One of the optimized RE allocation function does not work properly
and generates some decoding errors with a COTS UE. Let's use the
generic function for the moment, while analyzing what is wrong.
* adding merge request allow flag:
-- if false no forced merge to develop
-- if true and if branch is not develop, then forced merge to develop
* lowered the ru gains in CI conf files
* adding on-the-fly HTML summary within python main.py script
* adding 20 MHz testcases to FDD band 7 scenario
* adding GitLab status for
-- centOS build
-- FDD-band7 tests
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
- install package 'vim-common' to have 'xxd' needed by the compilation
process of the T tracer
- change the way xxd is called so that no file is created in case
of error
Adding a build stage on CentOS (only one variant : -w USRP --eNB)
Should not trigger a fail for the moment
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
-- stdbuf -o0 seems to be causing segmentation fault
-- using daemon instead of nohup
-- also repeating several time "got sync" message and flushing stdout/stderr
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
* added banner to pipeline script
* added parameters to Jenkins pipeline / python
* xmlfile and test stage
* eNB configuration files are located under ci-scripts/conf_files
* no more correction of MNC
* still correction of EPC / eNB IP addresses at runtime
* Initialize_eNB_args is back using the proper "-O " syntax
* Terminate stage is parallel as in FJU's original pipeline
* Log collection stage is back fully parallel
* Console as artifact an option
* Check if terminate calls ended well, to be repeated at the end
* Adding back iperf:
* selection of iperf vs iperf3
* adding some iperf testcases to test suite
* Better enb launch
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
-- a boolean parameter should be added to master eNB job
-- for slack: pipelineUsesSlack
if not present or false, no message will be sent
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
-- Jenkins Pipeline script:
-- Moved the node to a proper subnetwork
-- Added ansiColor option
-- XML testsuite file can now be an option to the main.py python script
-- Added some examples testsuite templates
-- main.py python file
-- Added an option when opening an SSH session: taking care of clients with copied SSH keys
in that case, verifing that we are on the correct host
-- BuildeNB method is closer from Fujitsu's original one
improvement on flexibility for code source path naming
better clean-up of workspace
-- LTEBOX EPC (init / termination / logCollection) methods improvements w/ a /tmp path
-- Checking of processes (EPC / eNB) is parallelized
also handles LTEBOX EPX vs OAI CN
-- UE methods added: attach, detach, reboot, ping
With complete parallelism for multiple UEs
iperf added but not tested
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
-- temporary into a separate jenkins job
-- parameters are not in the pipeline script file
they have to be set up manually in the Jenkins GUI configuration page
they are then check if present in pipeline script
-- call to python script
-- build on a server with RF HW capabilities
for the moment USRP B2x0
also handles merge request process for gitlab
-- since it is for CI, it is a clean build
-- connect to an EPC (OAI or LTEBOX)
-- terminates eNB / EPC
-- log collection
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
The power estimation of the UE has to be debugged (calibrated?).
In the meantime, using "--ue-rxgain 140" lets the UE and eNB
talk to each other in the basic simulator.
2. minor modifications in dlsch_demodulation.c for 2 eNB TX antennas, 1 RX antenna. Inconsistant indexing of some arrays in this case.
3. change in config_load_configmodule.c : removed exit(-1) when no configuration file is used. This makes unitary simulators exit. Effect on lte-XXsoftmodem TBC.
Cedric Roux