DB2 FOR LUW / DB2FORLUW / A50 - DB2

There is a tiny timing hole in our codes that could result in 
trap.  Buffer Manipulator(BM) still tries to access db2 agent's 
control block after they mark themselves complete, which could 
result in the trap, because once db2 agent gets notified that bm 
completes, it will clean up agent's control block. If bm happens 
to access the agent's control block after db2 agent cleans up 
the control block, a trap will be caused. 
 
     // Mark that the BM is done. 
     // 
---------------------------------------------------------------- 
--------- 
     if (pCB->childEntry != NULL) 
        pCB->childEntry->childState.markAsComplete(); 
 
      if ( pBRAgtCB->showAllBarStats ) 
      { 
        pdLogPrintf(PD_DEV | PD_LLEV_L, SQLT_sqluBMTerm, 0, 
__LINE__, PD_LEVEL_INF, 
                    "BufAddr  Read:  Success = " OSS_PRIdU32 ", 
Failure = " OSS_PRIdU32 "\n" 
                    "BufAddr  Write: Success = " OSS_PRIdU32 ", 
Failure = " OSS_PRIdU32 "\n" 
                    "Ms 
..................

**************************************************************** 
* USERS AFFECTED:                                              * 
* ALL                                                          * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* See Error Description                                        * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Upgrade to Db2 10.5 Fix Pack 9 or higher                     * 
****************************************************************

no local fix

First fixed in Db2 10.5 Fix Pack 9

not known / see Local fix

Date  - problem reported    :
Date  - problem closed      :
Date  - last modified       :

11.04.2017
29.09.2017
29.09.2017

9.0.