> > Of course not. You're not supposed to create it. It's created by the
> > kernel. If /dev/.devfsd doesn't exist, it means you haven't mounted
> > devfs on /dev.
>
> Clearly, there's something that I've not understood. Well, from what
> you're writting, I get it that devfsd and the stuff done by rc.devfs
> and in rc.sysinit:
>
> # If we're using devfs, start devfsd now - we need the old device names
> [ -e /dev/.devfsd -a -x /sbin/devfsd ] && /sbin/devfsd /dev
>
> are actually the devfs_d_aemon stuff and devfs (which is integrated
> into the kernel) does not need anything to run, but being mounted. The
> question is : how do you mount a devfs "pseudo-volume" on /dev?
There is a CONFIG option to make devfs mounted at boot time, or you
can do "devfs=mount" at the boot prompt, or you can do:
# mount -t devfs none /dev
Of course, it requires devfs to be compiled into your kernel.
> I've got the problem with the root_fs distributed with
> user-mode-linux. For example with the root_fs_redhat_6.2_big, I got:
Which kernel version is this?
> bash# cat /proc/devices
> Character devices:
> 1 mem
> 2 pty
> 3 ttyp
> 4 ttys/%d <--- aren't they devfs devices ?
> 5 serial/%d <--- aren't they devfs devices ?
Yes they are. And they shouldn't appear unless CONFIG_DEVFS_FS=y.
> bash# cat /proc/filesystems
> nodev proc
> nodev sockfs
> nodev tmpfs
> nodev pipefs
> ext2
> nodev devpts
>
> Why is there no 'nodev devfs' line here?
That's a bloody good question. Also check your dmesg output for devfs
messages:
% dmesg | fgrep devfs
> Perhaps it's the problem, but how to solve it?
For some reason, one part of your kernel thinks it has devfs, the
other part doesn't. Either you've got a strange patch applied, or
perhaps you didn't do a "make dep;make clean" after reconfiguring your
kernel.
Regards,
Richard....
Permanent: rgooch@xxxxxxxxxxxxx
Current: rgooch@xxxxxxxxxxxxxxx
|