01SF235_206_160 If your system is one of the models listed below, and has a Remote I/O (RIO) adapter F/C 1806, F/C 1807, or F/C 2888 installed, and has one or more I/O drawers attached to this adapter, a false failure has been identified which can potentially cause your system to fail and reboot. Due to this, it is recommended to plan to reboot the platform after installing this service pack to activate this fix and prevent a potential unplanned reboot. Affected sytems: i5: 9405-520 and 9406-520 with processor feature code 8325, 8327 or 8330 p5: 9115-505, 9111-285, 9131-52A, and 9133-551 System firmware changes that affect the model 575, 59x, other HMC-controlled servers, and clustered systems: In a system with a redundant service processor, a problem was fixed that caused one of the service processors to fail, and be removed from the configuration, if an interrupt was pending during initialization. In a system with a redundant service processor, a problem was fixed that caused one of the service processors to fail during a failover, and the hypervisor was then unable to reset the system. In a system with a redundant service processor, a problem was fixed that caused PCI bus errors to occur when a hardware reset of the service processor occurred at runtime, and the service processor was on a shared PCI bus. System firmware changes that affect certain p5 and i5 systems: DEFERRED and HIPER: A problem was fixed that caused the system to crash with SRC B131E500 ("I/O hub error"). This affects the following systems if Remote I/O (RIO) attachment card F/C 1806, F/C 1807, or F/C 2888 is installed, and one or more I/O drawers are attached to this adapter: i5: 9405-520 and 9406-520 with processor feature code 8325, 8327 or 8330 p5: 9115-505, 9111-285, 9131-52A, and 9133-551 HIPER: A problem was fixed that caused erroneous B113E500 errors. This affects the following systems: i5: 9405-520 and 9406-520 with processor feature code 8325, 8327 or 8330 p5: 9115-505, 9111-285, 9131-52A, and 9133-551 HIPER: A problem was fixed that intermittently caused a service processor performing a hardware reset to disturb other devices on the same PCI bus. This affects the following systems: i5: 9406-570 p5: 9117-570, 9118-575 HIPER: A problem was fixed that caused certain types of I/O adapters to fail at runtime. This affects the following systems: i5: 9405-520 and 9406-520 with processor feature code 8325, 8327 or 8330 p5: 9115-505, 9111-285, 9131-52A, and 9133-551 A problem was fixed that caused erroneous errors during boot on certain slots. This affects the following systems: i5: 9405-520 and 9406-520 with processor feature code 8325, 8327 or 8330 p5: 9115-505, 9111-285, 9131-52A, and 9133-551 01SF235_185_160 In firmware level SF235_180 (and by implication SF235_185), there were two fixes that were deferred: A fix that affected all p5, i5, and OpenPower systems: HIPER: An incorrect initialization setting in the shared memory interface (SMI) may cause a system checkstop, which could result in a system or partition crash. this is also known as the "SMI parity error" issue. In SF235_206, this fix is CONCURRENT. A fix that affected model 59x systems: HIPER: The initialization settings of the memory buffers were changed to fix an incorrect timeout setting in the non-cacheable unit (NCU). The current settings may cause a system checkstop, which could result in a system and partition crash. This is known as the "EMQ hang" problem. In SF235_206, this fix is CONCURRENT. Also in this release of the fix, the recovery time for the EMQ hang was improved to benefit HACMP configurations. The following changes affect all p5 and i5 systems: HIPER: A problem was fixed that caused the system to interpret a machine condition as requiring special attention handling, where no such special attention existed. This could result in a system checkstop. This problem is also know as the "Master Attention Alone" condition. HIPER: A problem was fixed that might keep the recovery of correctable memory errors from preventing the reoccurrence of those memory errors. It was indicated by error code B12xE500. This is known as the "bit steering" issue. DEFERRED and HIPER: Additional hardware initialization settings were changed to prevent system hangs with B7xx406E reference code. DEFERRED: A problem was fixed that caused a dump to stop before completion after a hardware failure with error code B1xxF22A. DEFERRED: A problem was fixed that caused the system to terminate with an error code of B182901D. A problem was fixed that caused the service processor to perform a reset/reload too frequently, which caused too many dump files to be generated. A problem was fixed that was causing SRC B1xxE500 (with SRC word 8 equal to DD02) when processor runtime diagnostics were called during a dump. A problem was fixed that caused a BA180010 error to be logged when a PCI-X DDR slot was empty. A problem was fixed that caused error codes 11001D60 and 11001D63 to be erroneously generated when powering down the system. A problem was fixed that kept an already-connected uninterruptible power supply (UPS) from being detected correctly after the service processor performed a reset/reload operation. A problem was fixed that caused the system to fail to boot if the white power button was pressed within 10 seconds after reaching standby, but before C1802000 was posted on the control (operator) panel. A problem was fixed that caused error code B157F22A to be posted after a main store dump. A problem was fixed that prevented informational errors in the advanced system management (ASM) error logs from being seen by the admin user. The firmware was changed to insure that error log entries related to ambient temperature problems (error codes such as 11007201, 11007203, and 11007205, for example) are sent up to the operating system. A problem was fixed that was preventing systems and partitions with multiple fiber channel adapters from booting. A problem was fixed that caused a reset/reload to fail with error code B181D13C when a certain type of hardware attention was being handled. System firmware changes that affect the model 575, 59x, other HMC-controlled servers, and clustered systems: DEFERRED and HIPER: A potential server checkstop issue exists in clusters using p5 models p575, p590, and p595, when these models are used to execute compute-intensive applications tuned with the November Power5 Fortran compiler. Clusters executing applications that are compiled for optimal performance by utilizing the software pre-fetch function are exposed to a server checkstop (B1xxE500). Hardware initialization settings were changed to fix this problem and the performance tools modified to take these new settings into account . DEFERRED and HIPER: An initalization problem was fixed that was causing the high-performance switch (HPS) network adapter to hit the adapter error threshold during run-time. HIPER: A problem was fixed that caused model 575 nodes to fail to boot with SRC B1817201. HIPER: On systems with a high performance switch (HPS) network adapter installed, a problem was fixed that caused a checkstop when memory was moved from a partition that did not own an HPS network adapter. HIPER: A problem was fixed that was causing the power control code to time out when querying the vital product data (VPD) on large system configurations. HIPER: In a system with a redundant service processor, a problem was fixed that caused erroneous registry errors with error code B1818A09 to be generated intermittently. HIPER: In a system with a redundant service processor, a problem was fixed that caused runtime fail-over to the secondary service processor to fail. HIPER: In a system with a redundant service processor, a problem was fixed that caused 11001Dxx error codes to be erroneously logged after a failover to the secondary service processor. HIPER: In a system with a redundant service processor, a problem was fixed that caused the system to fail to boot, and one of the service processors to become deconfigured. HIPER: In a system with a redundant service processor, a problem was fixed that caused the corruption of registry values on the primary service processor. This registry corruption was evidenced by the fact that the hypervisor reported the amount of configurable memory incorrectly. In a system with a redundant service processor, a problem was fixed that caused a firmware installation to fail with multiple occurrences of error code B1818A0E. The firmware was enhanced so that control (operator) panel function 30 displays the secondary service processor's IP address as well as the primary service processor's IP address. In a system with a redundant service processor, a problem was fixed that was causing error code B17CE433 to be erroneously generated after an administrative fail-over. A problem was fixed that caused error code HSCL1400 to be generated when a redundant service processor was installed with firmware release SF230 or earlier. In a system with a redundant service processor, a problem was fixed that was causing the secondary service processor to fail after too many write operations into the flash memory. In a system with a redundant service processor, a problem was fixed that was causing the system to reboot continuously while the secondary service processor was attempting to mount a file system. On a system in which a redundant service processor is installed, a problem was fixed that was preventing the system from booting (with error code B155B06F) when one service processor encountered a problem collecting vital product data (VPD). A problem was fixed that caused a system with only a primary service processor installed to report the state of the secondary as "broken" instead of "not installed"; this prevented a firmware installation on the primary service processor. A problem was fixed that caused the error logs to be cleared on the secondary service processor when it was rebooted. In a system with a redundant service processor, a problem was fixed that prevented a service processor fail-over from completing if a hardware error occurred before the fail-over. In a system with a redundant service processor, a problem was fixed that was causing a 11005000 or 11005001 error code to be logged during an administrative failover. In a system with a redundant service processor, a problem was fixed that was causing a partition's time-of-day (TOD) clock to jump forward a large amount of time after a failover to the secondary service processor. If this problem is encountered after this level of firmware is installed, the partition will have to be rebooted to correct the TOD clock. In a system with a redundant service processor, a problem was fixed that was causing an invalid error code (B181E436) to be generated during a concurrent firmware update. In a system with a redundant service processor, a problem was fixed that was preventing the system from being powered on or powered off after a surveillance failover during a power transition (with error code B181E664). A problem was fixed that was preventing a PCI adapter from configuring in a tower after concurrent maintenance on the slot. If this problem is encountered, the system must be rebooted for the adapter to be correctly configured after the repair. After this fix is concurrently activated, it will prevent the problem from reoccurring. A problem was fixed that caused some error log entries not to be seen on the HMC. A problem was fixed that caused multiple B176871C error codes to be generated. A problem was fixed that caused an error log entry with error code B1xxE661 to be erroneously generated after an HMC-initiated system firmware installation. A problem was fixed that was seen on an HSL Opticonnect cluster when there were more than 10 partitions on either system in the Opticonnect cluster; the failure occurred during a partition boot. - If you are currently affected by this problem, a server IPL must be performed to reactivate the low-level HSL Opticonnect connection between the servers. - This problem can be avoided by manually starting the partitions individually instead of using "autostart" to start the partitions. - Concurrently activating this fix will prevent the problem from occurring again. A problem was fixed that occurred during server boot on a server with a redundant service processor installed; the server hangs at progress code C700406E - If the server is currently hung at C700406E, this level of firmware can be installed disruptively from the HMC, and the problem will be resolved when the system is automatically rebooted after the installation. - If the server is not hung, this level of firmware can be applied concurrently. A problem was fixed that caused some error codes not to be sent to the control (operator) panel when i5/OS failed; instead, D200C2FF is left in the control panel display. The terminating SRC can be found in the ASMI error logs. This problem is only seen on i5 systems in default mode without an HMC attached. A problem was fixed that may be encountered during a reset/reload of the service processor; the symptom is a system failure with a B7000103 error code. If this problem is encountered, the system must be rebooted to recover. When this fix is concurrently activated, it will prevent the problem from reoccurring. System firmware changes that affect model 570 systems with more than one processor drawer: On model 570 systems with more than one processor drawer, a problem was fixed that was preventing enclosure VPD keywords from being changed in the second, third, and fourth processor drawers using the ASMI menus. The firmware was enhanced so that the progress codes that indicate discovery of the secondary service processor are displayed on the control (operator) panel. System firmware changes that affect p5 and i5 systems not in 24 inch racks: On a system that is not managed by an HMC, progress codes were being displayed on the potential firmware consoles; this caused the console selection message to scroll off the screen. A problem was fixed that caused error code 11002600 SRC to be generated during the power-down sequence, and a failure of the subsequent power-on with error code B150F22A. 01SF235_180_160 The following system firmware changes affect specific p5 and i5 systems not in 24 inch racks: Several problems were fixed that caused a failure to boot with SRCs B1817201, C1001014 and/or C1001020. Supports Double Data Rate (DDR) PCI adapters in DDR PCI slots in the following systems: 9111-520, 9131-52A, 9113-550, 9133-55A, 9111-285, 9405-520, and 9406-520. Running a DDR PCI adapter in a DDR PCI slot in one of these models with a firmware level prior to this one may result in a potential data error. IBM recommends that you install this level immediately if using a DDR PCI adapter. Full DDR adapter performance is not enabled in this level. Full performance is planned to be enabled in a service pack that will be released in the future. LIC code updated. None. The following system firmware changes affect models 590 and 595, other HMC-controlled systems, and clustered systems: Fixed a problem that was causing error code B1xx100A to be logged when an HMC was disconnected from a managed system. This problem is much more likely to occur on managed systems that have two or more HMCs attached, but it can also be seen on systems with only one HMC attached. LIC code updated. None. A problem was fixed that caused commands involving LEDs to time out on large, heavily-configured systems. LIC code updated. None. A new function was added to system firmware to allow power code downloads to be disabled and enabled. On eServer systems with a large number of towers, this allows customers to shorten the downtime after a system firmware installation by disabling the power code download to the towers, then enabling it at a convenient time. LIC code updated. None. 01SF235_160_160 The following system firmware changes affect all p5 and i5 systems: DEFERRED - The hardware initialization settings were changed to prevent system crashes at runtime under some circumstances. LIC code updated. None. DEFERRED - The initialization settings of the memory buffers were changed to eliminate spurious checkstops that were causing system and partition crashes. LIC code updated. None. A problem was fixed that causes the system attention LED to be in the incorrect state. LIC code updated. None. Fixed a problem that was causing the system attention light to come on even though no serviceable events had been logged. LIC code updated. None. A problem was fixed that was causing the System Management Services (SMS) menus to hang when the boot list was viewed or changed when two AIX images are on one hard file. LIC code updated. None. A problem was fixed that caused a 11003114 error code to be generated during power on. LIC code updated. None. A problem was fixed that caused a null pointer in the virtual I/O (VIO) firmware. LIC code updated. None. i5/OS only: A problem was fixed that was preventing an i5/OS partition from booting from an IOA in a DDR slot. LIC code updated. None. i5/OS only: A problem was fixed that was causing some adapters to fail to become operational in i5/OS in a DDR slot. LIC code updated. None. i5/OS only: A problem was fixed that was causing incorrect error reporting for adapters that do not support extended error handling (EEH) when installed in a DDR slot. LIC code updated. None. The following system firmware changes affect models 590 and 595, other HMC-controlled systems, and clustered systems: DEFERRED - A problem was fixed that prevented a VTERM session from being opened on the HMC in certain network configurations. LIC code updated. None. A problem was fixed that caused bulk power controller (BPC) connection errors in clustered systems by increasing the NETC SSL time-out value. LIC code updated. None. A problem was fixed that caused HMC-directed concurrent maintenance actions to fail with the message "CIM client detected: hard stop or user intervention. Try again or delay the repair.". LIC code updated. None. A problem was fixed that was causing platform dumps on large, richly-configured systems to result in an out-of-bounds memory access failure. LIC code updated. None. A change was made to shorten the time to execute DLPAR memory removal operations. LIC code updated. None. The following system firmware changes affect systems that are not HMC-controlled: A timing issue in the firmware was fixed that was causing firmware installations from the operating system to fail. LIC code updated. None. The following system firmware changes affect the Model 570: On model 570 systems with 12 or 16 processors, the maximum number of partitions had been erroneously set to 80 in SF235_160; this firmware level sets the maximum back to 120 partitions for 12 processors and 160 for 16 processors. LIC code updated. None. On the model 570 with three or four processor drawers, a problem was fixed that was causing the system to stop with error code A7004713 when installing SF235_160 or higher system firmware. LIC code updated. None. 01SF230_145_120 Fixed a problem that was causing the user data in a service processor error log entry generated by the bulk power controller to be truncated to 2048 bytes. LIC code updated. None. Fixes a problem that caused unnecessary Service Agent calls because SRC B1816009 was being generated when the white power button was pressed before the power-on sequence was complete. LIC code updated. None. On model 590 and 595 systems, fixed a problem that was causing DDR II memory cards in slots C10 and C12 to fail the memory test during a slow-mode boot. LIC code updated. None. Fixed a problem that was causing a platform dump to be incomplete under certain circumstances. LIC code updated. None. Fixed a problem that was causing a hypervisor dump to be corrupted. LIC code updated. None. Fixed a problem that caused the location codes to be incorrect after the planar is replaced in a pSeries 7311-D20 drawer. LIC code updated. None. The timestamp from the bulk power controller was added into the serviceable event log entry. LIC code updated. None. Fixed a problem that prevented the "nvunalias" open firmware command from working. LIC code updated. None. Fixed a problem that kept the physical I/O description field on the HMC GUI from being updated after a "remove FRU and add FRU" operation. LIC code updated. None. Fixed a problem that caused a service processor dump or a platform dump and SRC B1817201 or B7004400. LIC code updated. None. Changed the firmware so that a modem attached to a system port will reconfigure to enable call-in following a power failure. LIC code updated. None. Fixed several problems associated with firmware installations: - SRC B1814008 or B181F131 was generated on reboot - A service processor dump occurred during a firmware installation "lsmcode -c" command displayed all zeros for the firmware level on the T side after an installation LIC code updated. None. Fixed VPD-related command hangs (such as lsmcode, lsvpd, snap) LIC code updated. None. On i5 systems, added support for booting from a boot device attached to an IOP-less adapter. New function. None. Added support for the installation of two copies of the AIX operating system on one hard file. This includes support in the SMS menus for booting from either one. New function. None. Added support for feature code 1812, the GX Dual-port 4x InfiniBand Host Channel Adapter, on model 52A systems. Added support for feature code 1811, the GX Dual-port 4x InfiniBand Host Channel Adapter, on model 575 systems. New function. None. Added support for moving the CUoD resources from one machine to another machine. New function. None. Added support for the capability to setup a backup shared ethernet so that the shared ethernet function can fail-over to a standby shared ethernet if the primary fails. This function is part of the virtual IO server. New function. None. Added support for redundant service processors with dynamic failover in models 570, 590, and 595. New function. None. Added support for modifying the number of On/Off CoD activations during the active period. New function. None. Added support for unattended HMC installation over a network, and to back up full HMC disk images onto a network server. LIC code updated. None. Added support for new models: 9131-52A, 9111-285, 9133-55A, 9110-51A, 9115-505. LIC code updated to correct the data stream sent. None. Various enhancements to the ASCII version of the Advanced System Management Interface (ASMI) menus were made. LIC code updated None.