[PATCH] D150189: GlobalISel: Refactor unary FP op constant folding

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp:1307
+  case TargetOpcode::G_FPTRUNC: {
+    bool Unused;
+    LLT DstTy = MRI.getType(MI.getOperand(0).getReg());
----------------
Pierre-vh wrote:
> If I understand correctly this is a NFC, G_FPTRUNC was handled anyway by the `V.convert` call below, but now it's explicit?
Yes, the LLT-to-semantics is only used for fptrunc now.

I'm thinking of moving FPTRUNC handling out of here entirely. It's an annoying special case complicating another patch I'm working on 


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D150189/new/

https://reviews.llvm.org/D150189