[v4,0/2] improve quiesce time for large amount of namespaces

Message

Sagi Grimberg July 27, 2020, 10:07 p.m. UTC

This set improves the quiesce time when using a large set of
namespaces, which also improves I/O failover time in a multipath environment.

We improve for both non-blocking hctxs (e.g. pci, fc, rdma nvme transports)
and blocking hctxs (e.g. tcp nvme transport) by splitting queue quiesce to
blk_mq_quiesce_queue_async call_(s)rcu and blk_mq_quiesce_queue_async_wait
to wait for it to complete. These calls are meant to be called in parallel
for different queues.

Changes from v3:
- make hctx->rcu_sync dynamically allocated from the heap instead
  of a static member function

Changes from v2:
- made blk_mq_quiesce_queue_async operate on both blocking and
  non-blocking hctxs.
- removed separation between blocking vs. non-blocking queues
- dropeed patch from Chao
- dropped nvme-rdma test patch

Changes from v1:
- trivial typo fixes

Sagi Grimberg (2):
  blk-mq: add async quiesce interface
  nvme: improve quiesce time for large amount of namespaces

 block/blk-mq.c           | 46 ++++++++++++++++++++++++++++++++++++++++
 drivers/nvme/host/core.c |  4 +++-
 include/linux/blk-mq.h   |  4 ++++
 3 files changed, 53 insertions(+), 1 deletion(-)