Re: [PATCH 3/4] bonding: Fix work rearming

[Makito SHIOKAWA <mshiokawa@miraclelinux.com>]

This patch is supposing a case that bond_mii_monitor() is invoked in 
bond_open(), and after that, 0 is set to miimon via sysfs (see same place on 
other monitors).
Though message in bonding_store_miimon() says miimon value 1-INT_MAX rejected, 
but it looks like 0 can be accepted and monitor must be stopped in that case.

>> Change code not to rearm bond_mii_monitor() when value 0 is set for miimon.
>>
>> Signed-off-by: Makito SHIOKAWA <mshiokawa@miraclelinux.com>
>> ---
>>  drivers/net/bonding/bond_main.c |    3 ++-
>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> --- a/drivers/net/bonding/bond_main.c
>> +++ b/drivers/net/bonding/bond_main.c
>> @@ -2388,7 +2388,8 @@ void bond_mii_monitor(struct work_struct
>>  
>>  	delay = ((bond->params.miimon * HZ) / 1000) ? : 1;
>>  	read_unlock(&bond->lock);
>> -	queue_delayed_work(bond->wq, &bond->mii_work, delay);
>> +	if (bond->params.miimon)
>> +		queue_delayed_work(bond->wq, &bond->mii_work, delay);
>>  }
> 
> Maybe I miss something, but is this bond_mii_monitor() function
> supposed to be ever started if (!bond->params.miimon)? (IOW: isn't
> it enough to control this where the parameter is changed only?)
> 
> Regards,
> Jarek P.

--
Makito SHIOKAWA
MIRACLE LINUX CORPORATION