From: Daniel Lezcano <daniel.lezcano@linaro.org>
To: Nicolas Pitre <nicolas.pitre@linaro.org>
Cc: tglx@linutronix.de, peterz@infradead.org, rafael@kernel.org,
linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org,
vincent.guittot@linaro.org
Subject: Re: [RFC PATCH 2/2] sched: idle: IRQ based next prediction for idle period
Date: Thu, 7 Jan 2016 16:42:58 +0100 [thread overview]
Message-ID: <568E8782.9020401@linaro.org> (raw)
In-Reply-To: <alpine.LFD.2.20.1601061156120.1652@knanqh.ubzr>
On 01/06/2016 06:40 PM, Nicolas Pitre wrote:
> On Wed, 6 Jan 2016, Daniel Lezcano wrote:
>
>> Many IRQs are quiet most of the time, or they tend to come in bursts of
>> fairly equal time intervals within each burst. It is therefore possible
>> to detect those IRQs with stable intervals and guestimate when the next
>> IRQ event is most likely to happen.
>>
>> Examples of such IRQs may include audio related IRQs where the FIFO size
>> and/or DMA descriptor size with the sample rate create stable intervals,
>> block devices during large data transfers, etc. Even network streaming
>> of multimedia content creates patterns of periodic network interface IRQs
>> in some cases.
>>
>> This patch adds code to track the mean interval and variance for each IRQ
>> over a window of time intervals between IRQ events. Those statistics can
>> be used to assist cpuidle in selecting the most appropriate sleep state
>> by predicting the most likely time for the next interrupt.
>>
>> Because the stats are gathered in interrupt context, the core computation
>> is as light as possible.
>>
>> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
>
> There are still a few problems with this patch.
>
> Please see comments below.
[ ... ]
>> +/**
>> + * stats_variance - compute the variance
>> + *
>> + * @s: the statistic structure
>> + *
>> + * Returns an u64 corresponding to the variance, or zero if there is
>> + * no data
>> + */
>> +static u64 stats_variance(struct stats *s, u32 mean)
>> +{
>> + int i;
>> + u64 variance = 0;
>> +
>> + /*
>> + * The variance is the sum of the squared difference to the
>> + * average divided by the number of elements.
>> + */
>> + for (i = 0; i < STATS_NR_VALUES; i++) {
>> + s32 diff = s->values[i] - mean;
>> + variance += (u64)diff * diff;
>
> Strictly speaking, diff should be casted to s64 as it is a signed value
> that may actually be negative. Because of the strange C type promotion
> rules, the generated code appears correct (at least on ARM), but it
> would be clearer to use s64 anyway.
I don't get the connection in your explanation of why it should be a
s64. It is already a signed s32, s->values[] are s32 and mean is u32.
What would be the benefit to convert diff to s64 ?
> The product will end up being positive in all cases of course so
> variance may remain as a u64.
>
[ ... ]
>> +static ktime_t next_irq_event(void)
>> +{
>> + unsigned int irq, cpu = raw_smp_processor_id();
>> + ktime_t diff, next, min = { .tv64 = KTIME_MAX };
>
> To avoid exposing the ktime_t definition, you should use
> ktime_set(KTIME_SEC_MAX, 0) instead of { .tv64 = KTIME_MAX }.
Ok.
>> + struct wakeup *w;
>> + u32 interval, mean;
>> + u64 variance;
>> +
>> + /*
>> + * Lookup the interrupt array for this cpu and search for the
>> + * earlier expected interruption.
>> + */
>> + for (irq = 0; irq < NR_IRQS; irq++) {
>> +
>> + ktime_t now = ktime_get();
>
> ktime_get() is potentially non-trivial. It should be plenty good enough
> to call it only once outside the loop.
Ok.
>> + w = per_cpu(wakeups[irq], cpu);
>> +
>> + /*
>> + * The interrupt was not setup as a source of a wakeup
>> + * or the wakeup source is not considered at this
>> + * moment stable enough to do a prediction.
>> + */
>> + if (!w)
>> + continue;
>> +
>> + /*
>> + * No statistics available yet.
>> + */
>> + if (ktime_equal(w->timestamp, ktime_set(0, 0)))
>> + continue;
>> +
>> + diff = ktime_sub(now, w->timestamp);
>> +
>> + /*
>> + * There is no point attempting predictions on interrupts more
>> + * than 1 second apart. This has no benefit for sleep state
>> + * selection and increases the risk of overflowing our variance
>> + * computation. Reset all stats in that case.
>> + */
>> + if (unlikely(ktime_after(diff, ktime_set(1, 0)))) {
>> + stats_reset(&w->stats);
>> + continue;
>> + }
>
> The above is wrong. It is not computing the interval between successive
> interruts but rather the interval between the last interrupt occurrence
> and the present time (i.e. when we're about to go idle). This won't
> prevent interrupt intervals greater than one second from being summed
> and potentially overflowing the variance if this code is executed less
> than a second after one such IRQ interval. This test should rather be
> performed in sched_idle_irq().
Ok, I will move it.
>> + * If the mean value is null, just ignore this wakeup
>> + * source.
>> + */
>> + mean = stats_mean(&w->stats);
>> + if (!mean)
>> + continue;
>> +
>> + variance = stats_variance(&w->stats, mean);
>> + /*
>> + * We want to check the last interval is:
>> + *
>> + * mean - stddev < interval < mean + stddev
>> + *
>> + * That simplifies to:
>> + *
>> + * -stddev < interval - mean < stddev
>> + *
>> + * abs(interval - mean) < stddev
>> + *
>> + * The standard deviation is the sqrt of the variance:
>> + *
>> + * abs(interval - mean) < sqrt(variance)
>> + *
>> + * and we want to prevent to do an sqrt, so we square
>> + * the equation:
>> + *
>> + * (interval - mean)^2 < variance
>> + *
>> + * So if the latest value of the stats complies with
>> + * this condition, then the wakeup source is
>> + * considered predictable and can be used to predict
>> + * the next event.
>> + */
>> + interval = w->stats.values[(w->stats.w_ptr + 1) % STATS_NR_VALUES];
>
> But here (w->stats.w_ptr + 1) % STATS_NR_VALUES does not point at the
> last interval. It rather points at the second oldest.
> To make things simpler, you should rather pre-increment the pointer
> before updating the array in stats_add(), and here the value you want
> will directly be accessible with w->stats.values[w->stats.w_ptr].
Yeah, there is a glitch here. I will look at it.
>> + if ((u64)((interval - mean) * (interval - mean)) > variance)
>> + continue;
>
> Same comment as in stats_variance(): this should be s64.
>
>> + /*
>> + * Let's compute the next event: the wakeup source is
>> + * considered predictable, we add the average interval
>> + * time added to the latest interruption event time.
>> + */
>> + next = ktime_add_us(w->timestamp, stats_mean(&w->stats));
>> +
>> + /*
>> + * If the interrupt is supposed to happen before the
>> + * minimum time, then it becomes the minimum.
>> + */
>> + if (ktime_before(next, min))
>> + min = next;
>> + }
>> +
>> + /*
>> + * At this point, we have our prediction but the caller is
>> + * expecting the remaining time before the next event, so
>> + * compute the expected sleep length.
>> + */
>> + diff = ktime_sub(min, ktime_get());
>
> You should use the variable 'now' rather than asking for the current
> time again.
Yep.
>> + /*
>> + * The result could be negative for different reasons:
>> + * - the prediction is incorrect
>> + * - the prediction was too near now and expired while we were
>> + * in this function
>> + *
>> + * In both cases, we return KTIME_MAX as a failure to do a
>> + * prediction
>> + */
>> + if (ktime_compare(diff, ktime_set(0, 0)) <= 0)
>> + return (ktime_t){ .tv64 = KTIME_MAX };
>
> See my comment about ktime_t internals at the beginning of this
> function.
Ok.
Thanks for the review.
-- Daniel
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
next prev parent reply other threads:[~2016-01-07 15:43 UTC|newest]
Thread overview: 63+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-01-06 15:22 [RFC PATCH 0/2] IRQ based next prediction Daniel Lezcano
2016-01-06 15:22 ` [RFC PATCH 1/2] irq: Add a framework to measure interrupt timings Daniel Lezcano
2016-01-08 15:31 ` Thomas Gleixner
2016-01-12 11:42 ` Daniel Lezcano
2016-01-06 15:22 ` [RFC PATCH 2/2] sched: idle: IRQ based next prediction for idle period Daniel Lezcano
2016-01-06 17:40 ` Nicolas Pitre
2016-01-07 15:42 ` Daniel Lezcano [this message]
2016-01-12 19:27 ` Nicolas Pitre
2016-01-10 22:37 ` Daniel Lezcano
2016-01-10 22:46 ` Nicolas Pitre
2016-01-10 22:58 ` Daniel Lezcano
2016-01-10 23:13 ` Nicolas Pitre
2016-01-08 15:43 ` Thomas Gleixner
2016-01-12 12:41 ` Daniel Lezcano
2016-01-12 13:42 ` Thomas Gleixner
2016-01-12 14:16 ` Daniel Lezcano
2016-01-12 14:26 ` Thomas Gleixner
2016-01-12 14:52 ` Daniel Lezcano
2016-01-12 15:12 ` Thomas Gleixner
2016-01-12 16:04 ` Daniel Lezcano
2016-01-13 9:17 ` Thomas Gleixner
2016-01-18 13:21 ` Daniel Lezcano
2016-01-20 15:41 ` Thomas Gleixner
2016-01-20 16:00 ` [RFC V2 0/2] IRQ based next prediction Daniel Lezcano
2016-01-20 16:00 ` [RFC V2 1/2] irq: Add a framework to measure interrupt timings Daniel Lezcano
2016-01-20 17:55 ` Thomas Gleixner
2016-01-21 9:25 ` Daniel Lezcano
2016-01-21 10:27 ` Thomas Gleixner
2016-01-20 19:07 ` Peter Zijlstra
2016-01-20 19:57 ` Thomas Gleixner
2016-01-20 20:04 ` Nicolas Pitre
2016-01-20 20:20 ` Peter Zijlstra
2016-01-20 20:22 ` Thomas Gleixner
2016-01-21 9:50 ` Daniel Lezcano
2016-01-21 10:08 ` Peter Zijlstra
2016-01-21 12:38 ` Daniel Lezcano
2016-01-21 20:27 ` Thomas Gleixner
2016-01-21 13:52 ` Thomas Gleixner
2016-01-21 14:19 ` Daniel Lezcano
2016-01-21 18:56 ` Thomas Gleixner
2016-01-22 10:15 ` Peter Zijlstra
2016-01-21 9:26 ` Daniel Lezcano
2016-01-20 19:28 ` Peter Zijlstra
2016-01-21 9:53 ` Daniel Lezcano
2016-01-20 16:00 ` [RFC V2 2/2] sched: idle: IRQ based next prediction for idle period Daniel Lezcano
2016-01-20 17:46 ` Nicolas Pitre
2016-01-20 18:44 ` Peter Zijlstra
2016-01-21 10:03 ` Daniel Lezcano
2016-01-20 19:02 ` Peter Zijlstra
2016-01-20 19:17 ` Nicolas Pitre
2016-01-20 19:29 ` Peter Zijlstra
2016-01-20 19:34 ` Peter Zijlstra
2016-01-20 19:40 ` Peter Zijlstra
2016-01-20 19:57 ` Nicolas Pitre
2016-01-20 20:22 ` Peter Zijlstra
2016-01-20 19:49 ` Thomas Gleixner
2016-01-21 13:54 ` Daniel Lezcano
2016-01-21 14:12 ` Thomas Gleixner
2016-01-20 16:00 ` [RFC V2 0/2] IRQ based next prediction Daniel Lezcano
2016-01-20 16:00 ` [RFC V2 1/2] irq: Add a framework to measure interrupt timings Daniel Lezcano
2016-01-20 16:00 ` [RFC V2 2/2] sched: idle: IRQ based next prediction for idle period Daniel Lezcano
2016-01-20 20:14 ` Nicolas Pitre
2016-01-21 13:04 ` Daniel Lezcano
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=568E8782.9020401@linaro.org \
--to=daniel.lezcano@linaro.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=nicolas.pitre@linaro.org \
--cc=peterz@infradead.org \
--cc=rafael@kernel.org \
--cc=tglx@linutronix.de \
--cc=vincent.guittot@linaro.org \
/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
Powered by JetHome