Skip to contentSkip to conversation

iio: magnetometer: bmc150_magn: cleanup and formatting

PATCH Superseded Neel Bullywon 5 participants 12 messages Feb 10 – Feb 14 IIO

v6 of 8 Superseded Latest is v8 →

lore
Share & tools
lore
#0 lore raw

v6 cover iio: magnetometer: bmc150_magn: cleanup and formatting

This series cleans up the bmc150_magn driver in three patches:

1/3: Replace manual mutex lock/unlock with guard()/scoped_guard() 2/3: Replace msleep(5) with fsleep(5 * USEC_PER_MSEC) 3/3: Minor formatting cleanup for initializer lists and indentation

Changes in v6:
- Split fsleep change into its own patch (Andy).
- Use scoped_guard() instead of { guard() } for the IIO_CHAN_INFO_RAW case block in read_raw (Andy).
- Use 5 * USEC_PER_MSEC instead of 5000 for fsleep (Andy).
- Restyle bmc150_magn_samp_freq_table to group four entries per line with index comments (Andy).
- Remove trailing comma from scan_masks terminator (Andy).

Changes in v5:
- Split into two patches: functional and formatting (Jonathan).
- Use fsleep() instead of usleep_range() (Jonathan).
- Use scoped_guard() for short single-statement scopes (Jonathan).
- Add {} scope for guard() in case blocks (Jonathan).
- Leave trigger_handler unchanged (Jonathan).
- Drop all unnecessary reformatting (Jonathan).

Changes in v4:
- Extend guard() usage to all mutex_lock() instances in the driver.
- Replace msleep(5) with usleep_range(5000, 6000).

Changes in v3:
- Add Reviewed-by tags.

Changes in v2:
- Use guard() for mutex protection in bmc150_magn_data_rdy_trigger_set_state.

Neel Bullywon (3):
iio: magnetometer: bmc150_magn: use automated cleanup for mutex
iio: magnetometer: bmc150_magn: replace msleep with fsleep
iio: magnetometer: bmc150_magn: minor formatting cleanup

 drivers/iio/magnetometer/bmc150_magn.c | 157 ++++++++++---------------
 1 file changed, 64 insertions(+), 93 deletions(-)

base-commit: e7aa57247700733e52a8e2e4dee6a52c2a76de02

Signature · 1 line
2.44.0
#1 lore raw

PATCH v6 1/3 iio: magnetometer: bmc150_magn: use automated cleanup for mutex

Use guard() and scoped_guard() to replace manual mutex lock/unlock calls. This simplifies error handling and ensures RAII-style cleanup.

scoped_guard() is used in read_raw IIO_CHAN_INFO_RAW case for a multi-statement mutex-protected block, as well as in remove, runtime_suspend, and suspend where a short mutex-protected scope is needed for a single function call.

guard() is used in write_raw and trig_reen where the mutex scope extends to the end of the function or case block. Case blocks in write_raw are wrapped in braces to ensure clear scope for the cleanup guards.

The trigger_handler function is left unchanged as mixing guard() with goto error paths can be fragile.

---
 drivers/iio/magnetometer/bmc150_magn.c | 130 ++++++++++---------------
 1 file changed, 52 insertions(+), 78 deletions(-)
drivers/iio/magnetometer/bmc150_magn.c +52 -78
index 6a73f6e2f1f0..5eb710eb020d 100644
@@ -11,6 +11,7 @@
 
 #include <linux/module.h>
 #include <linux/i2c.h>
+#include <linux/cleanup.h>
 #include <linux/interrupt.h>
 #include <linux/delay.h>
 #include <linux/slab.h>
@@ -456,29 +457,22 @@ static int bmc150_magn_read_raw(struct iio_dev *indio_dev,
 	case IIO_CHAN_INFO_RAW:
 		if (iio_buffer_enabled(indio_dev))
 			return -EBUSY;
-		mutex_lock(&data->mutex);
-
-		ret = bmc150_magn_set_power_state(data, true);
-		if (ret < 0) {
-			mutex_unlock(&data->mutex);
-			return ret;
-		}
+		scoped_guard(mutex, &data->mutex) {
+			ret = bmc150_magn_set_power_state(data, true);
+			if (ret < 0)
+				return ret;
 
-		ret = bmc150_magn_read_xyz(data, values);
-		if (ret < 0) {
-			bmc150_magn_set_power_state(data, false);
-			mutex_unlock(&data->mutex);
-			return ret;
-		}
-		*val = values[chan->scan_index];
+			ret = bmc150_magn_read_xyz(data, values);
+			if (ret < 0) {
+				bmc150_magn_set_power_state(data, false);
+				return ret;
+			}
+			*val = values[chan->scan_index];
 
-		ret = bmc150_magn_set_power_state(data, false);
-		if (ret < 0) {
-			mutex_unlock(&data->mutex);
-			return ret;
+			ret = bmc150_magn_set_power_state(data, false);
+			if (ret < 0)
+				return ret;
 		}
-
-		mutex_unlock(&data->mutex);
 		return IIO_VAL_INT;
 	case IIO_CHAN_INFO_SCALE:
 		/*
@@ -530,45 +524,37 @@ static int bmc150_magn_write_raw(struct iio_dev *indio_dev,
 	case IIO_CHAN_INFO_SAMP_FREQ:
 		if (val > data->max_odr)
 			return -EINVAL;
-		mutex_lock(&data->mutex);
-		ret = bmc150_magn_set_odr(data, val);
-		mutex_unlock(&data->mutex);
-		return ret;
+		guard(mutex)(&data->mutex);
+		return bmc150_magn_set_odr(data, val);
 	case IIO_CHAN_INFO_OVERSAMPLING_RATIO:
 		switch (chan->channel2) {
 		case IIO_MOD_X:
 		case IIO_MOD_Y:
 			if (val < 1 || val > 511)
 				return -EINVAL;
-			mutex_lock(&data->mutex);
-			ret = bmc150_magn_set_max_odr(data, val, 0, 0);
-			if (ret < 0) {
-				mutex_unlock(&data->mutex);
-				return ret;
+			{
+				guard(mutex)(&data->mutex);
+				ret = bmc150_magn_set_max_odr(data, val, 0, 0);
+				if (ret < 0)
+					return ret;
+				return regmap_update_bits(data->regmap,
+							 BMC150_MAGN_REG_REP_XY,
+							 BMC150_MAGN_REG_REP_DATAMASK,
+							 BMC150_MAGN_REPXY_TO_REGVAL(val));
 			}
-			ret = regmap_update_bits(data->regmap,
-						 BMC150_MAGN_REG_REP_XY,
-						 BMC150_MAGN_REG_REP_DATAMASK,
-						 BMC150_MAGN_REPXY_TO_REGVAL
-						 (val));
-			mutex_unlock(&data->mutex);
-			return ret;
 		case IIO_MOD_Z:
 			if (val < 1 || val > 256)
 				return -EINVAL;
-			mutex_lock(&data->mutex);
-			ret = bmc150_magn_set_max_odr(data, 0, val, 0);
-			if (ret < 0) {
-				mutex_unlock(&data->mutex);
-				return ret;
+			{
+				guard(mutex)(&data->mutex);
+				ret = bmc150_magn_set_max_odr(data, 0, val, 0);
+				if (ret < 0)
+					return ret;
+				return regmap_update_bits(data->regmap,
+							 BMC150_MAGN_REG_REP_Z,
+							 BMC150_MAGN_REG_REP_DATAMASK,
+							 BMC150_MAGN_REPZ_TO_REGVAL(val));
 			}
-			ret = regmap_update_bits(data->regmap,
-						 BMC150_MAGN_REG_REP_Z,
-						 BMC150_MAGN_REG_REP_DATAMASK,
-						 BMC150_MAGN_REPZ_TO_REGVAL
-						 (val));
-			mutex_unlock(&data->mutex);
-			return ret;
 		default:
 			return -EINVAL;
 		}
@@ -782,9 +768,8 @@ static void bmc150_magn_trig_reen(struct iio_trigger *trig)
 	if (!data->dready_trigger_on)
 		return;
 
-	mutex_lock(&data->mutex);
+	guard(mutex)(&data->mutex);
 	ret = bmc150_magn_reset_intr(data);
-	mutex_unlock(&data->mutex);
 	if (ret)
 		dev_err(data->dev, "Failed to reset interrupt\n");
 }
@@ -794,32 +779,28 @@ static int bmc150_magn_data_rdy_trigger_set_state(struct iio_trigger *trig,
 {
 	struct iio_dev *indio_dev = iio_trigger_get_drvdata(trig);
 	struct bmc150_magn_data *data = iio_priv(indio_dev);
-	int ret = 0;
+	int ret;
+
+	guard(mutex)(&data->mutex);
 
-	mutex_lock(&data->mutex);
 	if (state == data->dready_trigger_on)
-		goto err_unlock;
+		return 0;
 
 	ret = regmap_update_bits(data->regmap, BMC150_MAGN_REG_INT_DRDY,
 				 BMC150_MAGN_MASK_DRDY_EN,
 				 state << BMC150_MAGN_SHIFT_DRDY_EN);
 	if (ret < 0)
-		goto err_unlock;
+		return ret;
 
 	data->dready_trigger_on = state;
 
 	if (state) {
 		ret = bmc150_magn_reset_intr(data);
 		if (ret < 0)
-			goto err_unlock;
+			return ret;
 	}
-	mutex_unlock(&data->mutex);
 
 	return 0;
-
-err_unlock:
-	mutex_unlock(&data->mutex);
-	return ret;
 }
 
 static const struct iio_trigger_ops bmc150_magn_trigger_ops = {
@@ -980,9 +961,8 @@ void bmc150_magn_remove(struct device *dev)
 	if (data->dready_trig)
 		iio_trigger_unregister(data->dready_trig);
 
-	mutex_lock(&data->mutex);
-	bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SUSPEND, true);
-	mutex_unlock(&data->mutex);
+	scoped_guard(mutex, &data->mutex)
+		bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SUSPEND, true);
 
 	regulator_bulk_disable(ARRAY_SIZE(data->regulators), data->regulators);
 }
@@ -995,10 +975,9 @@ static int bmc150_magn_runtime_suspend(struct device *dev)
 	struct bmc150_magn_data *data = iio_priv(indio_dev);
 	int ret;
 
-	mutex_lock(&data->mutex);
-	ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
-					 true);
-	mutex_unlock(&data->mutex);
+	scoped_guard(mutex, &data->mutex)
+		ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
+						 true);
 	if (ret < 0) {
 		dev_err(dev, "powering off device failed\n");
 		return ret;
@@ -1026,10 +1005,9 @@ static int bmc150_magn_suspend(struct device *dev)
 	struct bmc150_magn_data *data = iio_priv(indio_dev);
 	int ret;
 
-	mutex_lock(&data->mutex);
-	ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
-					 true);
-	mutex_unlock(&data->mutex);
+	scoped_guard(mutex, &data->mutex)
+		ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
+						 true);
 
 	return ret;
 }
@@ -1038,14 +1016,10 @@ static int bmc150_magn_resume(struct device *dev)
 {
 	struct iio_dev *indio_dev = dev_get_drvdata(dev);
 	struct bmc150_magn_data *data = iio_priv(indio_dev);
-	int ret;
-
-	mutex_lock(&data->mutex);
-	ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_NORMAL,
-					 true);
-	mutex_unlock(&data->mutex);
 
-	return ret;
+	guard(mutex)(&data->mutex);
+	return bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_NORMAL,
+					  true);
 }
 #endif
 
Signature · 1 line
2.44.0
#2 lore raw

Re: [PATCH v6 1/3] iio: magnetometer: bmc150_magn: use automated cleanup for mutex

Quoted 15 lines · Neel Bullywon

Use guard() and scoped_guard() to replace manual mutex lock/unlock calls. This simplifies error handling and ensures RAII-style cleanup.

scoped_guard() is used in read_raw IIO_CHAN_INFO_RAW case for a multi-statement mutex-protected block, as well as in remove, runtime_suspend, and suspend where a short mutex-protected scope is needed for a single function call.

guard() is used in write_raw and trig_reen where the mutex scope extends to the end of the function or case block. Case blocks in write_raw are wrapped in braces to ensure clear scope for the cleanup guards.

The trigger_handler function is left unchanged as mixing guard() with goto error paths can be fragile.

...

Quoted 6 lines

#include <linux/module.h>
#include <linux/i2c.h>
+#include <linux/cleanup.h>
#include <linux/interrupt.h>
#include <linux/delay.h>
#include <linux/slab.h>

With a given context I would place the new one just before the delay.h.

...

Quoted 30 lines
drivers/iio/magnetometer/bmc150_magn.c +10 -15
 	case IIO_CHAN_INFO_RAW:
 		if (iio_buffer_enabled(indio_dev))
 			return -EBUSY;
-		mutex_lock(&data->mutex);
-
-		ret = bmc150_magn_set_power_state(data, true);
-		if (ret < 0) {
-			mutex_unlock(&data->mutex);
-			return ret;
-		}
+		scoped_guard(mutex, &data->mutex) {
+			ret = bmc150_magn_set_power_state(data, true);
+			if (ret < 0)
+				return ret;
 
-		ret = bmc150_magn_read_xyz(data, values);
-		if (ret < 0) {
-			bmc150_magn_set_power_state(data, false);
-			mutex_unlock(&data->mutex);
-			return ret;
-		}
-		*val = values[chan->scan_index];
+			ret = bmc150_magn_read_xyz(data, values);
+			if (ret < 0) {
+				bmc150_magn_set_power_state(data, false);
+				return ret;
+			}
+			*val = values[chan->scan_index];
 
-		ret = bmc150_magn_set_power_state(data, false);

drivers/iio/magnetometer/bmc150_magn.c static int bmc150_magn_read_raw(struct iio_dev *indio_dev, · line 476

Quoted 10 lines
drivers/iio/magnetometer/bmc150_magn.c +3 -5
-		if (ret < 0) {
-			mutex_unlock(&data->mutex);
-			return ret;
+			ret = bmc150_magn_set_power_state(data, false);
+			if (ret < 0)
+				return ret;
 		}
-
-		mutex_unlock(&data->mutex);
 		return IIO_VAL_INT;

Actually looking again at this, I think we may use guard()(), but move the {} to the whole case:

	case IIO_CHAN_INFO_RAW: {
		if (iio_buffer_enabled(indio_dev))
			return -EBUSY;

		guard(mutex)(&data->mutex);
		...
	}

This will make the change much more cleaner.

...

drivers/iio/magnetometer/bmc150_magn.c static int bmc150_magn_write_raw(struct iio_dev *indio_dev, · line 549

Quoted 3 lines
drivers/iio/magnetometer/bmc150_magn.c +2 -0
 		case IIO_MOD_Z:
+			{
+				guard(mutex)(&data->mutex);

Why do you change only one case?! If the comment is given, it applies to all similar places. But see also above.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#3 lore raw

Re: [PATCH v6 1/3] iio: magnetometer: bmc150_magn: use automated cleanup for mutex

In reply to Neel Bullywon
kernel test robot noticed the following build errors:

Hi Neel,

kernel test robot noticed the following build errors:

[auto build test ERROR on e7aa57247700733e52a8e2e4dee6a52c2a76de02]

url:    https://github.com/intel-lab-lkp/linux/commits/Neel-Bullywon/iio-magnetometer-bmc150_magn-use-automated-cleanup-for-mutex/20260211-074324
base:   e7aa57247700733e52a8e2e4dee6a52c2a76de02
patch link:    https://lore.kernel.org/r/20260210233945.40975-2-neelb2403%40gmail.com
patch subject: [PATCH v6 1/3] iio: magnetometer: bmc150_magn: use automated cleanup for mutex
config: loongarch-allmodconfig (https://download.01.org/0day-ci/archive/20260214/[email protected]/config)
compiler: clang version 19.1.7 (https://github.com/llvm/llvm-project cd708029e0b2869e80abe31ddb175f7c35361f90)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20260214/[email protected]/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/

All errors (new ones prefixed by >>):

>> drivers/iio/magnetometer/bmc150_magn.c:561:2: error: cannot jump from switch statement to this case label
     561 |         default:
         |         ^
   drivers/iio/magnetometer/bmc150_magn.c:527:3: note: jump bypasses initialization of variable with __attribute__((cleanup))
     527 |                 guard(mutex)(&data->mutex);
         |                 ^
   include/linux/cleanup.h:414:15: note: expanded from macro 'guard'
     414 |         CLASS(_name, __UNIQUE_ID(guard))
         |                      ^
   include/linux/compiler.h:168:2: note: expanded from macro '__UNIQUE_ID'
     168 |         __PASTE(__UNIQUE_ID_,                                   \
         |         ^
   include/linux/compiler_types.h:16:23: note: expanded from macro '__PASTE'
      16 | #define __PASTE(a, b) ___PASTE(a, b)
         |                       ^
   include/linux/compiler_types.h:15:24: note: expanded from macro '___PASTE'
      15 | #define ___PASTE(a, b) a##b
         |                        ^
   <scratch space>:122:1: note: expanded from here
     122 | __UNIQUE_ID_guard_677
         | ^
   drivers/iio/magnetometer/bmc150_magn.c:529:2: error: cannot jump from switch statement to this case label
     529 |         case IIO_CHAN_INFO_OVERSAMPLING_RATIO:
         |         ^
   drivers/iio/magnetometer/bmc150_magn.c:527:3: note: jump bypasses initialization of variable with __attribute__((cleanup))
     527 |                 guard(mutex)(&data->mutex);
         |                 ^
   include/linux/cleanup.h:414:15: note: expanded from macro 'guard'
     414 |         CLASS(_name, __UNIQUE_ID(guard))
         |                      ^
   include/linux/compiler.h:168:2: note: expanded from macro '__UNIQUE_ID'
     168 |         __PASTE(__UNIQUE_ID_,                                   \
         |         ^
   include/linux/compiler_types.h:16:23: note: expanded from macro '__PASTE'
      16 | #define __PASTE(a, b) ___PASTE(a, b)
         |                       ^
   include/linux/compiler_types.h:15:24: note: expanded from macro '___PASTE'
      15 | #define ___PASTE(a, b) a##b
         |                        ^
   <scratch space>:122:1: note: expanded from here
     122 | __UNIQUE_ID_guard_677
         | ^

2 errors generated.

vim +561 drivers/iio/magnetometer/bmc150_magn.c

c91746a2361d75 Irina Tirdea   2015-04-29  515  
c91746a2361d75 Irina Tirdea   2015-04-29  516  static int bmc150_magn_write_raw(struct iio_dev *indio_dev,
c91746a2361d75 Irina Tirdea   2015-04-29  517  				 struct iio_chan_spec const *chan,
c91746a2361d75 Irina Tirdea   2015-04-29  518  				 int val, int val2, long mask)
c91746a2361d75 Irina Tirdea   2015-04-29  519  {
c91746a2361d75 Irina Tirdea   2015-04-29  520  	struct bmc150_magn_data *data = iio_priv(indio_dev);
c91746a2361d75 Irina Tirdea   2015-04-29  521  	int ret;
c91746a2361d75 Irina Tirdea   2015-04-29  522  
c91746a2361d75 Irina Tirdea   2015-04-29  523  	switch (mask) {
c91746a2361d75 Irina Tirdea   2015-04-29  524  	case IIO_CHAN_INFO_SAMP_FREQ:
5990dc9703679a Irina Tirdea   2015-04-29  525  		if (val > data->max_odr)
5990dc9703679a Irina Tirdea   2015-04-29  526  			return -EINVAL;
d65a37daae6aa8 Neel Bullywon  2026-02-10  527  		guard(mutex)(&data->mutex);
d65a37daae6aa8 Neel Bullywon  2026-02-10  528  		return bmc150_magn_set_odr(data, val);
5990dc9703679a Irina Tirdea   2015-04-29  529  	case IIO_CHAN_INFO_OVERSAMPLING_RATIO:
5990dc9703679a Irina Tirdea   2015-04-29  530  		switch (chan->channel2) {
5990dc9703679a Irina Tirdea   2015-04-29  531  		case IIO_MOD_X:
5990dc9703679a Irina Tirdea   2015-04-29  532  		case IIO_MOD_Y:
5990dc9703679a Irina Tirdea   2015-04-29  533  			if (val < 1 || val > 511)
5990dc9703679a Irina Tirdea   2015-04-29  534  				return -EINVAL;
d65a37daae6aa8 Neel Bullywon  2026-02-10  535  			{
d65a37daae6aa8 Neel Bullywon  2026-02-10  536  				guard(mutex)(&data->mutex);
5990dc9703679a Irina Tirdea   2015-04-29  537  				ret = bmc150_magn_set_max_odr(data, val, 0, 0);
d65a37daae6aa8 Neel Bullywon  2026-02-10  538  				if (ret < 0)
5990dc9703679a Irina Tirdea   2015-04-29  539  					return ret;
d65a37daae6aa8 Neel Bullywon  2026-02-10  540  				return regmap_update_bits(data->regmap,
5990dc9703679a Irina Tirdea   2015-04-29  541  							 BMC150_MAGN_REG_REP_XY,
1506f3cd0b8169 Hartmut Knaack 2015-07-17  542  							 BMC150_MAGN_REG_REP_DATAMASK,
d65a37daae6aa8 Neel Bullywon  2026-02-10  543  							 BMC150_MAGN_REPXY_TO_REGVAL(val));
d65a37daae6aa8 Neel Bullywon  2026-02-10  544  			}
5990dc9703679a Irina Tirdea   2015-04-29  545  		case IIO_MOD_Z:
5990dc9703679a Irina Tirdea   2015-04-29  546  			if (val < 1 || val > 256)
5990dc9703679a Irina Tirdea   2015-04-29  547  				return -EINVAL;
d65a37daae6aa8 Neel Bullywon  2026-02-10  548  			{
d65a37daae6aa8 Neel Bullywon  2026-02-10  549  				guard(mutex)(&data->mutex);
5990dc9703679a Irina Tirdea   2015-04-29  550  				ret = bmc150_magn_set_max_odr(data, 0, val, 0);
d65a37daae6aa8 Neel Bullywon  2026-02-10  551  				if (ret < 0)
5990dc9703679a Irina Tirdea   2015-04-29  552  					return ret;
d65a37daae6aa8 Neel Bullywon  2026-02-10  553  				return regmap_update_bits(data->regmap,
5990dc9703679a Irina Tirdea   2015-04-29  554  							 BMC150_MAGN_REG_REP_Z,
1506f3cd0b8169 Hartmut Knaack 2015-07-17  555  							 BMC150_MAGN_REG_REP_DATAMASK,
d65a37daae6aa8 Neel Bullywon  2026-02-10  556  							 BMC150_MAGN_REPZ_TO_REGVAL(val));
d65a37daae6aa8 Neel Bullywon  2026-02-10  557  			}
5990dc9703679a Irina Tirdea   2015-04-29  558  		default:
5990dc9703679a Irina Tirdea   2015-04-29  559  			return -EINVAL;
5990dc9703679a Irina Tirdea   2015-04-29  560  		}
c91746a2361d75 Irina Tirdea   2015-04-29 @561  	default:
c91746a2361d75 Irina Tirdea   2015-04-29  562  		return -EINVAL;
c91746a2361d75 Irina Tirdea   2015-04-29  563  	}
c91746a2361d75 Irina Tirdea   2015-04-29  564  }
c91746a2361d75 Irina Tirdea   2015-04-29  565  
Signature · 2 lines
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
maintainer
#4 lore raw

Re: [PATCH v6 1/3] iio: magnetometer: bmc150_magn: use automated cleanup for mutex

Quoted 3 lines · On Sat, 14 Feb 2026 07:31:03 +0800 kernel test robot <[email protected]> wrote:

Hi Neel,

kernel test robot noticed the following build errors:

Just replying to this to highlight what I the whole scope thing is about.

Quoted 65 lines

[auto build test ERROR on e7aa57247700733e52a8e2e4dee6a52c2a76de02]

url:    https://github.com/intel-lab-lkp/linux/commits/Neel-Bullywon/iio-magnetometer-bmc150_magn-use-automated-cleanup-for-mutex/20260211-074324
base:   e7aa57247700733e52a8e2e4dee6a52c2a76de02
patch link:    https://lore.kernel.org/r/20260210233945.40975-2-neelb2403%40gmail.com
patch subject: [PATCH v6 1/3] iio: magnetometer: bmc150_magn: use automated cleanup for mutex
config: loongarch-allmodconfig (https://download.01.org/0day-ci/archive/20260214/[email protected]/config)
compiler: clang version 19.1.7 (https://github.com/llvm/llvm-project cd708029e0b2869e80abe31ddb175f7c35361f90)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20260214/[email protected]/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/

All errors (new ones prefixed by >>):

Quoted 1 line
Quoted 1 line

drivers/iio/magnetometer/bmc150_magn.c:561:2: error: cannot jump from switch statement to this case label

     561 |         default:
         |         ^
   drivers/iio/magnetometer/bmc150_magn.c:527:3: note: jump bypasses initialization of variable with __attribute__((cleanup))
     527 |                 guard(mutex)(&data->mutex);
         |                 ^
   include/linux/cleanup.h:414:15: note: expanded from macro 'guard'
     414 |         CLASS(_name, __UNIQUE_ID(guard))
         |                      ^
   include/linux/compiler.h:168:2: note: expanded from macro '__UNIQUE_ID'
     168 |         __PASTE(__UNIQUE_ID_,                                   \
         |         ^
   include/linux/compiler_types.h:16:23: note: expanded from macro '__PASTE'
      16 | #define __PASTE(a, b) ___PASTE(a, b)
         |                       ^
   include/linux/compiler_types.h:15:24: note: expanded from macro '___PASTE'
      15 | #define ___PASTE(a, b) a##b
         |                        ^
   <scratch space>:122:1: note: expanded from here
     122 | __UNIQUE_ID_guard_677
         | ^
   drivers/iio/magnetometer/bmc150_magn.c:529:2: error: cannot jump from switch statement to this case label
     529 |         case IIO_CHAN_INFO_OVERSAMPLING_RATIO:
         |         ^
   drivers/iio/magnetometer/bmc150_magn.c:527:3: note: jump bypasses initialization of variable with __attribute__((cleanup))
     527 |                 guard(mutex)(&data->mutex);
         |                 ^
   include/linux/cleanup.h:414:15: note: expanded from macro 'guard'
     414 |         CLASS(_name, __UNIQUE_ID(guard))
         |                      ^
   include/linux/compiler.h:168:2: note: expanded from macro '__UNIQUE_ID'
     168 |         __PASTE(__UNIQUE_ID_,                                   \
         |         ^
   include/linux/compiler_types.h:16:23: note: expanded from macro '__PASTE'
      16 | #define __PASTE(a, b) ___PASTE(a, b)
         |                       ^
   include/linux/compiler_types.h:15:24: note: expanded from macro '___PASTE'
      15 | #define ___PASTE(a, b) a##b
         |                        ^
   <scratch space>:122:1: note: expanded from here
     122 | __UNIQUE_ID_guard_677
         | ^

2 errors generated.

vim +561 drivers/iio/magnetometer/bmc150_magn.c

Quoted 13 lines
c91746a2361d75 Irina Tirdea   2015-04-29  515  
c91746a2361d75 Irina Tirdea   2015-04-29  516  static int bmc150_magn_write_raw(struct iio_dev *indio_dev,
c91746a2361d75 Irina Tirdea   2015-04-29  517  				 struct iio_chan_spec const *chan,
c91746a2361d75 Irina Tirdea   2015-04-29  518  				 int val, int val2, long mask)
c91746a2361d75 Irina Tirdea   2015-04-29  519  {
c91746a2361d75 Irina Tirdea   2015-04-29  520  	struct bmc150_magn_data *data = iio_priv(indio_dev);
c91746a2361d75 Irina Tirdea   2015-04-29  521  	int ret;
c91746a2361d75 Irina Tirdea   2015-04-29  522  
c91746a2361d75 Irina Tirdea   2015-04-29  523  	switch (mask) {
c91746a2361d75 Irina Tirdea   2015-04-29  524  	case IIO_CHAN_INFO_SAMP_FREQ:
5990dc9703679a Irina Tirdea   2015-04-29  525  		if (val > data->max_odr)
5990dc9703679a Irina Tirdea   2015-04-29  526  			return -EINVAL;
d65a37daae6aa8 Neel Bullywon  2026-02-10  527  		guard(mutex)(&data->mutex);

It's hard to spot from the error message, but here is the bug. This is in the same scope as default label below that is hightlighted.
Thus wonderfully the __free() that hides under the guard() magic is initialized but because we don't go down this path to reach that default, we can the __free() (ultimately mutex_unlock())
on an unassigned pointer.

Hence scope for these really matters!

Jonathan

Quoted 39 lines
d65a37daae6aa8 Neel Bullywon  2026-02-10  528  		return bmc150_magn_set_odr(data, val);
5990dc9703679a Irina Tirdea   2015-04-29  529  	case IIO_CHAN_INFO_OVERSAMPLING_RATIO:
5990dc9703679a Irina Tirdea   2015-04-29  530  		switch (chan->channel2) {
5990dc9703679a Irina Tirdea   2015-04-29  531  		case IIO_MOD_X:
5990dc9703679a Irina Tirdea   2015-04-29  532  		case IIO_MOD_Y:
5990dc9703679a Irina Tirdea   2015-04-29  533  			if (val < 1 || val > 511)
5990dc9703679a Irina Tirdea   2015-04-29  534  				return -EINVAL;
d65a37daae6aa8 Neel Bullywon  2026-02-10  535  			{
d65a37daae6aa8 Neel Bullywon  2026-02-10  536  				guard(mutex)(&data->mutex);
5990dc9703679a Irina Tirdea   2015-04-29  537  				ret = bmc150_magn_set_max_odr(data, val, 0, 0);
d65a37daae6aa8 Neel Bullywon  2026-02-10  538  				if (ret < 0)
5990dc9703679a Irina Tirdea   2015-04-29  539  					return ret;
d65a37daae6aa8 Neel Bullywon  2026-02-10  540  				return regmap_update_bits(data->regmap,
5990dc9703679a Irina Tirdea   2015-04-29  541  							 BMC150_MAGN_REG_REP_XY,
1506f3cd0b8169 Hartmut Knaack 2015-07-17  542  							 BMC150_MAGN_REG_REP_DATAMASK,
d65a37daae6aa8 Neel Bullywon  2026-02-10  543  							 BMC150_MAGN_REPXY_TO_REGVAL(val));
d65a37daae6aa8 Neel Bullywon  2026-02-10  544  			}
5990dc9703679a Irina Tirdea   2015-04-29  545  		case IIO_MOD_Z:
5990dc9703679a Irina Tirdea   2015-04-29  546  			if (val < 1 || val > 256)
5990dc9703679a Irina Tirdea   2015-04-29  547  				return -EINVAL;
d65a37daae6aa8 Neel Bullywon  2026-02-10  548  			{
d65a37daae6aa8 Neel Bullywon  2026-02-10  549  				guard(mutex)(&data->mutex);
5990dc9703679a Irina Tirdea   2015-04-29  550  				ret = bmc150_magn_set_max_odr(data, 0, val, 0);
d65a37daae6aa8 Neel Bullywon  2026-02-10  551  				if (ret < 0)
5990dc9703679a Irina Tirdea   2015-04-29  552  					return ret;
d65a37daae6aa8 Neel Bullywon  2026-02-10  553  				return regmap_update_bits(data->regmap,
5990dc9703679a Irina Tirdea   2015-04-29  554  							 BMC150_MAGN_REG_REP_Z,
1506f3cd0b8169 Hartmut Knaack 2015-07-17  555  							 BMC150_MAGN_REG_REP_DATAMASK,
d65a37daae6aa8 Neel Bullywon  2026-02-10  556  							 BMC150_MAGN_REPZ_TO_REGVAL(val));
d65a37daae6aa8 Neel Bullywon  2026-02-10  557  			}
5990dc9703679a Irina Tirdea   2015-04-29  558  		default:
5990dc9703679a Irina Tirdea   2015-04-29  559  			return -EINVAL;
5990dc9703679a Irina Tirdea   2015-04-29  560  		}
c91746a2361d75 Irina Tirdea   2015-04-29 @561  	default:
c91746a2361d75 Irina Tirdea   2015-04-29  562  		return -EINVAL;
c91746a2361d75 Irina Tirdea   2015-04-29  563  	}
c91746a2361d75 Irina Tirdea   2015-04-29  564  }

c91746a2361d75 Irina Tirdea 2015-04-29 565

reviewer
#5 lore raw

Re: [PATCH v6 1/3] iio: magnetometer: bmc150_magn: use automated cleanup for mutex

In reply to Neel Bullywon
Quoted 35 lines · Neel Bullywon

Use guard() and scoped_guard() to replace manual mutex lock/unlock calls. This simplifies error handling and ensures RAII-style cleanup.

scoped_guard() is used in read_raw IIO_CHAN_INFO_RAW case for a multi-statement mutex-protected block, as well as in remove, runtime_suspend, and suspend where a short mutex-protected scope is needed for a single function call.

guard() is used in write_raw and trig_reen where the mutex scope extends to the end of the function or case block. Case blocks in write_raw are wrapped in braces to ensure clear scope for the cleanup guards.

The trigger_handler function is left unchanged as mixing guard() with goto error paths can be fragile.

---
 drivers/iio/magnetometer/bmc150_magn.c | 130 ++++++++++---------------
 1 file changed, 52 insertions(+), 78 deletions(-)
drivers/iio/magnetometer/bmc150_magn.c +1 -0
index 6a73f6e2f1f0..5eb710eb020d 100644
@@ -11,6 +11,7 @@
 
 #include <linux/module.h>
 #include <linux/i2c.h>
+#include <linux/cleanup.h>
 #include <linux/interrupt.h>
 #include <linux/delay.h>
 #include <linux/slab.h>
@@ -456,29 +457,22 @@ static int bmc150_magn_read_raw(struct iio_dev *indio_dev,
 	case IIO_CHAN_INFO_RAW:

drivers/iio/magnetometer/bmc150_magn.c static int bmc150_magn_read_raw(struct iio_dev *indio_dev, · line 460

Quoted 10 lines
drivers/iio/magnetometer/bmc150_magn.c +1 -7
 		if (iio_buffer_enabled(indio_dev))
 			return -EBUSY;
-		mutex_lock(&data->mutex);
-
-		ret = bmc150_magn_set_power_state(data, true);
-		if (ret < 0) {
-			mutex_unlock(&data->mutex);
-			return ret;
-		}
+		scoped_guard(mutex, &data->mutex) {

I am very wary of scoped_guard(), especially in case statements. There is a hidden for loop in the macro, so if break; is used inside of the scoped_guard() {}, it will only break out of the guard scope (for loop) and not the case.

It would be better if we could use guard() instead (in the style that Andy suggested).

Quoted 14 lines
drivers/iio/magnetometer/bmc150_magn.c +2 -3
 static const struct iio_trigger_ops bmc150_magn_trigger_ops = {
@@ -980,9 +961,8 @@ void bmc150_magn_remove(struct device *dev)
 	if (data->dready_trig)
 		iio_trigger_unregister(data->dready_trig);
 
-	mutex_lock(&data->mutex);
-	bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SUSPEND, true);
-	mutex_unlock(&data->mutex);
+	scoped_guard(mutex, &data->mutex)
+		bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SUSPEND, true);
 
 	regulator_bulk_disable(ARRAY_SIZE(data->regulators), data->regulators);
 }

drivers/iio/magnetometer/bmc150_magn.c static int bmc150_magn_runtime_suspend(struct device *dev) · line 980

Quoted 10 lines
drivers/iio/magnetometer/bmc150_magn.c +3 -4
@@ -995,10 +975,9 @@ static int bmc150_magn_runtime_suspend(struct device *dev)
 	struct bmc150_magn_data *data = iio_priv(indio_dev);
 	int ret;
 
-	mutex_lock(&data->mutex);
-	ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
-					 true);
-	mutex_unlock(&data->mutex);
+	scoped_guard(mutex, &data->mutex)
+		ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
+						 true);

This way of using scoped_guard() is OK since there is no { }.

Quoted 4 lines
drivers/iio/magnetometer/bmc150_magn.c +0 -0
 	if (ret < 0) {
 		dev_err(dev, "powering off device failed\n");
 		return ret;

drivers/iio/magnetometer/bmc150_magn.c static int bmc150_magn_suspend(struct device *dev) · line 1010

Quoted 10 lines
drivers/iio/magnetometer/bmc150_magn.c +3 -4
@@ -1026,10 +1005,9 @@ static int bmc150_magn_suspend(struct device *dev)
 	struct bmc150_magn_data *data = iio_priv(indio_dev);
 	int ret;
 
-	mutex_lock(&data->mutex);
-	ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
-					 true);
-	mutex_unlock(&data->mutex);
+	scoped_guard(mutex, &data->mutex)
+		ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SLEEP,
+						 true);

This one can be regular guard() and return directly.

Quoted 3 lines
drivers/iio/magnetometer/bmc150_magn.c +0 -0
 	return ret;
 }
#6 lore raw

PATCH v6 2/3 iio: magnetometer: bmc150_magn: replace msleep with fsleep

In reply to Neel Bullywon

Replace msleep(5) with fsleep(5 * USEC_PER_MSEC) to allow the kernel to select the most appropriate delay mechanism based on duration. Using USEC_PER_MSEC makes the unit conversion explicit.

---
 drivers/iio/magnetometer/bmc150_magn.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/iio/magnetometer/bmc150_magn.c +1 -1
index 5eb710eb020d..f254774d28e5 100644
@@ -681,7 +681,7 @@ static int bmc150_magn_init(struct bmc150_magn_data *data)
 	 * 3ms power-on time according to datasheet, let's better
 	 * be safe than sorry and set this delay to 5ms.
 	 */
-	msleep(5);
+	fsleep(5 * USEC_PER_MSEC);
 
 	ret = bmc150_magn_set_power_mode(data, BMC150_MAGN_POWER_MODE_SUSPEND,
 					 false);
Signature · 1 line
2.44.0
Reviewed-by on v6 2/3
#7 lore raw

Re: [PATCH v6 2/3] iio: magnetometer: bmc150_magn: replace msleep with fsleep

Quoted 3 lines · Neel Bullywon

Replace msleep(5) with fsleep(5 * USEC_PER_MSEC) to allow the kernel to select the most appropriate delay mechanism based on duration. Using USEC_PER_MSEC makes the unit conversion explicit.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#8 lore raw

PATCH v6 3/3 iio: magnetometer: bmc150_magn: minor formatting cleanup

In reply to Neel Bullywon

Improve initializer list style for bmc150_magn_samp_freq_table by moving the opening brace to its own line and grouping four entries per line with index comments.

Add spaces inside braces for initializer lists in the preset table for consistency.

Fix indentation of bmc150_magn_scan_masks array. No functional changes.

---
 drivers/iio/magnetometer/bmc150_magn.c | 25 +++++++++++--------------
 1 file changed, 11 insertions(+), 14 deletions(-)
drivers/iio/magnetometer/bmc150_magn.c +11 -14
index f254774d28e5..942c2f5e97e1 100644
@@ -149,14 +149,10 @@ struct bmc150_magn_data {
 static const struct {
 	int freq;
 	u8 reg_val;
-} bmc150_magn_samp_freq_table[] = { {2, 0x01},
-				    {6, 0x02},
-				    {8, 0x03},
-				    {10, 0x00},
-				    {15, 0x04},
-				    {20, 0x05},
-				    {25, 0x06},
-				    {30, 0x07} };
+} bmc150_magn_samp_freq_table[] = {
+	{ 2, 0x01 }, { 6, 0x02 }, { 8, 0x03 }, { 10, 0x00 },		/* 0 - 3 */
+	{ 15, 0x04 }, { 20, 0x05 }, { 25, 0x06 }, { 30, 0x07 },	    /* 4 - 7 */
+};
 
 enum bmc150_magn_presets {
 	LOW_POWER_PRESET,
@@ -170,10 +166,10 @@ static const struct bmc150_magn_preset {
 	u8 rep_z;
 	u8 odr;
 } bmc150_magn_presets_table[] = {
-	[LOW_POWER_PRESET] = {3, 3, 10},
-	[REGULAR_PRESET] =  {9, 15, 10},
-	[ENHANCED_REGULAR_PRESET] =  {15, 27, 10},
-	[HIGH_ACCURACY_PRESET] =  {47, 83, 20},
+	[LOW_POWER_PRESET] = { 3, 3, 10 },
+	[REGULAR_PRESET] = { 9, 15, 10 },
+	[ENHANCED_REGULAR_PRESET] = { 15, 27, 10 },
+	[HIGH_ACCURACY_PRESET] = { 47, 83, 20 },
 };
 
 #define BMC150_MAGN_DEFAULT_PRESET REGULAR_PRESET
@@ -641,8 +637,9 @@ static const struct iio_info bmc150_magn_info = {
 };
 
 static const unsigned long bmc150_magn_scan_masks[] = {
-					BIT(AXIS_X) | BIT(AXIS_Y) | BIT(AXIS_Z),
-					0};
+	BIT(AXIS_X) | BIT(AXIS_Y) | BIT(AXIS_Z),
+	0
+};
 
 static irqreturn_t bmc150_magn_trigger_handler(int irq, void *p)
 {
Signature · 1 line
2.44.0
#9 lore raw

Re: [PATCH v6 3/3] iio: magnetometer: bmc150_magn: minor formatting cleanup

Quoted 8 lines · Neel Bullywon

Improve initializer list style for bmc150_magn_samp_freq_table by moving the opening brace to its own line and grouping four entries per line with index comments.

Add spaces inside braces for initializer lists in the preset table for consistency.

Fix indentation of bmc150_magn_scan_masks array. No functional changes.

...

drivers/iio/magnetometer/bmc150_magn.c struct bmc150_magn_data { · line 154

Quoted 3 lines
drivers/iio/magnetometer/bmc150_magn.c +3 -0
+} bmc150_magn_samp_freq_table[] = {
+	{ 2, 0x01 }, { 6, 0x02 }, { 8, 0x03 }, { 10, 0x00 },		/* 0 - 3 */
+	{ 15, 0x04 }, { 20, 0x05 }, { 25, 0x06 }, { 30, 0x07 },	    /* 4 - 7 */

Problem with indentation of the comment in the last line.

Quoted 1 line

+};

Otherwise, LGTM.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
reviewer
#10 lore raw

Re: [PATCH v6 3/3] iio: magnetometer: bmc150_magn: minor formatting cleanup

In reply to Neel Bullywon
Quoted 25 lines · Neel Bullywon

Improve initializer list style for bmc150_magn_samp_freq_table by moving the opening brace to its own line and grouping four entries per line with index comments.

Add spaces inside braces for initializer lists in the preset table for consistency.

Fix indentation of bmc150_magn_scan_masks array. No functional changes.

---
 drivers/iio/magnetometer/bmc150_magn.c | 25 +++++++++++--------------
 1 file changed, 11 insertions(+), 14 deletions(-)
drivers/iio/magnetometer/bmc150_magn.c +0 -3
index f254774d28e5..942c2f5e97e1 100644
@@ -149,14 +149,10 @@ struct bmc150_magn_data {
 static const struct {
 	int freq;
 	u8 reg_val;
-} bmc150_magn_samp_freq_table[] = { {2, 0x01},
-				    {6, 0x02},
-				    {8, 0x03},

drivers/iio/magnetometer/bmc150_magn.c struct bmc150_magn_data { · line 156

Quoted 10 lines
drivers/iio/magnetometer/bmc150_magn.c +4 -5
-				    {10, 0x00},
-				    {15, 0x04},
-				    {20, 0x05},
-				    {25, 0x06},
-				    {30, 0x07} };
+} bmc150_magn_samp_freq_table[] = {
+	{ 2, 0x01 }, { 6, 0x02 }, { 8, 0x03 }, { 10, 0x00 },		/* 0 - 3 */
+	{ 15, 0x04 }, { 20, 0x05 }, { 25, 0x06 }, { 30, 0x07 },	    /* 4 - 7 */
+};
 

Just an opinion: I think it is easier to see the patterns in the numbers if we leave it as one struct per line. (Fixing the indent and padding is still an improvement.)

#11 lore raw

Re: [PATCH v6 3/3] iio: magnetometer: bmc150_magn: minor formatting cleanup

Quoted 1 line · David Lechner

On 2/10/26 5:39 PM, Neel Bullywon wrote:

...

Quoted 8 lines

drivers/iio/magnetometer/bmc150_magn.c struct bmc150_magn_data { · line 155

Quoted 5 lines
drivers/iio/magnetometer/bmc150_magn.c +4 -0
+} bmc150_magn_samp_freq_table[] = {
+	{ 2, 0x01 }, { 6, 0x02 }, { 8, 0x03 }, { 10, 0x00 },		/* 0 - 3 */
+	{ 15, 0x04 }, { 20, 0x05 }, { 25, 0x06 }, { 30, 0x07 },	    /* 4 - 7 */
+};

Just an opinion: I think it is easier to see the patterns in the numbers if we leave it as one struct per line. (Fixing the indent and padding is still an improvement.)

I proposed both, so I'm okay with that too.

Signature · 2 lines
With Best Regards,
Andy Shevchenko