HMM_SOURCE_IP: 172.18.0.188:41400.2025081230
HMM_ATTACHE_NUM: 0000
HMM_SOURCE_TYPE: SMTP
Received: from clientip-120.244.206.96 (unknown [172.18.0.188])
        by chinatelecom.cn (HERMES) with SMTP id A71D3280098;
        Mon, 20 Dec 2021 23:09:01 +0800 (CST)
X-189-SAVE-TO-SEND: sunshouxin@chinatelecom.cn
Received: from  ([172.18.0.188])
        by app0023 with ESMTP id 448f2ce83d55476ebb2bb0b912c8404f for eric.dumazet@gmail.com;
        Mon, 20 Dec 2021 23:09:04 CST
X-Transaction-ID: 448f2ce83d55476ebb2bb0b912c8404f
X-Real-From: sunshouxin@chinatelecom.cn
X-Receive-IP: 172.18.0.188
X-MEDUSA-Status: 0
Sender: sunshouxin@chinatelecom.cn
Message-ID: <e5815ae7-e716-95ab-72ef-2dbe6554968d@chinatelecom.cn>
Date:   Mon, 20 Dec 2021 23:09:00 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Thunderbird/91.4.0
Subject: Re: [PATCH v3] net: bonding: Add support for IPV6 ns/na
To:     Eric Dumazet <eric.dumazet@gmail.com>, j.vosburgh@gmail.com,
        vfalico@gmail.com, andy@greyhouse.net, davem@davemloft.net,
        kuba@kernel.org
Cc:     netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        huyd12@chinatelecom.cn
References: <20211217164829.31388-1-sunshouxin@chinatelecom.cn>
 <f35ef160-ffdb-e6de-6079-440cd72ab389@gmail.com>
From:   =?UTF-8?B?5a2Z5a6I6ZGr?= <sunshouxin@chinatelecom.cn>
In-Reply-To: <f35ef160-ffdb-e6de-6079-440cd72ab389@gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org


在 2021/12/20 23:03, Eric Dumazet 写道:
>
> On 12/17/21 8:48 AM, Sun Shouxin wrote:
>> Since ipv6 neighbor solicitation and advertisement messages
>> isn't handled gracefully in bonding6 driver, we can see packet
>> drop due to inconsistency bewteen mac address in the option
>> message and source MAC .
>>
>> Another examples is ipv6 neighbor solicitation and advertisement
>> messages from VM via tap attached to host brighe, the src mac
>> mighe be changed through balance-alb mode, but it is not synced
>> with Link-layer address in the option message.
>>
>> The patch implements bond6's tx handle for ipv6 neighbor
>> solicitation and advertisement messages.
>>
>>                          Border-Leaf
>>                          /        \
>>                         /          \
>>                      Tunnel1    Tunnel2
>>                       /              \
>>                      /                \
>>                    Leaf-1--Tunnel3--Leaf-2
>>                      \                /
>>                       \              /
>>                        \            /
>>                         \          /
>>                         NIC1    NIC2
>>                          \      /
>>                          server
>>
>> We can see in our lab the Border-Leaf receives occasionally
>> a NA packet which is assigned to NIC1 mac in ND/NS option
>> message, but actaully send out via NIC2 mac due to tx-alb,
>> as a result, it will cause inconsistency between MAC table
>> and ND Table in Border-Leaf, i.e, NIC1 = Tunnel2 in ND table
>> and  NIC1 = Tunnel1 in mac table.
>>
>> And then, Border-Leaf starts to forward packet destinated
>> to the Server, it will only check the ND table entry in some
>> switch to encapsulate the destination MAC of the message as
>> NIC1 MAC, and then send it out from Tunnel2 by ND table.
>> Then, Leaf-2 receives the packet, it notices the destination
>> MAC of message is NIC1 MAC and should forword it to Tunne1
>> by Tunnel3.
>>
>> However, this traffic forward will be failure due to split
>> horizon of VxLAN tunnels.
>>
>> Suggested-by: Hu Yadi <huyd12@chinatelecom.cn>
>> Reviewed-by: Jay Vosburgh<jay.vosburgh@canonical.com>
>> Reviewed-by: Eric Dumazet<eric.dumazet@gmail.com>
>
>
> ??? I do not recall giving my Reviewed-by:
Thanks your comment, I'll adjust it and send out V4 soon.
>
>
>> Reported-by: kernel test robot <lkp@intel.com>
>> Signed-off-by: Sun Shouxin <sunshouxin@chinatelecom.cn>
>> ---
>>   drivers/net/bonding/bond_alb.c | 131 +++++++++++++++++++++++++++++++++
>>   1 file changed, 131 insertions(+)
>>
>> diff --git a/drivers/net/bonding/bond_alb.c 
>> b/drivers/net/bonding/bond_alb.c
>> index 533e476988f2..b14017364594 100644
>> --- a/drivers/net/bonding/bond_alb.c
>> +++ b/drivers/net/bonding/bond_alb.c
>> @@ -22,6 +22,7 @@
>>   #include <asm/byteorder.h>
>>   #include <net/bonding.h>
>>   #include <net/bond_alb.h>
>> +#include <net/ndisc.h>
>>     static const u8 mac_v6_allmcast[ETH_ALEN + 2] __long_aligned = {
>>       0x33, 0x33, 0x00, 0x00, 0x00, 0x01
>> @@ -1269,6 +1270,119 @@ static int alb_set_mac_address(struct bonding 
>> *bond, void *addr)
>>       return res;
>>   }
>>   +/*determine if the packet is NA or NS*/
>> +static bool alb_determine_nd(struct icmp6hdr *hdr)
>> +{
>> +    if (hdr->icmp6_type == NDISC_NEIGHBOUR_ADVERTISEMENT ||
>> +        hdr->icmp6_type == NDISC_NEIGHBOUR_SOLICITATION) {
>> +        return true;
>> +    }
>> +
>> +    return false;
>> +}
>> +
>> +static void alb_change_nd_option(struct sk_buff *skb, void *data)
>> +{
>> +    struct nd_msg *msg = (struct nd_msg *)skb_transport_header(skb);
>> +    struct nd_opt_hdr *nd_opt = (struct nd_opt_hdr *)msg->opt;
>> +    struct net_device *dev = skb->dev;
>> +    struct icmp6hdr *icmp6h = icmp6_hdr(skb);
>> +    struct ipv6hdr *ip6hdr = ipv6_hdr(skb);
>> +    u8 *lladdr = NULL;
>> +    u32 ndoptlen = skb_tail_pointer(skb) - (skb_transport_header(skb) +
>> +                offsetof(struct nd_msg, opt));
>> +
>> +    while (ndoptlen) {
>> +        int l;
>> +
>> +        switch (nd_opt->nd_opt_type) {
>> +        case ND_OPT_SOURCE_LL_ADDR:
>> +        case ND_OPT_TARGET_LL_ADDR:
>> +        lladdr = ndisc_opt_addr_data(nd_opt, dev);
>> +        break;
>> +
>> +        default:
>> +        lladdr = NULL;
>> +        break;
>> +        }
>> +
>> +        l = nd_opt->nd_opt_len << 3;
>> +
>> +        if (ndoptlen < l || l == 0)
>> +            return;
>> +
>> +        if (lladdr) {
>> +            memcpy(lladdr, data, dev->addr_len);
>> +            icmp6h->icmp6_cksum = 0;
>> +
>> +            icmp6h->icmp6_cksum = csum_ipv6_magic(&ip6hdr->saddr,
>> +                                  &ip6hdr->daddr,
>> +                        ntohs(ip6hdr->payload_len),
>> +                        IPPROTO_ICMPV6,
>> +                        csum_partial(icmp6h,
>> +                                 ntohs(ip6hdr->payload_len), 0));
>> +        }
>> +        ndoptlen -= l;
>> +        nd_opt = ((void *)nd_opt) + l;
>> +    }
>> +}
>> +
>> +static u8 *alb_get_lladdr(struct sk_buff *skb)
>> +{
>> +    struct nd_msg *msg = (struct nd_msg *)skb_transport_header(skb);
>> +    struct nd_opt_hdr *nd_opt = (struct nd_opt_hdr *)msg->opt;
>> +    struct net_device *dev = skb->dev;
>> +    u8 *lladdr = NULL;
>> +    u32 ndoptlen = skb_tail_pointer(skb) - (skb_transport_header(skb) +
>> +                offsetof(struct nd_msg, opt));
>> +
>> +    while (ndoptlen) {
>> +        int l;
>> +
>> +        switch (nd_opt->nd_opt_type) {
>> +        case ND_OPT_SOURCE_LL_ADDR:
>> +        case ND_OPT_TARGET_LL_ADDR:
>> +            lladdr = ndisc_opt_addr_data(nd_opt, dev);
>> +            break;
>> +
>> +        default:
>> +            break;
>> +        }
>> +
>> +        l = nd_opt->nd_opt_len << 3;
>> +
>> +        if (ndoptlen < l || l == 0)
>> +            return NULL;
>> +
>> +        if (lladdr)
>> +            return lladdr;
>> +
>> +        ndoptlen -= l;
>> +        nd_opt = ((void *)nd_opt) + l;
>> +    }
>> +
>> +    return lladdr;
>> +}
>> +
>> +static void alb_set_nd_option(struct sk_buff *skb, struct bonding 
>> *bond,
>> +                  struct slave *tx_slave)
>> +{
>> +    struct ipv6hdr *ip6hdr;
>> +    struct icmp6hdr *hdr = NULL;
>> +
>> +    if (skb->protocol == htons(ETH_P_IPV6)) {
>> +        if (tx_slave && tx_slave !=
>> +            rcu_access_pointer(bond->curr_active_slave)) {
>> +            ip6hdr = ipv6_hdr(skb);
>> +            if (ip6hdr->nexthdr == IPPROTO_ICMPV6) {
>> +                hdr = icmp6_hdr(skb);
>> +                if (alb_determine_nd(hdr))
>> +                    alb_change_nd_option(skb, tx_slave->dev->dev_addr);
>> +            }
>> +        }
>> +    }
>> +}
>> +
>>   /************************ exported alb functions 
>> ************************/
>>     int bond_alb_initialize(struct bonding *bond, int rlb_enabled)
>> @@ -1415,6 +1529,7 @@ struct slave *bond_xmit_alb_slave_get(struct 
>> bonding *bond,
>>       }
>>       case ETH_P_IPV6: {
>>           const struct ipv6hdr *ip6hdr;
>> +        struct icmp6hdr *hdr = NULL;
>>             /* IPv6 doesn't really use broadcast mac address, but leave
>>            * that here just in case.
>> @@ -1446,6 +1561,21 @@ struct slave *bond_xmit_alb_slave_get(struct 
>> bonding *bond,
>>               break;
>>           }
>>   +        if (ip6hdr->nexthdr == IPPROTO_ICMPV6) {
>> +            hdr = icmp6_hdr(skb);
>> +            if (alb_determine_nd(hdr)) {
>> +                u8 *lladdr = NULL;
>> +
>> +                lladdr = alb_get_lladdr(skb);
>> +                if (lladdr) {
>> +                    if (!bond_slave_has_mac_rx(bond, lladdr)) {
>> +                        do_tx_balance = false;
>> +                        break;
>> +                    }
>> +                }
>> +            }
>> +        }
>> +
>>           hash_start = (char *)&ip6hdr->daddr;
>>           hash_size = sizeof(ip6hdr->daddr);
>>           break;
>> @@ -1489,6 +1619,7 @@ netdev_tx_t bond_alb_xmit(struct sk_buff *skb, 
>> struct net_device *bond_dev)
>>       struct slave *tx_slave = NULL;
>>         tx_slave = bond_xmit_alb_slave_get(bond, skb);
>> +    alb_set_nd_option(skb, bond, tx_slave);
>>       return bond_do_alb_xmit(skb, bond, tx_slave);
>>   }
>>
>> base-commit: 6441998e2e37131b0a4c310af9156d79d3351c16