[PATCH v5 0/3] power: domain: handle power supplies that need interrupts

[PATCH v5 0/3] power: domain: handle power supplies that need interrupts

[Martin Kepplinger]

hi Ulf, Lucas and all interested,

This (after a cleanup patch) makes available a new genpd flag
GENPD_FLAG_IRQ_ON in a relatively generic way: genpd providers can set
it when irqs are needed to manage power on/off. Since the main goal
here has been to fix systemd suspend/resume, adjusting these callbacks
is all that's being done when this flag gets set.

And since I'm working on imx8mq, the 3rd patch makes gpcv2 set this new
flag when a power domain has a power-supply descirbed in DT.
For i.MX8M* platforms, this should be ok. For other platforms this might
be useful too but needs to be tested.


revision history
----------------
v5: (thank you Lucas)
* simplify gpcv2 code: just set GENPD_FLAG_IRQ_ON when a power-supply is present

v4: (thank you Ulf and Lucas)
* split up genpd core and gpcv2 changes
* set callbacks inside of pm_genpd_init()
* make flag name and description a bit more generic
* print an error in __genpd_dev_pm_attach() if there a "mismatch"
https://lore.kernel.org/linux-arm-kernel/20220720043444.1289952-1-martin.kepplinger@puri.sm/T/#t

v3: (thank you Ulf)
* move DT parsing to gpcv2 and create a genpd flag that gets set
https://lore.kernel.org/linux-arm-kernel/20220718210302.674897-1-martin.kepplinger@puri.sm/

v2: (thank you Krzysztof)
* rewrite: find possible regulators' interrupts property in parents
  instead of inventing a new property.
https://lore.kernel.org/linux-arm-kernel/20220712121832.3659769-1-martin.kepplinger@puri.sm/

v1: (initial idea)
https://lore.kernel.org/linux-arm-kernel/20220711094549.3445566-1-martin.kepplinger@puri.sm/T/#t

Martin Kepplinger (3):
  PM: domain: fix indentation and use BIT macro for flags
  power: domain: handle genpd correctly when needing interrupts
  soc: imx: gpcv2: fix suspend/resume by setting GENPD_FLAG_IRQ_ON

 drivers/base/power/domain.c | 13 +++++++++++++
 drivers/soc/imx/gpcv2.c     |  3 +++
 include/linux/pm_domain.h   | 20 +++++++++++++-------
 3 files changed, 29 insertions(+), 7 deletions(-)

-- 
2.30.2

[PATCH v5 1/3] PM: domain: fix indentation and use BIT macro for flags

[Martin Kepplinger]

Use the BIT macro for flags and simply do 2 tags indentation.

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
---
 include/linux/pm_domain.h | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h
index ebc351698090..76bc9e3ef5ff 100644
--- a/include/linux/pm_domain.h
+++ b/include/linux/pm_domain.h
@@ -8,6 +8,7 @@
 #ifndef _LINUX_PM_DOMAIN_H
 #define _LINUX_PM_DOMAIN_H
 
+#include <linux/bits.h>
 #include <linux/device.h>
 #include <linux/ktime.h>
 #include <linux/mutex.h>
@@ -61,13 +62,13 @@
  *				components' next wakeup when determining the
  *				optimal idle state.
  */
-#define GENPD_FLAG_PM_CLK	 (1U << 0)
-#define GENPD_FLAG_IRQ_SAFE	 (1U << 1)
-#define GENPD_FLAG_ALWAYS_ON	 (1U << 2)
-#define GENPD_FLAG_ACTIVE_WAKEUP (1U << 3)
-#define GENPD_FLAG_CPU_DOMAIN	 (1U << 4)
-#define GENPD_FLAG_RPM_ALWAYS_ON (1U << 5)
-#define GENPD_FLAG_MIN_RESIDENCY (1U << 6)
+#define GENPD_FLAG_PM_CLK		BIT(0)
+#define GENPD_FLAG_IRQ_SAFE		BIT(1)
+#define GENPD_FLAG_ALWAYS_ON		BIT(2)
+#define GENPD_FLAG_ACTIVE_WAKEUP	BIT(3)
+#define GENPD_FLAG_CPU_DOMAIN		BIT(4)
+#define GENPD_FLAG_RPM_ALWAYS_ON	BIT(5)
+#define GENPD_FLAG_MIN_RESIDENCY	BIT(6)
 
 enum gpd_status {
 	GENPD_STATE_ON = 0,	/* PM domain is on */
-- 
2.30.2

[PATCH v5 2/3] power: domain: handle genpd correctly when needing interrupts

[Martin Kepplinger]

If for example the power-domains' power-supply node (regulator) needs
interrupts to work, the current setup with noirq callbacks cannot
work; for example a pmic regulator on i2c, when suspending, usually already
times out during suspend_noirq:

[   41.024193] buck4: failed to disable: -ETIMEDOUT

So fix system suspend and resume for these power-domains by using the
"outer" suspend/resume callbacks instead. Tested on the imx8mq-librem5 board,
but by looking at the dts, this will fix imx8mq-evk and possibly many other
boards too.

This is designed so that genpd providers just say "this genpd needs
interrupts" (by setting the flag) - without implying an implementation.

Initially system suspend problems had been discussed at
https://lore.kernel.org/linux-arm-kernel/20211002005954.1367653-8-l.stach@pengutronix.de/
which led to discussing the pmic that contains the regulators which
serve as power-domain power-supplies:
https://lore.kernel.org/linux-pm/573166b75e524517782471c2b7f96e03fd93d175.camel@puri.sm/T/

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
---
 drivers/base/power/domain.c | 13 +++++++++++++
 include/linux/pm_domain.h   |  5 +++++
 2 files changed, 18 insertions(+)

diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c
index 5a2e0232862e..ef77700e0def 100644
--- a/drivers/base/power/domain.c
+++ b/drivers/base/power/domain.c
@@ -130,6 +130,7 @@ static const struct genpd_lock_ops genpd_spin_ops = {
 #define genpd_is_active_wakeup(genpd)	(genpd->flags & GENPD_FLAG_ACTIVE_WAKEUP)
 #define genpd_is_cpu_domain(genpd)	(genpd->flags & GENPD_FLAG_CPU_DOMAIN)
 #define genpd_is_rpm_always_on(genpd)	(genpd->flags & GENPD_FLAG_RPM_ALWAYS_ON)
+#define genpd_irq_on(genpd)		(genpd->flags & GENPD_FLAG_IRQ_ON)
 
 static inline bool irq_safe_dev_in_sleep_domain(struct device *dev,
 		const struct generic_pm_domain *genpd)
@@ -2079,6 +2080,13 @@ int pm_genpd_init(struct generic_pm_domain *genpd,
 		genpd->dev_ops.start = pm_clk_resume;
 	}
 
+	if (genpd_irq_on(genpd)) {
+		genpd->domain.ops.suspend = genpd_suspend_noirq;
+		genpd->domain.ops.resume = genpd_resume_noirq;
+		genpd->domain.ops.suspend_noirq = NULL;
+		genpd->domain.ops.resume_noirq = NULL;
+	}
+
 	/* The always-on governor works better with the corresponding flag. */
 	if (gov == &pm_domain_always_on_gov)
 		genpd->flags |= GENPD_FLAG_RPM_ALWAYS_ON;
@@ -2769,6 +2777,11 @@ static int __genpd_dev_pm_attach(struct device *dev, struct device *base_dev,
 			goto err;
 		dev_gpd_data(dev)->default_pstate = pstate;
 	}
+
+	if (pd->domain.ops.suspend_noirq && (pd->flags & GENPD_FLAG_IRQ_ON))
+		dev_err(dev, "PM domain %s needs irqs but uses noirq suspend\n",
+			pd->name);
+
 	return 1;
 
 err:
diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h
index 76bc9e3ef5ff..03bb86e43550 100644
--- a/include/linux/pm_domain.h
+++ b/include/linux/pm_domain.h
@@ -61,6 +61,10 @@
  * GENPD_FLAG_MIN_RESIDENCY:	Enable the genpd governor to consider its
  *				components' next wakeup when determining the
  *				optimal idle state.
+ *
+ * GENPD_FLAG_IRQ_ON:		genpd needs irqs to be able to manage power
+ *				on/off. Use the outer suspend/resume callbacks
+ *				instead of noirq for example.
  */
 #define GENPD_FLAG_PM_CLK		BIT(0)
 #define GENPD_FLAG_IRQ_SAFE		BIT(1)
@@ -69,6 +73,7 @@
 #define GENPD_FLAG_CPU_DOMAIN		BIT(4)
 #define GENPD_FLAG_RPM_ALWAYS_ON	BIT(5)
 #define GENPD_FLAG_MIN_RESIDENCY	BIT(6)
+#define GENPD_FLAG_IRQ_ON		BIT(7)
 
 enum gpd_status {
 	GENPD_STATE_ON = 0,	/* PM domain is on */
-- 
2.30.2

[PATCH v5 3/3] soc: imx: gpcv2: fix suspend/resume by setting GENPD_FLAG_IRQ_ON

[Martin Kepplinger]

For boards that use power-domains' power-supplies that need interrupts
to work (like regulator over i2c), set GENPD_FLAG_IRQ_ON.
This will tell genpd to adjust accordingly. Currently it "only" sets the
correct suspend/resume callbacks.

This fixes suspend/resume on imx8mq-librem5 boards (tested) and
imx8mq-evk (by looking at dts) and possibly more.

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
---
 drivers/soc/imx/gpcv2.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/soc/imx/gpcv2.c b/drivers/soc/imx/gpcv2.c
index 6383a4edc360..e058aed76602 100644
--- a/drivers/soc/imx/gpcv2.c
+++ b/drivers/soc/imx/gpcv2.c
@@ -1337,6 +1337,9 @@ static int imx_pgc_domain_probe(struct platform_device *pdev)
 		regmap_update_bits(domain->regmap, domain->regs->map,
 				   domain->bits.map, domain->bits.map);
 
+	if (of_parse_phandle(domain->dev->of_node, "power-supply", 0))
+		domain->genpd.flags |= GENPD_FLAG_IRQ_ON;
+
 	ret = pm_genpd_init(&domain->genpd, NULL, true);
 	if (ret) {
 		dev_err(domain->dev, "Failed to init power domain\n");
-- 
2.30.2

Re: [PATCH v5 1/3] PM: domain: fix indentation and use BIT macro for flags

[Ulf Hansson]

On Thu, 21 Jul 2022 at 06:36, Martin Kepplinger
<martin.kepplinger@puri.sm> wrote:
>
> Use the BIT macro for flags and simply do 2 tags indentation.
>
> Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>

I prefer to keep the advantages with doing a git blame on this file,
over this cleanup. Unless you feel strongly about this I would
appreciate it if you can drop $subject patch from the series.

Kind regards
Uffe

> ---
>  include/linux/pm_domain.h | 15 ++++++++-------
>  1 file changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h
> index ebc351698090..76bc9e3ef5ff 100644
> --- a/include/linux/pm_domain.h
> +++ b/include/linux/pm_domain.h
> @@ -8,6 +8,7 @@
>  #ifndef _LINUX_PM_DOMAIN_H
>  #define _LINUX_PM_DOMAIN_H
>
> +#include <linux/bits.h>
>  #include <linux/device.h>
>  #include <linux/ktime.h>
>  #include <linux/mutex.h>
> @@ -61,13 +62,13 @@
>   *                             components' next wakeup when determining the
>   *                             optimal idle state.
>   */
> -#define GENPD_FLAG_PM_CLK       (1U << 0)
> -#define GENPD_FLAG_IRQ_SAFE     (1U << 1)
> -#define GENPD_FLAG_ALWAYS_ON    (1U << 2)
> -#define GENPD_FLAG_ACTIVE_WAKEUP (1U << 3)
> -#define GENPD_FLAG_CPU_DOMAIN   (1U << 4)
> -#define GENPD_FLAG_RPM_ALWAYS_ON (1U << 5)
> -#define GENPD_FLAG_MIN_RESIDENCY (1U << 6)
> +#define GENPD_FLAG_PM_CLK              BIT(0)
> +#define GENPD_FLAG_IRQ_SAFE            BIT(1)
> +#define GENPD_FLAG_ALWAYS_ON           BIT(2)
> +#define GENPD_FLAG_ACTIVE_WAKEUP       BIT(3)
> +#define GENPD_FLAG_CPU_DOMAIN          BIT(4)
> +#define GENPD_FLAG_RPM_ALWAYS_ON       BIT(5)
> +#define GENPD_FLAG_MIN_RESIDENCY       BIT(6)
>
>  enum gpd_status {
>         GENPD_STATE_ON = 0,     /* PM domain is on */
> --
> 2.30.2
>

Re: [PATCH v5 2/3] power: domain: handle genpd correctly when needing interrupts

[Ulf Hansson]

On Thu, 21 Jul 2022 at 06:37, Martin Kepplinger
<martin.kepplinger@puri.sm> wrote:
>
> If for example the power-domains' power-supply node (regulator) needs
> interrupts to work, the current setup with noirq callbacks cannot
> work; for example a pmic regulator on i2c, when suspending, usually already
> times out during suspend_noirq:
>
> [   41.024193] buck4: failed to disable: -ETIMEDOUT
>
> So fix system suspend and resume for these power-domains by using the
> "outer" suspend/resume callbacks instead. Tested on the imx8mq-librem5 board,
> but by looking at the dts, this will fix imx8mq-evk and possibly many other
> boards too.
>
> This is designed so that genpd providers just say "this genpd needs
> interrupts" (by setting the flag) - without implying an implementation.
>
> Initially system suspend problems had been discussed at
> https://lore.kernel.org/linux-arm-kernel/20211002005954.1367653-8-l.stach@pengutronix.de/
> which led to discussing the pmic that contains the regulators which
> serve as power-domain power-supplies:
> https://lore.kernel.org/linux-pm/573166b75e524517782471c2b7f96e03fd93d175.camel@puri.sm/T/
>
> Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
> ---
>  drivers/base/power/domain.c | 13 +++++++++++++
>  include/linux/pm_domain.h   |  5 +++++
>  2 files changed, 18 insertions(+)
>
> diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c
> index 5a2e0232862e..ef77700e0def 100644
> --- a/drivers/base/power/domain.c
> +++ b/drivers/base/power/domain.c
> @@ -130,6 +130,7 @@ static const struct genpd_lock_ops genpd_spin_ops = {
>  #define genpd_is_active_wakeup(genpd)  (genpd->flags & GENPD_FLAG_ACTIVE_WAKEUP)
>  #define genpd_is_cpu_domain(genpd)     (genpd->flags & GENPD_FLAG_CPU_DOMAIN)
>  #define genpd_is_rpm_always_on(genpd)  (genpd->flags & GENPD_FLAG_RPM_ALWAYS_ON)
> +#define genpd_irq_on(genpd)            (genpd->flags & GENPD_FLAG_IRQ_ON)
>
>  static inline bool irq_safe_dev_in_sleep_domain(struct device *dev,
>                 const struct generic_pm_domain *genpd)
> @@ -2079,6 +2080,13 @@ int pm_genpd_init(struct generic_pm_domain *genpd,
>                 genpd->dev_ops.start = pm_clk_resume;
>         }
>
> +       if (genpd_irq_on(genpd)) {
> +               genpd->domain.ops.suspend = genpd_suspend_noirq;
> +               genpd->domain.ops.resume = genpd_resume_noirq;
> +               genpd->domain.ops.suspend_noirq = NULL;
> +               genpd->domain.ops.resume_noirq = NULL;

Please move this a few lines above, just before we assign the _*noirq
callbacks. In this way you don't need to reset thosepointers.

> +       }
> +
>         /* The always-on governor works better with the corresponding flag. */
>         if (gov == &pm_domain_always_on_gov)
>                 genpd->flags |= GENPD_FLAG_RPM_ALWAYS_ON;
> @@ -2769,6 +2777,11 @@ static int __genpd_dev_pm_attach(struct device *dev, struct device *base_dev,
>                         goto err;
>                 dev_gpd_data(dev)->default_pstate = pstate;
>         }
> +
> +       if (pd->domain.ops.suspend_noirq && (pd->flags & GENPD_FLAG_IRQ_ON))
> +               dev_err(dev, "PM domain %s needs irqs but uses noirq suspend\n",
> +                       pd->name);
> +

This doesn't make sense, as it can never happen according to what we
do in pm_genpd_init().

What Lucas suggested in the other thread was to log a warning if a
device's (that gets attached to genpd) bus/driver has _*noirq
callbacks. That would make sense.

Thinking more about this, perhaps we should move to use the
_*late/early callbacks instead for genpd. This would decrease the
window of potential problematic consumers users (drivers/buses).

>         return 1;
>
>  err:
> diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h
> index 76bc9e3ef5ff..03bb86e43550 100644
> --- a/include/linux/pm_domain.h
> +++ b/include/linux/pm_domain.h
> @@ -61,6 +61,10 @@
>   * GENPD_FLAG_MIN_RESIDENCY:   Enable the genpd governor to consider its
>   *                             components' next wakeup when determining the
>   *                             optimal idle state.
> + *
> + * GENPD_FLAG_IRQ_ON:          genpd needs irqs to be able to manage power
> + *                             on/off. Use the outer suspend/resume callbacks
> + *                             instead of noirq for example.
>   */
>  #define GENPD_FLAG_PM_CLK              BIT(0)
>  #define GENPD_FLAG_IRQ_SAFE            BIT(1)
> @@ -69,6 +73,7 @@
>  #define GENPD_FLAG_CPU_DOMAIN          BIT(4)
>  #define GENPD_FLAG_RPM_ALWAYS_ON       BIT(5)
>  #define GENPD_FLAG_MIN_RESIDENCY       BIT(6)
> +#define GENPD_FLAG_IRQ_ON              BIT(7)
>
>  enum gpd_status {
>         GENPD_STATE_ON = 0,     /* PM domain is on */
> --
> 2.30.2
>

Kind regards
Uffe

Re: [PATCH v5 3/3] soc: imx: gpcv2: fix suspend/resume by setting GENPD_FLAG_IRQ_ON

[Lucas Stach]

Am Donnerstag, dem 21.07.2022 um 06:36 +0200 schrieb Martin Kepplinger:
> For boards that use power-domains' power-supplies that need interrupts
> to work (like regulator over i2c), set GENPD_FLAG_IRQ_ON.
> This will tell genpd to adjust accordingly. Currently it "only" sets the
> correct suspend/resume callbacks.
> 
> This fixes suspend/resume on imx8mq-librem5 boards (tested) and
> imx8mq-evk (by looking at dts) and possibly more.
> 
> Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
> ---
>  drivers/soc/imx/gpcv2.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/soc/imx/gpcv2.c b/drivers/soc/imx/gpcv2.c
> index 6383a4edc360..e058aed76602 100644
> --- a/drivers/soc/imx/gpcv2.c
> +++ b/drivers/soc/imx/gpcv2.c
> @@ -1337,6 +1337,9 @@ static int imx_pgc_domain_probe(struct platform_device *pdev)
>  		regmap_update_bits(domain->regmap, domain->regs->map,
>  				   domain->bits.map, domain->bits.map);
>  
> +	if (of_parse_phandle(domain->dev->of_node, "power-supply", 0))

We don't actually need to parse the phandle. For a simple presence
check of_property_read_bool() is enough.

Regards,
Lucas

> +		domain->genpd.flags |= GENPD_FLAG_IRQ_ON;
> +
>  	ret = pm_genpd_init(&domain->genpd, NULL, true);
>  	if (ret) {
>  		dev_err(domain->dev, "Failed to init power domain\n");

Re: [PATCH v5 2/3] power: domain: handle genpd correctly when needing interrupts

[Martin Kepplinger]

Am Donnerstag, dem 21.07.2022 um 13:16 +0200 schrieb Ulf Hansson:
> On Thu, 21 Jul 2022 at 06:37, Martin Kepplinger
> <martin.kepplinger@puri.sm> wrote:
> > 
> > If for example the power-domains' power-supply node (regulator)
> > needs
> > interrupts to work, the current setup with noirq callbacks cannot
> > work; for example a pmic regulator on i2c, when suspending, usually
> > already
> > times out during suspend_noirq:
> > 
> > [   41.024193] buck4: failed to disable: -ETIMEDOUT
> > 
> > So fix system suspend and resume for these power-domains by using
> > the
> > "outer" suspend/resume callbacks instead. Tested on the imx8mq-
> > librem5 board,
> > but by looking at the dts, this will fix imx8mq-evk and possibly
> > many other
> > boards too.
> > 
> > This is designed so that genpd providers just say "this genpd needs
> > interrupts" (by setting the flag) - without implying an
> > implementation.
> > 
> > Initially system suspend problems had been discussed at
> >  
> > https://lore.kernel.org/linux-arm-kernel/20211002005954.1367653-8-l.stach@pengutronix.de/
> > which led to discussing the pmic that contains the regulators which
> > serve as power-domain power-supplies:
> >  
> > https://lore.kernel.org/linux-pm/573166b75e524517782471c2b7f96e03fd93d175.camel@puri.sm/T/
> > 
> > Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
> > ---
> >  drivers/base/power/domain.c | 13 +++++++++++++
> >  include/linux/pm_domain.h   |  5 +++++
> >  2 files changed, 18 insertions(+)
> > 
> > diff --git a/drivers/base/power/domain.c
> > b/drivers/base/power/domain.c
> > index 5a2e0232862e..ef77700e0def 100644
> > --- a/drivers/base/power/domain.c
> > +++ b/drivers/base/power/domain.c
> > @@ -130,6 +130,7 @@ static const struct genpd_lock_ops
> > genpd_spin_ops = {
> >  #define genpd_is_active_wakeup(genpd)  (genpd->flags &
> > GENPD_FLAG_ACTIVE_WAKEUP)
> >  #define genpd_is_cpu_domain(genpd)     (genpd->flags &
> > GENPD_FLAG_CPU_DOMAIN)
> >  #define genpd_is_rpm_always_on(genpd)  (genpd->flags &
> > GENPD_FLAG_RPM_ALWAYS_ON)
> > +#define genpd_irq_on(genpd)            (genpd->flags &
> > GENPD_FLAG_IRQ_ON)
> > 
> >  static inline bool irq_safe_dev_in_sleep_domain(struct device
> > *dev,
> >                 const struct generic_pm_domain *genpd)
> > @@ -2079,6 +2080,13 @@ int pm_genpd_init(struct generic_pm_domain
> > *genpd,
> >                 genpd->dev_ops.start = pm_clk_resume;
> >         }
> > 
> > +       if (genpd_irq_on(genpd)) {
> > +               genpd->domain.ops.suspend = genpd_suspend_noirq;
> > +               genpd->domain.ops.resume = genpd_resume_noirq;
> > +               genpd->domain.ops.suspend_noirq = NULL;
> > +               genpd->domain.ops.resume_noirq = NULL;
> 
> Please move this a few lines above, just before we assign the _*noirq
> callbacks. In this way you don't need to reset thosepointers.

good point, thanks.

> 
> > +       }
> > +
> >         /* The always-on governor works better with the
> > corresponding flag. */
> >         if (gov == &pm_domain_always_on_gov)
> >                 genpd->flags |= GENPD_FLAG_RPM_ALWAYS_ON;
> > @@ -2769,6 +2777,11 @@ static int __genpd_dev_pm_attach(struct
> > device *dev, struct device *base_dev,
> >                         goto err;
> >                 dev_gpd_data(dev)->default_pstate = pstate;
> >         }
> > +
> > +       if (pd->domain.ops.suspend_noirq && (pd->flags &
> > GENPD_FLAG_IRQ_ON))
> > +               dev_err(dev, "PM domain %s needs irqs but uses
> > noirq suspend\n",
> > +                       pd->name);
> > +
> 
> This doesn't make sense, as it can never happen according to what we
> do in pm_genpd_init().
> 
> What Lucas suggested in the other thread was to log a warning if a
> device's (that gets attached to genpd) bus/driver has _*noirq
> callbacks. That would make sense.
> 
> Thinking more about this, perhaps we should move to use the
> _*late/early callbacks instead for genpd. This would decrease the
> window of potential problematic consumers users (drivers/buses).
> 

ok, I'll probably send another version without this. I'm leaving for
vacations for a few weeks soon, so I'd rather do the essentials only
here and take notes about what could be added later, if that's ok.

thank you very much!

                        martin


> >         return 1;
> > 
> >  err:
> > diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h
> > index 76bc9e3ef5ff..03bb86e43550 100644
> > --- a/include/linux/pm_domain.h
> > +++ b/include/linux/pm_domain.h
> > @@ -61,6 +61,10 @@
> >   * GENPD_FLAG_MIN_RESIDENCY:   Enable the genpd governor to
> > consider its
> >   *                             components' next wakeup when
> > determining the
> >   *                             optimal idle state.
> > + *
> > + * GENPD_FLAG_IRQ_ON:          genpd needs irqs to be able to
> > manage power
> > + *                             on/off. Use the outer
> > suspend/resume callbacks
> > + *                             instead of noirq for example.
> >   */
> >  #define GENPD_FLAG_PM_CLK              BIT(0)
> >  #define GENPD_FLAG_IRQ_SAFE            BIT(1)
> > @@ -69,6 +73,7 @@
> >  #define GENPD_FLAG_CPU_DOMAIN          BIT(4)
> >  #define GENPD_FLAG_RPM_ALWAYS_ON       BIT(5)
> >  #define GENPD_FLAG_MIN_RESIDENCY       BIT(6)
> > +#define GENPD_FLAG_IRQ_ON              BIT(7)
> > 
> >  enum gpd_status {
> >         GENPD_STATE_ON = 0,     /* PM domain is on */
> > --
> > 2.30.2
> > 
> 
> Kind regards
> Uffe