I'm curious about the two entries in /proc/mounts:
rootfs and /dev/root. Although I don't have much
experience with this, I've looked through some of the
kernel source code (e.g. fs/namespace.c or
init/do_mounts.c). All I can come up with (actually,
this is pretty much what I thought before I looked at
any code) is that rootfs is purely a kernel-generated
filesystem. The kernel creates the /dev/root device
there, and then mounts /dev/root at / (replacing
rootfs).
There were some mentions of 'initrd' too, so I'm not
sure if I'm confusing rootfs with an initrd. Could
anyone clear this up for me?
__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com
On 21 May 2003 04:14, Mark Dascher wrote:
> I'm curious about the two entries in /proc/mounts:
> rootfs and /dev/root. Although I don't have much
> experience with this, I've looked through some of the
> kernel source code (e.g. fs/namespace.c or
> init/do_mounts.c). All I can come up with (actually,
> this is pretty much what I thought before I looked at
> any code) is that rootfs is purely a kernel-generated
> filesystem. The kernel creates the /dev/root device
> there, and then mounts /dev/root at / (replacing
> rootfs).
Yes that's what happens. Guys are planning to be able to
*really* umount real (disk-based) root fs someday
and continue halt/reboot sequence using files (scripts etc)
on this virtual rootfs.
So there will be both 'early userspace' and 'late userspace' ;)
IMHO.
--
vda