PH30840: DFHSO0001 AN ABEND (CODE 0C4/AKEX) HAS OCCURRED AT OFFSET X'2790' IN MODULE DFHSOSE.

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Your CICS region intermittently receives:
  DFHSO0001  An abend (code 0C4/AKEX) has occurred at offset
  X'2790' in module DFHSOSE.
  The abend0c4 occurs in DFHSOSE + x'278A' as register 1 is zero.
  At the time, DFHSOSE is in routine DELETE_CERTIFICATE_DATA.
  The trace shows:
  QR AP 451C WBXN *EXC* ATTACH_ALIAS_ERROR *=01235DATA.
  QR AP 0746 ABAB *EXC* Percolate
  START_ABEND,39348018,NO,YES *=01235
  QR AP 1942 APLI *EXC* Abend-Percolate
  START_PROGRAM,DFHWBXN,NOCEDF,FULLAPI,EXEC,NO,38DE1918,00000000 ,
  0000
  
  SO SO 0803 SOSE *EXC* RECOVERY_ENTERED
  SECURE_SOC_CLOSE,DISASTER,
  
  The problem is caused by a race between a CWXN task trying to
  close a socket and CICS shutting down.  The CWXN task tries to
  close the socket after failing to attach an alias task.  At the
  same time the CSOL task is closing the inbound sockets because
  the region is shutting down.  Despite numerous checks throughout
  the close process it seems there is still a small window where
  CSOL can delete the socket from underneath the CWXN task.  This
  causes the DELETE_CERTIFICATE_DATA call to pick up a repository
  token from low core (because the socket's ssl_data_ptr is now
  0).  The token is not valid but DFHSOSE attempts to set it to 0
  and that causes the abend0c4.
  This APAR will prevent the DELETE_CERTIFICATE_DATA call being
  made with a bad token.
  
  Additional Keywords and Symptoms:
  ABEND0C4
  S0C4
  KIXREVJRL
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS Users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: Message DFHSO0001 is issued due to an   *
  *                      0C4 abend in DFHSOSE when CICS is       *
  *                      shutting down.                          *
  ****************************************************************
  CICS has a TCPIPSERVICE with PROTOCOL(HTTP) and SSL(YES).  A new
  request arrives on this TCPIPSERVICE and CWXN is attached to
  process it.  At the same time CICS starts to shut down.
  
  The CWXN task tries to attach the required web alias task, but
  finds the transaction is not enabled for shutdown.  This causes
  CWXN to abend AWB2 and then close the socket.
  
  At the same time the CSOL task has been called to quiesce and is
  in the process of closing all the sockets associated with the
  TCPIPSEVRICE, including the one used by the CWXN task.
  
  During the socket close process the CWXN task needs to switch to
  an S8 TCB.  While changing TCBs the CSOL task closes the same
  socket and freemains the associated control blocks.  Once the
  CWXN task has switched to the S8 TCB it calls DFHSOSE
  FUNCTION(SECURE_SOC_CLOSE).  This initially tries to delete any
  saved certificate data but uses a repository token taken from
  low core because the associated control block has already been
  freed.  After trying to delete the data, which fails, DFHSOSE
  sets the repository token in the freed control block to null
  which causes an 0C4 abend.
  
  Additional keywords:
  msgDFHSO0001  SO0001  ABEND0C4
  

Problem conclusion

• DFHSOCK has been changed to check that the socket is still valid
  after changing TCB mode during close processing.
  

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:

  UI73221 UI73222

Modules/Macros

• DFHSOCK
  

Fix information

• Fixed component name

  CICS TS Z/OS V5

• Fixed component ID

  5655Y0400

Applicable component levels

• R200 PSY UI73222

     UP20/12/19 P F012

• R300 PSY UI73221

     UP20/12/22 P F012

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.