On 11/20/25 6:49 PM, Dmitry Skorodumov wrote:

Make the addrs_lock be per port, not per ipvlan dev.

This appears to be a very minor problem though.
Since it's highly unlikely that ipvlan_add_addr() will
be called on 2 CPU simultaneously. But nevertheless,
this may cause:

1. False-negative of ipvlan_addr_busy(): one interface
iterated through all port->ipvlans + ipvlan->addrs
under some ipvlan spinlock, and another added IP
under its own lock. Though this is only possible
for IPv6, since looks like only ipvlan_addr6_event() can be
called without rtnl_lock.

2. Race since ipvlan_ht_addr_add(port) is called under
different ipvlan->addrs_lock locks

This should not affect performance, since add/remove IP
is a rare situation and spinlock is not locked on fast
paths.

Also, it's quite convenient to have addrs_lock on
ipvl_port, to dynamically prevent conflict of IPs
with addresses on main port.

CC: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Dmitry Skorodumov <redacted>

This really looks like a fix, that should go via the 'net' tree with a
suitable fixes tag.

/P