using the continueOnError option for test run steps is a convenient
way to allow all subsequent steps to run despite a test failure.
But this causes the job status to be 'SucceededWithIssues' and not
a full 'Failed' status.
This commit reverts the changes made in commit 81a9cc1f8 to reduce the
time for the LMK PLL to report lock status. The decision to revert the
change comes after an investigation found that reducing the overall time
to detect lock correlates with an increase in an error reported by the
TDC ('[ERROR] TDC measurements show a wide range of values! Check your
clock rates for incompatibilities. ... Uncaught exception in method
set_clock_source: TDC measurement out of expected range!'). Despite the
LMK PLL reporting lock status, our investigation revealed that it takes
additional time for the PLL to align the daughterboards' clocks closely
enough to pass the TDC measurement's range limit (i.e., no samples
exceeding the measurement mean +/- 500 ps). Reverting the change
increases the amount of time between achieving lock status and taking
the TDC measurements, thus greatly reducing the likelihood of the
reported error.
0cf8a60 introduces Python binding for the digital filter
classes. This does not compile under Windows because the
digital filter classes are header only classes but are
declared as UHD_API. They must be declared as UHD_API_HEADER.
See also 1b3ed20.
X310 differs from X410 in that it doesn't require the mgmt_addr to
function. This does a check to see if the field has been added,
instead of assuming that it has a value.
Signed-off-by: Steven Koo <steven.koo@ni.com>
This adds the ability to tune the recv and send frames as a parameter.
This is useful since some rates will fail with jitter in the system
with the default settings.
Signed-off-by: Steven Koo <steven.koo@ni.com>
This change increases the benchmark rate priority to high, which reduces
the jitter. This prevents UI interactions from causing overruns and
underruns.
Signed-off-by: Steven Koo <steven.koo@ni.com>
This was a test written in the early stages of re-writing UHD for the
4.0 release. It tests things that are no longer in use for UHD 4, thus,
the test can go as well.
This was removed from CMake ages ago, and hasn't been built in years. It
also no longer compiles. And since we do MPM on E310, it is also
completely pointless.
The example refactoring required usage of the get/set source API, which
isn't available on B2xx series. This patch tests for the existence of
said API, and disables its usage if appropriate.
The refactoring changed the behaviour of --bitbang: before, it would
terminate after one readback unless --repeat was specified, in which
case it would require a Ctrl-C (SIGINT). After the refactoring, it
always required a SIGINT. This changes the behaviour back to prior to
727141d, and will now only read back once, unless --repeat is provided.
This also fixes the bitbang devtest, which would go on indefinitely.
The atomic item size for the FFT block is always equal to the length of
the FFT, multiplied by bytes-per-sample (4). For now, this is all the
FFT block supports.
The example had organically grown and was getting hard to read, and also
had some known issues. Summary of fixes:
- Default GPIO bank and connector are now derived from the device. This
allows this example to pass without throwing an exception on E3xx and
X4xx series when using default arguments.
- The bitbang test is moved into its own code section, to make the rest
more readable.
- We move all the streamer-related code into a helper struct
- Some repetitive parts of the code are moved into their own functions
- The argument --require-loopback is added, which will fail tests if
GPIO pins are not correctly looped back externally
- --list-banks is renamed to --list_banks for consistency
In particular, the E3xx series have a single "daughterboard", and
a single radio block with two channels. Prior to this fix,
multi_usrp::get_gpio_banks() would return "INT0A", where the 'A' is to
distinguish the INT0 bank with other dboards. Since there are none, the
'A' is unnecessary.
When an assert statement fails, pytest will determine whether that should
be a fail or an xfail. By explicitly calling pytest.fail, we were
preventing pytest from doing that determination, so an individual
test case that had an xfail mark would have been a hard failure.
This commit fixes an issue in `replay_block_control_impl` with the
initialization of the `_cmd_fifo_spaces` vector member variable.
`std::vector<>::reserve()` only allocates memory for the vector items;
it does not resize the vector (i.e., instantiating an `std::vector<>`
and then calling `reserve()` leaves it with a size of 0). Attempting to
index a zero-sized vector causes some C++ debug mode runtimes to throw
an index out-of-range exception. The commit instantiates the vector
using the constructor variant that allocates memory sufficient for the
entries and default initializes them so that it can be indexed without
issue.
C++20 no longer allows redundant template parameter lists. The
template parameter list must be removed from the property_impl
constructor to allow C++20 projects to include UHD.
Signed-off-by: Clayton Smith <argilo@gmail.com>
Note: This function is currently unused. However, it is not correctly
implemented. During init, it would call the routing configuration
function instead of the init configuration function.
GNU Radio-based tests are not used anymore. The gr- prefix is removed
from the USRP test folder name and tests from the apps folder that do
not depend on GNU Radio are kept in the top level folder.
The dual_SFP test case parameter will be false for all B210 and
E320 test cases, but it still needs to be defined since the main
test_streaming() function requires a value for it.
This commit enables runs for X310, X310 TwinRx, E320, and N310.
By default, python's argparse module will expect a value to be
provided for each argument defined. Specifying an nargs value of '?'
for the pytest options allows us to use a single pytest invocation
in our pipeline configuration, even though some options don't apply
to some DUT types.
The frontend corrections for N320 (IQ imbalance, DC offset) require the
DB serial. However, there was an error in reading the DB serial in the
code that applies the corrections.
This causes the latest RFNoC protocol version to be used by default
and avoids the need to update YAML files every time the RFNoC
protocol version gets bumped.
