Re: [Qemu-devel] [PATCH v3 0/7] Implement reference count for BlockDriverState

[Stefan Hajnoczi <stefanha@redhat.com>]

On Wed, Jul 31, 2013 at 06:13:53PM +0800, Fam Zheng wrote:
> BlockDriverState lifecycle management is needed by future features such as
> image fleecing and blockdev-add. This series adds reference count to
> BlockDriverState.
> 
> The first two patches clean up two odd BlockDriverState use cases, so all code
> uses bdrv_new() to create BlockDriverState instance.
> 
> Then implemented bdrv_ref() and bdrv_unref() to operate on refcnt: Initially,
> refcnt is 1, which means bdrv_unref is effectively a bdrv_delete() here. So
> patch 04 has a search and replace to convert bdrv_delete to bdrv_unref, before
> bdrv_ref is used anywhere. 05~08 patches calls bdrv_ref for device attach,
> block-migration and nbd.
> 
> The rule is: Either bdrv_ref() or bdrv_new() must have a matching
> bdrv_unref() call, and the last matching bdrv_unref deletes the bs.
> 
> v3:
>     03: Removed unnecessary bdrv_close() call.
> 
> v2:
>     05: Removed: "block: use BlockDriverState refcnt for device attach/detach"
>     07: Fix xen_disk blk_disconnect() as it depended on device attach refcnt.
> 
> Fam Zheng (7):
>   vvfat: use bdrv_new() to allocate BlockDriverState
>   iscsi: use bdrv_new() instead of stack structure
>   block: implement reference count for BlockDriverState
>   block: make bdrv_delete() static
>   migration: omit drive ref as we have bdrv_ref now
>   xen_disk: simplify blk_disconnect with refcnt
>   nbd: use BlockDriverState refcnt
> 
>  block-migration.c         |  4 ++--
>  block.c                   | 44 +++++++++++++++++++++++++++++++++-----------
>  block/backup.c            |  2 +-
>  block/blkverify.c         |  4 ++--
>  block/cow.c               |  2 +-
>  block/iscsi.c             | 14 +++++++-------
>  block/mirror.c            |  2 +-
>  block/qcow.c              |  2 +-
>  block/qcow2.c             |  2 +-
>  block/qed.c               |  2 +-
>  block/sheepdog.c          |  6 +++---
>  block/snapshot.c          |  2 +-
>  block/stream.c            |  2 +-
>  block/vmdk.c              | 10 +++++-----
>  block/vvfat.c             |  6 +++---
>  blockdev-nbd.c            | 10 +---------
>  blockdev.c                | 14 +++++++-------
>  hw/block/xen_disk.c       | 13 ++++++-------
>  include/block/block.h     |  3 ++-
>  include/block/block_int.h |  1 +
>  nbd.c                     |  5 +++++
>  qemu-img.c                | 26 +++++++++++++-------------
>  qemu-io.c                 |  6 +++---
>  23 files changed, 101 insertions(+), 81 deletions(-)

One interesting thing is the interaction between the DriveInfo and BDS
lifecycle.  Both now have refcounts but drive_init()/drive_uninit() do
bdrv_ref()/bdrv_unref() so you can be sure that the BDS will not go away
if you hold a DriveInfo reference.

Thanks, applied to my block-next tree:
https://github.com/stefanha/qemu/commits/block-next

Stefan