On Thu, Jun 14, 2012 at 6:51 PM, Kevin Wolf <kw...@redhat.com> wrote:
> Am 13.06.2012 16:36, schrieb Dong Xu Wang:
>> add-cow file can't live alone, must together will image_file and
>> backing_file.
>> If we implement qemu-img convert operation for add-cow file format, we must
>> create image_file and backing_file manually, that will be confused for users,
>> so we just ignore add-cow format while doing converting.
>>
>> Signed-off-by: Dong Xu Wang <wdon...@linux.vnet.ibm.com>
>
> NACK.
>
> This stupid "let's drop the feature, it might confuse users" attitude is
> known from Gnome, but not from qemu.
>
> There's no technical reason to forbid it and anyone who manages to
> create a valid add-cow image will also be able to specify the very same
> options to a convert command. Also, having image format specific code in
> qemu-img is evil.
>
If I implement add-cow convert command, I am wondering which method should
I use:
1) create add-cow, and its backing_file, and its image_file, then we
need 3 files.
2) create add-cow(with all bitmap marked to allocated), and its
image_file, then we
need 2 files.
2) will be easier, I should let .add-cow file can live only with
image_file, without backing_file.
I think both 1) and 2) need add code to qemu-img.c. Or I will have to create
image_file automaticly in add_cow_create function.
Can you give some comments on how to implement convert? Thanks.
> Kevin
>
