Re: [chrony-dev] New online/offline command

2018-05-25 Thread Chris Perl
On Fri, May 25, 2018 at 5:28 AM, Miroslav Lichvar wrote: > > Ok, I went with "onoffline". If something better comes up before the > next release, we can change it. I'm not sure if you have constraints with regard to dashes in the name, or the length. But, to me, something

Re: [chrony-dev] Patch: avoid infinite select() loop on chronyc

2017-12-04 Thread Chris Perl
On Mon, Dec 4, 2017 at 9:02 AM, Chris Perl <cp...@janestreet.com> wrote: > > In one of my cases where I saved some strace output, it seems to show > what you suggest: Meaning a never ending string of EINVAL's. What happened to the timeout value, I'm not sure. -- To unsubscribe e

Re: [chrony-dev] Patch: avoid infinite select() loop on chronyc

2017-12-04 Thread Chris Perl
On Mon, Dec 4, 2017 at 5:14 AM, Miroslav Lichvar wrote: > > I suspect the real problem is with the timeout. If it was negative, > select() would return EINVAL and the loop would not terminate. I guess > this could happen if chronyd (or something else) stepped the system >

Re: [chrony-dev] Patch: avoid infinite select() loop on chronyc

2017-12-01 Thread Chris Perl
For what its worth, I've also experienced this issue (chronyc spinning on CPU with every select just returning -1), I just hadn't had the chance to dig in and figure out the cause. On Thu, Nov 30, 2017 at 4:48 PM, Gafton, Cristian wrote: > When chronyc attempts to chat over

Re: [chrony-dev] [PATCH v2 2/2] refclock: Add a new "tai" option

2017-10-11 Thread Chris Perl
On Wed, Oct 11, 2017 at 10:55 AM, Miroslav Lichvar <mlich...@redhat.com> wrote: > > On Tue, Oct 10, 2017 at 01:23:21PM -0400, Chris Perl wrote: > > + if (instance->tai) { > > +tai_offset = REF_GetTaiOffset(sample_time); > > It turned out, this doesn't act

Re: [chrony-dev] [PATCH 2/3] refclock: Include TAI offset when adding samples

2017-10-10 Thread Chris Perl
On Mon, Oct 9, 2017 at 5:34 AM, Miroslav Lichvar wrote: >> --- a/reference.c >> +++ b/reference.c >> @@ -1355,6 +1355,17 @@ int REF_IsLeapSecondClose(void) >> } >> >> /* == */ >> +int REF_GetTaiOffset(void) > > This function

[chrony-dev] [PATCH v2 0/2] Auto TAI-UTC offset for refclocks

2017-10-10 Thread Chris Perl
v2 for implementing a tai refclock option as discussed in [1]. [1] https://listengine.tuxfamily.org/chrony.tuxfamily.org/chrony-users/2017/09/msg00037.html Chris Perl (2): reference: add function to get current tai offset refclock: Add a new "tai" option conf.c

[chrony-dev] [PATCH v2 2/2] refclock: Add a new "tai" option

2017-10-10 Thread Chris Perl
This option is for indicating to chronyd that the reference clock is kept in TAI and that chrony should attempt to convert from TAI to UTC by using the timezone configured by the "leapsectz" directive. --- conf.c | 7 ++- doc/chrony.conf.adoc | 9 + refclock.c

Re: [chrony-dev] [PATCH 2/3] refclock: Include TAI offset when adding samples

2017-10-09 Thread Chris Perl
On Mon, Oct 9, 2017 at 8:41 AM, Miroslav Lichvar wrote: > On second thought, it would better to not accumulate samples when the > TAI offset is unknown, so something like this: > > if (instance->tai) { > tai_offset = REF_GetTaiOffset(sample_time); > if

[chrony-dev] [PATCH 0/3] Auto TAI-UTC offset for refclocks

2017-10-06 Thread Chris Perl
(Apologies if you receive this patch set twice, the first time I sent they didn't appear to make it to the list) This is a first pass at implementing a tai refclock option as discussed in [1]. [1] https://listengine.tuxfamily.org/chrony.tuxfamily.org/chrony-users/2017/09/msg00037.html Chris

[chrony-dev] [PATCH 3/3] doc: Document "tai" refclock option

2017-10-06 Thread Chris Perl
--- doc/chrony.conf.adoc | 9 + 1 file changed, 9 insertions(+) diff --git a/doc/chrony.conf.adoc b/doc/chrony.conf.adoc index 5c85959..55d23e1 100644 --- a/doc/chrony.conf.adoc +++ b/doc/chrony.conf.adoc @@ -552,6 +552,15 @@ Set the minimum number of samples kept for this source. This

[chrony-dev] [PATCH 1/3] refclock: Add a new "tai" option

2017-10-06 Thread Chris Perl
This option is for indicating to chronyd that the reference clock is kept in TAI and that chrony should attempt to convert from TAI to UTC by using the timezone configured by the "leapsectz" directive. --- conf.c | 7 ++- refclock.h | 1 + 2 files changed, 7 insertions(+), 1 deletion(-)

[chrony-dev] [PATCH 2/3] refclock: Include TAI offset when adding samples

2017-10-06 Thread Chris Perl
If the refclock has the "tai" option set, get the current TAI offset and include it when adding the sample. --- refclock.c | 28 ++-- reference.c | 11 +++ reference.h | 3 +++ 3 files changed, 40 insertions(+), 2 deletions(-) diff --git a/refclock.c

[chrony-dev] [PATCH 2/3] refclock: Include TAI offset when adding samples

2017-10-06 Thread Chris Perl
If the refclock has the "tai" option set, get the current TAI offset and include it when adding the sample. --- refclock.c | 28 ++-- reference.c | 11 +++ reference.h | 3 +++ 3 files changed, 40 insertions(+), 2 deletions(-) diff --git a/refclock.c

[chrony-dev] [PATCH 0/3] Auto TAI-UTC offset for refclocks

2017-10-06 Thread Chris Perl
First pass at implementing a tai refclock option as discussed in [1]. [1] https://listengine.tuxfamily.org/chrony.tuxfamily.org/chrony-users/2017/09/msg00037.html Chris Perl (3): refclock: Add a new "tai" option refclock: Include TAI offset when adding samples doc: Document &quo

[chrony-dev] [PATCH 3/3] doc: Document "tai" refclock option

2017-10-06 Thread Chris Perl
--- doc/chrony.conf.adoc | 9 + 1 file changed, 9 insertions(+) diff --git a/doc/chrony.conf.adoc b/doc/chrony.conf.adoc index 5c85959..55d23e1 100644 --- a/doc/chrony.conf.adoc +++ b/doc/chrony.conf.adoc @@ -552,6 +552,15 @@ Set the minimum number of samples kept for this source. This

[chrony-dev] Re:

2017-03-10 Thread Chris Perl
Sorry for the lack of a subject, I was too quick with my `git send-email'. On Fri, Mar 10, 2017 at 8:45 AM, Chris Perl <cp...@janestreet.com> wrote: > > Apologies if you receive this patch twice. I sent it the other day but never > saw it show up on the list via mail-archive.com.