Bug#556631: KSM support
On Tue, Nov 17, 2009 at 11:44:02AM +0100, Guido Trotter wrote:
> On Tue, Nov 17, 2009 at 10:37:03AM +0100, Aurelien Jarno wrote:
>
> Hi,
>
> > > But at:
> > > http://www.eglibc.org/archives/patches/msg00793.html
> > >
> > > I was told to wait for upstream GNU glibc support, for a routine merge.
> >
> > That's actually the reverse. You should submit the patch to upstream GNU
> > libc, so that it is merged later in EGLIBC.
> >
>
> Yes, that's what I meant.
Unless I missed something, I have seen no submission. I'll submit it
during the afternoon.
> > > KSM will be present in the 2.6.32 kernel, which is going to be released
> > > soon. It might be nice if Debian could include this patch earlier, since
> > > for us eglibc is the default libc, so that as soon as that kernel is out
> > > and the kvm/qemu people package a version with KSM support, the
> > > compilers will pick up the flag and KSM will work in unstable/squeeze.
> > > (qemu/kvm code has an #ifdef which means that they will be compiled
> > > without kvm if glibc doesn't have the flag)
> > >
> >
> > Given it changes the API, I would prefer to see it accepted upstream
> > first.
> >
>
> But given that it's in the upcoming kernel, and also a feature the newest Fedora
> is shipping with it
> (http://www.internetnews.com/dev-news/article.php/3848441/Fedora+12+Linux+Tackles+Virtualization.htm)
> is there any reason why this wouldn't reach upstream anyway?
I see no reason for it not to be accepted upstream. But waiting for
the upstream patch also means that we have the correct patch, including
all the architectures (even those in ports/).
I don't expect it to be long, it's probably a matter of days after the
patch has been submitted, probably even before 2.6.32 is released.
Then I have no problem to include such a patch.
--
Aurelien Jarno GPG: 1024D/F1BCDB73
aurelien@aurel32.net http://www.aurel32.net
Reply to: