Skip to contentSkip to conversation

Add continuous sampling with IIO buffers for Vybrid

PATCH New Sanchayan Maity 3 participants 8 messages Aug 11, 2015 – Aug 31, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 Add continuous sampling with IIO buffers for Vybrid

Hello,

This patch adds support for continuous sampling provided by the ADC block on Vybrid by leveraging the IIO triggered buffers infrastructure.

The patch has been tested on Colibri VF50 and VF61 on shawn's tree for-next branch with the patches [1] and [2] applied.

The below script was used for testing.

#!/bin/sh
echo 0 > /sys/bus/iio/devices/iio:device0/scan_elements/in_voltage8_en echo 0 > /sys/bus/iio/devices/iio:device0/scan_elements/in_voltage9_en echo 1 > /sys/bus/iio/devices/iio:device0/scan_elements/in_temp_en echo 1 > /sys/bus/iio/devices/iio:device0/scan_elements/in_timestamp_en echo 0 > /sys/bus/iio/devices/iio_sysfs_trigger/add_trigger /home/root/generic_buffer -n 4003b000.adc -t sysfstrig0 -l 512 -c 10 echo 0 > /sys/bus/iio/devices/iio:device0/scan_elements/in_timestamp_en echo 0 > /sys/bus/iio/devices/iio:device0/scan_elements/in_temp_en echo 0 > /sys/bus/iio/devices/iio:device0/scan_elements/in_voltage8_en echo 0 > /sys/bus/iio/devices/iio:device0/scan_elements/in_voltage9_en

Feedback and comments are most welcome.

Changes since v2:
1. Fix the wrong buffer size for statically allocated buffer
2. Drop the use of .address field from the iio_chan_spec
3. Use iio_buffer_enabled call inside the lock
4. Drop wrapper function around iio_trigered_* function calls
5. Drop Kconfig select of sysfs trigger
6. Drop Kconfig select IIO_TRIGGER as it is already selected by IIO_TRIGGERED_BUFFER

Changes since v1:
1. Use a fixed size buffer instead of kmalloc allocated during update scan mode
2. Remove a write to read only register ADC_HS (COCO bit)

Version 2 patch can be found here
http://www.gossamer-threads.com/lists/linux/kernel/2235178

Version 1 patch can be found here
http://www.spinics.net/lists/linux-iio/msg20053.html

[1]. https://lkml.org/lkml/2015/5/27/350
[2]. https://lkml.org/lkml/2015/7/14/395

Thanks & Regards,
Sanchayan Maity.

Sanchayan Maity (1):
iio: adc: vf610: Add IIO buffer support for Vybrid ADC

 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 102 +++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 97 insertions(+), 7 deletions(-)
Signature · 1 line
2.5.0
#1 lore raw

PATCH v3 1/1 iio: adc: vf610: Add IIO buffer support for Vybrid ADC

This patch adds support for IIO buffer to the Vybrid ADC driver. IIO triggered buffer infrastructure along with iio sysfs trigger is used to leverage continuous sampling support provided by the ADC block.

---
 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 102 +++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 97 insertions(+), 7 deletions(-)
drivers/iio/adc/Kconfig +2 -0
index 7c55658..660f790 100644
@@ -337,6 +337,8 @@ config TWL6030_GPADC
 config VF610_ADC
 	tristate "Freescale vf610 ADC driver"
 	depends on OF
+	select IIO_BUFFER
+	select IIO_TRIGGERED_BUFFER
 	help
 	  Say yes here to support for Vybrid board analog-to-digital converter.
 	  Since the IP is used for i.MX6SLX, the driver also support i.MX6SLX.
drivers/iio/adc/vf610_adc.c +95 -7

182 lines not on this page · raw message

Signature · 1 line
2.5.0
maintainer
#2 lore raw

Re: [PATCH v3] iio: adc: vf610: Add IIO buffer support for Vybrid ADC

Quoted 6 lines · Sanchayan Maity

This patch adds support for IIO buffer to the Vybrid ADC driver. IIO triggered buffer infrastructure along with iio sysfs trigger is used to leverage continuous sampling support provided by the ADC block.

Hi Sanchayan,

Very nearly there. One little point to do with the buffer handling. Basically I don't think you want anything in the preenable or postdisable hooks, just in the 'internal' ones.

Jonathan

Quoted 133 lines
---
 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 102 +++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 97 insertions(+), 7 deletions(-)
drivers/iio/adc/Kconfig +2 -0
index 7c55658..660f790 100644
@@ -337,6 +337,8 @@ config TWL6030_GPADC
 config VF610_ADC
 	tristate "Freescale vf610 ADC driver"
 	depends on OF
+	select IIO_BUFFER
+	select IIO_TRIGGERED_BUFFER
 	help
 	  Say yes here to support for Vybrid board analog-to-digital converter.
 	  Since the IP is used for i.MX6SLX, the driver also support i.MX6SLX.
drivers/iio/adc/vf610_adc.c +48 -4
index 23b8fb9..de62c48 100644
@@ -34,8 +34,11 @@
 #include <linux/err.h>
 
 #include <linux/iio/iio.h>
+#include <linux/iio/buffer.h>
 #include <linux/iio/sysfs.h>
-#include <linux/iio/driver.h>
+#include <linux/iio/trigger.h>
+#include <linux/iio/trigger_consumer.h>
+#include <linux/iio/triggered_buffer.h>
 
 /* This will be the driver name the kernel reports */
 #define DRIVER_NAME "vf610-adc"
@@ -170,6 +173,7 @@ struct vf610_adc {
 	u32 sample_freq_avail[5];
 
 	struct completion completion;
+	u16 buffer[8];
 };
 
 static const u32 vf610_hw_avgs[] = { 1, 4, 8, 16, 32 };
@@ -505,12 +509,24 @@ static const struct iio_chan_spec_ext_info vf610_ext_info[] = {
 	.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE) |	\
 				BIT(IIO_CHAN_INFO_SAMP_FREQ),	\
 	.ext_info = vf610_ext_info,				\
+	.scan_index = (_idx),			\
+	.scan_type = {					\
+		.sign = 'u',				\
+		.realbits = 12,				\
+		.storagebits = 16,			\
+	},						\
 }
 
 #define VF610_ADC_TEMPERATURE_CHAN(_idx, _chan_type) {	\
 	.type = (_chan_type),	\
 	.channel = (_idx),		\
 	.info_mask_separate = BIT(IIO_CHAN_INFO_PROCESSED),	\
+	.scan_index = (_idx),					\
+	.scan_type = {						\
+		.sign = 'u',					\
+		.realbits = 12,					\
+		.storagebits = 16,				\
+	},							\
 }
 
 static const struct iio_chan_spec vf610_adc_iio_channels[] = {
@@ -531,6 +547,7 @@ static const struct iio_chan_spec vf610_adc_iio_channels[] = {
 	VF610_ADC_CHAN(14, IIO_VOLTAGE),
 	VF610_ADC_CHAN(15, IIO_VOLTAGE),
 	VF610_ADC_TEMPERATURE_CHAN(26, IIO_TEMP),
+	IIO_CHAN_SOFT_TIMESTAMP(32),
 	/* sentinel */
 };
 
@@ -559,13 +576,20 @@ static int vf610_adc_read_data(struct vf610_adc *info)
 
 static irqreturn_t vf610_adc_isr(int irq, void *dev_id)
 {
-	struct vf610_adc *info = (struct vf610_adc *)dev_id;
+	struct iio_dev *indio_dev = (struct iio_dev *)dev_id;
+	struct vf610_adc *info = iio_priv(indio_dev);
 	int coco;
 
 	coco = readl(info->regs + VF610_REG_ADC_HS);
 	if (coco & VF610_ADC_HS_COCO0) {
 		info->value = vf610_adc_read_data(info);
-		complete(&info->completion);
+		if (iio_buffer_enabled(indio_dev)) {
+			info->buffer[0] = info->value;
+			iio_push_to_buffers_with_timestamp(indio_dev,
+					info->buffer, iio_get_time_ns());
+			iio_trigger_notify_done(indio_dev->trig);
+		} else
+			complete(&info->completion);
 	}
 
 	return IRQ_HANDLED;
@@ -613,8 +637,12 @@ static int vf610_read_raw(struct iio_dev *indio_dev,
 	case IIO_CHAN_INFO_RAW:
 	case IIO_CHAN_INFO_PROCESSED:
 		mutex_lock(&indio_dev->mlock);
-		reinit_completion(&info->completion);
+		if (iio_buffer_enabled(indio_dev)) {
+			mutex_unlock(&indio_dev->mlock);
+			return -EBUSY;
+		}
 
+		reinit_completion(&info->completion);
 		hc_cfg = VF610_ADC_ADCHC(chan->channel);
 		hc_cfg |= VF610_ADC_AIEN;
 		writel(hc_cfg, info->regs + VF610_REG_ADC_HC0);
@@ -694,6 +722,57 @@ static int vf610_write_raw(struct iio_dev *indio_dev,
 	return -EINVAL;
 }
 
+static int vf610_adc_buffer_postenable(struct iio_dev *indio_dev)
+{
+	struct vf610_adc *info = iio_priv(indio_dev);
+	unsigned int channel;
+	int ret;
+	int val;
+
+	ret = iio_triggered_buffer_postenable(indio_dev);
+	if (ret)
+		return ret;
+
+	val = readl(info->regs + VF610_REG_ADC_GC);
+	val |= VF610_ADC_ADCON;
+	writel(val, info->regs + VF610_REG_ADC_GC);
+
+	channel = find_first_bit(indio_dev->active_scan_mask,

drivers/iio/adc/vf610_adc.c static int vf610_write_raw(struct iio_dev *indio_dev, · line 750

Quoted 10 lines
drivers/iio/adc/vf610_adc.c +10 -0
+						indio_dev->masklength);
+
+	val = VF610_ADC_ADCHC(channel);
+	val |= VF610_ADC_AIEN;
+
+	writel(val, info->regs + VF610_REG_ADC_HC0);
+
+	return 0;
+}
+

These are suppose to be pair wise matched, so anything set up in postenable has to be torn down in predisable. Likewise, preenable setup torn down in postdisable.

The distinction is meant to be that a query will return that the buffer is enabled from a point between the pre and post enable and return that it is disabled from a point between the pre and post disable. Actually locks are held so there may no longer be a window where it matters, but best to keep to convention.

I think you want the postdisable -> predisable and then call the standard predisable from there.

Quoted 72 lines
drivers/iio/adc/vf610_adc.c +37 -3
+static int vf610_adc_buffer_postdisable(struct iio_dev *indio_dev)
+{
+	struct vf610_adc *info = iio_priv(indio_dev);
+	unsigned int hc_cfg = 0;
+	int val;
+
+	val = readl(info->regs + VF610_REG_ADC_GC);
+	val &= ~VF610_ADC_ADCON;
+	writel(val, info->regs + VF610_REG_ADC_GC);
+
+	hc_cfg |= VF610_ADC_CONV_DISABLE;
+	hc_cfg &= ~VF610_ADC_AIEN;
+
+	writel(hc_cfg, info->regs + VF610_REG_ADC_HC0);
+
+	return 0;
+}
+
+static const struct iio_buffer_setup_ops iio_triggered_buffer_setup_ops = {
+	.postenable = &vf610_adc_buffer_postenable,
+	.predisable = &iio_triggered_buffer_predisable,
+	.postdisable = &vf610_adc_buffer_postdisable,
+	.validate_scan_mask = &iio_validate_scan_mask_onehot,
+};
+
 static int vf610_adc_reg_access(struct iio_dev *indio_dev,
 			unsigned reg, unsigned writeval,
 			unsigned *readval)
@@ -753,7 +832,7 @@ static int vf610_adc_probe(struct platform_device *pdev)
 
 	ret = devm_request_irq(info->dev, irq,
 				vf610_adc_isr, 0,
-				dev_name(&pdev->dev), info);
+				dev_name(&pdev->dev), indio_dev);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed requesting irq, irq = %d\n", irq);
 		return ret;
@@ -806,15 +885,23 @@ static int vf610_adc_probe(struct platform_device *pdev)
 	vf610_adc_cfg_init(info);
 	vf610_adc_hw_init(info);
 
+	ret = iio_triggered_buffer_setup(indio_dev, &iio_pollfunc_store_time,
+					NULL, &iio_triggered_buffer_setup_ops);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "Couldn't initialise the buffer\n");
+		goto error_iio_device_register;
+	}
+
 	ret = iio_device_register(indio_dev);
 	if (ret) {
 		dev_err(&pdev->dev, "Couldn't register the device.\n");
-		goto error_iio_device_register;
+		goto error_adc_buffer_init;
 	}
 
 	return 0;
 
-
+error_adc_buffer_init:
+	iio_triggered_buffer_cleanup(indio_dev);
 error_iio_device_register:
 	clk_disable_unprepare(info->clk);
 error_adc_clk_enable:
@@ -829,6 +916,7 @@ static int vf610_adc_remove(struct platform_device *pdev)
 	struct vf610_adc *info = iio_priv(indio_dev);
 
 	iio_device_unregister(indio_dev);
+	iio_triggered_buffer_cleanup(indio_dev);
 	regulator_disable(info->vref);
 	clk_disable_unprepare(info->clk);
 
#3 lore raw

Re: [PATCH v3] iio: adc: vf610: Add IIO buffer support for Vybrid ADC

Hello Jonathan,

Quoted 162 lines · Jonathan Cameron
Quoted 6 lines · Sanchayan Maity

This patch adds support for IIO buffer to the Vybrid ADC driver. IIO triggered buffer infrastructure along with iio sysfs trigger is used to leverage continuous sampling support provided by the ADC block.

Hi Sanchayan,

Very nearly there. One little point to do with the buffer handling. Basically I don't think you want anything in the preenable or postdisable hooks, just in the 'internal' ones.

Jonathan

Quoted 133 lines
---
 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 102 +++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 97 insertions(+), 7 deletions(-)
drivers/iio/adc/Kconfig +2 -0
index 7c55658..660f790 100644
@@ -337,6 +337,8 @@ config TWL6030_GPADC
 config VF610_ADC
 	tristate "Freescale vf610 ADC driver"
 	depends on OF
+	select IIO_BUFFER
+	select IIO_TRIGGERED_BUFFER
 	help
 	  Say yes here to support for Vybrid board analog-to-digital converter.
 	  Since the IP is used for i.MX6SLX, the driver also support i.MX6SLX.
drivers/iio/adc/vf610_adc.c +48 -4
index 23b8fb9..de62c48 100644
@@ -34,8 +34,11 @@
 #include <linux/err.h>
 
 #include <linux/iio/iio.h>
+#include <linux/iio/buffer.h>
 #include <linux/iio/sysfs.h>
-#include <linux/iio/driver.h>
+#include <linux/iio/trigger.h>
+#include <linux/iio/trigger_consumer.h>
+#include <linux/iio/triggered_buffer.h>
 
 /* This will be the driver name the kernel reports */
 #define DRIVER_NAME "vf610-adc"
@@ -170,6 +173,7 @@ struct vf610_adc {
 	u32 sample_freq_avail[5];
 
 	struct completion completion;
+	u16 buffer[8];
 };
 
 static const u32 vf610_hw_avgs[] = { 1, 4, 8, 16, 32 };
@@ -505,12 +509,24 @@ static const struct iio_chan_spec_ext_info vf610_ext_info[] = {
 	.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE) |	\
 				BIT(IIO_CHAN_INFO_SAMP_FREQ),	\
 	.ext_info = vf610_ext_info,				\
+	.scan_index = (_idx),			\
+	.scan_type = {					\
+		.sign = 'u',				\
+		.realbits = 12,				\
+		.storagebits = 16,			\
+	},						\
 }
 
 #define VF610_ADC_TEMPERATURE_CHAN(_idx, _chan_type) {	\
 	.type = (_chan_type),	\
 	.channel = (_idx),		\
 	.info_mask_separate = BIT(IIO_CHAN_INFO_PROCESSED),	\
+	.scan_index = (_idx),					\
+	.scan_type = {						\
+		.sign = 'u',					\
+		.realbits = 12,					\
+		.storagebits = 16,				\
+	},							\
 }
 
 static const struct iio_chan_spec vf610_adc_iio_channels[] = {
@@ -531,6 +547,7 @@ static const struct iio_chan_spec vf610_adc_iio_channels[] = {
 	VF610_ADC_CHAN(14, IIO_VOLTAGE),
 	VF610_ADC_CHAN(15, IIO_VOLTAGE),
 	VF610_ADC_TEMPERATURE_CHAN(26, IIO_TEMP),
+	IIO_CHAN_SOFT_TIMESTAMP(32),
 	/* sentinel */
 };
 
@@ -559,13 +576,20 @@ static int vf610_adc_read_data(struct vf610_adc *info)
 
 static irqreturn_t vf610_adc_isr(int irq, void *dev_id)
 {
-	struct vf610_adc *info = (struct vf610_adc *)dev_id;
+	struct iio_dev *indio_dev = (struct iio_dev *)dev_id;
+	struct vf610_adc *info = iio_priv(indio_dev);
 	int coco;
 
 	coco = readl(info->regs + VF610_REG_ADC_HS);
 	if (coco & VF610_ADC_HS_COCO0) {
 		info->value = vf610_adc_read_data(info);
-		complete(&info->completion);
+		if (iio_buffer_enabled(indio_dev)) {
+			info->buffer[0] = info->value;
+			iio_push_to_buffers_with_timestamp(indio_dev,
+					info->buffer, iio_get_time_ns());
+			iio_trigger_notify_done(indio_dev->trig);
+		} else
+			complete(&info->completion);
 	}
 
 	return IRQ_HANDLED;
@@ -613,8 +637,12 @@ static int vf610_read_raw(struct iio_dev *indio_dev,
 	case IIO_CHAN_INFO_RAW:
 	case IIO_CHAN_INFO_PROCESSED:
 		mutex_lock(&indio_dev->mlock);
-		reinit_completion(&info->completion);
+		if (iio_buffer_enabled(indio_dev)) {
+			mutex_unlock(&indio_dev->mlock);
+			return -EBUSY;
+		}
 
+		reinit_completion(&info->completion);
 		hc_cfg = VF610_ADC_ADCHC(chan->channel);
 		hc_cfg |= VF610_ADC_AIEN;
 		writel(hc_cfg, info->regs + VF610_REG_ADC_HC0);
@@ -694,6 +722,57 @@ static int vf610_write_raw(struct iio_dev *indio_dev,
 	return -EINVAL;
 }
 
+static int vf610_adc_buffer_postenable(struct iio_dev *indio_dev)
+{
+	struct vf610_adc *info = iio_priv(indio_dev);
+	unsigned int channel;
+	int ret;
+	int val;
+
+	ret = iio_triggered_buffer_postenable(indio_dev);
+	if (ret)
+		return ret;
+
+	val = readl(info->regs + VF610_REG_ADC_GC);
+	val |= VF610_ADC_ADCON;
+	writel(val, info->regs + VF610_REG_ADC_GC);
+
+	channel = find_first_bit(indio_dev->active_scan_mask,

drivers/iio/adc/vf610_adc.c static int vf610_write_raw(struct iio_dev *indio_dev, · line 750

Quoted 10 lines
drivers/iio/adc/vf610_adc.c +10 -0
+						indio_dev->masklength);
+
+	val = VF610_ADC_ADCHC(channel);
+	val |= VF610_ADC_AIEN;
+
+	writel(val, info->regs + VF610_REG_ADC_HC0);
+
+	return 0;
+}
+

These are suppose to be pair wise matched, so anything set up in postenable has to be torn down in predisable. Likewise, preenable setup torn down in postdisable.

Quoted 8 lines

The distinction is meant to be that a query will return that the buffer is enabled from a point between the pre and post enable and return that it is disabled from a point between the pre and post disable. Actually locks are held so there may no longer be a window where it matters, but best to keep to convention.

I think you want the postdisable -> predisable and then call the standard predisable from there.

Ok. I guess I understood. So the postdisable code actually should go into predisable and while taking care of mutex locking.

The way I was reasoning wanting to enable the continuous sampling after the buffer is enabled and then after buffer is disabled, stopping it. With your explanation it makes sense that I should have it in predisable.

Thank you for the feedback. I will incoporate the above change and send the new version.

- Sanchayan.

Quoted 74 lines
Quoted 73 lines
drivers/iio/adc/vf610_adc.c +37 -3
+static int vf610_adc_buffer_postdisable(struct iio_dev *indio_dev)
+{
+	struct vf610_adc *info = iio_priv(indio_dev);
+	unsigned int hc_cfg = 0;
+	int val;
+
+	val = readl(info->regs + VF610_REG_ADC_GC);
+	val &= ~VF610_ADC_ADCON;
+	writel(val, info->regs + VF610_REG_ADC_GC);
+
+	hc_cfg |= VF610_ADC_CONV_DISABLE;
+	hc_cfg &= ~VF610_ADC_AIEN;
+
+	writel(hc_cfg, info->regs + VF610_REG_ADC_HC0);
+
+	return 0;
+}
+
+static const struct iio_buffer_setup_ops iio_triggered_buffer_setup_ops = {
+	.postenable = &vf610_adc_buffer_postenable,
+	.predisable = &iio_triggered_buffer_predisable,
+	.postdisable = &vf610_adc_buffer_postdisable,
+	.validate_scan_mask = &iio_validate_scan_mask_onehot,
+};
+
 static int vf610_adc_reg_access(struct iio_dev *indio_dev,
 			unsigned reg, unsigned writeval,
 			unsigned *readval)
@@ -753,7 +832,7 @@ static int vf610_adc_probe(struct platform_device *pdev)
 
 	ret = devm_request_irq(info->dev, irq,
 				vf610_adc_isr, 0,
-				dev_name(&pdev->dev), info);
+				dev_name(&pdev->dev), indio_dev);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed requesting irq, irq = %d\n", irq);
 		return ret;
@@ -806,15 +885,23 @@ static int vf610_adc_probe(struct platform_device *pdev)
 	vf610_adc_cfg_init(info);
 	vf610_adc_hw_init(info);
 
+	ret = iio_triggered_buffer_setup(indio_dev, &iio_pollfunc_store_time,
+					NULL, &iio_triggered_buffer_setup_ops);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "Couldn't initialise the buffer\n");
+		goto error_iio_device_register;
+	}
+
 	ret = iio_device_register(indio_dev);
 	if (ret) {
 		dev_err(&pdev->dev, "Couldn't register the device.\n");
-		goto error_iio_device_register;
+		goto error_adc_buffer_init;
 	}
 
 	return 0;
 
-
+error_adc_buffer_init:
+	iio_triggered_buffer_cleanup(indio_dev);
 error_iio_device_register:
 	clk_disable_unprepare(info->clk);
 error_adc_clk_enable:
@@ -829,6 +916,7 @@ static int vf610_adc_remove(struct platform_device *pdev)
 	struct vf610_adc *info = iio_priv(indio_dev);
 
 	iio_device_unregister(indio_dev);
+	iio_triggered_buffer_cleanup(indio_dev);
 	regulator_disable(info->vref);
 	clk_disable_unprepare(info->clk);
 
#4 lore raw

PATCH v4 1/1 iio: adc: vf610: Add IIO buffer support for Vybrid ADC

In reply to Sanchayan Maity

This patch adds support for IIO buffer to the Vybrid ADC driver. IIO triggered buffer infrastructure along with iio sysfs trigger is used to leverage continuous sampling support provided by the ADC block.

---

Changes since v3:
Fix iio_buffer_setup_ops for postenable and predisable functions to match pairwise. Before this the predisable work was being done in postdisable.

Changes since v2:
1. Fix the wrong buffer size for statically allocated buffer
2. Drop the use of .address field from the iio_chan_spec
3. Use iio_buffer_enabled call inside the lock
4. Drop wrapper function around iio_trigered_* function calls
5. Drop Kconfig select of sysfs trigger
6. Drop Kconfig select IIO_TRIGGER as it is already selected by IIO_TRIGGERED_BUFFER

Changes since v1:
1. Use a fixed size buffer instead of kmalloc allocated during update scan mode
2. Remove a write to read only register ADC_HS (COCO bit)

 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 105 +++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 100 insertions(+), 7 deletions(-)
drivers/iio/adc/Kconfig +2 -0
index 7c55658..660f790 100644
@@ -337,6 +337,8 @@ config TWL6030_GPADC
 config VF610_ADC
 	tristate "Freescale vf610 ADC driver"
 	depends on OF
+	select IIO_BUFFER
+	select IIO_TRIGGERED_BUFFER
 	help
 	  Say yes here to support for Vybrid board analog-to-digital converter.
 	  Since the IP is used for i.MX6SLX, the driver also support i.MX6SLX.
drivers/iio/adc/vf610_adc.c +98 -7

185 lines not on this page · raw message

Signature · 1 line
2.5.0
#5 lore raw

Re: [PATCH v4] iio: adc: vf610: Add IIO buffer support for Vybrid ADC

Hello Fugang Duan,

Can I have your ACK if you are ok with the changes made by the patch?

Thanks & Regards,
Sanchayan Maity.

Quoted 246 lines · Sanchayan Maity

This patch adds support for IIO buffer to the Vybrid ADC driver. IIO triggered buffer infrastructure along with iio sysfs trigger is used to leverage continuous sampling support provided by the ADC block.

---

Changes since v3:
Fix iio_buffer_setup_ops for postenable and predisable functions to match pairwise. Before this the predisable work was being done in postdisable.

Changes since v2:
1. Fix the wrong buffer size for statically allocated buffer
2. Drop the use of .address field from the iio_chan_spec
3. Use iio_buffer_enabled call inside the lock
4. Drop wrapper function around iio_trigered_* function calls
5. Drop Kconfig select of sysfs trigger
6. Drop Kconfig select IIO_TRIGGER as it is already selected by IIO_TRIGGERED_BUFFER

Changes since v1:
1. Use a fixed size buffer instead of kmalloc allocated during update scan mode
2. Remove a write to read only register ADC_HS (COCO bit)

 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 105 +++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 100 insertions(+), 7 deletions(-)
drivers/iio/adc/Kconfig +2 -0
index 7c55658..660f790 100644
@@ -337,6 +337,8 @@ config TWL6030_GPADC
 config VF610_ADC
 	tristate "Freescale vf610 ADC driver"
 	depends on OF
+	select IIO_BUFFER
+	select IIO_TRIGGERED_BUFFER
 	help
 	  Say yes here to support for Vybrid board analog-to-digital converter.
 	  Since the IP is used for i.MX6SLX, the driver also support i.MX6SLX.
drivers/iio/adc/vf610_adc.c +98 -7
index 23b8fb9..887a7ba 100644
@@ -34,8 +34,11 @@
 #include <linux/err.h>
 
 #include <linux/iio/iio.h>
+#include <linux/iio/buffer.h>
 #include <linux/iio/sysfs.h>
-#include <linux/iio/driver.h>
+#include <linux/iio/trigger.h>
+#include <linux/iio/trigger_consumer.h>
+#include <linux/iio/triggered_buffer.h>
 
 /* This will be the driver name the kernel reports */
 #define DRIVER_NAME "vf610-adc"
@@ -170,6 +173,7 @@ struct vf610_adc {
 	u32 sample_freq_avail[5];
 
 	struct completion completion;
+	u16 buffer[8];
 };
 
 static const u32 vf610_hw_avgs[] = { 1, 4, 8, 16, 32 };
@@ -505,12 +509,24 @@ static const struct iio_chan_spec_ext_info vf610_ext_info[] = {
 	.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE) |	\
 				BIT(IIO_CHAN_INFO_SAMP_FREQ),	\
 	.ext_info = vf610_ext_info,				\
+	.scan_index = (_idx),			\
+	.scan_type = {					\
+		.sign = 'u',				\
+		.realbits = 12,				\
+		.storagebits = 16,			\
+	},						\
 }
 
 #define VF610_ADC_TEMPERATURE_CHAN(_idx, _chan_type) {	\
 	.type = (_chan_type),	\
 	.channel = (_idx),		\
 	.info_mask_separate = BIT(IIO_CHAN_INFO_PROCESSED),	\
+	.scan_index = (_idx),					\
+	.scan_type = {						\
+		.sign = 'u',					\
+		.realbits = 12,					\
+		.storagebits = 16,				\
+	},							\
 }
 
 static const struct iio_chan_spec vf610_adc_iio_channels[] = {
@@ -531,6 +547,7 @@ static const struct iio_chan_spec vf610_adc_iio_channels[] = {
 	VF610_ADC_CHAN(14, IIO_VOLTAGE),
 	VF610_ADC_CHAN(15, IIO_VOLTAGE),
 	VF610_ADC_TEMPERATURE_CHAN(26, IIO_TEMP),
+	IIO_CHAN_SOFT_TIMESTAMP(32),
 	/* sentinel */
 };
 
@@ -559,13 +576,20 @@ static int vf610_adc_read_data(struct vf610_adc *info)
 
 static irqreturn_t vf610_adc_isr(int irq, void *dev_id)
 {
-	struct vf610_adc *info = (struct vf610_adc *)dev_id;
+	struct iio_dev *indio_dev = (struct iio_dev *)dev_id;
+	struct vf610_adc *info = iio_priv(indio_dev);
 	int coco;
 
 	coco = readl(info->regs + VF610_REG_ADC_HS);
 	if (coco & VF610_ADC_HS_COCO0) {
 		info->value = vf610_adc_read_data(info);
-		complete(&info->completion);
+		if (iio_buffer_enabled(indio_dev)) {
+			info->buffer[0] = info->value;
+			iio_push_to_buffers_with_timestamp(indio_dev,
+					info->buffer, iio_get_time_ns());
+			iio_trigger_notify_done(indio_dev->trig);
+		} else
+			complete(&info->completion);
 	}
 
 	return IRQ_HANDLED;
@@ -613,8 +637,12 @@ static int vf610_read_raw(struct iio_dev *indio_dev,
 	case IIO_CHAN_INFO_RAW:
 	case IIO_CHAN_INFO_PROCESSED:
 		mutex_lock(&indio_dev->mlock);
-		reinit_completion(&info->completion);
+		if (iio_buffer_enabled(indio_dev)) {
+			mutex_unlock(&indio_dev->mlock);
+			return -EBUSY;
+		}
 
+		reinit_completion(&info->completion);
 		hc_cfg = VF610_ADC_ADCHC(chan->channel);
 		hc_cfg |= VF610_ADC_AIEN;
 		writel(hc_cfg, info->regs + VF610_REG_ADC_HC0);
@@ -694,6 +722,60 @@ static int vf610_write_raw(struct iio_dev *indio_dev,
 	return -EINVAL;
 }
 
+static int vf610_adc_buffer_postenable(struct iio_dev *indio_dev)
+{
+	struct vf610_adc *info = iio_priv(indio_dev);
+	unsigned int channel;
+	int ret;
+	int val;
+
+	ret = iio_triggered_buffer_postenable(indio_dev);
+	if (ret)
+		return ret;
+
+	val = readl(info->regs + VF610_REG_ADC_GC);
+	val |= VF610_ADC_ADCON;
+	writel(val, info->regs + VF610_REG_ADC_GC);
+
+	channel = find_first_bit(indio_dev->active_scan_mask,
+						indio_dev->masklength);
+
+	val = VF610_ADC_ADCHC(channel);
+	val |= VF610_ADC_AIEN;
+
+	writel(val, info->regs + VF610_REG_ADC_HC0);
+
+	return 0;
+}
+
+static int vf610_adc_buffer_predisable(struct iio_dev *indio_dev)
+{
+	struct vf610_adc *info = iio_priv(indio_dev);
+	unsigned int hc_cfg = 0;
+	int val, ret;
+
+	val = readl(info->regs + VF610_REG_ADC_GC);
+	val &= ~VF610_ADC_ADCON;
+	writel(val, info->regs + VF610_REG_ADC_GC);
+
+	hc_cfg |= VF610_ADC_CONV_DISABLE;
+	hc_cfg &= ~VF610_ADC_AIEN;
+
+	writel(hc_cfg, info->regs + VF610_REG_ADC_HC0);
+
+	ret = iio_triggered_buffer_predisable(indio_dev);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static const struct iio_buffer_setup_ops iio_triggered_buffer_setup_ops = {
+	.postenable = &vf610_adc_buffer_postenable,
+	.predisable = &vf610_adc_buffer_predisable,
+	.validate_scan_mask = &iio_validate_scan_mask_onehot,
+};
+
 static int vf610_adc_reg_access(struct iio_dev *indio_dev,
 			unsigned reg, unsigned writeval,
 			unsigned *readval)
@@ -753,7 +835,7 @@ static int vf610_adc_probe(struct platform_device *pdev)
 
 	ret = devm_request_irq(info->dev, irq,
 				vf610_adc_isr, 0,
-				dev_name(&pdev->dev), info);
+				dev_name(&pdev->dev), indio_dev);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed requesting irq, irq = %d\n", irq);
 		return ret;
@@ -806,15 +888,23 @@ static int vf610_adc_probe(struct platform_device *pdev)
 	vf610_adc_cfg_init(info);
 	vf610_adc_hw_init(info);
 
+	ret = iio_triggered_buffer_setup(indio_dev, &iio_pollfunc_store_time,
+					NULL, &iio_triggered_buffer_setup_ops);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "Couldn't initialise the buffer\n");
+		goto error_iio_device_register;
+	}
+
 	ret = iio_device_register(indio_dev);
 	if (ret) {
 		dev_err(&pdev->dev, "Couldn't register the device.\n");
-		goto error_iio_device_register;
+		goto error_adc_buffer_init;
 	}
 
 	return 0;
 
-
+error_adc_buffer_init:
+	iio_triggered_buffer_cleanup(indio_dev);
 error_iio_device_register:
 	clk_disable_unprepare(info->clk);
 error_adc_clk_enable:
@@ -829,6 +919,7 @@ static int vf610_adc_remove(struct platform_device *pdev)
 	struct vf610_adc *info = iio_priv(indio_dev);
 
 	iio_device_unregister(indio_dev);
+	iio_triggered_buffer_cleanup(indio_dev);
 	regulator_disable(info->vref);
 	clk_disable_unprepare(info->clk);
 
Signature · 1 line
2.5.0
Acked-by Fugang Duan on v4 1/1
#6 lore raw

RE: [PATCH v4] iio: adc: vf610: Add IIO buffer support for Vybrid ADC

In reply to Sanchayan Maity

From: Sanchayan Maity <[email protected]> Sent: Monday, August 17, 2015 11:52 PM

Quoted 25 lines

To: jic23 at kernel.org; linux-iio at vger.kernel.org
Cc: stefan at agner.ch; Duan Fugang-B38611; pmeerw at pmeerw.net; linux- kernel at vger.kernel.org; linux-arm-kernel at lists.infradead.org; Sanchayan Maity
Subject: [PATCH v4] iio: adc: vf610: Add IIO buffer support for Vybrid ADC

This patch adds support for IIO buffer to the Vybrid ADC driver. IIO triggered buffer infrastructure along with iio sysfs trigger is used to leverage continuous sampling support provided by the ADC block.

---

Changes since v3:
Fix iio_buffer_setup_ops for postenable and predisable functions to match pairwise. Before this the predisable work was being done in postdisable.

Changes since v2:
1. Fix the wrong buffer size for statically allocated buffer 2. Drop the use of .address field from the iio_chan_spec 3. Use iio_buffer_enabled call inside the lock 4. Drop wrapper function around iio_trigered_* function calls 5. Drop Kconfig select of sysfs trigger 6. Drop Kconfig select IIO_TRIGGER as it is already selected by IIO_TRIGGERED_BUFFER

Quoted 9 lines

Changes since v1:
1. Use a fixed size buffer instead of kmalloc allocated during update scan mode 2. Remove a write to read only register ADC_HS (COCO bit)

 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 105
+++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 100 insertions(+), 7 deletions(-)

The version is fine for me. Thanks for your effort.

maintainer Applied
#7 lore raw

Re: [PATCH v4] iio: adc: vf610: Add IIO buffer support for Vybrid ADC

Quoted 36 lines · Duan Andy

From: Sanchayan Maity <[email protected]> Sent: Monday, August 17, 2015 11:52 PM

Quoted 25 lines

To: jic23 at kernel.org; linux-iio at vger.kernel.org
Cc: stefan at agner.ch; Duan Fugang-B38611; pmeerw at pmeerw.net; linux- kernel at vger.kernel.org; linux-arm-kernel at lists.infradead.org; Sanchayan Maity
Subject: [PATCH v4] iio: adc: vf610: Add IIO buffer support for Vybrid ADC

This patch adds support for IIO buffer to the Vybrid ADC driver. IIO triggered buffer infrastructure along with iio sysfs trigger is used to leverage continuous sampling support provided by the ADC block.

---

Changes since v3:
Fix iio_buffer_setup_ops for postenable and predisable functions to match pairwise. Before this the predisable work was being done in postdisable.

Changes since v2:
1. Fix the wrong buffer size for statically allocated buffer 2. Drop the use of .address field from the iio_chan_spec 3. Use iio_buffer_enabled call inside the lock 4. Drop wrapper function around iio_trigered_* function calls 5. Drop Kconfig select of sysfs trigger 6. Drop Kconfig select IIO_TRIGGER as it is already selected by IIO_TRIGGERED_BUFFER

Quoted 9 lines

Changes since v1:
1. Use a fixed size buffer instead of kmalloc allocated during update scan mode 2. Remove a write to read only register ADC_HS (COCO bit)

 drivers/iio/adc/Kconfig     |   2 +
 drivers/iio/adc/vf610_adc.c | 105
+++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 100 insertions(+), 7 deletions(-)
Quoted 3 lines

The version is fine for me. Thanks for your effort.

Applied to the togreg branch of iio.git - will be initially pushed out as testing for the autobuilders to play with it.

Thanks,

Jonathan

Quoted 5 lines

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