A potential deadly embrace caused by a disabled spin loop while waiting for 
something else to happen on another processor has already been addressed.  z/OS 
can detect that situation  and ABEND the process that is in the disabled spin 
loop.  If that process happens to be trying to take another processor offline, 
then that process needs to have some recovery logic in place to deal with this 
contingency.  Perhaps "stop trying to take the other CPU offline" or "ask the 
operator" are some logical choices. 



A deadly embrace in which both processes are enabled and waiting on each other 
is different. 

Bill Fairchild 
Franklin, TN 


----- Original Message -----
From: "Jon Perryman" <[email protected]> 
To: [email protected] 
Sent: Thursday, September 12, 2013 9:29:23 AM 
Subject: Re: Wait all CPU's enabled at least once 

I asked out of curiosity. I wanted to read how it worked. I wondered if it is 
really fast and how it avoided a deadly embrace. 

Jon Perryman.  


----- Original Message ----- 
> From: Elardus Engelbrecht <[email protected]> 
>  
> Jon Perryman wrote: 
> 
>> Is there an instruction that waits for all CPU's to be enabled at least 
> once? 
> 
> I saw Jim's reply and reread my PrOPs, but why do you want all CPU's to 
> be enabled? What type of CPU? 
> 
> Alternatively, what do you want to solve/achieve? 
> 
> Just curious of course if you don't mind please. 


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

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

Reply via email to