Database Info

[Resolved] Compiling Android USB Gadget drivers and insmod

Hi everyone,
I have been trying for quite a while now to compile HID Gadget support for my galaxy S3 and think i may be getting close. My current problem is that when i run insmod g_hid.ko i receive the following
Code:
insmod: init_module 'g_hid.ko' failed (Device or resource busy)
. The weird thing is that next time the screen turns off or if i attempt to insert it again the phone crashes (kernel panic i assume). I get this same error when i try and insmod g_zero.ko (although this one doesn't kill the phone). Can anyone tell me how i would attempt to debug this? Or even better, has anyone got g_hid working on there phone and if so, what modifications/kernel configs did you use?
Also, when compiling the kernel, in "USB gadget support" if i select any option other than Android Gadget in USB Gadget Drivers my kernel fails to build. Has anyone got any USB gadgets to compile and run and if so what procedure/modifications did you have to make. Currently i am compiling my kernel with android gadget selected as the usb gadget driver and then changing it to HID module and running make module_prepair and make module to get my .ko files (which may be why i am getting the above error).
If anyone can provide some insite and advice it would be greatly appreciated.
Thank you
Adrian

After months of trying, it appears i have got it solved. Unfortunately its not the most graceful solution and currently it disables MTP but i dont care. To fix it i searched every file in the kernel for mentions of CONFIG_USB_G_ANDROID and everywhere it was involved in an #if defined statement i added another section like so
Code:
#if defined(CONFIG_USB_ANDROID) || defined(CONFIG_USB_G_ANDROID) || defined(CONFIG_USB_S3C_OTGD)
In this case i added USB_S3C_OTG. As i said, it isn't pretty but it works.
Hope this helps someone in the future coz there is very little info on the net about this
Enjoy
Adrian

Code sample
I asked a question on Stack Overflow about exactly this issue. There is a small group of people who seem quite interested in this issue, and there really is not a lot of information on the net.
I would very much appreciate if you could explain in a little more detail what steps you took to get this to work for you. You could probably answer the Stack Overflow question more decisively than the current answers too: stack-overflow questions/9805731 - is-it-possible-to-program-android-to-act-as-physical-usb-keyboard. (I can't post the link as this forum blocks it.)
darcpyro said:
After months of trying, it appears i have got it solved. Unfortunately its not the most graceful solution and currently it disables MTP but i dont care. To fix it i searched every file in the kernel for mentions of CONFIG_USB_G_ANDROID and everywhere it was involved in an #if defined statement i added another section like so
Code:
#if defined(CONFIG_USB_ANDROID) || defined(CONFIG_USB_G_ANDROID) || defined(CONFIG_USB_S3C_OTGD)
In this case i added USB_S3C_OTG. As i said, it isn't pretty but it works.
Hope this helps someone in the future coz there is very little info on the net about this
Enjoy
Adrian
Click to expand...
Click to collapse

No worries, im happy to help. The funny thing is by reading the answers you got it started me on the right path to finding a solution. Im not sure how this will translate to other devices firmware but in the case of the galaxy S3 it was i went threw every file and added the
Code:
|| defined(CONFIG_USB_S3C_OTGD)
to all if defined statement that made any mention of CONFIG_USB_GADGET so that the code inside would be compiled. This was so that in the .config file for the kernel CONFIG_USB_GADGET would be set to yes and would then allow the hid driver to be compiled. Basically every time i got an error i saying something wasn't compiled or wasn't declared i went to that file, found where it was declared and added that if defined statement. I would upload my working source code but it would take weeks on my internet connection. If there is a specific file you want to see then let me know and i will post its contents.
Good luck
Adrian

darcpyro said:
No worries, im happy to help. The funny thing is by reading the answers you got it started me on the right path to finding a solution. Im not sure how this will translate to other devices firmware but in the case of the galaxy S3 it was i went threw every file and added the
Code:
|| defined(CONFIG_USB_S3C_OTGD)
to all if defined statement that made any mention of CONFIG_USB_GADGET so that the code inside would be compiled. This was so that in the .config file for the kernel CONFIG_USB_GADGET would be set to yes and would then allow the hid driver to be compiled. Basically every time i got an error i saying something wasn't compiled or wasn't declared i went to that file, found where it was declared and added that if defined statement. I would upload my working source code but it would take weeks on my internet connection. If there is a specific file you want to see then let me know and i will post its contents.
Good luck
Adrian
Click to expand...
Click to collapse
Hi Adrian
did you manage to use your android phone as a keyboard with USB_H_HID module ? I'm trying to do the same thing and I would be interested if you have more informations. For now, I had cyanogendmod source for Samsung Galaxy S III and I've compiled the default 10.1 ROM.
Yesterday, I tried to modify kernel options to include USB_H_HID in the kernel but the build fails on some android.c error.
See you
--
Thus0

Hi Thus0
I did manage to get it to work late last year (Just so you know that it can be done).From memory i basically ran grep over every file in the modules directory looking for CONFIG_USB_GADGET and then edited those files to say something like:
#if defined(CONFIG_USB_ANDROID) || defined(CONFIG_USB_G_ANDROID) || defined(CONFIG_USB_S3C_OTGD)
This way it would compile the USB Gadget modules as well. Let em know if you have any more questions and ill try and help.
Adrian

Huh – is it possible to make apk file for root devices?

Shir_man said:
Huh – is it possible to make apk file for root devices?
Click to expand...
Click to collapse
No sorry, as far as i know a custom kernal is compiled

darcpyro said:
No sorry, as far as i know a custom kernal is compiled
Click to expand...
Click to collapse
Is it possible to only use a driver and to load it with insmod in a stock rom? What modification require a custom kernel?

i was looking for a solution and ou,d something for the nexus 7
https://play.google.com/store/apps/details?id=remote.hid.keyboard.client
open xource here with kernel patch
https://www.google.com/url?q=https:...t&sa=D&usg=AFQjCNGztTd4-U4gHvd4DLzdg_qBxLb7gw
maybe someone can use it

The Theory of Everything (Building MTK from saucery) asop cm miui ubuntu

Welcome​
I have started this thread for the THEORETICAL development of the mt6732/mt6752 from source if such a thing happened to exist which of course it does not.
While compiling from source is pretty well documented :good: compiling MTK is not so well documented especially the mt6732/6752.
I have tried to keep this thread as ambiguous as possible and hopefully we will be left in peace to iron out any difficulties.
DO's:
I am a Total Noob myself to compiling from source but experienced enough to use the xda search box, Google and Youtube first before asking any questions. If your still confused after using the above then by all means ask here.
DON'T s:
If your a noob who should happen upon this thread then by all means read and learn but please respect the dev's by not asking random question without searching first :fingers-crossed:
SHARING:
Please only share things of a sensitive nature with recognised members who you know and via the PM. :good:
Lets just see how far we can push this Kernel
Recommended Reading:
[GUIDE]Building a Kernel from source{Mediatek}
Build Kernel MT6577 - Can't boot after build
How To Port CyanogenMod Android To Your Own Device
XDA:DevDB Information
k01q_e k01q_h, Kernel for all devices (see above for details)
Contributors
bigrammy
Kernel Special Features: Remains to be seen
Version Information
Status: Testing
Created 2015-02-25
Last Updated 2015-02-25

I am here, reporting for duty. If anyone wants an extra "potato" because he has too much "ketchup" for use feel free to ask me

Just to be clear
I am new to compiling from source in any shape or form
I believe the kernel to be not a problem and I know dev's are working on getting our phone on cm and maybe others :fingers-crossed:
But me being me I am very curious and would like to understand how we would go about doing what @varun.chitre15 managed to do for the mt6582 Here
I have the PC all setup for building now thanks to @carliv great guide Here and the cm and android tut's I also found this useful guide on youtube by Dave Bennet Here
Our device is not on the cm or google repo so how do we add it locally.
Do we need any special commands for mediatek
Could we use the mt6582 repo and substitute or mod the files
As you can see I have more questions than answers as normal :laugh:
I dont want to tread on any toes here or take over current developing but just want to learn as said in the OP there is a lack of mtk guides regarding this.
If I missed a clear mtk guide then please post the link to it. :good:

In short your looking at manifests. http://wiki.cyanogenmod.org/w/Doc:_Using_manifests
carliv (I think) posted the device config on github - link in your SPFlash thread somewhere.
Found it: https://github.com/carliv/device_elephone_p6000?files=1
Vendor files

I have compiled and flashed a kernel, I've been running it for 24+ hours with no obvious issues. It's honestly very easy to just get it to build if you don't try to make major changes.
I have (very lazily) tried to change a couple of things in the config to fix the known issues (OTG, compass): unfortunately I have no way to test the OTG function right now, while the compass did not magically start working. On the other hand, the notification light issue which is introduced by V8.4 is not strictly or exclusively kernel-dependent, since I am running V8.3 with my own kernel and the notification function is intact. That's all I can share at the moment.

xenonism said:
I have compiled and flashed a kernel, I've been running it for 24+ hours with no obvious issues. It's honestly very easy to just get it to build if you don't try to make major changes.
I have (very lazily) tried to change a couple of things in the config to fix the known issues (OTG, compass): unfortunately I have no way to test the OTG function right now, while the compass did not magically start working. On the other hand, the notification light issue which is introduced by V8.4 is not strictly or exclusively kernel-dependent, since I am running V8.3 with my own kernel and the notification function is intact. That's all I can share at the moment.
Click to expand...
Click to collapse
Can you switch on and post the /proc/config ?
Regarding the notification lights, I think v8.4 introduced the custom partition (might be wrong on that). Running grep -r "ro.notification.breath" /system/ the only result I got was services.odex (might have been settings.odex). I've bak(smali)ed it but couldn't see the difference between the two that would explain the change.

HypoTurtle said:
Can you switch on and post the /proc/config ?
Regarding the notification lights, I think v8.4 introduced the custom partition (might be wrong on that). Running grep -r "ro.notification.breath" /system/ the only result I got was services.odex (might have been settings.odex). I've bak(smali)ed it but couldn't see the difference between the two that would explain the change.
Click to expand...
Click to collapse
The config file is attached to the post, it's too big to paste it.
I have tried the new ROM which came out today, then flashed my kernel. I can't use either SIM card anymore. Flashed the boot.img that comes with the ROM - same. I guess I gotta go back to V8.3 for now.
The new ROM doesn't seem to be the same as the OTA: it reports as: Elephone_P6000_02_V8.0_20150206.
About the notification issues (which bothers me the most), I haven't had much time do to more experiments, but I was thinking this (which probably also led to my confusion*): there's a chance the functionality is not removed or shut down, at least in the intentions of the maker. After all, in V8.4 (and in the new ROM), when the phone is connected the light stays on, while notifications make it breath. While not a desirable behaviour (at least IMO), I wouldn't call it... a non-behaviour, so to say. So perhaps the functionality itself is intact but something is altering the way it works, for whatever reason. I also did some unpacking and grepping a few days ago, but I couldn't find anything useful.
* At some point I thought the issue was fixed because the light was breathing while connected to my PC, but it was probably because I had a notification to read.

xenonism said:
The config file is attached to the post, it's to big to paste it.
I have tried the new ROM which came out today, then flashed my kernel. I can't use either SIM card anymore. Flashed the boot.img that comes with the ROM - same. I guess I gotta go back to V8.3 for now.
The new ROM doesn't seem to be the same as the OTA: it reports as: Elephone_P6000_02_V8.0_20150206.
Click to expand...
Click to collapse
Lets not speculate too much - but perhaps there was a minor board change between the first and second preorders, notification could be a problem with granting notification access (in settings) - could this be a selinux issue? It would explain why things like Light manager work - as you grant them notification access.
For lost Imei - can you compare the custom partition to the one in the ota?

If anyone needs an easier way to grab the 'ketchup', my GitHub has it. Click on my blog link in my signature.

BachMinuetInG said:
If anyone needs an easier way to grab the 'ketchup', my GitHub has it. Click on my blog link in my signature.
Click to expand...
Click to collapse
Thanks bro,
Nice log
I was going to try use the sprout config as this is nice and clean Here when I have worked out how to do things that is.
My eyeballs are bleeding now with all this reading but from what I can see most of files are the same names so maybe we could just replace them with ours probably 98% ish
I did see one ROM some place for the mt6732/52 that had mt6582 references I just wish I could remember where I had seen it
Like I say I am a noob to this compiling and linux stuff so I maybe talking out of my ass :laugh:

bigrammy said:
Thanks bro,
Nice log
I was going to try use the sprout config as this is nice and clean Here when I have worked out how to do things that is.
My eyeballs are bleeding now with all this reading but from what I can see most of files are the same names so maybe we could just replace them with ours probably 98% ish
I did see one ROM some place for the mt6732/52 that had mt6582 references I just wish I could remember where I had seen it
Like I say I am a noob to this compiling and linux stuff so I maybe talking out of my ass :laugh:
Click to expand...
Click to collapse
I'm actually a noob too, and honestly I've only ever successfully built a fakeflash (temporary recovery) that didn't even work.

bigrammy said:
Thanks bro,
Nice log
I was going to try use the sprout config as this is nice and clean Here when I have worked out how to do things that is.
My eyeballs are bleeding now with all this reading but from what I can see most of files are the same names so maybe we could just replace them with ours probably 98% ish
I did see one ROM some place for the mt6732/52 that had mt6582 references I just wish I could remember where I had seen it
Like I say I am a noob to this compiling and linux stuff so I maybe talking out of my ass :laugh:
Click to expand...
Click to collapse
Can anyone actually make a guide noob friendly to build kernel from source? I got kernel with me locally zip file I want to build it please any help?

Tech N You said:
Can anyone actually make a guide noob friendly to build kernel from source? I got kernel with me locally zip file I want to build it please any help?
Click to expand...
Click to collapse
I think you can use the scripts in the root of the source code to build the kernel? make<something>.sh.
Make sure you're on Linux (Ubuntu preferred) and that you have all dependencies installed correctly. To execute the script, simply go to the Terminal, cd to the location, then type . make<something>.sh

Tech N You said:
Can anyone actually make a guide noob friendly to build kernel from source? I got kernel with me locally zip file I want to build it please any help?
Click to expand...
Click to collapse
Have a look at the README.
Does make menuconfig work here?

These few simple instructions from the readme file enable you to build a working kernel (at least in a Linux environment):
Code:
How to Build
kernel
======
1. Get the prebuilt cross compiler from AOSP website:
$ git clone https://android.googlesource.com/platform/prebuilts/gcc/linux-x86/arm/arm-eabi-4.6
2. Add required cross compiler to PATH:
$ export PATH=/YOUR_TOOLCHAIN_PATH/arm-eabi-4.6/bin:$PATH
$ export CROSS_COMPILE=arm-eabi-
3. Then use the following commands to build the kernel:
$ ./makeMtk k01q_e new k
make menuconfig can be made to work, but you need to set some parameters and I can't look into it right now.
You previously asked something about the custom partition, I need some guidance there as I am not familiar with the IMEI issue.

xenonism said:
These few simple instructions from the readme file enable you to build a working kernel (at least in a Linux environment):
Code:
How to Build
kernel
======
1. Get the prebuilt cross compiler from AOSP website:
$ git clone https://android.googlesource.com/platform/prebuilts/gcc/linux-x86/arm/arm-eabi-4.6
2. Add required cross compiler to PATH:
$ export PATH=/YOUR_TOOLCHAIN_PATH/arm-eabi-4.6/bin:$PATH
$ export CROSS_COMPILE=arm-eabi-
3. Then use the following commands to build the kernel:
$ ./makeMtk k01q_e new k
make menuconfig can be made to work, but you need to set some parameters and I can't look into it right now.
You previously asked something about the custom partition, I need some guidance there as I am not familiar with the IMEI issue.
Click to expand...
Click to collapse
Yea wasn't sure about menuconfig as mtk uses projectconfig rather than def_configs.
In the custom partition there are files like /custom/etc/firmware/modem.img etc. When messing with a Flyme port it was these files (and possible conflicts in /system) that caused an IMEI:nul.
FYI I opened the custom partitions on windows using an ext viewer after running the imgs through sgs2toext4.

Kernel building Mediatek
Tech N You said:
Can anyone actually make a guide noob friendly to build kernel from source? I got kernel with me locally zip file I want to build it please any help?
Click to expand...
Click to collapse
Not sure if you guys have seen or read this but it's a pretty comprehensive guide to building the mediatek kernel by @MasterAwesome and should really be compulsary for all kernel related things thread here http://forum.xda-developers.com/showthread.php?t=2754513
@HypoTurtle
Regarding the custom partition this is or could be a problem for us now and in the future and may require further investigation. The cm sprout branch has the modem.img in the (normal place /system/etc/firmware) but as you say ours is in the custom partition which is probably to protect it from bad /system flashes
Thing is I see no normal type link to it in the /system/etc/firmware so it must be linked some other way which may explain why all my port attempts failed as none of them used a custom partition (Asus_X002)
Maybe we will have to repartition the emmc to a standard config and alter the kernel (if the links are set via the kernel that is) for cm and other ports to work smoothly as I am unsure just how everything is linked up.
I have not had much experience with custom partitions so someone one know's of a good info source please link it. :good:
Hopefully Master @Santhosh M can figure out what's going on with the custom partition :fingers-crossed:

bigrammy said:
@HypoTurtle
Regarding the custom partition this is or could be a problem for us now and in the future and may require further investigation. The cm sprout branch has the modem.img in the (normal place /system/etc/firmware) but as you say ours is in the custom partition which is probably to protect it from bad /system flashes
Thing is I see no normal type link to it in the /system/etc/firmware so it must be linked some other way which may explain why all my port attempts failed as none of them used a custom partition (Asus_X002)
Click to expand...
Click to collapse
The partition is symlinked from .../by-name/custom (which is symlinked by the kernel from dev/block/mmcblk0p12) to /dev/customimg which is mouned after an e2fsck to /custom. /custom isn't linked to /system, it's just added to the global environment (init.environ.rc), will need to check on the environ, I'm on flyme and it has /custom/lib added to the library path (which doesn't exist).

HypoTurtle said:
The partition is symlinked from .../by-name/custom (which is symlinked by the kernel from dev/block/mmcblk0p12) to /dev/customimg which is mouned after an e2fsck to /custom. /custom isn't linked to /system, it's just added to the global environment (init.environ.rc), will need to check on the environ, I'm on flyme and it has /custom/lib added to the library path (which doesn't exist).
Click to expand...
Click to collapse
Haha thanks that explains a lot of weird things perfectly. :good:

What's the problem or what is the issue here.
Mediatek compiling guide ( by masterawesome ) that you have linked to is not actually practically this mtk kernel is done and is just way too complicatedly explained.
There is no defconfig stuff or pulling config.gz from phone in mtk. In this new source its just simple. Set up your toolchain path. Execute the makeMtk followed by the project no u want. Get zImage and patch it for mtk header and merge it with stock ramdisk. For this newer mtk chipsets repack has an extra stuff where u have to be careful of kernel command line parameters.
That's it the kernel stuff in mtk

"Vanilla" device tree

Hey folks,
i'm trying to get a "vanilla" aka unmodified device tree for the Shield Tablet K1. I know that you can sync over from nvidia's gitweb, tough the device tree is looking weird.
When i look at let's say cm's android_device_nvidia_shieldtablet it looks different. So i wonder how i would go from nvidias mutated tree to cm's layout without actually just forking it from cm.
Reason is, i want to port another custom rom over to the shield tablet, but i can't neither figure out how i'd do that with nvidias sources, nor how i would create a device tree similar to cm's, without any modifications.
I googled around for about 2-3 months now and still have no clue, so if anyone could give me a hint, i'd be very grateful!

What's wrong with the cm tree?
Unjustified Dev did the initial port of everything by hand. I've done maintainance manually since then. And to my knowledge, there's only been one other person to do a manual conversion (vartom). All custom ROMs derive from my tree. Should work in aosp as well. If there's something you need different from the cm tree, you can just add it on top. Or if something is broke in general, I need to know and fix it.

There's nothing "wrong", just not " clean", if you catch my drift. It's just a matter of reverting inwanted things, but yeah.. feels dirty.
Anyways, thanks for the info, man.

*shrugs* Okay, if you want to set up a new tree, nothing's stopping you. I wouldn't mind someone else knowing how to do it (pretty sure I'm the only active dev that has experience there), but it's a lot of parsing and research. I'll say that CMs trees are 95% unmodified from Nvidia's upstream, though. It's just rearranged into an aosp style tree. The kernel is a merge of the console and tablet since they were 98% identical anyways. I do my best to keep everything as clean as possible since I don't want to do through the work of making a lot of modifications every time a new release comes out... Engineers are lazy like that. I think the only things that aren't pure identical copies are the LTE init files and the unified device stuff (making the wifi only and lte models work in one ROM).

Nah, don't get me wrong there, i don't want to discredit you there.
I'm more talking about cm specific things like gello browser and stuff. Like i said, just a matter of a few changes to not include these. I'm just a beginner in any way, i can just follow instructions at best, i don't know c, and almost no java at all. So what you do is an astounishing task - it's just a personal preference coming and starting with nexus devices to tinker around, i have had the luxury of an AOSP tree, that's pretty much "my dilemma". I'm unexperienced and thus this might take a while for me. I've started to modify your device tree the day before yesterday, but i have something done wrong, as when i start compiling it'll ask if it should use " full_eng" config. (6.0.1 is used)
Again, thanks for putting me in the right direction, also for your efforts in maintaining our device.

What ROM are you trying to build? You could look at Carbonrom or Blissrom to see the rename changes needed to use the tree in a different ROM. Those should be similar across the board. I think the only CM specific package references would be gello and snap. Those commits could be reverted easily enough. Or if the ROM doesn't have them, I think they'd be ignored even if left as is.

Hi,
you can take a look at here
the base was the official cm tree, this is not vanilla but it s a K1 only tree.

Steel01 said:
What ROM are you trying to build? You could look at Carbonrom or Blissrom to see the rename changes needed to use the tree in a different ROM. Those should be similar across the board. I think the only CM specific package references would be gello and snap. Those commits could be reverted easily enough. Or if the ROM doesn't have them, I think they'd be ignored even if left as is.
Click to expand...
Click to collapse
I'm trying to port OmniROM https://docs.omnirom.org/Porting_Omni_To_Your_Device
Shouldn't be that difficult, i'm just doing something wrong.
kylon said:
Hi,
you can take a look at here
the base was the official cm tree, this is not vanilla but it s a K1 only tree.
Click to expand...
Click to collapse
Thank you, man.
I've forked it and will use it as a base.

Oh, omni. I build TWRP from omni. Take a look at the following two commits.
https://github.com/TeamWin/android_...mmit/c8e564a5ea44d963ab8d0e7829d9becd2ad5b0c0
https://github.com/TeamWin/android_...mmit/9b8772627795492f3380d2bf86680c09aada92c9
I haven't built the full ROM, but that should work. If you're using a K1 specific tree, the only difference should be in omni_shieldtablet.mk, instead of gsm.mk, use whatever omni has for tablet wifi only.

Steel01 said:
Oh, omni. I build TWRP from omni. Take a look at the following two commits.
https://github.com/TeamWin/android_...mmit/c8e564a5ea44d963ab8d0e7829d9becd2ad5b0c0
https://github.com/TeamWin/android_...mmit/9b8772627795492f3380d2bf86680c09aada92c9
I haven't built the full ROM, but that should work. If you're using a K1 specific tree, the only difference should be in omni_shieldtablet.mk, instead of gsm.mk, use whatever omni has for tablet wifi only.
Click to expand...
Click to collapse
Thank you, man. In fact i heavily orientaded on TWRP sources in this case. Also took a look at zombipop's repo.
Still whatever i do all i get is
[[email protected] omni]$ brunch shieldtablet
build/core/product_config.mk:241: *** No matches for product "omni_shieldtablet". Stop.
WARNING: Trying to fetch a device that's already there
Traceback (most recent call last):
File "build/tools/roomservice.py", line 352, in <module>
fetch_device(device)
File "build/tools/roomservice.py", line 320, in fetch_device
git_data = search_gerrit_for_device(device)
File "build/tools/roomservice.py", line 81, in search_gerrit_for_device
device_data = check_repo_exists(git_data, device)
File "build/tools/roomservice.py", line 58, in check_repo_exists
"exiting roomservice".format(device=device))
Exception: shieldtablet not found,exiting roomservice
build/core/product_config.mk:241: *** No matches for product "omni_shieldtablet". Stop.
** Don't have a product spec for: 'omni_shieldtablet'
** Do you have the right repo manifest?
No such item in brunch menu. Try 'breakfast'
Click to expand...
Click to collapse

Oh, brunch won't work because it wants to sync from the upstream server. Use lunch instead. Then run make to build.

Steel01 said:
Oh, brunch won't work because it wants to sync from the upstream server. Use lunch instead. Then run make to build.
Click to expand...
Click to collapse
Looks like a derp moment, i just had to modify the devicetree in another directory which is the actual working dir, instead on androidsrc/device/nvidia/shieldtablet directly. Like i said i just did something wrong.
It' compiling now, thank you for all the support!

ext2/3/4 filesystem create/fix tools (e2fstools pkg) static build for arm64 devices

e2fstools package, statically built for arm64 devices from Google's googlecode.com Oreo branch
NOTE: These will support ANY version of android's filesystems that use ext2 or ext3 or ext4. I built them from oreo branch as that one was the closest to compiling without any editing of the Makefiles. Don't let the name throw you off
I needed to run a modern fsck on a rooted device I have that I have been doing some hacking on.
The filesystem was in bad shape and the on-device e2fsck absolutely refused to check it while it was even mounted, even with the force option. Out of part desperation part determination I decided I would just build one myself. So I went and dug up the sources from googlecode.com, transfered them to my linux workstation, and after much fighting with gcc for cross compilation, finding a place in the google git repo where they actually build (hence going with orero), and having to tweak the c code even in a few places (mostly fixing includes and whatnot, no actual coding on my part), I succeeded.
As it says in the title, these are all the e2fstools binaries built from Google's googlecode.com source tree in their Oreo release branch. They are also compiled statically (no libraries are needed for them to function) so they should work absolutely fine on any device that has a 64bit arm (aarch64) processor. I have tested a handfull of them and they work perfectly fine on my Samsung Galaxy s8. In fact I used the fsck.ext4 binary to repair my system partition which it did perfectly.
https://www.dropbox.com/s/owb76hohnjzjdwe/e2fstools-oreo-aarch64-static.tar.gz?dl=0
Hope they come in as handy to someone else as they did me.
List of files follows:
Code:
[email protected]:~> tar tzf e2fstools-oreo-aarch64-static.tar.gz
e2fsbin/
e2fsbin/e2undo
e2fsbin/e2image
e2fsbin/badblocks
e2fsbin/mkfs.ext3
e2fsbin/fsck.ext4dev
e2fsbin/e2initrd_helper
e2fsbin/fsck.ext3
e2fsbin/e4crypt
e2fsbin/e4defrag
e2fsbin/mke2fs
e2fsbin/e2fsck
e2fsbin/fsck.ext4
e2fsbin/filefrag
e2fsbin/tune2fs
e2fsbin/e2freefrag
e2fsbin/uuidd
e2fsbin/e2label
e2fsbin/mkfs.ext2
e2fsbin/blkid
e2fsbin/logsave
e2fsbin/lsattr
e2fsbin/uuidgen
e2fsbin/findfs
e2fsbin/mklost+found
e2fsbin/dumpe2fs
e2fsbin/mkfs.ext4
e2fsbin/debugfs
e2fsbin/fsck.ext2
e2fsbin/mkfs.ext4dev
e2fsbin/resize2fs
e2fsbin/chattr
e2fsbin/fsck
PS: Now that I have a working arm64 cross compilation system setup, if anyone else desperately needs a working static binary for anything relatively simple to build (as in not 50 million dependencies for me to track down and install) and can send me or link me the sources for it and its dependencies, I would be happy to oblige.

Unfortunately, my device is armv7

buengeut said:
Unfortunately, my device is armv7
Click to expand...
Click to collapse
I forget, is that 32bit? If so i could fairly easily build new ones that are 32bit. Did you try them and they didnt work? If its 64bit they may work anyway, as these dont use any fancy instructions, and i did not build them even with -O1. I thought the main differences between the different "v"s of the same arch (32v64) was in libc and friends, of which these dont use as they are completely static so libc is built in.
I was even thinking of making this a flashable zip, but i hadnt bothered as no one replied to my initial post so i didnt think anyone cared ?️

partcyborg said:
I forget, is that 32bit? If so i could fairly easily build new ones that are 32bit. Did you try them and they didnt work? If its 64bit they may work anyway, as these dont use any fancy instructions, and i did not build them even with -O1. I thought the main differences between the different "v"s of the same arch (32v64) was in libc and friends, of which these dont use as they are completely static so libc is built in.
I was even thinking of making this a flashable zip, but i hadnt bothered as no one replied to my initial post so i didnt think anyone cared ?️
Click to expand...
Click to collapse
ARMv7 is ARM 32bit, 64bit program will not work on a 32bit system.
Can you build e2fstools for ARMv7.

Sadly your link is down.
would you care to reupload?

LNQ said:
Sadly your link is down.
would you care to reupload?
Click to expand...
Click to collapse
Sure. I'll even make it flashable this time ?

Thanks alot, good sir!

Legend! I'd been dealing with not being able to repair my /data errors for the past few days since my Oreo update. The current TWRP (3.2.3.4) still has an ancient e2fsck that doesn't support the ext4 quota option that LG formats their Oreo data partition with. The updated version in /system/bin works while the OS is running which is why I knew I had errors, but I couldn't get it to run in TWRP to fix them because of the incompatible library files.
I was all set to reinstall my Android SDK and compile another version with inbuilt libraries when I decided to do one last search and found this thread where you'd already done it. So thanks for saving me a few hours of downloading/installing/remembering how to compile for Android all over again
I guess I can also vouch for it working on an LG V20. I don't know if it's just LG's version of Oreo that needs the later e2fsck or others too, but anyone who can't repair errors on their ext4 partitions in TWRP due to needing a newer e2fsck version should be able to use this version. I just copied the file fsck.ext4dev to /cache and ran it from there on my data partition.

Your link is dropbox banned...
Any chance you could attach the zip here...?

quotient said:
Your link is dropbox banned...
Any chance you could attach the zip here...?
Click to expand...
Click to collapse
Seconded!
@partcyborg
Any chance you could put this file somewhere else? The above link is banned.
Thanks!

Hi,
I have a Oneplus 3 that after 3 years of use became sluggish as hell at file system access.
After reading thus document https://www.usenix.org/system/files/conference/hotstorage16/hotstorage16_ji.pdf
I would give a try to your e2fstools build.
Can you repost a link to download?
Thanks in advance

partcyborg said:
I forget, is that 32bit? If so i could fairly easily build new ones that are 32bit. Did you try them and they didnt work? If its 64bit they may work anyway, as these dont use any fancy instructions, and i did not build them even with -O1. I thought the main differences between the different "v"s of the same arch (32v64) was in libc and friends, of which these dont use as they are completely static so libc is built in.
I was even thinking of making this a flashable zip, but i hadnt bothered as no one replied to my initial post so i didnt think anyone cared ?️
Click to expand...
Click to collapse
Late to this thread, great work!
I have the exact same problem.
Could you please post flashable zip and/or instructions how to load this into android?
Thank you

I found a binary on website, tested on Nexus 5 (ARMv7L 32bits)
GitHub - FerryAr/e2fsprogs-arm: Static build of e2fsprogs for Android Magisk Module
Static build of e2fsprogs for Android Magisk Module - GitHub - FerryAr/e2fsprogs-arm: Static build of e2fsprogs for Android Magisk Module
github.com

partcyborg said:
e2fstools package, statically built for arm64 devices from Google's googlecode.com Oreo branch
NOTE: These will support ANY version of android's filesystems that use ext2 or ext3 or ext4. I built them from oreo branch as that one was the closest to compiling without any editing of the Makefiles. Don't let the name throw you off
I needed to run a modern fsck on a rooted device I have that I have been doing some hacking on.
The filesystem was in bad shape and the on-device e2fsck absolutely refused to check it while it was even mounted, even with the force option. Out of part desperation part determination I decided I would just build one myself. So I went and dug up the sources from googlecode.com, transfered them to my linux workstation, and after much fighting with gcc for cross compilation, finding a place in the google git repo where they actually build (hence going with orero), and having to tweak the c code even in a few places (mostly fixing includes and whatnot, no actual coding on my part), I succeeded.
As it says in the title, these are all the e2fstools binaries built from Google's googlecode.com source tree in their Oreo release branch. They are also compiled statically (no libraries are needed for them to function) so they should work absolutely fine on any device that has a 64bit arm (aarch64) processor. I have tested a handfull of them and they work perfectly fine on my Samsung Galaxy s8. In fact I used the fsck.ext4 binary to repair my system partition which it did perfectly.
https://www.dropbox.com/s/owb76hohnjzjdwe/e2fstools-oreo-aarch64-static.tar.gz?dl=0
Hope they come in as handy to someone else as they did me.
List of files follows:
Code:
[email protected]:~> tar tzf e2fstools-oreo-aarch64-static.tar.gz
e2fsbin/
e2fsbin/e2undo
e2fsbin/e2image
e2fsbin/badblocks
e2fsbin/mkfs.ext3
e2fsbin/fsck.ext4dev
e2fsbin/e2initrd_helper
e2fsbin/fsck.ext3
e2fsbin/e4crypt
e2fsbin/e4defrag
e2fsbin/mke2fs
e2fsbin/e2fsck
e2fsbin/fsck.ext4
e2fsbin/filefrag
e2fsbin/tune2fs
e2fsbin/e2freefrag
e2fsbin/uuidd
e2fsbin/e2label
e2fsbin/mkfs.ext2
e2fsbin/blkid
e2fsbin/logsave
e2fsbin/lsattr
e2fsbin/uuidgen
e2fsbin/findfs
e2fsbin/mklost+found
e2fsbin/dumpe2fs
e2fsbin/mkfs.ext4
e2fsbin/debugfs
e2fsbin/fsck.ext2
e2fsbin/mkfs.ext4dev
e2fsbin/resize2fs
e2fsbin/chattr
e2fsbin/fsck
PS: Now that I have a working arm64 cross compilation system setup, if anyone else desperately needs a working static binary for anything relatively simple to build (as in not 50 million dependencies for me to track down and install) and can send me or link me the sources for it and its dependencies, I would be happy to oblige.
Click to expand...
Click to collapse
Hi nice compilation job! I know it's not always easy to compile these binaries from source...
Can you please reupload you archive here as attachment? Thanks!

lebigmac said:
Hi nice compilation job! I know it's not always easy to compile these binaries from source...
Can you please reupload you archive here as attachment? Thanks!
Click to expand...
Click to collapse
There's a Magisk module that does cross-compilation specifically, you should get it.

By the way, the link for that drop box is dead.

Question attempt at learning how to compile TWRP for our device.

I am not an experienced dev, and only know enough to be dangerous. that said, i used twrpdtgen from https://github.com/twrpdtgen/twrpdtgen to generate what ought to be the berlna device tree with the necessary board files etc.
what do i have to do, specifically, to clone the twrp source via git/repo? can anybody here explain it like you realize i don't know what i am doing i'll figure git and repo out eventually, but this will go quicker if someone would enlighten me
which android version do i want to build twrp with? apparently 12.1 support isn't finished yet, so i should use 11 with aosp source?
note, for our deivce you need to use the script on upon boot.img as we've a/b format, grab from latest stock rom to gen yourself, or see attached. also, necessary to use WSL on windows with a linux distro, or a linux vm, python on windows will not work.
please feel free to use these if it furthers the cause of compiling twrp before i learn how to use git and repo and compile it.. hehe

I wanted to try my hand at this and start digging into git and repo. I got as far as getting all the files ready, setting up the tree where it needed to be, used ". build/envsetup.sh", then lunched with option 5, which is omni-berlna_eng, which is closest to the twrp minimal manifest build option.
After I lunch, the issue arrives with the fact that omni-berlna_eng uses a depreciated PRODUCT_STATIC_BOOT_CONTROL_HAL, whatever that means. So I tried using lunch option 2, which would be an aosp-arm64_eng, but that just gives me the error "#### failed to build some targets (1 seconds) ####".
As of now, I'm at a loss. I don't have the extra time to spend working on this to make it function, and I only get a few hours at a time to get a crack at it, but nothing seems to work.
Edit:
I should mention, this is my first attempt at anything android development. I got this phone as a secondary carry on and I've dirtyflashed Pixel Experience onto it using a method found for Moto G Power TONGA. Wanted to see if I could get TWRP running since dirty flashing killed my mms and calling services.