Re: [PATCH] 1/6: 64 bit jiffies

Tim Schmielau (tim@physik3.uni-rostock.de)
Sat, 11 May 2002 20:30:59 +0200 (CEST)


On Sat, 11 May 2002, Denis Vlasenko wrote:

> On 11 May 2002 08:25, Tim Schmielau wrote:
> > +static inline void init_jiffieswrap_timer(void)
> > +{
> > + init_timer(&jiffieswrap_timer);
> > + jiffieswrap_timer.expires = jiffies + CHECK_JIFFIESWRAP_INTERVAL;
> > + jiffieswrap_timer.function = check_jiffieswrap;
> > + add_timer(&jiffieswrap_timer);
> > +}
>
> I'm ignorant on the issue... does active timer mandate check for
> expiration at every timer tick?

No, timers are implemented in a highly efficient manner. The above
timer will just add O(1) cost to 4 table refills, meaning some 100 cycles
per quarter of a year.

> If yes, it is somewhat silly to use timer:
> such check would be more costly than
>
> if(!++jiffies) jiffies_hi++;
>
> (or similar) construct in timer int.
>
> BTW, I always liked above thing more that any other 64 jiffy solution.
> What's wrong with it?

It's slower than

jiffies_64++;

which went into Linus' tree yesterday :-)

Tim

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/