https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85185
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Target| |riscv
--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
It does look like a bug to me if riscv ends up with a reg containing sth
different than the value sub.a.
Using x86 is of course not a very good example so maybe check on a
PROMOTE_SUBREG target instead? It might be a bug in the riscv backend as well,
how does the
initial generated RTL look like?
