Date: Wed, 20 Apr 2005 14:29:59 -0700 From: Sam Leffler <sam@errno.com> To: Gleb Smirnoff <glebius@freebsd.org> Cc: net@freebsd.org Subject: Re: if_link_state_change() patch for review Message-ID: <4266C9D7.1010906@errno.com> In-Reply-To: <20050419171406.GA378@cell.sick.ru> References: <20050419064747.GC734@cell.sick.ru> <4264D430.D39B81D0@freebsd.org> <20050419120324.GA5862@cell.sick.ru> <4264F4BC.4F3B57AE@freebsd.org> <20050419121141.GB5862@cell.sick.ru> <4264F7D4.F5A9775F@freebsd.org> <42652357.8020604@errno.com> <20050419171406.GA378@cell.sick.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
Gleb Smirnoff wrote: > On Tue, Apr 19, 2005 at 08:27:19AM -0700, Sam Leffler wrote: > S> >change message will report either all state changes or those where at > S> >the time of reporting the state changed relative to the last report. > S> >The same assumption is true for the OpenIGPd we are working on at the > S> >moment. > S> > S> It is possible with the change to defer the messages to have multiple > S> changes coalesced. If an app is written to assume it receives notice of > S> every change and it uses this to track internal state then it can get > S> confused. The issue was whether or not to communicate any coalescing to > S> applications so they can recognize that it's happened. In lieu of doing > S> that I asked for a console printf so we could see if it ever happened in > S> practice. > > Yes, all the time we are speaking about a theoretical issue. Let's see > whether it can happen or not. I decided to go ahead with this change. > > S> >From an (routing) application point of view only effective state changes > S> >are interesting and only those should be provided. > S> > > S> > S> If an inteface does down, moves network, then comes back up and you only > S> get the up event then you will likely do the wrong thing unless you have > S> some other way of identifying what happened. I'm not convinced (yet) > S> this cannot happen so am being cautious. > > In this case you will receive additional messages, not only RTM_IFINFO. > Move network will also generate RTM_NEWADDR and RTM_DELADDR > Only if someone sets the address which won't be true in the case I'm thinking of: dhclient handling a wireless nic that re-associates with a different ap. But since the dhclient I'm thinking of isn't in the tree yet I'll look at the problem myself. Sam
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4266C9D7.1010906>