home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: alt.folklore.computers
- Path: sparky!uunet!spool.mu.edu!agate!linus!linus.mitre.org!jcmorris
- From: jcmorris@mwunix.mitre.org (Joe Morris)
- Subject: Re: ASSEMBLY LANGUAGE IBM
- Message-ID: <jcmorris.722525732@mwunix>
- Sender: news@linus.mitre.org (News Service)
- Nntp-Posting-Host: mwunix.mitre.org
- Organization: The MITRE Corporation
- References: <1992Nov18.134855.28580@geovision.gvc.com> <1992Nov19.230823.9763@spatula.rent.com> <30546@nntp_server.ems.cdc.com> <1992Nov22.032132.14404@spatula.rent.com>
- Date: Mon, 23 Nov 1992 13:35:32 GMT
- Lines: 44
-
- azog@spatula.rent.com (Billy D'Augustine) writes:
-
- >Well, for the longest time, the actual reference to the card reader
- >(2901? forget),
- ^^^^
- Probably a 2501. In general any machine numbered x9xx is a special-order
- (so-called "RPQ") product, and one numbered x5xx is a card handler. Of
- course, there are more than enough exceptions to this rule.
-
- > was removed the system startup (a DEFINE card), but
- >the physcial machine sat around for a few months. When we got rid of it,
- >we rewired channel 0 (where the cardreader and printers usually live,
- >due to being such slow I/O devices). Now we have 020 as a 3422 mag reel
- >drive, and 00E as a 4245 printer. Now, think about this addressing just
- >a little bit. What do you think will happen when you fire up a nice
- >tape sort on 020? The printer stops...
-
- Working as designed. A channel is (by definition) a single logical data
- path. A tape drive, once it begins operation, takes control of the
- channel until its last data transfer activity is complete. This is
- not unreasonable since few tape drives have any significant buffering
- capability; once the tape is running and requires channel service it
- cannot accept a delay caused by some other device hogging the channel.
-
- In general a system should not be configured with a burst mode device
- (like a tape drive) on a byte multiplexor channel...and on most systems
- channel zero is always a byte multiplexor.
-
- On some older systems you could install a special feature which
- converted the top 64 addresses on channel zero (0C0-0FF) into
- four so-called "selector subchannels" which operated independently
- of the lower channel zero addresses. With the decreased cost of
- channel hardware the need for this gimmick went away some time back.
-
- > All the password files (for CICS) are text files sitting
- >somewhere on the system. To add a new password, you have to assemble this
- >text file into a phase (IBMs name for an executable). The originating
- >text file is in no way encrypted or protected. On that same vein, a
- >user simply cannot change his password by himself.
-
- Sounds like the WIJIT system I knew and loved (NOT!) on a PDP-11/45
- a few years ago...
-
- Joe Morris
-