On Mon, 2016-11-21 at 11:01 +0000, Simon McVittie wrote: > On Mon, 21 Nov 2016 at 10:16:50 +0000, Luca Boccassi wrote: > > The context is a squashfs and liveboot/build based ISO. > ... > > I don't pick the mount utility in any of my scripts or configs, so I do > > not know when or where the change happened, could have been anytime > > since Jessie was released. But I am 100% sure that it's the util-linux > > one being used in this case, as proved by the fact that --move fixed the > > boot process. > > I believe this was considered to be a bug in live-boot-initramfs-tools > (and its fork open-infrastructure-system-boot): l-b-i-t and o-i-s-b added > /root/bin to $PATH, causing the to-be-booted system's util-linux /bin/mount > to be used instead of the initramfs's busybox or klibc mount. > A kernel and initramfs-tools maintainer stated in > <https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=823856#20> that this > is not supportable - I believe his position was that every command must > either run in the initramfs with $PATH entirely from the initramfs, > or run chrooted into the to-be-booted system with $PATH entirely from > the to-be-booted system, but never a mixture of the two. > > This was fixed as #823069 for l-b-i-t and as #832752 for o-i-s-b. Both > bugs were treated as having 'critical' severity. > > > Of course as you all pointed out this won't work with klibc mount and it > > needs to be fixed. But please if possible don't revert, as it will break > > again virtual filesystem based images. > > My understanding is that if you build your images > with live-boot-initramfs-tools (>= 1:20160511) or with > open-infrastructure-system-boot (>= 20160601-1), both of which are > available in testing, then reverting the udev change altogether > (consistently using mount -o move everywhere) should be harmless.
And you are indeed right! I had the previous version. I just checked with the latest and it works with the -o mount on the squashfs image. I do think it's nice to support all mount utilities available, but given it was a bug to use the util-linux one and there is a workaround personally I don't mind if the udev script is reverted to the previous behaviour. Thanks for pointing out my mistake, and sorry again for the noise. Kind regards, Luca Boccassi
signature.asc
Description: This is a digitally signed message part
