[PATCH] clk: sunxi: allow PLL6 clock to be reused

[PATCH] clk: sunxi: allow PLL6 clock to be reused

[Andre Przywara]

Currently we hard-code the base name for the PLL6 clock for both the
sun4i and sun6i variants in the driver (pll6 and pll6x2, respectively).
This unfortunately denies reusing this clock for the H3 and A64 PLL8
clock, which is the same otherwise.
Remove the hard-coded name in the source code and replace it with an
appropriate index into the clock-output-names array in the DT.
This is compatible with all current device trees, since they all carry
the hard-coded base name in there.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
---
Salut Maxime,

does that patch make sense?
It fixes the re-usability problem while still keeping compatibility
with existing DTs and also is pretty small.
If you agree with this, I'll send out a v3 of the A64 series including
this patch and the H3 PLL8 enablement later.
I tested this on my BananaPi, the clk_summary output was identical
with and without this patch.

Cheers,
Andre.

 drivers/clk/sunxi/clk-factors.c | 3 ++-
 drivers/clk/sunxi/clk-factors.h | 1 +
 drivers/clk/sunxi/clk-sunxi.c   | 4 ++--
 3 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/clk/sunxi/clk-factors.c b/drivers/clk/sunxi/clk-factors.c
index ddefe96..7daf01d 100644
--- a/drivers/clk/sunxi/clk-factors.c
+++ b/drivers/clk/sunxi/clk-factors.c
@@ -202,7 +202,8 @@ struct clk *sunxi_factors_register(struct device_node *node,
 	if (data->name)
 		clk_name = data->name;
 	else
-		of_property_read_string(node, "clock-output-names", &clk_name);
+		of_property_read_string_index(node, "clock-output-names",
+					      data->name_idx, &clk_name);
 
 	factors = kzalloc(sizeof(struct clk_factors), GFP_KERNEL);
 	if (!factors)
diff --git a/drivers/clk/sunxi/clk-factors.h b/drivers/clk/sunxi/clk-factors.h
index 1e63c5b..3a7da86 100644
--- a/drivers/clk/sunxi/clk-factors.h
+++ b/drivers/clk/sunxi/clk-factors.h
@@ -36,6 +36,7 @@ struct factors_data {
 	void (*getter)(struct factors_request *req);
 	void (*recalc)(struct factors_request *req);
 	const char *name;
+	int name_idx;
 };
 
 struct clk_factors {
diff --git a/drivers/clk/sunxi/clk-sunxi.c b/drivers/clk/sunxi/clk-sunxi.c
index 99f60ef..53f5927 100644
--- a/drivers/clk/sunxi/clk-sunxi.c
+++ b/drivers/clk/sunxi/clk-sunxi.c
@@ -530,14 +530,14 @@ static const struct factors_data sun4i_pll6_data __initconst = {
 	.enable = 31,
 	.table = &sun4i_pll5_config,
 	.getter = sun4i_get_pll5_factors,
-	.name = "pll6",
+	.name_idx = 2,
 };
 
 static const struct factors_data sun6i_a31_pll6_data __initconst = {
 	.enable = 31,
 	.table = &sun6i_a31_pll6_config,
 	.getter = sun6i_a31_get_pll6_factors,
-	.name = "pll6x2",
+	.name_idx = 1,
 };
 
 static const struct factors_data sun5i_a13_ahb_data __initconst = {
-- 
1.8.4

[PATCH] clk: sunxi: allow PLL6 clock to be reused

[Chen-Yu Tsai]

On Sun, Feb 21, 2016 at 5:39 PM, Andre Przywara <andre.przywara@arm.com> wrote:
> Currently we hard-code the base name for the PLL6 clock for both the
> sun4i and sun6i variants in the driver (pll6 and pll6x2, respectively).
> This unfortunately denies reusing this clock for the H3 and A64 PLL8
> clock, which is the same otherwise.
> Remove the hard-coded name in the source code and replace it with an
> appropriate index into the clock-output-names array in the DT.
> This is compatible with all current device trees, since they all carry
> the hard-coded base name in there.
>
> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
> ---
> Salut Maxime,
>
> does that patch make sense?
> It fixes the re-usability problem while still keeping compatibility
> with existing DTs and also is pretty small.
> If you agree with this, I'll send out a v3 of the A64 series including
> this patch and the H3 PLL8 enablement later.
> I tested this on my BananaPi, the clk_summary output was identical
> with and without this patch.

This look OK to me, except one minor detail below.

> Cheers,
> Andre.
>
>  drivers/clk/sunxi/clk-factors.c | 3 ++-
>  drivers/clk/sunxi/clk-factors.h | 1 +
>  drivers/clk/sunxi/clk-sunxi.c   | 4 ++--
>  3 files changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/clk/sunxi/clk-factors.c b/drivers/clk/sunxi/clk-factors.c
> index ddefe96..7daf01d 100644
> --- a/drivers/clk/sunxi/clk-factors.c
> +++ b/drivers/clk/sunxi/clk-factors.c
> @@ -202,7 +202,8 @@ struct clk *sunxi_factors_register(struct device_node *node,
>         if (data->name)
>                 clk_name = data->name;
>         else
> -               of_property_read_string(node, "clock-output-names", &clk_name);
> +               of_property_read_string_index(node, "clock-output-names",
> +                                             data->name_idx, &clk_name);
>
>         factors = kzalloc(sizeof(struct clk_factors), GFP_KERNEL);
>         if (!factors)
> diff --git a/drivers/clk/sunxi/clk-factors.h b/drivers/clk/sunxi/clk-factors.h
> index 1e63c5b..3a7da86 100644
> --- a/drivers/clk/sunxi/clk-factors.h
> +++ b/drivers/clk/sunxi/clk-factors.h
> @@ -36,6 +36,7 @@ struct factors_data {
>         void (*getter)(struct factors_request *req);
>         void (*recalc)(struct factors_request *req);
>         const char *name;
> +       int name_idx;

I would drop the .name field. It was a bad workaround
due to limitations of the factors clk code at the time
by me. We really shouldn't hard-code the name if we want
to reuse the driver.

Thanks
ChenYu

>  };
>
>  struct clk_factors {
> diff --git a/drivers/clk/sunxi/clk-sunxi.c b/drivers/clk/sunxi/clk-sunxi.c
> index 99f60ef..53f5927 100644
> --- a/drivers/clk/sunxi/clk-sunxi.c
> +++ b/drivers/clk/sunxi/clk-sunxi.c
> @@ -530,14 +530,14 @@ static const struct factors_data sun4i_pll6_data __initconst = {
>         .enable = 31,
>         .table = &sun4i_pll5_config,
>         .getter = sun4i_get_pll5_factors,
> -       .name = "pll6",
> +       .name_idx = 2,
>  };
>
>  static const struct factors_data sun6i_a31_pll6_data __initconst = {
>         .enable = 31,
>         .table = &sun6i_a31_pll6_config,
>         .getter = sun6i_a31_get_pll6_factors,
> -       .name = "pll6x2",
> +       .name_idx = 1,
>  };
>
>  static const struct factors_data sun5i_a13_ahb_data __initconst = {
> --
> 1.8.4
>

[PATCH] clk: sunxi: allow PLL6 clock to be reused

[Andre Przywara]

Hi Chen-Yu,

On 22/02/16 08:08, Chen-Yu Tsai wrote:
> On Sun, Feb 21, 2016 at 5:39 PM, Andre Przywara <andre.przywara@arm.com> wrote:
>> Currently we hard-code the base name for the PLL6 clock for both the
>> sun4i and sun6i variants in the driver (pll6 and pll6x2, respectively).
>> This unfortunately denies reusing this clock for the H3 and A64 PLL8
>> clock, which is the same otherwise.
>> Remove the hard-coded name in the source code and replace it with an
>> appropriate index into the clock-output-names array in the DT.
>> This is compatible with all current device trees, since they all carry
>> the hard-coded base name in there.
>>
>> Signed-off-by: Andre Przywara <andre.przywara@arm.com>
>> ---
>> Salut Maxime,
>>
>> does that patch make sense?
>> It fixes the re-usability problem while still keeping compatibility
>> with existing DTs and also is pretty small.
>> If you agree with this, I'll send out a v3 of the A64 series including
>> this patch and the H3 PLL8 enablement later.
>> I tested this on my BananaPi, the clk_summary output was identical
>> with and without this patch.
> 
> This look OK to me, except one minor detail below.
> 
>> Cheers,
>> Andre.
>>
>>  drivers/clk/sunxi/clk-factors.c | 3 ++-
>>  drivers/clk/sunxi/clk-factors.h | 1 +
>>  drivers/clk/sunxi/clk-sunxi.c   | 4 ++--
>>  3 files changed, 5 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/clk/sunxi/clk-factors.c b/drivers/clk/sunxi/clk-factors.c
>> index ddefe96..7daf01d 100644
>> --- a/drivers/clk/sunxi/clk-factors.c
>> +++ b/drivers/clk/sunxi/clk-factors.c
>> @@ -202,7 +202,8 @@ struct clk *sunxi_factors_register(struct device_node *node,
>>         if (data->name)
>>                 clk_name = data->name;
>>         else
>> -               of_property_read_string(node, "clock-output-names", &clk_name);
>> +               of_property_read_string_index(node, "clock-output-names",
>> +                                             data->name_idx, &clk_name);
>>
>>         factors = kzalloc(sizeof(struct clk_factors), GFP_KERNEL);
>>         if (!factors)
>> diff --git a/drivers/clk/sunxi/clk-factors.h b/drivers/clk/sunxi/clk-factors.h
>> index 1e63c5b..3a7da86 100644
>> --- a/drivers/clk/sunxi/clk-factors.h
>> +++ b/drivers/clk/sunxi/clk-factors.h
>> @@ -36,6 +36,7 @@ struct factors_data {
>>         void (*getter)(struct factors_request *req);
>>         void (*recalc)(struct factors_request *req);
>>         const char *name;
>> +       int name_idx;
> 
> I would drop the .name field. It was a bad workaround
> due to limitations of the factors clk code at the time
> by me. We really shouldn't hard-code the name if we want
> to reuse the driver.

I know what you mean (my first thought, too) and I totally agree, but we
need it still for PLL5, which does not carry the original name in the DT
output names.
So at least this workaround here does not work, I guess we have to come
up with something different - which would be a different patch.
I can take a look into this later.

Thanks,
Andre.

>>  };
>>
>>  struct clk_factors {
>> diff --git a/drivers/clk/sunxi/clk-sunxi.c b/drivers/clk/sunxi/clk-sunxi.c
>> index 99f60ef..53f5927 100644
>> --- a/drivers/clk/sunxi/clk-sunxi.c
>> +++ b/drivers/clk/sunxi/clk-sunxi.c
>> @@ -530,14 +530,14 @@ static const struct factors_data sun4i_pll6_data __initconst = {
>>         .enable = 31,
>>         .table = &sun4i_pll5_config,
>>         .getter = sun4i_get_pll5_factors,
>> -       .name = "pll6",
>> +       .name_idx = 2,
>>  };
>>
>>  static const struct factors_data sun6i_a31_pll6_data __initconst = {
>>         .enable = 31,
>>         .table = &sun6i_a31_pll6_config,
>>         .getter = sun6i_a31_get_pll6_factors,
>> -       .name = "pll6x2",
>> +       .name_idx = 1,
>>  };
>>
>>  static const struct factors_data sun5i_a13_ahb_data __initconst = {
>> --
>> 1.8.4
>>
> 

[PATCH] clk: sunxi: allow PLL6 clock to be reused

[Maxime Ripard]

On Mon, Feb 22, 2016 at 09:38:53AM +0000, Andre Przywara wrote:
> >> diff --git a/drivers/clk/sunxi/clk-factors.h b/drivers/clk/sunxi/clk-factors.h
> >> index 1e63c5b..3a7da86 100644
> >> --- a/drivers/clk/sunxi/clk-factors.h
> >> +++ b/drivers/clk/sunxi/clk-factors.h
> >> @@ -36,6 +36,7 @@ struct factors_data {
> >>         void (*getter)(struct factors_request *req);
> >>         void (*recalc)(struct factors_request *req);
> >>         const char *name;
> >> +       int name_idx;
> > 
> > I would drop the .name field. It was a bad workaround
> > due to limitations of the factors clk code at the time
> > by me. We really shouldn't hard-code the name if we want
> > to reuse the driver.
> 
> I know what you mean (my first thought, too) and I totally agree, but we
> need it still for PLL5, which does not carry the original name in the DT
> output names.
> So at least this workaround here does not work, I guess we have to come
> up with something different - which would be a different patch.
> I can take a look into this later.

Actually, it could be easily worked around. Always take the first
clock output name, take whatever is before '_', and you can remove the
name field entirely.

Jens did something along these lines here:
http://lists.infradead.org/pipermail/linux-arm-kernel/2015-October/379900.html

Maybe we should simply rebase this patch, and remove the part that
falls back on the name field.

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20160225/aba3505b/attachment.sig>

[PATCH] clk: sunxi: allow PLL6 clock to be reused

[André Przywara]

On 25/02/16 18:11, Maxime Ripard wrote:
> On Mon, Feb 22, 2016 at 09:38:53AM +0000, Andre Przywara wrote:
>>>> diff --git a/drivers/clk/sunxi/clk-factors.h b/drivers/clk/sunxi/clk-factors.h
>>>> index 1e63c5b..3a7da86 100644
>>>> --- a/drivers/clk/sunxi/clk-factors.h
>>>> +++ b/drivers/clk/sunxi/clk-factors.h
>>>> @@ -36,6 +36,7 @@ struct factors_data {
>>>>         void (*getter)(struct factors_request *req);
>>>>         void (*recalc)(struct factors_request *req);
>>>>         const char *name;
>>>> +       int name_idx;
>>>
>>> I would drop the .name field. It was a bad workaround
>>> due to limitations of the factors clk code at the time
>>> by me. We really shouldn't hard-code the name if we want
>>> to reuse the driver.
>>
>> I know what you mean (my first thought, too) and I totally agree, but we
>> need it still for PLL5, which does not carry the original name in the DT
>> output names.
>> So at least this workaround here does not work, I guess we have to come
>> up with something different - which would be a different patch.
>> I can take a look into this later.
> 
> Actually, it could be easily worked around. Always take the first
> clock output name, take whatever is before '_', and you can remove the
> name field entirely.

Yeah, I was thinking about that, too, but wanted to give an easy patch a
try first.

> Jens did something along these lines here:
> http://lists.infradead.org/pipermail/linux-arm-kernel/2015-October/379900.html

Indeed, I stumbled upon this by accident today. Looks great on the first
glance: using .self first, then the stub till the underscore. I can
review and test this, if needed.

> Maybe we should simply rebase this patch, and remove the part that
> falls back on the name field.

I am totally fine with this!

Thanks!
Andre.