Closed petmarb closed 3 years ago
No log ๐
same problem on los 17.2 with magisk 22.0 and riru 25.3.3. 11.6.0 was installed over 11.5.0.
yeah, I know, meaning I don't know how to get the logfile within a bootloop, because I had to delete the module within twrp and re-install the version before. just wanted to let you know.
Am 24. Mรคrz 2021 14:52:23 schrieb Rikka @.***>:
No log ๐
โ You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.
RikkaW , does 11.6.0 require Shizuku to be installed, or it can function without it?
Connect to a computer and use adb logcat, adb should be available in this type of "bootloop". Or use Magisk module to run logcat in post-fs-data.
look at the screenshot's last 2 lines, pls, is it ok?
sui_logcat.txt.zip here it is. after deleting v11.6.0 from /data/adb/modules phone starts fine.
@admzzz
com.android.settings
crash at the end, is the only thing suspicious. However, no dump since MagiskHide is enabled. Please disable MagiskHide and do logcat again.
Additionally, I found EdXposed, EdXposed is ridiculously low quality ๐. The crash is right after EdXposed. Need the log to confirm.
03-26 19:48:52.349 4149 4149 I EdXposed: Riru hooks installed
03-26 19:48:52.352 2023 2023 I NetworkStackClient: Network stack service connected
03-26 19:48:52.352 4128 4128 I EdXposed: Loading modules for com.qualcomm.qti.telephonyservice
03-26 19:48:52.353 4128 4128 I EdXposed-Bridge: Loading modules from /data/app/ru.YVKJNCuv.XNrCIyCEb-8OAGlypuQuh0ZjzIZ_jg4w==/base.apk
03-26 19:48:52.353 3769 3769 W ndroid.systemu: Unsupported class loader
03-26 19:48:52.354 4149 4149 F libc : Fatal signal 7 (SIGBUS), code 2 (BUS_ADRERR), fault addr 0x7464ba8000 in tid 4149 (ndroid.setting), pid 4149 (ndroid.setting)
@RikkaW , sui.txt.zip
04-23 13:19:55.926 1154 1154 E Riru : access /system/lib/libriru_sui.so failed with 2: No such file or directory
...
03-27 03:39:31.360 1153 1153 W Zygote : Class not found for preloading: android.app.-$$Lambda$ActivityThread$ZXDWm3IBeFmLnFVblhB-IOZCr9o
03-27 03:39:31.363 855 855 I SuiServer: service package is not started, wait 1s.
....
03-27 03:39:32.364 855 855 I SuiServer: service package is not started, wait 1s.
....
03-27 03:39:34.501 855 855 I SuiServer: uid for com.android.systemui is 10115
03-27 03:39:34.502 855 855 I SuiServer: uid for com.android.settings is 1000
03-27 03:39:34.502 855 855 W SuiServer: can't find com.google.android.gms, wait 1s
03-27 03:39:34.503 855 855 V SuiServer: binder rikka.sui.s1@e1565df
03-27 03:39:34.512 2203 3547 I SuiSystemServer: binder received
03-27 03:39:34.512 855 855 I SuiServer: send service to bridge
you were right about com.android settings: after switching off MagiskHide phone boots (almost) normally - with crashed settings' app (look at the screenshot).
and after deleting /data/adb/modules/Riru-sui phone boots without crashed settings' app.
what does it mean - "EdXposed is ridiculously low quality", and how one can fix/change it?
Disabling EdXposed should work.
@admzzz You can have a test to LSPosed, which we test to work normally with lastest sui.
@yujincheng08 , does LSPosed support all xposed's modules? is there any list of module's compatibility? ...compatibility list's link on https://github.com/LSPosed/LSPosed goes to the list of all xda exposed's threads...
@admzzz It supports all xposed modules that support EdXposed.
@yujincheng08 no way: all modules are not working due to "04-23 21:53:24.464 1067 1067 F art_apex: Device is not fsverity-enabled.".
LSPosed_Verbose_20210327_121511.log.zip
is there any way to fix it, or one had to switch back to edxposed in order modules to work?
@admzz As shown: it's Magisk's problem. Using the previous Magisk canary should fix it (not the latest because of https://github.com/topjohnwu/Magisk/issues/4098). And EdXposed also requires such sepolicies but it won't show warning but silently ignore it and let system modules unloaded: you can see from your previous EdXposed log: 03-27 03:39:32.278 2203 2203 E EdXposed: skip injecting into android because sepolicy was not loaded properly
@yujincheng08 , thank you kindly for your answers. after downgrading to magisk 20005 lsposed works (though line " is not fsverity-enabled." is still in the log and it is still partially activated), but not all modules works properly with it (mimnminguard or appsettings reborn, for example), but when i switch to edxposed these not working modules works fine. need time to check everything.
or maybe it would be better to downgrade to magisk 20003?
LSPosed never ever worked on my device. EdXpoded did always flawlessly and still does in it's latest version. Also Magisk does it's job as expected, with no drama at all - besides the fact that in my opinion the new gui is somewhat useless and I would wish back the old one. I don't think it's the right way to hit on many other things, when it is that simple, with downgrade to SUI 11.5.0 everything works again as before. Updating again -> bootloop So what?
@admzzz You need to reinstall LSPosed after downgrading to 22005 because the sepolicies are already missed by Magisk.
@admzzz It supports all xposed modules that support EdXposed.
really? GravityBox? Since when?
@petmarb Always working since 1.1.0. The devs of GravityBox also officially state it supports LSPosed. And EdXposed's YAHFA has a bug that prevents GravityBox from working on some devices with Android Q+.
@yujincheng08 , after deleting lsposed, phone's restart and reinstallation it from repository, it's status is activated, but some modules still do not work.
@RikkaW , you did not say a word about requested log, which i'd posted.
@admzzz As I said, bootloop is caused by EdXposed and thus this issue is closed because of irrelevance.
@RikkaW , jfyi: i'd updated magisk 22006 according to https://github.com/topjohnwu/Magisk/commit/aa020578955e0063aada2d618c6432faf92a0d33 and after installation of v11.6.0 got bootloop again.
so, the problem is not in (edxposed's) missing sepolicy rules for sure, as v11.5.0 works fine.
we never said bootloop is caused by sepolicy. And the crash point is definitely from edxposed. Stop using it.
@RikkaW , jfyi: i'd updated magisk 22006 according to topjohnwu/Magisk@aa02057 and after installation of v11.6.0 got bootloop again.
so, the problem is not in (edxposed's) missing sepolicy rules for sure, as v11.5.0 works fine.
This is totally because of EdXposed. I switched to LSPosed and it works. I don't get stuck in a bootloop anymore. So just shut up, and bite the bullet. You can test this by simply disabling EdXposed through your recovery.
In the other thread, RikkaW mentioned that this issue may crop up in future Riru modules. What steps can the user take to avoid this?
finally I found the time to give LSPosed another try in combination of all the other Riru modules, except for EdXpoded of course - rebooted, activated my beloved and very needed Xposed modules in LSPosed, rebooted again and ... what shall I say. besides the fact that now all Xposed modules work with no problem at all, I got an unbelievable dramatic speedboost all over the entire OS. It looks like EdXposed was slowing down the system, which I didn't even know. so, thank you very much RikkaW, you were totally right: kick off EdXposed for LSPosed in exchange, update everything and fall in love with your device again, thanks to RikkaW's great work!
Too bad LSposed devs don't want to implement global hooks for the modules which needs it... Ticking 100+ boxes for each module is driving me insane. ...and I can't use EdXposed with Sui...
that's not true. you don't have to tick hundreds of boxes. just mark the first, which should be the system framework to make the module work systemwide
That was the first thing I've tried. Unfortunately that didn't work for ButteredToast Revived.
Thanks so much for this; I had the same bootloop issue. Could you possibly add some kind of check for EdExposed in the install script to warn users of this bootloop?
on op6t with latest os, magisk, latest riru core + some other modules installed, updating sui to v. 11.6.0 causes reproducable bootloops. reinstalling v11.5.0 solved the problem