LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Ingo Molnar <mingo@elte.hu>
To: Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Cc: Linux Kernel <linux-kernel@vger.kernel.org>
Subject: Re: [BUG] scheduler: strange behavor with massive interactive processes
Date: Tue, 27 Mar 2007 21:14:20 +0200	[thread overview]
Message-ID: <20070327191420.GB29929@elte.hu> (raw)
In-Reply-To: <87r6rb1nbm.wl%takeuchi_satoru@jp.fujitsu.com>


* Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com> wrote:

> Hi Ingo and all,
> 
> When I was executing massive interactive processes, I found that some 
> of them occupy CPU time and the others hardly run.

yeah.

> I also attach the test program which easily recreates this problem.

thanks, this is really helpful - does the patch below improve the 
situation?

	Ingo

---------------------->
Subject: [patch] sched: improve starvation prevention
From: Ingo Molnar <mingo@elte.hu>

improve starvation prevention by guaranteeing STARVATION_LIMIT. This 
also simplifies the code a bit.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 kernel/sched.c |   12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

Index: linux/kernel/sched.c
===================================================================
--- linux.orig/kernel/sched.c
+++ linux/kernel/sched.c
@@ -235,7 +235,7 @@ struct rq {
 	 */
 	unsigned long nr_uninterruptible;
 
-	unsigned long expired_timestamp;
+	unsigned long switch_timestamp;
 	/* Cached timestamp set by update_cpu_clock() */
 	unsigned long long most_recent_timestamp;
 	struct task_struct *curr, *idle;
@@ -3103,9 +3103,9 @@ static inline int expired_starving(struc
 {
 	if (rq->curr->static_prio > rq->best_expired_prio)
 		return 1;
-	if (!STARVATION_LIMIT || !rq->expired_timestamp)
+	if (!STARVATION_LIMIT)
 		return 0;
-	if (jiffies - rq->expired_timestamp > STARVATION_LIMIT * rq->nr_running)
+	if (jiffies - rq->switch_timestamp > STARVATION_LIMIT)
 		return 1;
 	return 0;
 }
@@ -3218,8 +3218,6 @@ static void task_running_tick(struct rq 
 		p->time_slice = task_timeslice(p);
 		p->first_time_slice = 0;
 
-		if (!rq->expired_timestamp)
-			rq->expired_timestamp = jiffies;
 		if (!TASK_INTERACTIVE(p) || expired_starving(rq)) {
 			enqueue_task(p, rq->expired);
 			if (p->static_prio < rq->best_expired_prio)
@@ -3406,7 +3404,7 @@ need_resched_nonpreemptible:
 		idle_balance(cpu, rq);
 		if (!rq->nr_running) {
 			next = rq->idle;
-			rq->expired_timestamp = 0;
+			rq->switch_timestamp = jiffies;
 			goto switch_tasks;
 		}
 	}
@@ -3420,7 +3418,7 @@ need_resched_nonpreemptible:
 		rq->active = rq->expired;
 		rq->expired = array;
 		array = rq->active;
-		rq->expired_timestamp = 0;
+		rq->switch_timestamp = jiffies;
 		rq->best_expired_prio = MAX_PRIO;
 	}
 

  parent reply	other threads:[~2007-03-27 19:15 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-03-27  1:34 [BUG] scheduler: strange behavor with massive interactive processes Satoru Takeuchi
2007-03-27  5:04 ` Mike Galbraith
2007-03-28  8:38   ` Mike Galbraith
2007-03-28 11:45     ` Ingo Molnar
2007-03-28 11:51       ` Mike Galbraith
2007-03-31 10:15     ` Satoru Takeuchi
2007-03-31 10:29       ` Mike Galbraith
2007-03-27 19:14 ` Ingo Molnar [this message]
2007-03-28  1:16   ` Satoru Takeuchi
2007-03-31  8:16     ` Satoru Takeuchi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20070327191420.GB29929@elte.hu \
    --to=mingo@elte.hu \
    --cc=linux-kernel@vger.kernel.org \
    --cc=takeuchi_satoru@jp.fujitsu.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).