[Intel-wired-lan] [net] i40e: Fix receive buffer starvation for AF_XDP

[Jacob Keller]

On 1/5/2020 12:46 AM, Jeff Kirsher wrote:
> Magnus's fix to resolve a potential receive buffer starvation for AF_XDP
> got applied to both the i40e_xsk_umem_enable/disable() functions, when it
> should have only been applied to the "enable".  So clean up the undesired
> code in the disable function.
> 
> CC: Magnus Karlsson <magnus.karlsson at intel.com>
> Fixes: 1f459bdc2007 ("i40e: fix potential RX buffer starvation for AF_XDP")
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher at intel.com>
> ---
>  drivers/net/ethernet/intel/i40e/i40e_xsk.c | 5 -----
>  1 file changed, 5 deletions(-)
> 
> diff --git a/drivers/net/ethernet/intel/i40e/i40e_xsk.c b/drivers/net/ethernet/intel/i40e/i40e_xsk.c
> index 42058fad6a3c..1b13d8e0d67a 100644
> --- a/drivers/net/ethernet/intel/i40e/i40e_xsk.c
> +++ b/drivers/net/ethernet/intel/i40e/i40e_xsk.c
> @@ -114,11 +114,6 @@ static int i40e_xsk_umem_enable(struct i40e_vsi *vsi, struct xdp_umem *umem,

Umm.. Based on the commit message, you meant to remove only from the
disable function, not the enable one...

And indeed you actually did in October:

2c19e395e061 ("i40e: Fix receive buffer starvation for AF_XDP")
Date:   Mon Oct 7 15:07:24 2019 -0700

I think this patch should be dropped from your tree.

(Thanks Lihong for pointing this out to me!)

Thanks,
Jake

>  		err = i40e_queue_pair_enable(vsi, qid);
>  		if (err)
>  			return err;
> -
> -		/* Kick start the NAPI context so that receiving will start */
> -		err = i40e_xsk_wakeup(vsi->netdev, qid, XDP_WAKEUP_RX);
> -		if (err)
> -			return err;
>  	}
>  
>  	return 0;
>