> > The formatting is excellent - the filtering is non-existent. > > As I alluded earlier, the prime (extra) functionality I see as required > "in > > the field" is the ability to identify storage "leaks". An option to > > eliminate all "matched" get-free pairs would go a *LONG* way to achieving > > this goal. > > I believe IBM researched has dabbled with this, but it is a non-trivial task > to do by IBM or by others. Some of the complications that immediately come > to mind are being able to free 8 byte multiple of storage at a time > independent of the size of the original request, being able to free multiply > obtained adjacent areas with a single free request, free by subpool, and > region reset processing isn't traced.
Undoubtedly all true, however I see this as part of the "post processing" legwork that always has to be done. When some-one gets to the point of running a GFS trace, it's because (hopefully) they know they have a problem, and where. Hence the trace is likely to be against a specific ASID and probably known key(s) and subpool(s). What I would like implemented is a means of removing from the output all "matched" get/free records. Maybe something like; address, length, key and subpool on a freemain matches those same fields on a previous getmain. In my case, I also included TCB address in the criteria, but that may not always be a valid choice. With just this data removed (from the analysis, not the input) the data left to be manually assessed was *significantly* less - to the point of almost being the "smoking gun" without any more work. Even if this is not always the case, I can see something like this being *mighty* useful - ploughing through the formatted records of thousands of redundant entries that may be widely separated ain't any fun at all. Shane ... ---------------------------------------------------------------------- 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
