The PSW which is contained in the Fault Analyser Reports shows that
the AMODE at the time of error is 31 (see 99... at the leftmost part of
the instruction address), so the address in R2 is in fact treated as a
31 bit address. But the main task seems to be at RMODE 24 and
was probably AMODE 24. Could it be that the problem is that some
intermediate component switched the AMODE to 31 and didn't switch
it back to 24 on return?

Kind regards

Bernd



Am 17.03.2015 um 18:01 schrieb John McKown:
The problem is R2: R2:  4F1D6048 (Storage invalid)
If you have a dump that you can actually look at, see if there is
something "reasonable" at 001D6048. Otherwise, you'll need to try to
backtrack where this value came from. If the data at address  001D6048
"looks reasonable", then you need to figure out how the high order
byte got set ot x'4F'.

Well, I know nada about that package. The fact that the CSECT names
all start with @MEM is "suspicious". One question would be what
release of z/OS did it last run on successfully? And I have two
entries you could try with _might_ help. But, then again, they might
not. In the DIAGnn member of PARMLIB, try putting in two lines like:

VSM USEZOSV1R9RULES(YES)
CBLOC VIRTUAL24(IHALCCA,IHAPCCA)

http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/IEA2E2B0/31.6

and, less likely, in IEASYSnn you might try

CSCBLOC=BELOW

I've had really old programs blow up when the CSCB is above the line.
But this can have a negative impart on the use of common memory below
the line. Please read up on this.
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/iea2e2b0/54.8.14

The first two can be dynamically adjusted by putting them in a DIAGnn
member of PARMLIB and the issuing the command: T DIAG=nn
The latter requires an IPL to change.

On Tue, Mar 17, 2015 at 11:12 AM, Jerry Criste
<[email protected]> wrote:
Hello,

We are experiencing an S0C4 abend after upgrading to z/OS 1.13.  The 
application is known as KBMS, has been unsupported since 2000, and is critical 
to the business.  We are working on a replacement, but full implementation is 
still several months away.

We have experienced problems with the application in previous upgrades, but 
were able to address it by re-linking some of the programs (we do not have the 
source). Unfortunately that approach is not working this time, so I'm hoping 
someone can explain what might be happening and possible solutions.

Thanks in advance for any insight that is provided.

<snip>
Instructions around point of failure:

   Offset Hex            Instruction
   ------ -------------- ---------------------------------------------
      -10 90C8 C01C      STM     R12,R8,28(R12)
       -C 18FC           LR      R15,R12
       -A 41C0 C050      LA      R12,80(,R12)
       -6 18EB           LR      R14,R11
       -4 5820 F000      L       R2,0(,R15)
    ***** 5870 2005      L       R7,5(,R2)
       +4 D203 F014 7001 MVC     20(4,R15),1(R7)
       +A 5860 F004      L       R6,4(,R15)
       +E 1876           LR      R7,R6
      +10 4170 0007      LA      R7,7
      +14 1476           NR      R7,R6
      +16 1277           LTR     R7,R7

Program Status Word (PSW) . : 078D2000 9970FB58

General Purpose Registers:
   R0:  00030000 (651264 bytes of storage addressable)
   R1:  1970FCD8 (Module AICKBMSE CSECT @MEMFREE + X'0')
   R2:  4F1D6048 (Storage invalid)
   R3:  199DE000 (1069056 bytes of storage addressable)
   R4:  00000000 (2048 bytes of storage addressable)
   R5:  00000001 (2047 bytes of storage addressable)
   R6:  00000000 (2048 bytes of storage addressable)
   R7:  00000001 (2047 bytes of storage addressable)
   R8:  00000008 (2040 bytes of storage addressable)
   R9:  104D6000 (Storage invalid)
   R10: 199C6008 (1167352 bytes of storage addressable)
   R11: 1970FB48 (Module AICKBMSE CSECT @MEMALOC + X'0')
   R12: 199C86D8 (1157416 bytes of storage addressable)
   R13: 9970FE80 (Module AICKBMSE CSECT @MEMZER0 + X'20')
   R14: 1970FB48 (Module AICKBMSE CSECT @MEMALOC + X'0')
   R15: 199C8688 (1157496 bytes of storage addressable)


Virtual Storage Map
AREA                                                                     
ADDRESS                             SIZE
64-BIT HIGH PRIVATE AREA                         0002000000000000       8191P
64-BIT SHARED AREA                                   0000020000000000       510T
64-BIT COMMON AREA                                 000001EF80000000       
67583.9M
64-BIT LOW PRIVATE AREA                          0000000880000000       1948.00G
64-BIT JAVA AREA                                            0000000080000000    
   30720.0M
----------------------- 2 GIG BOUNDARY -----------------------
EXTENDED PRIVATE AREA                                            19700000       
   1641M
EXTENDED COMMON SERVICE AREA (ECSA)        09C28000          256864K
EXT MODIFIED LINK PACK AREA (MLPA)                09C27000             4K
EXT FIXED LINK PACK AREA (FLPA)                           09C24000            
12K
EXT PAGEABLE LINK PACK AREA (PLPA)                 05848000         69488K
EXTENDED SYSTEM QUEUE AREA (ESQA)               01B10000         62688K
EXTENDED NUCLEUS                                                      01000000  
       11328K
---------------------- 16 MEG BOUNDARY -----------------------
NUCLEUS                                                                         
    00FD3000           180K
SYSTEM QUEUE AREA (SQA)                                       00EB7000          
1136K
PAGEABLE LINK PACK AREA (PLPA)                          00CF8000          1788K
FIXED LINK PACK AREA (FLPA)                                    00000000         
    0K
MODIFIABLE LINK PACK AREA (MLPA)                    00000000             0K
COMMON SERVICE AREA (CSA)                                 00A00000          
3040K
PRIVATE AREA                                                                   
00006000         10216K
PREFIXED SAVE AREA AND SYSTEM                          00000000            24K


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to