On Sat, Nov 17, 2012 at 5:55 PM, Lindy Mayfield <lindy.mayfi...@sas.com>wrote:
> May I infer, then, that you are talking about "serious" issues where the > program better get it right the first time, especially when updating these > control blocks (1) that are being discussed at the moment. And simple > interfaces to storage, such as Rexx need have different requirements, thus > because of simple functions to read storage be exempt from this discussion? > I mean, that's all there is that programs like Rexx have, and they need > live with it, yet at the same time understand discussions such as these. > Lindy - Please consider the topic in a more general fashion. Any applications (problem or supervisor state) that has 2 or more threads of executions (TCBs) operating at the same time and share data structures (memory) need to agree when how to synchronise updates and read access to the share memory area. Otherwise, both threads may attempt to update the same memory location at the same time with different values. One of the values will be lost. This will negatively impact the way the application runs. Sam > > Kind regards > Lindy > > (1) Who can update these control blocks? I think from reading this only > z/OS can. (or should, you guys do what you want, seems like, then justify > it like I just did.) > > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On > Behalf Of John Gilmore > Sent: Saturday, November 17, 2012 8:41 PM > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: "New" way to do UCB lookups > > Lindy, > > I have received several private, offlist requests for elucidations of the > same sort you make. > > I suggest that you read the descriptions of the ENQ and DEQ macros in > > z/OS MVS: Assembler Services Reference, Volume 1 (ABEND-HSPSERV), > SA22-7606. > > Some of the computer science literature talks instead about serialization > and semaphores (Dijkstra's term, a metaphoric hijacking of the European > word for what Americans usually call 'traffic lights'). > > Also worth looking at are the discussions of the Test and Set (TS) > instruction and its more powerful brethren, the most recent of which is > Perform Locked Operation (PLO), in z/Architecture Principles of Operation, > SA22-7832. > > These notions are important ones. What needs to be emphasized about them > is that they all make use of a mixture of conventions and hardware help. > If you and I and others agree to adhere to a set of such conventions, then > we can safely share access to serially reusable resources. If not, not. > > John Gilmore, Ashland, MA 01721 - USA > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, send email > to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
