Vic wrote: > > > If you check all the PSA's you should be able to determine if > > the task is dispatched. > > I was also considering the PSA approach by comparing each PSATOLD with the > address of the TCB in question. If there's match, then its active. If > there's no match and the TCB is not waiting or paused can I assume that > it's dispatchable?
No, you have to simultaneously compare PSAAOLD with the address space you're interested in and PSATOLD with the task you are interested in because you will find TCBs at the same address in many address spaces. One "creative" way to do this is PLO.CL where you use the ASCB address as the compare value for PSAAOLD and load from PSATOLD. If the compare fails you just skip that PSA because the associated cpu isn't executing any task (or SRB) in that address space. But even if the TCB you're interested in happens to appear to be dispatched at the moment you looked, there's no guarantee at all that it is still running at the end of the instruction(s) you used to decide that it was running. That's the "feature" of asynchronous execution, there's just no way to know for sure what the state of any other unit of work is from moment to moment. > The TCBTTIME approach won't work for me since I need to know the status at > the time the sample is taken. If you are disabled at the point you begin your scan of the PSA's and the TCB you want really is active at that moment, then there is more likelihood that you will "catch it in the act". But, once again, no guarantees. You can get (legally) disabled by obtaining the CPU lock prior to the PSA scan and releasing it afterward. CC ---------------------------------------------------------------------- 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
