Re: [Ntp] NTPv5: big picture

[Magnus Danielson <[e-mail]>]

Philip,

On 2021-01-04 18:40, Philip Prindeville wrote:
>
>> On Jan 4, 2021, at 8:18 AM, Miroslav Lichvar <[e-mail]> wrote:
>>
>> On Thu, Dec 31, 2020 at 06:54:40PM -0800, Hal Murray wrote:
>>> Do we have a unifying theme?  Can you describe why we are working on NTPv5 in 
>>> one sentence?
>> There is a list of issues in NTPv4 I would like to see fixed in NTPv5:
>> https://trac.ietf.org/trac/ntp/wiki/NtpVersionFourIssues
>>
>> A major issue is that NTPv4 doesn't support short extension fields due
>> to conflicts with legacy MACs, so fixing all those issues by adding
>> new extending fields to NTPv4 seems impractical. Some things, e.g. the
>> timescale selection, makes more sense to have in the header.
>>
>>> Part of the motivation for this is to enable and encourage OSes to convert to 
>>> non-leaping time in the kernels.  Are there any subtle details in this area 
>>> that we should be aware of?  Who should we coordinate with?  ...
>> I don't think that should be the job of the NTP WG. The kernels will
>> need to support a leaping UTC timescale for as long as it is needed
>> for civil time.
>
> I disagree.  The kernel doesn’t inherently require UTC.  It could just as easily use TAI or some other format.
>
> The kernel needs to provide into user-space some format which is then convertible to UTC, for as long as UTC is in-use by applications.
>
> This could be handled by libc.

While I agree that kernel very well could operate in some suitable TAI
format, and the PTP scale would be the low-hanging fruit for that, as it
intends to be that blend of TAI and classical UNIX/POSIX time_t.

However, the only way I know with spreading in which TAI and UTC is
maintained is through the kernel interface provided by the NTP
nanokernel, and that is supported on effectively most Linuxes.

What can be done better is to use it properly and that libc would
provide the needed glue. If libc does that, then from the user's point,
if this is done completely in libc or shared between kernel and libc the
user do not care, as long as they get correct time-stamp and the machine
is setup to feed the needed data.

But then again, I do not think it's the NTP WG job do decide where it is
implemented, rather we shall make sure the NTP side of things can
support the needed features so that it works well.

Cheers,
Magnus