On Mon, Feb 04, 2019 at 11:31:01PM +0000, Vladimir Oltean wrote: > I don't see why letting userspace guess whether the extts event was > rising or falling is a good idea.
Actually, years ago the user/kernel interface foresaw this distinction, and there are flags that were meant to be used for this. However, I didn't follow through and strictly implement these for the few drivers that support ancillary signals. Patches on the netdev mailing are most welcome! However, in the end, the API will surely never capture the actual complexity of time and measurement systems in a satisfactory, generic way. Witness the failure of APIs like Comedi (and certain commercial systems which shall rename nameless). Having said that, in practice this is not too hard to set up. For example, when using the i210, I simply configure a 1/2 Hertz periodic output, knowing well that the device generates time stamps on both edges, and that this cannot be re-programmed. I can only recommend this. Don't guess. Get to know your HW before deciding on the final design. Thanks, Richard _______________________________________________ Linuxptp-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
