Re: [PATCH] rtc: core: let rtc_set_time properly populate element wday

[Heiner Kallweit <hkallweit1@gmail.com>]

Am 28.08.2017 um 09:54 schrieb Alexandre Belloni:
> On 27/08/2017 at 21:35:34 +0200, Heiner Kallweit wrote:
>> When having a look at the MCP794XX datasheet it became clear why the
>> mentioned weekday-related patch was submitted. This chip is quite
>> strange regarding the possible alarm match conditions:
>>
>> ALMxMSK<2:0>: Alarm Mask bits
>> 000 = Seconds match
>> 001 = Minutes match
>> 010 = Hours match (logic takes into account 12-/24-hour operation)
>> 011 = Day of week match
>> 100 = Date match
>> 101 = Reserved; do not use
>> 110 = Reserved; do not use
>> 111 = Seconds, Minutes, Hour, Day of Week, Date and Month
>>
>> When not having a proper weekday you get only either seconds or
>> minutes or hours or date match.
>>
>> So it's best to ensure that the weekday is properly populated.
>> However we can do this in the driver and don't have to touch the core.
>>
> 
> Can you do that only for mcp794xx? I received one earlier this month and
> I'll separate the driver from rtc-ds1307.c at some point.
> 
Sure, I'll submit a patch for this. Separating the driver for mcp794xx may
not be needed as it is not that different from the other chips supported
by this driver. Once I have my patch series proposal towards a ds1307_lib
in a little more clean state this should become clearer.

Regards, Heiner

> BTW, I've rebased and pushed some cleanup for this driver in:
> https://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux.git/log/?h=rtc-ds1307
> 
> Can you base yourself on that?
> 
OK

> Thanks,
>