On 3/9/20 10:36 AM, Daniel P. Berrangé wrote:
On Mon, Mar 09, 2020 at 04:21:12PM +0100, Kevin Wolf wrote:
Am 06.03.2020 um 23:51 hat Eric Blake geschrieben:
For qcow2 and qed, we want to encourage the use of -F always, as these formats can suffer from data corruption or security holes if backing format is probed. But for other formats, the backing format cannot be recorded. Making the user decide on a per-format basis whether to supply a backing format string is awkward, better is to just blindly accept a backing format argument even if it is ignored by the contraints of the format at hand.
Signed-off-by: Eric Blake <eblake@redhat.com>
I'm not sure if I agree with this reasoning. Accepting and silently ignoring -F could give users a false sense of security. If I specify a -F raw and QEMU later probes qcow2, that would be very surprising.
And if the user specifies "-F raw" and we probe qcow2, and the user does not realize this, they can become silently reliant on always probing qcow2. If we then honour the "-F raw" option in a later QEMU release, we'll break the behaviour they've relied on.
IMHO, we must not accept "-F fmt" unless we're in a position to honour it.
So I'm thinking: qemu-img create -f qcow -b backing.qcow -F qcow img.qcow => okay qemu-img create -f qcow -b backing.raw -F raw img.qcow => okay, slightly risky (if backing.raw is ever changed to be non-raw), but then again, backing files tend to be read-only (do we even support commit on qcow images, or do we limit that to qcow2?) qemu-img create -f qcow -b backing.qcow -F raw img.qcow => fails, due to mismatch qemu-img create -u -f qcow -b anything -F anything img.qcow $size => fails: we can't write -F into the image, nor can we open anything to probe its type to check that -F was correct qemu-img create -f qcow -b backing.qcow img.qcow => warns, but okay (we did not get -F, but the probe works out) qemu-img create -f qcow -b backing.raw img.qcow => likewise warns qemu-img create -f qcow -b backing.qcow2 img.qcow => error; new qcow images (which you should avoid where possible anyways) must be backed by only raw or qcow, going forward Other scenarios? Do the above ideas look reasonable? -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org