Re: [RFCv4] timerfd: add TFD_NOTIFY_CLOCK_SET to watch for clock changes

[Scott James Remnant <scott@xxxxxxxxxxxx>]

On Fri, Mar 11, 2011 at 11:56 AM, Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
> On Fri, 11 Mar 2011, Scott James Remnant wrote:
>> On Thu, Mar 10, 2011 at 12:25 AM, Andrew Morton
>> <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
>> > On Wed, 9 Mar 2011 18:01:09 -0800 Scott James Remnant <scott@xxxxxxxxxxxx> wrote:
>> >
>> >> > It would be helpful to know if the identified users of this feature
>> >> > actually find it useful and adequate. __I guess the most common
>> >> > application is the 1,001 desktop clock widgets. __Do you have any
>> >> > feedback from any of the owners of those?
>> >> >
>> >> cron is another obvious one (or init systems attempting to replace
>> >> cron). Having to wakeup and check the time every minute can be
>> >> non-conducive to power savings, it would be better if we could just
>> >> sleep until the next alarm and be woken up if the time changes in
>> >> between.
>> >>
>> >> (That being said, we also need to poll for and/or check for timezone
>> >> changes - but those are entirely userspace, so we can deal with that
>> >> separately)
>> >
>> > Sure, there will be lots of applications.
>> >
>> > But what I'm asking isn't "it is a good feature".  I'm asking "is the
>> > feature implemented well".  Ideally someone would get down and modify
>> > cron to use the interface in this patch.
>> >
>> So I've just been thinking today - and I'm actually not sure whether
>> this is needed at all for this case.
>>
>> A good cron implementation is going to set timers according to
>> CLOCK_REALTIME; in the case where the clock changes forwards, those
>> timers will fire as part of the clock changing already no? And in the
>> case where the clock changes backwards, you don't want to re-run old
>> ones anyway.
>>
>> Even the hourly/daily cases are actually at a fixed time, so would be
>> triggered - and a decent implementation wouldn't trigger a given
>> script more than once.
>
> Yeah, I was wondering about today as well. Though when you set back
> your clock several days, stuff might be surprised if it's not woken up
> for several days :)
>
I've checked the code, and more importantly, tested the
setting-forward example - timers do indeed fire at the point the clock
is wound forwards. This means there doesn't appear to be a utility for
this patch in the cron case.

In the wound back case, I believe that even current cron goes to some
effort to avoid firing events that have already happened?

Scott
--
To unsubscribe from this list: send the line "unsubscribe linux-api" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html