- Joined
- Dec 19, 2022
Yeah, but even in the early 90s boot drives were sufficiently large that you didn't need to do this any more, it's kind of silly to still be following this concept in the current year. We wouldn't really lose anything changing from /usr/bin et al to, say, /apps, especially not if you symlink the old directories to the new one (for a couple years at least). Same for /etc/ to /config/ or /run/media/$USER/$VOLNAME to /disks/$VOLNAME.Not really. /bin and /sbin were stuff needed to boot and were on the root drive/partition.
/usr was less critical stuff and would be available once the /usr partition was mounted.
/sbin was supposed to be critical statically linked stuff that would be usable in case you blew up the dynamic linker/libc/whatever. Obviously that's sort of morphed into "stuff for the superuser"
We didn't always have 4T boot drives.
My lawn something something.
Nobody cares about booting a second faster, people are opposing this change solely because it's Red Hat and Poettering. Which is silly, because it's a positive change.It's a necessary step to implement signed, read only root images, which he wants because it will make his laptop boot ever so slightly faster, and Rh wants because it grants them more control over their customers.
I want signed, read-only root images because it means SecureBoot compatibility and more security. As it is you can fairly easily replace someone's kernel image with a new one because running Linux means you either have to turn off a lot of these securities yourself, or your motherboard vendor did that for you. Even I have the technical know-how to inject something that makes the kernel fork an arbitrary process, imagine how much worse an actual hacker could do.