let's do some crowd sourced telemetry, please boost for reach.

Help me decide what i should be focusing on with regards to OnePlus 6/sdm845 #LinuxMobile development.

"mainline" => any Linux mobile distro shipping a close to mainline kernel

please only respond if one of the options below accurately describes you!

#kernel #postmarketos #linux

what issue do you care about the most?

this should only refer to Qualcomm devices newer than 2016 running close to mainline kernels and postmarketOS!

what would make you more likely to use a #LinuxMobile device?

@cas I got oneplus 6 because it looks like it will be very useful in future. I broke the android (still would like to recover the recovery and possibly android distro, too; plus the two red warnings during boot are annoying), tried Mobian and found Droidian was way better then expected.

My daily driver is Motorola Droid 4 with triple-boot. OnePlus is just physically too big ("incompatible with my horse").

I'd really like to see at least one camera to work. Fingerprint reader would be interesting. I'd really like to be able to power it down for storage (currently that does not work).

I still won't daily drive it -- would need bigger horse for that :-).

@cas Give me reliable calls/sms in device of about 100g, and I'll be very happy. Or give me camera and I'll carry op6 ... as a camera :-).

@cas all the above 🥳

@cas lack of free time!

@samueldr yeah same >,<

@cas @craftyguy + 1 camera. Anyway, thanks a lot for all your work

@cas does Linux mobile even have sandboxing like Android

@lucasmz short answer: not yet to the same extent.

it's a complicated one, because we're coming at this from a totally different angle it takes time to build these systems and gets apps to use them. Some very cool folks are putting a lot of work into portals (the tech flatpak permissions are built on). flatpak apps are sandboxed, but they're able to request access to a lot of things still that would make escaping the sandbox fairly trivial.

on the other hand, the app ecosystem is predominantly full of FOSS (in spirit) apps, they're open source and they're designed to treat their users as real people and not as ad-clicking data-producing machines. So we're fortunate enough that this problem hasn't reared it's ugly head yet (people aren't making malicious apps).

we have a lot of catching up to do, but most of the foundational work is done.

@cas The goal needs to be a "functionally complete" Linux-based mobile phone.

The UI and application ecosystem are developing nicely but there seems to be little available hardware that satisfies this need.

@cas Camera is a huge deal for me! I know it is friggin hard. But a smartphone without a usable camera can not replace my android phone. I would always have to carry an extra android 'camera phone'.

Really excited about what is happening in the linux mobile space currently though! Love what you are doing!

#LinuxMobile #PostmarketOS

@der_istvan @cas I have to agree. Top 3 reasons I'm not daily driving my OnePlus 6T at the moment:
1. No camera to send pictures to my wife.
2. No GPS navigation app to find my way.
3. I completely missed a call once and now I'm anxious.
Everything else is solved in my opinion! Like, seriously, good work!

@der_istvan @cas

I was going to say the same thing, although I kind of assumed that it would fall under "apps" ie "An app able to return reasonable photos (ie focused, minimal noice, properly exposed) under reasonable circumstances (ie reasonable light and stillness)

@BadassCoconut

preferably not from china

explain?

Just fell asleep at the computer with my hand on the keyboard.

Woke up to this:

@cas personal opinion here. It is not a single issue really, but the combination of **some** missing/lacking apps (camera and navigation), battery runtime (I do suffer from range anxiety, even with powerbanks in the backpack) and a feeling of sluggishness even on devices that are not underspecced (L5), possibly due to a combination of unfinished/unoptimized hw drivers and app optimization.
I find it very hard to name a **specific issue** that makes me sometimes want to grab the old Android.

@cas ohh, and we look forward to the day when gnome clocks can wake up our suspended phone :)

@neil @cas The goalposts are moving :-(. First it was cellphone, than it had to take photos, and now I'd like fingerrint reader, too...

@pavel @cas Yes. It's like there's a need for a Linux Mobile Driver project to develop an understanding of the hardware used in common phones and engineer drivers to address these features.

Assuming, of course, you can unlock devices to put a Linux Mobile OS on them. I think it's unlikely that a dedicated Linux Mobile Phone will appear from any mainline manufacturer so there's a need to repurpose like we do with laptops.

@neil @cas Unfortunately, PCs have working hw abstraction (you can just pretend it is VGA+IDE) and working BIOS/EC (so your battery charges even with no OS). Phones are just different :-(.

@neil @pavel I'm not sure what you think all the dozens of people contributing support for phones to the Linux kernel are doing... but this really isn't just a lack of organisation or something. these devices have so much complexity, it's just hard.

@cas @pavel I appreciate the efforts that are going into getting Linux onto mobile.

Is the real issue that so much phone hardware is closed and undocumented? I don't know that lobbying chip makers to make documentation freely available will produce results.

It's a hard problem to solve and I understand that but where are these efforts headed? Caleb asked recently for input related to what, I think, is a Broadcom device. Is that the targeted device because of particular support?

@neil @cas Closed and undocumented is part of a problem. That each phone is very different. PCs are, well, IBM PC clones, which means they are rather compatible with each other. Phones are not :-(.

@neil @pavel Qualcomm, not broadcom (but i can see why those are easy to mix up lol).

lack of documentation and source code certainly makes things difficult, but (weirdly) the strict requirements around automotive are actually helping us in a lot of ways here. As it means there's more of a direct profit motivation to be upstream and open.

I think lobbying efforts could help a lot, mostly just to ensure the bare minimum:

• the ability to run custom software (especially when the device goes EOL) and to do so with a (user managed) chain of trust (ie you can install your own signing keys)
• source availability for the stock OS (core parts at least) and associated tooling so that running a custom OS isn't unreasonably difficult.

The sheer complexity of modern smartphones is almost unfathomable. The Qualcomm landing team at Linaro (which I'm a part of) have been working on upstream Linux support for Qualcomm SoCs for literally a decade now, and devices like the OnePlus 6 being able to run mainline Linux with full modem support, call audio (with additional patches), accelerometer (with additional patches), battery/charging (with additional patches) is a testament to that work.

but this is making some huge assumptions about things like suspend working, that idle battery consumption is acceptable, that the GPU is able to ramp up and down its performance, that the UFS storage is able to negotiate enough bandwidth on the system bus to have fast access speeds (something we're actually struggling with).

it's hard to convey to someone without an embedded background just how difficult it is to do some of these things which we take for granted. I think people tend to assume that these devices are kinda like PCs, but they aren't at all. I mean, modern PCs (especially laptops) also aren't at all like the PC you might imagine in your head...

@cas @pavel Ah, Qualcomm, I stand corrected. Thank you for that, illuminating, post.

@cas Just started daily driving OP6 w PMOS edge/Gnome-mobile instead of L5 w Mobian (PureOS development on a pause atm). I like the experience, but miss the camera, app scaling and the ability to bring up OSK when it doesn't trigger. Anyway, extremely impresses by the work you do!

@tomasoqvist double-tap on the bottom bar should toggle keyboard visibility on recent versions of pmOS

@antonok Hmm, doesn't seem to work on my phone atm (if you by bottom bar mean the thin white line at the bottom of the screen). Will try again when I get new system updates.

@tomasoqvist ah, sorry - it's long-press, not double-press 😅

@antonok Long press works in Mobian and Phosh, but it doesn't seem to work with Gnome mobile on PMOS yet.

@tomasoqvist @antonok GNOME Mobile does not have a 'force the keyboard to show up' feature AFAIK. On stock GNOME, you can swipe up from the bottom to do that, but that gesture does another job on GNOME Mobile.

@linmob @antonok Maybe the phosh solution with a long press on the bottom bar could make its way into #gnomemobile eventually?

@tomasoqvist It's a separate project, so … no idea!
One known issue with the “long press the bar” implementation in Phosh is discoverability. That said, I really think GNOME Mobile needs a way to do force the keyboard to show up.
Cc: @verdre @tbernard

@antonok

@linmob @tomasoqvist @tbernard @antonok it's not that hard to add now that we have the new gesture framework :) https://gitlab.gnome.org/verdre/mobile-shell/-/commit/f16ed8f7302fe668dd9514d3d316aba0ac09b607