LoneCrusader

Install 98SE and then NUSB - nothing else. See if the problem persists when U98SESP3 is not involved. When you have 98SE set up and are ready to install NUSB all you need to do is open Control Panel / System and remove all USB devices from the Device Manager and then run NUSB. No need to delete or rename any files. Everything will be redetected and reinstalled on reboot. Don't connect any external USB devices to the system until AFTER NUSB has been installed and you have rebooted the system a second time.

Hmm... Have you tried running FIX95CPU again after you get this Windows Protection Error?

Could you please provide more detailed specs for your machine? How much RAM are you using?

I believe all you should need are USB2.INF, USBEHCI.SYS, USBHUB20.SYS, and USBPORT.SYS from NUSB to fix USB2 under ME. You MAY also want the XP version of USBCCGP.SYS if you have any composite devices that use it but it should still be considered experimental...

KernelEx is NOT for device drivers unfortunately. You may cause the driver setup program to run but if the drivers themselves are not Windows 9x compatible KernelEx will not make them work.

This is not necessary. Windows 9x can only see and use one core; any others are simply ignored. I have run Windows 95 on a Dual-Core machine without any need to make such a change. BTW, if you are committed to doing this I highly recommend rloew's various patches. They work 100% and he is always helpful if you need support.

In order to give a more precise explanation I would have to go back and retest some of the hardware setups I used during some driver testing I did recently... but to make a long story short, I was doing extensive tests under Windows 95. Several of the systems I used had no Windows 95 drivers and in those cases I sometimes ran VBE9X to escape 640x480x16. If I am remembering correctly, at one point I had two machines with video cards that have official Windows 98/ME drivers but no 95 drivers, these being an ATI Radeon 9800 XT and an ATI Radeon X800 Pro. The Radeon X800 Pro system would not run DOS Boxes in windowed mode without crashing with VBE9X, while the Radeon 9800 XT system was able to run DOS Boxes in windowed mode without issues. There may have been another instance of this but my memory is fuzzy at the moment and I only noted these things in passing as I was concentrating on the driver being tested.. also I didn't have much hope that VBE9X would be further updated but I am very pleased to find I was wrong!

+1 I actually started having trouble using it with eBay under XP the other day though... ridiculous. It has started the exact same behavior with regard to JavaScript that already plagues 9x - seems to prove that the problem was never 9x but poorly written garbage websites to begin with.

Any chance of getting the issues with windowed DOS Boxes addressed? They seem to work on some machines but not on others...

No, no, no. There is no "trouble" or "troubleshooting" to using Windows 9x on a wide range of hardware far superior to Pentium III. And no manufacturer is going to TELL you their machine is designed for an older system. They do what Microsoft wants them to do, and promote the "current" Microsoft OS of the time.

Apparently unrelated but maybe not so: can anyone point me to detailed information on (re)creation and structure of SYSTEM.1ST, located in the root of the bootable drive? Mine got corrupted a year or so ago and since then boot time increased very much. I wonder if the lack of this file triggers some hardware/software redetection at boot time and if this could help in creating/updating the CPU info displayed by SYSDM.CPL... I originally thought that the Microcode update/UPDATE.SYS had something do do with identifying the CPU as well, but it doesn't. For some reason I have yet to determine, changing UPDATE.SYS can cause no CPU Info to be displayed at all, but it does not actually create the CPU info displayed on the System Properties tab. It is only responsible for the "Update" related entries in the corresponding Registry key. From what I have been able to determine so far, when Microsoft issued SYSDM.CPL 4.10.2223 for Q216204, they actually completely broke the CPU ID function rather than properly fixing it. The KB article refers to "incorrectly identified Intel CPUs," but it appears their solution was to have NO identification of Intel CPUs at all rather than fix the bug. Since HotFixes are cumulative, the botched changes in 2223 that were only intended for Intel systems got carried over into 4.10.2224 for Q272620, which is a "good" HotFix that fixes a bug common to all systems. This probably explains why all CPU info disappears when this HotFix is installed to an existing system. The "Stepping Data" string is actually NOT correct for identifying an Intel CPU. It SHOULD print a friendly ID string, similar to the one displayed by the BIOS during POST. This "Stepping Data" sting is what Microsoft described in Q216204, and rather than fixing SYSDM.CPL to print the "friendly ID" they broke the ID entirely.

Update...This issue only seems to get stranger the more I work on it. The 98SE HotFix version of SYSDM.CPL that I tested in the quoted experiment worked fine for displaying the CPU info - on an AMD system during a clean installation ONLY. It is broken on an Intel system under ALL conditions.

Any Firefox beyond 3.6.28 is, IMO, garbage. I used to love Firefox, but the interface changes after the 3.x line and the idiotic "fast release" cycle have turned me off completely. I don't know for certain what has been changed after the 5.x versions, I did test some of these at one time, but I was so disgusted with it that I scrapped the testing and decided to keep 3.6.28 to the bitter end, lol. I was able to recover most of the buttons on those versions and re-enable the Status Bar with an Addon, so you may be able to recover some of that functionality by exploring all of the settings and searching for Addons... EDIT: Search for these addons: Status-4-Evar Firefox 3 theme for Firefox 4 Old Location Bar

Disabling multiple cores (or HyperThreading) on processors is NOT necessary. I have run Windows 95 on HyperThreading P4's and some later dual/multi core processors as well with no issues. The extra cores are simply ignored, and there is no reason to disable them, especially in a multi-boot environment where other OS'es can use them. If you want a "performance" Windows 98SE system, look for a system (desktop or laptop) with a motherboard based on the Intel 865/875 chipsets and with a 3+ GHz P4 processor. Virtually all systems with this configuration should have working drivers (or drivers should be available that can be adapted).

I wouldn't call this laptop 98SE friendly by any means... There aren't any existing 98SE compatible drivers for your hardware. But, if you are determined enough, you can probably get everything to work under 98SE except the Video (you'll be stuck at 640x480x16, maybe fixable with VBE9X depending on the intended use), Audio, and Networking functions. Bluetooth is a toss up if you can find a 98SE compatible stack that works with your hardware. It really depends on what you want to use 98SE on this machine for.

Reproduced this error on a VM. I have removed some unnecessary entries for testing and it prevents the error on my VM. This is the first such problem I have seen with an INF size smaller than 64KB. The real "limit" must lie somewhere in between 63KB and 64KB. I did some experiments the other day to try and pin down the exact "INF size limit" in bytes. Unfortunately, an exact byte limit does not seem to be the problem. Apparently, any text coming after a "comment" ( ; ) is either NOT counted toward whatever limit is being triggered, or is counted differently than text that is not commented, so the actual file size byte count may not be a reliable indicator of whether the INF will work or not. I removed similar amounts of "commented" and "uncommented" text from the earlier INF that had the problem, and got different results for each experiment despite the overall size reduction being similar. More experiments might be revealing in the end, but it looks like this may be a "trial and error" issue for any future changes to this project or other INF projects.

I know what you mean, I've seen it (and fought against it) on other forums as well. Rest assured, you will get none of that rubbish here. Welcome to MSFN! In your case you will obviously need a motherboard that does have Windows 98SE compatible drivers (more on that below). But first, realize that many of the so called "limitations" of Windows 9x have long been fixed. You do not have to limit your CPU speed (I have 3.4GHz Windows 95 systems). If you are willing to purchase some of RLoew's patches, then you can use up to 4GB of RAM and you can use HDD's larger than 137GB (this is the limit without a patch) and you can also use SATA drives. Now, on compatible hardware. Do you wish to build an Intel system or an AMD system, or does it matter? I am much more familiar with Intel systems, so I can help with those, but not so much with AMD. If Intel, then I recommend a motherboard based on the 865 or 875 chipsets. These chipsets have Windows 98SE compatible drivers and can use P4 Socket 478 CPUs up to 3.4GHz (slightly higher LGA775 P4's on 865/875 LGA775 boards, but these are more rare). These boards also typically support up to 4GB of DDR RAM and have AGP8x slots. Some have Gigabit LAN chips as well. Look up the Intel D865PERL and D875PBZ boards for examples, but there are many of these made by other manufacturers besides Intel. For video cards, I am much more familiar with ATI than NVidia, and the last 98-supported AGP 8x ATI card is the X850 XT Platinum Edition. The slightly older 9800 XT is still a very good card as well. PCI-E under 98SE is very "iffy" - it works for some and not for others, but in your case you will need proper driver support for your video card for games, and only a very small handful of PCI-E cards have 98 compatible drivers.

*EDIT: This works ONLY with an AMD processor. Does NOT work for an Intel processor.* This is possible, but requires full slipstreaming rather than a Service Pack. In the case of SYSDM.CPL and the CPU info, you can solve this one issue by placing a copy of the updated SYSDM.CPL (v4.10.2224 I believe?) in the \WIN98 folder that you install from (by copying it to the HDD and installing from there, or creating a new CD) along with the 98SE CAB files and then use this folder to do a new installation. The updated SYSDM.CPL will be used rather than extracting the older one from the CABs. (This works for any file that gets extracted from the CABs.)

This is a known bug when updating SYSDM.CPL after Windows 9x has been installed. As far as I know, there is no fix at the moment. It is NOT a bug in the new .CPL file, because if the new .CPL file is used during original SETUP, then the CPU info is displayed normally. The CPU info display under System Properties is very "finicky" and sometimes behaves strangely.

MSAA 1.3 SDK

Thanks! You're becoming the new MDGx for Windows 9x.

You might try using RLoew's DLLHOOK program to redirect the FlashWindowEx call to another function...

Qshare seems to be dead. I have uploaded it to zShare, but who knows how long that will last either. EDIT: New link: INTEL_INF_98SE.ZIP - 44.9 Kb

I'm sure there is, but that's not the right way to do it, and I have never tried to do so. Besides, once those devices are identified, it triggers the detection of others. That's what is supposed to happen. I suggest you start from the beginning. Install Windows 98SE, then install the Intel INF package I linked before doing anything else. If 98SE hangs on shutdown, give it a minute or so and press RESET and allow everything to continue. Sometimes you will see shutdown problems under 9x on newer systems. Once the drivers are installed, it should not be a problem, but if it still is I suggest you download and install the 98SE Shutdown HotFix.

You have to use the last 98SE compatible version of the Intel Chipset INF utility. The version provided by Intel for that board incorrectly claims 98 support. All of the other provided drivers do work.