On 08/24/2017 10:03 PM, Jacek Anaszewski wrote:
> Hi Cedric,
>
> On 08/24/2017 01:30 PM, Cédric Le Goater wrote:
>> On 08/12/2017 10:42 AM, Pavel Machek wrote:
>>> Hi!
>>>
>>>>>>>> +static u8 pca955x_read_input(struct i2c_client *client, int n)
>>>>>>>> +{
>>>>>>>> + return (u8)i2c_smbus_read_byte_data(client, n);
>>>>>>>> +}
>>>>>>>
>>>>>>> Is the cast needed? Should you attempt to handle errors?
>>>>>>
>>>>>> ah. this is a left over. I can fix in a resend.
>>>>>
>>>>> No need to resend just for this.
>>>>>
>>>>> Should you WARN_ON() if the read_byte_data returns < 0 or something?
>>>>>
>>>>> https://linuxtv.org/downloads/v4l-dvb-internals/device-drivers/API-i2c-smbus-read-byte-data.html
>>>>
>>>> I don't think so as this is just to read the gpio value.
>>>>
>>>> I suppose that the i2c layer will output some errors before
>>>> if it catches some invalid state.
>>>
>>> Normally its driver's job to output errors. (Same for write).
>>
>> ok. I can respin the patchset with an extra patch adding checks
>> for I2C errors (see below). In which order would you want that,
>> before the gpio support or after ?
>
> I'd prefer an incremental patch. The patch set has received 10
> days of testing and the merge window is imminent. Not the best
> moment to respin the for-next branch.
OK. The patch applies on top of the patchset. I will give it some
more testing before sending.
Thanks,
C.
