PK45310: DOUBLE FREEMAIN OF FEPI CONTROL BLOCKS DURING SHUTDOWN

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• During shutdown FEPI SCIP exit DFHSZXSC is being driven for an
  incoming UNBIND request after session cleanup processing has
  completed and the connection DFHSZDCD storage freed.  DFHSZXSC
  queues the connection up for processing once again, even though
  the storage has already been freed, and as the connection delete
  flags are still set on residually a second and failing attempt
  is made to free the storage. The root cause of this problem is
  different to PK37736 but the result was similar.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: Double FREEMAIN of FEPI DFHSZDCD        *
  *                      connection block resulting in message   *
  *                      DFHSM0002.                              *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A VTAM REQSESS macro is issued on behalf of a FEPI terminal to
  request a session with the target system. Normally the REQSESS
  would complete before the BIND is received, however in this case
  the BIND is received before the REQSESS completion exit is
  driven. When the BIND is received, the BIND exit, DFHSZXSC, is
  driven and sets SZD_CD_BINDR (BIND received). Then when the
  REQSESS exit, DFHSZ2QX, is driven, it sets bit SZD_CD_REQD
  (REQSESS accepted) but then incorrectly resets SZD_CD_BINDR,
  because it expects that the BIND exit will be driven later.
  When CICS and FEPI are later shutdown SZD_CD_BINDR being off, in
  combination with SZD_CD_REQD being set, is taken to mean that a
  session is in the process of being established. FEPI therefore
  issues an unnecessary TERMSESS TERMQ macro to terminate a
  nonexistent queued session. Completion of this request drives
  FEPI cleanup processing which results in the connection control
  block DFHSZDCD being scheduled for FREEMAIN.
  Before the DFHSZDCD is actually freed FEPI issues the valid
  TERMSESS TERMC request to terminate the active session. When
  this second TERMSESS completes the FREEMAIN of the DFHSZDCD has
  completed. However FEPI cleanup processing is once again
  scheduled resulting in a further FREEMAIN for the same DFHSZDCD
  storage. This second FREEMAIN fails with message DFHSM0002 SYSID
  A severe error (code X'0305') has occurred in module DFHSMGF.
  
  Additional keywords: MSGDFHSM0002 IMS 08390001 SNS08390001
                       SENSE08390001
  

Problem conclusion

• The FEPI REQSESS completion exit DFHSZ2QX has been changed so
  that it no longer resets flag SZD_CD_BINDR. Existing code in
  DFHSZ2QS has already initialized this flag when the REQSESS was
  issued.
  

Temporary fix

• FIX AVAILABLE BY PTF ONLY
  

Comments

APAR Information

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

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

  PK52529 UK31498

Modules/Macros

•    DFHSZ2QX
  

Fix information

• Fixed component name

  CICSTS 3.X Z/OS

• Fixed component ID

  5655M1500

Applicable component levels

• R400 PSY UK31498

     UP07/11/24 P F711

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.