Re: Problem with the UNDEL utlity under V5R3 -- MI400

Hi Njal,

UNDELM has one machine instruction that has to be changed with SST.  In
the source, scan for 'MOD:'  You should find

        SETSPP     SRCMSP, ADDR;
        CPYBLA     ADDR, PADDR;
/*MOD:  Mod preceding ins to copy PADDR to SRCMSP  */
/* 3B54 */                     /*  F86100C8  STD  3, 0XC8(1)  */

The F86100C8 is the new machine instruction.  However, it will probably
be different for your pgm because the compiler probably used a different
register.  This shows it using register 3, but I just recompiled it (on
V5R1) and it used register 23.

You can dump the pgm with SST, specifying "Format dis-assembled code"
and find the instruction by comparing MI instruction numbers from the MI
listing.  In the version that's currently on iseriesnetwork.com, it
should be instruction 1F9.

The dump for 1F8 (the SETSPP) and 1F9 (the CPYBLA) should look something
like this (ignoring the 2 address columns at the left):

  003750   3B1FFE48   ADDI 24,31,-440        1F8
  003754   FB0100C8   STD 24,0XC8(1)
  003758   EAFFFE30   LD 23,0XFE30(31)       1F9
  00375C   FAFFFE48   STD 23,0XFE48(31)

In this case the instruction at 00375C should be changed from FAFFFE48
to FAFF00C8:

  00375C   FAFF00C8   STD 23,0XC8(31)

If it's different, you can send me the dump for those instructions and
I'll tell you what the new instruction should be. You can email me
off-list if you want:

  davemck at zbiggroup dot com.

--Dave

Njal Fisketjon wrote:

We have been using the excellent UNDEL utility from time to time withjout any problems. Now with a new box running V5R3 and system value QVFYOBJRST set to 3, the UNDELM program wasn't restored.
I managed to compile the program, and patching it to system state
(security level 40), but it returns BADRRN no matter what we try.
Any suggestions?
Regards
Njal Fiskejon
http://www.fisketjon.no/nfindex.html