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:   Thu, 15 Aug 2019 11:24:42 -0400
From:   Joel Fernandes <joel@joelfernandes.org>
To:     Matthew Wilcox <willy@infradead.org>
Cc:     linux-kernel@vger.kernel.org,
        Greg Kroah-Hartman <gregkh@linuxfoundation.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>,
        "Paul E. McKenney" <paulmck@linux.ibm.com>,
        "Rafael J. Wysocki" <rafael@kernel.org>, rcu@vger.kernel.org,
        Steven Rostedt <rostedt@goodmis.org>, Tejun Heo <tj@kernel.org>
Subject: Re: [PATCH v3 -rcu] workqueue: Convert for_each_wq to use built-in
 list check
Message-ID: <20190815152442.GB12078@google.com>
References: <20190815141842.GB20599@google.com>
 <20190815145749.GA18474@bombadil.infradead.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20190815145749.GA18474@bombadil.infradead.org>
User-Agent: Mutt/1.10.1 (2018-07-13)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Aug 15, 2019 at 07:57:49AM -0700, Matthew Wilcox wrote:
> On Thu, Aug 15, 2019 at 10:18:42AM -0400, Joel Fernandes (Google) wrote:
> > list_for_each_entry_rcu now has support to check for RCU reader sections
> > as well as lock. Just use the support in it, instead of explicitly
> > checking in the caller.
> 
> ...
> 
> >  #define assert_rcu_or_wq_mutex_or_pool_mutex(wq)			\
> >  	RCU_LOCKDEP_WARN(!rcu_read_lock_held() &&			\
> >  			 !lockdep_is_held(&wq->mutex) &&		\
> 
> Can't you also get rid of this macro?

Could be. But that should be a different patch. I am only cleaning up the RCU
list lockdep checking in this series since the series introduces that
concept).  Please feel free to send a patch for the same.

Arguably, keeping the macro around also can be beneficial in the future.

> It's used in one place:
> 
> static struct pool_workqueue *unbound_pwq_by_node(struct workqueue_struct *wq,
>                                                   int node)
> {
>         assert_rcu_or_wq_mutex_or_pool_mutex(wq);
> 
>         /*
>          * XXX: @node can be NUMA_NO_NODE if CPU goes offline while a
>          * delayed item is pending.  The plan is to keep CPU -> NODE
>          * mapping valid and stable across CPU on/offlines.  Once that
>          * happens, this workaround can be removed.
>          */
>         if (unlikely(node == NUMA_NO_NODE))
>                 return wq->dfl_pwq;
> 
>         return rcu_dereference_raw(wq->numa_pwq_tbl[node]);
> }
> 
> Shouldn't we delete that assert and use
> 
> +	return rcu_dereference_check(wq->numa_pwq_tbl[node],
> +			lockdep_is_held(&wq->mutex) ||
> +			lockdep_is_held(&wq_pool_mutex));

Makes sense. This API also does sparse checking. Also hopefully no sparse
issues show up because rcu_dereference_check() but anyone such issues should
be fixed as well.

thanks,

 - Joel

>