diff mbox series

[03/20] posix-timers: Clarify timer_wait_running() comment

Message ID 20230425183312.985681995@linutronix.de
State New
Headers
Series posix-timers: Fixes and cleanups |

Commit Message

Thomas Gleixner April 25, 2023, 6:49 p.m. UTC 
  Explain it better and add the CONFIG_POSIX_CPU_TIMERS_TASK_WORK=y aspect
for completeness.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 kernel/time/posix-timers.c |   16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

Comments

Frederic Weisbecker May 9, 2023, 9:50 a.m. UTC | #1 
On Tue, Apr 25, 2023 at 08:49:00PM +0200, Thomas Gleixner wrote:
> Explain it better and add the CONFIG_POSIX_CPU_TIMERS_TASK_WORK=y aspect
> for completeness.
> 
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
diff mbox series

Patch

--- a/kernel/time/posix-timers.c
+++ b/kernel/time/posix-timers.c
@@ -835,10 +835,18 @@  static void common_timer_wait_running(st
 }
 
 /*
- * On PREEMPT_RT this prevent priority inversion against softirq kthread in
- * case it gets preempted while executing a timer callback. See comments in
- * hrtimer_cancel_wait_running. For PREEMPT_RT=n this just results in a
- * cpu_relax().
+ * On PREEMPT_RT this prevents priority inversion and a potential livelock
+ * against the ksoftirqd thread in case that ksoftirqd gets preempted while
+ * executing a hrtimer callback.
+ *
+ * See the comments in hrtimer_cancel_wait_running(). For PREEMPT_RT=n this
+ * just results in a cpu_relax().
+ *
+ * For POSIX CPU timers with CONFIG_POSIX_CPU_TIMERS_TASK_WORK=n this is
+ * just a cpu_relax(). With CONFIG_POSIX_CPU_TIMERS_TASK_WORK=y this
+ * prevents spinning on an eventually scheduled out task and a livelock
+ * when the task which tries to delete or disarm the timer has preempted
+ * the task which runs the expiry in task work context.
  */
 static struct k_itimer *timer_wait_running(struct k_itimer *timer,
 					   unsigned long *flags)