[PATCH] D71871: [X86] Enable STRICT_SINT_TO_FP/STRICT_UINT_TO_FP on X86 backend

[Pengfei Wang via Phabricator via llvm-commits]

pengfei marked 11 inline comments as done.
pengfei added inline comments.


================
Comment at: llvm/lib/CodeGen/SelectionDAG/LegalizeVectorOps.cpp:568
+                     : ISD::SIGN_EXTEND;
   for (unsigned j = 0; j != Op.getNumOperands(); ++j) {
+    if (IsStrict && j == 0) {
----------------
craig.topper wrote:
> Is a for loop needed here? Isn’t there only one non chain operand?
It has no difference between strict and normal node. Removed.


================
Comment at: llvm/test/CodeGen/X86/fp-intrinsics.ll:2204
 ; X87:       # %bb.0: # %entry
-; X87-NEXT:    subl $12, %esp
-; X87-NEXT:    .cfi_def_cfa_offset 16
----------------
craig.topper wrote:
> What change caused these test changes? It looks like most of this patch is vector related. What did I miss?
The target combine for the node STRICT_SINT_TO_FP/STRICT_UINT_TO_FP.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D71871/new/

https://reviews.llvm.org/D71871