hwmon: tps23861: add regmap_bulk_read() handling in port resistance read

[PATCH] hwmon: tps23861: add regmap_bulk_read() handling in port resistance read

Posted by Pavel Zhigulin 2 months, 3 weeks ago

regmap_bulk_read() was called without checking its return value
in function tps23861_port_resistance(). If the read failed,
the code proceeded with an uninitialized register value.

Add return value handling

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: fff7b8ab2255 ("hwmon: add Texas Instruments TPS23861 driver")
Signed-off-by: Pavel Zhigulin <Pavel.Zhigulin@kaspersky.com>
---
 drivers/hwmon/tps23861.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/hwmon/tps23861.c b/drivers/hwmon/tps23861.c
index 4cb3960d5170..8873b2e21064 100644
--- a/drivers/hwmon/tps23861.c
+++ b/drivers/hwmon/tps23861.c
@@ -457,13 +457,16 @@ static char *port_poe_plus_status_string(uint8_t poe_plus, unsigned int port)
 static int tps23861_port_resistance(struct tps23861_data *data, int port)
 {
 	unsigned int raw_val;
+	int rc = 0;
 	__le16 regval;

-	regmap_bulk_read(data->regmap,
-			 PORT_1_RESISTANCE_LSB + PORT_N_RESISTANCE_LSB_OFFSET * port,
-			 &regval,
-			 2);
+	rc = regmap_bulk_read(data->regmap,
+			      PORT_1_RESISTANCE_LSB + PORT_N_RESISTANCE_LSB_OFFSET * port,
+			      &regval,
+			      2);

+	if (rc < 0)
+		return 0;
 	raw_val = le16_to_cpu(regval);
 	switch (FIELD_GET(PORT_RESISTANCE_RSN_MASK, raw_val)) {
 	case PORT_RESISTANCE_RSN_OTHER:
--
2.43.0

Re: [PATCH] hwmon: tps23861: add regmap_bulk_read() handling in port resistance read

Posted by Guenter Roeck 2 months, 3 weeks ago

On 11/14/25 07:00, Pavel Zhigulin wrote:
> regmap_bulk_read() was called without checking its return value
> in function tps23861_port_resistance(). If the read failed,
> the code proceeded with an uninitialized register value.
> 
> Add return value handling
> 
> Found by Linux Verification Center (linuxtesting.org) with SVACE.
> 
> Fixes: fff7b8ab2255 ("hwmon: add Texas Instruments TPS23861 driver")
> Signed-off-by: Pavel Zhigulin <Pavel.Zhigulin@kaspersky.com>
> ---
>   drivers/hwmon/tps23861.c | 11 +++++++----
>   1 file changed, 7 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/hwmon/tps23861.c b/drivers/hwmon/tps23861.c
> index 4cb3960d5170..8873b2e21064 100644
> --- a/drivers/hwmon/tps23861.c
> +++ b/drivers/hwmon/tps23861.c
> @@ -457,13 +457,16 @@ static char *port_poe_plus_status_string(uint8_t poe_plus, unsigned int port)
>   static int tps23861_port_resistance(struct tps23861_data *data, int port)
>   {
>   	unsigned int raw_val;
> +	int rc = 0;
>   	__le16 regval;
> 
> -	regmap_bulk_read(data->regmap,
> -			 PORT_1_RESISTANCE_LSB + PORT_N_RESISTANCE_LSB_OFFSET * port,
> -			 &regval,
> -			 2);
> +	rc = regmap_bulk_read(data->regmap,
> +			      PORT_1_RESISTANCE_LSB + PORT_N_RESISTANCE_LSB_OFFSET * port,
> +			      &regval,
> +			      2);
> 
> +	if (rc < 0)
> +		return 0;
>   	raw_val = le16_to_cpu(regval);
>   	switch (FIELD_GET(PORT_RESISTANCE_RSN_MASK, raw_val)) {
>   	case PORT_RESISTANCE_RSN_OTHER:

tps23861_port_resistance() is called from tps23861_port_status_show(),
which has three unchecked calls to regmap_read(). It is not entirely obvious
to me why checking this specific call to regmap_bulk_read() would warrant
an error check but not the calls to regmap_read().

Also, I am not sure if it make sense to ignore the error.

Guenter