[v4,1/4] gpu: drm: meson: Use devm_regulator_*get_enable*()

  Simplify using the devm_regulator_get_enable_optional(). Also drop the
seemingly unused struct member 'hdmi_supply'.

Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>

---
v3 => v4:
- split meson part to own patch

RFCv1 => v2:
- Change also sii902x to use devm_regulator_bulk_get_enable()

Please note - this is only compile-tested due to the lack of HW. Careful
review and testing is _highly_ appreciated.
---
 drivers/gpu/drm/meson/meson_dw_hdmi.c | 23 +++--------------------
 1 file changed, 3 insertions(+), 20 deletions(-)
  

Hi Matti,

On Fri, Oct 21, 2022 at 04:18:01PM +0300, Matti Vaittinen wrote:
> Simplify using the devm_regulator_get_enable_optional(). Also drop the
> seemingly unused struct member 'hdmi_supply'.
> 
> Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>
> 
> ---
> v3 => v4:
> - split meson part to own patch
> 
> RFCv1 => v2:
> - Change also sii902x to use devm_regulator_bulk_get_enable()
> 
> Please note - this is only compile-tested due to the lack of HW. Careful
> review and testing is _highly_ appreciated.
> ---
>  drivers/gpu/drm/meson/meson_dw_hdmi.c | 23 +++--------------------
>  1 file changed, 3 insertions(+), 20 deletions(-)
> 
> diff --git a/drivers/gpu/drm/meson/meson_dw_hdmi.c b/drivers/gpu/drm/meson/meson_dw_hdmi.c
> index 5cd2b2ebbbd3..7642f740272b 100644
> --- a/drivers/gpu/drm/meson/meson_dw_hdmi.c
> +++ b/drivers/gpu/drm/meson/meson_dw_hdmi.c
> @@ -140,7 +140,6 @@ struct meson_dw_hdmi {
>  	struct reset_control *hdmitx_apb;
>  	struct reset_control *hdmitx_ctrl;
>  	struct reset_control *hdmitx_phy;
> -	struct regulator *hdmi_supply;
>  	u32 irq_stat;
>  	struct dw_hdmi *hdmi;
>  	struct drm_bridge *bridge;
> @@ -665,11 +664,6 @@ static void meson_dw_hdmi_init(struct meson_dw_hdmi *meson_dw_hdmi)
>  
>  }
>  
> -static void meson_disable_regulator(void *data)
> -{
> -	regulator_disable(data);
> -}
> -
>  static void meson_disable_clk(void *data)
>  {
>  	clk_disable_unprepare(data);
> @@ -723,20 +717,9 @@ static int meson_dw_hdmi_bind(struct device *dev, struct device *master,
>  	meson_dw_hdmi->data = match;
>  	dw_plat_data = &meson_dw_hdmi->dw_plat_data;
>  
> -	meson_dw_hdmi->hdmi_supply = devm_regulator_get_optional(dev, "hdmi");
> -	if (IS_ERR(meson_dw_hdmi->hdmi_supply)) {
> -		if (PTR_ERR(meson_dw_hdmi->hdmi_supply) == -EPROBE_DEFER)
> -			return -EPROBE_DEFER;
> -		meson_dw_hdmi->hdmi_supply = NULL;
> -	} else {
> -		ret = regulator_enable(meson_dw_hdmi->hdmi_supply);
> -		if (ret)
> -			return ret;
> -		ret = devm_add_action_or_reset(dev, meson_disable_regulator,
> -					       meson_dw_hdmi->hdmi_supply);
> -		if (ret)
> -			return ret;
> -	}
> +	ret = devm_regulator_get_enable_optional(dev, "hdmi");
> +	if (ret != -ENODEV)
> +		return ret;

As noted in the review of the series that introduced
devm_regulator_get_enable_optional(), the right thing to do is to
implement runtime PM in this driver to avoid wasting power.

>  
>  	meson_dw_hdmi->hdmitx_apb = devm_reset_control_get_exclusive(dev,
>  						"hdmitx_apb");
  

Hi,

On 21/10/2022 17:02, Laurent Pinchart wrote:
> Hi Matti,
> 
> On Fri, Oct 21, 2022 at 04:18:01PM +0300, Matti Vaittinen wrote:
>> Simplify using the devm_regulator_get_enable_optional(). Also drop the
>> seemingly unused struct member 'hdmi_supply'.
>>
>> Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>
>>
>> ---
>> v3 => v4:
>> - split meson part to own patch
>>
>> RFCv1 => v2:
>> - Change also sii902x to use devm_regulator_bulk_get_enable()
>>
>> Please note - this is only compile-tested due to the lack of HW. Careful
>> review and testing is _highly_ appreciated.
>> ---
>>   drivers/gpu/drm/meson/meson_dw_hdmi.c | 23 +++--------------------
>>   1 file changed, 3 insertions(+), 20 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/meson/meson_dw_hdmi.c b/drivers/gpu/drm/meson/meson_dw_hdmi.c
>> index 5cd2b2ebbbd3..7642f740272b 100644
>> --- a/drivers/gpu/drm/meson/meson_dw_hdmi.c
>> +++ b/drivers/gpu/drm/meson/meson_dw_hdmi.c
>> @@ -140,7 +140,6 @@ struct meson_dw_hdmi {
>>   	struct reset_control *hdmitx_apb;
>>   	struct reset_control *hdmitx_ctrl;
>>   	struct reset_control *hdmitx_phy;
>> -	struct regulator *hdmi_supply;
>>   	u32 irq_stat;
>>   	struct dw_hdmi *hdmi;
>>   	struct drm_bridge *bridge;
>> @@ -665,11 +664,6 @@ static void meson_dw_hdmi_init(struct meson_dw_hdmi *meson_dw_hdmi)
>>   
>>   }
>>   
>> -static void meson_disable_regulator(void *data)
>> -{
>> -	regulator_disable(data);
>> -}
>> -
>>   static void meson_disable_clk(void *data)
>>   {
>>   	clk_disable_unprepare(data);
>> @@ -723,20 +717,9 @@ static int meson_dw_hdmi_bind(struct device *dev, struct device *master,
>>   	meson_dw_hdmi->data = match;
>>   	dw_plat_data = &meson_dw_hdmi->dw_plat_data;
>>   
>> -	meson_dw_hdmi->hdmi_supply = devm_regulator_get_optional(dev, "hdmi");
>> -	if (IS_ERR(meson_dw_hdmi->hdmi_supply)) {
>> -		if (PTR_ERR(meson_dw_hdmi->hdmi_supply) == -EPROBE_DEFER)
>> -			return -EPROBE_DEFER;
>> -		meson_dw_hdmi->hdmi_supply = NULL;
>> -	} else {
>> -		ret = regulator_enable(meson_dw_hdmi->hdmi_supply);
>> -		if (ret)
>> -			return ret;
>> -		ret = devm_add_action_or_reset(dev, meson_disable_regulator,
>> -					       meson_dw_hdmi->hdmi_supply);
>> -		if (ret)
>> -			return ret;
>> -	}
>> +	ret = devm_regulator_get_enable_optional(dev, "hdmi");
>> +	if (ret != -ENODEV)
>> +		return ret;
> 
> As noted in the review of the series that introduced
> devm_regulator_get_enable_optional(), the right thing to do is to
> implement runtime PM in this driver to avoid wasting power.

While I agree, it's not really the same level of effort as this patch
should be functionally equivalent.

> 
>>   
>>   	meson_dw_hdmi->hdmitx_apb = devm_reset_control_get_exclusive(dev,
>>   						"hdmitx_apb");
> 

Neil
  

On 10/21/22 18:29, Neil Armstrong wrote:
> Hi,
> 
> On 21/10/2022 17:02, Laurent Pinchart wrote:
>> Hi Matti,
>>
>> On Fri, Oct 21, 2022 at 04:18:01PM +0300, Matti Vaittinen wrote:
>>> Simplify using the devm_regulator_get_enable_optional(). Also drop the
>>> seemingly unused struct member 'hdmi_supply'.
>>>
>>> Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>
>>>
>>> ---
>>> v3 => v4:
>>> - split meson part to own patch
>>>
>>> RFCv1 => v2:
>>> - Change also sii902x to use devm_regulator_bulk_get_enable()
>>>
>>> Please note - this is only compile-tested due to the lack of HW. Careful
>>> review and testing is _highly_ appreciated.
>>> ---

//Snip

>>
>> As noted in the review of the series that introduced
>> devm_regulator_get_enable_optional(), the right thing to do is to
>> implement runtime PM in this driver to avoid wasting power.
> 
> While I agree, it's not really the same level of effort as this patch
> should be functionally equivalent.
> 

Exactly. As I wrote, I do not have the HW to test this change. This 
intends to bring no functional changes. It is just a minor 
simplification while also making it harder to create a bug with the 
regulator control. (Registering the devm_action and leaving the handle 
to the regulator is more fragile than using this new API which does not 
give user the handle).

I am in no way against someone further improving the functionality here 
(by adding runtime PM) but this someone is not me. Yet, I don't see how 
this patch prevents runtime PM being implemented? The first thing that 
needs to be done is removing the devm-action assuming someone did 
implement power-saving by disabling the regulator(s) at runtime. After 
this patch is applied, the action removal is automatically a necessity 
in order to get the handle for regulator control.

I know this helper is not preferred by all but it is still safer than 
the current code which registers the action while allowing the regulator 
control.

Yours
	-- Matti

-- 
Matti Vaittinen
Linux kernel developer at ROHM Semiconductors
Oulu Finland

~~ When things go utterly wrong vim users can always type :help! ~~
  

Hi dee Ho Laurent & All,

On 10/21/22 18:29, Neil Armstrong wrote:
> Hi,
> 
> On 21/10/2022 17:02, Laurent Pinchart wrote:
>> Hi Matti,
>>
>> On Fri, Oct 21, 2022 at 04:18:01PM +0300, Matti Vaittinen wrote:
>>> Simplify using the devm_regulator_get_enable_optional(). Also drop the
>>> seemingly unused struct member 'hdmi_supply'.
>>>
>>> Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>
>>>
>>> ---
>>> v3 => v4:
>>> - split meson part to own patch
>>>
>>> RFCv1 => v2:
>>> - Change also sii902x to use devm_regulator_bulk_get_enable()
>>>
>>> Please note - this is only compile-tested due to the lack of HW. Careful
>>> review and testing is _highly_ appreciated.
>>> ---
>>>   drivers/gpu/drm/meson/meson_dw_hdmi.c | 23 +++--------------------
>>>   1 file changed, 3 insertions(+), 20 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/meson/meson_dw_hdmi.c 
>>> b/drivers/gpu/drm/meson/meson_dw_hdmi.c
>>> index 5cd2b2ebbbd3..7642f740272b 100644
>>> --- a/drivers/gpu/drm/meson/meson_dw_hdmi.c
>>> +++ b/drivers/gpu/drm/meson/meson_dw_hdmi.c
>>> @@ -140,7 +140,6 @@ struct meson_dw_hdmi {
>>>       struct reset_control *hdmitx_apb;
>>>       struct reset_control *hdmitx_ctrl;
>>>       struct reset_control *hdmitx_phy;
>>> -    struct regulator *hdmi_supply;
>>>       u32 irq_stat;
>>>       struct dw_hdmi *hdmi;
>>>       struct drm_bridge *bridge;
>>> @@ -665,11 +664,6 @@ static void meson_dw_hdmi_init(struct 
>>> meson_dw_hdmi *meson_dw_hdmi)
>>>   }
>>> -static void meson_disable_regulator(void *data)
>>> -{
>>> -    regulator_disable(data);
>>> -}
>>> -
>>>   static void meson_disable_clk(void *data)
>>>   {
>>>       clk_disable_unprepare(data);
>>> @@ -723,20 +717,9 @@ static int meson_dw_hdmi_bind(struct device 
>>> *dev, struct device *master,
>>>       meson_dw_hdmi->data = match;
>>>       dw_plat_data = &meson_dw_hdmi->dw_plat_data;
>>> -    meson_dw_hdmi->hdmi_supply = devm_regulator_get_optional(dev, 
>>> "hdmi");
>>> -    if (IS_ERR(meson_dw_hdmi->hdmi_supply)) {
>>> -        if (PTR_ERR(meson_dw_hdmi->hdmi_supply) == -EPROBE_DEFER)
>>> -            return -EPROBE_DEFER;
>>> -        meson_dw_hdmi->hdmi_supply = NULL;
>>> -    } else {
>>> -        ret = regulator_enable(meson_dw_hdmi->hdmi_supply);
>>> -        if (ret)
>>> -            return ret;
>>> -        ret = devm_add_action_or_reset(dev, meson_disable_regulator,
>>> -                           meson_dw_hdmi->hdmi_supply);
>>> -        if (ret)
>>> -            return ret;
>>> -    }
>>> +    ret = devm_regulator_get_enable_optional(dev, "hdmi");
>>> +    if (ret != -ENODEV)
>>> +        return ret;
>>
>> As noted in the review of the series that introduced
>> devm_regulator_get_enable_optional(), the right thing to do is to
>> implement runtime PM in this driver to avoid wasting power.
> 
> While I agree, it's not really the same level of effort as this patch
> should be functionally equivalent.

While we all agree that power savings gained by runtime PM would be 
great - I don't think we should stop minor improvements while waiting 
for that to magically happen ;)

I like the saying I first heard from Jonathan Cameron - "Don't let the 
perfect be an enemy of good".

After that being said, should I re-spin this or just drop it? (I am 
cleaning up my local git from old stuff, and these are about to vanish 
from my books).

Yours,
	-- Matti

-- 
Matti Vaittinen
Linux kernel developer at ROHM Semiconductors
Oulu Finland

~~ When things go utterly wrong vim users can always type :help! ~~
  

On 16/11/2022 13:02, Vaittinen, Matti wrote:
> Hi dee Ho Laurent & All,
> 
> On 10/21/22 18:29, Neil Armstrong wrote:
>> Hi,
>>
>> On 21/10/2022 17:02, Laurent Pinchart wrote:
>>> Hi Matti,
>>>
>>> On Fri, Oct 21, 2022 at 04:18:01PM +0300, Matti Vaittinen wrote:
>>>> Simplify using the devm_regulator_get_enable_optional(). Also drop the
>>>> seemingly unused struct member 'hdmi_supply'.
>>>>
>>>> Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>
>>>>
>>>> ---
>>>> v3 => v4:
>>>> - split meson part to own patch
>>>>
>>>> RFCv1 => v2:
>>>> - Change also sii902x to use devm_regulator_bulk_get_enable()
>>>>
>>>> Please note - this is only compile-tested due to the lack of HW. Careful
>>>> review and testing is _highly_ appreciated.
>>>> ---
>>>>    drivers/gpu/drm/meson/meson_dw_hdmi.c | 23 +++--------------------
>>>>    1 file changed, 3 insertions(+), 20 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/meson/meson_dw_hdmi.c
>>>> b/drivers/gpu/drm/meson/meson_dw_hdmi.c
>>>> index 5cd2b2ebbbd3..7642f740272b 100644
>>>> --- a/drivers/gpu/drm/meson/meson_dw_hdmi.c
>>>> +++ b/drivers/gpu/drm/meson/meson_dw_hdmi.c
>>>> @@ -140,7 +140,6 @@ struct meson_dw_hdmi {
>>>>        struct reset_control *hdmitx_apb;
>>>>        struct reset_control *hdmitx_ctrl;
>>>>        struct reset_control *hdmitx_phy;
>>>> -    struct regulator *hdmi_supply;
>>>>        u32 irq_stat;
>>>>        struct dw_hdmi *hdmi;
>>>>        struct drm_bridge *bridge;
>>>> @@ -665,11 +664,6 @@ static void meson_dw_hdmi_init(struct
>>>> meson_dw_hdmi *meson_dw_hdmi)
>>>>    }
>>>> -static void meson_disable_regulator(void *data)
>>>> -{
>>>> -    regulator_disable(data);
>>>> -}
>>>> -
>>>>    static void meson_disable_clk(void *data)
>>>>    {
>>>>        clk_disable_unprepare(data);
>>>> @@ -723,20 +717,9 @@ static int meson_dw_hdmi_bind(struct device
>>>> *dev, struct device *master,
>>>>        meson_dw_hdmi->data = match;
>>>>        dw_plat_data = &meson_dw_hdmi->dw_plat_data;
>>>> -    meson_dw_hdmi->hdmi_supply = devm_regulator_get_optional(dev,
>>>> "hdmi");
>>>> -    if (IS_ERR(meson_dw_hdmi->hdmi_supply)) {
>>>> -        if (PTR_ERR(meson_dw_hdmi->hdmi_supply) == -EPROBE_DEFER)
>>>> -            return -EPROBE_DEFER;
>>>> -        meson_dw_hdmi->hdmi_supply = NULL;
>>>> -    } else {
>>>> -        ret = regulator_enable(meson_dw_hdmi->hdmi_supply);
>>>> -        if (ret)
>>>> -            return ret;
>>>> -        ret = devm_add_action_or_reset(dev, meson_disable_regulator,
>>>> -                           meson_dw_hdmi->hdmi_supply);
>>>> -        if (ret)
>>>> -            return ret;
>>>> -    }
>>>> +    ret = devm_regulator_get_enable_optional(dev, "hdmi");
>>>> +    if (ret != -ENODEV)
>>>> +        return ret;
>>>
>>> As noted in the review of the series that introduced
>>> devm_regulator_get_enable_optional(), the right thing to do is to
>>> implement runtime PM in this driver to avoid wasting power.
>>
>> While I agree, it's not really the same level of effort as this patch
>> should be functionally equivalent.
> 
> While we all agree that power savings gained by runtime PM would be
> great - I don't think we should stop minor improvements while waiting
> for that to magically happen ;)
> 
> I like the saying I first heard from Jonathan Cameron - "Don't let the
> perfect be an enemy of good".
> 
> After that being said, should I re-spin this or just drop it? (I am
> cleaning up my local git from old stuff, and these are about to vanish
> from my books).

I'm ok with you, please re-spin it.

Neil

> 
> Yours,
> 	-- Matti
>