[Gb-ccb] CCB LEDs decoupled from ccbserver

[Martin Shepherd]

I have now modified the CCB code to keep the CCB front-panel LEDs
showing the status of the instrument, regardless of whether or not the
CCB manager is connected (previously the LEDs were only kept up to
date when the CCB manager was connected). Note that this only affects
a couple programs and a driver that run on the CCB computer.  The
manager doesn't need recompilation.

The changes were the following:

  1. The CCB server process now no longer controls the front-panel LEDs.

  2. I've modified the existing ccb_monitor_status program, such that
     it can optionally be run as a background daemon. In this mode it
     silently keeps the front-panel LEDs up to date.

  3. I've modified the GPIO device-driver to allow multiple programs to
     simultaneously read monitoring data from it, control its LEDs and
     control the firmware reload bit. Thus the CCB server can now be
     reading monitoring data from it, and control the firmware reload
     bit, while the ccb_monitor_status program is also reading
     monitoring data and updating the front-panel LEDs.

  4. Finally, I've modified the boot-time startup/shutdown script to
     run both the CCB server and the monitoring program as background
     daemons.

As far as I can tell, without being able to see the front-panel LEDs,
everything seems to be working fine. The boot-time script starts up
the two daemons okay, and these two daemons continue running, without
any complaints. The modified server is also returning sensible
monitoring data, according to ccb_demo_client. I have also checked
that I can run multiple instances of ccb_monitor_status, interatively,
and see the same monitoring data being reported by each of them.

Thus I have installed the new code as the default on the CCB computer,
and checked that everything still works okay after a reboot. If
there's time tomorrow, before or after the CCB goes up to the
telescope, could somebody please check that the front-panel LEDs
appear to be behaving as expected, especially when the manager is not
connected?

For example, when you first switch on the CCB at the telescope, and it
boots itself, check that the LEDs indicate sensible things, once the
reboot has completed. If you have more time, and an appropriate
power-supply, you could also then try reducing one of the power-supply
voltages until it is just out of range, to check that the power-OK LED
responds to this, and isn't just in some fixed startup state.

Thanks.

Martin