status update

[Francis Dupont]

> What I don't quite understand is the source(s)? of the conflict. Can't
> SO_BINDTODEVICE be used everywhere?

=> it can, and more simpler you can put an address before calling bind().
But dnsmasq doesn't do that, the -z just put the SO_REUSEADDR flag
which is incorrectly implemented on Linux as removing the conflict
check...

>     Set SO_BINDTODEVICE on DHCP sockets when doing DHCP on one interface
>     only. Fixes OpenSTack use-case.

=> I checked only the released version when the issue was discovered
so it is possible someone at dnsmasq fixed it (usually they are fast
to fix obvious issues).

I put this on my TODO list... currently I am playing with the Thales
HSM (small but expensive and fast :-) with the PKCS#11 20a4d2
(file name pkcs-11v2-20a4d2.pdf on the web is you want to look at it).

Regards

Francis Dupont <fdupont at isc.org>