Re: [PATCH net-next v2 3/3] bonding: convert pr_xxx() to pr_xxx_ratelimited() for arp interval

[Ding Tianhong <dingtianhong@huawei.com>]

On 2014/3/18 19:14, Joe Perches wrote:
> On Tue, 2014-03-18 at 04:01 -0700, Joe Perches wrote:
>> On Tue, 2014-03-18 at 18:43 +0800, Ding Tianhong wrote:
>>> The debug log in the arp interval should be rate limited, otherwise would
>>> occur spam the log, so convert them.
> []
>> Another way to do this is to use the
>> more global net_ratelimit() before
>> each existing pr_<level>.
>>
>> Not suggesting one or the other is
>> right or wrong here, it's just an option.
> 
> Another option is to use the net_<level>_ratelimited
> functions, but I'm not sure these are used much
> outside of net/.
> 
>>> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
>>
>>> @@ -2131,32 +2131,33 @@ static void bond_arp_send(struct net_device *slave_dev, int arp_op,
>>>  {
>>>  	struct sk_buff *skb;
>>>  
>>> -	pr_debug("arp %d on slave %s: dst %pI4 src %pI4\n",
>>> -		 arp_op, slave_dev->name, &dest_ip, &src_ip);
>>> +	pr_debug_ratelimited("arp %d on slave %s: dst %pI4 src %pI4\n",
>>> +			     arp_op, slave_dev->name, &dest_ip, &src_ip);
>>
>> 	if (net_ratelimit())
>> 		pr_debug(etc...)
> 
> or
> 
> 	net_dbg_ratelimited(etc...)
> 
> 
Thanks for these opinion. It looks like the net_ratelimit() is more reasonable.

It really need to spend time to distinguish which one is better here.
sometimes more choices more headaches. :)

Regards
Ding
> 
>