OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
From: AIX Service Mail Server (aixservaustin.ibm.com)
Date: Tue Aug 28 2001 - 02:22:00 CDT

  • Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

    APAR: IX85840 COMPID: 5765C3403 REL: 430
    ABSTRACT: DTTERM PRINTS GARBAGE IF STARTED FROM FILE MENU

    PROBLEM DESCRIPTION:
    Opening a dtterm using the NEW button
    from the FILE menu will create Garbage
    charecters in Ja_JP font. This is covered
    by defect 264446.

    PROBLEM SUMMARY:
    Closing apar since defect is being worked on.

    ------

    APAR: IY17138 COMPID: 5765C3403 REL: 430
    ABSTRACT: IN PRINTING FORTRAN POINTEE VARS, THE WRONG SYMBOL IS EVAL'D

    PROBLEM DESCRIPTION:
    printing FORTRAN pointee variables causes error message

    PROBLEM CONCLUSION:
    dup string passed into pointer/pointee coupling function

    ------

    APAR: IY18237 COMPID: 5765C3403 REL: 433
    ABSTRACT: CANNOT CREATE STATIC FILTER RULES FOR INTERFACE CSS0,PP0

    PROBLEM DESCRIPTION:
    Executing the command "/usr/sbin/genfilt -v 4 -a 'P' -s
    '10.10.10.10' -m '255.255.255.255' -g 'y' -c 'all' -o 'any'
    -p '0' -O 13 'any' -P '0' -r 'B' -w 'B' -l 'N' -f 'y' -t
    '0' -i 'css0'" will generate the "invalid interface css0"
    error message.

    PROBLEM CONCLUSION:
    Allow the filter rule creation on a specific interface, if
    the interface is either defined in the CuDv database or is
    active (meaning it is in the active interface list).

    ------

    APAR: IY18381 COMPID: 5765C3403 REL: 430
    ABSTRACT: ALT_DISK_INSTALL MKSYSB WON'T INSTALL

    PROBLEM DESCRIPTION:
    LED 555 from bootup of mksysb alt_disk_install.

    PROBLEM CONCLUSION:
    Code modifications to fix JFS2 stanzas.

    ------

    APAR: IY18402 COMPID: 5765C3403 REL: 430
    ABSTRACT: DTMAIL VACATION FEATURE NOT FUNCTIONING PROPERLY

    PROBLEM DESCRIPTION:
    The vacation feature in dtmail won't work

    PROBLEM CONCLUSION:
    Make sure /etc/shells contains the full pathnames of the shells
    when it is generated

    ------

    APAR: IY18493 COMPID: 5765C3403 REL: 430
    ABSTRACT: DROP AIX 5.0 QOS MANAGER ENHANCEMENTS BACK IN AIX 4.3.

    PROBLEM DESCRIPTION:
    AIX 5.0 QoS Manager enhancements not available in AIX 4.3.

    PROBLEM CONCLUSION:
    Drop AIX 5.0 QoS Manager into AIX 4.3.

    ------

    APAR: IY18496 COMPID: 5765C3403 REL: 430
    ABSTRACT: MEMORY EXERCISER TERMINATES WITH AN ERROR 15

    PROBLEM DESCRIPTION:
    The memory exerciser terminates with an error 15

    PROBLEM CONCLUSION:
    Fix exerciser setup so it would not exit prematurely thereby
    generating software termination signal 15, SIGTERM

    ------

    APAR: IY18604 COMPID: 5765C3403 REL: 430
    ABSTRACT: PVT_DRAG-N-DROP IS MISS RENDERING ON THIN SELECTIONS.

    PROBLEM DESCRIPTION:
    The Motif 2.1 DNDDemo program has a drawing bug if you try to
    select a rectangle with very small height or width.

    PROBLEM CONCLUSION:
    DNDDemo program code fix.

    ------

    APAR: IY18605 COMPID: 5765C3403 REL: 430
    ABSTRACT: PAPI_READ_FAILED WITH TWO RETURN CODES, RC = 1 AND RC = 3.

    PROBLEM DESCRIPTION:
    PAPI_READ_FAILED when executing qos policy add,remove commands.

    PROBLEM CONCLUSION:
    Check for SS_ISCONNECTED in qosin_pcblookup.

    ------

    APAR: IY18607 COMPID: 5765C3403 REL: 430
    ABSTRACT: ASSERT((QH)->DSOCKETS_HEAD.NEXT == &((QH)->DSOCKETS_HEAD)) IN C

    PROBLEM DESCRIPTION:
    crash in assert in close_qsocket from QOS_RESV_UNLOCK macro.

    PROBLEM CONCLUSION:
    Move the the QOSPCB_UNREF from final_close_resv() to QOS_RESV_
    UNLOCK. After this we got another illegal memory access thi
    time in close_qsocket. The net_free was in qos_classify or
    qos_isconnected both in the QOS_RESV_UNLOCK call. This was
    fixed in QOS_RESV_UNLOCK by checkng the fetch_and_add before
    releasing the lock.

    ------

    APAR: IY18705 COMPID: 5765C3403 REL: 430
    ABSTRACT: DIAGNOSTICS: SUPPORT ULTRA3 SCSI ADAPTER

    PROBLEM DESCRIPTION:
    Support Ultra3 SCSI adapter card

    PROBLEM CONCLUSION:
    Update Test Unit and Diagnostic Application code

    ------

    APAR: IY18723 COMPID: 5765C3403 REL: 430
    ABSTRACT: NEED ADDITIONAL INFORMATION FOR RAID CONFIG ERRORS

    PROBLEM DESCRIPTION:
    Menu goal didn't give enough info

    ------

    APAR: IY18938 COMPID: 5765C3403 REL: 430
    ABSTRACT: MODIFY HELP ID FOR TRANSMIT QUEUE SIZE

    PROBLEM DESCRIPTION:
    Token-Ring Smit and WebSM help has Transmit queue size limits
    that do not correspond to actual transmit queue size limits.

    PROBLEM CONCLUSION:
    Modified Help text to reflect the current values.

    ------

    APAR: IY18939 COMPID: 5765C3403 REL: 430
    ABSTRACT: CSTOKDD - UNABLE TO DISABLE GROUP ADDRESS

    PROBLEM DESCRIPTION:
    Can not add additional group addresses

    PROBLEM CONCLUSION:
    Modifying code to accept addresses to delete.

    ------

    APAR: IY19217 COMPID: 5765C3403 REL: 430
    ABSTRACT: ALT_DISK_INSTALL MKSYSB CAN LEAVE INVALID DEVICES

    PROBLEM DESCRIPTION:
    alt_disk_install mksysb can leave invalid devices
    when restoing mksysb. This can cause problems such
    as a hang when mounting non-existent filesystems.

    LOCAL FIX:
    Manually remove the device.

    PROBLEM SUMMARY:
    Added routine to alt_disk_install that will clean up invalid
    devices.

    ------

    APAR: IY19274 COMPID: 5765C3403 REL: 430
    ABSTRACT: MKSECLDAP -C GOT ERROR AT THE FIRST TIME

    PROBLEM DESCRIPTION:
    Running the client set up the first time with a -u
    flag for some user produces an error.

    PROBLEM CONCLUSION:
    Added LDAP load module definition to the methods.cfg
    file before processing any users.

    ------

    APAR: IY19331 COMPID: 5765C3403 REL: 430
    ABSTRACT: DBX MAY HANG ON RERUN

    PROBLEM DESCRIPTION:
    dbx may hang if debugee calls dlopen() using a non-existent
    shared obect and debugee is rerun'd.

    LOCAL FIX:
    manually restart dbx; don't use rerun subcommand.

    PROBLEM CONCLUSION:
    in startprog(), pthreads_reset() was called only when going
    from a core debug session to a live debug session. Change
    this to always call pthreads_reset().

    ------

    APAR: IY19343 COMPID: 5765C3403 REL: 430
    ABSTRACT: SMIT WRONGLY HANDLES LONG NAMED LOCALE

    PROBLEM DESCRIPTION:
    Once the customer installed some languages environment included
    long name locales, (e.g. fr_FR.8859-15), smitty did not install
    all selected locales.

    PROBLEM CONCLUSION:
    need to change the source code of sm_mle.add:
    since original source code used command "cut -f1 -d'.'" to
    ignore the all languages after it, need to modifed code to
    make sure getting all selected languages.

    ------

    APAR: IY19355 COMPID: 5765C3403 REL: 430
    ABSTRACT: CCOD3B59 ERROR IN IY16645 WITH ATM ADAPTER

    PROBLEM DESCRIPTION:
    Chatm recover message is not decoded in errpt

    PROBLEM CONCLUSION:
    Fix packaging to ship msg file.

    ------

    APAR: IY19380 COMPID: 5765C3403 REL: 430
    ABSTRACT: 155 MBPS FIBER AND UTP ATM DIAGNOSTICS

    PROBLEM DESCRIPTION:
    Diagnostic Support for 155 Mbps UTP and MMF ATM PCI Adapters

    ------

    APAR: IY19384 COMPID: 5765C3403 REL: 430
    ABSTRACT: MKSECLDAP COMMAND DOES NOT SETUP SSL RIGHT

    PROBLEM DESCRIPTION:
    1.Even if version 4 gskit is installed, the mksecldap command
    still returns an error saying that gskit is not installed
    during LDAP server/client SSL setup.
    2. After SSL setup of a LDAP server (LDAP version 3.2 and
    above), the slapd server process can not be started.

    PROBLEM CONCLUSION:
    1. update the mksecldap command to handle the new version 4
    gskit fileset.
    2. update the mksecldap command to place the SSL key password
    to the right place in slapd32.conf

    ------

    APAR: IY19385 COMPID: 5765C3403 REL: 430
    ABSTRACT: DBX CAN FAIL WITH -F OPTION

    PROBLEM DESCRIPTION:
    DBX can suffer a segmentation fault upon entry when the -F
    option is used.

    PROBLEM CONCLUSION:
    Removed code that attempted to close a file descriptor that
    has already been closed.

    ------

    APAR: IY19444 COMPID: 5765C3403 REL: 430
    ABSTRACT: DIAG -C -DSCRAID0 -T"ADAPELA"

    PROBLEM DESCRIPTION:
      The customer will see a message stating that the PCI SCSI
      RAID service aid for adapter error log was unable to write
      to a file.

    PROBLEM CONCLUSION:
      Whenever the service aid is run with no console mode, do not
      print a message about a failure in writing to a file.

    ------

    APAR: IY19555 COMPID: 5765C3403 REL: 430
    ABSTRACT: LID_PUT() NEEDS TO HANDLE ERRORS FROM LSP_GET_ALL() BETTER

    PROBLEM DESCRIPTION:
    Add some error checking to an IKE tunnel update.

    ------

    APAR: IY19557 COMPID: 5765C3403 REL: 430
    ABSTRACT: PERMISSIONS ON TASK GIFS ARE 777

    PROBLEM DESCRIPTION:
    Gif files are installed with 777 permissions.

    PROBLEM CONCLUSION:
    Change the .il file to have lesser permissions on the
    problematic files.

    ------

    APAR: IY19770 COMPID: 5765C3403 REL: 430
    ABSTRACT: UNTRANSLATION YES AND NO IN ENT0 PROPERTY

    PROBLEM DESCRIPTION:
    Yes and No are not translated in WebSM.

    PROBLEM CONCLUSION:
    Added Yes and No translation to ODM.

    ------

    APAR: IY19781 COMPID: 5765C3403 REL: 430
    ABSTRACT: 40X CD-ROM SCSI VIOLATION WITH PPR MESSAGES

    PROBLEM DESCRIPTION:
    Inability to install from or configure a 40X CD-ROM drive when
    attached to a SCSI adapter which uses parallel protocol
    request messages for negotiation.

    PROBLEM CONCLUSION:
    This is a temporary software workaround until update 40X FW
    code is made available. The solution involves preventing the
    adapter from using parallel protocol request messages for
    SCSI data transfer negotiation.

    ------

    APAR: IY19814 COMPID: 5765C3403 REL: 430
    ABSTRACT: CANT GET TO DIAG FUNCTION SELECTION MENU.

    PROBLEM DESCRIPTION:
      Usage error when running diag without any command line flags.

    PROBLEM CONCLUSION:
      The problem occured, because a single variable stored the
      value for whether or not advanced diagnostics was included.
      To resolve the problem, the values for the run time option
      and the command line flag (-A) had to be distinguished.
      By doing so, the usage error about specifying a device
      now occurs only when the -A flag is explicitly set.
      Similarly, a single variable stored the value for whether or
      not error log analysis was included. This was fixed by
      distinguishing between the value from the run time options
      and the value from the command line.

    ------

    APAR: IY20021 COMPID: 576565400 REL: 220
    ABSTRACT: FILEPLACE SHOWS WRONG OUTPUT FOR SPARSE FILES ON LARGE FS

    PROBLEM DESCRIPTION:
    fileplace shows wrong output for sparse files when a sparse
    file is appended and its size goes beyond 4MB.
    This happens with sparse files on Large Enabled filesystem.

    PROBLEM CONCLUSION:
    fixed the code to calculate the no. of fragments correctly
    if the size of sparse file is greater than 4MB.

    ------

    APAR: IY20041 COMPID: 5765C3403 REL: 430
    ABSTRACT: TIME/DATE NOT SUPPORTED IN PROTOCOL FAMILY

    PROBLEM DESCRIPTION:
    Cust will get the unexpected error message from date
    command

    PROBLEM CONCLUSION:
    Checking the value of errno for EPROTONOSUPPORT in correct
    way in service mode

    ------

    APAR: IY20043 COMPID: 5765C3403 REL: 430
    ABSTRACT: WRONG COLOR WITH SECONDARY COLOR IN 8BIT COLOR BUFFER

    PROBLEM DESCRIPTION:
    When rendering texture mapped primitives to a 8-bit (332)
    color component drawable with seperate specular color or
    secondary color enabled and dithering enabled, color
    components can overflow on devices using "Soft" OpenGL.
    Therefore, this problem only occurs on the GXT250 (high res,
    1280x1024), GXT150L, GXT150M, and GT3i.

    PROBLEM CONCLUSION:
    Correct color sum fragment processing to clamp to a reduced
    color range so that dithering will not produce an overflow.

    ------

    APAR: IY20108 COMPID: 576565400 REL: 220
    ABSTRACT: TPROF HOGS CPU

    PROBLEM DESCRIPTION:
    tprof hogs CPU when executed in real time mode and makes the
    system appearently in hung condition.
    Once the tprof execution is finished, system behaves normal.
    This problems occurs on uni-processor machines and when
    tprof is run on CPU intensive application.

    PROBLEM CONCLUSION:
    fixed the code to set the priority of tprof & running
    application correctly.

    ------

    APAR: IY20145 COMPID: 5765C3403 REL: 430
    ABSTRACT: MODIFY SMIT HELP ID'S FOR TRANSMIT QUEUE SIZE

    PROBLEM DESCRIPTION:
    Token-Ring Smit and WebSM help has Transmit queue size limits
    that do not correspond to actual transmit queue size limits.
    16-16384

    PROBLEM CONCLUSION:
    Modified Help text to reflect the current values.

    ------

    APAR: IY20155 COMPID: 5765C3403 REL: 430
    ABSTRACT: DBCS NUMBERS SHOW AS ALPHABETS

    PROBLEM DESCRIPTION:
    DBCS numbers are confused with DBCS alphabetic
    For example, DBCS '0' will be treated as DBCS 'P'.

    PROBLEM CONCLUSION:
    Calling setlocale in prntsrvr will solve this problem

    ------

    APAR: IY20205 COMPID: 5765C3403 REL: 430
    ABSTRACT: XTABLET TEST NOT WORKING PROPERLY AFTER CHANGEPOINTER ACTIVATED

    PROBLEM DESCRIPTION:
    Maximum axis values returned from XListInputDevices might be
    incorrect if previous application ran changed pointer device
    to tablet then back to mouse.

    PROBLEM CONCLUSION:
    Re-initialize axis values of tablet after changing pointer
    device back to mouse.

    ------

    APAR: IY20213 COMPID: 5765C3403 REL: 430
    ABSTRACT: FIX FOR PDIAGEX EXTRA CALLS TO THE SECOND LEVEL INTR HANDLER

    PROBLEM DESCRIPTION:
    customer may see SYSINTR ERR in the error log while running
    the diagnostics

    PROBLEM CONCLUSION:
    Implement the fix to prevent the extra calls to the second
    level interrupt handler

    ------

    APAR: IY20391 COMPID: 5765C3403 REL: 430
    ABSTRACT: MENU TITLE FOR ELA DISPLAYS TESTING

    PROBLEM DESCRIPTION:
      The customer will be unaware of the fact that Error Log
      Analysis is occuring on the floating point processor while
      running diagnostics in Problem Determination or ELA mode.

    PROBLEM CONCLUSION:
      Display a menu stating ANALYZING ERROR LOG before the menu
      stating TESTING in Problem Determination mode. Changed the
      menu that was displayed in ELA mode from the one stating
      TESTING to the menu stating ANALYZING ERROR LOG.

    ------

    APAR: IY20392 COMPID: 5765C3403 REL: 430
    ABSTRACT: FRU DESCRIPTION INACCURATE FOR SOFTWARE ON SRN 823-134

    PROBLEM DESCRIPTION:
      Here is what the customer will see:
    823-134: Unable to open the device.
      sio0 FRU: n/a 00-00
                   Use the SRN to determine the FRU part number.
      Software FRU: n/a 00-00-0M
                       External Diskette Drive

    PROBLEM CONCLUSION:
      The problem occured, because the diagnostic controller was
      using the right catalog but the wrong message set. By
      changing which message set is used, an appropriate message
      about software rather than an external diskette drive was
      displayed.

    ------

    APAR: IY20397 COMPID: 576565400 REL: 220
    ABSTRACT: TOPAS CORE DUMP

    PROBLEM DESCRIPTION:
    topas or any local consumer will Segfault and coredump
    when it tries to monitor more than 2000 processes

    PROBLEM CONCLUSION:
    Fixed the code for process instantiation such that it can
    handle > 4000 processes and also fixed the memory handler
    of the library to check the Boundry.

    ------

    APAR: IY20433 COMPID: 5765C3403 REL: 430
    ABSTRACT: BUG 4960-5(CRYPTO) RETURNS 1CC-101 NOT LISTED IN DIMBS

    PROBLEM DESCRIPTION:
    An incorrect number (1CA) will be displayed in the LED for
    the IBM Cryptographic Coprocessor (Feature Code 4960) during
    boot. Also an incorrect SRN (1CC) may appear in diagnostic
    error reports.

    PROBLEM CONCLUSION:
    Update the ODM with the correct LED value.

    ------

    APAR: IY20460 COMPID: 5765C3403 REL: 430
    ABSTRACT: ICA DEVICE 1410E601 UN-NATURAL ADDR IO

    PROBLEM DESCRIPTION:
    There is a memory read problem on unnatual memory addresses.
    The random number generator will fail on the
    IBM Cryptographic Acclerator adapter.
    Attempts to read via from attached memory to a
    16bit register on the 1410e601 IBM Cryptographic
    Accelerator adapter
    (memory attach + 1146), this garbled the register value.

    PROBLEM CONCLUSION:
    Read on a natual 32bit word aligned address and shift bits
    accordingly.

    ------

    APAR: IY20561 COMPID: 5765C3403 REL: 430
    ABSTRACT: IMPROVE PCI BUS I/O PERFORMANCE FOR DIAGNOSTICS DRIVER

    PROBLEM DESCRIPTION:
    The customer may experience Diagnostic failures in some
    adapters

    PROBLEM CONCLUSION:
    Will use different routines for load/store operations

    ------

    APAR: IY20564 COMPID: 5765C3403 REL: 430
    ABSTRACT: 43X DOC SEARCH SERVER

    PROBLEM DESCRIPTION:
    Users will see a message that the documentation library
    service could not be configured if the Liteserver is not
    installed when installing bos.docsearch.rte.

    PROBLEM CONCLUSION:
    Check if the Liteserver is installed before attempting
    to configure docsearch to use it.

    ------

    APAR: IY20567 COMPID: 5765C3403 REL: 430
    ABSTRACT: WRONG VRMF IN BOS IMAGE FOR BOS.RTE FILESETS

    PROBLEM DESCRIPTION:
    During BOS installs updates may unnecessarily install, or
    installs may fail due to incorrect bos.rte vrmf levels
    in the BOS image.

    PROBLEM CONCLUSION:
    For every bos.rte subsystem that has been updated to 4.3.3.75,
    change the fileset.add file, so that the bos image reflects
    the correct vrmf.

    ------

    APAR: IY20586 COMPID: 5765C3403 REL: 430
    ABSTRACT: NFSD STARTUP FAILS/SYSTEM MAY CRASH

    PROBLEM DESCRIPTION:
    This happens when nfs_socksize and/or nfs_tcp_socksize is set
    greater than sb_max. nfs_socksize and nfs_tcp_socksize should
    not be greater than sb_max and if that is so then nfsd will not
    start. But that is OK. The issue is it does not cleanup
    properly while exiting.
    This leads to nfs port being left open. Also if the nfsd was
    started in tcp mode (nfsd -p tcp) and nfs_socksize and
    nfs_tcp_socksize is greater then sb_max then when it fails to
    start it tries to cleanup udp sessions. This causes NULL
    pointer call and crashes the system.

    LOCAL FIX:
    Never set nfs_socksize and nfs_tcp_socksize greater
    than sb_max.

    PROBLEM SUMMARY:
    nfsd will not start also the system might crash if customer is
    starting nfsd with the option "-p tcp".

    PROBLEM CONCLUSION:
    Make sure we close nfs port if nfsd fails to startup. Also
    check if udp sessions has been started before cleaning up the
    udp sessions.

    ------

    APAR: IY20611 COMPID: 5765C3403 REL: 430
    ABSTRACT: CANNOT SET BREAKPOINTS IN SYSTEM CALLS

    PROBLEM DESCRIPTION:
    DBX can fail to set breakpoints in the glue code of glink
    imported functions. This symptom will only be observed when
    DBX has processed one or more CPP modules.

    PROBLEM CONCLUSION:
    Removed code that prevented glink functions from being
    properly resolved during breakpoint assignments.

    ------

    APAR: IY20613 COMPID: 5765C3403 REL: 430
    ABSTRACT: COPY TRACE TABLE OPTION NOT FUNCTIONING ON STANDALONE CD.

    PROBLEM DESCRIPTION:
      A SCSI CHECK CONDITION error occurs before the trace table
      gets copied.

    PROBLEM CONCLUSION:
      The copy trace table tries to extract and
    copy data to diskette in chunks of 1MB. If the adapter
    (through which the data flows to the device) capacity
    is less than 1MB than the extraction fails. The problem can
    be rectified by extracting data in chunks of adapter capacity
    until all the data is read and then transferring it to the
    diskette.

    ------

    APAR: IY20615 COMPID: 5765C3403 REL: 430
    ABSTRACT: COPY TRACE TABLE: SEGMENT NUMBER DECREASES BY 1 ON BAD/NO MEDIA

    PROBLEM DESCRIPTION:
      A SCSI BUSY error occurs when trying to copy the trace
      table to a good diskette after failed attempts.

    PROBLEM CONCLUSION:
      Keep the segment number to be copied to diskette the same
      when a failure occurs.

    ------

    APAR: IY20619 COMPID: 5765C3403 REL: 430
    ABSTRACT: DIAG TRACE FILE NOT CORRECT WHEN IN LOOP MODE

    PROBLEM DESCRIPTION:
    If diagnostic trace is activated, a reliable trace file may
    not be produced for the ATM adapter when running diagnostics.

    PROBLEM CONCLUSION:
    Modify code to ensure a good trace file is produced.

    ------

    APAR: IY20620 COMPID: 5765C3403 REL: 430
    ABSTRACT: REMOVE WORKAROUND FOR 40X CD-ROM SCSI VIOLATION

    PROBLEM DESCRIPTION:
    If a device rejects a parallel protocol request message during
    SCSI negotiation, the driver will no longer use that message.
    This can decrease overall data transfer speed/performance.

    PROBLEM CONCLUSION:
    Remove logic in the driver which currently prevents the
    SCSI adapter from using parallel protocol request messages
    for data transfer negotiation if the message is rejected.

    ------

    APAR: IY20660 COMPID: 5765D5101 REL: 121
    ABSTRACT: HAGSGLSM ERRORS IN ML0MEMBERSHIP

    PROBLEM DESCRIPTION:
    hagsglsm slow to generate the ml0membership.

    PROBLEM SUMMARY:
    Group Services Switch Membership daemon (HAGSGLSM)
    incorrectly handles the signals particularly
    during the Group Services ha_gs_dispatch() execution.
    For example, when a notification is arrived,
    the main thread in HAGSGLSM first blocks all signals
    (including "undefined/bad/interal traps") except SIGSEGV
    and calls ha_gs_dispatch(). Because of that,
    any signals can not be handled until the thread
    leaves ha_gs_dispatch(). Furthermore, blocking the
    undefined/bad/internal traps makes process unexpected
    from time to time noticed by experiment .
    Therefore, the fix to this behavior can be done by
    creating a dedicated thread for the signal handler.
    The signal handler will catch all asynchrous signals
    by masking out only defined asynchronous signals.

    PROBLEM CONCLUSION:
    The signals will be properly handled and thus hagsglsm
    daemon should not behave abnormally in terms of 'signals'.
    It also will not slow to form the memberships
    (css/ml0Membership) because of the slow processing
    of the signals.

    ------

    APAR: IY20682 COMPID: 5765C3403 REL: 430
    ABSTRACT: SMITTY NEEDS TO REPORT PIN ERRORS

    PROBLEM DESCRIPTION:
    If they set a pin that is larger than 8 chars and then attempt
    to use this pin they will find that they must use the previous
    pin because the new pin did not get saved.

    PROBLEM CONCLUSION:
    fix code
    workaround: do not enter pins longer than 8 or less that 4

    ------

    APAR: IY20689 COMPID: 576565400 REL: 220
    ABSTRACT: FILEMON ERROR - COULD NOT OPEN DEVICE

    PROBLEM DESCRIPTION:
    filemon gives the following error :
    "could not open device '<major=..,minor=..>' "

    PROBLEM SUMMARY:
    filemon sometimes gives following errors -
    "could not open device '<major=..,minor=..>' ".

    PROBLEM CONCLUSION:
    Fixed the filemon code to recognize device ids properly.

    ------

    APAR: IY20704 COMPID: 5765C3403 REL: 430
    ABSTRACT: ATM SIGNALING DAEMON ENHANCEMENTS.

    PROBLEM DESCRIPTION:
    ATM Signaling Daemon Enhancements.

    PROBLEM CONCLUSION:
    Changes for ATM Signaling Daemon Enhancements.

    ------

    APAR: IY20708 COMPID: 5765C3403 REL: 430
    ABSTRACT: MAN HANGS IN REMOTE SEARCH

    PROBLEM DESCRIPTION:
    When DOCUMENT_SERVER_MACHINE_NAME is set to an aliase name
    of the local host, man command tries for remote search and
    reports,
    Remote search is in progress..
    Remote search failed.
    Searching locally...
    and finally displays the command.

    PROBLEM CONCLUSION:
    man code was modified to take account of the aliases
    of the hosts. Also a new flag '-r' is added
    (for 43 release only as this flag is already present
    in 5.0 release onwards) which has to be specified with
    man command for remote search.

    ------

    APAR: IY20709 COMPID: 5765C3403 REL: 430
    ABSTRACT: TRCRPT REPORT INCORRECT SCB BIT STATUS.

    PROBLEM DESCRIPTION:
    trcrpt report incorrect scb bit status

    PROBLEM CONCLUSION:
    fix vmmsibits template

    ------

    APAR: IY20711 COMPID: 5765C3403 REL: 430
    ABSTRACT: SERVICE AID NEEDS TO SUPPORT RAW MODE WHEN DISPLAYING MICROCODE

    PROBLEM DESCRIPTION:
    When displaying microcode for some type of backplanes, it
    returns download microcode is not supported

    PROBLEM CONCLUSION:
    Display the microcode level for all backplanes

    ------

    APAR: IY20718 COMPID: 5765C3403 REL: 430
    ABSTRACT: FCPARRAY: ADD SUPPORT FOR SYMSM AND V4 ARRAYS

    PROBLEM DESCRIPTION:
    Limited functionality with LSI's SymSM GUI; unable
    to use V4 LSI arrays.

    PROBLEM CONCLUSION:
    Add support for access to both disk array controllers' UTM
    luns via /dev/dacN. Add config changes for V4 arrays and new
    uniquetype array/fcp/ibm-dac-V4.

    ------

    APAR: IY20719 COMPID: 5765C3403 REL: 430
    ABSTRACT: FLUTE: IOCTL PROCESSING PROBLEM

    PROBLEM DESCRIPTION:
    fcparray devices will not configure properly

    PROBLEM CONCLUSION:
    Make sure we only retarget buffer I/O for UTM

    ------

    APAR: IY20742 COMPID: 5765C3403 REL: 430
    ABSTRACT: ASSERT ON IPL IN ECP_CONFIG.

    PROBLEM DESCRIPTION:
    System crashes on boot in ecp_unit_dtor.

    PROBLEM SUMMARY:
    Fixed error path.

    ------

    APAR: IY20751 COMPID: 5765C3403 REL: 430
    ABSTRACT: SCSI TAPE DIAGNOSTICS DOES NOT DISPLAY INITIAL STARTUP SCREEN

    PROBLEM DESCRIPTION:
    Unable to detect when test begins.

    PROBLEM CONCLUSION:
    added a stand by menu.

    ------

    APAR: IY20752 COMPID: 5765C3403 REL: 430
    ABSTRACT: LSDEV WILL SHOW LSI-DERIVED DISK ARRAYS AS 2102-F10

    PROBLEM DESCRIPTION:
    All LSI-derived disk arrays will show up as 2102-F10
    in lsdev

    PROBLEM CONCLUSION:
    Match the inquiry string to controller type in cfgfdac, then
    write out the correct message_no and ctrl_type. Read ctrl_type
    in cfgfdar and use it to set message_no for dars and hdisks.
    Update message catalogs with strings for all current types.

    ------

    APAR: IY20754 COMPID: 5765C3403 REL: 430
    ABSTRACT: XTERM ENTRY IN UTMP IS NOT CORRECT

    PROBLEM DESCRIPTION:
    The entry created by xterm in the utmp file is not correct.

    PROBLEM CONCLUSION:
    Fix xterm to use the correct ut_id in the utmp entry.

    ------

    APAR: IY20759 COMPID: 5765C3403 REL: 430
    ABSTRACT: NFS MEMORY MAPPING PROBLEM FOR FILE SIZE MORE THAN 2GB

    PROBLEM DESCRIPTION:
    nfs memory mapping problem for file size more than 2gb

    PROBLEM CONCLUSION:
    checking the value of offset and length as a unsigned
    long.In past , value of offest and length is taken as
    unsigned long value.So file's offset and length is not able
    to calculate correctly over file size of 2GB and returns
    ENXIO error.ENXIO means no such device and files.

    ------

    APAR: IY20773 COMPID: 576565400 REL: 220
    ABSTRACT: PERFAGENT.TOOLS.README: INFORMATION ON TCBCHK ERRORS INCLUDED

    PROBLEM DESCRIPTION:
    Change in the perfagent.tools.README file to
    include the information on tcbchk errors and how
    to resolve.

    PROBLEM SUMMARY:
    Information about resolving the tcbchk errors for
    /etc/*hook* files generated by PTX , to be included in
    README.perfagent.tools file

    PROBLEM CONCLUSION:
    Added note in README.perfagent.tools file with the steps to
    resolve the errors.

    ------

    APAR: IY20776 COMPID: 5765C3403 REL: 430
    ABSTRACT: PDT REPORT: INCREASING DISK LIMIT TO 4096

    PROBLEM DESCRIPTION:
    PDT reports error when no. of disks exceed 2048.

    PROBLEM CONCLUSION:
    Increased the disk limit to 4096

    ------

    APAR: IY20780 COMPID: 576565400 REL: 220
    ABSTRACT: GENNAMES -F OPTION

    PROBLEM DESCRIPTION:
    gennames catches Filesystem and Volume information when it
    is not actually used by certain performance tools.

    PROBLEM CONCLUSION:
    Fixed the code such that by default gennames won't capture
    Filesystem / Volume information, and it capture those
    information when a user specify -f option in gennames.

    ------

    APAR: IY20819 COMPID: 5765C3403 REL: 430
    ABSTRACT: MEMORY LEAK IN GENMAJOR

    PROBLEM DESCRIPTION:
    Customer could see memory issues.

    PROBLEM CONCLUSION:
    Handled the mallocs in the routines.

    ------

    APAR: IY20820 COMPID: 5765C3403 REL: 430
    ABSTRACT: DBX DOESN'T SHOW CONST AND VOILATILE CLASS MEMBERS

    PROBLEM DESCRIPTION:
    Members of struct's and classes after enum declaration ignored

    PROBLEM CONCLUSION:
    An extraneous return statement

    ------

    APAR: IY20828 COMPID: 5765C3403 REL: 430
    ABSTRACT: UPDATE README FOR SNAPP ON 433

    PROBLEM DESCRIPTION:
    The README for the SNAPP functionality is not the latest
    version.

    PROBLEM CONCLUSION:
    Drop the latest version of the README.

    ------

    APAR: IY20829 COMPID: 5765C3403 REL: 430
    ABSTRACT: DIAGNOSTICS: SERVICE AID FOR SNAP

    PROBLEM DESCRIPTION:
    New diagnostic service aid invoking snap command

    PROBLEM CONCLUSION:
    Provide new diagnostic service aid

    ------

    APAR: IY20830 COMPID: 5765C3403 REL: 430
    ABSTRACT: LSMCODE SUPPORT FOR DVDRAM

    PROBLEM DESCRIPTION:
    The lsmcode command will not list the microcode level of
    DVDRAM drives, even though they support the operation.

    PROBLEM CONCLUSION:
    Alter lsmcode.c to check AIX Subtype if Class is CDROM to
    distinguish DVDRAM from CDROM-only devices (which don't
    support disply of microcode).

    ------

    APAR: IY20843 COMPID: 5765C3403 REL: 430
    ABSTRACT: FILE COMMAND SHOULD RECOGNISE AIX CORE FILES

    PROBLEM DESCRIPTION:
    file command unable to recognise AIX core files.
    Outputing default type- International Data or Text.

    PROBLEM SUMMARY:
    File command not able to recognise AIX corefiles.

    PROBLEM CONCLUSION:
    This APAR enables 'file' command to recognize AIX corefiles.

    ------

    APAR: IY20853 COMPID: 5765C3403 REL: 430
    ABSTRACT: DO NOT IMPORT CONCURRENT VOLUME GROUPS ON REINSTALL

    PROBLEM DESCRIPTION:
    bos_install attempts to import volume groups which are
    concurrent and fails.

    PROBLEM SUMMARY:
    When installing a mksysb or doing a preservation install, if
    there are concurrent volume groups, the install can cause
    the volume groups to be inaccessible by other systems.

    PROBLEM CONCLUSION:
    Do not import concurrent volume groups at install time.

    TEMPORARY FIX:
    Export concurrent volume groups on the system when creating
    backups or disconnect the concurrent volume groups on install
    of a system.

    ------

    APAR: IY20862 COMPID: 5765C3403 REL: 430
    ABSTRACT: INCORRECT PART NUMBER DISPLAYED FOR TAPE TEST MEDIA

    PROBLEM DESCRIPTION:
    Unable to see test tape part number in menuguoal

    PROBLEM CONCLUSION:
    In disp_mng, look for the part number and copied it to the
    menugoal buffer

    ------

    APAR: IY20867 COMPID: 5765C3403 REL: 430
    ABSTRACT: SECURITY: VULNARIBILITY IN _GETLAYOUTINITFUNC()

    PROBLEM DESCRIPTION:
    There is a possible security flaw in the layout service
    library that might allow a malicious user to gain
    root authority by manipulating the LANG environment
    variable

    PROBLEM CONCLUSION:
    Implement bounds checkout in GetLayoutInitFunc() so that
    overrunning the buffer is not possible...

    ------

    APAR: IY20868 COMPID: 5765C3403 REL: 430
    ABSTRACT: XLATE/XLATE64 BROKEN FOR STRIPED MIRRORS

    PROBLEM DESCRIPTION:
    Incorrect value returned from XLATE ioctl for striped
    mirrored logical volumes.

    PROBLEM CONCLUSION:
    Check for striped mirrored scheduling policy and use the
    mirror copy passed in from the caller.

    ------

    APAR: IY20883 COMPID: 5765C3403 REL: 430
    ABSTRACT: FC ADAPTER MAY BECOME INACCESSIBLE AFTER A PCI BUS ERROR.

    PROBLEM DESCRIPTION:
    FC adapter may become inaccessible after a PCI bus error.

    PROBLEM CONCLUSION:
    Alter adapter reset error recovery path.

    ------

    APAR: IY20903 COMPID: 5765C3403 REL: 430
    ABSTRACT: MULTIPLE APPLICATIONS CANNOT USE ICA AT THE SAME TIME.

    PROBLEM DESCRIPTION:
    under stress conditions of running des and random number
    generation, the ICA adapter will hang and not
    become usable until after reboot.

    PROBLEM CONCLUSION:
    have ica_random use the same des cmd lock

    ------

    APAR: IY20929 COMPID: 5765C3403 REL: 430
    ABSTRACT: MAKE COMMAND DUMPING CORE FOR A MISSING ')' IN THE TARGET FILE

    PROBLEM DESCRIPTION:
    make command dumping core for a missing ')' for the target
    in the Makefile" It should generate a error message instead
    of dumping core.

    PROBLEM SUMMARY:
     make command dumping core for a missing ')' in the Makefile

    PROBLEM CONCLUSION:
    code has been modified to fix the reported problem.

    ------

    APAR: IY20964 COMPID: 5765C3403 REL: 430
    ABSTRACT: RPC.MOUNTD CORE DUMPS

    PROBLEM DESCRIPTION:
    rpc.mountd core dumps. The stack traces recorded in the
    the errpt -a will show stack traces similar to
    splay A8
    free_y 11C
    free 104
    nisfree 58
    ho_byname 2C
    ho_byname 254
    gethostby 198
    gethostby 58
    mount 1C8
    mnt DC
    _svc_prog 1E8
    _svc_run_ 3A8
    svc_run 50
    main 558
    __start 8C
    OR
    free_y 2E4
    free_y 11C
    free 104
    _spinlock 1C
    _rec_mute AC
    fclose 11C
    endexport 18
    findentry 160
    mount 364
    mnt DC
    _svc_prog 1E8
    _svc_run_ 3A8
    _pthread_ D4
    The important part to note is the lower portion of the
    stack trace, ie,
    ...
    mount 364
    mnt DC
    _svc_prog 1E8
    _svc_run_ 3A8
    _pthread_ D4
    If this happens to be the case then this APAR will most
    likely fix the issue.

    PROBLEM SUMMARY:
    customer will have difficulty with nfs mounts.

    PROBLEM CONCLUSION:
    The idea is to undo the changes made in APAR IY12759, and
    implement it in development stream later if required,
    (it is more of a feature and requires extensive testing).

    ------

    APAR: IY21048 COMPID: 5765C3403 REL: 430
    ABSTRACT: CLOSE POSSIBILITY FOR SCSI CARDS TO READ STALE DATA

    PROBLEM DESCRIPTION:
    PowerPC architecture defines a storage model for ordering of
    storage accesses that allows for improved performance. However,
    this model requires that software explicitly order accesses to
    storage that is shared by threads or IO devices. Otherwise
    there exists the possibility for threads or IO devices to read
    stale data. Newer and faster systems that use this storage
    model extensively can surface these problems. After review of
    the SCSI adapter driver code, we have identified a few areas
    that need some ordering to prevent such problems.

    PROBLEM CONCLUSION:
    Target specific routines in the SCSI adapter device driver
    that require storage access to be ordered.

    ------

    APAR: IY21053 COMPID: 5765C3403 REL: 430
    ABSTRACT: DIAGNOSTIC CHANGES TO SUPPORT LACK OF LINE IN/OUT JACKS

    PROBLEM DESCRIPTION:
      Customer is prompted with messages about line-in/out jacks
      even though the system may not have those jacks.

    PROBLEM CONCLUSION:
      Menus make no mention of line-in/out jacks once the
      customer has stated that the system does not have them.
      The Test Units relying on line-in/out jacks are bypassed.

    ------

    APAR: IY21056 COMPID: 5765C3403 REL: 430
    ABSTRACT: SYSTEM CRASH DURING RMDEV OF ATM DEVICE

    PROBLEM DESCRIPTION:
    machine crash when try to rmdev ATM driver

    PROBLEM CONCLUSION:
    make sure recieve mpx close before remove code.

    ------

    APAR: IY21057 COMPID: 5765C3403 REL: 430
    ABSTRACT: GRAMATICAL ERROR IN MESSAGE FILE

    PROBLEM DESCRIPTION:
    Gramatical error in DVD-RAM Service Aid message file

    PROBLEM CONCLUSION:
    Small message file change

    ------

    APAR: IY21072 COMPID: 5765C3403 REL: 430
    ABSTRACT: ZAPDB NEEDS GUI UPDATE AND BINARY TRACE FUNCTIONALITY

    PROBLEM DESCRIPTION:
    Lack of gui output and inability to trace large executables.

    PROBLEM CONCLUSION:
    Add gui panels and add binary trace functionality.

    ------

    APAR: IY21080 COMPID: 5765C3403 REL: 430
    ABSTRACT: STICKY BIT NOT KEPT ON RESTORE FROM TAR ARCHIVE

    PROBLEM DESCRIPTION:
    Sticky bit set for directories is not kept when
    the directories are restored from the tar archive.

    PROBLEM SUMMARY:
    Sticky bit not kept on restore from tar archive.

    PROBLEM CONCLUSION:
    tar command's code is modified to resolve the problem.

    ------

    APAR: IY21095 COMPID: 5765C3403 REL: 430
    ABSTRACT: DEADLOCK IN PTY_INIT_SOCKET

    PROBLEM DESCRIPTION:
    The pty_init_socket routines holds the socket lock while
    it tries to acquire the pty lock. This is out of
    order and leads to a dead lock. The stack trace for all
    of the cpu's will show one cpu in pty_init_socket while
    another cpu will be, probably in an interrupt context,
    trying to acquire the socket lock but will already have
    the pty lock in the interrupted context.

    PROBLEM SUMMARY:
    The pty_init_socket routines holds the socket lock while
    it tries to acquire the pty lock. This is out of
    order and leads to a dead lock. The stack trace for all
    of the cpu's will show one cpu in pty_init_socket while
    another cpu will be, probably in an interrupt context,
    trying to acquire the socket lock but will already have
    the pty lock in the interrupted context.

    PROBLEM CONCLUSION:
    The pty lock is gained before the socket lock.

    ------

    APAR: IY21102 COMPID: 5765C3403 REL: 430
    ABSTRACT: IOCP SYSTEM DUMP

    PROBLEM DESCRIPTION:
    The iocp.ext can access a socket which has already been closed
    sometimes. This will usually lead to a system panic with the
    stack trace looking like:
    TACK TRACE:
       .so_pin_more_stack () f1679760
       .sosleep () f1679db0
       .soreceive () f1679e10
       . iocp.ext:iocp_recv_data_socket () f1679ea0
       . iocp.ext:iocp_process_completion_pkt () f1679ef0
       . iocp.ext:GetQueuedCompletionStatus () f1679f40
       .sys_call_ret () f1679fc0

    PROBLEM CONCLUSION:
    The queue of requests that have already completed but not have
    not yet been process has the references to the closed socket
    as well as the iocfd structure that has been freed as well.
    The solution is to run through the list completed requests
    when the request to close the socket comes down, clear any
    reference to the iocfd before freeing it, and let the socket
    close.

    ------

    APAR: IY21109 COMPID: 5765C3403 REL: 430
    ABSTRACT: SMIT: CHANGESHARE DOES NOT DISTINGUISH TEST VS ATEST2

    PROBLEM DESCRIPTION:
    Change Share panels (both File and Print) don't
    distinguish between TEST and ATEST2 as a sharename --
    if TEST is selected in the selection-box, then
    TEST, ATEST2, or xxxTESTyyy may be selected.
    Need to add SMIT code to require an exact match of the name.

    PROBLEM CONCLUSION:
    Fix the SMIT scripts to do EXACT match instead of SUB-STRING
    matching of share-names.

    ------

    APAR: IY21114 COMPID: 5765C3403 REL: 430
    ABSTRACT: NEED TO ZERO OUT PID LIST WHEN INITIALIZE IS CALLED

    PROBLEM DESCRIPTION:
    Applications forking many children that call C_Initialize
    will encounter an error (CKR_FUNCTION_FAILED) at about
    50 children.

    ------

    APAR: IY21120 COMPID: 5765C3403 REL: 430
    ABSTRACT: DSI IN SDLMONITOR

    PROBLEM DESCRIPTION:
    Prototyping on the OpnX trace entry causes the new compiler
    to extend the 4th parameter of the sdlmonitor() call into the
    5th and 6th parameters when they should have been zeroed.

    PROBLEM CONCLUSION:
    Correct the prototyping.

    ------

    APAR: IY21121 COMPID: 5765C3403 REL: 430
    ABSTRACT: SCENE NOT RENDERED AFTER WINDOW RESIZING AND REPOSITIONING

    PROBLEM DESCRIPTION:
    OpenGL shipped sample "list.c" doesn't work well under window
    resizes or moves.

    PROBLEM CONCLUSION:
    Fix test case to add defensive matrix manipulation code.

    ------

    APAR: IY21158 COMPID: 5765C3403 REL: 430
    ABSTRACT: CD DIAGS DOES NOT PROMPT FOR SPECIFIC SUPPLEMENTAL MEDIA

    PROBLEM DESCRIPTION:
      The customer will be prompted for supplemental media, but
      a device is not specified within the prompt.

    PROBLEM CONCLUSION:
      Whenever the user is processing supplemental media by way
      of running diagnostics, display the name of the expected
      device. If the user is running the "Process Supplemental
      Media" task, then do not display the name of a device.
      The resource description remains n/a on the resource menu
      and cannot be fixed within the diagnostic sub-system.

    ------

    APAR: IY21181 COMPID: 5765C3403 REL: 430
    ABSTRACT: LSPV COREDUMPS

    PROBLEM DESCRIPTION:
    ODM database will not be readbale, if customer updates
    the database when the filesystem on which ODM database resides
    is full. All the system databases reside on "/" filesytem.

    PROBLEM CONCLUSION:
    Error codes are not handled properly, when filesystem is full.

    ------

    APAR: IY21186 COMPID: 5765C3403 REL: 433
    ABSTRACT: ADDING STATIC ROUTES IN SMIT CALLS /ETC/METHODS/CFGIF

    PROBLEM DESCRIPTION:
    Using smit to add a static route will cause an interface
    (configured with ifconfig) to reconfigure to the value
    set in ODM.

    PROBLEM SUMMARY:
    Adding static route through smit was causing
    an interface (configured with ifconfig) to reconfigure its value
    set in ODM.

    ------

    APAR: IY21190 COMPID: 5765C3403 REL: 430
    ABSTRACT: ATM LAN EMULATION DOCUMENTATION - NOW SUPPORTS 622 MBPS

    PROBLEM DESCRIPTION:
    155mbps maximum rate now 622mbps.

    PROBLEM CONCLUSION:
    Correct 155mbps rate with new 622mbps rate.

    ------

    APAR: IY21191 COMPID: 5765C3403 REL: 430
    ABSTRACT: ATM LANE CANNOT CONFIGURE 622 ADAPTER SPEED ON VERSION 43

    PROBLEM DESCRIPTION:
    Version 43 LANE does not support the new ATM 622 adapters.

    PROBLEM CONCLUSION:
    Add LANE support for peak_rates of 622 Mbps.

    ------

    APAR: IY21201 COMPID: 5765C3403 REL: 430
    ABSTRACT: OSLEVEL PERFORMANCE PROBLEM

    PROBLEM DESCRIPTION:
    Oslevel takes longer to execute if there are no RMLs.

    PROBLEM CONCLUSION:
    Oslevel will not rebuild RML cache if there are no RMLs.

    ------

    APAR: IY21207 COMPID: 5765D5101 REL: 121
    ABSTRACT: HAGSGLSM DIES BECAUSE IT THINKS HAGSD DIED BUT IT DIDN'T

    PROBLEM DESCRIPTION:
    hagsglsm dies occasionally (exit code 5) because it thinks hags
    has died, when hags actually has not.

    PROBLEM SUMMARY:
    Group Services Switch Membership daemon (HAGSGLSM)
    incorrectly handles the return code of ha_gs_dispatch().
    The hagsglsm daemon exits if ha_gs_dispatch() != HA_GS_OK.
    However, there may be cases of ha_gs_dispatch() != HA_GS_OK
    even if the Group Services daemon did not die.
    The cases could happen if write() to socket fails from the
    library.
    The fix will
    1) remove all ambiguous errno handling in HAGS library
    (libha_gs.a).
    2) make HAGSGLSM to exit only when ha_gs_dispatch() ==
    HA_GS_NOT_OK
      in addition to the notification of HAGS daemon death.

    PROBLEM CONCLUSION:
    hagsglsmd daemon will not exit because of
    the return code of ha_gs_dispatch()!=HA_GS_OK,
    but will exit only when either it receives
    an notification of HAGS DAEMON STOPPED,
    or ha_gs_dispatch()==HA_GS_NOT_OK.

    ------

    APAR: IY21310 COMPID: 5765C3403 REL: 430
    ABSTRACT: FAILURE CREATING POSTSCRIPT QUEUES

    PROBLEM DESCRIPTION:
    There is a failure to create postscript queues from smit.

    PROBLEM CONCLUSION:
    The piodigest program has been fixed so postscript queues
    will be created.

    ------

    APAR: IY21349 COMPID: 5765C3403 REL: 430
    ABSTRACT: SYSTEM CRASHED WHEN DOING AN LVM I/O OPERATION

    PROBLEM DESCRIPTION:
    A system crash occurred on one 43 system when doing an
    LVM I/O operation.

    PROBLEM CONCLUSION:
    Check flags before referencing certain structures.

    ------

    APAR: IY21351 COMPID: 5765C3403 REL: 430
    ABSTRACT: BUSY ERROR WHEN MULTIPLE CRYPTOGRAPHIC ADAPTERS ARE INSTALLED

    PROBLEM DESCRIPTION:
    Busy error when multiple cryptographic cards are installed

    PROBLEM CONCLUSION:
    simple message change

    ------

    APAR: IY21355 COMPID: 5765C3403 REL: 430
    ABSTRACT: SCSI ADAPTER FAULT TOLERANT NEGOTIATION SCHEME

    PROBLEM DESCRIPTION:
    Inability to recognize and/or configure a SCSI device which
    does not properly adhere to the SCSI protocol standards with
    respect to SCSI extended messages. Typical scenario where
    this applies is with some SAFTE enclosure devices that do not
    properly handle extended messages used in synchronous
    negotiation. The SCSI adapter will attempt to negotiate with
    the device on the first outbound command (usually during
    device discovery at boot or config time), but be unable to
    complete the command because of a SCSI protocol violation.
    The user may see several SCSI bus errors in the system log
    and no device configured.

    PROBLEM CONCLUSION:
    Rework the SCSI adapter driver so that it is more fault
    tolerant of SCSI target devices. If a target SCSI device
    rejects the adapters extended message during negotiation, the
    driver will know not to use extended messages with that
    device. Thus, all data transfer with that particular device
    will happen at asynchronous, narrow transfer speed and width.

    ------

    APAR: IY21372 COMPID: 5765C3403 REL: 430
    ABSTRACT: UPDATE PCI FDDI ADAPTER FLASH TASK RETURNS ERROR MESSAGE

    PROBLEM DESCRIPTION:
    Error when using the standalone Diagnostics CD
    Getting an error saying " sh:/usr/sbin/restore: not found"

    PROBLEM CONCLUSION:
    The flash code was changed to use restbyname command instead
    of restore.

    ------

    APAR: IY21376 COMPID: 5765C3403 REL: 430
    ABSTRACT: BUSY ERROR ON ICA 2058 ADAPTER WHILE TESTING CRYPTO4758 ADAPTER

    PROBLEM DESCRIPTION:
    Diagnostic returns Device busy on cryptograhphic ICA 2058
    adapter when there is an application running on PCI 4758
    cryptographic adapter.

    PROBLEM CONCLUSION:
    Code was changed for both test unit on PCI 4758 and ICA 2058
    to correct this situation.

    ------

    APAR: IY21379 COMPID: 5765C3403 REL: 430
    ABSTRACT: IP TYPE TOC FRAGMENTS CAUSE INSTALL FAILURE

    PROBLEM DESCRIPTION:
    Attempting to install from a PTF package that contains an IP
    fragment in the .toc file results in an error similar to the
    following.
    installp: The format of the toc file is invalid.
    installp: The package name field required in a 4.1-formatted
              package is missing from the lpp_name file.

    PROBLEM CONCLUSION:
    Ignore IPP entries in the load toc functions in inu_toc.

    TEMPORARY FIX:
    Edit the .toc file and remove entries that have the following
    format.
    filename 4 R IP {
    }

    ------

    APAR: IY21380 COMPID: 5765C3403 REL: 430
    ABSTRACT: STATD SLOW AT STARTUP

    PROBLEM DESCRIPTION:
    When many clients are in use, statd can take a long time to
    start.

    PROBLEM CONCLUSION:
    Instead of contacting clients serially, contact them
    50 at a time.

    ------

    APAR: IY21408 COMPID: 5765C3403 REL: 430
    ABSTRACT: PTY IOCTL CODE HAS ERROR

    PROBLEM DESCRIPTION:
    By inspection, the code was found to have a bug so there has
    not been any external symptoms. The code does not do a break
    out from a case statement as it should.

    PROBLEM CONCLUSION:
    Add in the break.

    ------

    APAR: IY21475 COMPID: 5765C3403 REL: 430
    ABSTRACT: DOCSEARCH CONFIG FAILED IN BOS INSTALL

    PROBLEM DESCRIPTION:
    User may see an error message during install from
    bos.docsearch.rte. Message indicates that docsearch could
    not configure itself and will have to be done by the user.

    PROBLEM CONCLUSION:
    Place a test in the post_i and post_u scripts that detects
    if the server is installed before attempting to configure
    docsearch.

    ------

    APAR: IY21488 COMPID: 5765C3403 REL: 433
    ABSTRACT: JST:64BIT APP CANNOT OPEN PRINTER ON SYSTEM W 32BIT KERNEL

    PROBLEM DESCRIPTION:
    64bit application cannot open printer on 32 bit system.

    PROBLEM CONCLUSION:
    Fix d_opts.

    ------

    APAR: IY21537 COMPID: 5765D5101 REL: 121
    ABSTRACT: TOPOLOGY SERVICES DAEMON ATTEMPTING TO USE ALL CPU CYCLES

    PROBLEM DESCRIPTION:
    Topology Services daemon attempting to use all CPU cycles
    available. topsvcs dies on node after starting.

    PROBLEM SUMMARY:
    The RSCT Topology Services daemon may exit under
    some circumstances, due to excessive CPU use.
    In some scenarios, the select() system call invoked
    by one of the threads fails, and this causes the
    thread to retry the operation immediately. Since
    the system call fails repeatedly, the daemon
    starts using too much CPU. A "watchdog" thread
    in the daemon will detect the excessive CPU
    use and will terminate the process to prevent other
    processes in the system from being starved.

    PROBLEM CONCLUSION:
    The problem was fixed by introducing a time
    delay every time the select() system call
    fails.
    With the fix, the daemon should no longer
    exit due to excessive CPU consumption when the
    select() system call fails.

    ------

    APAR: IY21538 COMPID: 5765C3403 REL: 430
    ABSTRACT: DOWNLOADING TAPE MICROCODE RESULTS IN A SENSE DATA DUMP

    PROBLEM DESCRIPTION:
      The customer will be unable to update the microcode on the
      first attempt to a tape drive.

    PROBLEM CONCLUSION:
      Query the device twice before returning a failure.

    ------

    APAR: IY21539 COMPID: 5765C3403 REL: 430
    ABSTRACT: BUFFER OVERFLOW VULNERABILITY IN CDE DTPRINTINFO

    PROBLEM DESCRIPTION:
    There is a possible security threat due to the buffer overflow
    in dthelpview.

    PROBLEM CONCLUSION:
    The maximum length of the string that can be input into the
    search string text field is limited to 50 chars.

    ------

    APAR: IY21540 COMPID: 5765C3403 REL: 430
    ABSTRACT: MVP:PAX -RW CHANGES SETUID PERMISSIONS SOURCE FILES

    PROBLEM DESCRIPTION:
    When pax was executed with the -rw option the setuid
    permissions of the source files was changing.

    PROBLEM CONCLUSION:
    The code has been modified slightly to fix the problem.

    ------

    APAR: IY21541 COMPID: 5765C3403 REL: 430
    ABSTRACT: PRTCONF DISPLAYS IMPROPER 'MEMORY SIZE'

    PROBLEM DESCRIPTION:
    "prtconf" script displays improper "Memory Size"
    for machines having more than one memory cards.
    Only mem0 is taken into consideration while calculating
    the Memory Size for the machine.

    PROBLEM SUMMARY:
    "prtconf" displays improper "Memory Size" for
    machines having more than one memory card.

    PROBLEM CONCLUSION:
    Suitable code changes made to prtconf to
    consider all the memory cards in the machine
    while calculating the system "Memory Size".

    ------

    APAR: IY21542 COMPID: 5765C3403 REL: 430
    ABSTRACT: IOCP: SYSTEM PANIC

    PROBLEM DESCRIPTION:
    The can be a system panic in the iocp with the following
    stack.
       .so_pin_more_stack () f18b97c0
       .soreceive () f18b9e10
       . iocp.ext:iocp_recv_data_socket () f18b9ea0
       . iocp.ext:iocp_process_completion_pkt () f18b9ef0
       . iocp.ext:GetQueuedCompletionStatus () f18b9f40
       .sys_call_ret () f18b9fc0

    PROBLEM CONCLUSION:
    The iocfd structure was being released while there was at
    least one outstanding reference to it. This reference would
    then point to a socket that had already been closed. Access
    to the socket cause the panic.
    The solution was to add a use count to the iocfd structure so
    that it would not be freed (and the socket associated with it
    would not be closed) until all references had been removed.

    ------

    APAR: IY21543 COMPID: 5765C3403 REL: 430
    ABSTRACT: MWM CALLS XGRABBUTTON WITHOUT XUNGRABBUTTON

    PROBLEM DESCRIPTION:
    There may be a slow memory leak in the X server when running
    the mwm window manager.

    PROBLEM CONCLUSION:
    Ensure that each call to XGrabButton has a corresponding call
    to XUngrabButton.

    ------

    APAR: IY21544 COMPID: 5765C3403 REL: 430
    ABSTRACT: IMPROVE PIO PERFORMANCE

    PROBLEM DESCRIPTION:
    Existing PIO calls are using the MCA calls for error detection.
    There is no error check for PCI PIO calls, therefore there are
    instructions that are being used unnecissarily.

    PROBLEM CONCLUSION:
    Modified the PIO calls to remove the unnecessary instructions.

    ------

    APAR: IY21545 COMPID: 5765C3403 REL: 430
    ABSTRACT: LPPCHKS FROM REJECT OF RSCT.CORE.GUI 2.2.0.1

    PROBLEM DESCRIPTION:
    Doing a reject of an update that contains a filename
    w/ a $ (e.g., A$B.class) will delete A.class,
    causing lppchk to fail, and requiring that the
    base image for this fileset be reinstalled.

    PROBLEM CONCLUSION:
    Replace "rm -rf" w/ unlink to delete the file.

    ------

    APAR: IY21546 COMPID: 5765C3403 REL: 430
    ABSTRACT: AIXIM IS NOT WORKING PROPERLY

    PROBLEM DESCRIPTION:
    chg_first value returned in PreeditDrawCallback isn't correct
    while inputting single-byte and double-byte characters.

    PROBLEM CONCLUSION:
    Fix Xlib to return correct value.

    ------

    APAR: IY21547 COMPID: 5765C3403 REL: 430
    ABSTRACT: SUPPORTING HTX CDROM EXERCISER

    PROBLEM DESCRIPTION:
    the current IDE driver does not provide the IOCTL to support
    the HTX cdrom exerciser.

    PROBLEM CONCLUSION:
    implement the IOCTL provide HTX with ATAPI passthru capability

    ------

    APAR: IY21562 COMPID: 5765C3403 REL: 430
    ABSTRACT: JAVA APPLICATIONS MAY RECEIVE SIGILL AND COREDUMP

    PROBLEM DESCRIPTION:
    Java applications may receive SIGILL and coredump.

    PROBLEM SUMMARY:
    This APAR addresses a core dump in
    _XmAddHashEntry XmeTraitSet ArrowBG.ClassPartInitialize
    CallClassPartInit XtInitializeWidgetClass
    XtInitializeWidgetClass XtInitializeWidgetClass _XtCreateWidget
    XtCreateWidget
    Apply OSF Motif fixes from Motif 2.1.30. to current AIX 4.3.3
    Motif level.

    PROBLEM CONCLUSION:
    Apply OSF Motif fixes from Motif 2.1.30. to current AIX 4.3.3
    Motif level

    ------

    APAR: IY21592 COMPID: 5765C3403 REL: 430
    ABSTRACT: SYSTEM HANG WITH ATM CONNECTED BACK-TO-BACK

    PROBLEM DESCRIPTION:
    ILMI GetRequest for atmfAddressStatus sent every microsecond
    when connected back-to-back instead of on a switch.

    PROBLEM CONCLUSION:
    Add a timer to only send requests every few seconds when
    a Get Response for this MIB is returned with NO Such Name.

    ------

    APAR: IY21607 COMPID: 5765C3403 REL: 430
    ABSTRACT: LOCK TIMEOUT

    PROBLEM DESCRIPTION:
    Machine crash, lock time out error (we have not seen this on
    any of adapter we have in the field)

    PROBLEM CONCLUSION:
    change in the receive and adapter interface management
    to keep track on_adapter buffers to ensure freeing the proper
    lock.

    ------

    APAR: IY21608 COMPID: 5765C3403 REL: 430
    ABSTRACT: SYNC COMMANDS MAY BE REQUIRED

    PROBLEM DESCRIPTION:
    Possible system crash on systems with CPUs that have L3
    caches, due to stale data being used by the adapter.

    PROBLEM CONCLUSION:
    Issue a specific macro call in the device driver to force the
    flushing of data from the L3 cache onto the real memory.

    ------

    APAR: IY21610 COMPID: 5765C3403 REL: 430
    ABSTRACT: NO SERVICE ACTION DISPLAYED FOR CERTIFY FAILURES

    PROBLEM DESCRIPTION:
      The customer is notified that the certify
      successfully completed despite the existence of
      unrecoverable data errors.

    PROBLEM CONCLUSION:
      Check to see if the number of unrecoverable data errors is
      between one and three inclusive. If so, display a screen
      suggesting that the drive be backed up, formatted and
      certified, and then restored.

    ------

    APAR: IY21647 COMPID: 5765C3403 REL: 430
    ABSTRACT: UPDATE DIAGNOSTIC VERSION

    PROBLEM DESCRIPTION:
    Incorrect Diagnostic version number.

    PROBLEM CONCLUSION:
    Update the Diagnostic version number.

    ------

    APAR: IY21813 COMPID: 5765C3403 REL: 430
    ABSTRACT: BUTTONS FOR INCREMENT WORK AS DECREMENT IN CONFIG_ASSIST

    PROBLEM DESCRIPTION:
    When up-arrow key of Spin Box is pressed , it decrements
    the value instead of incrementing.

    PROBLEM CONCLUSION:
    File SpinButton.java is modified to handle all the key
    events properly.

    ------

    APAR: IY21814 COMPID: 5765C3403 REL: 430
    ABSTRACT: NEW RELEASE OF CEREADME FILE FOR DISPLAY SERVICE HINTS

    PROBLEM DESCRIPTION:
    The CE uses this file to better service RS/6000
    systems running AIX diagnostics.

    PROBLEM CONCLUSION:
    Latest release of the CEREADME file.

    ------

    APAR: IY21817 COMPID: 5765C3403 REL: 430
    ABSTRACT: 4.3.3 DOCSEARCH CONFIG FAILED IN BOS INSTALL

    PROBLEM DESCRIPTION:
    Docsearch displays an error message stating that due to a
    missing directory, the doc library was not configured. Users
    will have to manually configure the application.

    PROBLEM CONCLUSION:
    Place portions of the /usr/post_i script into a root/post_i.

    ------

    APAR: IY21823 COMPID: 5765C3403 REL: 430
    ABSTRACT: PVT_BUTTERFLY2 CAUSES GXT6000P TO HANG.

    PROBLEM DESCRIPTION:
    When an OpenGL application uses display lists and multi-draw
    arrays, a possibility exists that the adapter will hang when
    running to a GXT6000P or GXT6500P.

    PROBLEM CONCLUSION:
    Some of the optimizations done for CATIA generated a
    mismatch between the data being sent to the adapter and
    the mask that tells the adapter what is being sent. This
    leads to the adapter hang. This will be fixed.

    ------

    APAR: IY21840 COMPID: 5765C3403 REL: 430
    ABSTRACT: DIAGRPT -S FAILS BECAUSE ASL IS NOT INITIALIZED

    PROBLEM DESCRIPTION:
    diagrpt -s fails because:
    1. an sprintf is used instead of scanf to parse the input
    2. the tm_mon field of struct tm numbers months 0-11 not 1-12
    3. asl is not initialized before displaying the results

    PROBLEM SUMMARY:
    Running diagrpt -s <date> produces error message:
    1820-001 internal error: asl_init must be called
    before asl_screen

    PROBLEM CONCLUSION:
    Call asl_init before asl_screen.

    ------

    APAR: IY21845 COMPID: 5765C3403 REL: 430
    ABSTRACT: BOOT HANGS OR CRASHES WITH MANY GIGABIT ETHERNET ADAPTERS

    PROBLEM DESCRIPTION:
    boot hangs or crashes very soon after starting on
    systems with many gigabit ethernet adapters

    PROBLEM CONCLUSION:
    ensure that RAM filesystem is located correctly during
    boot image uncompression phase

    ------

    APAR: IY21850 COMPID: 5765D5101 REL: 121
    ABSTRACT: HAES: TOPSVCSD ASSARTED IN HB_NCT.C WHILE PROCESSING CORRUPTED

    PROBLEM DESCRIPTION:
    This problem had been found on HACMP/ES V4.2.2.
    HAES had a mechanizum that filtered a corrupting packets from
    other node. But in rare case, The corrupted packet was not disca
    rded but passed to other functions. So there was a possibilty
    that the function assarted while processing the packet.
     In our case, topsvcsd.log had following error message at that
    time.
    Assertion failed: amg_adapter_offset == adapter_offset,
    file ../../../ ../../../../src/ssp/availability/pts/pam
    /nct/Hb_nct.C, line 1248

    PROBLEM SUMMARY:
    It is possible for a hats packet to be corrupted in such a
    way that it is not detected by the checksum. In this case
    the corrupted packet ended up causing an assert in hats,
    which causes hats to die.

    PROBLEM CONCLUSION:
    A special check has been added to the code to detect this
    specific corruption pattern, and act accordingly.

    ------

    APAR: IY21870 COMPID: 5765C3403 REL: 430
    ABSTRACT: BOS.RTE.SECURITY BREAKS NIS+

    PROBLEM DESCRIPTION:
    The file /usr/lib/security/methods.cfg is overwritten during
    an upgrade.

    PROBLEM CONCLUSION:
    Added /usr/lib/security/methods.cfg to the NOUPDATE_LIST
    entry in the appropriate Makefile.

    ------

    APAR: IY21967 COMPID: 5765C3403 REL: 430
    ABSTRACT: OCCASIONALLY, QDAEMON KILLS ERRONEOUS PROCESSES

    PROBLEM DESCRIPTION:
    In rare occasions, qdaemon might kill a process whose
    pid resembles the qdaemon pid recorded in /var/spool/
    lpd/stat/pid.

    PROBLEM SUMMARY:
    In rare occasions, qdaemon might kill a process whose pid
    resembles the pid of qdaemon in /var/spool/lpd/stat/pid.

    PROBLEM CONCLUSION:
    The qdaemon code has been fixed to prevent qdaemon from
    sending a kill signal to the wrong process.

    ------

    APAR: IY21973 COMPID: 5765C3403 REL: 430
    ABSTRACT: SYSTEM CRASH WHEN CONFIGURING FIBRE CHANNEL ADAPTER

    PROBLEM DESCRIPTION:
    System crash when configuring fibre channel adapter.

    PROBLEM CONCLUSION:
    Fix usage of TRB timers during adapter post error recovery.

    ------

    APAR: IY21977 COMPID: 5765C3403 REL: 430
    ABSTRACT: FSCK CAN CORRUPT LOST+FOUND IN CERTAIN RARE CASES

    PROBLEM DESCRIPTION:
    When a catastrophically corrupted file system contains
    hundreds (variable, approximately 300 on a stardard JFS) of
    orphaned files, fsck will not properly constuct the lost+found
    directory.

    PROBLEM CONCLUSION:
    Correct fsck to correctly initialize directory blocks on
    fragment boundries.

    ------

    APAR: IY22012 COMPID: 5765C3403 REL: 430
    ABSTRACT: SRNS 803-646 OR 803-776 ON TOKEN-RING ADAPTERS

    PROBLEM DESCRIPTION:
    When running advanced diagnostics on some token-ring
    adapters, and if the adapter is set to autosense and not
    connected to a network, the diagnostics may report an
    803-646 or 803-776 SRNs.

    PROBLEM CONCLUSION:
    Modify code to allow more time for chdev to change
    speed from autosense to 16 due to SNMP activity that may
    be occurring.

    ------

    APAR: IY22014 COMPID: 5765C3403 REL: 430
    ABSTRACT: NIM DIAG BOOT MAY CALL OUT INVALID SRN FOR TOKEN-RING ADAPTERS

    PROBLEM DESCRIPTION:
    If booted in a NIM environment for diagnostics through an
    ethernet port, diagnostics for some token-ring adapters
    (14103e00) may fail with an SRN when there is nothing wrong
    with the adapter.

    PROBLEM CONCLUSION:
    Force a mkdev/rmdev on all 14103e00 token-ring adapters
    during the NIM diagnostic boot.

    ------

    APAR: IY22027 COMPID: 5765C3403 REL: 430
    ABSTRACT: UPDATE ZAPDB PACKAGING TO INSTALL BINARY TRACE REPLAY TOOL.

    PROBLEM DESCRIPTION:
    ZAPdbReplay tool not installed.

    PROBLEM CONCLUSION:
    Add ZAPdbReplay tool to packaging inventory list.

    ------

    APAR: IY22043 COMPID: 5765C3403 REL: 430
    ABSTRACT: ULTRA3 PCI SCSI DIAGS NON-DETECTION OF MISSING VPD

    PROBLEM DESCRIPTION:
    Adapters with this type of failure will not be called out by
    diagnostics. The adapter will continue to operate properly
    without displaying VPD information.

    PROBLEM CONCLUSION:
    Use the adapter description to reference part number/FRU
    number information.

    ------

    APAR: IY22044 COMPID: 5765C3403 REL: 430
    ABSTRACT: FIXES FOR FUTURE FC ADAPTER

    PROBLEM DESCRIPTION:
    Failures with I/O after some failover conditions,
    misconfiguration if UTM is configured on LUN 0,
    bad failover/recovery behavior, inability to reset just one
    LUN.

    PROBLEM CONCLUSION:
    Make autorecovery configurable. Fix Clear ACA problem with
    marking controllers failed. Improve controller-fail debugging.
    Check for UTM LUN on LUN 0 and adjust behavior to fix. Add
    LUN-level reset support.

    ------

    APAR: IY22115 COMPID: 5765C3403 REL: 430
    ABSTRACT: SCSI HOT SWAP MGR COMPLETES W/ FAULT LIGHT ON

    PROBLEM DESCRIPTION:
    Drives could not be set to normal.

    PROBLEM CONCLUSION:
    Added support for safte

    ------

    APAR: IY22121 COMPID: 5765C3403 REL: 430
    ABSTRACT: MOJAVE MODS TO ENABLE DLL

    PROBLEM DESCRIPTION:
    GXT4500P/GXT6500P diags will fail due to DLL EC

    PROBLEM CONCLUSION:
    Change GXT4500P/GXT6500P diags to support DLL enable

    ------

    APAR: IY22125 COMPID: 5765C3403 REL: 430
    ABSTRACT: ORACLE INSTANCE TERMINATED WHEN MEMORY IS PINNED

    PROBLEM DESCRIPTION:
    Application receiving an EIO

    PROBLEM CONCLUSION:
    Correct a faulty algoritm for npage
    Use ulong instead of int cast for a variable whose value
    is being shifted.

    ------

    APAR: IY22127 COMPID: 5765C3403 REL: 430
    ABSTRACT: SYSTEM CRASH IN ATM DEVICE DRIVER RUNNING HTX

    PROBLEM DESCRIPTION:
    system crach in atmcm_close because pointer being access
    after it's free.

    PROBLEM CONCLUSION:
    remove the pointer

    ------

    APAR: IY22129 COMPID: 5765C3403 REL: 430
    ABSTRACT: PAX OVERWRITES EXISTING FIFOS

    PROBLEM DESCRIPTION:
    The pax command with -p option would overwrite existing FIFO's
    with regular files.

    PROBLEM CONCLUSION:
    1. The pax command with -p option was modified not to overwrite
       FIFO's but instead give a warning message and continue its
       operation.
    2. The pax code was modified through 336827 for -p option to
       overwrite directories on AIX 43 and hence 43X track is also
       created for this defect.

    ------

    APAR: IY22135 COMPID: 5765D5101 REL: 121
    ABSTRACT: LIBNETMON:SIOCADDRT IOCTL() TO ADD ROUTES FAILS

    PROBLEM DESCRIPTION:
    libnetmon:siocaddrt ioctl() to add routes fails

    PROBLEM SUMMARY:
    A change in AIX has resulted in modifications to the
    way network routing is implemented. With the change,
    a host route table entry is created for each local
    adapter interface. The entry has the adapter's address as
    destination and the "loopback" address (127.0.0.1)
    as gateway. When running command
       netstat -rn
    lines like the following will appear in the command's
    output:
    9.114.149.65 127.0.0.1 UGH 0 29266 lo0
    These route entries create a problem for RSCT
    Topology Services when in the HACMP/ES environment,
    since this component needs to add and remove routes
    dynamically. Such route manipulations are needed
    when sending an "echo" message from a (say) standby
    adapter to a service adapter located in the same
    HACMP network, which is used to determine the adapter's
    liveness. The route added by AIX creates a conflict
    with the route to be added by Topology Services,
    resulting in the latter not being accepted by AIX.
    The following may be seen in the "netmon" log files,
    located at
    /var/ha/run/topsvcs.<cluster name>/nmDiagFile.<interface>
    (RSCT prior to version 1.2.1)
    or at
    /var/ha/log/nmDiag.nim.topsvcs.<interface>.<cluster name>
    (remaining releases):
    nmRouteAddHost:SIOCADDRT ioctl() failed, errno is 17.
    This may result in HACMP/ES falsely reporting adapters as
    dead, which will cause unnecessary (and possibly repeated)
    "adapter swap" events.

    PROBLEM CONCLUSION:
    The fix in Topology Services accounts for the
    possibility that the loopback routes are present.
    If an error is encountered while adding the
    temporary route, Topology Services will first
    retrieve information about the current route,
    remove it, then add the temporary route. After
    the packet is sent to the destination adapter, the
    temporary route is removed and the original route
    is restored.
    With the fix, no false adapter downs or spurious
    adapter swap events should occur in HACMP.

    ------

    APAR: IY22196 COMPID: 5765C3403 REL: 430
    ABSTRACT: PCI RACE PROTECTION FOR SECOND LEVEL INTR HANDLER

    PROBLEM DESCRIPTION:
    While running atm 155 diagnostics a SYSINTR ERR may log in
    the aix error log

    PROBLEM CONCLUSION:
    add pci race protection code in the second level intr handler

    ------

    APAR: IY22201 COMPID: 5765D5101 REL: 121
    ABSTRACT: NAS:HATS SHOULD HANDLE IFCONFIG'D ALIASES WHO HAVE NO IP ADDRESS

    PROBLEM DESCRIPTION:
    NAS:HATS should handle ifconfig'd aliases who have no IP address

    PROBLEM SUMMARY:
    The problem includes two parts. One is in the Topology
    Services netmon library, which only occurs in Linux, and the
    other is in the Topology Services daemon, which occurs in
    all platforms.
    When a node becomes singleton because it lost communication
    to other nodes or when it is newly started, Topology
    Services uses its netmon library to figure out if the
    network adapters are still working. Netmon library first
    finds all the adapters currently defined in the system.
    When an alias adapter is found, netmon needs to find the
    base adapter to get the information of the network
    activities occurred on the network adapter. The ioctl()
    system call used by netmon to get the IP address of the
    base adapter failed because the IP address of the base
    adapter was not defined.
    Because a network adapter cannot work without defining
    alias network adapter that does not have an IP address
    defined for its base adapter was considered an error.
    Netmon exits with an error code. The error causes Topology
    Services to consider the adapter as down. Since Linux 2.4
    allows defining an IP address for an alias network adapter
    without defining an IP address for its base adapter first,
    netmon library needs to be changed not to exit for this
    problem.
    In the Topology Services daemon, when a netmon library
    related
    problem is reported, an error message will be printed to
    its service log file. The parameters of the message
    catalogue of this message, however, is mis-aligned. The
    daemon
    would cause a segmentation violation when finding a problem
    with the netmon library and printing a message about it.

    PROBLEM CONCLUSION:
    Two changes are needed to fix this problem. On the netmon
    library side, since an alias network adapter cannot work
    without its base adapter correctly configured and started,
    netmon library should ignore the alias network adapter if
    its base adapter does not have an IP address defined.
    However, netmon should keep finding other network adapters
    without exiting. On the Topology Services daemon side, the
    message catalogue should be fixed so that the message
    parameters matche the message use in the code.

    ------

    APAR: IY22202 COMPID: 5765D5101 REL: 121
    ABSTRACT: REQUIRED UPDATED FOR RSCT VERSION 121

    PROBLEM DESCRIPTION:
    required updates for rsct version 121

    ------

    APAR: IY22204 COMPID: 5765D5101 REL: 121
    ABSTRACT: PERMIT LONGER PATH NAME FOR NETMON LOG FILE (HATS)

    PROBLEM DESCRIPTION:
    permit longer path name for netmon log file (hats)

    PROBLEM SUMMARY:
    permit longer path name for netmon log
    file (hats)

    PROBLEM CONCLUSION:
    permit longer path name for netmon log
    file (hats)

    ------

    APAR: IY22205 COMPID: 5765D5101 REL: 121
    ABSTRACT: HAEM DAEMON PRIORITY LIKE HATS_PRIORITY

    PROBLEM DESCRIPTION:
    To stop haem from core dumping, dev has recommended to implement
    the same mechanism that hags uses which is to set the priority
    of the daemon at hats_priority + 1.
    haem was not responding to the hags daemon at the given time (2
    mins). When haem detects this, he terminates himself, respawns
    and problem goes away until it can not respond again to hags for
    2 mins.
    haemd did not get enough cpu cycles because it runs with the

    PROBLEM SUMMARY:
    If the system is heavily loaded, there might be occassions
    that haem daemon is not getting enough resources to be able
    to respond to hags within two minutes. When this occurs,
    haemd terminates and respawns.

    PROBLEM CONCLUSION:
    The priority of the haemd process and its resource
    monitors are now equal to:
    (FixPri_Value attribute of the TS_Config class) + 1.
    In other words, haemd priority is set to hats_priority + 1
    (which is what hags is using). This will minimize the
    possibility of haem not being able to respond to hags within
    2 minutes since they are not both set to the same priority.

    ------

    APAR: IY22206 COMPID: 5765D5101 REL: 121
    ABSTRACT: LX: FFDC PUTS CONTROL CHARS IN SYSLOG: DIFFICULT TO READ

    PROBLEM DESCRIPTION:
    LX: FFDC puts control chars in syslog; difficult to read

    PROBLEM SUMMARY:
    By design, the FFDC logging utilities used the
    pseudo-control character sequence MI to indicate line
    breaks within the message being logged to the System Log.
    This imitated syslogd's fucntion of replacing newline
    characters in strings it processes with the M character
    sequence. Architects determined at a later time that a more
    readable field separator be used, since a control sequence
    might make a reader believe that the logged data was
    corrupted during recording.

    PROBLEM CONCLUSION:
    The apparant control character sequence MI has been
    replaced with the sequence ::: for readability

    ------

    APAR: IY22207 COMPID: 5765D5101 REL: 121
    ABSTRACT: LSSRC HAGSGLSM SHOWS INCORRECT OUTPUT

    PROBLEM DESCRIPTION:
    lssrc hagsglsm shows incorrect output

    PROBLEM SUMMARY:
    When a local adapter css0 or css1 is down,
    Group Services does not will join the group
    cssMembership or css1Membership respectively
    for the local node.
    Because the 'lssrc -ls hagsglsm' output
    currently print the membership information
    about the cssMembership or css1Membership,
    the count of memberships will be zero.
    However, unfortunately, if total number of
    the group, which has the local adapter down,
    is still greater than the total number of
    the other group, the output of 'lssrc -ls hagsglsm'
    will become confusion because the number
    of larger group is shown as zero.
    The fix is basically to subscribe the groups
    (cssMembership and css1Membership) to know
    and to print the global number of members.

    PROBLEM CONCLUSION:
    The output 'lssrc -ls hagsglsm' will now clearly
    show the number of members of cssMembership
    and css1Membership even if a local css0 or css1
    adapter down, and this removes the confusion
    particurly when the total number of members
    of the group (with the local adapter down)
    is bigger.

    ------

    APAR: IY22208 COMPID: 5765D5101 REL: 121
    ABSTRACT: HAGS RECEIVES A GARBAGE FFDC ID

    PROBLEM DESCRIPTION:
    HAGS receives a garbage FFDC ID

    PROBLEM SUMMARY:
    To reduce the possibility of passing FFDC ID with
    random characters to Group Services, an FFDC ID
    is cleared to contain a NULL string before passed
    to FFDC log facility.

    PROBLEM CONCLUSION:
    To reduce the possibility of passing FFDC ID with
    random characters to Group Services, an FFDC ID
    is cleared to contain a NULL string before passed
    to FFDC log facility.

    ------

    APAR: IY22209 COMPID: 5765D5101 REL: 121
    ABSTRACT: ERORR IN WRITING SYSLOG OR ERRPT

    PROBLEM DESCRIPTION:
    Erorr in writing syslog or errpt

    PROBLEM SUMMARY:
    When Group Services enhanced the logging facility
    of recording the important activities separately
    to keep the history for the longer period, a few
    irrelevant AIX log entries were also appeared.
    It is because of a few missing condition checking
    of whether the longer period log file is opened
    or not before the hags daemon tries to write
    the log messages.

    PROBLEM CONCLUSION:
    This fix will remove the erroneous AIX errlog entries
    which happened at the startup of hags daemon.

    ------

    APAR: IY22210 COMPID: 5765D5101 REL: 121
    ABSTRACT: 384WAY: HAGS CORE DUMPS DURING PERIODS OF STRESS

    PROBLEM DESCRIPTION:
    384way: hags core dumps during periods of stress

    PROBLEM SUMMARY:
    There are memory leaks in the logging statements
    which is called whenever a Group Services
    group membership changes.
    Therefore, after the long run of the hags daemon
    with many membership changes of any groups,
    the Group Services daemon may eventually consume
    all the virtual memory space which the daemon
    can use, and then the daemon will be coredumped.
    Once the memory leaks are fixed, the unexpected
    coredump of hags daemon will disappear.

    PROBLEM CONCLUSION:
    This fix will remove the memory leaks at the
    time of group membership changes, and thus
    prevent the eventual coredump of the hags
    daemon after the long run.

    ------

    APAR: IY22212 COMPID: 5765D5101 REL: 121
    ABSTRACT: HAGSGLSM CORE DURING SYSPAR_CTRL -G -K

    PROBLEM DESCRIPTION:
    hagsglsm core during syspar_ctrl -G -k

    PROBLEM SUMMARY:
    This problem happened due to the stopping timing
    between threads and the problem of handling
    the signal when the hagsglsm daemon process
    is asked to stop.
    The brief detail is as follows;
    The main thread, which also handles the signals,
    detects the SIGTERM signal and set a global
    variable and wake up the other threads.
    The other threads who are woken up will return
    from the thread run and become inactive.
    However, if one of the inactive threads was
    waiting for a locking primitive while
    the signal thread was using it, there could
    lead an invalid locking situation when
    the signal thread is releasing the lock primitive.
    So, the problem can be fixed by separating
    the signal thread from the main thread and
    by idling at the end of thread execution
    to prevent any invalid access to the lock primitives.

    PROBLEM CONCLUSION:
    The fix should not cause the occasional coredumps
    of the hagsglsm daemon when it is being stopped.

    ------

    APAR: IY22214 COMPID: 5765D5101 REL: 121
    ABSTRACT: 384WAY: HATS_NIM CORE DUMPS DURING PERIODS OF STRESS

    PROBLEM DESCRIPTION:
    384way: hats_nim core dumps during periods of stress

    PROBLEM SUMMARY:
    RSCT Topology Services has, besides a daemon,
    "Network Interface Module" (NIM) processes to monitor
    each of the local interfaces. If one such NIM process
    fails to provide timely response to an "are you alive?"
    request from the main daemon then the daemon considers
    the NIM process blocked and closes the connection to it.
    It has been observed that NIM processes could suffer
    a segmentation violation while on their way to being
    terminated (after the daemon had closed its connection
    to it). The problem has been traced back to the
    termination of the several threads used in the NIM
    implementation. The resulting core dump causes an
    AIX error log entry to be created.

    PROBLEM CONCLUSION:
    Two distinct fixes have been applied:
       - The threshold of the amount of blockage
         allowed in the Topology Services NIMs before
         the daemon considers the NIM process blocked
         has been increased, to allow for larger blocking
         times under heavy load
       - The termination procedure of the NIM process
         has been changed to remove known factors
         that cause the process to have a segmentation
         violation during termination
    With the fixes, the chance of the NIM process being
    abnormally terminated should be drastically reduced.

    ------

    APAR: IY22221 COMPID: 5765C3403 REL: 430
    ABSTRACT: NEED TO INITIALIZE FORKED PROCESSES PROPERLY

    PROBLEM DESCRIPTION:
    Customer may experience unusual and unexpected behavior when
    using PKCS11 applications that fork children processes that
    also use PKCS11

    ------

    APAR: IY22255 COMPID: 5765C3403 REL: 430
    ABSTRACT: SECURITY: VULNERABILITY IN LSMCODE

    PROBLEM DESCRIPTION:
      The customer will not receive a usage error when
      specifying an invalid type command line option for
      lsmcode.

    PROBLEM CONCLUSION:
      Check the type provided from the command line. If the
      type is not supported, then display a usage error.

    ------

    APAR: IY22258 COMPID: 5765C3403 REL: 430
    ABSTRACT: IDECDROM IOCTL IDEPASSTHRU RETURNS EINVAL

    PROBLEM DESCRIPTION:
    HTX fails to exercise IDE CD-ROM with EINVAL error.

    PROBLEM CONCLUSION:
    Fix the check for the ATAPI packet length-should be 12 or 16.
    Add default capability for 48X cdrom to read CD-DA.
    (digital audio)

    ------

    APAR: IY22259 COMPID: 5765C3403 REL: 430
    ABSTRACT: ADD 4330-09 ML TO FIXDATA

    PROBLEM DESCRIPTION:
    Customers will not be able to determine that the 4330-09
    Recommended Maintenance package is installed.

    PROBLEM CONCLUSION:
    Add the 4330-09 maintenance package stanza to the fixdata.

    ------

    APAR: IY22343 COMPID: 5765C3403 REL: 430
    ABSTRACT: HOT: ICA DOES NOT PERFORM BIG DES AND ICARSACRT CORRECTLY

    PROBLEM DESCRIPTION:
    i3des with large data fails

    PROBLEM CONCLUSION:
    don't require intr to see desdone and busmaster
    just busmaster

    ------

    APAR: IY22361 COMPID: 5765C3403 REL: 430
    ABSTRACT: UNKNOWN CLASS ERROR LOG ENTRY ON REBOOT WITH IDE CD-ROM

    PROBLEM DESCRIPTION:
    each the the system reboot, an error ATAIDE_ERR3 gets logged
    in the system's errlog. It mis-leads the customer thinking
    that there is a hardware problem.

    PROBLEM CONCLUSION:
    correct a mis-calculation.

    ------

    APAR: IY22373 COMPID: 5765C3403 REL: 430
    ABSTRACT: DRIVE WITH ROOTVG CAN BE REMOVED

    PROBLEM DESCRIPTION:
    Customer will not get an error message or denial when
    attempting to remove the device with system boot files

    PROBLEM CONCLUSION:
    Code has been added to detect if this is the boot device and
    does not allow customer to remove the device.

    ------

    APAR: IY22388 COMPID: 5765C3403 REL: 430
    ABSTRACT: BOS.NET.IPSEC.KEYMGT INSTALL FAILED ON 4.3.3

    PROBLEM DESCRIPTION:
    during install of bos.net.ipsec.keymgt the install process
    coredumps causing installation to abort

    PROBLEM CONCLUSION:
    Some of the pointers refered to by the ikegui command were
    not initialized properly. These were initialized properly
    to solve the problem

    ------

    APAR: IY22431 COMPID: 5765C3403 REL: 430
    ABSTRACT: SSA EPOW WORKAROUND TO RESET ADAPTER ON SUSPEND INTERRUPT

    PROBLEM DESCRIPTION:
    Add EPOW handling to better handle machine power loss when
    disks are left running.

    PROBLEM CONCLUSION:
    Add workaround in device driver to stop all I/O
    and reset the adapter on an EPOW suspend alert. If it is a
    power glitch, the driver will continue as normal on receipt
    of the EPOW resume interrupt.

    ------

    APAR: IY22539 COMPID: 5765C3403 REL: 430
    ABSTRACT: ULTRA2/3 INTEGRATED PCI SCSI CONTROLLER FAILS DIAGNOSTICS.

    PROBLEM DESCRIPTION:
    When running diagnostics on integrated SCSI Ultra 2 or Ultra 3
    adapters, a SCSI register error will be returned.

    PROBLEM CONCLUSION:
    The error will be returned for integrated
    Ultra 2 and Ultra 3 SCSI controllers that
    do not contain VPD.

    ------