MFC r263312:

Pull in r196939 from upstream llvm trunk (by Reid Kleckner):

  Reland "Fix miscompile of MS inline assembly with stack realignment"

  This re-lands commit r196876, which was reverted in r196879.

  The tests have been fixed to pass on platforms with a stack alignment
  larger than 4.

  Update to clang side tests will land shortly.

Pull in r196986 from upstream llvm trunk (by Reid Kleckner):

  Revert the backend fatal error from r196939

  The combination of inline asm, stack realignment, and dynamic allocas
  turns out to be too common to reject out of hand.

  ASan inserts empy inline asm fragments and uses aligned allocas.
  Compiling any trivial function containing a dynamic alloca with ASan is
  enough to trigger the check.

  XFAIL the test cases that would be miscompiled and add one that uses the
  relevant functionality.

Pull in r202930 from upstream llvm trunk (by Hans Wennborg):

  Check for dynamic allocas and inline asm that clobbers sp before building
  selection dag (PR19012)

  In X86SelectionDagInfo::EmitTargetCodeForMemcpy we check with MachineFrameInfo
  to make sure that ESI isn't used as a base pointer register before we choose to
  emit rep movs (which clobbers esi).

  The problem is that MachineFrameInfo wouldn't know about dynamic allocas or
  inline asm that clobbers the stack pointer until SelectionDAGBuilder has
  encountered them.

  This patch fixes the problem by checking for such things when building the
  FunctionLoweringInfo.

  Differential Revision: http://llvm-reviews.chandlerc.com/D2954

Together, these commits fix the problem encountered in the devel/emacs
port on the i386 architecture, where a combination of stack realignment,
alloca() and memcpy() could incidentally clobber the %esi register,
leading to segfaults in the temacs build-time utility.

See also: http://llvm.org/PR18171 and http://llvm.org/PR19012

Reported by:	ashish
PR:		ports/183064

MFC r263313:

Pull in r203311 from upstream llvm trunk (by Arnold Schwaighofer):

  ISel: Make VSELECT selection terminate in cases where the condition type has to
  be split and the result type widened.

  When the condition of a vselect has to be split it makes no sense widening the
  vselect and thereby widening the condition. We end up in an endless loop of
  widening (vselect result type) and splitting (condition mask type) doing this.
  Instead, split both the condition and the vselect and widen the result.

  I ran this over the test suite with i686 and mattr=+sse and saw no regressions.

  Fixes PR18036.

With this fix the original problem case from the graphics/rawtherapee
port (posted in http://llvm.org/PR18036 ) now compiles within ~97MB RSS.

Reported by:	mandree

MFC r263320:

Add separate patch files for all the customizations we have currently
applied to our copy of llvm/clang.  These can be applied in alphabetical
order to a pristine llvm/clang 3.4 release source tree, to result in the
same version used in FreeBSD.

This is intended to clearly document all the changes until now, which
mostly consist of cherry pickings from the respective upstream trunks,
plus a number of hand-written FreeBSD-specific ones.  Hopefully those
can eventually be cleaned up and sent upstream too.

1 files changed, 213 insertions, 0 deletions

diff --git a/contrib/llvm/patches/patch-r262261-llvm-r198480-sparc.diff b/contrib/llvm/patches/patch-r262261-llvm-r198480-sparc.diff
new file mode 100644
index 0000000..b3e1362
--- /dev/null
+++ b/contrib/llvm/patches/patch-r262261-llvm-r198480-sparc.diff
@@ -0,0 +1,213 @@
+Pull in r198480 from upstream llvm trunk (by Venkatraman Govindaraju):
+
+  [SparcV9]: Implement RETURNADDR and FRAMEADDR lowering in SPARC64. 
+
+Introduced here: http://svn.freebsd.org/changeset/base/262261
+
+Index: lib/Target/Sparc/SparcISelLowering.cpp
+===================================================================
+--- lib/Target/Sparc/SparcISelLowering.cpp
++++ lib/Target/Sparc/SparcISelLowering.cpp
+@@ -2415,7 +2415,8 @@ static SDValue getFLUSHW(SDValue Op, SelectionDAG
+   return Chain;
+ }
+ 
+-static SDValue LowerFRAMEADDR(SDValue Op, SelectionDAG &DAG) {
++static SDValue getFRAMEADDR(uint64_t depth, SDValue Op, SelectionDAG &DAG,
++                            const SparcSubtarget *Subtarget) {
+   MachineFrameInfo *MFI = DAG.getMachineFunction().getFrameInfo();
+   MFI->setFrameAddressIsTaken(true);
+ 
+@@ -2422,32 +2423,49 @@ static SDValue getFLUSHW(SDValue Op, SelectionDAG
+   EVT VT = Op.getValueType();
+   SDLoc dl(Op);
+   unsigned FrameReg = SP::I6;
++  unsigned stackBias = Subtarget->getStackPointerBias();
+ 
+-  uint64_t depth = Op.getConstantOperandVal(0);
++  SDValue FrameAddr;
+ 
+-  SDValue FrameAddr;
+-  if (depth == 0)
++  if (depth == 0) {
+     FrameAddr = DAG.getCopyFromReg(DAG.getEntryNode(), dl, FrameReg, VT);
+-  else {
+-    // flush first to make sure the windowed registers' values are in stack
+-    SDValue Chain = getFLUSHW(Op, DAG);
+-    FrameAddr = DAG.getCopyFromReg(Chain, dl, FrameReg, VT);
++    if (Subtarget->is64Bit())
++      FrameAddr = DAG.getNode(ISD::ADD, dl, VT, FrameAddr,
++                              DAG.getIntPtrConstant(stackBias));
++    return FrameAddr;
++  }
+ 
+-    for (uint64_t i = 0; i != depth; ++i) {
+-      SDValue Ptr = DAG.getNode(ISD::ADD,
+-                                dl, MVT::i32,
+-                                FrameAddr, DAG.getIntPtrConstant(56));
+-      FrameAddr = DAG.getLoad(MVT::i32, dl,
+-                              Chain,
+-                              Ptr,
+-                              MachinePointerInfo(), false, false, false, 0);
+-    }
++  // flush first to make sure the windowed registers' values are in stack
++  SDValue Chain = getFLUSHW(Op, DAG);
++  FrameAddr = DAG.getCopyFromReg(Chain, dl, FrameReg, VT);
++
++  unsigned Offset = (Subtarget->is64Bit()) ? (stackBias + 112) : 56;
++
++  while (depth--) {
++    SDValue Ptr = DAG.getNode(ISD::ADD, dl, VT, FrameAddr,
++                              DAG.getIntPtrConstant(Offset));
++    FrameAddr = DAG.getLoad(VT, dl, Chain, Ptr, MachinePointerInfo(),
++                            false, false, false, 0);
+   }
++  if (Subtarget->is64Bit())
++    FrameAddr = DAG.getNode(ISD::ADD, dl, VT, FrameAddr,
++                            DAG.getIntPtrConstant(stackBias));
+   return FrameAddr;
+ }
+ 
++
++static SDValue LowerFRAMEADDR(SDValue Op, SelectionDAG &DAG,
++                              const SparcSubtarget *Subtarget) {
++
++  uint64_t depth = Op.getConstantOperandVal(0);
++
++  return getFRAMEADDR(depth, Op, DAG, Subtarget);
++
++}
++
+ static SDValue LowerRETURNADDR(SDValue Op, SelectionDAG &DAG,
+-                               const SparcTargetLowering &TLI) {
++                               const SparcTargetLowering &TLI,
++                               const SparcSubtarget *Subtarget) {
+   MachineFunction &MF = DAG.getMachineFunction();
+   MachineFrameInfo *MFI = MF.getFrameInfo();
+   MFI->setReturnAddressIsTaken(true);
+@@ -2461,25 +2479,20 @@ static SDValue LowerRETURNADDR(SDValue Op, Selecti
+     unsigned RetReg = MF.addLiveIn(SP::I7,
+                                    TLI.getRegClassFor(TLI.getPointerTy()));
+     RetAddr = DAG.getCopyFromReg(DAG.getEntryNode(), dl, RetReg, VT);
+-  } else {
+-    // Need frame address to find return address of the caller.
+-    MFI->setFrameAddressIsTaken(true);
++    return RetAddr;
++  }
+ 
+-    // flush first to make sure the windowed registers' values are in stack
+-    SDValue Chain = getFLUSHW(Op, DAG);
+-    RetAddr = DAG.getCopyFromReg(Chain, dl, SP::I6, VT);
++  // Need frame address to find return address of the caller.
++  SDValue FrameAddr = getFRAMEADDR(depth - 1, Op, DAG, Subtarget);
+ 
+-    for (uint64_t i = 0; i != depth; ++i) {
+-      SDValue Ptr = DAG.getNode(ISD::ADD,
+-                                dl, MVT::i32,
+-                                RetAddr,
+-                                DAG.getIntPtrConstant((i == depth-1)?60:56));
+-      RetAddr = DAG.getLoad(MVT::i32, dl,
+-                            Chain,
+-                            Ptr,
+-                            MachinePointerInfo(), false, false, false, 0);
+-    }
+-  }
++  unsigned Offset = (Subtarget->is64Bit()) ? 120 : 60;
++  SDValue Ptr = DAG.getNode(ISD::ADD,
++                            dl, VT,
++                            FrameAddr,
++                            DAG.getIntPtrConstant(Offset));
++  RetAddr = DAG.getLoad(VT, dl, DAG.getEntryNode(), Ptr,
++                        MachinePointerInfo(), false, false, false, 0);
++
+   return RetAddr;
+ }
+ 
+@@ -2763,8 +2776,10 @@ LowerOperation(SDValue Op, SelectionDAG &DAG) cons
+   switch (Op.getOpcode()) {
+   default: llvm_unreachable("Should not custom lower this!");
+ 
+-  case ISD::RETURNADDR:         return LowerRETURNADDR(Op, DAG, *this);
+-  case ISD::FRAMEADDR:          return LowerFRAMEADDR(Op, DAG);
++  case ISD::RETURNADDR:         return LowerRETURNADDR(Op, DAG, *this,
++                                                       Subtarget);
++  case ISD::FRAMEADDR:          return LowerFRAMEADDR(Op, DAG,
++                                                      Subtarget);
+   case ISD::GlobalTLSAddress:   return LowerGlobalTLSAddress(Op, DAG);
+   case ISD::GlobalAddress:      return LowerGlobalAddress(Op, DAG);
+   case ISD::BlockAddress:       return LowerBlockAddress(Op, DAG);
+Index: test/CodeGen/SPARC/2011-01-11-FrameAddr.ll
+===================================================================
+--- test/CodeGen/SPARC/2011-01-11-FrameAddr.ll
++++ test/CodeGen/SPARC/2011-01-11-FrameAddr.ll
+@@ -2,6 +2,7 @@
+ ;RUN: llc -march=sparc -mattr=v9 < %s | FileCheck %s -check-prefix=V9
+ ;RUN: llc -march=sparc -regalloc=basic < %s | FileCheck %s -check-prefix=V8
+ ;RUN: llc -march=sparc -regalloc=basic -mattr=v9 < %s | FileCheck %s -check-prefix=V9
++;RUN: llc -march=sparcv9  < %s | FileCheck %s -check-prefix=SPARC64
+ 
+ 
+ define i8* @frameaddr() nounwind readnone {
+@@ -15,6 +16,13 @@ entry:
+ ;V9: save %sp, -96, %sp
+ ;V9: jmp %i7+8
+ ;V9: restore %g0, %fp, %o0
++
++;SPARC64-LABEL: frameaddr
++;SPARC64:       save %sp, -128, %sp
++;SPARC64:       add  %fp, 2047, %i0
++;SPARC64:       jmp %i7+8
++;SPARC64:       restore %g0, %g0, %g0
++
+   %0 = tail call i8* @llvm.frameaddress(i32 0)
+   ret i8* %0
+ }
+@@ -32,6 +40,14 @@ entry:
+ ;V9: ld [%fp+56], {{.+}}
+ ;V9: ld [{{.+}}+56], {{.+}}
+ ;V9: ld [{{.+}}+56], {{.+}}
++
++;SPARC64-LABEL: frameaddr2
++;SPARC64: flushw
++;SPARC64: ldx [%fp+2159],     %[[R0:[goli][0-7]]]
++;SPARC64: ldx [%[[R0]]+2159], %[[R1:[goli][0-7]]]
++;SPARC64: ldx [%[[R1]]+2159], %[[R2:[goli][0-7]]]
++;SPARC64: add %[[R2]], 2047, {{.+}}
++
+   %0 = tail call i8* @llvm.frameaddress(i32 3)
+   ret i8* %0
+ }
+@@ -48,6 +64,9 @@ entry:
+ ;V9-LABEL: retaddr:
+ ;V9: or %g0, %o7, {{.+}}
+ 
++;SPARC64-LABEL: retaddr
++;SPARC64:       or %g0, %o7, {{.+}}
++
+   %0 = tail call i8* @llvm.returnaddress(i32 0)
+   ret i8* %0
+ }
+@@ -66,18 +85,12 @@ entry:
+ ;V9: ld [{{.+}}+56], {{.+}}
+ ;V9: ld [{{.+}}+60], {{.+}}
+ 
+-;V8LEAF-LABEL: retaddr2:
+-;V8LEAF: ta 3
+-;V8LEAF: ld [%fp+56], %[[R:[goli][0-7]]]
+-;V8LEAF: ld [%[[R]]+56], %[[R1:[goli][0-7]]]
+-;V8LEAF: ld [%[[R1]]+60], {{.+}}
++;SPARC64-LABEL: retaddr2
++;SPARC64:       flushw
++;SPARC64: ldx [%fp+2159],     %[[R0:[goli][0-7]]]
++;SPARC64: ldx [%[[R0]]+2159], %[[R1:[goli][0-7]]]
++;SPARC64: ldx [%[[R1]]+2167], {{.+}}
+ 
+-;V9LEAF-LABEL: retaddr2:
+-;V9LEAF: flushw
+-;V9LEAF: ld [%fp+56], %[[R:[goli][0-7]]]
+-;V9LEAF: ld [%[[R]]+56], %[[R1:[goli][0-7]]]
+-;V9LEAF: ld [%[[R1]]+60], {{.+}}
+-
+   %0 = tail call i8* @llvm.returnaddress(i32 3)
+   ret i8* %0
+ }