From: Feng Tang <feng.tang@intel.com>
To: John Stultz <john.stultz@linaro.org>,
Thomas Gleixner <tglx@linutronix.de>,
Stephen Boyd <sboyd@kernel.org>,
linux-kernel@vger.kernel.org
Cc: Qais Yousef <qais.yousef@arm.com>,
Peter Zijlstra <peterz@infradead.org>,
andi.kleen@intel.com, Feng Tang <feng.tang@intel.com>
Subject: [PATCH] clocksource: don't run watchdog forever
Date: Tue, 2 Mar 2021 10:54:24 +0800 [thread overview]
Message-ID: <1614653665-20905-1-git-send-email-feng.tang@intel.com> (raw)
clocksource watchdog runs every 500ms, which creates some OS noise.
As the clocksource wreckage (especially for those that has per-cpu
reading hook) usually happens shortly after CPU is brought up or
after system resumes from sleep state, so add a time limit for
clocksource watchdog to only run for a period of time, and make
sure it run at least twice for each CPU.
Regarding performance data, there is no improvement data with the
micro-benchmarks we have like hackbench/netperf/fio/will-it-scale
etc. But it obviously reduces periodic timer interrupts, and may
help in following cases:
* When some CPUs are isolated to only run scientific or high
performance computing tasks on a NOHZ_FULL kernel, where there
is almost no interrupts, this could make it more quiet
* On a cluster which runs a lot of systems in parallel with
barriers there are always enough systems which run the watchdog
and make everyone else wait
Signed-off-by: Feng Tang <feng.tang@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
---
include/linux/clocksource.h | 7 +++++++
kernel/cpu.c | 3 +++
kernel/time/clocksource.c | 31 +++++++++++++++++++++++++++++--
3 files changed, 39 insertions(+), 2 deletions(-)
diff --git a/include/linux/clocksource.h b/include/linux/clocksource.h
index 86d143d..cf428a2 100644
--- a/include/linux/clocksource.h
+++ b/include/linux/clocksource.h
@@ -252,6 +252,13 @@ static inline void __clocksource_update_freq_khz(struct clocksource *cs, u32 khz
__clocksource_update_freq_scale(cs, 1000, khz);
}
+#ifdef CONFIG_CLOCKSOURCE_WATCHDOG
+extern void clocksource_kick_watchdog(void);
+#else
+static inline void clocksource_kick_watchdog(void) { }
+#endif
+
+
#ifdef CONFIG_ARCH_CLOCKSOURCE_INIT
extern void clocksource_arch_init(struct clocksource *cs);
#else
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 1b6302e..fdf3c69 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -32,6 +32,7 @@
#include <linux/relay.h>
#include <linux/slab.h>
#include <linux/percpu-rwsem.h>
+#include <linux/clocksource.h>
#include <trace/events/power.h>
#define CREATE_TRACE_POINTS
@@ -1286,6 +1287,8 @@ static int cpu_up(unsigned int cpu, enum cpuhp_state target)
}
err = _cpu_up(cpu, 0, target);
+
+ clocksource_kick_watchdog();
out:
cpu_maps_update_done();
return err;
diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c
index cce484a..aba985a 100644
--- a/kernel/time/clocksource.c
+++ b/kernel/time/clocksource.c
@@ -104,6 +104,7 @@ static DECLARE_WORK(watchdog_work, clocksource_watchdog_work);
static DEFINE_SPINLOCK(watchdog_lock);
static int watchdog_running;
static atomic_t watchdog_reset_pending;
+static unsigned long watchdog_stop_time; /* in jiffies */
static inline void clocksource_watchdog_lock(unsigned long *flags)
{
@@ -295,10 +296,16 @@ static void clocksource_watchdog(struct timer_list *unused)
next_cpu = cpumask_first(cpu_online_mask);
/*
- * Arm timer if not already pending: could race with concurrent
- * pair clocksource_stop_watchdog() clocksource_start_watchdog().
+ * Arm timer if not already pending or pass the check time window:
+ * could race with concurrent pair clocksource_stop_watchdog()
+ * clocksource_start_watchdog().
*/
if (!timer_pending(&watchdog_timer)) {
+ if (time_after(jiffies, watchdog_stop_time)) {
+ atomic_inc(&watchdog_reset_pending);
+ watchdog_running = 0;
+ goto out;
+ }
watchdog_timer.expires += WATCHDOG_INTERVAL;
add_timer_on(&watchdog_timer, next_cpu);
}
@@ -308,6 +315,16 @@ static void clocksource_watchdog(struct timer_list *unused)
static inline void clocksource_start_watchdog(void)
{
+ unsigned long check_ticks;
+
+ /*
+ * As all CPUs will be looped to run the timer, make sure each
+ * CPU can run the timer twice, and the check run for at least
+ * 10 minutes.
+ */
+ check_ticks = max_t(unsigned long, num_possible_cpus(), 600) * HZ;
+ watchdog_stop_time = jiffies + check_ticks;
+
if (watchdog_running || !watchdog || list_empty(&watchdog_list))
return;
timer_setup(&watchdog_timer, clocksource_watchdog, 0);
@@ -324,6 +341,15 @@ static inline void clocksource_stop_watchdog(void)
watchdog_running = 0;
}
+void clocksource_kick_watchdog(void)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&watchdog_lock, flags);
+ clocksource_start_watchdog();
+ spin_unlock_irqrestore(&watchdog_lock, flags);
+}
+
static inline void clocksource_reset_watchdog(void)
{
struct clocksource *cs;
@@ -618,6 +644,7 @@ void clocksource_resume(void)
cs->resume(cs);
clocksource_resume_watchdog();
+ clocksource_kick_watchdog();
}
/**
--
2.7.4
next reply other threads:[~2021-03-02 7:59 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-02 2:54 Feng Tang [this message]
2021-03-02 9:16 ` [PATCH] clocksource: don't run watchdog forever Peter Zijlstra
2021-03-02 12:06 ` Feng Tang
2021-03-03 15:50 ` Thomas Gleixner
2021-03-04 7:43 ` Feng Tang
2021-03-04 14:15 ` Thomas Gleixner
2021-03-05 2:30 ` Feng Tang
2021-03-25 8:34 ` Feng Tang
2021-03-25 11:40 ` Thomas Gleixner
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=1614653665-20905-1-git-send-email-feng.tang@intel.com \
--to=feng.tang@intel.com \
--cc=andi.kleen@intel.com \
--cc=john.stultz@linaro.org \
--cc=linux-kernel@vger.kernel.org \
--cc=peterz@infradead.org \
--cc=qais.yousef@arm.com \
--cc=sboyd@kernel.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;
as well as URLs for NNTP newsgroup(s).