OA67065: STATUS STOP PROCESSING CAN BE DELAYED DUE TO A RUNNING PRE-EMPTIBLE SRB, LEAVING OTHER WORK IN ASID NONDISPATCHABLE

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• In certain environments Status Stop processing may be delayed
  waiting for a preemptible SRB to relinquish control of the
  processor.  This can happen when there is no higher priority
  ready to run work waiting for access to the processor that the
  SRB is dispatched to. This can result in other TCBs and SRBs in
  the address space being left in a nondispatchable state for an
  extended period of time while waiting for the SRB to stop
  running.
  
  ANALYSIS:
  In the reported scenario, a TCB had issued a PurgeDQ which
  results in a Status Stop, SRBS SVC being issued. The Status SVC
  4F cannot complete until all other TCBs and SRBs in this address
  
  space have stopped running or terminated. A single pre-emptible
  SRB continued to run, despite multiple external timer
  interrupts, leaving the other TCBs and SRBs in the address space
  nondispatchable for an extended period of time.
  
  KNOWN IMPACT:
  This problem results in TCBs and SRBs in the address space being
  
  left nondispatchable for longer than they ordinarily should.
  Depending on the frequency of this occurring, this can result in
  increased lock contention, latch contention, or other problems
  related to delays due to elongated nondispatchability.
  
  VERIFICATION STEPS:
  System trace will show a TCB issuing an SVC 4F Status Stop, and
  repeated dispatches in IGC079, while an SRB in the same address
  space continues to run on another processor, even after the SRB
  encounters EXT  TIMR  00001005 interrupts and continues running.
  
  
  While this is occurring, the ASCB will have ASCBSSSS - Status
  stop SRB summary bit set.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * Users running z/OS HBB77C0 and above.                        *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * Status STOP SRB processing can be                            *
  * delayed due to a long running Enclave                        *
  * SRB, leaving other work in the                               *
  * address space nondispatchable.                               *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A TCB issued a PURGEDQ which results in a Status STOP SRB.
  In the address space, there was a long running Enclave SRB
  that never exited to the Dispatcher for any higher priority
  work. External interrupts did occur but the SRB stayed
  dispatched to the processor.  The PURGEDQ could not complete.
  This resulted in other work in the address space being left
  in a nondispatchable state while waiting for the Enclave SRB
  to stop running.
  

Problem conclusion

• Upon receiving an external interrupt for a time slice end,
  check if there is an Enclave SRB running, and if there is a
  Status STOP SRB being processed, then exit to the Dispatcher.
  This will allow work to get processed such as the PURGEDQ to
  complete.
  

Temporary fix

Comments

APAR Information

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

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

  UJ96471 UJ96472 UJ96473

Modules/Macros

• IEAVEEXT
  

Fix information

• Fixed component name

  SUPERVISOR CONT

• Fixed component ID

  5752SC1C5

Applicable component levels

• R7D0 PSY UJ96473

     UP24/12/24 P F412

• R7E0 PSY UJ96471

     UP24/12/24 P F412

• R7C0 PSY UJ96472

     UP24/12/24 P F412

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.