On Wed, Oct 21, 2020 at 5:28 PM Andrew Lunn <[email protected]> wrote: > > > Actually, I'd also be interested [for this PHY], to report a > > "significance impedance" detection, which is similar to the > > short-detection that is already done. > > You can add that as just another element of the enum. > > > At first, this report would sound like it could be interesting; but > > feel free to disagree with me. > > > > And there's also some "busy" indicator; as-in "unknown activity during > > diagnostics"; to-be-honest, I don't know what this is yet. > > The link partner did not go quiet. You can only do cable tests if the > partner is not sending frames or pulses. You will find most PHYs have > some sort of error status for this. For the Marvell driver, this is > MII_VCT7_RESULTS_INVALID. In that case, the Marvell driver returns > ETHTOOL_A_CABLE_RESULT_CODE_UNSPEC.
Good to know. So, then a quick question: would this patchset [well, a V2 of this] be ok in this form, for the initial cable-test support of this PHY? For other enhancements on the PHY's cable-test [that also require some new netlink attributes, I can do other patches, in the form of "new-netlink-attr, then driver change, and then ethtool update". > > Andrew
