Damnation

@Omores @MrMateczko I encountered this error when I tried to install 98SE on my Threadripper system, I never did find a solution to it unfortunately.

@Dave-H No, Integral Edition has not been updated since then, you have the latest. Make sure to replace the acpi.sys with the new compiled before running the optional patch integrator on the ISO.

@Dave-H Integral Edition could give you a smoother installation experience if you're installing XP from scratch (although it's good that you're testing @George King's XP2ESD too) Just make sure to swap out the older acpi.sys in Integrator Files\Patches\ACPI drivers with the newer compiled one before running the optional patch integrator script.

@Mov AX, 0xDEAD Do you think my implementation of Kisystemservice with a precompiled object is the right way to go about it?

@Just_AHD I can't say for sure, I've never used SP4. As a personal preference I like to use XP Integral Edition, it has all the POSReady July 2019 updates integrated already. Everything I want to use XP for seems to work well for me using that.

@Dietmar alright, thanks for testing. I'll work on implementing some more functions I'd stubbed properly.

@Just_AHD Try office 2010 without the unofficial XP SP4, might be conflicting.

@Dietmar can you test this to see if swapping windows7 ndis/netio/msrpc.sys on working XP install works for standard XP NDIS5 drivers? or does it fail?

@George King are you able to debug?

@George King That gets rid of the 0xDEADBEEF BSODs from ntoskrnl version mismatch - I got rid of it out of habit since I'd tried porting to windows 2000 earlier.

@George King I did an implementation of IoSynchronousCallDriver in x86 assembly, so no x64. see if it works or not. link: https://ufile.io/4un6rqyg edit: noticed and fixed a mistake https://ufile.io/w9bkpf69

@Dietmar @Mov AX, 0xDEAD @Dave-H It's the object already exists problem again - ba150748: \_SB,46AMLI_ERROR(c0000035): Object already exist CreateNameSpaceObject: object already exist - \_SB

@Dietmar these are the commands @Mov AX, 0xDEAD told me to run for a useful checked ACPI debug.

@Dietmar It's not the same error at least. I'm going to implement MmAllocatePagesForMdlEx, KeFreeCalloutStack, KeAllocateCalloutStack, and KeTestSpinLock next with a precompiled objects.

@Dietmar these files I modded earlier should still be good to use - https://ufile.io/5cxo9w60 can this windows 7 ndis.sys still work with NDIS5 drivers?

@Dietmar @George King @Mov AX, 0xDEAD I've gotten a KISystemservice implementation building with a precompiled object. I'll upload what I have in a bit, hopefully that'll help with the msrpc.sys BSOD. edit - upload here: https://ufile.io/fbdrjma3

@George King I'll try making workable code out of it, no guarantees. If you're willing to put up with 32bit only I can make an exact copy of this function in assembly, if you want?

@George King here, psuedocode will need reworking to be functional though. NTSTATUS IoSynchronousCallDriver_k8 ( PDEVICE_OBJECT DeviceObject, PIRP Irp ) { unsigned int local_0x18; // [esp-24] unsigned char local_0x14[20]; // [esp-20] unsigned long v1; // eax local_0x18 = (unsigned char)&local_0x18 & 0xFFFFFF00; local_0x14[0] = 0; local_0x18 = 1024; local_0x14[8] = &local_0x18[2]; local_0x14[4] = &local_0x18[2]; *(*(Irp + 96) + 4294967292) = &local_0x18; *(*(Irp + 96) + 4294967288) = &CmpCompleteFlushAndPurgeIrp; *(*(Irp + 96) + 4294967263) = 224; v1 = IofCallDriver( DeviceObject, Irp ); if( v1 == 259 ) { KeWaitForSingleObject( &local_0x18, 5, 0, 0, 0 ); v1 = *(Irp + 24); } return v1; }

@Dave-H BSOD A5 0x2001 - ACPI tried to evaluate the PIC control method but failed. This would explain why you're running out of hardware resources too, Programmable Interrupt Controller isn't working properly so you're running out of Interrupts. This is one of the unresolved BSODs mentioned on page 1 of this thread, so setting up windbg is a must to solve this. Sorry this is turning out to be so complicated for you @Dave-H

@Dave-H post a screenshot of the BSOD, maybe some of the parameters about the error have changed?

@George King @Dietmar @Dave-H So we're still getting a DDB error even when using our usual DDB fix, so we're gonna need to get windbg running to find the exact cause of this one. I left a post with debugging instructions here -

@Dave-H can you post a screenshot of the A5 BSOD please?

@Dave-H If you're willing to do a fresh install that would save a lot of hassle getting this working. If you must preserve some existing data for whatever reason, make sure you do a backup!

@Dave-H That's outdated. here - https://ufile.io/yuxh9ziz first try the one in 5.1.2600.7777.4\i386_free do an "ACPI Multiprocessor PC" install with this one. report any problems.

@Dave-H Try switching to "ACPI Multiprocessor PC" with the newer acpi.sys