[RFC] IR: Support atomicrmw FP ops with vector types (#86796)

Allow using atomicrmw fadd, fsub, fmin, and fmax with vectors of
floating-point type. AMDGPU supports atomic fadd for <2 x half> and <2 x
bfloat> on some targets and address spaces.

Note this only supports the proper floating-point operations; float
vector typed xchg is still not supported. cmpxchg still only supports
integers, so this inserts bitcasts for the loop expansion.

I have support for fp vector typed xchg, and vector of int/ptr
separately implemented but I don't have an immediate need for those
beyond feature consistency.

1 files changed, 3 insertions, 3 deletions

diff --git a/llvm/lib/CodeGen/AtomicExpandPass.cpp b/llvm/lib/CodeGen/AtomicExpandPass.cpp
index d5db79d..0aa89ea 100644
--- a/llvm/lib/CodeGen/AtomicExpandPass.cpp
+++ b/llvm/lib/CodeGen/AtomicExpandPass.cpp
@@ -562,9 +562,9 @@ static void createCmpXchgInstFun(IRBuilderBase &Builder, Value *Addr,
                                  Value *&Success, Value *&NewLoaded) {
   Type *OrigTy = NewVal->getType();
 
-  // This code can go away when cmpxchg supports FP types.
+  // This code can go away when cmpxchg supports FP and vector types.
   assert(!OrigTy->isPointerTy());
-  bool NeedBitcast = OrigTy->isFloatingPointTy();
+  bool NeedBitcast = OrigTy->isFloatingPointTy() || OrigTy->isVectorTy();
   if (NeedBitcast) {
     IntegerType *IntTy = Builder.getIntNTy(OrigTy->getPrimitiveSizeInBits());
     NewVal = Builder.CreateBitCast(NewVal, IntTy);
@@ -731,7 +731,7 @@ static PartwordMaskValues createMaskInstrs(IRBuilderBase &Builder,
   unsigned ValueSize = DL.getTypeStoreSize(ValueType);
 
   PMV.ValueType = PMV.IntValueType = ValueType;
-  if (PMV.ValueType->isFloatingPointTy())
+  if (PMV.ValueType->isFloatingPointTy() || PMV.ValueType->isVectorTy())
     PMV.IntValueType =
         Type::getIntNTy(Ctx, ValueType->getPrimitiveSizeInBits());