https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118594
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rsandifo at gcc dot gnu.org
--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I am thinking the best way to fix this is the following:
use simplify_subreg in force_subreg instead of simplify_gen_subreg.
If the op is a REG in outermode then generate the subreg.
Otherwise force to reg like when simplify_gen_subreg returns null.
The other option is to reject MEM inside validate_subreg.
The 3rd option is inside convert_mode_scalar to do a force reg before the
force_lowpart_subreg.
Both of these are GCC 16 code changes I think.
CCing Richard to see if he has any thoughts on the issue of
force_lowpart_subreg producing `(SUBREG (MEM/v ))` here.
