Re: Scheduler fixes
Chris Wedgwood (chris@cybernet.co.nz)
Thu, 25 Jun 1998 11:32:07 +1200
On Thu, Jun 25, 1998 at 12:15:05AM +0200, MOLNAR Ingo wrote:
>
> btw, we could do _exact_ process accounting if we want to, at the price of
> ~10-15 cycles per schedule(). we can just read the cycle counter (industry
> standard on most leading CPUs ;), add the delta to the process and do the
> math later, whenever someone (top) tries to access the values. This way we
> could even differentiate between IRQ cycles, kernel cycles, kernel thread
> cycles, idle cycles and user-space cycles. Is this an important and
> fundamental enough feature to justify those 10-15 cycles?
Two ways to look at it :-
- for most people, it not going to be useful most of the time
and
- its only 10-15 cycles, not much compared to a context switch.
I guess CONFIG_EXACT_PROCTIMES would make everyone happy (having this
dependent upon the processor selection and perhaps APM).
The values needn't even be available using getrusage(2), unless perhaps
someone adds a flag RUSAGE_EXACT for 'who'.
-Chris
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu