** Update ****************
************************
Posted a .zip with scripts for both Windows and *nix users to automate the process.
Linux:
-----
Unzip the contents of the attached ICS404root.zip anywhere on your computer and run the script aptly named "runme_root_script.sh". It should take care of the rest. Make sure you have USB Debugging enabled and you put the phone in Camera mode, not mass storage device.
Windows:
---------
Unzip ICS404root.zip wherever you want and then run "rootscript.bat". Make sure you have USB Debugging enabled and you put the phone in Camera mode, not mass storage device.
*************************
*************************
Credit to miloj for finding this technique on the Transformer. (See the thread noted below and be sure to thank him!) I modified it to work on our devices.
http://forum.xda-developers.com/showthread.php?t=1704209
I'll put together a script to automate this process shortly, but if you're antsy like me, here's the lowdown:
1. Download the following files:
su: http://db.tt/ShPzea6I
debugfs: http://db.tt/bGFh43LZ
2. Save the two files downloaded above on /sdcard. (ie: mount your sdcard in windows and copy them over, or "adb push" them to /sdcard).
**Make sure you have your phone on Mount Camera mode, not as a mass storage device; otherwise, you won't be able to access your /sdcard directory via adb. **
3. In a linux terminal/Windows command prompt:
Code:
adb shell
[email protected]_maserati:/ $ cd /sdcard
[email protected]_maserati:/ $ cp su /data/local/12m/
[email protected]_maserati:/ $ cp debugfs /data/local/12m/
[email protected]_maserati:/ $ cd /data/local/12m
[email protected]_maserati:/ $ chmod 755 debugfs
[email protected]_maserati:/ $ chmod 755 su
[email protected]_maserati:/ $ mv batch batch.bak
[email protected]_maserati:/ $ ln -s /dev/block/mmcblk1p20 batch
[email protected]_maserati:/ $ exit
adb reboot
4. While you are waiting for the phone to reboot, type the following into your terminal/command window:
Code:
adb wait-for-device shell
5. Once you're back into the android shell:
Code:
[email protected]_maserati:/ $ cd /data/local/12m
[email protected]_maserati:/ $ rm batch
[email protected]_maserati:/ $ mv batch.bak batch
[email protected]_maserati:/ $ /data/local/12m/debugfs -w /dev/block/mmcblk1p20
(The following is entered at the "debugfs:" prompt)
debugfs: # cd xbin
debugfs: # write /data/local/12m/su su
debugfs: # set_inode_field su mode 0104755
debugfs: # set_inode_field su uid 0
debugfs: # set_inode_field su gid 0
debugfs: # quit
[email protected]_maserati:/ $ cd /data/local/12m
[email protected]_maserati:/ $ rm su
[email protected]_maserati:/ $ rm debugfs
[email protected]_maserati:/ $ exit
adb reboot
Done deal. Now you've got the "su" binary pushed to your /system partition and set with the proper permissions for execution. Download the Superuser app from the market and you're good to go. Make sure you update the su binary within the Superuser app as well to make sure you're up to date.
Awesome! Were you able to upgrade to the latest leak and not lose root? Btw, what carrier are you on? I figured out how to get tethering fully functional on rogers but the process requires root...
Sent from my XT894 running ICS
You bet. I had to fastboot the leaked .208 update over top of the .206 update yesterday because I messed up my /system partition; I had used the OTA Rootkeeper to keep root permissions when upgrading from .219 but had foolishly disabled it right before I bungled everything up.
So to sum it up, this method didn't require anything to be done before updating to the .208 leak; since it has nothing to do with the technical details of the kernel itself, I'm fairly certain it should work for the .200 or .206 leaks as well. Root permissions were obtained from a completely stock system.
I'm in Canada with Bell but it doesn't matter because I imported the phone from the US; Verizon is the only carrier that has this phone. At any rate, this method is pretty universal, it is preying on a vulnerability present in the stock init.rc file and I bet it would work on other phones such as the RAZR as well.
So we can confirm this is 100% working with Fastbooting back and moving to 208? If so I will probably jump on this immediately.
I am trying to do this method but I cant adb to detect my phone. Im on the .208 leak. Can anybody help?
Have you enabled USB Debugging in the Settings->Developer Options menu?
Rick#2 said:
Have you enabled USB Debugging in the Settings->Developer Options menu?
Click to expand...
Click to collapse
Yep.
Not able to reboot, trying manually...
Code:
debugfs: /data/local/12m/su: Permission denied
debugfs: su: File not found by ext2_lookup
debugfs: su: File not found by ext2_lookup
debugfs: su: File not found by ext2_lookup
Had to reboot manually twice. This is the only error message I received. Tried Superuser, but it stops.
I'm on .200 btw.
droidian1441 said:
Yep.
Click to expand...
Click to collapse
I'm having the same issue. I'm on the 208 leak. I start command prompt in windows then type "adb shell" and I get the "device not found" message. I enabled usb debugging and my phone is connected as mass storage.
Likewise, Reboot requires su access, manual only. When I go and run the write command in debugfs permission denied. Any ideas what would cause this? Based on the code shown in the first post, SU had been already acquired(# vs $), which makes me wonder here.
Die Bruine said:
Not able to reboot, trying manually...
Code:
debugfs: /data/local/12m/su: Permission denied
debugfs: su: File not found by ext2_lookup
debugfs: su: File not found by ext2_lookup
debugfs: su: File not found by ext2_lookup
Had to reboot manually twice. This is the only error message I received. Tried Superuser, but it stops.
I'm on .200 btw.
Click to expand...
Click to collapse
Looks like you're doing something wrong with the debugfs command; you don't want to enter /data/local/12m/su at that prompt.
Running su from any partition other than /system will lead to a permissions error, so you don't want to bother trying to execute it from the /data/local/12m location.
(The following is entered at the "debugfs:" prompt, ie: after executing /data/local/12m/debugfs -w /dev/block/mmcblk1p20; see step 5.)
Code:
debugfs: # cd xbin
debugfs: # write /data/local/12m/su su
debugfs: # set_inode_field su mode 0104755
debugfs: # set_inode_field su uid 0
debugfs: # set_inode_field su gid 0
debugfs: # quit
Grizzy3 said:
I'm having the same issue. I'm on the 208 leak. I start command prompt in windows then type "adb shell" and I get the "device not found" message. I enabled usb debugging and my phone is connected as mass storage.
Click to expand...
Click to collapse
Ive got the same situation over here. I can stick without root, just the fact that I would have it again would be just the single reason to do it. Lol.
Sent from my DROID4 using Tapatalk 2
Code:
debugfs 1.42 (29-Nov-2011)
debugfs: cd xbin
cd xbin
debugfs: write /data/local/12m/su su
write /data/local/12m/su su
/data/local/12m/su: Permission denied
Rick, that's what we're putting in. From the code you posted it shows that you had root access already. Do you have any other suggestions on this? Because that's the in and out I get.
---------- Post added at 04:57 AM ---------- Previous post was at 04:53 AM ----------
Problem resolved. Need to run the following code:
Code:
chmod 755 debugfs
chmod 755 su
Then continue with rooting.
gdeeble said:
From the code you posted it shows that you had root access already.
Click to expand...
Click to collapse
Not sure where you're making this assumption from. I just wrote the "#" symbol in there to signal where to start entering commands... though I suppose you're correct in pointing out that the "#" shows up on a root prompt. A smarter choice probably would have been "$".
Trust me, I'm not an idiot. I wouldn't have gone through the hassle of writing up the guide in the first post if it didn't work.
Didn't mean it that way, just looked like it already had root, which was what confused me. But thanks again for this. :-D
Tried it again. This time no errors and the phone rebooted. But now Superuser keeps on FC .
Reinstalled superuser, updated and busybox. Now rooted! Thnx.
BTW, you might wanna update the OP. Do not batch the commands under windows. I tried several times. I think there is something wrong with the timing. Manually entering all the commands in a shell works. But putting them in a batch will enter them too fast for ADB to handle (under Windows shell) I guess.
Die Bruine said:
BTW, you might wanna update the OP. Do not batch the commands under windows. I tried several times.
Click to expand...
Click to collapse
I don't know, it seemed to work fine for me with the script I made. Anyways, glad it worked out for you.
Now that we can re-root as well as (somewhat convolutedly) fastboot ourselves back on track, we're good to go.
droidian1441 said:
Ive got the same situation over here. I can stick without root, just the fact that I would have it again would be just the single reason to do it. Lol.
Sent from my DROID4 using Tapatalk 2
Click to expand...
Click to collapse
As stated in the guide, you need to be in camera mode not mass storage.
Sent from my DROID4 using XDA
I was trying to do it manually last night before the OP posted the batch file, and it was not working because I was in MTP instead of PTP. SO make sure you use PTP.
Put your phone in camera (PTP) mode for the USB connection and it should work fine. Also, after it completes, download Superuser from the market.
I ran Titanium Backup after everything and it told me it needed to fix my su binary permissions or something like that... I let it do its thing... Either way, IT WORKED!!!!!
I put it in camera mode and made sure usb debugging is enabled. Then I ran the script for windows. Still getting the device not found error throughout. Really don't know what's going on.
I am in the position of having to manually apply the defult sepolicy patch, init,?*init_shell?* and?*recovery?*permissive, and as the title states when using the supolicy tool to modify my supplieded sepolicy it is not being produced and on closer inspection throwing an error. I have attached both the images and the sepolicy file I am trying to applie these change to.
Have I been doing something wrong or is the file corrupted??
If you need more info just ask
Note: when I first tried it inside an adb shell it reported a segumentation fault, but I was unable to reproduce that condition to be provided with as a screen shot.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
EDIT: I at least can say that the possibility of a corrupt file is now smaller becuse I am able to run dumpav and dump its contents to a txt file and then do afb pull back to pc. So amyone know any way to applie the defult P atchs needed to sysyemless root?
@Chainfire Since this is your binary files, you should know the most about it.
Commands to gain application root on emulator
Code:
adb shell df #Check Available Space
adb shell mount -o remount,rw /system
adb push su /system/bin/su
adb shell chmod 0755 /system/bin/su
adb push su /system/xbin/su
adb shell chmod 0755 /system/xbin/su
adb shell su --install
adb shell "su --daemon&"
adb install superuser.apk
adb install rootcheck.apk
I then proceed to patch the sepolicy file with the following commands
Code:
adb push sepolicy /data/local/tmp/sepolicy
adb shell su -c "supolicy --file /data/local/tmp/sepolicy /data/local/tmp/sepolicy_out" #There is no sepolicy_out file
adb shell su -c "chmod 0644 /data/local/tmp/sepolicy_out"
adb pull /data/local/tmp/sepolicy_out sepolicy_out
So what am I able to do?
Are you able to
A) Help me debug the problem
Or
B) Patch the sepolicy file and post the output back to me/here
Matt07211 said:
...
Click to expand...
Click to collapse
Works fine on my device... could be an x86 specific issue? Unfortunately I don't have any x86 devices to test with.
Thanks for that. Yes, I am trying to patch the policy for my armv8 (arm64) cpu device (just realised, but would trying to patch the sepolicy from one architecture using the supolicy for a different architecture have new the problem?, if so I feel dumb). And since I didn't have a spare device devce that met the requirements, I resorted to use the already setup emulator in my Windows installation.
I had proceeded to root and run the supolicy tool for which nothing out-putted (tried different directorys), I then created a new sub-directory, test, in /data/local/tmp and chmod it with read and write permissions. I tried again and failed, I then ran a dumpav in the sepolicy I was trying to patch and outputted it to /data/local/tmp/test/dumpav.txt which worked.
I am just wondering why it didn't work for me .
Thanks again for the sepolic_out file, I really do apperciate it.
Ah you're saying the segmentation fault occurred on the emulator? That's interesting. Might be reproducable on my end.
Note: look at my first image with cms in the foreground and near the bottom of the command window you should see the segfault message, around second last command or so.
To reproduce that segfault (hopefully):
1) https://software.intel.com/en-us/android/articles/android-44-kitkat-x86-emulator-system-image Download the system image from here (had to direct download instead of sdk as internet was running through profile and ask wouldn't work through it)
2) used the 2.78 SuperSu zip and run above commands to gain root
3) run above commands to try and modify sepolicy (it doesn't produce anything)
4) start an adb shell and then run the commands inside the shell. Know the outputs shown was segfault the first time running the commands, every time afterwards it would show the error in the above screenshots
If you figure out what cause the segfault can you please tell me ?
Matt07211 said:
Note: look at my first image with cms in the foreground and near the bottom of the command window you should see the segfault message, around second last command or so.
To reproduce that segfault (hopefully):
1) https://software.intel.com/en-us/android/articles/android-44-kitkat-x86-emulator-system-image Download the system image from here (had to direct download instead of sdk as internet was running through profile and ask wouldn't work through it)
2) used the 2.78 SuperSu zip and run above commands to gain root
3) run above commands to try and modify sepolicy (it doesn't produce anything)
4) start an adb shell and then run the commands inside the shell. Know the outputs shown was segfault the first time running the commands, every time afterwards it would show the error in the above screenshots
If you figure out what cause the segfault can you please tell me ?
Click to expand...
Click to collapse
Before I go do all this, can you make sure the issue persists with the v2.78 SR1 version from the BETA thread ? Some issues with supolicy were fixed in that release.
Started with a fresh emulator and the newest SuperSu and ran these commands to gain root (I am placing everything as described in update-binary in the right places just to eliminate one thing, missing dependencies)
Code:
adb shell df
adb shell mount -o remount,rw /system
adb push Superuser.apk /system/app/Superuser.apk
adb shell chmod 0644 /system/app/Superuser.apk
adb push install-recovery.sh /system/etc/install-recovery.sh
adb shell ln -s /system/etc/install-recovery.sh /system/bin/install-recovery.sh
adb shell chmod 0755 /system/etc/install-recovery.sh
adb push su /system/xbin/su
adb shell chmod 0755 /system/xbin/su
adb push su /system/bin/.ext/.su
adb shell chmod 0755 /system/bin/.ext/.su
adb push su /system/xbin/daemonsu
adb shell chmod 0755 /system/xbin/daemonsu
adb push su /system/xbin/sugote
adb shell chmod 0755 /system/xbin/sugote
adb push supolicy /system/xbin/supolicy
adb shell chmod 0755 /system/xbin/supolicy
adb push libsupol.so /system/lib/libsupol.so
adb shell chmod 0644 /system/lib/libsupol.so
adb push 99SuperSUDaemon /system/etc/init.d/99SuperSUDaemon
adb shell chmod 0755 /system/etc/init.d/99SuperSUDaemon
adb shell su --install
adb shell "su --daemon&"
adb install superuser.apk
adb install rootcheck.apk
No everything should be in place, and we now can eliminate one thing (supolicy not finding needed dependencies)
Opened up SuperSu and let it install/update binary (succesful)
I then proceeded to patch the sepolicy file like so
Code:
adb push sepolicy /data/local/tmp/sepolicy
adb shell su -c "supolicy --file /data/local/tmp/sepolicy /data/local/tmp/sepolicy_out"
I then did "ls" in the directory and no file out-putted. So I went into a shell and ran
Code:
supolicy --file /data/local/tmp/sepolicy /data/local/tmp/sepolicy_out
And it throw the error shown in the image below. First time running that command in shell it says stopped, but the second time it says stopped as well as segfault.
Keep in mind I am trying to patch a sepolicy file that originates from an armv8 cpu (arm64) on an x86 Intel emulator.
Any more info needed? I am happy to help @Chainfire
So, I think it has something to do with your emulator image (perhaps its too old ?)
I took SuperSU's ZIP file and extracted it, changed to that folder, then:
(note that my adb shell to my emulator image has # root by default)
Code:
adb push c:\download\sepolicy /data/local/tmp/sepolicy
adb push x86\. /data/local/tmp/.
adb shell
cd /data/local/tmp
chmod 0755 supolicy
LD_LIBRARY_PATH=/data/local/tmp:$LD_LIBRARY_PATH ./supolicy --file sepolicy sepolicy_out
exit
Resulting in:
Code:
supolicy v2.78 (ndk:x86) - Copyright (C) 2014-2016 - Chainfire
Patching policy [sepolicy] --> [sepolicy_out] ...
- Success
So, I'm really not sure what might be going on with your setup, but I don't think its SuperSU itself, but rather the emulator.
Note that to use supolicy --file, you only need supolicy and libsupol.so, you don't even need root.
Chainfire said:
So, I think it has something to do with your emulator image (perhaps its too old ?)
I took SuperSU's ZIP file and extracted it, changed to that folder, then:
(note that my adb shell to my emulator image has # root by default)
Resulting in:
So, I'm really not sure what might be going on with your setup, but I don't think its SuperSU itself, but rather the emulator.
Note that to use supolicy --file, you only need supolicy and libsupol.so, you don't even need root.
Click to expand...
Click to collapse
Hmm, I really don't know what is wrong, I will try exactly what you have done later today, to see If can reproduce the output. If it doesn't work then we can pin it down to the emulator itself. What emulator image did you use?
I also realise that so emulator are rooted in the sense that web shell has root acess, just wasn't sure what dependices supolicy had at the time.
Matt07211 said:
Hmm, I really don't know what is wrong, I will try exactly what you have done later today, to see If can reproduce the output. If it doesn't work then we can pin it down to the emulator itself. What emulator image did you use?
I also realise that so emulator are rooted in the sense that web shell has root acess, just wasn't sure what dependices supolicy had at the time.
Click to expand...
Click to collapse
I created an API 22 Google Nexus x86_64 AVD in Android Studio
I should be able to try that in about 20-30 mins after I download it, I was using api level 19, Intel's emulator image.
I ran these commands on the Intel api 19 x86 emulator image.
Code:
adb push libsupol.so /system/lib/libsupol.so
adb shell chmod 0644 /system/lib/libsupol.so
adb push /system/xbin/supolicy
adb shell chmod 0755 /system/xbin/supolicy
adb push supolicy /data/local/tmp/supolicy
adb shell chmod 0755 /data/local/tmp/supolicy
adb push sepolicy /data/local/tmp/sepolicy
adb shell
cd /data/local/tmp
chmod 0755 supolicy
LD_LIBRARY_PATH=/data/local/tmp:$LD_LIBRARY_PATH ./supolicy --file sepolicy sepolicy_out
and it results in the error(shown in screenshot)
Code:
libsepol.policydb_read: policydb magic number 0x464c457f does not match expected magic number 0xf97cff8c or 0xf97cff8d
-Failure!
I then tried it on the Intel x86_64 api 22 emulator image (running the same commands as the first one, resulting in a succes, with the file being outputted as the sepolicy_out.
So as you have stated @Chainfire , it looks like a problem with the emulator itself, and most likely not the supolicy tool.
Chainfire said:
So, I think it has something to do with your emulator image (perhaps its too old ?)
I took SuperSU's ZIP file and extracted it, changed to that folder, then:
(note that my adb shell to my emulator image has # root by default)
Code:
adb push c:\download\sepolicy /data/local/tmp/sepolicy
adb push x86\. /data/local/tmp/.
adb shell
cd /data/local/tmp
chmod 0755 supolicy
LD_LIBRARY_PATH=/data/local/tmp:$LD_LIBRARY_PATH ./supolicy --file sepolicy sepolicy_out
exit
Resulting in:
Code:
supolicy v2.78 (ndk:x86) - Copyright (C) 2014-2016 - Chainfire
Patching policy [sepolicy] --> [sepolicy_out] ...
- Success
So, I'm really not sure what might be going on with your setup, but I don't think its SuperSU itself, but rather the emulator.
Note that to use supolicy --file, you only need supolicy and libsupol.so, you don't even need root.
Click to expand...
Click to collapse
@Chainfire, I'm trying to patch sepolicy for a Samsung device running Nougat, so that Supersu can be installed in system mode. Could you confirm if the --sdk=24 parameter is required?
adb shell su -c "supolicy --file /data/local/tmp/sepolicy /data/local/tmp/sepolicy_out --sdk=24"
Thanks, appreciate your time.
ashyx said:
@Chainfire, I'm trying to patch sepolicy for a Samsung device running Nougat, so that Supersu can be installed in system mode. Could you confirm if the --sdk=24 parameter is required?
adb shell su -c "supolicy --file /data/local/tmp/sepolicy /data/local/tmp/sepolicy_out --sdk=24"
Thanks, appreciate your time.
Click to expand...
Click to collapse
Yes it is.
System mode hasn't been tested at all on 7.0 though. I'm not sure anybody has been able to get it to work at this point.
If you do, let me know and with the steps
Chainfire said:
Yes it is.
System mode hasn't been tested at all on 7.0 though. I'm not sure anybody has been able to get it to work at this point.
If you do, let me know and with the steps
Click to expand...
Click to collapse
Hmm wasn't aware of the lack of support for system mode in nougat, any plans to implement?
It seems system mode root renders the device unbootable according to reports from my tester.
Question, if I modify the supersu script to mount su.img from /system am I likely to hit issues?
Seems a strange query I know.
Reason is we have a Samsung device that for some reason will not boot from a source built custom Nougat kernel. Not sure if this is related to AVB yet or something else.
However we can get a half assed TWRP to boot with the stock kernel.
Only problem is, no matter what, only /system can be mounted and accessed with write permission due to permission denied issues with the rest of partitions. Pretty sure this is an SELinux issue.
Meaning systemless root cannot be installed as normal. No access to /data or /cache.
I can patch the boot.img ramdisk manually for systemless, but for root to work I would need to push su.img to system and mount it from there.
Is it possible to still mount su.img from system if I modify the ramdisk init as required?
The other avenue is to flash su.img to /data or /cache via ODIN.
If it was flashed to /cache would supersu automatically pick up its location and copy it to /data or would a flag need to be set?
Just trying to keep my options open here.
ashyx said:
Hmm wasn't aware of the lack of support for system mode in nougat, any plans to implement?
It seems system mode root renders the device unbootable according to reports from my tester.
Click to expand...
Click to collapse
It is on my list of things to test/fix, but that list is long and full of terrors.
Question, if I modify the supersu script to mount su.img from /system am I likely to hit issues?
Is it possible to still mount su.img from system if I modify the ramdisk init as required?
Click to expand...
Click to collapse
I think that could work, yes.
The other avenue is to flash su.img to /data or /cache via ODIN.
If it was flashed to /cache would supersu automatically pick up its location and copy it to /data or would a flag need to be set?
Just trying to keep my options open here.
Click to expand...
Click to collapse
SuperSU should pick it up from /cache. Alternatively, try SuperSU's FRP mode, which stores a copy of the needed files in the boot-image and re-creates /data/su.img as needed.
Chainfire said:
It is on my list of things to test/fix, but that list is long and full of terrors.
I think that could work, yes.
SuperSU should pick it up from /cache. Alternatively, try SuperSU's FRP mode, which stores a copy of the needed files in the boot-image and re-creates /data/su.img as needed.
Click to expand...
Click to collapse
Thanks, great info as always. Finally managed to root the damn thing by adding a short script to the init which copies su.img to cache.
However FRP mode sounds like a more elegant solution if I can work out how to implement it in the Ramdisk.
Much appreciate your input.
ashyx said:
So, I think it has something to do with your emulator image (perhaps its too old ?)
...
Could you confirm if the --sdk=24 parameter is required?
adb shell su -c "supolicy --file /data/local/tmp/sepolicy /data/local/tmp/sepolicy_out --sdk=24"
Thanks, appreciate your time.
Click to expand...
Click to collapse
Yea believe it was segfaulting due to the Android version, I think I was using KitKat and it wasn't working, bumped up to lollipop and above and it worked fine
Oh, the SDK parameter, never heard of it, what does it do? Geuss I'll Google that then.
ashyx said:
Thanks, great info as always. Finally managed to root the damn thing by adding a short script to the init which copies su.img to cache.
However FRP mode sounds like a more elegant solution if I can work out how to implement it in the Ramdisk.
Much appreciate your input.
Click to expand...
Click to collapse
I did the same thing for my device, add a little script to move it to data. Had no other way to get it to a locked down device without TWRP. Hehe. Good job
Can you please tell me how to manually patch init by supersu ?
I've googled a lot, but haven't found a way to manually patch init by supersu.
My model is Honor v10, there isn't a custom recovery, so i have to make a boot.img with supersu inside to get root.