[PATCHv4 0/4] nvme timeout updates

[keith.busch@intel.com (Keith Busch)]

The majority of the v3 was either merged or no longer necessary.

Patch 1 is still the same as before.

The remaining patches 2-4 are setting up the nvme core to handle
unfreezing queues and dispatching IO from a context that is safe to
handle IO timeouts. The pci nvme is the only one making use of that at
the momement, but I suspect fabrics can make something of it.

The patch retrying a failed reset on an admin init timeout has been
dropped. If we really need to do this, I think more discussion will help
so we are creating less arbitrary rules.

Keith Busch (4):
  nvme: Sync request queues on reset
  nvme: Start controller in own work queue
  nvme: Introduce frozen controller state
  nvme-pci: Use controller start work to dispath IO

 drivers/nvme/host/core.c | 54 ++++++++++++++++++++++++++++++++++++++++++++----
 drivers/nvme/host/nvme.h |  3 +++
 drivers/nvme/host/pci.c  | 20 ++++++------------
 3 files changed, 59 insertions(+), 18 deletions(-)

-- 
2.14.3