On 05/29/2015 01:27 PM, Guenter Roeck wrote:
Auto-detection for this chip is highly unreliable, and one of its I2C addresses can also be used by EEPROMs, increasing the risk for false positives even more. Drop auto-detection entirely to remove the risk. Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx> --- v2: Add text describing the supported I2C addresses, and that the driver must be instantiated explicitly. Validate supported vrm values in probe function.
[ ... ]
- static int atxp1_probe(struct i2c_client *client, const struct i2c_device_id *id) { @@ -309,6 +269,11 @@ static int atxp1_probe(struct i2c_client *client, /* Get VRM */ data->vrm = vid_which_vrm(); + if (data->vrm != 90 && data->vrm != 91) { + dev_err(&adapter->dev, "atxp1: Not supporting VRM %d.%d\n", + data->vrm / 10, data->vrm % 10);
s/&adapter->// Compile testing is always a good idea. Guenter _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors