IBM Mainframe Forum

by **sensuixel** » Tue Feb 22, 2011 3:10 pm

Hi

I'm a french junior System Adm. and i'm actually trying to build skills in HLASM.

I've coded a little program to retrieve the JCL PARM field and display it in a file, it almost works properly except for the length.

e.g if i code //APIASMX0 EXEC PGM=APIASMX0,PARM='4095'

The display will only be *** PARM = 40 instead of 4095, i only get the first 2 characters of the parm field.

Moreover i get ** ASMA033I Storage alignment for MSGPRM unfavorable at compilation, it's just an informationnal
message but still it bothers me a little.

note : INITL and RCNTL are our standard linkage and exit macro.

Am I missing something obvious ?

Select all

*---------------------------------------------------------------------- 00010000
*- PURPOSE : READ JCL PARM FIELD 00020092
*---------------------------------------------------------------------- 00030000
APIREC DSECT 00031038
APICNT DS CL131 BUFFER SYSTEM MAPPING FOR LOCATE MODE 00032099
* 00032199
APIX0 CSECT 00033024
INITL 10,EQU=R BASE REGISTER = 10, R15 = 15 00040039
USING APIREC,R5 ADDRESSAGE POUR DSECT VIA R5 00041099
*---------------------------------------------------------------------- 00050000
* MAIN 00060000
*---------------------------------------------------------------------- 00070000
BAL R6,READPARM RETRIEVE JCL PARM FIELD 00071025
BAL R6,OPENFIC OPEN SYSPRINT FILE 00071199
BAL R6,WRITEFIC WRITE PARM IN SYSPRINT 00071225
BAL R6,CLOSEFIC CLOSE SYSPRINT FILE 00071325
BAL R6,EXITPGM EXIT 00071425
* 00071599
*---------------------------------------------------------------------- 00071699
* READPARM : RETRIEVE JCL PARM 00071799
*---------------------------------------------------------------------- 00071899
READPARM EQU * 00071999
DS 0H 00072099
L R2,0(0,R1) R2 = adr of parm field 00072199
LH R9,0(0,R2) R9 = Length of parm field 00072299
L R3,0(0,R2) R3 = Content of parm field 00072399
ST R3,MSGPRM MSGPRM = content of parm field 00072499
LTR R9,R9 Was parm specified ? 00072599
BZ NOPARM If NO branch to NOPARM 00072699
BR R6 00072799
*---------------------------------------------------------------------- 00072822
* OPENFIC : OPEN SYSPRINT 00072992
*---------------------------------------------------------------------- 00073022
OPENFIC EQU * 00073122
OPEN (APIPRINT,(OUTPUT)) Open sysprint file as output 00073255
BR R6 00073355
* 00073922
*---------------------------------------------------------------------- 00074022
* NOPARM : WRITE ERROR MSG IN SYSPRINT 00074192
*---------------------------------------------------------------------- 00074222
NOPARM EQU * 00074350
BAL R6,OPENFIC 00074499
PUT APIPRINT 00074599
LR R5,R1 00074699
MVC 0(ERRMSG1,R5),ERRMSG 00074799
BAL R6,CLOSEFIC 00074899
RCNTL RC=4 00074999
* 00075022
*---------------------------------------------------------------------- 00075150
* WRITEFIC : WRITE SYSPRINT FILE 00075292
*---------------------------------------------------------------------- 00075350
WRITEFIC EQU * 00075450
PUT APIPRINT 00075550
LR R5,R1 00075699
MVC 0(LLINE1,R5),LINE1 00075799
BR R6 00075850
* 00075950
*---------------------------------------------------------------------- 00076050
* CLOSEFIC : CLOSE SYSPRINT 00076192
*---------------------------------------------------------------------- 00076250
CLOSEFIC EQU * 00076350
CLOSE APIPRINT 00076450
BR R6 00076599
* 00076650
*---------------------------------------------------------------------- 00076750
* EXITPGM : END OF PGM 00076850
*---------------------------------------------------------------------- 00076950
EXITPGM EQU * 00077050
RCNTL RC=0 00077199
* 00077250
*---------------------------------------------------------------------- 00077350
* WORKING STORAGE 00077450
*---------------------------------------------------------------------- 00077550
* 00077650
LINE1 EQU * 00077750
DC C'*** PARM = ' 00077899
MSGPRM DC CL120' ' 00077999
LLINE1 EQU *-LINE1 00078150
* 00078299
ERRMSG EQU * 00078399
DC C'*** RC = 4 NO PARM SPECIFIED' 00078499
DC CL103' ' 00078599
ERRMSG1 EQU *-ERRMSG 00078699
* 00078799
APIPRINT DCB DDNAME=APIPRINT,DSORG=PS,MACRF=(PL), X00078899
RECFM=FB,LRECL=131,BLKSIZE=1310 00078999
END APIX0 00079050

by **enrico-sorichetti** » Tue Feb 22, 2011 3:37 pm

why use a load and store, and not a simple mvc ?
if the length is at offset 0 the data will be at offset 2 ( at least )
the warning is about fullword alignment for load and store operations
once upon a time You would get a specification exception at execution time
then You might lose ( IIRC ) two cycles in accessing the data
right now You will get a warning that can be turned off using the proper assembler parms

and ... why use locate mode, much simpler to use move mode and issue the put from Your IO area

by **sensuixel** » Tue Feb 22, 2011 3:50 pm

Thank you !

It works perfectly, no more Informationnal message

I've replaced the Load and store operation with

Select all

MVC MSGPRM,0(R2)

So i get the parm field properly.

I used LOCATE mode just to get used to it, I also have a version with MOVE mode (easier to use i agree).

Thank you really much for your contribution.

by **stevexff** » Tue Feb 22, 2011 6:49 pm

I prefer to use symbolic names as much as possible, so I probably would have used a DSECT to map the PARM

Select all

PARM DSECT
PARM_LEN DS H halfword length
PARM_VAL DS CL100 passed parm
..
L R2,0(,R1)
USING PARM,R2

and possibly an EXecuted MVC based on the length to move it around.

by **enrico-sorichetti** » Tue Feb 22, 2011 6:57 pm

Select all

LH R9,0(0,R2) R9 = Length of parm field 00072299
L R3,0(0,R2) R3 = Content of parm field 00072399

I hinted about a wrong offset for the data

if the length is at offset 0 the data will be at offset 2 ( at least )

the TS disregarded it and coded

Select all

MVC MSGPRM,0(R2)

telling

So i get the parm field properly.

if he is happy with errors why should we make him unhappy by fixing things :ugeek:

by **steve-myers** » Tue Feb 22, 2011 8:17 pm

Select all

GETPARM CSECT DEFINE PROGRAM CSECT
USING *,12 ESTABLISH PROGRAM ADDRESSABILITY
SAVE (14,12),,* SAVE REGISTERS
LR 12,15 COPY ENTRY POINT ADDRESS TO REG 12
LA 15,SAVEAREA LOAD ADDRESS OF THE NEW SAVE AREA
ST 15,8(,13) ADD THE NEW SAVE AREA TO THE
ST 13,4(,15) SAVE AREA CHAIN
LR 13,15 ESTABLISH A NEW SAVE AREA POINTER
L 2,0(,1) LOAD ADDRESS OF PARM DATA
LH 3,0(,2) LOAD LENGTH OF PARM DATA
LA 2,2(,2) COMPUTE ADDRESS OF FIRST BYTE OF X
TEXT IN THE PARM DATA
LTR 3,3 TEST LENGTH OF PARM DATA
BNZ MOVEPARM BR IF NO PARM DATA
LA 2,ERRTEXT LOAD ADDRESS OF ERROR TEXT
LA 3,L'ERRTEXT LOAD LENGTH OF ERROR TEXT
MOVEPARM LA 14,PARMDATA LOAD ADDRESS OF MY COPY OF PARM DATA
LA 15,L'PARMDATA LOAD LENGTH OF MY COPY OF PARM DATA
ICM 3,B'1000',=C' ' LOAD THE FILL CHARACTER
MVCL 14,2 COPY PARM DATA TO MY STORAGE AREA
OPEN (ADCB,OUTPUT) OPEN THE DCB
PUT ADCB,PARMDATA WRITE THE PARM DATA
CLOSE ADCB CLOSE THE DCB
EXIT L 13,4(,13) LOAD ADDR OF THE PREVIOUS SAVE AREA
RETURN (14,12),T,RC=0 RESTORE REGISTERS AND RETURN
ADCB DCB DSORG=PS,MACRF=PM,RECFM=F,LRECL=100,BLKSIZE=100, X
DDNAME=SYSPRINT
SAVEAREA DS 18F OS/360 SAVE AREA
LTORG , DEFINE THE LITERAL POOL
ERRTEXT DC C'THERE IS NO PARM DATA!'
PARMDATA DS CL100 MY COPY OF PARM DATA
END GETPARM

I don't think I ever bothered to code a parm data DSECT like stevexff.

I suspect the fundamental problem (which I did not bother to check in a formal sense because the expansion of the non-standard entry macro is unknown. so I'm not sure what may have happened to register 1) in the original code is there is no equivalent to my LA 2,2(,2) before the data is copied to the program area.

As for the "unfavorable" storage alignment message. At one level it is nothing to worry about. The original System/360 had strict rules about storage alignment for operands. For the most part these storage alignment rules were lifted in System/370 around 1970, but there were warnings, which are still present, that ignoring the old storage alignment rules will cause a performance penalty. The alleged performance penalty is why the Assembler writes the warning message.

by **enrico-sorichetti** » Tue Feb 22, 2011 8:23 pm

I suspect the fundamental problem (which I did not bother to check in a formal sense because the expansion of the non-standard entry macro is unknown. so I'm not sure what may have happened to register 1) in the original code is there is no equivalent to my LA 2,2(,2) before the data is copied to the program area.

any half decently written prolog/entry macro will restore the R1 content to its original value !

by **sensuixel** » Tue Feb 22, 2011 8:30 pm

stevexff wrote:I prefer to use symbolic names as much as possible, so I probably would have used a DSECT to map the PARM
Select all
PARM DSECT
PARM_LEN DS H halfword length
PARM_VAL DS CL100 passed parm
..
L R2,0(,R1)
USING PARM,R2
and possibly an EXecuted MVC based on the length to move it around.

Thank you for your contribution, i'll try it but first I have to understand how the EX instruction works.

@steve-myers
You're right I don't have the equivalent of LA 2,2(,2) in my code.
I have to take a closer look at it.

by **sensuixel** » Tue Feb 22, 2011 8:36 pm

enrico-sorichetti wrote:
Select all
LH R9,0(0,R2) R9 = Length of parm field 00072299
L R3,0(0,R2) R3 = Content of parm field 00072399
I hinted about a wrong offset for the data
if the length is at offset 0 the data will be at offset 2 ( at least )
the TS disregarded it and coded
Select all
MVC MSGPRM,0(R2)
telling
So i get the parm field properly.

if he is happy with errors why should we make him unhappy by fixing things

Oups I didn't read tour answer thoroughly

You're right, when I activate HEX MODE in my file I do see the length of the PARM FIELD, it's not a clean ouput.

by **stevexff** » Tue Feb 22, 2011 9:03 pm

The EX instruction allows you to execute another out-of-line instruction (like an MVC for example), overriding the length field with the contents of a register. This allows you do a MVC where the length is not known until run time, like the length of your PARM field. steve-myers has worked around this by using MVCL, which also allows you to specify the length in a register, and works just as well as long as you happen to have four registers to spare; this isn't always the case

IBM Mainframe Forum

Retrieve JCL PARM properly

Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly

Re: Retrieve JCL PARM properly