[PATCH] mfd: kempld-core: Check for DMI definition before ACPI

[PATCH] mfd: kempld-core: Check for DMI definition before ACPI

[Michael Brunner]

Change the detection order to priorize DMI table entries over available
ACPI entries.

This makes it more easy for product developers to patch product specific
handling into the driver.
Furthermore it allows to simplify the implementation a bit and
especially to remove the need to force synchronous probing.

Based on the following commit introduced with v5.10-rc1:
commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")

Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
---
 drivers/mfd/kempld-core.c | 23 ++---------------------
 1 file changed, 2 insertions(+), 21 deletions(-)

diff --git a/drivers/mfd/kempld-core.c b/drivers/mfd/kempld-core.c
index 2c9295953c11..aa7f386646a1 100644
--- a/drivers/mfd/kempld-core.c
+++ b/drivers/mfd/kempld-core.c
@@ -125,7 +125,6 @@ static const struct kempld_platform_data kempld_platform_data_generic = {
 };
 
 static struct platform_device *kempld_pdev;
-static bool kempld_acpi_mode;
 
 static int kempld_create_platform_device(const struct dmi_system_id *id)
 {
@@ -501,8 +500,6 @@ static int kempld_probe(struct platform_device *pdev)
 		ret = kempld_get_acpi_data(pdev);
 		if (ret)
 			return ret;
-
-		kempld_acpi_mode = true;
 	} else if (kempld_pdev != pdev) {
 		/*
 		 * The platform device we are probing is not the one we
@@ -565,7 +562,6 @@ static struct platform_driver kempld_driver = {
 	.driver		= {
 		.name	= "kempld",
 		.acpi_match_table = ACPI_PTR(kempld_acpi_table),
-		.probe_type = PROBE_FORCE_SYNCHRONOUS,
 	},
 	.probe		= kempld_probe,
 	.remove		= kempld_remove,
@@ -884,7 +880,6 @@ MODULE_DEVICE_TABLE(dmi, kempld_dmi_table);
 static int __init kempld_init(void)
 {
 	const struct dmi_system_id *id;
-	int ret;
 
 	if (force_device_id[0]) {
 		for (id = kempld_dmi_table;
@@ -894,24 +889,10 @@ static int __init kempld_init(void)
 					break;
 		if (id->matches[0].slot == DMI_NONE)
 			return -ENODEV;
-	}
-
-	ret = platform_driver_register(&kempld_driver);
-	if (ret)
-		return ret;
-
-	/*
-	 * With synchronous probing the device should already be probed now.
-	 * If no device id is forced and also no ACPI definition for the
-	 * device was found, scan DMI table as fallback.
-	 *
-	 * If drivers_autoprobing is disabled and the device is found here,
-	 * only that device can be bound manually later.
-	 */
-	if (!kempld_pdev && !kempld_acpi_mode)
+	} else
 		dmi_check_system(kempld_dmi_table);
 
-	return 0;
+	return platform_driver_register(&kempld_driver);
 }
 
 static void __exit kempld_exit(void)
-- 
2.25.1

Re: [PATCH] mfd: kempld-core: Check for DMI definition before ACPI

[Guenter Roeck]

On 11/9/20 11:46 PM, Michael Brunner wrote:
> Change the detection order to priorize DMI table entries over available
> ACPI entries.
> 
> This makes it more easy for product developers to patch product specific
> handling into the driver.
> Furthermore it allows to simplify the implementation a bit and
> especially to remove the need to force synchronous probing.
> 
> Based on the following commit introduced with v5.10-rc1:
> commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")
> 
> Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
> ---
>  drivers/mfd/kempld-core.c | 23 ++---------------------
>  1 file changed, 2 insertions(+), 21 deletions(-)
> 
> diff --git a/drivers/mfd/kempld-core.c b/drivers/mfd/kempld-core.c
> index 2c9295953c11..aa7f386646a1 100644
> --- a/drivers/mfd/kempld-core.c
> +++ b/drivers/mfd/kempld-core.c
> @@ -125,7 +125,6 @@ static const struct kempld_platform_data kempld_platform_data_generic = {
>  };
>  
>  static struct platform_device *kempld_pdev;
> -static bool kempld_acpi_mode;
>  
>  static int kempld_create_platform_device(const struct dmi_system_id *id)
>  {
> @@ -501,8 +500,6 @@ static int kempld_probe(struct platform_device *pdev)
>  		ret = kempld_get_acpi_data(pdev);
>  		if (ret)
>  			return ret;
> -
> -		kempld_acpi_mode = true;
>  	} else if (kempld_pdev != pdev) {
>  		/*
>  		 * The platform device we are probing is not the one we
> @@ -565,7 +562,6 @@ static struct platform_driver kempld_driver = {
>  	.driver		= {
>  		.name	= "kempld",
>  		.acpi_match_table = ACPI_PTR(kempld_acpi_table),
> -		.probe_type = PROBE_FORCE_SYNCHRONOUS,
>  	},
>  	.probe		= kempld_probe,
>  	.remove		= kempld_remove,
> @@ -884,7 +880,6 @@ MODULE_DEVICE_TABLE(dmi, kempld_dmi_table);
>  static int __init kempld_init(void)
>  {
>  	const struct dmi_system_id *id;
> -	int ret;
>  
>  	if (force_device_id[0]) {
>  		for (id = kempld_dmi_table;
> @@ -894,24 +889,10 @@ static int __init kempld_init(void)
>  					break;
>  		if (id->matches[0].slot == DMI_NONE)
>  			return -ENODEV;
> -	}
> -
> -	ret = platform_driver_register(&kempld_driver);
> -	if (ret)
> -		return ret;
> -
> -	/*
> -	 * With synchronous probing the device should already be probed now.
> -	 * If no device id is forced and also no ACPI definition for the
> -	 * device was found, scan DMI table as fallback.
> -	 *
> -	 * If drivers_autoprobing is disabled and the device is found here,
> -	 * only that device can be bound manually later.
> -	 */
> -	if (!kempld_pdev && !kempld_acpi_mode)
> +	} else

	} else {
>  		dmi_check_system(kempld_dmi_table);
	}

Guenter

>  
> -	return 0;
> +	return platform_driver_register(&kempld_driver);
>  }
>  
>  static void __exit kempld_exit(void)
> 

Re: [PATCH] mfd: kempld-core: Check for DMI definition before ACPI

[Michael Brunner]

Hi Guenter,

thank you for the feedback! checkpatch didn't catch this.
I sent v2 of the patch.

Best regards,
  Michael

On Tue, 2020-11-10 at 06:39 -0800, Guenter Roeck wrote:
> On 11/9/20 11:46 PM, Michael Brunner wrote:
> > Change the detection order to priorize DMI table entries over
> > available
> > ACPI entries.
> > 
> > This makes it more easy for product developers to patch product
> > specific
> > handling into the driver.
> > Furthermore it allows to simplify the implementation a bit and
> > especially to remove the need to force synchronous probing.
> > 
> > Based on the following commit introduced with v5.10-rc1:
> > commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")
> > 
> > Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
> > ---
> >  drivers/mfd/kempld-core.c | 23 ++---------------------
> >  1 file changed, 2 insertions(+), 21 deletions(-)
> > 
> > diff --git a/drivers/mfd/kempld-core.c b/drivers/mfd/kempld-core.c
> > index 2c9295953c11..aa7f386646a1 100644
> > --- a/drivers/mfd/kempld-core.c
> > +++ b/drivers/mfd/kempld-core.c
> > @@ -125,7 +125,6 @@ static const struct kempld_platform_data
> > kempld_platform_data_generic = {
> >  };
> >  
> >  static struct platform_device *kempld_pdev;
> > -static bool kempld_acpi_mode;
> >  
> >  static int kempld_create_platform_device(const struct
> > dmi_system_id *id)
> >  {
> > @@ -501,8 +500,6 @@ static int kempld_probe(struct platform_device
> > *pdev)
> >  		ret = kempld_get_acpi_data(pdev);
> >  		if (ret)
> >  			return ret;
> > -
> > -		kempld_acpi_mode = true;
> >  	} else if (kempld_pdev != pdev) {
> >  		/*
> >  		 * The platform device we are probing is not the one we
> > @@ -565,7 +562,6 @@ static struct platform_driver kempld_driver = {
> >  	.driver		= {
> >  		.name	= "kempld",
> >  		.acpi_match_table = ACPI_PTR(kempld_acpi_table),
> > -		.probe_type = PROBE_FORCE_SYNCHRONOUS,
> >  	},
> >  	.probe		= kempld_probe,
> >  	.remove		= kempld_remove,
> > @@ -884,7 +880,6 @@ MODULE_DEVICE_TABLE(dmi, kempld_dmi_table);
> >  static int __init kempld_init(void)
> >  {
> >  	const struct dmi_system_id *id;
> > -	int ret;
> >  
> >  	if (force_device_id[0]) {
> >  		for (id = kempld_dmi_table;
> > @@ -894,24 +889,10 @@ static int __init kempld_init(void)
> >  					break;
> >  		if (id->matches[0].slot == DMI_NONE)
> >  			return -ENODEV;
> > -	}
> > -
> > -	ret = platform_driver_register(&kempld_driver);
> > -	if (ret)
> > -		return ret;
> > -
> > -	/*
> > -	 * With synchronous probing the device should already be probed
> > now.
> > -	 * If no device id is forced and also no ACPI definition for
> > the
> > -	 * device was found, scan DMI table as fallback.
> > -	 *
> > -	 * If drivers_autoprobing is disabled and the device is found
> > here,
> > -	 * only that device can be bound manually later.
> > -	 */
> > -	if (!kempld_pdev && !kempld_acpi_mode)
> > +	} else
> 
> 	} else {
> >  		dmi_check_system(kempld_dmi_table);
> 	}
> 
> Guenter
> 
> >  
> > -	return 0;
> > +	return platform_driver_register(&kempld_driver);
> >  }
> >  
> >  static void __exit kempld_exit(void)
> > 

Re: [PATCH] mfd: kempld-core: Check for DMI definition before ACPI

[Guenter Roeck]

On Tue, Nov 10, 2020 at 03:41:47PM +0000, Michael Brunner wrote:
> Hi Guenter,
> 
> thank you for the feedback! checkpatch didn't catch this.
> I sent v2 of the patch.
> 

It may have been "demoted" to --strict for whatever reason.

Guenter

> Best regards,
>   Michael
> 
> On Tue, 2020-11-10 at 06:39 -0800, Guenter Roeck wrote:
> > On 11/9/20 11:46 PM, Michael Brunner wrote:
> > > Change the detection order to priorize DMI table entries over
> > > available
> > > ACPI entries.
> > > 
> > > This makes it more easy for product developers to patch product
> > > specific
> > > handling into the driver.
> > > Furthermore it allows to simplify the implementation a bit and
> > > especially to remove the need to force synchronous probing.
> > > 
> > > Based on the following commit introduced with v5.10-rc1:
> > > commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")
> > > 
> > > Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
> > > ---
> > >  drivers/mfd/kempld-core.c | 23 ++---------------------
> > >  1 file changed, 2 insertions(+), 21 deletions(-)
> > > 
> > > diff --git a/drivers/mfd/kempld-core.c b/drivers/mfd/kempld-core.c
> > > index 2c9295953c11..aa7f386646a1 100644
> > > --- a/drivers/mfd/kempld-core.c
> > > +++ b/drivers/mfd/kempld-core.c
> > > @@ -125,7 +125,6 @@ static const struct kempld_platform_data
> > > kempld_platform_data_generic = {
> > >  };
> > >  
> > >  static struct platform_device *kempld_pdev;
> > > -static bool kempld_acpi_mode;
> > >  
> > >  static int kempld_create_platform_device(const struct
> > > dmi_system_id *id)
> > >  {
> > > @@ -501,8 +500,6 @@ static int kempld_probe(struct platform_device
> > > *pdev)
> > >  		ret = kempld_get_acpi_data(pdev);
> > >  		if (ret)
> > >  			return ret;
> > > -
> > > -		kempld_acpi_mode = true;
> > >  	} else if (kempld_pdev != pdev) {
> > >  		/*
> > >  		 * The platform device we are probing is not the one we
> > > @@ -565,7 +562,6 @@ static struct platform_driver kempld_driver = {
> > >  	.driver		= {
> > >  		.name	= "kempld",
> > >  		.acpi_match_table = ACPI_PTR(kempld_acpi_table),
> > > -		.probe_type = PROBE_FORCE_SYNCHRONOUS,
> > >  	},
> > >  	.probe		= kempld_probe,
> > >  	.remove		= kempld_remove,
> > > @@ -884,7 +880,6 @@ MODULE_DEVICE_TABLE(dmi, kempld_dmi_table);
> > >  static int __init kempld_init(void)
> > >  {
> > >  	const struct dmi_system_id *id;
> > > -	int ret;
> > >  
> > >  	if (force_device_id[0]) {
> > >  		for (id = kempld_dmi_table;
> > > @@ -894,24 +889,10 @@ static int __init kempld_init(void)
> > >  					break;
> > >  		if (id->matches[0].slot == DMI_NONE)
> > >  			return -ENODEV;
> > > -	}
> > > -
> > > -	ret = platform_driver_register(&kempld_driver);
> > > -	if (ret)
> > > -		return ret;
> > > -
> > > -	/*
> > > -	 * With synchronous probing the device should already be probed
> > > now.
> > > -	 * If no device id is forced and also no ACPI definition for
> > > the
> > > -	 * device was found, scan DMI table as fallback.
> > > -	 *
> > > -	 * If drivers_autoprobing is disabled and the device is found
> > > here,
> > > -	 * only that device can be bound manually later.
> > > -	 */
> > > -	if (!kempld_pdev && !kempld_acpi_mode)
> > > +	} else
> > 
> > 	} else {
> > >  		dmi_check_system(kempld_dmi_table);
> > 	}
> > 
> > Guenter
> > 
> > >  
> > > -	return 0;
> > > +	return platform_driver_register(&kempld_driver);
> > >  }
> > >  
> > >  static void __exit kempld_exit(void)
> > > 

[PATCH v2] mfd: kempld-core: Check for DMI definition before ACPI

[Michael Brunner]

Change the detection order to priorize DMI table entries over available
ACPI entries.

This makes it more easy for product developers to patch product specific
handling into the driver.
Furthermore it allows to simplify the implementation a bit and
especially to remove the need to force synchronous probing.

Based on the following commit introduced with v5.10-rc1:
commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")

v2: Fixed coding style as suggested by Guenther Roeck

Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
---
 drivers/mfd/kempld-core.c | 24 +++---------------------
 1 file changed, 3 insertions(+), 21 deletions(-)

diff --git a/drivers/mfd/kempld-core.c b/drivers/mfd/kempld-core.c
index 2c9295953c11..ecd26271b9a4 100644
--- a/drivers/mfd/kempld-core.c
+++ b/drivers/mfd/kempld-core.c
@@ -125,7 +125,6 @@ static const struct kempld_platform_data kempld_platform_data_generic = {
 };
 
 static struct platform_device *kempld_pdev;
-static bool kempld_acpi_mode;
 
 static int kempld_create_platform_device(const struct dmi_system_id *id)
 {
@@ -501,8 +500,6 @@ static int kempld_probe(struct platform_device *pdev)
 		ret = kempld_get_acpi_data(pdev);
 		if (ret)
 			return ret;
-
-		kempld_acpi_mode = true;
 	} else if (kempld_pdev != pdev) {
 		/*
 		 * The platform device we are probing is not the one we
@@ -565,7 +562,6 @@ static struct platform_driver kempld_driver = {
 	.driver		= {
 		.name	= "kempld",
 		.acpi_match_table = ACPI_PTR(kempld_acpi_table),
-		.probe_type = PROBE_FORCE_SYNCHRONOUS,
 	},
 	.probe		= kempld_probe,
 	.remove		= kempld_remove,
@@ -884,7 +880,6 @@ MODULE_DEVICE_TABLE(dmi, kempld_dmi_table);
 static int __init kempld_init(void)
 {
 	const struct dmi_system_id *id;
-	int ret;
 
 	if (force_device_id[0]) {
 		for (id = kempld_dmi_table;
@@ -894,24 +889,11 @@ static int __init kempld_init(void)
 					break;
 		if (id->matches[0].slot == DMI_NONE)
 			return -ENODEV;
-	}
-
-	ret = platform_driver_register(&kempld_driver);
-	if (ret)
-		return ret;
-
-	/*
-	 * With synchronous probing the device should already be probed now.
-	 * If no device id is forced and also no ACPI definition for the
-	 * device was found, scan DMI table as fallback.
-	 *
-	 * If drivers_autoprobing is disabled and the device is found here,
-	 * only that device can be bound manually later.
-	 */
-	if (!kempld_pdev && !kempld_acpi_mode)
+	} else {
 		dmi_check_system(kempld_dmi_table);
+	}
 
-	return 0;
+	return platform_driver_register(&kempld_driver);
 }
 
 static void __exit kempld_exit(void)
-- 
2.25.1

Re: [PATCH v2] mfd: kempld-core: Check for DMI definition before ACPI

[Guenter Roeck]

On Tue, Nov 10, 2020 at 03:23:21PM +0000, Michael Brunner wrote:
> Change the detection order to priorize DMI table entries over available
> ACPI entries.
> 
> This makes it more easy for product developers to patch product specific
> handling into the driver.
> Furthermore it allows to simplify the implementation a bit and
> especially to remove the need to force synchronous probing.
> 
> Based on the following commit introduced with v5.10-rc1:
> commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")
> 
> v2: Fixed coding style as suggested by Guenther Roeck
> 

Nit: change logs should be after '---'.

Other than that,

Reviewed-by: Guenter Roeck <linux@roeck-us.net>

Guenter

> Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
> ---
>  drivers/mfd/kempld-core.c | 24 +++---------------------
>  1 file changed, 3 insertions(+), 21 deletions(-)
> 
> diff --git a/drivers/mfd/kempld-core.c b/drivers/mfd/kempld-core.c
> index 2c9295953c11..ecd26271b9a4 100644
> --- a/drivers/mfd/kempld-core.c
> +++ b/drivers/mfd/kempld-core.c
> @@ -125,7 +125,6 @@ static const struct kempld_platform_data kempld_platform_data_generic = {
>  };
>  
>  static struct platform_device *kempld_pdev;
> -static bool kempld_acpi_mode;
>  
>  static int kempld_create_platform_device(const struct dmi_system_id *id)
>  {
> @@ -501,8 +500,6 @@ static int kempld_probe(struct platform_device *pdev)
>  		ret = kempld_get_acpi_data(pdev);
>  		if (ret)
>  			return ret;
> -
> -		kempld_acpi_mode = true;
>  	} else if (kempld_pdev != pdev) {
>  		/*
>  		 * The platform device we are probing is not the one we
> @@ -565,7 +562,6 @@ static struct platform_driver kempld_driver = {
>  	.driver		= {
>  		.name	= "kempld",
>  		.acpi_match_table = ACPI_PTR(kempld_acpi_table),
> -		.probe_type = PROBE_FORCE_SYNCHRONOUS,
>  	},
>  	.probe		= kempld_probe,
>  	.remove		= kempld_remove,
> @@ -884,7 +880,6 @@ MODULE_DEVICE_TABLE(dmi, kempld_dmi_table);
>  static int __init kempld_init(void)
>  {
>  	const struct dmi_system_id *id;
> -	int ret;
>  
>  	if (force_device_id[0]) {
>  		for (id = kempld_dmi_table;
> @@ -894,24 +889,11 @@ static int __init kempld_init(void)
>  					break;
>  		if (id->matches[0].slot == DMI_NONE)
>  			return -ENODEV;
> -	}
> -
> -	ret = platform_driver_register(&kempld_driver);
> -	if (ret)
> -		return ret;
> -
> -	/*
> -	 * With synchronous probing the device should already be probed now.
> -	 * If no device id is forced and also no ACPI definition for the
> -	 * device was found, scan DMI table as fallback.
> -	 *
> -	 * If drivers_autoprobing is disabled and the device is found here,
> -	 * only that device can be bound manually later.
> -	 */
> -	if (!kempld_pdev && !kempld_acpi_mode)
> +	} else {
>  		dmi_check_system(kempld_dmi_table);
> +	}
>  
> -	return 0;
> +	return platform_driver_register(&kempld_driver);
>  }
>  
>  static void __exit kempld_exit(void)
> -- 
> 2.25.1
> 

Re: [PATCH v2] mfd: kempld-core: Check for DMI definition before ACPI

[Lee Jones]

On Tue, 10 Nov 2020, Guenter Roeck wrote:

> On Tue, Nov 10, 2020 at 03:23:21PM +0000, Michael Brunner wrote:
> > Change the detection order to priorize DMI table entries over available
> > ACPI entries.
> > 
> > This makes it more easy for product developers to patch product specific
> > handling into the driver.
> > Furthermore it allows to simplify the implementation a bit and
> > especially to remove the need to force synchronous probing.
> > 
> > Based on the following commit introduced with v5.10-rc1:
> > commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")
> > 
> > v2: Fixed coding style as suggested by Guenther Roeck
> > 
> 
> Nit: change logs should be after '---'.

Right.  Please submit a v3 fixing this.

> Other than that,
> 
> Reviewed-by: Guenter Roeck <linux@roeck-us.net>

-- 
Lee Jones [李琼斯]
Senior Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog

Re: [PATCH v2] mfd: kempld-core: Check for DMI definition before ACPI

[Lee Jones]

On Tue, 10 Nov 2020, Michael Brunner wrote:

> Change the detection order to priorize DMI table entries over available
> ACPI entries.
> 
> This makes it more easy for product developers to patch product specific
> handling into the driver.
> Furthermore it allows to simplify the implementation a bit and
> especially to remove the need to force synchronous probing.


> Based on the following commit introduced with v5.10-rc1:
> commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")

This section should also be removed from the change log.

-- 
Lee Jones [李琼斯]
Senior Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog

Re: [PATCH v2] mfd: kempld-core: Check for DMI definition before ACPI

[Michael Brunner]

On Fri, 2020-11-13 at 10:16 +0000, Lee Jones wrote:
> On Tue, 10 Nov 2020, Michael Brunner wrote:
> 
> > Change the detection order to priorize DMI table entries over
> > available
> > ACPI entries.
> > 
> > This makes it more easy for product developers to patch product
> > specific
> > handling into the driver.
> > Furthermore it allows to simplify the implementation a bit and
> > especially to remove the need to force synchronous probing.
> > Based on the following commit introduced with v5.10-rc1:
> > commit e8299c7313af ("mfd: Add ACPI support to Kontron PLD driver")
> 
> This section should also be removed from the change log.

No problem, just sent v3.

Thanks,
  Michael

[PATCH v3] mfd: kempld-core: Check for DMI definition before ACPI

[Michael Brunner]

Change the detection order to priorize DMI table entries over available
ACPI entries.

This makes it more easy for product developers to patch product specific
handling into the driver.
Furthermore it allows to simplify the implementation a bit and
especially to remove the need to force synchronous probing.

Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
---

v3: Cleaned up comment, added Reviewed-by

 drivers/mfd/kempld-core.c | 24 +++---------------------
 1 file changed, 3 insertions(+), 21 deletions(-)

diff --git a/drivers/mfd/kempld-core.c b/drivers/mfd/kempld-core.c
index 2c9295953c11..ecd26271b9a4 100644
--- a/drivers/mfd/kempld-core.c
+++ b/drivers/mfd/kempld-core.c
@@ -125,7 +125,6 @@ static const struct kempld_platform_data kempld_platform_data_generic = {
 };
 
 static struct platform_device *kempld_pdev;
-static bool kempld_acpi_mode;
 
 static int kempld_create_platform_device(const struct dmi_system_id *id)
 {
@@ -501,8 +500,6 @@ static int kempld_probe(struct platform_device *pdev)
 		ret = kempld_get_acpi_data(pdev);
 		if (ret)
 			return ret;
-
-		kempld_acpi_mode = true;
 	} else if (kempld_pdev != pdev) {
 		/*
 		 * The platform device we are probing is not the one we
@@ -565,7 +562,6 @@ static struct platform_driver kempld_driver = {
 	.driver		= {
 		.name	= "kempld",
 		.acpi_match_table = ACPI_PTR(kempld_acpi_table),
-		.probe_type = PROBE_FORCE_SYNCHRONOUS,
 	},
 	.probe		= kempld_probe,
 	.remove		= kempld_remove,
@@ -884,7 +880,6 @@ MODULE_DEVICE_TABLE(dmi, kempld_dmi_table);
 static int __init kempld_init(void)
 {
 	const struct dmi_system_id *id;
-	int ret;
 
 	if (force_device_id[0]) {
 		for (id = kempld_dmi_table;
@@ -894,24 +889,11 @@ static int __init kempld_init(void)
 					break;
 		if (id->matches[0].slot == DMI_NONE)
 			return -ENODEV;
-	}
-
-	ret = platform_driver_register(&kempld_driver);
-	if (ret)
-		return ret;
-
-	/*
-	 * With synchronous probing the device should already be probed now.
-	 * If no device id is forced and also no ACPI definition for the
-	 * device was found, scan DMI table as fallback.
-	 *
-	 * If drivers_autoprobing is disabled and the device is found here,
-	 * only that device can be bound manually later.
-	 */
-	if (!kempld_pdev && !kempld_acpi_mode)
+	} else {
 		dmi_check_system(kempld_dmi_table);
+	}
 
-	return 0;
+	return platform_driver_register(&kempld_driver);
 }
 
 static void __exit kempld_exit(void)
-- 
2.25.1

Re: [PATCH v3] mfd: kempld-core: Check for DMI definition before ACPI

[Lee Jones]

On Mon, 16 Nov 2020, Michael Brunner wrote:

> Change the detection order to priorize DMI table entries over available
> ACPI entries.
> 
> This makes it more easy for product developers to patch product specific
> handling into the driver.
> Furthermore it allows to simplify the implementation a bit and
> especially to remove the need to force synchronous probing.
> 
> Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
> Reviewed-by: Guenter Roeck <linux@roeck-us.net>
> ---
> 
> v3: Cleaned up comment, added Reviewed-by
> 
>  drivers/mfd/kempld-core.c | 24 +++---------------------
>  1 file changed, 3 insertions(+), 21 deletions(-)

Nit: Just letting you know that checkpatch.pl complains about your
patches, since your From: address does not match your SoB one.

Patch applied though, thanks.

-- 
Lee Jones [李琼斯]
Senior Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog

Re: [PATCH v3] mfd: kempld-core: Check for DMI definition before ACPI

[Michael Brunner]

On Fri, 2020-11-27 at 07:53 +0000, Lee Jones wrote:
> On Mon, 16 Nov 2020, Michael Brunner wrote:
> 
> > Change the detection order to priorize DMI table entries over
> > available
> > ACPI entries.
> > 
> > This makes it more easy for product developers to patch product
> > specific
> > handling into the driver.
> > Furthermore it allows to simplify the implementation a bit and
> > especially to remove the need to force synchronous probing.
> > 
> > Signed-off-by: Michael Brunner <michael.brunner@kontron.com>
> > Reviewed-by: Guenter Roeck <linux@roeck-us.net>
> > ---
> > 
> > v3: Cleaned up comment, added Reviewed-by
> > 
> >  drivers/mfd/kempld-core.c | 24 +++---------------------
> >  1 file changed, 3 insertions(+), 21 deletions(-)
> 
> Nit: Just letting you know that checkpatch.pl complains about your
> patches, since your From: address does not match your SoB one.
> 
> Patch applied though, thanks.

Thanks!

Regarding the From - I guess it is because the upper-case letters. I
will check how to fix this for the next time. In the worst case I guess
I have to adapt the SoB.