Skip to contentSkip to conversation

hwmon: scmi: Driver improvements and fixes

PATCH In review a.shimko 3 participants 8 messages Aug 5, 2025 – Aug 6, 2025

Subsystems (2)
lore
Share & tools
lore

a.shimko

author
#0 lore raw

v2 cover hwmon: scmi: Driver improvements and fixes

This patch series introduces several improvements to the SCMI HWMON driver:

1. Better handling of unsupported sensor types with debug logging
2. Fixes redundant resource management in thermal registration
3. Enhanced error reporting using dev_err_probe()

The changes maintain backward compatibility while improving:
- Debugging capabilities
- Error reporting clarity
- Code maintenance

ChangeLog:
v1:
https://lore.kernel.org/linux-hwmon/[email protected]/
v2:
* fix wrong cover letter and patches count in subjects

Artem Shimko (3):
hwmon: scmi: Add default case with debug output
hwmon: scmi: Remove redundant devm_kfree call
hwmon: scmi: Enhance error reporting with dev_err_probe

 drivers/hwmon/scmi-hwmon.c | 45 +++++++++++++++++++++++++++-----------
 1 file changed, 32 insertions(+), 13 deletions(-)
Signature · 1 line
2.43.0
author
#1 lore raw

PATCH v2 1/3 hwmon: scmi: Add default case with debug output

From: Artem Shimko <[email protected]>

Improve handling of unsupported sensor types:
- Add default case in sensor type switch statement
- Log skipped sensors with debug information including:
* Sensor ID
* Sensor type
* Sensor name (if available)
- Use rate-limited dev_dbg for safety

Debug output format:
"Skipping unsupported sensor ID:%d Type:%d (%s)"

---
 drivers/hwmon/scmi-hwmon.c | 8 ++++++++
 1 file changed, 8 insertions(+)
drivers/hwmon/scmi-hwmon.c +8 -0
index 364199b332c0..a3b5b5c0ec25 100644
@@ -275,6 +275,10 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
 			nr_count[type]++;
 			break;
 		}
+		default:
+			dev_dbg(dev, "Skipping unsupported sensor ID:%d Type:%d (%s)\n",
+				i, sensor->type, sensor->name ? sensor->name : "unnamed");
+			continue;
 	}
 
 	if (nr_count[hwmon_temp])
@@ -323,6 +327,10 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
 			idx = --nr_count[type];
 			*(scmi_sensors->info[type] + idx) = sensor;
 			break;
+		default:
+			dev_dbg(dev, "Skipping unsupported sensor ID:%d Type:%d (%s)\n",
+				i, sensor->type, sensor->name ? sensor->name : "unnamed");
+			continue;
 		}
 	}
 
Signature · 1 line
2.43.0
maintainer
#2 lore raw

Re: [PATCH 1/3] hwmon: scmi: Add default case with debug output

Quoted 10 lines · a.shimko

From: Artem Shimko <[email protected]>

Improve handling of unsupported sensor types:
- Add default case in sensor type switch statement
- Log skipped sensors with debug information including:
* Sensor ID
* Sensor type
* Sensor name (if available)
- Use rate-limited dev_dbg for safety

The code doesn't actually do that, and it would be pointless and make the message useless.

Quoted 13 lines

Debug output format:
"Skipping unsupported sensor ID:%d Type:%d (%s)"

---
  drivers/hwmon/scmi-hwmon.c | 8 ++++++++
  1 file changed, 8 insertions(+)
drivers/hwmon/scmi-hwmon.c +0 -0
index 364199b332c0..a3b5b5c0ec25 100644

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 281

Quoted 7 lines
drivers/hwmon/scmi-hwmon.c +4 -0
@@ -275,6 +275,10 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
  			nr_count[type]++;
  			break;
  		}
+		default:
+			dev_dbg(dev, "Skipping unsupported sensor ID:%d Type:%d (%s)\n",
+				i, sensor->type, sensor->name ? sensor->name : "unnamed");
+			continue;
			break;

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 333

Quoted 11 lines
drivers/hwmon/scmi-hwmon.c +4 -0
  	}
 
  	if (nr_count[hwmon_temp])
@@ -323,6 +327,10 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
  			idx = --nr_count[type];
  			*(scmi_sensors->info[type] + idx) = sensor;
  			break;
+		default:
+			dev_dbg(dev, "Skipping unsupported sensor ID:%d Type:%d (%s)\n",
+				i, sensor->type, sensor->name ? sensor->name : "unnamed");
+			continue;
			break;
Quoted 3 lines
drivers/hwmon/scmi-hwmon.c +0 -0
  		}
  	}
 
#3 lore raw

Re: [PATCH 1/3] hwmon: scmi: Add default case with debug output

In reply to a.shimko
kernel test robot noticed the following build errors:

Hi a.shimko,

kernel test robot noticed the following build errors:

[auto build test ERROR on groeck-staging/hwmon-next]
[also build test ERROR on linus/master v6.16 next-20250806] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/a-shimko/hwmon-scmi-Add-default-case-with-debug-output/20250806-122638
base:   https://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging.git hwmon-next
patch link:    https://lore.kernel.org/r/20250805125003.12573-2-artyom.shimko%40gmail.com
patch subject: [PATCH 1/3] hwmon: scmi: Add default case with debug output
config: arc-randconfig-001-20250806 (https://download.01.org/0day-ci/archive/20250806/[email protected]/config)
compiler: arc-linux-gcc (GCC) 12.5.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20250806/[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/hwmon/scmi-hwmon.c: In function 'scmi_hwmon_probe':

>> drivers/hwmon/scmi-hwmon.c:278:17: error: 'default' label not within a switch statement
     278 |                 default:
         |                 ^~~~~~~

vim +/default +278 drivers/hwmon/scmi-hwmon.c

228
229 static int scmi_hwmon_probe(struct scmi_device *sdev)
230 {
231 int i, idx;
232 u16 nr_sensors;
233 enum hwmon_sensor_types type;
234 struct scmi_sensors *scmi_sensors;
235 const struct scmi_sensor_info *sensor;
236 int nr_count[hwmon_max] = {0}, nr_types = 0, nr_count_temp = 0;
237 const struct hwmon_chip_info *chip_info;
238 struct device *hwdev, *dev = &sdev->dev;
239 struct hwmon_channel_info *scmi_hwmon_chan;
240 const struct hwmon_channel_info **ptr_scmi_ci;
241 const struct scmi_handle *handle = sdev->handle;
242 struct scmi_protocol_handle *ph;
243
244 if (!handle)
245 return -ENODEV;
246
247 sensor_ops = handle->devm_protocol_get(sdev, SCMI_PROTOCOL_SENSOR, &ph);
248 if (IS_ERR(sensor_ops))
249 return PTR_ERR(sensor_ops);
250
251 nr_sensors = sensor_ops->count_get(ph);
252 if (!nr_sensors)
253 return -EIO;
254
255 scmi_sensors = devm_kzalloc(dev, sizeof(*scmi_sensors), GFP_KERNEL);
256 if (!scmi_sensors)
257 return -ENOMEM;
258
259 scmi_sensors->ph = ph;
260
261 for (i = 0; i < nr_sensors; i++) {
262 sensor = sensor_ops->info_get(ph, i);
263 if (!sensor)
264 return -EINVAL;
265
266 switch (sensor->type) {
267 case TEMPERATURE_C:
268 case VOLTAGE:
269 case CURRENT:
270 case POWER:
271 case ENERGY:
272 type = scmi_types[sensor->type];
273 if (!nr_count[type])
274 nr_types++;
275 nr_count[type]++;
276 break;
277 }

 > 278			default:

279 dev_dbg(dev, "Skipping unsupported sensor ID:%d Type:%d (%s)\n",
280 i, sensor->type, sensor->name ? sensor->name : "unnamed");
281 continue;
282 }
283
284 if (nr_count[hwmon_temp])
285 nr_count_temp = nr_count[hwmon_temp];
286
287 scmi_hwmon_chan = devm_kcalloc(dev, nr_types, sizeof(*scmi_hwmon_chan),
288 GFP_KERNEL);
289 if (!scmi_hwmon_chan)
290 return -ENOMEM;
291
292 ptr_scmi_ci = devm_kcalloc(dev, nr_types + 1, sizeof(*ptr_scmi_ci),
293 GFP_KERNEL);
294 if (!ptr_scmi_ci)
295 return -ENOMEM;
296
297 scmi_chip_info.info = ptr_scmi_ci;
298 chip_info = &scmi_chip_info;
299
300 for (type = 0; type < hwmon_max; type++) {
301 if (!nr_count[type])
302 continue;
303
304 scmi_hwmon_add_chan_info(scmi_hwmon_chan, dev, nr_count[type],
305 type, hwmon_attributes[type]);
306 *ptr_scmi_ci++ = scmi_hwmon_chan++;
307
308 scmi_sensors->info[type] =
309 devm_kcalloc(dev, nr_count[type],
310 sizeof(*scmi_sensors->info), GFP_KERNEL);
311 if (!scmi_sensors->info[type])
312 return -ENOMEM;
313 }
314
315 for (i = nr_sensors - 1; i >= 0 ; i--) {
316 sensor = sensor_ops->info_get(ph, i);
317 if (!sensor)
318 continue;
319
320 switch (sensor->type) {
321 case TEMPERATURE_C:
322 case VOLTAGE:
323 case CURRENT:
324 case POWER:
325 case ENERGY:
326 type = scmi_types[sensor->type];
327 idx = --nr_count[type];
328 *(scmi_sensors->info[type] + idx) = sensor;
329 break;
330 default:
331 dev_dbg(dev, "Skipping unsupported sensor ID:%d Type:%d (%s)\n",
332 i, sensor->type, sensor->name ? sensor->name : "unnamed");
333 continue;
334 }
335 }
336
337 hwdev = devm_hwmon_device_register_with_info(dev, "scmi_sensors",
338 scmi_sensors, chip_info,
339 NULL);
340 if (IS_ERR(hwdev))
341 return PTR_ERR(hwdev);
342
343 for (i = 0; i < nr_count_temp; i++) {
344 int ret;
345
346 sensor = *(scmi_sensors->info[hwmon_temp] + i);
347 if (!sensor)
348 continue;
349
350 /*
351 * Warn on any misconfiguration related to thermal zones but
352 * bail out of probing only on memory errors.
353 */
354 ret = scmi_thermal_sensor_register(dev, ph, sensor);
355 if (ret) {
356 if (ret == -ENOMEM)
357 return ret;
358 dev_warn(dev,
359 "Thermal zone misconfigured for %s. err=%d\n",
360 sensor->name, ret);
361 }
362 }
363
364 return 0;
365 }
366

Signature · 2 lines
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
author
#4 lore raw

PATCH v2 2/3 hwmon: scmi: Remove redundant devm_kfree call

In reply to a.shimko

From: Artem Shimko <[email protected]>

Fix potential resource management issue by:
- Removing unnecessary devm_kfree() call in error path
- Relying on devres automatic cleanup
- Preserving all error handling logic

Rationale:
- Memory was allocated with devm_kzalloc()
- devm_ thermal registration manages its own resources
- Double-free could occur during probe failure

---
 drivers/hwmon/scmi-hwmon.c | 2 --
 1 file changed, 2 deletions(-)
drivers/hwmon/scmi-hwmon.c +0 -2
index a3b5b5c0ec25..d03174922e65 100644
@@ -211,8 +211,6 @@ static int scmi_thermal_sensor_register(struct device *dev,
 	tzd = devm_thermal_of_zone_register(dev, th_sensor->info->id, th_sensor,
 					    &scmi_hwmon_thermal_ops);
 	if (IS_ERR(tzd)) {
-		devm_kfree(dev, th_sensor);
-
 		if (PTR_ERR(tzd) != -ENODEV)
 			return PTR_ERR(tzd);
 
Signature · 1 line
2.43.0
maintainer
#5 lore raw

Re: [PATCH 2/3] hwmon: scmi: Remove redundant devm_kfree call

Quoted 11 lines · a.shimko

From: Artem Shimko <[email protected]>

Fix potential resource management issue by:
- Removing unnecessary devm_kfree() call in error path
- Relying on devres automatic cleanup
- Preserving all error handling logic

Rationale:
- Memory was allocated with devm_kzalloc()
- devm_ thermal registration manages its own resources
- Double-free could occur during probe failure

The reason for calling devm_kfree() is to avoid holding the memory unnecessarily until the driver deregisters. There is no double free since the code calls devm_kfree(), not kfree().

The reasoning is wrong, and the patch does not add value.

Guenter

Quoted 19 lines
---
  drivers/hwmon/scmi-hwmon.c | 2 --
  1 file changed, 2 deletions(-)
drivers/hwmon/scmi-hwmon.c +0 -2
index a3b5b5c0ec25..d03174922e65 100644
@@ -211,8 +211,6 @@ static int scmi_thermal_sensor_register(struct device *dev,
  	tzd = devm_thermal_of_zone_register(dev, th_sensor->info->id, th_sensor,
  					    &scmi_hwmon_thermal_ops);
  	if (IS_ERR(tzd)) {
-		devm_kfree(dev, th_sensor);
-
  		if (PTR_ERR(tzd) != -ENODEV)
  			return PTR_ERR(tzd);
 
author
#6 lore raw

PATCH v2 3/3 hwmon: scmi: Enhance error reporting with dev_err_probe

In reply to a.shimko

From: Artem Shimko <[email protected]>

Replace error returns with dev_err_probe() throughout driver:
- Add descriptive error messages for all failure cases
- Include relevant context (sensor IDs, types etc)
- Standardize error reporting format

Improved messages include:
- "No valid sensor info for index %d"
- "Failed to allocate channel info array"
- "SCMI protocol ops not initialized"

---
 drivers/hwmon/scmi-hwmon.c | 35 ++++++++++++++++++++++++-----------
 1 file changed, 24 insertions(+), 11 deletions(-)
drivers/hwmon/scmi-hwmon.c +24 -11
index d03174922e65..081502418dfa 100644
@@ -240,26 +240,36 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
 	struct scmi_protocol_handle *ph;
 
 	if (!handle)
-		return -ENODEV;
+		return dev_err_probe(dev, -ENODEV, "SCMI device handle is NULL\n");
 
 	sensor_ops = handle->devm_protocol_get(sdev, SCMI_PROTOCOL_SENSOR, &ph);
-	if (IS_ERR(sensor_ops))
-		return PTR_ERR(sensor_ops);
+	if (IS_ERR_OR_NULL(sensor_ops)) {
+		if (IS_ERR(sensor_ops))
+			return dev_err_probe(dev, PTR_ERR(sensor_ops),
+					     "SCMI sensor protocol acquisition failed\n");
+		return dev_err_probe(dev, -EPROTO,
+				     "SCMI sensor protocol ops structure unexpectedly NULL\n");
+	}
+
+	if (!sensor_ops->info_get || !sensor_ops->count_get)
+		return dev_err_probe(dev, -ENOENT,
+				     "SCMI sensor protocol operations are not initialized\n");
 
 	nr_sensors = sensor_ops->count_get(ph);
 	if (!nr_sensors)
-		return -EIO;
+		return dev_err_probe(dev, -EIO, "No sensors found\n");
 
 	scmi_sensors = devm_kzalloc(dev, sizeof(*scmi_sensors), GFP_KERNEL);
 	if (!scmi_sensors)
-		return -ENOMEM;
+		return dev_err_probe(dev, -ENOMEM, "Failed to allocate scmi_sensors structure\n");
 
 	scmi_sensors->ph = ph;
 
 	for (i = 0; i < nr_sensors; i++) {
 		sensor = sensor_ops->info_get(ph, i);
 		if (!sensor)
-			return -EINVAL;
+			return dev_err_probe(dev, -EINVAL,
+					     "Failed to get sensor info for sensor %d\n", i);
 
 		switch (sensor->type) {
 		case TEMPERATURE_C:
@@ -285,12 +295,12 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
 	scmi_hwmon_chan = devm_kcalloc(dev, nr_types, sizeof(*scmi_hwmon_chan),
 				       GFP_KERNEL);
 	if (!scmi_hwmon_chan)
-		return -ENOMEM;
+		return dev_err_probe(dev, -ENOMEM, "Failed to allocate channel info array\n");
 
 	ptr_scmi_ci = devm_kcalloc(dev, nr_types + 1, sizeof(*ptr_scmi_ci),
 				   GFP_KERNEL);
 	if (!ptr_scmi_ci)
-		return -ENOMEM;
+		return dev_err_probe(dev, -ENOMEM, "Failed to allocate channel info pointers\n");
 
 	scmi_chip_info.info = ptr_scmi_ci;
 	chip_info = &scmi_chip_info;
@@ -307,7 +317,8 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
 			devm_kcalloc(dev, nr_count[type],
 				     sizeof(*scmi_sensors->info), GFP_KERNEL);
 		if (!scmi_sensors->info[type])
-			return -ENOMEM;
+			return dev_err_probe(dev, -ENOMEM,
+					     "Failed to allocate sensor info for type %d\n", type);
 	}
 
 	for (i = nr_sensors - 1; i >= 0 ; i--) {
@@ -336,7 +347,7 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
 						     scmi_sensors, chip_info,
 						     NULL);
 	if (IS_ERR(hwdev))
-		return PTR_ERR(hwdev);
+		return dev_err_probe(dev, PTR_ERR(hwdev), "Failed to register hwmon device\n");
 
 	for (i = 0; i < nr_count_temp; i++) {
 		int ret;
@@ -352,7 +363,9 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
 		ret = scmi_thermal_sensor_register(dev, ph, sensor);
 		if (ret) {
 			if (ret == -ENOMEM)
-				return ret;
+				return dev_err_probe(dev, ret,
+						     "Failed to allocate memory for thermal zone\n");
+
 			dev_warn(dev,
 				 "Thermal zone misconfigured for %s. err=%d\n",
 				 sensor->name, ret);
Signature · 1 line
2.43.0
maintainer
#7 lore raw

Re: [PATCH 3/3] hwmon: scmi: Enhance error reporting with dev_err_probe

Quoted 22 lines · a.shimko

From: Artem Shimko <[email protected]>

Replace error returns with dev_err_probe() throughout driver:
- Add descriptive error messages for all failure cases
- Include relevant context (sensor IDs, types etc)
- Standardize error reporting format

Improved messages include:
- "No valid sensor info for index %d"
- "Failed to allocate channel info array"
- "SCMI protocol ops not initialized"

---
  drivers/hwmon/scmi-hwmon.c | 35 ++++++++++++++++++++++++-----------
  1 file changed, 24 insertions(+), 11 deletions(-)
drivers/hwmon/scmi-hwmon.c +0 -0
index d03174922e65..081502418dfa 100644

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 243

Quoted 5 lines
drivers/hwmon/scmi-hwmon.c +1 -1
@@ -240,26 +240,36 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
  	struct scmi_protocol_handle *ph;
 
  	if (!handle)
-		return -ENODEV;
+		return dev_err_probe(dev, -ENODEV, "SCMI device handle is NULL\n");

-ENODEV is not supposed to log an error message.

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 246

Quoted 5 lines
drivers/hwmon/scmi-hwmon.c +1 -2
  	sensor_ops = handle->devm_protocol_get(sdev, SCMI_PROTOCOL_SENSOR, &ph);
-	if (IS_ERR(sensor_ops))
-		return PTR_ERR(sensor_ops);
+	if (IS_ERR_OR_NULL(sensor_ops)) {

This never returns NULL, and AFAIS no other driver checks for it. Why add this unnecessary check ?

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 256

Quoted 10 lines
drivers/hwmon/scmi-hwmon.c +10 -0
+		if (IS_ERR(sensor_ops))
+			return dev_err_probe(dev, PTR_ERR(sensor_ops),
+					     "SCMI sensor protocol acquisition failed\n");
+		return dev_err_probe(dev, -EPROTO,
+				     "SCMI sensor protocol ops structure unexpectedly NULL\n");
+	}
+
+	if (!sensor_ops->info_get || !sensor_ops->count_get)
+		return dev_err_probe(dev, -ENOENT,
+				     "SCMI sensor protocol operations are not initialized\n");

It is not this driver's responsibility to validate sensor_ops.

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 264

Quoted 10 lines
drivers/hwmon/scmi-hwmon.c +2 -2
  	nr_sensors = sensor_ops->count_get(ph);
  	if (!nr_sensors)
-		return -EIO;
+		return dev_err_probe(dev, -EIO, "No sensors found\n");
 
  	scmi_sensors = devm_kzalloc(dev, sizeof(*scmi_sensors), GFP_KERNEL);
  	if (!scmi_sensors)
-		return -ENOMEM;
+		return dev_err_probe(dev, -ENOMEM, "Failed to allocate scmi_sensors structure\n");

-ENOMEM is not supposed to log an error message.

Quoted 13 lines
drivers/hwmon/scmi-hwmon.c +2 -1
  	scmi_sensors->ph = ph;
 
  	for (i = 0; i < nr_sensors; i++) {
  		sensor = sensor_ops->info_get(ph, i);
  		if (!sensor)
-			return -EINVAL;
+			return dev_err_probe(dev, -EINVAL,
+					     "Failed to get sensor info for sensor %d\n", i);
 
  		switch (sensor->type) {
  		case TEMPERATURE_C:

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 298

Quoted 5 lines
drivers/hwmon/scmi-hwmon.c +1 -1
@@ -285,12 +295,12 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
  	scmi_hwmon_chan = devm_kcalloc(dev, nr_types, sizeof(*scmi_hwmon_chan),
  				       GFP_KERNEL);
  	if (!scmi_hwmon_chan)
-		return -ENOMEM;
+		return dev_err_probe(dev, -ENOMEM, "Failed to allocate channel info array\n");

Same as above.

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 303

Quoted 6 lines
drivers/hwmon/scmi-hwmon.c +1 -1
  	ptr_scmi_ci = devm_kcalloc(dev, nr_types + 1, sizeof(*ptr_scmi_ci),
  				   GFP_KERNEL);
  	if (!ptr_scmi_ci)
-		return -ENOMEM;
+		return dev_err_probe(dev, -ENOMEM, "Failed to allocate channel info pointers\n");

Same as above.

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 321

Quoted 10 lines
drivers/hwmon/scmi-hwmon.c +2 -1
  	scmi_chip_info.info = ptr_scmi_ci;
  	chip_info = &scmi_chip_info;
@@ -307,7 +317,8 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
  			devm_kcalloc(dev, nr_count[type],
  				     sizeof(*scmi_sensors->info), GFP_KERNEL);
  		if (!scmi_sensors->info[type])
-			return -ENOMEM;
+			return dev_err_probe(dev, -ENOMEM,
+					     "Failed to allocate sensor info for type %d\n", type);

And again.

Quoted 13 lines
drivers/hwmon/scmi-hwmon.c +1 -1
  	}
 
  	for (i = nr_sensors - 1; i >= 0 ; i--) {
@@ -336,7 +347,7 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
  						     scmi_sensors, chip_info,
  						     NULL);
  	if (IS_ERR(hwdev))
-		return PTR_ERR(hwdev);
+		return dev_err_probe(dev, PTR_ERR(hwdev), "Failed to register hwmon device\n");
 
  	for (i = 0; i < nr_count_temp; i++) {
  		int ret;

drivers/hwmon/scmi-hwmon.c static int scmi_hwmon_probe(struct scmi_device *sdev) · line 368

Quoted 7 lines
drivers/hwmon/scmi-hwmon.c +3 -1
@@ -352,7 +363,9 @@ static int scmi_hwmon_probe(struct scmi_device *sdev)
  		ret = scmi_thermal_sensor_register(dev, ph, sensor);
  		if (ret) {
  			if (ret == -ENOMEM)
-				return ret;
+				return dev_err_probe(dev, ret,
+						     "Failed to allocate memory for thermal zone\n");
+

And again.

Quoted 3 lines
drivers/hwmon/scmi-hwmon.c +0 -0
  			dev_warn(dev,
  				 "Thermal zone misconfigured for %s. err=%d\n",
  				 sensor->name, ret);