[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <cc3d1319-d67e-3031-9351-95b45af797d4@kernel.org>
Date: Fri, 11 Aug 2023 17:08:55 +0200
From: Jesper Dangaard Brouer <hawk@...nel.org>
To: Wei Fang <wei.fang@....com>,
Jesper Dangaard Brouer <hawk@...nel.org>,
"davem@...emloft.net" <davem@...emloft.net>,
"edumazet@...gle.com" <edumazet@...gle.com>,
"kuba@...nel.org" <kuba@...nel.org>,
"pabeni@...hat.com" <pabeni@...hat.com>,
Shenwei Wang <shenwei.wang@....com>,
Clark Wang <xiaoning.wang@....com>,
"ast@...nel.org" <ast@...nel.org>,
"daniel@...earbox.net" <daniel@...earbox.net>,
"john.fastabend@...il.com" <john.fastabend@...il.com>,
"larysa.zaremba@...el.com" <larysa.zaremba@...el.com>,
"aleksander.lobakin@...el.com" <aleksander.lobakin@...el.com>,
"jbrouer@...hat.com" <jbrouer@...hat.com>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Cc: dl-linux-imx <linux-imx@....com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"bpf@...r.kernel.org" <bpf@...r.kernel.org>
Subject: Re: [PATCH V5 net-next 1/2] net: fec: add XDP_TX feature support
On 11/08/2023 03.26, Wei Fang wrote:
>> If you add below code comment you can add my ACK in V6:
>>
> Okay, I will add the annotation to the code in V6. Thanks.
>
Great, one adjustment to my suggested comment below.
>> Acked-by: Jesper Dangaard Brouer <hawk@...nel.org>
>>
>>> @@ -1482,7 +1488,13 @@ fec_enet_tx_queue(struct net_device *ndev,
>> u16 queue_id, int budget)
>>> /* Free the sk buffer associated with this last transmit */
>>> dev_kfree_skb_any(skb);
>>> } else {
>>> - xdp_return_frame(xdpf);
>>> + if (txq->tx_buf[index].type == FEC_TXBUF_T_XDP_NDO) {
>>> + xdp_return_frame_rx_napi(xdpf);
>>> + } else {
>>> + struct page *page = virt_to_head_page(xdpf->data);
>>> +
>>
>> I think this usage of page_pool_put_page() with dma_sync_size=0 requires a
>> comment, else we will forget why this okay...
>> I suggest:
>>
>> /* PP dma_sync_size=0 as xmit already synced DMA for_device */
>>
I update my suggestion to:
/* PP dma_sync_size=0 as XDP_TX already synced DMA for_device */
Reading code path there is an simple "else" to reach this spot, and it
will be good to hint to code-reader that this code path deals with
XDP_TX completion handling.
You are of-cause free to come up with a better comment yourself.
>>> + page_pool_put_page(page->pp, page, 0, true);
>>> + }
>>>
>>> txq->tx_buf[index].xdp = NULL;
>>> /* restore default tx buffer type: FEC_TXBUF_T_SKB */
>> @@ -1541,7
>
--Jesper
Powered by blists - more mailing lists