PQ73078: ABEND0C1 IN TCPIP AFTER PURGEDQ

Obtain the fix for this APAR.

APAR status

• Closed as unreproducible in next release.

Error description

• During application termination a purgedq was done.  This caused
  suspended SRB work to be abended with an ABEND47B.  A TCPIP
  task for the application was suspended and began abend recovery.
  Because the abend was retryable the retry routine was executed.
  The program return from this routine should have returned to
  the socket call that was being processed at the time of the
  abend but the linkage stack still contained the address that
  pointed immediately after the suspension. This caused unexpected
  processing with a bad linkage stack and register save areas.
  This resulted in an ABEND0C1 when a bad register was used in
  a branch instruction.
  VERIFICATION STEPS:
    1) Task will be in SRB mode (ducb_tcb = 00000000)
    2) ducb_duae_suspended = '1' indicating recovery found the
       ducb suspended
    3) Review systrace for recovery retry (RCVY RTRY) entry
       before current abend followed closely by a PR entry where
       the from and returning address modules do not make sense
  

Local fix

• KEYWORDS: ABEND 0C1 47B
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All users of the IBM Communications Server   *
  *                 for z/OS Version 1 Release 4 IP              *
  ****************************************************************
  * PROBLEM DESCRIPTION: Unpredictable results may occur when    *
  *                      the TCPIP recovery routine requests a   *
  *                      retry for a retryable abend.  Specific  *
  *                      error reported is an abend0C1 following *
  *                      a retry from an abend47B.               *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  An SRB mode application unit of work was suspended during a
  socket call due to lock contention.  The specific module flow
  the socket call passed through established a retry label that
  would release any held locks and return an error code on the
  socket call.  The application space was stopped while the SRB
  was suspended, which resulted in a PURGEDQ being issued.  The
  suspended SRB failed with an abend47B and the TCPIP recovery
  routine requested a retry at the retry label.  The error code
  was set and the retry routine attempted returning to the caller
  via a PR instruction.  The current linkage stack entry was
  created by the PC for the SUSPEND macro instead of the entry
  to the socket call module.  The PR results in the PSW being
  set to the instruction following the PC for the SUSPEND.
  The execution flow in the reported case resulted in an abend0C1
  however the results are unpredictable.
  +-------------------------------------------------------------+
  + Please check our Communications Server for OS/390 homepages +
  + for common networking tips and fixes.  The URL for these    +
  + homepages can be found in Informational APAR II11334.       +
  +-------------------------------------------------------------+
  

Problem conclusion

Temporary fix

Comments

• The recovery logic has been altered to avoid retrying when the
  unit of work is suspended on a TCPIP lock.
  

APAR Information

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

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

  122PC5Ÿ UQ77757 104AC1Ÿ PQ75388

Modules/Macros

• EZBITDRC
  

Fix information

• Fixed component name

  TCP/IP V3 MVS

• Fixed component ID

  5655HAL00

Applicable component levels

• R140 PSY UQ77757

     UP03/07/15 P F307

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.