Magisk is a complete platform that focuses on hiding “unauthorized” modifications from the rest of the system. You can add any mods as a virtual overlay, it supports plugins etc. Root is only one of its benefits.
Here are things I’m currently using root for on LineageOS:
System-wide adblocking (hosts file with Adaway). Yes, there’s DNS-based adblocking, but sometimes that can be slow or not play nice with public wifi networks.
Battery management. I use AccA to limit my battery capacity and charge rate to keep it from wearing out. I might not care as much if the battery was easy to replace.
File management, as there’s an increasing amount of the filesystem blocked off without it.
Fair, but I guess my point was I haven’t personally needed those things in some time. At least, not enough to warrant exposing myself to vulnerability like before. Ad blocking especially, since you can do that via VPN or by using your VPN slot if you don’t want to use something like pihole. App + data backups are built into many custom roms as well, including the one I use (graphene).
Everyone’s security and convenience/feature needs are different of course. There is often a tradeoff. The way I see it, I wouldn’t give up having admin access on my PC for security, so why would I give it up on my phone?
Yes, it’s my main back up app. (The latest update kinda broke it though, I think first time happening, but the issue was spotted and will be fixed.) The only big limitation (which might have been fixed actually) is backing up and restoring app on work profile (anything other than user0).
The benefit to longevity from limiting battery charge even a little bit is pretty substantial. I recommend it to most people with the technical ability to do it.
If you’re really interested in the subject, here are the places to look around:
OpenWRT is the goto embedded Linux starring point. It is hard if you’re only familiar with desktop distros that use the Bash shell as the default. OpenWRT only has busybox with the Ash shell by default. OpenWRT is common on routers and is an entire distro that fits within 8-32 MB of flash memory on these devices or others like single board computers. Often, using an old router board just to hack around like it is an SBC is a great way to learn.
Try working your way through a Gentoo install on any old computer. Unlike Arch, Gentoo is a foundational distro that has a plethora of tutorial based content to guide you through all the different parts of an operating system. Gentoo packages very little for you in binary form. It shows you how to compile almost everything and allows you to learn how to customize and compile and package to your liking. You will likely need advanced-intermediate level skills to cope with the competence level that Gentoo assumes in documentation. Docs are written from the foundation up, but they assume infinite ability to learn and do not continue to hold your hand. By contrast, Arch de facto assumes you have a CS degree and have completed all courses on POSIX operating systems in how pacman is managed. Arch provides an excellent set of reference documentation, but is a fractal bottomless links chasm if you try to use it like a tutorial for contextualization like what Gentoo offers. If you ever wish to learn the next level and what Gentoo is actually packaging for you, Linux From Scratch (LFS) is a thing.
On a more practical side, base Debian is another foundational distro. It serves two primary purposes. It is the stable distro of choice. This means that most packages and libraries are frozen in time and only updated for security patches or packages that maintain backwards compatibility as an absolute priority. If you’re building some project using a bunch of high level Linux stuff but it needs to be safely online, doing so on a long term supported stable distro means, once the project is up and working, keeping the software up to date should not break anything. Debian is also the primary hacking distribution where hardware support for Linux is done. There are a bunch of Debian specific tools to get into and boot new hardware, called bootstrapping. This is how there is support for many hardware devices that have no public documentation or OEM support. The wonderful folks within this space make much of our world possible. I mention this one because a lot of the things happening with the initialization of hardware are better fundamentally documented in this space.
These are the places that I learned the basic lay of the land in this space. The boot up speed is a combination of the way the bootloader is configured, how the handles for hardware interfaces are initialized, how well the Linux kernel can trust these interfaces, and all of the software that is initialized before the user space.
Android does not require the end user to know anything about the device, networking, or OS best practices. It achieves this by eliminating the administrative user and any kernel packages that could modify the kernel or install an administrative binary. Then, Android makes all installed app developers full users on your device so that they may use their knowledge to configure all of the required interfaces and security. You ultimately have all of the same access as they do, but you are not the administrator or have any effective say over what they are or are not allowed to do on the device. There are a few measures to help block off some behaviors, but these are more like frivolous gestures to make you feel a little better rather than any kind of authority.
The reason your device gets depreciated and must be periodically replaced is because google packages the Android version of the Linux kernel with everything setup so that only the kernel hardware modules (drivers) required for the specific device need to be added at the last minute. These modules are only added in binary form at the last minute. The source code is never made public and these modules are not part of the mainline Linux kernel. This is the only reason your kernel is not updated regularly and is likely very VERY old with many security vulnerabilities. The manufacturer might recompile and send you an updated kernel if a CVE happens that enables remote code execution, but this is only likely if they have a substantial inventory of devices in the warehouse that have not already sold. It has nothing to do with you or ethical behavior. If the hardware supporting kernel modules code was merged with the mainline kernel, your devices would stay up to date with all the kernel security updates for decades automatically. If this sounds wrong, let me warn you now, saying so will put you in the Stallman camp where you will be labeled as a crazy extremist. This is the specific reason for Stallman’s insanity by his detractors. Stillman’s argument is that you don’t own your device.
These proprietary binary kernel modules are one of the primary aspects of boot speed. There is no telling what is happening on these levels when the device has proprietary binaries.
The system works with a bootloader that powers everything in a specific order and creates handles. The handles are passed to the kernel. The kernel initializes and starts running kernel space stuff. One of the main things it is doing is abstracting memory spaces.
If you’ve ever seen the earliest personal computers based on the microprocessor chips like the 6502 in an Apple II, they always had a RESET button. This is because a crash in the code crashed the actual hardware. In modern computers, your user space software only runs in virtual memory. This dies not require a reset because, while your software might still freeze, it is only running virtually. There is also a CPU scheduler that is handling interrupts (like key presses that can not wait, or background tasks) and power management works with this as well. When your software freezes, in theory, the kernel processes that are actually running on your hardware still get their time to run in kernel space priority on the CPU and their memory is protected from the virtual memory space of user software using virtualization.
Okay, all this bla bla bla is to say, if the device in question has no outside connection, and if the software can not change, and if the manufacturer is the one creating the bootloader AND kernel AND user space application all of this chain can be greatly simplified and bootup can happen lightning fast. This is called embedded Linux and is the most common form of Linux.
Android also has a system called Zygote. This preloads all of your apps when the user space loads. The user space on Android is actually like a single Linux application that runs on the Linux user space. The justification for Zygote loading everything in advance is because it makes everything load faster. Thus is what it says in documentation. Benchmarking shows that the difference is orders of magnitude smaller than your persistence of vision. In other words, it only exists to boot up the other dev users before you are loaded as the final product user. This is why you should not run any apps you do not exclusively trust. These app developers are like your bedmates but more intimately in contact with your person all the time. This is why everyone wants you to install their app. The google framework of Android is essentially a pimp and you are the product.
how is this even related to the question? this isn’t about any embedded linux, op stated it’s an android unit. and said it’s faster “than” other android phones and zygote applies to all so it’s unrelated also.
The vehicle likely does not have any protection systems or encryption. It likely has the bootloader integrated with the kernel. It is also running native or native like Android packages while altering zygote behavior so that extra applications are only loading in at execution time.
If you really want to understand the subject, intuitively grounding your understanding is a critical aspect of the processes. Telling people the simplified basics in isolation does not create useful understandings and assumes the person is on a similar foundational understanding. Someone that is genuinely curious, such as myself, but having no prior background can make use of such abstract overviews. Someone with total recall would likely find me pedantic. With abstracted intuitive thinking as a primary function, many people such as myself require such deeply embedded intuitive connections to make sense of the world with a deeper understanding of the connections involved. I retain no information in isolation in long term memory. This is neither right nor wrong in some asinine simple view of the world and the way people learn. If you find it odd, that is fine. Functional intuitive thinking is one of the rarer outlying personalities, but it is also the emulator function that can fake the rest with effort.
It is related, but on many levels, and useful to someone other than yourself. A binary perspective of learning and sharing of information is fundamentally incorrect.
Among the other things that have been said, Android auto often makes use of some tricks too. Things like hibernation that phones typically do not do (Probably the biggest one right here), Animations to hide loading time, loading some critical, but not latency sensitive services until after the boot. and some other misc service management stuff.
Yep but that’s not hibernate, that’s what happens when you leave your phone on bed side table for a couple hours when you sleep. You can get very low power under these, which may be comparable to your cars alarm system.
Android emulators though have the ability to suspend all activities to disk and come back in a few seconds.
The former is basically just a screen your phone is casting to. The latter is a lightweight (stripped down) Android fork designed to boot very quickly and do a couple things very well. It probably never really “turns off” since it still has a 12v connection even when the car is off (why your clock doesn’t reset).
Android on your phone is a much more general purpose operating system that runs on a (much more limited) battery. It isn’t designed to be turned on and off frequently.
Magisk had some incompatibility issues with lsposed-zygisk, which nearly resulted in losing my phone in a permanent bootloop, but I fixed it by flashing some images.
I now use Apatch on my Mi 10, running xiaomi.eu HyperOS 1
PS. I even tried KSU, by compiling a kernel for my phone after watching a ~40minute video on Portuguese with english automated subtitles, lol. Apatch has a much smoother procedure than that and kinda has the benefits of KSU.
One reason may be that they’re not actually off when the ignition is off, they’re just asleep like your phone is when the screen is off but it’s still powered on.
Holy jam, I saw this and had to comment. So idk about rooting a Motorola because it's not a "mainstream" Android (as in pretty much Samsung) nowadays. But perhaps Legacy in that situation. But I did have a solution for you if you were interested that doesn't require rooting (as far as I know). I was goobing around on the internet and found this man who was talking about the perfect minimalist phone. And figured when I gave a shit I would mess around with ADK (I believe it is) and this git-repo. But the lovely Jose Briones apparently also supports this. So give it a check if you want because I remember seeing Motorola explicitly being used in both the OG and the offshoot.
Anyways - less invasive phone while not being totally degoogled and the battery lasts longer. Hope you dig in cause I think you'll end up really liking it. GL!
You misunderstood. It’s not practical for YOU. It’s practical for Samsung who constantly needs a new gimmick to continue to hike the price of flagship devices higher.
? Just because it’s not practical for everyone doesn’t mean it’s “just a gimmick”.
Modern smartphones are freaking gigantic. Women’s pants pockets are comically tiny. Modern smartphones physically would not fit in the majority of my pants. I bought a Z Flip so I could fit a modern phone in my pants and not be forced to carry around a purse.
The length is important for women’s pants pockets, not necessarily the thickness. It is practical for me. No, it’s not practical for everyone.
I agree that modern smartphones are too big, but that’s what the vast majority of the options are nowadays. I don’t know if you can call it a “gimmick” anymore if literally all of them are like that. It’s getting harder and harder to find a reasonably sized phone. But believe me, if I didn’t have a folding phone, I’d be looking out for small phones like the Zen phone.
I bought a wallet case so its only as big as a wallet. Now I only have one object in my pockets instead of 2. Without the case its pretty slim folded up.
Yea, book style foldables (even though they still have the thickness, complexity and price tag) at least offer far more usefulness being able to deliver a near-tablet size screen when needed and a somewhat regular smartphone when not
These flip phones are nothing but a nostalgia cash grab IMO
I will add as someone who just got the z flip 6 as my first foldable, the biggest thing I’m enjoying and that the people around me are noticing, is that I don’t get “drug” into excess screen time just from looking at a notification. I’m more likely to dismiss it or deal with it quickly from the cover screen, or make a more conscious decision to open the phone up to engage, if that makes any sense.
I don’t think it’s recency bias or just trying to justify the purchase, and it’s probably fair to call it a really expensive solution to a self control problem, but it has changed the way I use my phone in a more significant manner. I also enjoy the form factor in the pocket, and for the security for NFC payment/taps, but that stuffs been covered plenty of places.
Until they can protect the hinge better, I wouldn’t be able to purchase one of these over a slab style phone. A grain of fine sand would ruin this phone’s life. I’m waiting on the time when foldable devices improve drastically enough for me to get one.
I used a non standard keyboard a long time ago and while it did deliver, the most important thing I learned is that non standard things tend to disappear and the time spent learning it was wasted.
My experience is it’s not worth it unless you’re looking for maybe a one-handed keyboard and want to type blindly (like, send a quick text while talking to someone without looking at your phone much).
Otherwise, I’m pretty happy with the default AOSP keyboard and Gboard. 70 WPM is more than good enough for texting, Slack, emails, and posting on social media.
I just Googled “typing test” and picked the first result which was typingtest.com. I don’t know if it’s good or accurate but it checks out for the score I expected.
A bit of an addendum: I think the limiting factor of a mobile keyboard is going to be that you have 1-2 fingers available for use rather than all 10.
A lot of those alternative keyboards seem to rely on some swiping gestures or drawing symbols. That’s good when you’re not looking at it as it’s very tolerant to misalignment, but if you already move fast and precisely, that feels like it would get in the way.
Looking at my fingers while I type, I’m already moving my thumbs as fast as I can without getting too sore: where it goes next doesn’t matter, it’s the same amount of time. Having to press and then move and release, in theory should only slow me down because by the time I finish the drag motion and lift the finger, I would have just moved my finger over the next key and pressed it. So that kind of layering is out of the window.
I can’t think of a way to type faster without involving mnemonics and chorded entry, but with two fingers you don’t gain that much.
Also I feel like the bottleneck there is how fast I can think of what to say next anyway.
android
Oldest
This magazine is from a federated server and may be incomplete. Browse more on the original instance.