Fri, Jan 22, 2016 at 05:21:28AM CET, wen.gang.w...@oracle.com wrote:
>
>
>在 2016年01月21日 16:35, Jiri Pirko 写道:
>>Thu, Jan 21, 2016 at 06:32:58AM CET, wen.gang.w...@oracle.com wrote:
>>>In a bonding setting, we determines fragment size according to MTU and
>>>PMTU associated to the bonding master. If the slave finds the fragment
>>>size is too big, it drops the fragment and calls 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_pmtu actually
>>>tries to update is the slave (skb->dev), not the master. Thus since no
>>>PMTU change happens on master, the fragment size for later packets doesn't
>>>change so all later fragments/packets are dropped too.
>>>
>>>The fix is letting build_skb_flow_key() take care of the transition of
>>>device index from bonding slave to the master. That makes the master become
>>>the target device that ip_rt_update_pmtu tries to update PMTU to.
>>>
>>>Signed-off-by: Wengang Wang <wen.gang.w...@oracle.com>
>>>---
>>>net/ipv4/route.c | 9 +++++++++
>>>1 file changed, 9 insertions(+)
>>>
>>>diff --git a/net/ipv4/route.c b/net/ipv4/route.c
>>>index 85f184e..7e766b5 100644
>>>--- a/net/ipv4/route.c
>>>+++ b/net/ipv4/route.c
>>>@@ -524,10 +524,19 @@ static void build_skb_flow_key(struct flowi4 *fl4,
>>>const struct sk_buff *skb,
>>>{
>>> const struct iphdr *iph = ip_hdr(skb);
>>> int oif = skb->dev->ifindex;
>>>+ struct net_device *master;
>>> u8 tos = RT_TOS(iph->tos);
>>> u8 prot = iph->protocol;
>>> u32 mark = skb->mark;
>>>
>>>+ if (netif_is_bond_slave(skb->dev)) {
>>>+ rcu_read_lock();
>>>+ master = netdev_master_upper_dev_get_rcu(skb->dev);
>>>+ if (master)
>>>+ oif = master->ifindex;
>>>+ rcu_read_unlock();
>>>+ }
>>This is certainly not correct as it should not be bond-specific but
>>rather generic.
>
>Then what you would suggest to fix it?
>>Note that you may have bond over bond or bridge over
>>bond or other scenarios, which this patch ignores.
>I don't think bond over bond is a good configuration. Do you have a real use
>case for that configuration?
Stacking of multiple master devices is absolutelly common.
You have to go in the upper tree all the way up, for all master device
types.
>
>thanks,
>wengang
>
