This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: better fp truncation sequence on i386
- To: Jan Hubicka <jh at suse dot cz>
- Subject: Re: better fp truncation sequence on i386
- From: Richard Henderson <rth at redhat dot com>
- Date: Fri, 22 Jun 2001 17:14:38 -0700
- Cc: gcc-patches at gcc dot gnu dot org, patches at x86-64 dot org, timothy dot c dot prince at intel dot com
- References: <20010622190533.C1375@atrey.karlin.mff.cuni.cz>
On Fri, Jun 22, 2001 at 07:05:33PM +0200, Jan Hubicka wrote:
> Overall it should be still better to make unconditional pre-reload splitter
> pass to add the USEs and use mode_switching pass to initialize the CW values.
> Unconditional pre-reload splitting can be usefull on number of other passes
> as well. Can be OK if I add one?
I don't think you'll need a pre-reload splitter for this. The reason
being that the USEs simply won't exist before mode switching. I would
like to see this done, however.
> * i386-protos.h (emit_i387_cw_initialization): Declare.
> * i386.md (movsi_insv_1): Remove '*'.
> (fix_trunc?f?i2): Revamp to use control word parameters.
> (fix_trunc?i_1): Likewise; update splitters.
> (fix splitter): Remove.
> * i386.c (MAX_386_STACK_LOCALS): set to 3.
> (emit_i387_cw_initialization): New function.
> (output_fix_trunc): Do not initialize control word;
> kill dead code; simplify.
Ok.
r~