LoneCrusader

You seem to have a lot of very strange problems running these older systems...

There is no guarantee your system will work with more than 512MB of RAM even with the SYSTEM.INI tweaks. They never worked for me on any system I tested them on. I did see 98SE run with slightly over 1.2GB one time when using an old version of the Unofficial Service Pack, but the system was highly unstable. If I were you I would forget anything over 512MB until you get everything set up and working smoothly. Then add it and see what changes. You may end up needing rloew's RAM patch if it doesn't work or makes your system unstable.

I was able to duplicate Mercury's results with the same VID&PID RALINK device and using rloew's latest WDMEX. Successfully connected to a WPA2 network under 9x using the Odyssey Client as well.

Most of the issues with 98 on a 915 or later chipset can't be fixed or helped by just having INF files. As I said the INF files only give proper names to the "generic" devices already installed by the OS during SETUP. Sometimes they will eliminate unknown devices such as "PCI System Management Bus" but all the INF does is give it a name - it doesn't load any files or set any configuration data for the device. Note that no .SYS, .VXD, .PDR, .MPD, or other actual "driver" files are contained in any of the Intel chipset packages. The only device in these INF's that actually causes a file to be loaded is the "LPC Controller" which is a legacy bridge providing compatibility with older BUS'es such as ISA. The LPC Controller entries cause ISAPNP.VXD to be loaded, but this file is already present and loaded to begin with. The real "compatibility" issue with 9xx and later chipsets under Windows 9x lies in how the motherboard BIOS is set up; what options it has and how it manages the system. For example I have an Intel-branded Intel-BIOS 965 chipset board that has some issues when trying to run 9x. However a Gigabyte-branded AWARD-BIOS 965 chipset board works fine. Now as we both know laptop BIOS'es are always "proprietary" as intended by their manufacturers and usually severely limit the options available to the end-user, this presents a problem when one wants to use the laptop in any fashion different from the one intended by the manufacturer. These proprietary BIOS'es tend to cause various conflicts, especially when running Windows 98 with ACPI enabled. Also, a big hurdle to compatibility with laptops using the 915 and later chipsets is the various other integrated devices. The ATI Mobility Radeon graphics chips are obviously problematic in some configurations (I've seen reports of the Dell version of this chip working with 9x but the HP version in my ZD8000 does not. It won't even work with VBE9X.) and I don't know if any available NVidia-based Mobile cards work any better, and in any case they are usually not user-replaceable. I believe I have actually found 9x drivers (not tested them all yet) for every single device in my ZD8000 except the video. SciTech Display Doctor does work on it, but I can't set the screen's native resolution of 1440x900 without it trying to go into some kind of "panning" mode, and when SDD is installed it breaks Plug & Play which means USB Flash Drives don't work when connected. Anyone who knows how to fix this problem with SDD please chime in... I have attached the 915 Chipset/ICH6 files extracted from the package I linked previously. This is the extent of Intel's support of the 915 Chipset under 98SE/ME. My unofficial INF's can be used to identify any other 9xx devices on the 915 laptops should they not be covered by the official files. ICH6INF.ZIP

You are the one who came in here acting as if you knew more about the subject than anyone else. If you have supposedly tried to get the package working then it is assumed that you have actually run it or know how to extract it properly. Apparently not, based on your statements. This is simply false, yet you were quite adamant about it. The support may be limited but it does exist. Another obviously false statement. If there were no INF files in the package it wouldn't work for any of the 8xx series chipsets either! This. Again, acting as if you know more about the subject than those who are trying to help you. I apologize if you were offended, but implying that I don't know what I am talking about when I try to give you information or implying that I am purposely misleading you is not very polite either.

Are you sure you know what you're doing? Of course there are INF files in the package! You can't just play around looking at the file with 7-Zip or whatever, you have to run it to extract them. Examine the "Command Line Flag Options" in the README if you need a command to extract it without installing it. No. Those Catalyst drivers support standard X600 devices, not Mobility X600 devices. There is a difference. They have been reported to work with Mobility devices in a couple of cases but these reports are isolated and not regularly reproducible.

The current Firefox 50.1.0 x86 setup will run properly under XP x64 SP2, however the 50.1.0 x64 version errors out and claims Windows 7 is required. I know that officially Mozilla only supports the x64 version on Windows 7 and up, but I wonder if this is simply an artificial limitation, especially in the case of Vista? I have seen reports of "later-than-officially-supported" 64-bit versions of Firefox running on XP x64 but not for 50.1.0 specifically yet. Waterfox 50.1.0 is still supporting XP x64 as well... Anyone know how to patch the x64 installer to stop this or what to compare between it and the x86 version to find out?

I've never tried to actually run the installer on a 915 chipset system so I'm not familiar with the behavior there. I know it works on the 875 chipset despite it not being specifically named 9x compatible (although the 875 is not much different from the 865, which is named-supported). You could try running the SETUP or extracting it to a folder and then manually use the 98SE Device Manager to update the driver for the "System Management BUS Controller"; pointing the driver dialog to ICH6CORE.INF and see what happens. 915.INF and 915M.INF will contain some of the devices as well. The security catalogs are pretty useless in all honesty. Probably all they would do in this case is verify the checksum of the INF file against the original one created by Intel. These are also unnecessary to make an OS run on a given set of hardware. Having a "working" chipset driver does not guarantee that the OS will run any smoother. In my experience Windows 98SE starts to have problems with ICH6 (9xx) and later chipsets when trying to run using ACPI. Usually it is necessary to run SETUP using the "/p i" switches to disable ACPI and thus prevent various instabilities. It is possible the USB problems you saw were caused by ACPI as I have seen this on newer systems. I have an HP ZD8000 laptop which uses the 915 chipset as well. I have had issues getting 9x to run properly on it but the main problem is the ATI Mobility Radeon X600 graphics chip rather than the motherboard chipset. In any case my unofficial Windows 98SE chipset drivers for 9xx and later Intel systems are here.

+1 I discussed this issue previously with rloew and he says there is nothing of interest in HAL.DLL from SP1. The problem is the issue of the potentially buggy drivers as exhibited by the USB issue, NOT whatever was changed in HAL.DLL. Reverting to the SP1 HAL.DLL will not prevent the potentially buggy driver problem so therefore it is no better a solution than patching the SP3 version or a later HotFix version if available.

Just what "instructions" does the 915 INF not have that, say, the 865 ones do? Intel's chipset drivers are essentially a set of glorified text files that only give proper names to their respective onboard devices. An OS can run just fine without these in any case. I don't remember anything being different about the 915 INF when I made my unofficial INF drivers for 9xx and later chipsets. You are correct however that the download linked by ~♥Aiko♥Chan♥~ does not support Windows 9x at all. This is a mistake made by Intel, either inadvertently or purposely. The last 9x compatible version is 6.3.0.1007 here. It does contain a 915 INF but 925 is not covered. Not sure about 910.

While you can make such an ISO yourself, it's still illegal to distribute it and it's against the forum rules here to ask others to do so. Despite the age of Windows 9x, it's still under copyright. You should be able to find enough information scattered around here to create such an ISO for yourself. Look for topics regarding MSBATCH.INF or "slipstreaming" in this section of the forum. I believe there is an old thread regarding "w98_slip" that should provide good information and I have also discussed slipstreaming at some points since. Topic locked. EDIT: Corrected name of old slipstreaming package

Are you using a plain vanilla installation of 98SE or are you using the Unofficial Service Pack or any other updates? Assuming a plain vanilla install, try installing this HotFix and see if it solves the problem. If not, try removing the device and reinstalling it with the Dell driver after the HotFix has been applied.

Have you tried these? Dell Link (I see the card has a Dell part number so this may be the best option. ME drivers should work under 98SE. If you can't run the .EXE on your computer try unpacking it with WinRAR, 7-Zip, etc.) DriverGuide Link 1 (Solve the CAPTCHA to prove you're not a robot and is should link you directly to the file. Good to see DriverGuide has stopped using forced .EXE installers.) DriverGuide Link 2 (Same)

Yes it does seem strange that it works as expected. ACPI problems usually begin at ICH6, and I have an Intel-branded ICH7 motherboard here that is completely useless with Windows 98SE unless SETUP /P I is used to disable ACPI. Without this it won't even boot. Much of this seems to be controlled by the BIOS so consider yourself lucky in that respect since proprietary (in your case Dell) BIOS'es and even moreso laptop BIOS'es in general are usually pretty worthless. As far as desktop BIOS'es go, Intel seems to be the worst, followed by AMI for 9x compatibility. AWARD seems to be the best.

Use Dependency Walker that you mentioned in another thread to examine VDMCONFIG.DLL and possibly other files in the package you are trying to install. If any of them report missing exports then it should help you track down what file is missing. If nothing else it should tell you what other DLL's VDMCONFIG.DLL is linked to so that you can be sure you have them. It's also possible you must register some other DLL before VDMCONFIG.DLL?

A quick search yielded very few results as expected. Any issues I did see reported with the newer HAL however all were related to the USB problem that is solved by using the 2K3 version of USBPORT.SYS. The only way to know will be to set up such a system and use it on a daily basis for some time. Which system drivers specifically? The USBPORT.SYS one is the only specific problem that I am aware of being reported. Previously you suggested replacing all of the .SYS files in XP with the ones from Server 2003. This is most likely overkill and could create more problems than it fixes... Such replacements should only be done on a case-by-case basis where an issue is discovered with the XP version and then the 2K3 version has been checked and verified to be free of missing NTOSKRNL functions first.

I've tested it successfully on a machine with 32GB of RAM using patched NTKRNLPA.EXE and HALMACPI.DLL from SP3 (5.1.2600.5512). Of course "successfully" here means that it booted to the desktop, reported the correct amount of RAM on the System Properties tab, otherwise behaved normally, and Shut Down without issues. I haven't had time to run other tests, use it on a daily basis, or experiment further with that particular setup yet so I don't know if other issues may be present. rloew did verify the problem reported with USBPORT.SYS and verify that replacing it with the Server 2003 version fixes it.

Is there any known reason to prefer an original pre-SP2 copy over a patched SP2 copy? I've only begun to look at this recently so I've read the relevant threads here but I'm not familiar with the subject beyond that. The version of NTOSKRNL.EXE on a given machine is chosen by SETUP based on the CPU and so forth. I assume that the version of HAL.DLL is chosen in the same manner? If that's the case would it not be possible that a different version of HAL other than HALMACPI.DLL will sometimes be chosen depending on the hardware? If I'm correct on this it would mean that other versions of HAL should also be checked for any updated files and checked for the proper locations to patch...

No ATI driver version, or device "INF installation path" from the ATI packages, or even a modified INF based on the working X700 report and the Windows XP INF for the X600 works with my HP ZD8000 Mobility Radeon X600 after hours upon hours of experimentation. I even replaced the VBIOS ROM of my system with the one from Dell but it causes the system to become unbootable. Had to recover the original BIOS with an emergency floppy. Looks like this is a dead end... (And we've gone way off the main topic here! lol...)

Links you sent me back then for this are now broken. (I still have the files though.) I tried using them with my HP ZD8000 laptop which has an ATI Mobility Radeon X600 chip (no ATI driver will work after hours upon hours of attempts). They seem to work OK but I always get a "Fatal Error! Unable to load nga_w32.dll!" on boot and whenever I try to access the SDD control panel. (Running 95 OSR2) I don't know if its related to this or not but after some time has passed the system sometimes locks up. Any idea if these drivers have problems on systems with large amounts (>512MB) of RAM? Are these files from the last SDD version (7 beta) or a different version? Most info I can find on this seems incomplete, and you seem to be the most well versed on this subject... I'd also like to request in your next version of VBE9x to include a 256MB version since they are now RAM-limited. Any updates in the works?

The onboard SATA controller MAY support IDE mode just fine but the option to change it is just hidden by the BIOS. Laptop BIOS are usually garbage with very few options. It may be possible to reenable the hidden options IF there are any... May be worth a visit to BIOS Mods...

Already tried that. No, as I mentioned much earlier there was no official 9x driver for the Mobility Radeon X600 although other X600 devices were supported. Your report of the Mobility X600 working seems to be the only one out there, although there is a report of an X700 working at least partially. So, now you seem to not remember specifically whether you actually had the X600 on the Dell Latitude D810 working or not..? Some searching on eBay reveals that the Mobility X600 device for the Dell D810 machines is removable and replaceable while the one for my HP ZD8000 is built on the motherboard. I wonder if the VBIOS of the two are interchangeable? But that's irrelevant if the 9x Catalyst drivers don't in fact work with the Dell either...

Do you still have this particular system and OS installation up and running? If so, I'd like to ask you some questions about it. What exact ATI package file did you download and from where? What is the exact VEN&DEV hardware ID of your Mobility Radeon X600 device? What exactly was this laptop? (Forgive me, I'm sure it's mentioned somewhere before but since so much time has passed let's state it again for the record.) I ask because I am attempting to get Windows 9x up and running on an HP ZD8000 laptop that uses the X600 graphics controller, and it absolutely refuses to work with any ATI package I have tried all the way from Catalyst 4.11Beta to Catalyst 6.2. As far as I can tell from Google you are the only person to ever get an ATI 9x driver working with the Mobility X600.

That's about the sum of it. There's just not enough people left with enough knowledge to undo the changes that cause incompatibility with older systems in newer code. (Although, and this is purely conjecture on my part, it might actually be surprising just how few changes would need to be made. I make this statement based on the fact that apparently the latest version 27.0.2 of PaleMoon can be hacked to run on Windows XP despite the continued insistence of it's developers that this is impossible.) IMO (and I'm no programmer), if anyone did make the attempt the way forward for 9x lies with K-Meleon. I've only used it in a couple of tests myself but it essentially takes Firefox and strips it of all of the "eye candy" and "window dressing" rubbish and uses only the webpage rendering engine while using the Windows API for all of the interface. In this process one loses some functionality and ease-of-use with Addons and such, but it would eliminate a raft of UI API calls not present on 9x from the equation and leave only the rendering engine to make compatible. Then all we would need would be someone good with the 9x UI to make a proper skin to dress up the final product. Tihiy? Most likely pipe dreams at this point, lol.

I wouldn't recommend anything later than Windows Vista for anyone or for anything personally, but that's just my opinion. (And I don't even really like Vista, but at least they tried to make it pretty...) Much as all of us here hate it, it has become almost impossible to use Windows 9x online these days. Over the past couple of years it's becoming more annoying to do so even under XP. One can still use 9x for many computing tasks but browsing online is not one of them and is not an enjoyable experience.