[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ntp-bugs
Subject:    [ntp:bugs] [Bug 2250] Leap warning bit in reply packets to clients is cleared shortly before the lea
From:       bugzilla-daemon () ntp ! org
Date:       2013-09-27 18:10:22
Message-ID: bug-2250-35-DYNNdpV012 () http ! bugs ! ntp ! org/
[Download RAW message or body]

https://bugs.ntp.org/show_bug.cgi?id=2250

--- Comment #9 from Juergen Perlinger <juergen.perlinger@t-online.de> 2013-09-27 18:10:22 UTC ---
(In reply to comment #8)
Hello Martin,

I spent a few days of my holiday to create something that does 'The Right
Thing'(tm). I carved the most of the leap second handling into a separate
module (with a unit test for it) and revamped the rest of the source code to
use this module. The special goodies:

- It's 2036 proof -- the current code isn't.
- Supports systems that do the leap transition under the hood ('electric mode')
  and crude do-it-yourself stepping.
- Includes a dead-band suppression for the 1st hour of a month.
- Permits more than 1 pending leap second ahead.

Unhappily, this involves changes in critical locations: ntp_proto.c and
ntp_crypto.c. (ntp_timer.c and npt_util.c are also affected, but that's nothing
DLM keeps Argos' eyes on. AFAIK, at least.)

We have two options here: Create a dumb solution that simply keeps the leap
indication for one more second, which also needs to modify ntp_proto.c a little
bit, or go the whole way. With a good code review on the modifications we
should be reasonable sure that this works. But I have to change the windows
clock stuff, too, to make this working -- this is next on my TODO list.

Harlan, what do you think? I'll probably have the whole thing working after
this weekend, but as I said, I want a proper review before even considering a
commit.

-- 
Configure bugmail: https://bugs.ntp.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.
_______________________________________________
bugs mailing list
bugs@lists.ntp.org
http://lists.ntp.org/listinfo/bugs
[prev in list] [next in list] [prev in thread] [next in thread]