State of the mobility SIG, Dec 2023


This will be a bit long, but I hope it will explain the current unfortunate state
of the mobility sig.

First a bit of history. The mobility sig was originally formed back in 2010.
The goal then was “… a group of Fedora contributors that are interested in
Fedora on small devices. Initially aimed at supporting NetBooks, Mobile Internet devices”
A good deal of progress on small devices like netbooks and tablets was made,
but then the members of the SIG moved on to other concerns. Then, in 2020 the
most recent version of the SIG formed. The goal was focused on getting Fedora
working well on the new pine64 pinephone, and then later when announced the new
pine64 pinephone pro.

In the last 3 years we have accomplished a lot:

The phosh stack has been packaged up

We have been producing Fedora Remixes with 3rd party kernels to provide support for pinephone and pinephone pro.

a Phosh spin has been created and officially produced

There was a effort to get a remix working for the oneplus 6, but still requiring a lot of manual tweaking.

Unfortunately, the active members of the sig are not able to move things forward right now.
The pinephone and pinephone pro are both old devices at this point. They were slow
when released, and thats not improved. Worse, upstreaming of patches to get either
of them working with a Fedora/Mainline kernel seems unlikely to us for a lot
of reasons, including that many of the 3rd party patches aren’t in a state that
upstream would accept and many authors of them have no particular desire to
upstream them.

The phosh spin is largely not useful in the current situation as the
Fedora mainline kernel barely boots on the pinephone pro (not at all on the
pinephone) and has no wireless or display. The image could be of use to
other aarch64 single board style devices, but since they usually don’t have
displays or modems, the phosh interface isn’t of particular insterest.

Until the advent of some performant mobile device to target or more
contributors who wish to drive things forward, I think we should just
mark the Mobility sig inactive at this point, including:

Retire the Phosh spin in f40
Hopefully find new maintainers for the phosh stack if there’s folks that want to keep them alive.
Mark the sig inactive on wiki and matrix.
respins will move to against a stable fedora release once a week or so, instead of against rawhide and more often.

Hopefully there is someday a brighter day for Fedora mobile devices.
Folks that want to try and revive the SIG and/or are able to see a good
path forward are invited to join


Is this something the Mobility SIG wants marketing help with? Depending on what the challenge is, we can come up with a time-constrained marketing plan to raise awareness about this SIG. Depending on the commitment level of the SIG members we can try regular posts for… 3 months? At the end of that time we can revisit whether we want to make the SIG as inactive or maybe it’s picked up enough activity to where we can keep it in the circulation of things we promote.

This isn’t the best example, but we’ve been trying to be intentional with supporting the Docs Team in increasing contributions. The Marketing Team wants to be available for teams and contributors who have something they want to promote.

At the same time I don’t think we want to spin our wheels on an effort if we realistically think it will wind down. What do you think?

Well, I don’t know that there’s a great path forward here.
On the one hand adervisting “come help us” and then telling people “we
have no idea what way forward there is here” seems… poor.

On the other hand, someone arriving may have some brilliant idea.

I don’t think the current sig folks have the energy to handle a large
influx of new folks, but I guess I shouldn’t speak for everyone. :slight_smile:

1 Like

I also think, less Spins who are complete is better than a ton of them who not really work or are incomplete.

Who knows maybe Risc-V will shorty change the game and make a new platform available.

Thanks @kevin for the State of the mobility SIG.

Same thing here, but it’s a shame nonetheless, would love to see Fedora being an important face in the development of GNOME Mobile and other smartphone endeavors.

Ok, well if other SIG members do want to have a run with marketing promotion to see how things go, we’re open. We can also include the state of the SIG in the messaging to manage expectations. But if you’d rather not, that’s fine too. I want to make sure we’re extending the help.

Me too! But until there is any phone hardware with enough upstream
support, it’s pretty difficult, at least in the Fedora context.

I think other projects with mobile artifacts are just all doing their
own (different) downstream kernels, which is fine, but it’s not good
long term and it’s not something thats really possible without remixing
in the fedora context.

Agreed! For the record this is what we got if someone wants to poke around and play with it (I’m not SIG but I have contributed in the past) - Fedora Mobility SIG · GitLab

Though as @kevin says the

  • Pinephone/Pinephone Pro builds were kindly built by yoda and @torbuntu with the so-called downstream “Megi Kernel”
  • Oneplus 6/T (SDM845) images were kindly built by Markus based on the PostmarketOS kernel

The rest of the stack was Fedora though! :slight_smile:

Also, surfacing these images from the repo above because demos are fun :smiley: