Re: FreeBSD/sparc64 on SunFire V215 - clock (?) problem

[KOT MATPOCKuH]

2009/9/21 Marius Strobl <marius@...>:

>> > Like I said, mainline FreeBSD doesn't support V215 so
>> > far. If you like you can give the following patch a try:
>> > http://people.freebsd.org/~marius/fire.diff
>> > It requires head r197164 or later.
>> I'm tried to apply this patch, also add "device ofw_console" to kernel
>> configuration, but system panic after:
> Please re-fetch fire.diff and try again.
System works like good. But I have a problem with clock in NVRAM (?).
The time in NVRAM's clock is not synced after reboot, and time
difference increases after each system restart.

For example:
14 Oct 18:31:27 ntpdate[963]: step time server 192.168.2.1 offset 343.494018 sec
15 Oct 10:22:53 ntpdate[962]: step time server 192.168.2.1 offset 345.509317 sec

After system boot I have ntpd running and have no problems with clock.

--
MATPOCKuH
_______________________________________________
freebsd-sparc64@... mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-sparc64
To unsubscribe, send any mail to "freebsd-sparc64-unsubscribe@..."

[Miles Nordin]

>>>>> "km" == KOT MATPOCKuH <matpockuh@...> writes:

    km> The time in NVRAM's clock is not synced after reboot,

    km> and time difference increases after each system restart.

    km> 14 Oct 18:31:27 ntpdate[963]: step time server 192.168.2.1
    km> offset 343.494018 sec

        [reboot]

    km> 15 Oct 10:22:53 ntpdate[962]: step time server 192.168.2.1
    km> offset 345.509317 sec

so you are finding two separate problems:

 1. if NVRAM clock were synced after reboot, the step on 15 Oct would
    be close to zero, not 300.

 2. accepting that bug exists, and the NVRAM time is never updated to
    system time, the step should be the same size on the 14th and
    15th.  Instead the nvram is gaining/losing two seconds in only 16
    hours which is much worse accuracy than you expect from the
    battery chip.

[Marius Strobl]

On Thu, Oct 15, 2009 at 10:24:30AM +0400, KOT MATPOCKuH wrote:

> 2009/9/21 Marius Strobl <marius@...>:
>
> >> > Like I said, mainline FreeBSD doesn't support V215 so
> >> > far. If you like you can give the following patch a try:
> >> > http://people.freebsd.org/~marius/fire.diff
> >> > It requires head r197164 or later.
> >> I'm tried to apply this patch, also add "device ofw_console" to kernel
> >> configuration, but system panic after:
> > Please re-fetch fire.diff and try again.
> System works like good. But I have a problem with clock in NVRAM (?).
> The time in NVRAM's clock is not synced after reboot, and time
> difference increases after each system restart.
>
> For example:
> 14 Oct 18:31:27 ntpdate[963]: step time server 192.168.2.1 offset 343.494018 sec
> 15 Oct 10:22:53 ntpdate[962]: step time server 192.168.2.1 offset 345.509317 sec
>
> After system boot I have ntpd running and have no problems with clock.
>

This issue isn't limited to V215 but also happens on some other
machines equipped with LOMs, though I haven't noticed their
clocks to increasingly differ between reboots so far. I think
what's going on is that the RTC is fed with the time from the
LOM during boot, so besides setting the LOM clock correctly
the only thing you can do until FreeBSD gains support for
talking to the various LOMs is to run ntpdate(8) during boot
(you can use ntpdate_enable together with ntpd_enable).

Marius

_______________________________________________
freebsd-sparc64@... mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-sparc64
To unsubscribe, send any mail to "freebsd-sparc64-unsubscribe@..."