Re: [PATCH 7/7] net: dsa: b53: Expose PTP timestamping ioctls to userspace

On Fri, 26 Nov 2021 at 18:31, Richard Cochran [off-list ref] wrote:

On Fri, Nov 26, 2021 at 09:42:32AM +0100, Kurt Kanzenbach wrote:

quoted

On Thu Nov 25 2021, Vladimir Oltean wrote:

quoted

Richard, when the request is PTP_V2_EVENT and the response is
PTP_V2_L2_EVENT, is that an upgrade or a downgrade?

It is a downgrade, isn't it?

Yes.  "Any kind of PTP Event" is a superset of "Any Layer-2 Event".

When userland asks for "any kind", then it wants to run PTP over IPv4,
IPv6, or Layer2, maybe even more than one at the same time.  If the
driver changes that to Layer2 only, then the PTP possibilities have
been downgraded.

Well, when I said that it's essentially the same pattern, this is what
I was talking about. The b53 driver downgrades everything and the
kitchen sink to HWTSTAMP_FILTER_PTP_V2_L2_EVENT, the ocelot driver to
HWTSTAMP_FILTER_PTP_V2_EVENT, and both are buggy for the same reason.
I don't see why you mention that there is an important difference
between HWTSTAMP_FILTER_PTP_V2_L2_EVENT and
HWTSTAMP_FILTER_PTP_V2_EVENT. I know there is, but the _pattern_ is
the same.
I'm still missing something obvious, aren't I?