The code which Peter Relson put into module CSVRBBLD in 1998 says:
XSBG64HE = 0; /* Make sure return in 64-bit
mode will work. If he BR's back, this will be
OK. If he BSM's back, he'll be in 24-bit
AMODE which is OK. */
/* We set the high order bit of GPR 14 so that the target
routine can tell if the caller is AMODE 24 or 31.
But as GPR 14 is also used as a return register, we cannot
do this for the AMODE 64 target: if we set the low order
bit, that would require the target to BSM back. If it issued
BR, that would program check. @L8A*/
IF ¬NewPRBP->PswINRB.PswAMODE64 THEN /* @L8A*/
IF EXSA31MS=ON THEN /* caller in 31/64 bit mode? @L8C*/
RBGRS14=RBGRS14|MODEBIT; /* yes, indicate so to exit rtn */
ELSE; /* caller in 24 bit mode @ZMC1694*/
ELSE DO; /* Target is AMODE 64 @L8A*/
; /* We cannot do anything except
provide a "normal" address here
so that the return in AMODE 64
will work OK @P6A*/
The LINK/LINKX documentation says:
14
Contains the return address the called module will return to. If the
high-order bit of this register is on, the issuer of the LINK or LINKX macro is
running in 31-bit mode; if off, the issuer is running in 24-bit mode.
15
Requested program's entry point address
When the target of the LINK or LINKX is AMODE(64), then reg 15 contains
xxxxxxxY
where Y is:
0 if the caller was AMODE 24
2 if the caller was AMODE 31
4 if the caller was AMODE 64
The ATTACH/ATTACHX documentation says:
14
Return address. Bit 0 is 0 if the subtask routine gets control in 24-bit
addressing mode; bit 0 is 1 if the subtask routine gets control in 31-bit
addressing mode.
15
When the subtask routine is to run in 24-bit or 31-bit addressing mode, the
entry point address of the subtask routine.
When the subtask routine is to run in 64-bit addressing mode, it is
expected to use relative branching and register 15 contains a value that can be
used to determine the addressing mode of the issuer of the ATTACH or ATTACHX
macro as follows:
Issuer AMODE 24: X'FFFFF000'
Issuer AMODE 31: X'FFFFF002'
Issuer AMODE 64: X'FFFFF004'
Jim Mulder
-----Original Message-----
From: IBM Mainframe Discussion List <[email protected]> On Behalf Of Jon
Perryman
Sent: Saturday, August 16, 2025 7:35 PM
To: [email protected]
Subject: Re: Running a program RMODE64
>That has nothing to do with whether the target module can be RMODE64.
R15 is the target module address (not R14).=20
>>Since RMODE is not documented for ATTACHX, RMODE64 is not supported.
>>Therefore, R14 will never be above the bar (64 bit address).
>False. That says you must not invoke ATTACHX in RMODE64.=20
R14 (return address) refers to the ATTACHX module (not target module) whi=
ch as we both agree can never be RMODE64. R14 can handle but can the addr=
ess be above the bar with the RMODE64 restriction?
>>R14 is "return address" which implies R14 will always be a valid addres=
s regardless of the AMODE.
You included this as part of the snip which needs some clarification. Are=
you saying that R14 might not be valid for the AMODE upon entry and that=
we should be using BSM instead of BR? Is IEFBR14 actually a BSM now? Thi=
s is going to be a hard habit to break.
>:>ROTFLOL, this is why RMODE capability shouldn't be documented as a
>requirement. Out of sight, out of mind.
>What does RMODE have to do with it? My statement was regarding AMODE.
>The R14 values for both 24 & 31 bit are appropriate for BSM and coincidentally
>BR.
IBM has geniuses developing simple yet effective solutions to problems. RMODE
plays a huge role in AMODE but is not always blatantly obvious.
For ATTACHX, what do we know about R14 (return address) because RMODE64 is
prohibited? Since the load module is required to be RMODE31 or RMODE24, R14
will always be a valid address below the bar (31-bit). It may be in a 64-bit
register but is guaranteed to be a valid address below the bar. IBM does not
document 64-bit R14 because it is not possible. You can argue it should be
documented but RMODE restriction tells us the valid R14 values.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
