sched: do not count frozen tasks toward load
authorNathan Lynch <ntl@pobox.com>
Thu, 9 Apr 2009 00:45:12 +0000 (19:45 -0500)
committerIngo Molnar <mingo@elte.hu>
Thu, 9 Apr 2009 05:37:02 +0000 (07:37 +0200)
Freezing tasks via the cgroup freezer causes the load average to climb
because the freezer's current implementation puts frozen tasks in
uninterruptible sleep (D state).

Some applications which perform job-scheduling functions consult the
load average when making decisions.  If a cgroup is frozen, the load
average does not provide a useful measure of the system's utilization
to such applications.  This is especially inconvenient if the job
scheduler employs the cgroup freezer as a mechanism for preempting low
priority jobs.  Contrast this with using SIGSTOP for the same purpose:
the stopped tasks do not count toward system load.

Change task_contributes_to_load() to return false if the task is
frozen.  This results in /proc/loadavg behavior that better meets
users' expectations.

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Nigel Cunningham <nigel@tuxonice.net>
Tested-by: Nigel Cunningham <nigel@tuxonice.net>
Cc: <stable@kernel.org>
Cc: containers@lists.linux-foundation.org
Cc: linux-pm@lists.linux-foundation.org
Cc: Matt Helsley <matthltc@us.ibm.com>
LKML-Reference: <20090408194512.47a99b95@manatee.lan>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
include/linux/sched.h

index 98e1fe51601df0066786500ba78e0648ec896287..b4c38bc8049cbbea17e0ca4f929f35df9cddbe1f 100644 (file)
@@ -205,7 +205,8 @@ extern unsigned long long time_sync_thresh;
 #define task_is_stopped_or_traced(task)        \
                        ((task->state & (__TASK_STOPPED | __TASK_TRACED)) != 0)
 #define task_contributes_to_load(task) \
-                               ((task->state & TASK_UNINTERRUPTIBLE) != 0)
+                               ((task->state & TASK_UNINTERRUPTIBLE) != 0 && \
+                                (task->flags & PF_FROZEN) == 0)
 
 #define __set_task_state(tsk, state_value)             \
        do { (tsk)->state = (state_value); } while (0)