Re: [PATCH net 3/3] net: lan743x: Address problems with wake option flags configuration sequences

From: Florian Fainelli
Date: Wed Mar 13 2024 - 16:39:46 EST


On 3/13/24 12:52, Andrew Lunn wrote:
So i think we need to agree on best practices for new drivers and
document what that is, probably in both the ethtool man page and
include/uapi/linux/ethtool.h

* WAKE_MAGIC on its own is O.K.
* WAKE_MAGICSECURE without WAKE_MAGIC is invalid. -EINVAL
* WAKE_MAGIC + WAKE_MAGICSECURE means only secure magic WoL.
* Each driver needs to enforce this, due to backwards compatibility.
Some may decide not to.

Can we agree on this?

Yes, that works for me as well.
--
Florian