time doesn't work?

Pyun YongHyeon pyunyh at gmail.com
Thu Feb 25 20:23:51 UTC 2010


On Thu, Feb 25, 2010 at 08:39:01PM +0100, Attilio Rao wrote:
> 2010/2/25 Pyun YongHyeon <pyunyh at gmail.com>:
> > On Thu, Feb 25, 2010 at 04:48:28PM +0100, Attilio Rao wrote:
> >> 2010/2/17 Pyun YongHyeon <pyunyh at gmail.com>:
> >> > On Wed, Feb 17, 2010 at 01:12:53PM -0800, Xin LI wrote:
> >> >> Hi,
> >> >>
> >> >> On Sun, Feb 14, 2010 at 10:20 AM, Gavin Atkinson
> >> >> <gavinFreeBSD.org at ury.york.ac.uk> wrote:
> >> >> >> /mnt: write failed, filesystem is full
> >> >> >> gzip: write: No space left on device
> >> >> >> gzip: output file: randomfile.gz wrong size (1673592832 != -1), deleting
> >> >> >> gzip: leaving original randomfile
> >> >> >> 0.000u 85.063s 1:25.10 99.9% ??0+0k 12440+12839io 7pf+0w
> >> >> >
> >> >> > Does reverting r202387, 202441 and 202534 make any difference?
> >> >>
> >> >> Yes, reverting these revisions makes everything back to normal
> >> >> (including top -P).
> >> >>
> >> >
> >> > I'm not sure this is also related with breakage of
> >> > "systat -vmstat 1" on amd64 CURRENT. systat(1) shows "The alternate
> >> > system clock has died! ??Reverting to ``pigs'' display." message and
> >> > does not work as expected.
> >> > When I run systat(1) on sparc64 CURRENT it worked as expected so I vaguely
> >> > guess it's related with attilio's change.(CCed)
> >>
> >> [CC'ing also others that may got this issue]
> >>
> >> Can you please upgrade to the newest CURRENT, try the attached patch
> >> and report if system choiches the right timer alone:
> >> http://www.freebsd.org/~attilio/Sandvine/STABLE_8/statclock_aliasing/statclock-fixup.diff
> >>
> >> with 'alone', I mean that you might strip from config all the helping
> >> tips (machdep.lapic_allclocks).
> >>
> >
> > With the patch above, systat(1) seems to work as expected.
> > One thing I see is dmesg output is
> >
> > RTC BIOS diagnostic error 11<memory_size>
> >
> > I'm not sure whether I had this message in previous kernel.
> 
> It is likely you had after the incriminating revision but not because
> the previous patch broke it, just because the previous patch does
> enable atrtc (while the old kernel didn't do this).
> 

Thank you for clarifying this.


More information about the freebsd-current mailing list