Re: Suggestion - Configurable Source Clock Type for Line Event Timestamping

[Jack Winch <sunt.un.morcov@xxxxxxxxx>]

> The background is in the commit:
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/gpio/gpiolib.c?id=f8850206e160bfe35de9ca2e726ab6d6b8cb77dd
> Also study the solution in IIO that started the discussion about
> all this:
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit?id=bc2b7dab629a51e8beb5fda4222c62a23b729f26


Thanks for the additional background information.  Funnily enough, I
had my sights on this part of the IIO functionality next.

> As Arnd stated in the thread from 2018:
> "most of these clocks make no sense at all for a random user
> space interface, mainly because I wouldn't trust user space
> programmers to make an informed decision which of those
> seven to use."
>
> So I suspect you actually managed to make a good argument
> for using the realtime clock, we didn't see that coming. :)


Well that's reassuring.  Considering doing this work is how I spent my
mid-teens, it looks like that time was well spent. :)

> If you really need the timestamp to be as close as possible
> to the actual event then I see why you want the kernel to
> make the timestamp in the hard IRQ handler already, but
> I just want to confirm that you really really need this.


Confirmed.  I see Kent has submitted a patch series for the chardev
and the v2 uAPI.  So I'll follow the patch threads from now on.  Thank
you all for this addition.

Cheers,
Jack