initrd, /dev/root, VFS, 2.4.19 heaven :-)

Bill Maltby - LFS Related lfsbill at
Thu Oct 24 04:51:17 PDT 2002

On Thu, 24 Oct 2002, martial daumas wrote:
>Le Wed, 23 Oct 2002 14:17:02 +0000 (UTC)
>richard at (Richard Lightman) a écris:
>> Get rid of that '-o 0 -g 0'. If I was root[...]

>> Get rid of everything after inserting modules.
>> The kernel will mount the root partition if /root/ is in the ramdisk
>> The kernel will mount /root/dev if you configured it to maount at boot.
>> The kernel will pivot_root for you. linuxrx cannot, as it is not process 1.
>> You can start devfsd from your init scripts.
>> Tell me if that does not work.
>Well, i've just woken up, I immediatly tested your method, while thinking "baaa, it

>when initrd is over etc...  And when i `mount` , i get this:
>rootfs on / type rootfs (rw)
>/dev/root.old on /initrd type ext2 (rw)
>proc on /initrd/proc type proc (rw)
>/dev/root on / type ext3 (rw)
>none on /dev type devfs (rw)
>proc on /proc type proc (rw)

Clean up: when init has been spawned, as you see from above, the
/dev/root.old is left mounted on /initrd. This allows things that were
running to keep on running. Since you are now into the "production" stage,
most (all?) of those things should be gone. You can now umount
/dev/root.old and free the memory used by the initrd (this is not done
automatically). Use

   blockdev --flushbufs <your-initrd-device>

and you will see 2/4/6... MB "reappear" for use.



Bill Maltby
billm at

Unsubscribe: send email to listar at
and put 'unsubscribe lfs-support' in the subject header of the message

More information about the lfs-support mailing list