[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANn89iLpON7eV9rHvErsoEu+GqDz18uYMv6M_4TLsh+WX9VQeg@mail.gmail.com>
Date: Fri, 21 Mar 2025 11:37:31 +0100
From: Eric Dumazet <edumazet@...gle.com>
To: Breno Leitao <leitao@...ian.org>
Cc: Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>,
Will Deacon <will@...nel.org>, Boqun Feng <boqun.feng@...il.com>, Waiman Long <longman@...hat.com>,
aeh@...a.com, linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
jhs@...atatu.com, kernel-team@...a.com, Erik Lundgren <elundgren@...a.com>,
"Paul E. McKenney" <paulmck@...nel.org>
Subject: Re: [PATCH] lockdep: Speed up lockdep_unregister_key() with expedited
RCU synchronization
On Fri, Mar 21, 2025 at 10:31 AM Breno Leitao <leitao@...ian.org> wrote:
>
> lockdep_unregister_key() is called from critical code paths, including
> sections where rtnl_lock() is held. For example, when replacing a qdisc
> in a network device, network egress traffic is disabled while
> __qdisc_destroy() is called for every network queue.
>
> If lockdep is enabled, __qdisc_destroy() calls lockdep_unregister_key(),
> which gets blocked waiting for synchronize_rcu() to complete.
>
> For example, a simple tc command to replace a qdisc could take 13
> seconds:
>
> # time /usr/sbin/tc qdisc replace dev eth0 root handle 0x1: mq
> real 0m13.195s
> user 0m0.001s
> sys 0m2.746s
>
> During this time, network egress is completely frozen while waiting for
> RCU synchronization.
>
> Use synchronize_rcu_expedited() instead to minimize the impact on
> critical operations like network connectivity changes.
Running 'critical operations' with LOCKDEP enabled kernels seems a bit
strange :)
Reviewed-by: Eric Dumazet <edumazet@...gle.com>
Powered by blists - more mailing lists