lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <4b9ae5bf-27fa-c6ac-6de2-0c8d0c3e6ab0@nvidia.com>
Date:   Tue, 31 Aug 2021 12:10:43 +0300
From:   Nikolay Aleksandrov <nikolay@...dia.com>
To:     Joachim Wiberg <troglobit@...il.com>,
        Nikolay Aleksandrov <razor@...ckwall.org>,
        netdev@...r.kernel.org
Cc:     roopa@...dia.com, dsahern@...il.com
Subject: Re: [PATCH iproute2-next 06/17] bridge: vlan: add global
 mcast_igmp_version option

On 31/08/2021 12:04, Nikolay Aleksandrov wrote:
> On 31/08/2021 12:02, Joachim Wiberg wrote:
>>
>> Hi Nik,
>>
>> awesome to see this patchset! :-)  I've begun setting things up here
>> for testing.  Just have a question about this:
>>
>> On Thu, Aug 26, 2021 at 16:05, Nikolay Aleksandrov <razor@...ckwall.org> wrote:
>>> Add control and dump support for the global mcast_igmp_version option
>>> which controls the IGMP version on the vlan (default 2).
>>
>> Why is the default IGMPv2?  Since we support IGMPv3, surely that should
>> be the default, with fallback to IGMPv2 when we detect end devices that
>> don't support v3?
>>
>> The snooping RFC refers back to the IGMPv3 RFC
>>
>>   https://datatracker.ietf.org/doc/html/rfc3376#section-7
>>
>> I noticed the default for MLD is also set to the older version v1, and
>> I'm guessing there's a reasoning behind both that I haven't yet grasped.
>>
>> Best regards
>>  /Joachim
>>
> 
> Hi,
> The reason is to be consistent with the bridge config. It already has IGMPv2 as default.
> I added IGMPv3 support much later and we couldn't change the default.
> I'd prefer not to surprise users and have different behaviour when they switch snooping
> from bridge-wide to per-vlan. It is a config option after all, distributions can choose
> to change their default when they create devices. :)
> 
> Cheers,
>  Nik
> 

Forgot to mention - the RFC fallback and compatibility logic is not implemented yet, so if
we set v3 we'll be stuck with it even if there are v2 participants. In fact it's a bit of
a mess when mixing v2 and v3 right now, their interop needs more work.

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ