Update clang, llvm, lld, lldb, compiler-rt and libc++ to 4.0.0 release:

MFC r309142 (by emaste):

Add WITH_LLD_AS_LD build knob

If set it installs LLD as /usr/bin/ld.  LLD (as of version 3.9) is not
capable of linking the world and kernel, but can self-host and link many
substantial applications. GNU ld continues to be used for the world and
kernel build, regardless of how this knob is set.

It is on by default for arm64, and off for all other CPU architectures.

Sponsored by:	The FreeBSD Foundation

MFC r310840:

Reapply 310775, now it also builds correctly if lldb is disabled:

Move llvm-objdump from CLANG_EXTRAS to installed by default

We currently install three tools from binutils 2.17.50: as, ld, and
objdump. Work is underway to migrate to a permissively-licensed
tool-chain, with one goal being the retirement of binutils 2.17.50.

LLVM's llvm-objdump is intended to be compatible with GNU objdump
although it is currently missing some options and may have formatting
differences. Enable it by default for testing and further investigation.
It may later be changed to install as /usr/bin/objdump, it becomes a
fully viable replacement.

Reviewed by:	emaste
Differential Revision:	https://reviews.freebsd.org/D8879

MFC r312855 (by emaste):

Rename LLD_AS_LD to LLD_IS_LD, for consistency with CLANG_IS_CC

Reported by:	Dan McGregor <dan.mcgregor usask.ca>

MFC r313559 | glebius | 2017-02-10 18:34:48 +0100 (Fri, 10 Feb 2017) | 5 lines

Don't check struct rtentry on FreeBSD, it is an internal kernel structure.
On other systems it may be API structure for SIOCADDRT/SIOCDELRT.

Reviewed by:	emaste, dim

MFC r314152 (by jkim):

Remove an assembler flag, which is redundant since r309124.  The upstream
took care of it by introducing a macro NO_EXEC_STACK_DIRECTIVE.

http://llvm.org/viewvc/llvm-project?rev=273500&view=rev

Reviewed by:	dim

MFC r314564:

Upgrade our copies of clang, llvm, lld, lldb, compiler-rt and libc++ to
4.0.0 (branches/release_40 296509).  The release will follow soon.

Please note that from 3.5.0 onwards, clang, llvm and lldb require C++11
support to build; see UPDATING for more information.

Also note that as of 4.0.0, lld should be able to link the base system
on amd64 and aarch64.  See the WITH_LLD_IS_LLD setting in src.conf(5).
Though please be aware that this is work in progress.

Release notes for llvm, clang and lld will be available here:
<http://releases.llvm.org/4.0.0/docs/ReleaseNotes.html>
<http://releases.llvm.org/4.0.0/tools/clang/docs/ReleaseNotes.html>
<http://releases.llvm.org/4.0.0/tools/lld/docs/ReleaseNotes.html>

Thanks to Ed Maste, Jan Beich, Antoine Brodin and Eric Fiselier for
their help.

Relnotes:	yes
Exp-run:	antoine
PR:		215969, 216008

MFC r314708:

For now, revert r287232 from upstream llvm trunk (by Daniil Fukalov):

  [SCEV] limit recursion depth of CompareSCEVComplexity

  Summary:
  CompareSCEVComplexity goes too deep (50+ on a quite a big unrolled
  loop) and runs almost infinite time.

  Added cache of "equal" SCEV pairs to earlier cutoff of further
  estimation. Recursion depth limit was also introduced as a parameter.

  Reviewers: sanjoy

  Subscribers: mzolotukhin, tstellarAMD, llvm-commits

  Differential Revision: https://reviews.llvm.org/D26389

This commit is the cause of excessive compile times on skein_block.c
(and possibly other files) during kernel builds on amd64.

We never saw the problematic behavior described in this upstream commit,
so for now it is better to revert it.  An upstream bug has been filed
here: https://bugs.llvm.org/show_bug.cgi?id=32142

Reported by:	mjg

MFC r314795:

Reapply r287232 from upstream llvm trunk (by Daniil Fukalov):

  [SCEV] limit recursion depth of CompareSCEVComplexity

  Summary:
  CompareSCEVComplexity goes too deep (50+ on a quite a big unrolled
  loop) and runs almost infinite time.

  Added cache of "equal" SCEV pairs to earlier cutoff of further
  estimation. Recursion depth limit was also introduced as a parameter.

  Reviewers: sanjoy

  Subscribers: mzolotukhin, tstellarAMD, llvm-commits

  Differential Revision: https://reviews.llvm.org/D26389

Pull in r296992 from upstream llvm trunk (by Sanjoy Das):

  [SCEV] Decrease the recursion threshold for CompareValueComplexity

  Fixes PR32142.

  r287232 accidentally increased the recursion threshold for
  CompareValueComplexity from 2 to 32.  This change reverses that
  change by introducing a separate flag for CompareValueComplexity's
  threshold.

The latter revision fixes the excessive compile times for skein_block.c.

MFC r314907 | mmel | 2017-03-08 12:40:27 +0100 (Wed, 08 Mar 2017) | 7 lines

Unbreak ARMv6 world.

The new compiler_rt library imported with clang 4.0.0 have several fatal
issues (non-functional __udivsi3 for example) with ARM specific instrict
functions. As temporary workaround, until upstream solve these problems,
disable all thumb[1][2] related feature.

MFC r315016:

Update clang, llvm, lld, lldb, compiler-rt and libc++ to 4.0.0 release.
We were already very close to the last release candidate, so this is a
pretty minor update.

Relnotes:	yes

MFC r316005:

Revert r314907, and pull in r298713 from upstream compiler-rt trunk (by
Weiming Zhao):

  builtins: Select correct code fragments when compiling for Thumb1/Thum2/ARM ISA.

  Summary:
  Value of __ARM_ARCH_ISA_THUMB isn't based on the actual compilation
  mode (-mthumb, -marm), it reflect's capability of given CPU.

  Due to this:
   - use  __tbumb__ and __thumb2__ insteand of __ARM_ARCH_ISA_THUMB
   - use '.thumb' directive consistently  in all affected files
   - decorate all thumb functions using
     DEFINE_COMPILERRT_THUMB_FUNCTION()

  ---------
  Note: This patch doesn't fix broken Thumb1 variant of __udivsi3 !

  Reviewers: weimingz, rengolin, compnerd

  Subscribers: aemerson, dim

  Differential Revision: https://reviews.llvm.org/D30938

Discussed with:	mmel

1 files changed, 192 insertions, 132 deletions

diff --git a/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp b/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
index 23c9352..5bf347e 100644
--- a/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
+++ b/contrib/llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
@@ -12,25 +12,48 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "AMDGPU.h"
 #include "AMDGPUInstrInfo.h"
-#include "AMDGPUIntrinsicInfo.h"
+#include "AMDGPURegisterInfo.h"
 #include "AMDGPUISelLowering.h" // For AMDGPUISD
 #include "AMDGPUSubtarget.h"
+#include "SIDefines.h"
+#include "SIInstrInfo.h"
+#include "SIRegisterInfo.h"
 #include "SIISelLowering.h"
 #include "SIMachineFunctionInfo.h"
+#include "llvm/ADT/APInt.h"
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/StringRef.h"
 #include "llvm/Analysis/ValueTracking.h"
 #include "llvm/CodeGen/FunctionLoweringInfo.h"
-#include "llvm/CodeGen/MachineFrameInfo.h"
-#include "llvm/CodeGen/PseudoSourceValue.h"
+#include "llvm/CodeGen/ISDOpcodes.h"
+#include "llvm/CodeGen/MachineFunction.h"
+#include "llvm/CodeGen/MachineRegisterInfo.h"
+#include "llvm/CodeGen/MachineValueType.h"
 #include "llvm/CodeGen/SelectionDAG.h"
 #include "llvm/CodeGen/SelectionDAGISel.h"
-#include "llvm/IR/DiagnosticInfo.h"
+#include "llvm/CodeGen/SelectionDAGNodes.h"
+#include "llvm/CodeGen/ValueTypes.h"
+#include "llvm/IR/BasicBlock.h"
+#include "llvm/IR/Instruction.h"
+#include "llvm/MC/MCInstrDesc.h"
+#include "llvm/Support/Casting.h"
+#include "llvm/Support/CodeGen.h"
+#include "llvm/Support/ErrorHandling.h"
+#include "llvm/Support/MathExtras.h"
+#include <cassert>
+#include <cstdint>
+#include <new>
+#include <vector>
 
 using namespace llvm;
 
 namespace llvm {
+
 class R600InstrInfo;
-}
+
+} // end namespace llvm
 
 //===----------------------------------------------------------------------===//
 // Instruction Selector Implementation
@@ -38,18 +61,6 @@ class R600InstrInfo;
 
 namespace {
 
-static bool isCBranchSCC(const SDNode *N) {
-  assert(N->getOpcode() == ISD::BRCOND);
-  if (!N->hasOneUse())
-    return false;
-
-  SDValue Cond = N->getOperand(1);
-  if (Cond.getOpcode() == ISD::CopyToReg)
-    Cond = Cond.getOperand(2);
-  return Cond.getOpcode() == ISD::SETCC &&
-         Cond.getOperand(0).getValueType() == MVT::i32 && Cond.hasOneUse();
-}
-
 /// AMDGPU specific code to select AMDGPU machine instructions for
 /// SelectionDAG operations.
 class AMDGPUDAGToDAGISel : public SelectionDAGISel {
@@ -58,16 +69,18 @@ class AMDGPUDAGToDAGISel : public SelectionDAGISel {
   const AMDGPUSubtarget *Subtarget;
 
 public:
-  AMDGPUDAGToDAGISel(TargetMachine &TM);
-  virtual ~AMDGPUDAGToDAGISel();
+  explicit AMDGPUDAGToDAGISel(TargetMachine &TM, CodeGenOpt::Level OptLevel)
+      : SelectionDAGISel(TM, OptLevel) {}
+  ~AMDGPUDAGToDAGISel() override = default;
+
   bool runOnMachineFunction(MachineFunction &MF) override;
   void Select(SDNode *N) override;
-  const char *getPassName() const override;
-  void PreprocessISelDAG() override;
+  StringRef getPassName() const override;
   void PostprocessISelDAG() override;
 
 private:
-  bool isInlineImmediate(SDNode *N) const;
+  SDValue foldFrameIndex(SDValue N) const;
+  bool isInlineImmediate(const SDNode *N) const;
   bool FoldOperand(SDValue &Src, SDValue &Sel, SDValue &Neg, SDValue &Abs,
                    const R600InstrInfo *TII);
   bool FoldOperands(unsigned, const R600InstrInfo *, std::vector<SDValue> &);
@@ -145,40 +158,46 @@ private:
 
   void SelectADD_SUB_I64(SDNode *N);
   void SelectDIV_SCALE(SDNode *N);
+  void SelectFMA_W_CHAIN(SDNode *N);
+  void SelectFMUL_W_CHAIN(SDNode *N);
 
   SDNode *getS_BFE(unsigned Opcode, const SDLoc &DL, SDValue Val,
                    uint32_t Offset, uint32_t Width);
   void SelectS_BFEFromShifts(SDNode *N);
   void SelectS_BFE(SDNode *N);
+  bool isCBranchSCC(const SDNode *N) const;
   void SelectBRCOND(SDNode *N);
   void SelectATOMIC_CMP_SWAP(SDNode *N);
 
   // Include the pieces autogenerated from the target description.
 #include "AMDGPUGenDAGISel.inc"
 };
+
 }  // end anonymous namespace
 
 /// \brief This pass converts a legalized DAG into a AMDGPU-specific
 // DAG, ready for instruction scheduling.
-FunctionPass *llvm::createAMDGPUISelDag(TargetMachine &TM) {
-  return new AMDGPUDAGToDAGISel(TM);
+FunctionPass *llvm::createAMDGPUISelDag(TargetMachine &TM,
+                                        CodeGenOpt::Level OptLevel) {
+  return new AMDGPUDAGToDAGISel(TM, OptLevel);
 }
 
-AMDGPUDAGToDAGISel::AMDGPUDAGToDAGISel(TargetMachine &TM)
-    : SelectionDAGISel(TM) {}
-
 bool AMDGPUDAGToDAGISel::runOnMachineFunction(MachineFunction &MF) {
   Subtarget = &MF.getSubtarget<AMDGPUSubtarget>();
   return SelectionDAGISel::runOnMachineFunction(MF);
 }
 
-AMDGPUDAGToDAGISel::~AMDGPUDAGToDAGISel() {
-}
+bool AMDGPUDAGToDAGISel::isInlineImmediate(const SDNode *N) const {
+  const SIInstrInfo *TII
+    = static_cast<const SISubtarget *>(Subtarget)->getInstrInfo();
+
+  if (const ConstantSDNode *C = dyn_cast<ConstantSDNode>(N))
+    return TII->isInlineConstant(C->getAPIntValue());
+
+  if (const ConstantFPSDNode *C = dyn_cast<ConstantFPSDNode>(N))
+    return TII->isInlineConstant(C->getValueAPF().bitcastToAPInt());
 
-bool AMDGPUDAGToDAGISel::isInlineImmediate(SDNode *N) const {
-  const SITargetLowering *TL
-      = static_cast<const SITargetLowering *>(getTargetLowering());
-  return TL->analyzeImmediate(N) == 0;
+  return false;
 }
 
 /// \brief Determine the register class for \p OpNo
@@ -187,8 +206,21 @@ bool AMDGPUDAGToDAGISel::isInlineImmediate(SDNode *N) const {
 /// determined.
 const TargetRegisterClass *AMDGPUDAGToDAGISel::getOperandRegClass(SDNode *N,
                                                           unsigned OpNo) const {
-  if (!N->isMachineOpcode())
+  if (!N->isMachineOpcode()) {
+    if (N->getOpcode() == ISD::CopyToReg) {
+      unsigned Reg = cast<RegisterSDNode>(N->getOperand(1))->getReg();
+      if (TargetRegisterInfo::isVirtualRegister(Reg)) {
+        MachineRegisterInfo &MRI = CurDAG->getMachineFunction().getRegInfo();
+        return MRI.getRegClass(Reg);
+      }
+
+      const SIRegisterInfo *TRI
+        = static_cast<const SISubtarget *>(Subtarget)->getRegisterInfo();
+      return TRI->getPhysRegClass(Reg);
+    }
+
     return nullptr;
+  }
 
   switch (N->getMachineOpcode()) {
   default: {
@@ -244,7 +276,7 @@ SDNode *AMDGPUDAGToDAGISel::glueCopyToM0(SDNode *N) const {
 static unsigned selectSGPRVectorRegClassID(unsigned NumVectorElts) {
   switch (NumVectorElts) {
   case 1:
-    return AMDGPU::SReg_32RegClassID;
+    return AMDGPU::SReg_32_XM0RegClassID;
   case 2:
     return AMDGPU::SReg_64RegClassID;
   case 4:
@@ -275,7 +307,11 @@ void AMDGPUDAGToDAGISel::Select(SDNode *N) {
   // DAG legalization, so we can fold some i64 ADDs used for address
   // calculation into the LOAD and STORE instructions.
   case ISD::ADD:
-  case ISD::SUB: {
+  case ISD::ADDC:
+  case ISD::ADDE:
+  case ISD::SUB:
+  case ISD::SUBC:
+  case ISD::SUBE: {
     if (N->getValueType(0) != MVT::i64 ||
         Subtarget->getGeneration() < AMDGPUSubtarget::SOUTHERN_ISLANDS)
       break;
@@ -283,6 +319,15 @@ void AMDGPUDAGToDAGISel::Select(SDNode *N) {
     SelectADD_SUB_I64(N);
     return;
   }
+  case AMDGPUISD::FMUL_W_CHAIN: {
+    SelectFMUL_W_CHAIN(N);
+    return;
+  }
+  case AMDGPUISD::FMA_W_CHAIN: {
+    SelectFMA_W_CHAIN(N);
+    return;
+  }
+
   case ISD::SCALAR_TO_VECTOR:
   case AMDGPUISD::BUILD_VERTICAL_VECTOR:
   case ISD::BUILD_VECTOR: {
@@ -498,7 +543,7 @@ bool AMDGPUDAGToDAGISel::isUniformBr(const SDNode *N) const {
          Term->getMetadata("structurizecfg.uniform");
 }
 
-const char *AMDGPUDAGToDAGISel::getPassName() const {
+StringRef AMDGPUDAGToDAGISel::getPassName() const {
   return "AMDGPU DAG->DAG Pattern Instruction Selection";
 }
 
@@ -563,6 +608,10 @@ bool AMDGPUDAGToDAGISel::SelectADDRIndirect(SDValue Addr, SDValue &Base,
   if ((C = dyn_cast<ConstantSDNode>(Addr))) {
     Base = CurDAG->getRegister(AMDGPU::INDIRECT_BASE_ADDR, MVT::i32);
     Offset = CurDAG->getTargetConstant(C->getZExtValue(), DL, MVT::i32);
+  } else if ((Addr.getOpcode() == AMDGPUISD::DWORDADDR) &&
+             (C = dyn_cast<ConstantSDNode>(Addr.getOperand(0)))) {
+    Base = CurDAG->getRegister(AMDGPU::INDIRECT_BASE_ADDR, MVT::i32);
+    Offset = CurDAG->getTargetConstant(C->getZExtValue(), DL, MVT::i32);
   } else if ((Addr.getOpcode() == ISD::ADD || Addr.getOpcode() == ISD::OR) &&
             (C = dyn_cast<ConstantSDNode>(Addr.getOperand(1)))) {
     Base = Addr.getOperand(0);
@@ -580,7 +629,12 @@ void AMDGPUDAGToDAGISel::SelectADD_SUB_I64(SDNode *N) {
   SDValue LHS = N->getOperand(0);
   SDValue RHS = N->getOperand(1);
 
-  bool IsAdd = (N->getOpcode() == ISD::ADD);
+  unsigned Opcode = N->getOpcode();
+  bool ConsumeCarry = (Opcode == ISD::ADDE || Opcode == ISD::SUBE);
+  bool ProduceCarry =
+      ConsumeCarry || Opcode == ISD::ADDC || Opcode == ISD::SUBC;
+  bool IsAdd =
+      (Opcode == ISD::ADD || Opcode == ISD::ADDC || Opcode == ISD::ADDE);
 
   SDValue Sub0 = CurDAG->getTargetConstant(AMDGPU::sub0, DL, MVT::i32);
   SDValue Sub1 = CurDAG->getTargetConstant(AMDGPU::sub1, DL, MVT::i32);
@@ -596,25 +650,70 @@ void AMDGPUDAGToDAGISel::SelectADD_SUB_I64(SDNode *N) {
                                        DL, MVT::i32, RHS, Sub1);
 
   SDVTList VTList = CurDAG->getVTList(MVT::i32, MVT::Glue);
-  SDValue AddLoArgs[] = { SDValue(Lo0, 0), SDValue(Lo1, 0) };
 
   unsigned Opc = IsAdd ? AMDGPU::S_ADD_U32 : AMDGPU::S_SUB_U32;
   unsigned CarryOpc = IsAdd ? AMDGPU::S_ADDC_U32 : AMDGPU::S_SUBB_U32;
 
-  SDNode *AddLo = CurDAG->getMachineNode( Opc, DL, VTList, AddLoArgs);
-  SDValue Carry(AddLo, 1);
-  SDNode *AddHi
-    = CurDAG->getMachineNode(CarryOpc, DL, MVT::i32,
-                             SDValue(Hi0, 0), SDValue(Hi1, 0), Carry);
+  SDNode *AddLo;
+  if (!ConsumeCarry) {
+    SDValue Args[] = { SDValue(Lo0, 0), SDValue(Lo1, 0) };
+    AddLo = CurDAG->getMachineNode(Opc, DL, VTList, Args);
+  } else {
+    SDValue Args[] = { SDValue(Lo0, 0), SDValue(Lo1, 0), N->getOperand(2) };
+    AddLo = CurDAG->getMachineNode(CarryOpc, DL, VTList, Args);
+  }
+  SDValue AddHiArgs[] = {
+    SDValue(Hi0, 0),
+    SDValue(Hi1, 0),
+    SDValue(AddLo, 1)
+  };
+  SDNode *AddHi = CurDAG->getMachineNode(CarryOpc, DL, VTList, AddHiArgs);
 
-  SDValue Args[5] = {
+  SDValue RegSequenceArgs[] = {
     CurDAG->getTargetConstant(AMDGPU::SReg_64RegClassID, DL, MVT::i32),
     SDValue(AddLo,0),
     Sub0,
     SDValue(AddHi,0),
     Sub1,
   };
-  CurDAG->SelectNodeTo(N, AMDGPU::REG_SEQUENCE, MVT::i64, Args);
+  SDNode *RegSequence = CurDAG->getMachineNode(AMDGPU::REG_SEQUENCE, DL,
+                                               MVT::i64, RegSequenceArgs);
+
+  if (ProduceCarry) {
+    // Replace the carry-use
+    CurDAG->ReplaceAllUsesOfValueWith(SDValue(N, 1), SDValue(AddHi, 1));
+  }
+
+  // Replace the remaining uses.
+  CurDAG->ReplaceAllUsesWith(N, RegSequence);
+  CurDAG->RemoveDeadNode(N);
+}
+
+void AMDGPUDAGToDAGISel::SelectFMA_W_CHAIN(SDNode *N) {
+  SDLoc SL(N);
+  //  src0_modifiers, src0,  src1_modifiers, src1, src2_modifiers, src2, clamp, omod
+  SDValue Ops[10];
+
+  SelectVOP3Mods0(N->getOperand(1), Ops[1], Ops[0], Ops[6], Ops[7]);
+  SelectVOP3Mods(N->getOperand(2), Ops[3], Ops[2]);
+  SelectVOP3Mods(N->getOperand(3), Ops[5], Ops[4]);
+  Ops[8] = N->getOperand(0);
+  Ops[9] = N->getOperand(4);
+
+  CurDAG->SelectNodeTo(N, AMDGPU::V_FMA_F32, N->getVTList(), Ops);
+}
+
+void AMDGPUDAGToDAGISel::SelectFMUL_W_CHAIN(SDNode *N) {
+  SDLoc SL(N);
+  //	src0_modifiers, src0,  src1_modifiers, src1, clamp, omod
+  SDValue Ops[8];
+
+  SelectVOP3Mods0(N->getOperand(1), Ops[1], Ops[0], Ops[4], Ops[5]);
+  SelectVOP3Mods(N->getOperand(2), Ops[3], Ops[2]);
+  Ops[6] = N->getOperand(0);
+  Ops[7] = N->getOperand(3);
+
+  CurDAG->SelectNodeTo(N, AMDGPU::V_MUL_F32_e64, N->getVTList(), Ops);
 }
 
 // We need to handle this here because tablegen doesn't support matching
@@ -628,14 +727,8 @@ void AMDGPUDAGToDAGISel::SelectDIV_SCALE(SDNode *N) {
   unsigned Opc
     = (VT == MVT::f64) ? AMDGPU::V_DIV_SCALE_F64 : AMDGPU::V_DIV_SCALE_F32;
 
-  // src0_modifiers, src0, src1_modifiers, src1, src2_modifiers, src2, clamp,
-  // omod
-  SDValue Ops[8];
-
-  SelectVOP3Mods0(N->getOperand(0), Ops[1], Ops[0], Ops[6], Ops[7]);
-  SelectVOP3Mods(N->getOperand(1), Ops[3], Ops[2]);
-  SelectVOP3Mods(N->getOperand(2), Ops[5], Ops[4]);
-  CurDAG->SelectNodeTo(N, Opc, VT, MVT::i1, Ops);
+  SDValue Ops[] = { N->getOperand(0), N->getOperand(1), N->getOperand(2) };
+  CurDAG->SelectNodeTo(N, Opc, N->getVTList(), Ops);
 }
 
 bool AMDGPUDAGToDAGISel::isDSOffsetLegal(const SDValue &Base, unsigned Offset,
@@ -779,6 +872,9 @@ bool AMDGPUDAGToDAGISel::SelectDS64Bit4ByteAligned(SDValue Addr, SDValue &Base,
   }
 
   // default case
+
+  // FIXME: This is broken on SI where we still need to check if the base
+  // pointer is positive here.
   Base = Addr;
   Offset0 = CurDAG->getTargetConstant(0, DL, MVT::i8);
   Offset1 = CurDAG->getTargetConstant(1, DL, MVT::i8);
@@ -825,7 +921,6 @@ bool AMDGPUDAGToDAGISel::SelectMUBUF(SDValue Addr, SDValue &Ptr,
       Ptr = N2;
       VAddr = N3;
     } else {
-
       // (add N0, C1) -> offset
       VAddr = CurDAG->getTargetConstant(0, DL, MVT::i32);
       Ptr = N0;
@@ -903,6 +998,12 @@ bool AMDGPUDAGToDAGISel::SelectMUBUFAddr64(SDValue Addr, SDValue &SRsrc,
   return SelectMUBUFAddr64(Addr, SRsrc, VAddr, SOffset, Offset, GLC, SLC, TFE);
 }
 
+SDValue AMDGPUDAGToDAGISel::foldFrameIndex(SDValue N) const {
+  if (auto FI = dyn_cast<FrameIndexSDNode>(N))
+    return CurDAG->getTargetFrameIndex(FI->getIndex(), FI->getValueType(0));
+  return N;
+}
+
 bool AMDGPUDAGToDAGISel::SelectMUBUFScratch(SDValue Addr, SDValue &Rsrc,
                                             SDValue &VAddr, SDValue &SOffset,
                                             SDValue &ImmOffset) const {
@@ -922,14 +1023,14 @@ bool AMDGPUDAGToDAGISel::SelectMUBUFScratch(SDValue Addr, SDValue &Rsrc,
     // Offsets in vaddr must be positive.
     ConstantSDNode *C1 = cast<ConstantSDNode>(N1);
     if (isLegalMUBUFImmOffset(C1)) {
-      VAddr = N0;
+      VAddr = foldFrameIndex(N0);
       ImmOffset = CurDAG->getTargetConstant(C1->getZExtValue(), DL, MVT::i16);
       return true;
     }
   }
 
   // (node)
-  VAddr = Addr;
+  VAddr = foldFrameIndex(Addr);
   ImmOffset = CurDAG->getTargetConstant(0, DL, MVT::i16);
   return true;
 }
@@ -1122,7 +1223,6 @@ bool AMDGPUDAGToDAGISel::SelectSMRDOffset(SDValue ByteOffsetNode,
 
 bool AMDGPUDAGToDAGISel::SelectSMRD(SDValue Addr, SDValue &SBase,
                                      SDValue &Offset, bool &Imm) const {
-
   SDLoc SL(Addr);
   if (CurDAG->isBaseWithConstantOffset(Addr)) {
     SDValue N0 = Addr.getOperand(0);
@@ -1327,36 +1427,53 @@ void AMDGPUDAGToDAGISel::SelectS_BFE(SDNode *N) {
   SelectCode(N);
 }
 
+bool AMDGPUDAGToDAGISel::isCBranchSCC(const SDNode *N) const {
+  assert(N->getOpcode() == ISD::BRCOND);
+  if (!N->hasOneUse())
+    return false;
+
+  SDValue Cond = N->getOperand(1);
+  if (Cond.getOpcode() == ISD::CopyToReg)
+    Cond = Cond.getOperand(2);
+
+  if (Cond.getOpcode() != ISD::SETCC || !Cond.hasOneUse())
+    return false;
+
+  MVT VT = Cond.getOperand(0).getSimpleValueType();
+  if (VT == MVT::i32)
+    return true;
+
+  if (VT == MVT::i64) {
+    auto ST = static_cast<const SISubtarget *>(Subtarget);
+
+    ISD::CondCode CC = cast<CondCodeSDNode>(Cond.getOperand(2))->get();
+    return (CC == ISD::SETEQ || CC == ISD::SETNE) && ST->hasScalarCompareEq64();
+  }
+
+  return false;
+}
+
 void AMDGPUDAGToDAGISel::SelectBRCOND(SDNode *N) {
   SDValue Cond = N->getOperand(1);
 
+  if (Cond.isUndef()) {
+    CurDAG->SelectNodeTo(N, AMDGPU::SI_BR_UNDEF, MVT::Other,
+                         N->getOperand(2), N->getOperand(0));
+    return;
+  }
+
   if (isCBranchSCC(N)) {
     // This brcond will use S_CBRANCH_SCC*, so let tablegen handle it.
     SelectCode(N);
     return;
   }
 
-  // The result of VOPC instructions is or'd against ~EXEC before it is
-  // written to vcc or another SGPR.  This means that the value '1' is always
-  // written to the corresponding bit for results that are masked.  In order
-  // to correctly check against vccz, we need to and VCC with the EXEC
-  // register in order to clear the value from the masked bits.
-
   SDLoc SL(N);
 
-  SDNode *MaskedCond =
-        CurDAG->getMachineNode(AMDGPU::S_AND_B64, SL, MVT::i1,
-                               CurDAG->getRegister(AMDGPU::EXEC, MVT::i1),
-                               Cond);
-  SDValue VCC = CurDAG->getCopyToReg(N->getOperand(0), SL, AMDGPU::VCC,
-                                     SDValue(MaskedCond, 0),
-                                     SDValue()); // Passing SDValue() adds a
-                                                 // glue output.
+  SDValue VCC = CurDAG->getCopyToReg(N->getOperand(0), SL, AMDGPU::VCC, Cond);
   CurDAG->SelectNodeTo(N, AMDGPU::S_CBRANCH_VCCNZ, MVT::Other,
                        N->getOperand(2), // Basic Block
-                       VCC.getValue(0),  // Chain
-                       VCC.getValue(1)); // Glue
-  return;
+                       VCC.getValue(0));
 }
 
 // This is here because there isn't a way to use the generated sub0_sub1 as the
@@ -1427,7 +1544,6 @@ void AMDGPUDAGToDAGISel::SelectATOMIC_CMP_SWAP(SDNode *N) {
 
 bool AMDGPUDAGToDAGISel::SelectVOP3Mods(SDValue In, SDValue &Src,
                                         SDValue &SrcMods) const {
-
   unsigned Mods = 0;
 
   Src = In;
@@ -1491,62 +1607,6 @@ bool AMDGPUDAGToDAGISel::SelectVOP3Mods0Clamp0OMod(SDValue In, SDValue &Src,
   return SelectVOP3Mods(In, Src, SrcMods);
 }
 
-void AMDGPUDAGToDAGISel::PreprocessISelDAG() {
-  MachineFrameInfo *MFI = CurDAG->getMachineFunction().getFrameInfo();
-
-  // Handle the perverse case where a frame index is being stored. We don't
-  // want to see multiple frame index operands on the same instruction since
-  // it complicates things and violates some assumptions about frame index
-  // lowering.
-  for (int I = MFI->getObjectIndexBegin(), E = MFI->getObjectIndexEnd();
-       I != E; ++I) {
-    SDValue FI = CurDAG->getTargetFrameIndex(I, MVT::i32);
-
-    // It's possible that we have a frame index defined in the function that
-    // isn't used in this block.
-    if (FI.use_empty())
-      continue;
-
-    // Skip over the AssertZext inserted during lowering.
-    SDValue EffectiveFI = FI;
-    auto It = FI->use_begin();
-    if (It->getOpcode() == ISD::AssertZext && FI->hasOneUse()) {
-      EffectiveFI = SDValue(*It, 0);
-      It = EffectiveFI->use_begin();
-    }
-
-    for (auto It = EffectiveFI->use_begin(); !It.atEnd(); ) {
-      SDUse &Use = It.getUse();
-      SDNode *User = Use.getUser();
-      unsigned OpIdx = It.getOperandNo();
-      ++It;
-
-      if (MemSDNode *M = dyn_cast<MemSDNode>(User)) {
-        unsigned PtrIdx = M->getOpcode() == ISD::STORE ? 2 : 1;
-        if (OpIdx == PtrIdx)
-          continue;
-
-        unsigned OpN = M->getNumOperands();
-        SDValue NewOps[8];
-
-        assert(OpN < array_lengthof(NewOps));
-        for (unsigned Op = 0; Op != OpN; ++Op) {
-          if (Op != OpIdx) {
-            NewOps[Op] = M->getOperand(Op);
-            continue;
-          }
-
-          MachineSDNode *Mov = CurDAG->getMachineNode(AMDGPU::V_MOV_B32_e32,
-                                                      SDLoc(M), MVT::i32, FI);
-          NewOps[Op] = SDValue(Mov, 0);
-        }
-
-        CurDAG->UpdateNodeOperands(M, makeArrayRef(NewOps, OpN));
-      }
-    }
-  }
-}
-
 void AMDGPUDAGToDAGISel::PostprocessISelDAG() {
   const AMDGPUTargetLowering& Lowering =
     *static_cast<const AMDGPUTargetLowering*>(getTargetLowering());