From mboxrd@z Thu Jan 1 00:00:00 1970 From: John Fastabend Subject: Re: [net-next PATCH v1 0/7] Managing the forwarding database(FDB) Date: Mon, 09 Apr 2012 15:32:04 -0700 Message-ID: <4F836364.6070406@intel.com> References: <20120409215419.3288.50790.stgit@jf-dev1-dcblab> <20120409151501.23ff83f6@nehalam.linuxnetplumber.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: roprabhu@cisco.com, mst@redhat.com, stephen.hemminger@vyatta.com, davem@davemloft.net, hadi@cyberus.ca, bhutchings@solarflare.com, jeffrey.t.kirsher@intel.com, netdev@vger.kernel.org, gregory.v.rose@intel.com, krkumar2@in.ibm.com, sri@us.ibm.com To: Stephen Hemminger Return-path: Received: from mga01.intel.com ([192.55.52.88]:21207 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757223Ab2DIWcF (ORCPT ); Mon, 9 Apr 2012 18:32:05 -0400 In-Reply-To: <20120409151501.23ff83f6@nehalam.linuxnetplumber.net> Sender: netdev-owner@vger.kernel.org List-ID: On 4/9/2012 3:15 PM, Stephen Hemminger wrote: > On Mon, 09 Apr 2012 15:00:11 -0700 > John Fastabend wrote: > >> The following series is a submission for net-next to allow >> embedded switches and other stacked devices other then the >> Linux bridge to manage a forwarding database. >> >> This was previously posted here (where it was deferred for >> more review and testing) >> >> http://lists.openwall.net/netdev/2012/03/19/26 >> >> This series adds macvlan support per discussions with Roopa >> and Michael. Also ixgbe was updated to support adding multicast >> addresses per request from Greg Rose. >> >> Finally cleanups in the generic dump routines were added >> for multicast to support ixgbe and macvlan use cases. >> >> Thanks to everyone for the helpful review and comments. As >> always any comments/feedback welcome. >> >> .John >> >> --- >> >> Greg Rose (1): >> ixgbe: UTA table incorrectly programmed >> >> John Fastabend (6): >> macvlan: add FDB bridge ops and new macvlan mode >> ixgbe: allow RAR table to be updated in promisc mode >> ixgbe: enable FDB netdevice ops >> net: add fdb generic dump routine >> net: addr_list: add exclusive dev_uc_add and dev_mc_add >> net: add generic PF_BRIDGE:RTM_ FDB hooks >> >> >> drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 121 ++++++++++---- >> drivers/net/macvlan.c | 60 ++++++- >> include/linux/if_link.h | 1 >> include/linux/neighbour.h | 3 >> include/linux/netdevice.h | 28 +++ >> include/linux/rtnetlink.h | 4 >> net/bridge/br_device.c | 3 >> net/bridge/br_fdb.c | 128 ++++----------- >> net/bridge/br_netlink.c | 12 - >> net/bridge/br_private.h | 15 +- >> net/core/dev_addr_lists.c | 97 ++++++++++-- >> net/core/rtnetlink.c | 209 +++++++++++++++++++++++++ >> 12 files changed, 508 insertions(+), 173 deletions(-) >> > > How do statistics work in this case? What if you wanted to do BRIDGE MIB? I expect we will need to do the same sort of thing for PF_BRIDGE:RTM_GETLINK to dump the statistics for an embedded or stacked bridge. This series only gets the forwarding database setup. All the other entries in the MIB still need to be populated. What I would like is to be able to write a BRIDGE MIB implementation and have it work with any of the bridging components in the linux kernel. But I don't think that should block this series. Make sense? Based on a quick scan I'm not sure the existing net/bridge code has one place to pull all the stats from anyways looks like you might have to read some sysfs entries AND do some netlink commands. I would have to check that though. It is on my todo list to create a bridge mib for some of the EVB cases. Of course if someone else got to it that would be great. .John