pappyN4

I have not tried since I only use 2003x86/64 for testing to see if a problem is with x64 vs x86 or with NT5.1 vs NT5.2. Is there a server 2003x86 update pack with all latest post SP2 updates?

"Installed INF does not contain digital signature" warning I got this warning after integrating the USB3 drivers. You will need to patch syssetup.dll file to get rid of it. Original patch is from @ Mattiwatti. I updated the x64 patch to match the one found in the 2019 updates. x86 syssetup.dll 5.1.2600.5512 @33bfe 8B F8 3B FE 75 20 -> 31 C0 31 FF 90 90 @33da6 8B F0 85 F6 75 17 -> 31 C0 31 F6 90 90 @35535 0F 84 9F 00 00 00 -> E9 A0 00 00 00 90 @37f03 3B C6 -> 31 C0 @37f0b 75 3E -> 90 90 x64 syssetup.dll 5.2.3790.4226 @423dc 75 4C 0F B6 0D 8C A2 08 00 -> 90 90 31 C9 90 90 90 90 90 @42431 75 4C 0F B6 0D 36 A2 08 00 -> 90 90 31 C9 90 90 90 90 90 @424cb 0F 85 D5 00 00 00 -> E9 D6 00 00 00 90 @425a6 40 8A 74 24 30 -> 31 F6 90 90 90 @3fe50 85 DB 0F 84 E3 00 00 00 -> 31 DB E9 E4 00 00 00 90 @41d2b 85 C0 8B F8 75 25 -> 31 C0 8B F8 90 90 @41fb6 85 C0 8B D8 75 1C -> 31 C0 8B D8 90 90 @4204b 0F 84 84 03 00 00 -> E9 85 03 00 00 90 @46bf9 85 C0 8B D8 75 47 -> 31 C0 8B D8 90 90 @571ab 0F 85 DB 00 00 00 -> E9 DC 00 00 00 90 @5728c 40 8A 74 24 40 -> 31 F6 90 90 90 @7204b 85 C0 75 1F -> 31 C0 90 90 @78a22 75 61 0F B6 0D 46 3C 05 00 -> E9 B9 00 00 00 90 90 90 90

@AndaluYes, my uasp restart issue is fixed with the test1 driver. Coexistence \Device\UaspPort is already changed in the stor_uas from test1. Only cosmetic issue remain is that stor_ua8->stor_uas->storport somewhere. So test1 driver is not fully "self contained", but still works. Uasp via stor_ua8 uses the patched stor_uas and since both XP/XP64 both need storport for ahci/nvme drivers, the two imports that stor_uas asks from storport exist. Would be cleaner if we could get that last storport dependency removed from stor_uas/stor_uas8.

USB Attached SCSI (UASP) When using the AMD USB drivers for the motherboard with a UASP capable device it defaults to using the default USB Mass Storage Device (usbstor), same as a standard USB stick. With the backported Win8 drivers it's different. If you do not have a UASP driver installed (USB\Class_08&SubClass_06&Prot_62) then it will not find a driver by default. You can force install by manually selecting Universal Serial Bus Controllers - USB Mass Storage Device, or install a uasp driver. Note that some motherboards may have a deadlock issue when using the uasp driver during reboots with the USB adapter still connected. If you select Safe Remove, reboot works fine, otherwise it doesn't finish the reboot. You will need to use a deadlock patched version of the driver if thats the case. The following test is with a JMicron USB to PCIE adapter, VID152D PID0583 and a XG5 256GB NVMe. The Gen2 port on the B450 chipset and the Gen1 port on the B450 and on the Ryzen2. Drivers are the default XP usbstor 5.2.3790.5829, VIA uasp 6.1.7600.4002, and win8 uaspstor 6.2.9200.16384.

It's nothing special. There was a win7 update a few months ago that had newer version for a couple of files. storport is 2kb bigger now. I updated the readme on github, I dont think its merged yet. https://github.com/pappyN4/NTOSKRNL_Emu/blob/master/README.md https://ufile.io/s8cp2lko

@canonkongGood to know. I tested daniel_k's usb3/uaspstor on Windows 7 and it worked fine at least on there. @Andalu Thanks for the testing. I havent tried the Etron, I have tried the VIA one from GeorgeKing drivers pack and from the Integral ISO. It works, but is slower. The Win8uasptor is super fast. MovAX sent me a copy of the 2003/w8_storport versions he had. Restart worked fine with both of those, but as soon as I put back the one using w7_storport restart stopped again. Must be something with my motherboard. I patched the deadlock on the win7_storport using the 2003 method from MovAX. It is not "pretty", since uaspstor needs the storpor8.sys extended function then the w7_storport. So two extra files for uasp driver to keep storports separate, but it does reboot now. Ideal would be to use w8_storport, but looking at it, it is not a simple thing, I would not be able to make one for x64 or update the Emu_extender readme to make it accessible for anyone to make. Here's x86 version using w7_storport. https://ufile.io/u28mfr1v There is still one little change left to make to make it completely self contained. Right now it looks to me that the renamed stor_uas still has contact with the original storport.sys somewhere. Need to figure out how to change that. It's probably simple but I'm too tired to see it now. Maybe @Mov AX, 0xDEAD can take a quick look? Or if you see a simpler way to do uasp/storpor8/w7_storport ?

Thanks. Just for clarity, the PS2 keyboard is only used when installing windows (at least until i figure out how to integrate win8 usb into DOSNET/HIVESYS/etc...) While I was testing the adapter, keyboard was USB, no PS2 port was being used. Did you use the adapter in another computer without the PS2 issue and it works? Same win7_storport, ntoskrn8,usbd8,storpor8 combination?

@Andalu JMicron, VID152D PID0583. JMS583Gen 2 to PCIe Gen3x2 Bridge

I bought a USB to NVME adapter to test out uasp. Can anyone confirm if win8_uaspstor+win7storport doesnt need any special fixes? I gave it a few crystaldiskmark tests and it seems ok to me. storport 6.1.7601.25735, uaspstor 6.2.9200.16384, with uasp security_cookie changed, patched to ntoskrn8,usbd8,storpor8, and wpprecorder. thanks edit: I did find one quirk. With usbstor, or the VIAuasp, restart/shutdown works normally. But with uaspstor from win8 I have to do "Safe remove" to be able to restart normally. If I dont, then Windows goes through all the login off steps, screen goes black, but stops right before the reboot. This is very similar to the restart issue when a PS2 keyboard is attached to the mobo, so might just be a quirk with my Gigabyte...

@George King If you want to test I can send it to you. I have not tested myself.

Samsung NVMe and SMART with CrystalDiskInfo The samsung nvme driver is able to display SMART data in XP/Win7. To get CrystalDiskInfo to work instead of crashing when you start it in XP/XP64 you will have stop it from using WMI. Open DiskInfo32 or DiskInfo64 with a hex editor, find where "\ \ . \ r o o t \ c i m v 2" is and change it to something like "\ \ . \ r o o t \ n o w m i". Tested on version 8.15.2 630069006D00760032 -> 6E006F0077006D0069 Samsung NVMe potential issues When I installed the latest nvidia 368.81 driver it caused the Samsung 3.3.0.2003 to stop working on XP64. @Andalu tested the same drivers on an Intel B250/GT730 and worked fine. After playing around with it, installing nvidia 355.98 appears to work/coexist with the Samsung NVMe driver on XP64. I've only tested on a GT710 so far but hopefully same result for my GTX750. Something to keep in mind if you have a amd/B450/gigabyte combo. On XP32 the Samsung driver did not work (code10), using the default non patched, or with the PAE patch /4GB. It did however work when PAE patched using /128GB or /all. Both with 355.98 and 368.81 Trim tests O&O Defrag 17.5 is the best from all the ones I tested. Works on SATA and NVMe SSDs. AMD/Samsung/nativeWin7 drivers were all able to TRIM. Can be run from CMD line so you can schedule it if you want. NVMe Trim ADataSSDToolbox doesnt see (Win7/samsung/SchtromOFA13 driver) TxBENCH see but no trim (Win7/samsung/SchtromOFA13 driver) SSDTool 0.95 see but trim didnt work (../../samsung) Naraeon5.4.0 win7 no see, SchtromOFA13 trim didnt work, samsung TRIM WORKS O&O Defrag 17.5 win7 TRIM WORKS, SchtromOFA13 see but no trim, samsung TRIM WORKS SSD Trim Naraeon5.4.0 AMD yes, Win7msahci yes, SchromAHCI no see, Win8ahci no see O&O Defrag 17.5 AMD yes, Win7msahci yes, SchromAHCI no trim, Win8ahci no trim used trimcheck0.7 to see if it worked (https://github.com/CyberShadow/trimcheck) *trimcheck x64 works also if Major/Minor changed from 6,0 *Naraeon has a "not a valid date and time error", doesnt affect functionality *O&O can be run from CMD line, needed to change Major/Minor from 6,0 down to 5,2 (or 5,1) to install Nvidia GPU driver & custom resolution Based on a few posts i found, Nvidia 355.98 is that last version where you can add custom resolution if for example you have a ultrawide screen or 4k monitor that you want to use. It looks like XP is limited to displayport HBR1 (8.64 Gbit/s). There is a bit of other stuff as well but roughly, horizontal pixels × vertical pixels × (3 × bits per color channel) × refresh rate in Hertz = Bits per second. Calculator https://tomverbeure.github.io/video_timings_calculator Ex: 2560 × 1600 × (3 × × 75 = 5898240000 bits/s / 1024³ = 6.86 Gbit/s 3840 x 2160 x (3 x x 40 = 7962624000 bits/s / 1024³ = 7.41 Gbit/s https://www.vogons.org/viewtopic.php?f=63&t=72471 http://wp.xin.at/archives/5616

@Andalu Glad that it works for you. Unfortunately for me, when I install nvidia 368.81 video card drivers the samsung nvme driver stops working with code10 on my B450 motherboard. If the video driver is not installed then it works fine. Win7nvme works, but that driver does not have SMART. Tried changing PCIe Gen version and PCIe Bifurcation settings in BIOS no luck. Older win7 23403 storport with MSI only patch or newest 25735 with MSI/smart patch still no good. Even tried older samsung driver, 3.3, 3.2, 3.1, and 2.4.7.0. Does not look like there is IRQ conflict or memory range overlap. Update: Got it to work when using a PAE 128GB or /all patched version. With no PAE patch, or PAE 4GB is when it does not work.

@Dietmar@Mov AX, 0xDEAD Thanks. I wanted to test a win8 driver/storport to see if it had same problem I was having with win7 backport. I updated the readme with newer version and some missing patches. https://github.com/pappyN4/NTOSKRNL_Emu/blob/master/README.md I did not add uaspstor as I do not have anything to test it with, not sure if it works when using win7_storport or storpor8.

catching up on TRIM in the deleted thread... Is there a backport win8 storport? Or is it just the storpor8.sys that was being discussed? NTOSKRNL_Emu has missing dependencies with win8 storport.

@Andalu Crystal Disk Info modded to stop WMI from loading. https://ufile.io/16asy5na I tried both 32 and 64bit version on my XP64. Both worked with my samsung nvme (didnt load NVMe size like for SSD/HDD but that info should be in the model names anyway). Now for the real question. Has anyone tried to use the samsung NVMe driver with a not samsung nvme? I only have the one so I cant test but I would be interested to see if the driver would work and pass through SMART for other manufactures. I tested the generic driver inf for samsung originally posted by Fernando and it worked for my samsung. Here's the same one for x86. (I leave it to you guys to mod the secnvme for x86).

https://itm4n.github.io/windows-dll-hijacking-clarified/ This explains better but Dependency Walker only checks if DLL is the import table. At runime a program can use LoadLibrary() to call other dlls. I looked through all the DLL from DepWalker that Crystal uses, but didnt see one that was "wbem" or "wmi" sounding... I tried replacing the 4 WMI related DLLs that Crystal needs into the windows sys folder. Dependency wise there was no issues, but Crystal still crashed. In IDA, there is a "Suspended on Library load/unload". Running it, it looks like ntdll.dll is where wbmprox first loads, but a bunch of other DLLs load in that same spot is how it looked to me in IDA. No idea how to look up where it was in ntdll.dll itself... Sumbled upon a "debug" setting that can be turned on in Crystal's ini. In there, the last entry before failure was "CMD_TYPE_NVME_INTEL". Opening up DiskInfo.exe, subroutine was findable with that name. A lot of IdentifyDevice, AddDevice calls, most ending up at a block with many inputs (all the Info/Add Devices Intel/Samsung/etc...) just before sub exits. Adding a breakpoint, it looks like Crystal gets through the NVME_INTEL block, and errors out at the MultiBlock.. Jumped out @4cf96 74 -> EB to skip "AddDiskNVMe - CMD_TYPE_NVME_INTEL". CDI starts at least, but nvme does not show up. Then tried skiping "DoIdentifyDeviceNVMeIntel" which calls ""AddDiskNVMe - CMD_TYPE_NVME_INTEL" @4ced8 74 77 -> 90 90 @4cf4b 0F 85 0A 01 00 00 -> E9 0B 01 00 00 90. Same things starts, but no nvme. 00001110 DoIdentifyDeviceNVMeStorageQuery 00001110 DoIdentifyDeviceNVMeIntelRst 00001110 DoIdentifyDeviceNVMeSamsung 00001110 DoIdentifyDeviceNVMeIntel 00001110 AddDiskNVMe - CMD_TYPE_NVME_INTEL Looked further up the logfile. Compared one with wbemprox removed to a normal one. no wbem / works 00000000 dlg.DoModal() 00000015 InitAta 00000015 CAtaSmart::Init WMI on - Start 00000015 CoInitializeSecurity() 00000031 CoCreateInstance() 00000031 NG:WMI Init 1 wbem / fails 00000016 dlg.DoModal() 00000032 InitAta 00000032 CAtaSmart::Init WMI on - Start 00000032 CoInitializeSecurity() 00000032 CoCreateInstance() 00000032 ConnectServer() 00000032 CoSetProxyBlanket() 00000047 OK:WMI Init 00000782 OK:Win32_IDEController 00000813 OK:Win32_SCSIController Searched around on on those keywords. ConnectServer() to a wmi server maybe? Ended up in two subroutines that sounded interesting. Saw "\\.\root\cimv2" which was familiar... Borrowed a quick vbscript to try wmi. Breaking the path stops it from working. strComputer = "." Set objWMIService = GetObject("winmgmts:\\.\root\CIMV2") Set colItems = objWMIService.ExecQuery("SELECT * FROM Win32_Process" & _ " WHERE Name = 'cscript.exe'" & " OR Name = 'wscript.exe'",,48) For Each objItem in colItems Wscript.Echo "-------------------------------------------" Wscript.Echo "CommandLine: " & objItem.CommandLine Wscript.Echo "Name: " & objItem.Name Next So in DiskInfo32 or DiskInfo64, find where "\ \ . \ r o o t \ c i m v 2" is with hex editor and change it to something like "\ \ . \ r o o t \ n o w m i". 630069006D00760032 -> 6E006F0077006D0069 Then let DiskInfo handle the error of a missing wmi the same way as if wbemprox file was missing.

Not much to go on in IDA to my untrained eyes. I assume all the dll loading happens at the very beginning of "start", errors out, then it quickly gets sent to trap. Error looks like it happens before we can do anything in Crystal. Playing with it there did nothing useful. ProcMon shows it using a couple of wbem dlls. Yet no dependency for it with Dependency Walker, or any mention of them in Crystal's dissembled EXE. I guess one of the known dlls causes the webm to be loaded without any explicitly defined imports. Something using LoadLibrary maybe... So the question is then. How does one use IDA or something else to trace where and in what DLL does wbemprox.dll first load?

IdaFree70 has fake limitation. If you change ida64.exe Major/minor Operating System version from 6.0 to 5.2 it works fine. No missing dependencies in DependencyWalker. x64 bit process.

@Andalu ...continuing from PM. Yeah, I tested renaming wbemprox.dll as a test yesterday it it fixed it for me as well on x64. I agree that renaming it is not the best since windows must need WBEM for other stuff. I looked at dropping the windows 7 wbem related files into the CrystalDiskInfo to load from there first, but its a lot a extra files, and some dependencies are missing. Easiest would be to simply stop Crystal from loading it at all, since it looks like it works well enough without it. Looked over the source code last night, there is a include <wbemcli.h> in AtaSmart.cpp and SlotSpeedGetter.cpp. I was thinking if maybe if I run it with IDA or something maybe we can track to where it breaks and hopefully jump over to the "file not found" case where it loads fine without it. Never used it that way before. Need to find an older version of IDA Free that works on XP. The one I have is to new.

If anyone is interested, there is a new storport for Windows 7 in KB5010404 x86 storport 6.1.7601.25735 msi patch hexposition: 4E1A change: 8B 8E 3C 01 00 00 -> B9 00 00 00 00 90 smart patch hexposition: 1030A 75 36 -> 90 90 hexposition: 10316 75 2A -> 90 90 ========================================================================== x64 storport 6.1.7601.25735 msi patch hexposition: 156AE change: 8B 83 C0 01 00 00 -> B8 00 00 00 00 90 smart patch hexposition: 192D5 change: 75 36 -> 90 90 hexposition: 192DE change: 75 2D -> 90 90

Ok, now for something new. Sadly doesnt seem to be much interest for x64 ACPIs. I started using it back in 2005, so I broke away from 16bit applications long time ago. I didnt actually use XP, I went from 2000 to XP64. Anyways... I'll start by saying I have no idea about patching ACPIs. I have no idea what each specific patch was made for. I dont know how to debug errors from ACPI. In the words of Mov AX, 0xDEAD, blindly turning off checks is not always the best idea. All I've done was compare the latest x86 5048 acpi (which worked on my ryzen), and tried to match existing patches and missing patches based on any calls or values by patterns and general structure/flow in IDA Free on the x64 5048. Some I feel confident I found the right one, others are reasonable, and one that is a total guess as it had nothing guess from. Feel free to open IDA and see for yourself. 5048 x86 "2020.10.24 - daniel_k, diderius6", all patches below are based on compare to original unpatched / patched file =Matched in x64 @27a80 00 00 00 -> A8 97 03 @27c58 28 92 03 -> 00 00 00 =Matched in x64 @d443 75 -> EB =MaybeMatch in x64 @15f15 8B F7 -> 89 FF -ex1 to x64 @283d6 7D -> EB -ex2 to x64 @d2b0 74 -> EB -exClose1 to x64 @f7e1 74 -> EB -exClose2 to x64 @4511 0F 84 39 01 00 00 -> E9 3A 01 00 00 90 -experiNotSure1 to x64 @317ac 75 -> EB @317c5 21 -> 00 5048 x64 "2019.6.30 (Ryzen) - diderius6" =Matched from x86 @4ae40 00 00 00 -> B8 C8 05 @4b1f0 18 BF 05 -> 00 00 00 =Matched from x86 @18ca4 0A -> 00 =MaybeMatch from x86 @26337 08 00 14 C0 -> 00 00 00 00 @26429 74 -> EB --end of origial 2019.06.30 patch ----------------------EXPERIMENT---------everything below this is untested guess from x86 -experi1 test @51bf1 79 -> EB -experi2 test @198dc 0F 84 7F -> E9 80 00 -experiClose1 test @1c666 74 -> EB -experiClose2 test @9638 0F 84 A8 01 00 00 -> E9 A9 01 00 00 90 -experiNotSure1 test @5fa63 74 -> EB test @5fa94 78 -> 00 experi1 in subroutine that has \\_S1,2,3,4,5 in the beginning shorty before the end, block just after a ZwPowerInformation and 2 ExFreePoolWithTag, x86 stop jump to 2001, 0A5, call: extern KeBugCheckEx:Dword so we do same, stop jump to __stdcall __noreturn extern KeBugCheckEx:qword =========== experi2 in function shortly after x64@18ca4,x86@d443 with 58080206h,4449485Fh push/mov, PNP0C0F is jumped so that it doesnt get to a 5349445Fh, which then goes to 10006h,0A5h ds:keBugCheckEx x64 version has a lot more paths to 0A5h, but same one jump to get away, good idea? I dont know. Another learning experience. This one I used the same E9 jump as from expericlose2, but this time I just tried with my own guess for the offset. I ended up with a bit longer jump then I wanted so I guess my math wasnt exactly correct, but was able to get it there eventually. =========== experiClose1 similar structure, compares 0FFFFFFFFh has 2 KeBugCheckEx, and ends with ExFreePoolWithTag first KeBugCheckEx is left alone, goes straight from head to bugcheck, other one near end is the one that is jumped out =========== experiClose2 similar structure, 4449555Fh,0Dh,0A5h to KeBugCheckEx at end, mentions 800h vs 80000000000h and 0A000h vs 0A00000000000h between x86 and x64 on x86 jumped to "retn" block, This one was a learning experience. Using the same E9 3A 01 landed in the wrong spot. Makes sense after looking up E9 (JMP) is based on offset distance entered. On the x64 where you want to go wasnt in the same spot. So used IDA free in text mode to see where the text jump landed with the wrong value from x86, in hex calculated the difference on how much further was needed, then added it to the JMP. So 31 01 (aka in hex calculator 0131 + 6F = 01A9), so A9 01. =========== experiNotSure1 not much to guess with, short subroutine, 2 KeBugCheckEx to avoid, total guess on this one Update https://www.mediafire.com/file/ietiycbz6eueacz/xp64_acpi5048_2022.04.12.rar/file

SATA driver speed tests (XP64 vs Win7x64) NVME driver speed tests (XP64 vs Win7x64) storport tests SMART tests ACPI tests

Build ISO with nLite WinXP64 ISO and updates en_win_xp_pro_x64_with_sp2_vl_x13-41611.iso 5eraph_Windows_XP_x64_Post-SP2_UpdatePack_ENU_v2019-08.7z; https://www.mediafire.com/folder/x6qqeye5y0a0x/x64 WinXP ISO and updates en_windows_xp_professional_with_service_pack_3_x86_cd_vl_x14-73974.iso XPSP3_QFE_UpdatePack_20210829_Pro.7z; https://www.4shared.com/web/directDownload/0t47E_Vhea/B8-gUmU2.8a5c8b6a84dcde59d76300e08ab95fac XPSP3_QFE_POSReady_Addon_20200528_5er.7z; https://www.4shared.com/web/directDownload/pCl2jGEqea/B8-gUmU2.7befe8c26d0ff79f226742fccad3d4c3 Note: order matters, in nlite have the UpdatePack run first, then the POSReady updates Note2: in POSReady updates, entries_POSUpdates.ini has a bug with nlite, you will need to comment out these lines ;Prevents Windows Setup copying drivers from driver.cab that the integrator failed to update ;drvindex.inf|ks.sys||2 ;drvindex.inf|termdd.sys||1 ;drvindex.inf|hsf_fs|hsf_fsks.sys|1 Compile NTOSKRNL_Emu; https://github.com/MovAX0xDEAD/NTOSKRNL_Emu Backport drivers/ create whatever patch files from above. Install nLite 1.4.9.3 extract Windows ISO to folder run nlite, customize as you want, make sure to import the update packs once you get to the create ISO stage stop ==============ACPI/NTOSKRNL/STORPORT/WIN8AHCI/WIN7NVME/WIN8USB3======================== open nlited ISO folder and extract SP3.cab (XP), SP2.cab (XP64) to a folder replace acpi.sys, add storport.sys, storpor8.sys, ntoskrn8.sys to folder use CabPack1.4 to recreate SP#.cab and replace cab file from ISO folder replace ACPI.SY_, add STORPORT.SY_, STORPOR8.SY_, NTOSKRN8.SY_ to I386 (XP), AMD64 (XP64) folder use cabsdk to compress eg: "CABARC -m LZX:21 N ACPI.SY_ ACPI.SYS" ------------------------------------------------ (I386 folder for XP) (AMD64 folder for XP64) in DOSNET.INF add (x64 already has a storport.sys line) in TXTSETUP.SIF add (x64 already has a storport.sys line) in HIVESYS.INF add for SETUPREG.HIV use offlinereg-win32.exe by Erwan to add REPLACE copy to AMD64 folder or I386 folder ================================================================= =================Other Patches===================================== replace SHELL32.DL_ patched version to AMD64 or I386 folder replace TCPIP.DL_ and TCPIP6.DL_ patched version to AMD64 or I386 folder for x86 replace sndrec32,disk,partmgr, and other files from AVX/PAE patch to the I386 folder or the SP#.cab as needed etc... =================Make ISO====================================== make ISO in nLite and test.

Patches shell32 missing icon notification patch (x86/x64) tcpip & tcpip6 max connection limit patch (x86/x64) WinXPPAE v3.5 patch not needed, XP64 is good up to 128GB of ram. AVX/AVX2 patch (only x86 version exists) GPT native support on data volume. With XP32 you have to copy disk.sys and partmgr.sys from server2003x86, but with XP64 its based on 5.2 (like 2003x64). sndrec32 2GB patch (only needed for x86, x64 does not need patch) Patch for applications 4GB patcher to add /LARGEADDRESSAWARE flag for EXEs. Otherwise 32bit process is limited to 2GB. This would allow it to use all 4GB. Useful for some games or apps that can use a lot of RAM. https://ntcore.com/?page_id=371

XP64 SP2 + post SP2 update pack Gigabyte B450 Aorus M F52 BIOS AGESA 1.0.0.6 (original F41 AGESA 1.0.0.3 also worked), (booted to XP64 on a different PC with F62d, so maybe Zen3 as well?) Ryzen Zen2 Matisse CPU Issues Restart does not work with XP/XP64 with any of the acpis when using a PS2 keyboard. Have to manually push the restart button. Shutdown works fine. Workaround, use USB keyboard after it is installed (or integrate backported USB3 driver) BIOS Gigabyte B450 Aorus M BIOS settings So I noticed that by default virtualization (AMD-V) aka "SVM" was disabled by default. BIOS > MIT > Advanced Freq Settings > Advanced CPU Settings > SVM Mode Enabled & CoolNQuiet Enabled BIOS > MIT > XMP Profile 1 BIOS > BIOS > CSM Support Enabled & Fastboot Disabled & Full Logo Disabled You can also enable a PPT limit on the CPU lower than default. Lose a little bit of performance for a nice drop in heat. BIOS > Peripherals > AMD Overclocking > PBO to Enabled BIOS > Peripherals > AMD CBS > Package Power Limit > Manual & ## Picture shows a nice spot for power/performace around 50W ACPI ACPI needed for new motherboards. First versions were based on longhorn 5048 with some BSODs patched. There was a x86 version based on the original XP acpi with new instructions patched in assembly as well (6666). Latest version for XP/XP64 (7777) is here, https://www.mediafire.com/file/wsqgptapdrxhidf/ACPI2.0_v4_x86+x64_5.1+5.2.7z/file Drivers GPU Drivers (Tested nvidia GT710, GTX750Ti, and amd HD5450 video cards) NVIDIA 368.81 7/10/2016 AMD 14.4 09/05/2014 Onboard LAN/Audio drivers Realtek Audio 5.10.0.7514 8/23/2016 Realtek LAN 5.836.125.2018 1/25/2018 AMD CPU CoolNQuiet Driver 1.3.2.53 8/22/2007 Download AMD Processor Driver 1.3.2.53 and turn on power management if you want CPU to clock down during idle. Chipset Drivers (works with native storport on XP64, for XP you will need to integrate backported storport from Advanced) DEV7901/43C8 FCH/400 Series SATA Controller Option1: AMD 1.2.1.321 Download an old 13.4 XP/XP64 chipset pack from AMD. Open up amd_sata.inf and add in these lines underneath existing DEVs (a non B450 may have different DEV_): %AMDSATA.DeviceDesc% = amd_sata_inst, PCI\VEN_1022&DEV_7901&CC_0106 %AMDSATA.DeviceDesc% = amd_sata_inst, PCI\VEN_1022&DEV_7904&CC_0106 %AMDSATA.DeviceDesc% = amd_sata_inst, PCI\VEN_1022&DEV_43C8&CC_0106[/spoiler][spoiler=DEV7901/43C8 FCH/400 Series SATA Controller Option2: StorAhci_V1.0_20200928]Download Schtrom's StorAhci SATA AHCI Driver from sourceforge. Pros: Universal driver. Not limited to just AMD controllers. Cons: Sequential speed is slower compared to AMD 1.2.1.321 Needs SMART patch on backported win7 storport to enable SMART. DEV7901/43C8 FCH/400 Series SATA Controller Option2: StorAhci_V1.0_20200928 Download Schtrom's StorAhci SATA AHCI Driver from sourceforge. Pros: Universal driver. Not limited to just AMD controllers. Cons: Sequential speed is slower compared to AMD 1.2.1.321 Needs SMART patch on backported win7 storport to enable SMART. NVMe OFA NVMe 1.3 Storport Miniport Option1: 20180719 Download Schtrom's OFA NVMe 1.3 Storport Miniport from sourceforge. You will need to integrate a new version of storport.sys for it to work. Version 4173 (KB943545) or newer. You can also use the postSP2 update pack. Tested with a Samsung PM981 and storport 5.2.3790.4485. -Version 1.3 works. Version 1.5 stalled in the TXT setup just before format partition page. -Use MBR for boot drive. Native XP/XP64 does not boot from GPT. -Use NTFS. FAT32 starts install but hangs in the GUI setup halfway through Installing Devices. -No SMART data from CrystalDiskInfo, SMART with HWiNFO. DEV43D5 400 Series USB 3.1 XHCI Controller 1.0.5.3 Download the Windows 7 chipset pack for the B450 from AMD. 1.0.5.3 USB driver from USB31_PT for the controller and the hub works with the motherboard USB ports. But you get a Code 39 error. To fix, you will have to copy usbd.sys to your \system32\drivers folder for a live system. You can also modify the inf file in two spots to also copy over the usbd.sys file when it copies over the driver sys file like the amdxhc31.sys or amdhub31.sys. Use can decompress the usbd.sy_ file from your ISO or extract latest version from an unofficial service pack. In CMD, "expand usbd.sy_" Chipset Drivers (advanced) NTOSKRNL Emu_Extender required -Download NTOSKRNL Emu_Extender from sourceforge. https://github.com/MovAX0xDEAD/NTOSKRNL_Emu -Follow the instructions and compile backported drivers. THe USB driver needs Kernel Mode Driver Framework 1.11 (WDF 1.11). -Follow the instructions from above and create updated wdf files. Useful Tools: -Hexplorer or equivalent hex editor -Dependency Walker -PEChecksum.exe DEV7901/43C8 FCH/400 Series SATA Controller Option3: AMD 1.2.1.402 -Download the Windows 7 chipset pack for the B450 from AMD. -Open up amd_sata.inf and change the two lines that have NTAMD64.6.1 -> NTAMD64 (NTx86 if doing the 32bit version) -Add in these lines underneath existing DEVs if not already there (non B450 board may have different DEV_): %AMDSATA.DeviceDesc% = amd_sata_inst, PCI\VEN_1022&DEV_7901&CC_0106 %AMDSATA.DeviceDesc% = amd_sata_inst, PCI\VEN_1022&DEV_7904&CC_0106 %AMDSATA.DeviceDesc% = amd_sata_inst, PCI\VEN_1022&DEV_43C8&CC_0106 -Open up amd_xata.sys with hexplorer or equivalent, find and change ntoskrnl.exe -> ntoskrn8.sys. -Apply PEChecksum.exe on amd_xata.sys DEV149C Matisse USB 3.0 Host Controller usbxhci 6.2.9200.xxxxx Backported generic Win 8 USB3 driver from NTOSKRNL_Emu works. Needs WDF1.11 backport as well. NVMe Generic Windows 7 NVMe Option2: stornvme 6.1.7601.23403 Backporting Windows 7 NVMe drive works. No SMART info. NVME Samsung Option3: 3.3.0.2003 Backport the Samsung NVME driver. secnvme.sys and secnvme.sys imports need to be changed to NTOSKRN8.SYS like with the other drivers. No SMART with CrystalDiskInfo but it does work with HWiNFO. I only have a Samsung NVME so havent tested if the generic works with other brands. Chipset Drivers (null) DEV790B FCH SMBus Controller Null driver. Download an old 13.4 XP/XP64 chipset pack from AMD or get Windows 7 for the B450 chipset. Open up SMBUSamd.inf and add in these two lines underneath existing DEVs: %AMD.DeviceDesc0% = AMDSMBus, PCI\VEN_1022&DEV_790B %AMD.DeviceDesc0% = AMDSMBus64, PCI\VEN_1022&DEV_790B DEV1481 Starship/Matisse IOMMU Null driver. Download an old 13.4 XP/XP64 chipset pack from AMD or get Windows 7 for the B450 chipset. Open up AmdIOV.inf and add in this line underneath existing DEVs: %IOMMU.DeviceDesc%=NULL_DRIVER,PCI\VEN_1022&DEV_1481 DEV1486 Starship/Matisse Cryptographic Coprocessor PSPCPP Install null driver to remove from missing devices. DEV1485/148A Starship/Matisse Reserved SPP/PCIe Dummy Function According to the readme this should also be a null driver. Unlike the SMBus and IOMMU, this one has a sys file that tries to install. Update inf to add DEV1485 and DEV148A. ACPI\AMDI0030 Install null driver to remove from missing devices Extra Stuff VMware and USB flash drive passthrough When using VMware, if you connect a USB flash drive and try to pass it through to the guest OS, you will need to use a USB port that is using the generic backported USB driver in the host OS. It does not work then using the AMD USB31_PT 1.0.5.3 chipset drivers from the motherboard. From the article it looks like other motherboard specific drivers will have the same issue with a few exceptions. Once you get the Win8 USB3 up and running, you can replace the AMD motherboard USB drivers. https://kb.vmware.com/s/article/2041591 Tested with: XP64host with VM10.0.7(last XP version) and Win7x64host with VM10.0.7 & 15.5.6(last Win7 version). Testing passthrough with a Kingston USB2.0 flash drive works fine. But when connecting a Kingston USB3.0 flash drive vmware gets a "unable to connect to ideal host", and the drive does not load properly in the WinXP guest OS. VMware and USB3.0 flash drive If you change the USB compatibility in vmware in Settings from 2.0 to 3.0, then install the backported generic drivers to the XP guest OS, it will work properly. Note that the vm image hardware compatibility mode has to be 8 or higher, if you are using an image set to 6.5-7.x mode or lower then the USB 3.0 drop down menu will not show up.