linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Mel Gorman <mgorman@suse.de>
To: Peter Zijlstra <peterz@infradead.org>
Cc: tglx@linutronix.de, mingo@kernel.org, juri.lelli@redhat.com,
	vincent.guittot@linaro.org, dietmar.eggemann@arm.com,
	rostedt@goodmis.org, bsegall@google.com, bristot@redhat.com,
	bsingharora@gmail.com, pbonzini@redhat.com, maz@kernel.org,
	linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
	riel@surriel.com, hannes@cmpxchg.org
Subject: Re: [PATCH 3/6] sched: Simplify sched_info_on()
Date: Wed, 12 May 2021 12:10:40 +0100	[thread overview]
Message-ID: <20210512111040.GC3672@suse.de> (raw)
In-Reply-To: <20210505111525.121458839@infradead.org>

On Wed, May 05, 2021 at 12:59:43PM +0200, Peter Zijlstra wrote:
> The situation around sched_info is somewhat complicated, it is used by
> sched_stats and delayacct and, indirectly, kvm.
> 
> If SCHEDSTATS=Y (but disabled by default) sched_info_on() is
> unconditionally true -- this is the case for all distro kernel configs
> I checked.
> 
> If for some reason SCHEDSTATS=N, but TASK_DELAY_ACCT=Y, then
> sched_info_on() can return false when delayacct is disabled,
> presumably because there would be no other users left; except kvm is.
> 
> Instead of complicating matters further by accurately accounting
> sched_stat and kvm state, simply unconditionally enable when
> SCHED_INFO=Y, matching the common distro case.
> 
> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> @@ -163,13 +158,12 @@ static inline void sched_info_reset_dequ
>   */
>  static inline void sched_info_dequeue(struct rq *rq, struct task_struct *t)
>  {
> -	unsigned long long now = rq_clock(rq), delta = 0;
> +	unsigned long long delta = 0;
>  
> -	if (sched_info_on()) {
> -		if (t->sched_info.last_queued)
> -			delta = now - t->sched_info.last_queued;
> +	if (t->sched_info.last_queued) {
> +		delta = rq_clock(rq) - t->sched_info.last_queued;
> +		t->sched_info.last_queued = 0;
>  	}
> -	sched_info_reset_dequeued(t);
>  	t->sched_info.run_delay += delta;
>  
>  	rq_sched_info_dequeue(rq, delta);

As delta is !0 iff t->sched_info.last_queued, why not this?

diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h
index 33ffd41935ba..37e33c0eeb7c 100644
--- a/kernel/sched/stats.h
+++ b/kernel/sched/stats.h
@@ -158,15 +158,14 @@ static inline void psi_sched_switch(struct task_struct *prev,
  */
 static inline void sched_info_dequeue(struct rq *rq, struct task_struct *t)
 {
-	unsigned long long delta = 0;
-
 	if (t->sched_info.last_queued) {
+		unsigned long long delta;
+
 		delta = rq_clock(rq) - t->sched_info.last_queued;
 		t->sched_info.last_queued = 0;
+		t->sched_info.run_delay += delta;
+		rq_sched_info_dequeue(rq, delta);
 	}
-	t->sched_info.run_delay += delta;
-
-	rq_sched_info_dequeue(rq, delta);
 }
 
 /*

> @@ -184,9 +178,10 @@ static void sched_info_arrive(struct rq
>  {
>  	unsigned long long now = rq_clock(rq), delta = 0;
>  
> -	if (t->sched_info.last_queued)
> +	if (t->sched_info.last_queued) {
>  		delta = now - t->sched_info.last_queued;
> -	sched_info_reset_dequeued(t);
> +		t->sched_info.last_queued = 0;
> +	}
>  	t->sched_info.run_delay += delta;
>  	t->sched_info.last_arrival = now;
>  	t->sched_info.pcount++;

Similarly

@@ -176,17 +175,18 @@ static inline void sched_info_dequeue(struct rq *rq, struct task_struct *t)
  */
 static void sched_info_arrive(struct rq *rq, struct task_struct *t)
 {
-	unsigned long long now = rq_clock(rq), delta = 0;
+	unsigned long long now = rq_clock(rq);
 
 	if (t->sched_info.last_queued) {
+		unsigned long long delta;
+
 		delta = now - t->sched_info.last_queued;
 		t->sched_info.last_queued = 0;
+		t->sched_info.run_delay += delta;
+		rq_sched_info_arrive(rq, delta);
 	}
-	t->sched_info.run_delay += delta;
 	t->sched_info.last_arrival = now;
 	t->sched_info.pcount++;
-
-	rq_sched_info_arrive(rq, delta);
 }
 
 /*

-- 
Mel Gorman
SUSE Labs

  parent reply	other threads:[~2021-05-12 11:10 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-05 10:59 [PATCH 0/6] sched,delayacct: Some cleanups Peter Zijlstra
2021-05-05 10:59 ` [PATCH 1/6] delayacct: Use sched_clock() Peter Zijlstra
2021-05-05 14:40   ` Rik van Riel
2021-05-06 13:59   ` Johannes Weiner
2021-05-06 14:17     ` Peter Zijlstra
2021-05-06 15:17       ` Johannes Weiner
2021-05-07 12:40   ` Balbir Singh
2021-05-12 10:28   ` [tip: sched/core] " tip-bot2 for Peter Zijlstra
2021-05-12 10:43   ` [PATCH 1/6] " Mel Gorman
2021-05-05 10:59 ` [PATCH 2/6] sched: Rename sched_info_{queued,dequeued} Peter Zijlstra
2021-05-05 14:39   ` Rik van Riel
2021-05-06 13:59   ` Johannes Weiner
2021-05-10  8:45   ` Balbir Singh
2021-05-12 10:28   ` [tip: sched/core] " tip-bot2 for Peter Zijlstra
2021-05-12 10:49   ` [PATCH 2/6] " Mel Gorman
2021-05-05 10:59 ` [PATCH 3/6] sched: Simplify sched_info_on() Peter Zijlstra
2021-05-06 14:03   ` Johannes Weiner
2021-05-12 10:28   ` [tip: sched/core] " tip-bot2 for Peter Zijlstra
2021-05-12 11:10   ` Mel Gorman [this message]
2021-05-12 11:32     ` [PATCH 3/6] " Peter Zijlstra
2021-05-12 12:51       ` Mel Gorman
2021-05-05 10:59 ` [PATCH 4/6] kvm: Select SCHED_INFO instead of TASK_DELAY_ACCT Peter Zijlstra
2021-05-05 11:37   ` Paolo Bonzini
2021-05-06 14:38   ` Marc Zyngier
2021-05-07 12:42   ` Balbir Singh
2021-05-12 10:28   ` [tip: sched/core] " tip-bot2 for Peter Zijlstra
2021-05-12 11:11   ` [PATCH 4/6] " Mel Gorman
2021-05-05 10:59 ` [PATCH 5/6] delayacct: Add static_branch in scheduler hooks Peter Zijlstra
2021-05-06 14:05   ` Johannes Weiner
2021-05-10  8:42   ` Balbir Singh
2021-05-12 10:28   ` [tip: sched/core] " tip-bot2 for Peter Zijlstra
2021-05-12 11:13   ` [PATCH 5/6] " Mel Gorman
2021-05-05 10:59 ` [PATCH 6/6] [RFC] delayacct: Default disabled Peter Zijlstra
2021-05-12 10:28   ` [tip: sched/core] " tip-bot2 for Peter Zijlstra
2021-05-12 11:35   ` [PATCH 6/6] [RFC] " Mel Gorman
2021-05-05 22:29 ` [PATCH 0/6] sched,delayacct: Some cleanups Balbir Singh
2021-05-06  9:13   ` Peter Zijlstra
2021-05-07 12:38     ` Balbir Singh
2021-05-12 11:34       ` Mel Gorman
2021-05-12 11:38         ` Peter Zijlstra
2021-05-12 12:23         ` Paul Wise
2021-05-12 13:00           ` Mel Gorman
2021-05-13  1:29             ` Paul Wise
2021-06-25  0:50         ` Paul Wise
2021-05-07  9:05 ` Thomas Gleixner
2021-05-10  7:08 ` Ingo Molnar
2021-05-10 12:05 ` [PATCH 7/6] delayacct: Add sysctl to enable at runtime Peter Zijlstra
2021-05-10 12:06   ` Peter Zijlstra
2021-05-12 10:28   ` [tip: sched/core] " tip-bot2 for Peter Zijlstra
2021-05-12 11:40   ` [PATCH 7/6] " Mel Gorman
2021-05-19  8:09     ` [tip: sched/core] delayacct: Document task_delayacct sysctl tip-bot2 for Mel Gorman

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=20210512111040.GC3672@suse.de \
    --to=mgorman@suse.de \
    --cc=bristot@redhat.com \
    --cc=bsegall@google.com \
    --cc=bsingharora@gmail.com \
    --cc=dietmar.eggemann@arm.com \
    --cc=hannes@cmpxchg.org \
    --cc=juri.lelli@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=maz@kernel.org \
    --cc=mingo@kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=peterz@infradead.org \
    --cc=riel@surriel.com \
    --cc=rostedt@goodmis.org \
    --cc=tglx@linutronix.de \
    --cc=vincent.guittot@linaro.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 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).