[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241029065129.503f51cb@kernel.org>
Date: Tue, 29 Oct 2024 06:51:29 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Jiri Pirko <jiri@...nulli.us>
Cc: netdev@...r.kernel.org, davem@...emloft.net, edumazet@...gle.com,
pabeni@...hat.com, donald.hunter@...il.com, vadim.fedorenko@...ux.dev,
arkadiusz.kubalewski@...el.com, saeedm@...dia.com, leon@...nel.org,
tariqt@...dia.com, maciejm@...dia.com
Subject: Re: [PATCH net-next v3 1/2] dpll: add clock quality level attribute
and op
On Tue, 29 Oct 2024 13:52:12 +0100 Jiri Pirko wrote:
> >I thought clock-id is basically clockid_t, IOW a reference.
> >I wish that the information about timekeepers was exposed
> >by the time subsystem rather than DPLL. Something like clock_getres().
>
> Hmm. From what I understand, the quality of the clock as it is defined
> by the ITU standard is an attribute of the DPLL device. DPLL device
> in our model is basically a board, which might combine oscillator,
> synchronizer and possibly other devices. The clock quality is determined
> by this combination and I understand that the ITU certification is also
> applied to this device.
>
> That's why it makes sense to have the clock quality as the DPLL
> attribute. Makes sense?
Hm, reading more carefully sounds like it's the quality of the holdover
clock. Can we say that in the documentation? "This mainly applies when
the dpll lock-status is not DPLL_LOCK_STATUS_LOCKED" is a bit of a mouthful.
I think I missed the "not" first time reading it.
Is it marked as multi-attr in case non-ITU clock quality is defined
later? Or is it legit to set to ITU bits at once?
Powered by blists - more mailing lists