From: "Paul E. McKenney" <paulmck@kernel.org>
To: Joel Fernandes <joel@joelfernandes.org>
Cc: Lai Jiangshan <laijs@linux.alibaba.com>,
linux-kernel@vger.kernel.org,
Peter Zijlstra <peterz@infradead.org>,
Josh Triplett <josh@joshtriplett.org>,
Steven Rostedt <rostedt@goodmis.org>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
Lai Jiangshan <jiangshanlai@gmail.com>,
rcu@vger.kernel.org
Subject: Re: [PATCH V2 1/7] rcu: use preempt_count to test whether scheduler locks is held
Date: Tue, 18 Feb 2020 19:59:20 -0800 [thread overview]
Message-ID: <20200219035920.GR2935@paulmck-ThinkPad-P72> (raw)
In-Reply-To: <20200219033147.GA103554@google.com>
On Tue, Feb 18, 2020 at 10:31:47PM -0500, Joel Fernandes wrote:
> On Sat, Nov 02, 2019 at 12:45:53PM +0000, Lai Jiangshan wrote:
> > Ever since preemption was introduced to linux kernel,
> > irq disabled spinlocks are always held with preemption
> > disabled. One of the reason is that sometimes we need
> > to use spin_unlock() which will do preempt_enable()
> > to unlock the irq disabled spinlock with keeping irq
> > disabled. So preempt_count can be used to test whether
> > scheduler locks is possible held.
> >
> > CC: Peter Zijlstra <peterz@infradead.org>
> > Signed-off-by: Lai Jiangshan <laijs@linux.alibaba.com>
> > ---
> > kernel/rcu/tree_plugin.h | 8 ++++++--
> > 1 file changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> > index 0982e9886103..aba5896d67e3 100644
> > --- a/kernel/rcu/tree_plugin.h
> > +++ b/kernel/rcu/tree_plugin.h
> > @@ -603,10 +603,14 @@ static void rcu_read_unlock_special(struct task_struct *t)
> > tick_nohz_full_cpu(rdp->cpu);
> > // Need to defer quiescent state until everything is enabled.
> > if (irqs_were_disabled && use_softirq &&
> > - (in_interrupt() ||
> > - (exp && !t->rcu_read_unlock_special.b.deferred_qs))) {
> > + (in_interrupt() || (exp && !preempt_bh_were_disabled))) {
> > // Using softirq, safe to awaken, and we get
> > // no help from enabling irqs, unlike bh/preempt.
> > + // in_interrupt(): raise_softirq_irqoff() is
> > + // guaranteed not to not do wakeup
> > + // !preempt_bh_were_disabled: scheduler locks cannot
> > + // be held, since spinlocks are always held with
> > + // preempt_disable(), so the wakeup will be safe.
>
> This means if preemption is disabled for any reason (other than scheduler
> locks), such as acquiring an unrelated lock that is not held by the
> scheduler, then the softirq would not be raised even if it was safe to
> do so. From that respect, it seems a step back no?
This patch was one of the things motivating me to turn tick on for
nohz_full CPUs that spend too long in the kernel. Given that change,
this patch can be (and recently was) made more straightforward. Prior to
the nohz_full change, Lai was kind of between a rock and a hard place
on this one. ;-)
Thanx, Paul
> thanks,
>
> - Joel
>
>
> > raise_softirq_irqoff(RCU_SOFTIRQ);
> > } else {
> > // Enabling BH or preempt does reschedule, so...
> > --
> > 2.20.1
> >
next prev parent reply other threads:[~2020-02-19 3:59 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-02 12:45 [PATCH V2 0/7] rcu: introduce percpu rcu_preempt_depth Lai Jiangshan
2019-11-02 12:45 ` [PATCH V2 1/7] rcu: use preempt_count to test whether scheduler locks is held Lai Jiangshan
2019-11-15 16:53 ` Paul E. McKenney
2020-02-19 3:31 ` Joel Fernandes
2020-02-19 3:59 ` Paul E. McKenney [this message]
2019-11-02 12:45 ` [PATCH V2 2/7] rcu: cleanup rcu_preempt_deferred_qs() Lai Jiangshan
2019-11-03 2:01 ` Boqun Feng
2019-11-03 5:01 ` Lai Jiangshan
2019-11-04 14:55 ` Paul E. McKenney
2019-11-05 2:09 ` Lai Jiangshan
2019-11-05 7:19 ` Paul E. McKenney
2019-11-11 14:32 ` Paul E. McKenney
2019-11-12 1:28 ` Lai Jiangshan
2020-02-17 23:23 ` Paul E. McKenney
2020-02-18 14:41 ` Steven Rostedt
2020-02-18 16:43 ` Paul E. McKenney
2019-11-15 16:55 ` Paul E. McKenney
2019-11-02 12:45 ` [PATCH V2 3/7] rcu: remove useless special.b.deferred_qs Lai Jiangshan
2020-02-17 23:23 ` Paul E. McKenney
2019-11-02 12:45 ` [PATCH V2 4/7] rcu: don't use negative ->rcu_read_lock_nesting Lai Jiangshan
2020-02-17 23:26 ` Paul E. McKenney
2019-11-02 12:45 ` [PATCH V2 5/7] rcu: wrap usages of rcu_read_lock_nesting Lai Jiangshan
2019-11-15 22:25 ` Paul E. McKenney
2019-11-02 12:45 ` [PATCH V2 6/7] rcu: clear the special.b.need_qs in rcu_note_context_switch() Lai Jiangshan
2019-11-16 15:46 ` Paul E. McKenney
2019-11-02 12:45 ` [PATCH V2 7/7] x86,rcu: use percpu rcu_preempt_depth Lai Jiangshan
2019-11-02 16:30 ` Borislav Petkov
2019-11-03 4:33 ` Lai Jiangshan
2019-11-04 9:25 ` Sebastian Andrzej Siewior
2019-11-04 11:41 ` Lai Jiangshan
2019-11-04 12:09 ` Sebastian Andrzej Siewior
2019-11-16 15:48 ` Paul E. McKenney
2019-11-18 2:02 ` Lai Jiangshan
2019-11-18 14:59 ` Paul E. McKenney
2019-11-19 1:59 ` Lai Jiangshan
2019-11-19 21:14 ` Paul E. McKenney
2019-11-20 2:47 ` Lai Jiangshan
2019-11-21 4:02 ` Paul E. McKenney
2019-11-02 15:05 ` [PATCH V2 0/7] rcu: introduce " Paul E. McKenney
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=20200219035920.GR2935@paulmck-ThinkPad-P72 \
--to=paulmck@kernel.org \
--cc=jiangshanlai@gmail.com \
--cc=joel@joelfernandes.org \
--cc=josh@joshtriplett.org \
--cc=laijs@linux.alibaba.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=peterz@infradead.org \
--cc=rcu@vger.kernel.org \
--cc=rostedt@goodmis.org \
/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;
as well as URLs for NNTP newsgroup(s).