postmarketOS at FOSDEM 2026 + Hackathon
29 points by achill
29 points by achill
I want so bad to use linux on a phone. I want it enough to suffer the laggy and sometimes buggy interfaces, but one thing they need to iron out is the basic function stability. It just is a total non-starter that a lot of the mobile linux OSes suggest that "calls may fail" and "SMS may not get delivered". That is the very bottom baseline I would need for any phone OS.
Unfortunately this is one of the hardest issues. I’m giving a shot at getting my Pixel 6a working and there’s just so much stuff that you need to get working first (and Google’s one of the best when it comes to mainlining stuff, because they have a 6.18 kernel for it!). Off the top of my head, for cellular:
And sure there are the android kernels (again you need to be lucky that the company actually fulfills their legal obligations here…) but those very very often just completely skip the normal Linux subsystems, so it’s not trivial to use them for upstream work. A lot of stuff in android is running on user space, too, and for that you’re really back to reverse engineering since they won’t be publishing sources for it at all.
And this is just device bring up, then you need to get all the stuff around cellular working which is different for seemingly every carrier, all this without a data sheet or official support from the carriers (and you’re lucky if they don’t have a whitelist for VoLTE!), and it’s so much different from regular Linux development that I won’t even begin to pretend to understand any of it.
Planning to work around issues with calling and SMS on Linux-based mobile operating systems is what led me to port my numbers to JMP. Now the only hurdle is mobile data, but I don't have to worry about things like VoLTE. Granted, the Jabber client situation isn't as good as on Android yet. Work is being done to improve that with projects like Dino and the Borogove SDK.
(Disclaimer: I now work with the co-op that runs JMP.)