VXA-320 FIRMWARE HISTORY ------------------------ Code Revision: v13233 Date: 03/10/09 Changes made to firmware since code revision V13231: Enhancement: Added Extended Diagnostics test. Enhancement: Final test is enhanced for ICR process Code Revision: v13231 Date: 09/12/08 Changes made to firmware since code revision V13222: Issue: Fix emergency eject when in media preservation mode with prevent media removal set. Solution: Fixed in Code. Issue: Power On hours can sometimes indicate an invalid number if two tasks attempt to update the power on time at exactly the same moment. Protection has been added to the code to prevent this problem. Solution: Fixed in code. Enhancement: Improvement of Hard FSC handling (1) Put POST failures, including failing Test ID, in Hard FSC Table (2) Add a count field to the Hard FSC Table entries. If a new entry is requested with the same FSC and the same media, just increment the count and do not create a new entry. Enhancement: Limit SCSI FILL command length (and lengths of other commands that access the R/W Buffers) to the size of those buffers to avoid splashing memory. Enhancement: Remove the check for the "Dead Bit" (Service Required Bit) prior to running the IST on a generic drive. Enhancement: IST abort via Eject button: (1) Ensure that the Batch Active flag is cleared so that the Eject is properly processed; (2) Ensure that the LED state is set to match the machine state after the Eject aborts the IST Issue: IST will timeout during load when X23 is loaded from PEOT. Solution: Fixed in code. Code Revision: v13222 Date: 06/02/08 Changes made to firmware since code revision V13221: Issue: IST (Internal Self Test) reported false Pass mode. Solution: Fixed in Code. Code Revision: v13221 Date: 02/11/08 Changes made to firmware since code revision V13220: Enhancement: Added Emergency Eject / Reset feature. This feature will override all normal drive operations and cause the tape drive to eject any data cartridge loaded in the drive. The tape drive will perform an Emergency Eject/Reset when the eject button is pressed and held for a minimum period of 20 seconds. Emergency Eject will override "Prevent Medium Removal" and not inhibit the data cartridge from subsequent use. Code Revision: V13220 Date: 10/10/2007 Changes made to firmware since code revision V13212: Issue: New tape pack tension can affect the ability to correctly position when coming out of Media Preservation mode. This mispositioning will result in a LBA not found condition when reading or a splice not found condition when writing. Solution: Fixed in code. Issue: YoYo during read/write search retry selectively being engaged Solution: This change simply retains whether a read issue was ever experienced and if so, ensures that the YoYo retry is employed as part of the handling on subsequent retry attempts. Issue: Logs sometimes contain an enormous number of entries for Test Unit Read commands. These were to have been filtered out so as not to ruin logs if the host is constantly polling. Solution: Change made to polling command counter. Issue: Delayed EOD gap write can result in buffer empties being written instead Solution: Made change to commanding the EOD gap prior to the final flush. Enhancements: Added high level write retry. Improvements made handling uncorrectable/permanent read error retries Provided the ability to perform an emergency eject via the drives eject button. -Press and hold the unload button for >20 seconds and then release it to force an emergency unload operation (unless a hardware error has occurred).LED's 1,2, and 3 will flash amber rapidly and LED 4 will be solid amber when the emergency unload operation is in progress. Improvements made to Sony media handling. Add machine type identifiers to Inqury String for new interfaces Code Revision: V13212 Date: 7/31/2007 Changes made to firmware since code revision V13208: Enhancement: Enhancing IST. IST Results to NVRAM & displayed in LEDs. Moving FRU definition to the test level. Also added further retries to combat bad test tape. IST run from library interface now checks the "dead bit" and will fail IST if it is set. Enhancement: Improved system directory write retry scheme Issue: After a reset where the initiator did not negotiate, the drive was not properly checking disconnect privilege and then disconnecting in a case where it should not have. Solution: Correct the disconnect privilege check. Issue: Hanging the bus due to the drive being disconnected while an emergency unload was initiated. Solution: Terminate SCSI operations before doing emergency re-boot to avoid bus hang Issue: Write delay timeout clashing with write pause timeout handling. Solution: Fixed in code. Issue: Log filled with TUR entries. When host issues constant Test Unit Readys, the log fills with them. Solution: Suppress Test Unit Ready command logging after a few to prevent the log from filling when TUR continuously issued. Issue: Servo unable to position to target tape count due to recorded region feedback Solution: Fixed in code. Issue: When cartridge left in loaded state for a long time servo deadman timer 1 asserted. Solution: Now this timer functions if tape path is in a position where the tape would be touching the drum. Issue: Fix unflushed immediate mark, introduced in v3206. Solution: Fixed in code. Issue: Threshold for broken tape too stringent, causing false broken tape indications. Solution: Relaxed threshold for broken tape in tape loop manager. Enhancement: Prevent V6/V10 media from being loaded and utilized. Enhancement: Make tool log summarizer available to tech support via the TRCLOG DLL. Added operations trace to tech support log summarization interface. Enhancement: Improved media handling. Enhancement: Enhanced detection/notification of worn/spent system directory media region. Enhancement: Improved handling of tape loops at EOT. Enhancement: Further refinement of FSC's used for servo failures to assist in FA of field service failures. Enhancement: Ensure tape ejected during load if logical format load attempt unable to be performed. Enhancement: Provide for PEOT media eject ability from media preservation mode. Issue: Recoverable/Unrecoverable media error indication in LED's lost if cleaning is required. Following the insertion of a piece of V media into a drive which is currently requiring cleaning, the cleaning indication overrides the posted unrecoverable error indication (amber, off, amber). Solution: Resetting the LED heartbeat when an unrecoverable/recoverable error is posted. Issue: Previously written tape left loaded in drive after a write buffer command fails with FSC 65 and is unreadable. Solution: Fixed in code, ensuring that dirty format gets a chance to run if we deem the media to have a corrupted format. --------------------------------------------------------------------------------- Code Revision: V13208 Date: 6/13/2007 Enhancement: Implemented drive Internal Self Test (IST). Test is started using eject button with results displayed in LEDs. Added IST command to library interface; for starting drive IST and getting last IST ending status. Enhancement: Add Single Ended SCSI connector test on second press of the eject button with no tape loaded. Results are all green (not SE) or all yellow (SE attached). Code Revision: V13206 Date: 11/29/2006 Changes made to firmware since code revision V13205: Enhancement: Enhanced media handling. Randomized timeout of automatic move away from BOT. Enhancement: This change is designed to enhance the ability of a drive to detect that media has been written before. Enhancement: Improved tape load mechanism applied voltage during retries. Issue: Servo error shutdowns can cause SCSI to experience an internal error, forcing a SCSI Bus reset. Solution: Fixed in code. Issue: A SCSI Reset issued to a drive when it is loaded with a blank tape may cause the mode to change to VXA1 mode. Solution: Fixed in code. Issue: LED can report incorrectly following dirty format recovery from Power on reset. Solution: LED indicate correct status. Enhancement: This change adds support for tape mapping. A map of tape region usage/failures is now being maintained in memory and in the tape system directory region. Implement media utilization management. Enhancement: This change corrects issue with random load positioning timer. Enhancement: Improved handling of General Read Error. Issue: On a load failure the original sense set up by the Load module is being overwritten by a less informative Load Fail status. Solution: Create new sense for a previous-load-failure situation only if there is not already sense set up by the load module. Enhancement: Improved tape tensioning. Enhancement: Improvement to Post system directory write validation. We will now wait to ensure that the VXA1 BOD's are intact following complete directory write completion. Issue: In some cases the cleaning light can come on too soon. Solution: Fixed in code Enhancement: This change will ensure that if a write command makes it into DPM at the point of write pause timeout and the ensuing cap off fails, the write command will be handled (discarded) appropriately. Enhancement: Media preservation after long period of inactivity. When we detect that the host has been inactive for a long period of time, we will now move out to end of tape and unload the tape path. Internally, this is referred to as taking the media off-line. On a subsequent host request, we will bring the media back on-line (returning it to its origin) prior to handling the request. Changing 10 minute timeout to 30 minutes. Issue: Code Revision: V13205 Date: 9/1/2006 Changes made to firmware since code revision V13204: Issue: Following the logical load, we would position to the same position every time. Over time, this would produce tape damage. Solution: Randomize post-load pause/stop position. Issue: Write abort can cause the SCSI chip to error out a with FIFO under-run. Solution: Fixed related code. Issue: A write failure shutdown can cause the SCSI chip to hangup. Solution: Fixed related code. Enhancement: Change cleaning frequency from 75 to 25 hours and cleaning tape duration to 10 seconds from 5. Enhancement: Improved handling of expired cleaning cartridge. Code Revision: V13204 Date: 8/1/2006 Changes made to firmware since code revision V13203: Issue: In rare cases the SCSI code may not disconnect when the drive goes into long write recovery retries. This 'no disconnect' for very long periods of time might cause some OSs to issue a SCSI bus reset. Solution: Fixed related code. Issue: After a servo shutdown the SCSI sequencer could lockup, forcing a SCSI bus reset recovery. Solution: Fixed related code. Code Revision: V13203 Date: 7/13/2006 Changes made to firmware since code revision V13200: Issue: False check condition and sense data generated when command was queued waiting for a load command. Solution: Fixed in code. Code Revision: V13200 Date: 2/28/06 Changes made to firmware since code revision V1310E: ***Note:*** Once 3200 code or higher is in the machine, it must stay at 3200 or above. New Read Channel Variable Support. Once loaded, old code can’t be reloaded. Issue: Some HBAs are not performing complete negotiation sequences when bringing up SCSI devices. Solution: Fixed related code. Issue: Drive could incorrectly report a broken tape when 1)cartridge is inserted into a drive and then pulled out while the mechanism is attempting to load the cartridge, 2) If a tape loop occurs at EOT, 3) during emergency unload operation. Solution: Fixed in related code. Issue: Drive could report a general hardware error due to data being in the buffer when we try to format on-the-fly. Solution: Fixed in code. Issue: SCSI Abort causes hang. A SCSI Abort message during dirty format recovery results in a machine which hangs the SCSI bus and will not communicate serially. Solution: Fixed in code. Issue: Remaining tape capacity was calculated based on where the media was physically located - not logically located. Solution: Fixed in code Issue: Occasional high rewrites due to loose index window. Solution: Fixed in code. Issue: Drive could experience SCSI timeout. FSC 50 with check condition SK 0B aborted command Solution: Fixed in code ---------------------------------------- Code Revision: V1310E Date: 31 Jan 2006 Changes made to firmware since code revision V13109: Issue: Drive could incorrectly report a servo error, FSC 5A. Solution: Fixed related code. Issue: On some SCSI controllers, the bus may hang during write. This is due to an issue with the controller handling an extended delay at the end of the Data Out phase before the drive switches to the Message In phase. Solution: The drive will disconnect immediately at the end of the Data Out phase. Enhancement: Added the Lock Write Density (LWD) option. With LWD turned on, Mode Select commands to set density to VXA-2 format will be ignored, and the drive will only write data from beginning of tape in VXA-3 format. If appending to a VXA-2 format tape, the appended data will remain in VXA-2 format. VXATool can be used to set LWD on or off. ------------------------------------------------------------------------ Code Revision: V13109 Date: 6 Dec 2005 Changes made to firmware since code revision V13106: Issue: A format recovery operation could hang and never complete. Solution: Fixed related code to complete the format recovery operation. Note: A format recovery operation occurs if the drive is powered off while writing data to tape. Issue: After tape motion and the drive scanner have been stopped due to inactivity, the next Read command may report a failure. Solution: Fixed related code. Issue: Read Block Limits command reported a minimum blocksize of 1 byte. Solution: Read Block Limits command reports a minimum blocksize of 4 bytes. Issue: A Long Erase command may stop and report good status without erasing the entire tape. Solution: Fixed related code. Issue: After sending a Prevent Media Removal followed by an Unload command, the drive reports incorrect sense data on following commands. Reported sense data is sense key 02, asc/ascq 04/01 (drive becoming ready) FSC = 4B. Solution: After sending a Prevent Media Removal followed by an Unload command, the drive reports sense key 02, asc/ascq 04/00 (drive not ready) FSC = 07. Issue: Drive could incorrectly report a broken tape condition. Solution: Fixed related code. Enhancement: Improved recovery during write and read operations to reduce the risk of media errors. Enhancement: Improved tape handling. Enhancement: The drive's firmware version will not be reported in the SCSI INQUIRY Device Identification Page (page code 83h) Bytes 22-29. These bytes will contain ASCII blanks (20h). Enhancement: This change affects a drive installed in the PacketLoader 1x10 1U library. Do not continue to display drive status as "writing" or "reading" after the drive has completed the write or read operation. Enhancement: This change only affects drives installed in a PacketLoader 1x10 1U library. The handling of incompatible media types has been changed. Drives installed in a library will not automatically eject an incompatible tape. Tape will remain in the drive until a command is received to unload it. Stand-alone drives (drives not installed in a library) will continue to automatically eject an incompatible media type. Incompatible media types are VXATapes in V6, V10, and V17 lengths. ------------------------------------------------------------------------ Code Revision: V13106 Date: 15 Sept 2005 Firmware V13106 is the initial code release for VXA-320.