reboot12
Content Type
Profiles
Forums
Events
Posts posted by reboot12
-
-
4 minutes ago, Dietmar said:
I still do not understand,
why now you get the latest Source files from csmwrap.efi .
Is it, because you connect now to another server via
No, they probably already corrected the repository and now git clones the latest version :-)
0 -
OK, git corrected and now downloads the current version
This is full tutorial how compile CSMWrap:
Download and install Ubuntu: https://releases.ubuntu.com/noble/ubuntu-24.04.2-live-server-amd64.iso
During installation, select Install OpenSSH server - this makes it easy to copy files from Ubuntu to Windows using WinSCP and use Ubuntu over SSH e.g. in Putty
sudo passwd root (enter password for user)
sudo passwd root (make new password for root)
sudo login root or reboot and login using root
nano /etc/ssh/sshd_configPermitRootLogin yes
PasswordAuthentication yesCtrl+o to save and Ctrl+x to exit
service ssh restart
apt update
apt install -y build-essential nasm curl xxd net-toolsgit clone https://github.com/FlyGoat/CSMWrap.git --recursive
cd CSMWrap
make seabios
makeifconfig - to check IP
using WinSCP make connection and connect to Ubuntu:New Site, File protocol: SCP, Host name: IP, root User name and Password, Save
0 -
No, I tried like this but errors:
wget https://github.com/FlyGoat/CSMWrap/archive/refs/tags/1.2.1.tar.gz tar -xvf 1.2.1.tar.gz mv CSMWrap-1.2.1 csmwrap cd csmwrap root@ubuntu:~/csmwrap# make seabios make -C seabios distclean make[1]: Entering directory '/root/csmwrap/seabios' make[1]: *** No rule to make target 'distclean'. Stop. make[1]: Leaving directory '/root/csmwrap/seabios' make: *** [GNUmakefile:271: seabios] Error 2 root@ubuntu:~/csmwrap# make make: *** No rule to make target 'nyu-efi/src/elf_x86_64_efi.lds', needed by 'bin-x86_64/csmwrap'. Stop.
F..k, but a mess
0 -
OK, but this compile version 1.1.0-12
0 -
OK, finally works
I was running all the time with CSMWrap and it turned out that with this it doesn't work on Haswell:
https://forums.mydigitallife.net/threads/winxppae-v3-5-winxp-sp2-issue.89363/#post-1881947Also, the ntkl64g.exe file has a bad checksum and doesn't work when trying to run using winload.exe - error 221:
setcsum ntkl64g.exe SetCSUM v1.01, (c)2000 Jeremy Collake http://www.collakesoftware.com mailto:collake@charter.net ---------------------------------------------------------------------- .\ntkl64g.exe = Header Checksum: 001FACDE Correct Checksum: 001FF99B Checksums do not match! Set correct checksum in file? [Y/N/Q] Y Recompute Header Checksum: 001FF99B Correct Checksum: 001FF99B Checksum set! Done!0 -
Which patch for WinXP SP2 ? I tried WinXPPAE_v35.7z but patch only ntkrnl*.exe
Search for differences 1. C:\WINDOWS\system32\ntkrnlpa.exe: 2 015 232 bytes 2. C:\WINDOWS\system32\ntkrnl2.exe: 2 015 232 bytes Offsets: hexadec. 138: D9 23 139: 22 1E 15BF1B: 07 00 15BF20: 01 02 160F0E: 74 EB 1B07CC: 10 00 1B07CD: 00 02 1B07E2: 1B 00 1B07EA: FC F8 1B07EE: 01 02 10 difference(s) found.
I try use patched files from Dibya128GB -> ntkrnlpa.exe + hal.dll but OS not boot and only black screen.
0 -
14 hours ago, NewcomerAl said:
For uefiseven I suggest to rename "csmwrap.efi" to "u7.original.efi" for chainload.
Or write a shell script.No, chainload not work:
UefiMain Found Windows Boot Manager at '\u7.original.efi' Launch Loaded 'PciRoot(0x0)/Pci(0x1F,0x2)/Sata(0x4,0xFFFF,0x0)/HD(1,MBR,0xD5AED5AE,0x800,0x32000)/\u7.original.efi' Launch Addresss behind FileImageHandle=C1B43718 Launch File does not match an EFI loader signature
but I fix and compile special version U7 no load Windows Boot Manager
- use verbose=1 in UefiSeven.ini to display this info on screen:
but CSMWrap replace C0000 memory using VBIOS PCIe or SeaVGABIOS. I wonder if you compile a special version of CSMWrap that doesn't change anything in memory under C0000 and if Windows would boot into the vga.sys driver ?
https://www.mediafire.com/file/o1mp9mbx11iyudm/u7nwbm.efi/file
1 -
3 hours ago, Dietmar said:
Do you have an idea, how to come back to the UEFI shell,
after using u7.efi
Simply not use u7.original.efi file - after u7 start set resolution and put fake int10 in C0000 memory but display error than not found u7.original.efi and after press Enter go back to shell:
FileExists Unable to open file '\u7.original.efi' for reading (error: Not Found) UefiMain Could not find Windows Boot Manager at '\u7.original.efi' UefiMain Press Enter to continue.
https://github.com/FlyGoat/CSMWrap/issues/37#issuecomment-2961137267
0 -
For new iGPU no drivers for xp then this is best way to use xp on modern desktop compi (for Haswell I have xp drivers so that's why I use the Multi-Monitor option)
The iGPU cannot be set as the Primary Display in the bios because then CSMWrap will load SeaVGABIOS and you can't install a graphics driver on it
0 -
To display the image on the monitor WinXP does not care what is in memory under C0000 if the graphics card manufacturer's driver is installed.
- Asus B85 with two graphics cards: iGPU Haswell (has legacy & GOP but this is not important), PCIe AMD6450 (has both legacy & GOP and this is very important)
- in bios settings CSM Disabled. Primary Display PCIE because CSMWrap can use VBIOS this card for xp32 on vga.sys if use ntldr loader. Multi-monitor Enabled for xp32 see both in Device Manager
What I did:
- boot PC in pure UEFI 64-bit to UEFI Shell > run CSMWrap 64-bit 1.2.1 > SeaBIOS boot xp from HDD using ntldr
- install AMD graphics driver
- reboot
- install Haswell graphics driver
- reboot
- now possible use any loader ntldr or bootmgr to boot xp - in xp display settings I set Primary display iGPU to make this pictures:
I make boot.bmp 640x480 4-bit logo and put in WINDOWS folder, then add to boot.ini switch /bootlogo /noguiboot - thank to this on screen stay this logo after xp boot if you use ntldr loader.
If the bootmgr loader is used, the aurora logo from winload.exe is left on the screen.
On both card I have 3D acceleration :-) and possible change primary monitor or enable extend desktop. In C0000 memory is always VBIOS from PCIe card
1 -
Is there really no solution on github?
Timeline cannot be loaded The timeline is currently unavailable due to a system error. Try reloading the page. Contact support if the problem persists.
The error occurs when I try to address a specific user by typing the @ character.
0 -
I tested the files on Haswell with iGPU, but they absloutely do not change anything compared to the original files.
Also try build PCIOpRom then load in UEFI Shell but command loadpcirom load only UEFI part (GOP) from file igpu.rom:
efirom -f 0x8086 -i 0x0412 -b vbios.rom -ec gop.efi -r 0x0000 --pci23 -l 0x030000 -o igpu.rom
I tried also boot MS-DOS 6.22 from USB, then using rambios.com load iGPU vbios.rom but screen make black then PC hangs:
rambios /i vbios.rom
0 -
On 6/7/2025 at 2:02 AM, Dietmar said:
After soso much crazy tests and hacks I succeed with the Nvidia GT 730 graphik card for to boot XP SP3 with full 32 bit 3D graphik
on the Arrow Lake Gigabyte board B860 DS3H under pure Uefi, means no csm at all.
Write what hacks you applied to boot xp on this motherboard with a PCIe GT 730 graphics card.
0 -
I think that rom builded using efirom can work and CSMWrap will be able to use the legacy part from the loaded rom file. In uefi bios legacy VBIOS iGPU is in RAW area in CSMCORE module and CSMWrap not know where. But if you use loadpcirom legacy+uefi.rom your graphics card then maybe CSMwrap see legacy part using PCI I/O protocol same as use legacy rom from PCIe graphics card.
0 -
No, efirom.exe is Windows CMD executable:
https://www.mediafire.com/file/luqx9gx86u9monv/EfiRom.zip/fileAfter you build rom then posiible load it in UEFI Shell using command loadpcirom
0 -
Yea, possible load pci rom builded using efirom:
UEFI Shell see efi rom but I don't know id CSMwrap see legacy part of rom
0 -
42 minutes ago, Dietmar said:
PS: Is there a possibility, at an UEFI shell,
to load the Legacy VGA Option Rom from that connected graphik card to C0000 in memory?
RAW OpROM (55 AA beginning) probably not but maybe after make efi rom image using efirom.exe tool possible - look at this:
https://msfn.org/board/topic/178182-xp-pro-x64-os-boot-nvme/page/3/#comment-1227538
On 10/18/2022 at 7:16 PM, reboot12 said:I built a new ROM (legacy+UEFI) using efirom.exe for my Samsung 970 EVO (144D A808):
efirom -f 0x144d -i 0xa808 -b legacya808.bin -ec ext.efi -r 0x0000 --pci23 -l 0x000002 -o image_A808.rom0 -
I test, test and test WinXP 32-bit SP2 with CSMWrap 1.2.1 on Asus B85M-E
Config:
- iGPU Haswell, monitor connected over HDMI-HDMI cable (has both legacy and GOP VBIOS)
- PCIe AMD6450 in second PCIe slot x16 work at x4, same monitor connected over DP-HDMI (second HDMI port on monitor) (has both legacy on PCIe card chip and GOP VBIOS inserted to motherboard bios)
- Monitor THOMSON 1360x768 native resolution with 3 HDMI port input
XP boot OK on vga.sys :
-
on Haswell need set Primary Display iGPU in bios, work on bootmgr+BCD+winload.exe 5472 > boot on SeaVGABIOS from CSMWrap:
XP desktop in 1024x768 32-bit on full monitor area if boot PC from bios Boot Menu then select HDD
XP desktop in 640x480 32-bit on top left corner monitor area if PC auto boot from HDD
not possible install Intel Haswell graphics driver - no free resources (yes there are Haswell iGPU drivers for XP and they work in normal legacy mode) -
on AMD need set Primary Display PCIE in bios, work on ntldr > boot on legacy VBIOS AMD6450
XP desktop in 1024x768 32-bit on full monitor area
possible install AMD drivers and works 3D
If I swap loaders:
- ntldr for Haswell iGPU then black screen and xp hangs but if set VGA Enable 1 before run CSMWrap (mm 0001003e 8 -pci) then boot OK but black screen
- winload.exe for PCIe AMD then stay on aurora boot screen but boot OK
Running bios Boot Menu set GOP to 1024x768 then xp works in same resolution. Instead Boot Menu possible use u7.efi 1.31 with UefiSeven.ini set to 1024x768 without u7.original.efi to make error in U7 then fallback to UEFI Shell - I use startup.nsh
@echo -off if exist fs0:\EFI\Boot\121.efi then set base "fs0:" endif if exist fs1:\EFI\Boot\121.efi then set base "fs1:" endif if exist fs2:\EFI\Boot\121.efi then set base "fs2:" endif if exist fs3:\EFI\Boot\121.efi then set base "fs3:" endif if exist fs4:\EFI\Boot\121.efi then set base "fs4:" endif if exist fs5:\EFI\Boot\121.efi then set base "fs5:" endif if exist fs6:\EFI\Boot\121.efi then set base "fs6:" endif if exist fs7:\EFI\Boot\121.efi then set base "fs7:" endif if exist fs8:\EFI\Boot\121.efi then set base "fs8:" endif if exist fs9:\EFI\Boot\121.efi then set base "fs9:" endif if exist fsa:\EFI\Boot\121.efi then set base "fsa:" endif if exist fsb:\EFI\Boot\121.efi then set base "fsb:" endif %base% u7.efi EFI\Boot\121.efi
0 -
2 hours ago, Dietmar said:
The Bios or XP notice (!), if there is a monitor connected to that card in Slot1 or not.
Try change in bios Primary Display from Auto to PCIE if you use PCIe graphics card. If you have second monitor connected to iGPU disconnect it.
0 -
1 hour ago, Dietmar said:
It has an Legacy VGA Option Rom.
With my post from above I only want you to know, how I succeed with full 32 bit graphik on the Arrow Lake board.
OK, but before you write on the forum about something new that you discovered, first write back to the previous post because it looks like you are ignoring which is very annoying
0 -
0
-
31 minutes ago, Dietmar said:
I just check again: The GT 730 does not have an Legacy VGA Option Rom.
I don't believe it. I asked you to take a screenshot in xp in SlicTool Kit (slic.exe):
https://www.mediafire.com/file/zuxxznjxjjnn3u1/memory_check.zip/fileAt the moment I'm not interested in whether the card works on ArrowLake with CSMwrap just whether it actually has VBIOS - connect card to other old motherboard run xp in legacy mode and please use slic tool kit.
0 -
16 minutes ago, Dietmar said:
The AMD has Legacy VGA Option Rom, the Nvidia GT 730 not.
It is impossible that such an old graphics card GT 730 does not have a Legacy vbios, after all, you booted the computer with this card in Legacy mode using CSMWrap - right?
https://msfn.org/board/topic/186793-csmwrap-boot-csm-on-uefi-only-systems/page/10/#findComment-1280463This is proof that it has vbios - show what the SLIC ToolKit memory dump at C0000 shows:
https://msfn.org/board/topic/186793-csmwrap-boot-csm-on-uefi-only-systems/page/9/#findComment-12804020 -
I noticed an interesting thing - I did the xp32 tests with CSMWrap described in the last few posts on the original xp sp2 files including acpi.sys and there were no problems with acpi. Today I wanted to test something, I set CSM Enabled in the bios, I run the system without CSMWrap and there is BSOD A5.
I replaced the acpi.sys outside of the running system with the ones from you - modded acpi.sys 5.2 and the system boots fine - but it seems to be related to the fact that I use winload.exe and it has a problem with the original acpi.sys - I seem to have written about it before (in 2019 and 2020) on BA about xp 64-bit + winload.exe:
https://www.betaarchive.com/forum/viewtopic.php?p=445453&sid=c0c23f5574a1881384f2de47f5f59987#p445453Very interesting that with CSMWrap work ok original acpi.sys + winload.exe
0
CSMwrap - boot CSM on UEFI only systems.
in Windows XP
Posted · Edited by reboot12
I compile CSMWrap replacing SeaVGABIOS to ValleyView Legacy video BIOS extracted from Jetway HPC070SC industrial computer:
https://github.com/FlyGoat/CSMWrap/issues/37#issuecomment-2921265627
Risky to use on other Laptops but this is Mobile/Desktop version so should work on desktop computers:
https://files.catbox.moe/6xj4hm.zip
User win98se on github test this version, "But the graphics are glitched, and after spinning for a while" but not damaged:
https://github.com/FlyGoat/CSMWrap/issues/39#issuecomment-2977152878