win32

BWC's 310.70 allows the stable use of Sid Meier's Railroads, but I'm having trouble with GTA: SA version 1.0 and City Life 2008 Edition. Instead of loading a game or even its cutscenes, I get an endless loop of D3D9.DLL calling EnumDisplayMonitors and GetMonitorInfoA in the case of the former. In the case of the latter, it does that for a bit but eventually fails with an "unable to create Direct3D device error". Indeed, 310.70 is listed as a beta driver, whereas the next one, 331.65 isn't. So I'm off to try that. And 331.65 fixes those games!

Unfortunately ntdllx3 is a dud. I tried to boot up with it and I got 0x6B BSoD with parameters 0xC00007B, 0x3, 0x0 and 0x0. Or I meant it was a dud, since I decided to take it, change SectionAlignment back to 00001000 from 00000200 and it works just as well as x2. And it wasn't in vain since it's still slightly smaller than BWC's last ntdll.

Unfortunately, I've discovered an issue with ntdllx2. When installing VMware Players 3.0.0 and 3.1.5, a total of roughly 12 services and drivers should be installed; but only one was installed (VMware Agent Service). VMware Agent Service is also the only service listed in the ServiceInstall and ServiceControl tables in vmware player.msi (which is copied to a folder labelled vmware-xxxxx in Documents and Settings\%USER%\Local Settings\Temp\ when the installer is run). I don't know how the rest of the services got installed (the rest do get installed on XP x64). The other services are also installed with BWC's latest version. Mind you, even though Player 3.x is supposed to support XP and up, the installer's contents reference Windows 2000 frequently, and it does some really deep checks to prevent itself from installing on 2000 (fcwin2k, NNN4NT5, appcomp,cmd all can't fool it, so the only workaround is dropping LaunchCondition table from the MSI). With what I know, it installs like it would on a Windows 2000 system, which it functionally no longer is. Luckily BWC's ntdll is a good temporary drop-in replacement yet I wonder why these installers only install one service and then leave the rest to something else which I can't find at present. but still, yours has yet to give me trouble with RtlEnterCriticalSection!

This should be a good starting point: HFSLIP does not rely on USP 5.1. Remember to: -only use nLite (for integrating AHCI drivers) before running HFSLIP -not integrate the .NET addons as they don't work very well -disable UAC while running HFSLIP on an NT6 system It should be possible to manually integrate the extended core files (only necessary on UEFI-era hardware usually) in a separate (or maybe in the same) HFSLIP run as it can replace files and rebuild the cabinet files. Though I need to refresh myself on this subject. The new v16e is stable as I'm using it on my main workstation with no instability after extensive usage of web browsers, 3D gaming, VMs (as a host) etc. Extended kernel is a little tricky though it should be possible through manual file replacement (it could be through HFSLIP as well). Problem is that there are so many options (SSE2 or no SSE optimizations, enabling sets of Vista functions, WineD3D modules etc.) and there are now new projects growing out of it (like a new ntdll.dll with more functions than BWC's latest file). You may just want to install after installing the OS.

My solution was intended moreso for those installing. Since you're not, you can simply copy usbd.sys to C:\WINNT\system32\drivers and hopefully the USB drivers will be able to start.

It does seem that the USB controller issue is down to this: on hardware with USB 2.0 (EHCI) controllers and no USB 1.1 (UHCI) controllers, Windows 2000 ignores the [USB.CopyFiles.NT] section of USB.INF. In the UniversalHCD and OpenHCD sections of the file, there is a CopyFile instruction for USB.CopyFiles.NT, whereas there isn't one for the EHCI section. USB.CopyFiles.NT includes usbd.sys and usbhub.sys; the latter is redundant due to usbhub20.sys, but usbhub20.sys relies on usbd.sys as shown by Dependency Walker. usbehci.sys relies on usbd.sys indirectly, through usbport.sys. So I believe the solution is to open USB.INF, look for the section [EHCI.CopyFiles.NT] and add usbd.sys immediately below it. [EHCI.CopyFiles.NT] usbd.sys usbehci.sys usbport.sys I have no way of testing this as my hardware is either too old or too new to experience this issue. Please tell me if it works. I think that the H57 problem would be fixed by this change as it seems to have only EHCI controllers; thus UHCI drivers would be useless as there are no such controllers in the system.

Download BlackWingCat's IMSM 7.6 AHCI driver (as it works even with base SP4), and then replace its INFs with the one from the Lenovo package.

The newest version has some Windows XP POSReady updates along with most of the Windows 2000 updates from 2003 to 2013 (except for redundant and/or buggy ones). And the USP 5.1 included buggy usb 1.1 drivers. And yet it still retains the LBA48 reg entry.

Yeah, just remove the NTx86 references everywhere. That first INTEL_HDC too; it's probably there because they had some intention of supporting Windows 2000 but didn't in the end. Use nLite to integrate the driver before running HFSLIP; apparently there is a way to integrate drivers with HFSLIP but no one has documented it of course.

Well, 2012R2 idles at 500-700 MB of RAM compared to nearly 1.5 GB for Windows 8.1. Given the tendency for the NT 6 servers to be relatively unbloated yet remaining very usable for desktop usage (not much to do after enabling desktop experience and WLAN; in fact it's even easier than to set up than Server 2003), I would say that Server 2016/2019 would be far better than regular Windows 10, and possibly better than LTSC (which I've managed to get it to idle at 1.1 GB). You could make Windows 10 much lighter using a variety of tools, but the intrusiveness of that OS makes it difficult to keep it the way the user wants, so you may as well get an OS configured light OOB, like Windows Server. I have 12 GB of RAM, and I like a game that many say is great with 16. Of course, most people play it on Windows 7/8.1/10. But the client versions use over 1.5 GB of RAM that could be better used with the game; so 2012R2 with RAM savings of over 800 MB could prove useful. Thankfully this game also supports XP x64 which uses even less RAM but that's a fluke, as most people (and at least some developers) think 64-bit computing started with Vista or 7. I just need to install a new instance of XP x64, and then after that I'll set up 2019 and see how it compares.

Solution: https://msfn.org/board/topic/181284-attempting-to-install-windows-2k-on-thinkpad-x201i-running-into-sata-driver-issues/?do=findComment&comment=1178399 You don't need extended core to install IMSM 7.6 AHCI drivers, but you do need it for IMSM 8.9. Unless you follow some procedure @tomasz86 mentioned in a post in the 2000 USP 5.2 thread from late April 2019, using nLite after HFSLIP will cause serious problems with the OS installation. IE6 will think it's IE5, DX9 will think it's DX7, and some files will be missing.

What tool states that it's invalid? My test modifications seem to be fine with CFF Explorer and stud_pe. I don't know much about this stuff, but putting the new sections at the end shouldn't affect the sections before it, I think... try maybe adjusting the virtual offset of the section that corresponds to the import address table? We should page @blackwingcat since he would obviously know about this stuff.

These were the drivers I was thinking of: http://web.archive.org/web/20030312015545/http://link-empire.servebeer.com/files/omega/Ati/rad_w9x_omega_2576.exe Released December 2002, so they reference the earlier R200 cards like the Radeon 9000. But your card was released in July 2004. Maybe there were minor changes that could have made the older drivers unusable? Yet, since it boots to a blank screen instead of producing a blue screen or just not starting with one of the user-mode components complaining about the lack of EnumDisplayMonitors, I thought that somehow that one would work better with Windows 95 than drivers from 2003 or 4.

The ATi Radeon 8500 works in Windows 95 according to @LoneCrusader, and apparently the same applies for the Radeon 9200. I'd like to say that 7xxx/8xxx/9000/9200/9500/9700 Radeons work, since they are listed as compatible with supposedly 95-compatible drivers from 2002. But my Mobility Radeon 7500 just boots to a blank screen with any 95-compatible driver installed.

Why don't you try Stud_PE? I tried adding sections to ntkrnlpa.exe from XP SP3 filled with random binary data and zeros and it worked. http://www.cgsoftlabs.ro/studpe.html

I think the issue is that in Add/Remove Programs, there is no entry to uninstall the hotfix nor is there a hidden $NtUninstallKBXXXXXX$ folder in the root system folder (X:\WINDOWS). I fiddled with some of WildBill's unofficial Windows 2000 hotfixes that have a modified update.exe that ignores catalogs, in order to make UMDF and IntelPPM updates (those attempts have failed). But after making significant edits to those updates I could still get the uninstaller to work. So maybe you can base your updates off those. You should look at Windows2000-KB2507618-x86-ENU.exe; that actually deals with ATM.

No, but I did have occasional problems at 77f87eeb (RtlEnterCriticalSection).

It's a great idea. Adding all of those functions will certainly help with Process Hacker 2.x, which relies on many ntdll functions exclusive to XP and above. I wish I knew how to do this myself. I gave your file a spin and unfortunately, it does pose instability. Serpent 52 2020-02-08 gets quite unstable and has difficulty launching at times (though enabling multiprocess mode helps). Many sites have difficulty loading. Even an attempt at loading the task manager gave me a 0xc142 error though it opened on the next attempt. This is what a Serpent instance/tab did when it loaded. You should also bump up the version number, as both BWC's latest ntdll and yours are 5.00.2195.7125.

That is the driver I'm using (ThinkPad T41 user). Then I thought of omega drivers and how well they worked on my Mobility Radeon X1300 with Windows 2000, so I'll try these out: https://retrosystemsrevival.blogspot.com/2019/06/windows-9x-ati-omega-drivers-2637.html I will be pleased if I can play Chip's Challenge without it locking up on me.

8.1 should be installed after Vista. Aside from that, there shouldn't be any special considerations.

I did omit it, since it wasn't present in the Windows 10 reg entry (since it has no map interface for time zones, I think that's why it doesn't have it; but many XP time zone entires don't have it either) but it would have been preserved on any NT 5.x install that had KB4507704 or previous Morocco time zone information.

https://mega.nz/file/510F0D5C#IA6TYIdyzURFlAZfGsU4XX1Q3cCyi70rfGEYe4KQlC8 OK, this one was simply taking the reg entry from Windows 10 (and removing the MUI references) since I still need to establish how the hex translates to time change date/times but it works.

I believe so. I managed to install the UMDF services and driver, but none of them were able to load. Loading wpdusb.sys on its own causes a BSOD.

I'd think that it would be considerably easier than 7 since Windows 8.1 comes with USB 3 and NVMe drivers.

Yes. That one probably lacks the UMDF components.