[ntp:questions] Re: What went wrong with the leap second
David L. Mills
mills at udel.edu
Thu Jan 5 04:17:58 UTC 2006
Here is how I tested the leapsecond code. You may notice in ntp-dev/util
a tg.c program. It generates fake IRIG and WWV/H audio signals that can
be piped out of one machine and into another. The code and options are
documented in the source file. When started without a time of day
argument, it uses the time from the system clock, but this can be
changed by an option. An option can be used to insert a leap second in
the audio stream and to light the leap warning; however, the IRIG format
has no provisions for that, so it works only for the WWV/H signal.
To see that the kernel follows the leap, use the WWV/H signal and set
the time option to fifteen minutes before midnight UTC on the last day
of June or December The fifteen minutes gives time for the WWV/H driver
to lock up. To see what happens if the kernel is not equipped for leap
insertion, disable the kernel (disable kernel) from the configuration file.
In all the testing with this thing, I never got unexpected behavior, but
since lots of others did experience pinball machine behavior, some wee
thing must have been overlooked. Workin' on it.
Rob van der Putten wrote:
> Hi there
> Maarten Wiltink wrote:
>>Too right. When do we try again?
> When the hard and software that does work is no longer available.
> Testing with a fake leap second is the only way to make sure that things
> actually do work.
More information about the questions