sidecar release

transceivers: fix I2C read bug (#1787)

#1768 did not properly account for the FIFO behavior of the FPGA's data
buffers. The "check the status byte" portion of the loop happened
outside the part where we read the buffer, and since the buffer was just
memory-mapped registers it could be repeatedly without consequence.
Since the data was now in a FIFO, I was inadvertently draining the FIFO
before the transaction was done. This PR consolidates the "is I2C done
yet" logic into the `get_i2c_status_and_read_buffer` so calling code can
just deal with the status register and the data buffer.

Fixes #1786