[email protected] wrote: > On (09/17/09 14:28), James Carlson wrote: >> [email protected] wrote: >>> I've added code that will ignore the RTM_LOSING if there's >>> no alternate default router available. >> It'd be nice if in.routed were smart enough to ignore RTM_LOSING when >> there are generally a lack of alternatives, and not just for RDISC. Any >> time we're about to do accelerated (or even perhaps _normal_) age-out of >> a route, and removal would mean that there's no longer a path to that >> subnet, then making sure it lingers as metric 15 instead would be a bit >> smarter. >> >> table.c >> 1240: consider printing the symbolic address rather than raw hex; >> use naddr_ntoa(). >> >> 2265,2296,2450: is a sowmini stop like a bus stop or a command? ;-} >> > > See > > http://cr.opensolaris.org/~sowmini/routed_fixes/ > > which now handles both rip and rdisc, for all prefixes (including > default)
Nice! Bug at line 1262: that should be "%s", not "%x". Nit at line 1256 -- the message should say something about why it's reporting a lack of a route. (got RTM_LOSING, but ...) Similar issue with 1270. Nits at 1263 and 1270: \n and trailing spaces probably not needed. -- James Carlson 42.703N 71.076W <[email protected]> _______________________________________________ networking-discuss mailing list [email protected]
