Re: [chrony-dev] Support for Multiple PPS Inputs on single PHC

[Matt Corallo <ntp-lists@xxxxxxxxxxxxxxx>]

I take that as kernel is not likely to change.

Since I got around to using this myself I hacked up a patch (next email). Its definitely the "how do 
I make this work" patch, not something where I spent a while digging into all of chrony's internals 
to make sure things match how it's "supposed to be done".

top seems to think there's no threads to worry about, so I assume the mutex can be dropped with no 
impact, but I left it cause I wasn't sure.

There's probably a cleaner way to do this, and I'm happy to clean up the patch from feedback if it 
makes sense, or y'all can replace with a more correct solution.

Thanks,
Matt

On 2/16/23 4:58 PM, Matt Corallo wrote:
> On 2/16/23 2:54 PM, Richard Cochran wrote:
> > On Thu, Feb 16, 2023 at 09:54:56AM -0800, Matt Corallo wrote:
> > > As for duplicating the output across sockets, ptp_chardev.c's `ptp_read` is
> > > pretty trivial - just pop the next sample off the queue and return it.
> > > Tweaking that to copy the sample into every reader is probably above my
> > > paygrade (and has a whole host of leak risk I'd probably screw up).
> > > `extts_fifo_show` appears to be functionally identical.
> >
> > Each extts in the fifo is delivered only once.  If there are multiple
> > readers, each reader will receive only some of the data.  This is
> > similar to how a pipe behaves.
>
> Right, sorry if the context wasn't clear, I only realized part of the message was removed in the 
> first reply after sending. The question from Miroslav was, basically, "would kernel accept something 
> to only get notified of extts pulses on a given channel, and, if so, how would we go about doing that".
>
> The "we get pulses from all extts channels on the same socket" thing is a bit annoying to munge into 
> chrony - it has the concept of "refclocks" which are a single clock, in this case a single pps pulse 
> generator. If you have two of them on the same PTP clock but coming in on different pins/channels it 
> doesn't have a way to express that outside of two refclocks. While we could take pulses from both 
> refclocks on one socket and shove them into some queue and have the refclocks pick that data up its 
> a bunch of complexity on the client side and not super clean in the current codebase.
>
> Thanks,
> Matt

--
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.