* All tests: runs if SMART=CS/OVERRIDE & BADBLOCKS=CS
* All drive tests: runs if SMART=CS/OVERRIDE & BADBLOCKS=CS
* I/O tests are readonly, blocksize=4M, and limited to 16 Gb
* Fix issue #23
* Renamed run_smart to run_nvme_smart to match rest of the code
* Mark NVMe/SMART status as NS earlier in a couple more cases
* Prevent crash when all tests are run and then aborted during Prime95
* The script waits for the self-test for polling-time+5min
* Unless it detects completion after 60 seconds
* Fixed missing linebreaks in progress panes
* BREAKING: TESTS data structure reworked to support per-dev status
* Renamed 'Health OK' to 'Quick Health OK'
* Better implies that we're not done testing
* Renamed SMART sections to NVMe/SMART
* Dropping SMART short-tests for NVMe drives
* Replaced smartmontools with smartmontool-svn
* Adds JSON exporting which makes parsing the data easier
* Using nvme-cli for NVMe drives instead of smartctl
* NVMe support in smartctl is still experimental and inconsistent
* New "Health OK" detection
* This is used to determine if a disk is okay to be tested
* e.g. an IDE disk w/out any SMART data now requires explicit overriding
* NVMe disks are set using the "critical_warning" attribute
* SMART disks are set using the "overall-health" attribute
