[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <1231912965.956.11.camel@marge.simson.net>
Date: Wed, 14 Jan 2009 07:02:45 +0100
From: Mike Galbraith <efault@....de>
To: Ingo Molnar <mingo@...e.hu>
Cc: Brian Rogers <brian@...w.org>, linux-kernel@...r.kernel.org,
Peter Zijlstra <a.p.zijlstra@...llo.nl>
Subject: Re: [BUG] How to get real-time priority using idle priority
On Wed, 2009-01-14 at 06:31 +0100, Ingo Molnar wrote:
> What about increasing the weight of SCHED_IDLE tasks from 1 to 2 or 3?
> That still makes them mega-nice (which is more than enough), but should
> make the math a lot less borderline.
The weight is currently 2, inv_weight is 1 << 31, which is correct.
I implemented the advance min_vruntime at nice 0 thing, and that seems
to work fine. SCHED_IDLE tasks slam right as they execute, but check
out how _far_ right they slam. This was with the build pinned to cpu 3,
so no migration issues. I'll try weight of 3.
cfs_rq[3]:
.exec_clock : 0.000000
.MIN_vruntime : 281842722.745713
.min_vruntime : 370490.315434
.max_vruntime : 281843966.887081
.spread : 1244.141368
.spread0 : 85467.080507
.nr_running : 6
.load : 12
.nr_spread_over : 0
rt_rq[3]:
.rt_nr_running : 0
.rt_throttled : 0
.rt_time : 0.000000
.rt_runtime : 950.000000
runnable tasks:
task PID tree-key switches prio exec-runtime sum-exec sum-sleep
----------------------------------------------------------------------------------------------------------
make 29977 281843914.902150 171 120 0 0 0.000000 0.000000 0.000000
make 30530 281842722.745713 6 120 0 0 0.000000 0.000000 0.000000
gcc 30584 281843921.282161 1 120 0 0 0.000000 0.000000 0.000000
sh 30590 281843966.887081 2 120 0 0 0.000000 0.000000 0.000000
gcc 30591 281843882.274316 1 120 0 0 0.000000 0.000000 0.000000
R cc1 30592 372734.946902 5 120 0 0 0.000000 0.000000 0.000000
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists