On Tue, Jul 04, 2017 at 10:55:04AM +0300, Sagi Grimberg wrote:

Before we either iterate on tags or cleanup the request queue
we must guarantee that the hw queues are stop and no inflight
.queue_rq is active. Thats what blk_mq_quiesce_queue is for, so
use it where appropriate.

queue freezing is used in cleanup path, and not required to
quiesce queue.

quiesce is required for canceling request via blk_mq_tagset_busy_iter()
for avoiding double release.

I think we should make it clear in comment log.

Sagi Grimberg (8):
  nvme-rdma: quiesce/unquiesce admin_q instead of start/stop its hw
    queues
  nvme-fc: quiesce/unquiesce admin_q instead of start/stop its hw queues
  nvme-loop: quiesce/unquiesce admin_q instead of start/stop its hw
    queues
  nvme-pci: quiesce/unquiesce admin_q instead of start/stop its hw
    queues
  nbd: quiesce request queues to make sure no submissions are inflight
  mtip32xx: quiesce request queues to make sure no submissions are
    inflight
  virtio_blk: quiesce/unquiesce live IO when entering PM states
  xen-blockfront: quiesce IO before device removal

 drivers/block/mtip32xx/mtip32xx.c |  8 ++++----
 drivers/block/nbd.c               |  4 ++--
 drivers/block/virtio_blk.c        |  4 ++--
 drivers/block/xen-blkfront.c      |  8 ++++----
 drivers/nvme/host/fc.c            |  8 +++++---
 drivers/nvme/host/pci.c           | 10 ++++++----
 drivers/nvme/host/rdma.c          |  7 ++++---
 drivers/nvme/target/loop.c        |  2 +-
 8 files changed, 28 insertions(+), 23 deletions(-)

-- 
2.7.4

-- 
Ming