Re: [PATCH v2 1/4] qapi: net: Add query-netdevs command

[Alexey Kirillov]

05.03.2020, 15:03, "Markus Armbruster" <address@hidden>:
> Alexey Kirillov <address@hidden> writes:
>
>>  Add a qmp command that provides information about currently attached
>>  network devices and their configuration.
>
> Closes a gap in QMP; appreciated!
>
>>  Signed-off-by: Alexey Kirillov <address@hidden>
>
> [...]
>>  diff --git a/qapi/net.json b/qapi/net.json
>>  index 1cb9a7d782..4f329a1de0 100644
>>  --- a/qapi/net.json
>>  +++ b/qapi/net.json
>>  @@ -750,3 +750,92 @@
>>   ##
>>   { 'event': 'FAILOVER_NEGOTIATED',
>>     'data': {'device-id': 'str'} }
>>  +
>>  +##
>>  +# @NetdevInfo:
>>  +#
>>  +# Configuration of a network device.
>>  +#
>>  +# @id: Device identifier.
>>  +#
>>  +# @type: Specify the driver used for interpreting remaining arguments.
>>  +#
>>  +# @peer: Connected network device.
>
> @peer is optional. I assume its present when the device is connected
> (frontend to backend or vice versa). Correct?
>

Yes, this field stores connected frontend/backend device @id.

>>  +#
>>  +# @queues-count: Number of queues.
>
> We use plain @queues elsewhere in the schema.
>

It can conflict with fields inside Netdev*Options, isn't it?

>>  +#
>>  +# @hub: hubid of hub, if connected to.
>
> How @hub is related to @peer is not quite obvious to me. Can you give
> an example where @hub is present?
>

NetdevHubPortOptions has an option @hubid. @hub gives that id, if
netdev is connected to the hub via hubport. As example:

HMP:

hub 0
 \ hub0port1: socket.0: index=0,type=socket,
 \ hub0port0: virtio-net-pci.0: 
index=0,type=nic,model=virtio-net-pci,macaddr=52:54:00:12:34:56

QMP:

[
  {
    "peer": "hub0port0",
    "netdev": "hub0port0",
    "hub": 0,
    "model": "virtio-net-pci",
    "macaddr": "52:54:00:12:34:56",
    "type": "nic",
    "queues-count": 1,
    "id": "virtio-net-pci.0"
  },
  {
    "peer": "hub0port1",
    "listen": "127.0.0.1:90",
    "hub": 0,
    "type": "socket",
    "queues-count": 1,
    "id": "socket.0"
  },
  {
    "peer": "socket.0",
    "netdev": "socket.0",
    "hub": 0,
    "hubid": 0,
    "type": "hubport",
    "queues-count": 1,
    "id": "hub0port1"
  },
  {
    "peer": "virtio-net-pci.0",
    "netdev": "virtio-net-pci.0",
    "hub": 0,
    "hubid": 0,
    "type": "hubport",
    "queues-count": 1,
    "id": "hub0port0"
  }
]

>>  +#
>>  +# @perm-mac: Original MAC address.
>
> What does "perm-" mean?
>
> It's optional. When exactly is it present?
>

@perm-mac is the permanent (original) MAC address. It only used
for nic, because most of nic realizations can change MAC at
runtime and/or reset it to default (permanent) value.

>>  +#
>>  +# Since: 5.0
>>  +##
>>  +{ 'union': 'NetdevInfo',
>>  + 'base': { 'id': 'str',
>>  + 'type': 'NetClientDriver',
>>  + '*peer': 'str',
>>  + 'queues-count': 'int',
>>  + '*hub': 'int',
>>  + '*perm-mac': 'str' },
>>  + 'discriminator': 'type',
>>  + 'data': {
>>  + 'nic': 'NetLegacyNicOptions',
>>  + 'user': 'NetdevUserOptions',
>>  + 'tap': 'NetdevTapOptions',
>>  + 'l2tpv3': 'NetdevL2TPv3Options',
>>  + 'socket': 'NetdevSocketOptions',
>>  + 'vde': 'NetdevVdeOptions',
>>  + 'bridge': 'NetdevBridgeOptions',
>>  + 'hubport': 'NetdevHubPortOptions',
>>  + 'netmap': 'NetdevNetmapOptions',
>>  + 'vhost-user': 'NetdevVhostUserOptions' } }
>
> This is a copy of union 'Netdev' with a few additional common members
> (@peer, @queues-count, @hub, @perm-mac). I can't see how to avoid the
> duplication without adding nesting on the wire.
>
>>  +
>>  +##
>>  +# @query-netdevs:
>>  +#
>>  +# Get a list of @NetdevInfo for all virtual network devices.
>>  +#
>>  +# Returns: a list of @NetdevInfo describing each virtual network device.
>>  +#
>>  +# Since: 5.0
>>  +#
>>  +# Example:
>>  +#
>>  +# -> { "execute": "query-netdevs" }
>>  +# <- { "return": [
>>  +# {
>>  +# "peer": "netdev0",
>>  +# "netdev": "netdev0",
>>  +# "perm-mac": "52:54:00:12:34:56"
>>  +# "model": "virtio-net-pci",
>>  +# "macaddr": "52:54:00:12:34:56",
>>  +# "queues-count": 1,
>>  +# "type": "nic",
>>  +# "id": "net0"
>>  +# },
>>  +# {
>>  +# "peer": "net0",
>>  +# "ipv6": true,
>>  +# "ipv4": true,
>>  +# "host": "10.0.2.2",
>>  +# "queues-count": 1,
>>  +# "ipv6-dns": "fec0::3",
>>  +# "ipv6-prefix": "fec0::",
>>  +# "net": "10.0.2.0/255.255.255.0",
>>  +# "ipv6-host": "fec0::2",
>>  +# "type": "user",
>>  +# "dns": "10.0.2.3",
>>  +# "hostfwd": [
>>  +# {
>>  +# "str": "tcp::20004-:22"
>>  +# }
>>  +# ],
>>  +# "ipv6-prefixlen": 64,
>>  +# "id": "netdev0",
>>  +# "restrict": false
>>  +# }
>>  +# ]
>>  +# }
>>  +#
>>  +##
>>  +{ 'command': 'query-netdevs', 'returns': ['NetdevInfo'] }
>
> Like HMP "info network" and -net, this mixes frontends ("type": "nic")
> and backends. Unlike query-chardev and query-block. Hmm.
>
> A long time ago, all we had was -net: "-net nic" for configuring
> frontends, "-net none" for suppressing a default frontend + backend, and
> "-net anything-else" for configuring backends. "info network" showed
> the stuff set up with -net.
>
> In v0.12, we got -device for configuring frontends, and -netdev for
> backends. -netdev is like -net less "none", "nic", and the hub
> weirdness. "info network" was extended to also show all this.
>
> In v2.12, we got -nic, replacing -net nic.
>
> Unless I'm missing something, -net is just for backward compatibility
> now.
>
> What's the use case for query-networks reporting frontends?

In my vision, new QMP command is the replacement for old
HMP command. It must provide information about all
network devices, mainly for recreate similar net topology.
Currently, there are no differrence between fronted and
backend devices in context of my command, because
all of them use the same interface in NetClientState.

>

-- 
Alexey Kirillov
Yandex.Cloud