> Assuming that the flag is set correctly from the original ipif during
 > the IPMP move, I'd suggest passing in ND_REACHABLE if ipif_addr_ready
 > is set, and ND_PROBE otherwise.

That seems feasible.  I'll explore it (in the new codebase) -- I'm not
too eager to fix things in the old codebase at this point.

 > (I don't know of any test failures related to this, and I certainly
 > did run the IPMP test suite, so maybe a new case is needed just for
 > this ....)

Yes, this will be tested in the new IPMP test suite.

-- 
meem

Reply via email to