Page MenuHomePhabricator

[ARM][DAGCOMBINE][FIX] PerformVMOVRRDCombine
ClosedPublic

Authored by dnsampaio on Jul 17 2019, 7:56 AM.

Details

Summary

PerformVMOVRRDCombine ommits adding a offset
of 4 to the PointerInfo, when converting a
f64 = load[M]
to
{i32, i32} = {load[M], load[M + 4]}

Which would allow the machine scheduller
to break dependencies with the second load.

  • pr42638

Diff Detail

Repository
rL LLVM

Event Timeline

dnsampaio created this revision.Jul 17 2019, 7:56 AM
ostannard accepted this revision.Jul 18 2019, 1:51 AM
ostannard added a subscriber: ostannard.

LGTM

This revision is now accepted and ready to land.Jul 18 2019, 1:51 AM
This revision was automatically updated to reflect the committed changes.