Sowmini.Varadhan at Sun.COM writes: > On (03/10/09 13:13), Paul Durrant wrote: > > > > How about deliberately making sure that the per-interface props. are > > differently names to global props. already available via ndd and giving > > per-interface props. the possibility of being undefined. Then: > > that still leaves a mess in the kernel (and even in the user-semantics). > Let's take the icmp error rate limit as an example. If I say > > # ipadm set-prop -p icmp_err_rate_limit=val1 bge0 > # ipadm set-prop -p icmp_err_rate_limit=val2 bge0 > > while supporting the > # ndd -[set,get] /dev/ip old_icmp_err_rate_limit <..> > > what is the kernel going to track for the older "global" definition? > No matter what I print for the global "ndd -get" value, it's going to > be a lie.
The fact that the user has deliberately asked to have a per-interface value for bge0 _means_ that the global value returned for a query needn't represent the value specified for that one interface. I don't think I see this as being terribly different from the existing ndd "ip_forwarding" bit. -- James Carlson, Solaris Networking <james.d.carlson at sun.com> Sun Microsystems / 35 Network Drive 71.232W Vox +1 781 442 2084 MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
