This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: need for instruction attributes (ia64)
- From: Richard Henderson <rth at redhat dot com>
- To: Jan Beulich <JBeulich at novell dot com>
- Cc: gcc at gcc dot gnu dot org
- Date: Tue, 11 Jan 2005 10:40:10 -0800
- Subject: Re: need for instruction attributes (ia64)
- References: <s1e3e810.023@emea1-mh.id2.novell.com>
On Tue, Jan 11, 2005 at 03:52:04PM +0100, Jan Beulich wrote:
> - whether define_insn and define_insn_and_split should generally have
> an itanium_class attribute, and if not, under what conditions this
> can/should be omitted
When we can be sure that the insn ought to be split, then we
don't need an attribute. In particular, there's usually not
one attribute that applies.
> - whether define_split (and define_insn_and_split if all of its
> alternatives split) should generally have a predicable attribute with
> value no ...
Splits don't have attributes.
> ... but the split code does not appear to ever get
> conditionalized, so the patterns seems simply useless, making the
> compiler bigger and perhaps slower)
Nor do they get duplicated for predication.
r~