Commit Graph

151 Commits

Author SHA1 Message Date
topjohnwu
59f78d7dfc Update to BusyBox 1.29.2 2018-08-13 01:30:15 +08:00
topjohnwu
d8405f0d05 Make recovery installed on on system_root devices normal 2018-08-12 00:16:59 +08:00
topjohnwu
0f34f0033c Switch to FrankeNDK for building native 2018-08-11 18:46:55 +08:00
topjohnwu
190646d50c Fix incorrect magisk metadata in ramdisk 2018-08-11 16:31:46 +08:00
topjohnwu
afcb3d8f34 Fix XZ decompression in magiskinit 2018-08-10 15:04:32 +08:00
topjohnwu
d4568aa0a7 Compress binaries and use xz-embedded in magiskinit 2018-08-10 05:57:12 +08:00
topjohnwu
1def9b301b Use xz-embedded for b64xz 2018-08-10 05:57:11 +08:00
topjohnwu
5bac442b18 Reorganize sources 2018-08-10 03:49:25 +08:00
topjohnwu
6add682705 Remove high compression mode 2018-08-10 03:49:25 +08:00
topjohnwu
8b50d84a05 Hide unnecessary error log 2018-08-09 15:10:00 +08:00
topjohnwu
d3858b81e2 Add new boot service: boot-complete 2018-08-09 14:52:44 +08:00
topjohnwu
bdff9769be Move remount,ro back to post-fs-data mode 2018-08-09 03:57:29 +08:00
topjohnwu
ada0f93686 Apply all sepolicy patches pre-init
Boot services tend to fail in the middle when the kernel loads a sepolicy live.
It seems that moving full patch (allow magisk * * *) to late_start is still not enough to fix service startup failures.
So screw it, apply all patched in magiskinit, which makes sure that all rules are only loaded in a single step.
The only down side is that some OEM with a HUGE set of secontexts (e.g. Samsung) might suffer a slightly longer boot time, which IS the reason why the rules are split to 2 parts in the first place.
2018-08-09 03:20:28 +08:00
topjohnwu
7b5d79d313 Kill all processes using the same UID of the target
To workaround OOS embryo optimization
2018-08-08 05:47:58 +08:00
topjohnwu
c38533e0f8 Prevent problematic modules causing device stuck in bootloop
If boot failed after 2 times, it will enable core only mode (which disables all modules)
2018-08-07 04:41:48 +08:00
topjohnwu
9159f86a9e Improvements to system_root devices booting as recovery 2018-08-07 02:20:40 +08:00
topjohnwu
a63696836c Proper addon.d-v2 support 2018-08-03 22:40:49 +08:00
topjohnwu
46aad00f16 Use buffer on stack 2018-08-03 21:30:44 +08:00
topjohnwu
252afe8932 Use mirror in post-fs-data scripts 2018-08-03 17:09:24 +08:00
topjohnwu
82e8375957 Respect filesystem type when mounting mirrors
Close #405
2018-08-03 04:45:07 +08:00
topjohnwu
169c0fe4af Stop use clashing names 2018-08-03 03:43:02 +08:00
topjohnwu
cd6918e6eb Stop altering PATH to mirror 2018-08-03 03:38:36 +08:00
topjohnwu
5be035fd44 Try logging a little harder 2018-08-03 01:58:56 +08:00
topjohnwu
f1edc8443c Make root shell always use dev_pts
Close #433
2018-08-02 20:29:18 +08:00
topjohnwu
d9564bd04c Delay full sepolicy patch loading time 2018-08-02 05:35:01 +08:00
topjohnwu
91818cfa1a Support compiling split cils via magiskpolicy CLI 2018-07-21 05:12:22 +08:00
topjohnwu
4ffc388491 Allow bootctl to run 2018-07-20 22:22:49 +08:00
topjohnwu
3ae959af95 Do not early mount on symlinks on half Treble devices
Fix #399
2018-07-19 17:43:37 +08:00
topjohnwu
e11e88a9c5 Huawei hardcodes the slot suffix, don't append the suffix twice 2018-07-19 03:12:28 +08:00
topjohnwu
7cec8baa55 Merge magiskpolicy into Magisk main repo 2018-07-18 18:45:21 +08:00
topjohnwu
e987db9fb5 Move magiskpolicy sources to native/jni/magiskpolicy 2018-07-18 18:44:43 +08:00
topjohnwu
c603b9084f Remove magiskpolicy as submodule 2018-07-18 18:43:36 +08:00
topjohnwu
492d6dfcf0 Merge MagiskSU into Magisk main repo 2018-07-18 18:40:13 +08:00
topjohnwu
a3e0f2dcc3 Remove MagiskSU as submodule 2018-07-18 18:40:10 +08:00
topjohnwu
cf211e26f4 Move MagiskSU sources to native/jni/su 2018-07-18 18:23:36 +08:00
topjohnwu
d08f326990 Log fatal errors in debug mode only 2018-07-18 03:25:36 +08:00
topjohnwu
adf95ce3a0 Read fstab from device tree
In previous versions, magiskinit will not early mount if /sepolicy is detected. However on OP5/5T latest betas, the devices are fully trebelized,
but for some reason the file /sepolicy still exists, making magiskinit think it is NOT a treble device and doesn't work properly.

So to properly fix this issue, I will have to use the "official" way - check fstab in device trees. Any block mentioned in the fstab in device trees
are supposed to be early mounted. Currently magiskinit will only mount system and vendor even if other partitions exists in the dtb fstab, since other
partitions are not used to construct sepolicy (currently).

These changes can also fix #373, since we dynamically detect PARTNAME from device trees.
2018-07-18 00:45:10 +08:00
topjohnwu
3c1aca114f Wrap accept4 since some device does not have it 2018-07-17 06:52:23 +08:00
topjohnwu
18d0fd9d2a MagiskInit optimizations 2018-07-17 06:18:36 +08:00
topjohnwu
2bde8a1975 EMUI's logcat will change output file's permission and break ADBD 2018-07-16 06:42:36 +08:00
topjohnwu
bf9927c7dd Sync source with AOSP 2018-07-14 04:34:48 +08:00
topjohnwu
f339a087a2 Let Magisk compile against SDK 16 2018-07-13 22:14:32 +08:00
topjohnwu
1affb91f17 Support compiling against lower SDK
Reduce even more size for static binaries
2018-07-13 05:41:29 +08:00
topjohnwu
c6fc0e587e Support hardlink based proc ns 2018-07-11 23:41:38 +08:00
topjohnwu
7ed2c077de Support deodexed ROM on Oreo 2018-07-07 01:37:04 +08:00
topjohnwu
1283167595 Maintain our own set of loop devices 2018-07-07 01:32:58 +08:00
topjohnwu
23c2e22910 Update image functions 2018-07-06 22:04:06 +08:00
topjohnwu
f44b2dbd45 Rename log_monitor -> log_daemon 2018-07-06 07:57:18 +08:00
topjohnwu
46ee2c3f4e Improve handshake between the 2 daemons 2018-07-06 07:51:17 +08:00
topjohnwu
5d5ec08566 Test the log buffers before running command 2018-07-06 07:32:16 +08:00