[PATCH 0/2] Block I/O while SG reset operation in progress

[PATCH 0/2] Block I/O while SG reset operation in progress

[James Smart]

As promised in a previous thread:
http://marc.theaimsgroup.com/?l=linux-scsi&m=113321277408600&w=2

The midlayer today allows for calls to the eh handlers via sg to occur
while i/o continues to be sent to the drive run abated.

This patch ensures that i/o is stopped while an eh handler is being
processed. It adds a new flag, set by the async reset callers, which
augments the host-in-reset checks and stops i/o.  The async reset
callers are already synchronized to hold off until the error thread is
no longer running.

There are 2 parts to this patch:
  - the patch to the midlayer to stop i/o while resetting
  - a patch to the lpfc driver which removes the barrier the driver
    had implemented in the past.

-- james s