Skip to contentSkip to conversation

staging: iio: ad9832: replace mlock with driver private lock

PATCH Applied Alison Schofield 2 participants 3 messages Mar 9, 2017 – Mar 11, 2017

Subsystems (4)

v1 of 2 Applied Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 staging: iio: ad9832: replace mlock with driver private lock

The IIO subsystem is redefining iio_dev->mlock to be used by the IIO core only for protecting device operating mode changes. ie. Changes between INDIO_DIRECT_MODE, INDIO_BUFFER_* modes.

In this driver, mlock was being used to protect hardware state changes. Replace it with a lock in the devices global data.

---

Jonathan et al,

If you still want to do this change, is it time to change the iio_dev->mlock comment back to [INTERN] in iio.h so that it doesn't get any new uses whilst we are migrating. Yes?

I'd like to give this as a task to the outreachy applicants. There are ~70 mlock usages spread amongst ~15 iio drivers in staging. Are you ready for a wave of these?

 drivers/staging/iio/frequency/ad9832.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/staging/iio/frequency/ad9832.c +4 -2
index 8d40c8e..c992648 100644
@@ -84,6 +84,7 @@
  * @freq_msg:		tuning word spi message
  * @phase_xfer:		tuning word spi transfer
  * @phase_msg:		tuning word spi message
+ * @lock		protect sensor state
  * @data:		spi transmit buffer
  * @phase_data:		tuning word spi transmit buffer
  * @freq_data:		tuning word spi transmit buffer
@@ -103,6 +104,7 @@ struct ad9832_state {
 	struct spi_message		freq_msg;
 	struct spi_transfer		phase_xfer[2];
 	struct spi_message		phase_msg;
+	struct mutex			lock;	/* protect sensor state */
 	/*
 	 * DMA (thus cache coherency maintenance) requires the
 	 * transfer buffers to live in their own cache lines.
@@ -177,7 +179,7 @@ static ssize_t ad9832_write(struct device *dev, struct device_attribute *attr,
 	if (ret)
 		goto error_ret;
 
-	mutex_lock(&indio_dev->mlock);
+	mutex_lock(&st->lock);
 	switch ((u32)this_attr->address) {
 	case AD9832_FREQ0HM:
 	case AD9832_FREQ1HM:
@@ -238,7 +240,7 @@ static ssize_t ad9832_write(struct device *dev, struct device_attribute *attr,
 	default:
 		ret = -ENODEV;
 	}
-	mutex_unlock(&indio_dev->mlock);
+	mutex_unlock(&st->lock);
 
 error_ret:
 	return ret ? ret : len;
Signature · 1 line
2.1.4
maintainer Applied
#1 lore raw
Quoted 8 lines · Alison Schofield

The IIO subsystem is redefining iio_dev->mlock to be used by the IIO core only for protecting device operating mode changes. ie. Changes between INDIO_DIRECT_MODE, INDIO_BUFFER_* modes.

In this driver, mlock was being used to protect hardware state changes. Replace it with a lock in the devices global data.

Applied to the togreg branch of iio.git and pushed out as testing for the autobuilders to play with it.

Thanks,

Jonathan

Quoted 11 lines

---

Jonathan et al,

If you still want to do this change, is it time to change the iio_dev->mlock comment back to [INTERN] in iio.h so that it doesn't get any new uses whilst we are migrating. Yes?

I'd like to give this as a task to the outreachy applicants. There are ~70 mlock usages spread amongst ~15 iio drivers in staging. Are you ready for a wave of these?

Yes. That was an easy question ;)

Make sure that they fully review the locking whilst at it though as chances are some of it may be less than ideal!

Jonathan

Quoted 44 lines

 drivers/staging/iio/frequency/ad9832.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/staging/iio/frequency/ad9832.c +4 -2
index 8d40c8e..c992648 100644
@@ -84,6 +84,7 @@
  * @freq_msg:		tuning word spi message
  * @phase_xfer:		tuning word spi transfer
  * @phase_msg:		tuning word spi message
+ * @lock		protect sensor state
  * @data:		spi transmit buffer
  * @phase_data:		tuning word spi transmit buffer
  * @freq_data:		tuning word spi transmit buffer
@@ -103,6 +104,7 @@ struct ad9832_state {
 	struct spi_message		freq_msg;
 	struct spi_transfer		phase_xfer[2];
 	struct spi_message		phase_msg;
+	struct mutex			lock;	/* protect sensor state */
 	/*
 	 * DMA (thus cache coherency maintenance) requires the
 	 * transfer buffers to live in their own cache lines.
@@ -177,7 +179,7 @@ static ssize_t ad9832_write(struct device *dev, struct device_attribute *attr,
 	if (ret)
 		goto error_ret;
 
-	mutex_lock(&indio_dev->mlock);
+	mutex_lock(&st->lock);
 	switch ((u32)this_attr->address) {
 	case AD9832_FREQ0HM:
 	case AD9832_FREQ1HM:
@@ -238,7 +240,7 @@ static ssize_t ad9832_write(struct device *dev, struct device_attribute *attr,
 	default:
 		ret = -ENODEV;
 	}
-	mutex_unlock(&indio_dev->mlock);
+	mutex_unlock(&st->lock);
 
 error_ret:
 	return ret ? ret : len;
#2 lore raw
Quoted 26 lines · Jonathan Cameron
Quoted 8 lines · Alison Schofield

The IIO subsystem is redefining iio_dev->mlock to be used by the IIO core only for protecting device operating mode changes. ie. Changes between INDIO_DIRECT_MODE, INDIO_BUFFER_* modes.

In this driver, mlock was being used to protect hardware state changes. Replace it with a lock in the devices global data.

Applied to the togreg branch of iio.git and pushed out as testing for the autobuilders to play with it.

Thanks,

Jonathan

Quoted 11 lines

---

Jonathan et al,

If you still want to do this change, is it time to change the iio_dev->mlock comment back to [INTERN] in iio.h so that it doesn't get any new uses whilst we are migrating. Yes?

I'd like to give this as a task to the outreachy applicants. There are ~70 mlock usages spread amongst ~15 iio drivers in staging. Are you ready for a wave of these?

Quoted 6 lines

Yes. That was an easy question ;)

Make sure that they fully review the locking whilst at it though as chances are some of it may be less than ideal!

Jonathan

Thanks Jonathan!
I hear you and I saw it when I selected this easy example ;) I'll certainly forewarn in the task description. alisons

Quoted 45 lines
Quoted 44 lines

 drivers/staging/iio/frequency/ad9832.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/staging/iio/frequency/ad9832.c +4 -2
index 8d40c8e..c992648 100644
@@ -84,6 +84,7 @@
  * @freq_msg:		tuning word spi message
  * @phase_xfer:		tuning word spi transfer
  * @phase_msg:		tuning word spi message
+ * @lock		protect sensor state
  * @data:		spi transmit buffer
  * @phase_data:		tuning word spi transmit buffer
  * @freq_data:		tuning word spi transmit buffer
@@ -103,6 +104,7 @@ struct ad9832_state {
 	struct spi_message		freq_msg;
 	struct spi_transfer		phase_xfer[2];
 	struct spi_message		phase_msg;
+	struct mutex			lock;	/* protect sensor state */
 	/*
 	 * DMA (thus cache coherency maintenance) requires the
 	 * transfer buffers to live in their own cache lines.
@@ -177,7 +179,7 @@ static ssize_t ad9832_write(struct device *dev, struct device_attribute *attr,
 	if (ret)
 		goto error_ret;
 
-	mutex_lock(&indio_dev->mlock);
+	mutex_lock(&st->lock);
 	switch ((u32)this_attr->address) {
 	case AD9832_FREQ0HM:
 	case AD9832_FREQ1HM:
@@ -238,7 +240,7 @@ static ssize_t ad9832_write(struct device *dev, struct device_attribute *attr,
 	default:
 		ret = -ENODEV;
 	}
-	mutex_unlock(&indio_dev->mlock);
+	mutex_unlock(&st->lock);
 
 error_ret:
 	return ret ? ret : len;