On Sun, 2005-13-11 at 00:04 +0100, Krzysztof Halasa wrote: > Jamal Hadi Salim <[EMAIL PROTECTED]> writes: > > > Issue 1: > > 1) There are routes that are added by the kernel. These are labelled as > > being added by the kernel. > > 2) Others maybe added by a dynamic routing daemon and these would be > > labelled as being inserted by such a daemon > > 3) Yet others are added by an admin using a command line tool such as > > iproute2, route, etc. And i can tell those as well. > > What's the difference WRT carrier?
If you consider "carrier" (by which i hope you mean oper state) is also status, then the only thing we would worry about is #1. That is consistent with how we treat admin status. I am lumping as well into #1 all routes created as a result of adding ip addresses (secondary/primary etc). > Are some of them usable more or less when the device is down? no route is usable if it is pointing to a next hop along a device that is admin or oper down - but that is not the question. The question is who is responsible for managing them? Should it not be whoever added them? As you can see from above there are 3 types of entities that could add them. > > > Routes can be made to > > prohibit, blackhole, throw packets or become unreachable. Is that what > > you are saying? > > Obviously, no. I mean route entry "inactive" flag (not existing yet) > which has exactly nothing to do with its destination. I dont know what that would buy you that a blackhole route could not give you: Packets will be dropped at the ip level way before they hit device level. And this is a lot more contentious than the little leeway i was suggesting with #1 above. You are asking for all routes even those that the kernel did not install to be treated this way. Right now you are not even getting a resolution to #1 ;-> >> I am not going to release this patch - until we get to some >> consensus. > > Just post it. How can we have consensus if others don't know clearly > our ideas? > Like i explained already: The patch extends oper state to add/delete routes on a netdevice that are owned by the kernel; i.e #1 above. I dont wanna distract the discussion with posting it. All these discussions have been very valuable - i think there is enough understanding of the difference of opinions that we need to close on them. I will post the issues and the different opinions on them - and hopefully we can close. cheers, jamal - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
