From: "Rafael J. Wysocki" <rjw@sisk.pl>
To: pm list <linux-pm@lists.linux-foundation.org>
Cc: Nigel Cunningham <nigel@nigel.suspend2.net>,
Oleg Nesterov <oleg@tv-sign.ru>, Pavel Machek <pavel@ucw.cz>
Subject: [RFC][PATCH -mm 2/2] Freezer: Use wait queue instead of busy looping
Date: Wed, 25 Jul 2007 14:09:39 +0200 [thread overview]
Message-ID: <200707251409.40592.rjw@sisk.pl> (raw)
In-Reply-To: <200707251401.48340.rjw@sisk.pl>
From: Rafael J. Wysocki <rjw@sisk.pl>
Use the observation that try_to_freeze() need not loop while waiting for the
freezing tasks to enter the refrigerator and make it use a wait queue.
The idea is that after sending freeze requests to the tasks regarded as
freezable try_to_freeze() can go to sleep and wait until at least one task
enters the refrigerator. The first task that does it wakes up try_to_freeze()
and the procedure is repeated. If the refrigerator is not entered by any tasks
before TIMEOUT expires, try_to_freeze() increases the counter of expired
timeouts and sends freeze requests to the remaining tasks. If the number of
expired timeouts becomes greater than MAX_WAITS, the freezing of tasks fails
(the counter of expired timeouts is reset whenever a task enters the
refrigerator).
This way, try_to_freeze() doesn't occupy the CPU unnecessarily when some
freezing tasks are waiting for I/O to complete and we have more fine grained
control over the freezing procedure.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Nigel Cunningham <nigel@nigel.suspend2.net>
---
kernel/power/process.c | 70 +++++++++++++++++++++++++++++++++++++++++--------
1 file changed, 59 insertions(+), 11 deletions(-)
Index: linux-2.6.23-rc1/kernel/power/process.c
===================================================================
--- linux-2.6.23-rc1.orig/kernel/power/process.c
+++ linux-2.6.23-rc1/kernel/power/process.c
@@ -13,11 +13,22 @@
#include <linux/module.h>
#include <linux/syscalls.h>
#include <linux/freezer.h>
+#include <linux/time.h>
/*
- * Timeout for stopping processes
+ * Time to wait until one or more tasks enter the refrigerator after sending
+ * freeze requests to them.
*/
-#define TIMEOUT (20 * HZ)
+#define TIMEOUT (HZ / 5)
+
+/*
+ * Each time after sending freeze requests to tasks the freezer will wait until
+ * some of them enter the refrigerater, but no longer than TIMEOUT. If TIMEOUT
+ * has been exceeded, the freezer increases the number of waits by one and
+ * repeats. If the number of waits becomes greater than MAX_WAITS, the
+ * freezing fails.
+ */
+#define MAX_WAITS 5
#define FREEZER_KERNEL_THREADS 0
#define FREEZER_USER_SPACE 1
@@ -43,6 +54,18 @@ static inline void frozen_process(void)
clear_freeze_flag(current);
}
+/*
+ * Wait queue head used by try_to_freeze_tasks() to wait for tasks to enter the
+ * refrigerator.
+ */
+static DECLARE_WAIT_QUEUE_HEAD(refrigerator_waitq);
+
+/*
+ * Used to signal try_to_freeze_tasks() that the refrigerator has been entered
+ * by a task.
+ */
+static int refrigerator_called;
+
/* Refrigerator is place where frozen processes are stored :-). */
void refrigerator(void)
{
@@ -50,6 +73,9 @@ void refrigerator(void)
processes around? */
long save;
+ refrigerator_called = 1;
+ wake_up(&refrigerator_waitq);
+
task_lock(current);
if (freezing(current)) {
frozen_process();
@@ -166,10 +192,16 @@ static void cancel_freezing(struct task_
static int try_to_freeze_tasks(int freeze_user_space)
{
struct task_struct *g, *p;
- unsigned long end_time;
- unsigned int todo;
+ unsigned int todo, waits;
+ unsigned long ret;
+ struct timeval start, end;
+ s64 elapsed_csecs64;
+ unsigned int elapsed_csecs;
+
+ do_gettimeofday(&start);
- end_time = jiffies + TIMEOUT;
+ refrigerator_called = 0;
+ waits = 0;
do {
todo = 0;
read_lock(&tasklist_lock);
@@ -189,11 +221,25 @@ static int try_to_freeze_tasks(int freez
todo++;
} while_each_thread(g, p);
read_unlock(&tasklist_lock);
- yield(); /* Yield is okay here */
- if (time_after(jiffies, end_time))
- break;
+
+ if (todo) {
+ ret = wait_event_timeout(refrigerator_waitq,
+ refrigerator_called, TIMEOUT);
+ if (!ret) {
+ if (++waits > MAX_WAITS)
+ break;
+ } else {
+ refrigerator_called = 0;
+ waits = 0;
+ }
+ }
} while (todo);
+ do_gettimeofday(&end);
+ elapsed_csecs64 = timeval_to_ns(&end) - timeval_to_ns(&start);
+ do_div(elapsed_csecs64, NSEC_PER_SEC / 100);
+ elapsed_csecs = elapsed_csecs64;
+
if (todo) {
/* This does not unfreeze processes that are already frozen
* (we have slightly ugly calling convention in that respect,
@@ -201,10 +247,9 @@ static int try_to_freeze_tasks(int freez
* but it cleans up leftover PF_FREEZE requests.
*/
printk("\n");
- printk(KERN_ERR "Freezing of %s timed out after %d seconds "
+ printk(KERN_ERR "Freezing of tasks failed after %d.%d seconds "
"(%d tasks refusing to freeze):\n",
- freeze_user_space ? "user space " : "tasks ",
- TIMEOUT / HZ, todo);
+ elapsed_csecs / 100, elapsed_csecs % 100, todo);
show_state();
read_lock(&tasklist_lock);
do_each_thread(g, p) {
@@ -215,6 +260,9 @@ static int try_to_freeze_tasks(int freez
task_unlock(p);
} while_each_thread(g, p);
read_unlock(&tasklist_lock);
+ } else {
+ printk("(elapsed %d.%d seconds) ", elapsed_csecs / 100,
+ elapsed_csecs % 100);
}
return todo ? -EBUSY : 0;
next prev parent reply other threads:[~2007-07-25 12:09 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-07-25 12:01 [RFC][PATCH -mm 0/2] Freezer: Use wait queue instead of busy looping Rafael J. Wysocki
2007-07-25 12:03 ` [RFC][PATCH -mm 1/2] Freezer: Be more verbose Rafael J. Wysocki
2007-07-25 12:27 ` Pavel Machek
2007-07-25 12:09 ` Rafael J. Wysocki [this message]
2007-07-25 12:28 ` [RFC][PATCH -mm 2/2] Freezer: Use wait queue instead of busy looping Pavel Machek
2007-07-25 12:55 ` Rafael J. Wysocki
2007-07-25 13:29 ` Oleg Nesterov
2007-07-25 14:03 ` Rafael J. Wysocki
2007-07-25 14:24 ` Oleg Nesterov
2007-07-26 12:24 ` Rafael J. Wysocki
2007-07-26 12:43 ` Rafael J. Wysocki
2007-07-31 8:01 ` Pavel Machek
2007-07-31 9:39 ` Rafael J. Wysocki
2007-07-31 10:00 ` Pavel Machek
2007-07-31 10:17 ` Rafael J. Wysocki
2007-07-31 10:08 ` Rafael J. Wysocki
2007-07-31 10:02 ` Pavel Machek
2007-07-31 22:25 ` [RFC][PATCH -mm 0/3] Freezer: Use wait queue instead of busy looping (updated) Rafael J. Wysocki
2007-07-31 22:26 ` [RFC][PATCH -mm 1/3] Freezer: Use wait queue instead of busy looping Rafael J. Wysocki
2007-08-01 7:59 ` Pavel Machek
2007-07-31 22:28 ` [RFC][PATCH -mm 2/3] Freezer: Measure the time of freezing tasks Rafael J. Wysocki
2007-08-01 8:28 ` Pavel Machek
2007-07-31 22:29 ` [RFC][PATCH -mm 3/3] Freezer: Replace the timeout Rafael J. Wysocki
2007-08-01 8:31 ` Pavel Machek
2007-08-01 10:43 ` Rafael J. Wysocki
2007-08-05 21:37 ` Pavel Machek
2007-08-05 22:38 ` Rafael J. Wysocki
2007-08-05 22:53 ` Pavel Machek
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=200707251409.40592.rjw@sisk.pl \
--to=rjw@sisk.pl \
--cc=linux-pm@lists.linux-foundation.org \
--cc=nigel@nigel.suspend2.net \
--cc=oleg@tv-sign.ru \
--cc=pavel@ucw.cz \
/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