From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Carlson Date: Fri, 15 Oct 2010 13:31:53 +0000 Subject: Re: [PATCH] pppoe: Missed check for destination addr in PADT frame Message-Id: <4CB857C9.9080703@workingcode.com> List-Id: References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-ppp@vger.kernel.org On 10/15/10 08:38, Leonid Lisovskiy wrote: > PADT h_source:11:95:fc:83:f3 h_dest: 00:0e:08:2f:da:c5 session_id: 1 > PADT h_source:11:95:fc:83:f3 h_dest: 00:14:85:03:da:49 session_id: 1 Wow, that's wacky. Thanks for the data point. I've never seen that behavior before. For what it's worth, I think there's nothing specifically wrong with that behavior, so I'd say it's a bit of a stretch calling it a bug. The session ID is for the server to use in demuxing; if it can demux without it, then using 1 for everything doesn't violate anything I can see. "Normal" clients should not be seeing unicast packets sent to anything other than their own MAC address -- they should get filtered out at the MAC layer before PPPoE ever has a chance to see them. So the reuse should be invisible. If you're not filtering for your own MAC address on unicast reception, I'd be a little concerned that there are other places where the protocols go awry, and not just PADT. This might be something to look into. -- James Carlson 42.703N 71.076W