block: Use a single global AioWait

When draining a block node, we recurse to its parent and for subtree
drains also to its children. A single AIO_WAIT_WHILE() is then used to
wait for bdrv_drain_poll() to become true, which depends on all of the
nodes we recursed to. However, if the respective child or parent becomes
quiescent and calls bdrv_wakeup(), only the AioWait of the child/parent
is checked, while AIO_WAIT_WHILE() depends on the AioWait of the
original node.

Fix this by using a single AioWait for all callers of AIO_WAIT_WHILE().

This may mean that the draining thread gets a few more unnecessary
wakeups because an unrelated operation got completed, but we already
wake it up when something _could_ have changed rather than only if it
has certainly changed.

Apart from that, drain is a slow path anyway. In theory it would be
possible to use wakeups more selectively and still correctly, but the
gains are likely not worth the additional complexity. In fact, this
patch is a nice simplification for some places in the code.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>

2 files changed, 6 insertions, 12 deletions

diff --git a/block/block-backend.c b/block/block-backend.c
index 551e4e1708..7b1ec5071b 100644
--- a/block/block-backend.c
+++ b/block/block-backend.c
@@ -88,7 +88,6 @@ struct BlockBackend {
      * Accessed with atomic ops.
      */
     unsigned int in_flight;
-    AioWait wait;
 };
 
 typedef struct BlockBackendAIOCB {
@@ -1298,7 +1297,7 @@ static void blk_inc_in_flight(BlockBackend *blk)
 static void blk_dec_in_flight(BlockBackend *blk)
 {
     atomic_dec(&blk->in_flight);
-    aio_wait_kick(&blk->wait);
+    aio_wait_kick();
 }
 
 static void error_callback_bh(void *opaque)
@@ -1599,9 +1598,8 @@ void blk_drain(BlockBackend *blk)
     }
 
     /* We may have -ENOMEDIUM completions in flight */
-    AIO_WAIT_WHILE(&blk->wait,
-            blk_get_aio_context(blk),
-            atomic_mb_read(&blk->in_flight) > 0);
+    AIO_WAIT_WHILE(blk_get_aio_context(blk),
+                   atomic_mb_read(&blk->in_flight) > 0);
 
     if (bs) {
         bdrv_drained_end(bs);
@@ -1620,8 +1618,7 @@ void blk_drain_all(void)
         aio_context_acquire(ctx);
 
         /* We may have -ENOMEDIUM completions in flight */
-        AIO_WAIT_WHILE(&blk->wait, ctx,
-                atomic_mb_read(&blk->in_flight) > 0);
+        AIO_WAIT_WHILE(ctx, atomic_mb_read(&blk->in_flight) > 0);
 
         aio_context_release(ctx);
     }
diff --git a/block/io.c b/block/io.c
index 8b81ff3913..bd9d688f8b 100644
--- a/block/io.c
+++ b/block/io.c
@@ -38,8 +38,6 @@
 /* Maximum bounce buffer for copy-on-read and write zeroes, in bytes */
 #define MAX_BOUNCE_BUFFER (32768 << BDRV_SECTOR_BITS)
 
-static AioWait drain_all_aio_wait;
-
 static void bdrv_parent_cb_resize(BlockDriverState *bs);
 static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs,
     int64_t offset, int bytes, BdrvRequestFlags flags);
@@ -557,7 +555,7 @@ void bdrv_drain_all_begin(void)
     }
 
     /* Now poll the in-flight requests */
-    AIO_WAIT_WHILE(&drain_all_aio_wait, NULL, bdrv_drain_all_poll());
+    AIO_WAIT_WHILE(NULL, bdrv_drain_all_poll());
 
     while ((bs = bdrv_next_all_states(bs))) {
         bdrv_drain_assert_idle(bs);
@@ -713,8 +711,7 @@ void bdrv_inc_in_flight(BlockDriverState *bs)
 
 void bdrv_wakeup(BlockDriverState *bs)
 {
-    aio_wait_kick(bdrv_get_aio_wait(bs));
-    aio_wait_kick(&drain_all_aio_wait);
+    aio_wait_kick();
 }
 
 void bdrv_dec_in_flight(BlockDriverState *bs)