[PATCH v11 12/14] cpuidle/poll_state: Wait for need-resched via tif_need_resched_relaxed_wait()

[Okanovic, Haris]

On Wed, 2026-04-08 at 17:55 +0530, Ankur Arora wrote:
> CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.
> 
> 
> 
> The inner loop in poll_idle() polls over the thread_info flags,
> waiting to see if the thread has TIF_NEED_RESCHED set. The loop
> exits once the condition is met, or if the poll time limit has
> been exceeded.
> 
> To minimize the number of instructions executed in each iteration,
> the time check is rate-limited. In addition, each loop iteration
> executes cpu_relax() which on certain platforms provides a hint to
> the pipeline that the loop busy-waits, allowing the processor to
> reduce power consumption.
> 
> Switch over to tif_need_resched_relaxed_wait() instead, since that
> provides exactly that.
> 
> However, since we want to minimize power consumption in idle, building
> of cpuidle/poll_state.c continues to depend on CONFIG_ARCH_HAS_CPU_RELAX
> as that serves as an indicator that the platform supports an optimized
> version of tif_need_resched_relaxed_wait() (via
> smp_cond_load_acquire_timeout()).
> 
> Cc: Rafael J. Wysocki <rafael at kernel.org>
> Cc: Daniel Lezcano <daniel.lezcano at linaro.org>
> Cc: linux-pm at vger.kernel.org
> Suggested-by: Rafael J. Wysocki <rafael at kernel.org>
> Acked-by: Rafael J. Wysocki (Intel) <rafael at kernel.org>
> Signed-off-by: Ankur Arora <ankur.a.arora at oracle.com>
> ---
>  drivers/cpuidle/poll_state.c | 21 +--------------------
>  1 file changed, 1 insertion(+), 20 deletions(-)
> 
> diff --git a/drivers/cpuidle/poll_state.c b/drivers/cpuidle/poll_state.c
> index c7524e4c522a..7443b3e971ba 100644
> --- a/drivers/cpuidle/poll_state.c
> +++ b/drivers/cpuidle/poll_state.c
> @@ -6,41 +6,22 @@
>  #include <linux/cpuidle.h>
>  #include <linux/export.h>
>  #include <linux/irqflags.h>
> -#include <linux/sched.h>
> -#include <linux/sched/clock.h>
>  #include <linux/sched/idle.h>
>  #include <linux/sprintf.h>
>  #include <linux/types.h>
> 
> -#define POLL_IDLE_RELAX_COUNT  200
> -
>  static int __cpuidle poll_idle(struct cpuidle_device *dev,
>                                struct cpuidle_driver *drv, int index)
>  {
> -       u64 time_start;
> -
> -       time_start = local_clock_noinstr();
> -
>         dev->poll_time_limit = false;
> 
>         raw_local_irq_enable();
>         if (!current_set_polling_and_test()) {
> -               unsigned int loop_count = 0;
>                 u64 limit;
> 
>                 limit = cpuidle_poll_time(drv, dev);
> 
> -               while (!need_resched()) {
> -                       cpu_relax();
> -                       if (loop_count++ < POLL_IDLE_RELAX_COUNT)
> -                               continue;
> -
> -                       loop_count = 0;
> -                       if (local_clock_noinstr() - time_start > limit) {
> -                               dev->poll_time_limit = true;
> -                               break;
> -                       }
> -               }
> +               dev->poll_time_limit = !tif_need_resched_relaxed_wait(limit);
>         }
>         raw_local_irq_disable();
> 
> --
> 2.31.1
> 

Hi Ankur,

Tested atop latest mainline d60bc1401 with the rest of your haltpoll
changes from separate thread:
~10% improvement in `perf sched bench pipe` micro and ~4-6% throughput
improvements in mysql,
postgresql, cassandra, and memcached in under-loaded configurations.
Tested on AWS Graviton3 and
Graviton4, ARM Neoverse V1 and V2 cores respectively.

I hope this series can merge soon. It's been stuck in review for more
than 2 years.

Tested-by: Haris Okanovic <harisokn at amazon.com>

Regards,
Haris Okanovic
AWS Graviton Software