Since them the code improved by a factor of roughly 8, and during my
short test yesterday evening I can offer:
stability 0.010PPM, offset 50ns, estimated error 2µs, all with a
Pentium-100 and a GPS running Linux-2.2.12.
I really wanted to say:
The current snapshot found at linux.kernel.org in
/pub/daemons/ntp/PPS offers these new features (also known as flame
bait):
I added a small experimental sysctl() interface to monitor kernel
time variables (kind of experimental).
I added a new variable "rtc_update" that allows to turn off periodic
updates of the RTC as well as setting the interval between updates.
The new logic allows to request updating the RTC once, and it is done
when setting the system time by default. Isn't that cool?
As the timezone is also available via sysctl() I can very well
imagine that the code could be extended to allow updating the RTC for
local time as well as UTC, all settable via sysctl(). I think this
sounds promising.
Other unimportant features (since pre1) are minor bug fixes (the
scope of time_tick) and new bugs (at least I can't proove that there
are none, so I better tell you ;-)
The terrible file name is PPSkit-0.8.1-0.9.0pre2.diff.gz, because it
is based on PPSkit-0.8.1.gz.
Please tell me what you think (I'm not subscribed here).
Regards,
Ulrich
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/