1959 Commits

Author SHA1 Message Date
topjohnwu
084d89fcce Create Utf8CStrBuffer type 2025-04-15 10:26:22 -07:00
topjohnwu
55b036c071 Introduce BorrowedDirectory 2025-04-15 10:26:22 -07:00
topjohnwu
30e79310ab Make pointers NonNull after error check 2025-04-15 00:18:48 -07:00
topjohnwu
f063fa5054 Cleanup xwrap implementation 2025-04-15 00:18:48 -07:00
topjohnwu
7bd901273c Provide richer error messages
Make sure most syscall/libc calls results are mapped to OsResult
that can produce more detailed error messages.
2025-04-15 00:18:48 -07:00
topjohnwu
cb08504fe5 Update cargo dependencies 2025-04-11 14:48:16 -07:00
topjohnwu
c0a1fb77be Code cleanup 2025-04-11 14:48:01 -07:00
LoveSy
4864c1112a no pty for -c by default, and add -i to force pty 2025-04-11 13:21:10 -07:00
LoveSy
9ddeab034b Fix wrong tty pump
See #1463
2025-04-11 13:21:10 -07:00
LoveSy
c4847ed288 Move pts to rust, and avoid using thread 2025-04-11 13:21:10 -07:00
topjohnwu
b8f1523fb2 Minor code reorg
[skip ci]
2025-04-08 17:20:22 -07:00
topjohnwu
9c7d359093 Optimize and format imports
[skip ci]
2025-04-08 09:57:09 -07:00
topjohnwu
eb54bc1fd7 Cleanup unused code 2025-04-08 02:33:52 -07:00
topjohnwu
d4a0286e13 Migrate magiskinit selinux.cpp to Rust 2025-04-08 02:33:52 -07:00
topjohnwu
1523ed9f78 Always go through rustup proxies 2025-04-01 10:01:35 -07:00
topjohnwu
2aba7247a9 Skip stub APK install on emulator
Reduce test flakiness
2025-03-26 13:15:12 -07:00
vvb2060
c2ece62e4c native: delete global 16k option
NDK 28 enable 16 KiB page size compatibility option by default, delete the global option to restore 4k alignment for 32-bit arch.
2025-03-26 13:15:12 -07:00
topjohnwu
50af14f2a3 Move all MagiskInit entrypoints into init.rs 2025-03-24 17:26:03 -07:00
topjohnwu
e0a356b319 Introduce mount helper methods 2025-03-24 17:26:03 -07:00
topjohnwu
c09a792958 Reorganize magiskinit code 2025-03-24 17:26:03 -07:00
topjohnwu
0bbfe7f44d Fix 2SI on legacy SAR devices 2025-03-24 17:26:03 -07:00
topjohnwu
a396abf565 Minor changes
[skip ci]
2025-03-22 01:16:51 -07:00
Wang Han
52ef1d1cb2
Simplify matching selinux context of child zygote (#8845) 2025-03-11 01:27:15 -07:00
John Wu
f14e3a89cc
Enable optimize_for_size for Rust std (#8844) 2025-03-10 15:50:43 -07:00
topjohnwu
95d3eac2e0 Cleanup xwrap functions 2025-03-09 01:10:41 -08:00
LoveSy
12a0870bc9 Replace bzip2 with bz2-rs 2025-03-08 14:37:17 -08:00
topjohnwu
6ff82c4e86 Introduce FsPathFollow
Make sure all operations of FsPath do not follow symlinks, and provide
a way to explicitly switch over to a set of operations that DO follow
symlinks by FsPath::follow_link.
2025-03-07 15:51:51 -08:00
LoveSy
c64de35375 Move magiskpolicy cli to argh 2025-03-07 14:29:30 -08:00
LoveSy
bd0e954fea Replace zlib with zlib-rs 2025-03-07 11:40:02 -08:00
topjohnwu
675471a49e Upgrade argh to stable release 2025-03-07 02:38:36 -08:00
topjohnwu
c90e73ccec Migration to Edition 2024 2025-03-07 02:35:25 -08:00
topjohnwu
a43c1267d8 Update Cargo.toml 2025-03-07 02:35:25 -08:00
vvb2060
e8958c6b5c get_secontext: ignore ENODATA 2025-03-06 20:03:36 -08:00
LoveSy
e8a3bf82c6 set exit code of log ExitOnError to -1 2025-03-06 20:03:15 -08:00
topjohnwu
c6c1a17ae6 Address several clippy warnings 2025-03-03 02:15:14 -08:00
David K.
0e82df9e10 Support zImage compression types other than gzip.
Instead of just searching for the gzip magic, it now incrementally searches the kernel for the first thing that `check_fmt_lg` doesn't report as `UNKNOWN`.
2025-03-02 13:35:13 -08:00
topjohnwu
c9eac0c438 Introduce new sepolicy strategy for legacy devices
The existing sepolicy patching strategy looks like this:

1. 2SI: use LD_PRELOAD to hijack `security_load_policy`
2. Split policy: devices using split policy implies it also needs to
   do early mount, which means fstab is stored in device tree.
   So we do the following:
   - Hijack the fstab node in the device tree in sysfs
   - Wait for init to mount selinuxfs for us
   - Hijack selinuxfs to intercept sepolicy loading
3. Monolithic policy: directly patch `/sepolicy`

Method #1 and #2 both has the magiskinit pre-init daemon handling
the sepolicy patching and loading process, while method #3 gives us
zero control over sepolicy loading process. Downsides:

a. Pre-init daemon bypasses the need to guess which sepolicy init
   will load, because the original init will literally send the stock
   sepolicy file directly to us with this approach.
b. If we want to add more features/functionalities during the sepolicy
   patching process, we will leave out devices using method #3

In order to solve these issues, we completely redesign the sepolicy
patching strategy for non-2SI devices. Instead of limiting usage of
pre-init daemon to early mount devices, we always intercept the
sepolicy loading process regardless of the Android version and device
setup. This will give us a unified implementation for sepolicy patching,
and will make it easier to develop further new features down the line.
2025-02-28 09:39:10 -08:00
topjohnwu
b6b34f7612 Fix overlay.d context preservation 2025-02-27 01:57:25 -08:00
LoveSy
e55c413261 Correctly handle truncated dtb 2025-02-23 20:31:55 -08:00
topjohnwu
019eb03823 Hide or remove mut constructors for Utf8CStr 2025-02-17 11:38:11 -08:00
topjohnwu
363410e1c0 Introduce cstr_buf helper functions 2025-02-17 11:32:21 -08:00
topjohnwu
fc2ef21660 Introduce path! macro for FsPath 2025-02-17 01:46:19 -08:00
topjohnwu
63231d97ce Properly handle db downgrades 2025-02-16 17:01:36 -08:00
topjohnwu
b73d4a7022 Fix log_ok() 2025-02-16 12:01:25 -08:00
topjohnwu
bf591fca12 Fix Utf8CString constructor and add more comments 2025-02-16 01:17:48 -08:00
topjohnwu
dcf027884d Update FsPathBuf 2025-02-15 18:27:45 -08:00
topjohnwu
584f3820fe Make all Utf8CStrWrite Utf8CStrBuf 2025-02-15 18:27:45 -08:00
topjohnwu
3c7c46307a Partially cleanup MagiskInit code 2025-02-15 18:27:45 -08:00
vvb2060
4d80361805 core: search for first available dir in PATH 2025-02-15 15:08:26 -08:00
LoveSy
9a74e19117 Add log_ok() for log().ok() 2025-02-14 14:24:13 -08:00