Re: [RFC][PATCH] new timeofday core subsystem (v.A0)

[Andi Kleen]

> I really think a tickless system, for other than UML systems, is a loosing 
> thing.  The accounting overhead on context switch (which increases as the 

You can run the accounting independently at much lower frequency (10ms is
perfectly fine as 2.4 has proven - i suspect even lower would be ok too) 
IMHO it should be a sysctl so that users can do a trade off between
power consumption and accounting accuracy.

And there is one important special that doesn't need any accounting
for user space at all: the idle loop. 
It still needs some way to account interrupts, but that could be done
in do_IRQ or also do with rather low frequency. 

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