This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR target/6512
From: Richard Henderson <rth@redhat.com>
Date: Tue, 30 Apr 2002 12:31:12 -0700
On Tue, Apr 30, 2002 at 02:19:05PM +0200, Jakub Jelinek wrote:
> I've tried patch P3 (changing e -> f constraints for e <- o and o <- e
> alternatives), but it did work at all, for the above insn
> reload just tried (set %f40 (mem ())); (set %f40 %f40).
Means you're missing a test in SECONDARY_RELOAD_NEEDED.
The P3 test is more correct. The aligned case is handled by
the e/W alternative.
Please note, this is almost surely the same bug as shown by
PR target/5628
If you fix PR target/6512 properly, you'll most likely end up
fixing PR target/5628 as well.
I am not %100 sure of this, which is why I haven't marked one
as a duplicate of the other in GNATS. They need to both be
verified independantly once a fix is decided upon.