[PATCH v2 0/3] pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

[PATCH v2 0/3] pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

[Yoshihiro Shimoda]

This patch series is based on renesas-drivers.git /
renesas-drivers-2019-07-30-v5.3-rc2 tag.

About R-Car PWM driver modification, it seems to need more time to achieve
output duty zero about suitable gpio vs pinctrl handling. But, I believe
this pinctrl patches could be applied into the mainline regardless
the R-Car PWM modification.

Changes from v1:
 - Spin-off the pinctrl patches (from 1/7 to 3/7).
 - Add Geert-san's Reviewed-by on 1/3 and 2/3.
 - Add Fixes tag on 2/3.
https://patchwork.kernel.org/project/linux-renesas-soc/list/?series=143129

Yoshihiro Shimoda (3):
  pinctrl: sh-pfc: add new flags into struct sh_pfc_pin_config
  pinctrl: sh-pfc: remove incomplete flag "cfg->type"
  pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

 drivers/pinctrl/sh-pfc/pinctrl.c | 45 ++++++++++++++++++++--------------------
 1 file changed, 23 insertions(+), 22 deletions(-)

-- 
2.7.4

[PATCH v2 1/3] pinctrl: sh-pfc: add new flags into struct sh_pfc_pin_config

[Yoshihiro Shimoda]

To clean/modify the code up later, this patch just adds new flags
"mux_set" and "gpio_enabled" into the struct sh_pfc_pin_config.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/pinctrl/sh-pfc/pinctrl.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/pinctrl/sh-pfc/pinctrl.c b/drivers/pinctrl/sh-pfc/pinctrl.c
index 2824be4..864da24 100644
--- a/drivers/pinctrl/sh-pfc/pinctrl.c
+++ b/drivers/pinctrl/sh-pfc/pinctrl.c
@@ -27,6 +27,8 @@
 
 struct sh_pfc_pin_config {
 	u32 type;
+	bool mux_set;
+	bool gpio_enabled;
 };
 
 struct sh_pfc_pinctrl {
@@ -364,7 +366,15 @@ static int sh_pfc_func_set_mux(struct pinctrl_dev *pctldev, unsigned selector,
 	for (i = 0; i < grp->nr_pins; ++i) {
 		ret = sh_pfc_config_mux(pfc, grp->mux[i], PINMUX_TYPE_FUNCTION);
 		if (ret < 0)
-			break;
+			goto done;
+	}
+
+	/* All group pins are configured, mark the pins as mux_set */
+	for (i = 0; i < grp->nr_pins; ++i) {
+		int idx = sh_pfc_get_pin_index(pfc, grp->pins[i]);
+		struct sh_pfc_pin_config *cfg = &pmx->configs[idx];
+
+		cfg->mux_set = true;
 	}
 
 done:
@@ -405,6 +415,7 @@ static int sh_pfc_gpio_request_enable(struct pinctrl_dev *pctldev,
 	}
 
 	cfg->type = PINMUX_TYPE_GPIO;
+	cfg->gpio_enabled = true;
 
 	ret = 0;
 
@@ -426,6 +437,7 @@ static void sh_pfc_gpio_disable_free(struct pinctrl_dev *pctldev,
 
 	spin_lock_irqsave(&pfc->lock, flags);
 	cfg->type = PINMUX_TYPE_NONE;
+	cfg->gpio_enabled = false;
 	spin_unlock_irqrestore(&pfc->lock, flags);
 }
 
-- 
2.7.4

Re: [PATCH v2 1/3] pinctrl: sh-pfc: add new flags into struct sh_pfc_pin_config

[Simon Horman]

On Thu, Aug 08, 2019 at 03:19:01PM +0900, Yoshihiro Shimoda wrote:
> To clean/modify the code up later, this patch just adds new flags
> "mux_set" and "gpio_enabled" into the struct sh_pfc_pin_config.
> 
> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

Reviewed-by: Simon Horman <horms+renesas@verge.net.au>

[PATCH v2 2/3] pinctrl: sh-pfc: remove incomplete flag "cfg->type"

[Yoshihiro Shimoda]

The old commit c58d9c1b26e3 ("sh-pfc: Implement generic pinconf
support") broke the cfg->type flag to PINMUX_TYPE_FUNCTION because
sh_pfc_pinconf_set() didn't call sh_pfc_reconfig_pin().
Now if we fix the cfg->type condition, it gets worse because:
 - Some drivers might be deferred so that .set_mux() will be called
   multiple times.
 - In such the case, the sh-pfc driver returns -EBUSY even if
   the group is the same, and then that driver fails to probe.

Since the pinctrl subsystem already has such conditions according
to @set_mux and @gpio_request_enable, this patch just remove
the incomplete flag from sh-pfc/pinctrl.c.

Fixes: c58d9c1b26e3 ("sh-pfc: Implement generic pinconf support")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/pinctrl/sh-pfc/pinctrl.c | 26 --------------------------
 1 file changed, 26 deletions(-)

diff --git a/drivers/pinctrl/sh-pfc/pinctrl.c b/drivers/pinctrl/sh-pfc/pinctrl.c
index 864da24..ab2aa93 100644
--- a/drivers/pinctrl/sh-pfc/pinctrl.c
+++ b/drivers/pinctrl/sh-pfc/pinctrl.c
@@ -26,7 +26,6 @@
 #include "../pinconf.h"
 
 struct sh_pfc_pin_config {
-	u32 type;
 	bool mux_set;
 	bool gpio_enabled;
 };
@@ -354,16 +353,6 @@ static int sh_pfc_func_set_mux(struct pinctrl_dev *pctldev, unsigned selector,
 	spin_lock_irqsave(&pfc->lock, flags);
 
 	for (i = 0; i < grp->nr_pins; ++i) {
-		int idx = sh_pfc_get_pin_index(pfc, grp->pins[i]);
-		struct sh_pfc_pin_config *cfg = &pmx->configs[idx];
-
-		if (cfg->type != PINMUX_TYPE_NONE) {
-			ret = -EBUSY;
-			goto done;
-		}
-	}
-
-	for (i = 0; i < grp->nr_pins; ++i) {
 		ret = sh_pfc_config_mux(pfc, grp->mux[i], PINMUX_TYPE_FUNCTION);
 		if (ret < 0)
 			goto done;
@@ -395,14 +384,6 @@ static int sh_pfc_gpio_request_enable(struct pinctrl_dev *pctldev,
 
 	spin_lock_irqsave(&pfc->lock, flags);
 
-	if (cfg->type != PINMUX_TYPE_NONE) {
-		dev_err(pfc->dev,
-			"Pin %u is busy, can't configure it as GPIO.\n",
-			offset);
-		ret = -EBUSY;
-		goto done;
-	}
-
 	if (!pfc->gpio) {
 		/* If GPIOs are handled externally the pin mux type need to be
 		 * set to GPIO here.
@@ -414,7 +395,6 @@ static int sh_pfc_gpio_request_enable(struct pinctrl_dev *pctldev,
 			goto done;
 	}
 
-	cfg->type = PINMUX_TYPE_GPIO;
 	cfg->gpio_enabled = true;
 
 	ret = 0;
@@ -436,7 +416,6 @@ static void sh_pfc_gpio_disable_free(struct pinctrl_dev *pctldev,
 	unsigned long flags;
 
 	spin_lock_irqsave(&pfc->lock, flags);
-	cfg->type = PINMUX_TYPE_NONE;
 	cfg->gpio_enabled = false;
 	spin_unlock_irqrestore(&pfc->lock, flags);
 }
@@ -450,7 +429,6 @@ static int sh_pfc_gpio_set_direction(struct pinctrl_dev *pctldev,
 	int new_type = input ? PINMUX_TYPE_INPUT : PINMUX_TYPE_OUTPUT;
 	int idx = sh_pfc_get_pin_index(pfc, offset);
 	const struct sh_pfc_pin *pin = &pfc->info->pins[idx];
-	struct sh_pfc_pin_config *cfg = &pmx->configs[idx];
 	unsigned long flags;
 	unsigned int dir;
 	int ret;
@@ -470,8 +448,6 @@ static int sh_pfc_gpio_set_direction(struct pinctrl_dev *pctldev,
 	if (ret < 0)
 		goto done;
 
-	cfg->type = new_type;
-
 done:
 	spin_unlock_irqrestore(&pfc->lock, flags);
 	return ret;
@@ -794,13 +770,11 @@ static int sh_pfc_map_pins(struct sh_pfc *pfc, struct sh_pfc_pinctrl *pmx)
 
 	for (i = 0; i < pfc->info->nr_pins; ++i) {
 		const struct sh_pfc_pin *info = &pfc->info->pins[i];
-		struct sh_pfc_pin_config *cfg = &pmx->configs[i];
 		struct pinctrl_pin_desc *pin = &pmx->pins[i];
 
 		/* If the pin number is equal to -1 all pins are considered */
 		pin->number = info->pin != (u16)-1 ? info->pin : i;
 		pin->name = info->name;
-		cfg->type = PINMUX_TYPE_NONE;
 	}
 
 	return 0;
-- 
2.7.4

Re: [PATCH v2 2/3] pinctrl: sh-pfc: remove incomplete flag "cfg->type"

[Simon Horman]

On Thu, Aug 08, 2019 at 03:19:02PM +0900, Yoshihiro Shimoda wrote:
> The old commit c58d9c1b26e3 ("sh-pfc: Implement generic pinconf
> support") broke the cfg->type flag to PINMUX_TYPE_FUNCTION because
> sh_pfc_pinconf_set() didn't call sh_pfc_reconfig_pin().
> Now if we fix the cfg->type condition, it gets worse because:
>  - Some drivers might be deferred so that .set_mux() will be called
>    multiple times.
>  - In such the case, the sh-pfc driver returns -EBUSY even if
>    the group is the same, and then that driver fails to probe.
> 
> Since the pinctrl subsystem already has such conditions according
> to @set_mux and @gpio_request_enable, this patch just remove
> the incomplete flag from sh-pfc/pinctrl.c.
> 
> Fixes: c58d9c1b26e3 ("sh-pfc: Implement generic pinconf support")
> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

Reviewed-by: Simon Horman <horms+renesas@verge.net.au>

[PATCH v2 3/3] pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

[Yoshihiro Shimoda]

R-Car PWM controller requires the gpio to output zero duty,
this patch allows to roll it back from gpio to mux when the gpio
is freed.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
---
 drivers/pinctrl/sh-pfc/pinctrl.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/drivers/pinctrl/sh-pfc/pinctrl.c b/drivers/pinctrl/sh-pfc/pinctrl.c
index ab2aa93..c9bdca5 100644
--- a/drivers/pinctrl/sh-pfc/pinctrl.c
+++ b/drivers/pinctrl/sh-pfc/pinctrl.c
@@ -26,6 +26,7 @@
 #include "../pinconf.h"
 
 struct sh_pfc_pin_config {
+	unsigned int mux_mark;
 	bool mux_set;
 	bool gpio_enabled;
 };
@@ -353,6 +354,16 @@ static int sh_pfc_func_set_mux(struct pinctrl_dev *pctldev, unsigned selector,
 	spin_lock_irqsave(&pfc->lock, flags);
 
 	for (i = 0; i < grp->nr_pins; ++i) {
+		int idx = sh_pfc_get_pin_index(pfc, grp->pins[i]);
+		struct sh_pfc_pin_config *cfg = &pmx->configs[idx];
+
+		/*
+		 * This driver cannot manage both gpio and mux when the gpio
+		 * pin is already enabled. So, this function failed.
+		 */
+		if (cfg->gpio_enabled)
+			return -EBUSY;
+
 		ret = sh_pfc_config_mux(pfc, grp->mux[i], PINMUX_TYPE_FUNCTION);
 		if (ret < 0)
 			goto done;
@@ -364,6 +375,7 @@ static int sh_pfc_func_set_mux(struct pinctrl_dev *pctldev, unsigned selector,
 		struct sh_pfc_pin_config *cfg = &pmx->configs[idx];
 
 		cfg->mux_set = true;
+		cfg->mux_mark = grp->mux[i];
 	}
 
 done:
@@ -417,6 +429,9 @@ static void sh_pfc_gpio_disable_free(struct pinctrl_dev *pctldev,
 
 	spin_lock_irqsave(&pfc->lock, flags);
 	cfg->gpio_enabled = false;
+	/* If mux is already set, this configures it here */
+	if (cfg->mux_set)
+		sh_pfc_config_mux(pfc, cfg->mux_mark, PINMUX_TYPE_FUNCTION);
 	spin_unlock_irqrestore(&pfc->lock, flags);
 }
 
-- 
2.7.4

Re: [PATCH v2 3/3] pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

[Geert Uytterhoeven]

On Thu, Aug 8, 2019 at 8:20 AM Yoshihiro Shimoda
<yoshihiro.shimoda.uh@renesas.com> wrote:
> R-Car PWM controller requires the gpio to output zero duty,
> this patch allows to roll it back from gpio to mux when the gpio
> is freed.
>
> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH v2 3/3] pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

[Geert Uytterhoeven]

On Thu, Aug 8, 2019 at 10:29 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> On Thu, Aug 8, 2019 at 8:20 AM Yoshihiro Shimoda
> <yoshihiro.shimoda.uh@renesas.com> wrote:
> > R-Car PWM controller requires the gpio to output zero duty,
> > this patch allows to roll it back from gpio to mux when the gpio
> > is freed.
> >
> > Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
>
> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

FTR: as the PWM work is put on hold, I have changed the commit
description to:

    pinctrl: sh-pfc: Rollback to mux if required when the gpio is freed

    Some drivers require switching between function and gpio at run-time.
    Allow to roll back from gpio to mux when the gpio is freed.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH v2 3/3] pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

[Simon Horman]

On Thu, Aug 08, 2019 at 03:19:03PM +0900, Yoshihiro Shimoda wrote:
> R-Car PWM controller requires the gpio to output zero duty,
> this patch allows to roll it back from gpio to mux when the gpio
> is freed.
> 
> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>

Reviewed-by: Simon Horman <horms+renesas@verge.net.au>

Re: [PATCH v2 0/3] pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed

[Geert Uytterhoeven]

Hi Shimoda-san,

On Thu, Aug 8, 2019 at 8:20 AM Yoshihiro Shimoda
<yoshihiro.shimoda.uh@renesas.com> wrote:
> This patch series is based on renesas-drivers.git /
> renesas-drivers-2019-07-30-v5.3-rc2 tag.
>
> About R-Car PWM driver modification, it seems to need more time to achieve
> output duty zero about suitable gpio vs pinctrl handling. But, I believe
> this pinctrl patches could be applied into the mainline regardless
> the R-Car PWM modification.

Thank you! I plan to apply this next week, after I've sent a first sh-pfc pull
request for v5.4, and let it soak in renesas-drivers.

> Changes from v1:
>  - Spin-off the pinctrl patches (from 1/7 to 3/7).
>  - Add Geert-san's Reviewed-by on 1/3 and 2/3.
>  - Add Fixes tag on 2/3.
> https://patchwork.kernel.org/project/linux-renesas-soc/list/?series=143129
>
> Yoshihiro Shimoda (3):
>   pinctrl: sh-pfc: add new flags into struct sh_pfc_pin_config
>   pinctrl: sh-pfc: remove incomplete flag "cfg->type"
>   pinctrl: sh-pfc: Rollback to mux if requires when the gpio is freed
>
>  drivers/pinctrl/sh-pfc/pinctrl.c | 45 ++++++++++++++++++++--------------------
>  1 file changed, 23 insertions(+), 22 deletions(-)

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds