RELEASE NOTES FOR version hpsa.6.0.0.124-1.01 of the hpsa driver for HPE Smart Array controllers


Make all code paths that send device resets use a consistent method for preventing multiple simultaneous resets, and for marking IO requests that will be affected by the reset.
 
If a device reset operation is in progress, delay any subsequently requested device rescan or discovery operations for few seconds, until reset operation in complete.

Before returning from a device reset operation, verify that device can accept and respond to new requests by issuing a test unit ready (TUR) command until the device responds positively.

Fix an error in code path that handles completions of IO requests that were in flight when a device reset operation occurred. Some requests failed to get completed, causing device to be taken offline by the operating system.

Correct LBA calculation for SCSI read(6) and write(6).

Replace expose_state flag with less complicated boolean expose_device, and change dependent code sections. This flag defines whether a device is exposed to the SCSI layer.

Generate a controller-level NMI during deaklock: a new method to assist in diagnosing controller lockups in case of a deadlocked firmware thread.

Treat failed Test Unit Ready (TUR) as offline device. Function hpsa_volume_offline was incorrectly returning (not offline) if test unit ready command fails.

Fix problems preventing certain failed volumes from being taken offline after a hot unplug event. Addresses VSAN hotplug certification issue.


