developing driver to use new hardware in GNU Radio

[Card, Stu]

(cross-posted to the 3 different sub-communities with relevant driver
architectures)

I would very much like to use GNU Radio as a real time signal analysis
environment with some Altera FPGA based hardware (not a general purpose
SDR, more a test probe in a comm system).

Searching for information on porting GNU Radio to new hardware, I found
https://lists.gnu.org/archive/html/discuss-gnuradio/2012-04/msg00535.html

That contains some good pointers, but was 4 years ago, so I wonder if there
are any other places I should be looking for information on how to write a
driver for my hardware, so I can drop a corresponding source block into GRC
and get started with my signal processing?

My signal analysis will typically start with a scrolling spectrogram, such
as is provided by fosphor (original GPU based gr-fosphor or later FPGA
based rfnoc-fosphor). We can do the high rate (FFT, magnitude, log) in the
FPGA: this would incline me to start with the rfnoc-fosphor software stack;
but my hardware is not Ettus, it does not have a UHD interface, it is not
Xilinx. So that pushes me towards writing a driver to go underneath
gr-osmosdr: then I think I also would need to extend gr-osmosdr itself to
make calls to that library, unless there is some sort of device driver
abstraction with a dispatcher to different actual drivers? Searching for a
framework that cleanly isolates the hardware specific device driver code
from the higher layers, I found  SoapySDR: reviewing the description on the
web, it looks very nice, and under active development; but as someone new
to much of this, I am not yet competent to assist with that development,
and am not clear on how complete and usable (by a n00b such as myself)
SoapySDR is yet?

Thanks for any tips!
-- Stu Card <stu.card at axenterprize.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osmocom.org/pipermail/osmocom-sdr/attachments/20160508/1ee1f036/attachment.htm>