On Wed, 2004-05-05 at 15:49, Jan-Benedict Glaw wrote:
> On Wed, 2004-05-05 14:44:49 -0400, Gregory Symons <[EMAIL PROTECTED]>
[snip]
> > And I am indeed running a 2.4.20 kernel. So maybe the configure script
> > needs to detect whether or not this flag is actually available (which I
> > thought was the point of the #ifdef), rather than trusting the fact that
> > the macro itself is defined. Otherwise, if IPV6 _is_ available, but
>
> Well, that depends...
>
> > IPV6_IPV6ONLY option is not, using the patch you sent would cause us to
> > skip any IPV6 addresses that we want to bind to.
>
> rsync (as other programs) should ./configure to adopt it's host system's
> libc API. That is, if libc offers IPV6_V6ONLY, rsync should accept that
> fact and use it {properly,if needed}. So if there's a (detectable) bug,
> let's complain about it, but please accept the presented API.
Which is why I didn't write my own patch:) I'm just starting to look at
stuff like this, so I don't necessarily know the best way to fix it. I'm
still working on the finding it part:).
> If really a workaround is needed, rsync should detect that error *at
> runtime*, complain loudly and exit (offering some
> --use-broken-linux-2.4.x-ipv6-stack) for allowing it to work properly.
>
> Runtime check? Right. Think about distributions that offer both 2.4.x
> and 2.6.x. For sure you won't like to deliver two rsync versions to your
> users:) ...and because IPV6_V6ONLY exists in 2.4.21 upwards, I think
> that's not a major problem. Hey, these kernels do have known root
> exploits, so it's time for an update!
Good point. That's the problem with gentoo... sometimes I lose sight of
the fact that not everyone's compiling their packages. I have to admit I
wasn't thinking that far ahead. And detecting the problem at runtime is
a perfectly acceptable way to handle the problem. My only concern is
that the patch as submitted will skip _every_ IPV6 socket on affected
systems, whether the user wanted to use them or not.
>
> So if there's a but, don't hide it, but work towards fixing it.
>
I agree. But at the same time, if that bug can be worked around, I think
it should be... not everyone can upgrade their
kernel/libraries/whatever.
I like the idea of dying loudly, though... it'd remind people to get off
those buggy kernel versions (which until today, I thought I had...
crossed it off the list, but never actually did it... good thing it's
just my home system, otherwise I'd be in big trouble:)
Greg
--
To unsubscribe or change options: http://lists.samba.org/mailman/listinfo/rsync
Before posting, read: http://www.catb.org/~esr/faqs/smart-questions.html
