On 4/24/20 10:18 AM, Peter Krempa wrote:
On Fri, Apr 24, 2020 at 09:52:00 -0500, Eric Blake wrote:
> On 4/24/20 4:24 AM, Peter Krempa wrote:
>> The implementation was never finished in libvirt. Remove it.
>>
>> Signed-off-by: Peter Krempa <pkrempa(a)redhat.com>
>> ---
>> src/util/virstoragefile.c | 19 ++-----------------
>> 1 file changed, 2 insertions(+), 17 deletions(-)
>
> Shouldn't we at least recognize if there is a qcow2 file with the extension
> header set, at which point we can error out and tell the user their image is
> unsupported, rather than trying to use it without the data file?
I don't think it's scalable to do so unless qemu exposes a whitelist of
supported extensions we can declare we support. Otherwise we are
band-aiding stuff we lear that exists and ignore any other extensions.
external data files have an incompatible feature bit set in addition to
the header extension. Maybe whitelisting header extensions is
prohibitive, but we should be able to recognize when a qcow2 file has an
incompatible feature bit set that we are not aware of, and assume that
since we don't know about the feature, we also don't know how to safely
pass that file to qemu.
That is, we could quickly reject qcow2 files with offset 72-79
containing bit 2 set, until we are ready to support external data files
in libvirt.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3226
Virtualization:
qemu.org |
libvirt.org