Uhd.tune_request_t & DBRSX[2]

I wonder if someone can help me?

I have been trying to understand ‘tuning’ and I note that UHD has a
couple of options for setting the centre frequency. The two methods:

tune_request_t( target_freq ) method1

and tune_request_t( target freq, lo_off)

when I try the former in GNURadio Companion for a source of UHD things
work as I expect but when I decide to put in a lo_off I get strange
results.

uhd.tune_request_t(804.75e6, 40e6) and I have enabled the dbsrx2_debug
flag to true to see how things are being driven and can see the
following when I execute the GRC, it says:

Target 844.75 and
actual 844.75 which seems correct (i.e. what I think I asked
for) but then I get:

setting of regs x00 through x07 and then
Warning:
The hardware does not support the requested RX frequency
Target 804.75
Actual: 868.75

but it receives the 804.75 signal (or thereabouts - tv audio so wideish
bandwidth)!

The s/w is calculating the correct integer and fractional values in N
(ext_div of 4 (HMC426), scaler 4 (since freq < 1125Mhz ) etc.

Some questions:

    1) why is 844.75 not valid ?

From the MAX2112 spec there are VASA/VASE signals which
allow the chip to search/select VCOs but cannot see
where these are set and assume that is the reason that
  1. why is the utility talking about 868.75?

  2. why, when it says actual is 868.75, is the system receiving
    signals on 804.75?

  3. if you can direct the tuner to a specific frequency with method1 ,
    why would you consider having the option to specify an lo_offset? would
    it allow an IF to be generated (e.g. 10.7 MHz)?

Kind Regards,

                     John