI), MOE);
+    }
+    case StackMaps::ConstantOp: {
+      ++MOI;
+      assert(MOI->isImm() && "Expected constant operand.");
+      int64_t Imm = MOI->getImm();
+      return std::make_pair(
+        Location(Location::Constant, sizeof(int64_t), 0, Imm), ++MOI);
+    }
+    }
+  }
+
+  // Otherwise this is a reg operand. The physical register number will
+  // ultimately be encoded as a DWARF regno. The stack map also records the size
+  // of a spill slot that can hold the register content. (The runtime can
+  // track the actual size of the data type if it needs to.)
+  assert(MOP.isReg() && "Expected register operand here.");
+  assert(TargetRegisterInfo::isPhysicalRegister(MOP.getReg()) &&
+         "Virtreg operands should have been rewritten before now.");
+  const TargetRegisterClass *RC =
+    TM.getRegisterInfo()->getMinimalPhysRegClass(MOP.getReg());
+  assert(!MOP.getSubReg() && "Physical subreg still around.");
+  return std::make_pair(
+    Location(Location::Register, RC->getSize(), MOP.getReg(), 0), ++MOI);
+}
+
+// Lower a stackmap of the form:
+// <id>, <shadowBytes>, ...
+static void LowerSTACKMAP(MCStreamer &OutStreamer,
+                          StackMaps &SM,
+                          const MachineInstr &MI)
+{
+  unsigned NumNOPBytes = MI.getOperand(1).getImm();
+  SM.recordStackMap(MI);
+  // Emit padding.
+  // FIXME: These nops ensure that the stackmap's shadow is covered by
+  // instructions from the same basic block, but the nops should not be
+  // necessary if instructions from the same block follow the stackmap.
+  for (unsigned i = 0; i < NumNOPBytes; ++i)
+    OutStreamer.EmitInstruction(MCInstBuilder(X86::NOOP));
+}
+
+// Lower a patchpoint of the form:
+// [<def>], <id>, <numBytes>, <target>, <numArgs>, <cc>, ...
+static void LowerPATCHPOINT(MCStreamer &OutStreamer,
+                            StackMaps &SM,
+                            const MachineInstr &MI) {
+  SM.recordPatchPoint(MI);
+
+  PatchPointOpers opers(&MI);
+  unsigned ScratchIdx = opers.getNextScratchIdx();
+  unsigned EncodedBytes = 0;
+  int64_t CallTarget = opers.getMetaOper(PatchPointOpers::TargetPos).getImm();
+  if (CallTarget) {
+    // Emit MOV to materialize the target address and the CALL to target.
+    // This is encoded with 12-13 bytes, depending on which register is used.
+    // We conservatively assume that it is 12 bytes and emit in worst case one
+    // extra NOP byte.
+    EncodedBytes = 12;
+    OutStreamer.EmitInstruction(MCInstBuilder(X86::MOV64ri)
+                                .addReg(MI.getOperand(ScratchIdx).getReg())
+                                .addImm(CallTarget));
+    OutStreamer.EmitInstruction(MCInstBuilder(X86::CALL64r)
+                                .addReg(MI.getOperand(ScratchIdx).getReg()));
+  }
+  // Emit padding.
+  unsigned NumBytes = opers.getMetaOper(PatchPointOpers::NBytesPos).getImm();
+  assert(NumBytes >= EncodedBytes &&
+         "Patchpoint can't request size less than the length of a call.");
+
+  for (unsigned i = EncodedBytes; i < NumBytes; ++i)
+    OutStreamer.EmitInstruction(MCInstBuilder(X86::NOOP));
+}
+
 void X86AsmPrinter::EmitInstruction(const MachineInstr *MI) {
-  X86MCInstLower MCInstLowering(Mang, *MF, *this);
+  X86MCInstLower MCInstLowering(*MF, *this);
   switch (MI->getOpcode()) {
   case TargetOpcode::DBG_VALUE:
-    if (isVerbose() && OutStreamer.hasRawTextSupport()) {
-      std::string TmpStr;
-      raw_string_ostream OS(TmpStr);
-      PrintDebugValueComment(MI, OS);
-      OutStreamer.EmitRawText(StringRef(OS.str()));
-    }
-    return;
+    llvm_unreachable("Should be handled target independently");
 
   // Emit nothing here but a comment if we can.
   case X86::Int_MemBarrier:
@@ -786,6 +896,24 @@ void X86AsmPrinter::EmitInstruction(const MachineInstr *MI) {
       .addExpr(DotExpr));
     return;
   }
+
+  case TargetOpcode::STACKMAP:
+    return LowerSTACKMAP(OutStreamer, SM, *MI);
+
+  case TargetOpcode::PATCHPOINT:
+    return LowerPATCHPOINT(OutStreamer, SM, *MI);
+
+  case X86::MORESTACK_RET:
+    OutStreamer.EmitInstruction(MCInstBuilder(X86::RET));
+    return;
+
+  case X86::MORESTACK_RET_RESTORE_R10:
+    // Return, then restore R10.
+    OutStreamer.EmitInstruction(MCInstBuilder(X86::RET));
+    OutStreamer.EmitInstruction(MCInstBuilder(X86::MOV64rr)
+      .addReg(X86::R10)
+      .addReg(X86::RAX));
+    return;
   }
 
   MCInst TmpInst;