> The SRB is sampling the address space so it tolerates statically error and > uncertainty. What I'd like to try and nail down is whether the TCB is dispatchable but > no CPU is available *or* is actually executing. Any ideas?
You -REALLY- just can't tell. WLM is the only thing that knows about delay reasons in any detail and for the most part, it too gets that information by sampling. If you keep a table of TCBs and you keep both the TCB address and the TTOKEN value as uniquifiers and you sample both the primary non-dispatch bit and TCBTTIME you can tell whether or not it was dispatchable when you looked and whether it is accumulating any cpu time. If you sample frequently enough then over time, a reasonably valid picture will emerge. The definition of "frequently enough" is a difficult subject on its own. Don't be misled by looking at TCBs in a dump. They are all in an unnatural state as a result of dump processing. It is fiendishly complicated these days to figure out what is actually going on simply by looking at the dispatchability indicators in the TCB. The field descriptions in the data areas manuals offer only vague clues at best. The only people who really know what's going on all work in Dutchess County NY. Also, note that TCBTTIME does not change while a task is executing on a cpu. The time is only incremented on undispatch events, e.g. by WAIT. If you see a change from one sample to the next then you know the TCB has been dispatched at least once since the last time you looked. > I was wondering, could the SRB examine the TCB queue hung of ASXBFTCB to > determine if the TCB is dispatchable? If you use TCBTCB as the "next" pointer, ASXBFTCB will get you ALL of the TCBs in the address space, regardless of dispatchability. Beware that the TCB chain can change while you're looking at it unless you hold the local lock. 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
