[libvirt] [PATCH v9 0/3] DEVICE_DELETED event

libvirt has a long-standing bug: when removing the device, it can request removal but does not know when the removal completes. Add an event so we can fix this in a robust way. First patch only adds the event with ID, second patch adds a path field. Split this way for ease of backport (stable downstreams without QOM would want to only take the first patch). Event without fields is still useful as management can use it to poll device list to figure out which device was removed. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> If there are no more comments I'll stick this on my pci branch. Changes from v8: - reorder qom destruction so no need to change unparent Changes from v7: - none, v7 was malformed series sent by mistake Changes from v6: - make empty event use data: {}, Markus prefers this Changes from v5: - Emit an empty event on unnamed devices in patch 1/3, as suggested by Markus Changes from v4: - Add extra triggers and extra fields as requested by Markus Changes from v3: - Document that we only emit events for devices with and ID, as suggested by Markus Changes from v2: - move event toward the end of device_unparent, so that parents are reported after their children, as suggested by Paolo Changes from v1: - move to device_unparent - address comments by Andreas and Eric -- Anthony Liguori Michael S. Tsirkin (3): qdev: DEVICE_DELETED event qom: call class destructor before unparent qmp: add path to device_deleted event QMP/qmp-events.txt | 18 ++++++++++++++++++ hw/qdev.c | 14 ++++++++++++++ include/monitor/monitor.h | 1 + monitor.c | 1 + qapi-schema.json | 4 +++- qom/object.c | 6 +++--- 6 files changed, 40 insertions(+), 4 deletions(-) -- MST

libvirt has a long-standing bug: when removing the device, it can request removal but does not know when the removal completes. Add an event so we can fix this in a robust way. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> --- QMP/qmp-events.txt | 16 ++++++++++++++++ hw/qdev.c | 11 +++++++++++ include/monitor/monitor.h | 1 + monitor.c | 1 + qapi-schema.json | 4 +++- 5 files changed, 32 insertions(+), 1 deletion(-) diff --git a/QMP/qmp-events.txt b/QMP/qmp-events.txt index b2698e4..24cf3e8 100644 --- a/QMP/qmp-events.txt +++ b/QMP/qmp-events.txt @@ -136,6 +136,22 @@ Example: Note: The "ready to complete" status is always reset by a BLOCK_JOB_ERROR event. +DEVICE_DELETED +----------------- + +Emitted whenever the device removal completion is acknowledged +by the guest. +At this point, it's safe to reuse the specified device ID. +Device removal can be initiated by the guest or by HMP/QMP commands. + +Data: + +- "device": device name (json-string, optional) + +{ "event": "DEVICE_DELETED", + "data": { "device": "virtio-net-pci-0" }, + "timestamp": { "seconds": 1265044230, "microseconds": 450486 } } + DEVICE_TRAY_MOVED ----------------- diff --git a/hw/qdev.c b/hw/qdev.c index 0b20280..5e8a89c 100644 --- a/hw/qdev.c +++ b/hw/qdev.c @@ -30,6 +30,8 @@ #include "qapi/error.h" #include "qapi/qmp/qerror.h" #include "qapi/visitor.h" +#include "qapi/qmp/qjson.h" +#include "monitor/monitor.h" int qdev_hotplug = 0; static bool qdev_hot_added = false; @@ -761,6 +763,7 @@ static void device_unparent(Object *obj) DeviceState *dev = DEVICE(obj); DeviceClass *dc = DEVICE_GET_CLASS(dev); BusState *bus; + QObject *event_data; while (dev->num_child_bus) { bus = QLIST_FIRST(&dev->child_bus); @@ -779,6 +782,14 @@ static void device_unparent(Object *obj) object_unref(OBJECT(dev->parent_bus)); dev->parent_bus = NULL; } + + if (dev->id) { + event_data = qobject_from_jsonf("{ 'device': %s }", dev->id); + } else { + event_data = qobject_from_jsonf("{ }"); + } + monitor_protocol_event(QEVENT_DEVICE_DELETED, event_data); + qobject_decref(event_data); } static void device_class_init(ObjectClass *class, void *data) diff --git a/include/monitor/monitor.h b/include/monitor/monitor.h index 87fb49c..b868760 100644 --- a/include/monitor/monitor.h +++ b/include/monitor/monitor.h @@ -39,6 +39,7 @@ typedef enum MonitorEvent { QEVENT_BLOCK_JOB_CANCELLED, QEVENT_BLOCK_JOB_ERROR, QEVENT_BLOCK_JOB_READY, + QEVENT_DEVICE_DELETED, QEVENT_DEVICE_TRAY_MOVED, QEVENT_SUSPEND, QEVENT_SUSPEND_DISK, diff --git a/monitor.c b/monitor.c index 112e920..2fdfb79 100644 --- a/monitor.c +++ b/monitor.c @@ -458,6 +458,7 @@ static const char *monitor_event_names[] = { [QEVENT_BLOCK_JOB_CANCELLED] = "BLOCK_JOB_CANCELLED", [QEVENT_BLOCK_JOB_ERROR] = "BLOCK_JOB_ERROR", [QEVENT_BLOCK_JOB_READY] = "BLOCK_JOB_READY", + [QEVENT_DEVICE_DELETED] = "DEVICE_DELETED", [QEVENT_DEVICE_TRAY_MOVED] = "DEVICE_TRAY_MOVED", [QEVENT_SUSPEND] = "SUSPEND", [QEVENT_SUSPEND_DISK] = "SUSPEND_DISK", diff --git a/qapi-schema.json b/qapi-schema.json index fdaa9da..080dc39 100644 --- a/qapi-schema.json +++ b/qapi-schema.json @@ -2354,7 +2354,9 @@ # Notes: When this command completes, the device may not be removed from the # guest. Hot removal is an operation that requires guest cooperation. # This command merely requests that the guest begin the hot removal -# process. +# process. Completion of the device removal process is signaled with a +# DEVICE_DELETED event. Guest reset will automatically complete removal +# for all devices. # # Since: 0.14.0 ## -- MST

"Michael S. Tsirkin" <mst@redhat.com> writes:
libvirt has a long-standing bug: when removing the device, it can request removal but does not know when the removal completes. Add an event so we can fix this in a robust way.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Markus Armbruster <armbru@redhat.com>

Make sure object is valid when destructor is called. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> --- qom/object.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/qom/object.c b/qom/object.c index 3d638ff..a0e3cbe 100644 --- a/qom/object.c +++ b/qom/object.c @@ -363,12 +363,12 @@ static void object_property_del_child(Object *obj, Object *child, Error **errp) void object_unparent(Object *obj) { object_ref(obj); - if (obj->parent) { - object_property_del_child(obj->parent, obj, NULL); - } if (obj->class->unparent) { (obj->class->unparent)(obj); } + if (obj->parent) { + object_property_del_child(obj->parent, obj, NULL); + } object_unref(obj); } -- MST

Il 21/03/2013 13:07, Michael S. Tsirkin ha scritto:
Make sure object is valid when destructor is called.
"Make sure the object still has a canonical path while the unparent callback is running". Otherwise, the series looks good. Paolo
Signed-off-by: Michael S. Tsirkin <mst@redhat.com> --- qom/object.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/qom/object.c b/qom/object.c index 3d638ff..a0e3cbe 100644 --- a/qom/object.c +++ b/qom/object.c @@ -363,12 +363,12 @@ static void object_property_del_child(Object *obj, Object *child, Error **errp) void object_unparent(Object *obj) { object_ref(obj); - if (obj->parent) { - object_property_del_child(obj->parent, obj, NULL); - } if (obj->class->unparent) { (obj->class->unparent)(obj); } + if (obj->parent) { + object_property_del_child(obj->parent, obj, NULL); + } object_unref(obj); }

On Thu, Mar 21, 2013 at 01:25:50PM +0100, Paolo Bonzini wrote:
Il 21/03/2013 13:07, Michael S. Tsirkin ha scritto:
Make sure object is valid when destructor is called.
"Make sure the object still has a canonical path while the unparent callback is running".
Otherwise, the series looks good.
Paolo
Worth respinning for?
Signed-off-by: Michael S. Tsirkin <mst@redhat.com> --- qom/object.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/qom/object.c b/qom/object.c index 3d638ff..a0e3cbe 100644 --- a/qom/object.c +++ b/qom/object.c @@ -363,12 +363,12 @@ static void object_property_del_child(Object *obj, Object *child, Error **errp) void object_unparent(Object *obj) { object_ref(obj); - if (obj->parent) { - object_property_del_child(obj->parent, obj, NULL); - } if (obj->class->unparent) { (obj->class->unparent)(obj); } + if (obj->parent) { + object_property_del_child(obj->parent, obj, NULL); + } object_unref(obj); }

Il 21/03/2013 13:30, Michael S. Tsirkin ha scritto:
On Thu, Mar 21, 2013 at 01:25:50PM +0100, Paolo Bonzini wrote:
Il 21/03/2013 13:07, Michael S. Tsirkin ha scritto:
Make sure object is valid when destructor is called.
"Make sure the object still has a canonical path while the unparent callback is running".
Otherwise, the series looks good.
Worth respinning for?
If it goes through the pci branch, of course not but still make the change. Otherwise, probably not. Paolo

Add QOM path to device deleted event. It now becomes useful to report it for devices which don't have an ID assigned. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> --- QMP/qmp-events.txt | 4 +++- hw/qdev.c | 7 +++++-- 2 files changed, 8 insertions(+), 3 deletions(-) diff --git a/QMP/qmp-events.txt b/QMP/qmp-events.txt index 24cf3e8..dcc826d 100644 --- a/QMP/qmp-events.txt +++ b/QMP/qmp-events.txt @@ -147,9 +147,11 @@ Device removal can be initiated by the guest or by HMP/QMP commands. Data: - "device": device name (json-string, optional) +- "path": device path (json-string) { "event": "DEVICE_DELETED", - "data": { "device": "virtio-net-pci-0" }, + "data": { "device": "virtio-net-pci-0", + "path": "/machine/peripheral/virtio-net-pci-0" }, "timestamp": { "seconds": 1265044230, "microseconds": 450486 } } DEVICE_TRAY_MOVED diff --git a/hw/qdev.c b/hw/qdev.c index 5e8a89c..0cdf568 100644 --- a/hw/qdev.c +++ b/hw/qdev.c @@ -764,6 +764,7 @@ static void device_unparent(Object *obj) DeviceClass *dc = DEVICE_GET_CLASS(dev); BusState *bus; QObject *event_data; + gchar *path = object_get_canonical_path(obj); while (dev->num_child_bus) { bus = QLIST_FIRST(&dev->child_bus); @@ -784,12 +785,14 @@ static void device_unparent(Object *obj) } if (dev->id) { - event_data = qobject_from_jsonf("{ 'device': %s }", dev->id); + event_data = qobject_from_jsonf("{ 'device': %s, 'path': %s }", + dev->id, path); } else { - event_data = qobject_from_jsonf("{ }"); + event_data = qobject_from_jsonf("{ 'path': %s }", path); } monitor_protocol_event(QEVENT_DEVICE_DELETED, event_data); qobject_decref(event_data); + g_free(path); } static void device_class_init(ObjectClass *class, void *data) -- MST

"Michael S. Tsirkin" <mst@redhat.com> writes:
Add QOM path to device deleted event. It now becomes useful to report it for devices which don't have an ID assigned.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Markus Armbruster <armbru@redhat.com>

Am 21.03.2013 13:07, schrieb Michael S. Tsirkin:
Changes from v8: - reorder qom destruction so no need to change unparent
Changes from v7: - none, v7 was malformed series sent by mistake
Changes from v6: - make empty event use data: {}, Markus prefers this
Changes from v5: - Emit an empty event on unnamed devices in patch 1/3, as suggested by Markus
Changes from v4: - Add extra triggers and extra fields as requested by Markus
Series looks good, Reviewed-by: Andreas Färber <afaerber@suse.de> Andreas
Changes from v3: - Document that we only emit events for devices with and ID, as suggested by Markus Changes from v2: - move event toward the end of device_unparent, so that parents are reported after their children, as suggested by Paolo Changes from v1: - move to device_unparent - address comments by Andreas and Eric
-- SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg
participants (4)
-
Andreas Färber
-
Markus Armbruster
-
Michael S. Tsirkin
-
Paolo Bonzini