This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Support for i386 builds of glibc?
On Mon, 4 Mar 2013, Carlos O'Donell wrote:
> On 03/04/2013 01:02 PM, Joseph S. Myers wrote:
> > On Mon, 4 Mar 2013, Carlos O'Donell wrote:
> >
> >> So the question I have to ask myself is: Are we willing to support
> >> i386 builds of glibc? Why? Why not?
> >
> > At least for Linux (I don't know about Hurd), I don't think it has ever
> > worked for NPTL (because of __sync_* operations that are only implemented
> > for i486 and above), and since the Linux kernel has removed i386 support I
> > don't think it would be worth trying to add such support back to glibc.
>
> Thanks, that's enough rationale for me to consider i386 dead.
>
> Would we want a hard error if you build with `-march=i386'?
If configured for i386, yes. If configured for i486 or above but with a
compiler that generates code for i386 (out-of-line __sync_* calls), add a
-march= option. See what I proposed in bug 10062 (and bug 10060 is
probably the same issue).
--
Joseph S. Myers
joseph@codesourcery.com