In ntp.conf(5), there is a wrong description: allan allan The argument becomes the new value for the minimum Allan intercept, which is a parameter of the PLL/FLL clock dis- cipline algorithm. The value in log2 seconds defaults to 7 (1024 s), which is also the lower limit. Yes. log2(1024) is not 7, but 10. There is also the same description in /usr/share/doc/ntp/miscopt.html, and they have been obsoleted. =46rom ntp-4.2.0 (currently contributed in FreeBSD) and later, this variable "allan" is set a real number (not an exponential number) and defaults to 1500. The default value is defined in src/contrib/ntp/ntpd/ntp_loopfilter.c. The old version (ntp-4.1.2) was: #define CLOCK_ALLAN 10 /* min Allan intercept (log2 s) */ and the recent version (ntp-4.2.0 and later) is: #define CLOCK_ALLAN 1500. /* compromise Allan intercept (s) */ The description has fixed in ntp-4.2.2: allan allan The argument becomes the new value for the Allan intercept, which is a parameter of the PLL/FLL clock discipline algorithm. The value is in seconds with default 1500 s, which is appropriate for most computer clocks. Fix: Here is a patch for CURRENT. How-To-Repeat: man ntp.conf
Responsible Changed From-To: freebsd-doc->roberto roberto maintains NTP. The ntp.conf.5 change may be applied now; however, the change to miscopt.html would take the file from the vendor branch. Assign to Ollivier.
For bugs matching the following criteria: Status: In Progress Changed: (is less than) 2014-06-01 Reset to default assignee and clear in-progress tags. Mail being skipped
Current manual reads the same. I am unable to verify if this is an issue, cc'ing current ntp maintainer. Reporter (18 years ago) said that this was fixed in ntp 4.2.2. Checking our current ntp 4.2.8p18 manual, it still reads exactly the same. Checking bugs.ntp.org, I couldn't find this issue.
(In reply to Alexander Ziaee from comment #3) FreeBSD does not use or display any HTML doc. The HTML docs are in contrib but unused. Probably makes more sense to copy (or merge) the original ntp.conf man page from contrib to the usr.bin copy. I'll look at doing that.
Looking at it before getting to $JOB work, I should be able to point to the mdoc files in contrib instead of using the FreeBSD maintained files. The FreeBSD maintained files are probably out of date now. This is true not only of the man page discussed in this PR but all ntp man pages. As I'm already in the middle of Makefile purgatory working on importing MIT KRB5 to replace our ancient Heimdal KRB5, this Makefile task should be relatively simple. We can use this PR to expand the scope of the todo to include all ntp man pages. This will certainly require a make universe test build.
Created attachment 254598 [details] Untested patch An untested patch. I'll test build on one of the universe machines first, then apply locally. After that it should probably go through a phabricator review.
Created attachment 254599 [details] Updated patch The previous patch wouldn't have built as the path also needed updating. Still untested, still uploading the sources to a universe15 machine.
Review at https://reviews.freebsd.org/D47319.
Comment on attachment 254599 [details] Updated patch >From e7f6237a48cb8f9c9444c75a6ae67f463820c935 Mon Sep 17 00:00:00 2001 >From: Cy Schubert <cy@FreeBSD.org> >Date: Mon, 28 Oct 2024 09:33:30 -0700 >Subject: [PATCH] ntp: Switch to upstream man pages > >The downstream (FreeBSD) ntp man pages are out of date. Switch to those >provided by ntp.org (nwtime.org). > >The FreeBSD maintained man pages will remain in the source tree for a >while until we are comfortable with the switch. > >PR: 104493 >--- > usr.sbin/ntp/doc/Makefile | 30 +++++++++++++++++++++++++++++- > 1 file changed, 29 insertions(+), 1 deletion(-) > >diff --git a/usr.sbin/ntp/doc/Makefile b/usr.sbin/ntp/doc/Makefile >index 2da65eccc1f0..c581a3b673a7 100644 >--- a/usr.sbin/ntp/doc/Makefile >+++ b/usr.sbin/ntp/doc/Makefile >@@ -24,9 +24,37 @@ FILES= access.html accopt.html assoc.html audio.html authentic.html \ > MAN= ntp.conf.5 ntp.keys.5 > MAN+= ntp-keygen.8 ntpd.8 ntpdate.8 ntpdc.8 ntpq.8 ntptime.8 sntp.8 > >+CLEANFILES= ntp.conf.5 ntp.keys.5 >+CLEANFILES+= ntp-keygen.8 ntpd.8 ntpdc.8 ntpq.8 sntp.8 >+ >+ntp.conf.5: ntp.conf.5mdoc >+ cp $< $@ >+ >+ntp.keys.5: ntp.keys.5mdoc >+ cp $< $@ >+ >+ntp-keygen.8: ntp-keygen.1ntp-keygenmdoc >+ cp $< $@ >+ >+ntpd.8: ntpd.1ntpdmdoc >+ cp $< $@ >+ >+ntpdc.8: ntpdc.1ntpdcmdoc >+ cp $< $@ >+ >+ntpq.8: ntpq.1ntpqmdoc >+ cp $< $@ >+ >+sntp.8: sntp.1sntpmdoc >+ cp $< $@ >+ > .PATH: ${SRCTOP}/contrib/ntp/html \ > ${SRCTOP}/contrib/ntp/util \ > ${SRCTOP}/contrib/ntp/ntpd \ >- ${SRCTOP}/contrib/ntp/ntpsnmpd >+ ${SRCTOP}/contrib/ntp/ntpdc \ >+ ${SRCTOP}/contrib/ntp/ntpdq \ >+ ${SRCTOP}/contrib/ntp/ntpsnmpd \ >+ ${SRCTOP}/contrib/ntp/sntp \ >+ ${SRCTOP}/contrib/ntp/util > > .include <bsd.prog.mk> >-- >2.47.0 >
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/src/commit/?id=c607b3258f15d4ea1710bc3cdbab05c9355c2828 commit c607b3258f15d4ea1710bc3cdbab05c9355c2828 Author: Cy Schubert <cy@FreeBSD.org> AuthorDate: 2024-10-28 16:33:30 +0000 Commit: Cy Schubert <cy@FreeBSD.org> CommitDate: 2024-11-12 03:10:04 +0000 ntp: Switch to upstream man pages The downstream (FreeBSD) ntp man pages are out of date. Switch to those provided by ntp.org (nwtime.org). PR: 104493 Differential Revision: https://reviews.freebsd.org/D47319 usr.sbin/ntp/doc/Makefile | 39 +- usr.sbin/ntp/doc/ntp-keygen.8 (gone) | 1223 ------------- usr.sbin/ntp/doc/ntp.conf.5 (gone) | 3277 ---------------------------------- usr.sbin/ntp/doc/ntp.keys.5 (gone) | 174 -- usr.sbin/ntp/doc/ntpd.8 (gone) | 908 ---------- usr.sbin/ntp/doc/ntpdc.8 (gone) | 809 --------- usr.sbin/ntp/doc/ntpq.8 (gone) | 1055 ----------- usr.sbin/ntp/doc/sntp.8 (gone) | 317 ---- 8 files changed, 38 insertions(+), 7764 deletions(-)
Hopefully, now it will be easier to get Kazuhiro's original patch in. Should I send a mail to ntp-questions asking if his description of the Allan algorithm is accurate?
(In reply to Alexander Ziaee from comment #11) Yes, you should, please. The "allan" variable has been reverted to a log2 variable since ntp-4.2.6. The default value is 11. ntp-4.2.8p18/ntpd/ntp_loopfilter.c: > #define CLOCK_ALLAN 11 /* Allan intercept (log2 s) */ The description about the variable in html/miscopt.html has been fixed since ntp-4.2.8p8. ntp-4.2.8p18/ChangeLog: > (4.2.8p8) 2016/06/02 Released by Harlan Stenn <stenn@ntp.org> (snip) > * [Bug 3054] miscopt.html documents the allan intercept in seconds. SWhite. ntp-4.2.8p18/html/miscopt.html: > tinker [allan allan | dispersion dispersion | freq freq | huffpuff > huffpuff | panic panic | step step | stepout stepout] (snip) > allan allan > Specifies the Allan intercept, which is a parameter of the > PLL/FLL clock discipline algorithm, in log2 seconds with > default 11 (2048 s). So ntpd/ntp.conf.def should be fixed on upstream like this: --- ntp-4.2.8p18/ntpd/ntp.conf.def.orig 2024-05-07 20:21:37.000000000 +0900 +++ ntp-4.2.8p18/ntpd/ntp.conf.def 2024-11-13 22:06:08.452433000 +0900 @@ -3135,7 +3135,7 @@ The argument becomes the new value for the minimum Allan intercept, which is a parameter of the PLL/FLL clock discipline algorithm. -The value in log2 seconds defaults to 7 (1024 s), which is also the lower +The value in log2 seconds defaults to 11 (2048 s), which is also the lower limit. .It Cm dispersion Ar dispersion The argument becomes the new value for the dispersion increase rate,
(In reply to WATANABE Kazuhiro from comment #12) I will send an email to my upstream contact.