All of lore.kernel.org
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: LKML <linux-kernel@vger.kernel.org>, RT <linux-rt-users@vger.kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Ingo Molnar <mingo@elte.hu>, Thomas Gleixner <tglx@linutronix.de>,
	Gregory Haskins <ghaskins@novell.com>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>
Subject: [patch 4/8] RT overloaded runqueues accounting
Date: Fri, 19 Oct 2007 14:42:58 -0400	[thread overview]
Message-ID: <20071019184336.669654901@goodmis.org> (raw)
In-Reply-To: 20071019184254.456160632@goodmis.org

[-- Attachment #1: rt-overload.patch --]
[-- Type: text/plain, Size: 1673 bytes --]

This patch adds an RT overload accounting system. When a runqueue has
more than one RT task queued, it is marked as overloaded. That is that it
is a candidate to have RT tasks pulled from it.

Along with the rt_overload counter (number of overloaded runqueues) a
bit mask of overloaded runqueues with RT tasks is set.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/sched_rt.c |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

Index: linux-test.git/kernel/sched_rt.c
===================================================================
--- linux-test.git.orig/kernel/sched_rt.c	2007-10-19 12:34:46.000000000 -0400
+++ linux-test.git/kernel/sched_rt.c	2007-10-19 12:35:58.000000000 -0400
@@ -25,10 +25,23 @@ static void update_curr_rt(struct rq *rq
 	curr->se.exec_start = rq->clock;
 }
 
+#ifdef CONFIG_SMP
+static __cacheline_aligned_in_smp atomic_t rt_overload;
+static cpumask_t rto_cpumask;
+#endif /* CONFIG_SMP*/
+
 static inline void inc_rt_tasks(struct task_struct *p, struct rq *rq)
 {
 	if (rt_task(p))
 		rq->rt_nr_running++;
+#ifdef CONFIG_SMP
+	if (rq->rt_nr_running == 2) {
+		cpu_set(rq->cpu, rto_cpumask);
+		/* Make sure this cpu is visible on rt_overloads */
+		smp_wmb();
+		atomic_inc(&rt_overload);
+	}
+#endif
 }
 
 static inline void dec_rt_tasks(struct task_struct *p, struct rq *rq)
@@ -37,6 +50,12 @@ static inline void dec_rt_tasks(struct t
 		WARN_ON(!rq->rt_nr_running);
 		rq->rt_nr_running--;
 	}
+#ifdef CONFIG_SMP
+	if (rq->rt_nr_running == 1) {
+		atomic_dec(&rt_overload);
+		cpu_clear(rq->cpu, rto_cpumask);
+	}
+#endif
 }
 
 static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup)

-- 

  parent reply	other threads:[~2007-10-19 18:44 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-10-19 18:42 [patch 0/8] New RT Task Balancing Steven Rostedt
2007-10-19 18:42 ` [patch 1/8] Add rt_nr_running accounting Steven Rostedt
2007-10-20 16:45   ` Dmitry Adamushko
2007-10-21  2:13     ` Steven Rostedt
2007-10-19 18:42 ` [patch 2/8] track highest prio queued on runqueue Steven Rostedt
2007-10-19 19:19   ` Steven Rostedt
2007-10-19 19:45   ` Gregory Haskins
2007-10-19 19:57     ` Steven Rostedt
2007-10-20 18:14   ` Dmitry Adamushko
2007-10-21  2:19     ` Steven Rostedt
2007-10-19 18:42 ` [patch 3/8] push RT tasks Steven Rostedt
2007-10-19 18:42 ` Steven Rostedt [this message]
2007-10-19 18:42 ` [patch 5/8] Move prototypes together Steven Rostedt
2007-10-19 18:43 ` [patch 6/8] pull RT tasks Steven Rostedt
2007-10-19 19:24   ` Peter Zijlstra
2007-10-19 19:35     ` Peter Zijlstra
2007-10-19 19:43       ` Steven Rostedt
2007-10-21  9:35   ` Dmitry Adamushko
2007-10-22 13:55     ` Steven Rostedt
2007-10-21 11:59   ` Dmitry Adamushko
2007-10-22 14:05     ` Steven Rostedt
2007-10-22 22:34       ` Dmitry Adamushko
2007-10-23  1:16         ` Steven Rostedt
2007-10-19 18:43 ` [patch 7/8] wake up balance RT Steven Rostedt
2007-10-19 18:43 ` [patch 8/8] disable CFS RT load balancing Steven Rostedt

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=20071019184336.669654901@goodmis.org \
    --to=rostedt@goodmis.org \
    --cc=a.p.zijlstra@chello.nl \
    --cc=akpm@linux-foundation.org \
    --cc=ghaskins@novell.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-rt-users@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=tglx@linutronix.de \
    --cc=torvalds@linux-foundation.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.