[ale] Shutdown process

[Chris Fowler]

I think I have it.  I'm beginning to work on this now.


If my RO FS has the following:


/sbin
/usr
/etc
/lost+found


and the ram image has


/dev
/usr
/usr3
/tmp


wil the / look like this after the kernel mounts /dev/ram7 as /


/dev
/usr
/usr3
/tmp
/sbin
/usr
/etc
/lost+found



My RO fs is not a image, it is actually part of a chip.


Thanks,
Chris
-----Original Message-----
From: byron at cc.gatech.edu [mailto:byron at cc.gatech.edu]
To: ale at ale.org
Sent: Wednesday, January 03, 2001 11:42 AM
To: ChrisF at computone.com
Cc: ale at ale.org
Subject: Re: [ale] Shutdown process



> I've read you email and understand but have one question.
> 
> if your script does this:
> 
> mount /dev/fla1 /usr -n -o ro
> gzip -cd ramdisk.gz > /dev/ram7
> mount /dev/ram7 to [WHAT]


It doesn't do the last command. That's the beauty of the initial ramdisk. it'll
automatically change the root device to the target specified as the root
parameter on the LILO argument line.


> 
> Can I remount on to /
> or do I need
> to do this
> 
> mkdir /system
> mount /dev/ram7 /system
> mount /dev/fla1 /system/usr
> chroot /system /system/sash 


Neither. Happens automatically. Take a read of Documentation/initrd.txt in
any /usr/src/linux kernel source for details.


> 
> My question is that if I already have a /
> how does /dev/ram7 become that /?


The kernel switches root from the initial ramdisk to the target root. I've used
this for mounting loopback filesystems as root too. Quite useful.


BAJ