Re: [RFC]: d-i, /dev, busybox

[Erik Andersen <andersen@codepoet.org>]

On Tue Feb 13, 2001 at 12:40:49PM -0800, Joey Hess wrote:
> Joey Hess wrote:
> > Erik Andersen wrote:
> > > No problem using ramfs.  When an initrd exits, the kernel does a pivot_root to
> > > whatever device is specified in /proc/sys/kernel/real-root-dev.  I am a bit
> > > embarrased to admit it, but I once built a special purpose embedded CD where it
> > > booted up, scanned a long list of devices, tried to mount them, and then looked
> > > for a magic file I had placed on the CD (you can't know apriori where you
> > > booted from).  If/when the file was found, I mounted it via loopback, echoed
> > > the major<<8+minor into /proc/sys/kernel/real-root-dev and then then kernel
> > > ended up with a read-only root-fs that was loopback mounted from the cdrom.
> > > You can do all sorts of goofy things this way.
> > 
> > Hm, but does ramfs have a device name associated with it? Ramfs's are
> > strange creatures..
> 
> s!device name!device major/minor!

If it is a nodev filesystem like proc ('cat /proc/filesystems'
to check) then, yes, we will need to explicitly rivot_root.

 -Erik

--
Erik B. Andersen   email:  andersee@debian.org, andersen@lineo.com
--This message was written using 73% post-consumer electrons--