From mboxrd@z Thu Jan 1 00:00:00 1970 From: Gao feng Subject: Re: [PATCH] net: remove the unnecessary list_del Date: Mon, 25 Feb 2013 17:04:23 +0800 Message-ID: <512B2917.4010909@cn.fujitsu.com> References: <512AF088.5040206@cn.fujitsu.com> <20130225.003931.1364384031279092217.davem@davemloft.net> <512B1A54.1030004@cn.fujitsu.com> <20130225.031256.202114640375754800.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org To: David Miller Return-path: Received: from cn.fujitsu.com ([222.73.24.84]:55951 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1758208Ab3BYJDn (ORCPT ); Mon, 25 Feb 2013 04:03:43 -0500 In-Reply-To: <20130225.031256.202114640375754800.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: On 2013/02/25 16:12, David Miller wrote: > From: Gao feng > Date: Mon, 25 Feb 2013 16:01:24 +0800 > >> I think it's better to make sure netdevice objects have valid pointers >> by calling list_del in unregister_netdevice_many. > > Please, for the sake of understanding, read the commit below. And > please make such necessary research in the future and reference any > such commits in your proposed patch, as well as explaining why those > commits in the past were wrong and why you're reverting of them is > correct. > > Just saying "this isn't necessary" is a very disappointing explanation > for something as very non-trivial as this is. Get it, sorry for the noise :( Will do more research before making patch. Thanks, the commit below is very useful. > > Thank you. > > commit ceaaec98ad99859ac90ac6863ad0a6cd075d8e0e > Author: Eric Dumazet > Date: Thu Feb 17 22:59:19 2011 +0000 > > net: deinit automatic LIST_HEAD > > commit 9b5e383c11b08784 (net: Introduce > unregister_netdevice_many()) left an active LIST_HEAD() in > rollback_registered(), with possible memory corruption. > > Even if device is freed without touching its unreg_list (and therefore > touching the previous memory location holding LISTE_HEAD(single), better > close the bug for good, since its really subtle. > > (Same fix for default_device_exit_batch() for completeness) > > Reported-by: Michal Hocko > Tested-by: Michal Hocko > Reported-by: Eric W. Biderman > Tested-by: Eric W. Biderman > Signed-off-by: Linus Torvalds > Signed-off-by: Eric Dumazet > CC: Ingo Molnar > CC: Octavian Purdila > CC: stable [.33+] > Signed-off-by: David S. Miller > > diff --git a/net/core/dev.c b/net/core/dev.c > index a18c164..8ae6631 100644 > --- a/net/core/dev.c > +++ b/net/core/dev.c > @@ -5066,6 +5066,7 @@ static void rollback_registered(struct net_device *dev) > > list_add(&dev->unreg_list, &single); > rollback_registered_many(&single); > + list_del(&single); > } > > unsigned long netdev_fix_features(unsigned long features, const char *name) > @@ -6219,6 +6220,7 @@ static void __net_exit default_device_exit_batch(struct list_head *net_list) > } > } > unregister_netdevice_many(&dev_kill_list); > + list_del(&dev_kill_list); > rtnl_unlock(); > } > > -- > To unsubscribe from this list: send the line "unsubscribe netdev" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html >