[PATCH 0/8] pwm: Get rid of pwm_[sg]et_chip_data()

[PATCH 0/8] pwm: Get rid of pwm_[sg]et_chip_data()

[Uwe Kleine-König]

Hello,

the semantic of pwm_[sg]et_chip_data() overlaps with that of the
standard dev_[sg]et_drvdata() functions. Also as noted by George Stark
there is a problem in pwm-sti that relies on chipdata being available
even after a pwm_put (which clears chipdata).

To improve the situation this series converts all drivers that make use
of pwm chipdata to use well-known driver data instead and drop chipdata
support from the pwm framework.

Best regards
Uwe

Uwe Kleine-König (8):
  pwm: berlin: Put channel config into driver data
  pwm: samsung: Put channel data into driver data
  pwm: jz4740: Put per-channel clk into driver data
  pwm: lp3943: Drop usage of pwm_[gs]et_chip_data()
  pwm: renesas: Drop usage of pwm_[gs]et_chip_data()
  pwm: sti: Reduce number of allocations and drop usage of chip_data
  pwm: cros-ec: Put per channel data into driver data
  pwm: Drop pwm_[sg]et_chip_data()

 drivers/pwm/core.c            | 31 -----------------------------
 drivers/pwm/pwm-berlin.c      | 37 ++++++-----------------------------
 drivers/pwm/pwm-cros-ec.c     | 31 +++++++----------------------
 drivers/pwm/pwm-jz4740.c      | 11 +++++++----
 drivers/pwm/pwm-lp3943.c      | 21 +++++++-------------
 drivers/pwm/pwm-renesas-tpu.c | 22 ++++++++++-----------
 drivers/pwm/pwm-samsung.c     | 20 +++++--------------
 drivers/pwm/pwm-sti.c         | 29 +++++++++++++--------------
 include/linux/pwm.h           | 14 -------------
 9 files changed, 57 insertions(+), 159 deletions(-)


base-commit: 92554cdd428fce212d2a71a06939e7cab90f7c77
-- 
2.39.2

[PATCH 3/8] pwm: jz4740: Put per-channel clk into driver data

[Uwe Kleine-König]

Stop using chip_data which is about to go away. Instead track the
per-channel clk in struct jz4740_pwm_chip.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
---
 drivers/pwm/pwm-jz4740.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/pwm/pwm-jz4740.c b/drivers/pwm/pwm-jz4740.c
index 3b7067f6cd0d..e0a57d71a60c 100644
--- a/drivers/pwm/pwm-jz4740.c
+++ b/drivers/pwm/pwm-jz4740.c
@@ -27,6 +27,7 @@ struct soc_info {
 struct jz4740_pwm_chip {
 	struct pwm_chip chip;
 	struct regmap *map;
+	struct clk *clk[];
 };
 
 static inline struct jz4740_pwm_chip *to_jz4740(struct pwm_chip *chip)
@@ -70,14 +71,15 @@ static int jz4740_pwm_request(struct pwm_chip *chip, struct pwm_device *pwm)
 		return err;
 	}
 
-	pwm_set_chip_data(pwm, clk);
+	jz->clk[pwm->hwpwm] = clk;
 
 	return 0;
 }
 
 static void jz4740_pwm_free(struct pwm_chip *chip, struct pwm_device *pwm)
 {
-	struct clk *clk = pwm_get_chip_data(pwm);
+	struct jz4740_pwm_chip *jz = to_jz4740(chip);
+	struct clk *clk = jz->clk[pwm->hwpwm];
 
 	clk_disable_unprepare(clk);
 	clk_put(clk);
@@ -123,7 +125,7 @@ static int jz4740_pwm_apply(struct pwm_chip *chip, struct pwm_device *pwm,
 {
 	struct jz4740_pwm_chip *jz4740 = to_jz4740(pwm->chip);
 	unsigned long long tmp = 0xffffull * NSEC_PER_SEC;
-	struct clk *clk = pwm_get_chip_data(pwm);
+	struct clk *clk = jz4740->clk[pwm->hwpwm];
 	unsigned long period, duty;
 	long rate;
 	int err;
@@ -229,7 +231,8 @@ static int jz4740_pwm_probe(struct platform_device *pdev)
 	if (!info)
 		return -EINVAL;
 
-	jz4740 = devm_kzalloc(dev, sizeof(*jz4740), GFP_KERNEL);
+	jz4740 = devm_kzalloc(dev, sizeof(*jz4740) + info->num_pwms * sizeof(jz4740->clk[0]),
+			      GFP_KERNEL);
 	if (!jz4740)
 		return -ENOMEM;
 
-- 
2.39.2

Re: [PATCH 3/8] pwm: jz4740: Put per-channel clk into driver data

[Philippe Mathieu-Daudé]

On 29/6/23 11:48, Uwe Kleine-König wrote:
> Stop using chip_data which is about to go away. Instead track the
> per-channel clk in struct jz4740_pwm_chip.
> 
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> ---
>   drivers/pwm/pwm-jz4740.c | 11 +++++++----
>   1 file changed, 7 insertions(+), 4 deletions(-)

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>

Re: [PATCH 3/8] pwm: jz4740: Put per-channel clk into driver data

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 2938 bytes --]

Hello Paul,

first of all, your mail is strange. I think the problem is that it
doesn't contain a Message-Id. The result is that I got it twice and in
return vger.kernel.org seems to have refused to take it. At least it's
neither in lore.kernel.org nor in
https://patchwork.ozlabs.org/project/linux-pwm/patch/20230629094839.757092-4-u.kleine-koenig@pengutronix.de/
.

On Thu, Jun 29, 2023 at 01:12:25PM +0200, Paul Cercueil wrote:
> Le 29 juin 2023 11:48, Uwe Kleine-König <u.kleine-koenig@pengutronix.de> a écrit :
> >
> > Stop using chip_data which is about to go away. Instead track the 
> > per-channel clk in struct jz4740_pwm_chip. 
> >
> > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> 
> > --- 
> > drivers/pwm/pwm-jz4740.c | 11 +++++++---- 
> > 1 file changed, 7 insertions(+), 4 deletions(-) 
> >
> > diff --git a/drivers/pwm/pwm-jz4740.c b/drivers/pwm/pwm-jz4740.c 
> > index 3b7067f6cd0d..e0a57d71a60c 100644 
> > --- a/drivers/pwm/pwm-jz4740.c 
> > +++ b/drivers/pwm/pwm-jz4740.c 
> > @@ -27,6 +27,7 @@ struct soc_info { 
> > struct jz4740_pwm_chip { 
> > struct pwm_chip chip; 
> > struct regmap *map; 
> > + struct clk *clk[]; 
> > }; 
> >
> > static inline struct jz4740_pwm_chip *to_jz4740(struct pwm_chip *chip) 
> > @@ -70,14 +71,15 @@ static int jz4740_pwm_request(struct pwm_chip *chip, struct pwm_device *pwm) 
> > return err; 
> > } 
> >
> > - pwm_set_chip_data(pwm, clk); 
> > + jz->clk[pwm->hwpwm] = clk; 
> >
> > return 0; 
> > } 
> >
> > static void jz4740_pwm_free(struct pwm_chip *chip, struct pwm_device *pwm) 
> > { 
> > - struct clk *clk = pwm_get_chip_data(pwm); 
> > + struct jz4740_pwm_chip *jz = to_jz4740(chip); 
> > + struct clk *clk = jz->clk[pwm->hwpwm]; 
> >
> > clk_disable_unprepare(clk); 
> > clk_put(clk); 
> > @@ -123,7 +125,7 @@ static int jz4740_pwm_apply(struct pwm_chip *chip, struct pwm_device *pwm, 
> > { 
> > struct jz4740_pwm_chip *jz4740 = to_jz4740(pwm->chip); 
> > unsigned long long tmp = 0xffffull * NSEC_PER_SEC; 
> > - struct clk *clk = pwm_get_chip_data(pwm); 
> > + struct clk *clk = jz4740->clk[pwm->hwpwm]; 
> > unsigned long period, duty; 
> > long rate; 
> > int err; 
> > @@ -229,7 +231,8 @@ static int jz4740_pwm_probe(struct platform_device *pdev) 
> > if (!info) 
> > return -EINVAL; 
> >
> > - jz4740 = devm_kzalloc(dev, sizeof(*jz4740), GFP_KERNEL); 
> > + jz4740 = devm_kzalloc(dev, sizeof(*jz4740) + info->num_pwms * sizeof(jz4740->clk[0]), 
> > +       GFP_KERNEL); 
> 
> LGTM, but please use struct_size() from <linux/overflow.h>.

Ah, I thought there is such a macro, but I didn't find it neither by
grepping nor by asking in #kernelnewbies. Thanks, will respin the series
in a few days.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH 3/8] pwm: jz4740: Put per-channel clk into driver data

[Paul Cercueil]

Le jeu. 29 juin 2023 à 16:07:45 +0200, Uwe Kleine-König 
<u.kleine-koenig@pengutronix.de> a écrit :
> Hello Paul,
> 
> first of all, your mail is strange. I think the problem is that it
> doesn't contain a Message-Id. The result is that I got it twice and in
> return vger.kernel.org seems to have refused to take it. At least it's
> neither in lore.kernel.org nor in
> https://patchwork.ozlabs.org/project/linux-pwm/patch/20230629094839.757092-4-u.kleine-koenig@pengutronix.de/
> .

Sorry about that, I'm at the Embedded Linux conference in Prague 
without my computer, so I just answered from my Android phone. I didn't 
know it would cause problems, I guess I'll refrain from doing that from 
now on.

Cheers,
-Paul

> 
> On Thu, Jun 29, 2023 at 01:12:25PM +0200, Paul Cercueil wrote:
>>  Le 29 juin 2023 11:48, Uwe Kleine-König 
>> <u.kleine-koenig@pengutronix.de> a écrit :
>>  >
>>  > Stop using chip_data which is about to go away. Instead track the
>>  > per-channel clk in struct jz4740_pwm_chip.
>>  >
>>  > Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
>>  > ---
>>  > drivers/pwm/pwm-jz4740.c | 11 +++++++----
>>  > 1 file changed, 7 insertions(+), 4 deletions(-)
>>  >
>>  > diff --git a/drivers/pwm/pwm-jz4740.c b/drivers/pwm/pwm-jz4740.c
>>  > index 3b7067f6cd0d..e0a57d71a60c 100644
>>  > --- a/drivers/pwm/pwm-jz4740.c
>>  > +++ b/drivers/pwm/pwm-jz4740.c
>>  > @@ -27,6 +27,7 @@ struct soc_info {
>>  > struct jz4740_pwm_chip {
>>  > struct pwm_chip chip;
>>  > struct regmap *map;
>>  > + struct clk *clk[];
>>  > };
>>  >
>>  > static inline struct jz4740_pwm_chip *to_jz4740(struct pwm_chip 
>> *chip)
>>  > @@ -70,14 +71,15 @@ static int jz4740_pwm_request(struct pwm_chip 
>> *chip, struct pwm_device *pwm)
>>  > return err;
>>  > }
>>  >
>>  > - pwm_set_chip_data(pwm, clk);
>>  > + jz->clk[pwm->hwpwm] = clk;
>>  >
>>  > return 0;
>>  > }
>>  >
>>  > static void jz4740_pwm_free(struct pwm_chip *chip, struct 
>> pwm_device *pwm)
>>  > {
>>  > - struct clk *clk = pwm_get_chip_data(pwm);
>>  > + struct jz4740_pwm_chip *jz = to_jz4740(chip);
>>  > + struct clk *clk = jz->clk[pwm->hwpwm];
>>  >
>>  > clk_disable_unprepare(clk);
>>  > clk_put(clk);
>>  > @@ -123,7 +125,7 @@ static int jz4740_pwm_apply(struct pwm_chip 
>> *chip, struct pwm_device *pwm,
>>  > {
>>  > struct jz4740_pwm_chip *jz4740 = to_jz4740(pwm->chip);
>>  > unsigned long long tmp = 0xffffull * NSEC_PER_SEC;
>>  > - struct clk *clk = pwm_get_chip_data(pwm);
>>  > + struct clk *clk = jz4740->clk[pwm->hwpwm];
>>  > unsigned long period, duty;
>>  > long rate;
>>  > int err;
>>  > @@ -229,7 +231,8 @@ static int jz4740_pwm_probe(struct 
>> platform_device *pdev)
>>  > if (!info)
>>  > return -EINVAL;
>>  >
>>  > - jz4740 = devm_kzalloc(dev, sizeof(*jz4740), GFP_KERNEL);
>>  > + jz4740 = devm_kzalloc(dev, sizeof(*jz4740) + info->num_pwms * 
>> sizeof(jz4740->clk[0]),
>>  > +       GFP_KERNEL);
>> 
>>  LGTM, but please use struct_size() from <linux/overflow.h>.
> 
> Ah, I thought there is such a macro, but I didn't find it neither by
> grepping nor by asking in #kernelnewbies. Thanks, will respin the 
> series
> in a few days.
> 
> Best regards
> Uwe
> 
> --
> Pengutronix e.K.                           | Uwe Kleine-König        
>     |
> Industrial Linux Solutions                 | 
> https://www.pengutronix.de/ |

Re: [PATCH 3/8] pwm: jz4740: Put per-channel clk into driver data

[Philippe Mathieu-Daudé]

On 29/6/23 16:07, Uwe Kleine-König wrote:

> On Thu, Jun 29, 2023 at 01:12:25PM +0200, Paul Cercueil wrote:
>> Le 29 juin 2023 11:48, Uwe Kleine-König <u.kleine-koenig@pengutronix.de> a écrit :
>>>
>>> Stop using chip_data which is about to go away. Instead track the
>>> per-channel clk in struct jz4740_pwm_chip.
>>>
>>> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
>>> ---
>>> drivers/pwm/pwm-jz4740.c | 11 +++++++----
>>> 1 file changed, 7 insertions(+), 4 deletions(-)


>>> @@ -229,7 +231,8 @@ static int jz4740_pwm_probe(struct platform_device *pdev)
>>> if (!info)
>>> return -EINVAL;
>>>
>>> - jz4740 = devm_kzalloc(dev, sizeof(*jz4740), GFP_KERNEL);
>>> + jz4740 = devm_kzalloc(dev, sizeof(*jz4740) + info->num_pwms * sizeof(jz4740->clk[0]),
>>> +       GFP_KERNEL);
>>
>> LGTM, but please use struct_size() from <linux/overflow.h>.
> 
> Ah, I thought there is such a macro, but I didn't find it neither by
> grepping nor by asking in #kernelnewbies. Thanks, will respin the series
> in a few days.

Oh, TIL too :) R-b stands in respin.