Mov AX, 0xDEAD

This is patch for W2003 src, for XP src tree first patching file must be /driver/amlinew/amlipriv.h

NDIS6 is complex beast and it's not about the symbols at all...

Off-topic: I bought today Compaq Prolinea 4100 (486 DX4/100Mhz) case for 8$, fastest among yours

@Dietmar Check with Task Manager, kernel may detect only one core/thread

@Dietmar There is no new acpi.sys updates from me, v7 is still last, compile yourself as usual Seems videodriver issue is not related with acpi, i see error on sxs.dl

@Dietmar Some strange with second PRT4: DSDT: PRT4 was processed few time in loop and no Name(_ADR,0x4) in log There is two PRT4 in dsdt: 1) EHC1.RHUB.GHUB.PRT4 Device (EHC1) { Name (_ADR, 0x00120000) Device (RHUB) { Name (_ADR, Zero) // _ADR: Address Device (GHUB) { Device (PRT1)... Device (PRT2)... Device (PRT3)... Device (PRT4) { Name (_ADR, 0x04) 2) XHC0.RHUB.PRT4 Device (XHC0) { Name (_ADR, 0x00100000) Device (RHUB) { Name (_ADR, Zero) Device (PRT1) Device (PRT5) Device (PRT2) Device (PRT6) Device (PRT3) Device (PRT7) Device (PRT4) { Name (_ADR, 0x04) Device (PRT8) 1) Try to remove ACPI0007 patch from acpi sources (namedobj.c) 2) Try to remove second PRT4 definition from DSDT

@Dietmar Hi, too much "ACPI0007 def #1" in DSDT/SSDT, as we know this is hack to convert Name (_HID, "ACPI0007") to Processor(...), hack has two cases with definition len 0x1A and 0x1B, probably you have another len and need retune hack again Next, ba91ca7a: AMLI_ERROR(c0140002): AML Stack overflow PushFrame: stack ran out of space , i think AML interpreter was stopped at this point and remain dsdt was not processed Also "ACPI0007 def #1" is from v6 patch, in v7 it is "Try ACPI0007 def #1" Can you share raw dsdt/ssdt tables ?

Hi @Dietmar Just make more debug log if it possible: Lite: ed Kd_ACPI_Mask 0xFFFFFFFF Full: !amli set spewon verboseon logon traceon

Hi Dietmar Does Standard PC support multi cpu in any way ? Don't know much about 0x7B vs ACPI issues, too hard to diagnose, boot from disk initiated by kernal as i remember, technicaly acpi can reject requests to disk, but it means only one bad thing - acpi is tottaly messed up on your system, useless info to find something

Confirm, DEV_8125 listed as compatible with Win10 kernel debugger driver, try LAN mode on target, you don't need USB on host anymore

Hi Gordo 1) USB - serial connector is OK for host, but not for target, i personaly use cheap "Silicon Labs CP210x usb<->com" on my XP laptop as host 2) Host can be even XP, main requirement is able to run WinDBG 6.3.9600, for XP i used patched windbg.exe(original exe misses some .dll dependies) Another option is LAN cable, what LAN chip on Asus B760M ? If original Intel USB3 Win7 driver rejects on your MB (code 10), you have nothing to do, try AMD or backported Win8 driver

Hi Kasumi Unfortunately kernel crashdump has no information what wrong with DSDT code on your bios. Currenly we have only one way - remote connection from other PC to save dump of ACPI.SYS exection flow.

Try old, but good SATA AHCI Controller 11.2.0.1006 by Intel, it is not storport-based, so even kernel patch not required. Probably it has problems with proper shutdowning SSD, it missed specific SSD (DIPM/HIPM) power commands With anyhing, sleeping is always bugged on XP, there is no cure and no knowledge what's wrong, i have same problem with my notebook(ivybridge, 2013, no need any patches for ACPI), in 80% i have blank screen and only in 20% succesfully power on

Hi Probably PAE (if active) mode messed crashdumping, if not active - recheck storport.sys + diskdump.sys + current disk controller driver.

Look at intel or amd CPU Programming documentation how to control SSE bit in MSR registers, "Internet" says some old BIOS have "SSE Enable/Disable" option, so it must be possible

For Windows sse2 is same like sse1, if you don't want sse2, disable sse too. There is no switch for windows kernel to disable sse, you need modded kernel, do it yourself

"if you don't understand what's going on, blame the ACPI driver"

bootvid doesn't work this way this .DLL needs to rewrite completely because it uses raw VGA register programming

If something I did for x32 is also suitable for x64, I usually do it too. But I'm not interested in doing something "unique" only for x64.

I think such processors will not be good because any win32 software will not run on it, all 28-year legacy software since win95 era is dropped, same crop like uefi-only booting Basicaly we can patch WinXP-64 to run on pure 64 cpu mode even in uefi environment, need rewrite legacy vga output like boot graphic, boot menu, bsod screen etc and initial memory manager

I dont have any new files, kdnet is not updated since 2021 because i don't found bugs, just need take kdnet10.dll from last Windows 10 releases if need support for some new chipsets This is as should be, kdnet based on windows10 files requires 3 files, check with DepWalker, they linked together. kdstub.dll is little converter between Windows8.1 API <->Windows10 API kdstub.c

Try /PCI_ID option, intel driver parse PCI_ID before adapter detection. I was unable to test driver because i don't have MB with intel lan chips

Sorry, i don't play with "Binary BLOB" games, ntoskrn8.sys is 2 000 000 bytes, what is inside ? ntoskrnl from win7 ?

ndis.sys/netio.sys/msrpc.sys from 5048 has many missed imports from kernel how it will works ?

@Dietmar, https://msfn.org/board/topic/183528-ndis6-support-for-xp/?do=findComment&comment=1244341