DFCT ID:	LSID100151807
Headline:	1800 PCI ERROR - No IRQ available forSystem which uses DL_Down instead of or in addition to PCI reset
Description:	Issue is seein with EF P8 code - On some servers if on soft reset or hard reset (through power button), BMC unit generates DL_DOWN interrupt and removes power. the customer‚s server uses DL_Down instead of or in addition to PCI reset user will see failure. Testing has shown that cache data will be lost roughly 1 in 5 times.
Version of Bug Reported:	11.0.1-0021
Version of Bug Fixed:	1.40.92-0746_PL-Ver-1.25.78.0
Steps to Reproduce:	Plug 1078 controller on the system which uses DL_Down instead of or in addition to PCI reset , up on boot user will see "1800 PCI ERROR - No IRQ available for PCI"
Resolution:	Fixed
Resolution Description:	On servers where DL-Down/DL_UP is part of reboot sequence after PCIRESET, controller does not post if MC is put in self refresh mode as part of DL_Down ISR. Instead do not reset Memory Controller as part of Module Reset in DL_Down ISR. This eliminates the need to put Memory Controller in Self Refresh mode. Instead let HW logic on interceptor put MC in Self-Refresh in case of power failure.
Customer List:	OEM -- OEM
Fix Impact:	Medium
Suggested Testing:	Data integrity across large file transfers tested during •     Soft reboot (invoked via Ctl+Alt+Del) •     Shut down (invoked by means of BMC switch) •     Power failure (invoked by yanking off power cable) Testing to be carried out on the following systems: •     System that generates PCI Reset but no DL_DOWN/DL_UP sequence during boot-up sequence •     System that generates DL_DOWN/DL_UP sequence but no PCI Reset during boot-up sequence •     System that generates PCI Reset followed by DL_DOWN/DL_UP sequence during boot-up sequence
Child Tasks:	LSID100151750

DFCT ID:	LSID100125563
Customer DFCT No:	HSA0238
Headline:	One drive command timeout causes Fatal FW Error.
Description:	One drive command timeout causes Fatal FW Error by "Driver detected possible F/W hang". <OEM Request> FW shouldn't fall into Fatal Firmware Error. Instead, make the no-response drive fail. <Note> -This issue is inherent in all programs such as Oconee2,Unicoi and EF. See attached <PR_HSA0238.doc> for more information.
Version of Bug Reported:	1.40.42-0615
Version of Bug Fixed:	1.40.92-0746_PL-Ver-1.25.78.0
Steps to Reproduce:	<Environments> RAID Controller: Eagle Ray (or ROMB w/1078) : Read policy --> Normal RAID Configuration: RAID0(PD0/OS) + RAID5(PD1-3) Drives: Fujitsu MAX3073RC PD1 is Error Injection Drive (MAX3073RC F/W:21BC) Error Type: No response after 5minutes from power-on. OS:Windows Server 2008 Enterprise x86 RAID Driver:3.9.0.32 <Steps to duplicate> a) Excecute continuous 8KB/Random Read access using IO meter for RAID5 drive. b) No response happens on PD1 during running IO meter c) After about 4 minutes the system stops with Fatal Firmware Error by Driver detected possinbe F/W hang. <=This is the problematic result. [NOTE] Issue rate is 100% when 8KB/Random Read access by IO meter. However with the other access pattern the issue happens intermitently.
Resolution:	Fixed
Resolution Description:	This problem should not be there with AF. In AF there is a change to drop the drive if it times out the same IO twice (as compared to 4 times in Unicoi or EF) So when a retried IO times out the drive will be marked as bad thereby avoiding the montask.
Customer Defect Track No:	HSA0238
Customer List:	OEM -- OEM
Fix Impact:	Medium
Suggested Testing:	use non-response disk, and run I/O meter
Child Tasks:	LSID100151819

DFCT ID:	LSID100151806
Headline:	NoMRRSChg environment variable needs to be included in EF code
Description:	NoMRRSChg env variable must be added so that stamping will turn off/on the c1078TimerMisc() function
Version of Bug Reported:	All previous versions
Version of Bug Fixed:	1.40.92-0746_PL-Ver-1.25.78.0
Steps to Reproduce:	None
Resolution:	Fixed
Resolution Description:	Provide support for ENV string NoMRRSChg to enable / disable dynamic readjusment of PCIE Maximum Read Request Size
Customer List:	LSI -- LSI
Fix Impact:	Low
Suggested Testing:	Stamp the image with string NoMRRSChg=1. FW log should print the message "Found Env String NoMRRSChg" User /host BIOS should be able to reduce Maximum Read Request Size to value less than 2048
Child Tasks:	LSID100151809

Task ID:	LSID100151819
Headline:	Immediate IO command time out fix
Description:	FW retries immediate I/O time out command with cached I/O path. In case timing wheel slots have many commands (typically more than 10 commands) at the time, and the disk becomes “non-response” condition, the timed out command will take more than 4 minutes to complete (with fail status). Fix is to modify the call back function for immediate I/O command. If the immediate command failed for the reason that is NOT for medium error, then the command should fail immediately without cached i/o path retry.
State:	Completed
Change Set Files:	0
References:	LSID100125563(DFCT)

Task ID:	LSID100151809
Headline:	Provide support for ENV string NoMRRSChg
Description:	Provide support for ENV string NoMRRSChg to enable / disable dynamic readjusment of PCIE Maximum Read Request Size
State:	Completed
Change Set Files:	0
References:	LSID100151806(DFCT)

Task ID:	LSID100151750
Headline:	Disable reset of Memory Controller
Description:	On servers where DL-Down/DL_UP is part of reboot sequence after PCIRESET, controller does not post if MC is put in self refresh mode as part of DL_Down ISR. Instead do not reset Memory Controller as part of Module Reset in DL_Down ISR. This eliminates the need to put Memory Controller in Self Refresh mode. Instead let HW logic on interceptor put MC in Self-Refresh in case of power failure.
State:	Completed
Change Set Files:	0
References:	LSID100151807(DFCT)

Task ID:	LSID100151822
Headline:	updating maintenance version from 8 to 9
Description:	updating maintenance version from 8 to 9
State:	Open
Change Set Files:	0
References:

Task ID:	LSID100151784
Headline:	write journal hole backport-4
Description:	write journal hole leads to data loss
State:	Completed
Change Set Files:	0
References:

Task ID:	LSID100151814
Headline:	Avoid invocation of ReadXor logic in R1
Description:	Hardware issue is reported from the field on OEM specific controller where ReadXor output corresponds to data from the first source buffer instead of XOR of first/second buffer. Workaround will: 1) Always use WriteXor instead of ReadXor for RAID 1 CC/BGI XOR operations 2) Whenever a CC/BGI inconsistency occurs, use the primary buffer as the data source for the mirror write instead of using the XOR-regenerated mirror data.
State:	Completed
Change Set Files:	0
References:

Task ID:	LSID100151842
Headline:	FW_SAS Release Version: 1.40.92-0746
Description:	FW_SAS Release Version: 1.40.92-0746
State:	Open
Change Set Files:	0
References: