mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Ingo Molnar <mingo@elte.hu>
To: Nick Piggin <piggin@cyberone.com.au>
Cc: Andrew Morton <akpm@osdl.org>,
	linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 3/5] 2.6.0 sched migrate comment
Date: Sat, 20 Dec 2003 21:26:11 +0100	[thread overview]
Message-ID: <20031220202611.GB32320@elte.hu> (raw)
In-Reply-To: <3FE468F5.7020501@cyberone.com.au>

[-- Attachment #1: Type: text/plain, Size: 578 bytes --]


* Nick Piggin <piggin@cyberone.com.au> wrote:

> Some comments were lost during minor surgery here...

this patch collides with John Hawkes' sched_clock() fix-patch which goes
in first. Also, the patch does more than just comment fixups. It changes
the order of tests, where a bug slipped in:

+       if (!idle && (delta <= cache_decay_ticks))

This will cause the cache-hot test to almost never trigger, which is
clearly incorrect. The correct test is:

        if (!idle && (delta <= JIFFIES_TO_NS(cache_decay_ticks)))

anyway, i've fixed it all up (patch attached).

	Ingo

[-- Attachment #2: sched-can-migrate-2.6.0-A1 --]
[-- Type: text/plain, Size: 1648 bytes --]

--- linux/kernel/sched.c.orig	
+++ linux/kernel/sched.c	
@@ -1193,25 +1193,25 @@ static inline void pull_task(runqueue_t 
 }
 
 /*
- * Previously:
- *
- * #define CAN_MIGRATE_TASK(p,rq,this_cpu)	\
- *	((!idle || (NS_TO_JIFFIES(now - (p)->timestamp) > \
- *		cache_decay_ticks)) && !task_running(rq, p) && \
- *			cpu_isset(this_cpu, (p)->cpus_allowed))
+ * can_migrate_task - may task p from runqueue rq be migrated to this_cpu?
  */
-
 static inline int
 can_migrate_task(task_t *tsk, runqueue_t *rq, int this_cpu, int idle)
 {
 	unsigned long delta = rq->timestamp_last_tick - tsk->timestamp;
 
-	if (!idle && (delta <= JIFFIES_TO_NS(cache_decay_ticks)))
-		return 0;
+	/*
+	 * We do not migrate tasks that are:
+	 * 1) running (obviously), or
+	 * 2) cannot be migrated to this CPU due to cpus_allowed, or
+	 * 3) are cache-hot on their current CPU.
+	 */
 	if (task_running(rq, tsk))
 		return 0;
 	if (!cpu_isset(this_cpu, tsk->cpus_allowed))
 		return 0;
+	if (!idle && (delta <= JIFFIES_TO_NS(cache_decay_ticks)))
+		return 0;
 	return 1;
 }
 
@@ -1273,13 +1273,6 @@ skip_bitmap:
 skip_queue:
 	tmp = list_entry(curr, task_t, run_list);
 
-	/*
-	 * We do not migrate tasks that are:
-	 * 1) running (obviously), or
-	 * 2) cannot be migrated to this CPU due to cpus_allowed, or
-	 * 3) are cache-hot on their current CPU.
-	 */
-
 	curr = curr->prev;
 
 	if (!can_migrate_task(tmp, busiest, this_cpu, idle)) {
@@ -1289,6 +1282,8 @@ skip_queue:
 		goto skip_bitmap;
 	}
 	pull_task(busiest, array, tmp, this_rq, this_cpu);
+
+	/* Only migrate one task if we are idle */
 	if (!idle && --imbalance) {
 		if (curr != head)
 			goto skip_queue;

  parent reply	other threads:[~2003-12-20 20:27 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-12-20 15:19 [PATCH 1/5] 2.6.0 fix preempt ctx switch accounting Nick Piggin
2003-12-20 15:20 ` [PATCH 2/5] 2.6.0 sched fork cleanup Nick Piggin
2003-12-20 15:21   ` [PATCH 3/5] 2.6.0 sched migrate comment Nick Piggin
2003-12-20 15:22     ` [PATCH 4/5] 2.6.0 sched style fixes Nick Piggin
2003-12-20 15:24       ` [PATCH 5/5] 2.6.0 sched affinity race Nick Piggin
2003-12-20 21:07       ` [PATCH 4/5] 2.6.0 sched style fixes Ingo Molnar
2003-12-20 20:26     ` Ingo Molnar [this message]
2003-12-20 23:19       ` [PATCH 3/5] 2.6.0 sched migrate comment Nick Piggin
2003-12-20 19:55   ` [PATCH 2/5] 2.6.0 sched fork cleanup Ingo Molnar
2003-12-20 23:17     ` Nick Piggin
2003-12-20 19:22 ` [PATCH 1/5] 2.6.0 fix preempt ctx switch accounting Ingo Molnar
2003-12-20 19:52   ` Rob Love
2003-12-20 20:07   ` Linus Torvalds
2003-12-20 21:32     ` Ingo Molnar
2003-12-20 23:15     ` Nick Piggin

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=20031220202611.GB32320@elte.hu \
    --to=mingo@elte.hu \
    --cc=akpm@osdl.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=piggin@cyberone.com.au \
    /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

all inboxes | Powered by JetHome®