ASoC: wm8903: add regulator handling

Message ID 20170320091352.4115-1-linus.walleij@linaro.org
State Not Applicable, archived
Headers show

Commit Message

Linus Walleij March 20, 2017, 9:13 a.m.
The WM8903 has four different voltage inputs: AVDD, CPVDD, DBVDD
and DCVDD. On the Qualcomm APQ8060 Dragonboard these are all
supplied from proper regulators and thus need activating and
binding.

This is a quick-and-dirty solution just grabbing and enabling the
regulator supplies on probe() and disabling them on remove() and
the errorpath. More elaborate power management is likely possible.

I assume the nVidia designs using this codec have some hard-wired
always-on power and will be happy with using the dummy regulators
for this. But someone from the nVidia camp should probably check
whether they can bind these to proper regulators instead.

We also amend the DT binding document. A small change like this
does not warrant a separate patch for augmenting these.

Cc: devicetree@vger.kernel.org
Cc: Mark Brown <broonie@kernel.org>
Cc: Liam Girdwood <lgirdwood@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
Mark: I was thinking about adding runtime PM for disabling
these regulators when unused, but I'm uncertain about the
interaction with DAPM in that regard. This atleast gives us
control over the supplies.
---
 Documentation/devicetree/bindings/sound/wm8903.txt | 13 +++++++++
 sound/soc/codecs/wm8903.c                          | 31 ++++++++++++++++++++++
 2 files changed, 44 insertions(+)

Comments

Charles Keepax March 20, 2017, 11:51 a.m. | #1
On Mon, Mar 20, 2017 at 10:13:52AM +0100, Linus Walleij wrote:
> The WM8903 has four different voltage inputs: AVDD, CPVDD, DBVDD
> and DCVDD. On the Qualcomm APQ8060 Dragonboard these are all
> supplied from proper regulators and thus need activating and
> binding.
> 
> This is a quick-and-dirty solution just grabbing and enabling the
> regulator supplies on probe() and disabling them on remove() and
> the errorpath. More elaborate power management is likely possible.
> 
> I assume the nVidia designs using this codec have some hard-wired
> always-on power and will be happy with using the dummy regulators
> for this. But someone from the nVidia camp should probably check
> whether they can bind these to proper regulators instead.
> 
> We also amend the DT binding document. A small change like this
> does not warrant a separate patch for augmenting these.
> 
> Cc: devicetree@vger.kernel.org
> Cc: Mark Brown <broonie@kernel.org>
> Cc: Liam Girdwood <lgirdwood@gmail.com>
> Cc: Stephen Warren <swarren@nvidia.com>
> Cc: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> Mark: I was thinking about adding runtime PM for disabling
> these regulators when unused, but I'm uncertain about the
> interaction with DAPM in that regard. This atleast gives us
> control over the supplies.

DAPM will hold a pm_runtime reference whilst the chip is active
so usually there isn't really much interaction to worry about.

I think the patch looks fine for a first pass adding the supplies
to me, someone can always add the suspend/resume and controlling
the supplies across that when they need it.

Acked-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>

Thanks,
Charles
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Linus Walleij March 22, 2017, 7:24 a.m. | #2
On Mon, Mar 20, 2017 at 12:51 PM, Charles Keepax
<ckeepax@opensource.wolfsonmicro.com> wrote:
> On Mon, Mar 20, 2017 at 10:13:52AM +0100, Linus Walleij wrote:

>> Mark: I was thinking about adding runtime PM for disabling
>> these regulators when unused, but I'm uncertain about the
>> interaction with DAPM in that regard. This atleast gives us
>> control over the supplies.
>
> DAPM will hold a pm_runtime reference whilst the chip is active
> so usually there isn't really much interaction to worry about.

OK, so since the codec is registered on i2c's client->dev I take
it I could use the runtime PM / suspend/resume callbacks on
that device, and it should just work.

> I think the patch looks fine for a first pass adding the supplies
> to me, someone can always add the suspend/resume and controlling
> the supplies across that when they need it.

I will make a try at this as a follow-on patch.

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Charles Keepax March 22, 2017, 9:06 a.m. | #3
On Wed, Mar 22, 2017 at 08:24:32AM +0100, Linus Walleij wrote:
> On Mon, Mar 20, 2017 at 12:51 PM, Charles Keepax
> <ckeepax@opensource.wolfsonmicro.com> wrote:
> > On Mon, Mar 20, 2017 at 10:13:52AM +0100, Linus Walleij wrote:
> 
> >> Mark: I was thinking about adding runtime PM for disabling
> >> these regulators when unused, but I'm uncertain about the
> >> interaction with DAPM in that regard. This atleast gives us
> >> control over the supplies.
> >
> > DAPM will hold a pm_runtime reference whilst the chip is active
> > so usually there isn't really much interaction to worry about.
> 
> OK, so since the codec is registered on i2c's client->dev I take
> it I could use the runtime PM / suspend/resume callbacks on
> that device, and it should just work.
> 

Yeah should do, wm8962 should provide a reasonably similar model
if you are looking for a handy example and feel free to drop me a
line if I can be of any help.

Thanks,
Charles
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Rob Herring March 24, 2017, 3:30 p.m. | #4
On Mon, Mar 20, 2017 at 10:13:52AM +0100, Linus Walleij wrote:
> The WM8903 has four different voltage inputs: AVDD, CPVDD, DBVDD
> and DCVDD. On the Qualcomm APQ8060 Dragonboard these are all
> supplied from proper regulators and thus need activating and
> binding.
> 
> This is a quick-and-dirty solution just grabbing and enabling the
> regulator supplies on probe() and disabling them on remove() and
> the errorpath. More elaborate power management is likely possible.
> 
> I assume the nVidia designs using this codec have some hard-wired
> always-on power and will be happy with using the dummy regulators
> for this. But someone from the nVidia camp should probably check
> whether they can bind these to proper regulators instead.
> 
> We also amend the DT binding document. A small change like this
> does not warrant a separate patch for augmenting these.
> 
> Cc: devicetree@vger.kernel.org
> Cc: Mark Brown <broonie@kernel.org>
> Cc: Liam Girdwood <lgirdwood@gmail.com>
> Cc: Stephen Warren <swarren@nvidia.com>
> Cc: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> Mark: I was thinking about adding runtime PM for disabling
> these regulators when unused, but I'm uncertain about the
> interaction with DAPM in that regard. This atleast gives us
> control over the supplies.
> ---
>  Documentation/devicetree/bindings/sound/wm8903.txt | 13 +++++++++

Acked-by: Rob Herring <robh@kernel.org>

>  sound/soc/codecs/wm8903.c                          | 31 ++++++++++++++++++++++
>  2 files changed, 44 insertions(+)
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Stephen Warren March 30, 2017, 7:42 p.m. | #5
On 03/20/2017 03:13 AM, Linus Walleij wrote:
> The WM8903 has four different voltage inputs: AVDD, CPVDD, DBVDD
> and DCVDD. On the Qualcomm APQ8060 Dragonboard these are all
> supplied from proper regulators and thus need activating and
> binding.
>
> This is a quick-and-dirty solution just grabbing and enabling the
> regulator supplies on probe() and disabling them on remove() and
> the errorpath. More elaborate power management is likely possible.
>
> I assume the nVidia designs using this codec have some hard-wired
> always-on power and will be happy with using the dummy regulators
> for this. But someone from the nVidia camp should probably check
> whether they can bind these to proper regulators instead.
>
> We also amend the DT binding document. A small change like this
> does not warrant a separate patch for augmenting these.

Tested-by: Stephen Warren <swarren@nvidia.com>

Sorry for the slow response.

Tested using NVIDIA springbank/seaboard board. Just in case anyone reads 
this later and wonders how: I will point out that in the baseline 
v4.11-rc4, the LCD (or perhaps just its backlight) doesn't work on this 
HW any more, but since the GUI login prompt plays a sound, I was still 
able to validate the WM8903 change.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/Documentation/devicetree/bindings/sound/wm8903.txt b/Documentation/devicetree/bindings/sound/wm8903.txt
index 94ec32c194bb..afc51caf1137 100644
--- a/Documentation/devicetree/bindings/sound/wm8903.txt
+++ b/Documentation/devicetree/bindings/sound/wm8903.txt
@@ -28,6 +28,14 @@  Optional properties:
     performed. If any entry has the value 0xffffffff, that GPIO's
     configuration will not be modified.
 
+  - AVDD-supply : Analog power supply regulator on the AVDD pin.
+
+  - CPVDD-supply : Charge pump supply regulator on the CPVDD pin.
+
+  - DBVDD-supply : Digital buffer supply regulator for the DBVDD pin.
+
+  - DCVDD-supply : Digital core supply regulator for the DCVDD pin.
+
 Pins on the device (for linking into audio routes):
 
   * IN1L
@@ -54,6 +62,11 @@  codec: wm8903@1a {
 	reg = <0x1a>;
 	interrupts = < 347 >;
 
+	AVDD-supply = <&fooreg_a>;
+	CPVDD-supply = <&fooreg_b>;
+	DBVDD-supply = <&fooreg_c>;
+	DCVDC-supply = <&fooreg_d>;
+
 	gpio-controller;
 	#gpio-cells = <2>;
 
diff --git a/sound/soc/codecs/wm8903.c b/sound/soc/codecs/wm8903.c
index 6e887c2c42b1..237eeb9a8b97 100644
--- a/sound/soc/codecs/wm8903.c
+++ b/sound/soc/codecs/wm8903.c
@@ -24,6 +24,7 @@ 
 #include <linux/pm.h>
 #include <linux/i2c.h>
 #include <linux/regmap.h>
+#include <linux/regulator/consumer.h>
 #include <linux/slab.h>
 #include <linux/irq.h>
 #include <linux/mutex.h>
@@ -115,10 +116,19 @@  static const struct reg_default wm8903_reg_defaults[] = {
 	{ 172, 0x0000 },    /* R172 - Analogue Output Bias 0 */
 };
 
+#define WM8903_NUM_SUPPLIES 4
+static const char *wm8903_supply_names[WM8903_NUM_SUPPLIES] = {
+	"AVDD",
+	"CPVDD",
+	"DBVDD",
+	"DCVDD",
+};
+
 struct wm8903_priv {
 	struct wm8903_platform_data *pdata;
 	struct device *dev;
 	struct regmap *regmap;
+	struct regulator_bulk_data supplies[WM8903_NUM_SUPPLIES];
 
 	int sysclk;
 	int irq;
@@ -2030,6 +2040,23 @@  static int wm8903_i2c_probe(struct i2c_client *i2c,
 
 	pdata = wm8903->pdata;
 
+	for (i = 0; i < ARRAY_SIZE(wm8903->supplies); i++)
+		wm8903->supplies[i].supply = wm8903_supply_names[i];
+
+	ret = devm_regulator_bulk_get(&i2c->dev, ARRAY_SIZE(wm8903->supplies),
+				      wm8903->supplies);
+	if (ret != 0) {
+		dev_err(&i2c->dev, "Failed to request supplies: %d\n", ret);
+		return ret;
+	}
+
+	ret = regulator_bulk_enable(ARRAY_SIZE(wm8903->supplies),
+				    wm8903->supplies);
+	if (ret != 0) {
+		dev_err(&i2c->dev, "Failed to enable supplies: %d\n", ret);
+		return ret;
+	}
+
 	ret = regmap_read(wm8903->regmap, WM8903_SW_RESET_AND_ID, &val);
 	if (ret != 0) {
 		dev_err(&i2c->dev, "Failed to read chip ID: %d\n", ret);
@@ -2160,6 +2187,8 @@  static int wm8903_i2c_probe(struct i2c_client *i2c,
 
 	return 0;
 err:
+	regulator_bulk_disable(ARRAY_SIZE(wm8903->supplies),
+			       wm8903->supplies);
 	return ret;
 }
 
@@ -2167,6 +2196,8 @@  static int wm8903_i2c_remove(struct i2c_client *client)
 {
 	struct wm8903_priv *wm8903 = i2c_get_clientdata(client);
 
+	regulator_bulk_disable(ARRAY_SIZE(wm8903->supplies),
+			       wm8903->supplies);
 	if (client->irq)
 		free_irq(client->irq, wm8903);
 	wm8903_free_gpio(wm8903);