I first wanna begin by saying thank you for the review and questions. This project means the world for me, and I wanna continue working on it for the future.
To your first question, my bugs are reviewed by Claude Sonnet and code is largely ported by me by using Linux drivers and viewing wifi drivers for reference when needed. The process of loading up a qemu process and testing over and over again is tedious and that is why I use claude Sonnet to help me rewrite the bugs that are small and persist that I cannot decipher on my own, I don't know which bugs are fixed by Sonnet but I'll make sure to document that in the future. I try to go by the guidelines as much as possible and I'll try to make it more streamlined and guideline appropriate in the future. The README, comments and commits were written by Sonnet because I was trying to save time, but I guess that wasn't the best thing to do.
Honestly I am not doing this for the grant, I'm doing this because I love NetBSD, I've been a fan since 8th grade basically. I'm sorry for the robotic voice I had in the previous email, I thought a professional voice would be better for my chances but I believe I made a mistake and showed you someone that I am not. Truth is, I'm a broke student trying to work on something I love in a easier way so I can work on it better and perfect what I am doing.
On Thu, 24 Sep 2026 06:11:43 +0200 (CEST)
"Toghrul. M" <toghrul.m%tuta.io@localhost> wrote:
Hi,
I'm developing a NetBSD driver for the Realtek RTL8922AE (rtw89) WiFi 7 adapter, and I'm running into a work problem. I only have one laptop, so I'm forced to develop/debug and test on the same machine. This makes it very hard to isolate driver bugs from environment issues, and it means a bad build can leave me without a working system to keep debugging on.
I'd like to request a grant to cover a dedicated test setup:- A Lenovo ThinkPad T430 to use as a secondary test machine (approx. $199)- An M.2 to mini-PCIe adapter/converter, since the T430 predates M.2 and the RTL8922AE is an M.2 card ( approx. $46)- A second RTL8922AE card to install in the test machine ( approx. $27-40, depending on whether I source from a more or less reputable seller)
This would let me keep a stable development environment on my main laptop while safely using the T430 as needed during testing.
A bit about the project:- Driver target: Realtek RTL8922AE (rtw89), a from-scratch net80211 driver ported from Linux's rtw89 driver, since NetBSD's existing rtwn(4) only goes up to 802.11ac- Current status: attach, BAR/MSI setup, chip identity, and the DMA/H2C command pipeline are all working and hardware-verified via VFIO passthrough. I'm now on firmware download (multi-firmware container parsing is in progress); firmware handshake, RX-to-net80211 wiring, and scan/auth/assoc are still ahead- Repo: https://codeberg.org/ToghrulMammadli/netbsd-rtw89- Estimated total cost: roughly $272-285 (laptop + adapter + card), plus shipping/tax
Let me know if you need more detail on the driver's progress or if there's a better process for hardware grant requests. I can provide receipts or report back on progress as a condition of the grant.
Hi,
Let me open with that this is a commendable project and I would be
happy to personally sponsor it. However I have a major concern:
Can you disclose how much of the work thus far has been assisted by
Claude Sonnet? The README file states intent to eventually upstream the
driver. Looking through the git commits they seem to have more than
trivial AI tool use.
My reason for asking is that it's important to establish that your
project workflow is compatible with the current commit guidelines as
published here: https://www.netbsd.org/developers/commit-guidelines.html
Otherwise it will be difficult to upstream the driver in the end.
Kind regards,
-Tobias