Re: [PATCH v5 net-next 1/5] net/ethtool: add netlink interface for the PLCA RS

[Piergiorgio Beruto]

On Wed, Dec 07, 2022 at 03:16:00PM +0100, Andrew Lunn wrote:
> > > TBH I can't parse the "ETHTOOL_A_PLCA_VERSION is reported as 0Axx
> > > where.." sentence. Specifically I'm confused about what the 0A is.
> > How about this: "When this standard is supported, the upper byte of
> > ``ETHTOOL_A_PLCA_VERSION`` shall be 0x0A (see Table A.1.0 — IDVER 
> > bits assignment).
> 
> I think the 0x0A is pointless and should not be included here. If the
> register does not contain 0x0A, the device does not follow the open
> alliance standard, and hence the lower part of the register is
> meaningless.
> 
> This is why i suggested -ENODEV should actually be returned on invalid
> values in this register.
I already integrated this change in v5 (returning -ENODEV). Give what you're
saying, I can just remove that sentence from the documentations. Agreed?

> 
> > > >   * struct ethtool_phy_ops - Optional PHY device options
> > > >   * @get_sset_count: Get number of strings that @get_strings will write.
> > > >   * @get_strings: Return a set of strings that describe the requested objects
> > > >   * @get_stats: Return extended statistics about the PHY device.
> > > > + * @get_plca_cfg: Return PLCA configuration.
> > > > + * @set_plca_cfg: Set PLCA configuration.
> > > 
> > > missing get status in kdoc
> > Fixed. Good catch.
> 
> Building with W=1 C=2 will tell you about kerneldoc issues. Ideally we
> want all network code to be clean with these two options.
Ok thanks. I probably need to upgrade my machine to achieve this. Will
do.