Re: [PATCH net 0/2] bridge/vxlan: fix reading neigh ha without synchronization

On 17/08/2026 18:25, Nikolay Aleksandrov wrote:
> On 17/08/2026 17:36, Nikolay Aleksandrov wrote:
>> Hi,
>> Neigh ha address must be read using the seqlock to get a stable snapshot.
>> Both the bridge and vxlan read it directly and can see partial updates.
>> I reproduced both issues with running neigh updates and exercising these
>> paths in parallel and saw partial addresses, e.g. updating between
>> neigh A: 02:00:00:00:00:00 neigh B: fe:ff:ff:ff:ff:ff was able to observe
>> 02:00:ff:ff:ff:ff and fe:ff:00:00:00:00 in packets. Noticed this initially
>> in the bridge, then checked vxlan and its arp/neigh_reduce functions have
>> the same bug, route_shortcircuit is doing the right thing already.
>>
>> Cheers,
>>   Nik
>>
>> Nikolay Aleksandrov (2):
>>    net: bridge: arp/nd proxy: fix reading neigh ha
>>    vxlan: fix reading neigh ha
>>
>>   drivers/net/vxlan/vxlan_core.c | 20 +++++++++++++-------
>>   net/bridge/br_arp_nd_proxy.c   | 24 ++++++++++++++----------
>>   2 files changed, 27 insertions(+), 17 deletions(-)
>>
> 
> (-CC bouncing emails)
> 
> Hmm perhaps it's better to use ETH_ALEN instead of MAX_ADDR_LEN to explicitly
> show the expected len everywhere, it cannot be different anyway.
> I'll do it for v2 after 24h.
> 
> Cheers,
>   Nik
> 
> 
> 

Sashiko brought up valid alignment questions (ether_addr_copy expects 2-byte align).
I'll fix that up before sending v2.

Thanks,
  Nik