🚀 go-pugleaf

Barry Watzman wrote:

> Masking interrupts -- even forever -- won't slow the clock down in 
> exactly the way that you suggest.  The clock chip keeps time completely 
> independently of the computer (imagine that it was an RS-232 serial 
> external clock -- the only thing that the computer does is read it, and 
> occasionally set it, but the clock runs completely independently of the 
> computer).  Although it's not an external serial clock, that's 
> effectively how it works.
> 
> The way that the computer keeps time is that there are two clocks -- the 
> actual "clock" ("hardware clock", if you prefer) and a "software clock" 
> maintained by the computer (BIOS and/or operating system) only while 
> it's running.  The software clock is what you see, what is used for 
> date/time stamping, what appears in the system tray.  It's updated by a 
> hardware generated interrupt that "ticks" 20 times per second 
> (approximately).  If you disable interrupts, you WILL lose time on the 
> software clock and it will not "catch up" on it's own when interrupts 
> are re-enabled.  HOWEVER, the software clock is reset every time that 
> you reboot the computer, so while it would temporarily lose time, it 
> would be reset the next time that you started the computer.
> 
> If you are losing time over a period of days, assuming that you do 
> reboot the computer from time to time, then it's the hardware clock -- 
> the real, hardware, battery powered, crystal controlled clock -- that's 
> losing time, and this has nothing to do with interrupts being on or off. 
> Turning off interrupts will cause a loss of time, but only until the 
> next reset of the computer.  Interrupts will not cause a permanent loss 
> of time.
> 
> 
Barry,

(1) it's only 18.2 Hz.  I have no idea why that frequency.
(2) some operating systems have a nasty habit of spotting that the two 
are out of step, and "correcting" the CMOS clock.  I don't recall which 
ones.


Andy.


-- 
My real mail is Andy at Champ dash family dot freeserve dot net