Hi,

> could you please tell the udev version running:
> dpkg -l udev

ii udev 0.114-2

> also stick an set -x on top of /usr/share/initramfs-tools/hooks/udev
> and set in the output you get then of mkinitramfs -o /tmp/foo

I'm not sure what you mean by "stick an set -x"... however if this is about
the file system rights of the hook scripts an

"ls -l /usr/share/initramfs-tools/hooks/" gives back:

-rwxr-xr-x 1 root root 498 2007-12-25 17:03 kernelextras
-rwxr-xr-x 1 root root 981 2008-06-04 17:21 keymap
-rwxr-xr-x 1 root root 444 2008-06-04 17:21 legacylvm
-rwxr-xr-x 1 root root 1316 2008-06-16 08:02 libsplashy
-rwxr-xr-x 1 root root 537 2008-06-16 08:02 splashy
-rwxr-xr-x 1 root root 844 2007-12-25 17:03 thermal
-rwxr-xr-x 1 root root 653 2007-08-06 06:12 udev
-rwxr-xr-x 1 root root 563 2008-07-07 13:50 udevhelper
-rwxr-xr-x 1 root root 600 2008-06-08 12:08 uswsusp

so they're all executable and always have been so the output of
mkinitramfs -o /tmp/foo shouldn't change.

Most interestingly in my opinion is the output of
update-initramfs -utv
which gives back a lot including lines like this:

/usr/share/initramfs-tools/hooks/keymap ignored: not alphanumeric or '_' file
/usr/share/initramfs-tools/hooks/udev ignored: not alphanumeric or '_' file
(The complete output of that is included in update-initramfs.log) which makes
clear that this is not an execution-rights issue.
Anyway these messages are issued by this code snippet
in /usr/share/initramfs-tools/script/functions:

set_initlist()
{
unset initlist
for si_x in ${initdir}/*; do
# skip empty dirs without warning
[ "${si_x}" = "${initdir}/*" ] && return

# only allow variable name chars
case ${si_x#${initdir}/} in
*[![:alnum:]_]*)
[ "${verbose}" = "y" ]
&& echo "$si_x ignored: not alphanumeric or '_' file"
continue
;;
esac

This causes the udev hook script to be skipped...

Cheers

Phil



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Oh sorry forgot to attach the update-initramfs.log in the previous message...
here it is

Phil

On Mon, 21 Jul 2008, Philipp Sternberg wrote:

> Oh sorry forgot to attach the update-initramfs.log in the previous message...
> here it is
>
> Phil

what is your default shell?

ls -l /bin/sh

obviously the code you are referring to is *working* on lots of boxes
so this looks like local breakage.

best regards

--
maks



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Tuesday, 22. July 2008, you wrote:
> what is your default shell?
> ls -l /bin/sh

Hi.. oh it's dash actually...

Ah... that looks much better!!!
I purged dash so that

ls -l /bin/sh gave

/bin/sh -> bash

afterwards.

That solved it!

(The ramdisk image now contains udev!)
Maybe there should be a package dependency forbidding dash to be installed or
(if that is possible to be set as default shell... Where is that configured
anyway??)

Anyway thanks a lot for your help and sorry for trying to set you on the wrong
track (that function-script thing) I did obviously not interpret the doings
of that stuff correctly...

Cheers

Phil



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org