Re: [PATCH net-next v1 1/1] net: phy: add remote fault support

From: Oleksij Rempel
Date: Thu Jun 16 2022 - 05:35:04 EST


On Wed, Jun 15, 2022 at 10:07:34PM +0200, Andrew Lunn wrote:
> On Tue, Jun 14, 2022 at 10:09:48PM -0700, Jakub Kicinski wrote:
> > On Wed, 15 Jun 2022 05:37:46 +0200 Andrew Lunn wrote:
> > > > Does this dovetail well with ETHTOOL_A_LINKSTATE_EXT_STATE /
> > > > ETHTOOL_A_LINKSTATE_EXT_SUBSTATE ?
> > > >
> > > > That's where people who read extended link state out of FW put it
> > > > (and therefore it's read only now).
> > >
> > > I did wonder about that. But this is to do with autoneg which is part
> > > of ksetting. Firmware hindered MAC drivers also support ksetting
> > > set/get. This patchset is also opening the door to more information
> > > which is passed via autoneg. It can also contain the ID the link peer
> > > PHY, etc. This is all part of 802.3, where as
> > > ETHTOOL_A_LINKSTATE_EXT_STATE tends to be whatever the firmware
> > > offers, not something covered by a standard.
> >
> > I see, yeah, I think you're right.
> >
> > But I'm missing the bigger picture. I'm unclear on who is supposed
> > to be setting the fault user space or kernel / device?
>
> It is also a bit unclear, but at the moment, i think user
> space. However, i can see the kernel making use of maybe RF TEST to
> ask the link peer to go quiet in order to perform a cable test.
>
> Oleksij, what are your use cases?

Currently I was thinking only about diagnostic:
- request transmit pause for cable testing
- request remote loopback for selftest. In this case I will need to use
vendor specific NextPage to request something like this.

> Maybe add something to patch 0/X indicating how you plan to make use of this?

I can move it from first patch if needed.

--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |