01SF222_079_079 HOT: GA3 SP2: B7000602 after Inband upgrade Load LIC None. 01SF222_078_078 LIC_INCORROUT IOA failure : adapter shows fscsi errors in AIX Load LIC None. 01SF222_077_077 FSP Shiplist PTR for Fix Pack 01SF222_078 FSP Shiplist PTR for Fix Pack 01SF222_078 None. FSP Shiplist PTR for Fix Pack 01SF222_066 FSP Shiplist PTR for Fix Pack 01SF222_066 None. During grub-testing runs there may be intermittent failures of B181F63B. This is due to the specific commands used by the GRUB tool and should not affect normal IPLs. Added a 2 second pause before checking the instruction completion bit. This will give the hardware adequate time to complete instructions. There is no workaround beyond attempting the operation again. This does not affect customer IPLs and only concerns manufacturing tests. 01SF222_075_075 FSP Shiplist PTR for Fix Pack 01SF222_077 FSP Shiplist PTR for Fix Pack 01SF222_077 None. The HMC will get a un-recoverable SRC when attempting to extract platform dump data if the fsp is not in the terminated or runtime states. The F029 SRC will be generated when the HMC attempts to extract a platform dump when the system is not in either terminate or run time state. the platform dump code changed to severity of the F029 SRC from un-recoverable to informational. B200813A SRC spotted in Error Logs. firmware microcode changed. None The failing address callout by PRD when a memory error occurs could be incorrect. CEC HW initialization values were changed to workaround this HW issue. Replace all DIMMs when a memory error is identified instead of the callout. The DIMMs that need to be replaced are system/platform dependent. When 2 (or more) nodes are sharing the same IO towers and one node detects an internal failure a second node may eventually lose the IO towers and the node also goes down. N/A N/A The correct memory FRU may not be called out when a memory failure occurs. A fix was added to program the hardware to capture the correct failing memory address. None. Field symptom that platform dump may run out of space. Remove nonessential rings from platform dump to help ensure that platform dump will not run out of space. Not applicable IPL "hangs" in the memdiags step of the IPL if a UE is found on ranks 0x00000008 to 0x0000000F of an 8GB dimm. Error log containing reason code: HFUS_MEMCTLR_MAINT_UE = F667 will repeat until system is powered down. One-line code change to fix bug. Without the code fix, the failing dimm must be identified and removed, replaced or manually garded (using ASM) before the system can IPL. See repeating F667 error log to identify failing dimm. If the system is IPLed for a lot of times, then due to this defect, memory leak occurs and FSP may reset/reload eventually. The problem was due to an error log handle was not processed properly in the "svpd_is_embedde_fru" function. The fix is to receive and delete this error log handle. There is no workaround for this problem. Need ability to bias voltage at one half percent intervals on H systems. Add one half percentage voltage bias capability to H systems. Not applicable The LEDs on the I/O drawers will not identify or rollup to the enclosure or the frame. A fix was made to the SPCN microcode. N/A B181F2DE does not have hardware callouts. But instead has two extra procedure callouts FSPSP28 FSPFFFF Change the code that determines what FRU a piece of hardware is on so it reports it correctly Run a slow mode IPL. If the IPL still fails with a F2DE ref code, call second level of support systems may fail at runtime in manufacturing when recoverable CEC HW errors occur. do not configure CEC HW to checkstop when recoverable errors occur in manufacturing. do not enable the manufacturing threshold policy flag MFG specific change. No Symptom. Increase thresholds as requested. No workaround avaliable. Patches provided to mfg. the FSP could generate error logs every few seconds when attempting to service a recoverable CEC HW error on systems that have processors with no memory installed behind them. configure CEC HW to ignore memory errors on processors that have no memory installed behind them. install memory behind all processors. Memory errors are detected running a lab testcase on DDR2 Power5 systems when the processor frequency is 1.65GHz. The solution is to change a register setting in the processor. The only known workaround is to change the hardware combination to something other than 1.65GHz with DDR2 memory. The current callout for and 1100871X does not call out a DCA problem. A fix was added to change the callout procedure for the 11008710 and 11008720 errors. If an 1100871X error occurs the first call out should be the DCA. After the reboot the fspremote tool hangs the fsp ipl. For fspremote tool used by mfg, the check is not look machine type if the state. is pre-standby. Run the fspremote tool after >5 mins after the reset of the fsp was issued. The code was incorrectly trying to optimize HSL traffic when an HSL link has failed. The code has been change to not attempt to optimize RIO traffic when an HSL link has failed. None. A reserved-uninitialized entry was found in the AIX device tree when there should not be. Code will be changed so that a reserved-uninitialized entry is not present in the AIX device tree when it should not be. None. Certain system configurations require Memory on Demand support. Changes will be made to support Memory on Demand for select system configurations. None. 01SF222_074_074 ASM option to program system vpd keywords, especially TM, is not working. When the anchor card is blank, ASM doesn't allow users to program TM. The error is invalid entry even though the data is typed correctly. A fix was added to allow customers to program blank system vpd keyworkds. So customer must have this firmware fix before they can program system vpd key words on blank anchor card using ASM. None. 01SF222_073_073 When the system is powered on, the front panel LED continues to blink rather than going on solidly. The firmware which controls the LED will be corrected to listen for the power on event. No workaround is available. The firmware will have to be updated when the logic is corrected. Enhanced FRU isolation. Updated Licensed Internal Code. None Serial console unresponsive following service processor reset/reload Licensed Internal Code updated. None For 590/595, resolved issue with System Attention LED not being activated when a component identification indicator was active. Updated Licensed Internal Code. None Improved error data and Platform dump data collection Update Licensed Internal code. None 590/595 Only: IPL failure if primary Service Processor fails during system power off. Update Licensed Internal code None Call home fials to attempt all destinations if the first location to succeed. Alter retry strategy to enable all locations regardless of previous fialures. N/A 590/595 Only: Corrects condition where one of the BPC's in a redundant pair will not become available after a code update. Update Licensed Internal Code. None Corrected condition preventing System attention LED from being turned off after a failure is resolved Update Licensed Internal Code. None Provide correct information for cases where Location Codes had the incorrect values. Update Licensed Internal Code. None Improved error recovery when updating system firmware from the HMC. Update Licensed Internal Code. Switch Side to P. This can be done using ASM Menus. Reset the fsp from the ASM menus, so that FSP will boot of the P-Side. Retry code update again. System Fails to re-IPL when dump date gathered exceeds file space. Update Licensed Internal Code None Loss of AC power as the Service Processor is rebooting during a microcoode update can cause unrecoverable corruption of the Service Processor's code which will require that it be replaced. Op panel repeatedly displays B18172** or B15172**. Systems without Op-panel will show No-Connection on the HMC GUI. Update Licensed Internal code. None Prevent condition where service processor may reset on a two frame 590 System. Update Licensed Internal code. None Condition where system Firmware code update erases Capacity on Demand activation codes. Update Licensed Internal code. None 01SF222_071_071 590/595 Only: Resolved erroneous reporting of reporting of faults relating to thermal monitoring, fan control, or communications problems between power controllers. Update Licensed Internal Code. None 590/595 Only: IPL Failure due to communication problems between power controllers within the power commuincation network. Update Licensed Internal Code. None Communication problem between system and HMC that results in hung Virtual Processors. Update Licensed Internal Code. None System is unable to dynamically deallocate 8GB DIMM and continue to boot when un uncorrectable error is detected during boot. Update Licensed Internal Code. None Displaying CEC properties through Hardware Management Console fails for iSeries Integrated xSeries Adapters. Update Licensed Internal Code. None. iSeries partition power down hang occurs if system loses state of data traffic on the HSL Opticonnect. Update Licensed Internal Code. None. System fails to quiesce properly or hangs during run-time due to a failed processor. Update Licensed Internal Code. None. Corrects condition where unreturned Trial Capacity on Demand memory is incorrectly reported as unreturned On/Off COD Memory when the trial period has expired. Update Licensed Internal code. None Improve IPL Performance when IPLing a large number of partitions at the same time. Update Licensed Internal code. None. iSeries only: With a UPS configured, message "CPI0961 UPS No Longer Attached" is generated six hours after an IPL. Update Licensed Internal Code. None. A partition's time appears to gain or lose time following a system IPL. This drift is in the range of 15 seconds per month. Update Licensed Internal Code. None. Support i5OS on pSeries. Update Licensed Internal Code. None If SMS option 5 "Default Boot List" is selected when booting, the machine will boot off of the stored boot list (option 6) instead of the default list. Update Licensed Internal Code. None. 01SF220_051_051 01SF220_050_050 Partition fails to reboot following snoop string detection Licensed internal code changed Reboot system using ASM interface. A timing problem as been identified in the power supply. A 11007120 service code may be posted. Licensed internal code changed N/A SRC B201E0AA found in platform error log Serial port handling code may close bad file descriptor N/A The Utility Failure detection may not report correctly to the host partitions. Licensed internal code changed N/A Power code may fail to allow the system to initialize. Licensed internal code changed N/A 01SF220_049_049 IPL fails with an SRC of B1xxF660. Licensed internal code changed Problem is intermittent so a reboot may be successful, problem will not affect runtime operation. Error code B155B058 may incorrectly tell you to replace the FSP Added code so that error will point to the correct piece of hardware to replace. Remove or swap out any recent hardware added to the system. Premature temperature warnings may be posted. Thresholds for temperature warnings are being raised to account for differences in sensor locations. N/A IPLing one partition participating in HSL Opticonnect causes HSL Opticonnect on other partitions to fail. Licensed internal code changed None. Tasks may loop processing machine check interrupts. Licensed internal code changed None. Vital product data in I/O drawers programmed by legacy pSeries hardware is not recognized by current hardware. Licensed internal code changed None. HMC went into a no connect state after attempting to re-IPL partition. Licensed internal code changed None. System does not recover following a reset/reload Licensed internal code changed None. System fails with SRC B181E500 after tower switch occurrs. Licensed internal code changed None. 01SF220_048_048 Incorrect processor time base frequency possible on RPA partitions. Fixed reported processor time based frequency for RPA partitions None. The IPL of a system connected in an HSL Opticonnect cluster may hang at SRCC7004080. The code was getting into a deadlock state during clustering initialization. The Licensed Internal Code has been corrected to avoid this deadlock. None. Adapter descriptions displayed on HMC are incorrect after resources are physically moved. Hypervisor logic will be more sensitive to IO resource changes. None. System profile fails to activate Hypervisor memory validation is not accounting for memory currently assigned to the partition. None. Reserve Processor request gets started successfully but the processors are not present in the Shared Pool. With this corrective package, Reserve Processors are fully integrated into the Shared Pool so their capacity can be accessed as expected. None. Systems participating in an HSL Opticonnect cluster with 3 systems on the same HSL ring may not connect. Licensed Interal Code Changed None. Resolved potential cause for system hang at checkpoint C7004078 when HSL Opticonnect starts. Licensed Internal Code Changed None. Corrected truncated serial number information displayed by lscfg command for I/O drawers or towers. Licensed Internal Code Changed None. Profile activation may fail when changing partition max memory. Licensed Internal Code Changed None. The server connection status displayed by the Server and Partition Management utility on the Hardware Management Console goes to Incomplete after a concurrent tower move. The server connection status displayed by the Server and Partition Management utility on the Hardware Management Console will not go to Incomplete after a concurrent tower move. None. A deadlock can occur causing console to not be able to comunicate with system. Licensed Internal Code Changed None. Opticonnect HSL links are not marked failed correctly when an optical cable is removed. Code updated to correctly mark links failed. None. Corrected potential cause for IPL hang at checkpoint C700406E Licensed Internal Code Changed None. Reserve CoD menus incorrectly display maximum number of processors supported instead of number which are physically installed Reserve CoD was using the maximum possible processors when calculating the number allowed to be activated. It has been changed to use the actual installed processors, which will produce results consistent with the number shown for available processors. None. Corrected potential cause for HMC Console going to incomplete state. Licensed Internal Code Changed None. Advanced POWER Virtualization: Current Shared Procs allowed to Violate Min and Max rules. Licensed Internal Code Changed None. OptiConnect High-Speed link (HSL) speeds may be set incorrectly. Licensed Internal Code updated to correctly initialize HSL links. None. OptiConnect High-Speed link (HSL) links fail to recover after cable replugged Licensed Internal Code updated to properly recover HSL links. None. IBM 4758 PCI Cryptographic Coprocessor does not identify in which slot is is physically plugged. Licensed Internal Code Changed None. Linux partition was failed to boot with error BA210000. Licensed Internal Code Changed None. Linux partitions boot slowly from Virtual SCSI. Licensed Internal Code Changed None. UPS detection is intermittent Licensed Internal Code Changed None. Cannot connect to ASM menus via network interface Licensed Internal Code Changed Reboot the service processor (FSP) Improved FRU isolation for error B181FB53 Added DIMM replacement procedure. None. B181F23E on IPL with missing SMP cable Licensed Internal Code changed to Allow system to IPL with only the first node present. None. ASM TTY interface hangs when web interface has non-English language selected Licensed Internal Code Changed Set web interface to English. Serial ports not functional of p570 systems equipped with multiple processor drawers (8, 12 and 16 way systems) Licensed Internal Code Changed Access ASM via HMC or graphics adapter. Corrected cause for false cooling fan faults when system in standby. (Affects 9111 systems only) Licensed Internal Code Changed None. Improved FRU isolation for error B123E500 Licensed Internal Code Changed None. 01SF220_047_047 B181F7BB error running mixed processors System configurations with 1.9Ghz processors may have a mix of High voltage and nominal voltage processors in the same system. Code updated to support this configuration. Do not mix high and nominal voltage processors in the same system. B1818A09 after power problem Code bug fixed to not report erroneous error. N/A 01SF220_046_046 01SF220_045_045 If an ethernet cable is not plugged into the service processor port, the service processor may not reach standby because the webserver dies causing the service processor to reboot and then hang. The service processor will configure the adapter even if the network ethernet cable is not plugged in for any port. Take a regular ethernet cable and connect it in a loop from HMC1 port to HMC2 port of the service processor 01SF220_044_044 Error logs are being incorrectly created during firmware downloads to I/O towers. Changed the error logs to informational logs only. N/A 01SF220_043_043 Dynamic memory remove from a partition fails. The code was incorrectly indicating that memory was unavailable for dynamic removal. The code has been changed to correctly indicate when memory is available for dynamic removal. None. Concurrent replace of an IOP fails. The code was not properly deallocating hypervisor resources belonging to the I/O resource being removed from the system. The code has been corrected. None. A logical partition exhibits time-of-day drift across a system IPL. Problem fixed in hypervisor code. None. Reserve Processors for Capacity On Demand are supposed to be idle until called upon for partition workloads. In some cases, all the reserved processors were put into use right away when they were not needed. The algorithm for deciding when the line of usage got far enough into the Reserve Processor capabilities was not accurate enough. None. A Trial CoD activation may be accepted while there are unreturned Trial resources. Trial activation will not be allowed while there are unreturned Trial resources of the same resource type. None. CoD history log entries are not being produced when Trial CoD expiration results in unreturned resources or when unreturned Trial or On/Off resources have all been reclaimed. CoD history log entries will be produced when Trial CoD expiration results in unreturned resources or when unreturned Trial or On/Off resources have all been reclaimed. None. Confirmation of a Capacity on Demand code for Trial resources may not work correctly depending on the number of available resources for conversion to Trial resources. The system hypervisor will be fixed to correctly handle CoD Trial resource confirmation requests. None. After replacement of card, unreturned CoD resources were reported as unreturned Trial instead of unreturned On/Off. When CoD resources become unreturned for reasons other than Trial CoD expiration, they will be treated as unreturned On/Off CoD resources. None. HMC displayed fields for Capacity on Demand memory amounts total to more memory than is installed in the server. Firmware is updated to correctly pass to the HMC for display the Capacity on Demand memory amounts. None. Entering an On/Off CoD termination code while there were unreturned On/Off CoD resources resulted in an extended error code of 0x'200A', which means that an On/Off request is active, rather than an error code of 0x'2018', which means the key failed because of unreturned resources. The failure code will be changed to indicate when a code is rejected due to unreturned resources. None. After activation and expiration of On/Off Processor resources, the entitled number of resources that LPAR should allow is not correct. The calcluations done after expiration of the On/Off Processor requests, as this testcase represented, were inaccurate. This inaccuracy was corrected and now the reported useable resource numbers are correctly being sent to the LPAR componentry. None. After a PCI bus error occurs that would require the PCI bus to be reset and recovered, about 2 seconds after the error occur, other IO error events may be received by the partition. Code has been fixed to remove the erroneous IO error events while the PCI bus is being recovered. None. Error log ID not the same for error entries fromt he same I/O error. Problem has been fixed so error entries due to the same error will get the same error log ID. None. Hardware recovery on an AIX or LINUX system does not always re-enable interrups. Leaving the devices in a non-functional state. Device recovery now restores devices to full function. None. Tower concurrent maintenence and DLPAR of slot holding an IOP is failing. The DLPAR and Tower concurrent maintenence operations succeed with this fix. None. The system may become unresponsive. The Licensed Internal Code will be changed so that the system will not become unresponsive. None. Error adding a cpu. Need to change the cache building scheme in firmware. None. VSCSIDisk not showing up. Changed calculation and the drives now appear. None. DIAG CD not IPLing in Slave CD/DVD Slot. Fixed. None. 01SF220_041_041 Hmc may not connect to system. Licensed Internal Code Changed None. System terminating with a A700xxxx error code after a reset/reload of the service processor. Removed erroneous reject of host intiated reset/reload. None. Increase resources in order to allow the use of more I/O adapters in a large configuration. Increase partiton resources. None. 01SF220_041_041 New firmware release New functions added N/A