Re: [PATCH net 1/4] bonding: implement xfrm state xdo_dev_state_free API

[Hangbin Liu <liuhangbin@gmail.com>]

On Mon, Jul 29, 2024 at 03:44:02PM +0300, Tariq Toukan wrote:
> From: Jianbo Liu <jianbol@nvidia.com>
> 
> Add this implementation for bonding, so hardware resources can be
> freed after xfrm state is deleted.
> 
> Fixes: 9a5605505d9c ("bonding: Add struct bond_ipesc to manage SA")
> Signed-off-by: Jianbo Liu <jianbol@nvidia.com>
> Reviewed-by: Cosmin Ratiu <cratiu@nvidia.com>
> Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
> ---
>  drivers/net/bonding/bond_main.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 1cd92c12e782..3b880ff2b82a 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -588,6 +588,15 @@ static void bond_ipsec_del_sa_all(struct bonding *bond)
>  	rcu_read_unlock();
>  }
>  
> +static void bond_ipsec_free_sa(struct xfrm_state *xs)
> +{
> +	struct net_device *real_dev = xs->xso.real_dev;

I think it's also good to check the bond/slave status like bond_ipsec_del_sa()
does, no?

> +
> +	if (real_dev && real_dev->xfrmdev_ops &&
> +	    real_dev->xfrmdev_ops->xdo_dev_state_free)
> +		real_dev->xfrmdev_ops->xdo_dev_state_free(xs);
> +}

And we should call this in bond_ipsec_del_sa_all() for each slave.

Thanks
Hangbin
> +
>  /**
>   * bond_ipsec_offload_ok - can this packet use the xfrm hw offload
>   * @skb: current data packet
> @@ -632,6 +641,7 @@ static bool bond_ipsec_offload_ok(struct sk_buff *skb, struct xfrm_state *xs)
>  static const struct xfrmdev_ops bond_xfrmdev_ops = {
>  	.xdo_dev_state_add = bond_ipsec_add_sa,
>  	.xdo_dev_state_delete = bond_ipsec_del_sa,
> +	.xdo_dev_state_free = bond_ipsec_free_sa,
>  	.xdo_dev_offload_ok = bond_ipsec_offload_ok,
>  };
>  #endif /* CONFIG_XFRM_OFFLOAD */
> -- 
> 2.44.0
>