Re: [PATCH v4 net 8/8] net: enetc: ensure enetc_xdp_xmit() calls enetc_update_tx_ring_tail()

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] net: enetc: XDP_XMIT_FLUSH is ignored when k == 0 causing TX ring stalls
--

commit a3de2d9d87edd1ec648a1023c638ce2d70118f57
Author: vladimir.oltean@nxp.com <vladimir.oltean@nxp.com>

net: enetc: ensure enetc_xdp_xmit() calls enetc_update_tx_ring_tail()

This commit changes the loop exit condition that triggers the TX doorbell
in enetc_xdp_xmit(). It replaces the redundant xdp_tx_frm_cnt variable with
the loop index k and updates the logic determining when to flush the TX
ring.

> diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
> --- a/drivers/net/ethernet/freescale/enetc/enetc.c
> +++ b/drivers/net/ethernet/freescale/enetc/enetc.c
[ ... ]
> @@ -1928,22 +1928,20 @@ int enetc_xdp_xmit(struct net_device *ndev, int num_frames,
>  						    &xdp_redirect_arr[i]);
>  			break;
>  		}
> -
> -		xdp_tx_frm_cnt++;
>  	}
>  
> -	if (unlikely((flags & XDP_XMIT_FLUSH) || k != xdp_tx_frm_cnt))
> +	if (unlikely(k && ((flags & XDP_XMIT_FLUSH) || k < num_frames)))
>  		enetc_update_tx_ring_tail(tx_ring);

[Severity: High]
If a caller batches XDP frames by making multiple ndo_xdp_xmit() calls
with flags = 0, and then makes a final call with flags = XDP_XMIT_FLUSH,
does this new condition safely handle the case where the final call fails
to enqueue any frames?

If the caller passes num_frames == 0, or if the first frame fails to map
(for example, due to a full TX ring or DMA error), the loop exits with
k = 0.

With the addition of the "k &&" guard, the entire condition evaluates to
false.

Could this skip enetc_update_tx_ring_tail() and leave previously queued
packets stalled in the TX ring without a doorbell ring, violating the
XDP_XMIT_FLUSH contract?

>  	enetc_unlock_mdio();

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261015105536.1692348-1-vladimir.oltean@nxp.com?part=8