[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Failure to load squashfs module



Daniel Baumann wrote:
> squashfs in-tree patches its magic number into the kernel mount routine,
> if it is built out-of-tree it's not. I hope one can either mount the
> image without having this sort of image-type-detection (by cheating on
> mount), otherwise we can't use squashfs out-of-tree which I really,
> really want to avoid under all circumstances.

I can verify that the out-of-tree squashfs module works with the
official kernel in etch, as my own live CD[0] uses it and mounts a
squashfs filesystem fine.  The relevant lines from my initramfs hook:

    force_load squashfs
    force_load unionfs
    force_load isofs
    force_load loop

And an excerpt from the initramfs script:

    cd /sys/block && for bd in hd* sd*; do
        mount -t iso9660 -o ro /dev/$bd ${isofs} >/dev/null 2>&1
        if [ $? = 0 ]; then
            if [ -e ${isofs}/boot/pod.squashfs ]; then
                mount -t squashfs -o loop,ro ${isofs}/boot/pod.squashfs ${squashfs}
                if [ $? != 0 ]; then
                    panic "ALERT!  unable to mount pod squashfs filesystem."
                fi

                mount -t unionfs -o noatime,dirs=${tmpfs}=rw:${squashfs}=ro none ${rootmnt}
                if [ $? != 0 ]; then
                    panic "ALERT!  unable to mount pod unionfs filesystem."
                fi

See the pod-source tarball in [0] for the complete scripts.

[0] http://debian.cc.gt.atl.ga.us/pod/

-- 
Robert Edmonds
edmonds@robertedmonds.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.debian-unofficial.org/pipermail/live/attachments/20060313/f193a3da/attachment.pgp

Reply to: