Adding this way of packet handling will bring a huge competition advantage
for chrony.
I think.

Here is some onfi about netlink practical usgee, in Russian, but you can
read it via Google translator.

Anyway, thanks in advance.

вт, 5 сент. 2023 г. в 17:03, Miroslav Lichvar <mlich...@redhat.com>:

> On Tue, Sep 05, 2023 at 04:33:11PM +0300, CpServiceSPb wrote:
> > > That would make more sense for security. However, it's not a simple
> thing
> > > to implement as peer associations use the server sockets too, so there
> > > would need to be some code selecting the right socket.
> > Maybe it is worth looking at NTP sources for aspects of the topic.
> > It supports multiple bindings as I know.
>
> The ntpd I/O code was designed around having a separate socket for
> each interface. It monitors changes in network configuration (on Linux
> using netlink socket). It needs that to be able to respond from the
> right address on multihomed hosts. The amount of code needed for that
> is huge and lot of it is system-specific.
>
> chronyd relies on the IP_PKTINFO socket option to get the destination
> address of requests. This is much simpler, but probably not available
> on all systems that ntpd currently supports.
>
> > I believe you will be able to implement correctly the functionality.
> > And will wait for the version to test it.
>
> I added it to my todo list, but don't get your hopes up.
>
> --
> Miroslav Lichvar
>
>
> --
> To unsubscribe email chrony-dev-requ...@chrony.tuxfamily.org with
> "unsubscribe" in the subject.
> For help email chrony-dev-requ...@chrony.tuxfamily.org with "help" in the
> subject.
> Trouble?  Email listmas...@chrony.tuxfamily.org.
>
>

Reply via email to