Related
Hello,
First of all, I am not attempting to compete with the other awesome images for the G Tablet available. I am doing this process to learn, and maybe contribute to the community.
I have grabbed the AOSP directly from Google's repo tree, and compiled it in two different ways: Generic build, modifying BoardConfig.mk and other files as necessary. I have also used the Device/Vendor files from the Cyanogen Beta 4 harmony repo tree, and compiled a harmony target with AOSP.
When I compiled the generic build (or the the targeted build), I made sure all proprietary files from the tablet had been extracted and replaced in my system.img. (I got this list from the extract-files.sh script in Cyanogen harmony repo).
I have found that using the stock boot.img from the original nvflash files will boot nearly any system.img (from my update.zip in clockwork). I used this original boot.img, combined with my system.img from the AOSP build (with the proprietary files in place), and made an updater-script which installed things. I found that the system.img was properly extracted to /system.
The problem that occurs -- when booting, the Viewsonic bootup screen will load the GTablet screen, but it will eventually loop back to the Viewsonic screen and report "deleting msc" prior to returning to Recovery Mode on it's own.
I have enabled ADB persistence in the boot ramdisk, and it works fine as long as it's not my system.img. When I use the AOSP system.img, adb devices shows the device, but adb shell/logcat fail until the device power cycles.
Does anyone know what I might be missing?
Also want to add:
I have tried the stock boot.img, which lacks a 'cmdline' parameter, and I have tried using the cmdline parameters from the Cyan Harmony boot.img specifications. I am not sure if this might have something to do with the issue or not.
kornyone said:
Hello,
First of all, I am not attempting to compete with the other awesome images for the G Tablet available. I am doing this process to learn, and maybe contribute to the community.
I have grabbed the AOSP directly from Google's repo tree, and compiled it in two different ways: Generic build, modifying BoardConfig.mk and other files as necessary. I have also used the Device/Vendor files from the Cyanogen Beta 4 harmony repo tree, and compiled a harmony target with AOSP.
When I compiled the generic build (or the the targeted build), I made sure all proprietary files from the tablet had been extracted and replaced in my system.img. (I got this list from the extract-files.sh script in Cyanogen harmony repo).
I have found that using the stock boot.img from the original nvflash files will boot nearly any system.img (from my update.zip in clockwork). I used this original boot.img, combined with my system.img from the AOSP build (with the proprietary files in place), and made an updater-script which installed things. I found that the system.img was properly extracted to /system.
The problem that occurs -- when booting, the Viewsonic bootup screen will load the GTablet screen, but it will eventually loop back to the Viewsonic screen and report "deleting msc" prior to returning to Recovery Mode on it's own.
I have enabled ADB persistence in the boot ramdisk, and it works fine as long as it's not my system.img. When I use the AOSP system.img, adb devices shows the device, but adb shell/logcat fail until the device power cycles.
Does anyone know what I might be missing?
Click to expand...
Click to collapse
Did you ever get your AOSP build to boot?
tjohnsonjr said:
Did you ever get your AOSP build to boot?
Click to expand...
Click to collapse
I did. It's in the Dev section now
I've got this side project going where I've decided to try to modify the stock kernel for my GT-I9000M.
I'm a Gentoo linux user and have cross-compiled kernels for embedded systems with or without initramfs and modules successfully many times. Rather than just grab and flash the latest mod from the net I want to understand exactly how to create my own.
I already have access via a chroot to a working Gentoo install, however that runs in a sandbox inside android userland, which can be irritating to say the least. I'm already able to create binaries for armv7a on the phone itself, which is a time and power-consuming novelty.
I setup a development environment on my main computer and made sure I could create working binaries for the armv7a.
I backed up all the bml partitions on my phone.
Next, I made sure I could access Download Mode and experimented with Heimdall until I was satisfied I was able to recover from any bad kernel flashes.
I altered the first boot logo in param.lfs with dd, but that's a whole other story.
Next, I downloaded Samsung's source for Froyo and followed the instructions to create a zImage.
aries_eur_defconfig selects the FM radio module and the Samsung TV out driver which are broken. No problem, I figured I could check that out later so I disabled those options for now. it compiles! great! But it didn't boot. I turned to the internet for ways to solve the problem.
There was a lot of conflicting information. I ended up trying a lot of different approaches, modules, firmware, version info, kernel command lines, fbcon... well I messed around.
Some time later, I find myself happily booting with my own kernel. The bootloader displayed my custom logo in param.lfs, but fbcon didn't work and I got a stock samsung gt-i9000 logo before android booted up! The phone worked, wifi worked, bluetooth worked, it was awesome (for 2.6.32.9 that is )!
So, great! Except I got overconfident and cocky. I made a slew of changes trying to get fbcon to work and to top off the evening I overwrote my working zImage with a broken kernel (without a backup of my working custom zImage or .config). I still have backups otherwise of course and haven't lost the functionality of my phone in any way, but now I'm just irritated.
Anybody else working on compiling from source on the GT-I9000M?
I pressed ahead and tried to recreate the conditions that got me a working kernel and as of yet I've been unsuccessful Luckily I've got a kmsg from the kernel when it booted. I can see the kernel command line on the custom kernel that booted was simply "console=ttySAC2,115200" although I do recall trying different values for init= ... The default init in the sources is /linuxrc which doesn't exist in initramfs.
There are also settings for the kernel command line and version information in arch/arm/mach-s5pv210/include/mach/param.h
Is there a known good method to compile from Samsung's sources?
I unpacked a known good zImage, played with it's initramfs, repacked it and it boots just fine.
I've tried again with a fresh copy of the gt-i9000 froyo-samsung branch:
- unpack initramfs from a known good 2.6.32.9 kernel
- make clean
- make aries_eur_defconfig
- make menuconfig, add initramfs, remove Samsung TV driver and FM radio module
- make
the resulting vmlinux is: ELF 32-bit LSB executable, ARM, version 1 (SYSV), statically linked, not stripped (I know, it's zImage not vmlinux but I'm just doing sanity checks)
I have a custom logo in param.lfs which always displays on boot. A second logo is compiled into the kernel, and I do see a change in the logo when my kernel boots (it shows GT-I9000 instead of GT-I9000M after the first logo). This leads me to believe that the kernel is booting... but the phone powers off after booting.
I've tried the recommended changes to default.prop to use adb to obtain the kmsg of the new kernel. This has not worked however, which could mean that the kernel doesn't get as far as initramfs.
the kernel command line from the kmsg I have from the one time that a custom kernel of mine booted is simply: "console=ttySAC2,115200 loglevel=4" . There's no mention of init, which I think should be init=/init although aries_eur_defconfig seems to think that init=/linuxrc is appropriate (linuxrc doesn't exist?) Regardless, it doesn't seem to matter what I specify for init.
Which leaves proprietary Samsung kernel modules. If they can't load, this could explain everything I'm experiencing. It isn't clear what specific version the proprietary modules are expecting.
Maybe this will help somebody else. I've got it working now.
Attached is my working config.
Hi guys!
Someone requested a tutorial on how to build the kernel source code, so I thought I might as well do it I'll give you a bit more information too.
Maybe this will kick start kernel development a little bit too. Who knows.
Anyway:
1. Set up your build environment as per this guide by AOSP here: http://source.android.com/source/initializing.html
This guide will be primarily aimed at Ubuntu, but it should be easy enough to do on other Linux PCs. Ubuntu is highly recommended. And don't ask about Windows.
2. Download the toolchains: http://code.google.com/p/rohan-kernel-evita/downloads/detail?name=toolchains.tar.gz&can=2
I've hosted both GCC 4.4 and GCC 4.6 (prebuilt) on my Google Code page. These are for Linux.
Download the toolchain package here: http://code.google.com/p/rohan-kernel-evita/downloads/detail?name=toolchains.tar.gz&can=2
Then un-tar the archive and extract it to your home folder so that the path to the toolchain directories are ~/toolchain/whatever-one-you-want
3. At this point you can decide which kernel source you want to download. There are really two sources.
The first choice is getting it from HTC Dev directly. This is good if you want to build your kernel for a Sense 4 ICS ROM. If you want to build for AOSP/JB, then scroll down below all the following HTC stuff.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
To use HTC's source:
Go to HTCdev.com and make an account. Then download the source code for our device (I'd recommend the "One X" source under carrier "AT&T" version "2.20". Let the zip file download and extract it to your home folder (so the path is /home/user/evita-ics..../
To build HTC's source, run these commands:
Code:
cd ~/evita-ics-whatever_the_directory_is_named
export ARCH=arm
make elite_defconfig
Second command is saying what type of architecture we want (we are compiling for ARM processors, so we want ARM)
Third command is saying to make the default config for our device (whose hardware is codenamed "elite")
Then to build the actual kernel:
Code:
make -j# ARCH=arm CROSS_COMPILE=~/toolchain/arm-eabi-4.4.3/bin/arm-eabi-
In the command above, there is "-j#". Replace the "#" with the number of CPUs you have.
Ask me about it if you need help. That should be it! Let the build go and in a few minutes you should have a zImage file located at ~/evita-ics-..../arch/arm/boot. That is the actual kernel.
To test out the zImage (kernel), connect your device via fastboot mode and type:
Code:
fastboot boot /path/to/zImage
Combined with the ramdisk, thats what makes the boot.img file. Eventually if you want to distribute your kernel, you should use an installer that injects the zImage into the boot.img or combine the zImage with a ramdisk to make a boot.img. I can do a tutorial on that later on as well.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
To get the "other" source:
This source is based off of HTC's source and includes other devices as well in it (the One S and Evo 4G LTE). Its good if you want to build JB AOSP kernels.
To download this kernel source do this:
Code:
mkdir ~/kernel
cd ~/kernel
git clone https://github.com/CyanogenMod/android_kernel_htc_msm8960.git -b android-msm-evita-3.0
This version also includes a number of optimizations and odd bug fixes present in the HTC version.
To build this second source, run these commands:
Code:
cd ~/kernel/android_kernel_htc_msm8960
export ARCH=arm
make elite_defconfig
Second command is saying what type of architecture we want (we are compiling for ARM processors, so we want ARM)
Third command is saying to make the default config for our device (whose hardware is codenamed "elite")
Then to build the actual kernel:
Code:
make -j# ARCH=arm CROSS_COMPILE=~/toolchain/arm-linux-androideabi-4.6/bin/arm-eabi-
In the command above, there is "-j#". Replace the "#" with the number of CPUs you have.
Ask me about it if you need help. That should be it! Let the build go and in a few minutes you should have a zImage file located at ~/evita-ics-..../arch/arm/boot. That is the actual kernel.
To test out the zImage (kernel), connect your device via fastboot mode and type:
Code:
fastboot boot /path/to/zImage
Combined with the ramdisk, thats what makes the boot.img file. Eventually if you want to distribute your kernel, you should use an installer that injects the zImage into the boot.img or combine the zImage with a ramdisk to make a boot.img. I can do a tutorial on that later on as well.
-----------------------------------------------------------------------------
The second option is also a bit better as it uses the 4.6 toolchain instead of GCC 4.4. HTC's source doesn't work with GCC 4.6 so it can't be used without changes that the second source has.
If you have any questions, feel free to post here, message me on twitter (@rohanXm), chat me on IRC (#HTC-One-XL) or PM me!
If this helped you, please consider hitting the donation link under my username on the left. Donations are never required but always appreciated.
Instead of downloading cm10 the readme inclided with the HTC source has directions for getting a tool chain which will compile the source.
Sent from my HTC One X using Tapatalk 2
Perfect. Now I just have to read.
Sent from my HTC One XL bumping it
rohan32 said:
Code:
repo init -u git://github.com/CyanogenMod/android.git -b jb
Click to expand...
Click to collapse
FYI, It looks like the branch name has changed. When I changed "jb" to "jellybean" the repo init command worked.
Rohan. You are one bad MF'er
Sent from my twin turbo'ed OneXL rocking rezound beats
rohan32 said:
2. You can either try to find a standalone package of the precompiled toolchain, or you are going to need to download a ROMs source. I'd recommend downloading a ROMs source since I've never found a good toolchain that worked for me. If you find one that works, post below
For now we will download CM10 since that seems like the defacto standard.
Click to expand...
Click to collapse
I downloaded the 2.20.502.7 kernel source and when I extracted it there was a file named evita_readme.txt which gave another, possibly more "official", location for downloading a toolchain:
--Please follow below command to download the official android toolchain: (arm-eabi-4.4.3)
git clone https://android.googlesource.com/platform/prebuilt
Click to expand...
Click to collapse
I just performed a build with this toolchain but got this error when I attempted to load zImage via fastboot:
c:\>fastboot flash boot zImage
sending 'boot' (5140 KB)...
OKAY [ 1.044s]
writing 'boot'...
FAILED (remote: image error! (BootMagic check fail))
finished. total time: 1.077s
Click to expand...
Click to collapse
I'm not sure if this is caused by the toolchain or if I screwed something up. Have you ever seen this error before?
EDIT: Ok I see what I did incorrectly. The zImage needs to be "Combined with the ramdisk". You wouldn't know how to perform this operation... would you?
denversc said:
I downloaded the 2.20.502.7 kernel source and when I extracted it there was a file named evita_readme.txt which gave another, possibly more "official", location for downloading a toolchain:
I just performed a build with this toolchain but got this error when I attempted to load zImage via fastboot:
I'm not sure if this is caused by the toolchain or if I screwed something up. Have you ever seen this error before?
EDIT: Ok I see what I did incorrectly. The zImage needs to be "Combined with the ramdisk". You wouldn't know how to perform this operation... would you?
Click to expand...
Click to collapse
You are trying to flash a zImage
To my knowledge, only booting zImages work on this device. For the time being just boot the zImage (use fastboot boot zImage)
I will make a tutorial on how to combine the zImage created with a ramdisk to make a boot.img when I get the chance
denversc said:
FYI, It looks like the branch name has changed. When I changed "jb" to "jellybean" the repo init command worked.
Click to expand...
Click to collapse
Oops, my bad! That was a mistake. AOKP uses -jb and CM uses -jellybean so I got them switched fixed now
Out of any device I've seen more people actually interested in helping and learning to help then any other community. I've watched noobs become less noobish.. I've watched skizz learn how to make themes. Hell I've learned 10 fold what I knew before this phone myself.
That is beautiful, and now such an informed, helpful post such as this.
Have great Sunday you guys! I think I might give this a shot!
Sent from my One X
I'll streamline this process in a bit, so that you don't need to download CM10
Edit: cleaned up post, now I'm uploading just the toolchains. Its a tar.gz archive, around 150mb. Much better than downloading the entire CM10 source
rohan32 said:
You are trying to flash a zImage. To my knowledge, only booting zImages work on this device. For the time being just boot the zImage (use fastboot boot zImage)
Click to expand...
Click to collapse
Thanks for your response, rohan. You are absolutely right: I was incorrectly attempting to flash zImage straight to the boot partition, and the error produced by flashboot was justified. I have since successfully "tested out" my compiled zImage by flashing it via "fastboot boot zImage" and it worked like a charm! I ported the modifications from sbryan's Blackout BeastMode kernel and I am now able to OC to 2106 MHz and UC to 192 MHz. It's been running solid for the past few hours at least
Of course, the kernel reverts back to the one stored in the boot partition after a reboot, and I want my shiny new kernel to "stick". I've done a bit of research on this topic and found some information about combining my zImage with a ramdisk into a "real" boot.img but have not yet been successful in creating a boot.img which does not bootloop after flashing it.
For example, I found an article on xda called Basic Kernel Kitchen for Minor Kernel Tweaking which points to a "kitchen sink" tool for creating a boot.img from a zImage and a ramdisk. My problem is that I don't know where to get or how to make a ramdisk. So I tried using the ramdisk from the boot.img of the ROM that I am currently running (CleanROM 4.5 DE) but just got into a bootloop. I've since been doing some yard work today so haven't gotten back to investigating further.
I also found another program named abootimg which can also produce a boot.img from a zImage abd a ramdisk, but when I tried it an error message about my zImage being "too big" was produced.
If it's not obvious yet, I am kind of fumbling around in the dark as compiling and deploying custom kernels is completely new to me! But this post was the most valuable resource I've come across in getting to this point. Thanks so much for writing it! I eagerly await your next article about creating the boot.img
---------- Post added at 06:05 PM ---------- Previous post was at 05:50 PM ----------
rohan32 said:
cleaned up post
Click to expand...
Click to collapse
Thanks for cleaning up the post rohan. I have a few follow-up questions/comments:
missing toolchains link -- the post says "Download the toolchains:" but there is no link to download anything... did you forget to paste the link?
official toolchain -- That's great that you uploaded the toolschains to save tonnes of bandwidth. The CM10 source was taking a VERY long time for me to grab. However, I imagine some people (like me) would prefer to get the toolchain from an "official" source. In the "evita_readme.txt" file of the kernel sources downloaded from HTC dev it instructs one to use the official sources from "git clone https://android.googlesource.com/platform/prebuilt". This is the toolchain that I used and it successfully built the zImage.
ko files -- I noticed in the ZIP file for Blackout BeastMode, in additional to installing the zImage it also puts a bunch of "ko" (kernel modules I believe) into the /system partition (eg. qce40.ko). Should I also be deploying .ko files from my build to the device?
Thanks!
denversc said:
Thanks for your response, rohan. You are absolutely right: I was incorrectly attempting to flash zImage straight to the boot partition, and the error produced by flashboot was justified. I have since successfully "tested out" my compiled zImage by flashing it via "fastboot boot zImage" and it worked like a charm! I ported the modifications from sbryan's Blackout BeastMode kernel and I am now able to OC to 2106 MHz and UC to 192 MHz. It's been running solid for the past few hours at least
Of course, the kernel reverts back to the one stored in the boot partition after a reboot, and I want my shiny new kernel to "stick". I've done a bit of research on this topic and found some information about combining my zImage with a ramdisk into a "real" boot.img but have not yet been successful in creating a boot.img which does not bootloop after flashing it.
For example, I found an article on xda called Basic Kernel Kitchen for Minor Kernel Tweaking which points to a "kitchen sink" tool for creating a boot.img from a zImage and a ramdisk. My problem is that I don't know where to get or how to make a ramdisk. So I tried using the ramdisk from the boot.img of the ROM that I am currently running (CleanROM 4.5 DE) but just got into a bootloop. I've since been doing some yard work today so haven't gotten back to investigating further.
I also found another program named abootimg which can also produce a boot.img from a zImage abd a ramdisk, but when I tried it an error message about my zImage being "too big" was produced.
If it's not obvious yet, I am kind of fumbling around in the dark as compiling and deploying custom kernels is completely new to me! But this post was the most valuable resource I've come across in getting to this point. Thanks so much for writing it! I eagerly await your next article about creating the boot.img
---------- Post added at 06:05 PM ---------- Previous post was at 05:50 PM ----------
Thanks for cleaning up the post rohan. I have a few follow-up questions/comments:
missing toolchains link -- the post says "Download the toolchains:" but there is no link to download anything... did you forget to paste the link?
official toolchain -- That's great that you uploaded the toolschains to save tonnes of bandwidth. The CM10 source was taking a VERY long time for me to grab. However, I imagine some people (like me) would prefer to get the toolchain from an "official" source. In the "evita_readme.txt" file of the kernel sources downloaded from HTC dev it instructs one to use the official sources from "git clone https://android.googlesource.com/platform/prebuilt". This is the toolchain that I used and it successfully built the zImage.
ko files -- I noticed in the ZIP file for Blackout BeastMode, in additional to installing the zImage it also puts a bunch of "ko" (kernel modules I believe) into the /system partition (eg. qce40.ko). Should I also be deploying .ko files from my build to the device?
Thanks!
Click to expand...
Click to collapse
Hey!
Sorry, set it to upload then got distracted Link posted
The reason why your boot.imgs bootloop is because there is a special ramdisk address that needs to be set when combining the ramdisk with the zImage, and most kitchens don't support this. You also need to set the address after setting the base value.
I will post a guide on how to make it a real boot.img whenever I get the chance.
I'm telling you people, Rohan is a BOSS. Most helpful dev I know.
Sent from my HTC One XL using xda app-developers app
rohan32 said:
Hey!
Sorry, set it to upload then got distracted Link posted
The reason why your boot.imgs bootloop is because there is a special ramdisk address that needs to be set when combining the ramdisk with the zImage, and most kitchens don't support this. You also need to set the address after setting the base value.
I will post a guide on how to make it a real boot.img whenever I get the chance.
Click to expand...
Click to collapse
I have the zImage thanks to your tutorial, but I want to know how to create the kernel zip, whenever you have time will be amazing if you can post a guide, I'm really looking forward to that guide, because i haven't been able to locate a guide that works
Sent from my HTC One XL using xda premium
rohan32 said:
Hi guys!
Second command is saying what type of architecture we want (we are compiling for ARM processors, so we want ARM)
Third command is saying to make the default config for our device (whose hardware is codenamed "elite")
Then to build the actual kernel:
Code:
make -j# ARCH=arm CROSS_COMPILE=~/toolchain/arm-linux-androideabi-4.6/bin/arm-eabi-
Click to expand...
Click to collapse
This is the correct make for "other source" kernel.
Code:
make -j# ARCH=arm CROSS_COMPILE=~/toolchain/arm-linux-androideabi-4.6/bin/[COLOR="Red"]arm-linux-androideabi-[/COLOR]
Can you make a tutorial on how to insert governors into a kernel?
Compiling problem
Hey man,
Please help me I followed your article but when I try copile with:
make -j2 ARCH=arm CROSS_COMPILE=~/toolchain/arm-eabi-4.4.3/bin/arm-eabi
I got something like this:
/home/martin/toolchain/arm-eabi-4.4.3/bin/../lib/gcc/arm-eabi/4.4.3/../../../../arm-eabi/bin/as: error while loading shared libraries: libz.so.1: cannot open shared object file: No such file or directory
Thanks
UPDATE:
now its work
I went deeper and install lib32z1 with "sudo apt-get install lib32z1"... now its work
My device repositories are not available on github, But I got device tree and vendor blobs by making changes in similar device repo. That reference device's kernel's lineageos_defconfig is situated in htc msm8974 kernel repo. So how can I get lineageos_defconfig for my device, and which other my device related kernel files(.dtsi or any other) I have to push in htc msm8974 repo and get those files to make things ready for build?
Please help......
I'm looking for a stock kernel that only patches Safetynet checking. Does this exist? If not, is it easy for me to "make it" myself?
I'm not sure if there's a prebuilt one, but building one yourself isn't too hard. The patch is at https://github.com/sultanxda/androi...bc05b16bbd33521c2fffaf491c5657a94bfcfc5.patch. You just follow the steps at http://source.android.com/source/building-kernels.html as usual with the following notes:
Use "kernel/msm" as the source location
Use "marlin_defconfig" as the build configuration
Apply the patch after running the git checkout command
Use the aarch64 prebuilts, not the arm ones
Cares said:
I'm looking for a stock kernel that only patches Safetynet checking. Does this exist? If not, is it easy for me to "make it" myself?
Click to expand...
Click to collapse
Depending on your motivation for a "stock kernel" you might try franco kernel. He doesn't seem to do anything that MIGHT introduce instability or strays very far from stock.
I can vouch for franco. He does minimal performance-only tweaks by default.
josephcsible said:
I'm not sure if there's a prebuilt one, but building one yourself isn't too hard. The patch is at https://github.com/sultanxda/androi...bc05b16bbd33521c2fffaf491c5657a94bfcfc5.patch. You just follow the steps at http://source.android.com/source/building-kernels.html as usual with the following notes:
Use "kernel/msm" as the source location
Use "marlin_defconfig" as the build configuration
Apply the patch after running the git checkout command
Use the aarch64 prebuilts, not the arm ones
Click to expand...
Click to collapse
Hey thanks for the tip. I went ahead and patched the no safetynet patch to the android-msm-marlin-3.18-nougat-mr1 kernel source and compiled it. I now have a Image.gz-dtb file which I zipped (I also just have a binary file named "Image"). What should I with those now, just flash those like I would one of the other kernels? And which file exactly? The gz file? or the just binary file named "Image"?
So essentially "fastboot flash kernel <file_name>"?
When I was compiling I got two warnings by the way:
drivers/soc/qcom/Kconfig:371:warning choice value used outside its choice group
drivers/soc/qcom/Kconfig:376:warning choice value used outside its choice group
Anything I should be concerned about? I've never done this before, but did a lot of reading before I went ahead and used to do some C coding back in the day, so it's not completely unknown to me.
Essentially, these are the steps I followed, after quickly installing Linux Mint:
Code:
Create a working directory in /home/$USER/ (I created /home/sakete/android)
Enter working directory
Download android kernal source
git clone https://android.googlesource.com/kernel/msm
Download prebuilt toolchain
git clone https://android.googlesource.com/platform/prebuilts/gcc/linux-x86/aarch64/aarch64-linux-android-4.9
cd aarch64-linux-android-4.9
export PATH=$(pwd)/prebuilts/gcc/linux-x86/aarch64/aarch64-linux-android-4.9
export CROSS_COMPILE=$(pwd)/bin/aarch64-linux-android-
export ARCH=arm64
export SUBARCH=arm64
Checkout specific kernel branch for Pixel/PixelXL (be in 'msm' folder)
git checkout android-msm-marlin-3.18-nougat-mr1
Get Safetynet Patch (still be in 'msm' folder)
git fetch https://github.com/sultanxda/android_kernel_oneplus_msm8996 cm-13.0-sultan
git cherry-pick abc05b16bbd33521c2fffaf491c5657a94bfcfc5
Build kernel (still be in 'msm' folder)
make clean
make mrproper
make marlin_defconfig
make -j$(grep -c ^processor /proc/cpuinfo)
I initially got some build errors, but running this command solved it: sudo apt-get install build-essential
This is a useful link for those of you who are interested in doing this as well: http://forum.xda-developers.com/showpost.php?p=69627576&postcount=7108
Hey if you can build it and post it here, that would be awesome. This is exactly what I'm looking for.
It will let me preemptively unlock my Verizon bootloader before flashing the latest OTA, while keeping Android Pay.
Has Google already posted the source for the 7.1.1 update kernel?
The source I pulled (android-msm-marlin-3.18-nougat-mr1) should be what's in the December update. It should be 7.1.1
Sakete said:
Hey thanks for the tip. I went ahead and patched the no safetynet patch to the android-msm-marlin-3.18-nougat-mr1 kernel source and compiled it. I now have a Image.gz-dtb file which I zipped (I also just have a binary file named "Image"). What should I with those now, just flash those like I would one of the other kernels? And which file exactly? The gz file? or the just binary file named "Image"?
Click to expand...
Click to collapse
The Image.gz-dtb file is the one you want.
Sakete said:
So essentially "fastboot flash kernel <file_name>"?
Click to expand...
Click to collapse
I've never done it like that, but that's apparently how Franco's kernel installs, so it's worth a shot I guess. Another way of doing it is to unpack the stock boot image with either pbatard's unmkbootimg or osm0sis's Android Image Kitchen, replace its kernel with your Image.gz-dtb, then repack and flash the new boot.img to the boot partitions.
Sakete said:
When I was compiling I got two warnings by the way:
drivers/soc/qcom/Kconfig:371:warning choice value used outside its choice group
drivers/soc/qcom/Kconfig:376:warning choice value used outside its choice group
Anything I should be concerned about? I've never done this before, but did a lot of reading before I went ahead and used to do some C coding back in the day, so it's not completely unknown to me.
Click to expand...
Click to collapse
Nothing you need to worry about.
josephcsible said:
The Image.gz-dtb file is the one you want.
I've never done it like that, but that's apparently how Franco's kernel installs, so it's worth a shot I guess. Another way of doing it is to unpack the stock boot image with either pbatard's unmkbootimg or osm0sis's Android Image Kitchen, replace its kernel with your Image.gz-dtb, then repack and flash the new boot.img to the boot partitions.
Nothing you need to worry about.
Click to expand...
Click to collapse
Great, thanks, I'll try flashing the kernel tomorrow night and will post it if successful.
Interestingly it seems that Pixel (sailfish) and Pixel XL (marlin) use the same kernel / kernel source? There at least doesn't seem to be a sailfish specific source. Will be interesting to see how it pans out tomorrow.
Sakete said:
Interestingly it seems that Pixel (sailfish) and Pixel XL (marlin) use the same kernel / kernel source? There at least doesn't seem to be a sailfish specific source. Will be interesting to see how it pans out tomorrow.
Click to expand...
Click to collapse
Indeed. The same kernel binary can run on both devices. (The ElementalX and Franco kernels don't even have separate builds for the two.)
Would you mind posting the image you built?
iPwn_ said:
Would you mind posting the image you built?
Click to expand...
Click to collapse
I'm at work now, will post it tonight.
Sent from my Pixel using Tapatalk
Stock Kernel + SafetyNet Patch applied
Well holy crap, it actually worked! Just flashed the kernel, set up android pay no problem! And everything else works just fine too.
Attached is a zip.
Steps to install (make sure you have adb and fastboot set up):
- Download file and unzip
- Reboot into bootloader (power down, hold Power + Volume Down)
- Attach device to computer
- Enter command: fastboot flash kernel <kernel_image>
- Enter command: fastboot reboot
- Disconnect device and wait for it to finish booting. That's it!
You're my hero.
Might be a lot to ask, but it would be dope if you maintained where you update the build every month for Google's latest release.
iPwn_ said:
You're my hero.
Might be a lot to ask, but it would be dope if you maintained where you update the build every month for Google's latest release.
Click to expand...
Click to collapse
I'm creating a thread in the dev section.
Edit: thread is up.
Cares said:
I'm looking for a stock kernel that only patches Safetynet checking. Does this exist? If not, is it easy for me to "make it" myself?
Click to expand...
Click to collapse
A bit late, but just for the record, ElementalX is just like stock with added features. If you don't use those features, you are essentially using the stock kernel.
I am thinking about going this route.. but I am not sure the process to flash a custom kernel on my Pixel.. would anyone be able to walk me through it? thanks!
Introduction
This thread contains the LineageOS 17.1 custom firmware images for the Unihertz Atom L, a rugged Android phone released by Unihertz in July 2020, and the accompanying LineageOS Recovery used for flashing the firmware.
Please note that this ROM is one of my side projects, for which I could provide zero warranty. By installing this ROM, you acknowledge that you take all the risks that come with installing custom firmwares on your devices, including but not limited to bricking your device, losing your data, etc. You are always suggested to keep backups and make sure you know how to flash back to official ROM before trying any custom ROMs.
Please find the download links in the Download section. The following sections are guides to installing the ROM.
WARNING: DO NOT try to install this on Atom XL. This is ONLY for the Atom L.
Working Features
- All basic features (Telephony, VoLTE, Audio, Camera, NFC, WiFi, Bluetooth, ....)
- Programmable PTT (red) button (Functionality can be set in Settings - System - Buttons, under the "Search Button" section)
- 48MP camera seems to be working (unlike on many other super resolution devices)
Known Issues
- VoLTE is working (at least for me) but sometimes quirky. If you find it somehow stopped working, usually turning it off and back on again (in Settings - Network - Mobile Network) will fix it. Putting the device to SELinux Permissive mode also fixes most of the VoLTE quirks but this is not recommended (a few quirks in Enforcing mode is better than having the whole device Permissive)
Unlocking
1. Boot your Atom L to the official OS
2. Go into Settings - About phone, tap "build number" several times to enable developer settings
3. Go to Settings - System - Developer Settings, enable OEM unlocking and ADB debugging
4. Run `adb reboot bootloader` on your PC (there is no way to enter bootloader directly, only possible through adb)
5. Run `adb flashing unlock` and comfirm unlock on device (THIS WILL WIPE ALL DATA)
6. Reboot and now you should see an unlocked warning during boot screen.
Installing LineageOS Recovery
For now the only working recovery is the LineageOS Recovery, because the device's kernel does not load the touch driver in recovery mode for whatever reason, rendering TWRP useless.
1. Download `lineage_recovery_XXX.img` and `vbmeta.zip`, unpack `vbmeta.zip` to get three .img files starting with `vbmeta`
2. Run `adb reboot bootloader` to put your device in bootloader mode
3. Run `fastboot flash --disable-verification --disable-verity vbmeta vbmeta.img`
4. Run `fastboot flash --disable-verification --disable-verity vbmeta_system vbmeta_system.img`
5. Run `fastboot flash --disable-verification --disable-verity vbmeta_vendor vbmeta_vendor.img`
6. Run `fastboot flash recovery lineage_recovery_XXX.img`
7. Run `fastboot reboot recovery` to reboot into the newly-installed LineageOS Recovery
The LineageOS Recovery is operated by volume keys as selection and power as confirmation (or entering sub-menus). To return to upper levels of menus from sub-menus, press volume up until the selection goes to the first item and then disappears, then press power (i.e. there's a hidden "Go Back" item at the very top of each sub-menu).
The recovery will show a verification failed prompt for most packages that are not signed with the AOSP keys. This is safe to ignore.
Installing LineageOS 17.1
The LineageOS image must be installed via LineageOS recovery.
1. Download `lineage-17.1-Atom_L-XXX.zip`
2. Reboot your device into recovery (`adb reboot recovery` or simply hold volume up while turning power on)
3. Wipe all data (factory reset) (THIS DELETES EVEN INTERNAL STORAGE)
4. Choose Apply Update, then Apply Update from ADB
5. Run `adb sideload lineage-17.1-Atom_L-XXX.zip` from your PC
6. Wait for the process to finish. (The recovery might prompt something about verification failure, just ignore it and continue anyway)
7. At this point, you can then sideload the LATEST Magisk and OpenGAPPS Nano at your will (note that the size of the system partition might only be enough for the `nano` variant of OpenGAPPS) (If installing Magisk / OpenGAPPS fails, you can try rebooting into recovery again in advanced menus, then try installing them again)
8. Reboot into system and enjoy (Note that Magisk might cause your device to boot loop once or two but it will eventually boot)
When updating to a newer build, you have to flash the new zip, and then re-flash whatever mod you have installed previously (Magisk / GAPPS).
Download Links
LineageOS:
lineage-17.1-Atom_L-20200828-peter-signed.zip: https://mega.nz/file/bAgh1BZA#jzMs_0e9NUR9NcALXWp51ZeWttM5rl_3K5T8Or9hAW0
- Synchronized updates from LineageOS upstream.
lineage-17.1-Atom_L-20200728-peter-signed.zip: https://mega.nz/file/vBwlmL5D#wpw8RovBHyVFCLFlhQ2H5QAIb0ECXkT4of0FRijiP6A
LineageOS Recovery:
lineage_recovery_20200728.img: https://mega.nz/file/yc4Dnbyb#yx0Ci9p3q9_lfAiXkGfgWDFnRJI-JSGrv3kyawkU3fw
vbmeta:
vbmeta.zip: https://mega.nz/file/nF51mBoY#ZNY4j92wc_6a1dXch3l5r-w4VFl9QjN7YJaRMKRoEGk
XDA:DevDB Information
LineageOS 17.1 for Unihertz Atom L, ROM for the Android General
Contributors
PeterCxy
Source Code: https://cgit.typeblog.net/android/device/unihertz/Atom_L/
ROM OS Version: Android 10
Version Information
Status: Alpha
Created 2020-07-28
Last Updated 2020-07-28
How different is the Atom XL?
PeterCxy said:
Introduction
WARNING: DO NOT try to install this on Atom XL. This is ONLY for the Atom L.
Unfortunately I've got the XL version which I thought only varied from the L by the presence of a UHF radio! Can you explain to me why its not a suitable candidate for your mods which sound very good!?
And before you ask, I only got this radio for hacking so I don't mind experimenting if that is required. Please let me know if I can help.
The Bitfarmer
Click to expand...
Click to collapse
tvroman said:
PeterCxy said:
Introduction
WARNING: DO NOT try to install this on Atom XL. This is ONLY for the Atom L.
Unfortunately I've got the XL version which I thought only varied from the L by the presence of a UHF radio! Can you explain to me why its not a suitable candidate for your mods which sound very good!?
And before you ask, I only got this radio for hacking so I don't mind experimenting if that is required. Please let me know if I can help.
The Bitfarmer
Click to expand...
Click to collapse
Because Unihertz publishes completely different firmware files for the L and XL, so the safest assumption is that there is more difference than just the UHF radio. If you want to risk it, then you CAN try using this ROM on the XL, as long as you know how to revert back to official if things go wrong. (But I cannot guarantee if the kernel image from L that this ROM uses will not cause serious issues like corrupted baseband or something on the XL)
My suggestion is that instead of trying this ROM directly on the XL, someone with XL can try to modify my device tree for L, replacing the kernel, dtbo images and other vendor blobs from the ones from XL, and then re-compile the ROM for XL. This would be the proper way to handle these two devices.
Click to expand...
Click to collapse
Going XL
Hi.
Great work. :good:
I want to built a ROM for the Atom XL myself. And because I'm no expert on this (for now) I'm in search of guides and hints on how to achieve my goal.
As far as I know the biggest problem with Unihertz is that they use a Mediatek chipset with which they are not allowed to provide the sourcecode of the kernel. Or at least you have to pay for it from Mediatek.
But there are some variants of the chipset (Helio P60; mt6771) used in other mobile phones (e.g. Nokia X5) for which I was able to find kernelsources on Github. Using these and the latest Android kernel from google I tried to compile a kernel as a starting point. I was able to extract the build.config directly from the phone which helped tremendously. This should at least get me to the point where I'm able to assemble a TWRP build. But I believe that I'm still missing some (vital?) drivers which are specific to the actual device. This includes I think the missing touchscreen driver that you mentioned is preventing the recovery to be useful.
So now I'm a little bit stuck, because most of the guides to arrange a LineageOS (or any other custom ROM) build tree I found require the sourcecode from the manufacturer which we don't have. All other guides to build from scratch were too generic for my current level of expertise.
Can you please share your approach to create this build?
If you don't want to do this in the open you could also PM me.
With kind regards
ADT
a-dead-trousers said:
Hi.
Great work. :good:
I want to built a ROM for the Atom XL myself. And because I'm no expert on this (for now) I'm in search of guides and hints on how to achieve my goal.
As far as I know the biggest problem with Unihertz is that they use a Mediatek chipset with which they are not allowed to provide the sourcecode of the kernel. Or at least you have to pay for it from Mediatek.
But there are some variants of the chipset (Helio P60; mt6771) used in other mobile phones (e.g. Nokia X5) for which I was able to find kernelsources on Github. Using these and the latest Android kernel from google I tried to compile a kernel as a starting point. I was able to extract the build.config directly from the phone which helped tremendously. This should at least get me to the point where I'm able to assemble a TWRP build. But I believe that I'm still missing some (vital?) drivers which are specific to the actual device. This includes I think the missing touchscreen driver that you mentioned is preventing the recovery to be useful.
So now I'm a little bit stuck, because most of the guides to arrange a LineageOS (or any other custom ROM) build tree I found require the sourcecode from the manufacturer which we don't have. All other guides to build from scratch were too generic for my current level of expertise.
Can you please share your approach to create this build?
If you don't want to do this in the open you could also PM me.
With kind regards
ADT
Click to expand...
Click to collapse
You don't need the kernel source code to build a working ROM -- just look at my device tree for Atom L. I think you can build a working ROM for the XL by just replacing the prebuilt kernel in my device tree with the one from Atom XL and also re-extracting the vendor blobs from XL using the script in my devcie tree, then rename everything to Atom XL instead of L. I don't know if the integrated amateur radio would still work though.
PeterCxy said:
You don't need the kernel source code to build a working ROM -- just look at my device tree for Atom L. I think you can build a working ROM for the XL by just replacing the prebuilt kernel in my device tree with the one from Atom XL and also re-extracting the vendor blobs from XL using the script in my devcie tree, then rename everything to Atom XL instead of L. I don't know if the integrated amateur radio would still work though.
Click to expand...
Click to collapse
I'm already on to that.
But I seem to have trouble extracting the prebuilt kernel. None of the tools I found gave me the exact files you have got (dtb.img, dtbo.img, Image.gz). What did you use?
The best I could get were "dtb", "kernel" and "dtborecovery" (without extensions) which roughly had the same size as yours.
Also, as far as I understand it, with your initial commit (without the modifications for Lineage itself) I should be able to at least compile a recovery image but I got an error regarding a missing dtb.img file in the "out" directory.
Something seems to be missing because, my dtb file is in the "device" directory and not being transfered into "out" during building.
I'm not sure that is because I have got a different naming scheme (renamig it didn't help) or I did something wrong with the extraction.
---------- Post added at 07:30 ---------- Previous post was at 07:14 ----------
Another question I have:
Are the vbmeta-files you used to flash the recovery the ones from the original firmeware zip from unihertz or did you get them from the lineage built?
And reguarding the rather smallish system partition:
I have an idea to bypass that by using the SPFlash Tool from Mediatek. As far as I understand the settings in the scatter-file this tool does a repartitioning of the internal storage. So we only need to "decrease" the userdata, "move" some partitions inbetween and "increase" the system. Only problem is, I couldn't find a partition designated as "system" in the scatter-file, only one big "super" and a "vbmeta-system" (which for my understaning is for verified boot) partition.
What do you think?
a-dead-trousers said:
I'm already on to that.
But I seem to have trouble extracting the prebuilt kernel. None of the tools I found gave me the exact files you have got (dtb.img, dtbo.img, Image.gz). What did you use?
The best I could get were "dtb", "kernel" and "dtborecovery" (without extensions) which roughly had the same size as yours.
Also, as far as I understand it, with your initial commit (without the modifications for Lineage itself) I should be able to at least compile a recovery image but I got an error regarding a missing dtb.img file in the "out" directory.
Something seems to be missing because, my dtb file is in the "device" directory and not being transfered into "out" during building.
I'm not sure that is because I have got a different naming scheme (renamig it didn't help) or I did something wrong with the extraction.
---------- Post added at 07:30 ---------- Previous post was at 07:14 ----------
Another question I have:
Are the vbmeta-files you used to flash the recovery the ones from the original firmeware zip from unihertz or did you get them from the lineage built?
And reguarding the rather smallish system partition:
I have an idea to bypass that by using the SPFlash Tool from Mediatek. As far as I understand the settings in the scatter-file this tool does a repartitioning of the internal storage. So we only need to "decrease" the userdata, "move" some partitions inbetween and "increase" the system. Only problem is, I couldn't find a partition designated as "system" in the scatter-file, only one big "super" and a "vbmeta-system" (which for my understaning is for verified boot) partition.
What do you think?
Click to expand...
Click to collapse
> None of the tools I found gave me the exact files you have got (dtb.img, dtbo.img, Image.gz). What did you use?
There is a tool called `unpack_bootimg` in the Android source code. Just run `make unpack_bootimg` in the root directory of the Android source tree and you will get one in the output directory. (btw I have renamed those extracted files so the names won't exactly match, but you need this tool to extract the correct images. All other tools won't work properly).
> my dtb file is in the "device" directory and not being transfered into "out" during building.
Because most tools other than `unpack_bootimg` extracts dtb incorrectly.
> Are the vbmeta-files you used to flash the recovery the ones from the original firmeware zip from unihertz or did you get them from the lineage built?
Those don't matter. Either will work as long as you flash it with the correct parameters as given in my post.
> And reguarding the rather smallish system partition
No don't do that. Android 10 does not use a separate system partition anymore, instead both system, vendor and product are sub-partitions in a huge super partition. When flashing a new ROM, the partitions are automatically resized to match the new image exactly, instead of leaving free space unused like before Android 10. That's why I need to reserve space in BoardConfig.mk for gapps to be installed correctly.
Still not able to build.
PeterCxy said:
There is a tool called `unpack_bootimg` in the Android source code. Just run `make unpack_bootimg` in the root directory of the Android source tree and you will get one in the output directory. (btw I have renamed those extracted files so the names won't exactly match, but you need this tool to extract the correct images. All other tools won't work properly).
Click to expand...
Click to collapse
I'm still getting an error:
Code:
FAILED: ninja: 'out/target/product/Atom_XL/dtb.img', needed by 'out/target/product/Atom_XL/boot.img', missing and no known rule to make it
Comparing your BoardConfig.mk with mine shows a slight difference in the offset and size values which could be associated with the different kernels of the phones.
But using "unpack_bootimg" I didn't get a value for "BOARD_KERNEL_OFFSET" like you have it in your config. Could this be the problem?
Your BoardConfig.mk
My BoardConfig.mk
Do you see anything else out of the ordinary?
(Because I'm doing everything what you did step-by-step the links point to the best matching commits)
Despite not being able to compile right now I tried to press on with integrating your changes in the hopes that it will be fixed somehow later on
So I'm currently stuck on this commit of yours:
Atom_L: import overlay from official vendor
Where did you get the "config.xml" and "power_profile.xml" from? The best thing I could find was a "power_profile.xml" inside "/vendor/overlay/FrameworkResOverlay/FrameworkResOverlay.apk" which seems to be a "compiled" version of the aforementioned xml-file.
a-dead-trousers said:
I'm still getting an error:
Code:
FAILED: ninja: 'out/target/product/Atom_XL/dtb.img', needed by 'out/target/product/Atom_XL/boot.img', missing and no known rule to make it
Comparing your BoardConfig.mk with mine shows a slight difference in the offset and size values which could be associated with the different kernels of the phones.
But using "unpack_bootimg" I didn't get a value for "BOARD_KERNEL_OFFSET" like you have it in your config. Could this be the problem?
Your BoardConfig.mk
My BoardConfig.mk
Do you see anything else out of the ordinary?
(Because I'm doing everything what you did step-by-step the links point to the best matching commits)
Despite not being able to compile right now I tried to press on with integrating your changes in the hopes that it will be fixed somehow later on
So I'm currently stuck on this commit of yours:
Atom_L: import overlay from official vendor
Where did you get the "config.xml" and "power_profile.xml" from? The best thing I could find was a "power_profile.xml" inside "/vendor/overlay/FrameworkResOverlay/FrameworkResOverlay.apk" which seems to be a "compiled" version of the aforementioned xml-file.
Click to expand...
Click to collapse
> Comparing your BoardConfig.mk with mine shows a slight difference in the offset and size values which could be associated with the different kernels of the phones.
TARGET_KERNEL_OFFSET should normally always be 0x00008000. Also, your other offset values seem to be wrong too -- those values from `unpack_bootimg` cannot be filled in directly to BoardConfig.mk. Instead, you need to subtract BOARD_KERNEL_BASE from them (e.g. BOARD_RAMDISK_OFFSET should be 0x55000000 - 0x40078000, which is 0x14f88000, the same as mine). In fact, I think those parameters should be exactly the same for XL and L. Other than that, I don't think I can see much of a problem about your makefiles.
However, note that not all of my historical commits represent a compilable state of the device tree. I'd suggest you start directly from the latest state and just replace whatever is relevant instead of starting over. And there should not be much that needs changing at all except device names, fingerprints and the proprietary vendor files.
> Where did you get the "config.xml" and "power_profile.xml" from
Exactly from those apks. Just decompile them using apktool.
PeterCxy said:
TARGET_KERNEL_OFFSET should normally always be 0x00008000. Also, your other offset values seem to be wrong too -- those values from `unpack_bootimg` cannot be filled in directly to BoardConfig.mk. Instead, you need to subtract BOARD_KERNEL_BASE from them (e.g. BOARD_RAMDISK_OFFSET should be 0x55000000 - 0x40078000, which is 0x14f88000, the same as mine). In fact, I think those parameters should be exactly the same for XL and L. Other than that, I don't think I can see much of a problem about your makefiles.
Click to expand...
Click to collapse
Still giving me errors.
So I tried a very unconventional approach: I just copied the file myself into the mentioned "out/target/product/Atom_XL" folder.
For now it's still compiling. Fingers crossed.
PeterCxy said:
However, note that not all of my historical commits represent a compilable state of the device tree. I'd suggest you start directly from the latest state and just replace whatever is relevant instead of starting over. And there should not be much that needs changing at all except device names, fingerprints and the proprietary vendor files.
Click to expand...
Click to collapse
I just reached your biggest commit yet.
Can you tell me how you got the list of needed files? I hope it's not through trial-and-error.
Except for the values in "setup-makefiles.sh" only the "proprietary-files.txt" seems to be device specific. Is there anything else I need to be aware of in this commit?
P.S.: I know it is tedious to go through your commits one by one but I want to learn something of it not just simply copying what you did. To get a feeling where the biggest pitfalls are and what you did to circumvent them.
a-dead-trousers said:
Still giving me errors.
So I tried a very unconventional approach: I just copied the file myself into the mentioned "out/target/product/Atom_XL" folder.
For now it's still compiling. Fingers crossed.
I just reached your biggest commit yet.
Can you tell me how you got the list of needed files? I hope it's not through trial-and-error.
Except for the values in "setup-makefiles.sh" only the "proprietary-files.txt" seems to be device specific. Is there anything else I need to be aware of in this commit?
P.S.: I know it is tedious to go through your commits one by one but I want to learn something of it not just simply copying what you did. To get a feeling where the biggest pitfalls are and what you did to circumvent them.
Click to expand...
Click to collapse
> Still giving me errors.
Looks like that dtb.img error was totally my fault -- it was due to my jerry-rigged solution of using prebuilt dtb image that conflicted with one of Lineage's update in August and I haven't built the ROM for a month. Now I have fixed it in the latest commit.
> Can you tell me how you got the list of needed files?
All of those files are for VoLTE support and I started with the list from a commit in Redmi Note 7 Pro's device tree that imported those VoLTE blobs, and then added what was missing one by one (when something is missing the Phone process will crash and you can see what got missing in the logs). I don't think the list will be any different on Atom XL so you can just use the one in my device tree.
Hi.
Thanks to you everything is running smoothly here. But what bugs me is that TWRP is not working on our devices.
Although for the Atom there is a possibility: https://forum.xda-developers.com/android/development/twrp-modded-to-unihertz-atom-t3885793
Before I want to go public with my build I wanted to solve this last "mystery".
So I tried to include it in my current source tree according to the (official?) guide but some errors prevented me from a successful build.
Naturally I asked for some guidance at the most reasonable places I know of but got nothing so far:
https://forum.xda-developers.com/showpost.php?p=83443611&postcount=4622
https://forum.xda-developers.com/showpost.php?p=83455271&postcount=4623
https://github.com/TeamWin/android_bootable_recovery/issues/70
I even tried different repositories (omnirom/android_bootable_recovery) and revisions (android-9.0) but these resulted in missing library "type" (static vs. shared) errors so I assume these are too old for LineageOS 17.1
What I want to know is how you managed to get TWRP to built for your device even though the touchscreen wasn't working?
Did you use your LineageOS source tree or one of the many "minimal" manifests? If so, which one would be the "best" to use?
wkr ADT
@PeterCxy and @a-dead-trousers
Thanks for all the work on this so far. I've got an Atom L and have gotten the ROM's PeterCxy posted running on them as in the OP. Do either of you have a quick step-by-step workflow of how you got all the Lineage sources set up and built into the various ROMs? I'd like to be able to build the ROMs from scratch and understand the process.
If I can get caught up to where you two are at with the builds, I can help debug, test and work through issues.
dirtylimerick said:
[MENTION=5351691] Do either of you have a quick step-by-step workflow of how you got all the Lineage sources set up and built into the various ROMs? I'd like to be able to build the ROMs from scratch and understand the process.
If I can get caught up to where you two are at with the builds, I can help debug, test and work through issues.
Click to expand...
Click to collapse
I documented my steps to setup up the build environment in the readme of my repo:
https://github.com/ADeadTrousers/android_device_Unihertz_Atom_XL
But leave out the TWRP part. It isn't working yet mostly because TeamWin/android_bootable_recovery and LineageOS/android_bootable_recovery are too similar.
To figure out all the bits and pieces needed for the device I followed the commit log of @PeterCxy build.
Hi, @PeterCxy.
Finally I was able to build a TWRP recovery and surprise, surprise the touchscreen isn't working.
But during my attempts to get a working TWRP build I came acros a guide that explains how to patch the kernel to get the touchscreen to work.
https://forum.hovatek.com/thread-27132.html
So I tried to follow it but failed to identify the "end" of the zipped Image-file (step 18) to remove the payload from the gz-file. Regardless of which of the null-bytes I use for cutting I always get a warning from 7-zip that there is still data at the end.
Do you know a better approach to achieve this whole patching? Maybe even come up with a scripting solution to easily apply this patch in later builds?
wkr ADT
a-dead-trousers said:
Hi, @PeterCxy.
Finally I was able to build a TWRP recovery and surprise, surprise the touchscreen isn't working.
But during my attempts to get a working TWRP build I came acros a guide that explains how to patch the kernel to get the touchscreen to work.
https://forum.hovatek.com/thread-27132.html
So I tried to follow it but failed to identify the "end" of the zipped Image-file (step 18) to remove the payload from the gz-file. Regardless of which of the null-bytes I use for cutting I always get a warning from 7-zip that there is still data at the end.
Do you know a better approach to achieve this whole patching? Maybe even come up with a scripting solution to easily apply this patch in later builds?
wkr ADT
Click to expand...
Click to collapse
There is no sane way to solve the problem without kernel source code. Basically the stock kernel just does not load the touch screen driver in recovery mode. That patching guide is pretty out of date and I imagine it won't work on most recent kernels. The only proper way is to pressure Unihertz to actually obey GPLv2 and release their kernel source code. Or maybe someone can try reverse-engineering the kernel, but at least I won't do it because it'll just be too much of a hassle.
PeterCxy said:
There is no sane way to solve the problem without kernel source code. Basically the stock kernel just does not load the touch screen driver in recovery mode. The only proper way is to pressure Unihertz to actually obey GPLv2 and release their kernel source code.
Click to expand...
Click to collapse
I'm with you on this one, but as long as we don't have the source code we need to resort to other means to achieve our goals.
PeterCxy said:
That patching guide is pretty out of date and I imagine it won't work on most recent kernels.
Click to expand...
Click to collapse
Yeah it's from way back in 2019
Anyway, with a little bit of tinkering I was able to modify my kernel to load the touchscreen driver in recovery mode.
Here is the device tree and the manifest i used.
I wouldn't recommend to use it in it's current state at all though because the fstab needs a little bit of tinkering. Everything seems to be either unordered or not mounted properly and I fear anything you do in there now will mess up the whole device. BUT I got the touchscreen goin for me which is nice.
PeterCxy said:
Or maybe someone can try reverse-engineering the kernel, but at least I won't do it because it'll just be too much of a hassle.
Click to expand...
Click to collapse
As soon as I have everything sorted out that needs to be fixed on my build (e.g. signing, radio, included gapps working properly, TWRP) I want to dig deeper into the kernel.
There are some devices with Helios P60 out there from other vendors which offer kernel sources.
P.S.: I also uploaded a HOW-TO in my device tree.
If you or someone else wants to try it. Also if you want to you can send me a "symbl.txt" (see to the HOW-TO) extracted from your device then I can do the patching for the Atom_L too.
a-dead-trousers said:
I'm with you on this one, but as long as we don't have the source code we need to resort to other means to achieve our goals.
Yeah it's from way back in 2019
Anyway, with a little bit of tinkering I was able to modify my kernel to load the touchscreen driver in recovery mode.
Here is the device tree and the manifest i used.
I wouldn't recommend to use it in it's current state at all though because the fstab needs a little bit of tinkering. Everything seems to be either unordered or not mounted properly and I fear anything you do in there now will mess up the whole device. BUT I got the touchscreen goin for me which is nice.
As soon as I have everything sorted out that needs to be fixed on my build (e.g. signing, radio, included gapps working properly, TWRP) I want to dig deeper into the kernel.
There are some devices with Helios P60 out there from other vendors which offer kernel sources.
P.S.: I also uploaded a HOW-TO in my device tree.
If you or someone else wants to try it. Also if you want to you can send me a "symbl.txt" (see to the HOW-TO) extracted from your device then I can do the patching for the Atom_L too.
Click to expand...
Click to collapse
Happy to hear that you were able to figure the touchscreen out. I tried to port TWRP at the very beginning when I started tinkering with the device but quickly grew frustrated and just ported Lineage Recovery instead. I guess I might try patching the kernel image too at some point later.
BTW, for TWRP to work with devices released after Android 10, I'm pretty sure you need an extra set of patches that are not yet fully merged to the main TWRP repository. I remember there's some guy providing another manifest with all the patches applied but I couldn't remember the name.
Hi.
I just officially announced my build for the Atom XL:
https://forum.xda-developers.com/android/development/rom-lineageos-17-1-unihertz-atom-xl-t4171407
Could you please put a link in your first post for those in search of the Atom XL and found your thread instead. Thanks.
wkr ADT
hi @PeterCxy.
During my daily usage of the phone I encountered a strage problem:
The audio jack isn't working. Plugging in some headphones I get this slight click in the earpieces when the circuits connect but nothing else happens. Neither a "headphone" icon in the status bar nor hearing anything coming from the headphones itself. The main speaker of the phone keeps playing the music. Using bluetooth everything is working as expected though. So I used logcat to see if something is coming up during plugging in but nothing "catchy" shows up in the logs. My guess is that some (vendor?) service is missing or not started during booting. Next I checked If something shows up on logcat during boot but I'm not sure for what to look exactly. There are quite a few errors and warnings though. In my despair I started to "fix" the "avc: denied" (SEPolicy) entries. Thats when I found a specific error reguarding VoLTE. Maybe this would fix the problems you had with VoLTE in enforcing mode:
https://github.com/ADeadTrousers/an..._Atom_XL/blob/master/sepolicy/private/init.te
(The line with "socket_device:sock_file")
My provider doesn't support VoLTE so I'm not sure if this helps or not. Maybe you could check it.
Anyway can you please tell me if your device's audio jack is working or not?
If you're (by some mysterious coincidence) not affected by this, can you at least give me some pointers for what to look for to get this fixed on my side.
The Internet Is not very helpful when searching for "android audio jack" or something similiar.
Thanks in advance.
wkr ADT