[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAEXW_YS-p9G=fLO+-vYx9RQS9oovHbVs48aPDfGcQrNWUdXwWA@mail.gmail.com>
Date: Mon, 18 Dec 2023 10:27:56 -0500
From: Joel Fernandes <joel@...lfernandes.org>
To: Frederic Weisbecker <frederic@...nel.org>
Cc: linux-kernel@...r.kernel.org, Lai Jiangshan <jiangshanlai@...il.com>,
"Paul E. McKenney" <paulmck@...nel.org>, Josh Triplett <josh@...htriplett.org>,
Steven Rostedt <rostedt@...dmis.org>, Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
Neeraj Upadhyay <neeraj.iitr10@...il.com>, rcu@...r.kernel.org
Subject: Re: [PATCH v2] srcu: Improve comments about acceleration leak
On Mon, Dec 18, 2023 at 7:13 AM Frederic Weisbecker <frederic@...nel.org> wrote:
>
> Le Sun, Dec 17, 2023 at 09:00:15PM -0500, Joel Fernandes a écrit :
> > "Acceleration can never fail because the state of gp_seq value used
> > for acceleration is <= the state of gp_seq used for advancing."
> >
> > Does that sound correct now?
>
> That can be confusing since acceleration relies on rcu_seq_snap() while
> advance relies on rcu_seq_current(). And rcu_seq_snap() returns a snapshot
> that may be above the subsequent rcu_seq_current() return value.
>
> So it should rather be something like:
>
> "The base current gp_seq value used to produce the snapshot has to
> be <= the gp_seq used for advancing."
Yeah "base current gp_seq" though probably equally confusing sounds a
bit better, so I'll just use that instead of "state of gp_seq".
With that can I add your Review tag?
- Joel
Powered by blists - more mailing lists