[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <512B1A54.1030004@cn.fujitsu.com>
Date: Mon, 25 Feb 2013 16:01:24 +0800
From: Gao feng <gaofeng@...fujitsu.com>
To: David Miller <davem@...emloft.net>
CC: netdev@...r.kernel.org
Subject: Re: [PATCH] net: remove the unnecessary list_del
On 2013/02/25 13:39, David Miller wrote:
> From: Gao feng <gaofeng@...fujitsu.com>
> Date: Mon, 25 Feb 2013 13:03:04 +0800
>
>> On 2013/02/25 12:54, David Miller wrote:
>>> From: Gao feng <gaofeng@...fujitsu.com>
>>> Date: Mon, 25 Feb 2013 12:41:56 +0800
>>>
>>>> These lists are used by unregister_netdevice_many,
>>>> they are local variables,will not be seen by others.
>>>> there is no need to delete them.
>>>>
>>>> Signed-off-by: Gao feng <gaofeng@...fujitsu.com>
>>>
>>> What about the devices on the list? The ones at the front and the
>>> rear of the list will have list pointers that point into no longer
>>> valid stack frame locations.
>>
>> These lists are only used by unregister_netdevice_many,
>> we don't access these lists again after unregister_netdevice_many,
>> so I think it doesn't have invalid stack frame locations problems.
>
> I do not see unregister_netdevice_many() list_del()'ing the devices
> on the list, therefore those netdevice objects have pointers into
> the stale stack frame.
>
Yes, I agree that this will make the netdevice objects have pointers
into the stale stack frame.
I check the codes,and found there are tons of codes that don't call
list_del after unregister_netdevice_many(). such as mroute_clean_tables,
ip6_tnl_destroy_tunnels...
I think it's better to make sure netdevice objects have valid pointers
by calling list_del in unregister_netdevice_many.
> You cannot make this change.
>
Will send another patch.
Thanks
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists