linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers
@ 2017-06-09  7:07 Eugen Hristev
  2017-06-09  7:07 ` [PATCH v4 1/4] ARM: dts: at91: sama5d2_xplained: enable ADTRG pin Eugen Hristev
                   ` (4 more replies)
  0 siblings, 5 replies; 10+ messages in thread
From: Eugen Hristev @ 2017-06-09  7:07 UTC (permalink / raw)
  To: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	jic23
  Cc: eugen.hristev

This patch implements the hardware triggers support and buffer management
for sama5d2.
The DT modifications ( [PATCH v4 1/4] ARM: dts: at91: sama5d2_xplained:
enable ADTRG pin) are for demonstration purposes of the feature,
setting the pinctrl for the ADC hw trigger pin,should go through
at91 maintainers.
As discussed on the mailing list, since we have a single external trigger
namely the ADTRG, but three different possible edges that can trigger it,
will have a single trigger in the IIO subsystem and three possible edges
which can be selected from the device tree.
Thus I created new bindings in the device tree for having the possible
edge types for the ADTRG pin.
The documentation for the bindings is in the patch
[PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add adtrg
 trigger binding
The modifications in the DT are in the patch
[PATCH v4 3/4] ARM: dts: at91: sama5d2: add ADC hw trigger
The new property must be set with the values from interrupt edge types, as this
pin is effectively an internal interrupt.

  Changes in v4:
 - Modified bindings to be similar with interrupts
 - Used iio claim direct to make sure buffer is not enabled when someone might
   try to use a software triggered conversion

  Changes in v3:
 - No longer have all three possible triggers registered in the subsystem and
   available in the sysfs. Only registering one trigger, and the possible edge
   is being read from the device tree. In the device tree, any possible edge
   can be set. There are three already made nodes for each possible edge which
   can be selected.
 - Fixed a bug where software triggered conversion could be started even if
   the hardware trigger was enabled.
 - Preallocate enough space for the buffer, considering all the channels
   and timestamp



  Changes in v2:
 - Moved buffer allocation and freeing into the preenable and postdisable
   callbacks.
   We have a total of scan bytes that can vary a lot depending on each channel
   enabled at a certain point.
 - made the at91 trigger list part of state structure
 - made the iio trigger list preallocated in state structure
 - moved irq enabling/disabling into the try_reenable callback
 - on trigger disable must write disable registries as well
 - Modified trigger name length to 64


Eugen Hristev (4):
  ARM: dts: at91: sama5d2_xplained: enable ADTRG pin
  Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding
  ARM: dts: at91: sama5d2: add ADC hw trigger edge type
  iio: adc: at91-sama5d2_adc: add hw trigger and buffer support

 .../bindings/iio/adc/at91-sama5d2_adc.txt          |   6 +
 arch/arm/boot/dts/at91-sama5d2_xplained.dts        |  16 +-
 arch/arm/boot/dts/sama5d2.dtsi                     |   1 +
 drivers/iio/adc/at91-sama5d2_adc.c                 | 236 ++++++++++++++++++++-
 4 files changed, 253 insertions(+), 6 deletions(-)

-- 
2.7.4

^ permalink raw reply	[flat|nested] 10+ messages in thread

* [PATCH v4 1/4] ARM: dts: at91: sama5d2_xplained: enable ADTRG pin
  2017-06-09  7:07 [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Eugen Hristev
@ 2017-06-09  7:07 ` Eugen Hristev
  2017-06-09  7:07 ` [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding Eugen Hristev
                   ` (3 subsequent siblings)
  4 siblings, 0 replies; 10+ messages in thread
From: Eugen Hristev @ 2017-06-09  7:07 UTC (permalink / raw)
  To: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	jic23
  Cc: eugen.hristev

Enable pinctrl for ADTRG pin (PD31) for ADC hardware trigger support.

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
---
 arch/arm/boot/dts/at91-sama5d2_xplained.dts | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)

diff --git a/arch/arm/boot/dts/at91-sama5d2_xplained.dts b/arch/arm/boot/dts/at91-sama5d2_xplained.dts
index 0bef9e0..04754b1 100644
--- a/arch/arm/boot/dts/at91-sama5d2_xplained.dts
+++ b/arch/arm/boot/dts/at91-sama5d2_xplained.dts
@@ -303,7 +303,7 @@
 				vddana-supply = <&vdd_3v3_lp_reg>;
 				vref-supply = <&vdd_3v3_lp_reg>;
 				pinctrl-names = "default";
-				pinctrl-0 = <&pinctrl_adc_default>;
+				pinctrl-0 = <&pinctrl_adc_default &pinctrl_adtrg_default>;
 				status = "okay";
 			};
 
@@ -322,6 +322,20 @@
 					bias-disable;
 				};
 
+				/*
+				 * The ADTRG pin can work on any edge type.
+				 * In here it's being pulled up, so need to
+				 * connect it to ground to get an edge e.g.
+				 * Trigger can be configured on falling, rise
+				 * or any edge, and the pull-up can be changed
+				 * to pull-down or left floating according to
+				 * needs.
+				 */
+				pinctrl_adtrg_default: adtrg_default {
+					pinmux = <PIN_PD31__ADTRG>;
+					bias-pull-up;
+				};
+
 				pinctrl_charger_chglev: charger_chglev {
 					pinmux = <PIN_PA12__GPIO>;
 					bias-disable;
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding
  2017-06-09  7:07 [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Eugen Hristev
  2017-06-09  7:07 ` [PATCH v4 1/4] ARM: dts: at91: sama5d2_xplained: enable ADTRG pin Eugen Hristev
@ 2017-06-09  7:07 ` Eugen Hristev
  2017-06-11 11:24   ` Jonathan Cameron
  2017-06-14 15:30   ` Rob Herring
  2017-06-09  7:07 ` [PATCH v4 3/4] ARM: dts: at91: sama5d2: add ADC hw trigger edge type Eugen Hristev
                   ` (2 subsequent siblings)
  4 siblings, 2 replies; 10+ messages in thread
From: Eugen Hristev @ 2017-06-09  7:07 UTC (permalink / raw)
  To: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	jic23
  Cc: eugen.hristev

Add property for the edge type of the hardware trigger pin ADTRG

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
---
  Changes in v4:
 - Modified bindings to be similar with interrupts

 Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt b/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
index 3223684..552e7a8 100644
--- a/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
+++ b/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
@@ -11,6 +11,11 @@ Required properties:
   - atmel,min-sample-rate-hz: Minimum sampling rate, it depends on SoC.
   - atmel,max-sample-rate-hz: Maximum sampling rate, it depends on SoC.
   - atmel,startup-time-ms: Startup time expressed in ms, it depends on SoC.
+  - atmel,trigger-edge-type: One of possible edge types for the ADTRG hardware
+  trigger pin. When the specific edge type is detected, the conversion will
+  start. Possible values are rising, falling, or both.
+  This property uses the IRQ edge types values: IRQ_TYPE_EDGE_RISING ,
+  IRQ_TYPE_EDGE_FALLING or IRQ_TYPE_EDGE_BOTH
 
 Example:
 
@@ -25,4 +30,5 @@ adc: adc@fc030000 {
 	atmel,startup-time-ms = <4>;
 	vddana-supply = <&vdd_3v3_lp_reg>;
 	vref-supply = <&vdd_3v3_lp_reg>;
+	atmel,trigger-edge-type = <IRQ_TYPE_EDGE_BOTH>;
 }
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* [PATCH v4 3/4] ARM: dts: at91: sama5d2: add ADC hw trigger edge type
  2017-06-09  7:07 [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Eugen Hristev
  2017-06-09  7:07 ` [PATCH v4 1/4] ARM: dts: at91: sama5d2_xplained: enable ADTRG pin Eugen Hristev
  2017-06-09  7:07 ` [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding Eugen Hristev
@ 2017-06-09  7:07 ` Eugen Hristev
  2017-06-09  7:07 ` [PATCH v4 4/4] iio: adc: at91-sama5d2_adc: add hw trigger and buffer support Eugen Hristev
  2017-06-14  6:57 ` [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Ludovic Desroches
  4 siblings, 0 replies; 10+ messages in thread
From: Eugen Hristev @ 2017-06-09  7:07 UTC (permalink / raw)
  To: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	jic23
  Cc: eugen.hristev

Added ADTRG edge type property as interrupt edge type value

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
---
  Changes in v4:
 - Modified bindings to be similar with interrupts

 arch/arm/boot/dts/sama5d2.dtsi | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/boot/dts/sama5d2.dtsi b/arch/arm/boot/dts/sama5d2.dtsi
index 8067c71..483a387 100644
--- a/arch/arm/boot/dts/sama5d2.dtsi
+++ b/arch/arm/boot/dts/sama5d2.dtsi
@@ -1267,6 +1267,7 @@
 				atmel,min-sample-rate-hz = <200000>;
 				atmel,max-sample-rate-hz = <20000000>;
 				atmel,startup-time-ms = <4>;
+				atmel,trigger-edge-type = <IRQ_TYPE_EDGE_RISING>;
 				status = "disabled";
 			};
 
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* [PATCH v4 4/4] iio: adc: at91-sama5d2_adc: add hw trigger and buffer support
  2017-06-09  7:07 [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Eugen Hristev
                   ` (2 preceding siblings ...)
  2017-06-09  7:07 ` [PATCH v4 3/4] ARM: dts: at91: sama5d2: add ADC hw trigger edge type Eugen Hristev
@ 2017-06-09  7:07 ` Eugen Hristev
  2017-06-11 11:34   ` Jonathan Cameron
  2017-06-14  6:57 ` [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Ludovic Desroches
  4 siblings, 1 reply; 10+ messages in thread
From: Eugen Hristev @ 2017-06-09  7:07 UTC (permalink / raw)
  To: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	jic23
  Cc: eugen.hristev

Added support for the external hardware trigger on pin ADTRG,
integrated the three possible edges into the driver
and created buffer management for data retrieval

Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
---
  Changes in v4:
 - Used iio claim direct to make sure buffer is not enabled when someone might
   try to use a software triggered conversion

  Changes in v3:
 - No longer have all three possible triggers registered in the subsystem and
   available in the sysfs. Only registering one trigger, and the possible edge
   is being read from the device tree. In the device tree, any possible edge
   can be set. There are three already made nodes for each possible edge which
   can be selected.
 - Fixed a bug where software triggered conversion could be started even if
   the hardware trigger was enabled.
 - Preallocate enough space for the buffer, considering all the channels
   and timestamp

  Changes in v2:
 - Moved buffer allocation and freeing into the preenable and postdisable
   callbacks.
   We have a total of scan bytes that can vary a lot depending on each channel
   enabled at a certain point.
 - made the at91 trigger list part of state structure
 - made the iio trigger list preallocated in state structure
 - moved irq enabling/disabling into the try_reenable callback
 - on trigger disable must write disable registries as well

 drivers/iio/adc/at91-sama5d2_adc.c | 236 ++++++++++++++++++++++++++++++++++++-
 1 file changed, 231 insertions(+), 5 deletions(-)

diff --git a/drivers/iio/adc/at91-sama5d2_adc.c b/drivers/iio/adc/at91-sama5d2_adc.c
index e10dca3..0d1c475 100644
--- a/drivers/iio/adc/at91-sama5d2_adc.c
+++ b/drivers/iio/adc/at91-sama5d2_adc.c
@@ -25,6 +25,10 @@
 #include <linux/wait.h>
 #include <linux/iio/iio.h>
 #include <linux/iio/sysfs.h>
+#include <linux/iio/buffer.h>
+#include <linux/iio/trigger.h>
+#include <linux/iio/trigger_consumer.h>
+#include <linux/iio/triggered_buffer.h>
 #include <linux/regulator/consumer.h>
 
 /* Control Register */
@@ -132,6 +136,17 @@
 #define AT91_SAMA5D2_PRESSR	0xbc
 /* Trigger Register */
 #define AT91_SAMA5D2_TRGR	0xc0
+/* Mask for TRGMOD field of TRGR register */
+#define AT91_SAMA5D2_TRGR_TRGMOD_MASK GENMASK(2, 0)
+/* No trigger, only software trigger can start conversions */
+#define AT91_SAMA5D2_TRGR_TRGMOD_NO_TRIGGER 0
+/* Trigger Mode external trigger rising edge */
+#define AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_RISE 1
+/* Trigger Mode external trigger falling edge */
+#define AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_FALL 2
+/* Trigger Mode external trigger any edge */
+#define AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_ANY 3
+
 /* Correction Select Register */
 #define AT91_SAMA5D2_COSR	0xd0
 /* Correction Value Register */
@@ -145,14 +160,28 @@
 /* Version Register */
 #define AT91_SAMA5D2_VERSION	0xfc
 
+#define AT91_SAMA5D2_HW_TRIG_CNT 3
+#define AT91_SAMA5D2_SINGLE_CHAN_CNT 12
+#define AT91_SAMA5D2_DIFF_CHAN_CNT 6
+
+/* Preferred number of bytes to hold conversion from all channels
+ * plus the timestamp
+ */
+#define AT91_BUFFER_MAX_BYTES ((AT91_SAMA5D2_SINGLE_CHAN_CNT +		\
+				AT91_SAMA5D2_DIFF_CHAN_CNT) * 2 + 8)
+
+#define AT91_BUFFER_MAX_HWORDS (AT91_BUFFER_MAX_BYTES / 2)
+
 #define AT91_SAMA5D2_CHAN_SINGLE(num, addr)				\
 	{								\
 		.type = IIO_VOLTAGE,					\
 		.channel = num,						\
 		.address = addr,					\
+		.scan_index = num,					\
 		.scan_type = {						\
 			.sign = 'u',					\
 			.realbits = 12,					\
+			.storagebits = 16,				\
 		},							\
 		.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),		\
 		.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE),	\
@@ -168,9 +197,11 @@
 		.channel = num,						\
 		.channel2 = num2,					\
 		.address = addr,					\
+		.scan_index = num + AT91_SAMA5D2_SINGLE_CHAN_CNT,	\
 		.scan_type = {						\
 			.sign = 's',					\
 			.realbits = 12,					\
+			.storagebits = 16,				\
 		},							\
 		.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),		\
 		.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE),	\
@@ -188,6 +219,12 @@ struct at91_adc_soc_info {
 	unsigned			max_sample_rate;
 };
 
+struct at91_adc_trigger {
+	char				*name;
+	unsigned int			trgmod_value;
+	unsigned int			edge_type;
+};
+
 struct at91_adc_state {
 	void __iomem			*base;
 	int				irq;
@@ -195,11 +232,14 @@ struct at91_adc_state {
 	struct regulator		*reg;
 	struct regulator		*vref;
 	int				vref_uv;
+	struct iio_trigger		*trig;
+	const struct at91_adc_trigger	*selected_trig;
 	const struct iio_chan_spec	*chan;
 	bool				conversion_done;
 	u32				conversion_value;
 	struct at91_adc_soc_info	soc_info;
 	wait_queue_head_t		wq_data_available;
+	u16				buffer[AT91_BUFFER_MAX_HWORDS];
 	/*
 	 * lock to prevent concurrent 'single conversion' requests through
 	 * sysfs.
@@ -207,6 +247,24 @@ struct at91_adc_state {
 	struct mutex			lock;
 };
 
+static const struct at91_adc_trigger at91_adc_trigger_list[] = {
+	{
+		.name = "external_rising",
+		.trgmod_value = AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_RISE,
+		.edge_type = IRQ_TYPE_EDGE_RISING,
+	},
+	{
+		.name = "external_falling",
+		.trgmod_value = AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_FALL,
+		.edge_type = IRQ_TYPE_EDGE_FALLING,
+	},
+	{
+		.name = "external_any",
+		.trgmod_value = AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_ANY,
+		.edge_type = IRQ_TYPE_EDGE_BOTH,
+	},
+};
+
 static const struct iio_chan_spec at91_adc_channels[] = {
 	AT91_SAMA5D2_CHAN_SINGLE(0, 0x50),
 	AT91_SAMA5D2_CHAN_SINGLE(1, 0x54),
@@ -226,8 +284,128 @@ static const struct iio_chan_spec at91_adc_channels[] = {
 	AT91_SAMA5D2_CHAN_DIFF(6, 7, 0x68),
 	AT91_SAMA5D2_CHAN_DIFF(8, 9, 0x70),
 	AT91_SAMA5D2_CHAN_DIFF(10, 11, 0x78),
+	IIO_CHAN_SOFT_TIMESTAMP(AT91_SAMA5D2_SINGLE_CHAN_CNT
+				+ AT91_SAMA5D2_DIFF_CHAN_CNT + 1),
+};
+
+static int at91_adc_configure_trigger(struct iio_trigger *trig, bool state)
+{
+	struct iio_dev *indio = iio_trigger_get_drvdata(trig);
+	struct at91_adc_state *st = iio_priv(indio);
+	u32 status = at91_adc_readl(st, AT91_SAMA5D2_TRGR);
+	u8 bit;
+
+	/* clear TRGMOD */
+	status &= ~AT91_SAMA5D2_TRGR_TRGMOD_MASK;
+
+	if (state)
+		status |= st->selected_trig->trgmod_value;
+
+	/* set/unset hw trigger */
+	at91_adc_writel(st, AT91_SAMA5D2_TRGR, status);
+
+	for_each_set_bit(bit, indio->active_scan_mask, indio->num_channels) {
+		struct iio_chan_spec const *chan = indio->channels + bit;
+
+		if (state) {
+			at91_adc_writel(st, AT91_SAMA5D2_CHER,
+					BIT(chan->channel));
+			at91_adc_writel(st, AT91_SAMA5D2_IER,
+					BIT(chan->channel));
+		} else {
+			at91_adc_writel(st, AT91_SAMA5D2_IDR,
+					BIT(chan->channel));
+			at91_adc_writel(st, AT91_SAMA5D2_CHDR,
+					BIT(chan->channel));
+		}
+	}
+
+	return 0;
+}
+
+static int at91_adc_reenable_trigger(struct iio_trigger *trig)
+{
+	struct iio_dev *indio = iio_trigger_get_drvdata(trig);
+	struct at91_adc_state *st = iio_priv(indio);
+
+	enable_irq(st->irq);
+	return 0;
+}
+
+static const struct iio_trigger_ops at91_adc_trigger_ops = {
+	.owner = THIS_MODULE,
+	.set_trigger_state = &at91_adc_configure_trigger,
+	.try_reenable = &at91_adc_reenable_trigger,
 };
 
+static struct iio_trigger *at91_adc_allocate_trigger(struct iio_dev *indio,
+						     char *trigger_name)
+{
+	struct iio_trigger *trig;
+	int ret;
+
+	trig = devm_iio_trigger_alloc(&indio->dev, "%s-dev%d-%s", indio->name,
+				      indio->id, trigger_name);
+	if (!trig)
+		return NULL;
+
+	trig->dev.parent = indio->dev.parent;
+	iio_trigger_set_drvdata(trig, indio);
+	trig->ops = &at91_adc_trigger_ops;
+
+	ret = devm_iio_trigger_register(&indio->dev, trig);
+	if (ret)
+		return NULL;
+
+	return trig;
+}
+
+static int at91_adc_trigger_init(struct iio_dev *indio)
+{
+	struct at91_adc_state *st = iio_priv(indio);
+
+	st->trig = at91_adc_allocate_trigger(indio, st->selected_trig->name);
+	if (!st->trig) {
+		dev_err(&indio->dev,
+			"could not allocate trigger\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static irqreturn_t at91_adc_trigger_handler(int irq, void *p)
+{
+	struct iio_poll_func *pf = p;
+	struct iio_dev *indio = pf->indio_dev;
+	struct at91_adc_state *st = iio_priv(indio);
+	int i = 0;
+	u8 bit;
+
+	for_each_set_bit(bit, indio->active_scan_mask, indio->num_channels) {
+		struct iio_chan_spec const *chan = indio->channels + bit;
+
+		st->buffer[i] = at91_adc_readl(st, chan->address);
+		i++;
+	}
+
+	iio_push_to_buffers_with_timestamp(indio, st->buffer, pf->timestamp);
+
+	iio_trigger_notify_done(indio->trig);
+
+	/* Needed to ACK the DRDY interruption */
+	at91_adc_readl(st, AT91_SAMA5D2_LCDR);
+
+	return IRQ_HANDLED;
+}
+
+static int at91_adc_buffer_init(struct iio_dev *indio)
+{
+	return devm_iio_triggered_buffer_setup(&indio->dev, indio,
+			&iio_pollfunc_store_time,
+			&at91_adc_trigger_handler, NULL);
+}
+
 static unsigned at91_adc_startup_time(unsigned startup_time_min,
 				      unsigned adc_clk_khz)
 {
@@ -293,14 +471,18 @@ static irqreturn_t at91_adc_interrupt(int irq, void *private)
 	u32 status = at91_adc_readl(st, AT91_SAMA5D2_ISR);
 	u32 imr = at91_adc_readl(st, AT91_SAMA5D2_IMR);
 
-	if (status & imr) {
+	if (!(status & imr))
+		return IRQ_NONE;
+
+	if (iio_buffer_enabled(indio)) {
+		disable_irq_nosync(irq);
+		iio_trigger_poll(indio->trig);
+	} else {
 		st->conversion_value = at91_adc_readl(st, st->chan->address);
 		st->conversion_done = true;
 		wake_up_interruptible(&st->wq_data_available);
-		return IRQ_HANDLED;
 	}
-
-	return IRQ_NONE;
+	return IRQ_HANDLED;
 }
 
 static int at91_adc_read_raw(struct iio_dev *indio_dev,
@@ -313,6 +495,11 @@ static int at91_adc_read_raw(struct iio_dev *indio_dev,
 
 	switch (mask) {
 	case IIO_CHAN_INFO_RAW:
+		/* we cannot use software trigger if hw trigger enabled */
+		ret = iio_device_claim_direct_mode(indio_dev);
+		if (ret)
+			return ret;
+
 		mutex_lock(&st->lock);
 
 		st->chan = chan;
@@ -344,6 +531,8 @@ static int at91_adc_read_raw(struct iio_dev *indio_dev,
 		at91_adc_writel(st, AT91_SAMA5D2_CHDR, BIT(chan->channel));
 
 		mutex_unlock(&st->lock);
+
+		iio_device_release_direct_mode(indio_dev);
 		return ret;
 
 	case IIO_CHAN_INFO_SCALE:
@@ -391,7 +580,8 @@ static int at91_adc_probe(struct platform_device *pdev)
 	struct iio_dev *indio_dev;
 	struct at91_adc_state *st;
 	struct resource	*res;
-	int ret;
+	int ret, i;
+	u32 edge_type;
 
 	indio_dev = devm_iio_device_alloc(&pdev->dev, sizeof(*st));
 	if (!indio_dev)
@@ -432,6 +622,27 @@ static int at91_adc_probe(struct platform_device *pdev)
 		return ret;
 	}
 
+	ret = of_property_read_u32(pdev->dev.of_node,
+				   "atmel,trigger-edge-type", &edge_type);
+	if (ret) {
+		dev_err(&pdev->dev,
+			"invalid or missing value for atmel,trigger-edge-type\n");
+		return ret;
+	}
+
+	st->selected_trig = NULL;
+
+	for (i = 0; i < AT91_SAMA5D2_HW_TRIG_CNT; i++)
+		if (at91_adc_trigger_list[i].edge_type == edge_type) {
+			st->selected_trig = &at91_adc_trigger_list[i];
+			break;
+		}
+
+	if (!st->selected_trig) {
+		dev_err(&pdev->dev, "invalid external trigger edge value\n");
+		return -EINVAL;
+	}
+
 	init_waitqueue_head(&st->wq_data_available);
 	mutex_init(&st->lock);
 
@@ -499,10 +710,25 @@ static int at91_adc_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, indio_dev);
 
+	ret = at91_adc_buffer_init(indio_dev);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "couldn't initialize the buffer.\n");
+		goto per_clk_disable_unprepare;
+	}
+
+	ret = at91_adc_trigger_init(indio_dev);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "couldn't setup the triggers.\n");
+		goto per_clk_disable_unprepare;
+	}
+
 	ret = iio_device_register(indio_dev);
 	if (ret < 0)
 		goto per_clk_disable_unprepare;
 
+	dev_info(&pdev->dev, "setting up trigger as %s\n",
+		 st->selected_trig->name);
+
 	dev_info(&pdev->dev, "version: %x\n",
 		 readl_relaxed(st->base + AT91_SAMA5D2_VERSION));
 
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 10+ messages in thread

* Re: [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding
  2017-06-09  7:07 ` [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding Eugen Hristev
@ 2017-06-11 11:24   ` Jonathan Cameron
  2017-06-11 11:25     ` Jonathan Cameron
  2017-06-14 15:30   ` Rob Herring
  1 sibling, 1 reply; 10+ messages in thread
From: Jonathan Cameron @ 2017-06-11 11:24 UTC (permalink / raw)
  To: Eugen Hristev
  Cc: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	Rob Herring, Mark Rutland, devicetree

On Fri, 9 Jun 2017 10:07:17 +0300
Eugen Hristev <eugen.hristev@microchip.com> wrote:

> Add property for the edge type of the hardware trigger pin ADTRG
> 
> Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Hi Eugen,

CC the device tree maintainers and list for devicetree binding changes.

I've added them.   This one is unusual enough it will need and Ack from
that direction (and I prefer to get them even for the trivial patches
when Rob and Mark aren't too busy!)

Jonathan
> ---
>   Changes in v4:
>  - Modified bindings to be similar with interrupts
> 
>  Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt b/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
> index 3223684..552e7a8 100644
> --- a/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
> +++ b/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
> @@ -11,6 +11,11 @@ Required properties:
>    - atmel,min-sample-rate-hz: Minimum sampling rate, it depends on SoC.
>    - atmel,max-sample-rate-hz: Maximum sampling rate, it depends on SoC.
>    - atmel,startup-time-ms: Startup time expressed in ms, it depends on SoC.
> +  - atmel,trigger-edge-type: One of possible edge types for the ADTRG hardware
> +  trigger pin. When the specific edge type is detected, the conversion will
> +  start. Possible values are rising, falling, or both.
> +  This property uses the IRQ edge types values: IRQ_TYPE_EDGE_RISING ,
> +  IRQ_TYPE_EDGE_FALLING or IRQ_TYPE_EDGE_BOTH
>  
>  Example:
>  
> @@ -25,4 +30,5 @@ adc: adc@fc030000 {
>  	atmel,startup-time-ms = <4>;
>  	vddana-supply = <&vdd_3v3_lp_reg>;
>  	vref-supply = <&vdd_3v3_lp_reg>;
> +	atmel,trigger-edge-type = <IRQ_TYPE_EDGE_BOTH>;
>  }

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding
  2017-06-11 11:24   ` Jonathan Cameron
@ 2017-06-11 11:25     ` Jonathan Cameron
  0 siblings, 0 replies; 10+ messages in thread
From: Jonathan Cameron @ 2017-06-11 11:25 UTC (permalink / raw)
  To: Eugen Hristev
  Cc: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	Rob Herring, Mark Rutland

On Sun, 11 Jun 2017 12:24:13 +0100
Jonathan Cameron <jic23@kernel.org> wrote:

> On Fri, 9 Jun 2017 10:07:17 +0300
> Eugen Hristev <eugen.hristev@microchip.com> wrote:
> 
> > Add property for the edge type of the hardware trigger pin ADTRG
> > 
> > Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>  
> Hi Eugen,
> 
> CC the device tree maintainers and list for devicetree binding changes.
> 
> I've added them.   This one is unusual enough it will need and Ack from
> that direction (and I prefer to get them even for the trivial patches
> when Rob and Mark aren't too busy!)
Sorry, I missed that you had cc'd the list - just noticed it in the
next patch and thought that was odd!

Anyhow, good work - we'll just wait for them to have a chance to look
at it :)
> 
> Jonathan
> > ---
> >   Changes in v4:
> >  - Modified bindings to be similar with interrupts
> > 
> >  Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt | 6 ++++++
> >  1 file changed, 6 insertions(+)
> > 
> > diff --git a/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt b/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
> > index 3223684..552e7a8 100644
> > --- a/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
> > +++ b/Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt
> > @@ -11,6 +11,11 @@ Required properties:
> >    - atmel,min-sample-rate-hz: Minimum sampling rate, it depends on SoC.
> >    - atmel,max-sample-rate-hz: Maximum sampling rate, it depends on SoC.
> >    - atmel,startup-time-ms: Startup time expressed in ms, it depends on SoC.
> > +  - atmel,trigger-edge-type: One of possible edge types for the ADTRG hardware
> > +  trigger pin. When the specific edge type is detected, the conversion will
> > +  start. Possible values are rising, falling, or both.
> > +  This property uses the IRQ edge types values: IRQ_TYPE_EDGE_RISING ,
> > +  IRQ_TYPE_EDGE_FALLING or IRQ_TYPE_EDGE_BOTH
> >  
> >  Example:
> >  
> > @@ -25,4 +30,5 @@ adc: adc@fc030000 {
> >  	atmel,startup-time-ms = <4>;
> >  	vddana-supply = <&vdd_3v3_lp_reg>;
> >  	vref-supply = <&vdd_3v3_lp_reg>;
> > +	atmel,trigger-edge-type = <IRQ_TYPE_EDGE_BOTH>;
> >  }  
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-iio" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v4 4/4] iio: adc: at91-sama5d2_adc: add hw trigger and buffer support
  2017-06-09  7:07 ` [PATCH v4 4/4] iio: adc: at91-sama5d2_adc: add hw trigger and buffer support Eugen Hristev
@ 2017-06-11 11:34   ` Jonathan Cameron
  0 siblings, 0 replies; 10+ messages in thread
From: Jonathan Cameron @ 2017-06-11 11:34 UTC (permalink / raw)
  To: Eugen Hristev
  Cc: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches

On Fri, 9 Jun 2017 10:07:19 +0300
Eugen Hristev <eugen.hristev@microchip.com> wrote:

> Added support for the external hardware trigger on pin ADTRG,
> integrated the three possible edges into the driver
> and created buffer management for data retrieval
> 
> Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
Hi Eugen,

I like this patch.  A few really minor bits and pieces inline, most of which
I simply missed on previous revisions (sorry about that!)

Big question is whether people will like the devicetree binding.

Thanks,

Jonathan
> ---
>   Changes in v4:
>  - Used iio claim direct to make sure buffer is not enabled when someone might
>    try to use a software triggered conversion
> 
>   Changes in v3:
>  - No longer have all three possible triggers registered in the subsystem and
>    available in the sysfs. Only registering one trigger, and the possible edge
>    is being read from the device tree. In the device tree, any possible edge
>    can be set. There are three already made nodes for each possible edge which
>    can be selected.
>  - Fixed a bug where software triggered conversion could be started even if
>    the hardware trigger was enabled.
>  - Preallocate enough space for the buffer, considering all the channels
>    and timestamp
> 
>   Changes in v2:
>  - Moved buffer allocation and freeing into the preenable and postdisable
>    callbacks.
>    We have a total of scan bytes that can vary a lot depending on each channel
>    enabled at a certain point.
>  - made the at91 trigger list part of state structure
>  - made the iio trigger list preallocated in state structure
>  - moved irq enabling/disabling into the try_reenable callback
>  - on trigger disable must write disable registries as well
> 
>  drivers/iio/adc/at91-sama5d2_adc.c | 236 ++++++++++++++++++++++++++++++++++++-
>  1 file changed, 231 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/iio/adc/at91-sama5d2_adc.c b/drivers/iio/adc/at91-sama5d2_adc.c
> index e10dca3..0d1c475 100644
> --- a/drivers/iio/adc/at91-sama5d2_adc.c
> +++ b/drivers/iio/adc/at91-sama5d2_adc.c
> @@ -25,6 +25,10 @@
>  #include <linux/wait.h>
>  #include <linux/iio/iio.h>
>  #include <linux/iio/sysfs.h>
> +#include <linux/iio/buffer.h>
> +#include <linux/iio/trigger.h>
> +#include <linux/iio/trigger_consumer.h>
> +#include <linux/iio/triggered_buffer.h>
>  #include <linux/regulator/consumer.h>
>  
>  /* Control Register */
> @@ -132,6 +136,17 @@
>  #define AT91_SAMA5D2_PRESSR	0xbc
>  /* Trigger Register */
>  #define AT91_SAMA5D2_TRGR	0xc0
> +/* Mask for TRGMOD field of TRGR register */
> +#define AT91_SAMA5D2_TRGR_TRGMOD_MASK GENMASK(2, 0)
> +/* No trigger, only software trigger can start conversions */
> +#define AT91_SAMA5D2_TRGR_TRGMOD_NO_TRIGGER 0
> +/* Trigger Mode external trigger rising edge */
> +#define AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_RISE 1
> +/* Trigger Mode external trigger falling edge */
> +#define AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_FALL 2
> +/* Trigger Mode external trigger any edge */
> +#define AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_ANY 3
> +
>  /* Correction Select Register */
>  #define AT91_SAMA5D2_COSR	0xd0
>  /* Correction Value Register */
> @@ -145,14 +160,28 @@
>  /* Version Register */
>  #define AT91_SAMA5D2_VERSION	0xfc
>  
> +#define AT91_SAMA5D2_HW_TRIG_CNT 3
> +#define AT91_SAMA5D2_SINGLE_CHAN_CNT 12
> +#define AT91_SAMA5D2_DIFF_CHAN_CNT 6
> +
> +/* Preferred number of bytes to hold conversion from all channels
Nitpick of the day
/*
 * Preferred...
> + * plus the timestamp
> + */
> +#define AT91_BUFFER_MAX_BYTES ((AT91_SAMA5D2_SINGLE_CHAN_CNT +		\
> +				AT91_SAMA5D2_DIFF_CHAN_CNT) * 2 + 8)
> +
> +#define AT91_BUFFER_MAX_HWORDS (AT91_BUFFER_MAX_BYTES / 2)
> +
>  #define AT91_SAMA5D2_CHAN_SINGLE(num, addr)				\
>  	{								\
>  		.type = IIO_VOLTAGE,					\
>  		.channel = num,						\
>  		.address = addr,					\
> +		.scan_index = num,					\
>  		.scan_type = {						\
>  			.sign = 'u',					\
>  			.realbits = 12,					\
> +			.storagebits = 16,				\
>  		},							\
>  		.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),		\
>  		.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE),	\
> @@ -168,9 +197,11 @@
>  		.channel = num,						\
>  		.channel2 = num2,					\
>  		.address = addr,					\
> +		.scan_index = num + AT91_SAMA5D2_SINGLE_CHAN_CNT,	\
>  		.scan_type = {						\
>  			.sign = 's',					\
>  			.realbits = 12,					\
> +			.storagebits = 16,				\
>  		},							\
>  		.info_mask_separate = BIT(IIO_CHAN_INFO_RAW),		\
>  		.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE),	\
> @@ -188,6 +219,12 @@ struct at91_adc_soc_info {
>  	unsigned			max_sample_rate;
>  };
>  
> +struct at91_adc_trigger {
> +	char				*name;
> +	unsigned int			trgmod_value;
> +	unsigned int			edge_type;
> +};
> +
>  struct at91_adc_state {
>  	void __iomem			*base;
>  	int				irq;
> @@ -195,11 +232,14 @@ struct at91_adc_state {
>  	struct regulator		*reg;
>  	struct regulator		*vref;
>  	int				vref_uv;
> +	struct iio_trigger		*trig;
> +	const struct at91_adc_trigger	*selected_trig;
>  	const struct iio_chan_spec	*chan;
>  	bool				conversion_done;
>  	u32				conversion_value;
>  	struct at91_adc_soc_info	soc_info;
>  	wait_queue_head_t		wq_data_available;
> +	u16				buffer[AT91_BUFFER_MAX_HWORDS];
>  	/*
>  	 * lock to prevent concurrent 'single conversion' requests through
>  	 * sysfs.
> @@ -207,6 +247,24 @@ struct at91_adc_state {
>  	struct mutex			lock;
>  };
>  
> +static const struct at91_adc_trigger at91_adc_trigger_list[] = {
> +	{
> +		.name = "external_rising",
> +		.trgmod_value = AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_RISE,
> +		.edge_type = IRQ_TYPE_EDGE_RISING,
> +	},
> +	{
> +		.name = "external_falling",
> +		.trgmod_value = AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_FALL,
> +		.edge_type = IRQ_TYPE_EDGE_FALLING,
> +	},
> +	{
> +		.name = "external_any",
> +		.trgmod_value = AT91_SAMA5D2_TRGR_TRGMOD_EXT_TRIG_ANY,
> +		.edge_type = IRQ_TYPE_EDGE_BOTH,
> +	},
> +};
> +
>  static const struct iio_chan_spec at91_adc_channels[] = {
>  	AT91_SAMA5D2_CHAN_SINGLE(0, 0x50),
>  	AT91_SAMA5D2_CHAN_SINGLE(1, 0x54),
> @@ -226,8 +284,128 @@ static const struct iio_chan_spec at91_adc_channels[] = {
>  	AT91_SAMA5D2_CHAN_DIFF(6, 7, 0x68),
>  	AT91_SAMA5D2_CHAN_DIFF(8, 9, 0x70),
>  	AT91_SAMA5D2_CHAN_DIFF(10, 11, 0x78),
> +	IIO_CHAN_SOFT_TIMESTAMP(AT91_SAMA5D2_SINGLE_CHAN_CNT
> +				+ AT91_SAMA5D2_DIFF_CHAN_CNT + 1),
> +};
> +
> +static int at91_adc_configure_trigger(struct iio_trigger *trig, bool state)
> +{
> +	struct iio_dev *indio = iio_trigger_get_drvdata(trig);
> +	struct at91_adc_state *st = iio_priv(indio);
> +	u32 status = at91_adc_readl(st, AT91_SAMA5D2_TRGR);
> +	u8 bit;
> +
> +	/* clear TRGMOD */
> +	status &= ~AT91_SAMA5D2_TRGR_TRGMOD_MASK;
> +
> +	if (state)
> +		status |= st->selected_trig->trgmod_value;
> +
> +	/* set/unset hw trigger */
> +	at91_adc_writel(st, AT91_SAMA5D2_TRGR, status);
> +
> +	for_each_set_bit(bit, indio->active_scan_mask, indio->num_channels) {
> +		struct iio_chan_spec const *chan = indio->channels + bit;
> +
> +		if (state) {
> +			at91_adc_writel(st, AT91_SAMA5D2_CHER,
> +					BIT(chan->channel));
> +			at91_adc_writel(st, AT91_SAMA5D2_IER,
> +					BIT(chan->channel));
> +		} else {
> +			at91_adc_writel(st, AT91_SAMA5D2_IDR,
> +					BIT(chan->channel));
> +			at91_adc_writel(st, AT91_SAMA5D2_CHDR,
> +					BIT(chan->channel));
> +		}
> +	}
> +
> +	return 0;
> +}
> +
> +static int at91_adc_reenable_trigger(struct iio_trigger *trig)
> +{
> +	struct iio_dev *indio = iio_trigger_get_drvdata(trig);
> +	struct at91_adc_state *st = iio_priv(indio);
> +
As I suggested below, I'd pull the ack in here as well.
> +	enable_irq(st->irq);
> +	return 0;
> +}
> +
> +static const struct iio_trigger_ops at91_adc_trigger_ops = {
> +	.owner = THIS_MODULE,
> +	.set_trigger_state = &at91_adc_configure_trigger,
> +	.try_reenable = &at91_adc_reenable_trigger,
>  };
>  
> +static struct iio_trigger *at91_adc_allocate_trigger(struct iio_dev *indio,
> +						     char *trigger_name)
> +{
> +	struct iio_trigger *trig;
> +	int ret;
> +
> +	trig = devm_iio_trigger_alloc(&indio->dev, "%s-dev%d-%s", indio->name,
> +				      indio->id, trigger_name);
> +	if (!trig)
> +		return NULL;
> +
> +	trig->dev.parent = indio->dev.parent;
> +	iio_trigger_set_drvdata(trig, indio);
> +	trig->ops = &at91_adc_trigger_ops;
> +
> +	ret = devm_iio_trigger_register(&indio->dev, trig);
> +	if (ret)
> +		return NULL;
Could use PTR_ERR and co to not eat the error returns but pass them up
and handle in probe...  With hind sight I should have done that with
all the iio_*alloc functions but tricky to fix now.
> +
> +	return trig;
> +}
> +
> +static int at91_adc_trigger_init(struct iio_dev *indio)
> +{
> +	struct at91_adc_state *st = iio_priv(indio);
> +
> +	st->trig = at91_adc_allocate_trigger(indio, st->selected_trig->name);
> +	if (!st->trig) {
> +		dev_err(&indio->dev,
> +			"could not allocate trigger\n");
> +		return -ENOMEM;
> +	}
> +
> +	return 0;
> +}
> +
> +static irqreturn_t at91_adc_trigger_handler(int irq, void *p)
> +{
> +	struct iio_poll_func *pf = p;
> +	struct iio_dev *indio = pf->indio_dev;
> +	struct at91_adc_state *st = iio_priv(indio);
> +	int i = 0;
> +	u8 bit;
> +
> +	for_each_set_bit(bit, indio->active_scan_mask, indio->num_channels) {
> +		struct iio_chan_spec const *chan = indio->channels + bit;
> +
> +		st->buffer[i] = at91_adc_readl(st, chan->address);
> +		i++;
> +	}
> +
> +	iio_push_to_buffers_with_timestamp(indio, st->buffer, pf->timestamp);
> +
> +	iio_trigger_notify_done(indio->trig);
> +
> +	/* Needed to ACK the DRDY interruption */
> +	at91_adc_readl(st, AT91_SAMA5D2_LCDR);
Hmm. I didn't notice this before.  Technically this should probably be in the
trigger try_reenable callback rather than within the poll function.

It's about the dataready trigger being reset rather than confirming that
we've read the data...  A bit of a corner case but nice to keep things
semantically as clean as possible.

Obviously we can't always do this as many devices use the fact the registers
were read to indicate the same thing, but lets do it nicely when we can ;)

Upshot is that it gets called immediately very soon after this point anyway.
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static int at91_adc_buffer_init(struct iio_dev *indio)
> +{
> +	return devm_iio_triggered_buffer_setup(&indio->dev, indio,
> +			&iio_pollfunc_store_time,
> +			&at91_adc_trigger_handler, NULL);
> +}
> +
>  static unsigned at91_adc_startup_time(unsigned startup_time_min,
>  				      unsigned adc_clk_khz)
>  {
> @@ -293,14 +471,18 @@ static irqreturn_t at91_adc_interrupt(int irq, void *private)
>  	u32 status = at91_adc_readl(st, AT91_SAMA5D2_ISR);
>  	u32 imr = at91_adc_readl(st, AT91_SAMA5D2_IMR);
>  
> -	if (status & imr) {
> +	if (!(status & imr))
> +		return IRQ_NONE;
> +
> +	if (iio_buffer_enabled(indio)) {
> +		disable_irq_nosync(irq);
> +		iio_trigger_poll(indio->trig);
> +	} else {
>  		st->conversion_value = at91_adc_readl(st, st->chan->address);
>  		st->conversion_done = true;
>  		wake_up_interruptible(&st->wq_data_available);
> -		return IRQ_HANDLED;
>  	}
> -
> -	return IRQ_NONE;
> +	return IRQ_HANDLED;
>  }
>  
>  static int at91_adc_read_raw(struct iio_dev *indio_dev,
> @@ -313,6 +495,11 @@ static int at91_adc_read_raw(struct iio_dev *indio_dev,
>  
>  	switch (mask) {
>  	case IIO_CHAN_INFO_RAW:
> +		/* we cannot use software trigger if hw trigger enabled */
> +		ret = iio_device_claim_direct_mode(indio_dev);
> +		if (ret)
> +			return ret;
> +
>  		mutex_lock(&st->lock);
>  
>  		st->chan = chan;
> @@ -344,6 +531,8 @@ static int at91_adc_read_raw(struct iio_dev *indio_dev,
>  		at91_adc_writel(st, AT91_SAMA5D2_CHDR, BIT(chan->channel));
>  
>  		mutex_unlock(&st->lock);
> +
> +		iio_device_release_direct_mode(indio_dev);
>  		return ret;
>  
>  	case IIO_CHAN_INFO_SCALE:
> @@ -391,7 +580,8 @@ static int at91_adc_probe(struct platform_device *pdev)
>  	struct iio_dev *indio_dev;
>  	struct at91_adc_state *st;
>  	struct resource	*res;
> -	int ret;
> +	int ret, i;
> +	u32 edge_type;
>  
>  	indio_dev = devm_iio_device_alloc(&pdev->dev, sizeof(*st));
>  	if (!indio_dev)
> @@ -432,6 +622,27 @@ static int at91_adc_probe(struct platform_device *pdev)
>  		return ret;
>  	}
>  
> +	ret = of_property_read_u32(pdev->dev.of_node,
> +				   "atmel,trigger-edge-type", &edge_type);
> +	if (ret) {
> +		dev_err(&pdev->dev,
> +			"invalid or missing value for atmel,trigger-edge-type\n");
> +		return ret;
> +	}
> +
> +	st->selected_trig = NULL;
> +
> +	for (i = 0; i < AT91_SAMA5D2_HW_TRIG_CNT; i++)
> +		if (at91_adc_trigger_list[i].edge_type == edge_type) {
> +			st->selected_trig = &at91_adc_trigger_list[i];
> +			break;
> +		}
> +
> +	if (!st->selected_trig) {
> +		dev_err(&pdev->dev, "invalid external trigger edge value\n");
> +		return -EINVAL;
> +	}
> +
>  	init_waitqueue_head(&st->wq_data_available);
>  	mutex_init(&st->lock);
>  
> @@ -499,10 +710,25 @@ static int at91_adc_probe(struct platform_device *pdev)
>  
>  	platform_set_drvdata(pdev, indio_dev);
>  
> +	ret = at91_adc_buffer_init(indio_dev);
> +	if (ret < 0) {
> +		dev_err(&pdev->dev, "couldn't initialize the buffer.\n");
> +		goto per_clk_disable_unprepare;
> +	}
> +
> +	ret = at91_adc_trigger_init(indio_dev);
> +	if (ret < 0) {
> +		dev_err(&pdev->dev, "couldn't setup the triggers.\n");
> +		goto per_clk_disable_unprepare;
> +	}
> +
>  	ret = iio_device_register(indio_dev);
>  	if (ret < 0)
>  		goto per_clk_disable_unprepare;
>  
> +	dev_info(&pdev->dev, "setting up trigger as %s\n",
> +		 st->selected_trig->name);
> +
I hope no one is silly enough to use this as ABI in any scripts! :(

>  	dev_info(&pdev->dev, "version: %x\n",
>  		 readl_relaxed(st->base + AT91_SAMA5D2_VERSION));
>  

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers
  2017-06-09  7:07 [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Eugen Hristev
                   ` (3 preceding siblings ...)
  2017-06-09  7:07 ` [PATCH v4 4/4] iio: adc: at91-sama5d2_adc: add hw trigger and buffer support Eugen Hristev
@ 2017-06-14  6:57 ` Ludovic Desroches
  4 siblings, 0 replies; 10+ messages in thread
From: Ludovic Desroches @ 2017-06-14  6:57 UTC (permalink / raw)
  To: Eugen Hristev
  Cc: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	jic23

Hi,

On Fri, Jun 09, 2017 at 10:07:15AM +0300, Eugen Hristev wrote:
> This patch implements the hardware triggers support and buffer management
> for sama5d2.
> The DT modifications ( [PATCH v4 1/4] ARM: dts: at91: sama5d2_xplained:
> enable ADTRG pin) are for demonstration purposes of the feature,
> setting the pinctrl for the ADC hw trigger pin,should go through
> at91 maintainers.
> As discussed on the mailing list, since we have a single external trigger
> namely the ADTRG, but three different possible edges that can trigger it,
> will have a single trigger in the IIO subsystem and three possible edges
> which can be selected from the device tree.
> Thus I created new bindings in the device tree for having the possible
> edge types for the ADTRG pin.
> The documentation for the bindings is in the patch
> [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add adtrg
>  trigger binding
> The modifications in the DT are in the patch
> [PATCH v4 3/4] ARM: dts: at91: sama5d2: add ADC hw trigger
> The new property must be set with the values from interrupt edge types, as this
> pin is effectively an internal interrupt.
>

If the reuse of IRQ macros in the trigger case is accepted, you can add

Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>

for the next version with the corrections suggested by Jonathan.

Regards

Ludovic
 
>   Changes in v4:
>  - Modified bindings to be similar with interrupts
>  - Used iio claim direct to make sure buffer is not enabled when someone might
>    try to use a software triggered conversion
> 
>   Changes in v3:
>  - No longer have all three possible triggers registered in the subsystem and
>    available in the sysfs. Only registering one trigger, and the possible edge
>    is being read from the device tree. In the device tree, any possible edge
>    can be set. There are three already made nodes for each possible edge which
>    can be selected.
>  - Fixed a bug where software triggered conversion could be started even if
>    the hardware trigger was enabled.
>  - Preallocate enough space for the buffer, considering all the channels
>    and timestamp
> 
> 
> 
>   Changes in v2:
>  - Moved buffer allocation and freeing into the preenable and postdisable
>    callbacks.
>    We have a total of scan bytes that can vary a lot depending on each channel
>    enabled at a certain point.
>  - made the at91 trigger list part of state structure
>  - made the iio trigger list preallocated in state structure
>  - moved irq enabling/disabling into the try_reenable callback
>  - on trigger disable must write disable registries as well
>  - Modified trigger name length to 64
> 
> 
> Eugen Hristev (4):
>   ARM: dts: at91: sama5d2_xplained: enable ADTRG pin
>   Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding
>   ARM: dts: at91: sama5d2: add ADC hw trigger edge type
>   iio: adc: at91-sama5d2_adc: add hw trigger and buffer support
> 
>  .../bindings/iio/adc/at91-sama5d2_adc.txt          |   6 +
>  arch/arm/boot/dts/at91-sama5d2_xplained.dts        |  16 +-
>  arch/arm/boot/dts/sama5d2.dtsi                     |   1 +
>  drivers/iio/adc/at91-sama5d2_adc.c                 | 236 ++++++++++++++++++++-
>  4 files changed, 253 insertions(+), 6 deletions(-)
> 
> -- 
> 2.7.4
> 

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding
  2017-06-09  7:07 ` [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding Eugen Hristev
  2017-06-11 11:24   ` Jonathan Cameron
@ 2017-06-14 15:30   ` Rob Herring
  1 sibling, 0 replies; 10+ messages in thread
From: Rob Herring @ 2017-06-14 15:30 UTC (permalink / raw)
  To: Eugen Hristev
  Cc: nicolas.ferre, alexandre.belloni, linux-iio, lars,
	linux-arm-kernel, devicetree, linux-kernel, ludovic.desroches,
	jic23

On Fri, Jun 09, 2017 at 10:07:17AM +0300, Eugen Hristev wrote:
> Add property for the edge type of the hardware trigger pin ADTRG
> 
> Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com>
> ---
>   Changes in v4:
>  - Modified bindings to be similar with interrupts
> 
>  Documentation/devicetree/bindings/iio/adc/at91-sama5d2_adc.txt | 6 ++++++
>  1 file changed, 6 insertions(+)

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

^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2017-06-14 15:30 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-06-09  7:07 [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Eugen Hristev
2017-06-09  7:07 ` [PATCH v4 1/4] ARM: dts: at91: sama5d2_xplained: enable ADTRG pin Eugen Hristev
2017-06-09  7:07 ` [PATCH v4 2/4] Documentation: dt: iio: at91-sama5d2_adc: add hw trigger edge binding Eugen Hristev
2017-06-11 11:24   ` Jonathan Cameron
2017-06-11 11:25     ` Jonathan Cameron
2017-06-14 15:30   ` Rob Herring
2017-06-09  7:07 ` [PATCH v4 3/4] ARM: dts: at91: sama5d2: add ADC hw trigger edge type Eugen Hristev
2017-06-09  7:07 ` [PATCH v4 4/4] iio: adc: at91-sama5d2_adc: add hw trigger and buffer support Eugen Hristev
2017-06-11 11:34   ` Jonathan Cameron
2017-06-14  6:57 ` [PATCH v4 0/4] iio: adc: sama5d2_adc hw triggers and buffers Ludovic Desroches

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).