From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andy Gospodarek Subject: Re: [PATCH] bonding: only use primary address for ARP Date: Wed, 30 Nov 2011 17:07:10 -0500 Message-ID: <20111130220709.GI25132@gospo.rdu.redhat.com> References: <20111124093715.GA25365@as68121.uab.ericsson.se> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netdev@vger.kernel.org, fubar@us.ibm.com To: Henrik Saavedra Persson Return-path: Received: from mx1.redhat.com ([209.132.183.28]:6098 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751628Ab1K3WHP (ORCPT ); Wed, 30 Nov 2011 17:07:15 -0500 Content-Disposition: inline In-Reply-To: <20111124093715.GA25365@as68121.uab.ericsson.se> Sender: netdev-owner@vger.kernel.org List-ID: On Thu, Nov 24, 2011 at 10:37:15AM +0100, Henrik Saavedra Persson wrote: > Only use the primary address of the bond device > for master_ip. This will prevent changing the ARP source > address in Active-Backup mode whenever a secondry address > is added to the bond device. > > Signed-off-by: Henrik Saavedra Persson > --- > drivers/net/bonding/bond_main.c | 33 ++++++--------------------------- > 1 files changed, 6 insertions(+), 27 deletions(-) > > diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c > index 6191e63..ec293b3 100644 > --- a/drivers/net/bonding/bond_main.c > +++ b/drivers/net/bonding/bond_main.c > @@ -2551,30 +2551,6 @@ out: > } > } > > -static __be32 bond_glean_dev_ip(struct net_device *dev) > -{ > - struct in_device *idev; > - struct in_ifaddr *ifa; > - __be32 addr = 0; > - > - if (!dev) > - return 0; > - > - rcu_read_lock(); > - idev = __in_dev_get_rcu(dev); > - if (!idev) > - goto out; > - > - ifa = idev->ifa_list; > - if (!ifa) > - goto out; > - > - addr = ifa->ifa_local; > -out: > - rcu_read_unlock(); > - return addr; > -} > - > static int bond_has_this_ip(struct bonding *bond, __be32 ip) > { > struct vlan_entry *vlan; > @@ -3316,6 +3292,10 @@ static int bond_inetaddr_event(struct notifier_block *this, unsigned long event, > struct bonding *bond; > struct vlan_entry *vlan; > > + /* we only care about primary address */ > + if(ifa->ifa_flags & IFA_F_SECONDARY) > + return NOTIFY_DONE; > + This change is good. We need to stop overwriting bond->master_ip when secondary addresses are added. > list_for_each_entry(bond, &bn->dev_list, bond_list) { > if (bond->dev == event_dev) { > switch (event) { > @@ -3323,7 +3303,7 @@ static int bond_inetaddr_event(struct notifier_block *this, unsigned long event, > bond->master_ip = ifa->ifa_local; > return NOTIFY_OK; > case NETDEV_DOWN: > - bond->master_ip = bond_glean_dev_ip(bond->dev); > + bond->master_ip = 0; > return NOTIFY_OK; > default: > return NOTIFY_DONE; > @@ -3339,8 +3319,7 @@ static int bond_inetaddr_event(struct notifier_block *this, unsigned long event, > vlan->vlan_ip = ifa->ifa_local; > return NOTIFY_OK; > case NETDEV_DOWN: > - vlan->vlan_ip = > - bond_glean_dev_ip(vlan_dev); > + vlan->vlan_ip = 0; > return NOTIFY_OK; > default: > return NOTIFY_DONE; Inspection indicates this is safe too. The only code that appears to check against master_up and vlan_ip appears to be in ARP monitoring code and it does not make much sense to use ARP monitoring when your host does not have an IP address. Signed-off-by: Andy Gospodarek