AioContext: speed up aio_notify

In many cases, the call to event_notifier_set in aio_notify is unnecessary.
In particular, if we are executing aio_dispatch, or if aio_poll is not
blocking, we know that we will soon get to the next loop iteration (if
necessary); the thread that hosts the AioContext's event loop does not
need any nudging.

The patch includes a Promela formal model that shows that this really
works and does not need any further complication such as generation
counts.  It needs a memory barrier though.

The generation counts are not needed because any change to
ctx->dispatching after the memory barrier is okay for aio_notify.
If it changes from zero to one, it is the right thing to skip
event_notifier_set.  If it changes from one to zero, the
event_notifier_set is unnecessary but harmless.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

1 files changed, 9 insertions, 0 deletions

diff --git a/include/block/aio.h b/include/block/aio.h
index d81250cf2b..433e7ff8b6 100644
--- a/include/block/aio.h
+++ b/include/block/aio.h
@@ -60,8 +60,14 @@ struct AioContext {
      */
     int walking_handlers;
 
+    /* Used to avoid unnecessary event_notifier_set calls in aio_notify.
+     * Writes protected by lock or BQL, reads are lockless.
+     */
+    bool dispatching;
+
     /* lock to protect between bh's adders and deleter */
     QemuMutex bh_lock;
+
     /* Anchor of the list of Bottom Halves belonging to the context */
     struct QEMUBH *first_bh;
 
@@ -83,6 +89,9 @@ struct AioContext {
     QEMUTimerListGroup tlg;
 };
 
+/* Used internally to synchronize aio_poll against qemu_bh_schedule.  */
+void aio_set_dispatching(AioContext *ctx, bool dispatching);
+
 /**
  * aio_context_new: Allocate a new AioContext.
  *