Yes, and I am, by your definition, one of the .01%.
But my point was, the contents of R0 may be relevant. Implying it
never matters is not quite right.
However I must agree with you that most people working on/with
z/OS will never need to work at this level. But depending on what
they are doing with VM and/or a z Linux distro....
BTW, I went and looked it up (I have quite a few copies of the
PoOP from different points in time).
This knowledge was once quite important to me because a long time
ago I worked on MDF at AMDAHL. And I was once working on an
emulator for S/370-XA but I think it was Don Higgins (regardless
he lived across the bay from me when I was in Tampa before
Amdahl) beat me to it. And I think that is where Hercules came
from (hazy memory these days).
If you get into the depths of MVS (as in MVS 3.8J or 3.8O I think
it was), you might need that info, so maybe it is more than 0.01%
that will be exposed to this during their lives.
Regards,
Steve Thompson
On 11/6/19 1:44 PM, Steve Smith wrote:
Prefixing is absolutely* irrelevant to this discussion, and to 99.99% of
assembler programmers for their entire lives. It's well-documented if
you're interested, but you probably aren't. Any discussion here is very
likely to be spurious, specious, and superficial.
*I guess it's "really" irrelevant too. :-)
sas
On Wed, Nov 6, 2019 at 1:30 PM Seymour J Metz <[email protected]> wrote:
The contents of R0 is irrelevant to forming the effective address of an RS
or RX instruction. the contents of the prefix register are only relevant
for real addresses.
Now,, if you're running DAT off and you use an instruction that takes an
address from a register, e.g., MVCL, then it is subject to prefixing. But
even there it's in terms of the storage assigned to the LPAR.
