Comment 16 for bug 435027

David Gibson (dwg) wrote :

Well, that approach sounds reasonable. However it seems odd to me to postpone a feature which will make it substantially easier to analyse and workaround bugs until after you've fixed the bugs.

I think the current mountall brokenness is the one I've reported in bug 44794. The "Can't remove /forcefsck" is because mountall dies due to bug 44794, then its post-stop action attempts to remove /forcefsck, but since mountall hasn't actually completed, / is still read-only. Seems like that post-stop should only be post a successful completion.

I can't check much else about mountall right now, because my latest kernels (both Ubuntu and home-built) currently don't get out of the initramfs, due to initramfs-tools bugs.