Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit fa18f7bd authored by KOSAKI Motohiro's avatar KOSAKI Motohiro Committed by Frederic Weisbecker
Browse files

posix-cpu-timers: don't account cpu timer after stopped thread runtime accounting 



When tsk->signal->cputimer->running is 1, signal->cputimer (i.e. per process
timer account) and tsk->sum_sched_runtime (i.e. per thread timer account)
increase at the same pace because update_curr() increases both accounting.

However, there is one exception. When thread exiting, __exit_signal() turns
over task's sum_shced_runtime to sig->sum_sched_runtime, but it doesn't stop
signal->cputimer accounting.

This inconsistency makes POSIX timer wake up too early. This patch fixes it.

Original-patch-by: default avatarOlivier Langlois <olivier@trillion01.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Acked-by: default avatarPeter Zijlstra <peterz@infradead.org>
Signed-off-by: default avatarOlivier Langlois <olivier@trillion01.com>
Signed-off-by: default avatarKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: default avatarFrederic Weisbecker <fweisbec@gmail.com>
parent a0b2062b

kernel/sched/stats.h

+36 −3
Original line number Diff line number Diff line
@@ -161,6 +161,39 @@ sched_info_switch(struct task_struct *prev, struct task_struct *next) 
 * on CONFIG_SCHEDSTATS.

 */



/**

 * cputimer_running - return true if cputimer is running

 *

 * @tsk:	Pointer to target task.

 */

static inline bool cputimer_running(struct task_struct *tsk)



{

	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;



	if (!cputimer->running)

		return false;



	/*

	 * After we flush the task's sum_exec_runtime to sig->sum_sched_runtime

	 * in __exit_signal(), we won't account to the signal struct further

	 * cputime consumed by that task, even though the task can still be

	 * ticking after __exit_signal().

	 *

	 * In order to keep a consistent behaviour between thread group cputime

	 * and thread group cputimer accounting, lets also ignore the cputime

	 * elapsing after __exit_signal() in any thread group timer running.

	 *

	 * This makes sure that POSIX CPU clocks and timers are synchronized, so

	 * that a POSIX CPU timer won't expire while the corresponding POSIX CPU

	 * clock delta is behind the expiring timer value.

	 */

	if (unlikely(!tsk->sighand))

		return false;



	return true;

}



/**

 * account_group_user_time - Maintain utime for a thread group.

 *
@@ -176,7 +209,7 @@ static inline void account_group_user_time(struct task_struct *tsk, 
{

	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;



	if (!cputimer->running)

	if (!cputimer_running(tsk))

		return;



	raw_spin_lock(&cputimer->lock);
@@ -199,7 +232,7 @@ static inline void account_group_system_time(struct task_struct *tsk, 
{

	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;



	if (!cputimer->running)

	if (!cputimer_running(tsk))

		return;



	raw_spin_lock(&cputimer->lock);
@@ -222,7 +255,7 @@ static inline void account_group_exec_runtime(struct task_struct *tsk, 
{

	struct thread_group_cputimer *cputimer = &tsk->signal->cputimer;



	if (!cputimer->running)

	if (!cputimer_running(tsk))

		return;



	raw_spin_lock(&cputimer->lock);