[PATCH] sunxi-rsb: Include OF based modalias in device uevent

[PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Stefan Brüns]

Include the OF-based modalias in the uevent sent when registering devices
on the sunxi RSB bus, so that user space has a chance to autoload the
kernel module for the device.

Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
the AXP803 PMIC is built as a module, it is not loaded and the system
ends up with an disfunctional MMC controller.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
---
 drivers/bus/sunxi-rsb.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
index 328ca93781cf..37cb57244cbe 100644
--- a/drivers/bus/sunxi-rsb.c
+++ b/drivers/bus/sunxi-rsb.c
@@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device *dev)
 	return drv->remove(to_sunxi_rsb_device(dev));
 }
 
+static int sunxi_rsb_device_uevent(struct device *dev,
+				   struct kobj_uevent_env *env)
+{
+	int ret;
+
+	ret = of_device_uevent_modalias(dev, env);
+	if (ret != -ENODEV)
+		return ret;
+
+	return 0;
+}
+
 static struct bus_type sunxi_rsb_bus = {
 	.name		= RSB_CTRL_NAME,
 	.match		= sunxi_rsb_device_match,
 	.probe		= sunxi_rsb_device_probe,
 	.remove		= sunxi_rsb_device_remove,
+	.uevent		= sunxi_rsb_device_uevent,
 };
 
 static void sunxi_rsb_dev_release(struct device *dev)
-- 
2.15.0

[PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Stefan Brüns]

Include the OF-based modalias in the uevent sent when registering devices
on the sunxi RSB bus, so that user space has a chance to autoload the
kernel module for the device.

Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
the AXP803 PMIC is built as a module, it is not loaded and the system
ends up with an disfunctional MMC controller.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: Stefan Br?ns <stefan.bruens@rwth-aachen.de>
---
 drivers/bus/sunxi-rsb.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
index 328ca93781cf..37cb57244cbe 100644
--- a/drivers/bus/sunxi-rsb.c
+++ b/drivers/bus/sunxi-rsb.c
@@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device *dev)
 	return drv->remove(to_sunxi_rsb_device(dev));
 }
 
+static int sunxi_rsb_device_uevent(struct device *dev,
+				   struct kobj_uevent_env *env)
+{
+	int ret;
+
+	ret = of_device_uevent_modalias(dev, env);
+	if (ret != -ENODEV)
+		return ret;
+
+	return 0;
+}
+
 static struct bus_type sunxi_rsb_bus = {
 	.name		= RSB_CTRL_NAME,
 	.match		= sunxi_rsb_device_match,
 	.probe		= sunxi_rsb_device_probe,
 	.remove		= sunxi_rsb_device_remove,
+	.uevent		= sunxi_rsb_device_uevent,
 };
 
 static void sunxi_rsb_dev_release(struct device *dev)
-- 
2.15.0

Re: [PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Maxime Ripard]

[-- Attachment #1: Type: text/plain, Size: 1803 bytes --]

Hi,

On Mon, Nov 27, 2017 at 01:17:25PM +0100, Stefan Brüns wrote:
> Include the OF-based modalias in the uevent sent when registering devices
> on the sunxi RSB bus, so that user space has a chance to autoload the
> kernel module for the device.
> 
> Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
> pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
> the AXP803 PMIC is built as a module, it is not loaded and the system
> ends up with an disfunctional MMC controller.
> 
> Cc: stable <stable@vger.kernel.org>
> Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
> ---
>  drivers/bus/sunxi-rsb.c | 13 +++++++++++++
>  1 file changed, 13 insertions(+)
> 
> diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
> index 328ca93781cf..37cb57244cbe 100644
> --- a/drivers/bus/sunxi-rsb.c
> +++ b/drivers/bus/sunxi-rsb.c
> @@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device *dev)
>  	return drv->remove(to_sunxi_rsb_device(dev));
>  }
>  
> +static int sunxi_rsb_device_uevent(struct device *dev,
> +				   struct kobj_uevent_env *env)
> +{
> +	int ret;
> +
> +	ret = of_device_uevent_modalias(dev, env);
> +	if (ret != -ENODEV)
> +		return ret;

A comment explaining why we need to ignore the ENODEV error code would
be great here.

> +	return 0;
> +}
> +
>  static struct bus_type sunxi_rsb_bus = {
>  	.name		= RSB_CTRL_NAME,
>  	.match		= sunxi_rsb_device_match,
>  	.probe		= sunxi_rsb_device_probe,
>  	.remove		= sunxi_rsb_device_remove,
> +	.uevent		= sunxi_rsb_device_uevent,

Any reason to not use of_device_uevent_modalias directly here?

Thanks!
Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Maxime Ripard]

Hi,

On Mon, Nov 27, 2017 at 01:17:25PM +0100, Stefan Br?ns wrote:
> Include the OF-based modalias in the uevent sent when registering devices
> on the sunxi RSB bus, so that user space has a chance to autoload the
> kernel module for the device.
> 
> Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
> pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
> the AXP803 PMIC is built as a module, it is not loaded and the system
> ends up with an disfunctional MMC controller.
> 
> Cc: stable <stable@vger.kernel.org>
> Signed-off-by: Stefan Br?ns <stefan.bruens@rwth-aachen.de>
> ---
>  drivers/bus/sunxi-rsb.c | 13 +++++++++++++
>  1 file changed, 13 insertions(+)
> 
> diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
> index 328ca93781cf..37cb57244cbe 100644
> --- a/drivers/bus/sunxi-rsb.c
> +++ b/drivers/bus/sunxi-rsb.c
> @@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device *dev)
>  	return drv->remove(to_sunxi_rsb_device(dev));
>  }
>  
> +static int sunxi_rsb_device_uevent(struct device *dev,
> +				   struct kobj_uevent_env *env)
> +{
> +	int ret;
> +
> +	ret = of_device_uevent_modalias(dev, env);
> +	if (ret != -ENODEV)
> +		return ret;

A comment explaining why we need to ignore the ENODEV error code would
be great here.

> +	return 0;
> +}
> +
>  static struct bus_type sunxi_rsb_bus = {
>  	.name		= RSB_CTRL_NAME,
>  	.match		= sunxi_rsb_device_match,
>  	.probe		= sunxi_rsb_device_probe,
>  	.remove		= sunxi_rsb_device_remove,
> +	.uevent		= sunxi_rsb_device_uevent,

Any reason to not use of_device_uevent_modalias directly here?

Thanks!
Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20171127/2f550d85/attachment.sig>

Re: [PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Stefan Brüns]

[-- Attachment #1: Type: text/plain, Size: 2443 bytes --]

On Monday, November 27, 2017 4:35:02 PM CET Maxime Ripard wrote:
> Hi,
> 
> On Mon, Nov 27, 2017 at 01:17:25PM +0100, Stefan Brüns wrote:
> > Include the OF-based modalias in the uevent sent when registering devices
> > on the sunxi RSB bus, so that user space has a chance to autoload the
> > kernel module for the device.
> > 
> > Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
> > pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
> > the AXP803 PMIC is built as a module, it is not loaded and the system
> > ends up with an disfunctional MMC controller.
> > 
> > Cc: stable <stable@vger.kernel.org>
> > Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
> > ---
> > 
> >  drivers/bus/sunxi-rsb.c | 13 +++++++++++++
> >  1 file changed, 13 insertions(+)
> > 
> > diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
> > index 328ca93781cf..37cb57244cbe 100644
> > --- a/drivers/bus/sunxi-rsb.c
> > +++ b/drivers/bus/sunxi-rsb.c
> > @@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device
> > *dev)> 
> >  	return drv->remove(to_sunxi_rsb_device(dev));
> >  
> >  }
> > 
> > +static int sunxi_rsb_device_uevent(struct device *dev,
> > +				   struct kobj_uevent_env *env)
> > +{
> > +	int ret;
> > +
> > +	ret = of_device_uevent_modalias(dev, env);
> > +	if (ret != -ENODEV)
> > +		return ret;
> 
> A comment explaining why we need to ignore the ENODEV error code would
> be great here.

Lazy answer - everyone else is doing the same, and nobody cared to add an 
explanation.

For *some* drivers, this is likely because the same device may be enumerated 
from e.g ACPI or OF, and for an ACPI device -ENODEV will be returned, as
dev->of_node is NULL.

For devices which are only usable in an OF context, this is bogus. Not sure 
about sunxi-rsb.

> > +	return 0;
> > +}
> > +
> > 
> >  static struct bus_type sunxi_rsb_bus = {
> >  
> >  	.name		= RSB_CTRL_NAME,
> >  	.match		= sunxi_rsb_device_match,
> >  	.probe		= sunxi_rsb_device_probe,
> >  	.remove		= sunxi_rsb_device_remove,
> > 
> > +	.uevent		= sunxi_rsb_device_uevent,
> 
> Any reason to not use of_device_uevent_modalias directly here?

*If* sunxi-rsb can be used without OF, then yes, otherwise no.

Regards,

Stefan

-- 
Stefan Brüns  /  Bergstraße 21  /  52062 Aachen
home: +49 241 53809034     mobile: +49 151 50412019

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

[PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Stefan Brüns]

On Monday, November 27, 2017 4:35:02 PM CET Maxime Ripard wrote:
> Hi,
> 
> On Mon, Nov 27, 2017 at 01:17:25PM +0100, Stefan Br?ns wrote:
> > Include the OF-based modalias in the uevent sent when registering devices
> > on the sunxi RSB bus, so that user space has a chance to autoload the
> > kernel module for the device.
> > 
> > Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
> > pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
> > the AXP803 PMIC is built as a module, it is not loaded and the system
> > ends up with an disfunctional MMC controller.
> > 
> > Cc: stable <stable@vger.kernel.org>
> > Signed-off-by: Stefan Br?ns <stefan.bruens@rwth-aachen.de>
> > ---
> > 
> >  drivers/bus/sunxi-rsb.c | 13 +++++++++++++
> >  1 file changed, 13 insertions(+)
> > 
> > diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
> > index 328ca93781cf..37cb57244cbe 100644
> > --- a/drivers/bus/sunxi-rsb.c
> > +++ b/drivers/bus/sunxi-rsb.c
> > @@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device
> > *dev)> 
> >  	return drv->remove(to_sunxi_rsb_device(dev));
> >  
> >  }
> > 
> > +static int sunxi_rsb_device_uevent(struct device *dev,
> > +				   struct kobj_uevent_env *env)
> > +{
> > +	int ret;
> > +
> > +	ret = of_device_uevent_modalias(dev, env);
> > +	if (ret != -ENODEV)
> > +		return ret;
> 
> A comment explaining why we need to ignore the ENODEV error code would
> be great here.

Lazy answer - everyone else is doing the same, and nobody cared to add an 
explanation.

For *some* drivers, this is likely because the same device may be enumerated 
from e.g ACPI or OF, and for an ACPI device -ENODEV will be returned, as
dev->of_node is NULL.

For devices which are only usable in an OF context, this is bogus. Not sure 
about sunxi-rsb.

> > +	return 0;
> > +}
> > +
> > 
> >  static struct bus_type sunxi_rsb_bus = {
> >  
> >  	.name		= RSB_CTRL_NAME,
> >  	.match		= sunxi_rsb_device_match,
> >  	.probe		= sunxi_rsb_device_probe,
> >  	.remove		= sunxi_rsb_device_remove,
> > 
> > +	.uevent		= sunxi_rsb_device_uevent,
> 
> Any reason to not use of_device_uevent_modalias directly here?

*If* sunxi-rsb can be used without OF, then yes, otherwise no.

Regards,

Stefan

-- 
Stefan Br?ns  /  Bergstra?e 21  /  52062 Aachen
home: +49 241 53809034     mobile: +49 151 50412019
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20171127/807caf5a/attachment-0001.sig>

Re: [PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Chen-Yu Tsai]

On Tue, Nov 28, 2017 at 1:19 AM, Stefan Brüns
<stefan.bruens@rwth-aachen.de> wrote:
> On Monday, November 27, 2017 4:35:02 PM CET Maxime Ripard wrote:
>> Hi,
>>
>> On Mon, Nov 27, 2017 at 01:17:25PM +0100, Stefan Brüns wrote:
>> > Include the OF-based modalias in the uevent sent when registering devices
>> > on the sunxi RSB bus, so that user space has a chance to autoload the
>> > kernel module for the device.
>> >
>> > Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
>> > pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
>> > the AXP803 PMIC is built as a module, it is not loaded and the system
>> > ends up with an disfunctional MMC controller.
>> >
>> > Cc: stable <stable@vger.kernel.org>
>> > Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
>> > ---
>> >
>> >  drivers/bus/sunxi-rsb.c | 13 +++++++++++++
>> >  1 file changed, 13 insertions(+)
>> >
>> > diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
>> > index 328ca93781cf..37cb57244cbe 100644
>> > --- a/drivers/bus/sunxi-rsb.c
>> > +++ b/drivers/bus/sunxi-rsb.c
>> > @@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device
>> > *dev)>
>> >     return drv->remove(to_sunxi_rsb_device(dev));
>> >
>> >  }
>> >
>> > +static int sunxi_rsb_device_uevent(struct device *dev,
>> > +                              struct kobj_uevent_env *env)
>> > +{
>> > +   int ret;
>> > +
>> > +   ret = of_device_uevent_modalias(dev, env);
>> > +   if (ret != -ENODEV)
>> > +           return ret;
>>
>> A comment explaining why we need to ignore the ENODEV error code would
>> be great here.
>
> Lazy answer - everyone else is doing the same, and nobody cared to add an
> explanation.
>
> For *some* drivers, this is likely because the same device may be enumerated
> from e.g ACPI or OF, and for an ACPI device -ENODEV will be returned, as
> dev->of_node is NULL.
>
> For devices which are only usable in an OF context, this is bogus. Not sure
> about sunxi-rsb.

sunxi-rsb (and Allwinner support in mainline in general) is OF only. With the
exception of mfd sub-devices, if someone is enumerating devices in some other
fashion, it should rightly blow up in their face.

>> > +   return 0;
>> > +}
>> > +
>> >
>> >  static struct bus_type sunxi_rsb_bus = {
>> >
>> >     .name           = RSB_CTRL_NAME,
>> >     .match          = sunxi_rsb_device_match,
>> >     .probe          = sunxi_rsb_device_probe,
>> >     .remove         = sunxi_rsb_device_remove,
>> >
>> > +   .uevent         = sunxi_rsb_device_uevent,
>>
>> Any reason to not use of_device_uevent_modalias directly here?
>
> *If* sunxi-rsb can be used without OF, then yes, otherwise no.

You can just use of_device_uevent_modalias, like the other OF platforms.

ChenYu

[PATCH] sunxi-rsb: Include OF based modalias in device uevent

[Chen-Yu Tsai]

On Tue, Nov 28, 2017 at 1:19 AM, Stefan Br?ns
<stefan.bruens@rwth-aachen.de> wrote:
> On Monday, November 27, 2017 4:35:02 PM CET Maxime Ripard wrote:
>> Hi,
>>
>> On Mon, Nov 27, 2017 at 01:17:25PM +0100, Stefan Br?ns wrote:
>> > Include the OF-based modalias in the uevent sent when registering devices
>> > on the sunxi RSB bus, so that user space has a chance to autoload the
>> > kernel module for the device.
>> >
>> > Fixes a regression caused by commit 3f241bfa60bd ("arm64: allwinner: a64:
>> > pine64: Use dcdc1 regulator for mmc0"). When the axp20x-rsb module for
>> > the AXP803 PMIC is built as a module, it is not loaded and the system
>> > ends up with an disfunctional MMC controller.
>> >
>> > Cc: stable <stable@vger.kernel.org>
>> > Signed-off-by: Stefan Br?ns <stefan.bruens@rwth-aachen.de>
>> > ---
>> >
>> >  drivers/bus/sunxi-rsb.c | 13 +++++++++++++
>> >  1 file changed, 13 insertions(+)
>> >
>> > diff --git a/drivers/bus/sunxi-rsb.c b/drivers/bus/sunxi-rsb.c
>> > index 328ca93781cf..37cb57244cbe 100644
>> > --- a/drivers/bus/sunxi-rsb.c
>> > +++ b/drivers/bus/sunxi-rsb.c
>> > @@ -173,11 +173,24 @@ static int sunxi_rsb_device_remove(struct device
>> > *dev)>
>> >     return drv->remove(to_sunxi_rsb_device(dev));
>> >
>> >  }
>> >
>> > +static int sunxi_rsb_device_uevent(struct device *dev,
>> > +                              struct kobj_uevent_env *env)
>> > +{
>> > +   int ret;
>> > +
>> > +   ret = of_device_uevent_modalias(dev, env);
>> > +   if (ret != -ENODEV)
>> > +           return ret;
>>
>> A comment explaining why we need to ignore the ENODEV error code would
>> be great here.
>
> Lazy answer - everyone else is doing the same, and nobody cared to add an
> explanation.
>
> For *some* drivers, this is likely because the same device may be enumerated
> from e.g ACPI or OF, and for an ACPI device -ENODEV will be returned, as
> dev->of_node is NULL.
>
> For devices which are only usable in an OF context, this is bogus. Not sure
> about sunxi-rsb.

sunxi-rsb (and Allwinner support in mainline in general) is OF only. With the
exception of mfd sub-devices, if someone is enumerating devices in some other
fashion, it should rightly blow up in their face.

>> > +   return 0;
>> > +}
>> > +
>> >
>> >  static struct bus_type sunxi_rsb_bus = {
>> >
>> >     .name           = RSB_CTRL_NAME,
>> >     .match          = sunxi_rsb_device_match,
>> >     .probe          = sunxi_rsb_device_probe,
>> >     .remove         = sunxi_rsb_device_remove,
>> >
>> > +   .uevent         = sunxi_rsb_device_uevent,
>>
>> Any reason to not use of_device_uevent_modalias directly here?
>
> *If* sunxi-rsb can be used without OF, then yes, otherwise no.

You can just use of_device_uevent_modalias, like the other OF platforms.

ChenYu