diff mbox series

[v2,2/2] cpuidle: psci: Allow PM domain to be initialized even if no OSI mode

Message ID 20200901121226.217568-3-ulf.hansson@linaro.org
State Superseded
Headers show
Series cpuidle: psci: Always create the PM domains | expand

Commit Message

Ulf Hansson Sept. 1, 2020, 12:12 p.m. UTC 
If the PSCI OSI mode isn't supported or fails to be enabled, the PM domain
topology with the genpd providers isn't initialized. This is perfectly fine
from cpuidle-psci point of view.

However, since the PM domain topology in the DTS files is a description of
the HW, no matter of whether the PSCI OSI mode is supported or not, other
consumers besides the CPUs may rely on it.

Therefore, let's always allow the initialization of the PM domain topology
to succeed, independently of whether the PSCI OSI mode is supported.
Consequentially we need to track if we succeed to enable the OSI mode, as
to know when a domain idlestate can be selected.

Note that, CPU devices are still not being attached to the PM domain
topology, unless the PSCI OSI mode is supported.

Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>

---

Changes in v2:
	- Assign the genpd ->power_off() callback, only when the PSCI OSI mode
	has been successfully enabled.

---
 drivers/cpuidle/cpuidle-psci-domain.c | 57 ++++++++++++++-------------
 1 file changed, 29 insertions(+), 28 deletions(-)

-- 
2.25.1

Comments

Sudeep Holla Sept. 1, 2020, 1:32 p.m. UTC | #1 
On Tue, Sep 01, 2020 at 02:12:26PM +0200, Ulf Hansson wrote:
> If the PSCI OSI mode isn't supported or fails to be enabled, the PM domain

> topology with the genpd providers isn't initialized. This is perfectly fine

> from cpuidle-psci point of view.

>

> However, since the PM domain topology in the DTS files is a description of

> the HW, no matter of whether the PSCI OSI mode is supported or not, other

> consumers besides the CPUs may rely on it.

>

> Therefore, let's always allow the initialization of the PM domain topology

> to succeed, independently of whether the PSCI OSI mode is supported.

> Consequentially we need to track if we succeed to enable the OSI mode, as

> to know when a domain idlestate can be selected.

>

> Note that, CPU devices are still not being attached to the PM domain

> topology, unless the PSCI OSI mode is supported.

>

> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>

> ---

>

> Changes in v2:

> 	- Assign the genpd ->power_off() callback, only when the PSCI OSI mode

> 	has been successfully enabled.

>

> ---

>  drivers/cpuidle/cpuidle-psci-domain.c | 57 ++++++++++++++-------------

>  1 file changed, 29 insertions(+), 28 deletions(-)

>

> diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c

> index b6ab0415f450..256e7e35b5af 100644

> --- a/drivers/cpuidle/cpuidle-psci-domain.c

> +++ b/drivers/cpuidle/cpuidle-psci-domain.c

> @@ -105,7 +105,7 @@ static void psci_pd_free_states(struct genpd_power_state *states,

>  	kfree(states);

>  }

>

> -static int psci_pd_init(struct device_node *np)

> +static int psci_pd_init(struct device_node *np, bool use_osi)

>  {

>  	struct generic_pm_domain *pd;

>  	struct psci_pd_provider *pd_provider;

> @@ -135,11 +135,14 @@ static int psci_pd_init(struct device_node *np)

>

>  	pd->free_states = psci_pd_free_states;

>  	pd->name = kbasename(pd->name);

> -	pd->power_off = psci_pd_power_off;

>  	pd->states = states;

>  	pd->state_count = state_count;

>  	pd->flags |= GENPD_FLAG_IRQ_SAFE | GENPD_FLAG_CPU_DOMAIN;

>

> +	/* Use the ->power_off() callback when OSI is enabled. */


IIUC, you did mention that we need to attach PD even when the OSI fails
as this could be shared domain. With the below conditional assignment
of power_off, I understand that if OSI fails and we have domains attached,
then it will be always ON ?


Other than that, this patch looks good.

Acked-by: Sudeep Holla <sudeep.holla@arm.com>


--
Regards,
Sudeep
Ulf Hansson Sept. 1, 2020, 1:49 p.m. UTC | #2 
On Tue, 1 Sep 2020 at 15:32, Sudeep Holla <sudeep.holla@arm.com> wrote:
>

> On Tue, Sep 01, 2020 at 02:12:26PM +0200, Ulf Hansson wrote:

> > If the PSCI OSI mode isn't supported or fails to be enabled, the PM domain

> > topology with the genpd providers isn't initialized. This is perfectly fine

> > from cpuidle-psci point of view.

> >

> > However, since the PM domain topology in the DTS files is a description of

> > the HW, no matter of whether the PSCI OSI mode is supported or not, other

> > consumers besides the CPUs may rely on it.

> >

> > Therefore, let's always allow the initialization of the PM domain topology

> > to succeed, independently of whether the PSCI OSI mode is supported.

> > Consequentially we need to track if we succeed to enable the OSI mode, as

> > to know when a domain idlestate can be selected.

> >

> > Note that, CPU devices are still not being attached to the PM domain

> > topology, unless the PSCI OSI mode is supported.

> >

> > Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>

> > ---

> >

> > Changes in v2:

> >       - Assign the genpd ->power_off() callback, only when the PSCI OSI mode

> >       has been successfully enabled.

> >

> > ---

> >  drivers/cpuidle/cpuidle-psci-domain.c | 57 ++++++++++++++-------------

> >  1 file changed, 29 insertions(+), 28 deletions(-)

> >

> > diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c

> > index b6ab0415f450..256e7e35b5af 100644

> > --- a/drivers/cpuidle/cpuidle-psci-domain.c

> > +++ b/drivers/cpuidle/cpuidle-psci-domain.c

> > @@ -105,7 +105,7 @@ static void psci_pd_free_states(struct genpd_power_state *states,

> >       kfree(states);

> >  }

> >

> > -static int psci_pd_init(struct device_node *np)

> > +static int psci_pd_init(struct device_node *np, bool use_osi)

> >  {

> >       struct generic_pm_domain *pd;

> >       struct psci_pd_provider *pd_provider;

> > @@ -135,11 +135,14 @@ static int psci_pd_init(struct device_node *np)

> >

> >       pd->free_states = psci_pd_free_states;

> >       pd->name = kbasename(pd->name);

> > -     pd->power_off = psci_pd_power_off;

> >       pd->states = states;

> >       pd->state_count = state_count;

> >       pd->flags |= GENPD_FLAG_IRQ_SAFE | GENPD_FLAG_CPU_DOMAIN;

> >

> > +     /* Use the ->power_off() callback when OSI is enabled. */

>

> IIUC, you did mention that we need to attach PD even when the OSI fails

> as this could be shared domain. With the below conditional assignment

> of power_off, I understand that if OSI fails and we have domains attached,

> then it will be always ON ?


From a genpd point of view, the corresponding PM domain can be powered
off (as long as all attached consumers allow that as well, of course).

Hmm, maybe we should add a GENPD_FLAG_ALWAYS_ON, to really enforce
that it stays on?

>

>

> Other than that, this patch looks good.

>

> Acked-by: Sudeep Holla <sudeep.holla@arm.com>


Thanks!

Kind regards
Uffe
Sudeep Holla Sept. 1, 2020, 2:11 p.m. UTC | #3 
On Tue, Sep 01, 2020 at 03:49:55PM +0200, Ulf Hansson wrote:
> On Tue, 1 Sep 2020 at 15:32, Sudeep Holla <sudeep.holla@arm.com> wrote:

> >

> > On Tue, Sep 01, 2020 at 02:12:26PM +0200, Ulf Hansson wrote:

> > > If the PSCI OSI mode isn't supported or fails to be enabled, the PM domain

> > > topology with the genpd providers isn't initialized. This is perfectly fine

> > > from cpuidle-psci point of view.

> > >

> > > However, since the PM domain topology in the DTS files is a description of

> > > the HW, no matter of whether the PSCI OSI mode is supported or not, other

> > > consumers besides the CPUs may rely on it.

> > >

> > > Therefore, let's always allow the initialization of the PM domain topology

> > > to succeed, independently of whether the PSCI OSI mode is supported.

> > > Consequentially we need to track if we succeed to enable the OSI mode, as

> > > to know when a domain idlestate can be selected.

> > >

> > > Note that, CPU devices are still not being attached to the PM domain

> > > topology, unless the PSCI OSI mode is supported.

> > >

> > > Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>

> > > ---

> > >

> > > Changes in v2:

> > >       - Assign the genpd ->power_off() callback, only when the PSCI OSI mode

> > >       has been successfully enabled.

> > >

> > > ---

> > >  drivers/cpuidle/cpuidle-psci-domain.c | 57 ++++++++++++++-------------

> > >  1 file changed, 29 insertions(+), 28 deletions(-)

> > >

> > > diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c

> > > index b6ab0415f450..256e7e35b5af 100644

> > > --- a/drivers/cpuidle/cpuidle-psci-domain.c

> > > +++ b/drivers/cpuidle/cpuidle-psci-domain.c

> > > @@ -105,7 +105,7 @@ static void psci_pd_free_states(struct genpd_power_state *states,

> > >       kfree(states);

> > >  }

> > >

> > > -static int psci_pd_init(struct device_node *np)

> > > +static int psci_pd_init(struct device_node *np, bool use_osi)

> > >  {

> > >       struct generic_pm_domain *pd;

> > >       struct psci_pd_provider *pd_provider;

> > > @@ -135,11 +135,14 @@ static int psci_pd_init(struct device_node *np)

> > >

> > >       pd->free_states = psci_pd_free_states;

> > >       pd->name = kbasename(pd->name);

> > > -     pd->power_off = psci_pd_power_off;

> > >       pd->states = states;

> > >       pd->state_count = state_count;

> > >       pd->flags |= GENPD_FLAG_IRQ_SAFE | GENPD_FLAG_CPU_DOMAIN;

> > >

> > > +     /* Use the ->power_off() callback when OSI is enabled. */

> >

> > IIUC, you did mention that we need to attach PD even when the OSI fails

> > as this could be shared domain. With the below conditional assignment

> > of power_off, I understand that if OSI fails and we have domains attached,

> > then it will be always ON ?

>

> From a genpd point of view, the corresponding PM domain can be powered

> off (as long as all attached consumers allow that as well, of course).

>


Can be powered off ? Or will be powered off ? It sounds wrong if it will
be powered off as the domain needs to be powered up for CPU to be running.
It sounds like we may be out of sync though not a big issue, it sounds
weird for me.

> Hmm, maybe we should add a GENPD_FLAG_ALWAYS_ON, to really enforce

> that it stays on?

>


Should that not be default for CPU power domains if OSI can't be enabled ?

--
Regards,
Sudeep
Ulf Hansson Sept. 1, 2020, 2:23 p.m. UTC | #4 
On Tue, 1 Sep 2020 at 16:11, Sudeep Holla <sudeep.holla@arm.com> wrote:
>

> On Tue, Sep 01, 2020 at 03:49:55PM +0200, Ulf Hansson wrote:

> > On Tue, 1 Sep 2020 at 15:32, Sudeep Holla <sudeep.holla@arm.com> wrote:

> > >

> > > On Tue, Sep 01, 2020 at 02:12:26PM +0200, Ulf Hansson wrote:

> > > > If the PSCI OSI mode isn't supported or fails to be enabled, the PM domain

> > > > topology with the genpd providers isn't initialized. This is perfectly fine

> > > > from cpuidle-psci point of view.

> > > >

> > > > However, since the PM domain topology in the DTS files is a description of

> > > > the HW, no matter of whether the PSCI OSI mode is supported or not, other

> > > > consumers besides the CPUs may rely on it.

> > > >

> > > > Therefore, let's always allow the initialization of the PM domain topology

> > > > to succeed, independently of whether the PSCI OSI mode is supported.

> > > > Consequentially we need to track if we succeed to enable the OSI mode, as

> > > > to know when a domain idlestate can be selected.

> > > >

> > > > Note that, CPU devices are still not being attached to the PM domain

> > > > topology, unless the PSCI OSI mode is supported.

> > > >

> > > > Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>

> > > > ---

> > > >

> > > > Changes in v2:

> > > >       - Assign the genpd ->power_off() callback, only when the PSCI OSI mode

> > > >       has been successfully enabled.

> > > >

> > > > ---

> > > >  drivers/cpuidle/cpuidle-psci-domain.c | 57 ++++++++++++++-------------

> > > >  1 file changed, 29 insertions(+), 28 deletions(-)

> > > >

> > > > diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c

> > > > index b6ab0415f450..256e7e35b5af 100644

> > > > --- a/drivers/cpuidle/cpuidle-psci-domain.c

> > > > +++ b/drivers/cpuidle/cpuidle-psci-domain.c

> > > > @@ -105,7 +105,7 @@ static void psci_pd_free_states(struct genpd_power_state *states,

> > > >       kfree(states);

> > > >  }

> > > >

> > > > -static int psci_pd_init(struct device_node *np)

> > > > +static int psci_pd_init(struct device_node *np, bool use_osi)

> > > >  {

> > > >       struct generic_pm_domain *pd;

> > > >       struct psci_pd_provider *pd_provider;

> > > > @@ -135,11 +135,14 @@ static int psci_pd_init(struct device_node *np)

> > > >

> > > >       pd->free_states = psci_pd_free_states;

> > > >       pd->name = kbasename(pd->name);

> > > > -     pd->power_off = psci_pd_power_off;

> > > >       pd->states = states;

> > > >       pd->state_count = state_count;

> > > >       pd->flags |= GENPD_FLAG_IRQ_SAFE | GENPD_FLAG_CPU_DOMAIN;

> > > >

> > > > +     /* Use the ->power_off() callback when OSI is enabled. */

> > >

> > > IIUC, you did mention that we need to attach PD even when the OSI fails

> > > as this could be shared domain. With the below conditional assignment

> > > of power_off, I understand that if OSI fails and we have domains attached,

> > > then it will be always ON ?

> >

> > From a genpd point of view, the corresponding PM domain can be powered

> > off (as long as all attached consumers allow that as well, of course).

> >

>

> Can be powered off ? Or will be powered off ? It sounds wrong if it will

> be powered off as the domain needs to be powered up for CPU to be running.

> It sounds like we may be out of sync though not a big issue, it sounds

> weird for me.


You are right, it's a bit weird.

>

> > Hmm, maybe we should add a GENPD_FLAG_ALWAYS_ON, to really enforce

> > that it stays on?

> >

>

> Should that not be default for CPU power domains if OSI can't be enabled ?


Yes, that makes perfect sense to me as well. Let me respin and fix it,
I will also add your acks/reviewed-by.

Kind regards
Uffe
diff mbox series

Patch

diff --git a/drivers/cpuidle/cpuidle-psci-domain.c b/drivers/cpuidle/cpuidle-psci-domain.c
index b6ab0415f450..256e7e35b5af 100644
--- a/drivers/cpuidle/cpuidle-psci-domain.c
+++ b/drivers/cpuidle/cpuidle-psci-domain.c
@@ -105,7 +105,7 @@  static void psci_pd_free_states(struct genpd_power_state *states,
 	kfree(states);
 }
 
-static int psci_pd_init(struct device_node *np)
+static int psci_pd_init(struct device_node *np, bool use_osi)
 {
 	struct generic_pm_domain *pd;
 	struct psci_pd_provider *pd_provider;
@@ -135,11 +135,14 @@  static int psci_pd_init(struct device_node *np)
 
 	pd->free_states = psci_pd_free_states;
 	pd->name = kbasename(pd->name);
-	pd->power_off = psci_pd_power_off;
 	pd->states = states;
 	pd->state_count = state_count;
 	pd->flags |= GENPD_FLAG_IRQ_SAFE | GENPD_FLAG_CPU_DOMAIN;
 
+	/* Use the ->power_off() callback when OSI is enabled. */
+	if (use_osi)
+		pd->power_off = psci_pd_power_off;
+
 	/* Use governor for CPU PM domains if it has some states to manage. */
 	pd_gov = state_count > 0 ? &pm_domain_cpu_gov : NULL;
 
@@ -190,7 +193,7 @@  static void psci_pd_remove(void)
 	}
 }
 
-static int psci_pd_init_topology(struct device_node *np, bool add)
+static int psci_pd_init_topology(struct device_node *np)
 {
 	struct device_node *node;
 	struct of_phandle_args child, parent;
@@ -203,9 +206,7 @@  static int psci_pd_init_topology(struct device_node *np, bool add)
 
 		child.np = node;
 		child.args_count = 0;
-
-		ret = add ? of_genpd_add_subdomain(&parent, &child) :
-			of_genpd_remove_subdomain(&parent, &child);
+		ret = of_genpd_add_subdomain(&parent, &child);
 		of_node_put(parent.np);
 		if (ret) {
 			of_node_put(node);
@@ -216,14 +217,20 @@  static int psci_pd_init_topology(struct device_node *np, bool add)
 	return 0;
 }
 
-static int psci_pd_add_topology(struct device_node *np)
+static bool psci_pd_try_set_osi_mode(void)
 {
-	return psci_pd_init_topology(np, true);
-}
+	int ret;
 
-static void psci_pd_remove_topology(struct device_node *np)
-{
-	psci_pd_init_topology(np, false);
+	if (!psci_has_osi_support())
+		return false;
+
+	ret = psci_set_osi_mode(true);
+	if (ret) {
+		pr_warn("failed to enable OSI mode: %d\n", ret);
+		return false;
+	}
+
+	return true;
 }
 
 static void psci_cpuidle_domain_sync_state(struct device *dev)
@@ -244,14 +251,14 @@  static int psci_cpuidle_domain_probe(struct platform_device *pdev)
 {
 	struct device_node *np = pdev->dev.of_node;
 	struct device_node *node;
+	bool use_osi;
 	int ret = 0, pd_count = 0;
 
 	if (!np)
 		return -ENODEV;
 
-	/* Currently limit the hierarchical topology to be used in OSI mode. */
-	if (!psci_has_osi_support())
-		return 0;
+	/* If OSI mode is supported, let's try to enable it. */
+	use_osi = psci_pd_try_set_osi_mode();
 
 	/*
 	 * Parse child nodes for the "#power-domain-cells" property and
@@ -261,7 +268,7 @@  static int psci_cpuidle_domain_probe(struct platform_device *pdev)
 		if (!of_find_property(node, "#power-domain-cells", NULL))
 			continue;
 
-		ret = psci_pd_init(node);
+		ret = psci_pd_init(node, use_osi);
 		if (ret)
 			goto put_node;
 
@@ -270,30 +277,24 @@  static int psci_cpuidle_domain_probe(struct platform_device *pdev)
 
 	/* Bail out if not using the hierarchical CPU topology. */
 	if (!pd_count)
-		return 0;
+		goto no_pd;
 
 	/* Link genpd masters/subdomains to model the CPU topology. */
-	ret = psci_pd_add_topology(np);
+	ret = psci_pd_init_topology(np);
 	if (ret)
 		goto remove_pd;
 
-	/* Try to enable OSI mode. */
-	ret = psci_set_osi_mode(true);
-	if (ret) {
-		pr_warn("failed to enable OSI mode: %d\n", ret);
-		psci_pd_remove_topology(np);
-		goto remove_pd;
-	}
-
 	pr_info("Initialized CPU PM domain topology\n");
 	return 0;
 
 put_node:
 	of_node_put(node);
 remove_pd:
-	if (pd_count)
-		psci_pd_remove();
+	psci_pd_remove();
 	pr_err("failed to create CPU PM domains ret=%d\n", ret);
+no_pd:
+	if (use_osi)
+		psci_set_osi_mode(false);
 	return ret;
 }