Commit Graph

71 Commits

Author SHA1 Message Date
topjohnwu
23c2e22910 Update image functions 2018-07-06 22:04:06 +08:00
topjohnwu
5d5ec08566 Test the log buffers before running command 2018-07-06 07:32:16 +08:00
topjohnwu
e7a2144def Optimize magiskhide to work with the log daemon 2018-07-04 01:52:23 +08:00
topjohnwu
7265450e2e Precise free space calculation for magisk.img
1. Introduce new applet: imgtool for better separation from the main program
2. Actually mount the image and check statvfs for free space in the image

This shall eliminate any possible module installation failure from image resizing issues.
2018-06-22 06:18:06 +08:00
topjohnwu
838f3cc01e Simplify hiding conditions 2018-06-17 05:16:52 +08:00
topjohnwu
a6958ac139 Fix adbd on GSIs 2018-06-17 01:25:27 +08:00
topjohnwu
d2e8ecc646 Fix error return code when patching dtb 2018-06-10 20:36:18 +08:00
topjohnwu
34dcf49fbc Update restorecon implementation 2018-06-03 14:43:03 +08:00
topjohnwu
b436bce565 Minor optimizations 2018-05-26 21:25:59 +08:00
topjohnwu
87f6018468 Massive sepolicy refactor 2018-04-15 03:18:18 +08:00
topjohnwu
5665e04014 Force using system binaries 2018-03-17 21:42:42 +08:00
topjohnwu
e4e5269836 Android P have no make_ext4fs, use mke2fs as fallback 2018-03-11 02:52:24 +08:00
topjohnwu
7da205f4c8 Round sizes to nearest integer 2018-02-12 04:56:50 +08:00
topjohnwu
bd7766b17e Prevent small memory leak 2018-02-11 21:55:57 +08:00
Andrew Gunnerson
70b7d73453 utils/cpio.c: Fix off-by-one error in cpio_vec_insert
Previously, if `cpio_vec_insert()` needed to replace a file and the file
already exists as the first entry, then a duplicate entry would get
created.

This fixes the bug I reported at:
https://forum.xda-developers.com/showpost.php?p=75449768&postcount=22647

Signed-off-by: Andrew Gunnerson <andrewgunnerson@gmail.com>
2018-02-11 18:50:42 +08:00
Andrew Gunnerson
5ad4702a5b utils/file.c: NULL terminate all files read into memory
Some functions, like `patch_init_rc()`, treat buffers read into memory
as a string instead of a byte buffer. Since the buffers weren't
NULL-terminated, this resulted in out-of-bounds reads and caused crashes
in certain conditions.

THis commit updates fd_full_read() to always NULL-terminate the buffers
so that they can be treated as strings when working with text files.

Signed-off-by: Andrew Gunnerson <andrewgunnerson@gmail.com>
2018-02-11 18:50:42 +08:00
topjohnwu
ebd509d92d Obfuscate socket name to prevent detection
Because why not
2018-02-11 03:59:54 +08:00
topjohnwu
599ae95251 Support moving files across filesystems 2018-02-02 04:47:16 +08:00
topjohnwu
bc2cac90fe Hardlink files recursively 2018-02-02 03:22:38 +08:00
topjohnwu
7cfc24d68f Add DHTB header support 2018-01-29 22:16:02 +08:00
topjohnwu
328fc44194 Rename module core to native 2018-01-27 09:11:28 +08:00