Mainframe S390 Assembler Programming
I'm not really a doctor, but i did see one on t.v. once
I'm not really a doctor, but i did see one on t.v. once
›› Defining data
Index
The define constant instruction define a piece of storage with a initial value. The word constant is misleading because this value can be changed at any time. This storage area will take up space in your program
label DC dtl'c'
The d defines a duplication factor and this is optional. If left out this will be assumed to be 1. You can use a 0 for the duplication value. For datatypes H, F and D this will for boundry alignement. The bytes skipped in the rocess to get the alignment correct will be set to zero.
The t defines the type of data the storage area will contain.
l Allows you to define an explicit length. Using a length value causes no boundary alignment to occur.
Some of the valid storage types are:
Duplication factor
The duplication factor allow you to tell the assembler to define multiple of the data items you defined. This is optional and if you leave this off the assembler will assume you specified 1. This duplication factor need to be a unsigned numeric value.
F - Fullword
This value is a signed, binary integer and the assembler will convert this value into a 32 bit two's complement value.
The range of numbers storable in a fullword is -2147483648 to +2147483447. This is -2^31 and +2^31.
Padding is on the left with zeros and truncation is not allowed.
H - Halfword
This is the same as a fullword but half the bits. It uses 16 bits or a halfword as the name implies. The maximum values to store in a halfword is -32768 and +32767. This is -2^15 and +2^15.
Padding is on the left with zeros and truncation is not allowed.
X - Hex
Data is allocated on a byte boundary and the minimum number of bytes required to store the value will be used. The data is stored as unsigned hex integer. 2 digits can fit into one byte. If you specify an uneven number of digits the assembler will pad your value on the left with a zero. The maximum number of characters you can store in this data type is 256 bytes. Valid data are hex digits 0-9 and A-F.
Padding is on the left with zeros and truncation will also be on the left.
B - Binary
Allow you to define your data in binary format. If you do not define the data on a 8 bit (1 byte) multiple the assembler will pad your data on the left with zeros.
C - Character data
Defines and valid EBCDIC character data. If you need to define special characters single quote and ampersand you have to define them twice '' or &&.
Padding and truncation is done on the right and the storage area is byte aligned.
P - Packed Decimal data
In packed format each byte represent 2 numbers and the last digit indicates the sign. The number range is from 1 - 31 digits
Padding and truncation is done on the left. Padding is with hexadecimal zeros.The storage area is byte aligned.
Z - Zoned Decimal data
Each byte consists of two pieces. The left half (4 bits) of the byte represent the zone and the right half (4 bits) represent the number.
Padding and truncation is done on the left. Padding is with hexadecimal zeros.The storage area is byte aligned and the data range is 1 to 16 bytes.
A - Address Constant
This is also known as a A-CON. This storage area consist of an absolute or relocatable expression. On difference with A_CON definitions is the constant (c) value is defined in parentheses instead of quotes.
Padding and truncation is done on the left. Padding is with hexadecimal zeros.The storage area is fullword aligned and is 1 to 4 bytes long. No alignment is done if explicit length are specified.
V - Address Constant
This is also known as a V-CON. This storage area will consist of an symbol. As with A_CON's the constant (c) value is defined in parentheses instead of quotes.
The instruction format looks like this: dVln(symbol)
Padding and truncation is done on the left. Padding is with hexadecimal zeros.The storage area is fullword aligned and is 1 to 4 bytes long. No alignment is done if explicit lengths are specified.
This data type is typically used when dealing with sub programs and is used to define a external address. More or sub programs in another section but the next example will show how to define a literal V-CON with the entry point of a sub program and branch to this sub program sitting the return address to R14.
L R15,=V(SUBPGM) /* Entry of SUBPGM stored in R15 */
BALR R14,R15 /* Go to address in R15 and set R14 */
/* to the return address. */
FCONST1 DC F'18' /* Define a fullword containing */
/* 00000012 */
FCONST2 DC F'-5' /* Area will contain FFFFFFFB */
/* The two's complement of +5 is stored */
FCONST2 DC FL3'32' /* Define a two's complement 3 byte no */
/* boundary aligned binary integer */
/* value. The storage area consist of */
/* '000020' */
BCONST1 DC B'1010' /* 00001010 will be stored */
XCONST1 DC X'ABCDE' /* This data stored is 0ABCDE. The 0 */
/* was added to complete the uneven byte*/
XCONST2 DC X'C1E2E2C5D5C2D4C5E1' /* Also store 'ASSEMBLER' */
XCONST3 DC XL3'678' /* Define 3 bytes and the storage will */
/* contain '000678' */
HCONST1 DC H'12' /* '000C' will be in this storage area */
HCONST2 DC 3H'12' /* '000C000C000C' will be stored in */
/* this area */
CCONST1 DC C'ASSEMBLER'/* Define storage area containing the */
/* word 'ASSEMBLER' */
CCONST2 DC CL5'DOG' /* Storage will consist of ' DOG' */
CCONST3 DC CL3'BIGDOG' /* Storage will consist of 'BIG'. */
/* Because the explicit length were */
/* specified truncation occurred. */
ACONST1 DC A(ACONST1) /* Store the address of ACONST1 in this */
/* storage area. */
ACONST2 DC A(*) /* Use the address in the location */
/* counter and store it in this area. */
/* This is the same as A(CONST2). */
ACONST3 DC A(HOME+4) /* Add 4 to the address where HOME is */
/* and store the address in ACONST3 */
ACONST4 DC AL3(HOME+8) /* Add 8 to the address where HOME is */
/* and store the address in ACONST4 but */
/* the address will only be 3 bytes long*/
The instruction format is very similar to the define constant instruction. The major difference is with DS you define or reserve the storage but the storage area does not get initialized to a initial value at definition time. If you do supply a constant value it would not be used. As with DC the duplication value can be set to zero to force boundary alignment. The bytes skipped will not be set to zero.
label DS dtl'c'
Refer to the define storage section for a description of the data types.
FSTG1 DS F /* Reserve a fullword (4 bytes) */
FSTG2 DS 2F /* Reserve 2 fullwords */
HSTG1 DS 50H /* Reserve 50 halfwords */
XSTG1 DS 5X /* 5 bytes reserved with no boundary */
/* alignment */
BSTG1 DS 5B /* Same as above but will describe the */
/* data content to humans more clearly */
/* Do this to make later debugging */
/* easier. */
XTG1 DS X'1111' /* The assembler will determine the */
/* length of the constant define the */
/* storage area with 2 bytes BUT the */
/* area will but be initialized to this*/
/* values. */
ALIGN DS 0F /* */
DS X'12345678' /* Define a data area the size of a */
/* fullword and force boundry alignment*/
/* of a fullword by using a zero */
/* duplication factor. */
Links