public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Venkatesh Pallipadi <venki@google.com>
To: Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@elte.hu>, "H. Peter Anvin" <hpa@zytor.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Balbir Singh <balbir@linux.vnet.ibm.com>,
	Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: linux-kernel@vger.kernel.org, Paul Turner <pjt@google.com>,
	Eric Dumazet <eric.dumazet@gmail.com>,
	Shaun Ruffell <sruffell@digium.com>,
	Yong Zhang <yong.zhang0@gmail.com>,
	Venkatesh Pallipadi <venki@google.com>
Subject: [PATCH 5/6] Account ksoftirq time as cpustat softirq
Date: Wed, 20 Oct 2010 15:49:01 -0700	[thread overview]
Message-ID: <1287614941-32325-6-git-send-email-venki@google.com> (raw)
In-Reply-To: <1287614941-32325-1-git-send-email-venki@google.com>

softirq time in ksoftirqd context is not accounted in ns granularity
per cpu softirq stats, as we want that to be a part of ksoftirqd
exec_runtime.

Accounting them separately, and using them while doing stats makes
/proc/stat include that time in cpu si time.

Signed-off-by: Venkatesh Pallipadi <venki@google.com>
---
 kernel/sched.c |   28 +++++++++++++++++++++-------
 1 files changed, 21 insertions(+), 7 deletions(-)

diff --git a/kernel/sched.c b/kernel/sched.c
index 7e812a6..cba8090 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -1941,6 +1941,7 @@ static void deactivate_task(struct rq *rq, struct task_struct *p, int flags)
  */
 static DEFINE_PER_CPU(u64, cpu_hardirq_time);
 static DEFINE_PER_CPU(u64, cpu_softirq_time);
+static DEFINE_PER_CPU(u64, cpu_ksoftirqd_time);
 
 static DEFINE_PER_CPU(u64, irq_start_time);
 static int sched_clock_irqtime;
@@ -1979,15 +1980,20 @@ void account_system_vtime(struct task_struct *curr)
 	delta = now - per_cpu(irq_start_time, cpu);
 	per_cpu(irq_start_time, cpu) = now;
 	/*
-	 * We do not account for softirq time from ksoftirqd here.
 	 * We want to continue accounting softirq time to ksoftirqd thread
 	 * in that case, so as not to confuse scheduler with a special task
 	 * that do not consume any time, but still wants to run.
+	 * So, softirq time in ksoftirqd is accounted separately and used
+	 * for softirq time reporting in /proc/stat.
 	 */
-	if (in_irq())
+	if (in_irq()) {
 		per_cpu(cpu_hardirq_time, cpu) += delta;
-	else if (in_serving_softirq() && !(is_ksoftirqd_context()))
-		per_cpu(cpu_softirq_time, cpu) += delta;
+	} else if (in_serving_softirq()) {
+		if (is_ksoftirqd_context())
+			per_cpu(cpu_ksoftirqd_time, cpu) += delta;
+		else
+			per_cpu(cpu_softirq_time, cpu) += delta;
+	}
 
 	local_irq_restore(flags);
 }
@@ -2025,7 +2031,9 @@ static int irqtime_account_si_update(void)
 	int ret = 0;
 
 	local_irq_save(flags);
-	latest_ns = __get_cpu_var(cpu_softirq_time);
+	latest_ns = __get_cpu_var(cpu_softirq_time) +
+			__get_cpu_var(cpu_ksoftirqd_time);
+
 	if (cputime64_gt(nsecs_to_cputime64(latest_ns), cpustat->softirq))
 		ret = 1;
 	local_irq_restore(flags);
@@ -3606,7 +3614,8 @@ void account_system_time(struct task_struct *p, int hardirq_offset,
  * no timer going off while we are on hardirq and hence we may never get an
  * oppurtunity to update it solely in system time.
  * p->stime and friends are only updated on system time and not on irq
- * softirq as those do not count in task exec_runtime any more.
+ * softirq as those do not count in task exec_runtime any more (except
+ * for ksoftirqd).
  */
 static void irqtime_account_process_tick(struct task_struct *p, int user_tick,
 						struct rq *rq)
@@ -3620,7 +3629,12 @@ static void irqtime_account_process_tick(struct task_struct *p, int user_tick,
 	} else if (user_tick) {
 		account_user_time(p, cputime_one_jiffy, one_jiffy_scaled);
 	} else if (irqtime_account_si_update()) {
-		cpustat->softirq = cputime64_add(cpustat->softirq, tmp);
+		if (is_ksoftirqd_context()) {
+			__account_system_time(p, cputime_one_jiffy,
+					one_jiffy_scaled, &cpustat->softirq);
+		} else {
+			cpustat->softirq = cputime64_add(cpustat->softirq, tmp);
+		}
 	} else if (p == rq->idle) {
 		account_idle_time(cputime_one_jiffy);
 	} else if (p->flags & PF_VCPU) { /* System time or guest time */
-- 
1.7.1


  parent reply	other threads:[~2010-10-20 22:49 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-10-20 22:48 [PATCH 0/5] Proper kernel irq time reporting -v0 Venkatesh Pallipadi
2010-10-20 22:48 ` [PATCH 1/6] Free up pf flag PF_KSOFTIRQD Venkatesh Pallipadi
2010-10-21  5:23   ` Eric Dumazet
2010-10-21 14:36     ` Venkatesh Pallipadi
2010-10-21 14:58       ` Eric Dumazet
2010-10-21 17:03         ` Venkatesh Pallipadi
2010-10-21 15:13   ` Christoph Lameter
2010-10-21 17:06     ` Venkatesh Pallipadi
2010-10-20 22:48 ` [PATCH 2/6] Add nsecs_to_cputime64 interface for asm-generic Venkatesh Pallipadi
2010-10-20 22:48 ` [PATCH 3/6] Refactor account_system_time separating id and actual update Venkatesh Pallipadi
2010-10-20 22:49 ` [PATCH 4/6] Export ns irqtimes from IRQ_TIME_ACCOUNTING through /proc/stat Venkatesh Pallipadi
2010-10-21 14:44   ` Peter Zijlstra
2010-10-21 19:25     ` Venkatesh Pallipadi
2010-10-22 12:23       ` Peter Zijlstra
2010-10-22 23:34         ` Venkatesh Pallipadi
2010-10-20 22:49 ` Venkatesh Pallipadi [this message]
2010-10-21 14:53   ` [PATCH 5/6] Account ksoftirq time as cpustat softirq Peter Zijlstra
2010-10-21 19:10     ` Venkatesh Pallipadi
2010-10-21 17:25 ` [PATCH 0/5] Proper kernel irq time reporting -v0 Shaun Ruffell

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=1287614941-32325-6-git-send-email-venki@google.com \
    --to=venki@google.com \
    --cc=balbir@linux.vnet.ibm.com \
    --cc=eric.dumazet@gmail.com \
    --cc=hpa@zytor.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=peterz@infradead.org \
    --cc=pjt@google.com \
    --cc=schwidefsky@de.ibm.com \
    --cc=sruffell@digium.com \
    --cc=tglx@linutronix.de \
    --cc=yong.zhang0@gmail.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