Received-SPF: pass (google.com: domain of linux-kernel+bounces-207034-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Sat, 8 Jun 2024 14:42:37 +0200
From: Oleg Nesterov <oleg@redhat.com>
To: Wei Fu <fuweid89@gmail.com>
Cc: Sudhanva.Huruli@microsoft.com, akpm@linux-foundation.org,
	apais@linux.microsoft.com, axboe@kernel.dk, boqun.feng@gmail.com,
	brauner@kernel.org, ebiederm@xmission.com, frederic@kernel.org,
	j.granados@samsung.com, jiangshanlai@gmail.com,
	joel@joelfernandes.org, josh@joshtriplett.org,
	linux-kernel@vger.kernel.org, mathieu.desnoyers@efficios.com,
	michael.christie@oracle.com, mjguzik@gmail.com,
	neeraj.upadhyay@kernel.org, paulmck@kernel.org,
	qiang.zhang1211@gmail.com, rachelmenge@linux.microsoft.com,
	rcu@vger.kernel.org, rostedt@goodmis.org, weifu@microsoft.com
Subject: Re: [RCU] zombie task hung in synchronize_rcu_expedited
Message-ID: <20240608124237.GC7947@redhat.com>
References: <20240607062531.GA30954@redhat.com>
 <20240607150443.2096063-1-fuweid89@gmail.com>
 <20240607212116.GA7258@redhat.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240607212116.GA7258@redhat.com>
User-Agent: Mutt/1.5.24 (2015-08-30)

On 06/07, Oleg Nesterov wrote:
>
> On 06/07, Wei Fu wrote:
> >
> > Yes. I applied your patch on v5.15.160 and run reproducer for 5 hours.
> > I didn't see this issue. Currently, it looks good!. I will continue that test
> > on this weekend.
>
> Great, thanks!
>
> > In last reply, you mentioned TIF_NOTIFY_SIGNAL related to busy-wait loop.
> > Would you please explain why flag-clear works here?
>
> Sure, I'll write the changelog with the explanation and send the patch on
> weekend. If it passes your testing.

Please see the patch I've sent. The changelog doesn't bother to describe this
particular problem because busy-waiting can obviously cause multiple problems,
especially without CONFIG_PREEMPT or if rt_task().

So let me add more details about this particular deadlock here.

The sub-namespace init task T spins in a tight loop calling kernel_wait4()
which returns -EINTR without sleeping because its child C has not exited
yet and signal_pending(T) is true due to TIF_NOTIFY_SIGNAL.

The exiting child C sleeps in synchronize_rcu() which hangs exactly because
T never calls schedule/rcu_note_context_switch, it can't be preempted because
CONFIG_PREEMPT is not enabled.

Note also that without PREEMPT_RCU __rcu_read_lock() is just preempt_disable()
which is nop without CONFIG_PREEMPT.

Oleg.