Since ipv6 neighbor solicitation and advertisement messages
isn't handled gracefully in bond6 driver, we can see packet
drop due to inconsistency between 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 bridge, the src mac
might 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.

Suggested-by: Hu Yadi <huyd12@chinatelecom.cn>
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: Sun Shouxin <sunshouxin@chinatelecom.cn>
---
v9->v10:
- add IPv6 header pull in alb_determine_nd.
- combine bond_xmit_alb_slave_get's IPv6 header
pull with alb_determine_nd's
---
 drivers/net/bonding/bond_alb.c | 40 ++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)

diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
index 533e476988f2..d9da6eb7f5c2 100644
--- a/drivers/net/bonding/bond_alb.c
+++ b/drivers/net/bonding/bond_alb.c
@@ -1269,6 +1269,37 @@ 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 bool alb_determine_nd(struct sk_buff *skb, struct bonding *bond)
+{
+	struct ipv6hdr *ip6hdr;
+	struct icmp6hdr *hdr;
+
+	if (!pskb_network_may_pull(skb, sizeof(*ip6hdr)))
+		return true;
+
+	ip6hdr = ipv6_hdr(skb);
+	if (ip6hdr->nexthdr == IPPROTO_ICMPV6) {
+		if (!pskb_may_pull(skb, sizeof(*ip6hdr) + sizeof(*hdr)))
+			return true;
+
+		hdr = icmp6_hdr(skb);
+		return __alb_determine_nd(hdr);
+	}
+
+	return false;
+}
+
 /************************ exported alb functions ************************/
 
 int bond_alb_initialize(struct bonding *bond, int rlb_enabled)
@@ -1348,8 +1379,11 @@ struct slave *bond_xmit_tlb_slave_get(struct bonding *bond,
 	/* Do not TX balance any multicast or broadcast */
 	if (!is_multicast_ether_addr(eth_data->h_dest)) {
 		switch (skb->protocol) {
-		case htons(ETH_P_IP):
 		case htons(ETH_P_IPV6):
+			if (alb_determine_nd(skb, bond))
+				break;
+			fallthrough;
+		case htons(ETH_P_IP):
 			hash_index = bond_xmit_hash(bond, skb);
 			if (bond->params.tlb_dynamic_lb) {
 				tx_slave = tlb_choose_channel(bond,
@@ -1432,10 +1466,12 @@ struct slave *bond_xmit_alb_slave_get(struct bonding *bond,
 			break;
 		}
 
-		if (!pskb_network_may_pull(skb, sizeof(*ip6hdr))) {
+		if (alb_determine_nd(skb, bond)) {
 			do_tx_balance = false;
 			break;
 		}
+
+		/* The IPv6 header is pulled by alb_determine_nd */
 		/* Additionally, DAD probes should not be tx-balanced as that
 		 * will lead to false positives for duplicate addresses and
 		 * prevent address configuration from working.

base-commit: dd81e1c7d5fb126e5fbc5c9e334d7b3ec29a16a0
-- 
2.27.0

On Tue, 25 Jan 2022 09:24:18 -0500 Sun Shouxin wrote:
> +/* 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 bool alb_determine_nd(struct sk_buff *skb, struct bonding *bond)
> +{
> +	struct ipv6hdr *ip6hdr;
> +	struct icmp6hdr *hdr;
> +
> +	if (!pskb_network_may_pull(skb, sizeof(*ip6hdr)))
> +		return true;
> +
> +	ip6hdr = ipv6_hdr(skb);
> +	if (ip6hdr->nexthdr == IPPROTO_ICMPV6) {

	if (ip6hdr->nexthdr != IPPROTO_ICMPV6)
		return false;

This way there's no need to indent the rest of the function.

> +		if (!pskb_may_pull(skb, sizeof(*ip6hdr) + sizeof(*hdr)))

What happened to the _network part? pskb_network_may_pull(), right?

> +			return true;
> +
> +		hdr = icmp6_hdr(skb);
> +		return __alb_determine_nd(hdr);

Why create a full helper for this condition? Why not just:

	return hdr->icmp6_type == NDISC_NEIGHBOUR_ADVERTISEMENT ||
	       hdr->icmp6_type == NDISC_NEIGHBOUR_SOLICITATION;


> +	}
> +
> +	return false;
> +}

Thanks your comment, I'll adjust it and send out v11 soon.


在 2022/1/28 10:47, Jakub Kicinski 写道:
> On Tue, 25 Jan 2022 09:24:18 -0500 Sun Shouxin wrote:
>> +/* 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 bool alb_determine_nd(struct sk_buff *skb, struct bonding *bond)
>> +{
>> +	struct ipv6hdr *ip6hdr;
>> +	struct icmp6hdr *hdr;
>> +
>> +	if (!pskb_network_may_pull(skb, sizeof(*ip6hdr)))
>> +		return true;
>> +
>> +	ip6hdr = ipv6_hdr(skb);
>> +	if (ip6hdr->nexthdr == IPPROTO_ICMPV6) {
> 	if (ip6hdr->nexthdr != IPPROTO_ICMPV6)
> 		return false;
>
> This way there's no need to indent the rest of the function.
>
>> +		if (!pskb_may_pull(skb, sizeof(*ip6hdr) + sizeof(*hdr)))
> What happened to the _network part? pskb_network_may_pull(), right?
>
>> +			return true;
>> +
>> +		hdr = icmp6_hdr(skb);
>> +		return __alb_determine_nd(hdr);
> Why create a full helper for this condition? Why not just:
>
> 	return hdr->icmp6_type == NDISC_NEIGHBOUR_ADVERTISEMENT ||
> 	       hdr->icmp6_type == NDISC_NEIGHBOUR_SOLICITATION;
>
>
>> +	}
>> +
>> +	return false;
>> +}