[PATCH v2 RESEND 2/5] drivers: acpi: acpi_apd: Remove unused device property "is-rv"

[PATCH v2 RESEND 2/5] drivers: acpi: acpi_apd: Remove unused device property "is-rv"

[Ajit Kumar Pandey]

Initially "is-rv" device property is added for 48MHz fixed clock
support on Raven or RV architecture. It's unused now as we moved
to clock config based selection to extend such support on others
architecture. This change removed unused code from acpi driver.

Signed-off-by: Ajit Kumar Pandey <AjitKumar.Pandey@amd.com>
---
 drivers/acpi/acpi_apd.c               | 3 ---
 include/linux/platform_data/clk-fch.h | 1 -
 2 files changed, 4 deletions(-)

diff --git a/drivers/acpi/acpi_apd.c b/drivers/acpi/acpi_apd.c
index 6e02448d15d9..6913e9712852 100644
--- a/drivers/acpi/acpi_apd.c
+++ b/drivers/acpi/acpi_apd.c
@@ -87,9 +87,6 @@ static int fch_misc_setup(struct apd_private_data *pdata)
 	if (ret < 0)
 		return -ENOENT;
 
-	if (!acpi_dev_get_property(adev, "is-rv", ACPI_TYPE_INTEGER, &obj))
-		clk_data->is_rv = obj->integer.value;
-
 	list_for_each_entry(rentry, &resource_list, node) {
 		clk_data->base = devm_ioremap(&adev->dev, rentry->res->start,
 					      resource_size(rentry->res));
diff --git a/include/linux/platform_data/clk-fch.h b/include/linux/platform_data/clk-fch.h
index b9f682459f08..850ca776156d 100644
--- a/include/linux/platform_data/clk-fch.h
+++ b/include/linux/platform_data/clk-fch.h
@@ -12,7 +12,6 @@
 
 struct fch_clk_data {
 	void __iomem *base;
-	u32 is_rv;
 };
 
 #endif /* __CLK_FCH_H */
-- 
2.25.1

[PATCH v2 RESEND 3/5] ACPI: APD: Add a fmw property clk-name

[Ajit Kumar Pandey]

Add a new device property to fetch clk-name from firmware.

Signed-off-by: Ajit Kumar Pandey <AjitKumar.Pandey@amd.com>
---
 drivers/acpi/acpi_apd.c               | 10 ++++++++++
 include/linux/platform_data/clk-fch.h |  1 +
 2 files changed, 11 insertions(+)

diff --git a/drivers/acpi/acpi_apd.c b/drivers/acpi/acpi_apd.c
index 6913e9712852..2b958b426b03 100644
--- a/drivers/acpi/acpi_apd.c
+++ b/drivers/acpi/acpi_apd.c
@@ -87,6 +87,16 @@ static int fch_misc_setup(struct apd_private_data *pdata)
 	if (ret < 0)
 		return -ENOENT;
 
+	if (!acpi_dev_get_property(adev, "clk-name", ACPI_TYPE_STRING, &obj)) {
+		clk_data->name = devm_kzalloc(&adev->dev, obj->string.length,
+					      GFP_KERNEL);
+
+		strcpy(clk_data->name, obj->string.pointer);
+	} else {
+		/* Set default name to mclk if entry missing in firmware */
+		clk_data->name = "mclk";
+	}
+
 	list_for_each_entry(rentry, &resource_list, node) {
 		clk_data->base = devm_ioremap(&adev->dev, rentry->res->start,
 					      resource_size(rentry->res));
diff --git a/include/linux/platform_data/clk-fch.h b/include/linux/platform_data/clk-fch.h
index 850ca776156d..11a2a23fd9b2 100644
--- a/include/linux/platform_data/clk-fch.h
+++ b/include/linux/platform_data/clk-fch.h
@@ -12,6 +12,7 @@
 
 struct fch_clk_data {
 	void __iomem *base;
+	char *name;
 };
 
 #endif /* __CLK_FCH_H */
-- 
2.25.1

Re: [PATCH v2 RESEND 2/5] drivers: acpi: acpi_apd: Remove unused device property "is-rv"

[Rafael J. Wysocki]

On Fri, Oct 29, 2021 at 8:33 AM Ajit Kumar Pandey
<AjitKumar.Pandey@amd.com> wrote:
>
> Initially "is-rv" device property is added for 48MHz fixed clock
> support on Raven or RV architecture. It's unused now as we moved
> to clock config based selection to extend such support on others
> architecture. This change removed unused code from acpi driver.
>
> Signed-off-by: Ajit Kumar Pandey <AjitKumar.Pandey@amd.com>

Is this and the [3/5] applicable without the [1/5]?

If so, I can pick them up once they are approved by some other AMD
folks, preferably Mario.

> ---
>  drivers/acpi/acpi_apd.c               | 3 ---
>  include/linux/platform_data/clk-fch.h | 1 -
>  2 files changed, 4 deletions(-)
>
> diff --git a/drivers/acpi/acpi_apd.c b/drivers/acpi/acpi_apd.c
> index 6e02448d15d9..6913e9712852 100644
> --- a/drivers/acpi/acpi_apd.c
> +++ b/drivers/acpi/acpi_apd.c
> @@ -87,9 +87,6 @@ static int fch_misc_setup(struct apd_private_data *pdata)
>         if (ret < 0)
>                 return -ENOENT;
>
> -       if (!acpi_dev_get_property(adev, "is-rv", ACPI_TYPE_INTEGER, &obj))
> -               clk_data->is_rv = obj->integer.value;
> -
>         list_for_each_entry(rentry, &resource_list, node) {
>                 clk_data->base = devm_ioremap(&adev->dev, rentry->res->start,
>                                               resource_size(rentry->res));
> diff --git a/include/linux/platform_data/clk-fch.h b/include/linux/platform_data/clk-fch.h
> index b9f682459f08..850ca776156d 100644
> --- a/include/linux/platform_data/clk-fch.h
> +++ b/include/linux/platform_data/clk-fch.h
> @@ -12,7 +12,6 @@
>
>  struct fch_clk_data {
>         void __iomem *base;
> -       u32 is_rv;
>  };
>
>  #endif /* __CLK_FCH_H */
> --
> 2.25.1
>

Re: [PATCH v2 RESEND 2/5] drivers: acpi: acpi_apd: Remove unused device property "is-rv"

[Ajit Kumar Pandey]

On 11/6/2021 3:52 AM, Stephen Boyd wrote:
> [CAUTION: External Email]
> 
> Quoting Rafael J. Wysocki (2021-11-05 08:02:26)
>> On Fri, Oct 29, 2021 at 8:33 AM Ajit Kumar Pandey
>> <AjitKumar.Pandey@amd.com> wrote:
>>>
>>> Initially "is-rv" device property is added for 48MHz fixed clock
>>> support on Raven or RV architecture. It's unused now as we moved
>>> to clock config based selection to extend such support on others
>>> architecture. This change removed unused code from acpi driver.
>>>
>>> Signed-off-by: Ajit Kumar Pandey <AjitKumar.Pandey@amd.com>
>>
>> Is this and the [3/5] applicable without the [1/5]?
>>
>> If so, I can pick them up once they are approved by some other AMD
>> folks, preferably Mario.
> 
> I'm still lost why we're putting this decision to have some clk or not
> into Kconfig.
> 
I agree it's kind of weird to select clock based on Kconfig but only 
reason to put Kconfig based selection is that this clk support required 
on some older platforms where BIOS or bootloader change isn't feasible.
We have only one consumer for this 48MHzclk i.e audio drivers and idea 
is that this config shouldn't be user-selectable one and automatically
select from audio machine driver which uses this clk.

RE: [PATCH v2 RESEND 2/5] drivers: acpi: acpi_apd: Remove unused device property "is-rv"

[Limonciello, Mario]

[Public]



> -----Original Message-----
> From: Rafael J. Wysocki <rafael@kernel.org>
> Sent: Friday, November 5, 2021 10:02
> To: Pandey, Ajit Kumar <AjitKumar.Pandey@amd.com>
> Cc: Stephen Boyd <sboyd@kernel.org>; linux-clk <linux-clk@vger.kernel.org>;
> Mukunda, Vijendar <Vijendar.Mukunda@amd.com>; Deucher, Alexander
> <Alexander.Deucher@amd.com>; Hiregoudar, Basavaraj
> <Basavaraj.Hiregoudar@amd.com>; Dommati, Sunil-kumar <Sunil-
> kumar.Dommati@amd.com>; Rafael J. Wysocki <rafael@kernel.org>; Len
> Brown <lenb@kernel.org>; open list:ACPI <linux-acpi@vger.kernel.org>; open
> list <linux-kernel@vger.kernel.org>; Limonciello, Mario
> <Mario.Limonciello@amd.com>
> Subject: Re: [PATCH v2 RESEND 2/5] drivers: acpi: acpi_apd: Remove unused
> device property "is-rv"
> 
> On Fri, Oct 29, 2021 at 8:33 AM Ajit Kumar Pandey
> <AjitKumar.Pandey@amd.com> wrote:
> >
> > Initially "is-rv" device property is added for 48MHz fixed clock
> > support on Raven or RV architecture. It's unused now as we moved
> > to clock config based selection to extend such support on others
> > architecture. This change removed unused code from acpi driver.
> >
> > Signed-off-by: Ajit Kumar Pandey <AjitKumar.Pandey@amd.com>
> 
> Is this and the [3/5] applicable without the [1/5]?
> 
> If so, I can pick them up once they are approved by some other AMD
> folks, preferably Mario.

I had some offline conversations with Ajit, and gave them some other
ideas instead of Kconfig based clock settings.  Let's wait for the result
of those tries.

> 
> > ---
> >  drivers/acpi/acpi_apd.c               | 3 ---
> >  include/linux/platform_data/clk-fch.h | 1 -
> >  2 files changed, 4 deletions(-)
> >
> > diff --git a/drivers/acpi/acpi_apd.c b/drivers/acpi/acpi_apd.c
> > index 6e02448d15d9..6913e9712852 100644
> > --- a/drivers/acpi/acpi_apd.c
> > +++ b/drivers/acpi/acpi_apd.c
> > @@ -87,9 +87,6 @@ static int fch_misc_setup(struct apd_private_data
> *pdata)
> >         if (ret < 0)
> >                 return -ENOENT;
> >
> > -       if (!acpi_dev_get_property(adev, "is-rv", ACPI_TYPE_INTEGER, &obj))
> > -               clk_data->is_rv = obj->integer.value;
> > -
> >         list_for_each_entry(rentry, &resource_list, node) {
> >                 clk_data->base = devm_ioremap(&adev->dev, rentry->res->start,
> >                                               resource_size(rentry->res));
> > diff --git a/include/linux/platform_data/clk-fch.h
> b/include/linux/platform_data/clk-fch.h
> > index b9f682459f08..850ca776156d 100644
> > --- a/include/linux/platform_data/clk-fch.h
> > +++ b/include/linux/platform_data/clk-fch.h
> > @@ -12,7 +12,6 @@
> >
> >  struct fch_clk_data {
> >         void __iomem *base;
> > -       u32 is_rv;
> >  };
> >
> >  #endif /* __CLK_FCH_H */
> > --
> > 2.25.1
> >