blockjob: Wake up BDS when job becomes idle

In the context of draining a BDS, the .drained_poll callback of block
jobs is called. If this returns true (i.e. there is still some activity
pending), the drain operation may call aio_poll() with blocking=true to
wait for completion.

As soon as the pending activity is completed and the job finally arrives
in a quiescent state (i.e. its coroutine either yields with busy=false
or terminates), the block job must notify the aio_poll() loop to wake
up, otherwise we get a deadlock if both are running in different
threads.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>

blockjob.c

@@ -221,6 +221,22 @@ int block_job_add_bdrv(BlockJob *job, const char *name, BlockDriverState *bs,
     return 0;
 }
 
+void block_job_wakeup_all_bdrv(BlockJob *job)
+{
+    GSList *l;
+
+    for (l = job->nodes; l; l = l->next) {
+        BdrvChild *c = l->data;
+        bdrv_wakeup(c->bs);
+    }
+}
+
+static void block_job_on_idle(Notifier *n, void *opaque)
+{
+    BlockJob *job = opaque;
+    block_job_wakeup_all_bdrv(job);
+}
+
 bool block_job_is_internal(BlockJob *job)
 {
     return (job->job.id == NULL);
@@ -416,6 +432,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
     job->finalize_completed_notifier.notify = block_job_event_completed;
     job->pending_notifier.notify = block_job_event_pending;
     job->ready_notifier.notify = block_job_event_ready;
+    job->idle_notifier.notify = block_job_on_idle;
 
     notifier_list_add(&job->job.on_finalize_cancelled,
                       &job->finalize_cancelled_notifier);
@@ -423,6 +440,7 @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver,
                       &job->finalize_completed_notifier);
     notifier_list_add(&job->job.on_pending, &job->pending_notifier);
     notifier_list_add(&job->job.on_ready, &job->ready_notifier);
+    notifier_list_add(&job->job.on_idle, &job->idle_notifier);
 
     error_setg(&job->blocker, "block device is in use by block job: %s",
                job_type_str(&job->job));

include/block/blockjob.h

@@ -70,6 +70,9 @@ typedef struct BlockJob {
     /** Called when the job transitions to READY */
     Notifier ready_notifier;
 
+    /** Called when the job coroutine yields or terminates */
+    Notifier idle_notifier;
+
     /** BlockDriverStates that are involved in this block job */
     GSList *nodes;
 } BlockJob;
@@ -118,6 +121,16 @@ int block_job_add_bdrv(BlockJob *job, const char *name, BlockDriverState *bs,
  */
 void block_job_remove_all_bdrv(BlockJob *job);
 
+/**
+ * block_job_wakeup_all_bdrv:
+ * @job: The block job
+ *
+ * Calls bdrv_wakeup() for all BlockDriverStates that have been added to the
+ * job. This function is to be called whenever child_job_drained_poll() would
+ * go from true to false to notify waiting drain requests.
+ */
+void block_job_wakeup_all_bdrv(BlockJob *job);
+
 /**
  * block_job_set_speed:
  * @job: The job to set the speed for.

include/qemu/job.h

@@ -156,6 +156,9 @@ typedef struct Job {
     /** Notifiers called when the job transitions to READY */
     NotifierList on_ready;
 
+    /** Notifiers called when the job coroutine yields or terminates */
+    NotifierList on_idle;
+
     /** Element of the list of jobs */
     QLIST_ENTRY(Job) job_list;
 

job.c

@@ -402,6 +402,11 @@ static void job_event_ready(Job *job)
     notifier_list_notify(&job->on_ready, job);
 }
 
+static void job_event_idle(Job *job)
+{
+    notifier_list_notify(&job->on_idle, job);
+}
+
 void job_enter_cond(Job *job, bool(*fn)(Job *job))
 {
     if (!job_started(job)) {
@@ -447,6 +452,7 @@ static void coroutine_fn job_do_yield(Job *job, uint64_t ns)
         timer_mod(&job->sleep_timer, ns);
     }
     job->busy = false;
+    job_event_idle(job);
     job_unlock();
     qemu_coroutine_yield();
 
@@ -865,6 +871,7 @@ static void coroutine_fn job_co_entry(void *opaque)
     assert(job && job->driver && job->driver->run);
     job_pause_point(job);
     job->ret = job->driver->run(job, &job->err);
+    job_event_idle(job);
     job->deferred_to_main_loop = true;
     aio_bh_schedule_oneshot(qemu_get_aio_context(), job_exit, job);
 }