system/runstate: add VM state change cb with return value

This patch adds the new VM state change cb type `VMChangeStateHandlerWithRet`,
which has return value for `VMChangeStateEntry`.

Thus, we can register a new VM state change cb with return value for device.
Note that `VMChangeStateHandler` and `VMChangeStateHandlerWithRet` are mutually
exclusive and cannot be provided at the same time.

This patch is the pre patch for 'vhost-user: return failure if backend crashes
when live migration', which makes the live migration aware of the loss of
connection with the vhost-user backend and aborts the live migration.

Virtio device will use VMChangeStateHandlerWithRet.

Signed-off-by: Haoqian He <haoqian.he@smartx.com>
Message-Id: <20250416024729.3289157-2-haoqian.he@smartx.com>
Tested-by: Lei Yang <leiyang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

system/cpus.c

@@ -299,14 +299,18 @@ static int do_vm_stop(RunState state, bool send_stop)
         if (oldstate == RUN_STATE_RUNNING) {
             pause_all_vcpus();
         }
-        vm_state_notify(0, state);
+        ret = vm_state_notify(0, state);
         if (send_stop) {
             qapi_event_send_stop();
         }
     }
 
     bdrv_drain_all();
-    ret = bdrv_flush_all();
+    /*
+     * Even if vm_state_notify() return failure,
+     * it would be better to flush as before.
+     */
+    ret |= bdrv_flush_all();
     trace_vm_stop_flush_all(ret);
 
     return ret;

system/runstate.c

@@ -297,6 +297,7 @@ void qemu_system_vmstop_request(RunState state)
 struct VMChangeStateEntry {
     VMChangeStateHandler *cb;
     VMChangeStateHandler *prepare_cb;
+    VMChangeStateHandlerWithRet *cb_ret;
     void *opaque;
     QTAILQ_ENTRY(VMChangeStateEntry) entries;
     int priority;
@@ -320,14 +321,15 @@ static QTAILQ_HEAD(, VMChangeStateEntry) vm_change_state_head =
 VMChangeStateEntry *qemu_add_vm_change_state_handler_prio(
         VMChangeStateHandler *cb, void *opaque, int priority)
 {
-    return qemu_add_vm_change_state_handler_prio_full(cb, NULL, opaque,
-                                                      priority);
+    return qemu_add_vm_change_state_handler_prio_full(cb, NULL, NULL,
+                                                      opaque, priority);
 }
 
 /**
  * qemu_add_vm_change_state_handler_prio_full:
  * @cb: the main callback to invoke
  * @prepare_cb: a callback to invoke before the main callback
+ * @cb_ret: the main callback to invoke with return value
  * @opaque: user data passed to the callbacks
  * @priority: low priorities execute first when the vm runs and the reverse is
  *            true when the vm stops
@@ -344,6 +346,7 @@ VMChangeStateEntry *qemu_add_vm_change_state_handler_prio(
 VMChangeStateEntry *
 qemu_add_vm_change_state_handler_prio_full(VMChangeStateHandler *cb,
                                            VMChangeStateHandler *prepare_cb,
+                                           VMChangeStateHandlerWithRet *cb_ret,
                                            void *opaque, int priority)
 {
     VMChangeStateEntry *e;
@@ -352,6 +355,7 @@ qemu_add_vm_change_state_handler_prio_full(VMChangeStateHandler *cb,
     e = g_malloc0(sizeof(*e));
     e->cb = cb;
     e->prepare_cb = prepare_cb;
+    e->cb_ret = cb_ret;
     e->opaque = opaque;
     e->priority = priority;
 
@@ -379,9 +383,10 @@ void qemu_del_vm_change_state_handler(VMChangeStateEntry *e)
     g_free(e);
 }
 
-void vm_state_notify(bool running, RunState state)
+int vm_state_notify(bool running, RunState state)
 {
     VMChangeStateEntry *e, *next;
+    int ret = 0;
 
     trace_vm_state_notify(running, state, RunState_str(state));
 
@@ -393,7 +398,17 @@ void vm_state_notify(bool running, RunState state)
         }
 
         QTAILQ_FOREACH_SAFE(e, &vm_change_state_head, entries, next) {
-            e->cb(e->opaque, running, state);
+            if (e->cb) {
+                e->cb(e->opaque, running, state);
+            } else if (e->cb_ret) {
+                /*
+                 * Here ignore the return value of cb_ret because
+                 * we only care about the stopping the device during
+                 * the VM live migration to indicate whether the
+                 * connection between qemu and backend is normal.
+                 */
+                e->cb_ret(e->opaque, running, state);
+            }
         }
     } else {
         QTAILQ_FOREACH_REVERSE_SAFE(e, &vm_change_state_head, entries, next) {
@@ -403,9 +418,19 @@ void vm_state_notify(bool running, RunState state)
         }
 
         QTAILQ_FOREACH_REVERSE_SAFE(e, &vm_change_state_head, entries, next) {
-            e->cb(e->opaque, running, state);
+            if (e->cb) {
+                e->cb(e->opaque, running, state);
+            } else if (e->cb_ret) {
+                /*
+                 * We should execute all registered callbacks even if
+                 * one of them returns failure, otherwise, some cleanup
+                 * work of the device will be skipped.
+                 */
+                ret |= e->cb_ret(e->opaque, running, state);
+            }
         }
     }
+    return ret;
 }
 
 static ShutdownCause reset_requested;