OA44060: WAIT084-4 WITH PSAPSTK EQUAL TO PSAPSAV

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• WAIT084-4 due to incorrect FRR stack for the program check FLIH.
  PSAPSTK at offset x'398' in the PSA for CP3 contains:
       07323020  07323020  0734C0B8  .
  The first word (PSAPSTK) and the second word (PSAPSAV) should
  never be equal which is why the wait state is loaded.  The
  sequence of events leading up to the wait state are as follows:
  1.  A CLKC external interrupts happens on one CP due to a TQE
      timer pop
  2.  On the other CP - the timer routine is given control and
      frees the TQE
  3.  Back on the original CP - we go to trace the CLKC interrupt
      in the system trace but the TQE is freed so we take a
      PIC11  ( this is the 'chance of a lifetime' timing issue -
      very low probability of that happening).
  4.  Program check first level interrupt handles the PIC11 who
      then calls RSM to resolve the page fault.  RSM needs a
      lock that is unavailable  (again -  very low probability of
      that happening) and another external interrupt comes in
      and uses the same save area as the CLKC external interrupt.
  The Program check first level interrupt handler detects the
  problem and issues the WAIT084-4 preventing any possibility of
  a PSA overlay.
  
  Verification Steps:
  1.  Check PSAPSTK (offset x'398' in PSA) and PSAPSAV (offset
      x'39C')  they should be equal.
  2.  Trace table will be similar to the following:
  03 0031 009CB300  PGM    011 00000000_0133C976  00040011
                               04043000 80000000
       *Pic11 will be in IEAVETRC
  03 0031 009CB300  EMS        00000000_012F96E2  00001201
                               05041000 80000000  811860BC
      *EMS will be in IEAVELKX
  03 0031 009CB300 *RCVY  PROG                    940C4000
  03 0031 009CB300 *RCVY  RTRY 040C0000 8133C97A  940C4000
  03 0031 009CB300  CLKC       00000000_0000B312  00001004
                               07851000 80000000
  03 0031 009CB300 *PGM    001 00000000_07B6B002  00020001
                               04042000 80000000
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users running z/OS HBB7780 and HBB7790.      *
  ****************************************************************
  * PROBLEM DESCRIPTION: A very rare situation resulted in       *
  *                      a PSA overlay and a WAIT084-04:         *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  The scenario that resulted in the PSA overlay is as follows:
  - System trace got a program check while processing a CLKC
    external interrupt.
  - The Program FLIH handling the program check had to call RSM
    to resolve a page fault, but it had to spin to wait for a
    lock.
  - During the wait, another external interrupt came in and
    used the same PSA save area as the CLKC external interrupt.
  
  The root cause of this problem was the program check
  occurring in System Trace processing.  On a cpu, system trace
  was handling a CLKC external interrupt and it was referencing
  a TQE (timer queue element).  However, on another cpu, the
  timer routine was given control and then freed the TQE.
  Back on the original cpu, system trace takes a PIC11 program
  check as it tries to reference the freed TQE.
  

Problem conclusion

• System trace will no longer reference the TQE while
  processing a CLKC external interrupt.
  
  
  DOCUMENTATION HOLD FOR APAR OA44060
    +-------------------------------------------------------+
  
    MVS Diagnosis:  Tools and Service Aids
  
    +--- LOCATION IN PUBLICATION ---------------------------+
    |                                                       |
    | System Trace                                          |
    |   Reading system trace output                         |
    |     CALL, CLKC, EMS, EXT, I/O, MCH, RST, and SS       |
    |     trace entries                                     |
    +-------------------------------------------------------+
  
  | - Make the following changes for the CLKC trace entry:
  
  | In the table that shows all the trace entries, for the
  | CLKC trace entry, remove the tqe-tcb- and tqe-asid
  | references.
  
  | In the section marked by: UNIQUE-1/UNIQUE-2/UNIQUE-3
  |                           UNIQUE-4/UNIQUE-5/UNIQUE-6
  |
  | Remove the tqe-tcb- and tqe-asid and their respective
  | definitions
  

Temporary fix

• *********
  * HIPER *
  *********
  

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

  UA72243 UA72244

Modules/Macros

• IEAVETRC
  

Fix information

• Fixed component name

  SYSTEM TRACE

• Fixed component ID

  5752SC142

Applicable component levels

• R780 PSY UA72243

     UP14/02/12 P F402 Ž

• R790 PSY UA72244

     UP14/02/12 P F402 Ž

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.