See http://stackoverflow.com/a/19177470

The current version of UHD is the most stable and consistent.
Supporting only the current API removes some ambiguity and
potential errant configurations such as those that have
occurred with GPSDO setup. The build process and maintenance
are simplified as well.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

B100 and E100 will react differently to overruns then the USRP2 or
N2xx devices. To handle such an overrun, we need to expect a short
packet with no error code followed by zero length packet with the
appropriate error code set. On USRP2, the event is contained
within a single packet.

To handle, if a short packet contains no error code, pull another
packet and check again for error code. If still no error code,
print error on "short packet" and continue. If there is an error
code, as expected, proceed normally and handle the error accord-
ingly.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

Behaviour varies by device / daughterboard combination
and a default value for one combination may not apply
to another. Don't set a default in kal - use NULL -
and only set the value in UHD if specified by the user.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

Signed-off-by: Thomas Tsou <ttsou@vt.edu>

UHD 003.004.000 is not an official release, though the
API changes are getting a notable amount of use. Ifdef new
interface changes for now and re-evaluate making 003.004.000
a minimum requirement at a later time.

This also allows use of the streamer interface in the main
branch, which addresses various reports of UHD failing to
send packets at start with the old, deprecated API.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

- adapt to API changes: use rx_streamer instead of m_dev->get_device()->recv
- use m_dev->set_clock_source() for external clock selection
- added KAL_DEVICE environment variable, used if set to pick device
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

Previous commit 7e246a6d



"uhd: use default 10MHz reference configuration unless specified"

did not have it's intended purpose in that the clock was being
set in all cases. We need to get out of the way entirely and
let the GPSDO do its own thing.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

The optional GPSDO automatically modifies the reference setting,
so this patch leaves the default setting alone. Devices without
the GPSDO will default to the interal reference by the UHD
driver.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

This setting was previously disabled because the clock rate
on the USRP2 is not adjustable to any useful value. With
USRP1/B100/E100 the setting is more meaningful at 52 MHz.

Default to the built-in device setting at 64 or 100 MHz and
allow setting the clock frequency from the command line.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

When no device type argument is specified, use the
first found UHD device regardless of type.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

To prevent API breaking in upcoming UHD 003.004.000
release.
Signed-off-by: Thomas Tsou <ttsou@vt.edu>

This commit adds a very simple progress report, not precise
but at least show to the user that something is happening
in the background.
Signed-off-by: Felipe Zimmerle <felipe@zimmerle.org>

The intention of this patches is to make the compiler happy and
get the code compiled. As far as i know C++ standard allows
the initialization of const static data members of any type of
integer inside the class description, but, well...
Signed-off-by: Felipe Zimmerle <felipe@zimmerle.org>

Commit cd274b3e (uhd: swap out the gnuradio based interface
for uhd) removed the overflow count in favor of simply
reporting boolean values. Without the count value in place,
the overflow value was never reset.

Reimplement the counter and allow NULL pointers as the
overrun argument

Simply because it's simpler.

Fixes the following that occurs with recent uhd changes.

UHDDevice.cpp:260: error: ‘struct uhd::gain_range_t’ has no member named ‘max’
UHDDevice.cpp:260: error: ‘struct uhd::gain_range_t’ has no member named ‘min’

This patch modifies the low level implementation to use uhd devices
instead of the usrp library from gnuadio. Also, edit the user
interface to reflect changes. Currently, only the USRP2 has been
tested.

http://thre.at/kalibrate/kal-v0.4.1.tar.bz2