Firmware Update: LSI Logic MegaRAID SCSI 320-X

LSI Logic has released new firmware for its MegaRAID SCSI 320-0X, 320-2X, and 320-4X RAID adapters. A number of critical firmware bugs prevented the delivery of this new generation of PCI-X SCSI RAID adapters. The showstoppers are now fixed. In the readme will be notified of the following changes:

Bug Fixes from Previous versions of Firmware

• 1. Rebuild does not start when inserting a replacement drive.
• 2. FW is returning Non-Coerced Drive Size on READ_CAP.
• 3. System freezing during drive failure in Raid50.
• 4. After creating multiple LDs, go into the properties of any LD except the first one. Select “delete” and click on “Go”. Then it asks “All data will be destroyed for Logical Drive Number: 1 Are you sure?” This number is zero-based where as the rest of Web BIOS the numbering starts from “1”. (WebBIOS displays wrong logical drive number on this screen, even though it deletes correct logical drive)
• 5. Multiple drive pull and reinsertions in Linux causes system to hang or crash.
• 6. Data Corruption using 128k WB DIO R5/ R0.
• 7. 2TB limit is imposed in new drive creation but not in Reconstruction.
• 8. In WebBIOS (1.78) on 0x controller the maximum SCSI capability is 160M.
• 9. Setting factory defaults sets the SCSI capability to FAST.
• 10. Duplicate Message Free
• 11. IO Stuck in Wait Queue
• 12. User is able to Create/Delete LD while CC/GBI is active.
• 13. System hangs on installing AS2.1/Windows.
• 14. If a media error is injected on a physical drive in a redundant array, the media error count remained even after the drive was removed and replaced with a new drive to rebuild (without media errors).
• 15. When using a consistent drive (BGI complete), and starting a check consistency from GAM then canceling the check consistency, the button stays grayed out, so user cannot start CC again.
• 16. Hot spare dead after canceling an auto-rebuild. If rebuild process is canceled on a hot spare, it is not possible to restart the rebuild because the hot spare has now in a Dead status.
• 17. Physical rebuild does not start in GAM when a drive is FAILED and then selected for manual rebuild.
• 18. WebBIOS did not show Check Consistency ERROR flag when a fatal media error was encountered.
• 19. Canceling check consistency started a background initialization even if the logical drive is consistent.
• 20. After a foreground initialization, it was not possible to start a check consistency (gray field).
• 21. Fujitsu drives show INVALID CDB in field message in GAM log.
• 22. GAM logs the media errors in error table only. Should be copied to the NVRAM log.

Enhancements

• 1. Adjust Patrol Read timing parameters. Tuned Patrol Read parameters changed the idle timer from 300 to 120 seconds and Increased the maximum permissible host commands from 1 to 2.
• 2. Added new drive failure reason. A new Drive failure reason is added, when the user aborts rebuild