[llvm] [AMDGPU] Remove RCP_IFLAG combine (PR #197426)

[Piotr Sobczak via llvm-commits]

================
@@ -15463,12 +15463,6 @@ SDValue SITargetLowering::performRcpCombine(SDNode *N,
                                  SDLoc(N), VT);
   }
 
-  if (VT == MVT::f32 && (N0.getOpcode() == ISD::UINT_TO_FP ||
-                         N0.getOpcode() == ISD::SINT_TO_FP)) {
-    return DCI.DAG.getNode(AMDGPUISD::RCP_IFLAG, SDLoc(N), VT, N0,
----------------
piotrAMD wrote:

I think Matt's reply was meant here: "we might want it, in strictfp functions, I suppose". So I'm keeping it.

https://github.com/llvm/llvm-project/pull/197426