TJA1100 100Base-T1 PHY features via ethtool?

TJA1100 100Base-T1 PHY features via ethtool?

[Michael Grzeschik]

[-- Attachment #1: Type: text/plain, Size: 1213 bytes --]

Hi David,

I use a special 100Base-T1 phy (NXP TJA1100 [1]) that has some features
like:

- enabling/disabling test modes
- fault detection
- switching managed/autonomous mode
- signal quality indication
- ...

I already implemented the support of the features with the
ethtool --get/set-phy-tunables features by adding ethtool_phy_tunables:

ETHTOOL_PHY_TEST_MODE
ETHTOOL_PHY_FAULT_DETECTION
ETHTOOL_PHY_MANAGED_MODE
ETHTOOL_PHY_SIGNAL_QUALITY

Before posting my series I wanted to ensure that this is the preferred
interface for those options.

I found a series from 2016 [2] that implements the userspace part for
the loopback feature of some phys, that did not get mainline so far
which makes me wonder if ethtool is still the way to go.

[1] https://www.nxp.com/docs/en/data-sheet/TJA1100.pdf
[2] https://www.spinics.net/lists/netdev/msg406614.html

Thanks,
Michael

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: TJA1100 100Base-T1 PHY features via ethtool?

[Florian Fainelli]

On 08/13/2018 12:35 PM, Michael Grzeschik wrote:
> Hi David,
> 
> I use a special 100Base-T1 phy (NXP TJA1100 [1]) that has some features
> like:
> 
> - enabling/disabling test modes
> - fault detection
> - switching managed/autonomous mode
> - signal quality indication
> - ...
> 
> I already implemented the support of the features with the
> ethtool --get/set-phy-tunables features by adding ethtool_phy_tunables:
> 
> ETHTOOL_PHY_TEST_MODE
> ETHTOOL_PHY_FAULT_DETECTION
> ETHTOOL_PHY_MANAGED_MODE
> ETHTOOL_PHY_SIGNAL_QUALITY
> 
> Before posting my series I wanted to ensure that this is the preferred
> interface for those options.

The tunable interface is there, but is very limited. A few months ago, I
had started proposing an interface to support PHY test modes [1] (the
standard IEEE 802.3 defined ones) but a lot of it should now be migrated
to the work that Michal is doing on the conversion of ethtool to netlink
[2].

[1]: https://lkml.org/lkml/2018/4/27/1172
[2]: https://www.spinics.net/lists/netdev/msg516233.html

> 
> I found a series from 2016 [2] that implements the userspace part for
> the loopback feature of some phys, that did not get mainline so far
> which makes me wonder if ethtool is still the way to go.

ethtool is being converted to netlink, and that will be a much more
flexible interface to work with since it is basically easily extensible
(unlike the current ethtool + ioctl approach).

Back when the patches were proposed, we just had mild disagreement on
the loopback terminology being used, and then nothing happened.

> 
> [1] https://www.nxp.com/docs/en/data-sheet/TJA1100.pdf
> [2] https://www.spinics.net/lists/netdev/msg406614.html
> 
> Thanks,
> Michael
> 


-- 
Florian

Re: TJA1100 100Base-T1 PHY features via ethtool?

[Michael Grzeschik]

[-- Attachment #1: Type: text/plain, Size: 2467 bytes --]

On Mon, Aug 13, 2018 at 12:53:35PM -0700, Florian Fainelli wrote:
> On 08/13/2018 12:35 PM, Michael Grzeschik wrote:
> > Hi David,
> > 
> > I use a special 100Base-T1 phy (NXP TJA1100 [1]) that has some features
> > like:
> > 
> > - enabling/disabling test modes
> > - fault detection
> > - switching managed/autonomous mode
> > - signal quality indication
> > - ...
> > 
> > I already implemented the support of the features with the
> > ethtool --get/set-phy-tunables features by adding ethtool_phy_tunables:
> > 
> > ETHTOOL_PHY_TEST_MODE
> > ETHTOOL_PHY_FAULT_DETECTION
> > ETHTOOL_PHY_MANAGED_MODE
> > ETHTOOL_PHY_SIGNAL_QUALITY
> > 
> > Before posting my series I wanted to ensure that this is the preferred
> > interface for those options.
> 
> The tunable interface is there, but is very limited. A few months ago, I
> had started proposing an interface to support PHY test modes [1] (the
> standard IEEE 802.3 defined ones) but a lot of it should now be migrated
> to the work that Michal is doing on the conversion of ethtool to netlink
> [2].
> 
> [1]: https://lkml.org/lkml/2018/4/27/1172
> [2]: https://www.spinics.net/lists/netdev/msg516233.html

The ethtool userspace tool is somehow odd to program on,
so switching to netlink is definitively a great idea!

> > 
> > I found a series from 2016 [2] that implements the userspace part for
> > the loopback feature of some phys, that did not get mainline so far
> > which makes me wonder if ethtool is still the way to go.
> 
> ethtool is being converted to netlink, and that will be a much more
> flexible interface to work with since it is basically easily extensible
> (unlike the current ethtool + ioctl approach).

Yes, netlink sounds absolutely more useful here.

> Back when the patches were proposed, we just had mild disagreement on
> the loopback terminology being used, and then nothing happened.

Right, thanks for clarification!

> 
> > 
> > [1] https://www.nxp.com/docs/en/data-sheet/TJA1100.pdf
> > [2] https://www.spinics.net/lists/netdev/msg406614.html
> > 
> > Thanks,
> > Michael
> > 
> 
> 
> -- 
> Florian
> 

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

RE: TJA1100 100Base-T1 PHY features via ethtool?

[Woojung.Huh]

Hi Florian & Michael,

> > ethtool is being converted to netlink, and that will be a much more
> > flexible interface to work with since it is basically easily extensible
> > (unlike the current ethtool + ioctl approach).
> 
> Yes, netlink sounds absolutely more useful here.
Is ethtool + netlink expected to be merged in net-next soon?
Couldn't find anything on the web except some experimental information.

Thanks.
Woojung