netdev - Re: [PATCH net-next 2/4] net: dev: don't set the same mac address for netdev

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <53905791.4060401@huawei.com>
Date:	Thu, 5 Jun 2014 19:42:09 +0800
From:	Ding Tianhong <dingtianhong@...wei.com>
To:	Toshiaki Makita <makita.toshiaki@....ntt.co.jp>, <kaber@...sh.net>,
	<davem@...emloft.net>, <edumazet@...gle.com>, <vyasevic@...hat.com>
CC:	<netdev@...r.kernel.org>
Subject: Re: [PATCH net-next 2/4] net: dev: don't set the same mac address
 for netdev

On 2014/6/5 18:51, Toshiaki Makita wrote:
> (2014/06/05 18:50), Ding Tianhong wrote:
>> On 2014/6/5 17:09, Toshiaki Makita wrote:
>>> (2014/06/05 15:50), Ding Tianhong wrote:
>>>> Most of netdev just like bond, team, vlan will set the mac address
>>>> and propagate to the upperdev or lowerdev regardless the mac address
>>>> is  same or not, I could not find that the same mac address could
>>>> make affect, so add equal check when set mac address.
>>>>
>>>> Signed-off-by: Ding Tianhong <dingtianhong@...wei.com>
>>>> ---
>>>>  net/core/dev.c | 2 ++
>>>>  1 file changed, 2 insertions(+)
>>>>
>>>> diff --git a/net/core/dev.c b/net/core/dev.c
>>>> index 5367bfb..4008a51 100644
>>>> --- a/net/core/dev.c
>>>> +++ b/net/core/dev.c
>>>> @@ -5570,6 +5570,8 @@ int dev_set_mac_address(struct net_device *dev, struct sockaddr *sa)
>>>>  		return -EINVAL;
>>>>  	if (!netif_device_present(dev))
>>>>  		return -ENODEV;
>>>> +	if (ether_addr_equal_64bits(dev->dev_addr, sa->sa_data))
>>>> +		return 0;
>>>>  	err = ops->ndo_set_mac_address(dev, sa);
>>>>  	if (err)
>>>>  		return err;
>>>>
>>>
>>> Bridge uses addr_assign_type to check if bridge_id can be propageted by
>>> bridge ports. If user set mac address, and even if it is the same as
>>> current one, bridge uses the fact that the mac address is set by user.
>>>
>>
>> OK
>>
>>> Although I'm not aware of a driver that needs calling of
>>> ndo_set_mac_address() for the same mac address, this change looks a bit
>>> risky to me.
>>> (For example, old bridge code needed this call because it managed
>>> BR_SET_MAC_ADDR in bridge flags.)
>>>
>> Except the old bridge, I still don't think any other driver need to call ndo_set_mac_address()
>> for the same mac address, if the dev_set_mac_address() don't do anything for the same address,
>> I think some drivers should ignore the same mac address themselves just like bonding, macvlan, vlan and so on.
> 
> Though I don't know why you think this is safe, looking over some
> drivers, br2684_mac_addr() seems to use a logic similar to old bridge's.
> 

Hm, I miss that, this improvement is risky for some drivers, I will remove this patch
and resend the series later. thanks. 

Ding

> Thanks,
> Toshiaki Makita
> 
> .
> 


--
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