Re: [PATCH net-next v5 2/2] bridge: vlan: signal if anything changed on vlan add

[Nikolay Aleksandrov <nikolay@cumulusnetworks.com>]

On 27/10/17 11:24, Toshiaki Makita wrote:
> On 2017/10/26 22:41, Nikolay Aleksandrov wrote:
>> Before this patch there was no way to tell if the vlan add operation
>> actually changed anything, thus we would always generate a notification
>> on adds. Let's make the notifications more precise and generate them
>> only if anything changed, so use the new bool parameter to signal that the
>> vlan was updated. We cannot return an error because there are valid use
>> cases that will be broken (e.g. overlapping range add) and also we can't
>> risk masking errors due to calls into drivers for vlan add which can
>> potentially return anything.
>>
>> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
> 
> Thanks for the explanation in the previous mail.
> 
>> ---
> ...
>> @@ -576,8 +592,11 @@ int br_vlan_add(struct net_bridge *br, u16 vid, u16 flags)
>>  			refcount_inc(&vlan->refcnt);
>>  			vlan->flags |= BRIDGE_VLAN_INFO_BRENTRY;
>>  			vg->num_vlans++;
>> +			*changed = true;
>>  		}
>> -		__vlan_add_flags(vlan, flags);
>> +		if (__vlan_add_flags(vlan, flags))
>> +			*changed = true;
>> +
>>  		return 0;
>>  	}
>>  
>> @@ -601,6 +620,7 @@ int br_vlan_add(struct net_bridge *br, u16 vid, u16 flags)
>>  		free_percpu(vlan->stats);
>>  		kfree(vlan);
> 
> We should not set changed=true in this failure case?
> 

Right, I don't know what I was thinking, clearly I missed the fact that we directly
clean up and return "ret" with an error. Very good catch, thanks! That should
go in the else branch, will send a v6 later with that fixed up.

Thanks,
 Nik

>>  	}
>> +	*changed = true;
>>  
>>  	return ret;
>>  }
> ...
>> @@ -1050,6 +1083,7 @@ int nbp_vlan_add(struct net_bridge_port *port, u16 vid, u16 flags)
>>  	ret = __vlan_add(vlan, flags);
>>  	if (ret)
>>  		kfree(vlan);
> 
> Likewise.
> 
>> +	*changed = true;
>>  
>>  	return ret;
>>  }
>>
>