[PATCH] D52768: [SelectionDAG] DAG combiner for fminnan and fmaxnan

[Heejin Ahn via Phabricator via llvm-commits]

aheejin added a comment.

So this change is not necessarily related to `minimum`/`maximum` new intrinsics, because they are pre-existing. So this means `FMINNAN`/`FMAXNAN` nodes have not been optimized in DAG combiner so far. Strange. Anyway apparently what this does is constant folding on DAG nodes, so I think we can come up with a test?



================
Comment at: lib/CodeGen/SelectionDAG/DAGCombiner.cpp:12110
+static SDValue visitFMinMax(SelectionDAG &DAG, SDNode *N,
+                            APFloat (*Op)(const APFloat&, const APFloat&)) {
   SDValue N0 = N->getOperand(0);
----------------
Is this line clang-formatted?


Repository:
  rL LLVM

https://reviews.llvm.org/D52768