Re: [PATCH] CPUidle: increment state counter for state actually entered

[Kevin Hilman <khilman@deeprootsystems.com>]

Kevin Hilman <khilman@deeprootsystems.com> writes:

> Currently, the count for state that the governor chooses is always
> incremented after the 'state_enter' hook is called.  However, the
> target's enter hook may choose a different state based on BM activity
> or other factors.
>
> This patch does the accounting use 'dev->last_state' state instead of
> the governor-chosen state.  If the target's enter_idle hook enters a
> different state than it was asked, it should update dev->last_state.
>

Oops, I sent this before I saw Venkatesh's earlier patch to do the
same.

Kevin

> ---
>  drivers/cpuidle/cpuidle.c |    6 +++---
>  1 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c
> index 5ce07b5..c1294f5 100644
> --- a/drivers/cpuidle/cpuidle.c
> +++ b/drivers/cpuidle/cpuidle.c
> @@ -67,10 +67,10 @@ static void cpuidle_idle_call(void)
>  	target_state = &dev->states[next_state];
>  
>  	/* enter the state and update stats */
> -	dev->last_residency = target_state->enter(dev, target_state);
>  	dev->last_state = target_state;
> -	target_state->time += (unsigned long long)dev->last_residency;
> -	target_state->usage++;
> +	dev->last_residency = target_state->enter(dev, target_state);
> +	dev->last_state->time += (unsigned long long)dev->last_residency;
> +	dev->last_state->usage++;
>  
>  	/* give the governor an opportunity to reflect on the outcome */
>  	if (cpuidle_curr_governor->reflect)
> -- 
> 1.6.0