[PATCH net] net: fec: correct rx_bytes statistic for the case SHIFT16 is set

[PATCH net] net: fec: correct rx_bytes statistic for the case SHIFT16 is set

[Wei Fang]

Two additional bytes in front of each frame received into the RX FIFO if
SHIFT16 is set, so we need to subtract the extra two bytes from pkt_len
to correct the statistic of rx_bytes.

Fixes: 3ac72b7b63d5 ("net: fec: align IP header in hardware")
Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
 drivers/net/ethernet/freescale/fec_main.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index 1edcfaee6819..3222359ac15b 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -1835,6 +1835,8 @@ fec_enet_rx_queue(struct net_device *ndev, u16 queue_id, int budget)
 		ndev->stats.rx_packets++;
 		pkt_len = fec16_to_cpu(bdp->cbd_datlen);
 		ndev->stats.rx_bytes += pkt_len;
+		if (fep->quirks & FEC_QUIRK_HAS_RACC)
+			ndev->stats.rx_bytes -= 2;
 
 		index = fec_enet_get_bd_index(bdp, &rxq->bd);
 		page = rxq->rx_skb_info[index].page;
-- 
2.34.1

Re: [PATCH net] net: fec: correct rx_bytes statistic for the case SHIFT16 is set

[Frank Li]

On Thu, Nov 06, 2025 at 10:14:21AM +0800, Wei Fang wrote:
> Two additional bytes in front of each frame received into the RX FIFO if
> SHIFT16 is set, so we need to subtract the extra two bytes from pkt_len
> to correct the statistic of rx_bytes.
>
> Fixes: 3ac72b7b63d5 ("net: fec: align IP header in hardware")
> Signed-off-by: Wei Fang <wei.fang@nxp.com>

Reviewed-by: Frank Li <Frank.Li@nxp.com>

> ---
>  drivers/net/ethernet/freescale/fec_main.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
> index 1edcfaee6819..3222359ac15b 100644
> --- a/drivers/net/ethernet/freescale/fec_main.c
> +++ b/drivers/net/ethernet/freescale/fec_main.c
> @@ -1835,6 +1835,8 @@ fec_enet_rx_queue(struct net_device *ndev, u16 queue_id, int budget)
>  		ndev->stats.rx_packets++;
>  		pkt_len = fec16_to_cpu(bdp->cbd_datlen);
>  		ndev->stats.rx_bytes += pkt_len;
> +		if (fep->quirks & FEC_QUIRK_HAS_RACC)
> +			ndev->stats.rx_bytes -= 2;
>
>  		index = fec_enet_get_bd_index(bdp, &rxq->bd);
>  		page = rxq->rx_skb_info[index].page;
> --
> 2.34.1
>

Re: [PATCH net] net: fec: correct rx_bytes statistic for the case SHIFT16 is set

[Jakub Kicinski]

On Thu,  6 Nov 2025 10:14:21 +0800 Wei Fang wrote:
>  		ndev->stats.rx_bytes += pkt_len;
> +		if (fep->quirks & FEC_QUIRK_HAS_RACC)
> +			ndev->stats.rx_bytes -= 2;

Orthogonal to this patch, but why not:

	ndev->stats.rx_bytes += pkt_len - sub_len;

? Is this driver intentionally counting FCS as bytes?

Re: [PATCH net] net: fec: correct rx_bytes statistic for the case SHIFT16 is set

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Thu,  6 Nov 2025 10:14:21 +0800 you wrote:
> Two additional bytes in front of each frame received into the RX FIFO if
> SHIFT16 is set, so we need to subtract the extra two bytes from pkt_len
> to correct the statistic of rx_bytes.
> 
> Fixes: 3ac72b7b63d5 ("net: fec: align IP header in hardware")
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> 
> [...]

Here is the summary with links:
  - [net] net: fec: correct rx_bytes statistic for the case SHIFT16 is set
    https://git.kernel.org/netdev/net/c/ad17e7e92a7c

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

RE: [PATCH net] net: fec: correct rx_bytes statistic for the case SHIFT16 is set

[Wei Fang]

> Subject: Re: [PATCH net] net: fec: correct rx_bytes statistic for the case SHIFT16
> is set
> 
> On Thu,  6 Nov 2025 10:14:21 +0800 Wei Fang wrote:
> >  		ndev->stats.rx_bytes += pkt_len;
> > +		if (fep->quirks & FEC_QUIRK_HAS_RACC)
> > +			ndev->stats.rx_bytes -= 2;
> 
> Orthogonal to this patch, but why not:
> 
> 	ndev->stats.rx_bytes += pkt_len - sub_len;
> 
> ? Is this driver intentionally counting FCS as bytes?

Yes, the FEC hardware does not remove the FCS from the frame, so the
frame received by the driver contains the FCS and will be included in the
statistics.