[PATCH v3 2/7] Implement drain_call_rcu and use it in hmp_device_del

[Maxim Levitsky]

This allows to preserve the semantics of hmp_device_del,
that the device is deleted immediatly which was changed by previos
patch that delayed this to RCU callback

Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Suggested-by: Stefan Hajnoczi <stefanha@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 include/qemu/rcu.h |  1 +
 qdev-monitor.c     | 22 +++++++++++++++++++
 util/rcu.c         | 55 ++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 78 insertions(+)

diff --git a/include/qemu/rcu.h b/include/qemu/rcu.h
index 570aa603eb..0e375ebe13 100644
--- a/include/qemu/rcu.h
+++ b/include/qemu/rcu.h
@@ -133,6 +133,7 @@ struct rcu_head {
 };
 
 extern void call_rcu1(struct rcu_head *head, RCUCBFunc *func);
+extern void drain_call_rcu(void);
 
 /* The operands of the minus operator must have the same type,
  * which must be the one that we specify in the cast.
diff --git a/qdev-monitor.c b/qdev-monitor.c
index 71ebce19df..a889fbc509 100644
--- a/qdev-monitor.c
+++ b/qdev-monitor.c
@@ -803,6 +803,18 @@ void qmp_device_add(QDict *qdict, QObject **ret_data, 
Error **errp)
         return;
     }
     dev = qdev_device_add(opts, errp);
+
+    /*
+     * Drain all pending RCU callbacks. This is done because
+     * some bus related operations can delay a device removal
+     * (in this case this can happen if device is added and then
+     * removed due to a configuration error)
+     * to a RCU callback, but user might expect that this interface
+     * will finish its job completely once qmp command returns result
+     * to the user
+     */
+    drain_call_rcu();
+
     if (!dev) {
         qemu_opts_del(opts);
         return;
@@ -894,6 +906,16 @@ void qmp_device_del(const char *id, Error **errp)
         }
 
         qdev_unplug(dev, errp);
+
+        /*
+         * Drain all pending RCU callbacks. This is done because
+         * some bus related operations can delay a device removal
+         * to a RCU callback, but user might expect that this interface
+         * will finish its job completely once qmp command returns result
+         * to the user
+         */
+
+        drain_call_rcu();
     }
 }
 
diff --git a/util/rcu.c b/util/rcu.c
index 60a37f72c3..c4fefa9333 100644
--- a/util/rcu.c
+++ b/util/rcu.c
@@ -293,6 +293,61 @@ void call_rcu1(struct rcu_head *node, void (*func)(struct 
rcu_head *node))
     qemu_event_set(&rcu_call_ready_event);
 }
 
+
+struct rcu_drain {
+    struct rcu_head rcu;
+    QemuEvent drain_complete_event;
+};
+
+static void drain_rcu_callback(struct rcu_head *node)
+{
+    struct rcu_drain *event = (struct rcu_drain *)node;
+    qemu_event_set(&event->drain_complete_event);
+}
+
+/*
+ * This function ensures that all pending RCU callbacks
+ * on the current thread are done executing
+
+ * drops big qemu lock during the wait to allow RCU thread
+ * to process the callbacks
+ *
+ */
+
+void drain_call_rcu(void)
+{
+    struct rcu_drain rcu_drain;
+    bool locked = qemu_mutex_iothread_locked();
+
+    memset(&rcu_drain, 0, sizeof(struct rcu_drain));
+    qemu_event_init(&rcu_drain.drain_complete_event, false);
+
+    if (locked) {
+        qemu_mutex_unlock_iothread();
+    }
+
+
+    /*
+     * RCU callbacks are invoked in the same order as in which they
+     * are registered, thus we can be sure that when 'drain_rcu_callback'
+     * is called, all RCU callbacks that were registered on this thread
+     * prior to calling this function are completed.
+     *
+     * Note that since we have only one global queue of the RCU callbacks,
+     * we also end up waiting for most of RCU callbacks that were registered
+     * on the other threads, but this is a side effect that shoudn't be
+     * assumed.
+     */
+
+    call_rcu1(&rcu_drain.rcu, drain_rcu_callback);
+    qemu_event_wait(&rcu_drain.drain_complete_event);
+
+    if (locked) {
+        qemu_mutex_lock_iothread();
+    }
+
+}
+
 void rcu_register_thread(void)
 {
     assert(rcu_reader.ctr == 0);
-- 
2.26.2