https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90202
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |ra
Target| |x86_64-*-* i?86-*-*
Status|UNCONFIRMED |RESOLVED
CC| |hjl.tools at gmail dot com
Resolution|--- |WORKSFORME
--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
The same pushes/moves appear for the AVX2 code. I think we have a duplicate
for this issue and it appears because we eventually have to align the local
stack to be able to spill ymm/zmm registers and we can't fully elide
the frame construction after RA.
Original issue worksfrme (works as designed), frame issue has a duplicate
somewhere.
