[RFC PATCH] rtc: rtc-at91rm9200: manage IMR depending on revision

Nicolas Ferre nicolas.ferre at atmel.com
Tue Apr 2 12:28:08 EDT 2013


On 04/02/2013 05:32 PM, Douglas Gilbert :
> On 13-04-02 09:06 AM, Nicolas Ferre wrote:
>> Signed-off-by: Nicolas Ferre <nicolas.ferre at atmel.com>
>> ---
>> Hi all,
>>
>> The funny thing is that I was writing exactly the same code as Johan's
>> when he posted his series.
>>
>> So, here is my single patch, with the comment about the readback
>> stolen from
>> Johan's, but without the way to determine with IP is buggy and which
>> one is
>> not...
>> After having dug the possibility to read the IP revision, I discovered
>> that it
>> is not possible to use this information ("version" register offset
>> changing
>> according to... IP version number: well done!).
>> In conclusion, I guess that the only way to determine if we need the
>> workaround
>> is to use the DT.
>> One remark though: if we use the compatibility string for this
>> purpose, I fear
>> that we would twist the meaning of this information: SoC using an
>> "atmel,at91sam9x5-rtc" compatible RTC will not necessarily be touched
>> by the
>> "non responding IMR" bug: at91sam9n12 or upcoming sama5d3 are not
>> affected for
>> instance, and we need to cling to "atmel,at91rm9200-rtc" for them...
>> I think that we can use this method for the moment and move to another
>> compatibility string later if it is needed.
> 
> Rather than have so many people working on rtc-at91rm9200.c,
> how about someone bring its "RTT" sibling into the DT
> world. I'm talking about drivers/rtc/rtc-at91sam9.c ...

I am currently trying to fix the issue that I have created by pushing a
boggus fix to Andrew's patch series (and "stable" incidentally).
So I am trying to find the best way to address this and:
- a correct
- a smallest possible
path or patch series (I admit that I prefer a single patch).

So, I am still posting rtc-at91rm9200.c patches and hoping from a
feedback. Once we have a good solution I will try to include it in 3.9
and the stable trees affected.

Best regards,
-- 
Nicolas Ferre



More information about the linux-arm-kernel mailing list