czw., 28 lip 2022 o 21:16 Vladimir Oltean [off-list ref] napisał(a):

On Thu, Jul 28, 2022 at 06:56:48PM +0200, Marcin Wojtas wrote:

quoted

There was a regression even for OF in v1, but after switching to
device_match_fwnode() it works indeed. Anyway patch v4 is imo useful,
I'll only reword the commit message.

Do you mean patch 4 or patch v4? If patch 4, of course it's useful, but

Patch 4/8 in v4 :) I'm working on it right now to submit asap.

not for avoiding a regression with OF (case in which I drop all my claims
made earlier about fw_find_net_device_by_node), but rather to actually

Change in the mvpp2 driver:
-       dev->dev.of_node = port_node;
+       device_set_node(&dev->dev, port_fwnode);
is desired and correct anyway, so as a low-cost change I think it can
be included in this series (which is in fact preparation-to-ACPI
support). I will update the commit message. accordingly.

get something working with actual ACPI (although perhaps not in this
series, you'll need to add ACPI IDs in the mv88e6xxx driver some time

v1 added all of this, but we agreed that ACPI-specific bits should be
sent separately later, after extending the ACPI Specification.

later as well, maybe you could focus this series just on converting DSA
to play nice with fwnodes). If you're already thinking about the v4 of
this patch set, I'll respond to that in a separate email shortly.