Windows 10 port for Allwinner A20 (Banana Pi P2) — feasibility / what would it take?

Question about which tools to use, bugs, the best way to implement a function, etc should go here. Don't forget to see if your question is answered in the wiki first! When in doubt post here.
Post Reply
Sesggl613
Posts: 1
Joined: Sun Aug 23, 2026 2:43 pm
Libera.chat IRC: Hi

Windows 10 port for Allwinner A20 (Banana Pi P2) — feasibility / what would it take?

Post by Sesggl613 »

I understand the major blockers going in:

• A20 is ARMv7, and Windows 10 ARM builds are 64-bit only
• No UEFI implementation exists for A20 (Windows requires UEFI to boot)
• No Windows driver stack exists for A20’s GPU/storage/wifi

Given those, I’m not expecting this to be simple or already done — but I’m curious:

• Has anyone in the retro-OS or ARM homebrew scene ever attempted this, even partially (e.g., got a Windows 10 ARM kernel to boot without full driver support)?
• Is a custom UEFI port for A20 something that’s been discussed/started anywhere (similar to how RPi got community UEFI)?
• Realistically, what would a from-scratch effort like this take — team size, skills needed, rough timeframe — for anyone who’s done bootloader/driver work on similar boards?

Not looking to be talked out of it — I’ve read the general “it’s not supported” answers already. Looking for anyone with actual firmware/kernel-level experience who can speak to what a real attempt would involve.
nullplan
Member
Member
Posts: 2034
Joined: Wed Aug 30, 2017 8:24 am

Re: Windows 10 port for Allwinner A20 (Banana Pi P2) — feasibility / what would it take?

Post by nullplan »

I don't know why you want to torture yourself with this quest, but unless you have or can procure an ARM32 (even just armv7-a) build of Windows 10, this quest is doomed from the start. This is because Windows is a closed-source operating system, so you don't get to just compile it for whatever new architecture you'd like.

Once you have your experimental armv7-a build of Windows, all you need to do is provide the kernel with the boot environment it expects. I suspect it may take a UEFI emulation of some description. I don't really know, since again, Windows is closed-source and you can't look inside.
Carpe diem!
nexos
Member
Member
Posts: 1090
Joined: Tue Feb 18, 2020 3:29 pm
Libera.chat IRC: nexos

Re: Windows 10 port for Allwinner A20 (Banana Pi P2) — feasibility / what would it take?

Post by nexos »

To be completely honest, such as endeavour would be borderline impossible. The only way this becomes feasible even the slightest is if you used leaked source code (which AFAIK W10 source hasn't been leaked), and even then you'd clearly be breaking the law so the whole thing would have to be kept under lock and key. I guess in theory you decompile windows into assembler, but it would be an absolute monstrous undertaking, akin to writing the entire Linux kernel by yourself. The reason for this is because you'd have to edit most of the lines of assembler in very subtle ways, without even really knowing what the code is doing, and then having to debug such a thing would be a disaster. So while not technically impossible, it's so close to being impossible without a massive team that it might as well be impossible.

The UEFI/driver part would be the easiest part. The WDK is well documented and you could realistically write drivers for the devices by yourself. Developing UEFI firmware for it wouldn't be to difficult if you made use of EDK2 (which I'm not sure if EDK2 still supports Armv7 or not).
"How did you do this?"
"It's very simple — you read the protocol and write the code." - Bill Joy
Projects: NexNix | libnex | nnpkg
vtsman
Posts: 7
Joined: Sat May 24, 2014 1:47 am
Libera.chat IRC: vtsman

Re: Windows 10 port for Allwinner A20 (Banana Pi P2) — feasibility / what would it take?

Post by vtsman »

I'll go against the grain here and say this isn't dead in the water, just because there is a leaked build of Windows 10 for armv7 which people have managed to install on the Surface RT. That said, I feel like this is the sort of project where if you have to ask, you're likely going to struggle early on.
nullplan
Member
Member
Posts: 2034
Joined: Wed Aug 30, 2017 8:24 am

Re: Windows 10 port for Allwinner A20 (Banana Pi P2) — feasibility / what would it take?

Post by nullplan »

vtsman wrote: Mon Aug 24, 2026 10:39 pm just because there is a leaked build of Windows 10 for armv7
Well, there you go. I said that was the requirement and there it is. Now the OP only has to find it and procure it without the FBI banging down their door, and they can get started.

There is one other option that came to me later: Emulation. If all you care about is running Windows 10 on your board, getting qemu-system-aarch64 to run on bare metal may be easier, and then you can run the official Windows 10 in that QEMU. Or alternatively, you can write your own emulator. Have fun! Emulating aarch64 on arm32 is probably not going to be performant, but this strategy will work if the emulation is good enough.
Carpe diem!
nexos
Member
Member
Posts: 1090
Joined: Tue Feb 18, 2020 3:29 pm
Libera.chat IRC: nexos

Re: Windows 10 port for Allwinner A20 (Banana Pi P2) — feasibility / what would it take?

Post by nexos »

There is one other option that came to me later: Emulation.
That's a actually not a bad idea. I would be afraid of what the performance of Windows 10 would be without hardware acceleration, but considering that I once read that someone got W10 to run on a 386, I imagine it could work
"How did you do this?"
"It's very simple — you read the protocol and write the code." - Bill Joy
Projects: NexNix | libnex | nnpkg
Post Reply