systemd files and tgtd outer loop

I'm using as-packaged in Ubuntu 24.04. But reading tgtd.c I was struck by s=
ome of the logic.=0A=
=0A=
while I get the purpose of sd_notify(), technically and from an ordering pe=
rspective, shouldn't that be  only when '-exec' is called later on and targ=
ets are defined? Or are we relying on one of the 'ExecPost' actions to blow=
 up and indicate the service is failed? But isn't that way too late? Downst=
ream dependencies have already fired.=0A=
=0A=
I'm assuming '=97forground' is chosen in a systemD context is so that logs =
will be picked up by systemD. So why not #ifdef USE_SYSTEMD to also emit th=
e PID to stdout, if pidfile is undefined?=0A=
=0A=
Also the choice of LOG_WARNING in log_init() is highly unusual. It should b=
e LOG_INFO if perhaps LOG_NOTICE. A routine message is never emitted at WAR=
N.=