qemu: add DEVICE_UNPLUG_GUEST_ERROR event support

The upcoming QEMU 6.2.0 implements a new event called
DEVICE_UNPLUG_GUEST_ERROR, a new event that reports generic device
unplug errors that were detected by the guest and reported back to QEMU.

This new event is going to be specially useful for pseries guests that
uses newer kernels (must have kernel commit 29c9a2699e71), which is the
case for Fedora 34 at this moment. These guests have the capability of
reporting CPU removal errors back to QEMU which, starting in 6.2.0, will
emit the DEVICE_UNPLUG_GUEST_ERROR event. Libvirt can use this event to
abort the device removal immediately instead of waiting for 'setvcpus'
timeout.

QEMU 6.2.0 is also going to emit DEVICE_UNPLUG_GUEST_ERROR for memory
hotunplug errors, both in pseries and ACPI guests. QEMU 6.1.0 reports
memory removal errors using the MEM_UNPLUG_ERROR event, which is going to
be deprecated by DEVICE_UNPLUG_GUEST_ERROR in 6.2.0. Given that
Libvirt wasn't handling the MEM_UNPLUG_ERROR event we don't need to
worry about it - adding support to DEVICE_UNPLUG_GUEST_ERROR will be
enough to cover all future cases.

This patch adds support to DEVICE_UNPLUG_GUEST_ERROR by adding the
minimal wiring required for Libvirt to be aware of it. The monitor
callback for this event will abort the pending removal operation of the
device reported by the "device" property of the event. Most of the heavy
lifting is already done by existing code that handles
QEMU_DOMAIN_UNPLUGGING_DEVICE_STATUS_GUEST_REJECTED, making our life
easier to abort the pending removal operation.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

src/qemu/qemu_monitor.c

@@ -1348,6 +1348,18 @@ qemuMonitorEmitDeviceDeleted(qemuMonitor *mon,
 }
 
 
+void
+qemuMonitorEmitDeviceUnplugErr(qemuMonitor *mon,
+                               const char *devPath,
+                               const char *devAlias)
+{
+    VIR_DEBUG("mon=%p", mon);
+
+    QEMU_MONITOR_CALLBACK(mon, domainDeviceUnplugError, mon->vm,
+                          devPath, devAlias);
+}
+
+
 void
 qemuMonitorEmitNicRxFilterChanged(qemuMonitor *mon,
                                   const char *devAlias)

src/qemu/qemu_monitor.h

@@ -294,6 +294,11 @@ typedef void (*qemuMonitorDomainDeviceDeletedCallback)(qemuMonitor *mon,
                                                        virDomainObj *vm,
                                                        const char *devAlias,
                                                        void *opaque);
+typedef void (*qemuMonitorDomainDeviceUnplugErrCallback)(qemuMonitor *mon,
+                                                         virDomainObj *vm,
+                                                         const char *devPath,
+                                                         const char *devAlias,
+                                                         void *opaque);
 typedef void (*qemuMonitorDomainNicRxFilterChangedCallback)(qemuMonitor *mon,
                                                             virDomainObj *vm,
                                                             const char *devAlias,
@@ -454,6 +459,7 @@ struct _qemuMonitorCallbacks {
     qemuMonitorDomainGuestCrashloadedCallback domainGuestCrashloaded;
     qemuMonitorDomainMemoryFailureCallback domainMemoryFailure;
     qemuMonitorDomainMemoryDeviceSizeChange domainMemoryDeviceSizeChange;
+    qemuMonitorDomainDeviceUnplugErrCallback domainDeviceUnplugError;
 };
 
 qemuMonitor *qemuMonitorOpen(virDomainObj *vm,
@@ -542,6 +548,9 @@ void qemuMonitorEmitGuestPanic(qemuMonitor *mon,
                                qemuMonitorEventPanicInfo *info);
 void qemuMonitorEmitDeviceDeleted(qemuMonitor *mon,
                                   const char *devAlias);
+void qemuMonitorEmitDeviceUnplugErr(qemuMonitor *mon,
+                                    const char *devPath,
+                                    const char *devAlias);
 void qemuMonitorEmitNicRxFilterChanged(qemuMonitor *mon,
                                        const char *devAlias);
 void qemuMonitorEmitSerialChange(qemuMonitor *mon,

src/qemu/qemu_monitor_json.c

@@ -113,6 +113,7 @@ static void qemuMonitorJSONHandlePRManagerStatusChanged(qemuMonitor *mon, virJSO
 static void qemuMonitorJSONHandleRdmaGidStatusChanged(qemuMonitor *mon, virJSONValue *data);
 static void qemuMonitorJSONHandleMemoryFailure(qemuMonitor *mon, virJSONValue *data);
 static void qemuMonitorJSONHandleMemoryDeviceSizeChange(qemuMonitor *mon, virJSONValue *data);
+static void qemuMonitorJSONHandleDeviceUnplugErr(qemuMonitor *mon, virJSONValue *data);
 
 typedef struct {
     const char *type;
@@ -129,6 +130,7 @@ static qemuEventHandler eventHandlers[] = {
     { "BLOCK_WRITE_THRESHOLD", qemuMonitorJSONHandleBlockThreshold, },
     { "DEVICE_DELETED", qemuMonitorJSONHandleDeviceDeleted, },
     { "DEVICE_TRAY_MOVED", qemuMonitorJSONHandleTrayChange, },
+    { "DEVICE_UNPLUG_GUEST_ERROR", qemuMonitorJSONHandleDeviceUnplugErr, },
     { "DUMP_COMPLETED", qemuMonitorJSONHandleDumpCompleted, },
     { "GUEST_CRASHLOADED", qemuMonitorJSONHandleGuestCrashloaded, },
     { "GUEST_PANICKED", qemuMonitorJSONHandleGuestPanic, },
@@ -1111,6 +1113,23 @@ qemuMonitorJSONHandleDeviceDeleted(qemuMonitor *mon, virJSONValue *data)
 }
 
 
+static void
+qemuMonitorJSONHandleDeviceUnplugErr(qemuMonitor *mon, virJSONValue *data)
+{
+    const char *device;
+    const char *path;
+
+    if (!(path = virJSONValueObjectGetString(data, "path"))) {
+        VIR_DEBUG("missing path in device unplug guest error event");
+        return;
+    }
+
+    device = virJSONValueObjectGetString(data, "device");
+
+    qemuMonitorEmitDeviceUnplugErr(mon, path, device);
+}
+
+
 static void
 qemuMonitorJSONHandleNicRxFilterChanged(qemuMonitor *mon, virJSONValue *data)
 {

src/qemu/qemu_process.c

@@ -1322,6 +1322,42 @@ qemuProcessHandleDeviceDeleted(qemuMonitor *mon G_GNUC_UNUSED,
 }
 
 
+static void
+qemuProcessHandleDeviceUnplugErr(qemuMonitor *mon G_GNUC_UNUSED,
+                                 virDomainObj *vm,
+                                 const char *devPath,
+                                 const char *devAlias,
+                                 void *opaque)
+{
+    virQEMUDriver *driver = opaque;
+    virObjectEvent *event = NULL;
+
+    virObjectLock(vm);
+
+    VIR_DEBUG("Device %s QOM path %s failed to be removed from domain %p %s",
+              devAlias, devPath, vm, vm->def->name);
+
+    /*
+     * DEVICE_UNPLUG_GUEST_ERROR will always contain the QOM path
+     * but QEMU will not guarantee that devAlias will be provided.
+     *
+     * However, given that all Libvirt devices have a devAlias, we
+     * can ignore the case where QEMU emitted this event without it.
+     */
+    if (!devAlias)
+        goto cleanup;
+
+    qemuDomainSignalDeviceRemoval(vm, devAlias,
+                                  QEMU_DOMAIN_UNPLUGGING_DEVICE_STATUS_GUEST_REJECTED);
+
+    event = virDomainEventDeviceRemovalFailedNewFromObj(vm, devAlias);
+
+ cleanup:
+    virObjectUnlock(vm);
+    virObjectEventStateQueue(driver->domainEventState, event);
+}
+
+
 /**
  *
  * Meaning of fields reported by the event according to the ACPI standard:
@@ -1891,6 +1927,7 @@ static qemuMonitorCallbacks monitorCallbacks = {
     .domainGuestCrashloaded = qemuProcessHandleGuestCrashloaded,
     .domainMemoryFailure = qemuProcessHandleMemoryFailure,
     .domainMemoryDeviceSizeChange = qemuProcessHandleMemoryDeviceSizeChange,
+    .domainDeviceUnplugError = qemuProcessHandleDeviceUnplugErr,
 };
 
 static void