Re: Steve's crashing 1.3 machine, cured?

Gerd Knorr (kraxel@cs.tu-berlin.de)
Wed, 24 Apr 1996 19:56:14 +0200 (MET DST)


On Wed, 17 Apr 1996, Jamie Lokier wrote:

> BTW, with initrd I presume it's possible to load the root filesystem as
> a module. If so, maybe even the new nfsroot is redundant now, because
> it could all be done using user-space programs. Talking of nfsroot,
> Alexey's module system would presumably allow the memory for the nfsroot
> code to be swapped out or unloaded after it has done its job.
>
The kernel needs to understand the initrd filesystem, so you need
the initrd filesystem compiled in the kernel. You can't boot the kernel
with *all* filesystems as modules. But SCSI-people can boot a kernel
without a harddisk-driver: The scsi modules can be insmoded from initrd.

BTW: ext2-fs does not work as module. Becauce of securelevel: It is'nt an
exported kernel symbol (for security reasons I assume), but ext2
needs it.

About nfsroot: Yes, you can do the bootp procedure from user-space using
initrd, but you need another filesystem (beside NFS) for the initrd
compiled into the kernel. For diskless boxes (with NFS filesystems only)
it is overkill to compile in a complete filesystem just for doing bootp
from user-space...

Gerd

--
Lost my .pinerc after a linux crash. Forgot to reconfigure my From: line. Sorry
If you got kraxel@felix.none.de bounced, retry kraxel@cs.tu-berlin.de. Thanks