From: "Paul E. McKenney" <paulmck@kernel.org>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Steven Rostedt <rostedt@goodmis.org>, linke li <lilinke99@qq.com>,
joel@joelfernandes.org, boqun.feng@gmail.com, dave@stgolabs.net,
frederic@kernel.org, jiangshanlai@gmail.com,
josh@joshtriplett.org, linux-kernel@vger.kernel.org,
mathieu.desnoyers@efficios.com, qiang.zhang1211@gmail.com,
quic_neeraju@quicinc.com, rcu@vger.kernel.org
Subject: Re: [PATCH] rcutorture: Fix rcu_torture_pipe_update_one()/rcu_torture_writer() data race and concurrency bug
Date: Wed, 6 Mar 2024 19:21:14 -0800 [thread overview]
Message-ID: <29a20fd4-ac5e-44a3-bc8a-9f77aa6a3cf9@paulmck-laptop> (raw)
In-Reply-To: <CAHk-=wi7rJ-eGq+xaxVfzFEgbL9tdf6Kc8Z89rCpfcQOKm74Tw@mail.gmail.com>
On Wed, Mar 06, 2024 at 06:49:38PM -0800, Linus Torvalds wrote:
> On Wed, 6 Mar 2024 at 18:43, Linus Torvalds
> <torvalds@linux-foundation.org> wrote:
> >
> > I dunno.
>
> Oh, and just looking at that patch, I still think the code is confused.
>
> On the reading side, we have:
>
> pipe_count = smp_load_acquire(&p->rtort_pipe_count);
> if (pipe_count > RCU_TORTURE_PIPE_LEN) {
> /* Should not happen, but... */
>
> where that comment clearly says that the pipe_count we read (whether
> with READ_ONCE() or with my smp_load_acquire() suggestion) should
> never be larger than RCU_TORTURE_PIPE_LEN.
I will fix that comment. It should not happen *if* RCU is working
correctly. It can happen if you have an RCU that is so broken that a
single RCU reader can span more than ten grace periods. An example of
an RCU that really is this broken can be selected using rcutorture's
torture_type=busted module parameter. No surprise, given that its
implementation of call_rcu() invokes the callback function directly and
its implementation of synchronize_rcu() is a complete no-op. ;-)
Of course, the purpose of that value of the torture_type module parameter
(along with all other possible values containing the string "busted")
is to test rcutorture itself.
> But the writing side very clearly did:
>
> i = rp->rtort_pipe_count;
> if (i > RCU_TORTURE_PIPE_LEN)
> i = RCU_TORTURE_PIPE_LEN;
> ...
> smp_store_release(&rp->rtort_pipe_count, ++i);
>
> (again, syntactically it could have been "i + 1" instead of my "++i" -
> same value), so clearly the writing side *can* write a value that is >
> RCU_TORTURE_PIPE_LEN.
>
> So while the whole READ/WRITE_ONCE vs smp_load_acquire/store_release
> is one thing that might be worth looking at, I think there are other
> very confusing aspects here.
With this change in that comment, are things better?
Thanx, Paul
------------------------------------------------------------------------
diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c
index 6b821a7037b03..0cb5452ecd945 100644
--- a/kernel/rcu/rcutorture.c
+++ b/kernel/rcu/rcutorture.c
@@ -2000,7 +2000,8 @@ static bool rcu_torture_one_read(struct torture_random_state *trsp, long myid)
preempt_disable();
pipe_count = READ_ONCE(p->rtort_pipe_count);
if (pipe_count > RCU_TORTURE_PIPE_LEN) {
- /* Should not happen, but... */
+ // Should not happen in a correct RCU implementation,
+ // happens quite often for torture_type=busted.
pipe_count = RCU_TORTURE_PIPE_LEN;
}
completed = cur_ops->get_gp_seq();
next prev parent reply other threads:[~2024-03-07 3:21 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-03-04 10:54 linke li
2024-03-04 16:19 ` Joel Fernandes
2024-03-04 17:14 ` Paul E. McKenney
2024-03-04 19:10 ` Joel Fernandes
2024-03-04 19:44 ` Paul E. McKenney
2024-03-04 20:13 ` Joel Fernandes
2024-03-04 20:47 ` Paul E. McKenney
2024-03-05 3:30 ` linke
2024-03-05 6:24 ` linke li
2024-03-06 15:37 ` Steven Rostedt
2024-03-06 17:36 ` Paul E. McKenney
2024-03-06 18:01 ` Steven Rostedt
2024-03-06 18:09 ` Paul E. McKenney
2024-03-06 18:20 ` Steven Rostedt
2024-03-06 18:43 ` Linus Torvalds
2024-03-06 18:55 ` Steven Rostedt
2024-03-06 19:01 ` Linus Torvalds
2024-03-06 19:27 ` Linus Torvalds
2024-03-06 19:47 ` Steven Rostedt
2024-03-06 20:06 ` Linus Torvalds
2024-03-07 13:20 ` Steven Rostedt
2024-03-07 16:12 ` Steven Rostedt
2024-03-06 19:27 ` Steven Rostedt
2024-03-06 19:46 ` Linus Torvalds
2024-03-06 20:20 ` Linus Torvalds
2024-03-07 2:29 ` Paul E. McKenney
2024-03-07 2:43 ` Linus Torvalds
2024-03-07 2:49 ` Linus Torvalds
2024-03-07 3:21 ` Paul E. McKenney [this message]
2024-03-07 3:06 ` Paul E. McKenney
2024-03-07 3:06 ` Mathieu Desnoyers
2024-03-07 3:37 ` Paul E. McKenney
2024-03-07 5:44 ` Joel Fernandes
2024-03-07 19:05 ` Paul E. McKenney
2024-03-07 13:53 ` Mathieu Desnoyers
2024-03-07 19:47 ` Paul E. McKenney
2024-03-07 19:53 ` Mathieu Desnoyers
2024-03-08 0:58 ` Paul E. McKenney
2024-03-07 20:00 ` Linus Torvalds
2024-03-07 20:57 ` Paul E. McKenney
2024-03-07 21:40 ` Julia Lawall
2024-03-07 22:09 ` Linus Torvalds
2024-03-08 0:55 ` 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=29a20fd4-ac5e-44a3-bc8a-9f77aa6a3cf9@paulmck-laptop \
--to=paulmck@kernel.org \
--cc=boqun.feng@gmail.com \
--cc=dave@stgolabs.net \
--cc=frederic@kernel.org \
--cc=jiangshanlai@gmail.com \
--cc=joel@joelfernandes.org \
--cc=josh@joshtriplett.org \
--cc=lilinke99@qq.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=qiang.zhang1211@gmail.com \
--cc=quic_neeraju@quicinc.com \
--cc=rcu@vger.kernel.org \
--cc=rostedt@goodmis.org \
--cc=torvalds@linux-foundation.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
all inboxes | Powered by JetHome®