This is the mail archive of the
mailing list for the GCC project.
Re: head: MIPS: A workaround for the R4000 divide/shift errata
- From: "Maciej W. Rozycki" <macro at ds2 dot pg dot gda dot pl>
- To: Richard Sandiford <rsandifo at redhat dot com>
- Cc: Eric Christopher <echristo at redhat dot com>, gcc-patches at gcc dot gnu dot org
- Date: Mon, 1 Mar 2004 12:17:46 +0100 (CET)
- Subject: Re: head: MIPS: A workaround for the R4000 divide/shift errata
- Organization: Technical University of Gdansk
- References: <Pine.LNX.firstname.lastname@example.org><email@example.com> <Pine.LNX.firstname.lastname@example.org><email@example.com> <mailpost.1077233012.25220@news-sj1-1><firstname.lastname@example.org> <email@example.com><firstname.lastname@example.org><Pine.LNX.email@example.com> <firstname.lastname@example.org><Pine.LNX.email@example.com><firstname.lastname@example.org> <email@example.com><firstname.lastname@example.org> <Pine.LNX.email@example.com><Pine.LNX.firstname.lastname@example.org> <email@example.com><Pine.LNX.firstname.lastname@example.org> <email@example.com>
On Sat, 28 Feb 2004, Richard Sandiford wrote:
> I hope you don't mind, but I decided to change the name of the option to
> -mfix-r4000 (and changed the names of the internal macros accordingly).
> -march=r4000 is the canonical option for selecting R4000 code and it
> would be confusing to have a different convention for -mfix-*.
It's OK, I've wondered about it, too.
> [ I know this isn't really consistent with the existing -mfix4300
> option, but then again, neither was -mfix-4000. -mfix4300 should
> probably have been named -mfix-vr4300 (to make the family explicit)
> but it's too late do anything about that now.
> Also, gas does have a -mfix-7000 option, but the negative of that
> is -mno-fix7000 (no extra dash), so I don't see that as a good
> example to follow. There's also -mfix-vr4122-bugs, which is a bit
> more consistent with the new name, except for the extra "-bugs".
> -mfix-r4000 seems more in the spirit of -mfix-sb1 though.
> All anal stuff, but I had a feeling someone would point this out
> to me unless I proved I'd done my homework. ;) ]
Hmm, how about adding consistent aliases and deprecating old options?
> I was going to add a comment about why the nop was needed even for
> -mcheck-zero-division. In the end, I figured it was easier just to
> force the division into the delay slot as previously discussed.
> Working on that now.
Please wait for a while. I have two more patches for this area to
reflect events that happened meanwhile.
+ Maciej W. Rozycki, Technical University of Gdansk, Poland +
+ e-mail: firstname.lastname@example.org, PGP key available +