2018-03-12 11:18:08

by Linus Lüssing

[permalink] [raw]
Subject: Re: Problem with bridge (mcast-to-ucast + hairpin) and Broadcom's 802.11f in their FullMAC fw

On Tue, Feb 27, 2018 at 11:08:20AM +0100, Rafał Miłecki wrote:
> I've problem when using OpenWrt/LEDE on a home router with Broadcom's
> FullMAC WiFi chipset.

Hi Rafał,

Thanks for reporting this issue!

> Can you see any solution for this problem? Is that an option to stop
> multicast-to-unicast from touching 802.11f packets? Some other ideas?
> Obviously I can't modify Broadcom's firmware and drop that obsoleted
> standard.

Just to avoid some potential confusion: This is more an issue of
hairpinning than it is an issue of multicast-to-unicast per se,
right?

That is, if you set this to 0 manually:
/sys/class/net/<ap-iface>/brport/multicast_to_unicast
Then the issue still occurs, right?

Regards, Linus


2018-03-12 21:49:52

by Rafał Miłecki

[permalink] [raw]
Subject: Re: Problem with bridge (mcast-to-ucast + hairpin) and Broadcom's 802.11f in their FullMAC fw

On 12 March 2018 at 12:08, Linus L=C3=BCssing <[email protected]> wr=
ote:
> On Tue, Feb 27, 2018 at 11:08:20AM +0100, Rafa=C5=82 Mi=C5=82ecki wrote:
>> I've problem when using OpenWrt/LEDE on a home router with Broadcom's
>> FullMAC WiFi chipset.
>
> Hi Rafa=C5=82,
>
> Thanks for reporting this issue!
>
>> Can you see any solution for this problem? Is that an option to stop
>> multicast-to-unicast from touching 802.11f packets? Some other ideas?
>> Obviously I can't modify Broadcom's firmware and drop that obsoleted
>> standard.
>
> Just to avoid some potential confusion: This is more an issue of
> hairpinning than it is an issue of multicast-to-unicast per se,
> right?
>
> That is, if you set this to 0 manually:
> /sys/class/net/<ap-iface>/brport/multicast_to_unicast
> Then the issue still occurs, right?

I can't really tell, I didn't go into details on bridge +
mcast-to-unicast + hairpin.

By default OpenWrt/LEDE sets
/sys/class/net/<ap-iface>/brport/multicast_to_unicast
to 1. Changing it to 0 (with a simple echo) doesn't /fix/ the problem.

--=20
Rafa=C5=82

2018-03-12 11:48:17

by Linus Lüssing

[permalink] [raw]
Subject: Re: Problem with bridge (mcast-to-ucast + hairpin) and Broadcom's 802.11f in their FullMAC fw

On Mon, Mar 12, 2018 at 12:08:56PM +0100, Linus Lüssing wrote:
> On Tue, Feb 27, 2018 at 11:08:20AM +0100, Rafał Miłecki wrote:
> > I've problem when using OpenWrt/LEDE on a home router with Broadcom's
> > FullMAC WiFi chipset.
>
> Hi Rafał,
>
> Thanks for reporting this issue!
>
> > Can you see any solution for this problem? Is that an option to stop
> > multicast-to-unicast from touching 802.11f packets? Some other ideas?
> > Obviously I can't modify Broadcom's firmware and drop that obsoleted
> > standard.
>
> Just to avoid some potential confusion: This is more an issue of
> hairpinning than it is an issue of multicast-to-unicast per se,
> right?
>
> That is, if you set this to 0 manually:
> /sys/class/net/<ap-iface>/brport/multicast_to_unicast
> Then the issue still occurs, right?

Btw., if in OpenWRT/LEDE you set 'option multicast_to_unicast 0'
in /etc/config/network for the bridge (and not bridge port)
then netifd should refrain from setting the bridge hairpinning and
AP isolation on wireless devices, too, if I remember correctly.

Can you confirm that the issue disappears for you then?

2018-03-12 21:52:16

by Rafał Miłecki

[permalink] [raw]
Subject: Re: Problem with bridge (mcast-to-ucast + hairpin) and Broadcom's 802.11f in their FullMAC fw

On 12 March 2018 at 12:48, Linus L=C3=BCssing <[email protected]> wr=
ote:
> On Mon, Mar 12, 2018 at 12:08:56PM +0100, Linus L=C3=BCssing wrote:
>> On Tue, Feb 27, 2018 at 11:08:20AM +0100, Rafa=C5=82 Mi=C5=82ecki wrote:
>> > I've problem when using OpenWrt/LEDE on a home router with Broadcom's
>> > FullMAC WiFi chipset.
>>
>> Hi Rafa=C5=82,
>>
>> Thanks for reporting this issue!
>>
>> > Can you see any solution for this problem? Is that an option to stop
>> > multicast-to-unicast from touching 802.11f packets? Some other ideas?
>> > Obviously I can't modify Broadcom's firmware and drop that obsoleted
>> > standard.
>>
>> Just to avoid some potential confusion: This is more an issue of
>> hairpinning than it is an issue of multicast-to-unicast per se,
>> right?
>>
>> That is, if you set this to 0 manually:
>> /sys/class/net/<ap-iface>/brport/multicast_to_unicast
>> Then the issue still occurs, right?
>
> Btw., if in OpenWRT/LEDE you set 'option multicast_to_unicast 0'
> in /etc/config/network for the bridge (and not bridge port)
> then netifd should refrain from setting the bridge hairpinning and
> AP isolation on wireless devices, too, if I remember correctly.
>
> Can you confirm that the issue disappears for you then?

Yes, absolutely. This reverts OpenWrt/LEDE to the old setup (no
mcast-to-ucast + hairpin) and it works. I was hoping we can make
mcast-to-ucast + hairpin work with Broadcom's 802.11f however instead
of moving OpenWrt/LEDE back to the old setup.

--=20
Rafa=C5=82