On Thu, Sep 10, 2015 at 08:49:26AM +0900, Krzysztof Kozlowski wrote: > On 10.09.2015 02:31, Andreas Dannenberg wrote: > > On Wed, Sep 09, 2015 at 01:17:11PM +0900, Krzysztof Kozlowski wrote: > >> On 09.09.2015 11:26, Andreas Dannenberg wrote: > >>> Krzysztof, good observation! In bq2425x_charger.c (formerly known as > >>> bq24257_charger.c :) that I worked on the unit used was uA. At that time > >>> I did a quick check and there didn't seem to be a clear standard whether > >>> to use the "micro" or "milli" units - different drivers use different > >>> units. However there seems to be a tendency for the TI drivers to prefer > >>> "milli" (bq2415x_charger.c, bq24735-charger.c) > >>> > >>> Personally I think "milli" units are more appropriate for chargers since > >>> they provide sufficient granularity and the numbers don't become too big > >>> (try typing a voltage in the Volt-range in uV, it's very easy to get the > >>> number of 0s wrong). However since the driver was already there I left > >>> that aspect alone to preserve compatibility. > >> > >> I am fine with both units but milli indeed seems easier to judge by fast > >> looking and less error-prone. Whatever you choose - choose the same one. :) > > > > Ok sounds good. If so, I could go ahead and change the units in the > > bq2425x_charger.c over to mA and mV? > > Wait, these are existing bindings (for bq24257). You cannot change > existing binding. > > Best regards, > Krzysztof > > > It would be a bit labor some and I > > also want to see what Laurentiu thinks but this way we could have most > > of those TI charger drivers use the same units (the new bq24261 driver > > Ram posted also uses mA/mV). Except bq25890_charger.c.... that would > > still use uA/uV.... > > > > Laurentiu -- what made you chose the "micro" units for bq24257_charger.c > > and bq25890_charger.c? When I started writing the BQ24257 driver, I had a look on what existed already. Somehow, I don't know why, I settled on bq24190_charger which has all the units in uA/uV. I thought it's the de-facto standard for charger drivers... :/ For bindings, I looked at the existing TI chips bindings and chose the same binding names, where possible. But for some dumb reason, I missed that the units already used where mainly in mA/mV and, honestly, I didn't even suspect units must be consistent from device to device. What happens if a chip needs a more fine grained setting? I agree on binding name consistency though: ti,charge-current (for example) should mean the same for all TI chargers. That said, is not much we can do on the mA/uA or mV/uV front... At least we can agree on using the same binding names. laurentiu -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
