Long term phase continuity for LOs

[KVasya]

What feature would you like to see and why?

As I doubt my request is technically plausible the post is mainly a question and possibly a feature request.

The problem I encountered with is as follows. I've got 2 HackRFOnes, synchronized in clocks, their reception being triggered by external source. I use standart utitility 'hackrf_transfer' to get signals from 2 boards. The boards are fed the same signal and show some random phase difference between signals, which is tolerable.
What torments me is that this phase difference varies between subsequent data receptions (i.e. 2 subsequent 'hackrf_transfer' launches).

I suspect the reason is that various LOs are turned off when hackrf_transfer finishes (thanks to energy saving improvements), and than start from random phases anew.
Could the LOs somehow please stay on? I think it could be a solution to the problem, because phase difference remains intact for very long times. Also, as I see from this diagram
https://github.com/greatscottgadgets/hackrf/assets/19487262/a163d1dc-4ed7-4563-b8ac-fbde058da146,
all LOs are derived from same clock source, so they should stay synchronized.

May be it could already be done via some hackrf_debug writing of 'power saving' registers?

[martinling]

Interesting idea. Perhaps the approach in #1314 could be extended to control whether the LOs should be switched off when the transceiver is in OFF mode.

The other approach, which you could use with current firmware, would be to write some form of server that keeps the two HackRFs in RX mode continuously, and a client program which would function like hackrf_transfer, but make requests to the already-running server program in order to maintain the operating state.