It's hard to distinguish multiple wireshark windows with trace.pcap
files, as only the file name and not the path is shown in the title.
Make it easier to distinguish them by adding the log name to the file
name.
Change-Id: I7007d7c1e74aef46f1942f4f6f675adebb890955
Fix the logic that is supposed to display a "q Enter to close" message
once and wait for user input, in case one of the Osmocom programs
crashed.
Explicitly read from /dev/tty, as it may not be connected as default
stdin anymore at this point. This happens when using the new code paths
with tmux as terminal, but also happened before recent refactorings for
me with all terminals.
Related: https://media.ccc.de/v/osmodevcall-20211210-osmith-osmo-dev-ttcn3#t=801
Change-Id: Iebb799493f76fd57f24b15c998ded2bd8e284e6b
Create one tmux session, with each Osmocom program in one window, if
tmux is chosen as "terminal". This allows using osmo-dev's run.sh over
ssh without X forwarding. Also the existing supported terminals
automatically scroll to the bottom as soon as there is new output,
making it hard to read logs.
Change-Id: Iaef08fb3cb3c6f58cd8490a631e177a03af9c974
Instead of passing a whole script to the terminal via command-line
argument, write it to a file. This is in preparation to use the same
script with tmux, see next patch.
Change-Id: I15760d706b20628421c10a16514120f442d5477f
Replace the previous approach of only storing the PIDs of the spawned
terminals, and then attempting to kill all other processes (the ones
running inside the spawned terminals and tcpdump) with "killall".
Instead, store PIDs of all spawned programs:
* Terminals
* Osmocom programs (new)
* Tcpdump (new)
Kill old processes at the start of run.sh as well, so even if run.sh was
killed before it could clean up, the Osmocom programs don't fail to
start anymore (as it was the case before).
Change-Id: I379ef71de2f1ab0ac4a28064e5cf9e275c4c61e9
Prepare to add "tmux" as new terminal. As discussed in code review, we
decided to drop support for auto-detection.
Change-Id: I0afb6b0242c399334a2c37f18a5d26d5beeabedc
Make STP_CN_IP and STP_RAN_IP configurable. If they are not set to the
same IP, run a second OsmoSTP and OsmoBSCNAT to connect both OsmoSTP.
Related: OS#2545
Change-Id: I8bfbd00bfeeebc6e1b87a8cd6333212db4380653
Put the terminal PIDs into a variable instead of running kill %1 %2 etc.
This makes it easier to conditionally spawn some of the programs.
Change-Id: Iaf6c175de2edf720009af13eace2b0ef83b0735b
Replace the old check with an exact check for the given rule. The old
check doesn't work on my system if docker is running and sets up other
masquerading rules.
Change -A to -I, so it is less likely to be affected by whatever other
rules may already be above in the chain.
Change-Id: I30c22fe4e7683611e8dada329dbf7d51f78b6016
This is what I did to get external MNCC with kamailio to work.
(As before, only change MSC_MNCC from "internal" to "external ..." in config_*
to enable use of kamailio as external MNCC.)
Change-Id: I18c991bf914dedeb69cb8a2fcde117633e51a899
Since commit [1], the MNCC socket path is configured in the .cfg file. There is
no need to pass the -M option to osmo-msc on the cmdline anymore.
Also, the udtrace example to launch osmo-msc in run.sh includes a '| tee' so
that the -M would be passed as option to tee and breaks that example.
[1] cbdd7180a6 / Ie6352c584c0fce57c617387160c285abf40a5f90
Change-Id: Id5c177a7fe98b752881ddd0d98152a2e30380d0e
Adopt "foreach" and "common_" features and generally take in the improved
templates as used at 35c3.
If your config file relies on the tmpl_std that is now removed, you can locally
resurrect it with
git checkout 2c6a55baa7 -- tmpl_std
but it's probably better to make a new copy of config_2g3g and insert your
local config there.
Change-Id: Ie6352c584c0fce57c617387160c285abf40a5f90
Oliver Smith
Neels Hofmeyr