FreeBSD change for openssh

Mike Karels karels at
Tue Dec 21 06:08:20 AEDT 2021

Following up on my own email; I wrote:

> Darren Tucker <dtucker at> wrote:
> > On Fri, 17 Dec 2021 at 10:11, Darren Tucker <dtucker at> wrote:
> > >
> > > On Fri, 17 Dec 2021 at 09:05, Mike Karels <karels at> wrote:
> > > [...]
> > > > I haven't seen any other comments.  Can you commit this?  If not, what
> > > > is the procedure to get it committed?
> > >
> > > Leaving aside the implementation details, the whole concept seems
> > > questionable to me.

> > To be specific: the concept I find questionable is repurposing (most
> > of) as global unicast addresses.

> Neither this change, nor the larger set of changes I'm making in FreeBSD,
> redefine the loopback network.  FreeBSD has had an IN_LOOPBACK macro for
> years, with essentially the same definition as the default I provided.
> The goal of my changes is to reduce the knowledge of the obsolete Class
> A/B/C network structure as much as possible.  It is true that this change
> makes it easier to change the system definition of the loopback network,
> but I am not doing that.  It seems to me to be preferable for programs like
> OpenSSH to use system definitions where possible, and not to hard-code
> things like the definition of the loopback network.

Where do we stand on this?  Any other opinions?


More information about the openssh-unix-dev mailing list