Jump to content

matthiaspaul

Member
  • Posts

    6
  • Joined

  • Last visited

  • Donations

    0.00 USD 
  • Country

    Germany

Everything posted by matthiaspaul

  1. The answer is yes and no. Yes, it is possible to run Windows 9x ontop of DR-DOS. And yes, it has several advantages over using MS-DOS (f.e. smaller memory footprint and much more advanced relocation methods, not only resulting in more free DOS memory, but also in more free Windows resources; better configurability and therefore higher flexibility, more advanced utilities). No, as Microsoft has artificially tied MS-DOS 7.xx and Windows 4.xx together and it is not trivial to pry these components apart. As one of the main developers of DR-DOS, I am the original author of a small system level device driver named "WinGlue" I wrote back in 1997/1998, when I worked with Caldera UK. The name "WinGlue" is both, a technical description of what it is, a thin layer of glue code, and a sarcastic play on words on the German word "Scheibenkleister" due to "dirty" coding methods I had to utilize in order to get it working. ;-) The interesting part was, that "WinGlue" basically just faked a number of undocumented interfaces and data structures, Windows expected to see during startup. With a resident memory footprint of just a few hundred bytes, "WinGlue" was able to convince Windows 95 to start up on either Novell DOS 7/Caldera OpenDOS 7.01 or Caldera DR-DOS 7.02. Of course, since "WinGlue" was only faking data structures, it was not free of trouble, but the very fact, that it worked at all, reproduceable and on many machines, had quite some impact on the law suit. Andrew Schulman, who at this time worked for Caldera as an external Windows consultant, came over to England to inspect my findings, and together we could research some of the remaining issues. Somewhat later Geoff Chappell was also assisting us in some questions. Demonstrating that it was possible to achieve the goal, it was decided to fork the kernel and directly add full MS-DOS 7.0 (and later 7.1) support into the DR-DOS kernel. The DOS 7 compatible fork was nicknamed DR-DOS "WinBolt" and has been worked on by a small group of developers, two original Digital Research veterans, Andy and Ian, and myself. The other fork, on which two of us participated as well heavily, became DR-DOS 7.03, which was still emulating PC DOS 6.1, but had optional dynamically loadable components such as DRFAT32 (a FAT32 redirector extension loadable on any DOS 3.31 or higher) and LONGNAME (LFN support for any plain DOS 7 and DR-DOS 7.xx - the shell COMMAND, which also works under other DOSes, would automatically enable LFN support as soon as the LFN API is present), both originally with DPMS (DOS Protected Mode Services), which was disabled in the published versions for some odd reasons. While DR-DOS 7.03 was released and sold in large numbers, DR-DOS "WinBolt" never left beta stage and remained unreleased. In hindsight, there have been lots of management failures and engineering had been hardly given the time to work on this. The most stupid decision was, when Caldera US found DR-DOS 7.03 good enough to finance their Linux business for several years to come, out of a sudden decided to close the UK office, where all the DR-DOS development took place, and lay off the OS experts, after they had done great work and brought out a successful product. Well, some of us continued to work on DR-DOS in customization projects for various DR-DOS OEMs, including variants with DOS 7 API level support, but basically the expertise was cast in the winds. Well, I continue to work on DR-DOS whenever I find the time for it (not much recently). Actually, there is a reason why I am writing this on a dual-core machine in a Windows 98 SE window... ;-) In the meanwhile, another fork has made great progress as Enhanced DR-DOS - this, however, is based on Caldera OpenDOS 7.01, which in turn was virtual identical to Novell DOS 7 update 10 or so, thus several man years of original DR-DOS development are missing from this source tree. Even if it supports some of the DOS 7 APIs, it still shows that it is internally based on an older kernel. Well, I'm not overly optimistic any more, but I still have not completely lost hope, that at some fine day in the future the current owner of the DR-DOS assets may decide that its commercial life is finally over and that it is due to open source the system. This would make it possible to reunificate the different code branches to create the most advanced DR-DOS ever for all its fans. Well, well... ;-) But back to the original question, yes, technically speaking it is possible to do this, but practically it is not for the time being, unfortunately... Greetings, Matthias
  2. Why don't you phone them up or send them an e-mail and ask? Or ask Commell, if they have a Czech distributor. ;-) Yes, the P4BWA is even more impressive, but I assume, it will be almost impossible to get working DOS or Windows 98 SE drivers for the 965 chipset. Yes, but ca. 6 months ago, when the P4LA Revision D (2.1) with Core 2 Duo support was released, the web site still didn't mention it for quite a time. Older revisions of the board did not support the Core 2 Duo. Greetings, Matthias
  3. I don't know a Commell distributor in the Czech Republic, but the German Commell distributor Spectra has offices in Austria and Switzerland as well, and they sell the Commell P4LA revision D mainboard. http://www.spectra.de/kontakt.cfm Spectra also lists alot of other industrial boards (in all kinds of form factors) with ISA slots, some of them also supporting Core 2 Duo processors. However, the listed specs are not always up to date, so check with the manufacturers. Actually, I first learnt about Commell's plans to release a P4LA revision with C2D support, when I contacted Commell earlier this year in order to discuss our company's special hardware requirements with them and they told me what they were just about to release. Even their own website didn't reflect C2D support at this time. ;-) http://www.commell.com.tw/Product/SBC/P4LA.HTM Hope it helps, Matthias
  4. Well, even the 8088, 8086, 80186, V20, V30 all were 16 bit processors similar to the 80286. (The 8080, a widespread processor at Digital Research CP/M times was 8 bit, though.) The 8088, used in the very first IBM PC had an 8 bit bus interface, but it was still a 16 bit processor from the software side. The 80286 was the first CPU to offer an Protected Mode and an address room of more than 1 MB (up to 16 MB IIRC). The 80386 was the first 32 bit processor in this CPU line. It provided a 32 bit extension of the Protected Mode and a new sub-mode of the Protected mode, the x86 Virtual Mode (actually the most powerful addition). Its virtual address room was 4 GB. The 80486 was the first processor to incorporate a numeric coprocessor (previously, coprocessor such as the 8087, 80187, 80287 were optional add-on chips and costing a mere fortune in the 500 - 1000 USD range ;-). Windows 3.0 in "386 Enhanced Mode" was the first Windows version taking advantage of the 32-bit Protected Mode, whilst the operating system remained a 16-Bit/32-Bit Real Mode/Protected Mode hybrid up to including Windows ME. (Several other operating systems used 16 bit and 32 bit Protected Mode earlier, including Digital Research Concurrent DOS 386 as well as IBM OS/2.) The major reasons for software developers to create 16-bit/32-bit hybrids were code efficiency (size and speed) and compatibility with existing applications or at least providing "easy" upgrade paths on source level. On machines with an average RAM population in the 1 to 8 MB range and 16 - 50 MHz clock speed, a 32 bit-only operating system would have been rather inefficient for mainstream applications. Greetings, Matthias
  5. I'm (mostly) successfully using an ATI All-in-Wonder X800XL PCIe DVB-T/PAL 256MB (article code 100-714502) graphics card with Radeon R430 GPU @ 400MHz in a Commell P4LA Rev. D BIOS 2.1 industrial mainboard with Intel 945G/ICH7R chipset and Intel Core2D E6600 SL9S8 2.4GHz CPU under Windows 98 SE. Partially citing from my yesterday's post: http://www.msfn.org/board/Modern_motherboa...st&p=695463 Since no Windows 98 SE drivers exist for the Intel GMA 950 shared memory onboard graphics (or at least I could not find any drivers supporting more than standard VGA 640x480x16 resolution), I installed this ATI graphics card into the board's PCIe 16x slot. This ATI graphics card was announced in 2005-07 and released ca. 2005-10. Apparently, it was withdrawn from the market in spring or summer 2006 and is quite difficult to source today - used or new. No AGP, and no 128 MB or 512 MB variants of this card exists. However, there is an US-variant with NTSC (instead of PAL), but without DVB-T support. I think, this video card is one of the latest ATI cards, for which Windows 98 SE drivers are still available: Catalyst 6.2 as of 2006-02-09. http://ati.amd.com/support/drivers/de/98me...onaiw-98me.html (the date on the ATI web-site is wrong) This driver suite officially supports only the Radeon 9800, 9700, 9600, 9500, 9200, 9100, 9000, 8500, 7500, 7200, 7000 and Xpress 200 series (incl. All-in-Wonder variants of them). However, it also provides beta support for the Radeon X850, X800, X600, X550 and X300 series (but not for the All-in-Wonder-variants of them). Trying to install these drivers, the automatic installation will detect an "incompatible card", but if you install the drivers manually, it is still possible to install the drivers. - Graphics Cards: -- Radeon X800 Series driver: 2006-02-03, ATI2DRAG.DRV/ATI2VXAG.VXD, ATI Tech. - Enhanced, 4.15.1.9165 -- Radeon X800 Series driver (secondary): 2006-02-03, ATI2DRAG.DRV+ATI2VXAG.VXD, ATI Tech. - Enhanced, 4.15.1.9165 - Audio, Video and Game Controllers: -- ATI WDM Rage Theater Video, 2004-04-14, ATIRVXX.SYS, ATI Technologies Inc, 6.14.10.6240 -- ATI WDM Teletext Decoder, 2004-04-14, ATIVTTXX.SYS, ATI Technologies Inc, 6.14.10.6240 -- ATI WDM TV Audio Crossbar, 2005-04-14, ATIVXSXX.SYS, ATI Technologies Inc, 6.14.10.6240 -- ! ATI WDM TV Tuner, 2004-04-14, ATIVTUXX.SYS, ATI Technologies Inc, 6.14.10.6240 - Universal Serial Bus Controllers: -- ATI Remote Wonder Controller, 2003-12-29, ATIRWVD.SYS, Jungo, 6.03 (from ATI Remote Wonder 3.04 driver package) -- ! ATI RemoteWonder Plus, 2005-03-02, X10UFX2.SYS, X10 Wireless Technology Inc, 3.0.0.180 (from ATI Remote Wonder 3.04 package) I can use the video card in all resolutions, with dual monitor support and as 3D graphics card. Hydravision works. DirectX 9.0c is installed. So far, I was not able to get the TV and radio tuner running under Windows 98 SE. And the driver for the Remote Wonder Plus remote control causes a non-fatal protection fault at startup. Any ideas on the reasons? Anyway, while it would be nice to get them running as well, I do count this as bonus, as I can use these features under the other OSes on the same machine - I just needed a relatively decent graphics card still working under Windows 98 SE and not limiting me too much under the other OSes. Greetings, Matthias PS. The mainboard is one of the very few modern boards still featuring an ISA slot. I plan to also insert an old "HGC" Hercules monochrom graphics card in order to have multi monitor support under plain DOS - for software development and debugging. So far, I have not tested, if this particular ATI card will peacefully coexist with a HGC card - well, it should, as this is part of the IBM PC specs, but some recent cards won't free the video segment B000-B7FF.
  6. Hi all. One of the reasons for running Windows 98 SE on new hardware is the best-possible support for DOS as well as driver support for older or special hardware, especially in the industrial / embedded world. However, often, old special purpose hardware does also require an ISA slot, which has become a rare gem on modern mainboards. Personally, I run DR-DOS 7.08 and Windows 98 SE (besides a host of other operating systems) on a Commell P4LA Rev D BIOS 2.1 industrial mainboard, equipped with Core2D E6600 SL9S8 2.4GHz and 4x1 GB RAM. http://www.commell.com.tw/Product/SBC/P4LA.HTM - Intel 945G PCI-Express chipset & ICH7R chipset (RAID + Matrix storage support) - CPU Intel Core2D, Pentium D, Pentium 4, Celeron D (no EE editions) with 800/1066 MHz FSB, LGA 775 socket - lots of BIOS config settings - 4x 240-pin DDR2 533/667 RAM slot for up to 4 GB of RAM (of course, Windows 98 SE can only use a fraction of it, but still) - 1x 16-bit ISA slot - 4x 32-bit PCI slots - 1x PCIe slot 4x - 1x PCIe slot 1x - 1x PCIe slot 16x - Onboard VGA (Intel GMA 950) with 224MB shared memory (no driver support for Windows 98 SE, so only standard VGA resolution under Windows) - PATA: 2x EIDE/ATAPI UDMA100 devices or a 40-pin DOM (disk-on-memory) - SATA: 4x SATA-II (modern OSes support all 6 channels simultanously, whilst only up to 4 of them can be used under Windows 98 SE. I'm quite confident that - with a little bit of tweaking - all six channels could be made work in DOS and thus in Windows 98 SE as well, although most probably not at full speed.) - Marvell E8053 Yukon Gigabit LAN (DOS and Windows 98 SE drivers exist, but I could not get 7.14.1.3 to run so far, so I use an Intel PRO 1000 GT PCI Gigabit LAN addon card. For me, the Yukon driver either hangs or crashes during Windows startup.) - Audio Realtek ALC880 HDaudio with S/P-DIF (no DOS and Windows 98 SE drivers exist to the best of my knowledge.) - 1x 32-pin DIP JEDEC DiskOnChip socket - 1x CF-Card slot Type II - 8x USB 2.0 - 5x RS232, one of them configurable as IrDA - 1x RS232/422/485 port - 1x LPT incl. ECP/EPP - 2x FDD (up to 2,88 ED disks, but without 3mode support) - 1x PS/2 keyboard - 1x PS/2 mouse - 8x user programmable digital I/O ports - configurable hardware watchdog Since no Windows 98 SE drivers exist for the GMA 950 onboard graphics (or at least I could not find any), I installed an ATI All-in-Wonder X800XL PCIe DVB-T/PAL 256MB (100-714502) graphics card into the PCIe 16x slot. I think, this video card is one of the latest, for which Windows 98 SE drivers are still available: Catalyst 6.2 as of 2006-02-09. This driver suite officially supports only the Radeon 9800, 9700, 9600, 9500, 9200, 9100, 9000, 8500, 7500, 7200, 7000 and Xpress 200 series (incl. All-in-Wonder variants of them). However, it also provides beta support for the Radeon X850, X800, X600, X550 and X300 series (but not for the All-in-Wonder-variants of them). Trying to install these drivers, the automatic installation will detect an "incompatible card", but if you install the drivers manually, it is still possible to use the video card in all resolutions, with dual monitor support and as 3D graphics card. So far, I was not able to get the TV Tuner running, and the driver for the Remote Wonder Plus remote control causes a non-fatal protection fault at startup. http://ati.amd.com/support/drivers/de/98me...onaiw-98me.html In regard to 945G chipset driver support. Intel does not officially provide Windows 98 SE drivers for them, but manually installing assorted generation 8xx drivers from various packages, I was able to find working drivers for most components. Only very few yellow "!" can be found in device manager: - Other components: -- ! PCI card -- ! PCI RAID controller (not essential for operation, see below) - Audio, Video, Game controllers: -- ! ATI WDM TV Tuner (I still have some hope to get it working) - System components: -- ! PCI standard ISA bridge (secondary) [the primary PCI standard ISA bridge works] (not essential for operation). EDIT: With the updated Intel drivers from http://windows98.ic.cz it now shows up as: "Intel® 82801GB/GR (ICH7 family) LPC Interface Controller - 27B8". The yellow "!" did not vanish, however... - Universal serial bus controllers: -- ! ATI RemoteWonder Plus (I still have some hope to get it working) Since RAID works on BIOS+hardware level, having no RAID drivers for Windows 98 SE is not much of a problem, unless someone would need maximum harddisk performance. In my case, my focus is more on reliability. Therefore I use RAID 1 for redundancy, and only a small section of the drive matrix as RAID 0 (for cache files). Recommended harddisks for RAID are Western Digital *YS RE2 SATA harddisks, where the 15s read timeout can be disabled using the TLER utility. In case there would be a RAID fault under Windows 98 SE, one would require another operating system such as Windows 2000 or XP in order to rebuild the RAID. However, basis setup does not require any drivers or tools at all and can be done from within the BIOS. I hope this information may help some people to find a mostly DOS and Windows 98 SE compatible modern, flexible, and highly reliable industrial grade mainboard with support for PCIe, PCI and ISA at the same time. Greetings, Matthias
×
×
  • Create New...