This is another alpha release mostly targeted at some openSeaChest_ZBD improvements.
- Updated error messages to include "or not supported" since on SATA we may not get the ATA sense data that is more descriptive about the actual error condition. This is an ATA limitation as without sense data it is not possible to know the difference between "Command not supported" and "Invalid field in command"
NOTE: SAT-6 helps with this significantly on ATA passthrough. We still need to implement the changes that clarify what the SATL does for us, but this will largely resolve this issue in the future. This work is not yet started.
Other changes in here:
- Fix FARM interpretation of Helium pressure statistic. There are 2 formats depending on which version of FARM and the code was only supporting one. This means that on any newer helium filled disk it will show as "tripped" even when it is within the limit/threshold due to the data formatting change in firmware.
- Fixed resetting "all" mode pages. If giving 3Fh-FFh, 3Fh, or xxh-FFh mode page numbers, this was outputting "failed" in the v26.03.2 release. This was due to a change introduced to catch bad/garbage return data from some devices and this broke all of these cases when we implemented this. This preserves that check for individual mode pages and allows these page codes to continue without validation. Note that the lack of validation on these "return all" pages is unlikely to be a major issue as devices not properly handling mode page requests do not usually implement this request. An update is planned to add some kind of validation in the future.
- Fixed passing SOURCE_DATE_EPOCH into VMActions builds
- Added a couple more OS's from VMActions to the CI and releases (Tribblix, openIndiana) with fixes for the stack protection flags used during compilation.
- Changed openSeaChest_PassthroughTest default SCSI command tests. This reduces how many things are tested as when this is used on USB many of the tests don't pass and output "this is not supported" anyways. This reduction means the test is more likely to get to the ATA/NVMe passthrough portion before it enters a bad state. A new option allows specifying levels of SCSI command testing from 1-5 to refine the test output data. It is recommended to repeat at the highest level possible when providing results for us to integrate.