This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: scheduling cc0
- From: Miles Bader <miles at lsi dot nec dot co dot jp>
- To: Paul Schlie <schlie at comcast dot net>
- Cc: Alexandre Oliva <aoliva at redhat dot com>, Ian Lance Taylor <ian at airs dot com>, Steven Bosscher <stevenb at suse dot de>, <gcc at gcc dot gnu dot org>
- Date: Fri, 04 Feb 2005 16:36:43 +0900
- Subject: Re: scheduling cc0
- References: <BE2797B0.8E4C%schlie@comcast.net>
- Reply-to: Miles Bader <miles at gnu dot org>
Paul Schlie <schlie@comcast.net> writes:
> - So it would appear that if there's a desire to transition CC0 based
> instruction descriptions from the implicit to explicit ones; intra-cycle
> parallel set & test semantic must be supported by GCC; as otherwise, CC0
> semantics can't truly be accurately described.
The fortunate thing is that the i386 is affected, and that's a pretty
obsessed-over architecture...
At least in gcc-3.4, i386.md uses clobbers of (reg:CC 17); does the info
missing because of the use of clobbers instead of sets cause many
problems?
Thanks,
-Miles
--
We are all lying in the gutter, but some of us are looking at the stars.
-Oscar Wilde