Lenovo YUM Repository
Release and Change History








Bundle Release Information:
Targeted Server Family (Machine Type): Flex System x240 (7162, 2588)
Targeted Operating System: RHEL6.8
YUM Repository Build Date: 2019_04_23

YUM Repository Release Version: 19A

Device:
IBM Flex System Storage Expansion Node Part Number: 68Y8588
Feature Code: na
Core System Device
Install Drivers on all Servers


Release History:
Release: 18B
Continued Support in 19A - Latest Release Change History Documents
Firmware Firmware Name Version
ServeRAID 6gb SAS/SATA Controller Firmware Update 6gb-23.34.0-0023
Drivers Driver Name Discreate Package Download Version
megaraid_sas
  kernel driver
kmod-megaraid_sas- 07.705.04.00_el6.8-1.x86_64.rpm 07.705.04.00_el6.8

Release: 17B
Firmware Firmware Name Version
ServeRAID 6gb SAS/SATA Controller Firmware Update 6gb-23.34.0-0023
Drivers Driver Name Discreate Package Download Version
megaraid_sas
  kernel driver
kmod-megaraid_sas- 07.702.06.00_el6.8-1.x86_64.rpm 07.702.06.00_el6.8


Release 18B Change History Documents

ServeRAID 6gb SAS/SATA Controller Firmware Update

´╗┐

ServeRAID 6GB SAS/SATA Controller Firmware Update

Adapters Supported:  ServeRAID M5115 SAS/SATA Controller (90Y4390)
		     Lenovo Storage Expansion Node (65Y8588)
		     ServeRAID M5110 SAS/SATA Controller for Lenovo System x (81Y4481)
		     ServeRAID M5110e SAS/SATA Controller for Lenovo System x 
		     ServeRAID M5120 SAS/SATA Controller for Lenovo System x (81Y4478)

NOTE TO SERVICE - Reference RETAIN #N/A


Version 23.34.0-0023 - High Impact/High Probability of Occurrence
--------------------
FW PACKAGE: 23.34.0-0023
MR FW:  3.460.145-8209
iMR FW:  3.460.144-8208
BIOS:  5.50.03.0
UEFI:  0x06110200
HII:  03.17.14.04

Fixes:
- Fixed an issue where malformed incoming host commands could cause card reset in DataCenter. (SCGCQ01240574)
- Fixed an issue where iMR Montask while installing on degraded R5. (SCGCQ01165519)
- Fixed an issue where MRS section in the DDF is not updated when a drive is moved from rebuilding to online state. (SCGCQ01255906)
- Fixed an issue where IN FAILED ARRAY bit is not being set when a VD is cleared. (SCGCQ01287457)
- Fixed an issue where cache windows of physical drives with invalid signatures were not being accounted properly which was lead to an infinite loop and ultimately kill adapter. (SCGCQ01260339 Port of Defect SCGCQ01058771)
- Fixed an issue where Users can't change default access policy of SVD when it blocked by SSC removal. (SCGCQ01260344 Port of Defect SCGCQ00814937)


Version 23.34.0-0019 - High Impact/High Probability of Occurrence
--------------------
FW PACKAGE: 23.34.0-0019
MR FW:  3.460.115-6465
iMR FW:  3.460.114-6464
BIOS:  5.50.03.0
UEFI:  0x06110200
HII:  03.17.14.04

Fixes:
 - Fixed an issue where copyback was being set to on after any reboot or OCR. (SCGCQ01157490) 
 - Fixed an issue where copyback was not automatically initiated when manually setting drives to Unconfigured_Good (SCGCQ01181295)
 - Fixed an issue where foreign config import fails with DDF corruption (SCGCQ01172248)
 - Fixed an issue where firmware was not resetting learn cycle failed and learn cycle timeout fields when a new learn cycle starts and completes. (SCGCQ01174449)
 - Fixed an issue where system stall after sense key 04/c4/01 in HSP disk and controller reset repeatedly (SCGCQ01174450 Defect - SCGCQ01142402) 
 - Fixed an issue where data mis-compare could occur during recovery of CacheCade pinned windows across OCR/Reboot or CacheCade pinned windows could be discarded while CacheCade degrade flush in progress (SCGCQ01163950 Defect - SCGCQ01118285)
 - Fixed an issue where virtual drive erase wouldnĺt complete if VD ID number is changed while the task is ongoing. (SCGCQ01182611)


Version 23.34.0-0018 -  High Impact/High Probability of Occurrence
--------------------
FW PACKAGE: 23.34.0-0018
MR FW:  3.460.105-6456
iMR FW:  3.460.104-6455
BIOS:  5.50.03.0
UEFI:  0x06110200
HII:  03.17.14.04

Fixes:
 - Disabled further usage of PI due to a potential data issue (SCGCQ01128078, SCGCQ01105478)
 - Fixed an issue where MegaRAID firmware crashes or hangs on systems with ASPM enabled on the host and disabled on the controller. (SCGCQ01123931)
- Fixed an issue where PI disabled before WJ flush (SCGCQ01135207)
- Fixed an issue where virtual drive secure erase resets after a reboot and hangs at 0%. (SCGCQ01137940)
- Fixed an issue where after reboot imported 4K drive R0 was not visible. (SCGCQ01156533)
- Fixed an issue where patrol read on PD with medium errors could not be stopped by the stop command. (SCGCQ01137941)

Version 23.34.0-0016 -  High Impact/High Probability of Occurrence
--------------------
FW Package: 23.34.0-0016

HII:  
03.17.14.04
UEFI Driver: 0x06110200
BIOS: 5.50.03.0
MR FW: 3.460.85-6425
iMR FW: 

 3.460.84-6424

Fixes:
 - Fixed an issue where enabling directpdmapping causes attached disk drives to be no longer be detected by storcli.  (SCGCQ01063721)
 - Fixed an issue where storcli commands and MSM are very slow when migratation activties are occurring.  (SCGCQ01086623)
 - Fixed an issue where Lenovo Info from storcli commands for bbu mismatches with the firmware log.  (SCGCQ01041511)

Version 23.34.0-0009 -  High Impact/High Probability of Occurrence
--------------------
FW Package: 23.24.0-0009

HII: 03.17.14.02 

UEFI Driver: 0x06110200

BIOS: 5.50.03.0

MR FW: 3.460.35-6104

iMR FW: 3.460.34-6103

 

Fixes:

 - Fixed the potential of a data miscompare that was observed only at a vendor test facility under very heavy IO and abnormal practices (SCGCQ00995594, Port of SCGCQ00976904)

 - Fixed an issue where error "Controller encountered fatal error and was reset" may occur under heavy IO and background operations (SCGCQ00894643)

 - Fixed an issue where VMWare may crash when a Guest Operating System is powered off (SCGCQ00872121, Port of SCGCQ00854914)

 - Fixed an issue where write-verify performance was lower than expected (SCGCQ00858151, SCGCQ00874193)

 - Fixed an issue where "I2C 4 PID 10 SCL STUCK Low! while I2C Rx receiving data intStatus 00008000" may appear in firmware log (SCGCQ01007592)

 - Fixed an issue where the Measured Voltage instead of Charging Voltage was displayed in event log (SCGCQ00876489)

 - Fixed an issue where Supercap "Battery Not Present" error may occur after shutdown or reboot during a learn cycle (SCGCQ00917115)

 - Fixed an issue where Write-through and Write-back cache policies can be changed via Windows Logical Disk Properties. (SCGCQ00856979)

 - Fixed an issue where enclosures may go missing when connected in daisy-chain after multiple expander reset cycles (SCGCQ00840725, Port of SCGCQ00818569)

 - Fixed an issue where MegaRAID Storage Mananger may display an error message if 'Schedule Consistency Check' is selected. (SCGCQ00852649, Port of SCGCQ00832908)

 - Fixed an issue where Patrol Read Properties window does not pop up on MegaRAID Storage Manager (SCGCQ00862478, Port of SCGCQ00797676)

 - Fixed an issue where MegaRAID Storage Manager shows messages with Error code 240 and no description (SCGCQ00970948, Port of SCGCQ00904466)

 - Fixed an issue where Amber LED may turn on erroneously when an enclosure's slot numbers start with 0 but bay numbers start with 1 (SCGCQ01019622, Port of SCGCQ00903181)

 - Fixed an issue where the Generic RAID1 profile option in HII is not greyed-out after creating 64 Virtual Drives (SCGCQ00826822, Port of SCGCQ00816137)

 - Fixed an issue where firmware may hang until system is reset (SCGCQ00928869, Port of SCGCQ00923286)

 - Fixed an issue where a Virtual Drive may disappear on reboot if it was created with drives from a previously-erased Virtual Drive (SCGCQ00937912, Port of SCGCQ00900237)

 - Fixed an issue where Functional Level Reset was not supported (SCGCQ00854343, Port of SCGCQ00840253)

 - Fixed an issue where an "L2 unrecoverable cache error" may occur when modifying Patrol read properties (SCGCQ00854668, Port of SCGCQ00847162)

 - Fixed an issue where a Hot Spare disk may become Unconfigured Good on reboot (SCGCQ00862477, Port of SCGCQ00803341)

 - Fixed an issue where SafeID may change on repeated queries using StorCLI (SCGCQ00949550, Port of SCGCQ00941577)

 - Fixed an issue where error "MonTask: line 1272 in file ../../raid/1078dma.c" may occur while running IO in Virtual Machines (SCGCQ01016071, Port of SCGCQ00822403)

 - Fixed an issue where incorrect connector information "connector: null x0" shows in MegaRAID Storage Manager (SCGCQ01016040, Port of SCGCQ00996578)

 - Fixed an issue where a concurrent Consistency Check cannot be started after resetting adapter to factory defaults (SCGCQ00839437, Port of SCGCQ00838830)

 - Fixed an issue where JBOD mode may be disabled if a Virtual Drives is offline (SCGCQ00871662, Port of SCGCQ00827731)

 

Enhancements:

 - Added protection for clearing Foreign Configuration if preserved cache exists (SCGCQ00880783, Port of SCGCQ00842327)

 - Enhanced cache restore operation for when restore data size is different than cache size (SCGCQ00860303, Port of SCGCQ00855038)

Version 23.33.0-0044 -  High Impact/High Probability of Occurrence
--------------------
Fixes:
 - Fixed an issue where controller may reset under heavy IO and background operations - (SCGCQ00894643) (RETAIN Tip H001154)
 - Fixed an issue where SuperCap "Battery Not Present" after shutdown and reboot during a learn cycle - (SCGCQ00917115)
 - Fixed an issue where a VD created with drives from a previously erased VD may disappear on next reboot - (SCGCQ00937912, Port of SCGCQ00900237)
 - Fixed an issue where firmware may incorrectly select backplane slots when updating drive status - (SCGCQ00980105)
 - Fixed an issue where MSM shows messages with Error Code 240 and no description - (SCGCQ00970948, Port of SCGCQ00904466)
 - Fixed an issue where SafeID may change with repeated queries using StorCLI - (SCGCQ00949550, Port of SCGCQ00941577)
 - Fixed an issue where a Foreign Configuration import may fail after reconstruction when using 4k sector drives - (SCGCQ00830109)
 - Fixed an issue where setting the Learn Cycle property of iBBU09 gives an I2C error - (SCGCQ00831315)
 - Fixed an issue where "Prepare for Removal" does not work on a PD when a drive erase is in progress - (SCGCQ00811219, Port of SCGCQ00703189)
 - Fixed an issue where EFI Driver Health fails when more than 64VDs are configured - (SCGCQ00831006, Port of SCGCQ00830918)
 - Fixed an issue where a controller may report Cache Memory size as 0 when no logical drives are present - (SCGCQ00837600, Port of SCGCQ00837162)
 - Fixed an issue where concurrent Consistency Checks cannot be started after issuing "setfactorydefaults" command - (SCGCQ00839437, Port of SCGCQ00838830)
 - Fixed an issue where Cache Memory size is not reporting 0 when there is no Cache Memory - (SCGCQ00838893, Port of SCGCQ00838193)
 - Fixed an issue where deleting a CacheCade VD can result in a firmware hang situation - (SCGCQ00838975, Port of SCGCQ00838953)
 - Fixed an issue where an incorrect error (MFI_STAT_PD_TYPE_WRONG (0x26) instead of MFI_STAT_LD_BLOCK_SIZE_MISMATCH (0x65)) is returned by the firmware when PDs with different block sizes are mixed - (SCGCQ00840251, Port of SCGCQ00836103)
 - Fixed an issue where BBU status reports Optimal but firmwre logs indicate the BBU has failed - (SCGCQ00838052, Port of SCGCQ00837834)
 - Fixed an issue where write-verify performance was lower than expected - (SCGCQ00858151)
 - Fixed an issue where Write-Through and Write-Back cache policies can be changed via Windows Logical Disk Properties - (SCGCQ00856979)
 - Fixed an issue where firmware reports no SMART error when a PFA (Predictive Failure) drive is attached - (SCGCQ00839455)
 - Fixed an issue where a status of 0xFF is erroneously returned from StoreLib - (SCGCQ00832673, Port of SCGCQ00827685)
 - Fixed an issue where an MSM displays an error message if "Schedule Consistency Check" is selected - (SCGCQ00852649, Port of SCGCQ00832908)
 - Fixed an issue where handling of Functional Level Reset (FLR) is not present - (SCGCQ00854343, Port of SCGCQ00840253)

Enhancements:
 - Changed the number of Sense Retry counts for a specific video application - (SCGCQ00841000)
 - Added ability to determine the revision ID for SuperCaps with no VPD programmed - (SCGCQ00828990)
 - Added support for SES RQST_OK deassertion - (SCGCQ00834321, Port of SCGCQ00817840)
 - Improved CPLD update capabilities - (SCGCQ00840631, Port of SCGCQ00758178)
 - Enhanced ability to flush cache with write-verify on RAID0 volumes with SATA drives - (SCGCQ00846777, Port of SCGCQ00842665)

Version 23.33.0-0043 -  High Impact/High Probability of Occurrence
--------------------
Fixes:
 - Fixed an issue where hard drive slot amber LED not functioning correctly.
 - Fixed an issue in iMR mode where "Controller encountered fatal error and was reset" errors can occur under heavy IOs and background operations (SCGCQ00894643) (RETAIN Tip H001154)
 - Fixed an issue where BBU status says optimal but firmware logs indicate that the BBU has failed (SCGCQ00838052) (RETAIN Tip H214857) 
 - Fixed an issue where importing a foreign configuration may fail after reconstruction on 4k sector drives (SCGCQ00830109)
 - Fixed an issue where the firmware may allow the OS to change VD cache settings (SCGCQ00842350)
 - Fixed an issue where MR firmware reports "no SMART" error when attaching a PFA drive (SCGCQ00839455)
 - Fixed an issue where the firmware may hang (enter MonTask) when a guest OS is powered on or off in VMWare (SCGCQ00854914)
 - Fixed an issue where setting the learn cycle property of iBBU09 gives an I2C error (SCGCQ00831315)
 - Fixed an issue where "Prepare for Removal" does not work on a PD when drive erase is in progress (SCGCQ00811219)
 - Fixed an issue where MSM displays "The requested command has invalid arguments" error message if 'Schedule Consistency Check' is selected (SCGCQ00852649)
 - Fixed an issue where cache memory size in iMR is not showing 0 (SCGCQ00838893)
 - Fixed an issue where an incorrect MFI_STAT_PD_TYPE_WRONG error is returned by the firmware when PDs with different block size are mixed (SCGCQ00840251)
 - Fixed an issue where the controller reports cache memory size as 0 when no logical drives are present (SCGCQ00837600)
 - Fixed an issue where concurrent consistency check cannot be started after issuing "setfactorydefaults" command (SCGCQ00839437)
 - Fixed an issue where uEFI may report LSI EFI driver "Unhealthy" or "Failed" when more than 64VDs are configured (SCGCQ00831006)
 - Fixed an issue where functional level reset was not properly showing as unsupported (SCGCQ00854343) 
 - Fixed an issue where the legacy CTRL-H utility may report the wrong PCIe slot number (SCGCQ00799339)
 - Fixed an issue where the legacy OpROM may show an expander as a JBOD drive in POST (SCGCQ00809682)
 - Fixed an issue where an encrypted JBOD may be presented to the system in encrypted form even though password was entered incorrectly (SCGCQ00808086)

Enhancements:
 - Added ability to determine SuperCap revision ID when VPD is not programmed (SCGCQ00828990)
 - Added support for SES RQST_OK deassertion for customer proprietary JBOD SEP (SCGCQ00834321)
 - Added feature to enable support for detecting CPLD code version (SCGCQ00840631)
 - Added feature to enable drive fault LED to be lit when a known-bad drive is connected to the controller (SCGCQ00858142)
 - Added feature to turn on drive Fault LED when the controller detects a drive that was previously marked Bad (SCGCQ00858140)


Version 23.33.0-0018 -  High Impact/High Probability of Occurrence
--------------------
Fixes:
- Incorporated MegaRAID FW stack 5.13
- Fixed an issue where drive temperature cannot be displayed for a drive in Failed or Foreign state (SCGCQ00693959)
- Fixed an issue where after creating an LD via StorCLI or MSM the firmware writes into saved power condition mode page (0x1a) of all PDs in the LD and enables their idle timers (SCGCQ00658701)
- Fixed an issue where Patrol Read would sporadically correct the same defective LBA repeatedly (SCGCQ00729591)
- Fixed an issue where importing a Foreign array onto a new controller may fail with "Cannot force emulate VD to 512E since enable512Esupport is false" error in firmware log (SCGCQ00721106)
- Fixed an issue where the controller may enter MegaMon sporadically (SCGCQ00776409)
- Fixed an issue where MSM may not report a Full Initialization Aborted error until system rebooted (SCGCQ00759100)
- Fixed an issue where the firmware may allow the OS to change VD cache settings (SCGCQ00842350)
- Fixed an issue where Link Speed may show UNKNOWN when attaching a 3G or 6G SATA drive (SCGCQ00803522)
- Fixed an issue where StorLib may return 0x00ff (fail) code even though operation succeeds when creating a VD out of an existing drive group (SCGCQ00827685)
- Fixed an issue where format process of SSD may take much longer than in previous firmware (SCGCQ00674388)
- Fixed an issue where controller must have cleared config to enable 240VD unless existing config only included single-drive RAID0 VDs (SCGCQ00559643)
- Fixed an issue where some commands would timeout with CDB=ff ff ff ff (SCGCQ00459760)
- Fixed an issue where slot ordering in iMR mode might differ from slot numbering in MR mode (SCGCQ00559274)
- Blocked CacheCade deletion when multiple CacheCade volumes are associated with Virtual Drives (SCGCQ00830977) (RETAIN TIP H214323)

Features/enhancements:
- Added support for 512e drives (SCGCQ00425363)
Enhanced Event Log to include CVPM health details after every Learn Cycle to allow closer monitoring of CVPM health (SCGCQ00819877)
- Implemented a retry for reads of CVPM data to avoid erroneous event 386 or 150 in the Event Log (SCGCQ00392439)
- Added ability for firmware to report the temperature for all SAS/SATA HDD/SSD (SCGCQ00490819)
- Implemented ability for firmware to read drive temperature immediately following SAS discovery (SCGCQ00602573)
- Added ability to clear the MaintainPdFailHistory (SCGCQ00474559)
 
Version 23.22.0-0024
--------------------
 -Initial release





Lenovo Data Center Group Linux OS Support Home Page
linux.lenovo.com

© 2018-2019 Lenovo. All rights reserved

ThinkSystem RAID and ServeRAID M Series SAS/SATA Controller Driver for RHEL 6 (07.705.04.00-3)

´╗┐

ThinkSystem RAID and ServeRAID M Series SAS/SATA Controller Driver for Linux

This product supports [RHEL 6.9, 6.8, 6.7 ]

Adapters Supported:
  - ThinkSystem RAID 930-8i 2GB Flash PCIe 12Gb Adapter
  - ThinkSystem RAID 930-16i 4GB Flash PCIe 12Gb Adapter
  - ThinkSystem RAID 930-16i 8GB Flash PCIe 12Gb Adapter
  - ThinkSystem RAID 930-24i 4GB Flash PCIe 12Gb Adapter
  - ThinkSystem RAID 930-8e 4GB Flash PCIe 12Gb Adapter
  - ThinkSystem RAID 930-4i 2GB Flash Flex Adapter
  - ThinkSystem RAID 730-8i 1GB Cache PCIe 12Gb Adapter
  - ThinkSystem RAID 530-8i PCIe 12Gb Adapter
  - ThinkSystem RAID 530-8i Dense Adapter
  - ThinkSystem RAID 530-4i 4GB Flex Adapter
  - ServeRAID M1215 SAS/SATA Controller for Lenovo System x (46C9114)
  - ServeRAID M1210e SAS/SATA Controller for Lenovo System x
  - ServeRAID M5225 SAS/SATA Controller for Lenovo System x (00AE938)
  - ServeRAID M5210 SAS/SATA Controller for Lenovo System X (46C9110)
  - ServeRAID M5110 SAS/SATA Controller for Lenovo System x (81Y4481)
  - ServeRAID M1115 SAS/SATA Controller for Lenovo System X (81Y4448)


NOTE TO SERVICE - Reference RETAIN #N/A

Version 07.705.04.00-3 - Moderate Impact/Low Probability of Occurrence
----------------------
Defects (7)
- Provided driver workaround to reset __GFP_WAIT flag while allocating IO request frames.(SCGCQ01394892)
- Driver will use 64 bit request descriptor for sending IO to Ventura firmware.(SCGCQ01514452)
- Set DMA mask after FW is in READY state in PCI resume path of megaraid_sas driver(SCGCQ01389431)
- Fixed few dev found and coverity reported defects(SCGCQ01392652)
- Selective stream detection logic in driver(SCGCQ01392643)
- During PCI remove, set unload flag after scsi_remove_host is called.(SCGCQ01392656)
- update LD map after populate drv_map, increase wait in IOCTL, fix index in error handling, rework in refiring DCMDs.(SCGCQ01415439)
- Added support for threaded interrupts.(SCGCQ01615015 port of SCGCQ01498747)
- Driver should go through stream detection logic after it's finalized which path(fastpath vs non fastpath) IO will go through(SCGCQ01483730 port of SCGCQ01468189)
- Use IOC_INIT frame once driver has verified IOC_INIT frame allocation was successful.(SCGCQ01455132 port of SCGCQ01430747)
- MR: OEL 7.4 and 6.x x86_64 RPMs do not have the correct kernel versions.(SCGCQ01611888 port of SCGCQ01607445)
- expose fw_outstanding and drv_outstanding variables to the OS level sys interface.(SCGCQ01475068 port of SCGCQ01019663)
- Updated correct supported kernel version of Ubuntu16.04 in driver's document- OS_support_list_mr7.5.txt.(SCGCQ01564604 port of SCGCQ01454090)
- Added NVME passthru support in MR linux driver.(SCGCQ01424605 port of SCGCQ01407028)
- Updated correct supported kernel version of Ubuntu16.04 in driver's document- OS_support_list_mr7.5.txt.(SCGCQ01455206 port of SCGCQ01454090)
- Added NVME passthru support in MR linux driver.(SCGCQ01455135 port of SCGCQ01407028)
- Updated OEL6_UEK and OEL7_UEK buildkit.(SCGCQ01443433 port of SCGCQ01439776)
- Setup IRQs only after Reply descriptors are allocated.(SCGCQ01514459 port of SCGCQ01506133)
- Use pci_{alloc/free}_irq_vectors for >=4.12 kernels.(SCGCQ01449816 port of SCGCQ01449816)
- Use IOC_INIT frame once driver has verified IOC_INIT frame allocation was successful.(SCGCQ01449786 port of SCGCQ01430747)
- Set DMA mask after FW is in READY state in PCI resume path of megaraid_sas driver.(SCGCQ01392672 port of SCGCQ01389431)

EnhancementRequests (10)
- MR7.4: Add SLES12 SP3 OS support.(SCGCQ01396756)
- Added support for RHEL7.4.(SCGCQ01455156)
- Added driver support for OEL7.4.(SCGCQ01396747)
- Increased JBOD FPIO timeout by 1 sec to workaround hardware bug which causes timeout 1 sec earlier than provided timeout value.(SCGCQ01449957)
- Added support for threaded interrupts.(SCGCQ01498747)
- MR7.4: OS driver support for RHEL 7.4.(SCGCQ01396738)
- Added support for SLES12 SP3.(SCGCQ01455167)
- Added driver binary support for Xenserver7.2.(SCGCQ01393249)
- Remove support for RHEL7.1, OEL7.2 and SLES12.(SCGCQ01449982)
- Removed support for few OS to align with OS support matrix policy.(SCGCQ01393194)



Version 7.703.07.00- Moderate Impact/Low Probability of Occurrence
---------------------
fixes:
- Send sense buffer upper 32bits inside IOC INIT frame.(SCGCQ01315987)
- MR: On OEL 6.x/7.x with UEK kernels - megaraid_sas.conf file from the kmod rpm does not match kernel version.(SCGCQ01319368)
- pass upper 32 bits of version buffer and system information buffer to FW inside IOC INIT frame.(SCGCQ01326366)
- Workaround changes for FW MFA frame pull issue.(SCGCQ01338087)
- Log dmesg if IOC INIT fails in IOMMU and RDPQ enabled enviornment for user information only.(SCGCQ01338097)
- Validated if any of DMA buffers pool which should be in same 4GB region does not cross 4GB boundary.(SCGCQ01338372)
- Memset to zero for MFI command frame for IOC INIT and use lower_32_bits() macro to get lower 32bits of DMA address passed to FW.(SCGCQ01341287)
- Removed Crusader ID(device ID- 0x0015) from supported PCI IDs of MR7.2 linux driver.(SCGCQ01350963)
- MR7.3: Code review defect - ioc_init_frame not freed during kernel module unload.(SCGCQ01384694)
- MR7.2: In megasas_issue_tm() SMID needs to be used for Task abort case only.(SCGCQ01315252)
- Incorrect processing of IOCTL frames for SMP/STP commands.(SCGCQ01383823)
- Set cmd_status inside DCMD/IO/header frame to MFI_STAT_OK to wake up all pended IOCTLs.(SCGCQ01298582)
- Removed print "Failed from megasas_complete_cmd" from PD CFQ sequence DCMD completion path.(SCGCQ01321565)
- Do not switch DMA mask to meet same 4gb requirement. IOMMU does not work well with such DMA mask switch.(SCGCQ01342578)
- Allocated MFI frame size and the size of MPT chain frame(Where MFI frame resides for MFI MPT passthru commands) should match.(SCGCQ01342650)
- Minimum possible value for module parameter- resetwaittime should be 1 secs instead of 0 secs.(SCGCQ01344137)
- Set DMA mask after FW is in READY state in PCI resume path of megaraid_sas driver.(SCGCQ01389431)
- MR7.3: Data miscompare found with WB secure VD RAID 1 with LSA sending IOCTLs in parallel.(SCGCQ01378684)
- Set cmd_status inside DCMD/IO/header frame to MFI_STAT_OK to wake up all pended IOCTLs.(SCGCQ01381770 port of SCGCQ01298582)
- MR: On OEL 6.x/7.x with UEK kernels - megaraid_sas.conf file from the kmod rpm does not match kernel version.(SCGCQ01320324 port of SCGCQ01319368)
- Use IOC_INIT frame once driver has verified IOC_INIT frame allocation was successful.(SCGCQ01455132 port of SCGCQ01430747)
- Updated correct supported kernel version of Ubuntu16.04 in driver's document- OS_support_list_mr7.5.txt.(SCGCQ01455206 port of SCGCQ01454090)
- Added NVME passthru support in MR linux driver.(SCGCQ01455135 port of SCGCQ01407028)
- Updated OEL6_UEK and OEL7_UEK buildkit.(SCGCQ01443433 port of SCGCQ01439776)
- Minimum possible value for module parameter- resetwaittime should be 1 secs instead of 0 secs.(SCGCQ01344142 port of SCGCQ01344137)

EnhancementRequests:
- Added support for 64bit DMA mask for invader.(SCGCQ01294199)
- MR7.3: add support for Citrix XenServer 7.1 OS.(SCGCQ01257971)
- Send current time to FW during IOC init.(SCGCQ01307072)
- Added support for RHEL7.4.(SCGCQ01455156)
- MR7.3: add support for OEL 6.9.(SCGCQ01289005)
- Added support for threaded interrupts.(SCGCQ01498747)
- Added support for 64bit DMA.(SCGCQ01301940)
- Added support for SLES12 SP3.(SCGCQ01455167)


Version 7.702.06.00- Moderate Impact/Low Probability of Occurrence
---------------------
Defects (47)
 - Reset MPT frame field cmd_completed while retrurning it back to free pool.(SCGCQ01176530)
 - Optimized R1 FP write code.(SCGCQ01191077)
 - Before populating driver RAID map copy memset drv_map to zero and ldTgtIdtoLd array to 0xff(SCGCQ01204087)
 - Initialize stream detection array properly.(SCGCQ01204096)
 - On SPARC/Big endian host, need to convert peerSMID to little endian to CPU before accessing(SCGCQ01216698)
 - Fixed coverity reported issues and changed name of few functions.(SCGCQ01219231)
 - Don't set tm_capable for JBDO when PD CFG SEQ map is not available(SCGCQ01226222)
 - Remove local defination FALSE/TRUE and use OS provided false/true.(SCGCQ01256130)
 - Used vmalloc for crash dump buffers and driver's local RAID maps(SCGCQ01270059)
 - Reduce size of fusion_context allocation during driver load by moving load_balance_info array out, and try vmalloc if kmalloc fails.(SCGCQ01183280)
 - Flooding of messages- "BRCM: IO does not end on page boundary; sge_len" while running RAW IOs.(SCGCQ01189815)
 - Fixed data corruption by asking OS not to do merging of IOs to avoid SGE holes(SCGCQ01231229)
 - Code review defect: create separate function for allocating RAID maps(SCGCQ01282924)
 - Removed Crusader ID(device ID- 0x0015) from supported PCI IDs of MR7.2 linux driver(SCGCQ01350963)
 - Updated OS to send page aligned SGEs by calling blk_queue_virt_boundary(SCGCQ01206279)
 - Use vmalloc followed by memset to zero instead of vzalloc in driver code.(SCGCQ01206316)
 - MR_ValidateMapInfo issue on bigendian platform (In some case Detect Ghost VDs)(SCGCQ01213952)
 - Fixed code review defects(SCGCQ01254323)
 - Removed code inside macros VD_EXT_DEBUG, SPAN_DEBUG and enahnce prints in OCR path(SCGCQ01270056)
 - FW fault happens as driver wrongly fill numSGE=0.(SCGCQ01274103)
 - Code optimization and enhanced prints for debugging(SCGCQ01224881)
 - Remove unsupported PCI Device IDs from the driver code(SCGCQ01234256)
 
 - Resolved conflict of PCI device ID-0x0015(SCGCQ01254317)
 - MR7.2: In megasas_issue_tm() SMID needs to be used for Task abort case only(SCGCQ01315252)
 - Save interfaceType for each sysPDs inside SCSI device's private data field which will persists until sysPD is removed/deleted.(SCGCQ01216464)
 - Code Optimization and enhanced prints(SCGCQ01223300)
 
 - MR 7.1: Linux OS crash when powering off enclosure during I/O(SCGCQ01229442)
 - Return pending IOs with status DID_REQUEUE instead of DID_RESET to SCSU layer before doing OCR(SCGCQ01231846)
 - MR7.2: Port changes from upstream to make in-house MR 7.2 code to be consistent with upstream code(SCGCQ01282321)
 - Add print for partially copied SCSI IO and fix few endianness issues(SCGCQ01224305)
 - Linux Driver 07.700.23.00/Getting megaraid_sas prints during startup of OS(SCGCQ01255074)
 - aen class range check to avoid kernel panic(SCGCQ01256126)
 - Corrected bitmap macros inside struct RAID_CONTEXT35(SCGCQ01269453)
 - Set cmd_status inside DCMD/IO/header frame to MFI_STAT_OK to wake up all pended IOCTLs.(SCGCQ01298582)
 - Fix few dev found issues(SCGCQ01226224)
 - On OEL SPARC (Big Endian), Linux MR 7.1 driver loading fails to initialize Ventura which has RDPQ support enabled(SCGCQ01229897)
 - Fixed few code collab defects/comments(SCGCQ01202237)
 - Remove throttle IOs prints in IO path and add sysfs parameter to reflect number of R1 FP writes(SCGCQ01204099)
 - Ventura MR-7.0.1 RHEL 7.2 driver: ERR messages when installing driver 7.700.20.00(SCGCQ01222990)
 - Use do_div() kernel API to do modulo operation of 64 bit dividend and 32bit divisor .(SCGCQ01243263)
 - megasas_complete_cmd_dpc_fusion() will be called with 1 sec delay inside function megasas_wait_for_outstanding_fusion(SCGCQ01256116)
 - Allocated MFI frame size and the size of MPT chain frame(Where MFI frame resides for MFI MPT passthru commands) should match(SCGCQ01342650)

 EnhancementRequests (17)
 - Big endian architecture is now support for Ventura generation of MR HBA.(SCGCQ01256016)
 - Fixed the problem of sync_irq not being called when MSI-x are disabled.(SCGCQ01206307)
 - Port changes from MR7.0 and MR6.13 critical fixes to MR7.1(SCGCQ01167939)
 - In megasas_build_ldio_fusion, code to set controller CPU affinity can be optimized(SCGCQ01216486)
 - Set device queue depth same as controller's can_queue for SCSI MQ mode(SCGCQ01271569)
 - Add SLES 11 32-bit build kit support to Ventura MR.(SCGCQ01214721)
 - Ported fixes from MR 6.13 to MR 7.1(SCGCQ01222522)
 - Added support for RHEL7.3(SCGCQ01224300)
 - Support SPARC/PPC/ARM (Big/Little Endian). Fix Endianess issue in MR Ventura driver.(SCGCQ01183275)
 - Added DCMD interface to read target properties for SCSI device(VDs and JBODs both)(SCGCQ01220142)
 - Added KMOD RPM support for OEL7.3 UEK.(SCGCQ01227476)
 - R1 Write IOPs improvement(SCGCQ01213232)
 - Extra code for bug in A0 ventura is removed. Optimized Descriptor write as Ventura required 32 bit in major IO Path operation.(SCGCQ01215347)
 - Remove Crusader device support from driver(SCGCQ01259175)
 - Don't re-issue Reset on the way back OCR(SCGCQ01215089)
 - Rework start_vm_build script used for VBAS build and new build infrastructure for DKMS RPMs(SCGCQ01261800)
-

Version 7.700.26.00- Moderate Impact/Low Probability of Occurrence
---------------------
Enhancement:
- Added support for ThinkSystem HBA and RAID.


Version 6.812.07.00 - Moderate Impact/Low Probability of Occurrence
---------------------
Fixes
 - Changed timeout value for DCMDs get_bios_data and get_pd_info 180 seconds to keep it aligned with other DCMDs' timeout value. (SCGCQ00884247)
 - Do not allow MFI frame to be freed from DCMD fire context when already an OCR is in progress. (SCGCQ00879190)
 - Fixed an issue where driver always worked in non RDPQ mode even the running firmware did not support RDPQ. (SCGCQ00872542)
 - Moved SCSI layer interaction code out of reset_mutex in AEN thread to prevent deadlock around reset_mutex. (SCGCQ00868731)
 - Convert supported JBOD count value returned by firmware in CPU format and then use the same value for further validation (SCGCQ00837088)
 - Fixed an issue where unevenSpan configuration of R1 would not have proper load balancing for FP READ. SCGCQ00646224
 - Added timeout value for wait_event. (SCGCQ00249742)
 - Enabled hostlockless flag of SCSI host for RHEL6 and clones for async IO submission from SCSI mid layer. (SCGCQ00884242)
 - Fixed an issue where double freeing of MFI frame in case of MFI IO timeout in polled mode could occur. (SCGCQ00837932)
 - Added synchronization between OCR thread and AEN thread, so that there should not be any PCI write from AEN thread during OCR (SCGCQ00836410)
 - Added IO timeout value for SCSI management commands for JBOD. (SCGCQ00750452)
 - Fixed a JBOD performance issue due to incorrect MSI-x index set in request desc. (SCGCQ00619694)
 - Changed state machine of firmware crash dump collection to prevent file system on VD being set to Read-only mode after ECC error injection. (SCGCQ00608705)
 - Fixed an issue where incorrect messages such as LSI SAS2.0 were printed for MR driver. (SCGCQ00298201_
 - Fixed an issue where the driver sent an incorrect region type "REGION_LOCK_EXCLUSIVE" set to FW for RAID level >=5
 - Changed error handling for DCMD failure cases to allow graceful exit for not fatal errors. (SCGCQ00868661)
 - Free task management frame after dumping the task management reply frame data to ensure that TM reply frame will have valid data returned by firmware. (SCGCQ00860038)
 - Release reset_mutex inside OCR function, if adapter is already killed to prevent a kernel panic. (SCGCQ00857740)
 - Increased timeout to 180 seconds for flush cache and shutdown controller DCMDs to prevent kernel panic when DCMDs fired during driver's shutdown. (SCGCQ00841491)
 - Fixed an issue where traversed LDs at multiple places up to maximum 256 were limited to 64. (SCGCQ00652795)
  - Fixed an issue where the driver crashed when msix_disable=1 was passed to the boot time parameter. (SCGCQ00383508)
 - Fixed an issue where kernel panic seen while deleting VD and background IOs running. (SCGCQ00803468)
 - Fixed an issue where the driver was not sending DCMD- MR_DCMD_CTRL_SET_CRASH_DUMP_PARAMS after OCR. (SCGCQ00601138)
 - Fixed an issue where the kernel panicked inside PCI subsystem because of IRQs not freed before disabling MSI-X interrupts (SCGCQ00827403)
 - Disabled SED drive JBOD support. (SCGCQ00751296)
 - Fixed multiple OCRs triggering hits Exception Core issue (SCGCQ00707592)
- Fixed an issue where the firmware faulted due to invalid regionlock Flags set by Driver. (SCGCQ00796461)
- Fixed an issue where subsequent reads of state value from hardware resulted different values and could lead to a timeout in and a non working adapter. (SCGCQ00709216)
- Fixed an issue where Offline VD caused boot hangs in SLES 11 (SCGCQ00284564)
 - Disabled SPAN_DEBUG and set get_arm_from_strip to print if firmware supports UnevenSpan. (SCGCQ00365846)
 - Added check to validate ldcount exposed in firmware RAID map for legacy firmware.  If ldcount is more than 64, fail RAID map validation. (SCGCQ00820349)
 - Fixed an issue where the Linux kernel crashed when AEN and LD Sync commands were issued frequently. (SCGCQ00712031)
 - Fixed an issue where duplicate disk(s) were seen on controller after loading Linux driver during OS installation, (SCGCQ00568992)
 - Fixed an issue where kernel panic could occur at driver load time on 3.0 kernels (i.e. SLES11 SP2, SLES11SP3). (SCGCQ00561505)
 - Changed driver to return SCSI command back to mid layer if the controller has any critical hardware errors. SCGCQ00362023
 - Fixed an issue where the driver did not do Load balance for R1 READ due to wrong stats counters. (SCGCQ00677516 Port of Defect - SCGCQ00674872)
 - Fixed potential duplicate IOCinit frames being fired by VMware5 VF driver during OCR. (SCGCQ00571967 PORT OF SCGCQ00571679)
 - Fixed OS crash issue on SCSI IO timeout (SCGCQ00631210 PORT OF SCGCQ00625137)
 -Fixed an issue where the kernel panicked due to MFI/MPT pass-through command timeout and later finished from Interrupt context. (SCGCQ00307867 PORT OF - SCGCQ00302080)
 - Fixed an issue where megasas_poll_wait_aen was cleared before poll syscall returns causing to lose events (SCGCQ00575192 PORT OF SCGCQ00574787)
 -Fixed an issue where data corruption observed on JBODs while running IOs with 32k block size. (SCGCQ00530992 PORT OF SCGCQ00523780)
 - Fixed MPC kernel panic during chassis power cycle if AEN completes and detach is called before event workqueue runs. (SCGCQ00647390 PORT OF SCGCQ00645836)
 - Fixed an issue where kernel crashed in some cases, where megaraid_sas driver is loaded but controller is not present. (SCGCQ00282968 PORT OF SCGCQ00282067)


Enhancements:  
 - Set queue depth for syspD based on interface type (SCGCQ00879700)
 - Updated module Meta data (module description and module author) with Avago replacing LSI. (SCGCQ00801961)
 - Set OS provided timeout value for sysPD non read-write IOs (SCGCQ00764795)
 - Performance tuning for SMP affinity is implemented for distros's which support irq_set_affinity_hint API. (SCGCQ00623697)
 - Optimized OCR path to enable interrupt and make HBA OPERATIONAL after driver has fired LD_MAP_GET_INFO DCMD. (SCGCQ00768402)
 - Added module parameter to turn on/off SMP affinity code. (SCGCQ00666706)
 - Driver will do load balance for N Drive Raid 1 volumes. (SCGCQ00634064)
 -Functionality in driver is added to differentiate between iMR and MR at firmware initialization time. (SCGCQ00361244)
 - Uneven Span support added in Driver. (SCGCQ00319104)
 - Driver code is updated such way that it can work for any block size drive, if the rest of the system has proper support for the same. (SCGCQ00259244)
 - Implemented online controller reset if controller is not responding at driver load time. (SCGCQ00820389)
 - Added SCSI residual bytes support to the megaraid_sas Linux driver (SCGCQ00347089)
 - Use sequence number for JBOD FastPath I/Os to eliminate command block after drive replacement (SCGCQ00806534)
 - Big endian architecture support added for MR Linux driver (SCGCQ00526911)
 -  Driver support is added for 3108 C0 chip combine reply queue ON (SCGCQ00354586)
 - Added support 240 VD feature (SCGCQ00622703)
 - Added Firmware crash dump collection support (SCGCQ00550369)
 - Added support for large I/O 1MB transfer size (SCGCQ00816313)
 - Changed the max sg table size to 64. Old setting was 70 (SCGCQ00653992)
 - Added driver support for task management for Jbod or VD (tmCapable) (SCGCQ00854635)
 - Added support for SED JBOD in MR Linux driver (SCGCQ00711555)
 - Changed code to complete IOCtls posted just before a kill adapter so user space applications do not hang. (SCGCQ00759916)


Version 6.810.09.00 - Moderate Impact/Low Probability of Occurrence
---------------------
 Fixes:

 - Fixed an issue where deadlock may occur when the Automatic Event Notification thread runs in parallel with error handler thread (SCGCQ00911332, Port of SCGCQ00868731)

 - Fixed an issue where driver will operated in RDPQ mode regardless of whether firmware has advertised support or not (SCGCQ00872542)

 - Fixed an issue where system may reset with error "BUG: unable to handle kernel NULL pointer dereference at 00000000000000d0" preceded by error "megasas: target reset FAILED!!" (SCGCQ00875223)

 - Fixed an issue where driver MFI frames could be freed twice (SCGCQ00879190)

 - Fixed an issue where some out-of-band commands had different timeouts than others (SCGCQ00884247)

 - Fixed an issue where performance was lower on async IO submissions from the SCSI mid layer (SCGCQ00884242)

 - Fixed an issue where a kernel panic may occur if the system is reset or fail to reboot (SCGCQ00859634, Port of SCGCQ00857740)

 - Fixed an issue where Task Management frames were not being dumped after the Tast Management Reply frame data (SCGCQ00860038)

 - Fixed an issue where a fatal error may occur when an out-of-band command times out during an Online Controller Reset (SCGCQ00882011, Port of SCGCQ00868661)

 - Fixed an issue where kernel may hang when many commands are timing out or many Automatic Event Notifications are being processed (SCGCQ00857245)

 - Fixed an issue where megaraid_sas module was not being included inside initrd/initramfs (SCGCQ00880965)

 - Fixed an issue where kernel may panic if driver does not setup IRQs for non-MSI-x interrupt mode (SCGCQ00870263, Port of SCGCQ00867873)

 - Fixed an issue where driver return code is based on command completion rather than command return code (SCGCQ00856103, Port of SCGCQ00856089)

 - Fixed an issue where Tape drives may not be detected by the Operating System (SCGCQ00914385, Port of SCGCQ00911702)

 - Fixed an issue where max_sectors per IO request may be set to more than firmware can handle (SCGCQ00917992, Port of SCGCQ00917399)
 
- Fixed an issue where an IERR may occurr on heavy IO with Failover (SCGCQ00864111, Port of SCGCQ00736612)
 
 
Enhancements:

 - Enhanced queue depths in driver for physical disks based on interface type (SCGCQ00879700)

 - Improved naming of internal functions (SCGCQ00866064)

 - Updated module parameter "rdpq_enable" default to "1" (enable RQPQ mode) (SCGCQ00871565)

 - Enhanced reply queue to support queue depth greater than 1K (SCGCQ00853521)

 - Added feature for driver to support Task Management for JBOD and Virtual Drives (SCGCQ00854635)

 - Added module parameter to disable 1M IO support and set default to enable 1M IO support (SCGCQ00911362)

 - Enhanced tasklet initialization so it would occur before setting up IRQs (SCGCQ00914393, Port of SCGCQ00911709)

 - Simplified HB handling for SR-IOV (SCGCQ00864113, Port of SCGCQ00700580)

 - Enhanced driver to update Online Controller Reset capability when it changes (SCGCQ00914389, Port of SCGCQ00908352)

 - Added feature to print Critical, Fatal and Dead firmware events into system logs (SCGCQ00868723, SCGCQ00882010, Port of SCGCQ00866176)

 - Enhanced HB handling to align with behavior on other Operating Systems (SCGCQ00864117, Port of SCGCQ00707323)

 - Removed excess dead code (SCGCQ00850012)

Version 6.808.16.00 - Moderate Impact/Low Probability of Occurrence
---------------------
Fixes:
- Fixed an issue where a deadlock may occur around reset_mutex when Automatic Event Notification threads run in parallel with error handlers - (SCGCQ00911332, Port of SCGCQ00868731)
- Fixed an issue where scanning of Tape drives does not occur depending on whether the drive state is set to systemPD or not - (SCGCQ00914385, Port of SCGCQ00911702)
- Fixed an issue where max_sectors may be set larger than what firmware can support - (SCGCQ00917992, Port of SCGCQ00917399)
- Fixed an issue where kernel may panic due to driver not setting up IRQs for non MSI-x interrupt mode - (SCGCQ00870263, Port of SCGCQ00867873)
- Fixed an issue where driver may halt firmware due to improper error handling for DCMD failure cases - (SCGCQ00882011, Port of SCGCQ00868661)
- Fixed an issue where driver may hang kernel due to deadlock - (SCGCQ00857245)
- Fixed an issue where system may crash due to timeout for flushing cache and shutting down controller DCMDs was not long enough - (SCGCQ00841491)
- Fixed an issue where deadlock may occur between Online Controller Reset and Automatic Event Notification processing function - (SCGCQ00846933)
- Fixed an issue where PCI writes from Automatic Event Notification thread were occurring during Online Controller Reset - (SCGCQ00836410)
- Fixed an issue where MFI frames linked list may corrupt due to double-freeing of frame in the case of an MFI IO timeout - (SCGCQ00837932)
- Fixed an issue where kernel may panic inside PCI subsystem due to IRQs not freed before disabling MSI-X interrupts - (SCGCQ00827403)
- Fixed an issue where ldcount was not validated for non-240VD firmware - (SCGCQ00820349)
- Fixed an issue where a command block would occur for JBODs with FastPath IOs after a drive is replaced - (SCGCQ00806534)
- Fixed an issue where 1MB IO support was not supported controller-wide - (SCGCQ00816313)

Enhancements:
- Set DMA alignment to 8 bytes for PI-enabled VDs - (SCGCQ00928796, Port of SCGCQ00928176)
- Added a parameter to support disabling 1MB IO support - (SCGCQ00911362)
- Added support for printing Critical, Fatal and Dead Firmware events into system logs - (SCGCQ00882010, Port of SCGCQ00866176)
- Added support for driver to perform an Online Controller Reset at driver load time if the controller is not responding - (SCGCQ00820389)

Version 6.808.14.00 - Moderate Impact/Low Probability of Occurrence
---------------------
Fixes
- Fixed an issue where controller may reset or "MonTask: line 1276" error will occur with JBOD present (SCGCQ00837088)
- Fixed a performance issue on RAID1 with 256k IO size (SCGCQ00674872)
- Fixed a performance issue on RAID1 reads with FastPath enabled (SCGCQ00646224)
- Fixed an issue where an Online Controller Reset may take an unusually lon time (SCGCQ00420374)
- Fixed an issue where a controller firmware crash may not be recognized by driver (SCGCQ00249742)
- Fixed an issue where an MFI frame may be freed twice on MFI command timeout (SCGCQ00837932)
- Fixed an issue where the firmware may fault in an Automatic Event Notification fires during an Online Controller Reset before the chip is in READY state (SCGCQ00836410)
- Fixed an issue where there was no IO timeout value for SCSI management commands for Secure JBOD (SCGCQ00750452)
- Fixed a performance issue on SSDs in JBOD mode (SCGCQ00619694)
- Fixed an issue where file system on a VD can go to read-only on ECC error detection (SCGCQ00608705)
- Fixed an issue where disks may not be found during installation after loading the driver (SCGCQ00299643)
- Fixed an issue where a kernel panic may occur when cable pulled during system reboot (SCGCQ00841491)
- Fixed an issue where 240VD support was not fully implemented (SCGCQ00652795)
- Fixed an issue where the driver may crash if msix_disable=1 is passed to kernel as boot parameter (SCGCQ00383508)
- Fixed an issue where IO errors may occur when connected to a Tape drive (SCGCQ00399947)
- Fixed an issue where kernel may hang due to megaraid_sas driver deadlock (SCGCQ00857245)
- Fixed an issue where kernel may panic when deleting VD and background IO is occurring (SCGCQ00803468)
- Fixed an issue where crash dump may be disabled after an Online Controller Reset (SCGCQ00601138)
- Fixed an issue in 32bit OSs where driver may not load on boot (SCGCQ00362423)
- Fixed an issue where deadlock may occur (SCGCQ00846933)
- Fixed an issue where a kernel panic may occur because IRQs are not freed before disabling MSI-X interrupts (SCGCQ00827403)
- Fixed an issue where multiple Online Controller Resets can trigger a Core Exception (SCGCQ00707592)
- Fixed an issue where crash dump may not be performed properly (SCGCQ00593374)
- Fixed an issue where firmware may fault with CacheCade (SCGCQ00796461)
- Fixed an issue where adapter may hang (SCGCQ00709216)
- Fixed an issue where an offline VD may cause boot hangs (SCGCQ00284564)
- Fixed an issue where kernel may panic (SCGCQ00820349)
- Fixed an issue where kernel may crash in megaraid_sas driver with CacheCade if there are many Automatic Event Notifications and LD sync operations occurring (SCGCQ00712031)
- Fixed an issue where system may hang with CacheCade (SCGCQ00687881)
- Fixed an issue where the same disk may show multiple times in OS (SCGCQ00568992)
- Fixed an issue where kernel may panic on Linux kernels version 3.0 or greater (SCGCQ00561505)
- Fixed an issue where invalid IO statistics may be provided if driver supports more than just 512b sector drives (SCGCQ00282615)
- Fixed an issue where driver fails to reboot system due to outstanding IOs in SCSI mid-layer when controller is in error state (SCGCQ00362023)

Enhancements
- Added ability for JBOD IOs going via firmware path to use timeout value provided by OS (SCGCQ00764795)
- Added a retry of IO commands as LDIO if raid map sync is failed with MFI_STAT_CONFIG_SEQ_MISMATCH (SCGCQ00712510)
- Added support for 4k sector drives if the system supports it (SCGCQ00259244)
- Added support for driver to perform an Online Controller Reset if the controller is not responding (SCGCQ00820389)
- Fixed an issue where system may hang on S3 resume (SCGCQ00758933)
- Added a feature to prevent command block with FastPath after JBOD drive replacement (SCGCQ00806534)
- Enhanced 1MB IO support (SCGCQ00816313)
- Added support for "megaraid_sas.disable_ext_io=1" boot parameter to disable 1M IO support (SCGCQ00911362)
- Added support for Secure JBOD (SCGCQ00711555, SCGCQ00753765)
- Fixed an issue where user space applications may hang if IOCtl is sent prior a kill adapter issued (SCGCQ00759916)
- Fixed an issue where OS may crash on a SCSI IO timeout (SCGCQ00631210)
- Fixed an issue where system may fail to reboot or will kernel panic on init 6 if controller has faulted (SCGCQ00859634)
- Fixed an issue where firmware may fault (SCGCQ00702536)
- Fixed an issue where controller may crash during system power cycle (SCGCQ00647390)
- Fixed an issue where kernel may crash when megaraid_sas driver is loaded but no MegaRAID controllers are present (SCGCQ00282968)


Version 6.806.08.00-1 - Moderate Impact/Low Probability of Occurrence
-------------------
Fixes:
- Incorporated MegaRAID 6.6 x86/x86_64 Linux driver stream
- Fixed an issue where performance may drop in sequential read 256kb IO (SCGCQ00677516)
- Fixed an issue which may lead to firmware crash and "MonTask: line 1257 in file ../../raid/2108vI2o.c" error in firmware log (SCGCQ00702536)

Features/enhancements:
- Added a retry of IO commands as LDIO if raid map sync is failed with MFI_STAT_CONFIG_SEQ_MISMATCH (SCGCQ00651680)
- Added support for 1MB IO if the firmware supports it as well (SCGCQ00671670)
- Added ability to take a VD offline / bring a VD online, similar to the equivalent actions for PDs (SCGCQ00521610)
- Added support for Self-Encrypting Drives in JBOD mode (SCGCQ00399739)
Note:  Installed device driver version is 6.806.08.00

Version 6.704.15.00 - Suggested
-------------------
 - Fixed an issue where uneven span configuration of RAID 1 will
   not have proper load balancing for Fast Path Read.

Version 6.602.03.00-2 - Initial release
---------------------





Lenovo Data Center Group Linux OS Support Home Page
linux.lenovo.com

© 2018-2019 Lenovo. All rights reserved