It’s just the basic posix shell syntax. It just looks weird because they are using lots of library functions and in-place substitutions. also apparently the function, to translate a system path to something grub will understand, is an ELF binary 0_o
It’s just the basic posix shell syntax. It just looks weird because they are using lots of library functions and in-place substitutions. also apparently the function, to translate a system path to something grub will understand, is an ELF binary 0_o
check the GRUB_CMDLINE_LINUX
key in /etc/default/grub
it should contain the info about any subvolume. if it does not then there might be another grub config hook that is used by Fedora to add that info.
If you want to be able to change subvol names without having to touch the grub config you might also want to switch to using subvolid
instead of subvol
keys on the kernel command line, because the id will stay the same after a rename (this could backfire though if you assign functions to certain names like “fallback” etc.).
Edit: found the hook that adds the kernel command line option for btrfs subvolumes. in /etc/grub.d/10_linux
there is this bit of code:
case x"$GRUB_FS" in
xbtrfs)
rootsubvol="`make_system_path_relative_to_its_root /`"
rootsubvol="${rootsubvol#/}"
if [ "x${rootsubvol}" != x ]; then
GRUB_CMDLINE_LINUX="rootflags=subvol=${rootsubvol} ${GRUB_CMDLINE_LINUX}"
fi;;
xzfs)
[skipped for brevity]
;;
esac
so it seems it is always looking at the subvol name of the currently mounted root fs.
this is extra tricky because they did not specify the exact kernel. mainline could be any of the kernels tagged as stable that you can build from linus’ git tree. i know that in the past you could run a mainline linux on intel 368 chips but today you probably can not because official support was dropped a while ago.
OpenWRT does not use liblzma or systemd so i think that one is pretty safe. I would also be surprised if Android included OpenSSH server binaries in that way.