17.04.2019 16:03, Guenter Roeck wrote:
> On 4/17/19 4:26 AM, Alexander Amelkin wrote:
>> Inspecting the OCC sources for P8 reveals that OCC may send
>> a special value 0xFFFF to indicate that a sensor read timeout
>> has occured, see
>>
> occurred

Yup. A typo. Will fix.

>
>> https://github.com/open-power/occ/blob/master_p8/src/occ/cmdh/cmdh_fsp_cmds.c#L395
>>
>> That situation wasn't handled in the driver. This patch adds invalid
>> temp value check for the sensor data format 1 and handles it the same
>> way as it is done for the format 2, where EREMOTEIO is reported for
>> this case.
>>
> ETIMEDOUT ? Though that is really a corner case, so I guess both are fine.

We just reused the error code used for the same case for format 2 in common.c:309 (inside occ_show_temp_2() function).

We thought it would be strange to report different codes for the same case in different format versions.

Besides, it's quite a remote I/O error indeed.

>
>> Signed-off-by: Alexander Soldatov <a.soldatov@yadro.com>
>> Signed-off-by: Alexander Amelkin <a.amelkin@yadro.com>
>> Reviewed-by: Alexander Amelkin <a.amelkin@yadro.com>
>> Cc: Edward A. James <eajames@us.ibm.com>
>> Cc: Joel Stanley <joel@jms.id.au>
>> ---
>>   drivers/hwmon/occ/common.c | 4 ++++
>>   1 file changed, 4 insertions(+)
>>
>> diff --git a/drivers/hwmon/occ/common.c b/drivers/hwmon/occ/common.c
>> index 4679acb..825631c 100644
>> --- a/drivers/hwmon/occ/common.c
>> +++ b/drivers/hwmon/occ/common.c
>> @@ -235,6 +235,10 @@ static ssize_t occ_show_temp_1(struct device *dev,
>>           val = get_unaligned_be16(&temp->sensor_id);
>>           break;
>>       case 1:
>> +        /* If a sensor timed out long enough,
>
> "timed out" is sufficient. "timed out long enough" is difficult to understand.
Agreed. That's a weird wording, but I double-checked before submitting that it was just a copy-paste of the wording from the OCC sources for this case. You're probably right though that it's better to fix it.
>
>> +           OCC returns 0xFFFF for that sensor.*/
>
> /*
>  * This is how multi-line comments look like
>  */
>
> Please run checkpatch on your patches.

Mea culpa. Didn't check. Will fix tomorrow.

>
> Thanks,
> Guenter
>
Thanks.

With best regards,
Alexander Amelkin,
Leading BMC Software Engineer, YADRO
https://yadro.com