Re: Using live-boot with mainline kernel
On 03/18/2011 01:57 PM, David Kuehling wrote:
> As mentioned before I'm currently trying to make live-boot work for a
> powerpcspe embedded system. Currently I'm testing this with a fat
> partition on a usb hard-disk attached to the system.
unionfs is going to disappear really soon in 3.x; unionmount support is
not yet in kernel mainline, there's experimental patches against
live-boot available[0] (and it's a quite high priority to get them
merged into live-boot soon).
or in other words: you will have to patch the kernel, either for using
aufs, or by using unionmount.
> Unfortunately live-boot is currently hardwired to work with either aufs
> or unionfs, which both are not in the mainline kernel.
...but, fortunately, in the debian kernel :) any patches to de-hardwire
it are most appreciated.
> - using union=unionfs-fuse could work with a standard kernel; however
> unionfs-fuse support is currently broken (I understand that you
> don't officialy support it)
it's not unsupported per se, just that it's not of any high priority
since we do have aufs in the debian kernel (which is a much faster and
more reliable solution anyway). however, patches to fix it for
unionfs-fuse are nevertheless very welcome too.
> - The 'toram' option already allows moving images into RAM. Together
> with options 'toram exposedroot skipunion', this *could* work. But
> it is broken by copy_live_to always re-mounting the tmpfs read-only
> (via 'mount -r -t move')
sort of the same as with unionfs-fuse, would be nice to be fixed, but
not of high priority, patches to fix it more than welcome.
> - The 'exposedroot' option currently relies on mounting writable tmpfs
> to /var/tmp, /var/run etc. Why does it have to use a unionfs to do
> that? Can't it just directly mount tmpfs over /var/tmp etc.,
> shadowing the underlying (empty) directories?
i've no idea. to be honest, i didn't completely get the idea of exposed
root yet anyhow.
> Any chances that a patch to make live-boot work with mainline kernel
> could get accepted?
patches are always welcome.
> Any preferences as to how to achive that? Using a
> writable ramdisk seems to be the most simple option. Maybe just make
> copy_live_to use a writable ramdisk whenever 'skipunion' is set?
wrt/ to the problem to boot without neither using aufs nor any of its
alternatives, i don't think there's a 'true and only' way to do it.
sometimes, multiple ways should be available at the same time for
different use cases. so, if you're going to work on that, regardless how
you're solving it, it would be nice to have that stuff in mainline
live-boot directly (pulling through git most appreciated, but sending
patches to the ml is fine too).
--
Address: Daniel Baumann, Burgunderstrasse 3, CH-4562 Biberist
Email: daniel.baumann@progress-technologies.net
Internet: http://people.progress-technologies.net/~daniel.baumann/
Reply to: