Re: [PATCH net-next] net: sched: remove unnecessay lock protection for skb_bad_txq/gso_skb

On 2021/3/17 2:43, Cong Wang wrote:

On Mon, Mar 15, 2021 at 4:42 PM David Miller [off-list ref] wrote:

quoted

From: Yunsheng Lin <redacted>
Date: Mon, 15 Mar 2021 17:30:10 +0800

quoted

Currently qdisc_lock(q) is taken before enqueuing and dequeuing
for lockless qdisc's skb_bad_txq/gso_skb queue, qdisc->seqlock is
also taken, which can provide the same protection as qdisc_lock(q).

This patch removes the unnecessay qdisc_lock(q) protection for
lockless qdisc' skb_bad_txq/gso_skb queue.

And dev_reset_queue() takes the qdisc->seqlock for lockless qdisc
besides taking the qdisc_lock(q) when doing the qdisc reset,
some_qdisc_is_busy() takes both qdisc->seqlock and qdisc_lock(q)
when checking qdisc status. It is unnecessary to take both lock
while the fast path only take one lock, so this patch also changes
it to only take qdisc_lock(q) for locked qdisc, and only take
qdisc->seqlock for lockless qdisc.

Since qdisc->seqlock is taken for lockless qdisc when calling
qdisc_is_running() in some_qdisc_is_busy(), use qdisc->running
to decide if the lockless qdisc is running.

Signed-off-by: Yunsheng Lin <redacted>

What about other things protected by this lock, such as statistics and qlen?

This change looks too risky to me.

They are per-cpu for pfifo_fast which sets TCQ_F_CPUSTATS too.

Did you mean qdisc_lock(q) are protecting per-cpu stats for
pfifo_fast too?

Thanks.

.