20 Commits

Author SHA1 Message Date
topjohnwu
8cc5f096a2 Some minor changes 2019-12-05 17:20:32 -05:00
topjohnwu
7681fde4d0 Record mounts to be cleaned up in a vector 2019-11-19 00:16:20 -05:00
topjohnwu
d3b7b41927 Fix kmsg logging in magiskinit 2019-11-18 17:18:56 -05:00
topjohnwu
a52a3e38ed Change some class names 2019-09-22 05:20:51 -04:00
topjohnwu
ee0cef06a6 Add support for A-only 2SI 2019-09-22 05:15:31 -04:00
topjohnwu
aa47966347 Fix raw_data move constructor 2019-07-16 23:30:54 -07:00
topjohnwu
7ba8202af5 Introduce new root overlay system 2019-07-16 01:08:28 -07:00
topjohnwu
db8dd9f186 Init code rearrangement 2019-06-30 11:39:13 -07:00
topjohnwu
f1112fdf37 Logical Resizable Android Partitions support
The way how logical partition, or "Logical Resizable Android Partitions"
as they say in AOSP source code, is setup makes it impossible to early
mount the partitions from the shared super partition with just
a few lines of code; in fact, AOSP has a whole "fs_mgr" folder which
consist of multiple complex libraries, with 15K lines of code just
to deal with the device mapper shenanigans.

In order to keep the already overly complicated MagiskInit more
managable, I chose NOT to go the route of including fs_mgr directly
into MagiskInit. Luckily, starting from Android Q, Google decided to
split init startup into 3 stages, with the first stage doing _only_
early mount. This is great news, because we can simply let the stock
init do its own thing for us, and we intercept the bootup sequence.

So the workflow can be visualized roughly below:

Magisk First Stage --> First Stage Mount --> Magisk Second Stage --+
   (MagiskInit)         (Original Init)         (MagiskInit)       +
                                                                   +
                                                                   +
     ...Rest of the boot... <-- Second Stage <-- Selinux Setup  <--+
      (__________________ Original Init ____________________)

The catch here is that after doing all the first stage mounting, /init
will pivot /system as root directory (/), leaving us impossible to
regain control after we hand it over. So the solution here is to patch
fstab in /first_stage_ramdisk on-the-fly to redirect /system to
/system_root, making the original init do all the hard work for
us and mount required early mount partitions, but skips the step of
switching root directory. It will also conveniently hand over execution
back to MagiskInit, which we will reuse the routine for patching
root directory in normal system-as-root situations.
2019-06-29 01:25:54 -07:00
topjohnwu
a462435f2f Load custom sepolicy 2019-06-25 21:34:02 -07:00
topjohnwu
bb67a837d3 Adjust class structures 2019-06-24 01:50:47 -07:00
topjohnwu
6cde695194 Remove Q dirty hacks in SARCompat 2019-06-24 01:31:42 -07:00
topjohnwu
a1a1ac0bbb Add sbin overlay to system-as-root 2019-06-24 01:21:33 -07:00
topjohnwu
9ec8bc2166 Boot MagiskInit as actual system-as-root
WIP, no customization. DO NOT USE YET!
2019-06-23 15:14:47 -07:00
topjohnwu
1058aeb04f Label current SAR impl as compat
The current system-as-root magiskinit implementation (converting
root directory in system partition to legacy rootfs setup) is now
considered as backwards compatible only.

The new implementation that is hide and Android Q friendly is coming soon.
2019-06-22 03:18:45 -07:00
topjohnwu
cfec0db947 Delay mounting sbin overlay 2019-06-22 03:14:33 -07:00
topjohnwu
5c7f69acaa Separate SAR and legacy implementation 2019-06-16 12:45:32 -07:00
topjohnwu
f1d9015e5f Move load kernel info out of class 2019-06-15 22:25:09 -07:00
topjohnwu
3b0f888407 Minor update for parsing uevent 2019-05-27 02:55:46 -07:00
topjohnwu
845d1e02b0 Separate magiskinit components 2019-05-27 00:29:43 -07:00