nbd: Grab aio context lock in more places

When iothreads are in use, the failure to grab the aio context results
in an assertion failure when trying to unlock things during blk_unref,
when trying to unlock a mutex that was not locked.  In short, all
calls to nbd_export_put need to done while within the correct aio
context.  But since nbd_export_put can recursively reach itself via
nbd_export_close, and recursively grabbing the context would deadlock,
we can't do the context grab directly in those functions, but must do
so in their callers.

Hoist the use of the correct aio_context from nbd_export_new() to its
caller qmp_nbd_server_add().  Then tweak qmp_nbd_server_remove(),
nbd_eject_notifier(), and nbd_esport_close_all() to grab the right
context, so that all callers during qemu now own the context before
nbd_export_put() can call blk_unref().

Remaining uses in qemu-nbd don't matter (since that use case does not
support iothreads).

Suggested-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20190917023917.32226-1-eblake@redhat.com>
Reviewed-by: Sergio Lopez <slp@redhat.com>

1 files changed, 18 insertions, 4 deletions

diff --git a/nbd/server.c b/nbd/server.c
index 31d624e146..d8d1e62455 100644
--- a/nbd/server.c
+++ b/nbd/server.c
@@ -1461,7 +1461,12 @@ static void blk_aio_detach(void *opaque)
 static void nbd_eject_notifier(Notifier *n, void *data)
 {
     NBDExport *exp = container_of(n, NBDExport, eject_notifier);
+    AioContext *aio_context;
+
+    aio_context = exp->ctx;
+    aio_context_acquire(aio_context);
     nbd_export_close(exp);
+    aio_context_release(aio_context);
 }
 
 NBDExport *nbd_export_new(BlockDriverState *bs, uint64_t dev_offset,
@@ -1480,12 +1485,11 @@ NBDExport *nbd_export_new(BlockDriverState *bs, uint64_t dev_offset,
      * NBD exports are used for non-shared storage migration.  Make sure
      * that BDRV_O_INACTIVE is cleared and the image is ready for write
      * access since the export could be available before migration handover.
+     * ctx was acquired in the caller.
      */
     assert(name);
     ctx = bdrv_get_aio_context(bs);
-    aio_context_acquire(ctx);
     bdrv_invalidate_cache(bs, NULL);
-    aio_context_release(ctx);
 
     /* Don't allow resize while the NBD server is running, otherwise we don't
      * care what happens with the node. */
@@ -1493,7 +1497,7 @@ NBDExport *nbd_export_new(BlockDriverState *bs, uint64_t dev_offset,
     if (!readonly) {
         perm |= BLK_PERM_WRITE;
     }
-    blk = blk_new(bdrv_get_aio_context(bs), perm,
+    blk = blk_new(ctx, perm,
                   BLK_PERM_CONSISTENT_READ | BLK_PERM_WRITE_UNCHANGED |
                   BLK_PERM_WRITE | BLK_PERM_GRAPH_MOD);
     ret = blk_insert_bs(blk, bs, errp);
@@ -1560,7 +1564,7 @@ NBDExport *nbd_export_new(BlockDriverState *bs, uint64_t dev_offset,
     }
 
     exp->close = close;
-    exp->ctx = blk_get_aio_context(blk);
+    exp->ctx = ctx;
     blk_add_aio_context_notifier(blk, blk_aio_attached, blk_aio_detach, exp);
 
     if (on_eject_blk) {
@@ -1593,6 +1597,12 @@ NBDExport *nbd_export_find(const char *name)
     return NULL;
 }
 
+AioContext *
+nbd_export_aio_context(NBDExport *exp)
+{
+    return exp->ctx;
+}
+
 void nbd_export_close(NBDExport *exp)
 {
     NBDClient *client, *next;
@@ -1687,9 +1697,13 @@ BlockBackend *nbd_export_get_blockdev(NBDExport *exp)
 void nbd_export_close_all(void)
 {
     NBDExport *exp, *next;
+    AioContext *aio_context;
 
     QTAILQ_FOREACH_SAFE(exp, &exports, next, next) {
+        aio_context = exp->ctx;
+        aio_context_acquire(aio_context);
         nbd_export_close(exp);
+        aio_context_release(aio_context);
     }
 }