Re: [Intel-wired-lan] [PATCH net v2 3/3] net: stmmac: skip queueMaxSDU check for AF_XDP

From
Jakub Kicinski <kuba@kernel.org>
Date
2026-08-18 18:19:43
Also in
bpf, intel-wired-lan, linux-arm-kernel, linux-rdma, lkml
Reply to
Stanislav Fomichev, 2026-08-18
On Tue, 18 Aug 2026 10:03:16 -0700 Stanislav Fomichev wrote:
quoted
Was completing the reserved descriptor considered instead of removing the
check, i.e. keeping the length test and doing:

	if (priv->est && priv->est->enable &&
	    priv->est->max_sdu[queue] &&
	    xdp_desc.len > priv->est->max_sdu[queue]) {
		priv->xstats.max_sdu_txq_drop[queue]++;
		xsk_tx_completed(pool, 1);
		continue;
	}

That would fix the ring accounting without dropping the enforcement.  
Don't think this suggestion would work. xsk_tx_completed(1) will complete
the first queued entry, no the current rejected one (completions are in order).

If it's not convincing, I can repost only patches 1 and 2. Or if it sounds
reasonable, can try to explain that part in the description.
Hm, I see. No strong preference but I'd probably just add a comment
above the checks saying that we are leaking there semi-intentionally
because it's an application error to send packets HW/schedule can't
handle. We expect user to fix their app. IOW if we can't handle this
well I think it's better to wedge the SW and bump the drop counter
for visibility, than risk HW silently misbehaving.
Thread (12 messages) 12 messages, 5 authors, last reply 2026-08-18
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help