Re: [PATCH nf-next v3 2/3] netfilter: flowtable: carry a priority into the offload

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] tc: mtk_ppe: silently ignores FLOW_ACTION_PRIORITY for TC flower offload
--

commit a587674714a18a7d90f4418de25fb5ffa3691fa3
Author: Julius Bairaktaris <julius@bairaktaris.de>

netfilter: flowtable: carry a priority into the offload

This commit updates the flowtable to store the skb->priority of the packet
that creates the flow, and emit it as FLOW_ACTION_PRIORITY during offload.
It also updates the airoha and mtk drivers to parse this action in their
shared flowtable and TC flower offload functions.

> diff --git a/drivers/net/ethernet/mediatek/mtk_ppe_offload.c b/drivers/net/ethernet/mediatek/mtk_ppe_offload.c
> index 99b28aaa7cc47..4ee99e8e4a34b 100644
> --- a/drivers/net/ethernet/mediatek/mtk_ppe_offload.c
> +++ b/drivers/net/ethernet/mediatek/mtk_ppe_offload.c

[ ... ]

> @@ -378,6 +378,7 @@ mtk_flow_offload_replace(struct mtk_eth *eth, struct flow_cls_offload *f,
>  		case FLOW_ACTION_REDIRECT:
>  			odev = act->dev;
>  			break;
> +		case FLOW_ACTION_PRIORITY:
>  		case FLOW_ACTION_CSUM:
>  			break;

[Severity: Medium]
Does this silently ignore explicit QoS queuing requests from TC flower?

Because mtk_flow_offload_replace() is shared for both netfilter flowtable
and TC flower offloads, accommodating flowtable packets by returning success
here causes the driver to also accept and silently discard "skbedit priority"
actions from TC rules. 

Hardware will silently drop the explicit QoS queuing request instead of
returning -EOPNOTSUPP, causing traffic to use default queues instead of
the requested priority queue without warning the user.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261004171636.3545085-1-julius@bairaktaris.de?part=2