add systemd watchdog support

[mo-krauti]

I added systemd watchdog support, so I can run the service reliably in a headless setup. Right now I am using a subprocess call to the system-notify cli tool on every GSMTAP file write.
This is my systemd service:
scat.service

[Unit]
Description=scat

[Service]
Type=simple
ExecStart=/usr/local/bin/scat-run.sh
WatchdogSec=60
Restart=always
RestartSec=5

[Install]
WantedBy=multi-user.target

and the script scat-run.sh ensures that every run saves to a new pcap file:

#!/bin/sh
ts=$(date +"%Y%m%d-%H%M%S")
exec /usr/local/bin/scat -u -i 4 -t sec -L "nas,rrc" -m e303 -F "/var/log/scat-$ts.pcap"

This works for me. Do you have interest in integrating this functionality upstream?
I could also use the systemd-watchdog python library for improved performance instead of calling the cli via a subprocess, but this would introduce a new dependency. We could make it optional like the fastcrc functionality.
Thank you for this great project.

[peremen]

There are some users who reported to run SCAT in headless setup, so definitely anything helping on this direction is useful. I will consider adding a configuration file support to specify the baseband type/device node for headless setup in the future.

I have some suggestions though:

• SCAT can also run on BSDs, macOS and Windows, which do not have systemd. I'd implement a runtime check for the systemd availability and only call it when systemd is available.
• Also, if you call the systemd-notify on write_cp(), there might be cases where a modem is receiving the message but not sending any control plane message. Examples include private LTE network where no Paging is expected. I need to think on where else also could be used.