On Sun, Jan 24, 2010 at 9:42 AM, Thomas Bächler <thomas@archlinux.org> wrote:
Okay, here are the changes:
initscripts: - hwclock improvements by Dan - some minor crypto fix on shutdown - moved the prekillall hook so it can be used better with splash Necessary for device-mapper/udev, as dm-device detection on startup has needs options introduced in 2009.11-1 (still in testing)
udev: - upstream update - remove resolve-modalias, replace with modprobe --resolve-alias
device-mapper/lvm2: - cleaned PKGBUILD, hopefully fix the dmeventd path mess - upstream update
cryptsetup: - upstream update, hopefully fixes lots of bugs
These have to be moved together (and with mkinitcpio) so that device-mapper still works fine with udev. This should solve FS#16735 finally - it has to be moved in any case, because this is better than the situation in core, even if some people will still have problems.
I want to move these 5 and mkinitcpio out of testing before I push the kill-klibc mkinitcpio in - the latter will also require changes in lvm2 and cryptsetup.
You'll also need to move the lilo-22.8-4 in testing with the new device-mapper. My i686 system using lilo and lvm2 booted fine.