[PATCH] D55506: [RFC v2] Allow target to handle STRICT floating-point nodes

[Kit Barton via Phabricator via llvm-commits]

kbarton added a comment.

Aside from two minor comments, I think this looks fine.
However, I don't think I'm qualified to give the final approval for this to land as I'm just starting to learn the background here.



================
Comment at: include/llvm/CodeGen/MachineInstr.h:835
 
+  /// Return true it this instruction could possibly raise a floating-point
+  /// exception.  This is the case if the instruction is a floating-point
----------------
it -> if


================
Comment at: include/llvm/CodeGen/SelectionDAGNodes.h:376
+  // intersection logic more straightforward.
+  bool NoExcept : 1;
+
----------------
Should this be renamed to NoFPExcept also, to remain consistent with the mayRaiseFPException flag?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D55506/new/

https://reviews.llvm.org/D55506