> > Also, I trust that you will not be taking these properties as-is.  For
 > > instance, it's cleaner to have a single echo-reply tunable that can be set
 > > to one or more values like "broadcast" than to have a sea of boolean
 > > tunables.
 > 
 > I've thought about this one.. on the one hand we don't want to create
 > a new sub-command, otoh, there's a whole family of these that can be
 > coalesced, as, perhaps
 > 
 >   ipadm set-prop \
 >         -p icmp_respond=<icmp_type>,<unicast|bcast|mcast>,<yes|no> \
 >         <interface> 
 > 
 > Note that redirects are really a special case of the above (icmp type 5
 > for ipv4).

Right.  To some degree this is a matter of personal taste and a privilege
that Girish gets as the RE for this work.  My broader point is that it'd
be a shame if we just moved the tunables verbatim into dladm without also
considering how we could clean them up.

-- 
meem

Reply via email to