From: Boqun Feng <boqun@kernel.org>
To: "Paul E. McKenney" <paulmck@kernel.org>
Cc: Joel Fernandes <joelagnelf@nvidia.com>,
Kumar Kartikeya Dwivedi <memxor@gmail.com>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
frederic@kernel.org, neeraj.iitr10@gmail.com, urezki@gmail.com,
boqun.feng@gmail.com, rcu@vger.kernel.org,
Tejun Heo <tj@kernel.org>,
bpf@vger.kernel.org, Alexei Starovoitov <ast@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
John Fastabend <john.fastabend@gmail.com>,
Andrea Righi <arighi@nvidia.com>, Zqiang <qiang.zhang@linux.dev>
Subject: Re: [PATCH] rcu: Use an intermediate irq_work to start process_srcu()
Date: Sat, 21 Mar 2026 12:45:27 -0700 [thread overview]
Message-ID: <ab71Vz9zmlVywWQT@tardis.local> (raw)
In-Reply-To: <609b5df1-aa06-46a9-8e93-0bf9eb8b7738@paulmck-laptop>
On Sat, Mar 21, 2026 at 12:31:04PM -0700, Paul E. McKenney wrote:
> On Sat, Mar 21, 2026 at 11:06:59AM -0700, Boqun Feng wrote:
> > On Sat, Mar 21, 2026 at 10:41:47AM -0700, Paul E. McKenney wrote:
> > [...]
> > > > > + raw_spin_lock_rcu_node(ssp->srcu_sup);
> > > > > + delay = srcu_get_delay(ssp);
> > > > > + raw_spin_unlock_rcu_node(ssp->srcu_sup);
> > > > >
> > > >
> > > > It was fixed differently in v2:
> > > >
> > > > https://lore.kernel.org/rcu/20260320222916.19987-1-boqun@kernel.org/
> > > >
> > > > I used _irqsave/_irqrestore just in case. Given it's an urgent fix,
> > > > overly careful code is probably fine ;-)
> > > >
> > > > Thanks for the testing and feedback.
> > >
> > > OK, I will try that one, thank you!
> > >
> > > FYI, with my change on your earlier version, SRCU-T got deadlocks between
> > > the pi-lock and the workqueue pool lock. Which might or might not be
> > > particularly urgent.
> > >
> >
> > I just checked my run yesterday, I also hit it. It's probably what
> > Zqiang has found:
> >
> > https://lore.kernel.org/rcu/4c23c66f86a2aff8f2d7b759f9dd257b82147a17@linux.dev/
> >
> > We have a queue_work_on() in srcu_schedule_cbs_sdp(), so
> >
> > srcu_torture_deferred_free():
> > raw_spin_lock_irqsave(->pi_lock,...);
> > call_srcu():
> > if (snp == snp_leaf && snp_seq != s) {
> > srcu_schedule_cbs_sdp(sdp, do_norm ? SRCU_INTERVAL : 0):
> > if (!delay)
> > queue_work_on(...)
> >
> > I was about to reply to Zqiang, fixing that could be a touch design
> > decision. Since it's a per srcu_data work ;-) NR_CPUS x irq_work
> > incoming.
>
> Just to be clear, SRCU-T is Tiny SRCU rather than Tree SRCU. So perhaps
> lower priority, though perhaps not lower irritation. ;-)
>
I see, there is a schedule_work() in srcutiny's
srcu_gp_start_if_needed(). But it couldn't cause deadlock on UP since
locks are (almost) no-op. Maybe we can make RCU torture only test it on
SMP?
Regards,
Boqun
> Thanx, Paul
next prev parent reply other threads:[~2026-03-21 19:45 UTC|newest]
Thread overview: 100+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-17 13:34 Next-level bug in SRCU implementation of RCU Tasks Trace + PREEMPT_RT Paul E. McKenney
2026-03-18 10:50 ` Sebastian Andrzej Siewior
2026-03-18 11:49 ` Paul E. McKenney
2026-03-18 14:43 ` Sebastian Andrzej Siewior
2026-03-18 15:43 ` Paul E. McKenney
2026-03-18 16:04 ` Sebastian Andrzej Siewior
2026-03-18 16:32 ` Paul E. McKenney
2026-03-18 16:42 ` Boqun Feng
2026-03-18 18:45 ` Paul E. McKenney
2026-03-18 16:47 ` Sebastian Andrzej Siewior
2026-03-18 18:48 ` Paul E. McKenney
2026-03-19 8:55 ` Sebastian Andrzej Siewior
2026-03-19 10:05 ` Paul E. McKenney
2026-03-19 10:43 ` Paul E. McKenney
2026-03-19 10:51 ` Sebastian Andrzej Siewior
2026-03-18 15:51 ` Boqun Feng
2026-03-18 18:42 ` Paul E. McKenney
2026-03-18 20:04 ` Joel Fernandes
2026-03-18 20:11 ` Kumar Kartikeya Dwivedi
2026-03-18 20:25 ` Joel Fernandes
2026-03-18 21:52 ` Boqun Feng
2026-03-18 21:55 ` Boqun Feng
2026-03-18 22:15 ` Boqun Feng
2026-03-18 22:52 ` Joel Fernandes
2026-03-18 23:27 ` Boqun Feng
2026-03-19 1:08 ` Boqun Feng
2026-03-19 9:03 ` Sebastian Andrzej Siewior
2026-03-19 16:27 ` Boqun Feng
2026-03-19 16:33 ` Sebastian Andrzej Siewior
2026-03-19 16:48 ` Boqun Feng
2026-03-19 16:59 ` Kumar Kartikeya Dwivedi
2026-03-19 17:27 ` Boqun Feng
2026-03-19 18:41 ` Kumar Kartikeya Dwivedi
2026-03-19 20:14 ` Boqun Feng
2026-03-19 20:21 ` Joel Fernandes
2026-03-19 20:39 ` Boqun Feng
2026-03-20 15:34 ` Paul E. McKenney
2026-03-20 15:59 ` Boqun Feng
2026-03-20 16:24 ` Paul E. McKenney
2026-03-20 16:57 ` Boqun Feng
2026-03-20 17:54 ` Joel Fernandes
2026-03-20 18:14 ` [PATCH] rcu: Use an intermediate irq_work to start process_srcu() Boqun Feng
2026-03-20 19:18 ` Joel Fernandes
2026-03-20 20:47 ` Andrea Righi
2026-03-20 20:54 ` Boqun Feng
2026-03-20 21:00 ` Andrea Righi
2026-03-20 21:02 ` Andrea Righi
2026-03-20 21:06 ` Boqun Feng
2026-03-20 22:29 ` [PATCH v2] " Boqun Feng
2026-03-23 21:09 ` Joel Fernandes
2026-03-23 22:18 ` Boqun Feng
2026-03-23 22:50 ` Joel Fernandes
2026-03-24 11:27 ` Frederic Weisbecker
2026-03-24 14:56 ` Joel Fernandes
2026-03-24 14:56 ` Alexei Starovoitov
2026-03-24 17:36 ` Boqun Feng
2026-03-24 18:40 ` Joel Fernandes
2026-03-24 19:23 ` Paul E. McKenney
2026-03-21 4:27 ` [PATCH] " Zqiang
2026-03-21 18:15 ` Boqun Feng
2026-03-21 10:10 ` Paul E. McKenney
2026-03-21 17:15 ` Boqun Feng
2026-03-21 17:41 ` Paul E. McKenney
2026-03-21 18:06 ` Boqun Feng
2026-03-21 19:31 ` Paul E. McKenney
2026-03-21 19:45 ` Boqun Feng [this message]
2026-03-21 20:07 ` Paul E. McKenney
2026-03-21 20:08 ` Boqun Feng
2026-03-22 10:09 ` Paul E. McKenney
2026-03-22 16:16 ` Boqun Feng
2026-03-22 17:09 ` Paul E. McKenney
2026-03-22 17:31 ` Boqun Feng
2026-03-22 17:44 ` Paul E. McKenney
2026-03-22 18:17 ` Boqun Feng
2026-03-22 19:47 ` Paul E. McKenney
2026-03-22 20:26 ` Boqun Feng
2026-03-23 7:50 ` Paul E. McKenney
2026-03-20 18:20 ` Next-level bug in SRCU implementation of RCU Tasks Trace + PREEMPT_RT Boqun Feng
2026-03-20 23:11 ` Paul E. McKenney
2026-03-21 3:29 ` Paul E. McKenney
2026-03-21 17:03 ` [RFC PATCH] rcu-tasks: Avoid using mod_timer() in call_rcu_tasks_generic() Boqun Feng
2026-03-23 15:17 ` Boqun Feng
2026-03-23 20:37 ` Joel Fernandes
2026-03-23 21:50 ` Kumar Kartikeya Dwivedi
2026-03-23 22:13 ` Boqun Feng
2026-03-20 16:15 ` Next-level bug in SRCU implementation of RCU Tasks Trace + PREEMPT_RT Boqun Feng
2026-03-20 16:24 ` Paul E. McKenney
2026-03-19 17:02 ` Sebastian Andrzej Siewior
2026-03-19 17:44 ` Boqun Feng
2026-03-19 18:42 ` Joel Fernandes
2026-03-19 20:20 ` Boqun Feng
2026-03-19 20:26 ` Joel Fernandes
2026-03-19 20:45 ` Joel Fernandes
2026-03-19 10:02 ` Paul E. McKenney
2026-03-19 14:34 ` Boqun Feng
2026-03-19 16:10 ` Paul E. McKenney
2026-03-18 23:56 ` Kumar Kartikeya Dwivedi
2026-03-19 0:26 ` Zqiang
2026-03-19 1:13 ` Boqun Feng
2026-03-19 2:47 ` Joel Fernandes
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ab71Vz9zmlVywWQT@tardis.local \
--to=boqun@kernel.org \
--cc=arighi@nvidia.com \
--cc=ast@kernel.org \
--cc=bigeasy@linutronix.de \
--cc=boqun.feng@gmail.com \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=frederic@kernel.org \
--cc=joelagnelf@nvidia.com \
--cc=john.fastabend@gmail.com \
--cc=memxor@gmail.com \
--cc=neeraj.iitr10@gmail.com \
--cc=paulmck@kernel.org \
--cc=qiang.zhang@linux.dev \
--cc=rcu@vger.kernel.org \
--cc=tj@kernel.org \
--cc=urezki@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox