Re: [PATCH] atm: lec use dev_change_mtu

[Yang Hongyang <yanghy@cn.fujitsu.com>]

David Miller wrote:
> From: Yang Hongyang <yanghy@cn.fujitsu.com>
> Date: Fri, 20 Mar 2009 15:45:20 +0800
> 
>> Stephen Hemminger wrote:
>>> Rather than calling device pointer directly (which is incorrect with
>>> net_device_ops), use the standard dev_change_mtu. Compile tested only.
>>>
>>> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
>>>
>>> --- a/net/atm/lec.c	2009-03-19 17:18:35.140652628 -0700
>>> +++ b/net/atm/lec.c	2009-03-19 17:19:19.500777355 -0700
>>> @@ -502,7 +502,7 @@ static int lec_atm_send(struct atm_vcc *
>>>  		priv->lane2_ops = NULL;
>>>  		if (priv->lane_version > 1)
>>>  			priv->lane2_ops = &lane2_ops;
>>> -		if (dev->change_mtu(dev, mesg->content.config.mtu))
>>> +		if (dev_set_mtu(dev, mesg->content.config.mtu))
>> I think it's not correct.This dev->change_mtu is:
>>  663 static int lec_change_mtu(struct net_device *dev, int new_mtu)
>>  664 {
>>  665         if ((new_mtu < 68) || (new_mtu > 18190))
>>  666                 return -EINVAL;
>>  667         dev->mtu = new_mtu;
>> 		^^^^^^^^^^^^^^^^maybe should use dev_set_mtu() here.
>>  668         return 0;
>>  669 }
>>
>> and it's not quite same as dev_set_mtu.
> 
> dev_set_mtu() calls netdev->ops->ndo_set_mtu() so it will call
> this function you are quoting.

got it.:)

> 
> 


-- 
Regards
Yang Hongyang