[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <a88f6b99-7195-eee3-0cd1-e5e31c9f946f@gmail.com>
Date: Tue, 26 Mar 2019 07:54:45 -0700
From: Eric Dumazet <eric.dumazet@...il.com>
To: Jesper Dangaard Brouer <brouer@...hat.com>,
Ioana Ciornei <ioana.ciornei@....com>
Cc: "davem@...emloft.net" <davem@...emloft.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Ioana Ciocoi Radulescu <ruxandra.radulescu@....com>,
Florian Westphal <fw@...len.de>,
Edward Cree <ecree@...arflare.com>
Subject: Re: [PATCH net-next] dpaa2-eth: use netif_receive_skb_list
On 03/26/2019 07:43 AM, Jesper Dangaard Brouer wrote:
> On Mon, 25 Mar 2019 13:42:39 +0000
> Ioana Ciornei <ioana.ciornei@....com> wrote:
>
>> Take advantage of the software Rx batching by using
>> netif_receive_skb_list instead of napi_gro_receive.
>>
>> Signed-off-by: Ioana Ciornei <ioana.ciornei@....com>
>> ---
>
> Nice to see more people/drivers using: netif_receive_skb_list()
>
> We should likely add a similar napi_gro_receive_list() function.
GRO is essentially processing one frame at a time, and recycling
one sk_buff. (one sk_buff per aggregated train)
Having to allocate one sk_buff per MSS just to provide a list
would likely be slower, as far as GRO is concerned.
Powered by blists - more mailing lists