[Meego-kernel] Doubt about a timer stat

[Arjan van de Ven]

On 12/22/2010 9:47 AM, Mathew, JohnX wrote:
>> -----Original Message-----
>> From: Arjan van de Ven [mailto:arjan at linux.intel.com]
>> Sent: Wednesday, December 22, 2010 5:48 PM
>> To: Tardy, Pierre
>> Cc: al; Mathew, JohnX; meego-kernel at lists.meego.com
>> Subject: Re: [Meego-kernel] Doubt about a timer stat
>> On 12/22/2010 7:52 AM, Tardy, Pierre wrote:
>>>> On 12/21/2010 11:10 PM, al wrote:
>>>>> Hi,
>>>>>
>>>>> I keep seeing the following timer stat -
>>>>> 1,     1 init             hrtimer_start_range_ns (hrtimer_wakeup)
>>>>>
>>>>> Can anyone please explain what init process cases this ?
>>>> which meego version is this on?
>>> On an old internal build from Oct, 14th, we are also seeing this timer. (15 sec timer from init process)
>>> John was supposed to look at this this week.
>>> John, do you have any news on this ?
>>> init will check every 15 seconds or so if there's zombies to reap.
>>> we could make it 30 if it shows up as a power issue.
> After using the ftrace - func_stack_trace I found that the fs/select.c - "select" system call is calling schedule_hrtimeout_range_clock each time the init process wakes up the system.
> So my assumption is that a userland application or the init process itself is using the "select" system call to wait on a file descriptor.
> Pierre/Arjan have you seen this behavior before?
> I am having trouble finding out which is the exact userland application it is....
>

it'll be init I suspect.


but.... why not run powertop 1.99 .. it'll tell you in quite a bit of 
detail who's waking the CPU up ;-)