| Message ID | 20240424-s4-pwm-v4-1-ee22effd40d0@amlogic.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Add support for Amlogic S4 PWM | expand |
On Wed 24 Apr 2024 at 18:28, Kelvin Zhang via B4 Relay <devnull+kelvin.zhang.amlogic.com@kernel.org> wrote: > From: Junyi Zhao <junyi.zhao@amlogic.com> > > This patch adds support for Amlogic S4 PWM. > > Signed-off-by: Junyi Zhao <junyi.zhao@amlogic.com> > Signed-off-by: Kelvin Zhang <kelvin.zhang@amlogic.com> > --- > drivers/pwm/pwm-meson.c | 37 +++++++++++++++++++++++++++++++++++++ > 1 file changed, 37 insertions(+) > > diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c > index ea96c5973488..6abc823745e4 100644 > --- a/drivers/pwm/pwm-meson.c > +++ b/drivers/pwm/pwm-meson.c > @@ -462,6 +462,35 @@ static int meson_pwm_init_channels_meson8b_v2(struct pwm_chip *chip) > return meson_pwm_init_clocks_meson8b(chip, mux_parent_data); > } > > +static int meson_pwm_init_channels_meson_s4(struct pwm_chip *chip) > +{ > + int i, ret; > + struct device *dev = pwmchip_parent(chip); > + struct device_node *np = dev->of_node; > + struct meson_pwm *meson = to_meson_pwm(chip); > + struct meson_pwm_channel *channel; > + > + for (i = 0; i < MESON_NUM_PWMS; i++) { > + channel = &meson->channels[i]; > + channel->clk = of_clk_get(np, i); > + if (IS_ERR(channel->clk)) { > + ret = PTR_ERR(channel->clk); > + dev_err_probe(dev, ret, "Failed to get clk\n"); > + goto err; > + } > + } > + > + return 0; > + > +err: > + while (--i >= 0) { > + channel = &meson->channels[i]; > + clk_put(channel->clk); Fine on error but leaks on module unload. Same as George, Add the devm variant of of_clk_get() if you must. Use devm_add_action_or_reset() otherwise Could please synchronize this series with George and deal with all the supported SoCs ? a1, s4, t7, c3 ... > + } > + > + return ret; > +} > + > static const struct meson_pwm_data pwm_meson8b_data = { > .parent_names = { "xtal", NULL, "fclk_div4", "fclk_div3" }, > .channels_init = meson_pwm_init_channels_meson8b_legacy, > @@ -500,6 +529,10 @@ static const struct meson_pwm_data pwm_meson8_v2_data = { > .channels_init = meson_pwm_init_channels_meson8b_v2, > }; > > +static const struct meson_pwm_data pwm_meson_s4_data = { > + .channels_init = meson_pwm_init_channels_meson_s4, > +}; > + > static const struct of_device_id meson_pwm_matches[] = { > { > .compatible = "amlogic,meson8-pwm-v2", > @@ -538,6 +571,10 @@ static const struct of_device_id meson_pwm_matches[] = { > .compatible = "amlogic,meson-g12a-ao-pwm-cd", > .data = &pwm_g12a_ao_cd_data > }, > + { > + .compatible = "amlogic,meson-s4-pwm", > + .data = &pwm_meson_s4_data > + }, > {}, > }; > MODULE_DEVICE_TABLE(of, meson_pwm_matches);
On 2024/4/24 18:32, Jerome Brunet wrote: > [ EXTERNAL EMAIL ] > > On Wed 24 Apr 2024 at 18:28, Kelvin Zhang via B4 Relay <devnull+kelvin.zhang.amlogic.com@kernel.org> wrote: > >> From: Junyi Zhao <junyi.zhao@amlogic.com> >> >> This patch adds support for Amlogic S4 PWM. >> >> Signed-off-by: Junyi Zhao <junyi.zhao@amlogic.com> >> Signed-off-by: Kelvin Zhang <kelvin.zhang@amlogic.com> >> --- >> drivers/pwm/pwm-meson.c | 37 +++++++++++++++++++++++++++++++++++++ >> 1 file changed, 37 insertions(+) >> >> diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c >> index ea96c5973488..6abc823745e4 100644 >> --- a/drivers/pwm/pwm-meson.c >> +++ b/drivers/pwm/pwm-meson.c >> @@ -462,6 +462,35 @@ static int meson_pwm_init_channels_meson8b_v2(struct pwm_chip *chip) >> return meson_pwm_init_clocks_meson8b(chip, mux_parent_data); >> } >> >> +static int meson_pwm_init_channels_meson_s4(struct pwm_chip *chip) >> +{ >> + int i, ret; >> + struct device *dev = pwmchip_parent(chip); >> + struct device_node *np = dev->of_node; >> + struct meson_pwm *meson = to_meson_pwm(chip); >> + struct meson_pwm_channel *channel; >> + >> + for (i = 0; i < MESON_NUM_PWMS; i++) { >> + channel = &meson->channels[i]; >> + channel->clk = of_clk_get(np, i); >> + if (IS_ERR(channel->clk)) { >> + ret = PTR_ERR(channel->clk); >> + dev_err_probe(dev, ret, "Failed to get clk\n"); >> + goto err; >> + } >> + } >> + >> + return 0; >> + >> +err: >> + while (--i >= 0) { >> + channel = &meson->channels[i]; >> + clk_put(channel->clk); > > Fine on error but leaks on module unload. > > Same as George, > > Add the devm variant of of_clk_get() if you must. > Use devm_add_action_or_reset() otherwise Hi jerom，but we have discussed before.devm variant such as follows： devm_clk_get_enable(struct device * dev, char * id) struct clk *devm_clk_get(struct device *dev, const char *id) struct clk *devm_clk_get_optional(struct device *dev, const char *id) after i check api parm ,these api's 2rd parm "id" is string not index. because dt binding have no name property. could we use devm？ > > Could please synchronize this series with George and deal with all the > supported SoCs ? a1, s4, t7, c3 ... > >> + } >> + >> + return ret; >> +} >> + >> static const struct meson_pwm_data pwm_meson8b_data = { >> .parent_names = { "xtal", NULL, "fclk_div4", "fclk_div3" }, >> .channels_init = meson_pwm_init_channels_meson8b_legacy, >> @@ -500,6 +529,10 @@ static const struct meson_pwm_data pwm_meson8_v2_data = { >> .channels_init = meson_pwm_init_channels_meson8b_v2, >> }; >> >> +static const struct meson_pwm_data pwm_meson_s4_data = { >> + .channels_init = meson_pwm_init_channels_meson_s4, >> +}; >> + >> static const struct of_device_id meson_pwm_matches[] = { >> { >> .compatible = "amlogic,meson8-pwm-v2", >> @@ -538,6 +571,10 @@ static const struct of_device_id meson_pwm_matches[] = { >> .compatible = "amlogic,meson-g12a-ao-pwm-cd", >> .data = &pwm_g12a_ao_cd_data >> }, >> + { >> + .compatible = "amlogic,meson-s4-pwm", >> + .data = &pwm_meson_s4_data >> + }, >> {}, >> }; >> MODULE_DEVICE_TABLE(of, meson_pwm_matches); > > > -- > Jerome
Hello, On Wed, Apr 24, 2024 at 12:32:36PM +0200, Jerome Brunet wrote: > > +err: > > + while (--i >= 0) { > > + channel = &meson->channels[i]; > > + clk_put(channel->clk); > > Fine on error but leaks on module unload. > > Same as George, > > Add the devm variant of of_clk_get() if you must. There shouldn't be a reason to still use of_clk_get(). I'd expect that devm_clk_get() should do the job and if not that's a bug. Best regards Uwe
On Wed 24 Apr 2024 at 15:51, Uwe Kleine-König <u.kleine-koenig@pengutronix.de> wrote: > [[PGP Signed Part:Undecided]] > Hello, > > On Wed, Apr 24, 2024 at 12:32:36PM +0200, Jerome Brunet wrote: >> > +err: >> > + while (--i >= 0) { >> > + channel = &meson->channels[i]; >> > + clk_put(channel->clk); >> >> Fine on error but leaks on module unload. >> >> Same as George, >> >> Add the devm variant of of_clk_get() if you must. > > There shouldn't be a reason to still use of_clk_get(). I'd expect that > devm_clk_get() should do the job and if not that's a bug. Getting a clock ressource by index instead of by name is a reason. > > Best regards > Uwe
Hello Jerome On 4/24/24 13:32, Jerome Brunet wrote: > > On Wed 24 Apr 2024 at 18:28, Kelvin Zhang via B4 Relay <devnull+kelvin.zhang.amlogic.com@kernel.org> wrote: > >> From: Junyi Zhao <junyi.zhao@amlogic.com> >> >> This patch adds support for Amlogic S4 PWM. >> >> Signed-off-by: Junyi Zhao <junyi.zhao@amlogic.com> >> Signed-off-by: Kelvin Zhang <kelvin.zhang@amlogic.com> >> --- >> drivers/pwm/pwm-meson.c | 37 +++++++++++++++++++++++++++++++++++++ >> 1 file changed, 37 insertions(+) >> >> diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c >> index ea96c5973488..6abc823745e4 100644 >> --- a/drivers/pwm/pwm-meson.c >> +++ b/drivers/pwm/pwm-meson.c >> @@ -462,6 +462,35 @@ static int meson_pwm_init_channels_meson8b_v2(struct pwm_chip *chip) >> return meson_pwm_init_clocks_meson8b(chip, mux_parent_data); >> } >> >> +static int meson_pwm_init_channels_meson_s4(struct pwm_chip *chip) >> +{ >> + int i, ret; >> + struct device *dev = pwmchip_parent(chip); >> + struct device_node *np = dev->of_node; >> + struct meson_pwm *meson = to_meson_pwm(chip); >> + struct meson_pwm_channel *channel; >> + >> + for (i = 0; i < MESON_NUM_PWMS; i++) { >> + channel = &meson->channels[i]; >> + channel->clk = of_clk_get(np, i); >> + if (IS_ERR(channel->clk)) { >> + ret = PTR_ERR(channel->clk); >> + dev_err_probe(dev, ret, "Failed to get clk\n"); >> + goto err; >> + } >> + } >> + >> + return 0; >> + >> +err: >> + while (--i >= 0) { >> + channel = &meson->channels[i]; >> + clk_put(channel->clk); > > Fine on error but leaks on module unload. > > Same as George, > > Add the devm variant of of_clk_get() if you must. > Use devm_add_action_or_reset() otherwise > > Could please synchronize this series with George and deal with all the > supported SoCs ? a1, s4, t7, c3 ... If the chipmaker eagers to support s4 himself we're ok :) But since I sent my patch first I think it'd be fair if this single patch have my tag: Co-Developed-by: George Stark <gnstark@salutedevices.com> I'll help to review the patch too. Jerome could we split support for all mentioned socs into different series? e.g. 1. Junyi finishes the driver's base patch and s4 dtsi patch 2. I send a1 dt-bindings and a1 dtsi patches 3. Someone later sends t7/c3 dt-bindings + dtsi The reason is to apply what we have on hand now due to meson-pwm is under heavy development more than a year already. > >> + } >> + >> + return ret; >> +} >> + >> static const struct meson_pwm_data pwm_meson8b_data = { >> .parent_names = { "xtal", NULL, "fclk_div4", "fclk_div3" }, >> .channels_init = meson_pwm_init_channels_meson8b_legacy, >> @@ -500,6 +529,10 @@ static const struct meson_pwm_data pwm_meson8_v2_data = { >> .channels_init = meson_pwm_init_channels_meson8b_v2, >> }; >> >> +static const struct meson_pwm_data pwm_meson_s4_data = { >> + .channels_init = meson_pwm_init_channels_meson_s4, >> +}; >> + >> static const struct of_device_id meson_pwm_matches[] = { >> { >> .compatible = "amlogic,meson8-pwm-v2", >> @@ -538,6 +571,10 @@ static const struct of_device_id meson_pwm_matches[] = { >> .compatible = "amlogic,meson-g12a-ao-pwm-cd", >> .data = &pwm_g12a_ao_cd_data >> }, >> + { >> + .compatible = "amlogic,meson-s4-pwm", >> + .data = &pwm_meson_s4_data >> + }, >> {}, >> }; >> MODULE_DEVICE_TABLE(of, meson_pwm_matches); > >
Hello Junyi, Kelvin
I'm sorry for the ping. Do you have plans to finish these patches?
Here is sample code how devm could be used to manage clk objects created
by of_clk_get:
struct clk_set_devres {
struct clk *clks[MESON_NUM_PWMS];
};
static void devm_clk_set_release(struct device *dev, void *res)
{
struct clk_set_devres *devres = res;
int num_clks = MESON_NUM_PWMS;
while (--num_clks >= 0)
clk_put(devres->clks[num_clks]);
}
static int meson_pwm_init_channels_s4(struct pwm_chip *chip)
{
struct device *dev = pwmchip_parent(chip);
struct meson_pwm *meson = to_meson_pwm(chip);
struct clk_set_devres *devres;
unsigned int i;
int res;
devres = devres_alloc(devm_clk_set_release,
sizeof(*devres), GFP_KERNEL);
if (!devres)
return -ENOMEM;
devres_add(dev, devres);
for (i = 0; i < MESON_NUM_PWMS; i++) {
devres->clks[i] = of_clk_get(dev->of_node, i);
if (IS_ERR(devres->clks[i])) {
res = PTR_ERR(devres->clks[i]);
dev_err_probe(dev, res, "Failed to get clk\n");
return res;
}
meson->channels[i].clk = devres->clks[i];
}
return 0;
}
On 4/24/24 14:44, Junyi Zhao wrote:
>
>
> On 2024/4/24 18:32, Jerome Brunet wrote:
>> [ EXTERNAL EMAIL ]
>>
>> On Wed 24 Apr 2024 at 18:28, Kelvin Zhang via B4 Relay
>> <devnull+kelvin.zhang.amlogic.com@kernel.org> wrote:
>>
>>> From: Junyi Zhao <junyi.zhao@amlogic.com>
>>>
>>> This patch adds support for Amlogic S4 PWM.
>>>
>>> Signed-off-by: Junyi Zhao <junyi.zhao@amlogic.com>
>>> Signed-off-by: Kelvin Zhang <kelvin.zhang@amlogic.com>
>>> ---
>>> drivers/pwm/pwm-meson.c | 37 +++++++++++++++++++++++++++++++++++++
>>> 1 file changed, 37 insertions(+)
>>>
>>> diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c
>>> index ea96c5973488..6abc823745e4 100644
>>> --- a/drivers/pwm/pwm-meson.c
>>> +++ b/drivers/pwm/pwm-meson.c
>>> @@ -462,6 +462,35 @@ static int
>>> meson_pwm_init_channels_meson8b_v2(struct pwm_chip *chip)
>>> return meson_pwm_init_clocks_meson8b(chip, mux_parent_data);
>>> }
>>>
>>> +static int meson_pwm_init_channels_meson_s4(struct pwm_chip *chip)
>>> +{
>>> + int i, ret;
>>> + struct device *dev = pwmchip_parent(chip);
>>> + struct device_node *np = dev->of_node;
>>> + struct meson_pwm *meson = to_meson_pwm(chip);
>>> + struct meson_pwm_channel *channel;
>>> +
>>> + for (i = 0; i < MESON_NUM_PWMS; i++) {
>>> + channel = &meson->channels[i];
>>> + channel->clk = of_clk_get(np, i);
>>> + if (IS_ERR(channel->clk)) {
>>> + ret = PTR_ERR(channel->clk);
>>> + dev_err_probe(dev, ret, "Failed to get clk\n");
>>> + goto err;
>>> + }
>>> + }
>>> +
>>> + return 0;
>>> +
>>> +err:
>>> + while (--i >= 0) {
>>> + channel = &meson->channels[i];
>>> + clk_put(channel->clk);
>>
>> Fine on error but leaks on module unload.
>>
>> Same as George,
>>
>> Add the devm variant of of_clk_get() if you must.
>> Use devm_add_action_or_reset() otherwise
> Hi jerom，but we have discussed before.devm variant such as follows：
> devm_clk_get_enable(struct device * dev, char * id)
> struct clk *devm_clk_get(struct device *dev, const char *id)
> struct clk *devm_clk_get_optional(struct device *dev, const char *id)
>
> after i check api parm ,these api's 2rd parm "id" is string not index.
> because dt binding have no name property. could we use devm？
>>
>> Could please synchronize this series with George and deal with all the
>> supported SoCs ? a1, s4, t7, c3 ...
>>
>>> + }
>>> +
>>> + return ret;
>>> +}
>>> +
>>> static const struct meson_pwm_data pwm_meson8b_data = {
>>> .parent_names = { "xtal", NULL, "fclk_div4", "fclk_div3" },
>>> .channels_init = meson_pwm_init_channels_meson8b_legacy,
>>> @@ -500,6 +529,10 @@ static const struct meson_pwm_data
>>> pwm_meson8_v2_data = {
>>> .channels_init = meson_pwm_init_channels_meson8b_v2,
>>> };
>>>
>>> +static const struct meson_pwm_data pwm_meson_s4_data = {
>>> + .channels_init = meson_pwm_init_channels_meson_s4,
>>> +};
>>> +
>>> static const struct of_device_id meson_pwm_matches[] = {
>>> {
>>> .compatible = "amlogic,meson8-pwm-v2",
>>> @@ -538,6 +571,10 @@ static const struct of_device_id
>>> meson_pwm_matches[] = {
>>> .compatible = "amlogic,meson-g12a-ao-pwm-cd",
>>> .data = &pwm_g12a_ao_cd_data
>>> },
>>> + {
>>> + .compatible = "amlogic,meson-s4-pwm",
>>> + .data = &pwm_meson_s4_data
>>> + },
>>> {},
>>> };
>>> MODULE_DEVICE_TABLE(of, meson_pwm_matches);
>>
>>
>> --
>> Jerome
diff --git a/drivers/pwm/pwm-meson.c b/drivers/pwm/pwm-meson.c index ea96c5973488..6abc823745e4 100644 --- a/drivers/pwm/pwm-meson.c +++ b/drivers/pwm/pwm-meson.c @@ -462,6 +462,35 @@ static int meson_pwm_init_channels_meson8b_v2(struct pwm_chip *chip) return meson_pwm_init_clocks_meson8b(chip, mux_parent_data); } +static int meson_pwm_init_channels_meson_s4(struct pwm_chip *chip) +{ + int i, ret; + struct device *dev = pwmchip_parent(chip); + struct device_node *np = dev->of_node; + struct meson_pwm *meson = to_meson_pwm(chip); + struct meson_pwm_channel *channel; + + for (i = 0; i < MESON_NUM_PWMS; i++) { + channel = &meson->channels[i]; + channel->clk = of_clk_get(np, i); + if (IS_ERR(channel->clk)) { + ret = PTR_ERR(channel->clk); + dev_err_probe(dev, ret, "Failed to get clk\n"); + goto err; + } + } + + return 0; + +err: + while (--i >= 0) { + channel = &meson->channels[i]; + clk_put(channel->clk); + } + + return ret; +} + static const struct meson_pwm_data pwm_meson8b_data = { .parent_names = { "xtal", NULL, "fclk_div4", "fclk_div3" }, .channels_init = meson_pwm_init_channels_meson8b_legacy, @@ -500,6 +529,10 @@ static const struct meson_pwm_data pwm_meson8_v2_data = { .channels_init = meson_pwm_init_channels_meson8b_v2, }; +static const struct meson_pwm_data pwm_meson_s4_data = { + .channels_init = meson_pwm_init_channels_meson_s4, +}; + static const struct of_device_id meson_pwm_matches[] = { { .compatible = "amlogic,meson8-pwm-v2", @@ -538,6 +571,10 @@ static const struct of_device_id meson_pwm_matches[] = { .compatible = "amlogic,meson-g12a-ao-pwm-cd", .data = &pwm_g12a_ao_cd_data }, + { + .compatible = "amlogic,meson-s4-pwm", + .data = &pwm_meson_s4_data + }, {}, }; MODULE_DEVICE_TABLE(of, meson_pwm_matches);