RikkaApps / Sui

Modern super user interface implementation on Android.
GNU General Public License v3.0
2.4k stars 129 forks source link

[Zygisk] (276479e) (Module #62) Sui does not run properly #18

Closed phorcys420 closed 2 years ago

phorcys420 commented 2 years ago

Running https://github.com/RikkaApps/Sui/commit/276479e026496af1e521f7448dd87efdeae5a062 (https://github.com/RikkaApps/Sui/actions/runs/1421584576)

The "Sui" entry in the system settings, long-press menu just takes me to Internet properties.

Screenshot_20211105-074410

Screenshot_20211105-074435

I can see that Sui's post-fs-data.sh has ran from the magisk log.

Apps that use Shizuku (WADB in this case) prompt me using the usual MagiskSU prompt and not Sui's.

/data/adb/sui exists and there are two files that are in it (sui.db & sui.db-journal), sui.db-journal is blank.

Magisk log: https://github.com/RikkaApps/Sui/files/7481161/magisk_log_2021-11-05T07_46_24Z.log

RikkaW commented 2 years ago
  1. Use debug version of Sui
  2. Magisk log does not help, please provide full boot log
phorcys420 commented 2 years ago

I installed the debug version, but how do I obtain said boot log ? logcat and dmesg show nothing about sui

RikkaW commented 2 years ago

Upload the log you grabbed.

Capture boot log: Start the logcat before the earliest logs are overwritten, so that each log can be captured. Set log buffer size to maximum, so you don't have to start logcat at the beginning. Or you can use some Magisk module which runs logcat in post-fs-data.sh.

phorcys420 commented 2 years ago

Sorry for delayed response. logcat.log

RikkaW commented 2 years ago

It seems that the Zygisk's companion is not working (no logs like * Launching zygiskd64), so it is a Magisk issue.

Can you try again with the latest Magisk CI build?

phorcys420 commented 2 years ago

Latest Magisk CI build only proposes me to patch for commit d232cba0 which I am already running (it's only canary). Also I am using LSPosed Zygisk with no issues.

RikkaW commented 2 years ago

After a quick look at Magisk's source code, I can confirm that it is a problem of Maigsk.

phorcys420 commented 2 years ago

What should I do then ? Also why does LSPosed run fine with zygisk ?

RikkaW commented 2 years ago

Wait for Magisk to fix it.

LSPosed uses the companion at a different time. Don't say why not change the timing, Sui does follow the document of Zygisk.

phorcys420 commented 2 years ago

I see, thanks for the heads up.

Gong-Mi commented 2 years ago

等待马吉斯克修复它。

LSPosed 在不同的时间使用同伴。别说为什么不改变时间, 苏确实遵循了齐吉斯克的文件。 貌似用过riru版本就可以用zy版本

Looks like after you use riru, you can magisk_log_2021-11-06T19_01_32Z.log

Gong-Mi commented 2 years ago

等待马吉斯克修复它。

LSPosed 在不同的时间使用同伴。别说为什么不改变时间, 苏确实遵循了齐吉斯克的文件。 貌似用过riru版本就可以用zy版本

Looks like after you use riru, you can magisk_log_2021-11-06T19_01_32Z.log

只是缓存

Okay, it's just a cache

RikkaW commented 2 years ago

Just learned that zygiskd is shared by all modules. So it is possible that LSPosed breaks zygiskd. Try to disable LSPosed? Never mind, john wu confirmed it's Magisk's fault.

Gong-Mi commented 2 years ago

~刚刚得知,所有模块都共享了 zygiskd。因此,有可能,LSPosed打破齐吉斯克德。尝试禁用 LSPosed?~没关系, 约翰 · 吴证实这是马吉斯克的错。

23013 ok Seems normal

phorcys420 commented 2 years ago

~Just learned that zygiskd is shared by all modules. So it is possible that LSPosed breaks zygiskd. Try to disable LSPosed?~ Never mind, john wu confirmed it's Magisk's fault.

Indeed. 23013: Disabling LSposed makes Sui run properly 23014: Both LSPosed and Sui work Closing this issue.