From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S1751878AbdBMGA2 (ORCPT <rfc822;w@1wt.eu>);
        Mon, 13 Feb 2017 01:00:28 -0500
Received: from LGEAMRELO11.lge.com ([156.147.23.51]:52110 "EHLO
        lgeamrelo11.lge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751638AbdBMGA1 (ORCPT
        <rfc822;linux-kernel@vger.kernel.org>);
        Mon, 13 Feb 2017 01:00:27 -0500
X-Original-SENDERIP: 156.147.1.151
X-Original-MAILFROM: byungchul.park@lge.com
X-Original-SENDERIP: 10.177.222.33
X-Original-MAILFROM: byungchul.park@lge.com
Date: Mon, 13 Feb 2017 15:00:04 +0900
From: Byungchul Park <byungchul.park@lge.com>
To: Al Viro <viro@ZenIV.linux.org.uk>
Cc: peterz@infradead.org, mingo@kernel.org, koverstreet@google.com,
        neilb@suse.de, nab@linux-iscsi.org, ying.huang@intel.com,
        oleg@redhat.com, asias@redhat.com, shli@kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] llist: Don't reinvent the wheel but use existing llist
 API
Message-ID: <20170213060004.GJ16086@X58A-UD3R>
References: <1486959013-26105-1-git-send-email-byungchul.park@lge.com>
 <20170213054326.GR13195@ZenIV.linux.org.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20170213054326.GR13195@ZenIV.linux.org.uk>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Mon, Feb 13, 2017 at 05:43:26AM +0000, Al Viro wrote:
> On Mon, Feb 13, 2017 at 01:10:13PM +0900, Byungchul Park wrote:
> > Although llist provides proper APIs, they are not used. Make them used.
> 
> > @@ -231,12 +231,10 @@ static void __fput(struct file *file)
> >  static void delayed_fput(struct work_struct *unused)
> >  {
> >  	struct llist_node *node = llist_del_all(&delayed_fput_list);
> > -	struct llist_node *next;
> > +	struct file *f;
> >  
> > -	for (; node; node = next) {
> > -		next = llist_next(node);
> > -		__fput(llist_entry(node, struct file, f_u.fu_llist));
> > -	}
> > +	llist_for_each_entry(f, node, f_u.fu_llist)
> > +		__fput(f);
> >  }
> 
> #define llist_for_each_entry(pos, node, member)                         \
>         for ((pos) = llist_entry((node), typeof(*(pos)), member);       \
>              &(pos)->member != NULL;                                    \
>              (pos) = llist_entry((pos)->member.next, typeof(*(pos)), member))
> 
> Now, think what happens after __fput() frees the damn thing.  In the
> step of that loop, that is.
> 
> That kind of pattern (find next, do something with the current, proceed to
> the next we'd found before) is a strong hint that this "do something"
> might remove the thing from the list, or outright destroy it.  Both
> file_table.c and namespace.c chunks are breaking exactly that kind of
> places.
> 
> Please, don't do this kind of conversions blindly.  There _is_ another
> iterating primitive for such places, but figuring out which one is right
> is not something you can do without understanding what the code is doing.

I'm really sorry. I made a mistake. I should have used the safe version in
the case.

> And no, blind replacement of all such loops with llist_for_each_entry_safe,
> just in case, is not a good idea either.