[PATCH] D55506: [RFC v2] Allow target to handle STRICT floating-point nodes
Ulrich Weigand via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Dec 18 02:29:37 PST 2018
uweigand marked an inline comment as done.
uweigand added inline comments.
================
Comment at: lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp:6520
+ SDNodeFlags Flags;
+ Flags.setFPExcept(true);
+ Result->setFlags(Flags);
----------------
andrew.w.kaylor wrote:
> This really shouldn't be unconditional. It's possible to have constrained intrinsics that are preserving the rounding mode but ignoring FP exceptions. At this point we still have that information.
Yes, of course, I'll add that for the final version. That's the main reason why I added the SDNode flag in the first place, instead of just checking for a STRICT_ node later on.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D55506/new/
https://reviews.llvm.org/D55506
More information about the llvm-commits
mailing list