Hi all,
this is an RFC series implementing an NBD server embedded inside QEMU.
This can be used in various cases, including migration with non-shared
storage.
Three new commands are introduced at the QMP level
{ 'command': 'nbd-server-start', 'data': { 'addr': 'IPSocketAddress' } }
{ 'command': 'nbd-server-add', 'data': {'device': 'str', '*writable': 'bool'}
}
{ 'command': 'nbd-server-stop' }
At the HMP level there is no nbd_server_add command. nbd_server_start
automatically exposes all of the VM's block devices, and an option -w
makes them writable.
The NBD server exports multiple devices on a single port; they can be
accessed using "nbd:HOST:PART:exportname=NAME".
The patches are mostly boring touching nbd.c. The part where I need
a second opinion and/or ack is patch 12 and 13. They fix the case of
a disk being unplugged while NBD export is active. To do this I add a
NotifierList to a BlockDriverState. Does this look okay, or is it too
ad hoc?
Paolo Bonzini (13):
nbd: add more constants
nbd: pass NBDClient to nbd_send_negotiate
nbd: do not leak nbd_trip coroutines when a connection is torn down
nbd: close all clients on deleting export
nbd: register named exports
nbd: negotiate with named exports
nbd: do not close BlockDriverState in nbd_export_close
qemu-sockets: publish dummy_opts
qmp: add NBD server commands
qemu-sockets: make inet_parse public
hmp: add NBD server commands
block: add close notifiers
nbd: add notifier to close exports when the image is closed
Makefile.objs | 5 +-
block.c | 19 +++-
block.h | 1 +
block_int.h | 2 +
blockdev-nbd.c | 131 ++++++++++++++++++++++
hmp-commands.hx | 29 +++++
hmp.c | 66 +++++++++++
hmp.h | 2 +
nbd.c | 311 +++++++++++++++++++++++++++++++++++++++++-----------
nbd.h | 6 +
qapi-schema.json | 69 ++++++++++++
qapi/opts-visitor.c | 48 ++++----
qemu-nbd.c | 1 +
qemu-sockets.c | 16 +--
qemu_socket.h | 3 +
qmp-commands.hx | 16 +++
16 file modificati, 621 inserzioni(+), 104 rimozioni(-)
create mode 100644 blockdev-nbd.c
--
1.7.11.2
