I built an AROS machine out of scrap!

Got to say it’s really cool that it works with hardware 3D acceleration.

I may want to try to do something with it later on, however I’ll need to get it a case.

I have a cold, and yeah sound like crap.  I’ll add specs later as I think my fever is kicking back in.  Sigh.  But yeah basically

  • P4 board / CPU / RAM $150 HKD
  • GTX-460 card $180 HKD
  • Sound Blaster Live! $20 HKD
  • 37GB IDE disk $20 HKD
  • PS/2 Keyboard $20 HKD
  • PS/2 Mouse $10 HKD
  • 700 Watt Power Supply $150 HKD

So yeah ~550 HKD or $70 USD.  Not bad.

So a little closer look at the hardware.  I’m lucky that there is an active used hardware market here in Hong Kong, elsewhere in the world you either have HAM radio events, ‘boot sales’, garage sales, or for the truly desperate, eBay, Yahoo auctions, and AliExpress.  My go to place here is of course the Capital Computer Centre, where they at least will test stuff before selling it.  I know I’m old fashioned but I like buying in person.


Intel P4 Mother Board

I had originally chosen this board to mess around with Darwin.  I wanted something new enough to have a P4, but old enough to still have an older ‘parallel’ EIDE controller port. And the Intel D945GNT, board certainly was up to that task.  Like ancient Darwin, AROS works best with either parallel disks, or SATA disks in older parallel emulation.  The markings on this board are a little hard to read as the bigger numbers are the product testing/radio compliance numbers, and the model number is a bunch of possible models as I guess they like to make so many variations on a single board.

close up of the D945GNT

Here is a close up, and E210882 is *NOT* the mode number.  Nothing like confusion.

The Intel D945GNT motherboards built in NIC, the Intel PRO/100

One of the big reasons for using the Intel board, is that it has an onboard NIC, and Intel of course uses Intel NIC’s so it has the very compatible Intel PRO/100 VE Desktop Adapter.

While you can get these on PCI cards, and use other boards, I figured since I was going to buy a board anyways, and once things get this old the people selling them really don’t care who made the board, but rather that this is an old P4 board, they all sell for the same price.

Another plus about this board is that it is new enough not to have AGP, but rather the new and exciting PCI Express.  This board as the Express x16, which of course is perfect for a ‘large’ GPU.  AROS has a port of the Gallium3D nouveau driver, making this perfect for a super cheap GTX 460.

I shopped around for a while and I found this accelerator, that although has no apparent labeling at least on the flip side it has the identification.  I really don’t know what is the fastest GPU you can get for AROS, but this one seems to work just fine, and it’s what Stephen Jones is using so that is what I went with.

Inno3D GTX-460

And looking under AROS, this is how the PCI resources show up for the video card.

NVIDIA GeForce GTX 460

This was an old card, and it looked like either the OEM didn’t put any stickers on it, or someone had taken them off.  Either way I don’t care, and it doesn’t matter as it works just fine.  Sure the GTX 1080 is over five times faster, but the open Gallium driver won’t work with it as Nvidia has done their best to break open stuff, and even if it did, you can’t buy a 1080 for less than a pizza.  At least not yet.

For some reason, I had begun collecting older and cheap Sound Blaster cards when I see them.  I wasn’t going to spend more than $50 HKD ($7 USD) for them, so I don’t have an Audigy cards yet, but I did have this Live card.  At the time I didn’t think it was anything special, although the EMU10k chip is desirable, and popular for much older systems.

Sound Blaster Live!

This card is the CT0100 model.  And it works great!


And this is how AROS sees the card

Sound Blaster Live! with EMU10k1 chip

The AROS HCL is a little confusing to me, but it all seems to work.  If it weren’t for the Stephen Jones video I wouldn’t have tried as it implies it won’t work.

I have to admit that it’s impressive.  For anyone interested you can find AROS Icarus Desktop here, and here is Stephen Jones’s video that inspired me to see this run in real life!

Started to re-build my MS-DOS machine


Good news, it actually works!  I was using the version 1.1 WAD, so honestly weird crashes really aren’t unexpected.  I haven’t looked much at what to do with audio, but I was really impressed compared to the Qauake II wars, it was really surprising to not only see DooM run on the first shot in real metal, but the keyboard works as well.  Well enough for me to pick a level, and get killed.

Naturally it doesn’t work under Windows, however it runs fine with MS-DOS mode.

Porting Catacomb3D to MS-DOS (DJGPP v1/GO32).

Catacomb 3-D for GO32

No really, it’s Catacomb 3-D: The Descent.  First ported to 32-bit SDL by NotStiller.  Me being the person I am, I fixed a slight bug regarding binary files on Windows, and MS-DOS, then cleaned up some of the C++ syntax (yuck!) making it far more C89 friendly.  And of course, hot off the heels of DooM for GO32 DPMI, I was able to get it to build and run using GCC 1.39 and GO32.

I know most people really won’t care, but I found it kind of interesting.  I should try to see if it’ll run on actual hardware, just as a comparison of tightly optimized Borland C++ / Assembly vs 100% pure C on DJGPP.  The best tech of 1991 for sure!

At current I just put the source up, you can git it here.

Null DooM, GCC 1.39, GO32 and DPMI


DooM via DJGPP v1 GO32

Around the time of the x68000 port of DooM, I was cutting down the DooM source for a null/portable version.  I never could get it to actually run either using EMX or  DJGPP 1.03, as I couldn’t get it to link to save my life with a constant never ending battle of unresolved symbols. After a while I just used what I had towards the x68000 version and concentrated on getting it up and running, and just shelved the null/portable effort.

Later on I wanted to get it running again as part of messing with another cross compiler, as DooM isn’t a trivial application to port and verify correct operation. And in the process of trying to get the null version to build and run on Windows using TDM GCC, I wanted to make sure it at least kept compiling with GCC v1.x.

Once more again I was able to compile individual files but unable to link.  But this time, I just looked at the diffs for binutils, I thought it should be somewhat easy to get hosted on Windows.  Although versions may point to binutils 1.0, I had to use binutils-1.9.tar.gz even though the diffs are against Mar 24 1991, and the source for 1.9 is dated April 17 1991.

My first effort gave me a linker that would happily link, but go32 would either refuse to run the executable, or just crash.  I was going to give up again, but I found mention in another file that DJGPP actually uses the linker from G++, the C++ compiler which was a separate thing in the late ’80s and early’90’s.  This time it worked, and I could link a trivial hello world style application!

Now that I finally had a cross linker actually working, I didn’t want to compile under emulation, so looking at the other diffs, they didn’t look too extensive. I went ahead ,and took DJGPP v1.06 and patched up the compiler & assembler to get a full cross toolchain.  And in no time, I had a null version of DooM running on MS-DOS well at least tested on DOSBox.

This was fun, and all but I didn’t see any easy way to do fun things like hook interrupts so I could get the keyboard & clock like any good MS-DOS program.  DPMI greatly eased this kind of stuff, so looking at the DJGPP history, DJGPP v1 version 1.10 actually adds preliminary DPMI support!  And in the next version, DPMI was much more better supported, however the binary format had changed from a.out to COFF as part of the move to v1.11. I was able to take the memory, and DPMI portions from the final v1.12 libc, and manually build and run them against the v1.06 library / dev tools.

And much to my surprise, it actually worked!  At least having the wrong format didn’t have any effect on how GO32 worked for me.

So feeling lazy, I snagged some of the support code from Maraakate’s revamp of DooM, just to make sure of the timer code, and the keyboard code, and again verified that I can build with the keyboard & timer ISR and I’m able to play the v1.9 shareware & commercial levels fine.  I haven’t done a thing to clean up or update the DooM source itself against all the dozens of bugs and issues with Ultimate DooM, or other games like Chex Quest etc.

I’m sure 99% of people wouldn’t care but you can download it here:

Download crossdjgppv1

Although I’m using DPMI to drive realtime events, if I looked further at the GO32 v1.06 environments I could either figure out how it operates it’s timer, or modify the extender directly to drive the PIC timer and keyboard as I need.  But overlooking that, the vintage 1991 software is more than capable of running DooM.

Installing Windows NT 3.1 on a physical computer

I have this P4 I got for super cheap in Hong Kong, that came with Windows 98 of all things. Naturally I want to load something more useful like Windows NT 3.1 onto it, so I did have to do some tweaking first.


The first thing is the hard disk. I was lucky in that this machine came with a 40GB disk, the Hitachi Desktar IC35L060AVV207-0.

HGST Deskstar 180GXP IC35L060AVV207-0

HGST Deskstar 180GXP IC35L060AVV207-0

Now what makes this disk great, is that it can be jumpered down to act like an 8GB disk, so that things like MS-DOS and older OS’s like Windows NT can recognize the disk.  Even nicer that the jumper settings are on the disk!

Disk jumpers

Disk jumpers

My board supports booting from IDE, which is nice as I could paritition and format the disk from MS-DOS 5.00, and make sure things were working fine.

However it really doesn’t matter as over on Beta Archive, The has made an ATAPI driver available for not only Windows NT 3.1, but also various beta versions as well!  You can find the post, and the links to download here! (mirror here).  Now you can install from the boot diskette & a driver diskette and load the rest of the OS from CD-ROM.


You will have patch the INITIAL.IN_ and SETUP.IN_ files to allow installation on any new processor.


STF_PROCESSOR = “” ? $(!LIBHANDLE) GetProcessor


STF_PROCESSOR = $(ProcessorID_I586)

You can leave the files expanded, but this is needed if your CPU is newer than a Pentium (Yes a Pentium 60/66 type processor, so that is Pentium Pro, Pentium II, Pentium III, Pentium 4 and beyond…).  But yes, this is great!  No need to try to dig up old SCSI cards, SCSI disks, and SCSI CD-ROM drives.


And much like Qemu and VMware, the AMD PCnet is a great go to PCI card, and I was able to find this IBM 11H8130 Type 8-Z 10BaseT PCI Network Card which works!

IBM 11H8130

IBM 11H8130

The card works great with 11265315.exe set of drivers, OR disk image pcnet.7z .  But for sure the key is the in the chipset!


AMD AM79C970AKC PCnet(tm)-PCI II

As this chipset, the AMD AM79C970AKC is the one that is explicitly listed as compatible.  This IBM card provides an AUI port, along with a 10baseT port.

Post install, service packs

Of course when installing Windows NT 3.1, you’ll want service pack 3, the last update to the OS.

Also don’t forget to replace NTLDR & NTDETECT.COM from a later version of Windows NT to allow access to more than 64MB of RAM.


Windows NT 3.1 will allow you to install on FAT, HPFS, and NTFS v1 partitions and disks.  The TCP/IP is a 3rd party, from Spider that does not support DHCP.  Outside of doing it just because, it really is better to go with NT 3.5 or 3.51 as they have better SMP support, are much faster, and have a much more robust network stack.


BSDI BSD/386 1.1

So this crossed my desk, from an anonymous source:


For those who like this kind of thing, here is a dmesg:

BSDI BSD/386 1.1 Kernel #0: Wed Mar 3 16:23:55 GMT 1999
cpu = Pentium (unknown speed) model 6, stepping 3
delay multiplier 8663
real mem = 68153344
avail mem = 65589248
buffer cache = 6774784
isa0 (root)
pccons0 at isa0 iobase 0x60 irq 1: color, 8 screens
com0 at isa0 iobase 0x3f8 irq 4: buffered
lp0 at isa0 iobase 0x378 irq 7
pe0 at isa0
xir0 at isa0 on lp0 (at 0x378)
fdc0 at isa0 iobase 0x3f0 irq 6 drq 2
fd0 at fdc0 slave 0: 1.44M HD 3.5
wdc0 at isa0 iobase 0x1f0 irq 14
wd0 at wdc0 slave 0
wdc1 at isa0 iobase 0x170 irq 15
npx0 at isa0 iobase 0xf0
vga0 at isa0 iobase 0x3c0 maddr 0xa0000-0xaffff
ne0 at isa0 iobase 0x300 irq 9: NE-2000, address 52:54:00:12:34:56
changing root device to wd0a
wd0: format error in bad-sector file

Yes it’s real!  For those who don’t remember history, after the Net/2 release there was a company called Berkeley Software Design Inc (BSDi) that provided a commercial port of Net/2 that also included source.  Add in the infamous 1-800-ITS-UNIX ad, and as they say the rest is history.

BSD/OS 1.1

BSD/OS 1.1

During this time frame it does get hard to track down as the name was in constant flux. BSDI, BSDi, BSD/OS, Internet Server…  Mix in the fun with 386BSD and you get all around naming confusion.

This version, 1.1 is from 1994.  The version timetable does get a tad bit confusing so here we go from what I can find:

1992, April – BSD/386 (BSDi) 0.3.1, first version
1992, June – BSD/386 (BSDi) 0.3.2
1993, March – BSD/386 (BSDi) 1.0
1994, Feb. – BSD/386 (BSDi) 1.1
1995, Jan. – BSD/OS (BSDi) 2.0
1995, June – BSD/OS (BSDi) 2.0.1
1996, Jan. – BSD/OS (BSDi) 2.1
1997, Feb. – BSD/OS (BSDi) 3.0
1998, March – BSD/OS (BSDi) 3.1
1998, Aug. – BSD/OS (BSDi) 4.0
1999, March – BSD/OS (BSDi) 4.0.1
1999, Dec. – BSD/OS (BSDi) 4.1
2000, Nov. – BSD/OS (BSDi) 4.2
2002, March – BSD/OS (Wind River) 4.3
2003, May – BSD/OS (Wind River) 5.0
2003, Oct. – BSD/OS (Wind River) 5.1

One can only hope that 0.3.1 from the apparent “300 customers” may eventually surface.

Fun source of the lawsuit meltdown C/O Computerworld 1992:

C/o Computerworld

C/o Computerworld

C/o Computerworld

C/o Computerworld

For anyone who want’s to relive the glory days, there is a qcow2 disk image suitable for Qemu floating around..

Word is you’d want to run it like this:

qemu-system-i386.exe -L pc-bios -net nic,model=ne2k_isa -net user -hda “bsdos-1.1(repack).qcow2” -redir tcp:4423::23

Porting Quake II to MS-DOS pt4

Bringing it all home for release day.

Bringing it all home for release day.

Since the last update we got some help in a few fields that have really fleshed out this ‘experimental’ port into a full fledged port.  First RayerR helped us with the fun of getting us onto the latest deployment of DJGPP, 2.05 (rc1).  It’s always nice to be in the latest available release.  Next in a passing comment, Ruslan Starodubov had mentioned that he had gotten a much older build of our QDOS to support the Intel HDA sound chipset via the  MPXPlay sound library.  I wrote to the author of MPXPlay, Pádár Attila asking for us to distribute his source in our project, and he granted permission.

So at this point things were looking good.  The only ‘feature’ that modern OS’s really held over us was the ability to dynamically load and unload game modules on the fly.  I had tried to use DLM, but it stripped the DPMI functionality out of the MS-DOS Extender making the port really useless.  I tried to build the newer DXE3 support but had no luck.  I suspect now my native tool chain was interfering with the build process. But Maraakate managed to get it to not only build, but to run!

Adding in DX3 support was relatively painless.  I first looked at DJGPP’s FAQ and downloaded the example code.  In the example code there was small helper functions to make unions and check the symbols.  If they didn’t exist a printf was spit out to alert you about it.  To resolve the issue you simply just add DXE_EXPORT to the other list of missing exports.

Compiling the game code was easy, again referring to the example I saw that basically they compiled it the same, but at link time you use DX3GEN and -U flag to ignore unresolved symbols.

The biggest head scratcher was the Sys_GetGameAPI failing to find GetGameAPI from the DX3.  After some piddling around I noticed that it listed GetGameAPI as _GetGameAPI inside the DX3 itself.  I added the underscore and it worked!

Other things that were relatively to easy to import was R1Q2’s HTTP downloading code.  Compiling CURL was kind of tricky because of the linking order, but thankfully neozeed figured it out quickly.

All of Yamagi’s Quake 2 updated game DLLs were all diff’d by hand using BeyondCompare to make sure I didn’t clash using some newer functions that weren’t available in DJGPP.  I also merged their Zaero code with their baseq2 code by comparing Zaeros code to the Quake 2 SDK, marking every thing that was changed.  The result is a really stable Zaero game code.  If you haven’t played Zaero check it out.  I think it’s a lot better than Rogue, but Xatrix is probably my favourite (even over stock Q2).

Other cool things I’m glad to get into the code was the GameSpy Browser.  It took me quite a bit of work to get it where it is, but it’s really nice to just be able to ping to a master server (a custom GameSpy emulator I wrote specifically for Q2DOS.  Source is not finalized yet, but will be available soon for those curious), pick a server and go!  All in DOS!

So here we are at the end of the journey.  Or at least safe enough for a 1.0 release.

To recap, we have:

* SVGA (LFB modes only)
* Mouse
* Keyboard
* SoundBlaster and Gravis UltraSound Family
* CD-ROM music
* OGG music
* Networking (You need a packet driver)
* Loading/unloading game DLLs in DX3 format.
* Intel HDA support -hda
* Mouse wheel support with -mwheel

And I should add, that it works GREAT on my MSI Z87 motherboard.

You can download Quake II for MS-DOS on bitbucket.  And as always the source is available here.

Don’t forget you can always make bootable USB stick with DOS, or CD-ROMs.

Continued in Part 5!

Windows 3.0 Debug Release 1.14

Well from popular request I finally got around to loading this up.  I went ahead with my favourite retro emulator, PCem for this, as it can nicely emulate an EGA display, unlike most emulators which do VGA, however when it comes to older versions of Microsoft products they really can detect the difference between EGA and VGA.

So to start off, I downloaded from the project page, this version of PCem, compiled it, and installed MS-DOS 4.01 , from April of 1989.  The Windows 3.0 Debug Release 1.14 itself is dated from February 22nd, 1989.  Which I figured is close enough to the time period.  I’m using the 486SX2/50 because I’m too impatient for the 386 speeds, but it does work fine on 386 or higher emulators.  It does NOT work with any 286 emulation. I’m also using the HIMEM.SYS from MS-DOS 4.01 vs the one with the Windows 3.0 (Alpha? Beta? Technical Preview?) since it is slightly newer.

There is no setup program per say, rather it just xcopies all the files to a directory, and from there you run ‘d.bat’ and away you go.  This version is hard coded to an EGA display, which again is the reason I went with PCem.  Once you start it up, you are greeted with:


Windows v3.0 Debug Release 1.14

And it identifies itself as Windows Version 2.1


Look at all the memory!

And first thing to notice is that on my setup with 8MB of ram, I have over 6MB of RAM free.  Compare this to regular Windows 2.1 which gives me 399Kb of ram in my current setup.

Windows 2.1 running in real mode

Windows 2.1 running in real mode

And with Windows/386 Version 2.1 it provides 383Kb of real memory, along with 6.7MB of EMS memory, as the Windows/386 Hypervisor includes EMS emulation.

Windows/386 memory

Windows/386 memory

Of course the major limitation of Windows 2 is that it runs in real mode, or in the case of Windows/386 an 8086VM.  As I mentioned a while back in a post about Windows 3.0,  This was game changing.

As now with Windows running in protected mode, all the memory in my PC is available to Windows, and I am using MS-DOS, with nothing special.

Besides the limitation of being EGA only, the Debug version of 3.0 is that there is no support for MS-DOS applications, as WINOLDAP.MOD is missing.

NO MS-DOS for you!

NO MS-DOS for you!

This is clearly an interim build of Windows 3.0 as mentioned in Murray Sargent’s MSDN blog Saving Windows from the OS/2 Bulldozer.  As mentioned from the article they began their work in the summer of 1988, so considering this is early 1989 it shows just how much progress they had made in getting Windows 2 to run in protected mode.  Along with Larry Osterman’s MSDN blog post Farewell to one of the great ones, which details how the Windows 3.0 skunkworks project was writing the new improved 386 hypervisor, and how Windows 3.0 got the green light, and changed the direction of not only Microsoft but the entire software industry.

I’ve been able to run most of the Windows 2.1 applets, however I’ve not been able to run Excel 2, or Word 1.  I suspect at this point that  only small memory model stuff from Windows 1 or 2 is capable of running.  Although at the same time, when 3.0 did ship, you really needed updated versions of Word 2 and Excel 3 to operate correctly.

Windows 3.0 Debug Release 1.14

Windows 3.0 Debug Release 1.14 on a 12MB system

The applets from Windows 2.1 seem to work a LOT better than the one from Windows/386 2.1 if that helps any.

This is an interesting peek at an exceptionally early build of Microsoft Windows.


PCem v9

PCem v9

From the main page:

PCem v9 released. Changes from v8.1 :

  • New machines – IBM PCjr
  • New graphics cards – Diamond Stealth 3D 2000 (S3 ViRGE/325), S3 ViRGE/DX
  • New sound cards – Innovation SSI-2001 (using ReSID-FP)
  • CPU fixes – Windows NT now works, OS/2 2.0+ works better
  • Fixed issue with port 3DA when in blanking, DOS 6.2/V now works
  • Re-written PIT emulation
  • IRQs 8-15 now handled correctly, Civilization no longer hangs
  • Fixed vertical axis on Amstrad mouse
  • Serial fixes – fixes mouse issues on Win 3.x and OS/2
  • New Windows keyboard code – should work better with international keyboards
  • Changes to keyboard emulation – should fix stuck keys
  • Some CD-ROM fixes
  • Joystick emulation
  • Preliminary Linux port

Thanks to HalfMinute, SA1988 and Battler for contributions towards this release.

Very excellent!

MS-DOS Player updates

Poorly translated from TAKEDA toshiya’s blog..

I has integrated source of i386 and i286 edition edition. 
In addition, in the i286 version, I added support for int 10h/16h. equivalent to 0.149 MAME, I was replaced with a 0.152 equivalent MAME core i386 i286 core. However, the i386 core, I have omit the TLB around.

Which is very cool, although I wasn’t sure about the MAME source code being open to other projects…?  I tried to contact the i86/i386 author vlinde but he then pulled his contact page.  I wanted to use i386 for something of my own, but the whole “Redistributions may not be sold, nor may they be used in a commercial product or activity.” really puts the damper on it.

I was able to get some simple XMS test program to run, but nothing of any substance.  No DOS4G/W or anything like that.  But if you re-build it specifying MS-DOS version 5.0, some of the MS-DOS utils and even command.com work!

The weird issue I had was running out of conventional RAM, because this program gives you nearly 1MB of conventional RAM… I was surprised, as I wasn’t expecting that much!