Does anyone here use, or intend to use, a Linux Phone?

[kedaha]

It's been three years since I last posted about this subject and things, as well as times, have changed. There are now several options on the market. However, I have an old Nokia n900, which I haven't used for years and which works perfectly. I was pleased to see that there is a new operating system for it called Maemo Leste, built from Devuan Beowulf (Debian Buster). So I am going to try it.
For free software enthusiasts, the obligatory use of proprietary apps for, e.g., the bank and the ubiquitous whatsapp, which one is pressured into using because everyone else uses it, really goes against the grain. It looks like the only way round this is to have two numbers.
But ridding oneself of proprietary software entirely is easier said than done but I well remember years ago being able to "migrate" from Windoze to Debian 4, Etch, a decision I've never regretted. So I'll start by seeing what I can do with my old Nokia.
Thanks for reading and for any replies.

[Head_on_a_Stick]

I still can't understand what advantages a "Linux" phone would offer over LineageOS or GrapheneOS. The security would be abysmal compared to those, Linux lacks coherent per-application sandboxing (and no, firejail doesn't count, user namespaces are a gaping wide hole in the system).

[Deb-fan]

^ Interesting Hoasinater.

As for the thread topic, abso-fraggin-lutely dude!!! Errrr, well sort of(not really.. no) Android is the closest to mobile gnu/Linux as I'll ever get. Am mostly tektarded as to everything Android can do. Honestly am probably never going to have the time or energy to learn about it. Nothing else mobile is ever going to get the development effort to match what's being done with Android. Yep ... it's known to have plenty of security issues, certainly but the ridiculous range of tech-crap people can do with Android/mobile is stunning too. For that matter folks are installing Android or hybrids on desktop to use apps etc.

Will support open source mobile(pretty much for selfish reasons.) If nix ever gets to the point of matching Android would prefer using it instead. Have to admit don't really know chit about the true state of Linux on mobile either though. After a decade of fanatically screwing with gnu/nix on desktop and still have tons to learn. Don't think I can ever muster the energy to try to tackle it either. Still thinking nobody will be able to match the momentum Google Inc has for Android.

[kedaha]

I was thinking more along the lines of wiki.debian.org/Mobile:

It would be great if Debian could run on many kinds of mobile devices: iPhones, Android capable HW, Windows Mobile HW, tablets, but this is very unlikely to happen anytime soon for the majority of devices.
Devices built to run FLOSS stacks are more promising alternatives, for example PinePhone and Librem 5.

[reinob]

We still have SailfishOS: it's GNU/Linux (with systemd :), where the only non-Linux thing is the android kernel and the various android hardware drivers.

I still use (seldom, but..) my Jolla 1 and a number of N900s.

[amyacker]

I use LineageOS with my Samsung. Works like a charm, fingerprint sensor and all, with constant updates.

[Deb-fan]

I really shouldn't have posted due to not having enough knowledge on the topic, mobile tech is bound to be a vast specialty area itself. Keep trying occasionally taking a stab into that field, haven't made much progress. At least with Android, it's running the Linux kernel. Am sure open source enthusiasts are going to do amazing things with mobile. Am also sure gnu/Linux proper is going to face the same challenges there as it always has. Major corps hamstringing it to ensure their interest$. Doesn't matter if someone has the most amazing software on the planet, if there isn't hardware capable of running it or using it well.

Google Inc has the influence and financial resources to put out compatible hardware, assure drivers and performance are there with Android and clearly have successfully hijacked a great deal of the efforts that open source devs will devote to mobile. The interest and demand for Android is already there, alternatives are going to receive a trickle of the time/effort compared, cause it's uphill duplicate work by this point. This thread is prompting an interest in researching the topic more(again.) Still can't ever count out gnu/Linux or the amazing people involved in the open source community either. Arghhh no matter which area of tech, still feel I've got too much to learn to have a real clue.

[Deb-fan]

Related brainfart: Imo something that's killing gnu/Linux regardless of form factor. Remember endlessly whining about this since started using gnu/Linux. Thinking about gnu/Linux + mobile vs Android brought it back to mind. Simply put gnu/Linux doesn't come preinstalled, much less on fully compatible hardware that's known to perform well.

Remember saying over and over through the years, why do people use Windows? When all this awesome/better software is available for free !?! Because it comes ready to go out-of-box on all the hardware they buy. Far as I know only Canonical Inc ever got to a point where gnu/Linux was preinstalled as an offering by any real tech-player. Dell tried selling systems with Ubuntu, it eventually failed of course. Situation with mobile, gnu/Linux has always been an ideal candidate for embedded devices esp, though again we're never going to see any real effort on preinstalled devices coming out for it. Though Canonical Inc set it's sights on that also and crashed and burned there too.

Google Inc is slaughtering gnu/Linux in terms of user-share, market penetration and popularity. In Android, for that matter ChromeOs(desktop), in the blink of an eye Google Inc was able to attract far more users for ChromeOs than there are users of every gnu/Linux distro combined AND growing. This obvious nix not coming readily preinstalled thing and impact it has, sadly is something that's unlikely to ever change. Personally feel gnu/Linux is the best thing going for everything and should have long since been dominating the desktop as well, now there's mobile and looks like it's going to be more of the same story to great extent.

Mentioned don't believe anyone should ever count gnu/nix out though. Personally really need to start doing more to give back and promote the platform. Obviously gnu/Linux is capable of amazing things but taking out ads to let people know it even exists isn't one of them.

[LE_746F6D617A7A69]

Linux lacks coherent per-application sandboxing

Uhm, Linux sandboxing capabilities are indeed horrible:
https://www.cvedetails.com/product/36/D ... ndor_id=23

A professional sandboxing offered by a serious OS like Android, a project worth billions of dollars, must be just perfect...
Well, no - not even close to perfection:
https://www.cvedetails.com/product/1999 ... or_id=1224


Sandboxing is a reletively new term, but it means nothing but process separation and control of usage of resources - which is precisely what the Linux kernel is doing.

[Head_on_a_Stick]

A professional sandboxing offered by a serious OS like Android, a project worth billions of dollars, must be just perfect...
Well, no - not even close to perfection:
https://www.cvedetails.com/product/1999 ... or_id=1224

Well yes, of course Android is far from perfect — Sturgeon's Law still applies. But it's *much* better than Linux.

[LE_746F6D617A7A69]

Well yes, of course Android is far from perfect — Sturgeon's Law still applies. But it's *much* better than Linux.

As far as I understand, You're claiming that free Linux distros are not suitable for mobile phones, because the standard Linux kernel has insufficient sandboxing capabilities. Could You please elaborate more on that? -> i.e. prove Your statement?
I mean, that CVE reports are showing that this is not the case... (Arbitrary code injection, bypassing sandboxes, etc)

I agree, that distributions like Debian are completely not ready atm for mobile phones, but not because of sandboxing problems - simply there's almost no single application ready for small touch screens. But IMO that could be relatively easily fixed - the problem is in missing firmware.

[Deb-fan]

^ Yeppers one of many things ChromeOs and Android enjoy gnu/nix never will(at least not as yet), hardware that's designed to work well with Google's offerings. With the sheer number of devices, esp Android anyone and everyone who isn't retarded is falling all over themselves to make a natively supported app. Plus the legions of developers flocking to freely contribute and create more apps for them, ah bunch of other junk. But seems it's a brave new world with tech like emulators and vm's, haven't as yet tried it but supposedly more than a few ways to run Android apps on gnu/Linux.

That's my dreamland perception of what gnu/Linux on mobile would be, a super streamlined nix install, specialized for the purpose with integrated emulation for using Android apps. Though of course it'll have to be minus the device support Google inc's stuff receives. Gnu/Linux on mobile will have to do the heavy lifting itself. No fully compatible, affordable and preinstalled hardware for gnu/nix mobile. Think it's another major barrier the platform has long suffered, users not only have to choose gnu/nix, they have to invest time-effort too. For a good chunk of tech-users and consumer electronics users, that's a bigtime deal breaker ... period. Given the ridiculous amount of users Android has now, oem's really have lesser to no motivation to even bother supporting gnu/linux.

Still given reasonable hardware support/performance think gnu/nix would be great on mobile anyway. Still no doubt ways to hack something together but with the size and scope of the open source communities surrounding ChromiumOs and Android not much reason to bother, even among us/those who feel nix is the best thing ever on desktop. In my uninformed and admittedly ignorant view would kind of like to see gnu/Linux focus on it's strong suit. Dominating the production/Enterprise areas of tech and expanding into desktop. Clearly one of the tech superpowers behind Google's explosive success, they readily get the device support gnu/Linux never has. Though in desktop that's changing somewhat and is bound to improve the more users gnu/nix attracts.

PS, Sunrats right, I talk/type too much, dayum ... Look at all that!

Oops ps2, Something else Google Inc's open source for fun n profit enjoys which gnu/nix doesn't, it's standardized and uniform, there's one gateway with the final word, yea or nay = Google Inc vs this fragmentation thing we keep hearing about in gnu/Linux, though think that's slated to change now too Redhat aka IBM, systemd, Wayland, pulseaudio, bunch of what they consider software chuff (or hobbist crap)being excised and left in the dust. Long established and disregarded best practices being enforced and open source projects, even major one's basically being told get up to this standard or go f-yaself. Hoping the end result will be overall positive. We'll see.

Really think it will, IBM didn't pour all those billions into buying Redhat for nothing. Between Google Inc and them, both maneuvering to rip M$ to pieces with open source. For fun and mostly for profit.

[Head_on_a_Stick]

As far as I understand, You're claiming that free Linux distros are not suitable for mobile phones, because the standard Linux kernel has insufficient sandboxing capabilities. Could You please elaborate more on that? -> i.e. prove Your statement?

No, I can't prove my opinion but neither can you prove that vanilla Linux is as good at sandboxing as Google's fork.

My opinion is based on what I have read (I'm no expert), here's a good example:

https://www.reddit.com/r/GrapheneOS/com ... dium=web2x

^ That's from Daniel Micay, the lead developer of GrapheneOS and the original maintainer of Arch's linux-hardened kernel version.

[LE_746F6D617A7A69]

My opinion is based on what I have read (I'm no expert), here's a good example:
https://www.reddit.com/r/GrapheneOS/com ... dium=web2x
^ That's from Daniel Micay, the lead developer of GrapheneOS and the original maintainer of Arch's linux-hardened kernel version.

Apparently, this guy has a serious problem:

The Linux kernel is a security disaster

When people tell you that Debian is secure, it's like someone trying to claim that Windows XP with partial security updates (via their extended support) would be secure.

I'm guessing that he have forked grsecurity kernel and now he is disappointed that Google didn't hired him...

Don't read articles written by some frustrated guy - learn the facts:
The Linux kernel implements 3 basic interfaces for full sandboxing of the applications (the "sandboxing" term is modern, but stupid) :
seccomp, cgroups and namespaces

The Android sandboxing framework is based on those kernel interfaces - there's no magic in it.

[Deb-fan]

Any project seriously attempting gnu/Linux proper aimed at mobile would have to involve using a highly modified custom Linux kernel for it, just like the techno-ninjas at Google Inc and the open source communities assoc with Android are doing. Agree yeah, Android is heavily relying on what's been done by Linux kernel devs, they'd have to be retarded not to take advantage of it and the people involved with Google are clearly some of the best of the best in the friggin world. They aren't stupid.

PS, though same thing for desktop nix too. Any serious distro's aimed at desktop are or should be using a kernel customized for the use, the whole "performance kernels" one setup for the task at hand. One of the endlessly amazing things open source makes possible. A lot of people just don't have the skill or know how to use what's available to best advantage. Google Inc however does.

[LE_746F6D617A7A69]

Any project seriously attempting gnu/Linux proper aimed at mobile would have to involve using a highly modified custom Linux kernel for it, just like the techno-ninjas at Google Inc

Right, so f.e. Google have modified the Linux kernel 2.6.x around year 2007 by contributing the cgroups code - and since then they are using it as core component for sandboxing apps on Android...

[Deb-fan]

^ Hey cool and thanks, of course didn't know such. Started to edit last post and add some stuff about Google contributing to the kernel and gawds know what all other open source too. Like to believe it's for altruistic reasons and think to some degree it actually is, though key motivator is almost surely looking out for their own best interests and ambitions. Wondering how much infuence/control they must have amassed over what happens with the Linux kernel by now? In ways that's kind of concerning to me, not that there's a dam thing I can do about it, whatever the situation and no matter what ... Google does contribute likely A LOT to open source.

Ah Google surely wouldn't hire me either, I'd be-get mad about it but honestly can't blame them, so whatcha gonna do? The people behind Google clearly don't just know how to maximize leverage and benefit from open source, friggers know how to cherry pick the best stuff from it and have the talent and resources to improve on whatever from it they've chosen. Generate massive interest among open sourcers to improve and expand on their projects seemingly at will. I can't help but admire them. Crushing competitors, left, right and center.

Score card:

Internet search: Google dominates.
Web browerage: Google Chrome dominates.
Mobile: Google Android DOMINATES!

[LE_746F6D617A7A69]

My point of view on Google is definitely not as optimistic as Yours... but that would deserve to start another thread...

The most fundamental problem with Android is that all user-accessible applications are written in Java and are running in customized JavaVM.
This is the main reason why Apple is kicking Google' ass in terms of performance and power consumption.

Google have realized that there's a lot Java code monkeys around, so from business point of view, the implementation of Dalvik was just excellent move - who cares about the end-users anyway? (or who cares about quality of code in Android apps?)
A 100MB application needed to connect the smartwatch? -> obviously a trojan/malware/spyware -> but of course, it is sandboxed

I really would like to have a mobile phone powered by some free distribution like Debian.
Having such an option, it would be easy to implement native support for touch navigation and optimize the GUIs for small screens.

[Deb-fan]

Appreciate you sharing some insights. I admire how Google operates but don't approve of everything they elect. Don't believe there's a whole lot I can do about the maneuvers tech-gaints like Google make. At least when it comes to Android for that matter ChromeOs and the Chrome browser, they're all tied to apparently thriving open source projects/communities + the kernel too. Not to mention gnu/Linux has to be powering overwhelming amounts of their backend tech. So think in ways Google is intimately joined to open source. Will protect it and has a vested interest in nurturing it(at least parts), if only out of self interest. Of course they want to make money too and are transparent about that fact.

Can see both pro's and cons surely. Am also sure efforts will continue in gnu/Linux on mobile, badly stunted by Android though. On both ends, oem's have pretty much zero motivation to bother giving any support to gnu/Linux with Android around and same for open source devs too now. At this point my veiw goes as follows, production environments or desktop, gnu/Linux. Mobile(for many types of consumer electronics, my own), Android.

Things like crappy power management-etc, hopefully the Android community continues making innovations, they surely will. Also am certain much can be done by anyone who gets familiar with the Os, I've drastically increased battery life, even on un-rooted Android devices and much better via rooting. That was only really basic poking around in Android, there were clearly A LOT more performance gains to be had fiddling around under it's hood but lost interest in it at the time. Only ever got to the point of scratching around on the surface of what's possible with it.

My prediction, think obviously Android's here to stay and will continue to dominate mobile from now until forevermore.

[Head_on_a_Stick]

Apparently, this guy has a serious problem:

The Linux kernel is a security disaster

When people tell you that Debian is secure, it's like someone trying to claim that Windows XP with partial security updates (via their extended support) would be secure.

Well the current upstream LTS kernel version is on 4.19.132 but Debian stable is still stuck on 4.19.118 so it does seem to be lagging behind somewhat.

I'm guessing that he have forked grsecurity kernel and now he is disappointed that Google didn't hired him...

He did manage to port the PaX kernel patches to Android and also port OpenBSD's malloc so he seems fairly capable to me

The Linux kernel implements 3 basic interfaces for full sandboxing of the applications (the "sandboxing" term is modern, but stupid) :
seccomp, cgroups and namespaces

The Android sandboxing framework is based on those kernel interfaces - there's no magic in it.

Yes, thank you, I am familiar with seccomp, cgroups and namespaces but my point is that Android uses those frameworks on a per-application level. Linux distributions only do that for a few discrete applications (like Chromium) and even then the separation is not complete.

As I said in my earlier post there really doesn't seem to be any point in porting Debian (or whatever) to Android when that already has it's own open source versions.