[ntp:questions] NTP & PPS, part 2 ;)

Martin Burnicki martin.burnicki at meinberg.de
Fri Dec 12 08:52:53 UTC 2014


Harlan Stenn wrote:
> William Unruh writes:
>
>> It is one of those pissing matches. Ntp thinks the os should provide
>> timepps.h, while the others think ntpd should provide it if they need
>> it. As a result the users get screwed while the self-righteous can
>> continute in glory. Just like with cdrecord.
>
> Bullshit.
>
> It's an OS-specific file that should be provided by the OS if the
> underlying API exists.
>
> We have a *sample* implementation that folks can see and perhaps use,
> but implementors can and frequently do make changes to that.
>
> We have a header file for the NANO time stuff, and the linux kernel
> folks decided to change those - that caused problem for years.
>
> If you disgree and think NTP should provide the file all the time, then:
>
> - how do you propose we find out if the underlying API is really
> provided in the currently-running kernel?
>
> - what do we do in the situation where the OS provides the localized
> (ie, correct) version of the header?  Do you *really* want to trust
> #include "..." v. #include <...>, and make sure that this is something
> that everybody who touches that part of the code understands, too?

IMO the best approach would be to detect this at runtime.

Martin



More information about the questions mailing list