Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Received-SPF: Pass (protection.outlook.com: domain of nvidia.com designates
 12.22.5.236 as permitted sender) receiver=protection.outlook.com;
 client-ip=12.22.5.236; helo=mail.nvidia.com;
Message-ID: <d0afa956-6852-2749-fce8-2a3e06cae556@nvidia.com>
Date:   Tue, 25 Jan 2022 10:51:32 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.4.0
Subject: Re: [PATCH v9] net: bonding: Add support for IPV6 ns/na to
 balance-alb/balance-tlb mode
Content-Language: en-US
To:     Sun Shouxin <sunshouxin@chinatelecom.cn>, <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>,
        <jay.vosburgh@canonical.com>, <huyd12@chinatelecom.cn>
References: <20220125023755.94837-1-sunshouxin@chinatelecom.cn>
From:   Nikolay Aleksandrov <nikolay@nvidia.com>
In-Reply-To: <20220125023755.94837-1-sunshouxin@chinatelecom.cn>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 25 Jan 2022 08:51:41.5430
 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 6475981e-8b65-45d9-05a6-08d9dfdfe847
X-MS-Exchange-CrossTenant-Id: 43083d15-7273-40c1-b7db-39efd9ccc17a
X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=43083d15-7273-40c1-b7db-39efd9ccc17a;Ip=[12.22.5.236];Helo=[mail.nvidia.com]
X-MS-Exchange-CrossTenant-AuthSource: BN8NAM11FT011.eop-nam11.prod.protection.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BN6PR12MB1585
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 25/01/2022 04:37, Sun Shouxin wrote:
> 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>
> ---
>  drivers/net/bonding/bond_alb.c | 38 +++++++++++++++++++++++++++++++++-
>  1 file changed, 37 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
> index 533e476988f2..ba7cc1a9bf6c 100644
> --- a/drivers/net/bonding/bond_alb.c
> +++ b/drivers/net/bonding/bond_alb.c
> @@ -1269,6 +1269,34 @@ 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;
> +
> +	ip6hdr = ipv6_hdr(skb);

You can't do this in bond_xmit_tlb_slave_get(), there's no check if the IPv6 header
is in the linear part before calling alb_determine_nd() there. You can combine
bond_xmit_alb_slave_get's IPv6 header pull with alb_determine_nd's, just please leave
a comment above the call in bond_xmit_alb_slave_get() that the IPv6 header is pulled by
alb_determine_nd.

> +	if (ip6hdr->nexthdr == IPPROTO_ICMPV6) {
> +		if (!pskb_may_pull(skb, sizeof(struct ipv6hdr) + sizeof(struct icmp6hdr)))
> +			return true;

This could invalidate current pointers to skb data, more below...

> +
> +		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 +1376,11 @@ 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,
> @@ -1446,6 +1477,11 @@ struct slave *bond_xmit_alb_slave_get(struct bonding *bond,
>  			break;
>  		}
>  
> +		if (alb_determine_nd(skb, bond)) {
> +			do_tx_balance = false;
> +			break;
> +		}
> +
>  		hash_start = (char *)&ip6hdr->daddr;
>  		hash_size = sizeof(ip6hdr->daddr);

... here you have to reload ip6hdr, but that can be avoided in a few different ways.
I.e. you could move alb_determine_nd() before the assignment of that ptr.

>  		break;
> 
> base-commit: dd81e1c7d5fb126e5fbc5c9e334d7b3ec29a16a0