- Written by Scott G Taylor
February 13, 2018 UPDATE
I've discovered (realized, really) an NFS issue. However this is not
specific to Mach386.
Because I am mounting mammoth-sized NFS filesystems currently, while I
look for and recover additional Mach386 data from my old media, I began
encountering segfaults and other strange behaviour in some programs.
I just realized that this is due to the filesystem itself. Many
programs simply aren't able to work with the filesystem, particularly
when creating temporary files... Like ex/vi does.
- Written by Scott G Taylor
February 12, 2018 UPDATE
I've fixed the SCSI Tape End-Of-File read problems.
The st.c driver has some very bizarre code interpreting the SCSI status blocks.
With that said, I found two issues:
1. When a short-record was read, it was not setting bp->b_resid
to the correct shorter length in certain cirumstances; and
2. When a tapemark was read, bp->b_resid was not being set to
indicate no bytes read.
I've tested the my changes extensively and I'm satisfied that SCSI Tape
functionality is working as it should now.
Current device probe configurations (in order):
|HD Drive 0||0x104||14|
|HD Drive 1||0x1f0||14|
|wd||SMC 80x3 Ethernet||0x280||3||0xd0000|
|3c501||3Com #C501 Ethernet||0x300||9|
|com||Serial Port 1||0x3f8||4|
|Serial Port 2||0x2f8||3|
|Serial Port 3||0x3e8||5|
|Serial Port 4||0x2e8||10|
- Written by Scott G Taylor
The following blog entry refers to files at ftp://ftp.mrynet.com/os/Mach2.5/MtXinu-binary-dist.
While not (yet) made accessible, this README entry describes the contents being assembled and subsequent maintenance of the MACH386 product.
February 1, 2018
These files contain the floppy images and scanned documentation for the
Mt XINU Mach386 binary distribution for "MACH and BERKELEY UNIX for i386".
This is the full 3.5" floppy distribution, and includes the Mach 3.0 CMU
product as well as the Mach 3.0 DUI (DOS server). The build/release
date of this product is March 23, 1992.
The usefulness of this software is questionable, for a number of reasons:
- Mt XINU only provided media for the base system. All
further updates, bug fixes and additional functionality,
were provided electronically only. There are many bugs
and issues in the base system, and no source is available.
- Mach386 only supports a very limited list of hardware,
primarily ISA bus-based hardware and early IBM-PC disk
- There are no source repositories of Mach 2.6 software, upon which
Mach386 is built. Neither the source nor the binaries
have been found. Mt XINU stopped selling software in the
mid 1990's, and no known archives exist. There was at least
one later release of Mach386, dated August 6 1992, that
would have had later software fixes and updates.
The following are included in this binary product distribution, as
received from Mt XINU in 1992:
This is a color advertising brochure describing the Mach386
This addendum describes known problems with this distribution.
This yet-to-be-scanned document provides the installation
and operation procedures of the distribution.
This document provides Mach 3.0 and the Mach 3.0 DUI
package installation and operation instructions.
This is a copy of a paper written by Mary R. Thomson, of
Carnegie Mellon University, describes set-up and booting of
the Mach 3.0 kernel on a Mach 2.5 base-installed system.
Mach386 Product Description for a later August 6, 1992, release.
The following 1.44MB 3.4" floppy images have been imaged. They comprise
the full Mt XINU Mach386 distribution dated March 23, 1992:
This is the installation bootstrap floppy image.
This is the installation bootstrap filesystem.
These floppies contain the base system distribution.
These floppies contain the Networking userland binaries,
libraries, development (header) files and on-line documentation.
These floppies contain the Mach386 on-line documentation
(man pages, etc).
These floppies provide X11 Window System binary and support
and documentation distribution.
These are the add-on Mach 3.0 and DUI distribution files.
The floppies are direct images of each of the distribution floppies.
They were extracted using the UNIX dd(1) command as follows:
# dd if=/dev/fd0 bs=512x18 of=imagename.DSK
The distribution floppies utilize standard PC-compatible 1.44MB
3.5" recording format that the DOS and Windows 'FORMAT' command
produces. They can be written directly to formatted floppies
with the same dd(1) command method:
# dd if=MtXinu-Image.dsk bs=512x18 of=/dev/fd0
or other equivalent means on other platforms.
All of these floppy images have been verified and test-installed
on a basic 500mhz i386 system using an IDE/ATA disk (BIOS-configured
for 1023/16/63 Cylinder/Head/Sector limits). An ongoing attempt
to obtain an ISA Western Digital/SMC WD80x3 ISA ethernet adapter
continues to be pursued to determine if networking support can be
This particular distribution has broken SCSI support, and is not
fixable since the electronic updates from Mt Xinu are no longer
available to update the distribution.
I personally used this Mach386 (with updates from Mt Xinu) from
1991 through 1998. Despite its development-quality short-comings,
it proved to be robust enough to manage for me as a professional
platform and as a personal platform back in the days of network
communications through dial-up mail-based USENET and early CERFnet
If there are any questions or if you have access to additional
Mt XINU Mach386 or MACH 2.5/2.6 source or other Mt XINU information,
please contact me.
Scott G. Taylor
bub7734 at gmail dot com
- Written by Joomla
February 3, 2018 UPDATE
I have found my last root dump and usr tar archives from 1996 which includes the
latest updates I was able to sup(8) from Mt Xinu in 1993. I will be assembling
additional distribution additions with the updated binaries and files. Included
in these updates from Mt Xinu are working SCSI support, additional userland and
GNU utilities and the Mt Xinu MSD 2.6 rebuildable/reconfigurable kernel.
I have also located the January 21, 1991 distribution floppies. However, there
are a number of unreadable floppy sectors in various locations of this set which
renders the entire distribution useless. Since the distribution sets are
effectively sequential chunks of compress(1)'d data, the loss of a block of
data in any of the floppy sets renders that set incomplete and a working
installation from it is impossible.
Page 4 of 5