From owner-freebsd-net Thu Apr 1 7:49:41 1999 Delivered-To: freebsd-net@freebsd.org Received: from labinfo.iet.unipi.it (labinfo.iet.unipi.it [131.114.9.5]) by hub.freebsd.org (Postfix) with SMTP id ED72F14FEF for ; Thu, 1 Apr 1999 07:49:36 -0800 (PST) (envelope-from luigi@labinfo.iet.unipi.it) Received: from localhost (luigi@localhost) by labinfo.iet.unipi.it (8.6.5/8.6.5) id PAA15763; Thu, 1 Apr 1999 15:28:57 +0200 From: Luigi Rizzo Message-Id: <199904011328.PAA15763@labinfo.iet.unipi.it> Subject: more problems with 'de' driver and multicast... To: net@freebsd.org Date: Thu, 1 Apr 1999 15:28:57 +0200 (MET DST) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 2384 Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi, i don't know if anyone cares at all, but i am having a whole lot of problems with the 'de' driver and multicast. Details below -- conclusion is either help me fix these problem, or be careful when using the 'de' driver with multicast at least. cheers luigi DETAILS: I am running a rather interesting experiment on a cluster of 28 diskless machines running FreeBSD3.1R off another 3.1R server. All machines are connected to a 100mbit hub, and use a DE-based card (some have a 21140A rev.2.2, others a 21143 rev.4.1 as detected by the "de" driver). The experiment consists in running on each machine an instance of an application using layered multicast congestion control -- the relevant part for this msg is that rlc issues very frequent ADD_MEMBERSHIP/DROP_MEMBERSHIP calls. First problem is that the machines with the 21143 cards after a while stop receiving data (apparently, both multicast and unicast traffic) and do not restart unless there is some spontaneous transmission originated the machine itself. I.e. i cannot ping them from the controlling machine but have to move to the console... Second problem is that all machines -- irrespective of the type of card -- i seem to fail quite often in generating the IGMP LEAVE message when they should (i.e. according to IGMPv2 when they are the last ones who replied to a query). I suspect in both case the problem lies in some problem in the driver, specifically in handling the modifications to the address filters. This part of the code has changed significantly from 2.2 to 3.x and i don't think the new code has been tested very well. I already had and fixed a related problem (which fortunately shows up only on machines running mrouted), and the weird behaviour of the 'de' counters in netstat (try a netstat -ni and see how often does the Opkt counter changes) further reduces my confidence in the driver. So, if someone has suggestions on how to fix the code... i have at least a reasonably good testbed to try patches. -----------------------------------+------------------------------------- Luigi RIZZO . EMAIL: luigi@iet.unipi.it . Dip. di Ing. dell'Informazione HTTP://www.iet.unipi.it/~luigi/ . Universita` di Pisa TEL/FAX: +39-050-568.533/522 . via Diotisalvi 2, 56126 PISA (Italy) -----------------------------------+------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message