Re: [PATCH net v4 1/2] bonding: fix initial last_rx vs ARP-monitor slack window

[Ramses]

> [Severity: Low]
> This isn't a bug, but is "until a real reply arrives" accurate? More than
> ARP replies can move last_rx forward.

Right, with arp_validate=none any received frame updates last_rx. v5
says "until it receives traffic".

> [Severity: Medium]
> This seed uses the arp_interval in effect at enslave time. What happens if
> ARP monitoring is turned on later, or the interval is raised?

That issue already existed before this series. The seed was already
 computed from arp_interval at enslave time and is not refreshed when the
 ARP monitor starts, from bond_open() or bond_option_arp_interval_set().
 This patch does not make it worse. Reseeding the slaves when the monitor
 starts touches different code paths, so I would rather do that in a separate
follow-up than grow the current patch series.

pw-bot: cr