Re: [PATCH net-next v3 2/2] netlink: specs: add ovs datapath and vport set operations

[Minxi Hou <houminxi@gmail.com>]

On Tue, 20 Oct 2026 16:24:42 -0400 Jakub Kicinski wrote:
> Should this doc also say that per-cpu-pids only takes effect when the
> same request sets OVS_DP_F_DISPATCH_UPCALL_PER_CPU in user-features?

I left that out on purpose. The doc covers the half that changes
state: a missing user-features attribute is treated as zero and clears
the features. per-cpu-pids is only read when that flag is set, and a
request without it succeeds with the pids dropped. The kernel already
behaves this way.

> Could these notifications be described in the specs, either here or
> in a follow-up?

A follow-up. new and del have the same gap in both specs: the handlers
send through ovs_notify(), so do() returns only an ack. Describing the
notifications means adding reply and notify to every write op, which is
more than this series set out to do.

> Should port-no be listed here too?

Not on set alone. lookup_vport() accepts a port number for new, del and
set, but the spec lists port-no only on del. Adding it to set would
make the three lookup keys disagree in a different way. I will fix all
three together, separately.

> Could options be dropped from this request, or could the doc mention
> these limits?

The doc now mentions them. The list stays: the handler reads both
attributes, rejects options with -EOPNOTSUPP, and rejects a type that
differs from the vport's type with -EINVAL. The list is what the
handler reads, not what it accepts.