diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c
index 88b5d6d67e6e..31641d95aa4d 100644
--- a/net/sunrpc/svc.c
+++ b/net/sunrpc/svc.c
@@ -402,6 +402,7 @@ struct svc_pool *svc_pool_for_cpu(struct svc_serv *serv)
 	struct svc_pool_map *m = &svc_pool_map;
 	int cpu = raw_smp_processor_id();
 	unsigned int pidx = 0;
+	unsigned int i;
 
 	if (serv->sv_nrpools <= 1)
 		return serv->sv_pools;
@@ -414,8 +415,34 @@ struct svc_pool *svc_pool_for_cpu(struct svc_serv *serv)
 		pidx = m->to_pool[cpu_to_node(cpu)];
 		break;
 	}
+	pidx %= serv->sv_nrpools;
 
-	return &serv->sv_pools[pidx % serv->sv_nrpools];
+	/*
+	 * It's possible to have a pool with no threads. Userland can just set
+	 * things up this way directly. Also, when threads are autodistributed
+	 * they are spread evenly across the pools, but when there are fewer
+	 * threads than pools some pools can end up with none.
+	 *
+	 * A transport enqueued on a threadless pool would never be picked up,
+	 * since each thread only services its own pool. Fall back to the next
+	 * populated pool, trading NUMA locality for a guarantee that the
+	 * transport is serviced.
+	 */
+	for (i = 0; i < serv->sv_nrpools; i++) {
+		struct svc_pool *pool = &serv->sv_pools[pidx];
+
+		/* This is set under the service mutex and rarely ever
+		 * changes. A data race here is harmless.
+		 */
+		if (data_race(pool->sp_nrthreads))
+			return pool;
+
+		if (++pidx >= serv->sv_nrpools)
+			pidx = 0;
+	}
+
+	/* No pool has any threads; nothing can service the transport. */
+	return &serv->sv_pools[pidx];
 }
 
 static int svc_rpcb_setup(struct svc_serv *serv, struct net *net)