On Thu, 11 Feb 2016 12:44:54 -0600, Paul Gilmartin wrote:

>I had thought (but I can't cite) that LPSW is sensitive to the difference 
>between
>a scrunched and a nonscrunched PSW (flag bit in a CR or in the PSW itself?)
>and capable of interpreting either correctly.

LPSW loads a 64-bit PSW. LPSWE loads a 128-bit PSW. This is clearly documented 
in the Principles of Operation.

>>​right. 64 bit z/Linux​ already allows execution above the 2 GiB limit
>>imposed by z/OS (and z/VSE, z/VM ??). Also, z/Linux does not have the
>>"bar". It has a "line" between < 2GiB (31 bit) and >= 2GiB (64 bit) because
>>it never had the "bit 0 is a flag bit" problem.
>>
>Some of those bits are set/interpreted by the hardware.  Linux can't sidestep
>them.  But it has little effect on problem state programs.

The bits in the PSW that specify the addressing mode are indeed applicable to 
Linux. 
I think that what John meant is that Linux has no need to support AMODE 24.

>Big-endian?  Little-endian?  I understand that AMODE 64 is indicated by the
>low order bit of the address (in the OPSW or returned by LOAD macro?);
>AMODE 31 by the high order bit provided that AMODE is not 64.

The AMODE of a module is returned to the issuer of LOAD as you described. Bit 
63=1 
indicates that it is AMODE 64. If not, bit 32 indicates that it is AMODE 31. 
Both bits 
equal 0 means that it is AMODE 24. These are the same specifications as used by 
BASSM and BSM.

-- 
Tom Marchant

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to