The image attached to:
https://www.mobileread.com/forums/sho...5&postcount=36
shows our very basic setup, the cpu with its hardware console, one 3270
Master operator's console and a bookshelf of reference manuals.
That nice "firm wood" bookcase is being emulated by documents posted on the Internet.
Referencing the image attached to:
https://www.mobileread.com/forums/sho...4&postcount=12
shows the system that we would like to install on the K5 devices.
After purchasing the optional, extra cost, swivel chair, the sysop would roll up to the cpu hardware console;
Open the access door in the pedestal, insert the 8 inch floppy disk, and start the micro-code upload to the main cpu.
With the current (day 1) version of this project, our K5/370 version of that is shown in the images attached to:
https://www.mobileread.com/forums/sho...73&postcount=2
https://www.mobileread.com/forums/sho...5&postcount=22
which show the "micro-code load and start-up" at the cpu's hardware console.
The pictured IBM/370 system used (a lot) of special cabling between devices.
The K5/370 system uses TCP/IP networking in replacement for the physical cabling.
In the IBM/370 system, the
Master operator's console would have been physically cabled before anything was powered on.
But in the K5/370 system, we have to 'attach by networking' the
Master operator's console to the cpu hardware.
The image attached to:
https://www.mobileread.com/forums/sho...0&postcount=34
Shows our
Master operator's console powered up and attached to the cpu hardware.
The Hercules emulator has a configuration file which is the details of all the equipment that is part of our K5/370.
One of the end goals of system K5/370 is to use the Wifi connection to the Kindle;
BUT . . . .
In that sub-system of the Kindle, the Kindle's IP address is dynamically assigned.
So for day 2, the config file will be set-up to use the USBnetworking sub-system of the Kindle, with it's static IP address.
If this was running on a desktop system, with a SCSI controller, Hercules could use the actual, physical, external SCSI devices.
But Amazon/lab126 did not include a SCSI controller in the Kindles.
So in the system K5/370, the physical storage devices are represented as 'image files' located in the user accessible USB storage area.
In the IBM/370 system, the sysop would have run around the room, physically installing the required disk packs in the removable pack disk drives.
In system K5/370, we
**JUST** have to drop the image files into an appropriate place in the USB storage area and then in the Hercules configure file define which files are what hardware devices.
The first (system bootstrap) was typically device 150 (or there abouts).
To IPL (Initial Program Load) the system, all that would be needed would be to issue (from the hardware console) a "reset and start" command sequence to device 140.
For a disk drive, this would start reading the platter at track 0, head 0, sector 0. Whatever was there, was what you where about to run, like it or not.
These systems could also be IPL'd from the magnetic tape drive, paper tape drive, punched card reader, grandma's flower pot, etc.
Anything with a device address.
But the 'Initial' program we want to load is VM/370 (the 'virtualbox' system of the mainframe world).
And that resides on (in) disk image files.
Notes:
VM/370 will also run on a K5/390 (or IBM/390) but the Kindles are a single core ARM processor.
There is no reason (at the moment) to be using it to emulate anything more advanced than a K5/370.
The public domain compilers are all IBM/360 (and VM/370 will run system 360 code) so the first shipped system will be K5/370.
Which starts Day-2 of the project.
Tweaking the config file (a whole lot).
And a link to be remembered:
http://mvs380.sourceforge.net/