zygote64 crash on Note8 Snapdragon(bootloop)

me2151 commented 6 years ago

Hello. I am attempting to use Xposed 88.1 on the Galaxy Note8(7.1.1 API 25 arm64) and I'm encountering a bootloop issue. Ive narrowed it down to the art files and managed to get some information using adb bugreport as I cannot get a logcat from the boot(adb offline).

Essentially this is where my boot fails:

10-19 20:40:04.680 19545 19545 I art : GenerateImage: /system/bin/dex2oat --image=/data/dalvik-cache/arm64/system@framework@boot.art --dex-file=/system/framework/SmpsManager.jar --dex-file=/system/framework/core-oj.jar --dex-file=/system/framework/core-libart.jar --dex-file=/system/framework/conscrypt.jar --dex-file=/system/framework/okhttp.jar --dex-file=/system/framework/core-junit.jar --dex-file=/system/framework/bouncycastle.jar --dex-file=/system/framework/ext.jar --dex-file=/system/framework/fram10-19 20:40:14.804 19545 19545 E art : Could not create image space with image file '/system/framework/boot.art'. Attempting to fall back to imageless running. Error was: Failed to generate image '/data/dalvik-cache/arm64/system@framework@boot.art': Failed execv(/system/bin/dex2oat --image=/data/dalvik-cache/arm64/system@framework@boot.art --dex-file=/system/framework/SmpsManager.jar --dex-file=/system/framework/core-oj.jar --dex-file=/system/framework/core-libart.jar --dex-file=/system/framework/consc10-19 20:40:14.804 19545 19545 E art : Attempted image: /system/framework/boot.art 10-19 20:40:15.600 19545 19545 W art : Skipping non-existent dex file '/system/framework/oem-services.jar'

That sadly was the most i saw error wise. It is impossible to get a logcat of the bootloop unfortunatly. But I removed the libart.so and the dexoat files from the zip for the framework and it BOOTS but obviously xposed installer says it isnt running

Here are tombstones from the failed boot https://gist.github.com/me2151/135b863463dd4b938f4a8d763f14faa4

EDIT: After trying again ive found that zygote64 is crashing which causes my bootloop. I added tombstones from the failed boot. Ive managed to get past the previous point of failure and now have this.

rovo89 commented 6 years ago

Unfortunately, the original bug report didn't contain the full error. So I can't say anything about that.

I'll have to check the new crash:

backtrace:
    #00 pc 00000000002f7a24  /system/lib64/libart.so (_ZNSt3__16__treeINS_12__value_typeINS_12basic_stringIcNS_11char_traitsIcEENS_9allocatorIcEEEEPN3art13SharedLibraryEEENS_19__map_value_compareIS7_SB_NS_4lessIS7_EELb1EEENS5_ISB_EEE4findIS7_EENS_15__tree_iteratorISB_PNS_11__tree_nodeISB_PvEElEERKT_+28)
    #01 pc 00000000002f4b6c  /system/lib64/libart.so (_ZN3art9JavaVMExt17LoadNativeLibraryEP7_JNIEnvRKNSt3__112basic_stringIcNS3_11char_traitsIcEENS3_9allocatorIcEEEEP8_jobjectP8_jstringPS9_+168)
    #02 pc 0000000000004698  /system/lib64/libopenjdkjvm.so (JVM_NativeLoad+284)
#03 pc 00000000736061c0 /data/dalvik-cache/arm64/system@framework@boot.oat (offset 0x356a000)

Last log line:

10-21 21:19:32.551 10302 10302 E SemAffinityControl: SemAffinityControl: registerfunction enter

Please also try whether v88.1 works any better for you.

me2151 commented 6 years ago

v88.1 produces same results. Currently HOPEFULLY pulling a logcat in a few minutes.

EDIT: Do you have any suggestions or tips? I just tried creating a binary that runs logcat and saves to a file in /tombstones. Replaced the bootanimation binary for it. Sadly doesnt write the file during boot but if I test outside of boot it works fine

me2151 commented 6 years ago

Finally got a logcat using the previously mentioned method. Its 12MB so its too big for paste bin so I uploaded to drive. https://drive.google.com/open?id=0B8CP3g3AqMuHeGZKMVpSeFIyaTA

rovo89 commented 6 years ago

Thanks! It basically confirms the same things as above. Looks like it's crashing while loading a native library. Could you please upload /data/dalvik-cache/arm64/system@framework@boot.oat? Then I can check what's at pc 0000000073be51c0 (from the new log).

me2151 commented 6 years ago

Here you go: This is the system@framework@boot.oat file from xposed installed bootloop. https://drive.google.com/file/d/0B8CP3g3AqMuHS1hFVThsTkdCajQ/view

ucxl commented 6 years ago

me2151, 老铁，还好你懂英文。。。我刷了xposed不能开机，瞬间不幸福了

me2151 commented 6 years ago

Any luck using the oat file?

rovo89 commented 6 years ago

00000000002f4b6c in https://github.com/rovo89/Xposed/issues/273#issuecomment-338560578 refers to library = libraries_->Get(path);

The same for 00000000002f4cec in https://github.com/rovo89/Xposed/issues/273#issuecomment-339089543.

And from https://github.com/rovo89/XposedBridge/issues/210:

10-16 21:03:34.416 F/DEBUG   ( 3956): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
10-16 21:03:34.416 F/DEBUG   ( 3956): Build fingerprint: 'samsung/greatqltezc/greatqltechn:7.1.1/NMF26X/N9500ZCU1AQI9:user/release-keys'
10-16 21:03:34.416 F/DEBUG   ( 3956): Revision: '11'
10-16 21:03:34.416 F/DEBUG   ( 3956): ABI: 'arm64'
10-16 21:03:34.416 F/DEBUG   ( 3956): pid: 759, tid: 759, name: main  >>> zygote64 <<<
10-16 21:03:34.416 F/DEBUG   ( 3956): signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x6f4c646573754e
10-16 21:03:34.416 F/DEBUG   ( 3956):     x0   636f4c6465737546  x1   0000007ff9c7f138  x2   0000007ff9c7f138  x3   0000000000000000
10-16 21:03:34.416 F/DEBUG   ( 3956):     x4   0000000000000000  x5   0000007ff9c7f150  x6   0000007fa2193b9e  x7   636176616a62696c
10-16 21:03:34.416 F/DEBUG   ( 3956):     x8   0000000000000000  x9   d5abb8b2f4cfdd06  x10  0000007fa6dddab0  x11  0000000000000003
10-16 21:03:34.416 F/DEBUG   ( 3956):     x12  0000000000000021  x13  0000007fa5e93080  x14  d5abb8b2f4cfdd06  x15  0000007fa20c1e58
10-16 21:03:34.416 F/DEBUG   ( 3956):     x16  0000007f95b45f38  x17  0000007fa1cc4c44  x18  0000000000000021  x19  636f4c6465737546
10-16 21:03:34.416 F/DEBUG   ( 3956):     x20  636f4c6465737546  x21  0000007fa2046300  x22  0000007fa20fbd40  x23  d5abb8b2f4cfdd06
10-16 21:03:34.416 F/DEBUG   ( 3956):     x24  0000000000000000  x25  000000000000001e  x26  0000007ff9c7f150  x27  0000007fa20cba00
10-16 21:03:34.416 F/DEBUG   ( 3956):     x28  0000007fa2033b80  x29  0000007ff9c7efd0  x30  0000007fa1cc4cf0
10-16 21:03:34.416 F/DEBUG   ( 3956):     sp   0000007ff9c7ef90  pc   0000007fa1cc7ba4  pstate 0000000020000000
10-16 21:03:34.522 F/DEBUG   ( 3956): 
10-16 21:03:34.522 F/DEBUG   ( 3956): backtrace:
10-16 21:03:34.522 F/DEBUG   ( 3956):     #00 pc 00000000002f7ba4  /system/lib64/libart.so (_ZNSt3__16__treeINS_12__value_typeINS_12basic_stringIcNS_11char_traitsIcEENS_9allocatorIcEEEEPN3art13SharedLibraryEEENS_19__map_value_compareIS7_SB_NS_4lessIS7_EELb1EEENS5_ISB_EEE4findIS7_EENS_15__tree_iteratorISB_PNS_11__tree_nodeISB_PvEElEERKT_+28)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #01 pc 00000000002f4cec  /system/lib64/libart.so (_ZN3art9JavaVMExt17LoadNativeLibraryEP7_JNIEnvRKNSt3__112basic_stringIcNS3_11char_traitsIcEENS3_9allocatorIcEEEEP8_jobjectP8_jstringPS9_+168)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #02 pc 0000000000004698  /system/lib64/libopenjdkjvm.so (JVM_NativeLoad+284)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #03 pc 00000000005df630  /system/framework/arm64/boot-core-oj.oat (offset 0x54f000) (java.lang.Runtime.nativeLoad+204)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #04 pc 00000000005df0d0  /system/framework/arm64/boot-core-oj.oat (offset 0x54f000) (java.lang.Runtime.doLoad+204)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #05 pc 00000000005e10ec  /system/framework/arm64/boot-core-oj.oat (offset 0x54f000) (java.lang.Runtime.loadLibrary0+1240)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #06 pc 0000000000602f14  /system/framework/arm64/boot-core-oj.oat (offset 0x54f000) (java.lang.System.loadLibrary+96)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #07 pc 000000000009426c  /system/framework/arm64/boot-conscrypt.oat (offset 0x76000) (com.android.org.conscrypt.NativeCryptoJni.init+56)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #08 pc 00000000000d3be8  /system/lib64/libart.so (art_quick_invoke_static_stub+600)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #09 pc 00000000000e0630  /system/lib64/libart.so (_ZN3art9ArtMethod6InvokeEPNS_6ThreadEPjjPNS_6JValueEPKc+252)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #10 pc 0000000000292f44  /system/lib64/libart.so (_ZN3art11interpreter34ArtInterpreterToCompiledCodeBridgeEPNS_6ThreadEPNS_9ArtMethodEPKNS_7DexFile8CodeItemEPNS_11ShadowFrameEPNS_6JValueE+312)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #11 pc 000000000028bf20  /system/lib64/libart.so (_ZN3art11interpreter6DoCallILb0ELb0EEEbPNS_9ArtMethodEPNS_6ThreadERNS_11ShadowFrameEPKNS_11InstructionEtPNS_6JValueE+592)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #12 pc 0000000000561aa8  /system/lib64/libart.so (MterpInvokeStatic+356)
10-16 21:03:34.522 F/DEBUG   ( 3956):     #13 pc 00000000000c6494  /system/lib64/libart.so (ExecuteMterpImpl+14612)

Same location here. That looks like com.android.org.conscrypt.NativeCryptoJni.init tries to load a native library and it fails with a crash. I assume that libraries_ isn't valid, which could mean that the whole JavaVMExt instance isn't valid.

Not sure if this is related or not:

10-16 21:03:34.344 E/System  (  759): Unable to open boot classpath entry: /system/framework/oem-services.jar
10-16 21:03:34.345 E/System  (  759): java.io.FileNotFoundException: File doesn't exist: /system/framework/oem-services.jar
10-16 21:03:34.345 E/System  (  759):   at java.util.zip.ZipFile.<init>(ZipFile.java:212)
10-16 21:03:34.345 E/System  (  759):   at java.util.zip.ZipFile.<init>(ZipFile.java:148)
10-16 21:03:34.345 E/System  (  759):   at java.util.jar.JarFile.<init>(JarFile.java:161)
10-16 21:03:34.345 E/System  (  759):   at java.util.jar.JarFile.<init>(JarFile.java:98)
10-16 21:03:34.345 E/System  (  759):   at libcore.io.ClassPathURLStreamHandler.<init>(ClassPathURLStreamHandler.java:47)
10-16 21:03:34.345 E/System  (  759):   at java.lang.VMClassLoader.createBootClassPathUrlHandlers(VMClassLoader.java:49)
10-16 21:03:34.345 E/System  (  759):   at java.lang.VMClassLoader.<clinit>(VMClassLoader.java:31)
10-16 21:03:34.345 E/System  (  759):   at java.lang.ClassLoader.findLoadedClass(ClassLoader.java:742)
10-16 21:03:34.345 E/System  (  759):   at java.lang.BootClassLoader.loadClass(ClassLoader.java:1403)
10-16 21:03:34.345 E/System  (  759):   at java.lang.ClassLoader.loadClass(ClassLoader.java:312)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderConfig.initProvider(ProviderConfig.java:248)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderConfig.-wrap0(ProviderConfig.java)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderConfig$2.run(ProviderConfig.java:214)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderConfig$2.run(ProviderConfig.java:207)
10-16 21:03:34.345 E/System  (  759):   at java.security.AccessController.doPrivileged(AccessController.java:41)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderConfig.doLoadProvider(ProviderConfig.java:206)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderConfig.getProvider(ProviderConfig.java:187)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderList.loadAll(ProviderList.java:281)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.ProviderList.removeInvalid(ProviderList.java:298)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.Providers.<clinit>(Providers.java:64)
10-16 21:03:34.345 E/System  (  759):   at sun.security.jca.GetInstance.getInstance(GetInstance.java:156)
10-16 21:03:34.345 E/System  (  759):   at java.security.Security.getImpl(Security.java:587)
10-16 21:03:34.345 E/System  (  759):   at java.security.MessageDigest.getInstance(MessageDigest.java:186)
10-16 21:03:34.345 E/System  (  759):   at de.robv.android.xposed.DexCreator.updateSignature(DexCreator.java:185)
10-16 21:03:34.345 E/System  (  759):   at de.robv.android.xposed.DexCreator.create(DexCreator.java:177)
10-16 21:03:34.345 E/System  (  759):   at de.robv.android.xposed.DexCreator.ensure(DexCreator.java:69)
10-16 21:03:34.345 E/System  (  759):   at de.robv.android.xposed.DexCreator.ensure(DexCreator.java:48)
10-16 21:03:34.345 E/System  (  759):   at de.robv.android.xposed.DexCreator.ensure(DexCreator.java:40)
10-16 21:03:34.345 E/System  (  759):   at de.robv.android.xposed.XposedBridge.initXResources(XposedBridge.java:123)
10-16 21:03:34.345 E/System  (  759):   at de.robv.android.xposed.XposedBridge.main(XposedBridge.java:79)

ucxl commented 6 years ago

new logcat file

logcat.txt

wanam commented 6 years ago

I checked my GS7 conscrypt sources with an other 7.1.1 deodexed Touchwiz Rom, and found that NativeCryptoJni.init() loads only one library "libjavacrypto".

package com.android.org.conscrypt;

class NativeCryptoJni
{
  public static void init()
  {
    System.loadLibrary("javacrypto");
  }
}

Unfortunately GS7 didn't get Android 7.1 yet, so i cannot debug this myself. We can try to hook bellow methods, this helped in the past to get out of boot-loop on TW Lollipop and Marshmallow, but i didn't think it would be necessary since many TouchWiz devices booted fine on Nougat without it:

com.android.org.conscrypt.NativeCrypto.isMdfEnforced() -> false
com.samsung.android.security.CCManager.isMdfEnforced() -> false

me2151 commented 6 years ago

I'll see if the deodex process does anything with oem-services.jar. I can assume that 88.2 does not have anything to do with this issue correct?

rovo89 commented 6 years ago

Yes, unfortunately not. The next step from my point of view would be to try @wanam's suggestions. If that doesn't help, we need a special build with more logging output in LoadNativeLibrary(), e.g. to log all the parameters and especially the address of libraries_.

me2151 commented 6 years ago

Alright. I'll figure out how to do @wanam's suggestion. Is there any documentation on it that i can follow? EDIT: Im not entirely sure on how and where to do this unfortunately. Is there anything I can follow to get that going or can you tell me how to do it real quick? EDIT2: found https://github.com/wanam/XposedBridge/commit/b9595e96bd204099bd8e6b1d9dbb89575fe1052a in @wanam's git history. Will be attempting

wanam commented 6 years ago

I have some internet connection issues, cannot upload full installer zip, so rename attached file to "XposedBridge.jar" and put it on "/system/framework" inside xposed installer zip, flash the zip and get the logcat if the issue persists. XposedBridge.zip

me2151 commented 6 years ago

currently installing the studio to build myself as well Already applied the commits to latest tree in my repo. Ill try this real quick though. Thanks

me2151 commented 6 years ago

Well... with that something new happened but I dont have a log because i need to recompile my binary for that which im doing right now. Phone still bootloops but it vibrated after sitting like 5 min. Will be posting a new logcat in a few minutes

wanam commented 6 years ago

You don't need to recompile the binaries, just use the official xposed 88.2 zip, replace "XposedBridge.jar" file inside the zip, flash it and try to get the logs.

Some of my old changes may not apply for tw nougat.

me2151 commented 6 years ago

Right. i have little to no adb access so i have custom binaries(commands) to fork a logcat and store it locally to the cache partition. Thats how I have to get a logcat since i cant adb logcat and live boot doesnt work

Well... thats different. Had FULL adb access. have a 39mb logcat so i think i let it run long enough. Pulling from device in a few

44MB logcat:https://drive.google.com/file/d/0B8CP3g3AqMuHVS1uZEFRQU9YYXM/view

ucxl commented 6 years ago

flash xposed-v88.2-sdk25-arm64, can not boot new locat file

logcat.txt

wanam commented 6 years ago

Okay let's try the opposite, surprisingly it seems that setting "isMdfEnforced()" to true forced Private mode to work, maybe other things i didn't test as well, looking at "conscrypt.jar" Samsung uses this check to add some OpenSSL crypto algorithms.

@rovo89 Other thing i noticed, PackageManager scans "/system/container" which is not supported on aosp version of "ScanSystemApps", this folder contains some Knox apps.

I'm also getting this error with no apparent issue: 10-31 08:59:09.122 3319 3319 E SemAffinityControl: SemAffinityControl: registerfunction enter

Please try attached file and post your logcat again. xposed-v88.3-sdk24-arm64-wanam-test1.zip

Edit: added sdk25 xposed-v88.3-sdk25-arm64-wanam-test1.zip

me2151 commented 6 years ago

Testing and getting logs now

Newest log using 88.3-sdk-wanam-test1 https://drive.google.com/file/d/0B8CP3g3AqMuHR3ZpdTAtY2QtTjA/view?usp=sharing

From what I can see it made absolutely no difference :(

me2151 commented 6 years ago

Any further ideas?

dazemc commented 6 years ago

Any progress being made?

enantiopure commented 6 years ago

Bump... Any forward momentum. Its a shame that the SD S8/8+ works, not Note SD Note 8

rovo89 commented 6 years ago

I haven't worked on this, there are currently other priorities for me.

wanam commented 6 years ago

Unfortunately the logs don't say much about where we should look at, do you know any GS7 Rom port that has the same issue, so i can reproduce it.

dazemc commented 6 years ago

If I can help anyway... I am pretty competent with debugging. That 88.3 zip that you uploaded- does that have the LoadNativeLibrary() like @rovo89 recommended?

me2151 commented 6 years ago

I have not actively worked on this for awhile but i am back on this. hopefully something nice will happen soon.

ucxl commented 6 years ago

update xposed-v89-sdk25-arm64, Note8 Snapdragon can not boot logcat.txt

me2151 commented 6 years ago

@wanam since you mentioned /system/container previously could not having the /system/container folder in the rom be causing the issue(we remove it for deknoxing)

wanam commented 6 years ago

I thought this was happening on a full stock snapdragon variant of TW 7.1, am I wrong?

me2151 commented 6 years ago

Technically because of the root process we have to inject system root into the stock rom without disturbing the boot.img we consider it a custom rom. We deodex the system and we deknox at the same time. I have attempted on a pure stock system(unrooted),odexed and all knox stuff and it still refused to boot but I could not get logs from that one. Im currently rebuilding the system with deodex and root but leaving all the knox stuff in the rom. Hopefully this changes something.

wanam commented 6 years ago

@me2151 If it works on the exynos variant of the same device, then we could just compare their framework, can you (or anyone else) upload the framework folder for both devices?

Edit: fyi you need to be rooted or at least having dm-verity disabled, otherwise it shouldn't boot up.

me2151 commented 6 years ago

It does work on the Exynos Variant. I will upload the frameworks(which do you want? Just the framework.jar or the touchwiz.jar also)

wanam commented 6 years ago

@me2151 No i mean the framework folder, or at least core-libart.jar and conscrypt.jar from deodexed rom.

me2151 commented 6 years ago

Will do. Im downloading a Exynos deodexed rom that has working xposed. Im currently deodexing the Snapdragon model as I type this so ill put both in a zip for you.

me2151 commented 6 years ago

Here are both frameworks @wanam It appears that Exynos doesnt have to deodex for working xposed and their roms that have it working also have /system/container removed. https://drive.google.com/open?id=1mc3TAJQ0f-oZWvpZscwt-PSe8ogGBWap

EDIT: Since we use systemroot is it possible that something is not playing nice with app_process? Nvm nothing to do with systemroot. Im literally grasping at straws now. I have no idea how to read stack traces that android makes.

EDIT2: so i decided to go through the framework files and i noticed that conscript and actually most of my framework files are 1kb. Scratch that. i was looking at odexed files. However the deodexed conscript.jar is quite different on each varient And oddly. the Exynos conscrypt.jar works fine on the Snapdragon model. I know im going about it the wrong way but I removed the Loading of javacrypto. Hopefully something else happens EDIT3: Last thing i can think of at this point is going with what @rovo89 stated before and make something with more logging. I have no more logcat at this point though. I have tombstones and possibly xposed logs(havent checked those yet). Ive pulled the symbols located in the libart.so file and it matches with things listed in the previous logs but thats where im stuck.

wanam commented 6 years ago

Can you deodex the exynos frm with same tool or post the odexed version of the snapdragon frm? can you also check if any of the system apps is xz compressed (the file name should end with xz).

dazemc commented 6 years ago

There's really only one tool to deodex

On Dec 24, 2017 02:24, "Mohamed Karami" notifications@github.com wrote:

Can you deodex the exynos frm with same tool or post the odexed version of the snapdragon frm? can you also check if any of the system apps is xz compressed (the file name should end with xz).

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rovo89/Xposed/issues/273#issuecomment-353776857, or mute the thread https://github.com/notifications/unsubscribe-auth/AE_WatX8o_bBYwlW8JGVYYZ-YuLHENicks5tDiZDgaJpZM4QBw3s .

dazemc commented 6 years ago

Don't bother if you don't even know different compression methods

On Dec 24, 2017 02:35, "Daazed McFarland" daazedjmcfarland@gmail.com wrote:

There's really only one tool to deodex

On Dec 24, 2017 02:24, "Mohamed Karami" notifications@github.com wrote:

Can you deodex the exynos frm with same tool or post the odexed version of the snapdragon frm? can you also check if any of the system apps is xz compressed (the file name should end with xz).

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rovo89/Xposed/issues/273#issuecomment-353776857, or mute the thread https://github.com/notifications/unsubscribe-auth/AE_WatX8o_bBYwlW8JGVYYZ-YuLHENicks5tDiZDgaJpZM4QBw3s .

me2151 commented 6 years ago

@wanam I was mistaken before. The framework files do appear to be deodexed for Exynos(atleast the conscrypt.jar and core-libart.jar are deodexed) there are no compressed apps. all plain apks. I did just see your edit about dm-verity. That is the basis of our root method. Our boot.imgs we use have dm-verity disabled from the factory. @dazemc PLEASE dont be rude. He's helping us of his own free will and can stop when ever he wants. We dont want that...

dazemc commented 6 years ago

@wanam Sorry about that! I admittedly was inebriated last night.

me2151 commented 6 years ago

@wanam going back to the OEM-Services.jar file error... theres no actual jar that ive found however there IS boot-oem-services.oat and boot-oem-services.art in the arm(64) folder there... That gets removed during my deodex. Possibly the cause of that error? @godofdeathftw ironically No that build did not boot but it got to the bootanimation. Im still working on a new deodex for my device using updated software If you are looking for the deodexed files the ones in https://androidfilehost.com/?fid=889964283620760222 work fine. Thats just an older firmware

dazemc commented 6 years ago

@wanam you can use this to deodex https://github.com/lcweik/smali and 'xz --decompress' or 'unxz'.

me2151 commented 6 years ago

@rovo89 @wanam we have gotten past the conscrypt errors! only thing that remains in the trace is libart mostly. Also no more oem-services.jar missing error logcat.txt

wanam commented 6 years ago

I checked the diff on core-libart and conscrypt for both snap and exynos variant, the only apparent change i noticed is that the snap variant has this change, and it was already added to xposed (PA rom also has it).

me2151 commented 6 years ago

@wanam was the PA version different than stock xposed? Actually... I just looked and its in the Nougat branch but not the Nougat-mr2 branch... Isnt mr2 branch sdk25? and non mr2 branch sdk24? if so thats the issue right?(im on sdk25) EDIT: nvm. i went further back and found it...

me2151 commented 6 years ago

@wanam I know i bug you alot. I do have another question though. https://github.com/rovo89/android_art/commit/bc2c39d94e5011bda5ef39bbbaecb4f41f369518 that commit there is the next commit. It looks like its making the 124 an optional thing. Wouldnt I need to either revert that or do something else to actually make https://android-review.googlesource.com/c/platform/libcore/+/360691 do something? EDIT: So I reverted https://github.com/rovo89/android_art/commit/bc2c39d94e5011bda5ef39bbbaecb4f41f369518 and got a new error." F art : art/runtime/jni_internal.cc:2186] Failed to register native method dalvik.system.DexFile.getOatFileCompilerFilter?(Ljava/lang/String;Ljava/lang/String;)Ljava/lang/String; in /system/framework/core-libart.jar 12-25 18:11:16.524 5416 5416 F art : art/runtime/jni_internal.cc:2191] ----- class 'Ldalvik/system/DexFile;' cl=0x0 ----- " about https://github.com/rovo89/android_art/commit/c4e139db5f38711b4b680d82b3fca0e0c1217dab I assume. Am I on the right track that at all?

Well after reverting the above commit I think im making progress... It still didnt boot but there was only 1 tombstone that had something in it(00-03 were empty and 04 had a zygote32 crash. 05-09 were also empty). Then when i reinstalled stock system(basically uninstalled xposed) ithad to optimize all my apps(normally it ever optimized like 65 apps after a fail. it did all 460) Nvm... back to the original issue.

me2151 commented 6 years ago

@rovo89 @wanam it was mentioned before that loadLibrary() is the point that fails correct? I found in my latest logcat a warning message:

12-30 20:07:06.954 20400 20400 W EffectsFactory: loadLibrary() failed to open /system/lib/libmyspaceplus.so 12-30 20:07:06.954 20400 20400 W EffectsFactory: loadEffect() could not get library myspaceplus 12-30 20:07:06.954 20400 20400 W EffectsFactory: loadEffect() could not get library myspaceplus 12-30 20:07:06.954 20400 20400 W EffectsFactory: loadEffect() could not get library myspaceplus

However it only shows up when trying to boot with xposed. a regular boot logcat does not have that...

It appears to be related to audio. Would that be our problem in anyway?

EDIT: For the previous thought that libraries_ is invalid. It IS valid. It is found in the stock libart.

rovo89 commented 6 years ago

Why did you close the issue? Is the problem solved?

rovo89 / Xposed

zygote64 crash on Note8 Snapdragon(bootloop) #273