Re: Why HZ on i386 is 100 ?

[Kent Borg]

On Wed, Apr 17, 2002 at 08:37:43AM +1000, Herbert Xu wrote:
> Why are we still measuring uptime using the tick variable? Ticks != time.
> Surely we should be recording the boot time somewhere (probably on a
> file system), and then comparing that with the current time?

It depends on the meaning of "is", er, opps, I mean: it depends on the
meaning of "uptime".

The notebook I am typing on at this moment was last booted just about
exactly 8 days ago (judging from the timestamp on /var/log/dmesg) but
in a cat-like way it spends a lot of its time asleep and so top
reports an uptime of only "4 days, 2:42".

Which is correct?  I suggest that the smaller number is closer to
correct because that is roughly the amount of time the system has
actually spent running.

-kb, the Kent who expects this question to get more complicated as the
new suspend gets more and more clever and if the kernel ever starts
seriously catnapping on its own.