Looking for a desktop environment suitable for use as a phone interface in FreeBSD installed on a phone.

[MrBSD]

reddit can be quite toxic

It can also be on point. Like in this case you quoted.

 

[MG]

And what device will all of this be installed on?

Without cellular baesband and/or SoC GPU drivers this whole idea is pointless. You need to start with the base system first.

The consumer industry is a problem. It's boarded up against the situation of the operator of the system controlling all buttons. This is why smartphones are pretty much limited to Android or iOS. Smartphones are an artificially separated class of computers that force the user in a sandbox with strict boundaries. You can't even run an ARM binary on your ARM computer...

 

[ZioMario]

And what device will all of this be installed on?

Without cellular baesband and/or SoC GPU drivers this whole idea is pointless. You need to start with the base system first.

I've already created the basic of the phone,porting the wi-fi and the GPU driver from Linux.

 

[MrBSD]

The consumer industry is a problem. It's boarded up against the situation of the operator of the system controlling all buttons. This is why smartphones are pretty much limited to Android or iOS. Smartphones are an artificially separated class of computers that force the user in a sandbox with strict boundaries. You can't even run an ARM binary on your ARM computer...

Linux phones exist. Like Liberty and Pine. Same can be done with FreeBSD. Should it be done? As a challenging hobby project, yes. As AI slop, no. Pure waste of tokens for absolutely nothing.

 

[ZioMario]

This is the Trump phone (more or less) :

The code that I've produced is working. So,I'm not sure if it is a SLOP. For sure it should be improved and I will do it over time,but for now It is pretty stable.

So,now,can we restart to talk about how the Desktop manager should be ? I have placed a lot of meat on the table. Not all the options I listed are worth to be developed.

 

[MrBSD]

The code that copy/pasted is working. So im sure its SLOP. For sure it should be improved and Claude will do it over time,but for now i hope its stable.

I fixed your post.

 

[ZioMario]

I fixed your post.

I don't like your correction. Claude doesn't work alone. I'm the one who gives it direction, yes, even though I don't know how to program, I understand the direction it's taking, the mistakes it's making, if the information it has is wrong, etc. I'm not a complete neophyte. And before all,the responsibility of what it does is mine. Even if I'm not a programmer.

 

[MrBSD]

I don't like your correction. Claude doesn't work alone. I'm the one who gives it direction, yes, even though I don't know how to program, I understand the direction he's taking, the mistakes he's making, if the information he has is wrong, etc. I'm not a complete neophyte.

Im done man...

 

[MG]

Linux phones exist. Like Liberty and Pine. Same can be done with FreeBSD. Should it be done? As a challenging hobby project, yes. As AI slop, no. Pure waste of tokens for absolutely nothing.

It shouldn't be a massive complicated procedure. If the hardware is known and supported, you can boot an USB image. FreeBSD can run on most CPU's but they are almost always rigged with a locked bootloader and if not, it gets no mobile connection for being an untrusted device that doesn't run the box kernel.

 

[NapoleonWils0n]

Is the cross posting a violation of some rule

Not at all just saying they tend to be quite toxic on reddit
when it comes to ai generated posts

tends to be the younger generation for some reason
was just a heads up that you are more likely to get grief than a sound technical answer

 

[MrBSD]

Older Motorola, Samsung and Pixel phones are well known and fully hacked. I remember old LG phones where you can dual boot Windows CE and Android. Unfortunately, such hardware is way too old for these kids that need 8K display resolution and 150 Megapixel octa lens camera. Which brings us to original conclusion. Its completely pointless and useless. In this very case, there is nothing to learn and nothing to gain.

 

[MG]

This is the Trump phone (more or less) :


View attachment 27087

The code that I've produced is working. So,I'm not sure if it is a SLOP. For sure it should be improved and I will do it over time,but for now It is pretty stable.

So,now,can we restart to talk about how the Desktop manager should be ? I have placed a lot of meat on the table. Not all the options I listed are worth to be developed.

You can dissolve the plugs with paint thinner and then isolate the exposed wires with cellulose spray. It gets much smaller of that.

 

[NapoleonWils0n]

All of the open source phone "desktop" are linux based
Or Android like grapheneos

eg Ubuntu touch Plasma mobile, sailfishos
and they havent made much of a in road into the phone market

The issue trying to get them to run is if they depend on systemd ( which wont work at all )
or sys calls that the linuxuxlator doesnt support

Or trying to strip everything out to get it to work natively on Freebsd

 

[MG]

All of the open source phone "desktop" are linux based
Or Android like grapheneos

eg Ubuntu touch Plasma mobile, sailfishos
and they havent made much of a in road into the phone market

The issue trying to get them to run is if they depend on systemd ( which wont work at all )
or sys calls that the linuxuxlator doesnt support

Or trying to strip everything out to get it to work natively on Freebsd

How does that dependency work? The hardware only agrees with a specific systemd executable that comes after the kernel?

 

[Beastie7]

The consumer industry is a problem.

The issue is that developing a smartphone operating system is hard. There's too many variables and constraints you have to consider.

We don't even have proper power aware CPU scheduling for ARM SoCs yet (especially for heterogeneous cores on high end SoCs), or dynamic service management. That alone is important for power efficiency - which is a huge part of having a usable smartphone. We're getting there though.

IMO, time and effort would be better spent on getting FreeBSD up to snuff on the low level parts first before even bothering with the upper layers like display servers and GUI toolkits. Fortunately, that part cannot be AI slopped either.

 

[Beastie7]

I'm the one who gives it direction, yes, even though I don't know how to program

You cannot AI slop your way into a smartphone OS. You need to learn how to code. That'd be a better long term investment of your time and energy.

 

[MG]

The issue is that developing a smartphone operating system is hard. There's too many variables and constraints you have to consider.

The situation doesn't change while on the charger... They made everything unreachable. No wired keyboard. network or UART and a OS-level touch interface that only listens when the original driver is loaded. Booting from internal storage is the only option. If you're lucky, you can replace the original kernel with brick-risk, which is also a scam. Computers don't do that when there's no physical damage.
The smartphone market is entirely falsified to rule out competition and speed up consumption of products with artificial obsoletion.

 

[Beastie7]

The situation doesn't change while on the charger... They made everything unreachable. No wired keyboard. network or UART and a OS-level touch interface that only listens when the original driver is loaded.

You do realize what a smartphone is, right? The status quo has always been the same, even before the iPhone. The main difference between an iPhone and a Palm Treo/Blackberry is how the UI is manipulated, and how applications are presented to the user. Smartphone OSs (which were called PDAs before the iPhone) are hard, period. ARM was conceived out of providing power efficient processors for these types of embedded devices. Modularity would never bode well with that sort of device. Your expectations are misplaced.

 

[MG]

You do realize what a smartphone is, right? The status quo has always been the same, even before the iPhone. The main difference between an iPhone and a Palm Treo/Blackberry is how the UI is manipulated, and how applications are presented to the user. Smartphone OSs (which were called PDAs before the iPhone) are hard, period. ARM was conceived out of providing power efficient processors for these types of embedded devices. Modularity would never bode well with that sort of device. Your expectations are misplaced.

Smartphones are a generation of computers. There are no significant differences with computers 40 years ago, except the size, speed and communication methods.

Serious pollution of the past 25 years is protected mode and ecosystems to lock out end-user.

 

[Beastie7]

Forest.. trees.

Anyway. No modern drivers/scheduling = no usable smartphone. That's not the industries fault, but a manpower, resources, and interest issue.

Disregarding all low level intricacies, iOS (which is in part, BSD derived ) does what it's designed to do, and does it well. Your average user isn't going to care what is or isn't open source beneath that layer. This is reality.

Good luck

 

[ZioMario]

Forest.. trees.

Anyway. No modern drivers/scheduling = no usable smartphone. That's not the industries fault, but a manpower, resources, and interest issue.

Disregarding all low level intricacies, iOS (which is in part, BSD derived ) does what it's designed to do, and does it well. Your average user isn't going to care what is or isn't open source beneath that layer. This is reality.

Good luck

What Claude suggest to do, in order :

1. Extract the data from the Allwinner BSP (energy model, OPPs, idle-state parameters, capacity-dmips-mhz) and check — with the commands given above — whether BL31 actually exposes CPU_SUSPEND/SYSTEM_SUSPEND. This costs an afternoon and decides whether items 3–6 in the earlier list are worth pursuing or are dead on arrival.
2. Write a replacement for powerd in userspace: cpuset(1) foreground/background partitioning, explicit per-cluster frequency, GPU and display gating. Weeks, not months, and it gives you a measurable baseline before touching the kernel.
3. Only afterwards, and only if the firmware gate has passed, capacity awareness in ULE — reimplemented, BSD-licensed, potentially upstreamable.

 

[balanga]

How does one go about installing an alternative OS on an Android phone?

 

[ZioMario]

How does one go about installing an alternative OS on an Android phone?

Not my favorite kind of project. I love FreeBSD and I also love to choose the hardware components and...I love to create new pieces of code never existed before for that OS...even if is Claude that does it. I don't care. New pieces of code will exists thanks to the fact that I love FreeBSD and it's my decision even if I will not write the code.

 

[MrBSD]

How does one go about installing an alternative OS on an Android phone?

If you have a Google Pixel, its very easy installing GrapheneOS on it. Everything else. especially Samsungs, are very tough eggs to crack because they have mechanisms in place to prevent you from doing that. You can literally brick your phone beyond repair. You should check XDA forums and see if your phone exist in android development section. Be extremely careful and read the instructions carefully.

 

[ZioMario]

Months ago I've bought the Google Pixel 7 pro and a long time project is about to install FreeBSD on that phone. I'm not interested in Graphene OS. I'm interested to experiment with FreeBSD. Neither Linux interests me. I use it only as an oracle to port the drivers.