Recent searches
Search options
Mobian started as a hobby project, triggered by the excitement of finally being able to hack one’s mobile phone at will. It is fair to say that the device that made it possible back then was the original PinePhone. However, after running once again into the same difficulties we experienced over and over, we’re wondering whether maintaining support for this device is still worth the effort…
https://blog.mobian.org/posts/2023/09/30/paperweight-dilemma/
To give you a sense of what we talk about: all patches in these folders are those we carry in Mobian.
https://salsa.debian.org/Mobian-team/devices/kernels/sunxi64-linux/-/tree/mobian-6.1/debian/patches
@mobian Does it make sense to join effort with @postmarketOS ? As far as I know, these guys are really pushing to mainline and upstream as many devices as possible, among them the pinephone. Maybe they could become the "one repo" for pinephone kernel ?
@mobian @postmarketOS my bad, it is also based on megi work it seems https://gitlab.com/postmarketOS/pmaports/-/blob/ebbbeca49c0bada7176e75b86c65a4a2be050884/device/main/linux-postmarketos-allwinner/APKBUILD#L35
@vmaurin @mobian @postmarketOS We already discuss and collaborate a lot, but even then, we're just a handful of volunteers with limited time and abilities
@mobian Same old story... Distros war, DE war... and now also the kernel which should be the COMMON cornerstone for everybody. 
With those project at the beginning there's always a lot of enthusiasm, confusion, then different ideas and approach lead to the disaster. Take the lesson from previous examples like Openmoko. Imho without a more "professional" approach I'm worry soon or late things will fall into oblivion. I bought my #Pinephone knowing that, but I hoped in a different destiny...
@KinmenRisingProject the hardware is what is dooming linux on the phone.
we need some basic standards from manufacterers for bootloaders, GPU drivers, etc.
i had lots of hope for my pinephone, and now i have sadly bought a Google Pixel. i hate it but has the best support for custom roms (GrapheneOS)
what i really want is a real linux phone with flatpaks.
@tootbrute the hw realated part is very disfficult to deal with... On a side the industries don't have interest for this market, on the other sometimes ppl build a project (for different reasons, some of which even right) relying on hw which is already obsolete or difficult to maintain; last the devs, some of whom are fixed in their own ideology & preconceptions, unable to speak & deal with other because they're always right. I can bring my experience with Openmoko's community to this chaos
@KinmenRisingProject yeah, it's almost like we need governments to legislate something like open hardware standards. they must publish the bootloader code so consumers can install whatever software they want on the device a mfg sells.
i heard the EU is doing something like this to help with right to repair. not sure how it'll work though. i just know i'm sick of throwing away phones because google stopped releasing android updates for it, or because i can't open the phone to change the battery.
@tootbrute @KinmenRisingProject To be honest, we already have most of the code we need to make things work (except for e.g. camera sensors), but making this code upstream-ready is just too much work for volunteers; what we need is upstream support from the SoC & phone vendors themselves: forks of the Linux kernel with millions of downstream changes (see https://not.mainline.space/) are just not usable in the long run.
@tootbrute @KinmenRisingProject It could happen through either government regulations or a policy change of the Android team (the ChromeOS team already mandates upstream support from Chromebook vendors), but I don't envision it coming for the foreseeable future...
@awai @tootbrute right, that's why I say there's the need of a sort of governance at a higher level to coordinate the efforts of the different communities. At the same time I'm aware this is almost a pure utopia, for various reasons.
@tootbrute @KinmenRisingProject the hardware part is what makes me think that Purism may have been right after all when they decided to spent quite some time on sourcing the right parts.
Sadly their high price point (now that devices are readily available) is just too high of a commitment :/
(I know that money flows back into software development but still…)
@gecko @tootbrute yes, different strategies, pros and cons for both
@gecko @tootbrute @KinmenRisingProject The real difference there comes from @purism having kernel devs on their payroll: in terms of hardware parts, IMHO they didn't make obviously better choices when looking only at the "software support" aspect (heck, even they ended up with a downstream driver for the wifi cards they initially shipped)
@awai @gecko @tootbrute @purism We all know that with money you can do more, especially when the work is longer and complex. But without, human resources should be optimized and you can only do this if there's a sort of governance (and ppl follow it). The development at the lowest lev.(fw etc) is the most critical point: fewer guys working on that (even because more difficult) means that, assuming you can't/ don't want to pay, they should collaborate (more) with each other. Same for upper levels
@KinmenRisingProject @awai @gecko @tootbrute @purism The only way to have people follow your governance is either:
a) to pay them; or
b) to find someone already aligned with what you want to do, and with ability to do so.
Everything that happens for free comes from b). It can lead to strong structures based on volunteer work and mutual trust, but it's *inherently* chaotic, especially when the community is still relatively small.
@awai @gecko @tootbrute @KinmenRisingProject @purism I think this was a bit of a "misunderstanding" between the card vendor and Purism who will look into upstreaming the missing driver bits 
(fueled by a takeover of the card vendor by another company)
@KinmenRisingProject @mobian If you want "professional", someone needs to pay for it. Otherwise you're fully at mercy of your community becoming big and diverse enough for someone with the right skillset to step in and do the needed stuff before the interest fades. And even when they do, they do it on their own terms - you can't demand anything.
Ever wondered where the price difference between PinePhone and Librem 5 comes from? Hint: there's no trouble with maintaining the Librem 5 kernel tree.
@dos @mobian We already know that. But professional >< enterprise. It isn't just a matter of cost, but also of will to give up your own private garden and join a field where you play following common rules. Even developing for yourself has a cost: your time, your money, stress, etc. Let's say it all: many open source fail because they are a total chaos. The approach/model of the most used LInux/BSD/etc keeps their ethic (not perfect,ok) and they do work well. Because there's organization. (...)
@dos @mobian Is that difficult to get 2-3 pine distros working together to achieve the same scope for such "small" project like a phone? Freedom doesn't mean total chaos. Focus on what everybody wants (a working phone). A final consideration: what would the Linux kernel be nowadays without a governance? Tnx for your contribution to the discussion, btw 
@KinmenRisingProject @dos @mobian
Mobian does collaborate with distros that *want* to collaborate. And a phone is by no means a "small" project. I would invite you to look at the difff between mainline and the OG PP and see just how much it differs.
@KinmenRisingProject @dos @mobian the thing is, this kind of work is incredibly time consuming. It's just hard. And there's actually way less incentive to do it for a device like the pinephone.
Pine64 won't pay you, but will directly profit from your work. Due to the immense popularity of the device you will receive a huge number of messages from people asking when feature X will be ready or trying to use you for tech support.
last but not least, i'm not sure how much attention upstream sunxi support actually gets, whos maintaining and testing?
compare that to working on a Qualcomm platform where your work is likely to benefit many different devices which are already out in the wild (rather than having to buy more future e-waste). There is a growing upstream community, and you can actually get paid.
the pinephone issue isn't for lack of trying, there just isn't anyone who wants to work on it because you just don't get anything in return except for users complaining about things and a company using your work to sell more products.
@cas @dos @mobian It's not an exclusive of Pine64: how many manufacturers do pay everybody working on on their hw for making floss/foss fw,drivers, sw? When somebody creates a device, somebody need to invest money. PIne64 like everybody else payed (and pays) for project, components dev. etc. & a company to exists must earn money. What you get in exchange is a device where you can install LInux.
I agree about time consuming activities (I 
the same in a previous post of this thread) but (...)
@cas @dos @mobian if you do this to get 
, well than we are talking about jobs, smth different. About different platforms I'm not competent to say what is worth to do and what is not. I just point out that pine
to contribute 4 free
@KinmenRisingProject @cas @mobian Both Pi and Arduino employ developers that act as stewards for software support for their platforms. Even Openmoko did.
I did my fair share of unpaid development work in the mobile GNU/Linux world across the last two decades; I also had an opportunity to get paid for such work, so I know both sides well:) I wouldn't be able to do half the things I did if had to earn for living with something else as well, and the community is still small enough for it to matter.
@KinmenRisingProject @cas @mobian Once we get to the size and level of organizational maturity of Debian, Linux or similar hugely successful FLOSS projects, then things will mostly work on their own: if there's something that needs to be done in Linux, eventually someone pops up and does it, paid or not (though more likely to be paid, in fact!). We're nowhere close to that yet, I still regularly stumble upon issues that turn out to be trivial to fix but that simply nobody looked close at so far.
@dos @KinmenRisingProject @cas @mobian Fully agree on that! Partly due to the small size of this community, we lack the people with the right combination of personal (or corporate) interest, skills and available time. Which is pretty much why this particular kernel ended up being maintained by a single person, drastically reducing the bus factor.
Also, a phone is nowhere near a "small" project, that's actually one of the most complex piece of hardware...
@mobian
The repo you linked has 22 patches in the pinephone dir.
https://salsa.debian.org/Mobian-team/devices/kernels/rockchip-linux/-/tree/mobian-6.6/debian/patches/device-tree has 38 patches.
Those patches are all related to the DeviceTree, describing the hardware!
If every 'distro' has their own view of the actual hardware, I can't see how any real progress can be made.
Then you get f.e. https://forum.pine64.org/showthread.php?tid=17686 where the first Q in OP is:
"Are you able to get audio to work during phone calls?"
This is just 1 example; took me <5m to find.
This is sad and VERY worrisome :-(
@mobian I know this is about the original PinePhone, but to the casual reader it might appear that Mobian and the PinePhone ecosystem as a whole are left with an unmaintained kernel. Appendix II seems very hopeful that the community has learned from these mistakes moving forward. A title change such as "The Paperweight Problem and the Original PinePhone" might help convey this.
@mobian Beside of lacking quality control and other bad experiences (my Pinebook Pro didn't even last two months) that lack of core software support is one of the main reasons I will never ever buy another Pine64 device! If you sell devices to run Linux you don't have to go all the way like Purism and develope everything in house but you should at least have one or two people who work on the core to make life easier for distros and ensure certain issues actually get fixed for everyone! :/
@mobian In the long run their devices aren't nice to run Linux on unless you only care for Manjaro which seems to have some sort of deal with Pine64 and they will end as additional E-Waste. Yes I can reuse my Pinephone as minicomputer and that's great but it's not a very good use for most of the parts in that device and the fact that E.g. the screen or modem won't end in the trash only means it can't be recycled ether while I continue to use it as minicomputer! :/
@gamey tbh this blogpost from @mobian about the difficulties of maintaining the kernel for the original pinephone kinda reminds of the article @drewdevault wrote sometime ago:
https://drewdevault.com/2022/01/18/Pine64s-weird-priorities.html