lkml.org 
[lkml]   [1999]   [Jan]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Scheduler problems
Hi,

On Tue, 19 Jan 1999 20:43:12 +0100 (CET), MOLNAR Ingo
<mingo@chiara.csoma.elte.hu> said:

> On Tue, 19 Jan 1999, Stephen C. Tweedie wrote:

>> Ingo, any thoughts? Seems that if for any reason all wakeup interrupts
>> arrive at the same CPU which is running the SCHED_FIFO task, then we
>> will _never_ perform any successful wakeups without first exhausting the
>> other CPU's current timeslice.

> yes i have a patch for this but it's not straightforward enough for 2.2.
> (ie. a 50k patch) It has to do things like a full goodness() run for every
> CPU in wakeup(), which is less than ideal.

Ingo, before I read this I was thinking about it a little and yes, this
was one of the issues which worried me. However, is there any reason
why we can't just use the last known goodness of the process currently
running on each CPU, rather than recalculating the current goodness?

Goodness can only decrease over time as the quantum expires, so the only
danger would be that we miss a reschedule event, not that we reschedule
too often. That's certainly better than never getting a reschedule
right at all. Given that the default scheduler keeps the goodness of
interactive processes higher at all times than the goodness of any
CPU-bound processes, this behaviour would probably work pretty well
without requiring a per-CPU goodness recalculate for every wakeup.

--Stephen

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

\
 
 \ /
  Last update: 2005-03-22 13:49    [W:0.062 / U:0.292 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site