Mailinglist Archive: opensuse (4547 mails)

< Previous Next >
Re: [SLE] realtime clock probs-solved ...
  • From: "Carlos E. R." <robin1.listas@xxxxxxxxxx>
  • Date: Wed, 5 May 2004 02:45:57 +0200 (CEST)
  • Message-id: <Pine.LNX.4.53.0405050201380.13860@xxxxxxxxxxxxxxxx>

The Tuesday 2004-05-04 at 22:43 +0200, Anders Johansson wrote:

> > No, it is not accessed at all, for many reasons. One, it is a very slow
> > access. Two, the CPU maintains its own clock, and doesn't need it.
> /usr/src/linux/arch/i386/kernel/time.c
> do_timer_interrupt()
> It is accessed, around once every 11 minutes it seems, though I'm not entirely
> certain under what circumstances STA_UNSYNC is set, I didn't trace it that
> deeply.

Huh? Ah, Well; it can be periodically updated, on certain circumstances:

* If we have an externally synchronized Linux clock, then update
* CMOS clock accordingly every ~11 minutes. Set_rtc_mmss() has to be
* called as close as possible to 500 ms before the new second starts.

That may be of use when the kernel got synchronized by ntpd, for example.

> In any case, /dev/rtc has many uses in a running system, even if not by the
> kernel itself

Of course it has :-)

> > Read my howto, it is documented there, and my sources for so saying.
> I only read one source.

I never read only one source for any writeup :-)

> > And some (or all) of those chips are also powered from the board 5 Vcc
> > while powered up.
> OK, that is interesting. Then maybe this debate is entirely pointless :)

Carlos Robinson

< Previous Next >