net/virtio: add failover support

This patch adds support to handle failover device pairs of a virtio-net
device and a (vfio-)pci device, where the virtio-net acts as the standby
device and the (vfio-)pci device as the primary.

The general idea is that we have a pair of devices, a (vfio-)pci and a
emulated (virtio-net) device. Before migration the vfio device is
unplugged and data flows to the emulated device, on the target side
another (vfio-)pci device is plugged in to take over the data-path. In the
guest the net_failover module will pair net devices with the same MAC
address.

To achieve this we need:

1. Provide a callback function for the should_be_hidden DeviceListener.
   It is called when the primary device is plugged in. Evaluate the QOpt
   passed in to check if it is the matching primary device. It returns
   if the device should be hidden or not.
   When it should be hidden it stores the device options in the VirtioNet
   struct and the device is added once the VIRTIO_NET_F_STANDBY feature is
   negotiated during virtio feature negotiation.

   If the virtio-net devices are not realized at the time the (vfio-)pci
   devices are realized, we need to connect the devices later. This way
   we make sure primary and standby devices can be specified in any
   order.

2. Register a callback for migration status notifier. When called it
   will unplug its primary device before the migration happens.

3. Register a callback for the migration code that checks if a device
   needs to be unplugged from the guest.

Signed-off-by: Jens Freimann <jfreimann@redhat.com>
Message-Id: <20191029114905.6856-11-jfreimann@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This commit is contained in:
Jens Freimann 2019-10-29 12:49:04 +01:00 committed by Michael S. Tsirkin
parent ea45cb8d84
commit 9711cd0dfc
5 changed files with 384 additions and 0 deletions

View File

@ -1435,6 +1435,7 @@ S: Odd Fixes
F: hw/net/ F: hw/net/
F: include/hw/net/ F: include/hw/net/
F: tests/virtio-net-test.c F: tests/virtio-net-test.c
F: docs/virtio-net-failover.rst
T: git https://github.com/jasowang/qemu.git net T: git https://github.com/jasowang/qemu.git net
Parallel NOR Flash devices Parallel NOR Flash devices

View File

@ -0,0 +1,68 @@
========================
QEMU virtio-net standby (net_failover)
========================
This document explains the setup and usage of virtio-net standby feature which
is used to create a net_failover pair of devices.
The general idea is that we have a pair of devices, a (vfio-)pci and a
virtio-net device. Before migration the vfio device is unplugged and data flows
through the virtio-net device, on the target side another vfio-pci device is
plugged in to take over the data-path. In the guest the net_failover kernel
module will pair net devices with the same MAC address.
The two devices are called primary and standby device. The fast hardware based
networking device is called the primary device and the virtio-net device is the
standby device.
Restrictions
------------
Currently only PCIe devices are allowed as primary devices, this restriction
can be lifted in the future with enhanced QEMU support. Also, only networking
devices are allowed as primary device. The user needs to ensure that primary
and standby devices are not plugged into the same PCIe slot.
Usecase
-------
Virtio-net standby allows easy migration while using a passed-through fast
networking device by falling back to a virtio-net device for the duration of
the migration. It is like a simple version of a bond, the difference is that it
requires no configuration in the guest. When a guest is live-migrated to
another host QEMU will unplug the primary device via the PCIe based hotplug
handler and traffic will go through the virtio-net device. On the target
system the primary device will be automatically plugged back and the
net_failover module registers it again as the primary device.
Usage
-----
The primary device can be hotplugged or be part of the startup configuration
-device virtio-net-pci,netdev=hostnet1,id=net1,mac=52:54:00:6f:55:cc, \
bus=root2,failover=on
With the parameter failover=on the VIRTIO_NET_F_STANDBY feature will be enabled.
-device vfio-pci,host=5e:00.2,id=hostdev0,bus=root1,failover_pair_id=net1
failover_pair_id references the id of the virtio-net standby device. This
is only for pairing the devices within QEMU. The guest kernel module
net_failover will match devices with identical MAC addresses.
Hotplug
-------
Both primary and standby device can be hotplugged via the QEMU monitor. Note
that if the virtio-net device is plugged first a warning will be issued that it
couldn't find the primary device.
Migration
---------
A new migration state wait-unplug was added for this feature. If failover primary
devices are present in the configuration, migration will go into this state.
It will wait until the device unplug is completed in the guest and then move into
active state. On the target system the primary devices will be automatically hotplugged
when the feature bit was negotiated for the virtio-net standby device.

View File

@ -12,6 +12,7 @@
*/ */
#include "qemu/osdep.h" #include "qemu/osdep.h"
#include "qemu/atomic.h"
#include "qemu/iov.h" #include "qemu/iov.h"
#include "qemu/main-loop.h" #include "qemu/main-loop.h"
#include "qemu/module.h" #include "qemu/module.h"
@ -21,6 +22,10 @@
#include "net/tap.h" #include "net/tap.h"
#include "qemu/error-report.h" #include "qemu/error-report.h"
#include "qemu/timer.h" #include "qemu/timer.h"
#include "qemu/option.h"
#include "qemu/option_int.h"
#include "qemu/config-file.h"
#include "qapi/qmp/qdict.h"
#include "hw/virtio/virtio-net.h" #include "hw/virtio/virtio-net.h"
#include "net/vhost_net.h" #include "net/vhost_net.h"
#include "net/announce.h" #include "net/announce.h"
@ -28,11 +33,15 @@
#include "qapi/error.h" #include "qapi/error.h"
#include "qapi/qapi-events-net.h" #include "qapi/qapi-events-net.h"
#include "hw/qdev-properties.h" #include "hw/qdev-properties.h"
#include "qapi/qapi-types-migration.h"
#include "qapi/qapi-events-migration.h"
#include "hw/virtio/virtio-access.h" #include "hw/virtio/virtio-access.h"
#include "migration/misc.h" #include "migration/misc.h"
#include "standard-headers/linux/ethtool.h" #include "standard-headers/linux/ethtool.h"
#include "sysemu/sysemu.h" #include "sysemu/sysemu.h"
#include "trace.h" #include "trace.h"
#include "monitor/qdev.h"
#include "hw/pci/pci.h"
#define VIRTIO_NET_VM_VERSION 11 #define VIRTIO_NET_VM_VERSION 11
@ -746,9 +755,99 @@ static inline uint64_t virtio_net_supported_guest_offloads(VirtIONet *n)
return virtio_net_guest_offloads_by_features(vdev->guest_features); return virtio_net_guest_offloads_by_features(vdev->guest_features);
} }
static void failover_add_primary(VirtIONet *n, Error **errp)
{
Error *err = NULL;
n->primary_device_opts = qemu_opts_find(qemu_find_opts("device"),
n->primary_device_id);
if (n->primary_device_opts) {
n->primary_dev = qdev_device_add(n->primary_device_opts, &err);
if (err) {
qemu_opts_del(n->primary_device_opts);
}
if (n->primary_dev) {
n->primary_bus = n->primary_dev->parent_bus;
if (err) {
qdev_unplug(n->primary_dev, &err);
qdev_set_id(n->primary_dev, "");
}
}
} else {
error_setg(errp, "Primary device not found");
error_append_hint(errp, "Virtio-net failover will not work. Make "
"sure primary device has parameter"
" failover_pair_id=<virtio-net-id>\n");
}
if (err) {
error_propagate(errp, err);
}
}
static int is_my_primary(void *opaque, QemuOpts *opts, Error **errp)
{
VirtIONet *n = opaque;
int ret = 0;
const char *standby_id = qemu_opt_get(opts, "failover_pair_id");
if (standby_id != NULL && (g_strcmp0(standby_id, n->netclient_name) == 0)) {
n->primary_device_id = g_strdup(opts->id);
ret = 1;
}
return ret;
}
static DeviceState *virtio_net_find_primary(VirtIONet *n, Error **errp)
{
DeviceState *dev = NULL;
Error *err = NULL;
if (qemu_opts_foreach(qemu_find_opts("device"),
is_my_primary, n, &err)) {
if (err) {
error_propagate(errp, err);
return NULL;
}
if (n->primary_device_id) {
dev = qdev_find_recursive(sysbus_get_default(),
n->primary_device_id);
} else {
error_setg(errp, "Primary device id not found");
return NULL;
}
}
return dev;
}
static DeviceState *virtio_connect_failover_devices(VirtIONet *n,
DeviceState *dev,
Error **errp)
{
DeviceState *prim_dev = NULL;
Error *err = NULL;
prim_dev = virtio_net_find_primary(n, &err);
if (prim_dev) {
n->primary_device_id = g_strdup(prim_dev->id);
n->primary_device_opts = prim_dev->opts;
} else {
if (err) {
error_propagate(errp, err);
}
}
return prim_dev;
}
static void virtio_net_set_features(VirtIODevice *vdev, uint64_t features) static void virtio_net_set_features(VirtIODevice *vdev, uint64_t features)
{ {
VirtIONet *n = VIRTIO_NET(vdev); VirtIONet *n = VIRTIO_NET(vdev);
Error *err = NULL;
int i; int i;
if (n->mtu_bypass_backend && if (n->mtu_bypass_backend &&
@ -790,6 +889,28 @@ static void virtio_net_set_features(VirtIODevice *vdev, uint64_t features)
} else { } else {
memset(n->vlans, 0xff, MAX_VLAN >> 3); memset(n->vlans, 0xff, MAX_VLAN >> 3);
} }
if (virtio_has_feature(features, VIRTIO_NET_F_STANDBY)) {
qapi_event_send_failover_negotiated(n->netclient_name);
atomic_set(&n->primary_should_be_hidden, false);
failover_add_primary(n, &err);
if (err) {
n->primary_dev = virtio_connect_failover_devices(n, n->qdev, &err);
if (err) {
goto out_err;
}
failover_add_primary(n, &err);
if (err) {
goto out_err;
}
}
}
return;
out_err:
if (err) {
warn_report_err(err);
}
} }
static int virtio_net_handle_rx_mode(VirtIONet *n, uint8_t cmd, static int virtio_net_handle_rx_mode(VirtIONet *n, uint8_t cmd,
@ -2650,6 +2771,150 @@ void virtio_net_set_netclient_name(VirtIONet *n, const char *name,
n->netclient_type = g_strdup(type); n->netclient_type = g_strdup(type);
} }
static bool failover_unplug_primary(VirtIONet *n)
{
HotplugHandler *hotplug_ctrl;
PCIDevice *pci_dev;
Error *err = NULL;
hotplug_ctrl = qdev_get_hotplug_handler(n->primary_dev);
if (hotplug_ctrl) {
pci_dev = PCI_DEVICE(n->primary_dev);
pci_dev->partially_hotplugged = true;
hotplug_handler_unplug_request(hotplug_ctrl, n->primary_dev, &err);
if (err) {
error_report_err(err);
return false;
}
} else {
return false;
}
return true;
}
static bool failover_replug_primary(VirtIONet *n, Error **errp)
{
HotplugHandler *hotplug_ctrl;
PCIDevice *pdev = PCI_DEVICE(n->primary_dev);
if (!pdev->partially_hotplugged) {
return true;
}
if (!n->primary_device_opts) {
n->primary_device_opts = qemu_opts_from_qdict(
qemu_find_opts("device"),
n->primary_device_dict, errp);
}
if (n->primary_device_opts) {
if (n->primary_dev) {
n->primary_bus = n->primary_dev->parent_bus;
}
qdev_set_parent_bus(n->primary_dev, n->primary_bus);
n->primary_should_be_hidden = false;
qemu_opt_set_bool(n->primary_device_opts,
"partially_hotplugged", true, errp);
hotplug_ctrl = qdev_get_hotplug_handler(n->primary_dev);
if (hotplug_ctrl) {
hotplug_handler_pre_plug(hotplug_ctrl, n->primary_dev, errp);
hotplug_handler_plug(hotplug_ctrl, n->primary_dev, errp);
}
if (!n->primary_dev) {
error_setg(errp, "virtio_net: couldn't find primary device");
}
}
return *errp != NULL;
}
static void virtio_net_handle_migration_primary(VirtIONet *n,
MigrationState *s)
{
bool should_be_hidden;
Error *err = NULL;
should_be_hidden = atomic_read(&n->primary_should_be_hidden);
if (!n->primary_dev) {
n->primary_dev = virtio_connect_failover_devices(n, n->qdev, &err);
if (!n->primary_dev) {
return;
}
}
if (migration_in_setup(s) && !should_be_hidden &&
n->primary_dev) {
if (failover_unplug_primary(n)) {
vmstate_unregister(n->primary_dev, qdev_get_vmsd(n->primary_dev),
n->primary_dev);
qapi_event_send_unplug_primary(n->primary_device_id);
atomic_set(&n->primary_should_be_hidden, true);
} else {
warn_report("couldn't unplug primary device");
}
} else if (migration_has_failed(s)) {
/* We already unplugged the device let's plugged it back */
if (!failover_replug_primary(n, &err)) {
if (err) {
error_report_err(err);
}
}
}
}
static void virtio_net_migration_state_notifier(Notifier *notifier, void *data)
{
MigrationState *s = data;
VirtIONet *n = container_of(notifier, VirtIONet, migration_state);
virtio_net_handle_migration_primary(n, s);
}
static int virtio_net_primary_should_be_hidden(DeviceListener *listener,
QemuOpts *device_opts)
{
VirtIONet *n = container_of(listener, VirtIONet, primary_listener);
bool match_found;
bool hide;
n->primary_device_dict = qemu_opts_to_qdict(device_opts,
n->primary_device_dict);
if (n->primary_device_dict) {
g_free(n->standby_id);
n->standby_id = g_strdup(qdict_get_try_str(n->primary_device_dict,
"failover_pair_id"));
}
if (device_opts && g_strcmp0(n->standby_id, n->netclient_name) == 0) {
match_found = true;
} else {
match_found = false;
hide = false;
g_free(n->standby_id);
n->primary_device_dict = NULL;
goto out;
}
n->primary_device_opts = device_opts;
/* primary_should_be_hidden is set during feature negotiation */
hide = atomic_read(&n->primary_should_be_hidden);
if (n->primary_device_dict) {
g_free(n->primary_device_id);
n->primary_device_id = g_strdup(qdict_get_try_str(
n->primary_device_dict, "id"));
if (!n->primary_device_id) {
warn_report("primary_device_id not set");
}
}
out:
if (match_found && hide) {
return 1;
} else if (match_found && !hide) {
return 0;
} else {
return -1;
}
}
static void virtio_net_device_realize(DeviceState *dev, Error **errp) static void virtio_net_device_realize(DeviceState *dev, Error **errp)
{ {
VirtIODevice *vdev = VIRTIO_DEVICE(dev); VirtIODevice *vdev = VIRTIO_DEVICE(dev);
@ -2680,6 +2945,16 @@ static void virtio_net_device_realize(DeviceState *dev, Error **errp)
n->host_features |= (1ULL << VIRTIO_NET_F_SPEED_DUPLEX); n->host_features |= (1ULL << VIRTIO_NET_F_SPEED_DUPLEX);
} }
if (n->failover) {
n->primary_listener.should_be_hidden =
virtio_net_primary_should_be_hidden;
atomic_set(&n->primary_should_be_hidden, true);
device_listener_register(&n->primary_listener);
n->migration_state.notify = virtio_net_migration_state_notifier;
add_migration_state_change_notifier(&n->migration_state);
n->host_features |= (1ULL << VIRTIO_NET_F_STANDBY);
}
virtio_net_set_config_size(n, n->host_features); virtio_net_set_config_size(n, n->host_features);
virtio_init(vdev, "virtio-net", VIRTIO_ID_NET, n->config_size); virtio_init(vdev, "virtio-net", VIRTIO_ID_NET, n->config_size);
@ -2802,6 +3077,13 @@ static void virtio_net_device_unrealize(DeviceState *dev, Error **errp)
g_free(n->mac_table.macs); g_free(n->mac_table.macs);
g_free(n->vlans); g_free(n->vlans);
if (n->failover) {
g_free(n->primary_device_id);
g_free(n->standby_id);
qobject_unref(n->primary_device_dict);
n->primary_device_dict = NULL;
}
max_queues = n->multiqueue ? n->max_queues : 1; max_queues = n->multiqueue ? n->max_queues : 1;
for (i = 0; i < max_queues; i++) { for (i = 0; i < max_queues; i++) {
virtio_net_del_queue(n, i); virtio_net_del_queue(n, i);
@ -2839,6 +3121,23 @@ static int virtio_net_pre_save(void *opaque)
return 0; return 0;
} }
static bool primary_unplug_pending(void *opaque)
{
DeviceState *dev = opaque;
VirtIODevice *vdev = VIRTIO_DEVICE(dev);
VirtIONet *n = VIRTIO_NET(vdev);
return n->primary_dev ? n->primary_dev->pending_deleted_event : false;
}
static bool dev_unplug_pending(void *opaque)
{
DeviceState *dev = opaque;
VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(dev);
return vdc->primary_unplug_pending(dev);
}
static const VMStateDescription vmstate_virtio_net = { static const VMStateDescription vmstate_virtio_net = {
.name = "virtio-net", .name = "virtio-net",
.minimum_version_id = VIRTIO_NET_VM_VERSION, .minimum_version_id = VIRTIO_NET_VM_VERSION,
@ -2848,6 +3147,7 @@ static const VMStateDescription vmstate_virtio_net = {
VMSTATE_END_OF_LIST() VMSTATE_END_OF_LIST()
}, },
.pre_save = virtio_net_pre_save, .pre_save = virtio_net_pre_save,
.dev_unplug_pending = dev_unplug_pending,
}; };
static Property virtio_net_properties[] = { static Property virtio_net_properties[] = {
@ -2909,6 +3209,7 @@ static Property virtio_net_properties[] = {
true), true),
DEFINE_PROP_INT32("speed", VirtIONet, net_conf.speed, SPEED_UNKNOWN), DEFINE_PROP_INT32("speed", VirtIONet, net_conf.speed, SPEED_UNKNOWN),
DEFINE_PROP_STRING("duplex", VirtIONet, net_conf.duplex_str), DEFINE_PROP_STRING("duplex", VirtIONet, net_conf.duplex_str),
DEFINE_PROP_BOOL("failover", VirtIONet, failover, false),
DEFINE_PROP_END_OF_LIST(), DEFINE_PROP_END_OF_LIST(),
}; };
@ -2934,6 +3235,7 @@ static void virtio_net_class_init(ObjectClass *klass, void *data)
vdc->legacy_features |= (0x1 << VIRTIO_NET_F_GSO); vdc->legacy_features |= (0x1 << VIRTIO_NET_F_GSO);
vdc->post_load = virtio_net_post_load_virtio; vdc->post_load = virtio_net_post_load_virtio;
vdc->vmsd = &vmstate_virtio_net_device; vdc->vmsd = &vmstate_virtio_net_device;
vdc->primary_unplug_pending = primary_unplug_pending;
} }
static const TypeInfo virtio_net_info = { static const TypeInfo virtio_net_info = {

View File

@ -18,6 +18,7 @@
#include "standard-headers/linux/virtio_net.h" #include "standard-headers/linux/virtio_net.h"
#include "hw/virtio/virtio.h" #include "hw/virtio/virtio.h"
#include "net/announce.h" #include "net/announce.h"
#include "qemu/option_int.h"
#define TYPE_VIRTIO_NET "virtio-net-device" #define TYPE_VIRTIO_NET "virtio-net-device"
#define VIRTIO_NET(obj) \ #define VIRTIO_NET(obj) \
@ -43,6 +44,7 @@ typedef struct virtio_net_conf
int32_t speed; int32_t speed;
char *duplex_str; char *duplex_str;
uint8_t duplex; uint8_t duplex;
char *primary_id_str;
} virtio_net_conf; } virtio_net_conf;
/* Coalesced packets type & status */ /* Coalesced packets type & status */
@ -187,6 +189,16 @@ struct VirtIONet {
AnnounceTimer announce_timer; AnnounceTimer announce_timer;
bool needs_vnet_hdr_swap; bool needs_vnet_hdr_swap;
bool mtu_bypass_backend; bool mtu_bypass_backend;
QemuOpts *primary_device_opts;
QDict *primary_device_dict;
DeviceState *primary_dev;
BusState *primary_bus;
char *primary_device_id;
char *standby_id;
bool primary_should_be_hidden;
bool failover;
DeviceListener primary_listener;
Notifier migration_state;
}; };
void virtio_net_set_netclient_name(VirtIONet *n, const char *name, void virtio_net_set_netclient_name(VirtIONet *n, const char *name,

View File

@ -160,6 +160,7 @@ typedef struct VirtioDeviceClass {
*/ */
int (*post_load)(VirtIODevice *vdev); int (*post_load)(VirtIODevice *vdev);
const VMStateDescription *vmsd; const VMStateDescription *vmsd;
bool (*primary_unplug_pending)(void *opaque);
} VirtioDeviceClass; } VirtioDeviceClass;
void virtio_instance_init_common(Object *proxy_obj, void *data, void virtio_instance_init_common(Object *proxy_obj, void *data,