Skip to contentSkip to conversation

twl6030-gpadc support for twl6037

PATCH Superseded Belisko Marek 11 participants 38 messages Jul 20, 2015 – Oct 13, 2015 2 messages not in the archive

Subsystems (3)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

Hi all,

I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs for gpadc but functionality is missing). My main focus is to add support for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support should be straightforward but dts file used in omap5-uevm using palmas driver and not twl6030. So it's completely confusing to me how to proceed or which direction is correct. Seems I'm missing some parts of puzzle ;) Thanks for any hints.

BR,

marek

Signature · 10 lines
as simple and primitive as possible
-------------------------------------------------
Marek Belisko - OPEN-NANDRA
Freelance Developer

Ruska Nova Ves 219 | Presov, 08005 Slovak Republic
Tel: +421 915 052 184
skype: marekwhite
twitter: #opennandra
web: http://open-nandra.com
#1 lore raw

Adding also twl6030-gpadc driver authors to loop

Quoted 29 lines · Belisko Marek

Hi all,

I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs for gpadc but functionality is missing). My main focus is to add support for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support should be straightforward but dts file used in omap5-uevm using palmas driver and not twl6030. So it's completely confusing to me how to proceed or which direction is correct. Seems I'm missing some parts of puzzle ;) Thanks for any hints.

BR,

marek

--
as simple and primitive as possible
-------------------------------------------------
Marek Belisko - OPEN-NANDRA
Freelance Developer

Ruska Nova Ves 219 | Presov, 08005 Slovak Republic
Tel: +421 915 052 184
skype: marekwhite
twitter: #opennandra
web: http://open-nandra.com

#2 lore raw
Quoted 18 lines · Belisko Marek

Adding also twl6030-gpadc driver authors to loop

Quoted 14 lines · Belisko Marek

Hi all,

I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs for gpadc but functionality is missing). My main focus is to add support for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support should be straightforward but dts file used in omap5-uevm using palmas driver and not twl6030. So it's completely confusing to me how to proceed or which direction is correct. Seems I'm missing some parts of puzzle ;) Thanks for any hints.

Phoenix is twl6030, twl6032

Palmas is twl6035, twl6037 and maybe others.

Phoenix and Palmas are two totally different chips but marketing gave them numbers near each other for unknown reasons.

Graeme

#3 lore raw
Quoted 21 lines · Graeme Gregory
Quoted 18 lines · Belisko Marek

Adding also twl6030-gpadc driver authors to loop

Quoted 14 lines · Belisko Marek

Hi all,

I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs for gpadc but functionality is missing). My main focus is to add support for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support should be straightforward but dts file used in omap5-uevm using palmas driver and not twl6030. So it's completely confusing to me how to proceed or which direction is correct. Seems I'm missing some parts of puzzle ;) Thanks for any hints.

Quoted 7 lines

Phoenix is twl6030, twl6032

Palmas is twl6035, twl6037 and maybe others.

Phoenix and Palmas are two totally different chips but marketing gave them numbers near each other for unknown reasons.

Also this might be useful.

https://github.com/slimlogic/linux/commit/312a6a1c8013fd056431c31567b43d8eef40f333

A pre IIO palmas ADC driver which I never got around to upstreaming because TIpocalypse left me with no access to a palmas chip.

Graeme

#4 lore raw

Am 22.07.2015 um 21:41 schrieb Graeme Gregory <[email protected]>:

Quoted 9 lines

=20
=20

Quoted 6 lines · Graeme Gregory

=20
=20

Quoted 3 lines · Belisko Marek

Adding also twl6030-gpadc driver authors to loop
=20
On Mon, Jul 20, 2015 at 8:57 AM, Belisko Marek =

<[email protected]>

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines · wrote:

Hi all,
=20
I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs =

for

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

gpadc but functionality is missing). My main focus is to add =

support

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support =

should

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

be straightforward but dts file used in omap5-uevm using palmas =

driver

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

and not twl6030. So it's completely confusing to me how to proceed =

or

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

which direction is correct. Seems I'm missing some parts of puzzle =

;)

Quoted 9 lines
Quoted 9 lines
Quoted 2 lines
Quoted 2 lines

Thanks for any hints.
=20

Phoenix is twl6030, twl6032
=20
Palmas is twl6035, twl6037 and maybe others.
=20
Phoenix and Palmas are two totally different chips but marketing gave them numbers near each other for unknown reasons.
=20

Quoted 3 lines

Also this might be useful.
=20
=

https://github.com/slimlogic/linux/commit/312a6a1c8013fd056431c31567b43d8e=
ef40f333

That is basically what we are looking for.

We will give it a try asap and let you know results.

Quoted 2 lines

=20
A pre IIO palmas ADC driver which I never got around to upstreaming

Well, unless it already exists somewhere, it should be converterd to iio = by updating the
driver structure to follow iio/adc/twl6030-gpadc.

Quoted 1 line

because TIpocalypse left me with no access to a palmas chip.

The OMAP5432EVM by SVtronics [1] comes with a twl6037 (and it is supported by mainline device tree omap5-uevm.dts). And we hope that the pyra-handheld.com hardware is finished soon. For hardware bring-up it would be good to already have the gpadc available - so we have some hen-and-egg situation here.

BR and thanks,
Nikolaus

[1]: http://www.svtronics.com/5432

#5 lore raw

Re: gpadc iio support for tlw6037/palmas [was: twl6030-gpadc support for twl6037]

In reply to Graeme Gregory

Hi Graeme,

Am 22.07.2015 um 21:41 schrieb Graeme Gregory <gg-kDsPt+C1G03kYMGBc/[email protected]>:

Quoted 31 lines · Graeme Gregory
Quoted 21 lines · Belisko Marek
Quoted 18 lines

Adding also twl6030-gpadc driver authors to loop

Quoted 14 lines · Belisko Marek

Hi all,

I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs for gpadc but functionality is missing). My main focus is to add support for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support should be straightforward but dts file used in omap5-uevm using palmas driver and not twl6030. So it's completely confusing to me how to proceed or which direction is correct. Seems I'm missing some parts of puzzle ;) Thanks for any hints.

Quoted 7 lines

Phoenix is twl6030, twl6032

Palmas is twl6035, twl6037 and maybe others.

Phoenix and Palmas are two totally different chips but marketing gave them numbers near each other for unknown reasons.

Quoted 6 lines

Also this might be useful.

https://github.com/slimlogic/linux/commit/312a6a1c8013fd056431c31567b43d8eef40f333

A pre IIO palmas ADC driver which I never got around to upstreaming because TIpocalypse left me with no access to a palmas chip.

We managed to integrate your code into our 4.1 kernel (needed minor fixes mainly for irq initialization) and it appears to work on the omap5432evm. Especially channels in6_channel and in7_channel in /sys/bus/platform/drivers/palmas-gpadc/48070000.i2c:palmas@48:gpadc show reasonable (and slightly fluctuating) volt values.

Now the question is how to proceed. We could either write some drivers/iio/adc/palmas-gpadc.c wrapper around your driver, or could try this one which appears to be a pure iio implementation:

https://android.googlesource.com/kernel/tegra.git/+/28f245bb8f590f61bb244f1b4d03f4b4919d32ea/Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt
https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc/palmas_gpadc.c

Do you think the latter has a chance to be accepted upstream?

The reason we ask is that we neither want to write new code if something exists nor base our work on something that can not be upstreamed.

We have added the author(s) of this tegra/android driver so they can please also comment on their plans for submitting the driver upstream.

BR and thank you,
Nikolaus and Marek

#6 lore raw

Re: gpadc iio support for tlw6037/palmas [was: twl6030-gpadc support for twl6037]

Quoted 50 lines · Dr. H. Nikolaus Schaller

Hi Graeme,

Am 22.07.2015 um 21:41 schrieb Graeme Gregory <gg-kDsPt+C1G03kYMGBc/[email protected]>:

Quoted 31 lines · Graeme Gregory
Quoted 21 lines · Belisko Marek
Quoted 18 lines

Adding also twl6030-gpadc driver authors to loop

Quoted 14 lines · Belisko Marek

Hi all,

I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs for gpadc but functionality is missing). My main focus is to add support for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support should be straightforward but dts file used in omap5-uevm using palmas driver and not twl6030. So it's completely confusing to me how to proceed or which direction is correct. Seems I'm missing some parts of puzzle ;) Thanks for any hints.

Quoted 7 lines

Phoenix is twl6030, twl6032

Palmas is twl6035, twl6037 and maybe others.

Phoenix and Palmas are two totally different chips but marketing gave them numbers near each other for unknown reasons.

Quoted 6 lines

Also this might be useful.

https://github.com/slimlogic/linux/commit/312a6a1c8013fd056431c31567b43d8eef40f333

A pre IIO palmas ADC driver which I never got around to upstreaming because TIpocalypse left me with no access to a palmas chip.

We managed to integrate your code into our 4.1 kernel (needed minor fixes mainly for
irq initialization) and it appears to work on the omap5432evm. Especially channels
in6_channel and in7_channel in /sys/bus/platform/drivers/palmas-gpadc/48070000.i2c:palmas@48:gpadc show reasonable (and slightly fluctuating) volt values.

Quoted 5 lines

Now the question is how to proceed. We could either write some drivers/iio/adc/palmas-gpadc.c
wrapper around your driver, or could try this one which appears to be a pure iio implementation:

Pure IIO is better, one of the reasons my driver never got sent upstream is I knew IIO was due "real soon now" but I lost access to the chip and documentation before IIO landed in the kernel.

Quoted 5 lines

https://android.googlesource.com/kernel/tegra.git/+/28f245bb8f590f61bb244f1b4d03f4b4919d32ea/Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt
https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc/palmas_gpadc.c

Do you think the latter has a chance to be accepted upstream?

I see nothing with a quick scan that looks controversial.

Quoted 4 lines

The reason we ask is that we neither want to write new code if something exists nor base
our work on something that can not be upstreamed.

I would base your product on the nvidia driver and if they are not planning to upstream then submit yourself.

Graeme

#7 lore raw

Re: gpadc iio support for tlw6037/palmas [was: twl6030-gpadc support for twl6037]

Quoted 62 lines · Graeme Gregory
Quoted 50 lines · Dr. H. Nikolaus Schaller

Hi Graeme,

Am 22.07.2015 um 21:41 schrieb Graeme Gregory <[email protected]>:

Quoted 31 lines · Graeme Gregory
Quoted 21 lines · Belisko Marek
Quoted 18 lines

Adding also twl6030-gpadc driver authors to loop

Quoted 14 lines · Belisko Marek

Hi all,

I'm trying to figure out what is relation between palmas mfd driver and twl-core for twl603x devices.
twl-core driver seems to be older than palmas driver but seems both export same functionality (except that palmas have defined structs for gpadc but functionality is missing). My main focus is to add support for twl6037 to gpadc iio driver in iio/adc/twl6030-gpadc which is using twl-core for twl6030 and twl6032. Adding twl6037 support should be straightforward but dts file used in omap5-uevm using palmas driver and not twl6030. So it's completely confusing to me how to proceed or which direction is correct. Seems I'm missing some parts of puzzle ;) Thanks for any hints.

Quoted 7 lines

Phoenix is twl6030, twl6032

Palmas is twl6035, twl6037 and maybe others.

Phoenix and Palmas are two totally different chips but marketing gave them numbers near each other for unknown reasons.

Quoted 6 lines

Also this might be useful.

https://github.com/slimlogic/linux/commit/312a6a1c8013fd056431c31567b43d8eef40f333

A pre IIO palmas ADC driver which I never got around to upstreaming because TIpocalypse left me with no access to a palmas chip.

We managed to integrate your code into our 4.1 kernel (needed minor fixes mainly for
irq initialization) and it appears to work on the omap5432evm. Especially channels
in6_channel and in7_channel in /sys/bus/platform/drivers/palmas-gpadc/48070000.i2c:palmas@48:gpadc show reasonable (and slightly fluctuating) volt values.

Quoted 5 lines

Now the question is how to proceed. We could either write some drivers/iio/adc/palmas-gpadc.c
wrapper around your driver, or could try this one which appears to be a pure iio implementation:

Pure IIO is better, one of the reasons my driver never got sent upstream is I knew IIO was due "real soon now" but I lost access to the chip and documentation before IIO landed in the kernel.

Quoted 6 lines
Quoted 5 lines

https://android.googlesource.com/kernel/tegra.git/+/28f245bb8f590f61bb244f1b4d03f4b4919d32ea/Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt
https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc/palmas_gpadc.c

Do you think the latter has a chance to be accepted upstream?

I see nothing with a quick scan that looks controversial.

There are a few odd corners (based on a 20 second look). What is the dual channel stuff about?

Best bet might be to do a quick cleanup pass then post it as an RFC to linux-iio. Then we can take a proper look at it.

Quoted 7 lines
Quoted 5 lines

The reason we ask is that we neither want to write new code if something exists nor base
our work on something that can not be upstreamed.

I would base your product on the nvidia driver and if they are not planning to upstream then submit yourself.

Makes sense.

Quoted 7 lines

Graeme

--
To unsubscribe from this list: send the line "unsubscribe linux-iio" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#8 lore raw

cover Add Palmas iio gpadc

In reply to Graeme Gregory

Add iio driver for the TI Palmas (twl6035, 6037) including device tree bindings. It enables the gpadc for the OMAP5 uevm.

This patch series is based on original code taken from Android Tegra kernels:
(https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc)

Tegra code was developed by:
Pradeep Goudagunta <[email protected]>
Laxman Dewangan <[email protected]>

Edited and extended for mainline by:
H. Nikolaus Schaller <[email protected]>
Marek Belisko <[email protected]>

H. Nikolaus Schaller (2):
iio:adc: add iio driver for Palmas (twl6035/7) gpadc
ARM: dts: omap5-uevm: enable iio gpadc for Palmas

Marek Belisko (1):
iio:adc:palmas: add DT support

.../devicetree/bindings/iio/adc/palmas-gpadc.txt   |  67 ++
arch/arm/boot/dts/omap5-uevm.dts                   |  22 +
drivers/iio/adc/Kconfig                            |   9 +
drivers/iio/adc/Makefile                           |   1 +
drivers/iio/adc/palmas_gpadc.c                     | 883 +++++++++++++++++++++
include/linux/mfd/palmas.h                         |  59 +-
6 files changed, 1037 insertions(+), 4 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt create mode 100644 drivers/iio/adc/palmas_gpadc.c

Signature · 1 line
2.5.1
#9 lore raw

v2 cover Add Palmas iio gpadc

V2 changes:
- worked in review comments by Peter Meerwald
- made channels 1 and 3 report (raw) temperature values
- moved PALMAS_GPADC_TRIMINVALID to driver core
- worked in review comments by Jonathan Cameron
- iio_map removed

Add iio driver for the TI Palmas (twl6035, 6037) including device tree bindings. It enables the gpadc for the OMAP5 uevm.

This patch series is based on original code taken from Android Tegra kernels:
(https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc)

Tegra code was developed by:
Pradeep Goudagunta <[email protected]>
Laxman Dewangan <[email protected]>

Edited and extended for mainline by:
H. Nikolaus Schaller <[email protected]>
Marek Belisko <[email protected]>

H. Nikolaus Schaller (2):
iio:adc: add iio driver for Palmas (twl6035/7) gpadc
ARM: dts: omap5-uevm: enable iio gpadc for Palmas

Marek Belisko (1):
iio:adc:palmas: add DT support

.../devicetree/bindings/iio/adc/palmas-gpadc.txt   |  46 ++
arch/arm/boot/dts/omap5-uevm.dts                   |   9 +
drivers/iio/adc/Kconfig                            |   9 +
drivers/iio/adc/Makefile                           |   1 +
drivers/iio/adc/palmas_gpadc.c                     | 860 +++++++++++++++++++++
include/linux/mfd/palmas.h                         |  75 +-
6 files changed, 976 insertions(+), 24 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt create mode 100644 drivers/iio/adc/palmas_gpadc.c

Signature · 1 line
2.5.1

A message that is not in the archive (sent off-list or not yet received): [email protected]

#11 lore raw

PATCH v2 1/3 iio:adc: add iio driver for Palmas (twl6035/7) gpadc

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 818 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  75 ++--
4 files changed, 879 insertions(+), 24 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +0 -0
index eb0cd89..05a0368 100644
@@ -242,6 +242,15 @@ config NAU7802

To compile this driver as a module, choose M here: the module will be called nau7802.

+config PALMAS_GPADC
+ tristate "TI Palmas General Purpose ADC"
+ depends on MFD_PALMAS
+ help
+ Say yes here to build support for Texas Instruments Palmas.
+
+ Palmas series pmic chip (twl6035/6037) is used in smartphones and
+ tablets and supports a 16 channel general purpose ADC.
+
config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +818 -0

818 lines not on this page · raw message

include/linux/mfd/palmas.h +0 -0
index bb270bd..2fcd70e 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
			    struct regulator_config config);
};

+struct palmas_adc_wakeup_property {
+ int adc_channel_number;
+ int adc_high_threshold;
+ int adc_low_threshold;
+};
+

struct palmas_gpadc_platform_data {
	/* Channel 3 current source is only enabled during conversion */

- int ch3_current;
+ int ch3_current; /* 0: off; 1: 10uA; 2: 400uA; 3: 800 uA */

	/* Channel 0 current source can be used for battery detection.
	 * If used for battery detection this will cause a permanent current
	 * consumption depending on current level set here.
	 */

- int ch0_current;
+ int ch0_current; /* 0: off; 1: 5uA; 2: 15uA; 3: 20 uA */
+ bool extended_delay; /* use extended delay for conversion */

	/* default BAT_REMOVAL_DAT setting on device probe */
	int bat_removal;

	/* Sets the START_POLARITY bit in the RT_CTRL register */
	int start_polarity;

+
+ int auto_conversion_period_ms;
+ struct palmas_adc_wakeup_property *adc_wakeup1_data;
+ struct palmas_adc_wakeup_property *adc_wakeup2_data;
};

struct palmas_reg_init {

@@ -404,28 +415,7 @@ struct palmas_gpadc_calibration {
	s32 offset_error;
};

-struct palmas_gpadc {
- struct device *dev;
- struct palmas *palmas;
-
- int ch3_current;
- int ch0_current;
-
- int gpadc_force;
-
- int bat_removal;
-
- struct mutex reading_lock;
- struct completion irq_complete;
-
- int eoc_sw_irq;
-
- struct palmas_gpadc_calibration *palmas_cal_tbl;
-
- int conv0_channel;
- int conv1_channel;
- int rt_channel;
-};
+#define PALMAS_DATASHEET_NAME(_name) "palmas-gpadc-chan-"#_name

struct palmas_gpadc_result {
	s32 raw_code;

@@ -519,6 +509,43 @@ enum palmas_irqs {
PALMAS_NUM_IRQ,
};

+/* Palmas GPADC Channels */
+enum {
+ PALMAS_ADC_CH_IN0,
+ PALMAS_ADC_CH_IN1,
+ PALMAS_ADC_CH_IN2,
+ PALMAS_ADC_CH_IN3,
+ PALMAS_ADC_CH_IN4,
+ PALMAS_ADC_CH_IN5,
+ PALMAS_ADC_CH_IN6,
+ PALMAS_ADC_CH_IN7,
+ PALMAS_ADC_CH_IN8,
+ PALMAS_ADC_CH_IN9,
+ PALMAS_ADC_CH_IN10,
+ PALMAS_ADC_CH_IN11,
+ PALMAS_ADC_CH_IN12,
+ PALMAS_ADC_CH_IN13,
+ PALMAS_ADC_CH_IN14,
+ PALMAS_ADC_CH_IN15,
+ PALMAS_ADC_CH_MAX,
+};
+
+/* Palmas GPADC Channel0 Current Source */
+enum {
+ PALMAS_ADC_CH0_CURRENT_SRC_0,
+ PALMAS_ADC_CH0_CURRENT_SRC_5,
+ PALMAS_ADC_CH0_CURRENT_SRC_15,
+ PALMAS_ADC_CH0_CURRENT_SRC_20,
+};
+
+/* Palmas GPADC Channel3 Current Source */
+enum {
+ PALMAS_ADC_CH3_CURRENT_SRC_0,
+ PALMAS_ADC_CH3_CURRENT_SRC_10,
+ PALMAS_ADC_CH3_CURRENT_SRC_400,
+ PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+

struct palmas_pmic {
	struct palmas *palmas;
	struct device *dev;
Signature · 1 line
2.5.1
#12 lore raw

PATCH v2 2/3 iio:adc:palmas: add DT support

In reply to a message not in the archive

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]

---
.../devicetree/bindings/iio/adc/palmas-gpadc.txt | 46 +++++++++++++++++++
drivers/iio/adc/palmas_gpadc.c | 52 +++++++++++++++++++---

2 files changed, 93 insertions(+), 5 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +46 -0
new file mode 100644
index 0000000..2149afe
@@ -0,0 +1,46 @@
+* Palmas general purpose ADC IP block devicetree bindings
+
+Channels list:
+	0 battery type
+	1 battery temp NTC (optional current source)
+	2 GP
+	3 temp (with ext. diode, optional current source)
+	4 GP
+	5 GP
+	6 VBAT_SENSE
+	7 VCC_SENSE
+	8 Backup Battery voltage
+	9 external charger (VCHG)
+	10 VBUS
+	11 DC-DC current probe (how does this work?)
+	12 internal die temp
+	13 internal die temp
+	14 USB ID pin voltage
+	15 test network
+
+Required properties:
+- compatible : Must be "ti,palmas-gpadc".
+
+Optional sub-nodes:
+ti,channel0-current-microamp: Channel 0 current in uA.
+	Values are rounded to derive 0uA, 5uA, 15uA, 20uA.
+ti,channel3-current-microamp: Channel 3 current in uA.
+	Valid are rounded to derive 0uA, 10uA, 400uA, 800uA.
+ti,enable-extended-delay: Enable extended delay.
+
+Example:
+
+pmic {
+	compatible = "ti,twl6035-pmic", "ti,palmas-pmic";
+	...
+	gpadc {
+		compatible = "ti,palmas-gpadc";
+		interrupts = <18 0
+			      16 0
+			      17 0>;
+		ti,channel0-current-microamp = <5>;
+		ti,channel3-current-microamp = <10>;
+		};
+	};
+	...
+};
drivers/iio/adc/palmas_gpadc.c +0 -0
index 6805d81..1297c81 100644
@@ -21,6 +21,8 @@

#include <linux/pm.h>
#include <linux/mfd/palmas.h>
#include <linux/completion.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
#include <linux/iio/iio.h>
#include <linux/iio/machine.h>
#include <linux/iio/driver.h>
@@ -460,6 +462,34 @@ static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
};

+static int palmas_gpadc_get_adc_dt_data(struct platform_device *pdev,
+ struct palmas_gpadc_platform_data **gpadc_pdata)
+{
+ struct device_node *np = pdev->dev.of_node;
+ struct palmas_gpadc_platform_data *gp_data;
+ int ret;
+ u32 pval;
+
+ gp_data = devm_kzalloc(&pdev->dev, sizeof(*gp_data), GFP_KERNEL);
+ if (!gp_data)
+ return -ENOMEM;
+
+ ret = of_property_read_u32(np, "ti,channel0-current-microamp", &pval);
+ if (!ret)
+ gp_data->ch0_current = pval;
+
+ ret = of_property_read_u32(np, "ti,channel3-current-microamp", &pval);
+ if (!ret)
+ gp_data->ch3_current = pval;
+
+ gp_data->extended_delay = of_property_read_bool(np,
+ "ti,enable-extended-delay");
+
+ *gpadc_pdata = gp_data;
+
+ return 0;
+}
+
static int palmas_gpadc_probe(struct platform_device *pdev)

{
	struct palmas_gpadc *adc;
@@ -469,12 +499,17 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
	int ret, i;

	pdata = dev_get_platdata(pdev->dev.parent);

- if (!pdata || !pdata->gpadc_pdata) {
- dev_err(&pdev->dev, "No platform data\n");
- return -ENODEV;
- }

- gpadc_pdata = pdata->gpadc_pdata;
+ if (pdata && pdata->gpadc_pdata)
+ gpadc_pdata = pdata->gpadc_pdata;
+
+ if (!gpadc_pdata && pdev->dev.of_node) {
+ ret = palmas_gpadc_get_adc_dt_data(pdev, &gpadc_pdata);
+ if (ret < 0)
+ return ret;
+ }
+ if (!gpadc_pdata)
+ return -EINVAL;

	indio_dev = devm_iio_device_alloc(&pdev->dev, sizeof(*adc));
	if (!indio_dev) {

@@ -790,6 +825,12 @@ static const struct dev_pm_ops palmas_pm_ops = {
palmas_gpadc_resume)
};

+static const struct of_device_id of_palmas_gpadc_match_tbl[] = {
+ { .compatible = "ti,palmas-gpadc", },
+ { /* end */ }
+};
+MODULE_DEVICE_TABLE(of, of_palmas_gpadc_match_tbl);
+
static struct platform_driver palmas_gpadc_driver = {
.probe = palmas_gpadc_probe,
.remove = palmas_gpadc_remove,
@@ -797,6 +838,7 @@ static struct platform_driver palmas_gpadc_driver = {
.name = MOD_NAME,
.owner = THIS_MODULE,
.pm = &palmas_pm_ops,
+ .of_match_table = of_palmas_gpadc_match_tbl,

	},
};
Signature · 1 line
2.5.1
#13 lore raw

Re: [PATCH v2 2/3] iio:adc:palmas: add DT support

Quoted 38 lines · H. Nikolaus Schaller

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]

---
.../devicetree/bindings/iio/adc/palmas-gpadc.txt | 46 +++++++++++++++++++
drivers/iio/adc/palmas_gpadc.c | 52 +++++++++++++++++++---

2 files changed, 93 insertions(+), 5 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +18 -0
new file mode 100644
index 0000000..2149afe
@@ -0,0 +1,46 @@
+* Palmas general purpose ADC IP block devicetree bindings
+
+Channels list:
+	0 battery type
+	1 battery temp NTC (optional current source)
+	2 GP
+	3 temp (with ext. diode, optional current source)
+	4 GP
+	5 GP
+	6 VBAT_SENSE
+	7 VCC_SENSE
+	8 Backup Battery voltage
+	9 external charger (VCHG)
+	10 VBUS
+	11 DC-DC current probe (how does this work?)
+	12 internal die temp
+	13 internal die temp
+	14 USB ID pin voltage

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 28

Quoted 10 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +10 -0
+	15 test network
+
+Required properties:
+- compatible : Must be "ti,palmas-gpadc".
+
+Optional sub-nodes:
+ti,channel0-current-microamp: Channel 0 current in uA.
+	Values are rounded to derive 0uA, 5uA, 15uA, 20uA.
+ti,channel3-current-microamp: Channel 3 current in uA.
+	Valid are rounded to derive 0uA, 10uA, 400uA, 800uA.

It's only possible to configure channels 0 and 3 in this manner?

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 29

Quoted 1 line
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +1 -0
+ti,enable-extended-delay: Enable extended delay.

What is this? When would I select it? Why does it belong in the DT rather than being a runtime option?

Quoted 6 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +6 -0
+Example:
+
+pmic {
+	compatible = "ti,twl6035-pmic", "ti,palmas-pmic";
+	...
+	gpadc {

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 46

Quoted 10 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +10 -0
+		compatible = "ti,palmas-gpadc";
+		interrupts = <18 0
+			      16 0
+			      17 0>;
+		ti,channel0-current-microamp = <5>;
+		ti,channel3-current-microamp = <10>;
+		};
+	};
+	...
+};

I thought you needed #iio-cells for encoding the channel?

Mark.

#14 lore raw

Re: [PATCH v2 2/3] iio:adc:palmas: add DT support

Am 05.10.2015 um 13:17 schrieb Mark Rutland <[email protected]>:

Quoted 43 lines · H. Nikolaus Schaller
Quoted 42 lines

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]

---
.../devicetree/bindings/iio/adc/palmas-gpadc.txt | 46 +++++++++++++++++++
drivers/iio/adc/palmas_gpadc.c | 52 +++++++++++++++++++---

2 files changed, 93 insertions(+), 5 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +22 -0
new file mode 100644
index 0000000..2149afe
@@ -0,0 +1,46 @@
+* Palmas general purpose ADC IP block devicetree bindings
+
+Channels list:
+	0 battery type
+	1 battery temp NTC (optional current source)
+	2 GP
+	3 temp (with ext. diode, optional current source)
+	4 GP
+	5 GP
+	6 VBAT_SENSE
+	7 VCC_SENSE
+	8 Backup Battery voltage
+	9 external charger (VCHG)
+	10 VBUS
+	11 DC-DC current probe (how does this work?)
+	12 internal die temp
+	13 internal die temp
+	14 USB ID pin voltage
+	15 test network
+
+Required properties:
+- compatible : Must be "ti,palmas-gpadc".
Quoted 8 lines

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 28

Quoted 6 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +6 -0
+
+Optional sub-nodes:
+ti,channel0-current-microamp: Channel 0 current in uA.
+	Values are rounded to derive 0uA, 5uA, 15uA, 20uA.
+ti,channel3-current-microamp: Channel 3 current in uA.
+	Valid are rounded to derive 0uA, 10uA, 400uA, 800uA.

It's only possible to configure channels 0 and 3 in this manner?

Yes. The other channels have no built-in current source, i.e. these channels are special.

Quoted 5 lines

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 29

Quoted 2 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +1 -0
+ti,enable-extended-delay: Enable extended delay.

What is this? When would I select it? Why does it belong in the DT rather than being a runtime option?

The chip allows to extend the time window between channel selection and sampling by 400µs (according to data sheet). But for all channels and not each one. As far as I understand, this is - depending on hardware setup - to get more stable ADC readings.

Most channels have a fixed function (e.g. battery voltage, USB VBUS, temperature NTC) so it can't be arbitrarily chosen and depends on how the Palmas is embedded (it is a PMIC with a bank of ADCs and not a general purpose ADC chip).

So I think it is really a hardware dependent parameter and not something the user should be able to change.

Quoted 11 lines
Quoted 11 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +10 -0
+Example:
+
+pmic {
+	compatible = "ti,twl6035-pmic", "ti,palmas-pmic";
+	...
+	gpadc {
+		compatible = "ti,palmas-gpadc";
+		interrupts = <18 0
+			      16 0
+			      17 0>;
Quoted 8 lines

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 46

Quoted 6 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +6 -0
+		ti,channel0-current-microamp = <5>;
+		ti,channel3-current-microamp = <10>;
+		};
+	};
+	...
+};

I thought you needed #iio-cells for encoding the channel?

Yes, we forgot (we don't have a client for it in our setup yet - so did not test :)

Quoted 2 lines

Mark.

Thanks,
Nikolaus Schaller

#15 lore raw

PATCH v2 3/3 ARM: dts: omap5-uevm: enable iio gpadc for Palmas

In reply to a message not in the archive

---
arch/arm/boot/dts/omap5-uevm.dts | 9 +++++++++

1 file changed, 9 insertions(+)
arch/arm/boot/dts/omap5-uevm.dts +0 -0
index 3b16e8f..c78ee2f 100644
@@ -342,6 +342,15 @@
 
			ti,ldo6-vibrator;

+ gpadc {
+ compatible = "ti,palmas-gpadc";
+ interrupts = <18 0
+ 16 0
+ 17 0>;
+ ti,channel0-current-microamp = <5>;
+ ti,channel3-current-microamp = <10>;
+ };
+

			regulators {
				smps123_reg: smps123 {
					/* VDD_OPP_MPU */
Signature · 1 line
2.5.1
#16 lore raw

v2 cover Add Palmas iio gpadc

In reply to H. Nikolaus Schaller

V2 changes:
- worked in review comments by Peter Meerwald
- made channels 1 and 3 report (raw) temperature values
- moved PALMAS_GPADC_TRIMINVALID to driver core
- worked in review comments by Jonathan Cameron
- iio_map removed

Add iio driver for the TI Palmas (twl6035, 6037) including device tree bindings. It enables the gpadc for the OMAP5 uevm.

This patch series is based on original code taken from Android Tegra kernels:
(https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc)

Tegra code was developed by:
Pradeep Goudagunta <[email protected]>
Laxman Dewangan <[email protected]>

Edited and extended for mainline by:
H. Nikolaus Schaller <[email protected]>
Marek Belisko <[email protected]>

H. Nikolaus Schaller (2):
iio:adc: add iio driver for Palmas (twl6035/7) gpadc
ARM: dts: omap5-uevm: enable iio gpadc for Palmas

Marek Belisko (1):
iio:adc:palmas: add DT support

 .../devicetree/bindings/iio/adc/palmas-gpadc.txt   |  46 ++
 arch/arm/boot/dts/omap5-uevm.dts                   |   9 +
 drivers/iio/adc/Kconfig                            |   9 +
 drivers/iio/adc/Makefile                           |   1 +
 drivers/iio/adc/palmas_gpadc.c                     | 860 +++++++++++++++++++++
 include/linux/mfd/palmas.h                         |  75 +-
 6 files changed, 976 insertions(+), 24 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt
 create mode 100644 drivers/iio/adc/palmas_gpadc.c
Signature · 1 line
2.5.1
#17 lore raw

PATCH v2 1/3 iio:adc: add iio driver for Palmas (twl6035/7) gpadc

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

---
 drivers/iio/adc/Kconfig        |   9 +
 drivers/iio/adc/Makefile       |   1 +
 drivers/iio/adc/palmas_gpadc.c | 818 +++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/palmas.h     |  75 ++--
 4 files changed, 879 insertions(+), 24 deletions(-)
 create mode 100644 drivers/iio/adc/palmas_gpadc.c
drivers/iio/adc/Kconfig +9 -0
index eb0cd89..05a0368 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Say yes here to build support for Texas Instruments Palmas.
+
+	  Palmas series pmic chip (twl6035/6037) is used in smartphones and
+	  tablets and supports a 16 channel general purpose ADC.
+
 config QCOM_SPMI_IADC
 	tristate "Qualcomm SPMI PMIC current ADC"
 	depends on SPMI
drivers/iio/adc/Makefile +1 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o
 obj-$(CONFIG_MCP3422) += mcp3422.o
 obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o
 obj-$(CONFIG_NAU7802) += nau7802.o
+obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o
 obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o
 obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o
 obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o
drivers/iio/adc/palmas_gpadc.c new file +818 -0

818 lines not on this page · raw message

include/linux/mfd/palmas.h +51 -24

106 lines not on this page · raw message

Signature · 1 line
2.5.1
#18 lore raw

[PATCH] iio:adc: fix platform_no_drv_owner.cocci warnings

drivers/iio/adc/palmas_gpadc.c:798:3-8: No need to set .owner here. The core will do it.

drivers/iio/adc/palmas_gpadc.c:798:3-8: No need to set .owner here. The core will do it.

Remove .owner field if calls are used which set it automatically

Generated by: scripts/coccinelle/api/platform_no_drv_owner.cocci

CC H. Nikolaus Schaller <[email protected]>

---

 palmas_gpadc.c |    1 -
 1 file changed, 1 deletion(-)
drivers/iio/adc/palmas_gpadc.c +0 -1
@@ -795,7 +795,6 @@ static struct platform_driver palmas_gpa
 	.remove = palmas_gpadc_remove,
 	.driver = {
 		.name = MOD_NAME,
-		.owner = THIS_MODULE,
 		.pm = &palmas_pm_ops,
 	},
 };

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#19 lore raw

Re: [PATCH v2 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

In reply to H. Nikolaus Schaller
coccinelle warnings: (new ones prefixed by >>)

Hi Nikolaus,

[auto build test WARNING on v4.3-rc4 -- if it's inappropriate base, please ignore]

coccinelle warnings: (new ones prefixed by >>)

>> drivers/iio/adc/palmas_gpadc.c:798:3-8: No need to set .owner here. The core will do it.

Please review and possibly fold the followup patch.

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

Acked-by on v2 1/3
#20 lore raw

Re: [PATCH v2 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

In reply to H. Nikolaus Schaller
Quoted 10 lines · H. Nikolaus Schaller

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

maintainer Reviewed-by on v2 1/3
#21 lore raw

Re: [PATCH v2 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

In reply to H. Nikolaus Schaller
Quoted 9 lines · H. Nikolaus Schaller

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

I'm pretty much fine with this. However, there are some edits within the existing mfd support so I will want acks for that or for the driver to go through the MFD tree (note that as it touched that, even if only a header, Lee and Samuel should have been cc'd).

One thing that slightly confuses me is there seems to be hints of support in the mfd driver already... I can't find any sign of the child device however so I guess this is fine from that point of view.

Quoted 978 lines
---
 drivers/iio/adc/Kconfig        |   9 +
 drivers/iio/adc/Makefile       |   1 +
 drivers/iio/adc/palmas_gpadc.c | 818 +++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/palmas.h     |  75 ++--
 4 files changed, 879 insertions(+), 24 deletions(-)
 create mode 100644 drivers/iio/adc/palmas_gpadc.c
drivers/iio/adc/Kconfig +9 -0
index eb0cd89..05a0368 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Say yes here to build support for Texas Instruments Palmas.
+
+	  Palmas series pmic chip (twl6035/6037) is used in smartphones and
+	  tablets and supports a 16 channel general purpose ADC.
+
 config QCOM_SPMI_IADC
 	tristate "Qualcomm SPMI PMIC current ADC"
 	depends on SPMI
drivers/iio/adc/Makefile +1 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o
 obj-$(CONFIG_MCP3422) += mcp3422.o
 obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o
 obj-$(CONFIG_NAU7802) += nau7802.o
+obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o
 obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o
 obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o
 obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o
drivers/iio/adc/palmas_gpadc.c new file +818 -0
new file mode 100644
index 0000000..6805d81
@@ -0,0 +1,818 @@
+/*
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */
+
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
+#define MOD_NAME "palmas-gpadc"
+#define PALMAS_ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define PALMAS_TO_BE_CALCULATED 0
+#define PALMAS_GPADC_TRIMINVALID	-1
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */
+	int x1;	/* lower ideal code */
+	int x2;	/* higher ideal code */
+	int v1;	/* expected lower volt reading */
+	int v2;	/* expected higher volt reading */
+	u8 trim1_reg;	/* register number for lower trim */
+	u8 trim2_reg;	/* register number for upper trim */
+	int gain;	/* calculated from above (after reading trim regs) */
+	int offset;	/* calculated from above (after reading trim regs) */
+	int gain_error;	/* calculated from above (after reading trim regs) */
+	bool is_uncalibrated;	/* if channel has calibration data */
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_uncalibrated) \
+	[PALMAS_ADC_CH_##_chan] = { \
+		.x1 = _x1, \
+		.x2 = _x2, \
+		.v1 = _v1, \
+		.v2 = _v2, \
+		.gain = PALMAS_TO_BE_CALCULATED, \
+		.offset = PALMAS_TO_BE_CALCULATED, \
+		.gain_error = PALMAS_TO_BE_CALCULATED, \
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1, \
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,  \
+		.is_uncalibrated = _is_uncalibrated \
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+/**
+ * struct palmas_gpadc - the palmas_gpadc structure
+ * @ch0_current:	channel 0 current source setting
+ *    0: 0 uA
+ *    1: 5 uA
+ *    2: 15 uA
+ *    3: 20 uA
+ * @ch3_current:	channel 3 current source setting
+ *    0: 0 uA
+ *    1: 10 uA
+ *    2: 400 uA
+ *    3: 800 uA
+ * @extended_delay:	enable the gpadc extended delay mode
+ * @auto_conversion_period:	define the auto_conversion_period
+ *
+ * This is the palmas_gpadc structure to store run-time information
+ * and pointers for this driver instance.
+ */
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;
+	u8				ch0_current;
+	u8				ch3_current;
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+
+	return ret;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	complete(&adc->conv_completion);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_dbg(adc->dev, "Threshold interrupt %d occurs\n", irq);
+	palmas_disable_auto_conversion(adc);
+
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc,
+							bool mask)
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/* gain error calculation */
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/* gain calculation */
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/* offset Calculation */
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {
+		dev_err(adc->dev, "SELECT_SW_START write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				PALMAS_ADC_CONVERSION_TIMEOUT);
+	if (ret == 0) {
+		dev_err(adc->dev, "conversion not completed\n");
+		return -ETIMEDOUT;
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "SW_CONV0_LSB read failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = val & 0xFFF;
+
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (!adc->adc_info[adc_chan].is_uncalibrated)
+		val  = (val*1000 - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	if (val < 0) {
+		dev_err(adc->dev, "Mismatch with calibration\n");
+		return 0;
+	}
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)
+		return -EINVAL;
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start conversion failed\n");
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+
+	return ret;
+}
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type, chan_info)	\
+{							\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),	\
+	.type = _type,					\
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |	\
+			BIT(chan_info),			\
+	.indexed = 1,					\
+	.channel = PALMAS_ADC_CH_##chan,		\
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *gpadc_pdata = NULL;
+	struct iio_dev *indio_dev;
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+
+	gpadc_pdata = pdata->gpadc_pdata;
+
+	indio_dev = devm_iio_device_alloc(&pdev->dev, sizeof(*adc));
+	if (!indio_dev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	adc = iio_priv(indio_dev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, indio_dev);
+
+	adc->auto_conversion_period = gpadc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
+	if (adc->irq < 0) {
+		dev_err(adc->dev,
+			"get virq failed: %d\n", adc->irq);
+		ret = adc->irq;
+		goto out;
+	}
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %d\n", adc->irq, ret);
+		goto out;
+	}
+
+	if (gpadc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, gpadc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %d\n",
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (gpadc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, gpadc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %d\n",
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+
+	/* set the current source 0 (value 0/5/15/20 uA => 0..3) */
+	if (gpadc_pdata->ch0_current <= 1)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (gpadc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (gpadc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	/* set the current source 3 (value 0/10/400/800 uA => 0..3) */
+	if (gpadc_pdata->ch3_current <= 1)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (gpadc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (gpadc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = gpadc_pdata->extended_delay;
+
+	indio_dev->name = MOD_NAME;
+	indio_dev->dev.parent = &pdev->dev;
+	indio_dev->info = &palmas_gpadc_iio_info;
+	indio_dev->modes = INDIO_DIRECT_MODE;
+	indio_dev->channels = palmas_gpadc_iio_channel;
+	indio_dev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(indio_dev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
+		if (!(adc->adc_info[i].is_uncalibrated))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);
+
+	return 0;
+
+out_irq_auto1_free:
+	if (gpadc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (gpadc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out:
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *indio_dev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(indio_dev);
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_disable(&pdev->dev);
+	iio_device_unregister(indio_dev);
+	free_irq(adc->irq, adc);
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {
+		if (((1000 * (1 << i)) / 32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {
+		int polarity;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			polarity = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			polarity = PALMAS_GPADC_THRES_CONV0_MSB_THRES_CONV0_POL;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | polarity);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int polarity;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			polarity = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			polarity = PALMAS_GPADC_THRES_CONV1_MSB_THRES_CONV1_POL;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | polarity);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0)
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0)
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *indio_dev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(indio_dev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *indio_dev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(indio_dev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +51 -24
index bb270bd..2fcd70e 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
-	int ch3_current;
+	int ch3_current;	/* 0: off; 1: 10uA; 2: 400uA; 3: 800 uA */
 
 	/* Channel 0 current source can be used for battery detection.
 	 * If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
 	 */
-	int ch0_current;
+	int ch0_current;	/* 0: off; 1: 5uA; 2: 15uA; 3: 20 uA */
+	bool extended_delay;	/* use extended delay for conversion */
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -404,28 +415,7 @@ struct palmas_gpadc_calibration {
 	s32 offset_error;
 };
 
-struct palmas_gpadc {
-	struct device *dev;
-	struct palmas *palmas;
-
-	int ch3_current;
-	int ch0_current;
-
-	int gpadc_force;
-
-	int bat_removal;
-
-	struct mutex reading_lock;
-	struct completion irq_complete;
-
-	int eoc_sw_irq;
-
-	struct palmas_gpadc_calibration *palmas_cal_tbl;
-
-	int conv0_channel;
-	int conv1_channel;
-	int rt_channel;
-};
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;
@@ -519,6 +509,43 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palmas GPADC Channels */
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,
+	PALMAS_ADC_CH_MAX,
+};
+
+/* Palmas GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};
+
+/* Palmas GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
Acked-by
#22 lore raw

Re: [PATCH v2 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

Quoted 16 lines · Jonathan Cameron
Quoted 9 lines · H. Nikolaus Schaller

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

I'm pretty much fine with this. However, there are some edits within the existing mfd support so I will want acks for that or for the driver to go through the MFD tree (note that as it touched that, even if only a header, Lee and Samuel should have been cc'd).

Quoted 10 lines

One thing that slightly confuses me is there seems to be hints of support in the mfd driver already... I can't find any sign of the child device however so I guess this is fine from that point of view.

Quoted 5 lines
---
 drivers/iio/adc/Kconfig        |   9 +
 drivers/iio/adc/Makefile       |   1 +
 drivers/iio/adc/palmas_gpadc.c | 818 +++++++++++++++++++++++++++++++++++++++++
 include/linux/mfd/palmas.h     |  75 ++--
Quoted 974 lines
Quoted 973 lines
 4 files changed, 879 insertions(+), 24 deletions(-)
 create mode 100644 drivers/iio/adc/palmas_gpadc.c
drivers/iio/adc/Kconfig +9 -0
index eb0cd89..05a0368 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Say yes here to build support for Texas Instruments Palmas.
+
+	  Palmas series pmic chip (twl6035/6037) is used in smartphones and
+	  tablets and supports a 16 channel general purpose ADC.
+
 config QCOM_SPMI_IADC
 	tristate "Qualcomm SPMI PMIC current ADC"
 	depends on SPMI
drivers/iio/adc/Makefile +1 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o
 obj-$(CONFIG_MCP3422) += mcp3422.o
 obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o
 obj-$(CONFIG_NAU7802) += nau7802.o
+obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o
 obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o
 obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o
 obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o
drivers/iio/adc/palmas_gpadc.c new file +818 -0
new file mode 100644
index 0000000..6805d81
@@ -0,0 +1,818 @@
+/*
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */
+
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
+#define MOD_NAME "palmas-gpadc"
+#define PALMAS_ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define PALMAS_TO_BE_CALCULATED 0
+#define PALMAS_GPADC_TRIMINVALID	-1
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */
+	int x1;	/* lower ideal code */
+	int x2;	/* higher ideal code */
+	int v1;	/* expected lower volt reading */
+	int v2;	/* expected higher volt reading */
+	u8 trim1_reg;	/* register number for lower trim */
+	u8 trim2_reg;	/* register number for upper trim */
+	int gain;	/* calculated from above (after reading trim regs) */
+	int offset;	/* calculated from above (after reading trim regs) */
+	int gain_error;	/* calculated from above (after reading trim regs) */
+	bool is_uncalibrated;	/* if channel has calibration data */
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_uncalibrated) \
+	[PALMAS_ADC_CH_##_chan] = { \
+		.x1 = _x1, \
+		.x2 = _x2, \
+		.v1 = _v1, \
+		.v2 = _v2, \
+		.gain = PALMAS_TO_BE_CALCULATED, \
+		.offset = PALMAS_TO_BE_CALCULATED, \
+		.gain_error = PALMAS_TO_BE_CALCULATED, \
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1, \
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,  \
+		.is_uncalibrated = _is_uncalibrated \
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+/**
+ * struct palmas_gpadc - the palmas_gpadc structure
+ * @ch0_current:	channel 0 current source setting
+ *    0: 0 uA
+ *    1: 5 uA
+ *    2: 15 uA
+ *    3: 20 uA
+ * @ch3_current:	channel 3 current source setting
+ *    0: 0 uA
+ *    1: 10 uA
+ *    2: 400 uA
+ *    3: 800 uA
+ * @extended_delay:	enable the gpadc extended delay mode
+ * @auto_conversion_period:	define the auto_conversion_period
+ *
+ * This is the palmas_gpadc structure to store run-time information
+ * and pointers for this driver instance.
+ */
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;
+	u8				ch0_current;
+	u8				ch3_current;
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+
+	return ret;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	complete(&adc->conv_completion);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_dbg(adc->dev, "Threshold interrupt %d occurs\n", irq);
+	palmas_disable_auto_conversion(adc);
+
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc,
+							bool mask)
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/* gain error calculation */
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/* gain calculation */
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/* offset Calculation */
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {
+		dev_err(adc->dev, "SELECT_SW_START write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				PALMAS_ADC_CONVERSION_TIMEOUT);
+	if (ret == 0) {
+		dev_err(adc->dev, "conversion not completed\n");
+		return -ETIMEDOUT;
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "SW_CONV0_LSB read failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = val & 0xFFF;
+
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (!adc->adc_info[adc_chan].is_uncalibrated)
+		val  = (val*1000 - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	if (val < 0) {
+		dev_err(adc->dev, "Mismatch with calibration\n");
+		return 0;
+	}
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)
+		return -EINVAL;
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start conversion failed\n");
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+
+	return ret;
+}
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type, chan_info)	\
+{							\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),	\
+	.type = _type,					\
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |	\
+			BIT(chan_info),			\
+	.indexed = 1,					\
+	.channel = PALMAS_ADC_CH_##chan,		\
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_TEMP, IIO_CHAN_INFO_RAW),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE, IIO_CHAN_INFO_PROCESSED),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *gpadc_pdata = NULL;
+	struct iio_dev *indio_dev;
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+
+	gpadc_pdata = pdata->gpadc_pdata;
+
+	indio_dev = devm_iio_device_alloc(&pdev->dev, sizeof(*adc));
+	if (!indio_dev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	adc = iio_priv(indio_dev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, indio_dev);
+
+	adc->auto_conversion_period = gpadc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
+	if (adc->irq < 0) {
+		dev_err(adc->dev,
+			"get virq failed: %d\n", adc->irq);
+		ret = adc->irq;
+		goto out;
+	}
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %d\n", adc->irq, ret);
+		goto out;
+	}
+
+	if (gpadc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, gpadc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %d\n",
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (gpadc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, gpadc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %d\n",
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+
+	/* set the current source 0 (value 0/5/15/20 uA => 0..3) */
+	if (gpadc_pdata->ch0_current <= 1)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (gpadc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (gpadc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	/* set the current source 3 (value 0/10/400/800 uA => 0..3) */
+	if (gpadc_pdata->ch3_current <= 1)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (gpadc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (gpadc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = gpadc_pdata->extended_delay;
+
+	indio_dev->name = MOD_NAME;
+	indio_dev->dev.parent = &pdev->dev;
+	indio_dev->info = &palmas_gpadc_iio_info;
+	indio_dev->modes = INDIO_DIRECT_MODE;
+	indio_dev->channels = palmas_gpadc_iio_channel;
+	indio_dev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(indio_dev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
+		if (!(adc->adc_info[i].is_uncalibrated))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);
+
+	return 0;
+
+out_irq_auto1_free:
+	if (gpadc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (gpadc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out:
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *indio_dev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(indio_dev);
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_disable(&pdev->dev);
+	iio_device_unregister(indio_dev);
+	free_irq(adc->irq, adc);
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {
+		if (((1000 * (1 << i)) / 32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {
+		int polarity;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			polarity = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			polarity = PALMAS_GPADC_THRES_CONV0_MSB_THRES_CONV0_POL;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | polarity);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int polarity;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			polarity = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			polarity = PALMAS_GPADC_THRES_CONV1_MSB_THRES_CONV1_POL;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | polarity);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0)
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0)
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *indio_dev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(indio_dev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *indio_dev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(indio_dev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +51 -24
index bb270bd..2fcd70e 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
-	int ch3_current;
+	int ch3_current;	/* 0: off; 1: 10uA; 2: 400uA; 3: 800 uA */
 
 	/* Channel 0 current source can be used for battery detection.
 	 * If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
 	 */
-	int ch0_current;
+	int ch0_current;	/* 0: off; 1: 5uA; 2: 15uA; 3: 20 uA */
+	bool extended_delay;	/* use extended delay for conversion */
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -404,28 +415,7 @@ struct palmas_gpadc_calibration {
 	s32 offset_error;
 };
 
-struct palmas_gpadc {
-	struct device *dev;
-	struct palmas *palmas;
-
-	int ch3_current;
-	int ch0_current;
-
-	int gpadc_force;
-
-	int bat_removal;
-
-	struct mutex reading_lock;
-	struct completion irq_complete;
-
-	int eoc_sw_irq;
-
-	struct palmas_gpadc_calibration *palmas_cal_tbl;
-
-	int conv0_channel;
-	int conv1_channel;
-	int rt_channel;
-};
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;
@@ -519,6 +509,43 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palmas GPADC Channels */
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,
+	PALMAS_ADC_CH_MAX,
+};
+
+/* Palmas GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};
+
+/* Palmas GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
Signature · 8 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#23 lore raw

PATCH v2 2/3 iio:adc:palmas: add DT support

In reply to H. Nikolaus Schaller

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]
---
 .../devicetree/bindings/iio/adc/palmas-gpadc.txt   | 46 +++++++++++++++++++
 drivers/iio/adc/palmas_gpadc.c                     | 52 +++++++++++++++++++---
 2 files changed, 93 insertions(+), 5 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +46 -0
new file mode 100644
index 0000000..2149afe
@@ -0,0 +1,46 @@
+* Palmas general purpose ADC IP block devicetree bindings
+
+Channels list:
+	0 battery type
+	1 battery temp NTC (optional current source)
+	2 GP
+	3 temp (with ext. diode, optional current source)
+	4 GP
+	5 GP
+	6 VBAT_SENSE
+	7 VCC_SENSE
+	8 Backup Battery voltage
+	9 external charger (VCHG)
+	10 VBUS
+	11 DC-DC current probe (how does this work?)
+	12 internal die temp
+	13 internal die temp
+	14 USB ID pin voltage
+	15 test network
+
+Required properties:
+- compatible : Must be "ti,palmas-gpadc".
+
+Optional sub-nodes:
+ti,channel0-current-microamp: Channel 0 current in uA.
+	Values are rounded to derive 0uA, 5uA, 15uA, 20uA.
+ti,channel3-current-microamp: Channel 3 current in uA.
+	Valid are rounded to derive 0uA, 10uA, 400uA, 800uA.
+ti,enable-extended-delay: Enable extended delay.
+
+Example:
+
+pmic {
+	compatible = "ti,twl6035-pmic", "ti,palmas-pmic";
+	...
+	gpadc {
+		compatible = "ti,palmas-gpadc";
+		interrupts = <18 0
+			      16 0
+			      17 0>;
+		ti,channel0-current-microamp = <5>;
+		ti,channel3-current-microamp = <10>;
+		};
+	};
+	...
+};
drivers/iio/adc/palmas_gpadc.c +47 -5

83 lines not on this page · raw message

Signature · 1 line
2.5.1
Acked-by on v2 2/3
#24 lore raw

Re: [PATCH v2 2/3] iio:adc:palmas: add DT support

Quoted 9 lines · H. Nikolaus Schaller

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]

---

#25 lore raw

PATCH v2 3/3 ARM: dts: omap5-uevm: enable iio gpadc for Palmas

In reply to H. Nikolaus Schaller
---
 arch/arm/boot/dts/omap5-uevm.dts | 9 +++++++++
 1 file changed, 9 insertions(+)
arch/arm/boot/dts/omap5-uevm.dts +9 -0
index 3b16e8f..c78ee2f 100644
@@ -342,6 +342,15 @@
 
 			ti,ldo6-vibrator;
 
+			gpadc {
+				compatible = "ti,palmas-gpadc";
+				interrupts = <18 0
+					      16 0
+					      17 0>;
+				ti,channel0-current-microamp = <5>;
+				ti,channel3-current-microamp = <10>;
+			};
+
 			regulators {
 				smps123_reg: smps123 {
 					/* VDD_OPP_MPU */
Signature · 1 line
2.5.1

A message that is not in the archive (sent off-list or not yet received): [email protected]

#27 lore raw

PATCH 1/3 iio:adc: add iio driver for Palmas (twl6035/7) gpadc

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 797 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  59 ++-
4 files changed, 862 insertions(+), 4 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +0 -0
index eb0cd89..f6df9db 100644
@@ -242,6 +242,15 @@ config NAU7802

To compile this driver as a module, choose M here: the module will be called nau7802.

+config PALMAS_GPADC
+ tristate "TI Palmas General Purpose ADC"
+ depends on MFD_PALMAS
+ help

+	  Palmas series pmic chip by texas Instruments (twl6035/6037)
+	  is used in smartphones and tablets and supports a 16 channel
+	  general purpose ADC. Add iio driver to read different channel
+	  of the GPADCs.

+
config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +797 -0

797 lines not on this page · raw message

include/linux/mfd/palmas.h +0 -0
index bb270bd..60acfa2 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
			    struct regulator_config config);
};

+struct palmas_adc_wakeup_property {
+ int adc_channel_number;
+ int adc_high_threshold;
+ int adc_low_threshold;
+};
+

struct palmas_gpadc_platform_data {
	/* Channel 3 current source is only enabled during conversion */
	int ch3_current;
-
	/* Channel 0 current source can be used for battery detection.

- * If used for battery detection this will cause a permanent current
+ * If used for battery detection this will cause a permanent current
* consumption depending on current level set here.
- */
+ */

	int ch0_current;

+ bool extended_delay;

	/* default BAT_REMOVAL_DAT setting on device probe */
	int bat_removal;

	/* Sets the START_POLARITY bit in the RT_CTRL register */
	int start_polarity;

+
+ struct iio_map *iio_maps;
+ int auto_conversion_period_ms;
+ struct palmas_adc_wakeup_property *adc_wakeup1_data;
+ struct palmas_adc_wakeup_property *adc_wakeup2_data;
};

struct palmas_reg_init {

@@ -403,7 +414,7 @@ struct palmas_gpadc_calibration {
	s32 gain_error;
	s32 offset_error;
};
-
+/*
struct palmas_gpadc {
	struct device *dev;
	struct palmas *palmas;
@@ -426,6 +437,9 @@ struct palmas_gpadc {
	int conv1_channel;
	int rt_channel;
};

+*/
+
+#define PALMAS_DATASHEET_NAME(_name) "palmas-gpadc-chan-"#_name

struct palmas_gpadc_result {
	s32 raw_code;

@@ -519,6 +533,42 @@ enum palmas_irqs {
PALMAS_NUM_IRQ,
};

+/* Palma GPADC Channels */
+enum {
+ PALMAS_ADC_CH_IN0,
+ PALMAS_ADC_CH_IN1,
+ PALMAS_ADC_CH_IN2,
+ PALMAS_ADC_CH_IN3,
+ PALMAS_ADC_CH_IN4,
+ PALMAS_ADC_CH_IN5,
+ PALMAS_ADC_CH_IN6,
+ PALMAS_ADC_CH_IN7,
+ PALMAS_ADC_CH_IN8,
+ PALMAS_ADC_CH_IN9,
+ PALMAS_ADC_CH_IN10,
+ PALMAS_ADC_CH_IN11,
+ PALMAS_ADC_CH_IN12,
+ PALMAS_ADC_CH_IN13,
+ PALMAS_ADC_CH_IN14,
+ PALMAS_ADC_CH_IN15,
+ PALMAS_ADC_CH_MAX,
+};
+
+/* Palma GPADC Channel0 Current Source */
+enum {
+ PALMAS_ADC_CH0_CURRENT_SRC_0,
+ PALMAS_ADC_CH0_CURRENT_SRC_5,
+ PALMAS_ADC_CH0_CURRENT_SRC_15,
+ PALMAS_ADC_CH0_CURRENT_SRC_20,
+};
+/* Palma GPADC Channel3 Current Source */
+enum {
+ PALMAS_ADC_CH3_CURRENT_SRC_0,
+ PALMAS_ADC_CH3_CURRENT_SRC_10,
+ PALMAS_ADC_CH3_CURRENT_SRC_400,
+ PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+

struct palmas_pmic {
	struct palmas *palmas;
	struct device *dev;

@@ -2999,6 +3049,7 @@ enum usb_irq_events {
#define PALMAS_GPADC_TRIM14 0x0D
#define PALMAS_GPADC_TRIM15 0x0E
#define PALMAS_GPADC_TRIM16 0x0F
+#define PALMAS_GPADC_TRIMINVALID -1

/* TPS659038 regen2_ctrl offset iss different from palmas */
#define TPS659038_REGEN2_CTRL 0x12

Signature · 1 line
2.5.1
#28 lore raw

Re: [PATCH 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

Quoted 6 lines

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

several minor comments below
probably the mfd part should be split out in a separate patch

Quoted 17 lines

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 797 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  59 ++-
4 files changed, 862 insertions(+), 4 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +0 -0
index eb0cd89..f6df9db 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the

drivers/iio/adc/Kconfig config NAU7802 · line 252

Quoted 10 lines
drivers/iio/adc/Kconfig +8 -0
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Palmas series pmic chip by texas Instruments (twl6035/6037)
+	  is used in smartphones and tablets and supports a 16 channel
+	  general purpose ADC. Add iio driver to read different channel
+	  of the GPADCs.

PMIC
_T_exas

drop meaningless "Add iio driver to read different channel of the GPADCs."

Quoted 23 lines
drivers/iio/adc/Kconfig +1 -0
+

config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +1 -0
new file mode 100644
index 0000000..17abb28
@@ -0,0 +1,797 @@
+/*

drivers/iio/adc/palmas_gpadc.c line 11

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */

newline

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>

drivers/iio/adc/palmas_gpadc.c line 29

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
+#define MOD_NAME "palmas-gpadc"
+#define ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define TO_BE_CALCULATED 0

add prefix PALMAS_

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */
+	int x1;
+	int x2;
+	int v1;

drivers/iio/adc/palmas_gpadc.c line 45

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	int v2;
+	u8 trim1_reg;
+	u8 trim2_reg;
+	int gain;
+	int offset;
+	int gain_error;
+	bool is_correct_code;
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_correct_code)\

whitespace and indentation

Quoted 111 lines
drivers/iio/adc/palmas_gpadc.c new file +111 -0
+[PALMAS_ADC_CH_##_chan] = {						\
+		.x1 = _x1,						\
+		.x2 = _x2,						\
+		.v1 = _v1,						\
+		.v2 = _v2,						\
+		.gain = TO_BE_CALCULATED,				\
+		.offset = TO_BE_CALCULATED,				\
+		.gain_error = TO_BE_CALCULATED,				\
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1,			\
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,			\
+		.is_correct_code = _is_correct_code			\
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;
+	u8				ch0_current;
+	u8				ch3_current;
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;

drivers/iio/adc/palmas_gpadc.c line 166

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	complete(&adc->conv_completion);
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_info(adc->dev, "Threshold interrupt %d occurs\n", irq);

is the message needed except for debugging?

Quoted 117 lines
drivers/iio/adc/palmas_gpadc.c new file +117 -0
+	palmas_disable_auto_conversion(adc);
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc, int mask)
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;

drivers/iio/adc/palmas_gpadc.c line 293

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/*Gain error Calculation*/

ugly, add whitespace /* bla */

Quoted 16 lines
drivers/iio/adc/palmas_gpadc.c new file +16 -0
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/*Gain Calculation*/
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/*offset Calculation*/
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{

drivers/iio/adc/palmas_gpadc.c line 319

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADC_SW_START write failed: %d\n", ret);

generally, the ADC_ prefix is not logged

Quoted 5 lines
drivers/iio/adc/palmas_gpadc.c new file +5 -0
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				ADC_CONVERSION_TIMEOUT);

drivers/iio/adc/palmas_gpadc.c line 334

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	if (ret == 0) {
+		dev_err(adc->dev, "ADC conversion not completed\n");
+		ret = -ETIMEDOUT;
+		return ret;
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADCDATA read failed: %d\n", ret);

SW_CONV0_LSB read failed

drivers/iio/adc/palmas_gpadc.c line 338

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+		return ret;
+	}
+
+	ret = (val & 0xFFF);

parenthesis not needed, there are plenty of redundant parenthesis below...

Quoted 32 lines
drivers/iio/adc/palmas_gpadc.c new file +32 -0
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (((val*1000) - adc->adc_info[adc_chan].offset) < 0) {
+		dev_err(adc->dev, "No Input Connected\n");
+		return 0;
+	}
+
+	if (!(adc->adc_info[adc_chan].is_correct_code))
+		val  = ((val*1000) - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)
+		return -EINVAL;
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {

drivers/iio/adc/palmas_gpadc.c line 380

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start coversion failed\n");

typo: conversion

Quoted 21 lines
drivers/iio/adc/palmas_gpadc.c new file +21 -0
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+}

drivers/iio/adc/palmas_gpadc.c line 411

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type)			\
+{									\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),			\
+	.type = _type,							\

_type is always IIO_VOLTAGE, no need for parameter

Quoted 31 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |			\
+			BIT(IIO_CHAN_INFO_PROCESSED),			\
+	.indexed = 1,							\
+	.channel = PALMAS_ADC_CH_##chan,				\
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *adc_pdata;
+	struct iio_dev *iodev;

drivers/iio/adc/palmas_gpadc.c line 452

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+	adc_pdata = pdata->gpadc_pdata;
+
+	iodev = iio_device_alloc(sizeof(*adc));

use devm_iio_device_alloc()
indio_dev is the preferred name instead of iodev

Quoted 12 lines
drivers/iio/adc/palmas_gpadc.c new file +12 -0
+	if (!iodev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	if (adc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "iio_map_array_register failed\n");
+			goto out;
+		}
+	}

drivers/iio/adc/palmas_gpadc.c line 474

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	adc = iio_priv(iodev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, iodev);
+
+	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);

palmas_irq_get_virq() can fail, check needed

drivers/iio/adc/palmas_gpadc.c line 481

Quoted 7 lines
drivers/iio/adc/palmas_gpadc.c new file +7 -0
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %dn", adc->irq, ret);

format string shoule end in %d\n I guess

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+		goto out_unregister_map;
+	}
+
+	if (adc_pdata->adc_wakeup1_data) {

drivers/iio/adc/palmas_gpadc.c line 495

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %dn",

\n

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (adc_pdata->adc_wakeup2_data) {

drivers/iio/adc/palmas_gpadc.c line 511

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %dn",

\n

Quoted 346 lines
drivers/iio/adc/palmas_gpadc.c new file +286 -0
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+
+	if (adc_pdata->ch0_current == 0)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (adc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (adc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	if (adc_pdata->ch3_current == 0)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (adc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (adc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = adc_pdata->extended_delay;
+
+	iodev->name = MOD_NAME;
+	iodev->dev.parent = &pdev->dev;
+	iodev->info = &palmas_gpadc_iio_info;
+	iodev->modes = INDIO_DIRECT_MODE;
+	iodev->channels = palmas_gpadc_iio_channel;
+	iodev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(iodev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
+		if (!(adc->adc_info[i].is_correct_code))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);
+
+	return 0;
+
+out_irq_auto1_free:
+	if (adc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (adc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out_unregister_map:
+	if (adc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+out:
+	iio_device_free(iodev);
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	struct palmas_platform_data *pdata = dev_get_platdata(pdev->dev.parent);
+
+	if (pdata->gpadc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+	iio_device_unregister(iodev);
+	free_irq(adc->irq, adc);
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+	iio_device_free(iodev);
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {
+		if (((1000 * (1 << i))/32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {
+		int is_high;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int is_high;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0) {
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +18 -4
index bb270bd..60acfa2 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
 	int ch3_current;
-
 	/* Channel 0 current source can be used for battery detection.
-	 * If used for battery detection this will cause a permanent current
+	* If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
-	 */
+		 */
 	int ch0_current;
+	bool extended_delay;
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	struct iio_map *iio_maps;
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -403,7 +414,7 @@ struct palmas_gpadc_calibration {
 	s32 gain_error;
 	s32 offset_error;
 };
-
+/*
 struct palmas_gpadc {
 	struct device *dev;
 	struct palmas *palmas;
@@ -426,6 +437,9 @@ struct palmas_gpadc {
 	int conv1_channel;
 	int rt_channel;
 };
+*/
+
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;

include/linux/mfd/palmas.h enum palmas_irqs { · line 536

Quoted 4 lines
include/linux/mfd/palmas.h +1 -0
@@ -519,6 +533,42 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palma GPADC Channels */

Palmas, here and below

Quoted 47 lines
include/linux/mfd/palmas.h +36 -0
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,
+	PALMAS_ADC_CH_MAX,
+};
+
+/* Palma GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};
+/* Palma GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
@@ -2999,6 +3049,7 @@ enum usb_irq_events {
 #define PALMAS_GPADC_TRIM14					0x0D
 #define PALMAS_GPADC_TRIM15					0x0E
 #define PALMAS_GPADC_TRIM16					0x0F
+#define PALMAS_GPADC_TRIMINVALID				-1
 
 /* TPS659038 regen2_ctrl offset iss different from palmas */
 #define TPS659038_REGEN2_CTRL					0x12
Signature · 6 lines
Peter Meerwald
+43-664-2444418 (mobile)
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#29 lore raw

Re: [PATCH 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

Am 23.09.2015 um 15:26 schrieb Peter Meerwald <[email protected]>:

Quoted 8 lines
Quoted 6 lines

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

several minor comments below

thanks!
I have worked in for a V2 but will wait a little for further comments until I post it.

Quoted 1 line

probably the mfd part should be split out in a separate patch

well, I think those changes should stay bundled in a single patch since they depend on each other.

Quoted 488 lines
Quoted 18 lines

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 797 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  59 ++-
4 files changed, 862 insertions(+), 4 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +0 -0
index eb0cd89..f6df9db 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the

drivers/iio/adc/Kconfig config NAU7802 · line 252

Quoted 10 lines
drivers/iio/adc/Kconfig +8 -0
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Palmas series pmic chip by texas Instruments (twl6035/6037)
+	  is used in smartphones and tablets and supports a 16 channel
+	  general purpose ADC. Add iio driver to read different channel
+	  of the GPADCs.

PMIC
_T_exas

drop meaningless "Add iio driver to read different channel of the GPADCs."

Quoted 23 lines
drivers/iio/adc/Kconfig +1 -0
+

config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +1 -0
new file mode 100644
index 0000000..17abb28
@@ -0,0 +1,797 @@
+/*

drivers/iio/adc/palmas_gpadc.c line 11

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */

newline

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>

drivers/iio/adc/palmas_gpadc.c line 29

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
+#define MOD_NAME "palmas-gpadc"
+#define ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define TO_BE_CALCULATED 0

add prefix PALMAS_

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */
+	int x1;
+	int x2;
+	int v1;

drivers/iio/adc/palmas_gpadc.c line 45

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	int v2;
+	u8 trim1_reg;
+	u8 trim2_reg;
+	int gain;
+	int offset;
+	int gain_error;
+	bool is_correct_code;
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_correct_code)\

whitespace and indentation

Quoted 111 lines
drivers/iio/adc/palmas_gpadc.c new file +111 -0
+[PALMAS_ADC_CH_##_chan] = {						\
+		.x1 = _x1,						\
+		.x2 = _x2,						\
+		.v1 = _v1,						\
+		.v2 = _v2,						\
+		.gain = TO_BE_CALCULATED,				\
+		.offset = TO_BE_CALCULATED,				\
+		.gain_error = TO_BE_CALCULATED,				\
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1,			\
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,			\
+		.is_correct_code = _is_correct_code			\
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;
+	u8				ch0_current;
+	u8				ch3_current;
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;

drivers/iio/adc/palmas_gpadc.c line 166

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	complete(&adc->conv_completion);
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_info(adc->dev, "Threshold interrupt %d occurs\n", irq);

is the message needed except for debugging?

Quoted 117 lines
drivers/iio/adc/palmas_gpadc.c new file +117 -0
+	palmas_disable_auto_conversion(adc);
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc, int mask)
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;

drivers/iio/adc/palmas_gpadc.c line 293

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/*Gain error Calculation*/

ugly, add whitespace /* bla */

Quoted 16 lines
drivers/iio/adc/palmas_gpadc.c new file +16 -0
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/*Gain Calculation*/
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/*offset Calculation*/
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{

drivers/iio/adc/palmas_gpadc.c line 319

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADC_SW_START write failed: %d\n", ret);

generally, the ADC_ prefix is not logged

Quoted 5 lines
drivers/iio/adc/palmas_gpadc.c new file +5 -0
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				ADC_CONVERSION_TIMEOUT);

drivers/iio/adc/palmas_gpadc.c line 334

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	if (ret == 0) {
+		dev_err(adc->dev, "ADC conversion not completed\n");
+		ret = -ETIMEDOUT;
+		return ret;
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADCDATA read failed: %d\n", ret);

SW_CONV0_LSB read failed

drivers/iio/adc/palmas_gpadc.c line 338

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+		return ret;
+	}
+
+	ret = (val & 0xFFF);

parenthesis not needed, there are plenty of redundant parenthesis below...

Quoted 32 lines
drivers/iio/adc/palmas_gpadc.c new file +32 -0
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (((val*1000) - adc->adc_info[adc_chan].offset) < 0) {
+		dev_err(adc->dev, "No Input Connected\n");
+		return 0;
+	}
+
+	if (!(adc->adc_info[adc_chan].is_correct_code))
+		val  = ((val*1000) - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)
+		return -EINVAL;
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {

drivers/iio/adc/palmas_gpadc.c line 380

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start coversion failed\n");

typo: conversion

Quoted 25 lines
drivers/iio/adc/palmas_gpadc.c new file +25 -0
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+}
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 411

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type)			\
+{									\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),			\
+	.type = _type,							\

_type is always IIO_VOLTAGE, no need for parameter

here we forgot a minor change to define IIO_TEMP for some channels.

Quoted 514 lines
Quoted 32 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |			\
+			BIT(IIO_CHAN_INFO_PROCESSED),			\
+	.indexed = 1,							\
+	.channel = PALMAS_ADC_CH_##chan,				\
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *adc_pdata;
+	struct iio_dev *iodev;

drivers/iio/adc/palmas_gpadc.c line 452

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+	adc_pdata = pdata->gpadc_pdata;
+
+	iodev = iio_device_alloc(sizeof(*adc));

use devm_iio_device_alloc()
indio_dev is the preferred name instead of iodev

Quoted 12 lines
drivers/iio/adc/palmas_gpadc.c new file +12 -0
+	if (!iodev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	if (adc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "iio_map_array_register failed\n");
+			goto out;
+		}
+	}

drivers/iio/adc/palmas_gpadc.c line 474

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	adc = iio_priv(iodev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, iodev);
+
+	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);

palmas_irq_get_virq() can fail, check needed

drivers/iio/adc/palmas_gpadc.c line 481

Quoted 7 lines
drivers/iio/adc/palmas_gpadc.c new file +7 -0
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %dn", adc->irq, ret);

format string shoule end in %d\n I guess

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+		goto out_unregister_map;
+	}
+
+	if (adc_pdata->adc_wakeup1_data) {

drivers/iio/adc/palmas_gpadc.c line 495

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %dn",

\n

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (adc_pdata->adc_wakeup2_data) {

drivers/iio/adc/palmas_gpadc.c line 511

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %dn",

\n

Quoted 346 lines
drivers/iio/adc/palmas_gpadc.c new file +286 -0
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+
+	if (adc_pdata->ch0_current == 0)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (adc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (adc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	if (adc_pdata->ch3_current == 0)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (adc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (adc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = adc_pdata->extended_delay;
+
+	iodev->name = MOD_NAME;
+	iodev->dev.parent = &pdev->dev;
+	iodev->info = &palmas_gpadc_iio_info;
+	iodev->modes = INDIO_DIRECT_MODE;
+	iodev->channels = palmas_gpadc_iio_channel;
+	iodev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(iodev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
+		if (!(adc->adc_info[i].is_correct_code))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);
+
+	return 0;
+
+out_irq_auto1_free:
+	if (adc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (adc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out_unregister_map:
+	if (adc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+out:
+	iio_device_free(iodev);
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	struct palmas_platform_data *pdata = dev_get_platdata(pdev->dev.parent);
+
+	if (pdata->gpadc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+	iio_device_unregister(iodev);
+	free_irq(adc->irq, adc);
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+	iio_device_free(iodev);
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {
+		if (((1000 * (1 << i))/32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {
+		int is_high;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int is_high;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0) {
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +18 -4
index bb270bd..60acfa2 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
 	int ch3_current;
-
 	/* Channel 0 current source can be used for battery detection.
-	 * If used for battery detection this will cause a permanent current
+	* If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
-	 */
+		 */
 	int ch0_current;
+	bool extended_delay;
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	struct iio_map *iio_maps;
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -403,7 +414,7 @@ struct palmas_gpadc_calibration {
 	s32 gain_error;
 	s32 offset_error;
 };
-
+/*
 struct palmas_gpadc {
 	struct device *dev;
 	struct palmas *palmas;
@@ -426,6 +437,9 @@ struct palmas_gpadc {
 	int conv1_channel;
 	int rt_channel;
 };
+*/
+
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;

include/linux/mfd/palmas.h enum palmas_irqs { · line 536

Quoted 4 lines
include/linux/mfd/palmas.h +1 -0
@@ -519,6 +533,42 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palma GPADC Channels */

Palmas, here and below

Quoted 44 lines
include/linux/mfd/palmas.h +36 -0
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,
+	PALMAS_ADC_CH_MAX,
+};
+
+/* Palma GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};
+/* Palma GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
@@ -2999,6 +3049,7 @@ enum usb_irq_events {
 #define PALMAS_GPADC_TRIM14					0x0D
 #define PALMAS_GPADC_TRIM15					0x0E
 #define PALMAS_GPADC_TRIM16					0x0F
+#define PALMAS_GPADC_TRIMINVALID				-1
 
Quoted 8 lines
Quoted 3 lines
include/linux/mfd/palmas.h +0 -0
 /* TPS659038 regen2_ctrl offset iss different from palmas */
 #define TPS659038_REGEN2_CTRL					0x12
Signature · 2 lines
Peter Meerwald
+43-664-2444418 (mobile)

BR and tnx,
Nikolaus Schaller

maintainer
#30 lore raw

Re: [PATCH 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

In reply to H. Nikolaus Schaller
Quoted 9 lines · H. Nikolaus Schaller

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

Various bits inline.

Jonathan

Quoted 68 lines

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 797 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  59 ++-
4 files changed, 862 insertions(+), 4 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +9 -0
index eb0cd89..f6df9db 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Palmas series pmic chip by texas Instruments (twl6035/6037)
+	  is used in smartphones and tablets and supports a 16 channel
+	  general purpose ADC. Add iio driver to read different channel
+	  of the GPADCs.
+

config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +22 -0
new file mode 100644
index 0000000..17abb28
@@ -0,0 +1,797 @@
+/*
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>

drivers/iio/adc/palmas_gpadc.c line 32

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
+#define MOD_NAME "palmas-gpadc"
+#define ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define TO_BE_CALCULATED 0
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */

Full docs on this would be appreciated.

Quoted 38 lines
drivers/iio/adc/palmas_gpadc.c new file +38 -0
+	int x1;
+	int x2;
+	int v1;
+	int v2;
+	u8 trim1_reg;
+	u8 trim2_reg;
+	int gain;
+	int offset;
+	int gain_error;
+	bool is_correct_code;
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_correct_code)\
+[PALMAS_ADC_CH_##_chan] = {						\
+		.x1 = _x1,						\
+		.x2 = _x2,						\
+		.v1 = _v1,						\
+		.v2 = _v2,						\
+		.gain = TO_BE_CALCULATED,				\
+		.offset = TO_BE_CALCULATED,				\
+		.gain_error = TO_BE_CALCULATED,				\
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1,			\
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,			\
+		.is_correct_code = _is_correct_code			\
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),

drivers/iio/adc/palmas_gpadc.c line 80

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;

As there are some non obvious elements in here (such as the next two) kernel-doc for the whole stucture would be good.

Quoted 60 lines
drivers/iio/adc/palmas_gpadc.c new file +60 -0
+	u8				ch0_current;
+	u8				ch3_current;
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}

drivers/iio/adc/palmas_gpadc.c line 150

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}

return ret and drop the return above. Coccicheck will moan at you about this.

drivers/iio/adc/palmas_gpadc.c line 158

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return 0;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	complete(&adc->conv_completion);

blank line.

drivers/iio/adc/palmas_gpadc.c line 166

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_info(adc->dev, "Threshold interrupt %d occurs\n", irq);

Could indicate this to userspace... other than through the logs.

drivers/iio/adc/palmas_gpadc.c line 167

Quoted 1 line
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+	palmas_disable_auto_conversion(adc);

blank line generally before all function returns..

drivers/iio/adc/palmas_gpadc.c line 171

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc, int mask)

mask is boolean. Make it explicitly so for readability.

Quoted 147 lines
drivers/iio/adc/palmas_gpadc.c new file +147 -0
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/*Gain error Calculation*/
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/*Gain Calculation*/
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/*offset Calculation*/
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {

drivers/iio/adc/palmas_gpadc.c line 328

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		dev_err(adc->dev, "ADC_SW_START write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				ADC_CONVERSION_TIMEOUT);
+	if (ret == 0) {
+		dev_err(adc->dev, "ADC conversion not completed\n");
+		ret = -ETIMEDOUT;
+		return ret;

return -ETIMEDOUT; Might be worth you setting up to do sparse, smatch and
coccicheck runs on the code as they'll pick up on a lot of issues like this.

drivers/iio/adc/palmas_gpadc.c line 338

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADCDATA read failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = (val & 0xFFF);

nitpick : blank line here.

drivers/iio/adc/palmas_gpadc.c line 348

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (((val*1000) - adc->adc_info[adc_chan].offset) < 0) {
+		dev_err(adc->dev, "No Input Connected\n");
+		return 0;
+	}

Interesting, but should this not be caught for raw channel reads as well?

drivers/iio/adc/palmas_gpadc.c line 349

Quoted 1 line
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+

Umm. what is is_correct_code? Should be documented somewhere

Quoted 17 lines
drivers/iio/adc/palmas_gpadc.c new file +16 -0
+	if (!(adc->adc_info[adc_chan].is_correct_code))
+		val  = ((val*1000) - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)

= given I think your channels are 0 indexed?

drivers/iio/adc/palmas_gpadc.c line 372

Quoted 7 lines
drivers/iio/adc/palmas_gpadc.c new file +7 -0
+		return -EINVAL;
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:

I'd be tempted to split the two code paths here as that will be slightly easier to read (if longer).

I'm also highly suspicious of any driver that does processed output and has a return type of IIO_VAL_INT. Are you processed values in the units documented in Documentation/ABI/testing/sysfs-bus-iio? They could be, but I would like a comment saying that.

Quoted 29 lines
drivers/iio/adc/palmas_gpadc.c new file +29 -0
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start coversion failed\n");
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+}

drivers/iio/adc/palmas_gpadc.c line 411

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type)			\
+{									\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),			\
+	.type = _type,							\

Right now they are all voltage channels, why have that specifiable in this macro?

drivers/iio/adc/palmas_gpadc.c line 413

Quoted 2 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |			\
+			BIT(IIO_CHAN_INFO_PROCESSED),			\

Hmm. This is very very rarely justified. Either the driver has nice linear relationship between the raw value and the processed one, in which case you should be providing *_RAW, *_OFFSET and *_SCALE and leaving the maths to userspace (or the in kernel wrappers), or they are non linear in which case only the processed value is of interest and the raw one not directly so. (the only exception to this is light sensors where the processed channel is often a computed channel from several input _raw readings).

drivers/iio/adc/palmas_gpadc.c line 415

Quoted 2 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.indexed = 1,							\
+	.channel = PALMAS_ADC_CH_##chan,				\

Given this maps back to the value of chan, just put that in as a number and loose the enum. Channel is used in the naming so it doesn't make sense to obscure that behind an enum anyway.

Quoted 91 lines
drivers/iio/adc/palmas_gpadc.c new file +91 -0
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *adc_pdata;
+	struct iio_dev *iodev;
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+	adc_pdata = pdata->gpadc_pdata;
+
+	iodev = iio_device_alloc(sizeof(*adc));
+	if (!iodev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	if (adc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "iio_map_array_register failed\n");
+			goto out;
+		}
+	}
+
+	adc = iio_priv(iodev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, iodev);
+
+	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %dn", adc->irq, ret);
+		goto out_unregister_map;
+	}
+
+	if (adc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %dn",
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (adc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,

drivers/iio/adc/palmas_gpadc.c line 516

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %dn",
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+

I've requested kernel-doc above for ch0_current, but if that doesn't make it clear what matic is going on here then some comments here would be good.

Quoted 31 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	if (adc_pdata->ch0_current == 0)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (adc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (adc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	if (adc_pdata->ch3_current == 0)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (adc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (adc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = adc_pdata->extended_delay;
+
+	iodev->name = MOD_NAME;
+	iodev->dev.parent = &pdev->dev;
+	iodev->info = &palmas_gpadc_iio_info;
+	iodev->modes = INDIO_DIRECT_MODE;
+	iodev->channels = palmas_gpadc_iio_channel;
+	iodev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(iodev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;

drivers/iio/adc/palmas_gpadc.c line 557

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
+		if (!(adc->adc_info[i].is_correct_code))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);

There is no match for this in the remove... Should there be one? (not an interface I know anything about!)

Quoted 24 lines
drivers/iio/adc/palmas_gpadc.c new file +24 -0
+
+	return 0;
+
+out_irq_auto1_free:
+	if (adc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (adc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out_unregister_map:
+	if (adc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+out:
+	iio_device_free(iodev);
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	struct palmas_platform_data *pdata = dev_get_platdata(pdev->dev.parent);

drivers/iio/adc/palmas_gpadc.c line 591

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	if (pdata->gpadc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+	iio_device_unregister(iodev);
+	free_irq(adc->irq, adc);
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+	iio_device_free(iodev);

Could use demv_iio_device_alloc and not need the free here or on error path. Given location of irq frees you could do devm allocations of them as well which would cut out a fair bit of code without reordering anything.

nit: blank line here.

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP

drivers/iio/adc/palmas_gpadc.c line 605

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {

spacing around the /

Quoted 17 lines
drivers/iio/adc/palmas_gpadc.c new file +17 -0
+		if (((1000 * (1 << i))/32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {

drivers/iio/adc/palmas_gpadc.c line 632

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		int is_high;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			is_high = BIT(7);

BIT(7) is a bit random, so I'd suggest defining an appropriate macro for it this register field.

Quoted 56 lines
drivers/iio/adc/palmas_gpadc.c new file +56 -0
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int is_high;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;

drivers/iio/adc/palmas_gpadc.c line 698

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}

nitpick. Blank line here please.

Quoted 31 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	return 0;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0) {
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;

drivers/iio/adc/palmas_gpadc.c line 739

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);

nitpick. Blank line here please.

Quoted 72 lines
drivers/iio/adc/palmas_gpadc.c new file +58 -0
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +6 -0
index bb270bd..60acfa2 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+

include/linux/mfd/palmas.h struct palmas_pmic_driver_data { · line 148

Quoted 10 lines
include/linux/mfd/palmas.h +2 -3
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
 	int ch3_current;
-
 	/* Channel 0 current source can be used for battery detection.
-	 * If used for battery detection this will cause a permanent current
+	* If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
-	 */
+		 */

Some oddness occured here in patch creation. Do check your own patches as we all end up with stuff like this from time to time that should be picked up on beofre sending out. It just adds noise.

Quoted 26 lines
include/linux/mfd/palmas.h +7 -1
 	int ch0_current;
+	bool extended_delay;
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	struct iio_map *iio_maps;
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -403,7 +414,7 @@ struct palmas_gpadc_calibration {
 	s32 gain_error;
 	s32 offset_error;
 };
-
+/*
 struct palmas_gpadc {
 	struct device *dev;
 	struct palmas *palmas;

include/linux/mfd/palmas.h struct palmas_gpadc { · line 440

Quoted 4 lines
include/linux/mfd/palmas.h +1 -0
@@ -426,6 +437,9 @@ struct palmas_gpadc {
 	int conv1_channel;
 	int rt_channel;
 };
+*/

So there's a commented out bit of code in this patch? Sort that out before the next version.

Quoted 17 lines
include/linux/mfd/palmas.h +10 -0
+
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;
@@ -519,6 +533,42 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palma GPADC Channels */
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,

include/linux/mfd/palmas.h enum palmas_irqs { · line 553

Quoted 10 lines
include/linux/mfd/palmas.h +10 -0
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,

This does rather feel like an enum that doesn't add anything as the enum values = the index given at the end anyway...

include/linux/mfd/palmas.h enum palmas_irqs { · line 563

Quoted 10 lines
include/linux/mfd/palmas.h +10 -0
+	PALMAS_ADC_CH_MAX,
+};
+
+/* Palma GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};

Nitpick: new line here.

Quoted 20 lines
include/linux/mfd/palmas.h +9 -0
+/* Palma GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
@@ -2999,6 +3049,7 @@ enum usb_irq_events {
 #define PALMAS_GPADC_TRIM14					0x0D
 #define PALMAS_GPADC_TRIM15					0x0E
 #define PALMAS_GPADC_TRIM16					0x0F
+#define PALMAS_GPADC_TRIMINVALID				-1
 
 /* TPS659038 regen2_ctrl offset iss different from palmas */
 #define TPS659038_REGEN2_CTRL					0x12
#31 lore raw

Re: [PATCH 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

Hi,

Am 27.09.2015 um 17:21 schrieb Jonathan Cameron <[email protected]>:

Quoted 11 lines · H. Nikolaus Schaller
Quoted 9 lines

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

Various bits inline.

thanks for the valuable comments!

Will work into V2.

Nikolaus

Quoted 1,023 lines

Jonathan

Quoted 68 lines

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 797 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  59 ++-
4 files changed, 862 insertions(+), 4 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +9 -0
index eb0cd89..f6df9db 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Palmas series pmic chip by texas Instruments (twl6035/6037)
+	  is used in smartphones and tablets and supports a 16 channel
+	  general purpose ADC. Add iio driver to read different channel
+	  of the GPADCs.
+

config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +22 -0
new file mode 100644
index 0000000..17abb28
@@ -0,0 +1,797 @@
+/*
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>

drivers/iio/adc/palmas_gpadc.c line 32

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
+
+#define MOD_NAME "palmas-gpadc"
+#define ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define TO_BE_CALCULATED 0
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */

Full docs on this would be appreciated.

Quoted 38 lines
drivers/iio/adc/palmas_gpadc.c new file +38 -0
+	int x1;
+	int x2;
+	int v1;
+	int v2;
+	u8 trim1_reg;
+	u8 trim2_reg;
+	int gain;
+	int offset;
+	int gain_error;
+	bool is_correct_code;
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_correct_code)\
+[PALMAS_ADC_CH_##_chan] = {						\
+		.x1 = _x1,						\
+		.x2 = _x2,						\
+		.v1 = _v1,						\
+		.v2 = _v2,						\
+		.gain = TO_BE_CALCULATED,				\
+		.offset = TO_BE_CALCULATED,				\
+		.gain_error = TO_BE_CALCULATED,				\
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1,			\
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,			\
+		.is_correct_code = _is_correct_code			\
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),

drivers/iio/adc/palmas_gpadc.c line 80

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;

As there are some non obvious elements in here (such as the next two) kernel-doc for the whole stucture would be good.

Quoted 60 lines
drivers/iio/adc/palmas_gpadc.c new file +60 -0
+	u8				ch0_current;
+	u8				ch3_current;
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}

drivers/iio/adc/palmas_gpadc.c line 150

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}

return ret and drop the return above. Coccicheck will moan at you about this.

drivers/iio/adc/palmas_gpadc.c line 158

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return 0;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	complete(&adc->conv_completion);

blank line.

drivers/iio/adc/palmas_gpadc.c line 166

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_info(adc->dev, "Threshold interrupt %d occurs\n", irq);

Could indicate this to userspace... other than through the logs.

drivers/iio/adc/palmas_gpadc.c line 167

Quoted 1 line
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+	palmas_disable_auto_conversion(adc);

blank line generally before all function returns..

drivers/iio/adc/palmas_gpadc.c line 171

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc, int mask)

mask is boolean. Make it explicitly so for readability.

Quoted 147 lines
drivers/iio/adc/palmas_gpadc.c new file +147 -0
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/*Gain error Calculation*/
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/*Gain Calculation*/
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/*offset Calculation*/
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {

drivers/iio/adc/palmas_gpadc.c line 328

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		dev_err(adc->dev, "ADC_SW_START write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				ADC_CONVERSION_TIMEOUT);
+	if (ret == 0) {
+		dev_err(adc->dev, "ADC conversion not completed\n");
+		ret = -ETIMEDOUT;
+		return ret;

return -ETIMEDOUT; Might be worth you setting up to do sparse, smatch and
coccicheck runs on the code as they'll pick up on a lot of issues like this.

drivers/iio/adc/palmas_gpadc.c line 338

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADCDATA read failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = (val & 0xFFF);

nitpick : blank line here.

drivers/iio/adc/palmas_gpadc.c line 348

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (((val*1000) - adc->adc_info[adc_chan].offset) < 0) {
+		dev_err(adc->dev, "No Input Connected\n");
+		return 0;
+	}

Interesting, but should this not be caught for raw channel reads as well?

drivers/iio/adc/palmas_gpadc.c line 349

Quoted 1 line
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+

Umm. what is is_correct_code? Should be documented somewhere

Quoted 17 lines
drivers/iio/adc/palmas_gpadc.c new file +16 -0
+	if (!(adc->adc_info[adc_chan].is_correct_code))
+		val  = ((val*1000) - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)

= given I think your channels are 0 indexed?

drivers/iio/adc/palmas_gpadc.c line 372

Quoted 7 lines
drivers/iio/adc/palmas_gpadc.c new file +7 -0
+		return -EINVAL;
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:

I'd be tempted to split the two code paths here as that will be slightly easier to read (if longer).

I'm also highly suspicious of any driver that does processed output and has a return type of IIO_VAL_INT. Are you processed values in the units documented in Documentation/ABI/testing/sysfs-bus-iio? They could be, but I would like a comment saying that.

Quoted 29 lines
drivers/iio/adc/palmas_gpadc.c new file +29 -0
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start coversion failed\n");
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+}

drivers/iio/adc/palmas_gpadc.c line 411

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type)			\
+{									\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),			\
+	.type = _type,							\

Right now they are all voltage channels, why have that specifiable in this macro?

drivers/iio/adc/palmas_gpadc.c line 413

Quoted 2 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |			\
+			BIT(IIO_CHAN_INFO_PROCESSED),			\

Hmm. This is very very rarely justified. Either the driver has nice linear relationship between the raw value and the processed one, in which case you should be providing *_RAW, *_OFFSET and *_SCALE and leaving the maths to userspace (or the in kernel wrappers), or they are non linear in which case only the processed value is of interest and the raw one not directly so. (the only exception to this is light sensors where the processed channel is often a computed channel from several input _raw readings).

drivers/iio/adc/palmas_gpadc.c line 415

Quoted 2 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.indexed = 1,							\
+	.channel = PALMAS_ADC_CH_##chan,				\

Given this maps back to the value of chan, just put that in as a number and loose the enum. Channel is used in the naming so it doesn't make sense to obscure that behind an enum anyway.

Quoted 91 lines
drivers/iio/adc/palmas_gpadc.c new file +91 -0
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *adc_pdata;
+	struct iio_dev *iodev;
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+	adc_pdata = pdata->gpadc_pdata;
+
+	iodev = iio_device_alloc(sizeof(*adc));
+	if (!iodev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	if (adc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "iio_map_array_register failed\n");
+			goto out;
+		}
+	}
+
+	adc = iio_priv(iodev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, iodev);
+
+	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %dn", adc->irq, ret);
+		goto out_unregister_map;
+	}
+
+	if (adc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %dn",
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (adc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,

drivers/iio/adc/palmas_gpadc.c line 516

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %dn",
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+

I've requested kernel-doc above for ch0_current, but if that doesn't make it clear what matic is going on here then some comments here would be good.

Quoted 31 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	if (adc_pdata->ch0_current == 0)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (adc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (adc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	if (adc_pdata->ch3_current == 0)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (adc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (adc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = adc_pdata->extended_delay;
+
+	iodev->name = MOD_NAME;
+	iodev->dev.parent = &pdev->dev;
+	iodev->info = &palmas_gpadc_iio_info;
+	iodev->modes = INDIO_DIRECT_MODE;
+	iodev->channels = palmas_gpadc_iio_channel;
+	iodev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(iodev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;

drivers/iio/adc/palmas_gpadc.c line 557

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
+		if (!(adc->adc_info[i].is_correct_code))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);

There is no match for this in the remove... Should there be one? (not an interface I know anything about!)

Quoted 24 lines
drivers/iio/adc/palmas_gpadc.c new file +24 -0
+
+	return 0;
+
+out_irq_auto1_free:
+	if (adc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (adc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out_unregister_map:
+	if (adc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+out:
+	iio_device_free(iodev);
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	struct palmas_platform_data *pdata = dev_get_platdata(pdev->dev.parent);

drivers/iio/adc/palmas_gpadc.c line 591

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	if (pdata->gpadc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+	iio_device_unregister(iodev);
+	free_irq(adc->irq, adc);
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+	iio_device_free(iodev);

Could use demv_iio_device_alloc and not need the free here or on error path. Given location of irq frees you could do devm allocations of them as well which would cut out a fair bit of code without reordering anything.

nit: blank line here.

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP

drivers/iio/adc/palmas_gpadc.c line 605

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {

spacing around the /

Quoted 17 lines
drivers/iio/adc/palmas_gpadc.c new file +17 -0
+		if (((1000 * (1 << i))/32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {

drivers/iio/adc/palmas_gpadc.c line 632

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		int is_high;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			is_high = BIT(7);

BIT(7) is a bit random, so I'd suggest defining an appropriate macro for it this register field.

Quoted 56 lines
drivers/iio/adc/palmas_gpadc.c new file +56 -0
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int is_high;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;

drivers/iio/adc/palmas_gpadc.c line 698

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}

nitpick. Blank line here please.

Quoted 31 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	return 0;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0) {
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;

drivers/iio/adc/palmas_gpadc.c line 739

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);

nitpick. Blank line here please.

Quoted 72 lines
drivers/iio/adc/palmas_gpadc.c new file +58 -0
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +6 -0
index bb270bd..60acfa2 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+

include/linux/mfd/palmas.h struct palmas_pmic_driver_data { · line 148

Quoted 10 lines
include/linux/mfd/palmas.h +2 -3
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
 	int ch3_current;
-
 	/* Channel 0 current source can be used for battery detection.
-	 * If used for battery detection this will cause a permanent current
+	* If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
-	 */
+		 */

Some oddness occured here in patch creation. Do check your own patches as we all end up with stuff like this from time to time that should be picked up on beofre sending out. It just adds noise.

Quoted 26 lines
include/linux/mfd/palmas.h +7 -1
 	int ch0_current;
+	bool extended_delay;
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	struct iio_map *iio_maps;
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -403,7 +414,7 @@ struct palmas_gpadc_calibration {
 	s32 gain_error;
 	s32 offset_error;
 };
-
+/*
 struct palmas_gpadc {
 	struct device *dev;
 	struct palmas *palmas;

include/linux/mfd/palmas.h struct palmas_gpadc { · line 440

Quoted 4 lines
include/linux/mfd/palmas.h +1 -0
@@ -426,6 +437,9 @@ struct palmas_gpadc {
 	int conv1_channel;
 	int rt_channel;
 };
+*/

So there's a commented out bit of code in this patch? Sort that out before the next version.

Quoted 17 lines
include/linux/mfd/palmas.h +10 -0
+
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;
@@ -519,6 +533,42 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palma GPADC Channels */
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,

include/linux/mfd/palmas.h enum palmas_irqs { · line 553

Quoted 10 lines
include/linux/mfd/palmas.h +10 -0
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,

This does rather feel like an enum that doesn't add anything as the enum values = the index given at the end anyway...

include/linux/mfd/palmas.h enum palmas_irqs { · line 563

Quoted 10 lines
include/linux/mfd/palmas.h +10 -0
+	PALMAS_ADC_CH_MAX,
+};
+
+/* Palma GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};

Nitpick: new line here.

Quoted 20 lines
include/linux/mfd/palmas.h +9 -0
+/* Palma GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
@@ -2999,6 +3049,7 @@ enum usb_irq_events {
 #define PALMAS_GPADC_TRIM14					0x0D
 #define PALMAS_GPADC_TRIM15					0x0E
 #define PALMAS_GPADC_TRIM16					0x0F
+#define PALMAS_GPADC_TRIMINVALID				-1
 
 /* TPS659038 regen2_ctrl offset iss different from palmas */
 #define TPS659038_REGEN2_CTRL					0x12
#32 lore raw

Re: [PATCH 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

In reply to Jonathan Cameron

Am 27.09.2015 um 17:21 schrieb Jonathan Cameron <[email protected]>:

Quoted 11 lines · H. Nikolaus Schaller
Quoted 9 lines

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

Various bits inline.

Thanks again!

Worked into V2 (coming right after this mail).
Comments below, where/why we have not exactly followed your feedback.

BR,
Nikolaus

Quoted 73 lines

Jonathan

Quoted 71 lines

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 797 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  59 ++-
4 files changed, 862 insertions(+), 4 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +9 -0
index eb0cd89..f6df9db 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Palmas series pmic chip by texas Instruments (twl6035/6037)
+	  is used in smartphones and tablets and supports a 16 channel
+	  general purpose ADC. Add iio driver to read different channel
+	  of the GPADCs.
+

config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +25 -0
new file mode 100644
index 0000000..17abb28
@@ -0,0 +1,797 @@
+/*
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 32

Quoted 7 lines
drivers/iio/adc/palmas_gpadc.c new file +7 -0
+
+#define MOD_NAME "palmas-gpadc"
+#define ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define TO_BE_CALCULATED 0
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */

Full docs on this would be appreciated.

Is mostly defined in the Palmas data sheet but I have added some comments.

Quoted 42 lines
Quoted 42 lines
drivers/iio/adc/palmas_gpadc.c new file +42 -0
+	int x1;
+	int x2;
+	int v1;
+	int v2;
+	u8 trim1_reg;
+	u8 trim2_reg;
+	int gain;
+	int offset;
+	int gain_error;
+	bool is_correct_code;
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_correct_code)\
+[PALMAS_ADC_CH_##_chan] = {						\
+		.x1 = _x1,						\
+		.x2 = _x2,						\
+		.v1 = _v1,						\
+		.v2 = _v2,						\
+		.gain = TO_BE_CALCULATED,				\
+		.offset = TO_BE_CALCULATED,				\
+		.gain_error = TO_BE_CALCULATED,				\
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1,			\
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,			\
+		.is_correct_code = _is_correct_code			\
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 80

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;

As there are some non obvious elements in here (such as the next two) kernel-doc for the whole stucture would be good.

I don't know what all of them are doing, so I have only added the next two and some more.

Quoted 3 lines

drivers/iio/adc/palmas_gpadc.c line 82

Quoted 3 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	u8				ch0_current;
+	u8				ch3_current;

If I understand them correctly, they just store 2 bits each written into the current source registers. The value is only calculated in the probe function and internal to the driver. The bit pattern required is defined by the data sheet.

All TWL4030/TWL603x have such current sources. In the DT, just uA are specified and this is to temporarily store the bit pattern until it is sent to the chip. So it is quite a deep driver internal, but should be obvious to everyone who has the data sheet.

Quoted 79 lines
Quoted 58 lines
drivers/iio/adc/palmas_gpadc.c new file +58 -0
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}

drivers/iio/adc/palmas_gpadc.c line 150

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}

return ret and drop the return above. Coccicheck will moan at you about this.

drivers/iio/adc/palmas_gpadc.c line 158

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return 0;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	complete(&adc->conv_completion);

blank line.

Quoted 9 lines

drivers/iio/adc/palmas_gpadc.c line 166

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_info(adc->dev, "Threshold interrupt %d occurs\n", irq);

Could indicate this to userspace... other than through the logs.

Since I am not the original author and don't know how to present that to user space, I would leave this open for future improvements.

Quoted 181 lines

drivers/iio/adc/palmas_gpadc.c line 167

Quoted 2 lines
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+	palmas_disable_auto_conversion(adc);

blank line generally before all function returns..

drivers/iio/adc/palmas_gpadc.c line 171

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc, int mask)

mask is boolean. Make it explicitly so for readability.

Quoted 147 lines
drivers/iio/adc/palmas_gpadc.c new file +147 -0
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/*Gain error Calculation*/
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/*Gain Calculation*/
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/*offset Calculation*/
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {

drivers/iio/adc/palmas_gpadc.c line 328

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		dev_err(adc->dev, "ADC_SW_START write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				ADC_CONVERSION_TIMEOUT);
+	if (ret == 0) {
+		dev_err(adc->dev, "ADC conversion not completed\n");
+		ret = -ETIMEDOUT;
+		return ret;

return -ETIMEDOUT; Might be worth you setting up to do sparse, smatch and
coccicheck runs on the code as they'll pick up on a lot of issues like this.

drivers/iio/adc/palmas_gpadc.c line 338

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADCDATA read failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = (val & 0xFFF);

nitpick : blank line here.

Quoted 11 lines

drivers/iio/adc/palmas_gpadc.c line 348

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (((val*1000) - adc->adc_info[adc_chan].offset) < 0) {
+		dev_err(adc->dev, "No Input Connected\n");
+		return 0;
+	}

Interesting, but should this not be caught for raw channel reads as well?

It appears that this function is only called for calibrated values.

The condition appears to detect that the ADC has reported a value smaller than the calibration allows.

This indicates to be some error condition (maybe a floating ADC input?). But I don't know enough of the Palmas details to judge this. So I have only moved it behind the (optional) calibration calculation and detect if either would report negative values. So it essentially clamps voltage reports at 0V.

Quoted 2 lines

drivers/iio/adc/palmas_gpadc.c line 349

Quoted 1 line
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+

Umm. what is is_correct_code? Should be documented somewhere

have renamed it to is_uncalibrated which means that there is no register on the chip to read out calibration data for that channel.

Quoted 18 lines
Quoted 18 lines
drivers/iio/adc/palmas_gpadc.c new file +16 -0
+	if (!(adc->adc_info[adc_chan].is_correct_code))
+		val  = ((val*1000) - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)

= given I think your channels are 0 indexed?

drivers/iio/adc/palmas_gpadc.c new file +1 -0
+		return -EINVAL;
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 372

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:

I'd be tempted to split the two code paths here as that will be slightly easier to read (if longer).

IMHO would duplicate code (not sure if gcc can detect). Therefore I have left it as is.

Quoted 4 lines

I'm also highly suspicious of any driver that does processed output and has a return type of IIO_VAL_INT. Are you processed values in the units documented in Documentation/ABI/testing/sysfs-bus-iio?

"Units after application of scale and offset are millivolts."

The driver reports millivolts on OMAP5 uEVM, e.g. VBUS = 5013 (mV).

Quoted 35 lines

They could be, but I would like a comment saying that.

Quoted 33 lines
drivers/iio/adc/palmas_gpadc.c new file +33 -0
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start coversion failed\n");
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+}
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 411

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type)			\
+{									\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),			\
+	.type = _type,							\

Right now they are all voltage channels, why have that specifiable in this macro?

There are two channels that are or can be temperature values. Presented as voltages across an NTC/Diode. Since the NTC is outside the Palmas chip the conversion into temperature is left to user-space. The type value is in preparation for this and I have changed it to IIO_TEMPERATURE.

Quoted 9 lines

drivers/iio/adc/palmas_gpadc.c line 413

Quoted 2 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |			\
+			BIT(IIO_CHAN_INFO_PROCESSED),			\

Hmm. This is very very rarely justified. Either the driver has nice linear relationship between the raw value and the processed one, in which case you should be providing *_RAW, *_OFFSET and *_SCALE and leaving the maths to userspace (or the in kernel wrappers), or they are non linear in which case only the processed value is of interest and the raw one not directly so. (the only exception to this is light sensors where the processed channel is often a computed channel from several input _raw readings).

If I understand the driver correctty (I am not the original author), the calibration is indeed linear, but translating the calibration data into *_RAW, *_OFFSET and *_SCALE is different from how Texas suggests to do the calculation (maybe avoiding rounding and truncation errors).

Well, we could simply remove the _RAW values - but IMHO it is sometimes good to be able to see what is going on. They can still be ignored by the consumer.

Quoted 6 lines

drivers/iio/adc/palmas_gpadc.c line 415

Quoted 3 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.indexed = 1,							\
+	.channel = PALMAS_ADC_CH_##chan,				\

Given this maps back to the value of chan, just put that in as a number and loose the enum. Channel is used in the naming so it doesn't make sense to obscure that behind an enum anyway.

Well, enum constants allow to check against typos and that it matches the channel numbers defined in the header.

So PALMAS_ADC_CHAN_IIO(IN24, ...) would make gcc warn while. PALMAS_ADC_CHAN_IIO(24, ...) would fail in operation.

Of course we have no typo, because it is already checked to compile :)

So I understand by which good C coding practise it was probably introduced and am undecided if it is good to keep or not.

Quoted 96 lines
Quoted 96 lines
drivers/iio/adc/palmas_gpadc.c new file +96 -0
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *adc_pdata;
+	struct iio_dev *iodev;
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+	adc_pdata = pdata->gpadc_pdata;
+
+	iodev = iio_device_alloc(sizeof(*adc));
+	if (!iodev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	if (adc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "iio_map_array_register failed\n");
+			goto out;
+		}
+	}
+
+	adc = iio_priv(iodev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, iodev);
+
+	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %dn", adc->irq, ret);
+		goto out_unregister_map;
+	}
+
+	if (adc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %dn",
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (adc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %dn",
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 516

Quoted 5 lines
drivers/iio/adc/palmas_gpadc.c new file +5 -0
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+

I've requested kernel-doc above for ch0_current, but if that doesn't make it clear what matic is going on here then some comments here would be good.

I think the best location to make that clear is the DT bindings where you can set the values (in microamperes and not encoded values).

Basically you specify in the adc_pdata the uA you want and this function does a quantization to perpare 2 bits that can be written into the control registers later on.

Quoted 35 lines
Quoted 35 lines
drivers/iio/adc/palmas_gpadc.c new file +35 -0
+	if (adc_pdata->ch0_current == 0)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (adc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (adc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	if (adc_pdata->ch3_current == 0)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (adc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (adc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = adc_pdata->extended_delay;
+
+	iodev->name = MOD_NAME;
+	iodev->dev.parent = &pdev->dev;
+	iodev->info = &palmas_gpadc_iio_info;
+	iodev->modes = INDIO_DIRECT_MODE;
+	iodev->channels = palmas_gpadc_iio_channel;
+	iodev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(iodev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 557

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+		if (!(adc->adc_info[i].is_correct_code))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);

There is no match for this in the remove... Should there be one? (not an interface I know anything about!)

Same for me and I have no idea how to test. But it looks reasonable to call device_wakeup_disable(). Therefore I have added.

Quoted 29 lines
Quoted 29 lines
drivers/iio/adc/palmas_gpadc.c new file +29 -0
+
+	return 0;
+
+out_irq_auto1_free:
+	if (adc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (adc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out_unregister_map:
+	if (adc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+out:
+	iio_device_free(iodev);
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	struct palmas_platform_data *pdata = dev_get_platdata(pdev->dev.parent);
+
+	if (pdata->gpadc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+	iio_device_unregister(iodev);
+	free_irq(adc->irq, adc);
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 591

Quoted 5 lines
drivers/iio/adc/palmas_gpadc.c new file +5 -0
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+	iio_device_free(iodev);

Could use demv_iio_device_alloc and not need the free here or on error path. Given location of irq frees you could do devm allocations of them as well which would cut out a fair bit of code without reordering anything.

has also been proposed by Peter Meerwald and has been reworked.

Quoted 38 lines

nit: blank line here.

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP

drivers/iio/adc/palmas_gpadc.c line 605

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {

spacing around the /

Quoted 21 lines
drivers/iio/adc/palmas_gpadc.c new file +21 -0
+		if (((1000 * (1 << i))/32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {
+		int is_high;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 632

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			is_high = BIT(7);

BIT(7) is a bit random, so I'd suggest defining an appropriate macro for it this register field.

There is indeed a proper macro in palmas.h

Quoted 219 lines
Quoted 57 lines
drivers/iio/adc/palmas_gpadc.c new file +56 -0
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int is_high;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;

drivers/iio/adc/palmas_gpadc.c line 698

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}

nitpick. Blank line here please.

Quoted 31 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	return 0;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0) {
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;

drivers/iio/adc/palmas_gpadc.c line 739

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);

nitpick. Blank line here please.

Quoted 72 lines
drivers/iio/adc/palmas_gpadc.c new file +58 -0
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +6 -0
index bb270bd..60acfa2 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+

include/linux/mfd/palmas.h struct palmas_pmic_driver_data { · line 148

Quoted 10 lines
include/linux/mfd/palmas.h +2 -3
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
 	int ch3_current;
-
 	/* Channel 0 current source can be used for battery detection.
-	 * If used for battery detection this will cause a permanent current
+	* If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
-	 */
+		 */

Some oddness occured here in patch creation. Do check your own patches as we all end up with stuff like this from time to time that should be picked up on beofre sending out. It just adds noise.

Quoted 24 lines
include/linux/mfd/palmas.h +7 -1
 	int ch0_current;
+	bool extended_delay;
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	struct iio_map *iio_maps;
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -403,7 +414,7 @@ struct palmas_gpadc_calibration {
 	s32 gain_error;
 	s32 offset_error;
 };
-
+/*
 struct palmas_gpadc {
 	struct device *dev;
Quoted 8 lines

include/linux/mfd/palmas.h struct palmas_gpadc { · line 440

Quoted 6 lines
include/linux/mfd/palmas.h +1 -0
 	struct palmas *palmas;
@@ -426,6 +437,9 @@ struct palmas_gpadc {
 	int conv1_channel;
 	int rt_channel;
 };
+*/

So there's a commented out bit of code in this patch? Sort that out before the next version.

We forgot to delete...

Quoted 22 lines
Quoted 22 lines
include/linux/mfd/palmas.h +14 -0
+
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;
@@ -519,6 +533,42 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palma GPADC Channels */
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
Quoted 8 lines

include/linux/mfd/palmas.h enum palmas_irqs { · line 553

Quoted 6 lines
include/linux/mfd/palmas.h +6 -0
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,

This does rather feel like an enum that doesn't add anything as the enum values = the index given at the end anyway...

As said it adds some (questionable) check by the compiler,

Quoted 1 line

include/linux/mfd/palmas.h enum palmas_irqs { · line 554

Quoted 1 line
include/linux/mfd/palmas.h +1 -0
+	PALMAS_ADC_CH_MAX,

and this constant is indirectly defined.

Quoted 31 lines

include/linux/mfd/palmas.h enum palmas_irqs { · line 563

Quoted 9 lines
include/linux/mfd/palmas.h +9 -0
+};
+
+/* Palma GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};

Nitpick: new line here.

Quoted 20 lines
include/linux/mfd/palmas.h +9 -0
+/* Palma GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
@@ -2999,6 +3049,7 @@ enum usb_irq_events {
 #define PALMAS_GPADC_TRIM14					0x0D
 #define PALMAS_GPADC_TRIM15					0x0E
 #define PALMAS_GPADC_TRIM16					0x0F
+#define PALMAS_GPADC_TRIMINVALID				-1
 
 /* TPS659038 regen2_ctrl offset iss different from palmas */
 #define TPS659038_REGEN2_CTRL					0x12
maintainer
#33 lore raw

Re: [PATCH 1/3] iio:adc: add iio driver for Palmas (twl6035/7) gpadc

Quoted 13 lines · H. Nikolaus Schaller

Am 27.09.2015 um 17:21 schrieb Jonathan Cameron <[email protected]>:

Quoted 10 lines · H. Nikolaus Schaller
Quoted 9 lines

This driver code was found as:

https://android.googlesource.com/kernel/tegra/+/aaabb2e045f31e5a970109ffdaae900dd403d17e/drivers/staging/iio/adc

Fixed various compilation issues and test this driver on omap5 evm.

Quoted 6 lines
Quoted 1 line

Various bits inline.

Thanks again!

Worked into V2 (coming right after this mail).
Comments below, where/why we have not exactly followed your feedback.

Responses inline, but we haven't disagreed on anything important so none of it really matters!

Jonathan

Quoted 240 lines

BR,
Nikolaus

Quoted 73 lines

Jonathan

Quoted 71 lines

---

drivers/iio/adc/Kconfig        |   9 +
drivers/iio/adc/Makefile       |   1 +
drivers/iio/adc/palmas_gpadc.c | 797 +++++++++++++++++++++++++++++++++++++++++
include/linux/mfd/palmas.h     |  59 ++-
4 files changed, 862 insertions(+), 4 deletions(-)

create mode 100644 drivers/iio/adc/palmas_gpadc.c

drivers/iio/adc/Kconfig +9 -0
index eb0cd89..f6df9db 100644
@@ -242,6 +242,15 @@ config NAU7802
 	  To compile this driver as a module, choose M here: the
 	  module will be called nau7802.
 
+config PALMAS_GPADC
+	tristate "TI Palmas General Purpose ADC"
+	depends on MFD_PALMAS
+	help
+	  Palmas series pmic chip by texas Instruments (twl6035/6037)
+	  is used in smartphones and tablets and supports a 16 channel
+	  general purpose ADC. Add iio driver to read different channel
+	  of the GPADCs.
+

config QCOM_SPMI_IADC
tristate "Qualcomm SPMI PMIC current ADC"
depends on SPMI

drivers/iio/adc/Makefile +0 -0
index a096210..716f112 100644
@@ -26,6 +26,7 @@ obj-$(CONFIG_MCP320X) += mcp320x.o

obj-$(CONFIG_MCP3422) += mcp3422.o obj-$(CONFIG_MEN_Z188_ADC) += men_z188_adc.o obj-$(CONFIG_NAU7802) += nau7802.o +obj-$(CONFIG_PALMAS_GPADC) += palmas_gpadc.o obj-$(CONFIG_QCOM_SPMI_IADC) += qcom-spmi-iadc.o obj-$(CONFIG_QCOM_SPMI_VADC) += qcom-spmi-vadc.o obj-$(CONFIG_ROCKCHIP_SARADC) += rockchip_saradc.o

drivers/iio/adc/palmas_gpadc.c new file +25 -0
new file mode 100644
index 0000000..17abb28
@@ -0,0 +1,797 @@
+/*
+ * palmas-adc.c -- TI PALMAS GPADC.
+ *
+ * Copyright (c) 2013, NVIDIA Corporation. All rights reserved.
+ *
+ * Author: Pradeep Goudagunta <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ */
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/i2c.h>
+#include <linux/pm.h>
+#include <linux/mfd/palmas.h>
+#include <linux/completion.h>
+#include <linux/iio/iio.h>
+#include <linux/iio/machine.h>
+#include <linux/iio/driver.h>
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 32

Quoted 7 lines
drivers/iio/adc/palmas_gpadc.c new file +7 -0
+
+#define MOD_NAME "palmas-gpadc"
+#define ADC_CONVERSION_TIMEOUT	(msecs_to_jiffies(5000))
+#define TO_BE_CALCULATED 0
+
+struct palmas_gpadc_info {
+/* calibration codes and regs */

Full docs on this would be appreciated.

Is mostly defined in the Palmas data sheet but I have added some comments.

Quoted 42 lines
Quoted 42 lines
drivers/iio/adc/palmas_gpadc.c new file +42 -0
+	int x1;
+	int x2;
+	int v1;
+	int v2;
+	u8 trim1_reg;
+	u8 trim2_reg;
+	int gain;
+	int offset;
+	int gain_error;
+	bool is_correct_code;
+};
+
+#define PALMAS_ADC_INFO(_chan, _x1, _x2, _v1, _v2, _t1, _t2, _is_correct_code)\
+[PALMAS_ADC_CH_##_chan] = {						\
+		.x1 = _x1,						\
+		.x2 = _x2,						\
+		.v1 = _v1,						\
+		.v2 = _v2,						\
+		.gain = TO_BE_CALCULATED,				\
+		.offset = TO_BE_CALCULATED,				\
+		.gain_error = TO_BE_CALCULATED,				\
+		.trim1_reg = PALMAS_GPADC_TRIM##_t1,			\
+		.trim2_reg = PALMAS_GPADC_TRIM##_t2,			\
+		.is_correct_code = _is_correct_code			\
+	}
+
+static struct palmas_gpadc_info palmas_gpadc_info[] = {
+	PALMAS_ADC_INFO(IN0, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN1, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN2, 2064, 3112, 1260, 1900, 3, 4, false),
+	PALMAS_ADC_INFO(IN3, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN4, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN5, 2064, 3112, 630, 950, 1, 2, false),
+	PALMAS_ADC_INFO(IN6, 2064, 3112, 2520, 3800, 5, 6, false),
+	PALMAS_ADC_INFO(IN7, 2064, 3112, 2520, 3800, 7, 8, false),
+	PALMAS_ADC_INFO(IN8, 2064, 3112, 3150, 4750, 9, 10, false),
+	PALMAS_ADC_INFO(IN9, 2064, 3112, 5670, 8550, 11, 12, false),
+	PALMAS_ADC_INFO(IN10, 2064, 3112, 3465, 5225, 13, 14, false),
+	PALMAS_ADC_INFO(IN11, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN12, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN13, 0, 0, 0, 0, INVALID, INVALID, true),
+	PALMAS_ADC_INFO(IN14, 2064, 3112, 3645, 5225, 15, 16, false),
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 80

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+	PALMAS_ADC_INFO(IN15, 0, 0, 0, 0, INVALID, INVALID, true),
+};
+
+struct palmas_gpadc {
+	struct device			*dev;
+	struct palmas			*palmas;

As there are some non obvious elements in here (such as the next two) kernel-doc for the whole stucture would be good.

I don't know what all of them are doing, so I have only added the next two and some more.

Quoted 3 lines

drivers/iio/adc/palmas_gpadc.c line 82

Quoted 3 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	u8				ch0_current;
+	u8				ch3_current;

If I understand them correctly, they just store 2 bits each written into the current source registers. The value is only calculated in the probe function and internal to the driver. The bit pattern required is defined by the data sheet.

All TWL4030/TWL603x have such current sources. In the DT, just uA are specified and this is to temporarily store the bit pattern until it is sent to the chip. So it is quite a deep driver internal, but should be obvious to everyone who has the data sheet.

Quoted 79 lines
Quoted 58 lines
drivers/iio/adc/palmas_gpadc.c new file +58 -0
+	bool				extended_delay;
+	int				irq;
+	int				irq_auto_0;
+	int				irq_auto_1;
+	struct palmas_gpadc_info	*adc_info;
+	struct completion		conv_completion;
+	struct palmas_adc_wakeup_property wakeup1_data;
+	struct palmas_adc_wakeup_property wakeup2_data;
+	bool				wakeup1_enable;
+	bool				wakeup2_enable;
+	int				auto_conversion_period;
+};
+
+/*
+ * GPADC lock issue in AUTO mode.
+ * Impact: In AUTO mode, GPADC conversion can be locked after disabling AUTO
+ *	   mode feature.
+ * Details:
+ *	When the AUTO mode is the only conversion mode enabled, if the AUTO
+ *	mode feature is disabled with bit GPADC_AUTO_CTRL.  AUTO_CONV1_EN = 0
+ *	or bit GPADC_AUTO_CTRL.  AUTO_CONV0_EN = 0 during a conversion, the
+ *	conversion mechanism can be seen as locked meaning that all following
+ *	conversion will give 0 as a result.  Bit GPADC_STATUS.GPADC_AVAILABLE
+ *	will stay at 0 meaning that GPADC is busy.  An RT conversion can unlock
+ *	the GPADC.
+ *
+ * Workaround(s):
+ *	To avoid the lock mechanism, the workaround to follow before any stop
+ *	conversion request is:
+ *	Force the GPADC state machine to be ON by using the GPADC_CTRL1.
+ *		GPADC_FORCE bit = 1
+ *	Shutdown the GPADC AUTO conversion using
+ *		GPADC_AUTO_CTRL.SHUTDOWN_CONV[01] = 0.
+ *	After 100us, force the GPADC state machine to be OFF by using the
+ *		GPADC_CTRL1.  GPADC_FORCE bit = 0
+ */
+static int palmas_disable_auto_conversion(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV1 |
+			PALMAS_GPADC_AUTO_CTRL_SHUTDOWN_CONV0,
+			0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL update failed: %d\n", ret);
+		return ret;
+	}

drivers/iio/adc/palmas_gpadc.c line 150

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	udelay(100);
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_CTRL1,
+			PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "GPADC_CTRL1 update failed: %d\n", ret);
+		return ret;
+	}

return ret and drop the return above. Coccicheck will moan at you about this.

drivers/iio/adc/palmas_gpadc.c line 158

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return 0;
+}
+
+static irqreturn_t palmas_gpadc_irq(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	complete(&adc->conv_completion);

blank line.

Quoted 9 lines

drivers/iio/adc/palmas_gpadc.c line 166

Quoted 8 lines
drivers/iio/adc/palmas_gpadc.c new file +8 -0
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t palmas_gpadc_irq_auto(int irq, void *data)
+{
+	struct palmas_gpadc *adc = data;
+
+	dev_info(adc->dev, "Threshold interrupt %d occurs\n", irq);

Could indicate this to userspace... other than through the logs.

Quoted 2 lines

Since I am not the original author and don't know how to present that to user space, I would leave this open for future improvements.

Fair enough.

Quoted 240 lines
Quoted 182 lines

drivers/iio/adc/palmas_gpadc.c line 167

Quoted 3 lines
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+	palmas_disable_auto_conversion(adc);

blank line generally before all function returns..

drivers/iio/adc/palmas_gpadc.c line 171

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return IRQ_HANDLED;
+}
+
+static int palmas_gpadc_start_mask_interrupt(struct palmas_gpadc *adc, int mask)

mask is boolean. Make it explicitly so for readability.

Quoted 147 lines
drivers/iio/adc/palmas_gpadc.c new file +147 -0
+{
+	int ret;
+
+	if (!mask)
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW, 0);
+	else
+		ret = palmas_update_bits(adc->palmas, PALMAS_INTERRUPT_BASE,
+					PALMAS_INT3_MASK,
+					PALMAS_INT3_MASK_GPADC_EOC_SW,
+					PALMAS_INT3_MASK_GPADC_EOC_SW);
+	if (ret < 0)
+		dev_err(adc->dev, "GPADC INT MASK update failed: %d\n", ret);
+
+	return ret;
+}
+
+static int palmas_gpadc_enable(struct palmas_gpadc *adc, int adc_chan,
+			       int enable)
+{
+	unsigned int mask, val;
+	int ret;
+
+	if (enable) {
+		val = (adc->extended_delay
+			<< PALMAS_GPADC_RT_CTRL_EXTEND_DELAY_SHIFT);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+					PALMAS_GPADC_RT_CTRL,
+					PALMAS_GPADC_RT_CTRL_EXTEND_DELAY, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "RT_CTRL update failed: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_MASK |
+			PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_MASK |
+			PALMAS_GPADC_CTRL1_GPADC_FORCE);
+		val = (adc->ch0_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH0_SHIFT);
+		val |= (adc->ch3_current
+			<< PALMAS_GPADC_CTRL1_CURRENT_SRC_CH3_SHIFT);
+		val |= PALMAS_GPADC_CTRL1_GPADC_FORCE;
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"Failed to update current setting: %d\n", ret);
+			return ret;
+		}
+
+		mask = (PALMAS_GPADC_SW_SELECT_SW_CONV0_SEL_MASK |
+			PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		val = (adc_chan | PALMAS_GPADC_SW_SELECT_SW_CONV_EN);
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, mask, val);
+		if (ret < 0) {
+			dev_err(adc->dev, "SW_SELECT update failed: %d\n", ret);
+			return ret;
+		}
+	} else {
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT, 0);
+		if (ret < 0)
+			dev_err(adc->dev, "SW_SELECT write failed: %d\n", ret);
+
+		ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_CTRL1,
+				PALMAS_GPADC_CTRL1_GPADC_FORCE, 0);
+		if (ret < 0) {
+			dev_err(adc->dev, "CTRL1 update failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	return ret;
+}
+
+static int palmas_gpadc_read_prepare(struct palmas_gpadc *adc, int adc_chan)
+{
+	int ret;
+
+	ret = palmas_gpadc_enable(adc, adc_chan, true);
+	if (ret < 0)
+		return ret;
+
+	return palmas_gpadc_start_mask_interrupt(adc, 0);
+}
+
+static void palmas_gpadc_read_done(struct palmas_gpadc *adc, int adc_chan)
+{
+	palmas_gpadc_start_mask_interrupt(adc, 1);
+	palmas_gpadc_enable(adc, adc_chan, false);
+}
+
+static int palmas_gpadc_calibrate(struct palmas_gpadc *adc, int adc_chan)
+{
+	int k;
+	int d1;
+	int d2;
+	int ret;
+	int gain;
+	int x1 =  adc->adc_info[adc_chan].x1;
+	int x2 =  adc->adc_info[adc_chan].x2;
+	int v1 = adc->adc_info[adc_chan].v1;
+	int v2 = adc->adc_info[adc_chan].v2;
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim1_reg, &d1);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	ret = palmas_read(adc->palmas, PALMAS_TRIM_GPADC_BASE,
+				adc->adc_info[adc_chan].trim2_reg, &d2);
+	if (ret < 0) {
+		dev_err(adc->dev, "TRIM read failed: %d\n", ret);
+		goto scrub;
+	}
+
+	/*Gain error Calculation*/
+	k = (1000 + (1000 * (d2 - d1)) / (x2 - x1));
+
+	/*Gain Calculation*/
+	gain = ((v2 - v1) * 1000) / (x2 - x1);
+
+	adc->adc_info[adc_chan].gain_error = k;
+	adc->adc_info[adc_chan].gain = gain;
+	/*offset Calculation*/
+	adc->adc_info[adc_chan].offset = (d1 * 1000) - ((k - 1000) * x1);
+
+scrub:
+	return ret;
+}
+
+static int palmas_gpadc_start_conversion(struct palmas_gpadc *adc, int adc_chan)
+{
+	unsigned int val;
+	int ret;
+
+	init_completion(&adc->conv_completion);
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_SELECT,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0,
+				PALMAS_GPADC_SW_SELECT_SW_START_CONV0);
+	if (ret < 0) {

drivers/iio/adc/palmas_gpadc.c line 328

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+		dev_err(adc->dev, "ADC_SW_START write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = wait_for_completion_timeout(&adc->conv_completion,
+				ADC_CONVERSION_TIMEOUT);
+	if (ret == 0) {
+		dev_err(adc->dev, "ADC conversion not completed\n");
+		ret = -ETIMEDOUT;
+		return ret;

return -ETIMEDOUT; Might be worth you setting up to do sparse, smatch and
coccicheck runs on the code as they'll pick up on a lot of issues like this.

drivers/iio/adc/palmas_gpadc.c line 338

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_bulk_read(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_SW_CONV0_LSB, &val, 2);
+	if (ret < 0) {
+		dev_err(adc->dev, "ADCDATA read failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = (val & 0xFFF);

nitpick : blank line here.

Quoted 11 lines

drivers/iio/adc/palmas_gpadc.c line 348

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	return ret;
+}
+
+static int palmas_gpadc_get_calibrated_code(struct palmas_gpadc *adc,
+						int adc_chan, int val)
+{
+	if (((val*1000) - adc->adc_info[adc_chan].offset) < 0) {
+		dev_err(adc->dev, "No Input Connected\n");
+		return 0;
+	}

Interesting, but should this not be caught for raw channel reads as well?

It appears that this function is only called for calibrated values.

The condition appears to detect that the ADC has reported a value smaller than the calibration allows.

This indicates to be some error condition (maybe a floating ADC input?). But I don't know enough of the Palmas details to judge this. So I have only moved it behind the (optional) calibration calculation and detect if either would report negative values. So it essentially clamps voltage reports at 0V.

Quoted 2 lines

drivers/iio/adc/palmas_gpadc.c line 349

Quoted 1 line
drivers/iio/adc/palmas_gpadc.c new file +1 -0
+

Umm. what is is_correct_code? Should be documented somewhere

have renamed it to is_uncalibrated which means that there is no register on the chip to read out calibration data for that channel.

Quoted 18 lines
Quoted 18 lines
drivers/iio/adc/palmas_gpadc.c new file +16 -0
+	if (!(adc->adc_info[adc_chan].is_correct_code))
+		val  = ((val*1000) - adc->adc_info[adc_chan].offset) /
+					adc->adc_info[adc_chan].gain_error;
+
+	val = (val * adc->adc_info[adc_chan].gain) / 1000;
+	return val;
+}
+
+static int palmas_gpadc_read_raw(struct iio_dev *indio_dev,
+	struct iio_chan_spec const *chan, int *val, int *val2, long mask)
+{
+	struct  palmas_gpadc *adc = iio_priv(indio_dev);
+	int adc_chan = chan->channel;
+	int ret = 0;
+
+	if (adc_chan > PALMAS_ADC_CH_MAX)

= given I think your channels are 0 indexed?

drivers/iio/adc/palmas_gpadc.c new file +1 -0
+		return -EINVAL;
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 372

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+
+	mutex_lock(&indio_dev->mlock);
+
+	switch (mask) {
+	case IIO_CHAN_INFO_RAW:
+	case IIO_CHAN_INFO_PROCESSED:

I'd be tempted to split the two code paths here as that will be slightly easier to read (if longer).

IMHO would duplicate code (not sure if gcc can detect). Therefore I have left it as is.

Quoted 1 line
Quoted 7 lines
Quoted 3 lines

I'm also highly suspicious of any driver that does processed output and has a return type of IIO_VAL_INT. Are you processed values in the units documented in Documentation/ABI/testing/sysfs-bus-iio?

"Units after application of scale and offset are millivolts."

The driver reports millivolts on OMAP5 uEVM, e.g. VBUS = 5013 (mV).

cool

Quoted 60 lines
Quoted 36 lines

They could be, but I would like a comment saying that.

Quoted 33 lines
drivers/iio/adc/palmas_gpadc.c new file +33 -0
+		ret = palmas_gpadc_read_prepare(adc, adc_chan);
+		if (ret < 0)
+			goto out;
+
+		ret = palmas_gpadc_start_conversion(adc, adc_chan);
+		if (ret < 0) {
+			dev_err(adc->dev,
+			"ADC start coversion failed\n");
+			goto out;
+		}
+
+		if (mask == IIO_CHAN_INFO_PROCESSED)
+			ret = palmas_gpadc_get_calibrated_code(
+							adc, adc_chan, ret);
+
+		*val = ret;
+
+		ret = IIO_VAL_INT;
+		goto out;
+	}
+
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+
+out:
+	palmas_gpadc_read_done(adc, adc_chan);
+	mutex_unlock(&indio_dev->mlock);
+	return ret;
+}
+
+static const struct iio_info palmas_gpadc_iio_info = {
+	.read_raw = palmas_gpadc_read_raw,
+	.driver_module = THIS_MODULE,
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 411

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+};
+
+#define PALMAS_ADC_CHAN_IIO(chan, _type)			\
+{									\
+	.datasheet_name = PALMAS_DATASHEET_NAME(chan),			\
+	.type = _type,							\

Right now they are all voltage channels, why have that specifiable in this macro?

There are two channels that are or can be temperature values. Presented as voltages across an NTC/Diode. Since the NTC is outside the Palmas chip the conversion into temperature is left to user-space. The type value is in preparation for this and I have changed it to IIO_TEMPERATURE.

Quoted 9 lines

drivers/iio/adc/palmas_gpadc.c line 413

Quoted 2 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW) |			\
+			BIT(IIO_CHAN_INFO_PROCESSED),			\

Hmm. This is very very rarely justified. Either the driver has nice linear relationship between the raw value and the processed one, in which case you should be providing *_RAW, *_OFFSET and *_SCALE and leaving the maths to userspace (or the in kernel wrappers), or they are non linear in which case only the processed value is of interest and the raw one not directly so. (the only exception to this is light sensors where the processed channel is often a computed channel from several input _raw readings).

Quoted 7 lines

If I understand the driver correctty (I am not the original author), the calibration is indeed linear, but translating the calibration data into *_RAW, *_OFFSET and *_SCALE is different from how Texas suggests to do the calculation (maybe avoiding rounding and truncation errors).

Well, we could simply remove the _RAW values - but IMHO it is sometimes good to be able to see what is going on. They can still be ignored by the consumer.

hmm.. I'm not totally happy with having both, but will let it go I think.

Quoted 11 lines
Quoted 7 lines

drivers/iio/adc/palmas_gpadc.c line 415

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +2 -0
+	.indexed = 1,							\
+	.channel = PALMAS_ADC_CH_##chan,				\

Given this maps back to the value of chan, just put that in as a number and loose the enum. Channel is used in the naming so it doesn't make sense to obscure that behind an enum anyway.

Well, enum constants allow to check against typos and that it matches the channel numbers defined in the header.

Quoted 7 lines

So PALMAS_ADC_CHAN_IIO(IN24, ...) would make gcc warn while. PALMAS_ADC_CHAN_IIO(24, ...) would fail in operation.

Of course we have no typo, because it is already checked to compile :)

So I understand by which good C coding practise it was probably introduced and am undecided if it is good to keep or not.

hmm. not writing silly bugs seems like a better idea to me, or convoluted tricky to check code. I'm not that fussed however.

Quoted 549 lines
Quoted 97 lines
Quoted 97 lines
drivers/iio/adc/palmas_gpadc.c new file +96 -0
+}
+
+static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
+	PALMAS_ADC_CHAN_IIO(IN0, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN1, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN2, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN3, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN4, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN5, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN6, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN7, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN8, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN9, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN10, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN11, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN12, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN13, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN14, IIO_VOLTAGE),
+	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
+};
+
+static int palmas_gpadc_probe(struct platform_device *pdev)
+{
+	struct palmas_gpadc *adc;
+	struct palmas_platform_data *pdata;
+	struct palmas_gpadc_platform_data *adc_pdata;
+	struct iio_dev *iodev;
+	int ret, i;
+
+	pdata = dev_get_platdata(pdev->dev.parent);
+	if (!pdata || !pdata->gpadc_pdata) {
+		dev_err(&pdev->dev, "No platform data\n");
+		return -ENODEV;
+	}
+	adc_pdata = pdata->gpadc_pdata;
+
+	iodev = iio_device_alloc(sizeof(*adc));
+	if (!iodev) {
+		dev_err(&pdev->dev, "iio_device_alloc failed\n");
+		return -ENOMEM;
+	}
+
+	if (adc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "iio_map_array_register failed\n");
+			goto out;
+		}
+	}
+
+	adc = iio_priv(iodev);
+	adc->dev = &pdev->dev;
+	adc->palmas = dev_get_drvdata(pdev->dev.parent);
+	adc->adc_info = palmas_gpadc_info;
+	init_completion(&adc->conv_completion);
+	dev_set_drvdata(&pdev->dev, iodev);
+
+	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
+	ret = request_threaded_irq(adc->irq, NULL,
+		palmas_gpadc_irq,
+		IRQF_ONESHOT | IRQF_EARLY_RESUME, dev_name(adc->dev),
+		adc);
+	if (ret < 0) {
+		dev_err(adc->dev,
+			"request irq %d failed: %dn", adc->irq, ret);
+		goto out_unregister_map;
+	}
+
+	if (adc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+			sizeof(adc->wakeup1_data));
+		adc->wakeup1_enable = true;
+		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
+		ret = request_threaded_irq(adc->irq_auto_0, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-0", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto0 irq %d failed: %dn",
+				adc->irq_auto_0, ret);
+			goto out_irq_free;
+		}
+	}
+
+	if (adc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+				sizeof(adc->wakeup2_data));
+		adc->wakeup2_enable = true;
+		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
+		ret = request_threaded_irq(adc->irq_auto_1, NULL,
+				palmas_gpadc_irq_auto,
+				IRQF_ONESHOT | IRQF_EARLY_RESUME,
+				"palmas-adc-auto-1", adc);
+		if (ret < 0) {
+			dev_err(adc->dev, "request auto1 irq %d failed: %dn",
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 516

Quoted 5 lines
drivers/iio/adc/palmas_gpadc.c new file +5 -0
+				adc->irq_auto_1, ret);
+			goto out_irq_auto0_free;
+		}
+	}
+

I've requested kernel-doc above for ch0_current, but if that doesn't make it clear what matic is going on here then some comments here would be good.

I think the best location to make that clear is the DT bindings where you can set the values (in microamperes and not encoded values).

Basically you specify in the adc_pdata the uA you want and this function does a quantization to perpare 2 bits that can be written into the control registers later on.

Quoted 35 lines
Quoted 35 lines
drivers/iio/adc/palmas_gpadc.c new file +35 -0
+	if (adc_pdata->ch0_current == 0)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
+	else if (adc_pdata->ch0_current <= 5)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
+	else if (adc_pdata->ch0_current <= 15)
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
+	else
+		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
+
+	if (adc_pdata->ch3_current == 0)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
+	else if (adc_pdata->ch3_current <= 10)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
+	else if (adc_pdata->ch3_current <= 400)
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
+	else
+		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
+
+	adc->extended_delay = adc_pdata->extended_delay;
+
+	iodev->name = MOD_NAME;
+	iodev->dev.parent = &pdev->dev;
+	iodev->info = &palmas_gpadc_iio_info;
+	iodev->modes = INDIO_DIRECT_MODE;
+	iodev->channels = palmas_gpadc_iio_channel;
+	iodev->num_channels = ARRAY_SIZE(palmas_gpadc_iio_channel);
+
+	ret = iio_device_register(iodev);
+	if (ret < 0) {
+		dev_err(adc->dev, "iio_device_register() failed: %d\n", ret);
+		goto out_irq_auto1_free;
+	}
+
+	device_set_wakeup_capable(&pdev->dev, 1);
+	for (i = 0; i < PALMAS_ADC_CH_MAX; i++) {
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 557

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+		if (!(adc->adc_info[i].is_correct_code))
+			palmas_gpadc_calibrate(adc, i);
+	}
+
+	if (adc->wakeup1_enable || adc->wakeup2_enable)
+		device_wakeup_enable(&pdev->dev);

There is no match for this in the remove... Should there be one? (not an interface I know anything about!)

Same for me and I have no idea how to test. But it looks reasonable to call device_wakeup_disable(). Therefore I have added.

Quoted 29 lines
Quoted 29 lines
drivers/iio/adc/palmas_gpadc.c new file +29 -0
+
+	return 0;
+
+out_irq_auto1_free:
+	if (adc_pdata->adc_wakeup2_data)
+		free_irq(adc->irq_auto_1, adc);
+out_irq_auto0_free:
+	if (adc_pdata->adc_wakeup1_data)
+		free_irq(adc->irq_auto_0, adc);
+out_irq_free:
+	free_irq(adc->irq, adc);
+out_unregister_map:
+	if (adc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+out:
+	iio_device_free(iodev);
+	return ret;
+}
+
+static int palmas_gpadc_remove(struct platform_device *pdev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(&pdev->dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	struct palmas_platform_data *pdata = dev_get_platdata(pdev->dev.parent);
+
+	if (pdata->gpadc_pdata->iio_maps)
+		iio_map_array_unregister(iodev);
+	iio_device_unregister(iodev);
+	free_irq(adc->irq, adc);
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 591

Quoted 5 lines
drivers/iio/adc/palmas_gpadc.c new file +5 -0
+	if (adc->wakeup1_enable)
+		free_irq(adc->irq_auto_0, adc);
+	if (adc->wakeup2_enable)
+		free_irq(adc->irq_auto_1, adc);
+	iio_device_free(iodev);

Could use demv_iio_device_alloc and not need the free here or on error path. Given location of irq frees you could do devm allocations of them as well which would cut out a fair bit of code without reordering anything.

has also been proposed by Peter Meerwald and has been reworked.

Quoted 38 lines

nit: blank line here.

Quoted 4 lines
drivers/iio/adc/palmas_gpadc.c new file +4 -0
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP

drivers/iio/adc/palmas_gpadc.c line 605

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+static int palmas_adc_wakeup_configure(struct palmas_gpadc *adc)
+{
+	int adc_period, conv;
+	int i;
+	int ch0 = 0, ch1 = 0;
+	int thres;
+	int ret;
+
+	adc_period = adc->auto_conversion_period;
+	for (i = 0; i < 16; ++i) {

spacing around the /

Quoted 21 lines
drivers/iio/adc/palmas_gpadc.c new file +21 -0
+		if (((1000 * (1 << i))/32) < adc_period)
+			continue;
+	}
+	if (i > 0)
+		i--;
+	adc_period = i;
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_COUNTER_CONV_MASK,
+			adc_period);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}
+
+	conv = 0;
+	if (adc->wakeup1_enable) {
+		int is_high;
+
+		ch0 = adc->wakeup1_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN;
Quoted 8 lines

drivers/iio/adc/palmas_gpadc.c line 632

Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c new file +6 -0
+		if (adc->wakeup1_data.adc_high_threshold > 0) {
+			thres = adc->wakeup1_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup1_data.adc_low_threshold;
+			is_high = BIT(7);

BIT(7) is a bit random, so I'd suggest defining an appropriate macro for it this register field.

There is indeed a proper macro in palmas.h

Quoted 219 lines
Quoted 57 lines
drivers/iio/adc/palmas_gpadc.c new file +56 -0
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV0_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV0_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	if (adc->wakeup2_enable) {
+		int is_high;
+
+		ch1 = adc->wakeup2_data.adc_channel_number;
+		conv |= PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN;
+		if (adc->wakeup2_data.adc_high_threshold > 0) {
+			thres = adc->wakeup2_data.adc_high_threshold;
+			is_high = 0;
+		} else {
+			thres = adc->wakeup2_data.adc_low_threshold;
+			is_high = BIT(7);
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_LSB, thres & 0xFF);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_LSB write failed: %d\n", ret);
+			return ret;
+		}
+
+		ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+				PALMAS_GPADC_THRES_CONV1_MSB,
+				((thres >> 8) & 0xF) | is_high);
+		if (ret < 0) {
+			dev_err(adc->dev,
+				"THRES_CONV1_MSB write failed: %d\n", ret);
+			return ret;
+		}
+	}
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, (ch1 << 4) | ch0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;

drivers/iio/adc/palmas_gpadc.c line 698

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+	}
+
+	ret = palmas_update_bits(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_CTRL,
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV1_EN |
+			PALMAS_GPADC_AUTO_CTRL_AUTO_CONV0_EN, conv);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_CTRL write failed: %d\n", ret);
+		return ret;
+	}

nitpick. Blank line here please.

Quoted 31 lines
drivers/iio/adc/palmas_gpadc.c new file +31 -0
+	return 0;
+}
+
+static int palmas_adc_wakeup_reset(struct palmas_gpadc *adc)
+{
+	int ret;
+
+	ret = palmas_write(adc->palmas, PALMAS_GPADC_BASE,
+			PALMAS_GPADC_AUTO_SELECT, 0);
+	if (ret < 0) {
+		dev_err(adc->dev, "AUTO_SELECT write failed: %d\n", ret);
+		return ret;
+	}
+
+	ret = palmas_disable_auto_conversion(adc);
+	if (ret < 0) {
+		dev_err(adc->dev, "Disable auto conversion failed: %d\n", ret);
+		return ret;
+	}
+	return 0;
+}
+
+static int palmas_gpadc_suspend(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;

drivers/iio/adc/palmas_gpadc.c line 739

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c new file +10 -0
+
+	ret = palmas_adc_wakeup_configure(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		enable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		enable_irq_wake(adc->irq_auto_1);

nitpick. Blank line here please.

Quoted 72 lines
drivers/iio/adc/palmas_gpadc.c new file +58 -0
+	return 0;
+}
+
+static int palmas_gpadc_resume(struct device *dev)
+{
+	struct iio_dev *iodev = dev_to_iio_dev(dev);
+	struct palmas_gpadc *adc = iio_priv(iodev);
+	int wakeup = adc->wakeup1_enable || adc->wakeup2_enable;
+	int ret;
+
+	if (!device_may_wakeup(dev) || !wakeup)
+		return 0;
+
+	ret = palmas_adc_wakeup_reset(adc);
+	if (ret < 0)
+		return ret;
+
+	if (adc->wakeup1_enable)
+		disable_irq_wake(adc->irq_auto_0);
+
+	if (adc->wakeup2_enable)
+		disable_irq_wake(adc->irq_auto_1);
+
+	return 0;
+};
+#endif
+
+static const struct dev_pm_ops palmas_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(palmas_gpadc_suspend,
+				palmas_gpadc_resume)
+};
+
+static struct platform_driver palmas_gpadc_driver = {
+	.probe = palmas_gpadc_probe,
+	.remove = palmas_gpadc_remove,
+	.driver = {
+		.name = MOD_NAME,
+		.owner = THIS_MODULE,
+		.pm = &palmas_pm_ops,
+	},
+};
+
+static int __init palmas_gpadc_init(void)
+{
+	return platform_driver_register(&palmas_gpadc_driver);
+}
+module_init(palmas_gpadc_init);
+
+static void __exit palmas_gpadc_exit(void)
+{
+	platform_driver_unregister(&palmas_gpadc_driver);
+}
+module_exit(palmas_gpadc_exit);
+
+MODULE_DESCRIPTION("palmas GPADC driver");
+MODULE_AUTHOR("Pradeep Goudagunta<[email protected]>");
+MODULE_ALIAS("platform:palmas-gpadc");
+MODULE_LICENSE("GPL v2");
include/linux/mfd/palmas.h +6 -0
index bb270bd..60acfa2 100644
@@ -133,21 +133,32 @@ struct palmas_pmic_driver_data {
 			    struct regulator_config config);
 };
 
+struct palmas_adc_wakeup_property {
+	int adc_channel_number;
+	int adc_high_threshold;
+	int adc_low_threshold;
+};
+

include/linux/mfd/palmas.h struct palmas_pmic_driver_data { · line 148

Quoted 10 lines
include/linux/mfd/palmas.h +2 -3
 struct palmas_gpadc_platform_data {
 	/* Channel 3 current source is only enabled during conversion */
 	int ch3_current;
-
 	/* Channel 0 current source can be used for battery detection.
-	 * If used for battery detection this will cause a permanent current
+	* If used for battery detection this will cause a permanent current
 	 * consumption depending on current level set here.
-	 */
+		 */

Some oddness occured here in patch creation. Do check your own patches as we all end up with stuff like this from time to time that should be picked up on beofre sending out. It just adds noise.

Quoted 24 lines
include/linux/mfd/palmas.h +7 -1
 	int ch0_current;
+	bool extended_delay;
 
 	/* default BAT_REMOVAL_DAT setting on device probe */
 	int bat_removal;
 
 	/* Sets the START_POLARITY bit in the RT_CTRL register */
 	int start_polarity;
+
+	struct iio_map *iio_maps;
+	int auto_conversion_period_ms;
+	struct palmas_adc_wakeup_property *adc_wakeup1_data;
+	struct palmas_adc_wakeup_property *adc_wakeup2_data;
 };
 
 struct palmas_reg_init {
@@ -403,7 +414,7 @@ struct palmas_gpadc_calibration {
 	s32 gain_error;
 	s32 offset_error;
 };
-
+/*
 struct palmas_gpadc {
 	struct device *dev;
Quoted 8 lines

include/linux/mfd/palmas.h struct palmas_gpadc { · line 440

Quoted 6 lines
include/linux/mfd/palmas.h +1 -0
 	struct palmas *palmas;
@@ -426,6 +437,9 @@ struct palmas_gpadc {
 	int conv1_channel;
 	int rt_channel;
 };
+*/

So there's a commented out bit of code in this patch? Sort that out before the next version.

We forgot to delete...

Quoted 22 lines
Quoted 22 lines
include/linux/mfd/palmas.h +14 -0
+
+#define PALMAS_DATASHEET_NAME(_name)	"palmas-gpadc-chan-"#_name
 
 struct palmas_gpadc_result {
 	s32 raw_code;
@@ -519,6 +533,42 @@ enum palmas_irqs {
 	PALMAS_NUM_IRQ,
 };
 
+/* Palma GPADC Channels */
+enum {
+	PALMAS_ADC_CH_IN0,
+	PALMAS_ADC_CH_IN1,
+	PALMAS_ADC_CH_IN2,
+	PALMAS_ADC_CH_IN3,
+	PALMAS_ADC_CH_IN4,
+	PALMAS_ADC_CH_IN5,
+	PALMAS_ADC_CH_IN6,
+	PALMAS_ADC_CH_IN7,
+	PALMAS_ADC_CH_IN8,
+	PALMAS_ADC_CH_IN9,
Quoted 8 lines

include/linux/mfd/palmas.h enum palmas_irqs { · line 553

Quoted 6 lines
include/linux/mfd/palmas.h +6 -0
+	PALMAS_ADC_CH_IN10,
+	PALMAS_ADC_CH_IN11,
+	PALMAS_ADC_CH_IN12,
+	PALMAS_ADC_CH_IN13,
+	PALMAS_ADC_CH_IN14,
+	PALMAS_ADC_CH_IN15,

This does rather feel like an enum that doesn't add anything as the enum values = the index given at the end anyway...

As said it adds some (questionable) check by the compiler,

Quoted 1 line

include/linux/mfd/palmas.h enum palmas_irqs { · line 554

Quoted 1 line
include/linux/mfd/palmas.h +1 -0
+	PALMAS_ADC_CH_MAX,

and this constant is indirectly defined.

Quoted 31 lines

include/linux/mfd/palmas.h enum palmas_irqs { · line 563

Quoted 9 lines
include/linux/mfd/palmas.h +9 -0
+};
+
+/* Palma GPADC Channel0 Current Source */
+enum {
+	PALMAS_ADC_CH0_CURRENT_SRC_0,
+	PALMAS_ADC_CH0_CURRENT_SRC_5,
+	PALMAS_ADC_CH0_CURRENT_SRC_15,
+	PALMAS_ADC_CH0_CURRENT_SRC_20,
+};

Nitpick: new line here.

Quoted 20 lines
include/linux/mfd/palmas.h +9 -0
+/* Palma GPADC Channel3 Current Source */
+enum {
+	PALMAS_ADC_CH3_CURRENT_SRC_0,
+	PALMAS_ADC_CH3_CURRENT_SRC_10,
+	PALMAS_ADC_CH3_CURRENT_SRC_400,
+	PALMAS_ADC_CH3_CURRENT_SRC_800,
+};
+
 struct palmas_pmic {
 	struct palmas *palmas;
 	struct device *dev;
@@ -2999,6 +3049,7 @@ enum usb_irq_events {
 #define PALMAS_GPADC_TRIM14					0x0D
 #define PALMAS_GPADC_TRIM15					0x0E
 #define PALMAS_GPADC_TRIM16					0x0F
+#define PALMAS_GPADC_TRIMINVALID				-1
 
 /* TPS659038 regen2_ctrl offset iss different from palmas */
 #define TPS659038_REGEN2_CTRL					0x12
#34 lore raw

PATCH 2/3 iio:adc:palmas: add DT support

In reply to a message not in the archive

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]

---
.../devicetree/bindings/iio/adc/palmas-gpadc.txt | 67 +++++++++++
drivers/iio/adc/palmas_gpadc.c | 130 +++++++++++++++++----

2 files changed, 175 insertions(+), 22 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +67 -0
new file mode 100644
index 0000000..a5a33ba
@@ -0,0 +1,67 @@
+* Palmas general purpose ADC IP block devicetree bindings
+
+Channels list:
+	0 battery type
+        1 battery temp NTC
+	2 GP
+	3 temp (with ext. diode)
+	4 GP
+	5 GP
+	6 VBAT_SENSE
+	7 VCC_SENSE
+	8 Backup Battery voltage
+	9 external charger (VCHG)
+	10 VBUS
+	11 DC-DC current probe (how does this work?)
+	12 internal die temp
+	13 internal die temp
+	14 USB ID pin voltage
+	15 test network
+
+Required properties:
+- compatible : Must be "ti,palmas-gpadc".
+
+Optional sub-nodes:
+ti,channel0-current-microamp: Channel 0 current in uA.
+	Valid values 0uA, 5uA, 15uA, 20uA.
+ti,channel3-current-microamp: Channel 3 current in uA.
+	Valid value 0uA, 10uA, 400uA, 800uA.
+ti,enable-channel3-dual-current: Enable dual current on channel 3.
+ti,enable-extended-delay: Enable extended delay.
+
+Optional sub-node:
+The Palmas ADC node has optional subnode to define the iio mapping.
+It is the name with "iio_map". This node has again subnode to define
+the property of the channel. The sub subnode has following properties:
+- ti,adc-channel-number: ADC channel number.
+- ti,adc-consumer-device: Consumer device name.
+- ti,adc-consumer-channel: ADC consumer channel name.
+
+Example:
+
+pmic {
+	compatible = "ti,twl6035-pmic", "ti,palmas-pmic";
+	...
+	gpadc {
+		compatible = "ti,palmas-gpadc";
+		interrupts = <18 0
+			      16 0
+			      17 0>;
+		ti,channel0-current-microamp = <5>;
+		ti,channel3-current-microamp = <10>;
+		iio_map {
+			ch1 {
+				ti,adc-channel-number = <1>;
+				ti,adc-consumer-device = "generic-adc-thermal.0";
+				ti,adc-consumer-channel ="battery-temp-channel";
+			};
+
+			ch6 {
+				ti,adc-channel-number = <6>;
+				ti,adc-consumer-device = "palmas-battery";
+				ti,adc-consumer-channel ="vbat_channel";
+			};
+		};
+	};
+	...
+};
drivers/iio/adc/palmas_gpadc.c +0 -0
index 17abb28..bc4db43 100644
@@ -20,6 +20,8 @@

#include <linux/pm.h>
#include <linux/mfd/palmas.h>
#include <linux/completion.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
#include <linux/iio/iio.h>
#include <linux/iio/machine.h>
#include <linux/iio/driver.h>

@@ -434,20 +436,97 @@ static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
};

+static int palmas_gpadc_get_adc_dt_data(struct platform_device *pdev,
+ struct palmas_gpadc_platform_data **gpadc_pdata)
+{
+ struct device_node *np = pdev->dev.of_node;
+ struct palmas_gpadc_platform_data *gp_data;
+ struct device_node *map_node;
+ struct device_node *child;
+ struct iio_map *palmas_iio_map;
+ int ret;
+ u32 pval;
+ int nmap, nvalid_map;
+
+ gp_data = devm_kzalloc(&pdev->dev, sizeof(*gp_data), GFP_KERNEL);
+ if (!gp_data)
+ return -ENOMEM;
+
+ ret = of_property_read_u32(np, "ti,channel0-current-microamp", &pval);
+ if (!ret)
+ gp_data->ch0_current = pval;
+
+ ret = of_property_read_u32(np, "ti,channel3-current-microamp", &pval);
+ if (!ret)
+ gp_data->ch3_current = pval;
+
+ gp_data->extended_delay = of_property_read_bool(np,
+ "ti,enable-extended-delay");
+
+ map_node = of_get_child_by_name(np, "iio_map");
+ if (!map_node) {
+ dev_warn(&pdev->dev, "IIO map table not found\n");
+ goto done;
+ }
+
+ nmap = of_get_child_count(map_node);
+ if (!nmap)
+ goto done;
+
+ nmap++;
+ palmas_iio_map = devm_kzalloc(&pdev->dev,
+ sizeof(*palmas_iio_map) * nmap, GFP_KERNEL);
+ if (!palmas_iio_map)
+ goto done;
+
+ nvalid_map = 0;
+ for_each_child_of_node(map_node, child) {
+ ret = of_property_read_u32(child, "ti,adc-channel-number",
+ &pval);
+ if (!ret && pval < ARRAY_SIZE(palmas_gpadc_iio_channel))
+ palmas_iio_map[nvalid_map].adc_channel_label =
+ palmas_gpadc_iio_channel[pval].datasheet_name;
+ of_property_read_string(child, "ti,adc-consumer-device",
+ &palmas_iio_map[nvalid_map].consumer_dev_name);
+ of_property_read_string(child, "ti,adc-consumer-channel",
+ &palmas_iio_map[nvalid_map].consumer_channel);
+ dev_dbg(&pdev->dev,
+ "Channel %s consumer dev %s and consumer channel %s\n",
+ palmas_iio_map[nvalid_map].adc_channel_label,
+ palmas_iio_map[nvalid_map].consumer_dev_name,
+ palmas_iio_map[nvalid_map].consumer_channel);
+ nvalid_map++;
+ }
+ palmas_iio_map[nvalid_map].adc_channel_label = NULL;
+ palmas_iio_map[nvalid_map].consumer_dev_name = NULL;
+ palmas_iio_map[nvalid_map].consumer_channel = NULL;
+
+ gp_data->iio_maps = palmas_iio_map;
+
+done:
+ *gpadc_pdata = gp_data;
+ return 0;
+}
+
static int palmas_gpadc_probe(struct platform_device *pdev)

{
	struct palmas_gpadc *adc;
	struct palmas_platform_data *pdata;
-	struct palmas_gpadc_platform_data *adc_pdata;
+	struct palmas_gpadc_platform_data *gpadc_pdata = NULL;
	struct iio_dev *iodev;
	int ret, i;

	pdata = dev_get_platdata(pdev->dev.parent);

- if (!pdata || !pdata->gpadc_pdata) {
- dev_err(&pdev->dev, "No platform data\n");
- return -ENODEV;
+ if (pdata && pdata->gpadc_pdata)
+ gpadc_pdata = pdata->gpadc_pdata;
+
+ if (!gpadc_pdata && pdev->dev.of_node) {
+ ret = palmas_gpadc_get_adc_dt_data(pdev, &gpadc_pdata);
+ if (ret < 0)

+			return ret;
	}

- adc_pdata = pdata->gpadc_pdata;
+ if (!gpadc_pdata)
+ return -EINVAL;

	iodev = iio_device_alloc(sizeof(*adc));
	if (!iodev) {
@@ -455,8 +534,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
		return -ENOMEM;
	}

- if (adc_pdata->iio_maps) {
- ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+ if (gpadc_pdata->iio_maps) {

+		ret = iio_map_array_register(iodev, gpadc_pdata->iio_maps);
		if (ret < 0) {
			dev_err(&pdev->dev, "iio_map_array_register failed\n");
			goto out;
@@ -470,7 +549,7 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
	init_completion(&adc->conv_completion);
	dev_set_drvdata(&pdev->dev, iodev);

- adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;

+	adc->auto_conversion_period = gpadc_pdata->auto_conversion_period_ms;
	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
	ret = request_threaded_irq(adc->irq, NULL,
		palmas_gpadc_irq,
@@ -482,8 +561,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
		goto out_unregister_map;
	}

- if (adc_pdata->adc_wakeup1_data) {
- memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+ if (gpadc_pdata->adc_wakeup1_data) {
+ memcpy(&adc->wakeup1_data, gpadc_pdata->adc_wakeup1_data,

			sizeof(adc->wakeup1_data));
		adc->wakeup1_enable = true;
		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
@@ -498,8 +577,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
		}
	}

- if (adc_pdata->adc_wakeup2_data) {
- memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+ if (gpadc_pdata->adc_wakeup2_data) {
+ memcpy(&adc->wakeup2_data, gpadc_pdata->adc_wakeup2_data,

				sizeof(adc->wakeup2_data));
		adc->wakeup2_enable = true;
		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
@@ -514,25 +593,25 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
		}
	}

- if (adc_pdata->ch0_current == 0)
+ if (gpadc_pdata->ch0_current == 0)

		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
-	else if (adc_pdata->ch0_current <= 5)
+	else if (gpadc_pdata->ch0_current <= 5)
		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
-	else if (adc_pdata->ch0_current <= 15)
+	else if (gpadc_pdata->ch0_current <= 15)
		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
	else
		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;

- if (adc_pdata->ch3_current == 0)
+ if (gpadc_pdata->ch3_current == 0)

		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
-	else if (adc_pdata->ch3_current <= 10)
+	else if (gpadc_pdata->ch3_current <= 10)
		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
-	else if (adc_pdata->ch3_current <= 400)
+	else if (gpadc_pdata->ch3_current <= 400)
		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
	else
		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;

- adc->extended_delay = adc_pdata->extended_delay;
+ adc->extended_delay = gpadc_pdata->extended_delay;

	iodev->name = MOD_NAME;
	iodev->dev.parent = &pdev->dev;
@@ -559,15 +638,15 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
	return 0;

out_irq_auto1_free:
- if (adc_pdata->adc_wakeup2_data)
+ if (gpadc_pdata->adc_wakeup2_data)

		free_irq(adc->irq_auto_1, adc);
out_irq_auto0_free:
-	if (adc_pdata->adc_wakeup1_data)
+	if (gpadc_pdata->adc_wakeup1_data)
		free_irq(adc->irq_auto_0, adc);
out_irq_free:
	free_irq(adc->irq, adc);
out_unregister_map:
-	if (adc_pdata->iio_maps)
+	if (gpadc_pdata->iio_maps)
		iio_map_array_unregister(iodev);
out:
	iio_device_free(iodev);

@@ -769,6 +848,12 @@ static const struct dev_pm_ops palmas_pm_ops = {
palmas_gpadc_resume)
};

+static const struct of_device_id of_palmas_gpadc_match_tbl[] = {
+ { .compatible = "ti,palmas-gpadc", },
+ { /* end */ }
+};
+MODULE_DEVICE_TABLE(of, of_palmas_gpadc_match_tbl);
+
static struct platform_driver palmas_gpadc_driver = {
.probe = palmas_gpadc_probe,
.remove = palmas_gpadc_remove,
@@ -776,6 +861,7 @@ static struct platform_driver palmas_gpadc_driver = {
.name = MOD_NAME,
.owner = THIS_MODULE,
.pm = &palmas_pm_ops,
+ .of_match_table = of_palmas_gpadc_match_tbl,

	},
};
Signature · 1 line
2.5.1
maintainer
#35 lore raw

Re: [PATCH 2/3] iio:adc:palmas: add DT support

Quoted 8 lines · H. Nikolaus Schaller

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]

Various commments inline. You've run into the region of previous arguments over driver bindings... Mark Rutland, one for you to comment on as I know you love our bindings ;) and yes we haven't done anything about them in two years or so since you last moaned about them to me :(

Quoted 12 lines

---
.../devicetree/bindings/iio/adc/palmas-gpadc.txt | 67 +++++++++++
drivers/iio/adc/palmas_gpadc.c | 130 +++++++++++++++++----

2 files changed, 175 insertions(+), 22 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +0 -0
new file mode 100644
index 0000000..a5a33ba

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 5

Quoted 5 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +5 -0
@@ -0,0 +1,67 @@
+* Palmas general purpose ADC IP block devicetree bindings
+
+Channels list:
+	0 battery type
+        1 battery temp NTC

tab vs space fun.

Quoted 48 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +48 -0
+	2 GP
+	3 temp (with ext. diode)
+	4 GP
+	5 GP
+	6 VBAT_SENSE
+	7 VCC_SENSE
+	8 Backup Battery voltage
+	9 external charger (VCHG)
+	10 VBUS
+	11 DC-DC current probe (how does this work?)
+	12 internal die temp
+	13 internal die temp
+	14 USB ID pin voltage
+	15 test network
+
+Required properties:
+- compatible : Must be "ti,palmas-gpadc".
+
+Optional sub-nodes:
+ti,channel0-current-microamp: Channel 0 current in uA.
+	Valid values 0uA, 5uA, 15uA, 20uA.
+ti,channel3-current-microamp: Channel 3 current in uA.
+	Valid value 0uA, 10uA, 400uA, 800uA.
+ti,enable-channel3-dual-current: Enable dual current on channel 3.
+ti,enable-extended-delay: Enable extended delay.
+
+Optional sub-node:
+The Palmas ADC node has optional subnode to define the iio mapping.
+It is the name with "iio_map". This node has again subnode to define
+the property of the channel. The sub subnode has following properties:
+- ti,adc-channel-number: ADC channel number.
+- ti,adc-consumer-device: Consumer device name.
+- ti,adc-consumer-channel: ADC consumer channel name.
+
+Example:
+
+pmic {
+	compatible = "ti,twl6035-pmic", "ti,palmas-pmic";
+	...
+	gpadc {
+		compatible = "ti,palmas-gpadc";
+		interrupts = <18 0
+			      16 0
+			      17 0>;
+		ti,channel0-current-microamp = <5>;
+		ti,channel3-current-microamp = <10>;
+		iio_map {
+			ch1 {

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 63

Quoted 10 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +10 -0
+				ti,adc-channel-number = <1>;
+				ti,adc-consumer-device = "generic-adc-thermal.0";
+				ti,adc-consumer-channel ="battery-temp-channel";
+			};
+
+			ch6 {
+				ti,adc-channel-number = <6>;
+				ti,adc-consumer-device = "palmas-battery";
+				ti,adc-consumer-channel ="vbat_channel";
+			};

See comments below. There is an existing iio-consumer binding. Various uses of it have caused complaints from the device tree guys in the past. I'd be interested to get their feedback on this use case.

Quoted 65 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +4 -0
+		};
+	};
+	...
+};
drivers/iio/adc/palmas_gpadc.c +46 -0
index 17abb28..bc4db43 100644
@@ -20,6 +20,8 @@
 #include <linux/pm.h>
 #include <linux/mfd/palmas.h>
 #include <linux/completion.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
 #include <linux/iio/iio.h>
 #include <linux/iio/machine.h>
 #include <linux/iio/driver.h>
@@ -434,20 +436,97 @@ static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
 	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
 };
 
+static int palmas_gpadc_get_adc_dt_data(struct platform_device *pdev,
+	struct palmas_gpadc_platform_data **gpadc_pdata)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct palmas_gpadc_platform_data *gp_data;
+	struct device_node *map_node;
+	struct device_node *child;
+	struct iio_map *palmas_iio_map;
+	int ret;
+	u32 pval;
+	int nmap, nvalid_map;
+
+	gp_data = devm_kzalloc(&pdev->dev, sizeof(*gp_data), GFP_KERNEL);
+	if (!gp_data)
+		return -ENOMEM;
+
+	ret = of_property_read_u32(np, "ti,channel0-current-microamp", &pval);
+	if (!ret)
+		gp_data->ch0_current = pval;
+
+	ret = of_property_read_u32(np, "ti,channel3-current-microamp", &pval);
+	if (!ret)
+		gp_data->ch3_current = pval;
+
+	gp_data->extended_delay = of_property_read_bool(np,
+					"ti,enable-extended-delay");
+
+	map_node = of_get_child_by_name(np, "iio_map");
+	if (!map_node) {
+		dev_warn(&pdev->dev, "IIO map table not found\n");
+		goto done;
+	}
+
+	nmap = of_get_child_count(map_node);
+	if (!nmap)
+		goto done;
+
+	nmap++;
+	palmas_iio_map = devm_kzalloc(&pdev->dev,
+				sizeof(*palmas_iio_map) * nmap, GFP_KERNEL);
+	if (!palmas_iio_map)
+		goto done;
+
+	nvalid_map = 0;

drivers/iio/adc/palmas_gpadc.c static const struct iio_chan_spec palmas_gpadc_iio_channel[] = { · line 492

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c +10 -0
+	for_each_child_of_node(map_node, child) {
+		ret = of_property_read_u32(child, "ti,adc-channel-number",
+					&pval);
+		if (!ret && pval < ARRAY_SIZE(palmas_gpadc_iio_channel))
+			palmas_iio_map[nvalid_map].adc_channel_label =
+				palmas_gpadc_iio_channel[pval].datasheet_name;
+		of_property_read_string(child, "ti,adc-consumer-device",
+				&palmas_iio_map[nvalid_map].consumer_dev_name);
+		of_property_read_string(child, "ti,adc-consumer-channel",
+				&palmas_iio_map[nvalid_map].consumer_channel);

Hmm. These aren't TI specific, and there are already IIO bindings documented for consumers at Documentation/devicetree/bindings/iio/iio-bindings.txt.

They have caused some issues in the past due to binding to devices that are linux specific convenience functions. I fear that is sort of the case here with your thermal binding, but we can see what response we get.

Mark Rutland is IIRC the most vocal on this :)

The current plan for iio-hwmon which caused all the controversy is to look at moving the decision to create the mapping entirely into userspace via configfs and hence make it a clear policy decision rather than claiming that changing user interface is a hardware mapping.

Quoted 5 lines
drivers/iio/adc/palmas_gpadc.c +5 -0
+		dev_dbg(&pdev->dev,
+			"Channel %s consumer dev %s and consumer channel %s\n",
+				palmas_iio_map[nvalid_map].adc_channel_label,
+				palmas_iio_map[nvalid_map].consumer_dev_name,
+				palmas_iio_map[nvalid_map].consumer_channel);

drivers/iio/adc/palmas_gpadc.c static const struct iio_chan_spec palmas_gpadc_iio_channel[] = { · line 507

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c +10 -0
+		nvalid_map++;
+	}
+	palmas_iio_map[nvalid_map].adc_channel_label = NULL;
+	palmas_iio_map[nvalid_map].consumer_dev_name = NULL;
+	palmas_iio_map[nvalid_map].consumer_channel = NULL;
+
+	gp_data->iio_maps = palmas_iio_map;
+
+done:
+	*gpadc_pdata = gp_data;

blank line here.

drivers/iio/adc/palmas_gpadc.c static const struct iio_chan_spec palmas_gpadc_iio_channel[] = { · line 515

Quoted 9 lines
drivers/iio/adc/palmas_gpadc.c +4 -1
+	return 0;
+}
+
 static int palmas_gpadc_probe(struct platform_device *pdev)
 {
 	struct palmas_gpadc *adc;
 	struct palmas_platform_data *pdata;
-	struct palmas_gpadc_platform_data *adc_pdata;
+	struct palmas_gpadc_platform_data *gpadc_pdata = NULL;

This rename should not be here. It adds a lot of churn to the patch and makes it harder to review. Fix it in the previous patch.

Quoted 136 lines
drivers/iio/adc/palmas_gpadc.c +23 -18
 	struct iio_dev *iodev;
 	int ret, i;
 
 	pdata = dev_get_platdata(pdev->dev.parent);
-	if (!pdata || !pdata->gpadc_pdata) {
-		dev_err(&pdev->dev, "No platform data\n");
-		return -ENODEV;
+	if (pdata && pdata->gpadc_pdata)
+		gpadc_pdata = pdata->gpadc_pdata;
+
+	if (!gpadc_pdata && pdev->dev.of_node) {
+		ret = palmas_gpadc_get_adc_dt_data(pdev, &gpadc_pdata);
+		if (ret < 0)
+			return ret;
 	}
-	adc_pdata = pdata->gpadc_pdata;
+	if (!gpadc_pdata)
+		return -EINVAL;
 
 	iodev = iio_device_alloc(sizeof(*adc));
 	if (!iodev) {
@@ -455,8 +534,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		return -ENOMEM;
 	}
 
-	if (adc_pdata->iio_maps) {
-		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+	if (gpadc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, gpadc_pdata->iio_maps);
 		if (ret < 0) {
 			dev_err(&pdev->dev, "iio_map_array_register failed\n");
 			goto out;
@@ -470,7 +549,7 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 	init_completion(&adc->conv_completion);
 	dev_set_drvdata(&pdev->dev, iodev);
 
-	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->auto_conversion_period = gpadc_pdata->auto_conversion_period_ms;
 	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
 	ret = request_threaded_irq(adc->irq, NULL,
 		palmas_gpadc_irq,
@@ -482,8 +561,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		goto out_unregister_map;
 	}
 
-	if (adc_pdata->adc_wakeup1_data) {
-		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+	if (gpadc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, gpadc_pdata->adc_wakeup1_data,
 			sizeof(adc->wakeup1_data));
 		adc->wakeup1_enable = true;
 		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
@@ -498,8 +577,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		}
 	}
 
-	if (adc_pdata->adc_wakeup2_data) {
-		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+	if (gpadc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, gpadc_pdata->adc_wakeup2_data,
 				sizeof(adc->wakeup2_data));
 		adc->wakeup2_enable = true;
 		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
@@ -514,25 +593,25 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		}
 	}
 
-	if (adc_pdata->ch0_current == 0)
+	if (gpadc_pdata->ch0_current == 0)
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
-	else if (adc_pdata->ch0_current <= 5)
+	else if (gpadc_pdata->ch0_current <= 5)
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
-	else if (adc_pdata->ch0_current <= 15)
+	else if (gpadc_pdata->ch0_current <= 15)
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
 	else
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
 
-	if (adc_pdata->ch3_current == 0)
+	if (gpadc_pdata->ch3_current == 0)
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
-	else if (adc_pdata->ch3_current <= 10)
+	else if (gpadc_pdata->ch3_current <= 10)
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
-	else if (adc_pdata->ch3_current <= 400)
+	else if (gpadc_pdata->ch3_current <= 400)
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
 	else
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
 
-	adc->extended_delay = adc_pdata->extended_delay;
+	adc->extended_delay = gpadc_pdata->extended_delay;
 
 	iodev->name = MOD_NAME;
 	iodev->dev.parent = &pdev->dev;
@@ -559,15 +638,15 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 	return 0;
 

out_irq_auto1_free:

drivers/iio/adc/palmas_gpadc.c +1 -1
-	if (adc_pdata->adc_wakeup2_data)
+	if (gpadc_pdata->adc_wakeup2_data)
		free_irq(adc->irq_auto_1, adc);

out_irq_auto0_free:

drivers/iio/adc/palmas_gpadc.c +1 -1
-	if (adc_pdata->adc_wakeup1_data)
+	if (gpadc_pdata->adc_wakeup1_data)
		free_irq(adc->irq_auto_0, adc);
out_irq_free:
	free_irq(adc->irq, adc);

out_unregister_map:

drivers/iio/adc/palmas_gpadc.c +1 -1
-	if (adc_pdata->iio_maps)
+	if (gpadc_pdata->iio_maps)
		iio_map_array_unregister(iodev);
out:
	iio_device_free(iodev);
drivers/iio/adc/palmas_gpadc.c +7 -0
@@ -769,6 +848,12 @@ static const struct dev_pm_ops palmas_pm_ops = {
 				palmas_gpadc_resume)
 };
 
+static const struct of_device_id of_palmas_gpadc_match_tbl[] = {
+	{ .compatible = "ti,palmas-gpadc", },
+	{ /* end */ }
+};
+MODULE_DEVICE_TABLE(of, of_palmas_gpadc_match_tbl);
+
 static struct platform_driver palmas_gpadc_driver = {
 	.probe = palmas_gpadc_probe,
 	.remove = palmas_gpadc_remove,
@@ -776,6 +861,7 @@ static struct platform_driver palmas_gpadc_driver = {
 		.name = MOD_NAME,
 		.owner = THIS_MODULE,
 		.pm = &palmas_pm_ops,
+		.of_match_table = of_palmas_gpadc_match_tbl,
 	},
 };
 
#36 lore raw

Re: [PATCH 2/3] iio:adc:palmas: add DT support

Am 27.09.2015 um 17:37 schrieb Jonathan Cameron <[email protected]>:

Quoted 85 lines · H. Nikolaus Schaller
Quoted 8 lines

From: Marek Belisko <[email protected]>

Code was found at:
https://android.googlesource.com/kernel/tegra/+/a90856a6626d502d42c6e7abccbdf9d730b36270%5E%21/#F1

[Fixed minor typos + add channels list to documentation]

Various commments inline. You've run into the region of previous arguments over driver bindings... Mark Rutland, one for you to comment on as I know you love our bindings ;) and yes we haven't done anything about them in two years or so since you last moaned about them to me :(

Quoted 12 lines

---
.../devicetree/bindings/iio/adc/palmas-gpadc.txt | 67 +++++++++++
drivers/iio/adc/palmas_gpadc.c | 130 +++++++++++++++++----

2 files changed, 175 insertions(+), 22 deletions(-)

create mode 100644 Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +0 -0
new file mode 100644
index 0000000..a5a33ba

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 5

Quoted 5 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +5 -0
@@ -0,0 +1,67 @@
+* Palmas general purpose ADC IP block devicetree bindings
+
+Channels list:
+	0 battery type
+        1 battery temp NTC

tab vs space fun.

Quoted 53 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +53 -0
+	2 GP
+	3 temp (with ext. diode)
+	4 GP
+	5 GP
+	6 VBAT_SENSE
+	7 VCC_SENSE
+	8 Backup Battery voltage
+	9 external charger (VCHG)
+	10 VBUS
+	11 DC-DC current probe (how does this work?)
+	12 internal die temp
+	13 internal die temp
+	14 USB ID pin voltage
+	15 test network
+
+Required properties:
+- compatible : Must be "ti,palmas-gpadc".
+
+Optional sub-nodes:
+ti,channel0-current-microamp: Channel 0 current in uA.
+	Valid values 0uA, 5uA, 15uA, 20uA.
+ti,channel3-current-microamp: Channel 3 current in uA.
+	Valid value 0uA, 10uA, 400uA, 800uA.
+ti,enable-channel3-dual-current: Enable dual current on channel 3.
+ti,enable-extended-delay: Enable extended delay.
+
+Optional sub-node:
+The Palmas ADC node has optional subnode to define the iio mapping.
+It is the name with "iio_map". This node has again subnode to define
+the property of the channel. The sub subnode has following properties:
+- ti,adc-channel-number: ADC channel number.
+- ti,adc-consumer-device: Consumer device name.
+- ti,adc-consumer-channel: ADC consumer channel name.
+
+Example:
+
+pmic {
+	compatible = "ti,twl6035-pmic", "ti,palmas-pmic";
+	...
+	gpadc {
+		compatible = "ti,palmas-gpadc";
+		interrupts = <18 0
+			      16 0
+			      17 0>;
+		ti,channel0-current-microamp = <5>;
+		ti,channel3-current-microamp = <10>;
+		iio_map {
+			ch1 {
+				ti,adc-channel-number = <1>;
+				ti,adc-consumer-device = "generic-adc-thermal.0";
+				ti,adc-consumer-channel ="battery-temp-channel";
+			};
+
Quoted 8 lines

Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt line 63

Quoted 5 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +5 -0
+			ch6 {
+				ti,adc-channel-number = <6>;
+				ti,adc-consumer-device = "palmas-battery";
+				ti,adc-consumer-channel ="vbat_channel";
+			};

See comments below. There is an existing iio-consumer binding. Various uses of it have caused complaints from the device tree guys in the past. I'd be interested to get their feedback on this use case.

We just were not aware what it really means. iiomapping was available through the pdata of the original driver, but we have no real use-case.

Quoted 82 lines
Quoted 65 lines
Documentation/devicetree/bindings/iio/adc/palmas-gpadc.txt new file +4 -0
+		};
+	};
+	...
+};
drivers/iio/adc/palmas_gpadc.c +46 -0
index 17abb28..bc4db43 100644
@@ -20,6 +20,8 @@
 #include <linux/pm.h>
 #include <linux/mfd/palmas.h>
 #include <linux/completion.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
 #include <linux/iio/iio.h>
 #include <linux/iio/machine.h>
 #include <linux/iio/driver.h>
@@ -434,20 +436,97 @@ static const struct iio_chan_spec palmas_gpadc_iio_channel[] = {
 	PALMAS_ADC_CHAN_IIO(IN15, IIO_VOLTAGE),
 };
 
+static int palmas_gpadc_get_adc_dt_data(struct platform_device *pdev,
+	struct palmas_gpadc_platform_data **gpadc_pdata)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct palmas_gpadc_platform_data *gp_data;
+	struct device_node *map_node;
+	struct device_node *child;
+	struct iio_map *palmas_iio_map;
+	int ret;
+	u32 pval;
+	int nmap, nvalid_map;
+
+	gp_data = devm_kzalloc(&pdev->dev, sizeof(*gp_data), GFP_KERNEL);
+	if (!gp_data)
+		return -ENOMEM;
+
+	ret = of_property_read_u32(np, "ti,channel0-current-microamp", &pval);
+	if (!ret)
+		gp_data->ch0_current = pval;
+
+	ret = of_property_read_u32(np, "ti,channel3-current-microamp", &pval);
+	if (!ret)
+		gp_data->ch3_current = pval;
+
+	gp_data->extended_delay = of_property_read_bool(np,
+					"ti,enable-extended-delay");
+
+	map_node = of_get_child_by_name(np, "iio_map");
+	if (!map_node) {
+		dev_warn(&pdev->dev, "IIO map table not found\n");
+		goto done;
+	}
+
+	nmap = of_get_child_count(map_node);
+	if (!nmap)
+		goto done;
+
+	nmap++;
+	palmas_iio_map = devm_kzalloc(&pdev->dev,
+				sizeof(*palmas_iio_map) * nmap, GFP_KERNEL);
+	if (!palmas_iio_map)
+		goto done;
+
+	nvalid_map = 0;

drivers/iio/adc/palmas_gpadc.c static const struct iio_chan_spec palmas_gpadc_iio_channel[] = { · line 492

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c +10 -0
+	for_each_child_of_node(map_node, child) {
+		ret = of_property_read_u32(child, "ti,adc-channel-number",
+					&pval);
+		if (!ret && pval < ARRAY_SIZE(palmas_gpadc_iio_channel))
+			palmas_iio_map[nvalid_map].adc_channel_label =
+				palmas_gpadc_iio_channel[pval].datasheet_name;
+		of_property_read_string(child, "ti,adc-consumer-device",
+				&palmas_iio_map[nvalid_map].consumer_dev_name);
+		of_property_read_string(child, "ti,adc-consumer-channel",
+				&palmas_iio_map[nvalid_map].consumer_channel);

Hmm. These aren't TI specific, and there are already IIO bindings documented for consumers at Documentation/devicetree/bindings/iio/iio-bindings.txt.

They have caused some issues in the past due to binding to devices that are linux specific convenience functions. I fear that is sort of the case here with your thermal binding, but we can see what response we get.

Quoted 6 lines

Mark Rutland is IIRC the most vocal on this :)

The current plan for iio-hwmon which caused all the controversy is to look at moving the decision to create the mapping entirely into userspace via configfs and hence make it a clear policy decision rather than claiming that changing user interface is a hardware mapping.

we have removed iio mapping completely. It appears that it is only used by 5 other drivers and none of the cousins of the Palmas (twl4030, twl6030, tps65950 etc.).

Quoted 17 lines
Quoted 6 lines
drivers/iio/adc/palmas_gpadc.c +5 -0
+		dev_dbg(&pdev->dev,
+			"Channel %s consumer dev %s and consumer channel %s\n",
+				palmas_iio_map[nvalid_map].adc_channel_label,
+				palmas_iio_map[nvalid_map].consumer_dev_name,
+				palmas_iio_map[nvalid_map].consumer_channel);

drivers/iio/adc/palmas_gpadc.c static const struct iio_chan_spec palmas_gpadc_iio_channel[] = { · line 507

Quoted 10 lines
drivers/iio/adc/palmas_gpadc.c +10 -0
+		nvalid_map++;
+	}
+	palmas_iio_map[nvalid_map].adc_channel_label = NULL;
+	palmas_iio_map[nvalid_map].consumer_dev_name = NULL;
+	palmas_iio_map[nvalid_map].consumer_channel = NULL;
+
+	gp_data->iio_maps = palmas_iio_map;
+
+done:
+	*gpadc_pdata = gp_data;

blank line here.

Quoted 11 lines

drivers/iio/adc/palmas_gpadc.c static const struct iio_chan_spec palmas_gpadc_iio_channel[] = { · line 515

Quoted 9 lines
drivers/iio/adc/palmas_gpadc.c +4 -1
+	return 0;
+}
+
 static int palmas_gpadc_probe(struct platform_device *pdev)
 {
 	struct palmas_gpadc *adc;
 	struct palmas_platform_data *pdata;
-	struct palmas_gpadc_platform_data *adc_pdata;
+	struct palmas_gpadc_platform_data *gpadc_pdata = NULL;

This rename should not be here. It adds a lot of churn to the patch and makes it harder to review. Fix it in the previous patch.

Well, the first patch was intended to stay as close to the original as possible, but here it is indeed very reasonable to start with renamed variables.

Quoted 138 lines
Quoted 137 lines
drivers/iio/adc/palmas_gpadc.c +23 -18
 	struct iio_dev *iodev;
 	int ret, i;
 
 	pdata = dev_get_platdata(pdev->dev.parent);
-	if (!pdata || !pdata->gpadc_pdata) {
-		dev_err(&pdev->dev, "No platform data\n");
-		return -ENODEV;
+	if (pdata && pdata->gpadc_pdata)
+		gpadc_pdata = pdata->gpadc_pdata;
+
+	if (!gpadc_pdata && pdev->dev.of_node) {
+		ret = palmas_gpadc_get_adc_dt_data(pdev, &gpadc_pdata);
+		if (ret < 0)
+			return ret;
 	}
-	adc_pdata = pdata->gpadc_pdata;
+	if (!gpadc_pdata)
+		return -EINVAL;
 
 	iodev = iio_device_alloc(sizeof(*adc));
 	if (!iodev) {
@@ -455,8 +534,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		return -ENOMEM;
 	}
 
-	if (adc_pdata->iio_maps) {
-		ret = iio_map_array_register(iodev, adc_pdata->iio_maps);
+	if (gpadc_pdata->iio_maps) {
+		ret = iio_map_array_register(iodev, gpadc_pdata->iio_maps);
 		if (ret < 0) {
 			dev_err(&pdev->dev, "iio_map_array_register failed\n");
 			goto out;
@@ -470,7 +549,7 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 	init_completion(&adc->conv_completion);
 	dev_set_drvdata(&pdev->dev, iodev);
 
-	adc->auto_conversion_period = adc_pdata->auto_conversion_period_ms;
+	adc->auto_conversion_period = gpadc_pdata->auto_conversion_period_ms;
 	adc->irq = palmas_irq_get_virq(adc->palmas, PALMAS_GPADC_EOC_SW_IRQ);
 	ret = request_threaded_irq(adc->irq, NULL,
 		palmas_gpadc_irq,
@@ -482,8 +561,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		goto out_unregister_map;
 	}
 
-	if (adc_pdata->adc_wakeup1_data) {
-		memcpy(&adc->wakeup1_data, adc_pdata->adc_wakeup1_data,
+	if (gpadc_pdata->adc_wakeup1_data) {
+		memcpy(&adc->wakeup1_data, gpadc_pdata->adc_wakeup1_data,
 			sizeof(adc->wakeup1_data));
 		adc->wakeup1_enable = true;
 		adc->irq_auto_0 =  platform_get_irq(pdev, 1);
@@ -498,8 +577,8 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		}
 	}
 
-	if (adc_pdata->adc_wakeup2_data) {
-		memcpy(&adc->wakeup2_data, adc_pdata->adc_wakeup2_data,
+	if (gpadc_pdata->adc_wakeup2_data) {
+		memcpy(&adc->wakeup2_data, gpadc_pdata->adc_wakeup2_data,
 				sizeof(adc->wakeup2_data));
 		adc->wakeup2_enable = true;
 		adc->irq_auto_1 =  platform_get_irq(pdev, 2);
@@ -514,25 +593,25 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 		}
 	}
 
-	if (adc_pdata->ch0_current == 0)
+	if (gpadc_pdata->ch0_current == 0)
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_0;
-	else if (adc_pdata->ch0_current <= 5)
+	else if (gpadc_pdata->ch0_current <= 5)
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_5;
-	else if (adc_pdata->ch0_current <= 15)
+	else if (gpadc_pdata->ch0_current <= 15)
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_15;
 	else
 		adc->ch0_current = PALMAS_ADC_CH0_CURRENT_SRC_20;
 
-	if (adc_pdata->ch3_current == 0)
+	if (gpadc_pdata->ch3_current == 0)
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_0;
-	else if (adc_pdata->ch3_current <= 10)
+	else if (gpadc_pdata->ch3_current <= 10)
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_10;
-	else if (adc_pdata->ch3_current <= 400)
+	else if (gpadc_pdata->ch3_current <= 400)
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_400;
 	else
 		adc->ch3_current = PALMAS_ADC_CH3_CURRENT_SRC_800;
 
-	adc->extended_delay = adc_pdata->extended_delay;
+	adc->extended_delay = gpadc_pdata->extended_delay;
 
 	iodev->name = MOD_NAME;
 	iodev->dev.parent = &pdev->dev;
@@ -559,15 +638,15 @@ static int palmas_gpadc_probe(struct platform_device *pdev)
 	return 0;
 

out_irq_auto1_free:

drivers/iio/adc/palmas_gpadc.c +1 -1
-	if (adc_pdata->adc_wakeup2_data)
+	if (gpadc_pdata->adc_wakeup2_data)
		free_irq(adc->irq_auto_1, adc);

out_irq_auto0_free:

drivers/iio/adc/palmas_gpadc.c +1 -1
-	if (adc_pdata->adc_wakeup1_data)
+	if (gpadc_pdata->adc_wakeup1_data)
		free_irq(adc->irq_auto_0, adc);
out_irq_free:
	free_irq(adc->irq, adc);

out_unregister_map:

drivers/iio/adc/palmas_gpadc.c +1 -1
-	if (adc_pdata->iio_maps)
+	if (gpadc_pdata->iio_maps)
		iio_map_array_unregister(iodev);
out:
	iio_device_free(iodev);
drivers/iio/adc/palmas_gpadc.c +7 -0
@@ -769,6 +848,12 @@ static const struct dev_pm_ops palmas_pm_ops = {
 				palmas_gpadc_resume)
 };
 
+static const struct of_device_id of_palmas_gpadc_match_tbl[] = {
+	{ .compatible = "ti,palmas-gpadc", },
+	{ /* end */ }
+};
+MODULE_DEVICE_TABLE(of, of_palmas_gpadc_match_tbl);
+
 static struct platform_driver palmas_gpadc_driver = {
 	.probe = palmas_gpadc_probe,
 	.remove = palmas_gpadc_remove,
@@ -776,6 +861,7 @@ static struct platform_driver palmas_gpadc_driver = {
 		.name = MOD_NAME,
 		.owner = THIS_MODULE,
 		.pm = &palmas_pm_ops,
+		.of_match_table = of_palmas_gpadc_match_tbl,
 	},
 };
 
#37 lore raw

PATCH 3/3 ARM: dts: omap5-uevm: enable iio gpadc for Palmas

In reply to a message not in the archive

---
arch/arm/boot/dts/omap5-uevm.dts | 22 ++++++++++++++++++++++

1 file changed, 22 insertions(+)
arch/arm/boot/dts/omap5-uevm.dts +0 -0
index 3b16e8f..0d4c8ff 100644
@@ -342,6 +342,28 @@
 
			ti,ldo6-vibrator;

+ gpadc {
+ compatible = "ti,palmas-gpadc";
+ interrupts = <18 0
+ 16 0
+ 17 0>;
+ ti,channel0-current-microamp = <5>;
+ ti,channel3-current-microamp = <10>;
+ iio_map {
+ ch1 {
+ ti,adc-channel-number = <1>;
+ ti,adc-consumer-device = "generic-adc-thermal.0";
+ ti,adc-consumer-channel ="battery-temp-channel";
+ };
+
+ ch6 {
+ ti,adc-channel-number = <6>;
+ ti,adc-consumer-device = "palmas-battery";
+ ti,adc-consumer-channel ="vbat_channel";
+ };
+ };
+ };
+

			regulators {
				smps123_reg: smps123 {
					/* VDD_OPP_MPU */
Signature · 1 line
2.5.1
#38 lore raw

Re: [PATCH 3/3] ARM: dts: omap5-uevm: enable iio gpadc for Palmas

* H. Nikolaus Schaller <[email protected]> [150923 05:53]:

Missing description? BTW, this I want to queue separately as I have patches in works to support omap5 variants with omap5-board-common.dtsi to avoid duplicating things as we get omap5 better supported for things like regulators.

Regards,

Tony

Quoted 41 lines

---
arch/arm/boot/dts/omap5-uevm.dts | 22 ++++++++++++++++++++++

1 file changed, 22 insertions(+)
arch/arm/boot/dts/omap5-uevm.dts +22 -0
index 3b16e8f..0d4c8ff 100644
@@ -342,6 +342,28 @@
 
 			ti,ldo6-vibrator;
 
+			gpadc {
+				compatible = "ti,palmas-gpadc";
+				interrupts = <18 0
+					      16 0
+					      17 0>;
+				ti,channel0-current-microamp = <5>;
+				ti,channel3-current-microamp = <10>;
+				iio_map {
+					ch1 {
+						ti,adc-channel-number = <1>;
+						ti,adc-consumer-device = "generic-adc-thermal.0";
+						ti,adc-consumer-channel ="battery-temp-channel";
+					};
+
+					ch6 {
+						ti,adc-channel-number = <6>;
+						ti,adc-consumer-device = "palmas-battery";
+						ti,adc-consumer-channel ="vbat_channel";
+					};
+				};
+			};
+
 			regulators {
 				smps123_reg: smps123 {
 					/* VDD_OPP_MPU */
Signature · 1 line
2.5.1
#39 lore raw

Re: [PATCH 3/3] ARM: dts: omap5-uevm: enable iio gpadc for Palmas

Hi,

Am 23.09.2015 um 18:56 schrieb Tony Lindgren <[email protected]>:

Quoted 3 lines · H. Nikolaus Schaller

Missing description?

well, description is “ARM: dts: omap5-uevm: enable iio gpadc for Palmas"

and I didn’t catch that the commit is otherwise empty. Sorry.

Quoted 4 lines

BTW, this I want to queue separately
as I have patches in works to support omap5 variants with omap5-board-common.dtsi to avoid duplicating things as we get omap5 better supported for things like regulators.

Yes, it should work for all Palmas based OMAP5 boards and a omap5-board-common.dtsi would be the right place (after it comes to existence).

Quoted 38 lines

Regards,

Tony

Quoted 33 lines

---
arch/arm/boot/dts/omap5-uevm.dts | 22 ++++++++++++++++++++++

1 file changed, 22 insertions(+)
arch/arm/boot/dts/omap5-uevm.dts +20 -0
index 3b16e8f..0d4c8ff 100644
@@ -342,6 +342,28 @@
 
 			ti,ldo6-vibrator;
 
+			gpadc {
+				compatible = "ti,palmas-gpadc";
+				interrupts = <18 0
+					      16 0
+					      17 0>;
+				ti,channel0-current-microamp = <5>;
+				ti,channel3-current-microamp = <10>;
+				iio_map {
+					ch1 {
+						ti,adc-channel-number = <1>;
+						ti,adc-consumer-device = "generic-adc-thermal.0";
+						ti,adc-consumer-channel ="battery-temp-channel";
+					};
+
+					ch6 {
+						ti,adc-channel-number = <6>;
+						ti,adc-consumer-device = "palmas-battery";
+						ti,adc-consumer-channel ="vbat_channel";
+					};
+				};
Quoted 8 lines
Quoted 8 lines
arch/arm/boot/dts/omap5-uevm.dts +2 -0
+			};
+
 			regulators {
 				smps123_reg: smps123 {
 					/* VDD_OPP_MPU */
Signature · 1 line
2.5.1

BR,
Nikolaus