From: John Heffner <[EMAIL PROTECTED]> Date: Thu, 01 Dec 2005 11:53:44 -0500
> Yes it does, and arguably correctly so. As I see it this really comes > down to a question of cached metrics scope. I had a discussion recently > about this with Fernando Gont. See the thread "Improvement for the > current PMTUD mechanism" at > <http://www1.ietf.org/mail-archive/web/pmtud/current/threads.html>. Thanks for the interesting info John. I had glanced at the UDP stuff, and remembered how we can't validate those PMTUs, but I didn't connect this to the ability to "pollute" metrics TCP will end up using thus making any TCP checks futile. I'm actually happy about this, because in all honesty I didn't want to add the code necessary to support the PMTU protections in section 7.2 of Gont's draf. It adds a bunch of state manipulations in the fast paths, and as shown gives zero protection in the end. This gives further credence to BSD's hostcache which makes it use PMTU metrics only learned by TCP. I still dislike the reduced granularity of such a scheme, since as we all know ipsec routes can have wildly different metrics and can be keyed by things like port numbers. - 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
