Solaris 8 64-bit C compilation fails to add .register declaration
Vic Abell
abe@purdue.edu
Tue Jan 11 12:06:00 GMT 2000
Nicolas,
You write:
>
> On Tue, Jan 11, 2000 at 02:07:53AM +0100, Martin v. Loewis wrote:
> > > I believe this non-emission constitutes a bug.
> >
> > Thanks for your bug report. This is a bug, indeed. It is not the only
> > one.
> >
> > > I found instructions for building a 64-bit-capable GCC for Solaris 8
> > > on UltraSPARCs
> >
> > Don't follow such instructions. They mislead you by giving the
> > impression that gcc supports sparcv9 (aka sparc64). It does not. Don't
> > use gcc on that architecture, at least not in 64-bit mode. It is not
> > supported, and has numerous bugs.
> >
> > Regards,
> > Martin
> >
> > P.S. If I only could get the FAQ-o-matic to work for me again, so I
> > could add this to the FAQ...
>
> Thanks for your quick reply.
>
> Note also that the LSOF FAQ for LSOF 4.47_W provides some instructions
> for building GCC for building 64-bit UltraSPARC binaries.
>
> Between the changlogs for GCC 2.95.2, the LSOF FAQ and the instructions
> I found on Usenet I came to believe that this was doable.
>
> I guess the GCC and LSOF FAQs need updating then.
>
> Thanks,
>
> Nico
>
> PS: Note I'm CC'ing this mail to LSOF's maintainer, Vic Abell <abe@purdue.edu>.
As of the lsof 4.47 release, I had built 64 bit gcc's for gcc
versions 2.95, 2.95.1, 2.95.2, and 2.96. I then used each one to
build an lsof binary; all lsof binaries worked. That's the most
I can say for the effort; that's the most the lsof FAQ says; and
I defer to experts who caution against using such a gcc for
anything else.
The lsof FAQ does recommend consulting gcc experts. :-)
I'll add additional warnings to the FAQ.
Thanks,
Vic
More information about the Gcc-bugs
mailing list