linux-pm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Gautham R Shenoy <ego@linux.vnet.ibm.com>
To: Nicholas Piggin <npiggin@gmail.com>
Cc: Gautham R Shenoy <ego@linux.vnet.ibm.com>,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Daniel Lezcano <daniel.lezcano@linaro.org>,
	Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>,
	linux-pm@vger.kernel.org,
	Akshay Adiga <akshay.adiga@linux.vnet.ibm.com>
Subject: Re: [RFC] cpuidle: menu: nearby timer use lightest state; allow state 0 to be disabled
Date: Tue, 6 Jun 2017 15:34:23 +0530	[thread overview]
Message-ID: <20170606100423.GA19040@in.ibm.com> (raw)
In-Reply-To: <20170606111555.2881763e@roar.ozlabs.ibm.com>

On Tue, Jun 06, 2017 at 11:15:55AM +1000, Nicholas Piggin wrote:
> > 
> > Are you able to observe any functional difference with this patch ?
> 
> Yes for some tests, but it did have a bug where state0 was still
> being used despite other states being available. This patch really
> disables state0 (unless all states are disabled, then it falls back
> to 0 again).

Nice. I agree with this approach. We need to document that if the
lower states are disabled, then the menu governor selection logic
would promote to a higher cpuidle state even when the predicted
residency is small. So,folks who are worried about latency shouldn't
be disabling the lower states in the first place.

Acked-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>

> 
> ---
>  drivers/cpuidle/governors/menu.c | 20 +++++++++++++++-----
>  1 file changed, 15 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/cpuidle/governors/menu.c b/drivers/cpuidle/governors/menu.c
> index b2330fd69e34..61b64c2b2cb8 100644
> --- a/drivers/cpuidle/governors/menu.c
> +++ b/drivers/cpuidle/governors/menu.c
> @@ -286,6 +286,8 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev)
>  	struct device *device = get_cpu_device(dev->cpu);
>  	int latency_req = pm_qos_request(PM_QOS_CPU_DMA_LATENCY);
>  	int i;
> +	int first_idx;
> +	int idx;
>  	unsigned int interactivity_req;
>  	unsigned int expected_interval;
>  	unsigned long nr_iowaiters, cpu_load;
> @@ -335,11 +337,11 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev)
>  		if (data->next_timer_us > polling_threshold &&
>  		    latency_req > s->exit_latency && !s->disabled &&
>  		    !dev->states_usage[CPUIDLE_DRIVER_STATE_START].disable)
> -			data->last_state_idx = CPUIDLE_DRIVER_STATE_START;
> +			first_idx = CPUIDLE_DRIVER_STATE_START;
>  		else
> -			data->last_state_idx = CPUIDLE_DRIVER_STATE_START - 1;
> +			first_idx = CPUIDLE_DRIVER_STATE_START - 1;
>  	} else {
> -		data->last_state_idx = CPUIDLE_DRIVER_STATE_START;
> +		first_idx = 0;
>  	}
> 
>  	/*
> @@ -359,20 +361,28 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev)
>  	 * Find the idle state with the lowest power while satisfying
>  	 * our constraints.
>  	 */
> -	for (i = data->last_state_idx + 1; i < drv->state_count; i++) {
> +	idx = -1;
> +	for (i = first_idx; i < drv->state_count; i++) {
>  		struct cpuidle_state *s = &drv->states[i];
>  		struct cpuidle_state_usage *su = &dev->states_usage[i];
> 
>  		if (s->disabled || su->disable)
>  			continue;
> +		if (idx == -1)
> +			idx = i; /* first enabled state */
>  		if (s->target_residency > data->predicted_us)
>  			break;
>  		if (s->exit_latency > latency_req)
>  			break;
> 
> -		data->last_state_idx = i;
> +		idx = i;
>  	}
> 
> +	if (idx == -1)
> +		idx = 0; /* No states enabled. Must use 0. */
> +
> +	data->last_state_idx = idx;
> +
>  	return data->last_state_idx;
>  }
> 
> -- 
> 

      reply	other threads:[~2017-06-06 10:04 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-05-30 16:26 [RFC] cpuidle: menu: nearby timer use lightest state; allow state 0 to be disabled Nicholas Piggin
2017-06-05  7:40 ` Nicholas Piggin
2017-06-05 12:58   ` Rafael J. Wysocki
2017-06-05 15:19 ` Gautham R Shenoy
2017-06-06  1:15   ` Nicholas Piggin
2017-06-06 10:04     ` Gautham R Shenoy [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20170606100423.GA19040@in.ibm.com \
    --to=ego@linux.vnet.ibm.com \
    --cc=akshay.adiga@linux.vnet.ibm.com \
    --cc=daniel.lezcano@linaro.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=npiggin@gmail.com \
    --cc=rjw@rjwysocki.net \
    --cc=svaidy@linux.vnet.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).