/tech/ - Technology★

UNIX braindamage

No, because signal handling and killing pid 1 kills all processes

Systemd-init is only an init system. Systemd-init is the only process that's PID1. All the other parts of systemd are optional and run in their own processes that are not PID1.

>>1060243

systemd haters BTFO once again

>>1060261

You're actually wrong. Only systemd-init runs as PID1.

It also initializes new processes, which fork() off of it.

>>1060238 (OP)

The mac os init is in PID 0, and the old mac os' (pre X) init was in the ROM itself.

>>1060238 (OP)

It only needs to keep running to reap orphaned zombies. In linux at least, all exited processes must be acknowledged by user space by its parent calling wait(). Processes whose parents have exited themselves get reparented to pid 1, so init needs to execute anything you want to bring up the system and then do while (1) wait(&status);

You could just do it the old fashioned way. Have /etc/rc be a shell script that does what you want it to and use dalias' init: http://ewontfix.com/14/ - though if you do I'd advise linking it statically against musl to make it as robust as he claims.

>>1060269

I don't know what's running on your systems but there's at least 22M of code loaded into my systemd pid1 (standard Fedora 29 install)

; ps aux | grep systemd | sed 1q
root         1  0.0  0.0 171944 13812 ?        Ss   Mar31   0:21 /usr/lib/systemd/systemd --switched-root --system --deserialize 33
; du -hcL /usr/lib/systemd/systemd `{ldd /usr/lib/systemd/systemd | grep -v linux-vdso | cut -d'>' -f2 | cut -d'(' -f1}
1.6M	/usr/lib/systemd/systemd
2.9M	/usr/lib/systemd/libsystemd-shared-239.so
64K	/lib64/librt.so.1
300K	/lib64/libseccomp.so.2
188K	/lib64/libselinux.so.1
388K	/lib64/libmount.so.1
80K	/lib64/libpam.so.0
136K	/lib64/libaudit.so.1
100K	/lib64/libkmod.so.2
104K	/lib64/libgcc_s.so.1
256K	/lib64/libpthread.so.0
2.7M	/lib64/libc.so.6
28K	/lib64/libcap.so.2
60K	/lib64/libacl.so.1
356K	/lib64/libcryptsetup.so.12
1.2M	/lib64/libgcrypt.so.20
40K	/lib64/libip4tc.so.0
132K	/lib64/libidn2.so.0
188K	/lib64/liblzma.so.5
120K	/lib64/liblz4.so.1
372K	/lib64/libblkid.so.1
224K	/lib64/ld-linux-x86-64.so.2
536K	/lib64/libpcre2-8.so.0
32K	/lib64/libdl.so.2
40K	/lib64/libuuid.so.1
28K	/lib64/libcap-ng.so.0
104K	/lib64/libz.so.1
32K	/lib64/libattr.so.1
368K	/lib64/libdevmapper.so.1.02
612K	/lib64/libssl.so.1.1
3.1M	/lib64/libcrypto.so.1.1
40K	/lib64/libargon2.so.0
76K	/lib64/libjson-c.so.4
144K	/lib64/libgpg-error.so.0
292K	/lib64/libpcap.so.1
1.7M	/lib64/libunistring.so.2
728K	/lib64/libsepol.so.1
272K	/lib64/libudev.so.1
2.0M	/lib64/libm.so.6
22M	total

>>1060243

SystemD haters DESTROYED with facts and logic.

>>1060241

This, basically, it's to deal with zombie processes.

>>1060314

like the shit you intentionally want running like screen?

remember when systemd killed this?

<it still does

the only reason it doesn't anymore is it received so much backlash that all the distro's that rely on systemd overrided the default configuration.

>>1060289

>and the old mac os' (pre X) init was in the ROM itself.

Mac OS pre-X wasn't even *nix though so can you really say it had an "init" at all?

Because init also integrates a service manager for some fucking reason.