I've done my research on the topic and am indeed aware that Realtek SoCs in general aren't supported by OpenWrt for a myriad of reasons, and that this is a fool's errand and I should definitely not do it because it'll be a massive headache and not worth it at all.
However, I purchased a D-Link DIR-842 recently without being informed it was the R1 model, and frankly can't be bothered to return a router for the third time because the specific hardware revision is incompatible with OpenWrt. The fact that others have made at least some decent headway into supporting 819x devices also gives me hope that, with some head-scratching and a lot of work, I could get at least the basics working on this.
The stock firmware allows telnet access for local addresses (from which I can theoretically access the entire filesystem), and also displays the unfiltered system logs, which would certainly help with figuring out how to make the device work.
I suppose my first question on this topic is how I would go about starting it, which I do still intend to do despite knowing how bad an idea it is. I have at least some experience in software and can write a decent amount of code, though I'm not particularly familiar with OpenWrt development due to only having used it for a few months.
Is the written offer for the open source code packaged? First step is to ask your manufacturer for exactly that.
The good news
- rtl8197fs is 'normal' mips 24Kc and not lexra
so at least the fundamental toolchain and kernel-arch side are covered
The bad news:
- 8/64 is formally below minimum system requirements for OpenWrt beyond 24.10.x (and new development only goes to main), unless there are devices using this SOC on the market (and in sight of getting OpenWrt support alongside), it would be a very hard sell for OpenWrt to add a new target whose only supported device does not meet minimum system requirements.
- adding a new target is several notches harder than 'just' adding new device support for an already supported target, you not only have to get your device 'work' somehow, but the required SOC/ driver patches will have to go to the mainline kernel in parallel (so you need to get them good enough to pass that bar)
- getting it ported will be your hobby for the next months++ (double digit+), don't expect it to be easy
- you would be starting pretty much from zero, as there is nothing for this SOC already present in OpenWrt and the wireless side has very little real-world testing in AP mode
The low-spec issue (flash/ RAM) is a quite major deterrent to get others motivated to help you in your endeavour, modern Realtek PCIe wireless drivers (rtw88/ rtw89) are working quite well (only tested in STA mode, but AP mode another can of worms to consider) though.
If you want to retain your sanity - or if you have any questions, you should go with a fully supported device (exceeding minimum system requirements) instead, it will be cheaper than starting this project. If you want a real challenge and know what you're doing, have a bite - it's your time.
--
very personally speaking, I wouldn't even consider starting the development for a completely new target unless my device had at the very least 32/256 (with 128/512 devices around as well).
Disclaimer: I'm not an OpenWrt developer and can't speak for the project, the opinions expressed above are mine alone and not coordinated with anyone else.
I did completely miss the lower memory and storage capacity of the device, which makes it even worse that there was absolutely zero information on the device version. With that compounding the already very difficult task of adding a whole new target, I suppose it's a better choice to not do this and return the device. I'd have at least tried to go through the trouble of it if it could run the current version of OpenWrt, but that definitely just ruins any chance it might've had.
This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.