SM6FBQ / HackRF clone - History

Don't update MAX2837 freq_cache unless set_freq() is successful.

Fix long-standing example projects breakage due to tuning and SGPIO API and dependency changes.

Change M0 linking to use newlib-nano. snprintf() is a pig when the floating point junk gets pulled in on an M0...

Moved M0 memory region to larger first AHB region (32K), moved USB bulk buffers to smaller AHB RAM area.

Change USB bulk endpoint for baseband data to have only one queue item. Since there are only two baseband buffers, and one is transferring with the codec and the other with USB, enqueueing more than one buffer at a time would result in transferring an incomplete buffer.

Adjust SGPIO GPDMA trigger slice data to a single clock width pulse. Previously, it was 3 clocks long with a 4 clock period, which *seemed* to address GPDMA data drop-outs at maximum baseband speed (20Msps complex).

Adjust tuning API to use a single 64-bit integer for frequency in Hz, since the Cortex-M4F has good support for uint64_t.