Merge llvm-project 12.0.1 release and follow-up fixes

Merge llvm-project main llvmorg-12-init-17869-g8e464dd76bef

This updates llvm, clang, compiler-rt, libc++, libunwind, lld, lldb and
openmp to llvmorg-12-init-17869-g8e464dd76bef, the last commit before the
upstream release/12.x branch was created.

PR:		255570

(cherry picked from commit e8d8bef961a50d4dc22501cde4fb9fb0be1b2532)

Merge llvm-project 12.0.0 release

This updates llvm, clang, compiler-rt, libc++, libunwind, lld, lldb and
openmp to llvmorg-12.0.0-0-gd28af7c654d8, a.k.a. 12.0.0 release.

PR:		255570

(cherry picked from commit d409305fa3838fb39b38c26fc085fb729b8766d5)

Disable strict-fp for powerpcspe, as it does not work properly yet

Merge commit 5c18d1136665 from llvm git (by Qiu Chaofan)

  [SPE] Disable strict-fp for SPE by default

  As discussed in PR50385, strict-fp on PowerPC SPE has not been
  handled well. This patch disables it by default for SPE.

  Reviewed By: nemanjai, vit9696, jhibbits

  Differential Revision: https://reviews.llvm.org/D103235

PR:		255570

(cherry picked from commit 715df83abc049b23d9acddc81f2480bd4c056d64)

Apply upstream libc++ fix to allow building with devel/xxx-xtoolchain-gcc

Merge commit 52e9d80d5db2 from llvm git (by Jason Liu):

  [libc++] add `inline` for __open's definition in ifstream and ofstream

  Summary:

  When building with gcc on AIX, it seems that gcc does not like the
  `always_inline` without the `inline` keyword.
  So adding the inline keywords in for __open in ifstream and ofstream.
  That will also make it consistent with __open in basic_filebuf
  (it seems we added `inline` there before for gcc build as well).

  Differential Revision: https://reviews.llvm.org/D99422

PR:		255570

(cherry picked from commit d099db25464b826c5724cf2fb5b22292bbe15f6e)

Undefine HAVE_(DE)REGISTER_FRAME in llvm's config.h on arm

Otherwise, the lli tool (enable by WITH_CLANG_EXTRAS) won't link on arm,
stating that __register_frame is undefined. This function is normally
provided by libunwind, but explicitly not for the ARM Exception ABI.

Reported by:	oh
PR:		255570

(cherry picked from commit f336b45e943c7f9a90ffcea1a6c4c7039e54c73c)

Merge llvm-project 12.0.1 rc2

This updates llvm, clang, compiler-rt, libc++, libunwind, lld, lldb and
openmp to llvmorg-12.0.1-rc2-0-ge7dac564cd0e, a.k.a. 12.0.1 rc2.

PR:		255570

(cherry picked from commit 23408297fbf3089f0388a8873b02fa75ab3f5bb9)

Revert libunwind change to fix backtrace segfault on aarch64

Revert commit 22b615a96593 from llvm git (by Daniel Kiss):

  [libunwind] Support for leaf function unwinding.

  Unwinding leaf function is useful in cases when the backtrace finds a
  leaf function for example when it caused a signal.
  This patch also add the support for the DW_CFA_undefined because it marks
  the end of the frames.

  Ryan Prichard provided code for the tests.

  Reviewed By: #libunwind, mstorsjo

  Differential Revision: https://reviews.llvm.org/D83573

  Reland with limit the test to the x86_64-linux target.

Bisection has shown that this particular upstream commit causes programs
using backtrace(3) on aarch64 to segfault. This affects the lang/rust
port, for instance. Until we can upstream to fix this problem, revert
the commit for now.

Reported by:	mikael
PR:		256864

(cherry picked from commit 5866c369e4fd917c0d456f0f10b92ee354b82279)

Merge llvm-project 12.0.1 release

This updates llvm, clang, compiler-rt, libc++, libunwind, lld, lldb and
openmp to llvmorg-12.0.1-0-gfed41342a82f, a.k.a. 12.0.1 release.

PR:		255570

(cherry picked from commit 4652422eb477731f284b1345afeefef7f269da50)

compilert-rt: build out-of-line LSE atomics helpers for aarch64

Both clang >= 12 and gcc >= 10.1 now default to -moutline-atomics for
aarch64. This requires a bunch of helper functions in libcompiler_rt.a,
to avoid link errors like "undefined symbol: __aarch64_ldadd8_acq_rel".

(Note: of course you can use -mno-outline-atomics as a workaround too,
but this would negate the potential performance benefit of the faster
LSE instructions.)

Bump __FreeBSD_version so ports maintainers can easily detect this.

PR:		257392

(cherry picked from commit cc55ee8009a550810d38777fd6ace9abf3a2f6b4)

1 files changed, 32 insertions, 2 deletions

diff --git a/contrib/llvm-project/llvm/lib/CodeGen/GlobalISel/CSEMIRBuilder.cpp b/contrib/llvm-project/llvm/lib/CodeGen/GlobalISel/CSEMIRBuilder.cpp
index 88173dc4d302..2c86f06a602d 100644
--- a/contrib/llvm-project/llvm/lib/CodeGen/GlobalISel/CSEMIRBuilder.cpp
+++ b/contrib/llvm-project/llvm/lib/CodeGen/GlobalISel/CSEMIRBuilder.cpp
@@ -13,6 +13,7 @@
 
 #include "llvm/CodeGen/GlobalISel/CSEMIRBuilder.h"
 #include "llvm/CodeGen/GlobalISel/GISelChangeObserver.h"
+#include "llvm/IR/DebugInfoMetadata.h"
 
 using namespace llvm;
 
@@ -41,8 +42,14 @@ CSEMIRBuilder::getDominatingInstrForID(FoldingSetNodeID &ID,
   if (MI) {
     CSEInfo->countOpcodeHit(MI->getOpcode());
     auto CurrPos = getInsertPt();
-    if (!dominates(MI, CurrPos))
+    auto MII = MachineBasicBlock::iterator(MI);
+    if (MII == CurrPos) {
+      // Move the insert point ahead of the instruction so any future uses of
+      // this builder will have the def ready.
+      setInsertPt(*CurMBB, std::next(MII));
+    } else if (!dominates(MI, CurrPos)) {
       CurMBB->splice(CurrPos, CurMBB, MI);
+    }
     return MachineInstrBuilder(getMF(), MI);
   }
   return MachineInstrBuilder();
@@ -61,6 +68,11 @@ void CSEMIRBuilder::profileDstOp(const DstOp &Op,
   case DstOp::DstType::Ty_RC:
     B.addNodeIDRegType(Op.getRegClass());
     break;
+  case DstOp::DstType::Ty_Reg: {
+    // Regs can have LLT&(RB|RC). If those exist, profile them as well.
+    B.addNodeIDReg(Op.getReg());
+    break;
+  }
   default:
     B.addNodeIDRegType(Op.getLLTTy(*getMRI()));
     break;
@@ -70,6 +82,9 @@ void CSEMIRBuilder::profileDstOp(const DstOp &Op,
 void CSEMIRBuilder::profileSrcOp(const SrcOp &Op,
                                  GISelInstProfileBuilder &B) const {
   switch (Op.getSrcOpKind()) {
+  case SrcOp::SrcType::Ty_Imm:
+    B.addNodeIDImmediate(static_cast<int64_t>(Op.getImm()));
+    break;
   case SrcOp::SrcType::Ty_Predicate:
     B.addNodeIDImmediate(static_cast<int64_t>(Op.getPredicate()));
     break;
@@ -115,7 +130,7 @@ bool CSEMIRBuilder::checkCopyToDefsPossible(ArrayRef<DstOp> DstOps) {
   if (DstOps.size() == 1)
     return true; // always possible to emit copy to just 1 vreg.
 
-  return std::all_of(DstOps.begin(), DstOps.end(), [](const DstOp &Op) {
+  return llvm::all_of(DstOps, [](const DstOp &Op) {
     DstOp::DstType DT = Op.getDstOpKind();
     return DT == DstOp::DstType::Ty_LLT || DT == DstOp::DstType::Ty_RC;
   });
@@ -131,6 +146,21 @@ CSEMIRBuilder::generateCopiesIfRequired(ArrayRef<DstOp> DstOps,
     if (Op.getDstOpKind() == DstOp::DstType::Ty_Reg)
       return buildCopy(Op.getReg(), MIB.getReg(0));
   }
+
+  // If we didn't generate a copy then we're re-using an existing node directly
+  // instead of emitting any code. Merge the debug location we wanted to emit
+  // into the instruction we're CSE'ing with. Debug locations arent part of the
+  // profile so we don't need to recompute it.
+  if (getDebugLoc()) {
+    GISelChangeObserver *Observer = getState().Observer;
+    if (Observer)
+      Observer->changingInstr(*MIB);
+    MIB->setDebugLoc(
+        DILocation::getMergedLocation(MIB->getDebugLoc(), getDebugLoc()));
+    if (Observer)
+      Observer->changedInstr(*MIB);
+  }
+
   return MIB;
 }