Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Wed, 19 Jun 2019 14:23:03 +1000
From:   Nicholas Piggin <npiggin@gmail.com>
Subject: Re: [PATCH v2 1/1] cpuidle-powernv : forced wakeup for stop states
To:     Abhishek Goel <huntbag@linux.vnet.ibm.com>,
        linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org,
        linuxppc-dev@lists.ozlabs.org
Cc:     daniel.lezcano@linaro.org, dja@axtens.net, ego@linux.vnet.ibm.com,
        mpe@ellerman.id.au, rjw@rjwysocki.net
References: <20190617095648.18847-1-huntbag@linux.vnet.ibm.com>
        <20190617095648.18847-2-huntbag@linux.vnet.ibm.com>
In-Reply-To: <20190617095648.18847-2-huntbag@linux.vnet.ibm.com>
MIME-Version: 1.0
User-Agent: astroid/0.14.0 (https://github.com/astroidmail/astroid)
Message-Id: <1560917320.mk5nn6r8jw.astroid@bobo.none>
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Abhishek Goel's on June 17, 2019 7:56 pm:
> Currently, the cpuidle governors determine what idle state a idling CPU
> should enter into based on heuristics that depend on the idle history on
> that CPU. Given that no predictive heuristic is perfect, there are cases
> where the governor predicts a shallow idle state, hoping that the CPU wil=
l
> be busy soon. However, if no new workload is scheduled on that CPU in the
> near future, the CPU may end up in the shallow state.
>=20
> This is problematic, when the predicted state in the aforementioned
> scenario is a shallow stop state on a tickless system. As we might get
> stuck into shallow states for hours, in absence of ticks or interrupts.
>=20
> To address this, We forcefully wakeup the cpu by setting the
> decrementer. The decrementer is set to a value that corresponds with the
> residency of the next available state. Thus firing up a timer that will
> forcefully wakeup the cpu. Few such iterations will essentially train the
> governor to select a deeper state for that cpu, as the timer here
> corresponds to the next available cpuidle state residency. Thus, cpu will
> eventually end up in the deepest possible state.
>=20
> Signed-off-by: Abhishek Goel <huntbag@linux.vnet.ibm.com>
> ---
>=20
> Auto-promotion
>  v1 : started as auto promotion logic for cpuidle states in generic
> driver
>  v2 : Removed timeout_needed and rebased the code to upstream kernel
> Forced-wakeup
>  v1 : New patch with name of forced wakeup started
>  v2 : Extending the forced wakeup logic for all states. Setting the
> decrementer instead of queuing up a hrtimer to implement the logic.
>=20
>  drivers/cpuidle/cpuidle-powernv.c | 38 +++++++++++++++++++++++++++++++
>  1 file changed, 38 insertions(+)
>=20
> diff --git a/drivers/cpuidle/cpuidle-powernv.c b/drivers/cpuidle/cpuidle-=
powernv.c
> index 84b1ebe212b3..bc9ca18ae7e3 100644
> --- a/drivers/cpuidle/cpuidle-powernv.c
> +++ b/drivers/cpuidle/cpuidle-powernv.c
> @@ -46,6 +46,26 @@ static struct stop_psscr_table stop_psscr_table[CPUIDL=
E_STATE_MAX] __read_mostly
>  static u64 default_snooze_timeout __read_mostly;
>  static bool snooze_timeout_en __read_mostly;
> =20
> +static u64 forced_wakeup_timeout(struct cpuidle_device *dev,
> +				 struct cpuidle_driver *drv,
> +				 int index)
> +{
> +	int i;
> +
> +	for (i =3D index + 1; i < drv->state_count; i++) {
> +		struct cpuidle_state *s =3D &drv->states[i];
> +		struct cpuidle_state_usage *su =3D &dev->states_usage[i];
> +
> +		if (s->disabled || su->disable)
> +			continue;
> +
> +		return (s->target_residency + 2 * s->exit_latency) *
> +			tb_ticks_per_usec;
> +	}
> +
> +	return 0;
> +}

It would be nice to not have this kind of loop iteration in the
idle fast path. Can we add a flag or something to the idle state?

> +
>  static u64 get_snooze_timeout(struct cpuidle_device *dev,
>  			      struct cpuidle_driver *drv,
>  			      int index)
> @@ -144,8 +164,26 @@ static int stop_loop(struct cpuidle_device *dev,
>  		     struct cpuidle_driver *drv,
>  		     int index)
>  {
> +	u64 dec_expiry_tb, dec, timeout_tb, forced_wakeup;
> +
> +	dec =3D mfspr(SPRN_DEC);
> +	timeout_tb =3D forced_wakeup_timeout(dev, drv, index);
> +	forced_wakeup =3D 0;
> +
> +	if (timeout_tb && timeout_tb < dec) {
> +		forced_wakeup =3D 1;
> +		dec_expiry_tb =3D mftb() + dec;
> +	}

The compiler probably can't optimise away the SPR manipulations so try
to avoid them if possible.

> +
> +	if (forced_wakeup)
> +		mtspr(SPRN_DEC, timeout_tb);

This should just be put in the above 'if'.

> +
>  	power9_idle_type(stop_psscr_table[index].val,
>  			 stop_psscr_table[index].mask);
> +
> +	if (forced_wakeup)
> +		mtspr(SPRN_DEC, dec_expiry_tb - mftb());

This will sometimes go negative and result in another timer interrupt.

It also breaks irq work (which can be set here by machine check I
believe.

May need to implement some timer code to do this for you.

static void reset_dec_after_idle(void)
{
	u64 now;
        u64 *next_tb;

	if (test_irq_work_pending())
		return;
	now =3D mftb;
	next_tb =3D this_cpu_ptr(&decrementers_next_tb);

	if (now >=3D *next_tb)
		return;
	set_dec(*next_tb - now);
	if (test_irq_work_pending())
		set_dec(1);
}

Something vaguely like that. See timer_interrupt().

Thanks,
Nick
=