Re: [PATCH v3 2/5] pinctrl: st: Enhance the controller to manage unavailable registers

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 






On 03/10/2014 10:17 AM, Lee Jones wrote:
From: Giuseppe Cavallaro <peppe.cavallaro@xxxxxx>

This patch adds a new logic inside the st pinctrl to manage
an unsupported scenario: some sysconfig are not available!

This is the case of STiH407 where, although documented, the
following registers from SYSCFG_FLASH have been removed from the SoC.

SYSTEM_CONFIG3040
    Output Enable pad control for all PIO Alternate Functions
and
SYSTEM_ CONFIG3050
    Pull Up pad control for all PIO Alternate Functions

Without managing this condition an imprecise external abort
will be detect.

To do this the patch also reviews the st_parse_syscfgs
and other routines to manipulate the registers only if
actually available.
In any case, for example the st_parse_syscfgs detected
an error condition but no action was made in the
st_pctl_probe_dt.

Signed-off-by: Maxime Coquelin <maxime.coquelin@xxxxxx>
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@xxxxxx>

These two SOBs need reordering.
Right,
this will be changed here and everywhere else in the series.


---
  drivers/pinctrl/pinctrl-st.c | 106 +++++++++++++++++++++++++------------------
  1 file changed, 61 insertions(+), 45 deletions(-)

diff --git a/drivers/pinctrl/pinctrl-st.c b/drivers/pinctrl/pinctrl-st.c
index 9fb66aa..1721611 100644
--- a/drivers/pinctrl/pinctrl-st.c
+++ b/drivers/pinctrl/pinctrl-st.c
@@ -410,25 +410,27 @@ static void st_pinconf_set_config(struct st_pio_control *pc,
  	unsigned int oe_value, pu_value, od_value;
  	unsigned long mask = BIT(pin);

-	regmap_field_read(output_enable, &oe_value);
-	regmap_field_read(pull_up, &pu_value);
-	regmap_field_read(open_drain, &od_value);
-
-	/* Clear old values */
-	oe_value &= ~mask;
-	pu_value &= ~mask;
-	od_value &= ~mask;
-
-	if (config & ST_PINCONF_OE)
-		oe_value |= mask;
-	if (config & ST_PINCONF_PU)
-		pu_value |= mask;
-	if (config & ST_PINCONF_OD)
-		od_value |= mask;
-
-	regmap_field_write(output_enable, oe_value);
-	regmap_field_write(pull_up, pu_value);
-	regmap_field_write(open_drain, od_value);
+	if (output_enable) {
+		regmap_field_read(output_enable, &oe_value);
+		oe_value &= ~mask;
+		if (config & ST_PINCONF_OE)
+			oe_value |= mask;
+		regmap_field_write(output_enable, oe_value);
+	}
+	if (pull_up) {
+		regmap_field_read(pull_up, &pu_value);
+		pu_value &= ~mask;
+		if (config & ST_PINCONF_PU)
+			pu_value |= mask;
+		regmap_field_write(pull_up, pu_value);
+	}
+	if (open_drain) {
+		regmap_field_read(open_drain, &od_value);
+		od_value &= ~mask;
+		if (config & ST_PINCONF_OD)
+			od_value |= mask;
+		regmap_field_write(open_drain, od_value);
+	}

Nice change.

Nit: For consistency with the changes below, please consider placing
new lines between the 3 outer checks.
Done


  }


<snip>

-static void st_pinconf_get_direction(struct st_pio_control *pc,
-	int pin, unsigned long *config)
+static void st_pinconf_get_direction(struct st_pio_control *pc, int pin,
+				     unsigned long *config)

Unrelated change?
Yes this is unrelated.
I removed this change. It will be sent later with maybe other cosmetic changes.


  {
  	unsigned int oe_value, pu_value, od_value;

Is it worth checking for (!config) here?

That would be better indeed.
But since that lack of safety check was already present before this change, it should be handled in a separate patch.

I'll add this in my todo list.


-	regmap_field_read(pc->oe, &oe_value);
-	regmap_field_read(pc->pu, &pu_value);
-	regmap_field_read(pc->od, &od_value);
+	if (pc->oe) {
+		regmap_field_read(pc->oe, &oe_value);
+		if (oe_value & BIT(pin))
+			ST_PINCONF_PACK_OE(*config);
+	}

-	if (oe_value & BIT(pin))
-		ST_PINCONF_PACK_OE(*config);
-	if (pu_value & BIT(pin))
-		ST_PINCONF_PACK_PU(*config);
-	if (od_value & BIT(pin))
-		ST_PINCONF_PACK_OD(*config);
+	if (pc->pu) {
+		regmap_field_read(pc->pu, &pu_value);
+		if (pu_value & BIT(pin))
+			ST_PINCONF_PACK_PU(*config);
+	}

+	if (pc->od) {
+		regmap_field_read(pc->od, &od_value);
+		if (od_value & BIT(pin))
+			ST_PINCONF_PACK_OD(*config);
+	}
  }

Nice.

  static int st_pinconf_get_retime_packed(struct st_pinctrl *info,
@@ -1105,8 +1116,21 @@ static int st_pctl_dt_setup_retime(struct st_pinctrl *info,
  	return -EINVAL;
  }

-static int st_parse_syscfgs(struct st_pinctrl *info,
-		int bank, struct device_node *np)
+
+static struct regmap_field *st_pc_get_value(struct device *dev,
+					    struct regmap *regmap, int bank,
+					    int data, int lsb, int msb)
+{
+	struct reg_field reg = REG_FIELD((data + bank) * 4, lsb, msb);
+
+	if (data < 0)
+		return NULL;

What happens is data < 0 and it's used in REG_FIELD?
Nothing bad, but I agree this is not crystal clear.


Would it make more sense to make this check before calling REG_FIELD?
Yes, it will be done in the v4.

<snip>

Thanks,
Maxime
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]
  Powered by Linux