On 12/09/2021 22:06, John Thomson wrote:
Hi Vadym,
On Wed, 8 Sep 2021, at 09:56, Vadym Kochan wrote:
Hi Jan,
Jan Lübbe <jlu@xxxxxxxxxxxxxx> writes:
…
I think it would be useful to have a way to express this setup for systems with
many interfaces, but am unsure of where this should be described. Maybe a "mac-
address-offset" property in the generic ethernet controller binding?
Regards,
Jan
May be something like eth_address_provider should be introduced in
net/ethernet/ ?
This provider can provide something like eth_provider_address_next() which
will consider "mac-address-num" (or other specific fields).
A patch series proposed the devicetree property
mac-address-increment, but it did not get support at the time:
Please have a look at some recent nvmem patches
https://lkml.org/lkml/2021/9/8/270 that adds support for vendor specific
post-processing of nvmem-cells.
Am hoping that increment usecase (along with other variants) should also
be dealt in similar way.
--srini
of_net: add mac-address-increment support
https://lore.kernel.org/all/20200920095724.8251-4-ansuelsmth@xxxxxxxxx/
dt-bindings: net: Document use of mac-address-increment
https://lore.kernel.org/all/20200920095724.8251-5-ansuelsmth@xxxxxxxxx/
Cheers,
