Re: [chrony-dev] preparing for 1.25-pre1

[Miroslav Lichvar <mlichvar@xxxxxxxxxx>]

On Sat, Jan 29, 2011 at 10:01:53PM -0600, John Hasler wrote:
> 1) There's a regression which prevents Chrony from building on hppa and
>    ia64 due to CHRONY_IOC_ missing from io_linux.h.  I'll upload a patch
>    soon.

Regression since which version?

What if we started using <linux/ioctl.h> instead of maintaining our
own copy of the ioctl macros for every arch? A configure check would
enable the rtc support only when the system header provides all we
need.

> 2) When chronyd is using IPv6 chronyc mutilates the reference ID into a
>    bogus IPv4 address.  I think that this can be fixed by having chronyc
>    interpret the union in which it receives the ref ID as containing an
>    IPv4 address when the rightmost 96 bits are zero and as IPv6
>    otherwise because no valid IPv6 unicast address can have this form.
>    It's an ugly hack, though.

Is this about the tracking output? Reference ID is always 32 bits and
is equal to the source's IP address only for IPv4, for IPv6 it's the
first 32 bits of MD5 sum of the address. It's printed in quad-dotted
notation, similarly to what ntpd/ntpq does.

But chronyc knows both refid and IP address (and if it's IPv4 or IPv6)
if you want to change the format, no hacks should be necessary.

-- 
Miroslav Lichvar

---
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.