There is a small race between do_fork() and sched_move_task(), which is trying
to move the child.
do_fork() sched_move_task()
--------------------------------+---------------------------------
copy_process()
sched_fork()
task_fork_fair()
-> vruntime of the child is initialized
based on that of the parent.
-> we can see the child in "tasks" file now.
task_rq_lock()
task_move_group_fair()
->child.se.vruntime -= (old)cfs_rq->min_vruntime
->child.se.vruntime += (new)cfs_rq->min_vruntime
task_rq_unlock()
wake_up_new_task()
...
enqueue_entity()
child.se->vruntime += cfs_rq->min_vruntime
As a result, vruntime of the child becomes far bigger than min_vruntime,
if (new)cfs_rq->min_vruntime >> (old)cfs_rq->min_vruntime.
This patch fixes this problem by just ignoring such process in task_move_group_fair(),
because the vruntime has already been normalized in task_fork_fair().
Signed-off-by: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
---
kernel/sched_fair.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index 5c9e679..df145a9 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -4922,10 +4922,10 @@ static void task_move_group_fair(struct task_struct *p, int on_rq)
* to another cgroup's rq. This does somewhat interfere with the
* fair sleeper stuff for the first placement, but who cares.
*/
- if (!on_rq)
+ if (!on_rq && p->state != TASK_RUNNING)
p->se.vruntime -= cfs_rq_of(&p->se)->min_vruntime;
set_task_rq(p, task_cpu(p));
- if (!on_rq)
+ if (!on_rq && p->state != TASK_RUNNING)
p->se.vruntime += cfs_rq_of(&p->se)->min_vruntime;
}
#endif
--
1.7.1
--
To unsubscribe from this list: send the line "unsubscribe cgroups" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
