Re: [PATCH] bridge: mask forwarding of IEEE 802 local multicast groups

[Nick Carter <ncarter100@gmail.com>]

On 1 July 2011 23:37, David Lamparter <equinox@diac24.net> wrote:
> On Fri, Jul 01, 2011 at 10:21:44PM +0100, Nick Carter wrote:
>> Introduce sysfs ../bridge/group_fwd_mask attribute so users can
>> configure which group mac addresses are forwarded.
>>
>> These diffs do not change the default behaviour of bridge.ko.  By
>> changing the group_fwd_mask value users can select any combination of
>> the 01-80-C2-00-00-00 - 01-80-C2-00-00-0F addresses to be forwarded.
>>
>> Signed-off-by: Nick Carter <ncarter100@gmail.com>
>
> Hm. Makes it very easy to shoot yourself in the foot, but...
>
>> +     br->group_fwd_mask = 0;
>
> ... but the default is safe, so we're following the unix philosophy of
> handing out the gun to shoot yourself :)
>
> Reviewed-by: David Lamparter <equinox@diac24.net>
>
>
> PS:
>> @@ -166,6 +166,9 @@ struct sk_buff *br_handle_frame(struct sk_buff *skb)
>>               if (p->br->stp_enabled == BR_NO_STP && dest[5] == 0)
>>                       goto forward;
>
> Not sure if we still need the STP one when we have the mask, it should
> probably go away for consistency. It would change existing behaviour
> though... either way it can be a separate patch.
Yes I agree.  Probably best to wait until the user space tools are
updated to support the new mask, so users can easily configure a
workaround to the change in behaviour.  I'm happy to update brctl to
support the mask, I pretty much have the diffs already.
Nick

>
>> +             if (p->br->group_fwd_mask & (1 << dest[5]))
>> +                     goto forward;
>> +
>>               if (NF_HOOK(NFPROTO_BRIDGE, NF_BR_LOCAL_IN, skb, skb->dev,
>>                           NULL, br_handle_local_finish))
>>                       return NULL;    /* frame consumed by filter */
>