PK93158: CICS HANGS IN DFHIRP WAITING FOR SUDBECB. MSGBUF ADDRESS IS ZERO.

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• You have cancelled a task that had an MRO session
  connected to a partner region and your CICS region is
  hung.
  
  An SDUMP of the hung CICS region shows the QR KTCB in a
  MVS wait out of module DFHIRP during disconnect
  processing in routine IRWAITLX, waiting for the SUDBECB
  to be posted.
  
  At the time of the hang, your joblog shows dump request:
  
  IEA794I SVC DUMP HAS CAPTURED:
  DUMPID=010 REQUESTED BY JOB (CICSJOB)
  DUMP TITLE=CICS SYSTEM DUMP REQUESTED BY DFHIRP'S MESSAGE EXIT
             EXIT FRR
  
  Viewing this dump using IPCS VERBX LOGDATA shows
  abend S00C4 in IXCS1MS1 due to the message buffer address
  in R2 being zeros.
  
  DFHIRP has invalidly passed a buffer address of zeros to XCF
  causing IXCS1MS1 to abend.  Thus the CICS/XCF message EXIT
  does not complete SWITCH processing, leaving the QR TCB
  in an MVS wait.
  
  Additional Symptoms and Keywords:
  KIXREVEAA
  
  The task that was cancelled and is currently running on the QR
  TCB has KE STACK:
  KE_NUM @STACK   LEN  TYPE ADDRESS  LINK REG OFFS ERROR NAME
  
  018F   2F0BF020 0130 Bot  AD401A00 AD401DAC 03AC       DFHKETA
  018F   2F0BF150 0330 Dom  AD419DC0 AD419FC8 0208       DFHDSKE
  018F   2F0BF480 0530 Dom  AD43F6B8 AD4428F0 3238       DFHXMTA
                       Int     +3160 AD440B60 14A8
                                        RMXM_BACKOUT_TRANSACTION
  018F   2F0BF9B0 0A30 Dom  AD4C9B18 AD4CC36A 2852       DFHRMUW
                       Int     +0B48 AD4C9D12 01FA
                                            RMUW_BACKOUT_UOW_RTN
  018F   2F0C03E0 06B0 Sub  AD4A0CD0 AD4A10D4 0404       DFHRMLSO
  018F   2F0C0A90 0350 Dom  ADEEA9E8 ADEEBF18 1530       DFHCRIU
                       Int  +0E62 ADEEAC70 0288
                                                  PERFORM_COMMIT
                       Int     +139A ADEEB9C6 0FDE       TERM
  018F   2F0C0DE0 04E0 Lifo 2DED2900 ADED4F66 2666       DFHZISP
  018F   2F0C12C0 0348 Lifo 2E0F64A8 00000000 0000       DFHZIS2
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: A CICS region that uses MRO/XCF         *
  *                      hangs in DFHIRP waiting for the         *
  *                      SUDBECB ECB.                            *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A task running in a CICS region that uses MRO/XCF is cancelled
  after issuing a function ship request. During the disconnect
  of the MRO session, the message buffer address is set to zeros,
  but before the length is set to zeros, the QR TCB is
  interrupted, by XCF, to drive the IXCMSGO message exit. As a
  non-zero length and a zero address is passed to the message
  exit, it abends OC4 leaving the QR TCB in a permanent WAIT.
  Additional keywords: abend0C4 abendS0C4 SCCBSLTH SCCBAREA
  

Problem conclusion

• DFHIRP has been changed so that a valid message buffer
  address is passed on the XCF IXCMSGI call.
  

Temporary fix

• *********
              * HIPER *
              *********
  FIX AVAILABLE BY PTF ONLY
  

Comments

APAR Information

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

  PK92797

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

  UK50179

Modules/Macros

• DFHIRP   DFHIRPSW
  

Fix information

• Fixed component name

  CICS TS Z/OS V4

• Fixed component ID

  5655S9700

Applicable component levels

• R600 PSY UK50179

     UP09/09/19 P F909

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.