This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: bit extract adjust_address versus scheduler
- From: Richard Henderson <rth at redhat dot com>
- To: Alexandre Oliva <aoliva at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 1 May 2003 18:29:03 -0700
- Subject: Re: bit extract adjust_address versus scheduler
- References: <orissv56wp.fsf@free.redhat.lsd.ic.unicamp.br>
On Thu, May 01, 2003 at 06:54:14AM -0300, Alexandre Oliva wrote:
> It turned out that, even after fixing the MIPS-specific code, the
> problem was still there, because the generic code insisted in
> converting MEMs to byte_mode.
This is indeed a problem, but I don't know that BLKmode is
any better. It does seem like every backend will have to
do something with this so that they get the proper mode for
whatever store it is they want.
I'm willing to defer the problem to the backends though,
and use VOIDmode here for the address adjustments. In most
cases that'll keep the BLKmode mark that was already there.
I don't know if this can ever be invoked for other MEMs,
but it doesn't seem impossible.
r~