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, 41 insertions, 31 deletions

diff --git a/contrib/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp b/contrib/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp
index 491ccc4..5ae93284 100644
--- a/contrib/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp
+++ b/contrib/llvm/lib/Target/Hexagon/HexagonISelDAGToDAG.cpp
@@ -50,15 +50,13 @@ class HexagonDAGToDAGISel : public SelectionDAGISel {
 
   // Keep a reference to HexagonTargetMachine.
   const HexagonTargetMachine& TM;
-  const HexagonInstrInfo *TII;
   DenseMap<const GlobalValue *, unsigned> GlobalAddressUseCountMap;
 public:
-  explicit HexagonDAGToDAGISel(const HexagonTargetMachine &targetmachine,
+  explicit HexagonDAGToDAGISel(HexagonTargetMachine &targetmachine,
                                CodeGenOpt::Level OptLevel)
     : SelectionDAGISel(targetmachine, OptLevel),
       Subtarget(targetmachine.getSubtarget<HexagonSubtarget>()),
-      TM(targetmachine),
-      TII(static_cast<const HexagonInstrInfo*>(TM.getInstrInfo())) {
+      TM(targetmachine) {
     initializeHexagonDAGToDAGISelPass(*PassRegistry::getPassRegistry());
   }
   bool hasNumUsesBelowThresGA(SDNode *N) const;
@@ -92,14 +90,14 @@ public:
   bool SelectAddr(SDNode *Op, SDValue Addr, SDValue &Base, SDValue &Offset);
 
   SDNode *SelectLoad(SDNode *N);
-  SDNode *SelectBaseOffsetLoad(LoadSDNode *LD, DebugLoc dl);
-  SDNode *SelectIndexedLoad(LoadSDNode *LD, DebugLoc dl);
+  SDNode *SelectBaseOffsetLoad(LoadSDNode *LD, SDLoc dl);
+  SDNode *SelectIndexedLoad(LoadSDNode *LD, SDLoc dl);
   SDNode *SelectIndexedLoadZeroExtend64(LoadSDNode *LD, unsigned Opcode,
-                                        DebugLoc dl);
+                                        SDLoc dl);
   SDNode *SelectIndexedLoadSignExtend64(LoadSDNode *LD, unsigned Opcode,
-                                        DebugLoc dl);
-  SDNode *SelectBaseOffsetStore(StoreSDNode *ST, DebugLoc dl);
-  SDNode *SelectIndexedStore(StoreSDNode *ST, DebugLoc dl);
+                                        SDLoc dl);
+  SDNode *SelectBaseOffsetStore(StoreSDNode *ST, SDLoc dl);
+  SDNode *SelectIndexedStore(StoreSDNode *ST, SDLoc dl);
   SDNode *SelectStore(SDNode *N);
   SDNode *SelectSHL(SDNode *N);
   SDNode *SelectSelect(SDNode *N);
@@ -180,7 +178,7 @@ inline SDValue XformUToUM1Imm(unsigned Imm) {
 /// createHexagonISelDag - This pass converts a legalized DAG into a
 /// Hexagon-specific DAG, ready for instruction scheduling.
 ///
-FunctionPass *llvm::createHexagonISelDag(const HexagonTargetMachine &TM,
+FunctionPass *llvm::createHexagonISelDag(HexagonTargetMachine &TM,
                                          CodeGenOpt::Level OptLevel) {
   return new HexagonDAGToDAGISel(TM, OptLevel);
 }
@@ -385,7 +383,7 @@ static bool OffsetFitsS11(EVT MemType, int64_t Offset) {
 // lowering for GlobalAddress nodes has already turned it into a
 // CONST32.
 //
-SDNode *HexagonDAGToDAGISel::SelectBaseOffsetLoad(LoadSDNode *LD, DebugLoc dl) {
+SDNode *HexagonDAGToDAGISel::SelectBaseOffsetLoad(LoadSDNode *LD, SDLoc dl) {
   SDValue Chain = LD->getChain();
   SDNode* Const32 = LD->getBasePtr().getNode();
   unsigned Opcode = 0;
@@ -396,7 +394,7 @@ SDNode *HexagonDAGToDAGISel::SelectBaseOffsetLoad(LoadSDNode *LD, DebugLoc dl) {
     EVT LoadedVT = LD->getMemoryVT();
     int64_t Offset = cast<GlobalAddressSDNode>(Base)->getOffset();
     if (Offset != 0 && OffsetFitsS11(LoadedVT, Offset)) {
-      MVT PointerTy = TLI.getPointerTy();
+      MVT PointerTy = getTargetLowering()->getPointerTy();
       const GlobalValue* GV =
         cast<GlobalAddressSDNode>(Base)->getGlobal();
       SDValue TargAddr =
@@ -433,7 +431,7 @@ SDNode *HexagonDAGToDAGISel::SelectBaseOffsetLoad(LoadSDNode *LD, DebugLoc dl) {
 
 SDNode *HexagonDAGToDAGISel::SelectIndexedLoadSignExtend64(LoadSDNode *LD,
                                                            unsigned Opcode,
-                                                           DebugLoc dl)
+                                                           SDLoc dl)
 {
   SDValue Chain = LD->getChain();
   EVT LoadedVT = LD->getMemoryVT();
@@ -444,8 +442,11 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedLoadSignExtend64(LoadSDNode *LD,
   SDValue N1 = LD->getOperand(1);
   SDValue CPTmpN1_0;
   SDValue CPTmpN1_1;
+
   if (SelectADDRriS11_2(N1, CPTmpN1_0, CPTmpN1_1) &&
       N1.getNode()->getValueType(0) == MVT::i32) {
+    const HexagonInstrInfo *TII =
+      static_cast<const HexagonInstrInfo*>(TM.getInstrInfo());
     if (TII->isValidAutoIncImm(LoadedVT, Val)) {
       SDValue TargetConst = CurDAG->getTargetConstant(Val, MVT::i32);
       SDNode *Result_1 = CurDAG->getMachineNode(Opcode, dl, MVT::i32, MVT::i32,
@@ -497,7 +498,7 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedLoadSignExtend64(LoadSDNode *LD,
 
 SDNode *HexagonDAGToDAGISel::SelectIndexedLoadZeroExtend64(LoadSDNode *LD,
                                                            unsigned Opcode,
-                                                           DebugLoc dl)
+                                                           SDLoc dl)
 {
   SDValue Chain = LD->getChain();
   EVT LoadedVT = LD->getMemoryVT();
@@ -508,8 +509,11 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedLoadZeroExtend64(LoadSDNode *LD,
   SDValue N1 = LD->getOperand(1);
   SDValue CPTmpN1_0;
   SDValue CPTmpN1_1;
+
   if (SelectADDRriS11_2(N1, CPTmpN1_0, CPTmpN1_1) &&
       N1.getNode()->getValueType(0) == MVT::i32) {
+    const HexagonInstrInfo *TII =
+      static_cast<const HexagonInstrInfo*>(TM.getInstrInfo());
     if (TII->isValidAutoIncImm(LoadedVT, Val)) {
       SDValue TargetConstVal = CurDAG->getTargetConstant(Val, MVT::i32);
       SDValue TargetConst0 = CurDAG->getTargetConstant(0, MVT::i32);
@@ -572,7 +576,7 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedLoadZeroExtend64(LoadSDNode *LD,
 }
 
 
-SDNode *HexagonDAGToDAGISel::SelectIndexedLoad(LoadSDNode *LD, DebugLoc dl) {
+SDNode *HexagonDAGToDAGISel::SelectIndexedLoad(LoadSDNode *LD, SDLoc dl) {
   SDValue Chain = LD->getChain();
   SDValue Base = LD->getBasePtr();
   SDValue Offset = LD->getOffset();
@@ -586,6 +590,8 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedLoad(LoadSDNode *LD, DebugLoc dl) {
   bool zextval = (LD->getExtensionType() == ISD::ZEXTLOAD);
 
   // Figure out the opcode.
+  const HexagonInstrInfo *TII =
+    static_cast<const HexagonInstrInfo*>(TM.getInstrInfo());
   if (LoadedVT == MVT::i64) {
     if (TII->isValidAutoIncImm(LoadedVT, Val))
       Opcode = Hexagon::POST_LDrid;
@@ -667,7 +673,7 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedLoad(LoadSDNode *LD, DebugLoc dl) {
 
 SDNode *HexagonDAGToDAGISel::SelectLoad(SDNode *N) {
   SDNode *result;
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   LoadSDNode *LD = cast<LoadSDNode>(N);
   ISD::MemIndexedMode AM = LD->getAddressingMode();
 
@@ -682,7 +688,7 @@ SDNode *HexagonDAGToDAGISel::SelectLoad(SDNode *N) {
 }
 
 
-SDNode *HexagonDAGToDAGISel::SelectIndexedStore(StoreSDNode *ST, DebugLoc dl) {
+SDNode *HexagonDAGToDAGISel::SelectIndexedStore(StoreSDNode *ST, SDLoc dl) {
   SDValue Chain = ST->getChain();
   SDValue Base = ST->getBasePtr();
   SDValue Offset = ST->getOffset();
@@ -694,6 +700,8 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedStore(StoreSDNode *ST, DebugLoc dl) {
 
   // Offset value must be within representable range
   // and must have correct alignment properties.
+  const HexagonInstrInfo *TII =
+    static_cast<const HexagonInstrInfo*>(TM.getInstrInfo());
   if (TII->isValidAutoIncImm(StoredVT, Val)) {
     SDValue Ops[] = {Base, CurDAG->getTargetConstant(Val, MVT::i32), Value,
                      Chain};
@@ -751,7 +759,7 @@ SDNode *HexagonDAGToDAGISel::SelectIndexedStore(StoreSDNode *ST, DebugLoc dl) {
 
 
 SDNode *HexagonDAGToDAGISel::SelectBaseOffsetStore(StoreSDNode *ST,
-                                                   DebugLoc dl) {
+                                                   SDLoc dl) {
   SDValue Chain = ST->getChain();
   SDNode* Const32 = ST->getBasePtr().getNode();
   SDValue Value = ST->getValue();
@@ -769,7 +777,7 @@ SDNode *HexagonDAGToDAGISel::SelectBaseOffsetStore(StoreSDNode *ST,
       EVT StoredVT = ST->getMemoryVT();
       int64_t Offset = cast<GlobalAddressSDNode>(Base)->getOffset();
       if (Offset != 0 && OffsetFitsS11(StoredVT, Offset)) {
-        MVT PointerTy = TLI.getPointerTy();
+        MVT PointerTy = getTargetLowering()->getPointerTy();
         const GlobalValue* GV =
           cast<GlobalAddressSDNode>(Base)->getGlobal();
         SDValue TargAddr =
@@ -805,7 +813,7 @@ SDNode *HexagonDAGToDAGISel::SelectBaseOffsetStore(StoreSDNode *ST,
 
 
 SDNode *HexagonDAGToDAGISel::SelectStore(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   StoreSDNode *ST = cast<StoreSDNode>(N);
   ISD::MemIndexedMode AM = ST->getAddressingMode();
 
@@ -818,7 +826,7 @@ SDNode *HexagonDAGToDAGISel::SelectStore(SDNode *N) {
 }
 
 SDNode *HexagonDAGToDAGISel::SelectMul(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
 
   //
   // %conv.i = sext i32 %tmp1 to i64
@@ -902,7 +910,7 @@ SDNode *HexagonDAGToDAGISel::SelectMul(SDNode *N) {
 
 
 SDNode *HexagonDAGToDAGISel::SelectSelect(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   SDValue N0 = N->getOperand(0);
   if (N0.getOpcode() == ISD::SETCC) {
     SDValue N00 = N0.getOperand(0);
@@ -969,7 +977,7 @@ SDNode *HexagonDAGToDAGISel::SelectSelect(SDNode *N) {
 
 
 SDNode *HexagonDAGToDAGISel::SelectTruncate(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   SDValue Shift = N->getOperand(0);
 
   //
@@ -1082,7 +1090,7 @@ SDNode *HexagonDAGToDAGISel::SelectTruncate(SDNode *N) {
 
 
 SDNode *HexagonDAGToDAGISel::SelectSHL(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   if (N->getValueType(0) == MVT::i32) {
     SDValue Shl_0 = N->getOperand(0);
     SDValue Shl_1 = N->getOperand(1);
@@ -1158,7 +1166,7 @@ SDNode *HexagonDAGToDAGISel::SelectSHL(SDNode *N) {
 // We want to preserve all the lower 8-bits and, not just 1 LSB bit.
 //
 SDNode *HexagonDAGToDAGISel::SelectZeroExtend(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   SDNode *IsIntrinsic = N->getOperand(0).getNode();
   if ((IsIntrinsic->getOpcode() == ISD::INTRINSIC_WO_CHAIN)) {
     unsigned ID =
@@ -1201,7 +1209,7 @@ SDNode *HexagonDAGToDAGISel::SelectZeroExtend(SDNode *N) {
 // and lowering to the actual intrinsic.
 //
 SDNode *HexagonDAGToDAGISel::SelectIntrinsicWOChain(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   unsigned ID = cast<ConstantSDNode>(N->getOperand(0))->getZExtValue();
   unsigned IntrinsicWithPred = doesIntrinsicContainPredicate(ID);
 
@@ -1209,6 +1217,8 @@ SDNode *HexagonDAGToDAGISel::SelectIntrinsicWOChain(SDNode *N) {
   // as at least one of the operands.
   if (IntrinsicWithPred) {
     SmallVector<SDValue, 8> Ops;
+    const HexagonInstrInfo *TII =
+      static_cast<const HexagonInstrInfo*>(TM.getInstrInfo());
     const MCInstrDesc &MCID = TII->get(IntrinsicWithPred);
     const TargetRegisterInfo *TRI = TM.getRegisterInfo();
 
@@ -1251,7 +1261,7 @@ SDNode *HexagonDAGToDAGISel::SelectIntrinsicWOChain(SDNode *N) {
 // Map floating point constant values.
 //
 SDNode *HexagonDAGToDAGISel::SelectConstantFP(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   ConstantFPSDNode *CN = dyn_cast<ConstantFPSDNode>(N);
   APFloat APF = CN->getValueAPF();
   if (N->getValueType(0) == MVT::f32) {
@@ -1271,7 +1281,7 @@ SDNode *HexagonDAGToDAGISel::SelectConstantFP(SDNode *N) {
 // Map predicate true (encoded as -1 in LLVM) to a XOR.
 //
 SDNode *HexagonDAGToDAGISel::SelectConstant(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   if (N->getValueType(0) == MVT::i1) {
     SDNode* Result;
     int32_t Val = cast<ConstantSDNode>(N)->getSExtValue();
@@ -1310,7 +1320,7 @@ SDNode *HexagonDAGToDAGISel::SelectConstant(SDNode *N) {
 // Map add followed by a asr -> asr +=.
 //
 SDNode *HexagonDAGToDAGISel::SelectAdd(SDNode *N) {
-  DebugLoc dl = N->getDebugLoc();
+  SDLoc dl(N);
   if (N->getValueType(0) != MVT::i32) {
     return SelectCode(N);
   }
@@ -1662,7 +1672,7 @@ bool HexagonDAGToDAGISel::foldGlobalAddressImpl(SDValue &N, SDValue &R,
                 !hasNumUsesBelowThresGA(GA))
             return false;
         R = CurDAG->getTargetGlobalAddress(GA->getGlobal(),
-                                          Const->getDebugLoc(),
+                                          SDLoc(Const),
                                           N.getValueType(),
                                           GA->getOffset() +
                                           (uint64_t)Const->getSExtValue());