I've just pushed the following changes to master, which I believe address all the remaining RC bugs in the Debian BTS. I downgraded one bug that I couldn't reproduce. I've tested the result with a variety of configurations using LVM, md-RAID, cryptsetup, virtio, emulated PATA, emulated USB storage, real SATA and real USB storage. Everything works with no need to specify rootdelay. lvm2 needs to support the new local-block script phase, and I already made an NMU for that. mdadm apparently doesn't as its udev hook is sufficient. However it might be useful to add a sanity check to its postinst to check that INITRDSTART is still set correctly. Dependencies for md-RAID are very simple-minded and don't work at all with virtio or with multiple drivers for the component devices. But that's not a regression from 0.116. I intend to upload 0.119 to unstable with these changes (and the others already committed) on Wednesday so please point out any bugs as soon as possible. Ben. Ben Hutchings (10): mkinitramfs: Use default xz compression level rather than -8 hooks/fsck: Remove dependency on hwclock by disabling e2fsck time check NEWS: Begin a NEWS entry for the fsck and /usr-mount changes NEWS: Document the possible need to update mdadm INITRDSTART control: Ensure cryptsetup is upgraded to support mounting /usr Allow passing arguments to hooks and scripts local: Call local-block boot scripts to prepare additional block devices control: Add versioned Breaks on lvm2 without a local-block boot script local: Invoke local-block scripts from the loop in local_device_setup local: Use max(30, rootdelay) as timeout for block device to appear -- Ben Hutchings Anthony's Law of Force: Don't force it, get a larger hammer.
Attachment:
signature.asc
Description: This is a digitally signed message part