After a couple weeks of daily-driving Sailfish OS on the new Jolla Phone 2 (2026), here's my take on the European, privacy-first alternative to the Android and iOS duopoly. Design, hardware, camera...
The kernel depends on the hardware. Trying to get hardware with mainline kernel support is hard and expensive. Mainlining isn’t the trivial task you make it out to be, otherwise postmarket would not struggle that much. Or tell me which device just works with postmarket OS?
Sailfish OS can work with mainline Linux kernels, it has been ported to the Pine Phone for example.
It is not trivial, it can be. The company designing the phone can choose components with mainline kernel support over others.
If you choose an SoC with decent support and, for example, a vibration module or modem with decent mainline support, you just have to write a device tree.
If you build a phone, you know what regulators exist, what voltages they have and where they connect.
It’s not the type of reverse engineering I have to do myself, the phone company knows their own device.
And even if they task another company to do it, they can just write in their contract “provide us with the following” and they get what they need.
It’s a lot easier for a phone company than for people doing it in their spare time by reverse engineering things.
Yes, you can choose components and then you get something like the Pinephone. Do I really need to go on why there are substantial downsides connected to that?dann können Of course, if you sell 10 mio phones things get easier but that is ou of reach for any such project by a long shot.
Or which mainline Linux phone with acceptable hardware am I unaware of?
Oh, there are many acceptable phones, as people use Linux phones and this community also exists. There are many Linux phone users out there that accept a phone and choose to use it.
No need to use a Pinephone, we have more modern SoC that are well supported, such as SM6350, SDM845, SDM670, SC7280…
EDIT: Also, what I forgot to mention is that you really don’t need to sell 10 million phones to build usable and decent hardware. See for example Fairphone, they have sold less than a million.
Name me a concrete phone model that has no significant compromises in hardware functionality. Maybe I missed those when I had a look into it.
The Postmarket wiki only shows a handful of devices where Call functionality is classified with “Yes”, rather than “Partial” or even “No”. Other than Pine Phone and Purism Librem that is only the Shift6mq from 2020 and four Xiaomi devices with the newest one being also from 2020, all of those have core functionality that does not fully work. Maybe that wiki is horrendously out of date but then this doesn’t do good service to give newcomers a good idea where the project stands.
BTW: I am not saying btw that postmarket isn’t interesting nonetheless to many people and a great project.
“more urgent tasks”
Jolla could just mainline and open source and they’d get more developer power, the devs just prefer devices that are more open
Also, I’m mainlining stuff myself, without being a phone company. Surely they’d have a much easier time.
The kernel depends on the hardware. Trying to get hardware with mainline kernel support is hard and expensive. Mainlining isn’t the trivial task you make it out to be, otherwise postmarket would not struggle that much. Or tell me which device just works with postmarket OS?
Sailfish OS can work with mainline Linux kernels, it has been ported to the Pine Phone for example.
It is not trivial, it can be. The company designing the phone can choose components with mainline kernel support over others.
If you choose an SoC with decent support and, for example, a vibration module or modem with decent mainline support, you just have to write a device tree.
If you build a phone, you know what regulators exist, what voltages they have and where they connect.
It’s not the type of reverse engineering I have to do myself, the phone company knows their own device.
And even if they task another company to do it, they can just write in their contract “provide us with the following” and they get what they need.
It’s a lot easier for a phone company than for people doing it in their spare time by reverse engineering things.
Yes, you can choose components and then you get something like the Pinephone. Do I really need to go on why there are substantial downsides connected to that?dann können Of course, if you sell 10 mio phones things get easier but that is ou of reach for any such project by a long shot.
Or which mainline Linux phone with acceptable hardware am I unaware of?
Oh, there are many acceptable phones, as people use Linux phones and this community also exists. There are many Linux phone users out there that accept a phone and choose to use it.
No need to use a Pinephone, we have more modern SoC that are well supported, such as SM6350, SDM845, SDM670, SC7280…
EDIT: Also, what I forgot to mention is that you really don’t need to sell 10 million phones to build usable and decent hardware. See for example Fairphone, they have sold less than a million.
Name me a concrete phone model that has no significant compromises in hardware functionality. Maybe I missed those when I had a look into it.
The Postmarket wiki only shows a handful of devices where Call functionality is classified with “Yes”, rather than “Partial” or even “No”. Other than Pine Phone and Purism Librem that is only the Shift6mq from 2020 and four Xiaomi devices with the newest one being also from 2020, all of those have core functionality that does not fully work. Maybe that wiki is horrendously out of date but then this doesn’t do good service to give newcomers a good idea where the project stands.
BTW: I am not saying btw that postmarket isn’t interesting nonetheless to many people and a great project.