In <[EMAIL PROTECTED]>, on 10/26/2006
at 12:00 PM, David Cole <[EMAIL PROTECTED]> said:
>In MVT days (i.e. pre-DAT days) Type-3 and Type-4 SVCs (and I think
>certain machine check error handling routines, but I'm not sure) were
> loaded into a fixed set of 1K buffers.
The SVC transient areas were 1 KiB. There was also an error recovery
transient area, whose size I don't recall.
>As interrupts would occur, the SVC code (if any) being used by the
>interrupted task would be overlayed in the buffers by SVC code
>(if any) being used by the interrupting task. I.e. the code would be
>"refreshed".
That's not the nomenclature IBM used. IBM used "refresh" to refer to
the reloading of the transient area after a memory parity error.
>So the SVC modules would require the REFR attribute.
Not every SVC module was linked REFR.
>It used to be that reentrant code could be self-modifying as long as
> it did so only temporarily while the hardware was disabled for
>interrupts.
Or you used ENQ. Or you used some other serialization technique. I've
thought for decades that it is inadvisable to write such code.
--
Shmuel (Seymour J.) Metz, SysProg and JOAT
ISO position; see <http://patriot.net/~shmuel/resume/brief.html>
We don't care. We don't have to care, we're Congress.
(S877: The Shut up and Eat Your spam act of 2003)
----------------------------------------------------------------------
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
