https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65648
Vladimir Makarov <vmakarov at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |vmakarov at gcc dot gnu.org
--- Comment #4 from Vladimir Makarov <vmakarov at gcc dot gnu.org> ---
(In reply to Jeffrey A. Law from comment #2)
> Well, I think the real issue here is LRA deciding to move insn 12 regardless
> of whether or not the MD file gets a scratch register the recommend way or
> not.
>
> Note, the insns below are not consecutive in the .ira dump, had me rather
> confused for a few minutes.
The insn looks moved as the bug is in new rematerialization sub-pass of LRA.
I'll try to fix this until Wednesday.
