Hercules 1.62 announcement

Jay Maynard

Jun 3, 2000

What's new in release 1.62
Release date: 3 June 2000

Still more multiprocessor improvements (by Jan Jaeger)
Dynamic CPU reconfiguration (by Jan Jaeger)
Basic vector facility (by Jan Jaeger)
Floating point version 6 (by Peter Kuschnerus)
READ AND RESET BUFFERED LOG CCW (X'A4') support (by Jay Maynard)
WRITE SPECIAL CKD CCW (X'01') support (by Jay Maynard)
FBA DASD model reporting fixes (by Jay Maynard)

I've added RPMs for Intel x86 and Alpha processors, and links to it all from
the main page. The distribution tarball is still available, as before. I've
moved the HTML documentation to its own directory, for ease of maintenance.

The vector facility is a very basic form, and quite incomplete, but the
framework is there. The WRITE SPECIAL CKD CCW is the first half of track
overflow support; it writes the flag that the read routines will need to
check to step to the next track and keep reading.

Links to the downloads are on the main page, at
http://jmaynard.home.texas.net/hercules .

On a personal note, the URL for the Hercules page will change in about a
month and a half, hopefully. I'm moving, and getting DSL in the bargain -
but have to change ISPs, since texas.net's DSL pricing and policies are
among the worst I've seen. I'll give as much notice as I can before that
happens.

2:53 pm

1.63 announcement

Jay Maynard

Jun 19, 2000

While I was moving, the rest of the world was coding! There's a lot of stuff
in this one. Without further ado:

What's new in release 1.63
Release date: 18 June 2000

3270 CCW processing improvements (by Jan Jaeger)
OSTAILOR generalization, and new pgmtrace panel command (by Jan Jaeger)
VM IUCV instruction correction and DIAGNOSE improvements (by Jan Jaeger)
CPU timer and clock comparator improvements (by Jan Jaeger, after a
suggestion by Willem Koynenberg)
3480 READ BLOCK ID and LOCATE CCW support (by Brandon Hill)
Networking support via virtual CTCA (by Willem Koynenberg)
Restructured CPU execution, by function call instead of switch statement
(by Mike Noel)
Support for IEBCOPY sequential output datasets in dasdload (by Ronen Tzur)
New dasdls command lists the VTOC of a CKD DASD volume (by Malcolm Beattie)
New AWSTAPE handling commands: tapesplit, tapemap (by Jay Maynard)
make install target to install in /usr/bin (by Jay Maynard)

In addition, the Hercules page has moved again. Sorry about the rapid
succession of moves, but changing ISPs necessitated it. The good news is
that the new page is on a machine under my control, not hosted by someone
else, so I can do whatever I want! Muahahaha!

The new URL is http://www.conmicro.cx/hercules .

There's still more work to be done; the track overflow and timer resolution
improvements I have been thinking about and working on sporadically didn't
make it into this release, but should be in 1.64. (I hope!) I'm still moving
stuff to the new apartment, and wil be for a while yet, but all of the
important stuff (TV, computers, microwave...) are here now, so I can move
the rest at my leisure.

Thanks for all of the improvements...without you, we wouldn't be anywhere
near as far along as we are now.

5:17 am

Hercules 1.64 announcement

Jay Maynard

Jul 5, 2000

What's new in release 1.64
Release date: 4 July 2000
* Added track overflow processing for CKD DASD (by Jay Maynard)
* Makefile change to allow RPM building with RPM_BUILD_ROOT (by
David Barth)
* Added NetBSD build definitions to makefile (by Soren Jorvang)
* Moved version definition to version.h and removed makefile
dependency for source modules (by Jay Maynard)
* Package change: tarball now explodes into hercules-<version>
subdirectory (by Jay Maynard, suggested by Soren Jorvang)
* Fix backward going TOD clock (by Jan Jaeger)
* Suppress superflous HHC701/HHC702 messages (by Jan Jaeger)
* Rework cpu.c to decode instructions by macro (by Jan Jaeger)
* Bypass bug in IBM telnet client (by Jan Jaeger)

This version should IPL MVS 3.8 successfully, unless something else is
discovered that gets in the way...

As always, get your copy from http://www.conmicro.cx/hercules , in tarball
or RPM form...

3:17 am

Hercules 1.65 announcement

Jay Maynard

Jul 22, 2000

What's new in release 1.65
Release date: 22 July 2000
* Track overflow processing fixes (by Jay Maynard, suggested by
Valery Pogonchenko)
* Added TOD clock update to STCK, STCKE, DIAG 204, and TRACE
processing (by Jay Maynard)
* Fixed READ DEVICE CHARACTERISTICS alternate track values for 3380
and 3390 (by petermac@...)
* Skeletal CMPSC instruction (by Bernard van der Helm)
* Added support for 3340 and 3375 DASD (by Jay Maynard, with help
from Rick Fochtman and David Cole)
* Corrected interval timer update increment (by Jay Maynard)
* float.c optimization for new instruction decode and execution (by
Peter Kuschnerus)
* Fix program check on TIC ccw (by Jan Jaeger)
* Fix program check on NOP ccw (by Jan Jaeger)
* Instruction decode & execution restructure (by Jan Jaeger)
* Added -fomit-frame-pointer to compiles for improved performance
(by Jan Jaeger)
* Fix STCKE instruction (by Bernard van der Helm)

8:36 pm

Hercules 1.66 announcement

Jay Maynard

Aug 3, 2000

What's new in release 1.66
Release date: 3 August 2000
* Simplify logmsg and DEVTRACE macro definitions (Jay Maynard)
* Prevent incorrect length indication on CONTROL NOP CCW (Jay
Maynard)
* Complete 370 HIO processing (Jay Maynard)
* Correct nullification of TPI and TSCH (Jan Jaeger)
* Add device locking to MSCH (Jan Jaeger)
* Correct TPROT instruction (Jan Jaeger)
* Correct address wrapping on assist instructions (Jan Jaeger)
* Change interrupt logic to use longjmp on all interrupts (Jan
Jaeger)
* Clear remainder of ASTE when loading ASTE with ASF=0 in
translate_asn (Jan Jaeger)
* Add (incomplete) PLO instruction (Jan Jaeger)
* Fix CLCL interruption problem (Jan Jaeger)
* Fix addresswrap in MVO (Jan Jaeger)
* Make ED and EDMK perform a trial run (Jan Jaeger)
* Fix address wraparound in MVO (Jan Jaeger)
* Fix CR15 corruption in form_stack_entry, fix nullification in
form_stack_entry and unstack_registers (Jan Jaeger)
* Fix loss of interrupts in PR (Jan Jaeger)

One other note, not included in the above: I changed the optimization in the
makefile from -O3 to -O2; -O3 is broken on the Alpha, and there was no
measurable difference in execution speed on the x86 between the two. I plan
to leave it that way unless there's a really convincing argument otherwise.

11:28 pm