After reviewing the error_recovery paths to address the abort on the
io to cover timeouts on connecting ios, it became clear that some of the
thing in the way it was coded were unnecessary.

This patch set cleans up those items and refactors. It then addresses the
failure case of io errors during reconnect which wants to take the create
failure cases, but there is no status back from the core routines to
trigger that path.

James Smart (4):
  nvme-fc: remove err_work work item
  nvme-fc: eliminate terminate_io use by nvme_fc_error_recovery
  nvme-fc: remove nvme_fc_terminate_io()
  nvme_fc: track error_recovery while connecting

 drivers/nvme/host/fc.c | 270 +++++++++++++++++------------------------
 1 file changed, 109 insertions(+), 161 deletions(-)

-- 
2.26.2