[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1dbae52d-6aba-467a-a864-24eeb3f96449@intel.com>
Date: Thu, 5 Jan 2023 18:35:05 +0100
From: Alexander Lobakin <alexandr.lobakin@...el.com>
To: Gerhard Engleder <gerhard@...leder-embedded.com>
CC: <davem@...emloft.net>, <kuba@...nel.org>, <edumazet@...gle.com>,
<pabeni@...hat.com>, Saeed Mahameed <saeed@...nel.org>,
<netdev@...r.kernel.org>
Subject: Re: [PATCH net-next v3 8/9] tsnep: Add RX queue info for XDP support
From: Gerhard Engleder <gerhard@...leder-embedded.com>
Date: Wed Jan 04 2023 20:41:31 GMT+0100
> Register xdp_rxq_info with page_pool memory model. This is needed for
> XDP buffer handling.
>
> Signed-off-by: Gerhard Engleder <gerhard@...leder-embedded.com>
> Reviewed-by: Saeed Mahameed <saeed@...nel.org>
> ---
> drivers/net/ethernet/engleder/tsnep.h | 6 ++--
> drivers/net/ethernet/engleder/tsnep_main.c | 34 +++++++++++++++++-----
> 2 files changed, 31 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/net/ethernet/engleder/tsnep.h b/drivers/net/ethernet/engleder/tsnep.h
> index 0e7fc36a64e1..0210dab90f71 100644
> --- a/drivers/net/ethernet/engleder/tsnep.h
> +++ b/drivers/net/ethernet/engleder/tsnep.h
> @@ -133,17 +133,19 @@ struct tsnep_rx {
> u32 dropped;
> u32 multicast;
> u32 alloc_failed;
> +
> + struct xdp_rxq_info xdp_rxq;
> };
>
> struct tsnep_queue {
> struct tsnep_adapter *adapter;
> char name[IFNAMSIZ + 9];
>
> + struct napi_struct napi;
> +
> struct tsnep_tx *tx;
> struct tsnep_rx *rx;
>
> - struct napi_struct napi;
> -
I'd leave a word in the commit message that you're moving ::napi to
improve structure cacheline span. Or even do that in a separate commit
with some pahole output to make it clear why you do that.
> int irq;
> u32 irq_mask;
> void __iomem *irq_delay_addr;
[...]
> @@ -1253,6 +1266,7 @@ int tsnep_netdev_open(struct net_device *netdev)
> {
> struct tsnep_adapter *adapter = netdev_priv(netdev);
> int i;
> + unsigned int napi_id;
Reverse Christmas Tree variable style is already messed up here, maybe
you could fix it inplace while at it or at least not make it worse? :D
> void __iomem *addr;
> int tx_queue_index = 0;
> int rx_queue_index = 0;
[...]
Thanks,
Olek
Powered by blists - more mailing lists