This is the mail archive of the
mailing list for the GCC project.
Re: head: MIPS: Workaround an R4000SC/MC branch/divide erratum
"Maciej W. Rozycki" <firstname.lastname@example.org> writes:
> On Wed, 3 Mar 2004, Richard Sandiford wrote:
>> The only difference from the previous on-list version is in the
>> change to the comment above mips_output_division.
> I don't think so -- I didn't made any changes to the comment.
Note that I said "on-list". Like I said, the patch I committed was
exactly what you sent me. I just wanted to give a heads-up to others
on the gcc-patches list that the comment had changed since the last
> BTW, I've just wondered if we should obey "-fno-delayed-branch" here, for
> consistency if not anything else.
IMO, no. Fixed-form patterns have unconditionally used branch delay
slots for a long time. E.g. the !TARGET_64BIT versions of the shift
patterns. Not that I'm strongly against changing it...