On 25 Oct 2006 08:04:55 -0700, in bit.listserv.ibm-main
(Message-ID:<[EMAIL PROTECTED]>)
[EMAIL PROTECTED] (Robert Wright) wrote:
REFR as an attribute, does this also imply that the code is
self-relocating so that if the code is reloaded at a
different place, it
is still able to continue execution (of the task in
flight)? So PRE DAT
days, this would be an overlay type program, right?
REFR was intended to let the system know that it could
re-fetch a module in
circumstances such as storage going bad or overlays being
detected that
might impact the module. As things like the ability to
load from DCBs that
had been opened temporarily became prevalent, it became
less practical, and
storage was becoming more reliable in the same time
frame. It's no longer
supported but the attribute has been retained.
I also take REFR as a hint that the code can go into
the LPALIB concatenation. It's my understanding that those
modules are paged-in via CLPA and never paged out. When
the storage is needed, the frame can be stolen. When the
code is needed, it can be paged in again from the original.
So, rather than being protection from bad memory,
it's protection from the I/O of page-out.
(I once tried a PMR or Share requirement (I forget
which) that all of IBM's LPA-eligible modules be linked
REFR, along with error messages for any non-REFR modules
loaded via CLPA. It got shot down because the modules are
owned by too many different groups. I guess the idea that
all of those groups are owned by IBM, who *should* have the
last say in QA, was too subtle.)
--
I cannot receive mail at the address this was sent from.
To reply directly, send to ar23hur "at" intergate "dot" com
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
