This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Bootstrap is broken on i[345]86-linux
- From: "H. J. Lu" <hjl at lucon dot org>
- To: Jan Hubicka <jh at suse dot cz>
- Cc: FX Coudert <fxcoudert at gmail dot com>, michael dot meissner at amd dot com, gcc-patches List <gcc-patches at gcc dot gnu dot org>, rth at redhat dot com, ubizjak at gmail dot com
- Date: Sun, 29 Apr 2007 17:14:05 -0700
- Subject: Re: Bootstrap is broken on i[345]86-linux
- References: <19c433eb0704180058k1efd94ccqe5eda2120726e2ee@mail.gmail.com> <A94F7CB8-977D-4729-ADD4-0A412462D9DE@gmail.com> <20070429134753.GA18052@kam.mff.cuni.cz>
On Sun, Apr 29, 2007 at 03:47:53PM +0200, Jan Hubicka wrote:
> > ping again, again and again. Bootstrap has been broken on i[345]86-
> > linux with default options for more than a month. It's sad that
> > noone, including i386 maintainers, cares enough to review an already
> > existing patch or pressure the original committers (H. J. Lu and
> > Michael Meissner) into doing something, maybe disabling decimal float
> > on i[345]86-linux.
>
> The problem is that the proposed patch is more a workaround, rather
> than that a generic fix is needed in push expanding machinery. (as
> mentioned earlier in thread, emit_move_insn knows how to decompose
> decimal into integer, but push code doesn't)
>
> I can write it, but my hdd just died today so I will need some time to
> recover data from backups. So if someone beats me, it would be nice.
> I will try to get into it ASAP.
>
emit_single_push_insn calls emit_move_insn to push dfp. emit_move_insn
calls emit_move_via_integer to push dfp via integer, which generates
(set (reg/f:SI 67) (pre_dec:SI (reg/f:SI 7 sp)))
It leads to an unrecognizable insn error. I am not sure if
emit_move_via_integer properly handles moving from Y to X and X
satifies push_operand.
H.J.