[PATCH v3 2/2] nvme: improve quiesce time for large amount of namespaces

[Sagi Grimberg]

nvme currently will synchronize queue quiesce for each namespace at once.
This can slow down failover time (which first quiesce all ns queues) if we
have a large amount of namespaces. Instead, we want to use an async interface
and do the namespaces quiesce in parallel rather than serially.

Signed-off-by: Sagi Grimberg <sagi at grimberg.me>
---
 drivers/nvme/host/core.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index c16bfdff2953..45f1559ee160 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -4535,7 +4535,9 @@ void nvme_stop_queues(struct nvme_ctrl *ctrl)
 
 	down_read(&ctrl->namespaces_rwsem);
 	list_for_each_entry(ns, &ctrl->namespaces, list)
-		blk_mq_quiesce_queue(ns->queue);
+		blk_mq_quiesce_queue_async(ns->queue);
+	list_for_each_entry(ns, &ctrl->namespaces, list)
+		blk_mq_quiesce_queue_async_wait(ns->queue);
 	up_read(&ctrl->namespaces_rwsem);
 }
 EXPORT_SYMBOL_GPL(nvme_stop_queues);
-- 
2.25.1