Re: [Qemu-devel] [PATCH v2 0/4]: QMP: add BLOCK_MEDIUM_EJECT event

[Luiz Capitulino <lcapitulino@redhat.com>]

On Fri, 17 Feb 2012 11:49:21 +0100
Paolo Bonzini <pbonzini@redhat.com> wrote:

> On 02/15/2012 07:42 PM, Luiz Capitulino wrote:
> > There's been one non-rfc patch and one or two rfc ones. This is v2 of the
> > non-rfc one.
> > 
> > I think this version does what Kevin and Markus were asking: the event is
> > emitted whenever the tray moves, be it the guest or HMP/QMP commands.
> > 
> > In a previous email I said that I'd be reviving an old series that breaks the
> > eject and change commands into multiple tray commands (tray-open/close,
> > medium-insert/remove etc), but that's not so trivial anymore as the tray state
> > moved to device models (still possible of course, but requires a bit more
> > work). So I decided to do it the way patch 4/4 does it.
> > 
> >  QMP/qmp-events.txt |   17 ++++++++++
> >  block.c            |   84 +++++++++++++++++++++++++++++++++------------------
> >  block.h            |    8 ++--
> >  block/raw-posix.c  |    6 ++--
> >  block/raw.c        |    4 +-
> >  block_int.h        |    2 +-
> >  hw/ide/atapi.c     |    2 +-
> >  hw/ide/core.c      |    8 ++--
> >  hw/scsi-disk.c     |    8 ++--
> >  hw/virtio-blk.c    |    6 ++--
> >  monitor.c          |    3 ++
> >  monitor.h          |    1 +
> >  12 files changed, 97 insertions(+), 52 deletions(-)
> 
> We can seek perfection, but we can also do that incrementally.  We're
> discussing the code more than the actual behavior, so
> 
> Acked-by: Paolo Bonzini <pbonzini@redhat.com>

I'll submit v3 with a few renames and the change proposed by Markus, which is
not to call bdrv_eject() in the device models if the tray didn't move.