Skip to contentSkip to conversation

hwmon: lm90: Fix DT channel constraints and reject unsupported channel 2

PATCH Applied Flaviu Nistor 4 participants 9 messages Aug 26 – Aug 30

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v2 cover hwmon: lm90: Fix DT channel constraints and reject unsupported channel 2

The intention of this patch series is to fix DT channel constrains in the dt-bindings and inprove the driver to reject unsupported channel 2.

dt-bindings:
Forbid temperature-offset-millicelsius on channel@0 (local sensor). Rejecting of assigning temperature-offset-millicelsius for channel@0 is already implemented in the driver.
Reject channel@2 for compatibles that have only one remote sensor.

driver:
If the sensor have only one remote channel, reject a possible unsupported channel 2 and return an error, similar to miss-assigment of temperature -offset-millicelsius property to channel@0.

Flaviu Nistor (2):
dt-bindings: hwmon: national,lm90: Fix channel constraints for
temperature offset
hwmon: (lm90) Reject channel 2 on chips with only one remote sensor

 .../bindings/hwmon/national,lm90.yaml         | 28 +++++++++++++++++--
 drivers/hwmon/lm90.c                          |  5 ++++
 2 files changed, 30 insertions(+), 3 deletions(-)
Signature · 1 line
2.34.1
#1 lore raw

PATCH v2 1/2 dt-bindings: hwmon: national,lm90: Fix channel constraints for temperature offset

Limit temperature-offset-millicelsius to remote channels only, since channel 0 is local and this property does not apply to it. Channel 2 is only valid on devices with two remote sensors, so reject channel 2 for compatibles that do not support a second remote channel.

---
Changes in v2:
- Use else statement in if/then block as suggested by Conor Dooley.
- Link to v1: https://lore.kernel.org/all/[email protected]/

 .../bindings/hwmon/national,lm90.yaml         | 28 +++++++++++++++++--
 1 file changed, 25 insertions(+), 3 deletions(-)
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +25 -3
index 164068ba069d..a8b2a24501b3 100644
@@ -113,6 +113,23 @@ allOf:
       properties:
         ti,extended-range-enable: false
 
+  - if:
+      not:
+        properties:
+          compatible:
+            contains:
+              enum:
+                - adi,adt7481
+                - dallas,max6695
+                - dallas,max6696
+    then:
+      patternProperties:
+        "^channel@[0-1]$":
+          properties:
+            reg:
+              enum: [0, 1]
+        "channel@2": false
+
   - if:
       properties:
         compatible:
@@ -134,6 +151,11 @@ allOf:
         "^channel@([0-2])$":
           properties:
             temperature-offset-millicelsius: false
+    else:
+      patternProperties:
+        "channel@0":
+          properties:
+            temperature-offset-millicelsius: false
 
   - if:
       properties:
@@ -149,7 +171,7 @@ allOf:
               - onnn,nct1008
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127750
@@ -172,7 +194,7 @@ allOf:
               - winbond,w83l771
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127875
@@ -186,7 +208,7 @@ allOf:
               - ti,tmp461
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127937
Signature · 1 line
2.34.1
#2 lore raw

Re: [PATCH v2 1/2] dt-bindings: hwmon: national,lm90: Fix channel constraints for temperature offset

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Low] Applying `enum: [0, 1]` directly to the `reg` property breaks dtbs validation because `reg`…

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] Applying enum: [0, 1] directly to the reg property breaks dtbs validation because reg evaluates as an array. --

commit 962bb439f62765fdf82aa8ab6bf085e5158b3bb7
Author: Flaviu Nistor <[email protected]>

dt-bindings: hwmon: national,lm90: Fix channel constraints for temperature offset

This commit restricts temperature-offset-millicelsius to remote channels and rejects channel 2 for chips with only one remote sensor. It adds conditional schema rules to limit the allowed channels based on the compatible string.

Quoted 13 lines
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +5 -0
index 164068ba069d7..a8b2a24501b35 100644
@@ -113,6 +113,23 @@ allOf:
       properties:
         ti,extended-range-enable: false
 
+  - if:
+      not:
+        properties:
+          compatible:
+            contains:

Documentation/devicetree/bindings/hwmon/national,lm90.yaml allOf: · line 130

Quoted 10 lines
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +10 -0
+              enum:
+                - adi,adt7481
+                - dallas,max6695
+                - dallas,max6696
+    then:
+      patternProperties:
+        "^channel@[0-1]$":
+          properties:
+            reg:
+              enum: [0, 1]

[Severity: Low]
Does this schema constraint fail validation?

The dt-schema tooling treats reg as an array based on the core schemas. Since this block applies enum: [0, 1] directly to reg, an array value like [0] or [1] will not match the scalar integers 0 or 1 provided in the enum constraint. As a result, valid device tree nodes containing channel@0 or channel@1 could fail make dtbs_check with a type mismatch error.

Should this be constrained inside an items block instead, such as
items: enum: [0, 1]?

Quoted 1 line
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +1 -0
+        "channel@2": false
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
#3 lore raw

Re: [PATCH v2 1/2] dt-bindings: hwmon: national,lm90: Fix channel constraints for temperature offset

In reply to Flaviu Nistor
Quoted 28 lines · Flaviu Nistor

Limit temperature-offset-millicelsius to remote channels only, since channel 0 is local and this property does not apply to it. Channel 2 is only valid on devices with two remote sensors, so reject channel 2 for compatibles that do not support a second remote channel.

---
Changes in v2:
- Use else statement in if/then block as suggested by Conor Dooley.
- Link to v1: https://lore.kernel.org/all/[email protected]/

 .../bindings/hwmon/national,lm90.yaml         | 28 +++++++++++++++++--
 1 file changed, 25 insertions(+), 3 deletions(-)
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +6 -0
index 164068ba069d..a8b2a24501b3 100644
@@ -113,6 +113,23 @@ allOf:
       properties:
         ti,extended-range-enable: false
 
+  - if:
+      not:
+        properties:
+          compatible:
+            contains:
+              enum:

Documentation/devicetree/bindings/hwmon/national,lm90.yaml allOf: · line 131

Quoted 10 lines
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +9 -0
+                - adi,adt7481
+                - dallas,max6695
+                - dallas,max6696
+    then:
 
+      patternProperties:
+        "^channel@[0-1]$":
+          properties:
+            reg:
+              enum: [0, 1]

Is this needed? If channel@2 is false, then you won't be able to get to this point to begin with, right?

Conor.

Quoted 47 lines
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +2 -0
+        "channel@2": false
+
   - if:
       properties:

compatible:

Documentation/devicetree/bindings/hwmon/national,lm90.yaml +8 -3
@@ -134,6 +151,11 @@ allOf:
         "^channel@([0-2])$":
           properties:
             temperature-offset-millicelsius: false
+    else:
+      patternProperties:
+        "channel@0":
+          properties:
+            temperature-offset-millicelsius: false
 
   - if:
       properties:
@@ -149,7 +171,7 @@ allOf:
               - onnn,nct1008
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127750
@@ -172,7 +194,7 @@ allOf:
               - winbond,w83l771
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127875
@@ -186,7 +208,7 @@ allOf:
               - ti,tmp461
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127937
Signature · 1 line
2.34.1
Acked-by
#4 lore raw

Re: [PATCH v2 1/2] dt-bindings: hwmon: national,lm90: Fix channel constraints for temperature offset

Quoted 34 lines · Conor Dooley
Quoted 32 lines · Flaviu Nistor

Limit temperature-offset-millicelsius to remote channels only, since channel 0 is local and this property does not apply to it. Channel 2 is only valid on devices with two remote sensors, so reject channel 2 for compatibles that do not support a second remote channel.

---
Changes in v2:
- Use else statement in if/then block as suggested by Conor Dooley.
- Link to v1: https://lore.kernel.org/all/[email protected]/

 .../bindings/hwmon/national,lm90.yaml         | 28 +++++++++++++++++--
 1 file changed, 25 insertions(+), 3 deletions(-)
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +10 -0
index 164068ba069d..a8b2a24501b3 100644
@@ -113,6 +113,23 @@ allOf:
       properties:
         ti,extended-range-enable: false
 
+  - if:
+      not:
+        properties:
+          compatible:
+            contains:
+              enum:
+                - adi,adt7481
+                - dallas,max6695
+                - dallas,max6696
+    then:
Quoted 8 lines

Documentation/devicetree/bindings/hwmon/national,lm90.yaml allOf: · line 130

Quoted 5 lines
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +5 -0
+      patternProperties:
+        "^channel@[0-1]$":
+          properties:
+            reg:
+              enum: [0, 1]

Is this needed? If channel@2 is false, then you won't be able to get to this point to begin with, right?

Hmm, actually keep it. It'd fail dtbs_check but I don't think dt_binding_check fails for channel@0 + reg = <2>

pw-bot: not-applicable

Quoted 50 lines

Conor.

Quoted 47 lines
Documentation/devicetree/bindings/hwmon/national,lm90.yaml +10 -3
+        "channel@2": false
+
   - if:
       properties:
         compatible:
@@ -134,6 +151,11 @@ allOf:
         "^channel@([0-2])$":
           properties:
             temperature-offset-millicelsius: false
+    else:
+      patternProperties:
+        "channel@0":
+          properties:
+            temperature-offset-millicelsius: false
 
   - if:
       properties:
@@ -149,7 +171,7 @@ allOf:
               - onnn,nct1008
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127750
@@ -172,7 +194,7 @@ allOf:
               - winbond,w83l771
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127875
@@ -186,7 +208,7 @@ allOf:
               - ti,tmp461
     then:
       patternProperties:
-        "^channel@([0-2])$":
+        "^channel@([1-2])$":
           properties:
             temperature-offset-millicelsius:
               maximum: 127937
Signature · 1 line
2.34.1
maintainer Applied
#5 lore raw

Re: [PATCH v2 1/2] dt-bindings: hwmon: national,lm90: Fix channel constraints for temperature offset

In reply to Flaviu Nistor
Quoted 7 lines · Flaviu Nistor

Limit temperature-offset-millicelsius to remote channels only, since channel 0 is local and this property does not apply to it. Channel 2 is only valid on devices with two remote sensors, so reject channel 2 for compatibles that do not support a second remote channel.

Applied.

Thanks,
Guenter

#6 lore raw

PATCH v2 2/2 hwmon: (lm90) Reject channel 2 on chips with only one remote sensor

In reply to Flaviu Nistor

Validate firmware channel definitions against chip capabilities and return -EINVAL when channel 2 is configured on devices with 1 remote channel.

---
 drivers/hwmon/lm90.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/hwmon/lm90.c +5 -0
index 1c603272538a..3faeb2c6ab01 100644
@@ -2712,6 +2712,11 @@ static int lm90_probe_channel(struct i2c_client *client,
 		return -EINVAL;
 	}
 
+	if (id == 2 && !(data->flags & LM90_HAVE_TEMP3)) {
+		dev_err(dev, "channel %d is not supported for this chip in %pfw\n", id, child);
+		return -EINVAL;
+	}
+
 	err = fwnode_property_read_string(child, "label", &data->channel_label[id]);
 	if (err == -ENODATA || err == -EILSEQ) {
 		dev_err(dev, "invalid label property in %pfw\n", child);
Signature · 1 line
2.34.1
#7 lore raw

Re: [PATCH v2 2/2] hwmon: (lm90) Reject channel 2 on chips with only one remote sensor

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Low] Incorrect format specifier '%d' used for 'u32' variable. --

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] Incorrect format specifier '%d' used for 'u32' variable.
--

commit dbab4ee8e2d548c97ffd31b1a3d640af352905b1
Author: Flaviu Nistor <[email protected]>

hwmon: (lm90) Reject channel 2 on chips with only one remote sensor

This commit validates firmware channel definitions against chip capabilities and returns -EINVAL when channel 2 is configured on devices with 1 remote channel.

drivers/hwmon/lm90.c static int lm90_probe_channel(struct i2c_client *client, · line 2716

Quoted 10 lines
drivers/hwmon/lm90.c +2 -0
index 1c603272538aa..3faeb2c6ab01b 100644
@@ -2712,6 +2712,11 @@ static int lm90_probe_channel(struct i2c_client *client,
 		return -EINVAL;
 	}
 
+	if (id == 2 && !(data->flags & LM90_HAVE_TEMP3)) {
+		dev_err(dev, "channel %d is not supported for this chip in %pfw\n", id, child);

[Severity: Low]
Does this code use the correct format specifier for the id variable?

While passing an unsigned 32-bit integer to a %d specifier works for small values (since id is guaranteed to be exactly 2 on this path), it is semantically incorrect for a u32 variable and violates format string contracts, leading to potential compiler warnings (-Wformat).

Would it be better to use %u as the format specifier for a u32?

Quoted 4 lines
drivers/hwmon/lm90.c +3 -0
+		return -EINVAL;
+	}
+
 	err = fwnode_property_read_string(child, "label", &data->channel_label[id]);
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=2
maintainer Applied
#8 lore raw

Re: [PATCH v2 2/2] hwmon: (lm90) Reject channel 2 on chips with only one remote sensor

In reply to Flaviu Nistor
Quoted 4 lines · Flaviu Nistor

Validate firmware channel definitions against chip capabilities and return -EINVAL when channel 2 is configured on devices with 1 remote channel.

Applied.

Thanks,
Guenter