LoneCrusader

No, we should not post any disassembly here. If we end up finding a solution it should be OK to post patches that can be applied to the original files as they were distributed however.

Excellent. While you're at it go ahead and grab the .PDB's for HDAUDIO.SYS and the other .EXE and DLL files in the package in case we need them later. Now let's hope rloew can use the debug symbols to make some sense out of the code in the files since we don't have any source code to work with. When he and I worked on the RLUSB driver for Windows 95 we had source code to work with. This is more like shooting in the dark.

ATI brand - X850 XT Platinum Edition is the last with working drivers NVidia brand - 6800 Ultra is the last with official drivers; 7950GT is last that works with unofficial drivers 7950 GT AGP cards are EXTREMELY rare however. Best bet is to find a 256MB 7800GS AGP 7xxx series NVidia cards will probably need the Shutdown patch applied to the driver. Cards with 512MB of RAM may need rloew's NVidia patch to make all RAM available.

If anyone can obtain and provide the debug symbols (HDAUDBUS.PDB) for KB888111 and/or KB835221 for Windows 2000 it might be helpful. I've tried searching but no luck so far...

WarCraft III on a Pentium III? I've seen lag with this game when running it on a Pentium 4 3.06GHz with 2GB of RAM and using an ATI Radeon 9800XT 256MB video card. Of course I was playing on a highly customized map with lots of enemies, but the point is that it can still lag even with these specs. I wouldn't attempt that particular game on anything less than a P4. Also possibly worth nothing that a while back one of the official Blizzard patches broke the game under 98 for me; I don't know if this was ever fixed or whether it affected others besides myself. Needless to say their support was less than helpful when I reported the issue. Anyway, this is getting off-topic. If you are looking for an AGP card for the earliest versions of the AGP spec (i.e. not AGP 2.0 or 3.0) then I doubt there are many (if any) around that will fall into the 512MB category to begin with.

This makes no sense. All of the features added to 98SE by NUSB etc. already exist under Windows XP by default. Why would you ever need to use a third-party app or a script to manage removable devices under XP when it's already built in... Also, there are plenty of XP compatible USB3 drivers. Only Intel refused to produce an XP-compatible driver AFAIK.

Best as in "adequate for normal computing" or best as in "performance?" If for normal computing, just about any motherboard based on the Intel 865 or 875 Chipset will do fine and support up to 3+GHz P4. If for performance, here are a few standouts. Intel D875PBZ Pic (Last 9x-compatible Intel-branded board. Be sure to get true Intel retail version not OEM "Wasp" version) Soyo P4I875P Dragon 2 Platinum Pic (Beware of BAD CAPS; these are excellent boards but will likely need CAPS replaced) DFI Lanparty 875P-T Pic (Very rare LGA775 board with 98SE compatible chipset) MSI 865PE Neo3-F Pic (More common LGA775 board)

The Etron USB3 drivers don't even work under Windows 2000, so I wouldn't have much hope they will work on 9x. WDM under XP sometimes behaves differently even for the same functions. I don't think any XP files have been found to work under 9x other than possibly USBCCGP.SYS and even it has not been extensively tested because it requires a specific set of devices that use it and those same devices may work without it, depending instead on USBHUB/USBHUB20.

I'm sorry you feel that way. It's too bad, since you do have some programming/debugging knowledge you might have been able to help considerably in sorting this out. I have no direct knowledge of such things or patching files; I can only do research, very basic hexediting, and test what is given to me. I feel like it's a worthy investment but not everyone will agree. These days, nowhere else will you find such an experiment as WDMEX. Who else is left with enough knowledge to even make the attempt? Petr, Tihiy, Xeno86, PassingBy, LLXX, and others I'm sure I'm forgetting to mention have all moved on. Fatal Exception 0E at 0028:C026BEE6 in VXD NTKERN(01) + 0000053A Interesting that the BSOD error moves from the HDAUDBUS driver back into NTKERN with the older version.

Those functions are missing from Windows 9x but are included in WDMEX. More below. Yes, it's there. The ME USB1 drivers depend on KS.SYS but work under 98 so the 98 version is sufficient for them. The ME version of KS.SYS may or may not be useful for HD Audio if we get that far. Here's what we know so far: HDAUDBUS.SYS is missing these functions (a long list ) under Windows 98SE without any WDMSTUB/WDMEX loaded: MmAllocatePagesForMdl MmFreePagesFromMdl MmAllocateContiguousMemorySpecifyCache MmFreeContiguousMemorySpecifyCache MmMapLockedPagesSpecifyCache ExFreePoolWithTag PsGetVersion MmGetSystemRoutineAddress KdDebuggerEnabled IoFreeWorkItem IoQueueWorkItem IoAllocateWorkItem ZwSetSecurityObject ObOpenObjectByPointer IoDeviceObjectType RtlGetDaclSecurityDescriptor RtlGetSaclSecurityDescriptor RtlGetGroupSecurityDescriptor RtlGetOwnerSecurityDescriptor RtlLengthSecurityDescriptor SeCaptureSecurityDescriptor SeExports RtlAddAccessAllowedAce RtlLengthSid RtlAbsoluteToSelfRelativeSD RtlSetDaclSecurityDescriptor RtlCreateSecurityDescriptor rloew's WDMEX covers these functions but we have no way to know if HDAUDBUS.SYS is expecting a different return value for any of these. I experimented further and found that a "HKR,,DevLoader,,*NTKERN" line must be present in the INF or the driver will do absolutely nothing. It will "load" without errors in the Device Manager this way even without WDMEX loaded, so it can't be doing anything and it's no wonder there is no enumeration. "Enumeration" is NOT the problem based on this; so we need to see what is the cause of the BSOD I mentioned before we can know anything else. I put the required line back in my INF and now I get this BSOD when Windows 98SE boots and attempts to load the driver: Fatal Exception 0E at 0028:FF097160 in VXD hdaudbus(01) + 000050C0 This can be dismissed but then the system hangs here before the desktop loads and must be reset.

Why would anyone even bother to go there? Apparently one can be branded a "troll" simply for disagreeing with the attitude or outlook of the prevailing majority. Good riddance.

Have you tried adding the HotFix for KB254660 or using IntelliPoint to update your mouse drivers? Might help, might not. I've not encountered this problem before so I don't know what else to try. I'll have to have a look at the INF and experiment with it. Thanks for the info! WDMEX isn't free or currently publicly available; you would need to contact rloew about it if you're interested. rloew may also be able to provide a replacement for WDMCHECK; it's interesting that WDMCHECK is crashing out when run on HDAUDBUS.SYS. I just meant that I had compared the INF files for the 9x-compatible WDM AC'97 drivers to the 2K INF files for the HDAUDIO drivers to see how driver loading was handled, because this would need to be modified for Windows 9x to use the 2K INF files. The AC'97 driver INF's do not seem to use an "HKR,,DevLoader,,*NTKERN" line while other driver INF's such as USBSTOR do have this. I don't know if it's relevant or not yet. Yes; when Microsoft added USB support to 95 OSR2 a new VMM32.VXD, KERNEL32.DLL, KRNL386.EXE, and several other core system VXD's are replaced. (See my XUSBSUPP package.) NTKERN.VXD is added which contains an "embryonic" implementation of WDM under Windows 95. Worth mentioning here that rloew and I were able to create a hybrid version of Windows 95 OSR2 + the Windows 98SE version of NTKERN using WDMEX and some other custom modifications. It allows one to use USB2 drivers, USB-HID devices, FireWire (1394 Bus and Storage) and possibly other 98-compatilbe drivers under 95 OSR2. The main problem we face is that even if one implements all WDM functions that are missing from a 2K driver under 98 using WDMSTUB/WDMEX there is still no guarantee of success. WDM functions are sometimes implemented differently and expect differing return values under different operating systems. One could spend a lot of time working on this with no idea whether or not it will all be wasted. Yes, very interesting. Have a link? Find anything else like this? If anyone can manage to track down a copy of the "IHV Enabling Kit" mentioned in these links, it might be the holy grail for this issue... ftp://ftp.micro-square.nl/Drivers/Audio/CMedia CMI9880/v5.12.01.0008/XP_2003_64bit/README Microsoft x64 prerelease HD AUdio drivers.txt http://microsoft.public.development.device.drivers.narkive.com/2cX5zfSm/uaa-hd-audio-function-driver-sample-code

What's wrong with the VMware driver? I know there is one that works under 95 because I have used it. This is possible under XP (and 2K I assume) but not under Windows 9x (that I am aware of ). Under Windows 9x one must dig through HKLM\Enum key of the registry or use a third-party tool do get specific hardware ID's.

Hello, Welcome to MSFN! Sounds like you already have a fair amount of experience in working with Windows 9x and newer systems. I've recently been experimenting with Windows 95 and 98 on a Gigabyte GA-X79-UD4 board so I know what you're dealing with. I know these aren't directly related to your question, but I wanted to ask. What's wrong with your USB mouse? It should work... And, what did you change in BIOSINFO.INF? It's common practice here to run SETUP with the "/p i" switches to disable ACPI, but another solution is always interesting. On the subject of HD Audio, I have also tried to get the HDAUDBUS.SYS driver loaded with the help of our resident expert rloew and his WDMEX.VXD (an expanded replacement for WDMSTUB.SYS). So far I managed to get the driver to load without errors in the Device Manager but it also doesn't function properly. It does not enumerate the \HDAUDIO "child devices" tree as it should. So far we don't know why, and I don't know what else to try. rloew may be able to elaborate further on the subject. As far as getting the driver to load with the INF, I removed all of the mentions of NT and all of the NT-specific sections from it and added the following two lines, based on my examination of the last version of the RealTek AC'97 WDM driver package INFs. HKR,,Driver,,hdaudbus.sys HKR,,NTMPDriver,,hdaudbus.sys If this line is present, I get a BSOD during boot. I don't know if it is necessary or not. HKR,,DevLoader,,*NTKERN Since you are getting an error in the Device Manager it sounds like your WDMSTUB.SYS is not covering all of the necessary WDM functions used by HDAUDBUS.SYS. EDIT: Struck out incorrect or superseded information. HKR,,DevLoader,,*NTKERN line IS REQUIRED. Without it driver does nothing; with it produces a crash. Enumeration is not the current problem based on these findings.

You probably just need a 9x compatible version of VMware tools, which provides a display driver. Google "VMware tools Windows 95" and you will find some links; supposedly this one should work.

I can verify that the mods on the linked page will get a MSI GeForce GTX 980 Gaming 4G up and running under XP with the 368.81 driver package. No further testing done besides installing the driver, but that's a big step in itself. Now if VBE9x or some other method will just get rid of the glitches it presents under 9x...

I remember on an earlier version of the forum software, there was an option to email yourself a copy of a PM conversation. Very useful for technical conversations where you may need the info without having to login to the forum or for helping clean out your PM inbox. Is this feature still around somewhere or has it disappeared?

In all honesty guys, any attempt to extend XP's kernel and API functionality will most likely require disabling/circumventing Windows File Protection and possibly other "security" functions. I would simply accept this as a given prerequisite and just move forward... It comes down to a choice between "do I want to leave things as they are for security" or "do I want to extend the possibilities of XP?" I think you're worrying too much about this point. Most users who would be willing to use a "KernelEx" type project are power users who are certainly familiar with the risks of running "non-secure" systems. In the end security comes down to the user, regardless of what operating system they're using.

Oh, the PaleMoon guys are being Microsoft's toadies again? Hating on older systems because they want an excuse to stop supporting them? Big surprise. I saw this coming even before the artificial POSReady block brouhaha earlier this year. EDIT: Wow, wth? My post is considered older than the one I quoted?

If I were you I would dig a little deeper into those reports and want to see some video and/or screenshot evidence that the ATI cards are really messing up EGA games. Also, just because one ATI card or series of cards may have had issues does not mean that all ATI cards have the same problem. I wouldn't rule them out just because someone had something bad to say about them, there are too many variables, not least of which can be user error.

While I hold the good old Voodoo cards in high respect there's no reason to have to go back that far in order to find a working PCI graphics card. I know ATI made PCI versions of several of their 7xxx, 8xxx, and 9xxx series cards. I would probably look for a PCI version of the Radeon 8500, as I have used this particular card in several Win95 machines and I know the disc that came with the cards has working 95 drivers... I don't see any of those on eBay currently but there are some 7xxx and 9xxx series: http://www.ebay.com/itm/152225129722 http://www.ebay.com/itm/182243441276 http://www.ebay.com/itm/381761732402

Dibya, while we all wish you success with your projects it might be better if you started a new thread (or used one of your older ones) in the XP forum for them so that everything can be kept together and organized. Right now you have discussion of your mods going on in three or more different threads and it will soon become difficult to keep track of what information is where. If you need to consult with other developers or project maintainers in their threads you can always include a link back to your main thread that contains the relevant information.

I'm not familiar with that board, but from a quick Google search it appears that the 3.0 revision of the board can support Wolfdale (a BIOS version 3.2 and a beta BIOS version 3.22 were mentioned as well so it may need this even on the 3.0 board?). Looks like it has a AMI BIOS though.. I wouldn't choose this board for a 9x setup myself but to each his own. If I wanted Socket 775 + AGP I would look for a DFI LanParty 875P-T or 865PE-T (these are rare!), but unless you need AGP there's no reason not to use a PCI-E board and an NVidia 7xxx card.

As far as I know you can't have "too large" a power supply, as the motherboard and devices will only draw as much as they need. The problems come when the devices draw more than the PSU can put out. I would trash the Celeron (known to gamers as CelerOWNED ) processor and replace it with the highest-clocked Pentium III that is compatible with that board. You will see a very significant increase in performance. Virtually any Pentium 4 (Socket 478/some Socket 775) motherboard is great for running Windows 9x as most of them have compatible drivers and some can use over 3GHz processors and 4GB of RAM (you'll need rloew's patch for that much RAM though). Boards newer than that can be used in some cases but require more effort and more non-free patches. See the sticky thread we have devoted to motherboards for more info (I know it's a lot to read and somewhat disorganized, but we hope it will be improved in the future). You've come to the right place for Windows 9x support. You won't find any haters here that will waste your time telling you it's old and to run something else.

This card apparently uses the "memory banking system" that rloew and I discovered while experimenting with 512MB and larger nVidia cards. With rloew's NVSIZE patch you can make all 256MB of the cards memory usable under 98SE.