[PATCH v2 0/5] Open vSwitch support in LNST

[Ondrej Lichtner]

From: Ondrej Lichtner <olichtne(a)redhat.com&gt;

Second version of the patchset that brings OvS support to LNST. The difference
is the way we define vlans - instead of using the vlan_tag/vlan_tags attributes
use the <vlan> element like this:
<vlan tag="1">
    <slaves>
        <slave id="port1"/>
        <slave id="port2"/>
    </slaves>
</vlan>

If a single port is present in more than one vlan it becomes a trunk port.

The previous example recipe therefore becomes this:
<lnstrecipe>
    <network>
        <host id="1">
            <params>
                <param name="os" value="fedora20"/>
            </params>
            <interfaces>
                <eth id="t1" label="ttnet1"/>
                <eth id="t2" label="ttnet2"/>
                <ovs_bridge id="xyz">
                    <define>
                        <alias name="aa" value="bb"/>
                    </define>
                    <slaves>
                        <slave id="t1"/>
                        <slave id="t2"/>
                    </slaves>
                    <vlan tag="1">
                        <slaves>
                            <slave id="t1"/>
                            <slave id="t2"/>
                        </slaves>
                    </vlan>
                    <bond id="mfg">
                        <slaves>
                            <slave id="t1"/>
                            <slave id="t2"/>
                        </slaves>
                    </bond>
                </ovs_bridge>
            </interfaces>
        </host>
        <host id="2">
            <interfaces>
                <eth id="t1" label="ttnet1">
                    <addresses>
                        <address value="192.168.100.215/24"/>
                    </addresses>
                </eth>
            </interfaces>
        </host>
        <host id="3">
            <interfaces>
                <eth id="t2" label="ttnet2">
                    <addresses>
                        <address value="192.168.100.216/24"/>
                    </addresses>
                </eth>
            </interfaces>
        </host>
    </network>

    <task>
        <run host="2" module="IcmpPing" timeout="30">
            <options>
                <option name="addr" value="{ip(3,t2)}"/>
                <option name="count" value="40"/>
                <option name="interval" value="0"/>
                <option name="limit_rate" value="95"/>
            </options>
        </run>
    </task>
</lnstrecipe>

Ondrej Lichtner (5):
  schema-recipe.rng: add element ovs_bridge
  RecipeParser: add support for ovs_bridge element
  Controller: add support for ovs_conf
  {Nm, Net}ConfigDevice: add support for OvS bridges
  InterfaceManager: generate ovs_bridge interface names

 lnst/Controller/Machine.py           | 10 ++++-
 lnst/Controller/NetTestController.py |  3 ++
 lnst/Controller/RecipeParser.py      | 79 ++++++++++++++++++++++++++++++---
 lnst/Slave/InterfaceManager.py       |  2 +-
 lnst/Slave/NetConfigDevice.py        | 86 +++++++++++++++++++++++++++++++++++-
 lnst/Slave/NmConfigDevice.py         | 19 +++++++-
 schema-recipe.rng                    | 78 ++++++++++++++++++++++++++++++++
 7 files changed, 268 insertions(+), 9 deletions(-)

-- 
1.8.5.3

[PATCH v2 1/5] schema-recipe.rng: add element ovs_bridge

[Ondrej Lichtner]

From: Ondrej Lichtner <olichtne(a)redhat.com&gt;

This element represents an Open vSwitch bridge. This element has a
mandatory child element <slaves> that defines which previously defined
interfaces should be connected to the new bridge as ports.

To create VLANs on this bridge we can use the <vlan> elements. This vlan
element is different from the element that defines a linux vlan
interface. This element defines a virtual lan identified by it's
"tag" (a mandatory attribute) that contains ports that are defined in
the <slaves> element in the usual way. It is important to note that one
port can be in multiple vlans, this makes the port a VLAN trunk.

Finally the bridge can define zero or more bonds by child elements
<bond>. These are similar to normal linux bonds, however at the moment
they don't accept any options, only the specification of it's slaves.
These bond slaves MUST be a subset of the bridges slaves!

As usual the <define> tag can be used anywhere.

An example use in a recipe:
<ovs_bridge id="ovs_br">
    <slaves>
        <slave id="t1"/>
        <slave id="t2"/>
    </slaves>
    <vlan tag="1">
        <slaves>
            <slave id="t1"/>
        </slaves>
    </vlan>
    <bond id="mfg">
        <slaves>
            <slave id="t1"/>
            <slave id="t2"/>
        </slaves>
    </bond>
</ovs_bridge>

Signed-off-by: Ondrej Lichtner <olichtne(a)redhat.com&gt;
---
 schema-recipe.rng | 78 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 78 insertions(+)

diff --git a/schema-recipe.rng b/schema-recipe.rng
index 129d7e7..b623a99 100644
--- a/schema-recipe.rng
+++ b/schema-recipe.rng
@@ -134,6 +134,7 @@
                             <ref name="vlan"/>
                             <ref name="macvlan"/>
                             <ref name="team"/>
+                            <ref name="ovs_bridge"/>
                         </choice>
                     </zeroOrMore>
                 </element>
@@ -175,6 +176,83 @@
         </element>
     </define>
 
+    <define name="ovs_bridge">
+        <element name="ovs_bridge">
+            <attribute name="id"/>
+
+            <interleave>
+                <optional>
+                    <ref name="define"/>
+                </optional>
+
+                <element name="slaves">
+                    <interleave>
+                        <optional>
+                            <ref name="define"/>
+                        </optional>
+
+                        <oneOrMore>
+                            <element name="slave">
+                                <attribute name="id"/>
+                            </element>
+                        </oneOrMore>
+                    </interleave>
+                </element>
+
+                <zeroOrMore>
+                    <element name="bond">
+                        <attribute name="id"/>
+
+                        <interleave>
+                            <optional>
+                                <ref name="define"/>
+                            </optional>
+
+                            <element name="slaves">
+                                <interleave>
+                                    <optional>
+                                        <ref name="define"/>
+                                    </optional>
+
+                                    <oneOrMore>
+                                        <element name="slave">
+                                            <attribute name="id"/>
+                                        </element>
+                                    </oneOrMore>
+                                </interleave>
+                            </element>
+                        </interleave>
+                    </element>
+                </zeroOrMore>
+                <zeroOrMore>
+                    <element name="vlan">
+                        <attribute name="tag"/>
+
+                        <interleave>
+                            <optional>
+                                <ref name="define"/>
+                            </optional>
+
+                            <element name="slaves">
+                                <interleave>
+                                    <optional>
+                                        <ref name="define"/>
+                                    </optional>
+
+                                    <oneOrMore>
+                                        <element name="slave">
+                                            <attribute name="id"/>
+                                        </element>
+                                    </oneOrMore>
+                                </interleave>
+                            </element>
+                        </interleave>
+                    </element>
+                </zeroOrMore>
+            </interleave>
+        </element>
+    </define>
+
     <define name="vlan">
         <element name="vlan">
             <attribute name="id"/>
-- 
1.8.5.3

[PATCH v2 2/5] RecipeParser: add support for ovs_bridge element

[Ondrej Lichtner]

From: Ondrej Lichtner <olichtne(a)redhat.com&gt;

The slaves of the ovs_bridge are translated to the internal
representation in the same way as slaves of other devices. However
information about vlan tags and bonds is stored in a new "ovs_conf"
field. I added this field to make it relatively simple to add new
features to the recipe schema of ovs_bridges.

In addition to this, the new code checks for conflicts such as
definition of the same bond, enslaving the same port in different bonds
or referencing non-existing ports.

Signed-off-by: Ondrej Lichtner <olichtne(a)redhat.com&gt;
---
 lnst/Controller/RecipeParser.py | 79 ++++++++++++++++++++++++++++++++++++++---
 1 file changed, 74 insertions(+), 5 deletions(-)

diff --git a/lnst/Controller/RecipeParser.py b/lnst/Controller/RecipeParser.py
index e3b6df9..c4fe61f 100644
--- a/lnst/Controller/RecipeParser.py
+++ b/lnst/Controller/RecipeParser.py
@@ -86,9 +86,6 @@ class RecipeParser(XmlParser):
         iface["id"] = self._get_attribute(iface_tag, "id")
         iface["type"] = iface_tag.tag
 
-        if iface["type"] == "eth":
-            iface["network"] = self._get_attribute(iface_tag,
"label")
-
         # params
         params_tag = iface_tag.find("params")
         params = self._process_params(params_tag)
@@ -106,8 +103,9 @@ class RecipeParser(XmlParser):
                     addr = self._get_content(addr_tag)
                 iface["addresses"].append(addr)
 
-
-        if iface["type"] in ["bond", "bridge",
"vlan", "macvlan", "team"]:
+        if iface["type"] == "eth":
+            iface["network"] = self._get_attribute(iface_tag,
"label")
+        elif iface["type"] in ["bond", "bridge",
"vlan", "macvlan", "team"]:
             # slaves
             slaves_tag = iface_tag.find("slaves")
             if slaves_tag is not None and len(slaves_tag) > 0:
@@ -129,6 +127,77 @@ class RecipeParser(XmlParser):
             opts = self._proces_options(opts_tag)
             if len(opts) > 0:
                 iface["options"] = opts
+        elif iface["type"] == "ovs_bridge":
+            slaves_tag = iface_tag.find("slaves")
+            iface["slaves"] = XmlCollection(slaves_tag)
+            ovsb_slaves = []
+
+            iface["ovs_conf"] = XmlData(slaves_tag)
+            for slave_tag in slaves_tag:
+                slave = XmlData(slave_tag)
+                slave["id"] = str(self._get_attribute(slave_tag,
"id"))
+                ovsb_slaves.append(slave["id"])
+
+                iface["slaves"].append(slave)
+
+            vlan_elems = iface_tag.findall("vlan")
+            if len(vlan_elems) > 0:
+                vlans = iface["ovs_conf"]["vlans"] =
XmlData(slaves_tag)
+            for vlan in vlan_elems:
+                vlan_tag = str(self._get_attribute(vlan, "tag"))
+                if vlan_tag in vlans:
+                    msg = "VLAN '%s' already defined for "\
+                          "this ovs_bridge." % vlan_tag
+                    raise RecipeError(msg, vlan)
+
+                vlans[vlan_tag] = XmlData(vlan)
+                vlans[vlan_tag]["slaves"] = XmlCollection(vlan)
+                vlan_slaves = vlans[vlan_tag]["slaves"]
+
+                slaves_tag = vlan.find("slaves")
+                for slave_tag in slaves_tag:
+                    slave_id = str(self._get_attribute(slave_tag, "id"))
+                    if slave_id not in ovsb_slaves:
+                        msg = "No port with id '%s' defined for "\
+                              "this ovs_bridge." % slave_id
+                        raise RecipeError(msg, slave_tag)
+
+                    if slave_id in vlan_slaves:
+                        msg = "Port '%s' already a member of vlan %s"\
+                              % (slave_id, vlan_tag)
+                        raise RecipeError(msg, slave_tag)
+                    else:
+                        vlan_slaves.append(slave_id)
+
+            bonded_slaves = {}
+            bond_elems = iface_tag.findall("bond")
+            if len(bond_elems) > 0:
+                bonds = iface["ovs_conf"]["bonds"] =
XmlData(slaves_tag)
+            for bond_tag in bond_elems:
+                bond_id = str(self._get_attribute(bond_tag, "id"))
+                if bond_id in bonds:
+                    msg = "Bond with id '%s' already defined for "\
+                          "this ovs_bridge." % bond_id
+                    raise RecipeError(msg, bond_tag)
+                bonds[bond_id] = XmlData(bond_tag)
+                bond_slaves = bonds[bond_id]["slaves"] =
XmlCollection(bond_tag)
+
+                slaves_tag = bond_tag.find("slaves")
+                for slave_tag in slaves_tag:
+                    slave_id = str(self._get_attribute(slave_tag, "id"))
+                    if slave_id not in ovsb_slaves:
+                        msg = "No port with id '%s' defined for "\
+                              "this ovs_bridge." % slave_id
+                        raise RecipeError(msg, slave_tag)
+
+                    if slave_id in bonded_slaves:
+                        msg = "Port with id '%s' already in bond with id
'%s'"\
+                              % (slave_id, bonded_slaves[slave_id])
+                        raise RecipeError(msg, slave_tag)
+                    else:
+                        bonded_slaves[slave_id] = bond_id
+
+                    bond_slaves.append(slave_id)
 
         return iface
 
-- 
1.8.5.3

[PATCH v2 3/5] Controller: add support for ovs_conf

[Ondrej Lichtner]

From: Ondrej Lichtner <olichtne(a)redhat.com&gt;

This commit adds Controller support for the ovs_conf field introduced in
the previous commit. The data is now propagated into the Interface
object representing the ovs_bridge and will be sent to the slave when
the configure() method is called.

Signed-off-by: Ondrej Lichtner <olichtne(a)redhat.com&gt;
---
 lnst/Controller/Machine.py           | 10 +++++++++-
 lnst/Controller/NetTestController.py |  3 +++
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/lnst/Controller/Machine.py b/lnst/Controller/Machine.py
index a849e3b..e2f1d3e 100644
--- a/lnst/Controller/Machine.py
+++ b/lnst/Controller/Machine.py
@@ -393,6 +393,8 @@ class Interface(object):
 
         self._master = None
 
+        self._ovs_conf = None
+
     def get_id(self):
         return self._id
 
@@ -454,6 +456,12 @@ class Interface(object):
     def get_address(self, num):
         return self._addresses[num].split('/')[0]
 
+    def set_ovs_conf(self, ovs_conf):
+        self._ovs_conf = ovs_conf
+
+    def get_ovs_conf(self):
+        return self._ovs_conf
+
     def get_prefix(self, num):
         try:
             return self._addresses[num].split('/')[1]
@@ -465,7 +473,7 @@ class Interface(object):
                   "addresses": self._addresses, "slaves":
self._slaves.keys(),
                   "options": self._options,
                   "slave_options": self._slave_options,
-                  "master": None}
+                  "master": None, "ovs_conf": self._ovs_conf}
 
         if self._master != None:
             config["master"] = self._master.get_id()
diff --git a/lnst/Controller/NetTestController.py b/lnst/Controller/NetTestController.py
index 40f35b3..2453042 100644
--- a/lnst/Controller/NetTestController.py
+++ b/lnst/Controller/NetTestController.py
@@ -284,6 +284,9 @@ class NetTestController:
             for opt in iface_xml_data["options"]:
                 iface.set_option(opt["name"], opt["value"])
 
+        if "ovs_conf" in iface_xml_data:
+            iface.set_ovs_conf(iface_xml_data["ovs_conf"])
+
     def _prepare_tasks(self):
         self._tasks = []
         for task_data in self._recipe["tasks"]:
-- 
1.8.5.3

[PATCH v2 4/5] {Nm, Net}ConfigDevice: add support for OvS bridges

[Ondrej Lichtner]

From: Ondrej Lichtner <olichtne(a)redhat.com&gt;

This commit adds the classes NmConfigDeviceOvsBridge and
NetConfigDeviceOvsBridge that represent the configuration of an Open
vSwitch bridge. The Nm class only defines the is_nm_managed method that
is set to return False and check the devices slaves, since NM does not
support OvS configuration.

The Net class takes care of the OvS configuration by calling appropriate
shell commands, at the moment it can handle the creation and destruction
of an OvS bridge, add/remove ports and set their vlan tags and finally
add/remove bonds. This is sufficient for basic support of Open vSwitch.

Signed-off-by: Ondrej Lichtner <olichtne(a)redhat.com&gt;
---
 lnst/Slave/NetConfigDevice.py | 86 ++++++++++++++++++++++++++++++++++++++++++-
 lnst/Slave/NmConfigDevice.py  | 19 +++++++++-
 2 files changed, 103 insertions(+), 2 deletions(-)

diff --git a/lnst/Slave/NetConfigDevice.py b/lnst/Slave/NetConfigDevice.py
index e279d66..39267a6 100644
--- a/lnst/Slave/NetConfigDevice.py
+++ b/lnst/Slave/NetConfigDevice.py
@@ -294,13 +294,97 @@ class NetConfigDeviceTeam(NetConfigDeviceGeneric):
         dbus_option = "-D" if self._should_enable_dbus() else ""
         exec_cmd("teamdctl %s %s port remove %s" % (dbus_option, dev_name,
port_name))
 
+class NetConfigDeviceOvsBridge(NetConfigDeviceGeneric):
+    _modulename = "openvswitch"
+    _moduleload = True
+
+    @classmethod
+    def type_init(self):
+        super(NetConfigDeviceOvsBridge, self).type_init()
+        exec_cmd("mkdir -p /var/run/openvswitch/")
+        exec_cmd("ovsdb-server --detach --pidfile "\
+                              "--remote=punix:/var/run/openvswitch/db.sock",
+                              die_on_err=False)
+        exec_cmd("ovs-vswitchd --detach --pidfile", die_on_err=False)
+
+    def _add_ports(self):
+        slaves = self._dev_config["slaves"]
+        vlans = self._dev_config["ovs_conf"]["vlans"]
+
+        br_name = self._dev_config["name"]
+
+        for slave_id in slaves:
+            slave_dev = self._if_manager.get_mapped_device(slave_id)
+            slave_name = slave_dev.get_name()
+
+            vlan_tags = []
+            for tag, vlan in vlans.iteritems():
+                if slave_id in vlan["slaves"]:
+                    vlan_tags.append(tag)
+            if len(vlan_tags) == 0:
+                tags = ""
+            elif len(vlan_tags) == 1:
+                tags = " tag=%s" % vlan_tags[0]
+            elif len(vlan_tags) > 1:
+                tags = " trunks=" + ",".join(vlan_tags)
+            exec_cmd("ovs-vsctl add-port %s %s%s" % (br_name, slave_name,
tags))
+
+    def _del_ports(self):
+        slaves = self._dev_config["slaves"]
+
+        br_name = self._dev_config["name"]
+
+        for slave_id in slaves:
+            slave_dev = self._if_manager.get_mapped_device(slave_id)
+            slave_name = slave_dev.get_name()
+
+            exec_cmd("ovs-vsctl del-port %s %s" % (br_name, slave_name))
+
+    def _add_bonds(self):
+        br_name = self._dev_config["name"]
+
+        bonds = self._dev_config["ovs_conf"]["bonds"]
+        for bond_id, bond in bonds.iteritems():
+            ifaces = ""
+            for slave in bond["slaves"]:
+                ifaces += " %s" % slave
+            exec_cmd("ovs-vsctl add-bond %s %s %s" % (br_name, bond_id,
ifaces))
+
+    def _del_bonds(self):
+        br_name = self._dev_config["name"]
+
+        bonds = self._dev_config["ovs_conf"]["bonds"]
+        for bond_id, bond in bonds.iteritems():
+            exec_cmd("ovs-vsctl del-port %s %s" % (br_name, bond_id))
+
+    def configure(self):
+        dev_cfg = self._dev_config
+
+        br_name = dev_cfg["name"]
+        exec_cmd("ovs-vsctl add-br %s" % br_name)
+
+        self._add_ports()
+
+        self._add_bonds()
+
+    def deconfigure(self):
+        dev_cfg = self._dev_config
+
+        self._del_bonds()
+
+        self._del_ports()
+
+        br_name = dev_cfg["name"]
+        exec_cmd("ovs-vsctl del-br %s" % br_name)
+
 type_class_mapping = {
     "eth": NetConfigDeviceEth,
     "bond": NetConfigDeviceBond,
     "bridge": NetConfigDeviceBridge,
     "macvlan": NetConfigDeviceMacvlan,
     "vlan": NetConfigDeviceVlan,
-    "team": NetConfigDeviceTeam
+    "team": NetConfigDeviceTeam,
+    "ovs_bridge": NetConfigDeviceOvsBridge
 }
 
 def NetConfigDevice(dev_config, if_manager):
diff --git a/lnst/Slave/NmConfigDevice.py b/lnst/Slave/NmConfigDevice.py
index 40f4b70..460fd26 100644
--- a/lnst/Slave/NmConfigDevice.py
+++ b/lnst/Slave/NmConfigDevice.py
@@ -781,13 +781,30 @@ class NmConfigDeviceTeam(NmConfigDeviceGeneric):
         self._rm_slaves()
         self._rm_team()
 
+class NmConfigDeviceOvsBridge(NmConfigDeviceGeneric):
+    #Not supported by NetworkManager
+    @classmethod
+    def is_nm_managed(cls, dev_config, if_manager):
+        managed = False
+
+        for slave_id in get_slaves(dev_config):
+            slave_dev = if_manager.get_mapped_device(slave_id)
+            slave_config = slave_dev.get_conf_dict()
+            if is_nm_managed(slave_config, if_manager) != managed:
+                msg = "Mixing NM managed and not managed devices in a "\
+                        "master-slave relationship is not allowed!"
+                raise Exception(msg)
+
+        return managed
+
 type_class_mapping = {
     "eth": NmConfigDeviceEth,
     "bond": NmConfigDeviceBond,
     "bridge": NmConfigDeviceBridge,
     "macvlan": NmConfigDeviceMacvlan,
     "vlan": NmConfigDeviceVlan,
-    "team": NmConfigDeviceTeam
+    "team": NmConfigDeviceTeam,
+    "ovs_bridge": NmConfigDeviceOvsBridge
 }
 
 def is_nm_managed(dev_config, if_manager):
-- 
1.8.5.3

[PATCH v2 5/5] InterfaceManager: generate ovs_bridge interface names

[Ondrej Lichtner]

From: Ondrej Lichtner <olichtne(a)redhat.com&gt;

Add the "ovs_bridge" interface type to the assign_name method. The ovs
bridges will get names generated as if they were normal bridges.

Signed-off-by: Ondrej Lichtner <olichtne(a)redhat.com&gt;
---
 lnst/Slave/InterfaceManager.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lnst/Slave/InterfaceManager.py b/lnst/Slave/InterfaceManager.py
index a69e4d4..107601c 100644
--- a/lnst/Slave/InterfaceManager.py
+++ b/lnst/Slave/InterfaceManager.py
@@ -162,7 +162,7 @@ class InterfaceManager(object):
                     return dev.get_name()
         elif dev_type == "bond":
             return self._assign_name_generic("t_bond")
-        elif dev_type == "bridge":
+        elif dev_type == "bridge" or dev_type == "ovs_bridge":
             return self._assign_name_generic("t_br")
         elif dev_type == "macvlan":
             return self._assign_name_generic("t_macvlan")
-- 
1.8.5.3

[Jiri Pirko]

Mon, Mar 31, 2014 at 01:25:38PM CEST, olichtne(a)redhat.com wrote:
> From: Ondrej Lichtner <olichtne(a)redhat.com&gt;
>
> Second version of the patchset that brings OvS support to LNST. The difference
> is the way we define vlans - instead of using the vlan_tag/vlan_tags attributes
> use the <vlan> element like this:
> <vlan tag="1">
>     <slaves>
>         <slave id="port1"/>
>         <slave id="port2"/>
>     </slaves>
> </vlan>
>
> If a single port is present in more than one vlan it becomes a trunk port.
>
> The previous example recipe therefore becomes this:
> <lnstrecipe>
>     <network>
>         <host id="1">
>             <params>
>                 <param name="os" value="fedora20"/>
>             </params>
>             <interfaces>
>                 <eth id="t1" label="ttnet1"/>
>                 <eth id="t2" label="ttnet2"/>
>                 <ovs_bridge id="xyz">
>                     <define>
>                         <alias name="aa" value="bb"/>
>                     </define>
>                     <slaves>
>                         <slave id="t1"/>
>                         <slave id="t2"/>
>                     </slaves>
>                     <vlan tag="1">
>                         <slaves>
>                             <slave id="t1"/>
>                             <slave id="t2"/>
>                         </slaves>
>                     </vlan>
>                     <bond id="mfg">
>                         <slaves>
>                             <slave id="t1"/>
>                             <slave id="t2"/>
>                         </slaves>
>                     </bond>
>                 </ovs_bridge>
>             </interfaces>
>         </host>
>         <host id="2">
>             <interfaces>
>                 <eth id="t1" label="ttnet1">
>                     <addresses>
>                         <address value="192.168.100.215/24"/>
>                     </addresses>
>                 </eth>
>             </interfaces>
>         </host>
>         <host id="3">
>             <interfaces>
>                 <eth id="t2" label="ttnet2">
>                     <addresses>
>                         <address value="192.168.100.216/24"/>
>                     </addresses>
>                 </eth>
>             </interfaces>
>         </host>
>     </network>
>
>     <task>
>         <run host="2" module="IcmpPing" timeout="30">
>             <options>
>                 <option name="addr" value="{ip(3,t2)}"/>
>                 <option name="count" value="40"/>
>                 <option name="interval" value="0"/>
>                 <option name="limit_rate" value="95"/>
>             </options>
>         </run>
>     </task>
> </lnstrecipe>
>
> Ondrej Lichtner (5):
>   schema-recipe.rng: add element ovs_bridge
>   RecipeParser: add support for ovs_bridge element
>   Controller: add support for ovs_conf
>   {Nm, Net}ConfigDevice: add support for OvS bridges
>   InterfaceManager: generate ovs_bridge interface names
>
>  lnst/Controller/Machine.py           | 10 ++++-
>  lnst/Controller/NetTestController.py |  3 ++
>  lnst/Controller/RecipeParser.py      | 79 ++++++++++++++++++++++++++++++---
>  lnst/Slave/InterfaceManager.py       |  2 +-
>  lnst/Slave/NetConfigDevice.py        | 86 +++++++++++++++++++++++++++++++++++-
>  lnst/Slave/NmConfigDevice.py         | 19 +++++++-
>  schema-recipe.rng                    | 78 ++++++++++++++++++++++++++++++++
>  7 files changed, 268 insertions(+), 9 deletions(-)
>
> -- 
> 1.8.5.3
>
> _______________________________________________
> LNST-developers mailing list
> LNST-developers(a)lists.fedorahosted.org
> https://lists.fedorahosted.org/mailman/listinfo/lnst-developers 

applied. Thanks!