From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ding Tianhong Subject: Re: [PATCH] bonding: Inactive slaves should keep inactive flag's value to 1 in tlb and alb mode. Date: Wed, 26 Mar 2014 08:53:14 +0800 Message-ID: <533224FA.6040307@huawei.com> References: <1395716407-31601-1-git-send-email-zheng.x.li@oracle.com> <5330FB34.4090005@huawei.com> <53314017.8040104@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: , , , , , To: "zheng.li" Return-path: In-Reply-To: <53314017.8040104@oracle.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On 2014/3/25 16:36, zheng.li wrote: > =E4=BA=8E 2014=E5=B9=B403=E6=9C=8825=E6=97=A5 11:42, Ding Tianhong =E5= =86=99=E9=81=93: >> On 2014/3/25 11:00, Zheng Li wrote: >>> In bond mode tlb and alb, inactive slaves should keep inactive flag= to >>> 1 to refuse to receive broadcast packets. Now, active slave send br= oadcast packets >>> (for example ARP requests) which will arrive inactive slaves on sam= e host from switch, >>> but inactive slave's inactive flag is zero that cause bridge receiv= e the broadcast >>> packets to produce a wrong entry in forward table. Typical situatio= n is domu send some >>> ARP request which go out from dom0 bond's active slave, then the AR= P broadcast request >>> packets go back to inactive slave from switch, because the inactive= slave's inactive >>> flag is zero, kernel will receive the packets and pass them to brid= ge, that cause dom0's >>> bridge map domu's MAC address to port of bond, bridge should map do= mu's MAC to port of vif. >>> >>> Signed-off-by: Zheng Li >>> --- >>> drivers/net/bonding/bond_main.c | 2 +- >>> 1 files changed, 1 insertions(+), 1 deletions(-) >>> >>> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/= bond_main.c >>> index e5628fc..8761df6 100644 >>> --- a/drivers/net/bonding/bond_main.c >>> +++ b/drivers/net/bonding/bond_main.c >>> @@ -3062,7 +3062,7 @@ static int bond_open(struct net_device *bond_= dev) >>> && (slave !=3D bond->curr_active_slave)) { >>> bond_set_slave_inactive_flags(slave, >>> BOND_SLAVE_NOTIFY_NOW); >>> - } else { >>> + } else if (!bond_is_lb(bond)) { >>> bond_set_slave_active_flags(slave, >>> BOND_SLAVE_NOTIFY_NOW); >>> } >>> >> I think you did not fix the problem completely, the state monitor wi= ll change the status for the slaves >> and the inactive slave still could receive the broadcast. >=20 > Had tested, it can fix the issue, verified by our QA. > Default set slave of bond as inactive when add a slave to bond, when > link UP, just set one slave as current active slave and clear its > inactive flag, the inactive slave's inactive flag will keep the value= of 1. >=20 >=20 Ok, I found that in the mii monitor, it will only change the backup sta= te, no problem, it looks good to me. Ding >> >> Regards >> Ding >> >> -- >> To unsubscribe from this list: send the line "unsubscribe netdev" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> >=20 >=20 >=20 > . >=20