Only Microsoft digital signed bootmgfw.efi can be loaded = part of Secure Boot. Other self-made or modded *.efi will not be loaded by BIOS under Secure Boot enabled.
"FlashBoot Pro has an undocumented feature: sideloading patched ACPI tables from ACPITABL.DAT in EFI system partition, so you can sidestep that hurdle, but after doing that, battery charge indicator in Windows 7 does not work. You will not know the charge state of your battery in Windows 7 on this tablet." Maybe make it documented Also, possible to allow Flashboot to work with custom isos?
Agreed, the trick is slipstreaming the video solution. UEFI7 has been known to work, and the Flashboot post above tells us how to use his solution. One of my earliest attempts was to use a Win 10 installation USB with the Win 7 install.wim. This only solved the problem of Win 7 not wanting to install in UEFI on the Optiplex 7050, but does not solve the video issue. Currently, my video solution has been to use a graphics card with a combined Win 7/8 driver package, but I've been thinking about the alternatives in here. The most popular youtube vids have been freezing up, so I wonder if their coding doesn't play well with my graphics card's 7-year-old driver. Since only their popular vids freeze, I wonder if their website selectively borrows the viewer's system resources?
Thanks so much! But are those Resizable BAR cards a bit pricey? They appear to be more expensive than what I paid for my Optiplex 7050s. Maybe there's a cheaper solution (or I don't need a solution) if I figure out why the popular youtube vids freeze? Instead of my cheap GPU, it might have to do with the rapidly fading browser support for Win 7? I appreciate the forum sensitizing me to the security vulnerabilities of virtualizing. So I won't use the Win 10 virtual system for everything - only for safe things that no longer work in Win 7.
That's another possible answer, my youtube issue might be solved by running a Pi-hole DNS server on my network. Maybe a 32-bit laptop that I no longer have a use for could be my network's Pi-hole server.
I think it's very hardware dependent, for some FP will work, but for others like the Microsoft line, the ACPI tables need to be patched to prevent the A5 BSOD.
Agreed, that would be my guess. I'm assuming there would be more class 2 UEFI, easier to work with, in desktops than in smaller devices. Maybe more firmware GUI support for getting secure-boot keys to make UEFI7 work in a desktop? My video solution was a cheap graphics card with a combined Win 7/8 driver package. I'm guessing the firmware may have been of assistance in that working? Since I've never experienced BSOD in Win 7, I haven't read much about it.
FlashBoot’s tiny VGA BIOS differs from a full BIOS. It implements only INT 10h, supports just one or two video modes (1024×768 and the panel’s native LCD resolution), and never reprograms the hardware. When it was created in 2018, there were no comparable open-source projects to draw from, so I started with VGA mini-BIOS of QEMU’s UEFI firmware (assembly code) and hand-optimized it for size until it fit within 512 bytes — the upper half of the 16-bit interrupt vector table. Unlike other comparable projects, FlashBoot's mini-BIOS does not try to convert a chunk of base memory into read-only mapping, it reuses IVT instead. FlashBoot’s patched Windows 7 UEFI loader then substitutes several immediate constants in this 16-bit code. Windows 7’s VGA miniport driver executes the result inside a built-in kernel-mode VM, so the CPU never actually enters 16-bit mode; it effectively treats the code as bytecode. The remaining magic happens at ExitBootServices(): the UEFI loader patches ntoskrnl.exe and vga.sys of Windows 7 in memory to complete the setup.
Reusing IVT memory space has a benefit: it will not be affected by subsequent physical memory writting on 0000_0000 to 0000_0800, which is the hal.dll's x86LowMemory array in its memory space.