[PATCH 0/8] Armada XP pinctrl consolidation and ix4-300d fixes

[Benoit Masson]

Le 15 oct. 2014 à 16:41, Andrew Lunn <andrew at lunn.ch> a écrit :

>>> All the 3 branch works WHEN APPLYING A0 patch (below), with both my custom kernel config and the arch/arm/configs/mvebu_v7_defconfig
>> Just one thing though, one of my cron behavior change with 3.17, when I use the cmd  "/sbin/hwclock --adjust -f /dev/rtc1" I get this new error I never had before:
>> "hwclock: select() to /dev/rtc1 to wait for clock tick timed out: Success"
>> 
>> where rtc1 if the i2c rtc connected to the i2C bus, can this be related to the i2c patch that would slow down i2c operation or somthing ? 
> 
> Hi Benoit
> 
> It won't fix your problem, but if the built in RTC is not usable, it
> would be good to disable it in the .dts file. The i2c RTC will then
> become RTC0.
> 
> 
> The actual "error" message is odd:
> "hwclock: select() to /dev/rtc1 to wait for clock tick timed out: Success"
> 
Yes this is the wired thing is that everything works, even getting the date via "cat /sys/class/i2c-adapter/i2c-0/0-0051/rtc/rtc1/time" is ok
reading the rtc driver code it seems this message is shown when multiple resources try to access shared rtc... but why now and not before ?

So I was wondering if the i2c_quirk() applied by the batch has any effect with this ?

Anyhow I'll handle the message for now and see if there are performance impact on this...

Benoit

> Success means it actually worked, not an error. Does the RTC contain
> the correct time afterwards?
> 
>    Andrew