Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Sat, 6 Oct 2018 18:47:29 -0400
From:   "Theodore Y. Ts'o" <tytso@mit.edu>
To:     Joel Fernandes <joel@joelfernandes.org>
Cc:     "Paul E. McKenney" <paulmck@linux.ibm.com>,
        linux-kernel@vger.kernel.org, Jonathan Corbet <corbet@lwn.net>,
        Josh Triplett <josh@joshtriplett.org>,
        Lai Jiangshan <jiangshanlai@gmail.com>,
        linux-doc@vger.kernel.org,
        Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        Steven Rostedt <rostedt@goodmis.org>, pantin@google.com
Subject: Re: [PATCH RFC 0/5] rcu doc updates for whatisRCU and checklist
Message-ID: <20181006224729.GA7130@thunk.org>
Mail-Followup-To: "Theodore Y. Ts'o" <tytso@mit.edu>,
        Joel Fernandes <joel@joelfernandes.org>,
        "Paul E. McKenney" <paulmck@linux.ibm.com>,
        linux-kernel@vger.kernel.org, Jonathan Corbet <corbet@lwn.net>,
        Josh Triplett <josh@joshtriplett.org>,
        Lai Jiangshan <jiangshanlai@gmail.com>, linux-doc@vger.kernel.org,
        Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        Steven Rostedt <rostedt@goodmis.org>, pantin@google.com
References: <20181005231815.170433-1-joel@joelfernandes.org>
 <20181005234628.GB2548@thunk.org>
 <20181006034540.GM2674@linux.ibm.com>
 <20181006044057.GA41279@joelaf.mtv.corp.google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20181006044057.GA41279@joelaf.mtv.corp.google.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Oct 05, 2018 at 09:40:57PM -0700, Joel Fernandes wrote:
> 
> Could you annotate this pointer (sbi->s_qf_names) with __rcu so it can be
> checked by sparse for proper usage? Its also point #16 in the checklist.txt
> RCU document. I enclosed a diff to do this below.

Sure.

> I also saw a bunch of places in super.c where the pointer isn't accessed from
> an rcu read section or rcu_dereference, but it was a quick look so sorry if I
> missed something. If its true, then are you planning to convert these to use
> rcu_dereference and wrapped by an rcu_read_lock/unlock as well?

These are places where we are holding s_umount, so there's no way
s_qf_names can change out from under us.  So the conversion should be
to use rcu_dereference_protected().

> > +		to_free[i] = rcu_dereference_protected(sbi->s_qf_names[i],
> > +						       &sb->s_umount);
> 
> Also should this be the following?
> 		to_free[i] = rcu_dereference_protected(sbi->s_qf_names[i],
> 					       lockdep_is_held(&sb->s_umount));

Yep, good catch, thanks.

						- Ted