On Fri, Aug 09, 2013 at 06:01:53PM +0800, Fam Zheng wrote: > [resend to the correct list] > > 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. > > v4: > 08: Added, let block job use BDS reference. > 02: Fix leak of bs.opaque > > 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.
Sorry, can't merge this because it breaks qemu-iotests 041 and 055: $ ./check -qcow2 055 041 Please always run qemu-iotests before submitting patches. Stefan
