I don't have much of value to add to the conversation, but since it seems that user feedback would help with getting this merged into mainline OpenWRT, I've just flashed it from the stock Firmware's web UI and things appear to be working.
I've pieced the process together reading this thread, and I'm not terribly familiar with OpenWRT, so for people in the same boat who want to try this, you'll want to do something along these lines:
- Make sure you have the required packages installed
- Clone "flashable-1Gbit" the branch of this repo:
git clone -b wax206/flashable-1Gbit https://github.com/boretom/openwrt-fork.git
- In there, run
./scripts/feeds update -a
, followed by./scripts/feeds install -a
- Paste the contents of this post by patient0 into
.config
, then runmake defconfig
followed bymake oldconfig
(*) - Set up the build toolchain by running
make -j "$(nproc)" toolchain/install
- this took several minutes on my 32 core workstation, go grab your favourite hot beverage in the meantime - Finally, you can build the image by running
make -j "$(nproc)"
- you'll find the build output in thebin/targets/mediatek/mt7622
subdirectory - Upload
openwrt-mediatek-mt7622-netgear_wax206-squashfs-factory.img
from that directory using the stock firmware's firmware upgrade dialogue. It'll complain about version info, you can safely ignore that. - After the device has rebooted, connect your laptop/PC to one of the LAN ports (not the WAN port). You'll receive an address in the 192.168.1.0/24 subnet and be able to SSH into OpenWrt (
ssh root@192.168.1.1
). From there, you can follow the OpenWrt docs/wiki, e.g. https://openwrt.org/docs/guide-user/installation/generic.flashing#installation_checklist
(*): I'm not sure if these exact instructions are correct/required in this exact order, but it's what worked for me.
If something goes wrong, patient0 documented a recovery procedure here - but for me, everything went smoothly, just figuring out the process took a bit of effort.
I didn't really get to test anything yet, but judging by others' experiences, I expect no major problems. I wouldn't be opposed to getting one of those multi-speed SFP+ to RJ45 transceivers and help test link speed negotiation for the 1G/2.5G issue, when/if somebody comes up with a solution (I'm far from being an embedded or networking developer so I'm not much use outside of gathering data).