Re: [PATCH v3 net 5/7] net: enetc: pad short XDP frames coming from devmap

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] net: double counting of dropped XDP frames
--

commit eb0b831b4274ed0cdbc903bad8f5a21c4a5499be
Author: Vladimir Oltean <vladimir.oltean@nxp.com>

net: enetc: pad short XDP frames coming from devmap

This patch adds length checks and padding for short XDP frames transmitted
through enetc_xdp_xmit(). It introduces xdp_frame_pad() to pad frames to
the minimum Ethernet size and drops frames that cannot be expanded.

> diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
> index bbad942041f5a..8a9ba168eab1e 100644
> --- a/drivers/net/ethernet/freescale/enetc/enetc.c
> +++ b/drivers/net/ethernet/freescale/enetc/enetc.c
> @@ -1838,15 +1838,23 @@ int enetc_xdp_xmit(struct net_device *ndev, int num_frames,
>  	prefetchw(ENETC_TXBD(*tx_ring, tx_ring->next_to_use));
>  
>  	for (k = 0; k < num_frames; k++) {
[ ... ]
> +		if (unlikely(xdp_frame_pad(xdpf) ||
> +			     xdpf->len < ENETC_MIN_BUFF_SIZE)) {
> +			tx_ring->stats.xdp_tx_drops++;
> +			break;
> +		}
> +
>  		xdp_tx_bd_cnt = enetc_xdp_frame_to_xdp_tx_swbd(tx_ring,

[Severity: Medium]
Will this cause the dropped frame at index k to be counted twice?

When xdp_frame_pad() fails or the length is still below ENETC_MIN_BUFF_SIZE,
the drop counter is explicitly incremented before breaking out of the loop
in enetc_xdp_xmit().

However, looking at the end of enetc_xdp_xmit(), the function 
unconditionally evaluates the remainder of the frames:

    tx_ring->stats.xdp_tx += k;
    tx_ring->stats.xdp_tx_drops += num_frames - k;

Since k is the index of the frame that just failed, num_frames - k includes
this same frame again. As a result, the frame at index k appears to be
counted once inside the loop, and a second time at the end of the function.

Could we simply break out of the loop here and let the end of the function
handle adding it to the drop count, similar to how it handles the other
failure paths?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260915222735.1016937-1-vladimir.oltean@nxp.com?part=5