Why would anyone use the definition of ADDR2 shown below?
Why the question arose follows.
A recent problem reported to the z390 project
(Don Higgins' z390 assembler project) about an invalid
displacement generated in the "L 3,ADDR3" instruction
shown below. The problem boiled down to a z390 error
processing the statement "ADDR2 DC A(DWD+DWD)", which
does contain a complex relocatable expression. Sample
source provided by the user reporting the problem follows.
CPXRELEX CSECT
USING CPXRELEX,15
L 1,ADDR1
L 2,ADDR2
L 3,ADDR3
BR 14
ORG CPXRELEX+X'100'
DWD DS D
ADDR1 DC A(DWD)
ADDR2 DC A(DWD+DWD) Complex relocatable expression
ADDR3 DC A(DWD+1)
END
HLASM correctly assembles the sample source.
Running the program on z/OS, with the "BR 14" replaced by
"DC H'0'", shows (in the summary dump) registers 1, 2, and 3
set correctly (R1 contains the address of DWD, R2 = 2 * R1, and
R3 = R1 + 1).
The problem has been fixed in z390.
Thanks for any insight.
