Add libcall support for G_FMINNUM, G_FMAXNUM, G_FSQRT, G_FRINT, G_FNEARBYINT.
Add 128-bit support to all simple libcalls.
Paths
| Differential D75516
[GlobalISel] Add missing floating point and s128 libcall support ClosedPublic Authored by gargaroff on Mar 3 2020, 4:40 AM.
Details Summary Add libcall support for G_FMINNUM, G_FMAXNUM, G_FSQRT, G_FRINT, G_FNEARBYINT. Add 128-bit support to all simple libcalls.
Diff Detail Event Timeline
Comment Actions Sorry, still struggling with Phabricator... Apply correct patch, addressing review comments. This revision is now accepted and ready to land.Mar 4 2020, 12:44 PM Comment Actions
Nevermind, I just got commit access and was able to commit this myself.
Revision Contents
Diff 248120 llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
llvm/unittests/CodeGen/GlobalISel/LegalizerHelperTest.cpp
|
Should there be a table for this? The nested ternary operators are becoming unwieldy