LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Karsten Wiese <fzu@wemgehoertderstaat.de>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: Andrew Morton <akpm@osdl.org>,
	LKML <linux-kernel@vger.kernel.org>, Ingo Molnar <mingo@elte.hu>,
	John Stultz <johnstul@us.ibm.com>,
	Arjan van de Veen <arjan@infradead.org>,
	Roman Zippel <zippel@linux-m68k.org>
Subject: [PATCH] high_res_timers: precisely update_process_times; Re: [patch 36/46] tick-management: dyntick / highres functionality
Date: Sun, 28 Jan 2007 03:03:35 +0100	[thread overview]
Message-ID: <200701280303.36441.fzu@wemgehoertderstaat.de> (raw)
In-Reply-To: <20070123211208.550932000@localhost.localdomain>

Am Dienstag, 23. Januar 2007 23:01 schrieb Thomas Gleixner:
> 
> Add functions to provide dynamic ticks and high resolution timers.
> The code which keeps track of jiffies and handles the long idle
> periods is shared between tick based and high resolution timer based
> dynticks. The dyntick functionality can be disabled on the kernel
> commandline. Provide also the infrastructure to support high resolution
> timers.

cpufreq_ondemand didn't work here on kernels 2.6.20-rc4-rt1 and above.
Following patch against 20-rc6-rt2 helps.

      Karsten
-------------------------------------------------------------------------------
From: Karsten Wiese <fzu@wemgehoertderstaat.de>

high_res_timers: precisely update_process_times

Sometimes tick_sched_timer() calls tick_do_update_jiffies64() and
jiffies is updated by !=1.
Cope with these situations by splitting update_process_times() into
__update_process_times() and tick() and
exchanging call to update_process_times() from tick_sched_timer()
by calls to the splits.
Fixes cpufreq_ondemand going nuts here. 

Signed-off-by: Karsten Wiese <fzu@wemgehoertderstaat.de>


--- rc6-rt2/include/linux/sched.h	2007-01-26 14:42:55.000000000 +0100
+++ rc6-rt2-kw/include/linux/sched.h	2007-01-28 02:02:29.000000000 +0100
@@ -264,6 +264,13 @@ long io_schedule_timeout(long timeout);
 extern void cpu_init (void);
 extern void trap_init(void);
 extern void update_process_times(int user);
+#ifdef CONFIG_HIGH_RES_TIMERS
+extern void __update_process_times(int user_tick, unsigned long ticks);
+extern void tick(int user_tick);
+#define NO_HIGH_RES_TIMERS_STATIC
+#else
+#define NO_HIGH_RES_TIMERS_STATIC static
+#endif
 extern void scheduler_tick(void);
 
 #ifdef CONFIG_GENERIC_HARDIRQS
diff -pur rc6-rt2/kernel/time/tick-sched.c rc6-rt2-kw/kernel/time/tick-sched.c
--- rc6-rt2/kernel/time/tick-sched.c	2007-01-26 14:42:55.000000000 +0100
+++ rc6-rt2-kw/kernel/time/tick-sched.c	2007-01-28 01:45:14.000000000 +0100
@@ -41,9 +41,9 @@ struct tick_sched *tick_get_tick_sched(i
 /*
  * Must be called with interrupts disabled !
  */
-static void tick_do_update_jiffies64(ktime_t now)
+static unsigned long tick_do_update_jiffies64(ktime_t now)
 {
-	unsigned long ticks = 1;
+	unsigned long ticks = 0;
 	ktime_t delta;
 
 	/* Reevalute with xtime_lock held */
@@ -51,7 +51,7 @@ static void tick_do_update_jiffies64(kti
 
 	delta = ktime_sub(now, last_jiffies_update);
 	if (delta.tv64 >= tick_period.tv64) {
-
+		ticks = 1;
 		delta = ktime_sub(delta, tick_period);
 		last_jiffies_update = ktime_add(last_jiffies_update,
 						tick_period);
@@ -68,6 +68,7 @@ static void tick_do_update_jiffies64(kti
 		do_timer(ticks);
 	}
 	write_sequnlock(&xtime_lock);
+	return ticks;
 }
 
 /*
@@ -423,32 +424,36 @@ static enum hrtimer_restart tick_sched_t
 	ktime_t now = ktime_get();
 
 	/* Check, if the jiffies need an update */
-	tick_do_update_jiffies64(now);
+	unsigned long ticks = tick_do_update_jiffies64(now);
 
 	/*
 	 * Do not call, when we are not in irq context and have
 	 * no valid regs pointer
 	 */
 	if (regs) {
-		/*
-		 * When we are idle and the tick is stopped, we have to touch
-		 * the watchdog as we might not schedule for a really long
-		 * time. This happens on complete idle SMP systems while
-		 * waiting on the login prompt. We also increment the "start of
-		 * idle" jiffy stamp so the idle accounting adjustment we do
-		 * when we go busy again does not account too much ticks.
-		 */
-		if (ts->tick_stopped) {
-			touch_softlockup_watchdog();
-			ts->idle_jiffies++;
+		if (ticks) {
+			/*
+			 * When we are idle and the tick is stopped, we have to touch
+			 * the watchdog as we might not schedule for a really long
+			 * time. This happens on complete idle SMP systems while
+			 * waiting on the login prompt. We also increment the "start of
+			 * idle" jiffy stamp so the idle accounting adjustment we do
+			 * when we go busy again does not account too much ticks.
+			 */
+			if (ts->tick_stopped) {
+				touch_softlockup_watchdog();
+				ts->idle_jiffies++;
+				__update_process_times(user_mode(regs), 1);
+			} else
+				__update_process_times(user_mode(regs), ticks);
 		}
 		/*
-		 * update_process_times() might take tasklist_lock, hence
+		 * tick() might take tasklist_lock, hence
 		 * drop the base lock. sched-tick hrtimers are per-CPU and
 		 * never accessible by userspace APIs, so this is safe to do.
 		 */
 		spin_unlock(&base->lock);
-		update_process_times(user_mode(regs));
+		tick(user_mode(regs));
 //		profile_tick(CPU_PROFILING);
 		spin_lock(&base->lock);
 	}
diff -pur rc6-rt2/kernel/timer.c rc6-rt2-kw/kernel/timer.c
--- rc6-rt2/kernel/timer.c	2007-01-26 14:42:55.000000000 +0100
+++ rc6-rt2-kw/kernel/timer.c	2007-01-28 02:03:01.000000000 +0100
@@ -1312,20 +1312,24 @@ static void update_wall_time(void)
 	update_vsyscall(&xtime, clock);
 }
 
-/*
- * Called from the timer interrupt handler to charge one tick to the current 
- * process.  user_tick is 1 if the tick is user time, 0 for system.
- */
-void update_process_times(int user_tick)
+NO_HIGH_RES_TIMERS_STATIC
+void __update_process_times(int user_tick, unsigned long ticks)
 {
-	int cpu = smp_processor_id();
 	struct task_struct *p = current;
 
 	/* Note: this timer irq context must be accounted for as well. */
 	if (user_tick)
-		account_user_time(p, jiffies_to_cputime(1));
+		account_user_time(p, jiffies_to_cputime(ticks));
 	else
-		account_system_time(p, HARDIRQ_OFFSET, jiffies_to_cputime(1));
+		account_system_time(p, HARDIRQ_OFFSET, jiffies_to_cputime(ticks));
+}
+
+NO_HIGH_RES_TIMERS_STATIC
+void tick(int user_tick)
+{
+	int cpu = smp_processor_id();
+	struct task_struct *p = current;
+
 	scheduler_tick();
 	run_local_timers();
 	if (rcu_pending(cpu))
@@ -1334,6 +1338,16 @@ void update_process_times(int user_tick)
 }
 
 /*
+ * Called from the timer interrupt handler to charge one tick to the current 
+ * process.  user_tick is 1 if the tick is user time, 0 for system.
+ */
+void update_process_times(int user_tick)
+{
+	__update_process_times(user_tick, 1);
+	tick(user_tick);
+}
+
+/*
  * Nr of active tasks - counted in fixed-point numbers
  */
 static unsigned long count_active_tasks(void)

  reply	other threads:[~2007-01-28  2:03 UTC|newest]

Thread overview: 78+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-01-23 22:00 [patch 00/46] High resolution timer / dynamic tick update Thomas Gleixner
2007-01-23 22:00 ` [patch 01/46] Add irq flag to disable balancing for an interrupt Thomas Gleixner
2007-01-23 22:00 ` [patch 02/46] Add a functions to handle interrupt affinity setting Thomas Gleixner
2007-01-23 22:00 ` [patch 03/46] [RFC] HZ free ntp Thomas Gleixner
2007-01-23 22:00 ` [patch 04/46] Uninline jiffies.h functions Thomas Gleixner
2007-01-23 22:01 ` [patch 05/46] Thomas Gleixner
2007-01-23 22:01 ` [patch 06/46] Fix timeout overflow with jiffies Thomas Gleixner
2007-01-23 22:01 ` [patch 07/46] GTOD: persistent clock support Thomas Gleixner
2007-01-23 22:01 ` [patch 08/46] i386: use GTOD " Thomas Gleixner
2007-01-23 22:01 ` [patch 09/46] i386 Remove useless code in tsc.c Thomas Gleixner
2007-01-23 22:01 ` [patch 10/46] Simplify the registration of clocksources Thomas Gleixner
2007-01-23 22:01 ` [patch 11/46] x86: rewrite SMP TSC sync code Thomas Gleixner
2007-01-23 22:01 ` [patch 12/46] clocksource: replace is_continuous by a flag field Thomas Gleixner
2007-01-24 11:23   ` [patch] clocksource: fixup is_continous changes in vmitime.c Ingo Molnar
2007-01-24 11:53     ` Thomas Gleixner
2007-01-23 22:01 ` [patch 13/46] clocksource: fixup is_continous changes on ARM Thomas Gleixner
2007-01-23 22:01 ` [patch 14/46] clocksource: fixup is_continous changes on AVR32 Thomas Gleixner
2007-01-23 22:01 ` [patch 15/46] clocksource: fixup is_continous changes on S390 Thomas Gleixner
2007-01-23 22:01 ` [patch 16/46] clocksource: fixup is_continous changes on MIPS Thomas Gleixner
2007-01-23 22:01 ` [patch 17/46] clocksource: Remove the update callback Thomas Gleixner
2007-01-23 22:01 ` [patch 18/46] clocksource: Add verification (watchdog) helper Thomas Gleixner
2007-01-24 15:42   ` [patch] clocksource: add verification (watchdog) helper, fix Ingo Molnar
2007-01-23 22:01 ` [patch 19/46] Mark TSC on GeodeLX reliable Thomas Gleixner
2007-01-23 22:01 ` [patch 20/46] uninline irq_enter() Thomas Gleixner
2007-01-23 22:01 ` [patch 21/46] Fix cascade lookup of next_timer_interrupt Thomas Gleixner
2007-01-23 22:01 ` [patch 22/46] Extend next_timer_interrupt() to use a reference jiffie Thomas Gleixner
2007-01-23 22:01 ` [patch 23/46] hrtimers: namespace and enum cleanup Thomas Gleixner
2007-01-23 22:01 ` [patch 24/46] hrtimers: namespace and enum cleanup vs. git-input Thomas Gleixner
2007-01-23 22:01 ` [patch 25/46] hrtimers: cleanup locking Thomas Gleixner
2007-01-23 22:01 ` [patch 26/46] hrtimers; add state tracking Thomas Gleixner
2007-01-23 22:01 ` [patch 27/46] hrtimers: clean up callback tracking Thomas Gleixner
2007-01-23 22:01 ` [patch 28/46] hrtimers: move and add documentation Thomas Gleixner
2007-01-23 22:01 ` [patch 29/46] ACPI: fix missing include for UP Thomas Gleixner
2007-01-23 22:01 ` [patch 30/46] ACPI keep track of timer broadcasting Thomas Gleixner
2007-01-23 22:01 ` [patch 31/46] Allow early access to the power management timer Thomas Gleixner
2007-01-23 22:01 ` [patch 32/46] i386, apic: clean up the APIC code Thomas Gleixner
2007-01-23 22:01 ` [patch 33/46] clockevents: add core functionality Thomas Gleixner
2007-01-23 22:01 ` [patch 34/46] tick-management: " Thomas Gleixner
2007-01-23 22:01 ` [patch 35/46] tick-management: broadcast functionality Thomas Gleixner
2007-01-23 22:01 ` [patch 36/46] tick-management: dyntick / highres functionality Thomas Gleixner
2007-01-28  2:03   ` Karsten Wiese [this message]
2007-01-23 22:01 ` [patch 37/46] clockevents: i383 drivers Thomas Gleixner
2007-01-23 22:01 ` [patch 38/46] i386 rework local apic timer calibration Thomas Gleixner
2007-01-23 22:01 ` [patch 39/46] i386 prepare for dyntick Thomas Gleixner
2007-01-23 22:01 ` [patch 40/46] i386 prepare nmi watchdog for dynticks Thomas Gleixner
2007-01-23 22:01 ` [patch 41/46] i386: enable dynticks in kconfig Thomas Gleixner
2007-01-23 22:01 ` [patch 42/46] hrtimers: add high resolution timer support Thomas Gleixner
2007-01-23 22:01 ` [patch 43/46] hrtimers: prevent possible itimer DoS Thomas Gleixner
2007-01-23 22:01 ` [patch 44/46] Add debugging feature /proc/timer_stat Thomas Gleixner
2007-01-23 22:01 ` [patch 45/46] Add debugging feature /proc/timer_list Thomas Gleixner
2007-01-23 22:01 ` [patch 46/46] Add SysRq-Q to print timer_list debug info Thomas Gleixner
2007-01-24  2:16 ` [patch 00/46] High resolution timer / dynamic tick update Daniel Walker
2007-01-24  2:23   ` Andrew Morton
2007-01-24  3:25     ` Daniel Walker
2007-01-24  7:07   ` Ingo Molnar
2007-01-24  9:30     ` Daniel Walker
2007-01-24  9:51       ` Ingo Molnar
2007-01-24 10:23         ` Daniel Walker
2007-01-24 10:29           ` Ingo Molnar
2007-01-24 10:53             ` Daniel Walker
2007-01-24 11:04               ` Ingo Molnar
2007-01-24 11:13           ` Thomas Gleixner
2007-01-24 15:53             ` Daniel Walker
     [not found]               ` <20070124160046.GA24798@elte.hu>
2007-01-24 17:21                 ` Daniel Walker
     [not found]                 ` <1169655076.19471.241.camel@imap.mvista.com>
2007-01-24 19:38                   ` Ingo Molnar
2007-01-24 20:09                     ` Daniel Walker
2007-01-24 20:13                       ` Ingo Molnar
2007-01-24 19:57       ` john stultz
2007-01-24 20:51         ` Daniel Walker
2007-01-24 21:23           ` john stultz
2007-01-24 21:37             ` Daniel Walker
2007-01-25  6:10           ` Ingo Molnar
2007-01-25  6:37           ` Ingo Molnar
2007-01-25  6:32         ` Ingo Molnar
2007-01-25 16:38           ` Daniel Walker
2007-01-28  2:17 ` Andrew Morton
2007-01-29 21:31   ` john stultz
2007-01-29 21:45     ` john stultz

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=200701280303.36441.fzu@wemgehoertderstaat.de \
    --to=fzu@wemgehoertderstaat.de \
    --cc=akpm@osdl.org \
    --cc=arjan@infradead.org \
    --cc=johnstul@us.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=tglx@linutronix.de \
    --cc=zippel@linux-m68k.org \
    --subject='[PATCH] high_res_timers: precisely update_process_times; Re: [patch 36/46] tick-management: dyntick / highres functionality' \
    /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

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).