https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79779
Vladimir Makarov <vmakarov at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |vmakarov at gcc dot gnu.org
--- Comment #1 from Vladimir Makarov <vmakarov at gcc dot gnu.org> ---
It is hard for me to understand what is the target. The test is for m68k. M68k
uses always reload. So the target is not m68k.
I guess x86-64 was used (I have no idea what was a reason for doing this). In
this case, LRA gives the correct diagnostic. You can not have 2 different
outputs of class "d" because there is only one x86/x86-64 hard register of such
class.
So I believe the PR should be closed as not-a-bug.
