Andrew,
Thanks your your comments, I'm currently preparing/testing a revised patch based on your suggestions, which I will post later this week.
A couple of comments inline.
Steve
> I wouldn't bother with EXPERIMENTAL personally. It seems a farily
> pointless thing.
>
OK, I copied one of the other hwmon drivers, many of which are marked experimental.
>> +static int ads7828_attach_adapter(struct i2c_adapter *adapter);
>> +static int ads7828_detect(struct i2c_adapter *adapter, int address, int
>> kind);
>> +static void ads7828_init_client(struct i2c_client *client);
>> +static int ads7828_detach_client(struct i2c_client *client);
>> +static struct ads7828_data *ads7828_update_device(struct device *dev);
>> +static u16 ads7828_read_value(struct i2c_client *client, u8 reg);
>
> I do dislike all these forward declarations, but they're all needed here
> give the order of the functions. Maybe from my Pascal-on-pdp11 days..
>
OK, this is due to re-using the driver structure/style of other hwmon drivers, will try to improve.
>> +static int ads7828_attach_adapter(struct i2c_adapter *adapter)
>> +{
>> + if (!(adapter->class & I2C_CLASS_HWMON))
>> + return 0;
>
> Can this happen?
Hmmm, this code exists in pretty much all of the other hwmon drivers, so I will leave it in.
I think it relates to I2C vs ISA connected devices, to avoid detecting something with the correct ID on the wrong bus?
