[ovs-dev] [PATCH ovs v3 2/2] netdev-dpdk: Add dpdkvdpa port
Noa Ezra
noae at mellanox.com
Thu Oct 17 11:16:56 UTC 2019
dpdkvdpa netdev works with 3 components:
vhost-user socket, vdpa device: real vdpa device or a VF and
representor of "vdpa device".
In order to add a new vDPA port, add a new port to existing bridge
with type dpdkvdpa and vDPA options:
ovs-vsctl add-port br0 vdpa0 -- set Interface vdpa0 type=dpdkvdpa
options:vdpa-socket-path=<sock path>
options:vdpa-accelerator-devargs=<VF pci id>
options:dpdk-devargs=<vdpa pci id>,representor=[id]
On this command OVS will create a new netdev:
1. Register vhost-user-client device.
2. Open and configure VF dpdk port.
3. Open and configure representor dpdk port.
The new netdev will use netdev_rxq_recv() function in order to receive
packets from VF and push to vhost-user and receive packets from
vhost-user and push to VF.
Signed-off-by: Noa Ezra <noae at mellanox.com>
Reviewed-by: Oz Shlomo <ozsh at mellanox.com>
---
Documentation/automake.mk | 1 +
Documentation/topics/dpdk/index.rst | 1 +
Documentation/topics/dpdk/vdpa.rst | 90 ++++++++++++++++++++
NEWS | 1 +
lib/netdev-dpdk.c | 162 ++++++++++++++++++++++++++++++++++++
vswitchd/vswitch.xml | 25 ++++++
6 files changed, 280 insertions(+)
create mode 100644 Documentation/topics/dpdk/vdpa.rst
diff --git a/Documentation/automake.mk b/Documentation/automake.mk
index cd68f3b..ee574bc 100644
--- a/Documentation/automake.mk
+++ b/Documentation/automake.mk
@@ -43,6 +43,7 @@ DOC_SOURCE = \
Documentation/topics/dpdk/ring.rst \
Documentation/topics/dpdk/vdev.rst \
Documentation/topics/dpdk/vhost-user.rst \
+ Documentation/topics/dpdk/vdpa.rst \
Documentation/topics/fuzzing/index.rst \
Documentation/topics/fuzzing/what-is-fuzzing.rst \
Documentation/topics/fuzzing/ovs-fuzzing-infrastructure.rst \
diff --git a/Documentation/topics/dpdk/index.rst b/Documentation/topics/dpdk/index.rst
index cf24a7b..c1d4ea7 100644
--- a/Documentation/topics/dpdk/index.rst
+++ b/Documentation/topics/dpdk/index.rst
@@ -41,3 +41,4 @@ The DPDK Datapath
/topics/dpdk/pdump
/topics/dpdk/jumbo-frames
/topics/dpdk/memory
+ /topics/dpdk/vdpa
diff --git a/Documentation/topics/dpdk/vdpa.rst b/Documentation/topics/dpdk/vdpa.rst
new file mode 100644
index 0000000..34c5300
--- /dev/null
+++ b/Documentation/topics/dpdk/vdpa.rst
@@ -0,0 +1,90 @@
+..
+ Copyright (c) 2019 Mellanox Technologies, Ltd.
+
+ Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+ You may obtain a copy of the License at:
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+ Unless required by applicable law or agreed to in writing, software
+ distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+ WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+ License for the specific language governing permissions and limitations
+ under the License.
+
+ Convention for heading levels in Open vSwitch documentation:
+
+ ======= Heading 0 (reserved for the title in a document)
+ ------- Heading 1
+ ~~~~~~~ Heading 2
+ +++++++ Heading 3
+ ''''''' Heading 4
+
+ Avoid deeper levels because they do not render well.
+
+
+===============
+DPDK VDPA Ports
+===============
+
+In user space there are two main approaches to communicate with a guest (VM),
+using virtIO ports (e.g. netdev type=dpdkvhoshuser/dpdkvhostuserclient) or
+SR-IOV using phy ports (e.g. netdev type = dpdk).
+Phy ports allow working with port representor which is attached to the OVS and
+a matching VF is given with pass-through to the guest.
+HW rules can process packets from up-link and direct them to the VF without
+going through SW (OVS) and therefore using phy ports gives the best
+performance.
+However, SR-IOV architecture requires that the guest will use a driver which is
+specific to the underlying HW. Specific HW driver has two main drawbacks:
+1. Breaks virtualization in some sense (guest aware of the HW), can also limit
+the type of images supported.
+2. Less natural support for live migration.
+
+Using virtIO port solves both problems, but reduces performance and causes
+losing of some functionality, for example, for some HW offload, working
+directly with virtIO cannot be supported.
+
+We created a new netdev type- dpdkvdpa. dpdkvdpa port solves this conflict.
+The new netdev is basically very similar to regular dpdk netdev but it has some
+additional functionally.
+This port translates between phy port to virtIO port, it takes packets from
+rx-queue and send them to the suitable tx-queue and allows to transfer packets
+from virtIO guest (VM) to a VF and vice versa and benefit both SR-IOV and
+virtIO.
+
+Quick Example
+-------------
+
+Configure OVS bridge and ports
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+you must first create a bridge and add ports to the switch.
+Since the dpdkvdpa port is configured as a client, the vdpa-socket-path must be
+configured by the user.
+VHOST_USER_SOCKET_PATH=/path/to/socket
+
+ $ ovs-vsctl add-br br0-ovs -- set bridge br0-ovs datapath_type=netdev
+ $ ovs-vsctl add-port br0-ovs pf -- set Interface pf \
+ type=dpdk options:dpdk-devargs=<pf pci id>
+ $ ovs-vsctl add-port br0 vdpa0 -- set Interface vdpa0 type=dpdkvdpa \
+ options:vdpa-socket-path=VHOST_USER_SOCKET_PATH \
+ options:vdpa-accelerator-devargs=<vf pci id> \
+ options:dpdk-devargs=<pf pci id>,representor=[id]
+
+Once the ports have been added to the switch, they must be added to the guest.
+
+Adding vhost-user ports to the guest (QEMU)
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Attach the vhost-user device sockets to the guest. To do this, you must pass
+the following parameters to QEMU:
+
+ -chardev socket,id=char1,path=$VHOST_USER_SOCKET_PATH,server
+ -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce
+ -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1
+
+QEMU will wait until the port is created successfully in OVS to boot the VM.
+In this mode, in case the switch will crash, the vHost ports will reconnect
+automatically once it is brought back.
diff --git a/NEWS b/NEWS
index f5a0b8f..6f315c6 100644
--- a/NEWS
+++ b/NEWS
@@ -542,6 +542,7 @@ v2.6.0 - 27 Sep 2016
* Remove dpdkvhostcuse port type.
* OVS client mode for vHost and vHost reconnect (Requires QEMU 2.7)
* 'dpdkvhostuserclient' port type.
+ * 'dpdkvdpa' port type.
- Increase number of registers to 16.
- ovs-benchmark: This utility has been removed due to lack of use and
bitrot.
diff --git a/lib/netdev-dpdk.c b/lib/netdev-dpdk.c
index bc20d68..16ddf58 100644
--- a/lib/netdev-dpdk.c
+++ b/lib/netdev-dpdk.c
@@ -47,6 +47,7 @@
#include "dpif-netdev.h"
#include "fatal-signal.h"
#include "netdev-provider.h"
+#include "netdev-dpdk-vdpa.h"
#include "netdev-vport.h"
#include "odp-util.h"
#include "openvswitch/dynamic-string.h"
@@ -137,6 +138,9 @@ typedef uint16_t dpdk_port_t;
/* Legacy default value for vhost tx retries. */
#define VHOST_ENQ_RETRY_DEF 8
+/* Size of VDPA custom stats. */
+#define VDPA_CUSTOM_STATS_SIZE 4
+
#define IF_NAME_SZ (PATH_MAX > IFNAMSIZ ? PATH_MAX : IFNAMSIZ)
static const struct rte_eth_conf port_conf = {
@@ -461,6 +465,8 @@ struct netdev_dpdk {
int rte_xstats_ids_size;
uint64_t *rte_xstats_ids;
);
+
+ struct netdev_dpdk_vdpa_relay *relay;
};
struct netdev_rxq_dpdk {
@@ -1346,6 +1352,30 @@ netdev_dpdk_construct(struct netdev *netdev)
return err;
}
+static int
+netdev_dpdk_vdpa_construct(struct netdev *netdev)
+{
+ struct netdev_dpdk *dev;
+ int err;
+
+ err = netdev_dpdk_construct(netdev);
+ if (err) {
+ VLOG_ERR("netdev_dpdk_construct failed. Port: %s\n", netdev->name);
+ goto out;
+ }
+
+ ovs_mutex_lock(&dpdk_mutex);
+ dev = netdev_dpdk_cast(netdev);
+ dev->relay = netdev_dpdk_vdpa_alloc_relay();
+ if (!dev->relay) {
+ err = ENOMEM;
+ }
+
+ ovs_mutex_unlock(&dpdk_mutex);
+out:
+ return err;
+}
+
static void
common_destruct(struct netdev_dpdk *dev)
OVS_REQUIRES(dpdk_mutex)
@@ -1428,6 +1458,19 @@ dpdk_vhost_driver_unregister(struct netdev_dpdk *dev OVS_UNUSED,
}
static void
+netdev_dpdk_vdpa_destruct(struct netdev *netdev)
+{
+ struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
+
+ ovs_mutex_lock(&dpdk_mutex);
+ netdev_dpdk_vdpa_destruct_impl(dev->relay);
+ rte_free(dev->relay);
+ ovs_mutex_unlock(&dpdk_mutex);
+
+ netdev_dpdk_destruct(netdev);
+}
+
+static void
netdev_dpdk_vhost_destruct(struct netdev *netdev)
{
struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
@@ -1878,6 +1921,47 @@ out:
}
static int
+netdev_dpdk_vdpa_set_config(struct netdev *netdev, const struct smap *args,
+ char **errp)
+{
+ struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
+ const char *vdpa_accelerator_devargs =
+ smap_get(args, "vdpa-accelerator-devargs");
+ const char *vdpa_socket_path =
+ smap_get(args, "vdpa-socket-path");
+ int err = 0;
+
+ if ((vdpa_accelerator_devargs == NULL) || (vdpa_socket_path == NULL)) {
+ VLOG_ERR("netdev_dpdk_vdpa_set_config failed."
+ "Required arguments are missing for VDPA port %s",
+ netdev->name);
+ goto free_relay;
+ }
+
+ err = netdev_dpdk_set_config(netdev, args, errp);
+ if (err) {
+ VLOG_ERR("netdev_dpdk_set_config failed. Port: %s", netdev->name);
+ goto free_relay;
+ }
+
+ err = netdev_dpdk_vdpa_config_impl(dev->relay, dev->port_id,
+ vdpa_socket_path,
+ vdpa_accelerator_devargs);
+ if (err) {
+ VLOG_ERR("netdev_dpdk_vdpa_config_impl failed. Port %s",
+ netdev->name);
+ goto free_relay;
+ }
+
+ goto out;
+
+free_relay:
+ rte_free(dev->relay);
+out:
+ return err;
+}
+
+static int
netdev_dpdk_ring_set_config(struct netdev *netdev, const struct smap *args,
char **errp OVS_UNUSED)
{
@@ -2273,6 +2357,23 @@ netdev_dpdk_rxq_recv(struct netdev_rxq *rxq, struct dp_packet_batch *batch,
return 0;
}
+static int
+netdev_dpdk_vdpa_rxq_recv(struct netdev_rxq *rxq,
+ struct dp_packet_batch *batch,
+ int *qfill)
+{
+ struct netdev_dpdk *dev = netdev_dpdk_cast(rxq->netdev);
+ int fwd_rx;
+ int ret;
+
+ fwd_rx = netdev_dpdk_vdpa_rxq_recv_impl(dev->relay, rxq->queue_id);
+ ret = netdev_dpdk_rxq_recv(rxq, batch, qfill);
+ if ((ret == EAGAIN) && fwd_rx) {
+ return 0;
+ }
+ return ret;
+}
+
static inline int
netdev_dpdk_qos_run(struct netdev_dpdk *dev, struct rte_mbuf **pkts,
int cnt, bool should_steal)
@@ -2854,6 +2955,29 @@ netdev_dpdk_vhost_get_custom_stats(const struct netdev *netdev,
}
static int
+netdev_dpdk_vdpa_get_custom_stats(const struct netdev *netdev,
+ struct netdev_custom_stats *custom_stats)
+{
+ struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
+ int err = 0;
+
+ ovs_mutex_lock(&dev->mutex);
+
+ custom_stats->size = VDPA_CUSTOM_STATS_SIZE;
+ custom_stats->counters = xcalloc(custom_stats->size,
+ sizeof *custom_stats->counters);
+ err = netdev_dpdk_vdpa_get_custom_stats_impl(dev->relay,
+ custom_stats);
+ if (err) {
+ VLOG_ERR("netdev_dpdk_vdpa_get_custom_stats_impl failed."
+ "Port %s\n", netdev->name);
+ }
+
+ ovs_mutex_unlock(&dev->mutex);
+ return err;
+}
+
+static int
netdev_dpdk_get_features(const struct netdev *netdev,
enum netdev_features *current,
enum netdev_features *advertised,
@@ -4237,6 +4361,31 @@ netdev_dpdk_vhost_reconfigure(struct netdev *netdev)
}
static int
+netdev_dpdk_vdpa_reconfigure(struct netdev *netdev)
+{
+ struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
+ int err;
+
+ err = netdev_dpdk_reconfigure(netdev);
+ if (err) {
+ VLOG_ERR("netdev_dpdk_reconfigure failed. Port %s", netdev->name);
+ goto out;
+ }
+
+ ovs_mutex_lock(&dev->mutex);
+ err = netdev_dpdk_vdpa_update_relay(dev->relay, dev->dpdk_mp->mp,
+ dev->up.n_rxq);
+ if (err) {
+ VLOG_ERR("netdev_dpdk_vdpa_update_relay failed. Port %s",
+ netdev->name);
+ }
+
+ ovs_mutex_unlock(&dev->mutex);
+out:
+ return err;
+}
+
+static int
netdev_dpdk_vhost_client_reconfigure(struct netdev *netdev)
{
struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
@@ -4456,6 +4605,18 @@ static const struct netdev_class dpdk_vhost_client_class = {
.rxq_enabled = netdev_dpdk_vhost_rxq_enabled,
};
+static const struct netdev_class dpdk_vdpa_class = {
+ .type = "dpdkvdpa",
+ NETDEV_DPDK_CLASS_COMMON,
+ .construct = netdev_dpdk_vdpa_construct,
+ .destruct = netdev_dpdk_vdpa_destruct,
+ .rxq_recv = netdev_dpdk_vdpa_rxq_recv,
+ .set_config = netdev_dpdk_vdpa_set_config,
+ .reconfigure = netdev_dpdk_vdpa_reconfigure,
+ .get_custom_stats = netdev_dpdk_vdpa_get_custom_stats,
+ .send = netdev_dpdk_eth_send
+};
+
void
netdev_dpdk_register(void)
{
@@ -4463,4 +4624,5 @@ netdev_dpdk_register(void)
netdev_register_provider(&dpdk_ring_class);
netdev_register_provider(&dpdk_vhost_class);
netdev_register_provider(&dpdk_vhost_client_class);
+ netdev_register_provider(&dpdk_vdpa_class);
}
diff --git a/vswitchd/vswitch.xml b/vswitchd/vswitch.xml
index 9a743c0..9e94950 100644
--- a/vswitchd/vswitch.xml
+++ b/vswitchd/vswitch.xml
@@ -2640,6 +2640,13 @@
<dd>
A pair of virtual devices that act as a patch cable.
</dd>
+
+ <dt><code>dpdkvdpa</code></dt>
+ <dd>
+ The dpdk vDPA port allows forwarding bi-directional traffic between
+ SR-IOV virtual functions (VFs) and VirtIO devices in virtual
+ machines (VMs).
+ </dd>
</dl>
</column>
</group>
@@ -3156,6 +3163,24 @@ ovs-vsctl add-port br0 p0 -- set Interface p0 type=patch options:peer=p1 \
</p>
</column>
+ <column name="options" key="vdpa-socket-path"
+ type='{"type": "string"}'>
+ <p>
+ The value specifies the path to the socket associated with a VDPA
+ port that will be created by QEMU.
+ Only supported by dpdkvdpa interfaces.
+ </p>
+ </column>
+
+ <column name="options" key="vdpa-accelerator-devargs"
+ type='{"type": "string"}'>
+ <p>
+ The value specifies the PCI address associated with the virtual
+ function.
+ Only supported by dpdkvdpa interfaces.
+ </p>
+ </column>
+
<column name="options" key="dq-zero-copy"
type='{"type": "boolean"}'>
<p>
--
1.8.3.1
More information about the dev
mailing list