From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wengang Wang Subject: Re: [PATCH] net: take care of bonding in build_skb_flow_key (v3) Date: Thu, 21 Jan 2016 10:40:34 +0800 Message-ID: <56A04522.9010300@oracle.com> References: <1453267933-25381-1-git-send-email-wen.gang.wang@oracle.com> <569F2806.70608@gmail.com> <569F397B.7010808@oracle.com> <569F3D19.8020307@gmail.com> <569F579F.1060708@oracle.com> <569F59D8.2060200@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: jay.vosburgh@canonical.com To: zhuyj , netdev@vger.kernel.org Return-path: Received: from userp1040.oracle.com ([156.151.31.81]:41625 "EHLO userp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751499AbcAUChQ (ORCPT ); Wed, 20 Jan 2016 21:37:16 -0500 In-Reply-To: <569F59D8.2060200@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: =E5=9C=A8 2016=E5=B9=B401=E6=9C=8820=E6=97=A5 17:56, zhuyj =E5=86=99=E9= =81=93: > On 01/20/2016 05:47 PM, Wengang Wang wrote: >> >> >> =E5=9C=A8 2016=E5=B9=B401=E6=9C=8820=E6=97=A5 15:54, zhuyj =E5=86=99= =E9=81=93: >>> On 01/20/2016 03:38 PM, Wengang Wang wrote: >>>> >>>> >>>> =E5=9C=A8 2016=E5=B9=B401=E6=9C=8820=E6=97=A5 14:24, zhuyj =E5=86=99= =E9=81=93: >>>>> On 01/20/2016 01:32 PM, Wengang Wang wrote: >>>>>> In a bonding setting, we determines fragment size according to=20 >>>>>> MTU and >>>>>> PMTU associated to the bonding master. If the slave finds the=20 >>>>>> fragment >>>>>> size is too big, it drops the fragment and calls=20 >>>>>> ip_rt_update_pmtu(), >>>>>> passing _skb_ and _pmtu_, trying to update the path MTU. >>>>>> Problem is that the target device that function ip_rt_update_pmt= u=20 >>>>>> actually >>>>>> tries to update is the slave (skb->dev), not the master. Thus=20 >>>>>> since no >>>>>> PMTU change happens on master, the fragment size for later=20 >>>>>> packets doesn't >>>>>> change so all later fragments/packets are dropped too. >>>>>> >>>>>> The fix is letting build_skb_flow_key() take care of the=20 >>>>>> transition of >>>>>> device index from bonding slave to the master. That makes the=20 >>>>>> master become >>>>>> the target device that ip_rt_update_pmtu tries to update PMTU to= =2E >>>>>> >>>>>> Signed-off-by: Wengang Wang >>>>>> --- >>>>>> net/ipv4/route.c | 13 ++++++++++++- >>>>>> 1 file changed, 12 insertions(+), 1 deletion(-) >>>>>> >>>>>> diff --git a/net/ipv4/route.c b/net/ipv4/route.c >>>>>> index 85f184e..c59fb0d 100644 >>>>>> --- a/net/ipv4/route.c >>>>>> +++ b/net/ipv4/route.c >>>>>> @@ -523,10 +523,21 @@ static void build_skb_flow_key(struct=20 >>>>>> flowi4 *fl4, const struct sk_buff *skb, >>>>>> const struct sock *sk) >>>>>> { >>>>>> const struct iphdr *iph =3D ip_hdr(skb); >>>>>> - int oif =3D skb->dev->ifindex; >>>>>> + struct net_device *master =3D NULL; >>>>>> u8 tos =3D RT_TOS(iph->tos); >>>>>> u8 prot =3D iph->protocol; >>>>>> u32 mark =3D skb->mark; >>>>>> + int oif; >>>>>> + >>>>>> + if (skb->dev->flags & IFF_SLAVE) { >>>>>> + rtnl_lock(); >>>>>> + master =3D netdev_master_upper_dev_get(skb->dev); >>>>>> + rtnl_unlock(); >>>>> update_pmtu is called very frequently. Is it appropriate to use=20 >>>>> rtnl_lock here? >>>> By "very frequently", how frequently it is expected? And what=20 >>>> situation can cause that? >>>> For my case, the update_pmtu is called only once. >>> ip_tunnel_xmit >>> >> Can you please explain with more details? > > dev_queue_xmit->ipip_tunnel_xmit->ip_tunnel_xmit->tnl_update_pmtu->=20 > skb_dst(skb)->ops->update_pmtu =46or ipip, yes seems update_pmtu is called in line for each call of=20 queue_xmit. Do you know if it's a good configuration for ipip + bondin= g? Other's comment and suggestion? thanks, wengang