linux-kernel - Re: [PATCH net] net: bridge: locally receive all multicast packets if IFF

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:
 <TYCPR01MB8437DA180663B75AECDEB24798D62@TYCPR01MB8437.jpnprd01.prod.outlook.com>
Date: Mon, 10 Mar 2025 22:13:17 +0800
From: Shengyu Qu <wiagn233@...look.com>
To: Felix Fietkau <nbd@....name>, Nikolay Aleksandrov <razor@...ckwall.org>,
 netdev@...r.kernel.org, Roopa Prabhu <roopa@...dia.com>,
 "David S. Miller" <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
 Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
 Simon Horman <horms@...nel.org>
Cc: wiagn233@...look.com, bridge@...ts.linux.dev, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net] net: bridge: locally receive all multicast packets if
 IFF_ALLMULTI is set

Ping for next version.

在 2025/2/17 20:58, Felix Fietkau 写道:
> On 17.02.25 12:54, Nikolay Aleksandrov wrote:
>> On 2/17/25 13:26, Felix Fietkau wrote:
>>> If multicast snooping is enabled, multicast packets may not always 
>>> end up on
>>> the local bridge interface, if the host is not a member of the multicast
>>> group. Similar to how IFF_PROMISC allows all packets to be received 
>>> locally,
>>> let IFF_ALLMULTI allow all multicast packets to be received.
>>>
>>> Signed-off-by: Felix Fietkau <nbd@....name>
>>> ---
>>>  net/bridge/br_input.c | 2 ++
>>>  1 file changed, 2 insertions(+)
>>>
>>> diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c
>>> index 232133a0fd21..7fa2da6985b5 100644
>>> --- a/net/bridge/br_input.c
>>> +++ b/net/bridge/br_input.c
>>> @@ -155,6 +155,8 @@ int br_handle_frame_finish(struct net *net, 
>>> struct sock *sk, struct sk_buff *skb
>>>              pkt_type = BR_PKT_MULTICAST;
>>>              if (br_multicast_rcv(&brmctx, &pmctx, vlan, skb, vid))
>>>                  goto drop;
>>> +            if (br->dev->flags & IFF_ALLMULTI)
>>> +                local_rcv = true;
>>>          }
>>>      }
>>>
>>
>> This doesn't look like a bug fix, IMO it should be for net-next.
>>
>> Also you might miss a mcast stat increase, see the multicast code
>> below, the only case that this would cover is the missing "else"
>> branch of:
>>                         if ((mdst && mdst->host_joined) ||
>>                              br_multicast_is_router(brmctx, skb)) {
>>                                  local_rcv = true;
>>                                  DEV_STATS_INC(br->dev, multicast);
>>                          }
>>
>> So I'd suggest to augment the condition and include this ALLMULTI 
>> check there,
>> maybe with a comment to mention that all other cases are covered by 
>> the current
>> code so people are not surprised.
> Will do, thanks.
> 
>> By the way what is the motivation for supporting this flag? I mean you 
>> can
>> make the bridge mcast router and it will receive all mcast anyway.
> 
> OpenWrt uses a user space daemon for DHCPv6/RA/NDP handling, and in 
> relay mode it sets the ALLMULTI flag in order to receive all relevant 
> queries on the network.
> This works for normal network interfaces and non-snooping bridges, but 
> not snooping bridges (unless, as you pointed out, multicast routing is 
> enabled).
> 
> - Felix


Download attachment "OpenPGP_0xE3520CC91929C8E7.asc" of type "application/pgp-keys" (6869 bytes)

Download attachment "OpenPGP_signature.asc" of type "application/pgp-signature" (841 bytes)