| Message ID | 20231024205805.19314-1-luizluca@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next,1/2] net: dsa: realtek: support reset controller | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Clearly marked for net-next |
| netdev/apply | fail | Patch does not apply to net-next |
On 10/24/23 13:58, Luiz Angelo Daros de Luca wrote: > The 'reset-gpios' will not work when the switch reset is controlled by a > reset controller. > > Although the reset is optional and the driver performs a soft reset > during setup, if the initial reset state was asserted, the driver will > not detect it. > > This is an example of how to use the reset controller: > > switch { > compatible = "realtek,rtl8366rb"; > > resets = <&rst 8>; > reset-names = "switch"; > > ... > } > > The reset controller will take precedence over the reset GPIO. > > Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com> > --- > drivers/net/dsa/realtek/realtek-mdio.c | 36 +++++++++++++++++++++----- > drivers/net/dsa/realtek/realtek-smi.c | 34 +++++++++++++++++++----- > drivers/net/dsa/realtek/realtek.h | 6 +++++ > 3 files changed, 63 insertions(+), 13 deletions(-) > > diff --git a/drivers/net/dsa/realtek/realtek-mdio.c b/drivers/net/dsa/realtek/realtek-mdio.c > index 292e6d087e8b..600124c58c00 100644 > --- a/drivers/net/dsa/realtek/realtek-mdio.c > +++ b/drivers/net/dsa/realtek/realtek-mdio.c > @@ -140,6 +140,23 @@ static const struct regmap_config realtek_mdio_nolock_regmap_config = { > .disable_locking = true, > }; > > +static int realtek_mdio_hwreset(struct realtek_priv *priv, bool active) > +{ > +#ifdef CONFIG_RESET_CONTROLLER > + if (priv->reset_ctl) { > + if (active) > + return reset_control_assert(priv->reset_ctl); > + else > + return reset_control_deassert(priv->reset_ctl); > + } > +#endif Empty stubs are provided when CONFIG_RESET_CONTROLLER is disabled, and if you switch to using devm_reset_control_get() then you will get a NULL reset_control reference which will be a no-op for all of those operations.
> Empty stubs are provided when CONFIG_RESET_CONTROLLER is disabled Nice! I'll drop the "#ifdef"s. > if you switch to using devm_reset_control_get() then you will get a NULL > reset_control reference which will be a no-op for all of those operations. I'm already using devm_reset_control_get(). Maybe you copied the wrong name? Did you mean devm_reset_control_get_optional()? It is, indeed, what I needed. Thanks. Regards, Luiz
On 10/24/23 15:02, Luiz Angelo Daros de Luca wrote: >> Empty stubs are provided when CONFIG_RESET_CONTROLLER is disabled > > Nice! I'll drop the "#ifdef"s. > >> if you switch to using devm_reset_control_get() then you will get a NULL >> reset_control reference which will be a no-op for all of those operations. > > I'm already using devm_reset_control_get(). Maybe you copied the wrong > name? Did you mean devm_reset_control_get_optional()? > It is, indeed, what I needed. Thanks. Yes I copy/pasted what I had just been searching for and instead intended to mention that devm_reset_control_get_optional() would do what you need. Thanks!
On Tue, Oct 24, 2023 at 07:02:33PM -0300, Luiz Angelo Daros de Luca wrote: > > Empty stubs are provided when CONFIG_RESET_CONTROLLER is disabled > > Nice! I'll drop the "#ifdef"s. > > > if you switch to using devm_reset_control_get() then you will get a NULL > > reset_control reference which will be a no-op for all of those operations. > > I'm already using devm_reset_control_get(). Maybe you copied the wrong > name? Did you mean devm_reset_control_get_optional()? > It is, indeed, what I needed. Thanks. Please also wait for a review on the device tree binding change before posting a v2.
Hi Linus, > - /* TODO: if power is software controlled, set up any regulators here */ > +#ifdef CONFIG_RESET_CONTROLLER > + priv->reset_ctl = devm_reset_control_get(dev, "switch"); > + if (IS_ERR(priv->reset_ctl)) { > + dev_err(dev, "failed to get switch reset control\n"); > + return PTR_ERR(priv->reset_ctl); > + } > +#endif I'm dropping this TODO as I think it means something like this reset control, right? Regards, Luiz
On Tue, Oct 24, 2023 at 07:17:10PM -0300, Luiz Angelo Daros de Luca wrote: > Hi Linus, > > > - /* TODO: if power is software controlled, set up any regulators here */ > > +#ifdef CONFIG_RESET_CONTROLLER > > + priv->reset_ctl = devm_reset_control_get(dev, "switch"); > > + if (IS_ERR(priv->reset_ctl)) { > > + dev_err(dev, "failed to get switch reset control\n"); > > + return PTR_ERR(priv->reset_ctl); > > + } > > +#endif > > I'm dropping this TODO as I think it means something like this reset > control, right? No, a regulator is a different thing to a reset controller. A regulator is used to control the power to the device. Andrew
On Tue, Oct 24, 2023 at 05:58:04PM -0300, Luiz Angelo Daros de Luca wrote: > The 'reset-gpios' will not work when the switch reset is controlled by a > reset controller. > > Although the reset is optional and the driver performs a soft reset > during setup, if the initial reset state was asserted, the driver will > not detect it. > > This is an example of how to use the reset controller: > > switch { > compatible = "realtek,rtl8366rb"; > > resets = <&rst 8>; > reset-names = "switch"; > > ... > } Mix of tabs and spaces here. Also, examples belong to the dt-schema. > > The reset controller will take precedence over the reset GPIO. > > Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com> > --- > drivers/net/dsa/realtek/realtek-mdio.c | 36 +++++++++++++++++++++----- > drivers/net/dsa/realtek/realtek-smi.c | 34 +++++++++++++++++++----- > drivers/net/dsa/realtek/realtek.h | 6 +++++ > 3 files changed, 63 insertions(+), 13 deletions(-) > > diff --git a/drivers/net/dsa/realtek/realtek-mdio.c b/drivers/net/dsa/realtek/realtek-mdio.c > index 292e6d087e8b..600124c58c00 100644 > --- a/drivers/net/dsa/realtek/realtek-mdio.c > +++ b/drivers/net/dsa/realtek/realtek-mdio.c > @@ -140,6 +140,23 @@ static const struct regmap_config realtek_mdio_nolock_regmap_config = { > .disable_locking = true, > }; > > +static int realtek_mdio_hwreset(struct realtek_priv *priv, bool active) > +{ > +#ifdef CONFIG_RESET_CONTROLLER > + if (priv->reset_ctl) { > + if (active) > + return reset_control_assert(priv->reset_ctl); > + else > + return reset_control_deassert(priv->reset_ctl); > + } > +#endif > + > + if (priv->reset) > + gpiod_set_value(priv->reset, active); > + > + return 0; > +} > + This "bool active" artificially unifies two discrete code paths in the same function, where the callers are not the same and the implementation is not the same (given a priv->reset_ctl presence), separated by an "if". Would it make more sense to have discrete functions, each with its unique caller, like this? static int realtek_reset_assert(struct realtek_priv *priv) { if (priv->reset_ctl) return reset_control_assert(priv->reset_ctl); if (priv->reset) gpiod_set_value(priv->reset, 1); return 0; } static int realtek_reset_deassert(struct realtek_priv *priv) { if (priv->reset_ctl) return reset_control_deassert(priv->reset_ctl); if (priv->reset) gpiod_set_value(priv->reset, 0); return 0; } Also, you return int but ignore error values everywhere. I guess it would make more sense to return void, but print warnings within the reset functions if the calls to the reset control fail. > static int realtek_mdio_probe(struct mdio_device *mdiodev) > { > struct realtek_priv *priv; > @@ -194,20 +211,26 @@ static int realtek_mdio_probe(struct mdio_device *mdiodev) > > dev_set_drvdata(dev, priv); > > - /* TODO: if power is software controlled, set up any regulators here */ I'm not sure if "power" and "reset" are the same thing... > priv->leds_disabled = of_property_read_bool(np, "realtek,disable-leds"); > > +#ifdef CONFIG_RESET_CONTROLLER > + priv->reset_ctl = devm_reset_control_get(dev, "switch"); > + if (IS_ERR(priv->reset_ctl)) { > + dev_err(dev, "failed to get switch reset control\n"); > + return PTR_ERR(priv->reset_ctl); ret = PTR_ERR(priv->reset_ctl); return dev_err_probe(dev, err, "failed to get reset control\n"); This suppresses -EPROBE_DEFER prints. > + } > +#endif > + > priv->reset = devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_LOW); > if (IS_ERR(priv->reset)) { > dev_err(dev, "failed to get RESET GPIO\n"); > return PTR_ERR(priv->reset); > } > - > - if (priv->reset) { > - gpiod_set_value(priv->reset, 1); > + if (priv->reset_ctl || priv->reset) { > + realtek_mdio_hwreset(priv, 1); > dev_dbg(dev, "asserted RESET\n"); > msleep(REALTEK_HW_STOP_DELAY); > - gpiod_set_value(priv->reset, 0); > + realtek_mdio_hwreset(priv, 0); > msleep(REALTEK_HW_START_DELAY); > dev_dbg(dev, "deasserted RESET\n"); > } > @@ -246,8 +269,7 @@ static void realtek_mdio_remove(struct mdio_device *mdiodev) > dsa_unregister_switch(priv->ds); > > /* leave the device reset asserted */ > - if (priv->reset) > - gpiod_set_value(priv->reset, 1); > + realtek_mdio_hwreset(priv, 1); nitpick: "bool" arguments should take "true" or "false". > } > > static void realtek_mdio_shutdown(struct mdio_device *mdiodev) > diff --git a/drivers/net/dsa/realtek/realtek-smi.c b/drivers/net/dsa/realtek/realtek-smi.c > index bfd11591faf4..751159d71223 100644 > --- a/drivers/net/dsa/realtek/realtek-smi.c > +++ b/drivers/net/dsa/realtek/realtek-smi.c > @@ -408,6 +408,23 @@ static int realtek_smi_setup_mdio(struct dsa_switch *ds) > return ret; > } > > +static int realtek_smi_hwreset(struct realtek_priv *priv, bool active) > +{ > +#ifdef CONFIG_RESET_CONTROLLER > + if (priv->reset_ctl) { > + if (active) > + return reset_control_assert(priv->reset_ctl); > + else > + return reset_control_deassert(priv->reset_ctl); > + } > +#endif > + > + if (priv->reset) > + gpiod_set_value(priv->reset, active); > + > + return 0; > +} What is the reason for duplicating realtek_mdio_hwreset()?
> On Tue, Oct 24, 2023 at 05:58:04PM -0300, Luiz Angelo Daros de Luca wrote: > > The 'reset-gpios' will not work when the switch reset is controlled by a > > reset controller. > > > > Although the reset is optional and the driver performs a soft reset > > during setup, if the initial reset state was asserted, the driver will > > not detect it. > > > > This is an example of how to use the reset controller: > > > > switch { > > compatible = "realtek,rtl8366rb"; > > > > resets = <&rst 8>; > > reset-names = "switch"; > > > > ... > > } > > Mix of tabs and spaces here. > Also, examples belong to the dt-schema. OK > > > > > The reset controller will take precedence over the reset GPIO. > > > > Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com> > > --- > > drivers/net/dsa/realtek/realtek-mdio.c | 36 +++++++++++++++++++++----- > > drivers/net/dsa/realtek/realtek-smi.c | 34 +++++++++++++++++++----- > > drivers/net/dsa/realtek/realtek.h | 6 +++++ > > 3 files changed, 63 insertions(+), 13 deletions(-) > > > > diff --git a/drivers/net/dsa/realtek/realtek-mdio.c b/drivers/net/dsa/realtek/realtek-mdio.c > > index 292e6d087e8b..600124c58c00 100644 > > --- a/drivers/net/dsa/realtek/realtek-mdio.c > > +++ b/drivers/net/dsa/realtek/realtek-mdio.c > > @@ -140,6 +140,23 @@ static const struct regmap_config realtek_mdio_nolock_regmap_config = { > > .disable_locking = true, > > }; > > > > +static int realtek_mdio_hwreset(struct realtek_priv *priv, bool active) > > +{ > > +#ifdef CONFIG_RESET_CONTROLLER > > + if (priv->reset_ctl) { > > + if (active) > > + return reset_control_assert(priv->reset_ctl); > > + else > > + return reset_control_deassert(priv->reset_ctl); > > + } > > +#endif > > + > > + if (priv->reset) > > + gpiod_set_value(priv->reset, active); > > + > > + return 0; > > +} > > + > > This "bool active" artificially unifies two discrete code paths in the > same function, where the callers are not the same and the implementation > is not the same (given a priv->reset_ctl presence), separated by an "if". > > Would it make more sense to have discrete functions, each with its > unique caller, like this? > > static int realtek_reset_assert(struct realtek_priv *priv) > { > if (priv->reset_ctl) > return reset_control_assert(priv->reset_ctl); > > if (priv->reset) > gpiod_set_value(priv->reset, 1); > > return 0; > } > > static int realtek_reset_deassert(struct realtek_priv *priv) > { > if (priv->reset_ctl) > return reset_control_deassert(priv->reset_ctl); > > if (priv->reset) > gpiod_set_value(priv->reset, 0); > > return 0; > } Sure. It is better. > Also, you return int but ignore error values everywhere. I guess it > would make more sense to return void, but print warnings within the > reset functions if the calls to the reset control fail. > > > static int realtek_mdio_probe(struct mdio_device *mdiodev) > > { > > struct realtek_priv *priv; > > @@ -194,20 +211,26 @@ static int realtek_mdio_probe(struct mdio_device *mdiodev) > > > > dev_set_drvdata(dev, priv); > > > > - /* TODO: if power is software controlled, set up any regulators here */ > > I'm not sure if "power" and "reset" are the same thing... > > > priv->leds_disabled = of_property_read_bool(np, "realtek,disable-leds"); > > > > +#ifdef CONFIG_RESET_CONTROLLER > > + priv->reset_ctl = devm_reset_control_get(dev, "switch"); > > + if (IS_ERR(priv->reset_ctl)) { > > + dev_err(dev, "failed to get switch reset control\n"); > > + return PTR_ERR(priv->reset_ctl); > > ret = PTR_ERR(priv->reset_ctl); > return dev_err_probe(dev, err, "failed to get reset control\n"); > > This suppresses -EPROBE_DEFER prints. OK > > > + } > > +#endif > > + > > priv->reset = devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_LOW); > > if (IS_ERR(priv->reset)) { > > dev_err(dev, "failed to get RESET GPIO\n"); > > return PTR_ERR(priv->reset); > > } > > - > > - if (priv->reset) { > > - gpiod_set_value(priv->reset, 1); > > + if (priv->reset_ctl || priv->reset) { > > + realtek_mdio_hwreset(priv, 1); > > dev_dbg(dev, "asserted RESET\n"); > > msleep(REALTEK_HW_STOP_DELAY); > > - gpiod_set_value(priv->reset, 0); > > + realtek_mdio_hwreset(priv, 0); > > msleep(REALTEK_HW_START_DELAY); > > dev_dbg(dev, "deasserted RESET\n"); > > } > > @@ -246,8 +269,7 @@ static void realtek_mdio_remove(struct mdio_device *mdiodev) > > dsa_unregister_switch(priv->ds); > > > > /* leave the device reset asserted */ > > - if (priv->reset) > > - gpiod_set_value(priv->reset, 1); > > + realtek_mdio_hwreset(priv, 1); > > nitpick: "bool" arguments should take "true" or "false". OK > > > } > > > > static void realtek_mdio_shutdown(struct mdio_device *mdiodev) > > diff --git a/drivers/net/dsa/realtek/realtek-smi.c b/drivers/net/dsa/realtek/realtek-smi.c > > index bfd11591faf4..751159d71223 100644 > > --- a/drivers/net/dsa/realtek/realtek-smi.c > > +++ b/drivers/net/dsa/realtek/realtek-smi.c > > @@ -408,6 +408,23 @@ static int realtek_smi_setup_mdio(struct dsa_switch *ds) > > return ret; > > } > > > > +static int realtek_smi_hwreset(struct realtek_priv *priv, bool active) > > +{ > > +#ifdef CONFIG_RESET_CONTROLLER > > + if (priv->reset_ctl) { > > + if (active) > > + return reset_control_assert(priv->reset_ctl); > > + else > > + return reset_control_deassert(priv->reset_ctl); > > + } > > +#endif > > + > > + if (priv->reset) > > + gpiod_set_value(priv->reset, active); > > + > > + return 0; > > +} > > What is the reason for duplicating realtek_mdio_hwreset()? Both interface modules, realtek-smi and realtek-mdio, do not share code, except for the realtek.h header file. I don't know if it is worth it to put the code in a new shared module. What is the best practice here? Create a realtek_common.c linked to both modules? Regards, Luiz
diff --git a/drivers/net/dsa/realtek/realtek-mdio.c b/drivers/net/dsa/realtek/realtek-mdio.c index 292e6d087e8b..600124c58c00 100644 --- a/drivers/net/dsa/realtek/realtek-mdio.c +++ b/drivers/net/dsa/realtek/realtek-mdio.c @@ -140,6 +140,23 @@ static const struct regmap_config realtek_mdio_nolock_regmap_config = { .disable_locking = true, }; +static int realtek_mdio_hwreset(struct realtek_priv *priv, bool active) +{ +#ifdef CONFIG_RESET_CONTROLLER + if (priv->reset_ctl) { + if (active) + return reset_control_assert(priv->reset_ctl); + else + return reset_control_deassert(priv->reset_ctl); + } +#endif + + if (priv->reset) + gpiod_set_value(priv->reset, active); + + return 0; +} + static int realtek_mdio_probe(struct mdio_device *mdiodev) { struct realtek_priv *priv; @@ -194,20 +211,26 @@ static int realtek_mdio_probe(struct mdio_device *mdiodev) dev_set_drvdata(dev, priv); - /* TODO: if power is software controlled, set up any regulators here */ priv->leds_disabled = of_property_read_bool(np, "realtek,disable-leds"); +#ifdef CONFIG_RESET_CONTROLLER + priv->reset_ctl = devm_reset_control_get(dev, "switch"); + if (IS_ERR(priv->reset_ctl)) { + dev_err(dev, "failed to get switch reset control\n"); + return PTR_ERR(priv->reset_ctl); + } +#endif + priv->reset = devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_LOW); if (IS_ERR(priv->reset)) { dev_err(dev, "failed to get RESET GPIO\n"); return PTR_ERR(priv->reset); } - - if (priv->reset) { - gpiod_set_value(priv->reset, 1); + if (priv->reset_ctl || priv->reset) { + realtek_mdio_hwreset(priv, 1); dev_dbg(dev, "asserted RESET\n"); msleep(REALTEK_HW_STOP_DELAY); - gpiod_set_value(priv->reset, 0); + realtek_mdio_hwreset(priv, 0); msleep(REALTEK_HW_START_DELAY); dev_dbg(dev, "deasserted RESET\n"); } @@ -246,8 +269,7 @@ static void realtek_mdio_remove(struct mdio_device *mdiodev) dsa_unregister_switch(priv->ds); /* leave the device reset asserted */ - if (priv->reset) - gpiod_set_value(priv->reset, 1); + realtek_mdio_hwreset(priv, 1); } static void realtek_mdio_shutdown(struct mdio_device *mdiodev) diff --git a/drivers/net/dsa/realtek/realtek-smi.c b/drivers/net/dsa/realtek/realtek-smi.c index bfd11591faf4..751159d71223 100644 --- a/drivers/net/dsa/realtek/realtek-smi.c +++ b/drivers/net/dsa/realtek/realtek-smi.c @@ -408,6 +408,23 @@ static int realtek_smi_setup_mdio(struct dsa_switch *ds) return ret; } +static int realtek_smi_hwreset(struct realtek_priv *priv, bool active) +{ +#ifdef CONFIG_RESET_CONTROLLER + if (priv->reset_ctl) { + if (active) + return reset_control_assert(priv->reset_ctl); + else + return reset_control_deassert(priv->reset_ctl); + } +#endif + + if (priv->reset) + gpiod_set_value(priv->reset, active); + + return 0; +} + static int realtek_smi_probe(struct platform_device *pdev) { const struct realtek_variant *var; @@ -457,18 +474,24 @@ static int realtek_smi_probe(struct platform_device *pdev) dev_set_drvdata(dev, priv); spin_lock_init(&priv->lock); - /* TODO: if power is software controlled, set up any regulators here */ +#ifdef CONFIG_RESET_CONTROLLER + priv->reset_ctl = devm_reset_control_get(dev, "switch"); + if (IS_ERR(priv->reset_ctl)) { + dev_err(dev, "failed to get switch reset control\n"); + return PTR_ERR(priv->reset_ctl); + } +#endif priv->reset = devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_LOW); if (IS_ERR(priv->reset)) { dev_err(dev, "failed to get RESET GPIO\n"); return PTR_ERR(priv->reset); } - if (priv->reset) { - gpiod_set_value(priv->reset, 1); + if (priv->reset_ctl || priv->reset) { + realtek_smi_hwreset(priv, 1); dev_dbg(dev, "asserted RESET\n"); msleep(REALTEK_HW_STOP_DELAY); - gpiod_set_value(priv->reset, 0); + realtek_smi_hwreset(priv, 0); msleep(REALTEK_HW_START_DELAY); dev_dbg(dev, "deasserted RESET\n"); } @@ -518,8 +541,7 @@ static void realtek_smi_remove(struct platform_device *pdev) of_node_put(priv->slave_mii_bus->dev.of_node); /* leave the device reset asserted */ - if (priv->reset) - gpiod_set_value(priv->reset, 1); + realtek_smi_hwreset(priv, 1); } static void realtek_smi_shutdown(struct platform_device *pdev) diff --git a/drivers/net/dsa/realtek/realtek.h b/drivers/net/dsa/realtek/realtek.h index 4fa7c6ba874a..ad61e5c13f96 100644 --- a/drivers/net/dsa/realtek/realtek.h +++ b/drivers/net/dsa/realtek/realtek.h @@ -12,6 +12,9 @@ #include <linux/platform_device.h> #include <linux/gpio/consumer.h> #include <net/dsa.h> +#ifdef CONFIG_RESET_CONTROLLER +#include <linux/reset.h> +#endif #define REALTEK_HW_STOP_DELAY 25 /* msecs */ #define REALTEK_HW_START_DELAY 100 /* msecs */ @@ -48,6 +51,9 @@ struct rtl8366_vlan_4k { struct realtek_priv { struct device *dev; +#ifdef CONFIG_RESET_CONTROLLER + struct reset_control *reset_ctl; +#endif struct gpio_desc *reset; struct gpio_desc *mdc; struct gpio_desc *mdio;
The 'reset-gpios' will not work when the switch reset is controlled by a reset controller. Although the reset is optional and the driver performs a soft reset during setup, if the initial reset state was asserted, the driver will not detect it. This is an example of how to use the reset controller: switch { compatible = "realtek,rtl8366rb"; resets = <&rst 8>; reset-names = "switch"; ... } The reset controller will take precedence over the reset GPIO. Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com> --- drivers/net/dsa/realtek/realtek-mdio.c | 36 +++++++++++++++++++++----- drivers/net/dsa/realtek/realtek-smi.c | 34 +++++++++++++++++++----- drivers/net/dsa/realtek/realtek.h | 6 +++++ 3 files changed, 63 insertions(+), 13 deletions(-)