From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH net-next] bonding: allow bond in mode balance-alb to work properly in bridge -try4.2 Date: Wed, 27 May 2009 16:39:22 +0200 Message-ID: <4A1D509A.5020205@cosmosbay.com> References: <20090313183303.GF3436@psychotron.englab.brq.redhat.com> <20090326155205.GA28868@psychotron.englab.brq.redhat.com> <20090526151717.GB11147@psychotron.englab.brq.redhat.com> <20090527135351.GD1652@psychotron.englab.brq.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: netdev@vger.kernel.org, jgarzik@pobox.com, davem@davemloft.net, shemminger@linux-foundation.org, bridge@lists.linux-foundation.org, fubar@us.ibm.com, bonding-devel@lists.sourceforge.net, kaber@trash.net, mschmidt@redhat.com, andy@greyhouse.net To: Jiri Pirko Return-path: Received: from gw1.cosmosbay.com ([212.99.114.194]:59199 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759849AbZE0Ok0 convert rfc822-to-8bit (ORCPT ); Wed, 27 May 2009 10:40:26 -0400 In-Reply-To: <20090527135351.GD1652@psychotron.englab.brq.redhat.com> Sender: netdev-owner@vger.kernel.org List-ID: Jiri Pirko a =E9crit : > [PATCH net-next] bonding: allow bond in mode balance-alb to work prop= erly in bridge -try4.2 >=20 > (updated) > changes v4.1 -> v4.2 > - use skb->pkt_type =3D=3D PACKET_HOST compare rather then comparing = skb dest addr > against skb->dev->dev_addr >=20 > Hi all. >=20 > The problem is described in following bugzilla: > https://bugzilla.redhat.com/show_bug.cgi?id=3D487763 >=20 > Basically here's what's going on. In every mode, bonding interface us= es the same > mac address for all enslaved devices (except fail_over_mac). Only bal= ance-alb > will simultaneously use multiple MAC addresses across different slave= s. When you > put this kind of bond device into a bridge it will only add one of ma= c adresses > into a hash list of mac addresses, say X. This mac address is marked = as local. > But this bonding interface also has mac address Y. Now then packet ar= rives with > destination address Y, this address is not marked as local and the pa= cked looks > like it needs to be forwarded. This packet is then lost which is wron= g. >=20 > Notice that interfaces can be added and removed from bond while it is= in bridge. >=20 > *** > When the multiple addresses for bridge port approach failed to solve = this issue > due to STP I started to think other way to solve this. I returned to = previous > solution but tweaked one. >=20 > This patch solves the situation in the bonding without touching bridg= e code. > For every incoming frame to bonding the destination address is compar= ed to > current address of the slave device from which tha packet came. If th= ese two > match destination address is replaced by mac address of the master. T= his address > is known by bridge so it is delivered properly. Note that the compars= ion is not > made directly, it's used skb->pkt_type =3D=3D PACKET_HOST instead. Th= is is "set" > previously in eth_type_trans(). >=20 > I experimentally tried that this works as good as searching through t= he slave > list (v4 of this patch). >=20 > I was forced to create a new header because I need to use > compare_ether_addr_64bits() (defined in linux/etherdevice.h) in > linux/netdevice.h. I've hit some cross include issues. I think that i= t's good > to have skb_bond_should_drop() in a separate file anyway. >=20 > Jirka >=20 >=20 > Signed-off-by: Jiri Pirko >=20 > diff --git a/include/linux/bonding.h b/include/linux/bonding.h > new file mode 100644 > index 0000000..e50939d > --- /dev/null > +++ b/include/linux/bonding.h > @@ -0,0 +1,78 @@ > +/* > + * include/linux/bonding.h > + * > + * Copyright (C) 2009 Jiri Pirko > + * > + * This program is free software; you can redistribute it and/or mod= ify > + * it under the terms of the GNU General Public License version 2 > + * as published by the Free Software Foundation. > + * > + * Bonding device helpers. > + */ > + > +#ifndef _LINUX_BONDING_H > +#define _LINUX_BONDING_H > + > +#ifdef __KERNEL__ > + > +#include > +#include > +#include > +#include > +#include > +#include > + > +static inline void skb_bond_set_mac_by_master(struct sk_buff *skb, > + struct net_device *master) > +{ > + unsigned char *dest =3D eth_hdr(skb)->h_dest; > + > + if (compare_ether_addr_64bits(dest, master->dev_addr) && > + (skb->pkt_type =3D=3D PACKET_HOST)) > + memcpy(dest, master->dev_addr, ETH_ALEN); Just overwriting the dest would be faster, and avoids to include , maybe a new include file could be avoided ? If it is already the master->dev_addr, then memcpy() is a no-op If it wasnt the master->dev_addr, then memcpy() does what you wanted. You can also give a hint to gcc as h_dest is guaranteed to be 16 bit al= igned static inline void skb_bond_set_mac_by_master(struct sk_buff *skb, struct net_device *master) { if (skb->pkt_type =3D=3D PACKET_HOST) { u16 *dest =3D (u16 *)eth_hdr(skb)->h_dest; memcpy(dest, master->dev_addr, ETH_ALEN); } } Compiler will emit better code for memcpy() on some arches. (not on x86, as it already does one 32bit and one 16bit move)