* add new UE feature
* enable in srsue.conf.templ
* add new table for maximum rates
* add config scenario to enable SIB option for QAM64
Change-Id: I6ac2c9989a761e91b93d76c2507f55f0140b202d
Due to the integration of DL-QAM256 another table for DL max rates is needed.
Therefore, I added the parameter 'qam256' to the feature list in the resource.cfg.
The patch also enables the correct UE settings in the config file.
Change-Id: I2d34395449cdcfb31db66ea887d9adbee551e757
make sure to insert cell specific TAC, PCI and root seq ind
into cell config and do not depend on cell index for a particular
enb. This causes issues in multi-eNB setups.
Change-Id: I6642128a449a0562dd23f7fa393ff48ae2503006
ENB was using its own address as a remote address when the broker is
used, which actually only holds true for the first ENB (since the broker
is currently run on the same host/address of the first ENB).
Change-Id: I5c83075d140632846c6d357152963754d16b4633
The remotely run script is moved into a new subdir called "external",
where external utils to be used by osmo-gsm-tester (exernal to its own
process) are placed.
It needs to be in another directory because python files in obj/ are
loaded at startup of osmo-gsm-tester to dynamically load schemas.
Change-Id: I633a85294694f2c6efd58535729e9b8af166b3ff
It may happen that the non-emergency call MT leg is still not properly
released when the emergency call MT leg is to be assigned a TCH, meaning
BSC will fail with an Assignment Failure upon Assignment request from
MSC.
The test sometimes passed and sometimes didn't, due to above mentioned
race condition. Let's relax a bit the test expectancies to have it
always passing, while still verifying preemption happens, and MT phone
is reached.
Related: OS#4549
Change-Id: I3697227cac56a1327f9ea08c5c2f52568e8d2a8a
First step towards support everal ENBs and installing the remote script
at runtime:
* The gnuradio broker class is moved to its own file, to decouple it
from RFemu.
* The ENBs are registered earlier in the test so that the GrBroker knows
when to start (delay start + setup until all ENBs have been configured).
* Handle ENBs internally as a list.
Change-Id: I4f1095bbc7ed0a816fe47caef44f7becadd9d737
this patch adds the basic notion of FDD and TDD duplexing modes
to the eNB object. So far we've always assume FDD.
Since only Amarisoft eNB supports TDD, the required config
template changes, etc. are only applied there.
The patch also adds a scenario to enable the default TDD config.
Change-Id: I37216b5bfdf527d221913283b6c41d3c8fd6b500
Don't wait until adding it to the subscriber data base, since the msisdn
may be needed beforehand (for instance in follow up patch to be passed
to osmo-msc configuration to route emergency calls).
Change-Id: I0d6902687e4beb3d6cdcefd4e343f21496100504
srsENB currently creates 1 zmq stream (1 tx, 1 rx) for each cell (2 if
MIMO is enabled). Each cell transceives on a given EARFCN (and several
cells can transmit on same EARFCN).
However, for handover test purposes, we want to join all cells operating
on the same EARFCN to transceive on the same ZMQ conn, so that an srsUE
can interact with them at the same time (same as if the medium was shared).
Furthermore, we want to set different gains on each of those paths
before merging them in order to emulate RF conditions like handover.
In order to do so, a new element called the Broker is introduced, which
is placed in between ENBs and UEs ZMQ conenctions, multiplexing the
connections on the ENB side towards the UE side.
A separate process for the broker is run remotely (ENB run host) which
listens on a ctrl socket for commands. An internal Broker class is used
in osmo-gsm-tester to interact with the remote script, for instance to
configure the ports, start and stop the remote process, send commands to
it, etc.
On each ENB, when the rfemu "gnuradio_zmq" rfemu implementation is selected
in configuration, it will configure its zmq connections and the UE ones to
go over the Broker.
As a result, that means the UE zmq port configuration is expected to be
different than when no broker is in used, since there's the multiplexing
per EARFCN in between.
In this commit, only 1 ENB is supported, but multi-enb support is
planned in the future.
The handover test passes in the docker setup with this config:
"""
OSMO_GSM_TESTER_OPTS="-T -l dbg -s 4g:srsue-rftype@zmq+srsenb-rftype@zmq+" \
"mod-enb-nprb@6+mod-enb-ncells@2+mod-enb-cells-2ca+suite-4g@10,2+" \
"mod-enb-meas-enable -t =handover.py"
"""
and in resources.conf (or scenario), added:
"""
enb:
...
cell_list:
- dl_rfemu:
type: gnuradio_zmq
- dl_rfemu:
type: gnuradio_zmq
"""
Note that since the broker is used, there's not need for mod-srsue-ncarriers@2
since the broker is joining the 2 enb cells into 1 stream on the UE side.
Change-Id: I6282cda400558dcb356276786d91e6388524c5b1
The getter method was named the same as the itnernal field, and hence
when used it would fail since the intenral field would be sleect and
fail to be called.
Change-Id: I2f631eb6256eb0e065f41d5b7531395c4a054cd8
the logic for the RF emulator on the command line is, like the
Amarisoft equivalent, to set a negative gain value to express
attenuation. Therefore, the sign needs to be inverted.
Change-Id: I7b6ac611a9f711bbdc1570602bfd797a5346ab86
pass-through the option so they can be used in templates
just concatenate with rf_dev_args for srsLTE eNB/UE, arguments
parsing will handle them
Change-Id: I3818026c159780f29968888f547163cdf730afad
the get_num_phy_errors() uses the external KPI analyzer module
to calculate the number of PHY layers errors in a vector of log metrics
Change-Id: I0c8de87e7f83f986e288585640ce56847959c497
we've added the modifier to configure Amarisoft eNB channel
model. This patch enables it for srsENB.
Change-Id: I30e65d0431b2d2792986128287caf8b23a22b2c1
this avoids a false negative detection when the UE attach takes a bit
longer and the first seconds all zeros are reported in the CSV
the HO test, for example, would fail in such a case as it expects
no zero TP over the course of the experiment.
Change-Id: I96dab17bb19249504dedda6659aed5eac0a65a26
the WS is closed if the eNB process dies. this patch will log an
event but not stop the experiment.
Change-Id: I0fafc7e38954400540375d51813e96500700f50c
we've noticed that the Iperf client times out too soon in
some experiment. It seems that this happens with ZMQ if
the RAT is too slow (slower than realtime) and the eNB buffers
too much data. This causes iperf to still receive DL data
for example long after the configured runtime.
Change-Id: I9b823ab22812c96920d54fb7342c8d4c578d5270
Since a while ago, we avoid by default using the --logfile param to log
the json output to a different file, in order to stay compatible with
older versions of iperf3.
However, the local logfile was still being created, and was not
overwritten with the remote file with content because there's none
created remotely (due to reasons explained above).
Change-Id: I218ec28dc65febc7bd083d6d15854319bf4f3daf
when phy.signal=1 is set in the log_options, the eNB will dump
a binary file that needs to be copied back as well
Change-Id: I63d13cb74572ac065d5a3fcfb1632552d0a6dbc6
this avoids issues when the channel emulation stops in the middle
of a high attenuation values and the device is left in this state.
Change-Id: I9aeafc68bf81b8932aac4cb28040b9aea6bfcbdb
This object allows to run the "stress" cmdline program in the background
on the specifies run_node, to simulate system load.
To run stress, a test can simply do:
stress = tenv.stress(epc.run_node())
stress.start(cpu_workers=2, mem_workers=1, io_workers=1)
And the test environment will remember to stop it when the test finishes.
Change-Id: I21023e6c64c48109f294291bfe3d8d8f4e1de038
this method uses the kpi_analyzer module for analyzing
stdout, CSV metrics and the logfile (if present).
if the module can't be loaded, no KPI will be added.
Change-Id: I28226a375f9ac4e08424c488062ae6a74a19af92
New patchelf 0.11 release is out and it has fixes hitting us in both 0.9
and 0.10 on different binaries. So let's simply requiring having
patchelf 0.11 in all cases and be done with it (actually, if not running
4G binaries, using 0.9 is enough).
Related: OS#4389
Change-Id: I2701e46e7710db4c923a68e6b3a2f3c7b126f356
every call to verify_metrics would copy back metrics again. see log here:
23:29:22.149299 run srsue(pid=31521): Terminating (SIGINT)
23:29:22.368263 run srsue(pid=31521): Terminated: ok {rc=0}
23:29:28.401035 run scp-back-metrics(pid=32265): Launched
23:29:29.423083 run scp-back-metrics(pid=32265): Terminated: ok {rc=0}
23:29:29.490275 run scp-back-metrics(pid=32271): Launched
23:29:30.513852 run scp-back-metrics(pid=32271): Terminated: ok {rc=0}
23:29:30.550862 run scp-back-metrics(pid=32277): Launched
23:29:31.573962 run scp-back-metrics(pid=32277): Terminated: ok {rc=0}
this patch only copies back metrics only once. also make sure that, similar
to logs, metrics are always copied back at the end of the run even if they
are not used for metrics checking. this is needed to run the plotting
script.
Change-Id: Ie7c156a5b2c8768bf90ff8e86515976e0d454c25
In order to set up the channel emulator we need to calculate the total
number of RF channels that a eNB has from within a test.
Change-Id: If641993033eb5dcdc6e2d6afdc8426ed322241ca
this is helpful to run some action for the exact run time of an iperf3
session. note that if no time spec is given when calling
"prepare_test_proc()", a modifier could set the runtime of an DL experiment
to an arbitrary value so we need to have a mechanism to read
the actual time from within a test
Change-Id: I484a0add322ffd269a1e9df72a00cf348950b268
10s was too short and has still caused tests to fail because they were
stopped just before the iperf finished.
30s seems to be enought and has passed multiple times so far
Change-Id: I80b732a060356aa98882ff0e8b5debaf623ba0fb
this refactor no longer enforces blocking operation of the process.
Instead it returns the process object to the caller who
can now run either proc.launch() for non-blocking operation
or proc.launch_sync() for blocking mode.
The non-block mode allos doing other stuff in the background,
for example controlling the rfemu while running a ping.
Change-Id: Ia6372e55a8829f722e40db537d9dfd63a94d1be9
process object always used timeout=300 while runnig wth launch_sync().
Let's allow replacing that value beforehand so that iperf3 can
pre-configure the process object and caller doesn't need to care about
calculating expected time.
Change-Id: I7f6c5078f648013515919aa35ebcdb3ef157b5e4
Setting the log.ctx manually is not needed anymore and it's actually
harmful since all palces where it was used, a log.Origin already in path
was being passed, causing a origin loop.
Change-Id: I0511b9f7bc59e3c7f2269ff3155d0c95db58d063
Use tailored sample rate only use for 6, 50 and 75 PRB. For all other bandwidths,
we use the full LTE rate and downsample.
With 5.76e6 as base sample rate, for example, we had some issue during the tests like:
[zmq] Error: tx time is 0.178 ms in the past (7833600 < 7835648)
[zmq] Error: tx time is 0.011 ms in the past (7835520 < 7835648)
Using a higher sample rate seems to work better.
Change-Id: I53678587b2c5e0e5dccb1b55a328ad2f1a97fc4c
the overhead with 6 PRB and MIMO is a bit higher when compared
to other PRBs resulting in lower achievable throughput
Change-Id: I63888435553bba4f7be88cc745e24472921a7fb4
rename UE helper functions to query RRC connected/idle state
and adapt tests accordingly
also add helper to retrieve assinged IP address
Change-Id: I6cd057e34b4df6a1a73695355dd6406d7e039546
It turned out that the Amarisoft and SRS eNB scheduler produces
slightly different maximum data rates for both UL and DL.
Change-Id: I30fa7006906d101c53ba586fb06bced3945aa960
This way tests which require a very specific config file can override
specific template files used by object classes.
Change-Id: I65d1b1e826d2d430ee83810d998b98d0ccaa07cd
when calculating the dl and ul bitrate, it is required to take the
sum of all active component carriers since they are reported seperately
inside the CSV file
Change-Id: If30ec3f22ce5871f911433e6a6997e9c6e7ca210
specifiying the rx/tx_freq for the eNB is optional. But when
it's done it needs to be correct. There seems to be an issue
when mapping the freq[0-3] between srsENB and Amarisoft eNB.
So it's safe to just drop the config here altogether.
Change-Id: I70f5776173c2df5894031f68df2f6c3a900b79f8
tgz files in trials can be categorized in subdirectories, allowing to
select different bianry files at runtime based on the target run node
which is going to run them. This way for instance one can have a binary
linked against libs for eg. CentOS under run_label "centos/" or an ARM
target under "arm", and then use "run_label: arm" on the resource using
it.
Change-Id: Iaf2e97da3aff693395f44f0e93b184d4846cf6da
we've only waited when running with ZMQ but in fact we need
to always wait because the UE may need some time to
gracefully exit, i.e. send the detach request and shut down correctly.
this is checked in cleanup() and verity_metrics(), so also before copying back
the CSV metrics. But we only wait once.
Change-Id: I47e46b8ccce41c9a62d2d6866260d22c927e710d
this allows to sum dl_brate and ul_brate and perform
the known operations, like max, average, etc. on both
UL and DL bitrate combined
For examples, with this CSV ..
time;cc;pci;earfcn;rsrp;pl;cfo;dl_mcs;dl_snr;dl_turbo;dl_brate;dl_bler;ul_ta;ul_mcs;ul_buff;ul_brate;ul_bler;rf_o;rf_u;rf_l;is_attached
1999;0;1;3400;-67;59;913;27;29;0.50;10432.000000;0;2.6;20;0.0;10848.000000;0;0.0;0.0;0.0;1.0
2998;0;1;3400;-67;59;912;28;29;0.50;2136.000000;0;2.6;20;0.0;1680.000000;0;0.0;0.0;0.0;1.0
3997;0;1;3400;-69;61;909;28;29;0.50;2136.000000;0;2.6;20;0.0;1680.000000;0;0.0;0.0;0.0;1.0
.. one could calculate the average of DL and UL bit rate with:
ue.verify_metric(1e6, operation='avg', metric='dl_brate+ul_brate', criterion='gt', window=test_duration)
and veriy that it's average is greater than 1 Mbit/s
Change-Id: I5c7c80bb107fa6b93b215176e6ebbb5dc8594860
this allows to calculate the rolling average over a specific
window in time and take the maximum of that
this is useful to get average value for 30s UDP traffic for
example from the UE metrics
Change-Id: I34bbfe08dbc1f27b86c9805f54649d44d697fa18
It is known that sometimes srsENB hangs until it is killed -9, specially
when using ZMQ backend. Let's use RemoteProcessSafeExit in order to make
sure it is killed in an acceptable time (srs binaries use some
preventive sigalarm 6 seconds auto-kill procedure, hence we use 7
seconds) before next test is started and potentially try to re-use the
same ENB and fails due to previous one still running.
Change-Id: I905bd753c7822feccf1c1bb59752698f1d1b85f0
The API was doing far more stuff than its name indicated. Even more
important stuff, like making sure the process is killed at the end with
-9 after ssh connection is dropped.
Change-Id: If043ecab509b34b0922824d73db916196274ec64
smpplib modules are used in several places, so it's not enough to import
them in the constructor since they are only imported for the function
scope.
Change-Id: I6511e28c7ef6aa86e85a7e03aa10a67d87896588
numpy is used in a few places in the class code so we need to jhave it
available in a bigger scope, while still only loading when needed.
Change-Id: Iea66e623e1c980a62d691e20dacb00df99fdd78f
we need to make the rf_port a param in the cell list because for CA with
two cells on different EARFNCs, the RF port is different
for two cells without CA on the same EARFCN the RF port is the same
so we need to be able to configure it
Change-Id: I32e83cc34811588757104c5ce1cae7277402addd
This test currently only works with AmarisoftENB and srsUE, because:
* srsENB has no handover support yet
* AmarisoftUE doesn't provide any known way to verify handovers.
As a result, it usually runs with following filters and modifiers:
amarisoftenb-rftype@uhd+srsue-rftype@uhd+mod-enb-ncells@2+mod-enb-meas-enable
The get_counter() API may change in the future based on information
available from other implementations.
Change-Id: I0dff4af7e3a89804a03842ad4a4d697bca3a9773
Andre Puschmann
Pau Espin