>I think the default exspin order is spin-abend-term generating
>abend071 rsn10, 30, 20. Second pass through exspin recovery
>is allowed to retry, third pass through not. That usually takes
>care of all spin loops, unless they were caused by something
>that makes the component that is *validly* spinning spin again
>immediately due to other errors.
>Restart interrupts from the HMC cause reason codes C and less,
>and the unit of work is aleays terminated (according to the book).
I was not talking about spin-loop recovery but about the RESTART
interrupt in general. I was pretty sure I am correct but your and
Kees' responses made me start doubting, so I searched my archives.
I knew I tried this once and I had documented it. Here are my
findings from a while ago (but I assume this has not changed):
- I ran a little looping program, basically a LM followed by a Branch
back to the LM. I named it PHUNLOOP.
- I then initiated the RESTART interrrupt. Since my program was *not*
running disabled, I had to try more than once to catch my program.
I caught *MASTER*, TCP/IP and others before I finally caught
PHUNLOOP.
- Here is what appeared on the master console:
IEE107I THE FOLLOWING MESSAGE IS ISSUED BY SYNCHRONOUS WTO/R SERVICE
*00 BLW004A RESTART INTERRUPT DURING PHUNLOOP PHUNLOOP
ASID=0023 MODE=TASK PSW=078D0000 91900FA8
SYSTEM NON-DISPATCHABILITY INDICATOR IS OFF
REPLY ABEND TO ABEND INTERRUPTED PROGRAM,
RESUME TO RESUME INTERRUPTED PROGRAM,
REPAIR TO PERFORM REPAIR ACTIONS.
- Replying with RESUME lets the program continue to waste CPU cycles :-)
- Replying with ABEND abends the program with a S071 Reason=00000000
- Not responding in time (the WTOR times out) leads to the same abend.
--
Peter Hunkeler
Credit Suisse
----------------------------------------------------------------------
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
