netdev - Re: bonding: Do we need netlink events for LACP status?

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <15143.1710040289@famine>
Date: Sat, 09 Mar 2024 19:11:29 -0800
From: Jay Vosburgh <jay.vosburgh@...onical.com>
To: Hangbin Liu <liuhangbin@...il.com>
cc: netdev@...r.kernel.org
Subject: Re: bonding: Do we need netlink events for LACP status?

Hangbin Liu <liuhangbin@...il.com> wrote:

>On Mon, Mar 04, 2024 at 10:33:27PM -0800, Jay Vosburgh wrote:
>> Hangbin Liu <liuhangbin@...il.com> wrote:
>> 
>> >A customer asked to add netlink event notifications for LACP bond state
>> >changes. With this, the network monitor could get the LACP state of bonding
>> >and port interfaces, and end user may change the down link port based
>> >on the current LACP state. Do you think if this is a reasonable case
>> >and do able? If yes, I will add it to my to do list.
>> 
>> 	I think I'm going to need some more detail here.
>> 
>> 	To make sure I understand, the suggestion here is to add netlink
>> notifications for transitions in the LACP mux state machine (ATTACHED,
>> COLLECTING, DISTRIBUTING, et al), correct?  If not, then what
>
>Yes, the LACP mux state. Maybe also including the port channel info.
>
>> specifically do you mean?
>> 
>> 	Also, what does "change the down link port" mean?
>
>If a port is down, or in attached state for a period, which means the end
>of port is not in a channel, or the switch crash. The admin could detect this
>via the LACP state notification and remove the port from bond, adding other
>ports to the bond, etc.

	Generally speaking, I don't see an issue with adding these type
of netlink events, as in normal usage the volume will be low.

	Looking at the code, I think it would be a matter of adding the
new IFLA_BOND_LACP_STUFF labels, updating bond_fill_slave_info() and
maybe bond_fill_info() to populate the netlink message for those new
IFLAs.  Add a call to call_netdevice_notifiers() in ad_mux_machine()
when the state changes, using a new event type that would need to be
handled by rtnetlink_event().

	-J

---
	-Jay Vosburgh, jay.vosburgh@...onical.com