All of my background apps get killed when I turn off the screen. This is particularly annoying for my most-used messaging app, **WhatsApp**. I've looked at the obvious "**Battery Optimization**" option: Based on advice found elsewhere, I've checked that the background process limit in **developer options** is the default (I've never changed this AFAIK): And yet, it gets killed in the same manner as battery optimization would. If I look at running background processes: **Before Poweroff** When I then press the power button, and power it back on after < 10s, the list is the same. But if I wait longer than that: **After Poweroff** ...whatsapp has disappeared. This happens for all apps I install that have some sort of useful background feature - the Battery Optimization setting doesn't seem to do its job anymore. What else can I try to diagnose this? **NB0**: Android 7.0 on DOOGEE S60 **NB1**: This just started happening some time ago, i.e., it _did_ work propertly when I got the phone. **NB2**: I know how to use adb **NB3**: I'm also fond of trying new apps and have ~400 apps installed. Ideally, I don't uninstall/disable _all_ of them 1-by-1 to see which one is the culprit...

I have a Xiaomi smartphone, the OS is Xiaomi HyperOS, which I believe is a variant of Google's vanilla Android OS. I used adb to open a shell, and found that there are two "suspicious" processes tftp_server and tcpdump running all the time:

vendor_rfs    1779     1 0 20:50:18 ?     00:01:42 tftp_server
system        2289     1 0 20:50:19 ?     00:00:45 tcpdump -i any -W 3 -C 5 -s 134 -w /data/vendor/wlan_logs/tcpdump.pcap -Z system

I know tftp service is usually for firmware upgrade in embedded systems, and tcpdump is for capturing packets, especially when debugging code related to network protocols. What is the purpose of these 2 processes?

I enabled developer options on my Samsung Galaxy Grand Duos and then changed "Limit background process" option from *Standard limit* to *No background process*. After that my phone is working fine, with no hanging problems or no slowness as before. But my doubt is, will this change affect any process, updates, applications or anything which might cause an error or failure?

I'm using Frida over USB debug and when I run frida-ps -U, I see zygiskd64, zygiskd32 and zygote appearing in the process list. This alarms me as some of my apps might be detecting those as an anti-root measure. Obviously, I'm rooted on Android 12 with Magisk, and modules Play Integrity, Shamiko and Lsposed installed. Is it possible that it's only me seeing the zygisk process but Magisk effectively filtered out (hidden) those from the apps are can apps see the same thing when they emulate running process? If so, is there a way to hide (or at least rename) those processes? Of course apps have very diverse range of anti-root measures, some are easily bypassed some are stronger, but I'm sure this is one of them.

Is there a way to list all Intents and all processes associated with them? Specifics: Droid X, 2.2 stock ROM, rooted. I'm perfectly fine with any approach, as long as it gets a guaranteed **complete** list: - A script or a series of commands in Terminal Emulator is fine. No limitation on supposed required technical proficiency (assume that the reader is a qualified long-term Unix power user). As noted, the phone is rooted. - An app is OK (ideally, it should be able to export the list, but even Autostarts-like GUI display is OK) Caveat: if you recommend an app (e.g. Autostarts ), I would request some assurance that it does, indeed, list **all** registered Intent receivers, not merely some sub-set - either as documentation/statement from the author or some other evidence (e.g. analysis of which APIs the code uses). - Worst case scenario, I'll take an API specifics

I have a non-rooted phone. I just realized why Android is bad at memory management. I'm monitoring running progress in ADB shell with top. I can see the memory is about 100% used. Image resized. Click for full-size How do I stop unnecessary processes with ADB shell? Also, I can't scroll down top command. It shows weird char ^[[b. I think that's because the shell uses sh instead of bash. Any idea how to free up memory with ADB? Maybe pkill or what.

On my rooted Android, I have a screenshot from Termux after running sudo htop. There is a logcat process running with PID 925. I want to stop this process. So far, I've tried running su -c kill 925 which is the Process ID of the logcat process I'm trying to stop but have been unsuccessful. How to stop this process?

I have an Android device and I want to know either: 1. whether there are currently any apps running in multiple processes 2. whether there are any apps installed with the ability to run in multiple processes Is either of these possible with adb?

There are lots of discussion on how to see which app is requesting network connection. I have tried the following apps: tcapturepacket wifi monitor connection tracker However, none of them was useful. For example, for a short period of time, I uploaded a 30KB pcap file which can be viewed here . Still I haven't find out which app is requesting connection and to where it is trying to connect. I also tried to look at adb logcat, but didn't find useful information. Maybe something has been missed here. Any guess?

I have an app spamming my logcat messages and I assume it is responsible for slow operation of my device. The logs all look like this and there are thousands: 2022-01-18 18:02:19.204 646-646/? E/audit: type=1400 audit(1642345453.765:312963): avc: denied { read } for pid=1173 comm=87654F456789656Z646C1234567890 name="gpuclk" dev="sysfs" ino=74938 scontext=u:r:system_server:s0 tcontext=u:object_r:vendor_sysfs_kgsl_gpuclk:s0 tclass=file permissive=0 SEPF_SV-X825_11_0010 audit_filtered I used adb shell ps -A to list my processes and the spamming PID is the following line: vendor_audit 646 1 2142068 2592 0 0 S auditd Also I checked the other PID mentioned in the log, it's this one: system 1173 673 11627288 282980 0 0 S system_server *Note: I do not know what all these numbers in the log output are, I assume some random device identifiers and similar. Just to be on the safe side, I changed them before uploading, but I kept the length the same. I left the PID numbers* Device is a Galaxy Tab S5e with the preinstalled system, not rooted.

My camera doesn't work on a custom ROM (Android 11), I searched for a fix online, and it says that I have to restart the camera service, they said to run this command on ADB: kill -HUP ps -A | grep camera | awk '{print $2}' but when I hit enter it gave me an error of:

/system/bin/sh: kill: ps: arguments must be jobs or process IDs   
/system/bin/sh: kill: -A: arguments must be jobs or process IDs

--- I ran ps -A | grep camera It outputs

cameraserver    375    1   25432   2584 binder_thread_read a9825cec S android.hardware.camera.provider@2.4-service 
camera          426    1   73104   2432 poll_schedule_timeout aea69eac S mm-qcamera-daemon

Can anyone help me to solve this problem?

How to root camon X (CA7) or application for root access. How can i get root access? Whenever i try it root it, it keeps saying no root access

I noticed that on my Xiaomi 11i with the Snapdragon 888 processor, it is supposed to be one of the current top Android phones, with 8 cores and that only the Apple A14 Bionic has slightly higher performance than it. However, if I run two Pokemon Go using its Dual App split screen, then I noticed, if we look inside a gym, and a Pokemon (Venonat or Exeggutor) is swinging left and right, then every few seconds, on the top or bottom screen, the Pokemon would have a slight 0.2 second pause, so both apps are choppy every few seconds. If I use the app Activity Monitor, right after I boot up the phone so that no other app is running, I did see that every 6 to 10 seconds, all the cores of the processors would ramped up to 100%, so it looks like it is doing some work in the background. I noticed if I run one single instance of Pokemon Go, then it is smoother. But my question is, how can 8 cores not able to handle some slowing animating graphics. If anything, the swinging left and right repeatedly can even be "set it and forget it" -- everything done by the GPU (Adreno 660). So trying to solve it, I tried to tap the Version number of the system for a few times and went to Developer's Options, and from there, I set the background app to be: Running only 1 process at a time. So that should mean I have 7 cores running now for the 2 apps on screen? But after rebooting the phone and trying again, the same issue was still there. Is there a way to make it able to take 2 cores (or 3 cores) for each app, or just make them run at top performance and let any background process run with the remaining 2 cores? Or another way is, if there is a way to set the background processes only use up to 50% of the cores, so the main apps can always have 8 cores to use. **Update:** an insight to finding a solution: I noticed if we tap on a Pokemon and try to catch it, the Pokemon, if it is a flying Pokemon with their wings constantly flapping in the dual app, has really smooth animation and there is no tiny pauses ever. It may appear that when we look inside of a gym, usually six Pokemon are shown inside, so if both apps are looking inside of a gym, there can be 12 Pokemon moving inside. Even if 10 Pokemon have tiny movements only and two Pokemon have more apparent movements, it seemed it was possible that 12 threads were used, and so if there are 12 threads and 8 cores, maybe that's the reason that there can be tiny pauses. I am not sure if we can change the Settings on the phone to solve this, as these movements are very lightweight.

On Android 9 or less, I use a shell script which continuously reads from /proc/net/xt_qtaguid/stats and greps by app UID to get app network usage. However, this method is deprecated on android 10: https://source.android.com/devices/tech/datausage/kernel-overview Is there any other way I can get the same information? Or, how can I use the new eBPF network traffic tool to get this information?

I have an Allwinner A10 tablet clone. While working the tablet "freezes" regularly but not too frequently for a strech of time between 2 and 15 seconds. Described differently: the tablet is useable but the nice "fluid" workflow is abruptly interrupted once in a while... Situations: - e.g. when opening an additional tab in firefox when already having a few open, - when working with the tablet while downloading a file, these freezes are more frequent - etc Using the process overlay option from developer settings I realised that during this time the "nandd" process is consuming CPU almost exclusively. It does not appear to max out the CPU but during this time none of the other processes consume more than minute amounts of CPU time. Can somebody tell me: 1. whether this is "normal" behaviour 2. whether an update to CyanogenMod 10 is likely to help 3. where issue is likely to lie: with slow nand hardware or poor nand driver implementation. Many thanks.

When I plug my phone into my computer, I can see the logcat output in Android Studio. I recently noticed that whenever I start entering a password in a field designated for passwords (the ones that display '*' for any character), the phone prints a very concerning message to the console: D/HwAutofillHelper: inputType is password, now record Since it is an Honor Phone, I suppose Hw is short for Huawei? Nevertheless, I'd like to know if there is a way to find out what process produced this output and what app it belongs to?

I was searching weather a mobile processor be Overclocked or not so i came across an answer saying no because processor operates on PLL(phase locked loop) which if set to a particular frequency can't be changed. But i have heard about Android Overclocking with some custom kernels and rooted devices. So i need some expert opinion on this matter.

I am testing one of my application on android's (Android version 7.0) termux terminal emulator. I want to spawn some zombie processes. On GNU/Linux, if I open a terminal emulator and write:

ruby -e '10.times { fork { exit! } } && sleep '

This will create 10 zombie processes. To check about zombie processes, I write:

ruby -e "puts Dir['/proc/**'].select { |x| File.split(x).then { |y| y.to_i.to_s == y } }.then { |a| %).split == ?Z.freeze} } Zombies)> }"

Which outputs in the format:

Active Processes: 189 (10 Zombies)

But on Termux, I can't create zombie processes. When I am forking a process, the process count doesn't increase (as reported by htop and the ruby program). Even if I try to spawn processes, it won't spawn any. I have total 11 processes showing on Termux! What's so different about android? Is there any way to test my application using termux? Or are there any other applications that will allow me to do such stuff?

I have a rooted device, with superuser rights granted to a few apps I trust. I have often found my /system or /vendor partition mounted read/write, and some files chmoded to 0777. Similarly directories/files with random names are created/deleted (e.g. /sdcard/.a03Degpoif), .nomedia files are created inside them to hide their contents from gallery etc. logcat/dmesg contains no traces of all this happening in background. This doesn't happen periodically but in a slapdash way, internet connectivity isn't a condition either. I have tried disabling/uninstalling apps, looking for any suspicious processes running, but couldn't identify the culprit(s). Is there a way to figure out who is doing this?

In the Settings > Apps screen, under "Running", a few apps are listed as having as active (i.e. a service is running) despite having zero running processes and taking no memory. Why are they in this state? Are they waiting for some kind of event?