[PATCH net-next 3/3] net: ethtool: tsconfig: Re-configure hwtstamp upon provider change

[Kory Maincent]

On Thu, 16 Oct 2025 10:01:53 +0200
Maxime Chevallier <maxime.chevallier at bootlin.com> wrote:

> Hi Köry,
> 
> On 15/10/2025 14:45, Kory Maincent wrote:
> > On Wed, 15 Oct 2025 12:27:23 +0200
> > Maxime Chevallier <maxime.chevallier at bootlin.com> wrote:
> >   
> >> When a hwprov timestamping source is changed, but without updating the
> >> timestamping parameters, we may want to reconfigure the timestamping
> >> source to enable the new provider.
> >>
> >> This is especially important if the same HW unit implements 2 providers,
> >> a precise and an approx one. In this case, we need to make sure we call
> >> the hwtstamp_set operation for the newly selected provider.  
> > 
> > This is a design choice.
> > Do we want to preserve the hwtstamp config if only the hwtstamp source is
> > changed from ethtool?
> > If we want to configure the new source to the old source config we will also
> > need to remove this condition:
> > https://elixir.bootlin.com/linux/v6.17.1/source/net/ethtool/tsconfig.c#L339
> >  
> 
> What I get from the ethtool output is that the ts config is per-source.
> Re-applying the old config to the new source may not work if the new one
> doesn't have the same capabilities.
> 
> > 
> > I do not really have a strong opinion on this, let's discuss which behavior
> > we prefer.  
> 
> Well if we want to support different timestamp providers provided by the same
> HW block (same MAC or even same PHY), then we need a way to notify the
> provider when the timestamp provider gets selected and unselected.
> 
> Otherwise there's no way for the provider to know it has been re-enabled,
> unless we perform a config change at the same time.

Oh right, indeed, we need a call to ndo_hwtstamp_set to tell the provider to
change the qualifier configured. I missed that.
This could even be a fix but as it is used nowhere it won't fix anything.

Acked-by: Kory Maincent <kory.maincent at bootlin.com>

Thank you!

Regards,
-- 
Köry Maincent, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com