mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: kosaki.motohiro@gmail.com
To: linux-kernel@vger.kernel.org
Cc: Olivier Langlois <olivier@trillion01.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Frederic Weisbecker <fweisbec@gmail.com>,
	Ingo Molnar <mingo@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Subject: [PATCH 4/8] posix-cpu-timers: timer functions should use timer time instead of clock time
Date: Sun, 26 May 2013 17:35:45 -0400	[thread overview]
Message-ID: <1369604149-13016-7-git-send-email-kosaki.motohiro@gmail.com> (raw)
In-Reply-To: <1369604149-13016-1-git-send-email-kosaki.motohiro@gmail.com>

From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>

For process timers, we use cpu_clock_sample_group() and cpu_timer_sample_group()
correctly. However, for thread timers, we always use cpu_clock_sample(). This is
wrong because a cpu_clock_sample() accounts uncommitted delta_exec too. And this
is inconsistent against run_posix_cpu_tiemrs().

This is not big matter because the following workaround code in
posix_cpu_timer_get() hides the timer miscalculation issue. However, it makes
rq lock held wrongly and it would be better to fix it.

  if (cpu_time_before(timer->it_clock, now, timer->it.cpu.expires)) {
    ...
  } else {
    /*
     * The timer should have expired already, but the firing
     * hasn't taken place yet.  Say it's just about to expire.
     */
     itp->it_value.tv_nsec = 1;
     itp->it_value.tv_sec = 0;

Cc: Olivier Langlois <olivier@trillion01.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
---
 kernel/posix-cpu-timers.c |   37 +++++++++++++++++++++++++++----------
 1 files changed, 27 insertions(+), 10 deletions(-)

diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c
index d068808..1b33c77 100644
--- a/kernel/posix-cpu-timers.c
+++ b/kernel/posix-cpu-timers.c
@@ -221,11 +221,10 @@ posix_cpu_clock_set(const clockid_t which_clock, const struct timespec *tp)
 }
 
 
-/*
- * Sample a per-thread clock for the given task.
- */
-static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p,
-			    union cpu_time_count *cpu)
+static int do_cpu_clock_timer_sample(const clockid_t which_clock,
+				     struct task_struct *p,
+				     bool add_delta,
+				     union cpu_time_count *cpu)
 {
 	switch (CPUCLOCK_WHICH(which_clock)) {
 	default:
@@ -237,12 +236,30 @@ static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p,
 		cpu->cpu = virt_ticks(p);
 		break;
 	case CPUCLOCK_SCHED:
-		cpu->sched = task_sched_runtime(p, true);
+		cpu->sched = task_sched_runtime(p, add_delta);
 		break;
 	}
 	return 0;
 }
 
+/*
+ * Sample a per-thread clock for the given task.
+ */
+static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p,
+			    union cpu_time_count *cpu)
+{
+	return do_cpu_clock_timer_sample(which_clock, p, true, cpu);
+}
+
+/*
+ * Sample a per-thread timer clock for the given task.
+ */
+static int cpu_timer_sample(const clockid_t which_clock, struct task_struct *p,
+			    union cpu_time_count *cpu)
+{
+	return do_cpu_clock_timer_sample(which_clock, p, false, cpu);
+}
+
 static void update_gt_cputime(struct task_cputime *a, struct task_cputime *b)
 {
 	if (b->utime > a->utime)
@@ -748,7 +765,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags,
 	 * check if it's already passed.  In short, we need a sample.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(timer->it_clock, p, &val);
+		cpu_timer_sample(timer->it_clock, p, &val);
 	} else {
 		cpu_timer_sample_group(timer->it_clock, p, &val);
 	}
@@ -804,7 +821,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int flags,
 		 * we need exact current time.
 		 */
 		if (CPUCLOCK_PERTHREAD(timer->it_clock))
-			now = val;
+			cpu_clock_sample(timer->it_clock, p, &now);
 		else
 			cpu_clock_sample_group(timer->it_clock, p, &now);
 		cpu_time_add(timer->it_clock, &new_expires, now);
@@ -894,7 +911,7 @@ static void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp)
 	 * Sample the clock to take the difference with the expiry time.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(timer->it_clock, p, &now);
+		cpu_timer_sample(timer->it_clock, p, &now);
 		clear_dead = p->exit_state;
 	} else {
 		read_lock(&tasklist_lock);
@@ -1203,7 +1220,7 @@ void posix_cpu_timer_schedule(struct k_itimer *timer)
 	 * Fetch the current sample and update the timer's expiry time.
 	 */
 	if (CPUCLOCK_PERTHREAD(timer->it_clock)) {
-		cpu_clock_sample(timer->it_clock, p, &now);
+		cpu_timer_sample(timer->it_clock, p, &now);
 		bump_cpu_timer(timer, now);
 		if (unlikely(p->exit_state)) {
 			clear_dead_task(timer, now);
-- 
1.7.1


  parent reply	other threads:[~2013-05-26 21:37 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-05-26 21:35 [PATCH v5 0/8] posix timers fixlet kosaki.motohiro
2013-05-26 21:35 ` [PATCH 1/8] posix-cpu-timers: don't account cpu timer after stopped thread runtime accounting kosaki.motohiro
2013-05-26 21:35 ` [PATCH 2/8] posix-cpu-timers: fix acounting delta_exec twice kosaki.motohiro
2013-06-03 19:07   ` Frederic Weisbecker
2013-05-26 21:35 ` kosaki.motohiro
2013-05-26 21:35 ` [PATCH 3/8] posix-cpu-timers: fix wrong timer initialization kosaki.motohiro
2013-06-18 14:20   ` Frederic Weisbecker
2013-06-18 15:12     ` KOSAKI Motohiro
2013-06-19 20:56       ` Frederic Weisbecker
2013-05-26 21:35 ` kosaki.motohiro [this message]
2013-05-26 21:35 ` [PATCH 5/8] posix-cpu-timers: check_thread_timers() uses task_sched_runtime() kosaki.motohiro
2013-05-26 21:35 ` [PATCH 6/8] sched: task_sched_runtime introduce micro optimization kosaki.motohiro
2013-06-18 14:27   ` Frederic Weisbecker
2013-06-18 15:17     ` KOSAKI Motohiro
2013-06-18 15:28       ` KOSAKI Motohiro
2013-06-18 17:18       ` Frederic Weisbecker
2013-06-20  8:41         ` KOSAKI Motohiro
2013-05-26 21:35 ` [PATCH 7/8] posix-cpu-timers: cleanup cpu_{clock,timer}_sample{,_group} kosaki.motohiro
2013-05-26 21:35 ` [PATCH 8/8] posix-cpu-timers: fix cputimer initialization mistake for {u,s}time kosaki.motohiro
2013-05-31 16:26 ` [PATCH v5 0/8] posix timers fixlet KOSAKI Motohiro

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=1369604149-13016-7-git-send-email-kosaki.motohiro@gmail.com \
    --to=kosaki.motohiro@gmail.com \
    --cc=fweisbec@gmail.com \
    --cc=kosaki.motohiro@jp.fujitsu.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=olivier@trillion01.com \
    --cc=peterz@infradead.org \
    --cc=tglx@linutronix.de \
    /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®