Re: [PATCH] qemu-img: Support bitmap --merge into backing image

[Eric Blake]

On 9/11/20 3:31 AM, Max Reitz wrote:
> On 09.09.20 14:33, Eric Blake wrote:
> > If you have the chain 'base.qcow2 <- top.qcow2' and want to merge a
> > bitmap from top into base, qemu-img was failing with:
> >
> > qemu-img: Could not open 'top.qcow2': Could not open backing file: Failed to get shared 
> > "write" lock
> > Is another process using the image [base.qcow2]?
> >
> > The easiest fix is to not open the entire backing chain of the source
> > image, so that we aren't worrying about competing BDS visiting the
> > backing image as both a read-only backing of the source and the
> > writeable destination.
> >
> > Fixes: http://bugzilla.redhat.com/1877209
> > Reported-by: Eyal Shenitzky <eshenitz@redhat.com>
> > Signed-off-by: Eric Blake <eblake@redhat.com>
> > ---
> >   qemu-img.c                 |  3 +-
> >   tests/qemu-iotests/291     | 12 ++++++++
> >   tests/qemu-iotests/291.out | 56 ++++++++++++++++++++++++++++++++++++++
> >   3 files changed, 70 insertions(+), 1 deletion(-)
> >
> > diff --git a/qemu-img.c b/qemu-img.c
> > index eb2fc1f86243..b15098a2f9b3 100644
> > --- a/qemu-img.c
> > +++ b/qemu-img.c
> > @@ -4755,7 +4755,8 @@ static int img_bitmap(int argc, char **argv)
> >       }
> >       bs = blk_bs(blk);
> >       if (src_filename) {
> > -        src = img_open(false, src_filename, src_fmt, 0, false, false, false);
> > +        src = img_open(false, src_filename, src_fmt, BDRV_O_NO_BACKING,
> > +                       false, false, false);
>
> Why not do the same for the destination BB?

Yeah, that should work, too.

> >           if (!src) {
> >               goto out;
> >           }
> > diff --git a/tests/qemu-iotests/291 b/tests/qemu-iotests/291
> > index 1e0bb76959bb..4f837b205655 100755
> > --- a/tests/qemu-iotests/291
> > +++ b/tests/qemu-iotests/291
>
> [...]
>
> > @@ -107,6 +116,9 @@ $QEMU_IMG map --output=json --image-opts \
> >   nbd_server_start_unix_socket -r -f qcow2 -B b2 "$TEST_IMG"
> >   $QEMU_IMG map --output=json --image-opts \
> >       "$IMG,x-dirty-bitmap=qemu:dirty-bitmap:b2" | _filter_qemu_img_map
> > +nbd_server_start_unix_socket -r -f qcow2 -B b3 "$TEST_IMG"
>
> Why not look into $TEST_IMG.base to see specifically whether the bitmap
> is there?

We did just that, several lines earlier, with the qemu-img info 
--backing-chain.

> (I also am quite surprised that it’s even possible to export bitmaps
> from backing nodes, but, well.)

I actually ought to call that out in the commit message.  It used to be 
that we were inconsistent on what we could see from the backing chain (a 
filter would make it so we can't), but as soon as your filter patches 
land, then we _do_ want to always be able to find a bitmap from the 
backing chain (incremental backup depends on that: we create an overlay 
disk to run the block-copy job as a filter, and _want_ to expose that 
overlay image with the bitmap it inherits from the original image).  So 
being able to export bitmaps from a backing node is normally a feature; 
and it is only in 'qemu-img bitmap' where we don't want accidental 
inheritance to get in the way from what we are actually merging.

I'll send a v2.

--
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org