Re: [v6,net-next,8/8] bng_en: Add support for TPA events

[Jakub Kicinski <kuba@kernel.org>]

This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.

For local reproduction steps see https://netdev-ai.bots.linux.dev/ai-local.html
---
bng_en: Add support for TPA events

This commit enables TPA (Transport Protocol Aggregation) functionality
in the VNIC and adds functions to handle TPA events for LRO/GRO
processing.

> +static void bnge_tpa_agg(struct bnge_net *bn, struct bnge_rx_ring_info *rxr,
> +			 struct rx_agg_cmp *rx_agg)
> +{
> +	u16 agg_id = TPA_AGG_AGG_ID(rx_agg);
> +	struct bnge_tpa_info *tpa_info;
> +
> +	agg_id = bnge_lookup_agg_idx(rxr, agg_id);
> +	tpa_info = &rxr->rx_tpa[agg_id];
> +
> +	tpa_info->agg_arr[tpa_info->agg_count++] = *rx_agg;
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^

The agg_arr is allocated with MAX_SKB_FRAGS entries, but there is no
bounds check before writing to it. The bnxt driver has a BUG_ON guard
at this location:

    BUG_ON(tpa_info->agg_count >= MAX_SKB_FRAGS);

Is there a reason this check was omitted? While the check in
bnge_tpa_end() catches agg_bufs > MAX_SKB_FRAGS, that happens after
the aggregation completions have already been stored. If hardware
misbehaves and sends more aggregation completions than expected, could
this overflow agg_arr[]?

> +}
-- 
pw-bot: cr