IBM Mainframe Forum

by **michel123** » Wed Nov 23, 2011 4:56 pm

Hello,

I have a concern when the linkage of two related programs through a call.
I have a program P1 with a call to an entry point GDATA placed in a program SP2.

In P1 :
LA R13,SAVE1
CALL GDATE,(PARAM)

In SP2 :
SP2 CSECT
ENTRY GDATE
YREGS
*-------------------------------------------------------------------
GDATE SAVE (14,12),,SP2-&SYSDATE-&SYSTIME
LR R12,R15 R12=R15 : ADRESSE POINT D'ENTREE
USING SP2,R12
ST R13,SAVE2+4 ADRESSE SAVE APPELANT CHEZ NOUS
LA R11,SAVE2 R11 POINTE SUR NOTRE SAVE
ST R11,8(R13) ADRESSE NOTRE SAVE CHEZ APPELANT
DROP R15

compilation de SP2 :
//L.SYSIN DD
ENTRY GDATE
NAME SP2(R)
/*

compilation de P1
//L.SYSIN DD *
INCLUDE SYSLMOD(P1)
INCLUDE SYSLMOD(SP2)
NAME P1(R)
/*

The execution gives me a 0C1 !

by **BillyBoyo** » Wed Nov 23, 2011 5:16 pm

Try to use the Code tags to preserve formatted data. Highlight the text. Click on the Code button. Click on the Preview button to see what it will look like on the forum. Amend as necessary. Click on Submit when happy.

Presumably the S0C1 is on the call?

Can you post the output from the linkedit steps please?

by **michel123** » Wed Nov 23, 2011 5:27 pm

I gives again the code :

Prgramm P1 :

Select all

TRAIT DS 0H
*
* L R15,=V(GDATE)
* LA R1,=A(PARAM)
LA R13,SAVE1
* BALR R14,R15
CALL GDATE,(PARAM)

Program P2 :

Select all

SP2 CSECT
PRINT GEN SPARE US THE MACRO EXPANSIONS
ENTRY GDATE
YREGS
*--------------------------------------------------------------------*
GDATE SAVE (14,12),,SP2-&SYSDATE-&SYSTIME
LR R12,R15 R12=R15 : ADRESSE POINT D'ENTREE
USING SP2,R12
ST R13,SAVE2+4 ADRESSE SAVE APPELANT CHEZ NOUS
LA R11,SAVE2 R11 POINTE SUR NOTRE SAVE
ST R11,8(R13) ADRESSE NOTRE SAVE CHEZ APPELANT
DROP R15
*--------------------------------------------------------------------*
*
INIT DS 0H

And :

Select all

IEA995I SYMPTOM DUMP OUTPUT
SYSTEM COMPLETION CODE=0C1 REASON CODE=00000001
TIME=05.44.08 SEQ=10271 CPU=0000 ASID=0026
PSW AT TIME OF ERROR 078D0000 00007CF4 ILC 2 INTC 01
ACTIVE LOAD MODULE ADDRESS=00007A38 OFFSET=000002BC
NAME=P1
DATA AT PSW 00007CEE - E000100C 00000000 00009240
AR/GR 0: 80C972CA/00000000_00000000 1: 00000000/00000000_00007CB4
2: 00000000/00000000_00007B44 3: 00000000/00000000_009D09D4
4: 00000000/00000000_009D09B0 5: 00000000/00000000_009FD098
6: 00000000/00000000_009C0FE0 7: 00000000/00000000_FD000000
8: 00000000/00000000_009F91C8 9: 00000000/00000000_009CED08
A: 00000000/00000000_00000000 B: 00000000/00000000_00007D80
C: 00000000/00000000_00007C68 D: 00000000/00000000_00007C0C
E: 00000000/00000000_00007DD6 F: 00000000/00000003_00000000
END OF SYMPTOM DUMP
IEF472I PG8997X P1 - COMPLETION CODE - SYSTEM=0C1 USER=0000 REASON=00000001

by **steve-myers** » Wed Nov 23, 2011 5:56 pm

First, it is essentially impossible to place dump type data or Binder module map output in code tags. Don't try.

Second, with no Binder map, the indicative dump you tried to give us is basically useless.

Third, at least in concept, the entry linkage for GDATE is incorrect, though it may be correct in fact. Register 15 contains the address of GDATE, not SP2. It appears they are the same here so it should be OK.

Fourth, we have no idea how addressability in the program that called GDATE is setup, so that may well be the root cause of your problem, not GDATE in SP2.

by **BillyBoyo** » Wed Nov 23, 2011 7:30 pm

Have you located the where the PSW is pointing in your program? It doesn't look like the "unresolved module" sort of S0C1.

by **michel123** » Wed Nov 23, 2011 10:14 pm

In point 4 of Steve, I have compiled as follows:

first the program SP2 :

Select all

//ASSEMBLE EXEC ASMACL
//C.SYSIN DD DISP=SHR,DSN=PG8997.ASM.SOURCE(SP2)
//C.SYSLIB DD DISP=SHR,DSN=SYS1.MACLIB
// DD DISP=SHR,DSN=SYS1.MODGEN
// DD DISP=SHR,DSN=PG8997.ASM.COPY
//L.SYSLMOD DD DISP=SHR,DSN=PG8997.BATCH.LOADLIB
//L.SYSIN DD *
ENTRY GDATE
NAME SP2(R)
/*

Then the program P1 :

Select all

//ASSEMBLE EXEC ASMACL
//C.SYSIN DD DISP=SHR,DSN=PG8997.ASM.SOURCE(P1)
//C.SYSLIB DD DISP=SHR,DSN=SYS1.MACLIB
// DD DISP=SHR,DSN=SYS1.MODGEN
//L.SYSLMOD DD DISP=SHR,DSN=PG8997.BATCH.LOADLIB
//L.SYSIN DD *
INCLUDE SYSLMOD(P1)
INCLUDE SYSLMOD(SP2)
NAME P1(R)
/*

by **steve-myers** » Thu Nov 24, 2011 1:51 am

In the Binder control statements in the JCL to assemble and link P1 you should explicitly specify an entry point using an ENTRY statement as you did for SP2. It probably should be -

ENTRY P1

The INCLUDE SYSLMOD(P1) statement is unneeded, and most likely will cause an error. It's uneeded because the JCL has the effect of bringing the object output from the assembly into the Binder before any of the control statements are processed. It can cause an error in two ways:

The very first time you run the JCL, presumably there is no P1 module, so the INCLUDE statement will fail.
Afterwards, the INCLUDE statement will load modules P1 and SP2 (which may be in load module P1). Since the Binder already knows about SP2 it will effectively ignore the INCLUDE SYSLMOD(SP2) statement. More accurately, it will ignore module SP2 in the SP2 load module specified by the INCLUDE statement.

IBM Mainframe Forum

How linker program?

How linker program?

Re: How linker program?

Re: How linker program?

Re: How linker program?

Re: How linker program?

Re: How linker program?

Re: How linker program?