ksm: React on changing "sleep_millisecs" parameter faster

Submitted by Kirill Tkhai on Dec. 12, 2018, 11:15 a.m.


Message ID 154461329269.5296.5046385991304182077.stgit@localhost.localdomain
State New
Series "ksm: React on changing "sleep_millisecs" parameter faster"
Headers show

Commit Message

Kirill Tkhai Dec. 12, 2018, 11:15 a.m.
[This commit is from mm tree. It may be useful for svarog@,
 when he changes ksm politics in dependence of VMs number]

ksm thread unconditionally sleeps in ksm_scan_thread()
after each iteration:


The timeout is configured in /sys/kernel/mm/ksm/sleep_millisecs.

In case of user writes a big value by a mistake, and the thread
enters into schedule_timeout_interruptible(), it's not possible
to cancel the sleep by writing a new smaler value; the thread
is just sleeping till timeout expires.

The patch fixes the problem by waking the thread each time
after the value is updated.

This also may be useful for debug purposes; and also for userspace
daemons, which change sleep_millisecs value in dependence of
system load.

Link: http://lkml.kernel.org/r/154454107680.3258.3558002210423531566.stgit@localhost.localdomain
Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Acked-by: Cyrill Gorcunov <gorcunov@gmail.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
 mm/ksm.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

Patch hide | download patch | download mbox

diff --git a/mm/ksm.c b/mm/ksm.c
index 1fb8051b22e4..5ba6b218cdb6 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -269,6 +269,7 @@  static unsigned long ksm_run = KSM_RUN_STOP;
 static void wait_while_offlining(void);
 static DECLARE_WAIT_QUEUE_HEAD(ksm_thread_wait);
+static DECLARE_WAIT_QUEUE_HEAD(ksm_iter_wait);
 static DEFINE_MUTEX(ksm_thread_mutex);
 static DEFINE_SPINLOCK(ksm_mmlist_lock);
@@ -2315,6 +2316,8 @@  static int ksmd_should_run(void)
 static int ksm_scan_thread(void *nothing)
+	unsigned int sleep_ms;
 	set_user_nice(current, 5);
@@ -2328,8 +2331,10 @@  static int ksm_scan_thread(void *nothing)
 		if (ksmd_should_run()) {
-			schedule_timeout_interruptible(
-				msecs_to_jiffies(ksm_thread_sleep_millisecs));
+			sleep_ms = READ_ONCE(ksm_thread_sleep_millisecs);
+			wait_event_interruptible_timeout(ksm_iter_wait,
+				sleep_ms != READ_ONCE(ksm_thread_sleep_millisecs),
+				msecs_to_jiffies(sleep_ms));
 		} else {
 				ksmd_should_run() || kthread_should_stop());
@@ -2850,6 +2855,7 @@  static ssize_t sleep_millisecs_store(struct kobject *kobj,
 		return -EINVAL;
 	ksm_thread_sleep_millisecs = msecs;
+	wake_up_interruptible(&ksm_iter_wait);
 	return count;