Dima Krasner
2016-08-26 23:11:46 UTC
Hi everyone,
I'm Dima (more info about me at dimakrasner.com) and I'm new here, so ... hi.
For years I've been following the ARM scene, since I'm very interested in libre hardware projects.
Chinese ARM netbooks with WonerMedia chips (8850, 8880, etc') are very common in sites like AliExpress and Ebay and they're pretty darn cheap. Performance, RAM amount limit and decoder performance should be reasonable, compared to other cheap ARM boards.
I see the WM8950 has been evaluated, but the wiki doesn't say what's wrong with it. As far as I can tell it's yet another Mali-400 board.
Upstream kernel has partial support for newer WonderMedia chips, but there is a small community at https://groups.google.com/forum/#!forum/vt8500-wm8505-linux-kernel and as far as I know older ones (like the 8650) should work to some degree with current upstream kernels, some rotten code floating around multiple kernel trees needs to be upstreamed (possibly, after refactoring) and there is an old, public U-boot fork, at least for the 8650.
As far as I see, these are technical ecosystem issues that can be solved by a small group of volunteers with kernel, upstreaming and embedded development experience, even without any vendor support. Therefore, if the 8650 still under mass production, it could be an EOMA68 candidate.
Although I understand while the A20 is preferred to WonderMedia chips (it makes life easier on the software front, probably more than any other non-x86 SoC), I wonder, what makes them unacceptable in RYF hardware? Is there any technical issue that I've missed?
Cheers,
Dima
I'm Dima (more info about me at dimakrasner.com) and I'm new here, so ... hi.
For years I've been following the ARM scene, since I'm very interested in libre hardware projects.
Chinese ARM netbooks with WonerMedia chips (8850, 8880, etc') are very common in sites like AliExpress and Ebay and they're pretty darn cheap. Performance, RAM amount limit and decoder performance should be reasonable, compared to other cheap ARM boards.
I see the WM8950 has been evaluated, but the wiki doesn't say what's wrong with it. As far as I can tell it's yet another Mali-400 board.
Upstream kernel has partial support for newer WonderMedia chips, but there is a small community at https://groups.google.com/forum/#!forum/vt8500-wm8505-linux-kernel and as far as I know older ones (like the 8650) should work to some degree with current upstream kernels, some rotten code floating around multiple kernel trees needs to be upstreamed (possibly, after refactoring) and there is an old, public U-boot fork, at least for the 8650.
As far as I see, these are technical ecosystem issues that can be solved by a small group of volunteers with kernel, upstreaming and embedded development experience, even without any vendor support. Therefore, if the 8650 still under mass production, it could be an EOMA68 candidate.
Although I understand while the A20 is preferred to WonderMedia chips (it makes life easier on the software front, probably more than any other non-x86 SoC), I wonder, what makes them unacceptable in RYF hardware? Is there any technical issue that I've missed?
Cheers,
Dima