snprintf() with "%02X" and "%04X" specifiers on an int argument
triggers -Wformat-truncation (and thus a build failure with
CONFIG_WERROR) because GCC assumes the int can hold a value wider
than the destination buffer.

The value comes from i2c_smbus_read_byte_data() and
i2c_smbus_read_word_data(), which return at most 8 and 16 bits
respectively, so mask the value to make that explicit and silence
the warning.

Signed-off-by: hanzhijian <hanzhijian1991@gmail.com>
---
 drivers/hwmon/pmbus/ibm-cffps.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/hwmon/pmbus/ibm-cffps.c b/drivers/hwmon/pmbus/ibm-cffps.c
index aad94bcb9..dd7ff3a19 100644
--- a/drivers/hwmon/pmbus/ibm-cffps.c
+++ b/drivers/hwmon/pmbus/ibm-cffps.c
@@ -166,7 +166,7 @@ static ssize_t ibm_cffps_debugfs_read(struct file *file, char __user *buf,
 				if (rc < 0)
 					goto unlock;
 
-				snprintf(&data[i * 2], 3, "%02X", rc);
+				snprintf(&data[i * 2], 3, "%02X", rc & 0xff);
 			}
 
 			rc = i * 2;
@@ -177,7 +177,7 @@ static ssize_t ibm_cffps_debugfs_read(struct file *file, char __user *buf,
 				if (rc < 0)
 					goto unlock;
 
-				snprintf(&data[i * 4], 5, "%04X", rc);
+				snprintf(&data[i * 4], 5, "%04X", rc & 0xffff);
 			}
 
 			rc = i * 4;
-- 
2.43.0

On 8/21/26 04:57, hanzhijian wrote:
> snprintf() with "%02X" and "%04X" specifiers on an int argument
> triggers -Wformat-truncation (and thus a build failure with
> CONFIG_WERROR) because GCC assumes the int can hold a value wider
> than the destination buffer.
> 
> The value comes from i2c_smbus_read_byte_data() and
> i2c_smbus_read_word_data(), which return at most 8 and 16 bits
> respectively, so mask the value to make that explicit and silence
> the warning.
> 

So the perceived overflow never happens, and the patch would be
purely to make the compiler happy. I am not going to apply such patches.

Guenter

> Signed-off-by: hanzhijian <hanzhijian1991@gmail.com>
> ---
>   drivers/hwmon/pmbus/ibm-cffps.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/hwmon/pmbus/ibm-cffps.c b/drivers/hwmon/pmbus/ibm-cffps.c
> index aad94bcb9..dd7ff3a19 100644
> --- a/drivers/hwmon/pmbus/ibm-cffps.c
> +++ b/drivers/hwmon/pmbus/ibm-cffps.c
> @@ -166,7 +166,7 @@ static ssize_t ibm_cffps_debugfs_read(struct file *file, char __user *buf,
>   				if (rc < 0)
>   					goto unlock;
>   
> -				snprintf(&data[i * 2], 3, "%02X", rc);
> +				snprintf(&data[i * 2], 3, "%02X", rc & 0xff);
>   			}
>   
>   			rc = i * 2;
> @@ -177,7 +177,7 @@ static ssize_t ibm_cffps_debugfs_read(struct file *file, char __user *buf,
>   				if (rc < 0)
>   					goto unlock;
>   
> -				snprintf(&data[i * 4], 5, "%04X", rc);
> +				snprintf(&data[i * 4], 5, "%04X", rc & 0xffff);
>   			}
>   
>   			rc = i * 4;

On 8/21/26 06:59, Guenter Roeck wrote:
> So the perceived overflow never happens, and the patch would be
> purely to make the compiler happy. I am not going to apply such patches.

Understood, thanks for the review.  I'll drop it.

hanzhijian