- Since official KSU has exposed kernel_umount API and provided meta module for handling mount operations, susfs can finally handover try_umount and auto add_try_umount to KSU and its meta module
- Synced with upstream, see e1e58ed737
Co-authored-by: simonpunk <simonpunk2016@gmail.com>
Reason:
- AUTO_ADD_TRY_UMOUNT_FOR_BIND_MOUNT is also causing a bit more performance overheads and still it cannot catch all the sus mounts in all situations. Actually it can easily be done in boot-completed.sh, and it should be more accurate, see module templates for more details.
- Official KernelSU also allows ksud to add custom path to try_umount list as well, users can use their own way to add only the desired sus mounts to try_umount list, but remember to disable susfs ADD_TRY_UMOUNT in kernel if users want to use the official one.
- There are less use cases for umount_for_zygote_system_process, and sometimes enabling this may cause bootloop with some modules enabled, instead user can use busybox nsenter to umount the sus mounts for specific process later by themmselves.
Co-authored-by: simonpunk <simonpunk2016@gmail.com>
Reason:
- add_try_umount, AUTO_ADD_SUS_KSU_DEFAULT_MOUNT and AUTO_ADD_SUS_BIND_MOUNT are only used to flag the sus_mount bit on target inode, but actually none of the hooks are checking for this bit, so this is why these features should be no longer needed, and we only need to check for sus mnt_id to determine if it is a sus mount. Plus it takes more CPU cycle to check for the path and set the sus_mount flag during boot time
Co-authored-by: simonpunk <simonpunk2016@gmail.com>
* use sys_enter tracepoint for sucompat
* update sucompat rules
* clean tif mark
* mark tif after load allow list
* clear all tif first, then mark target
* Fix shell su
* allow when escape
* fix bugs
* kernel: Resolve logical inconsistencies
---------
Co-authored-by: Ylarod <me@ylarod.cn>
Co-authored-by: weishu <twsxtd@gmail.com>
* refactor: replace throne tracker with ksud token
* use snprintf
* refactor: new supercall impl
- Import the sukisu command
* disable seccomp for supercall users
* kernel: fmt clear
* kernel: Enable macro protection for sulog
- Only enabled on kernel versions greater than 5.10.245
* kernel: Refactor kprobe hooks and implement LSM hooks for improved security handling
* debug mode
* kernel: Add functionality to generate and validate authentication tokens for cmd_su
* kernel: Simplified manual SU command processing for code
* kernel: replace renameat hook with fsnotify
* Revert "refactor: replace throne tracker with ksud token"
This reverts commit aa2cbbf.
* kernel: fix compile
* kernel: fix compile below 6.0
* Fix compile err; Add become_manager
* kernel: install fd for manager automaticlly
- extend to import the corresponding command
* manager: new supercall impl
* temp changes for ksud
* ksud: fix compile
* fix wrong opcode
* kernel: fix compile
* kernel: Fixed hook type and KPM status retrieval errors
* kernel: Fixed potential null pointer issue with current->mm in kernel version 5.10
When calling get_full_comm() within system call hooks, current->mm may be null (prctl). A fallback mechanism for current->comm must be added beforehand to prevent null pointer dereferences when accessing mm->arg_start/arg_end.
Signed-off-by: ShirkNeko <109797057+ShirkNeko@users.noreply.github.com>
* ksud: fix cargo check
* manager: Fixed an issue where the KSUD release and user-mode scanning switch failed to function correctly.
- kernel: fix spin lock mutual
kernel: Fixed potential null pointer issue with current->mm in kernel version 5.10
When calling get_full_comm() within system call hooks, current->mm may be null (prctl). A fallback mechanism for current->comm must be added beforehand to prevent null pointer dereferences when accessing mm->arg_start/arg_end.
kernel: try introduce like susfs's method to fix prctl delay
* seccomp: allow reboot
* use u32
* update clang-format
* 4 spaces save the world
* ksud: Fix build on macOS
* manager: bump minimal supported kernel.
- When get_hook_type is empty, display “Unknown”.
* Fix ksud build (#2841)
* try fix ksud
* fix for macos
* remove any
* Fix ksud build, take 3
* try fix allowlist
* bring lsm hook back
* fix: a lot again
* Fix ksud build, take 4 (#2846)
Remove init_driver_fd function for non-linux/android targets
* manager: Return to the native method via KSUd installation
* Merge with susfs-mian format
---------
Signed-off-by: ShirkNeko <109797057+ShirkNeko@users.noreply.github.com>
Co-authored-by: Ylarod <me@ylarod.cn>
Co-authored-by: weishu <twsxtd@gmail.com>
Co-authored-by: AlexLiuDev233 <wzylin11@outlook.com>
Co-authored-by: Wang Han <416810799@qq.com>
kernel & KernelSU: More Optimization on SUS_MOUNT and AUTO_ADD_TRY_UMOUNT_FOR_BIND_MOUNT; Dropped HAS_MAGIC_MOUNT feature in Kconfig
- In order to stop more performance loss, the changes here is not to monitor any ksu mounts anymore during/after boot-completed stage,
since monitoring the free of mnt_id and reordering the mnt_id takes some CPU cycles already.
- HAS_MAGIC_MOUNT is not needed anymore and catching magic mounts can be simplifed by checking whether
1. It is bind mount
2. Its mnt_devname is "KSU"
3. It is a directory
4. mounted path begins with "/debug_ramdisk/workdir/"
- Some sus bind mounts still cannot be catched with these changes, like revanced youtube module,
since it bind mounts after the user data is decrypted, at this point user may need to manually
add the mounted path to try_umount.
Co-authored-by: simonpunk <simonpunk2016@gmail.com>
Run throne_tracker() in kthread instead of blocking the caller.
Prevents full lockup during installation and removing the manager.
By default, first run remains synchronous for compatibility purposes
(FDE, FBEv1, FBEv2)
Features:
- looks and waits for manager UID in /data/system/packages.list
- run track_throne() in a kthread after the first synchronous run
- prevent duplicate thread creation with a single-instance check
- spinlock-on-d_lock based polling adressing possible race conditions.
Race conditions adressed
- single instance kthread lock, smp_mb()
- track_throne_function, packages.list, spinlock-on-d_lock based polling
- is_manager_apk, apk, spinlock-on-d_lock based polling
This is a squash of:
https://github.com/tiann/KernelSU/pull/2632
Original skeleton based on:
`kernelsu: move throne_tracker() to kthread`
`kernelsu: check locking before accessing files and dirs during searching manager`
`kernelsu: look for manager UID in /data/system/packages.list, not /data/system/packages.list.tmp`
0b05e927...8783badd
Co-Authored-By: backslashxx <118538522+backslashxx@users.noreply.github.com>
Co-Authored-By: Yaroslav Zviezda <10716792+acroreiser@users.noreply.github.com>
Signed-off-by: backslashxx <118538522+backslashxx@users.noreply.github.com>
- In order to reduce the extra patch code and increase the rate of hunk succeed, we do not rename the symbols in KernelSU drvier but instead do it all in kernel/kallsym.c
- Make susfs_starts_with() global as it maybe useful in future
Co-authored-by: simonpunk <simonpunk2016@gmail.com>
Tracepoint is a predefined hook point in the kernel, compared to Kprobe,
it is more stable and has lower performance overhead, although compatibility
is relatively poor, it is still worth trying
By the way, we have also included the config definitions related to hook types
in Kconfig, to enhance cleanliness
These patches is based on https://github.com/backslashxx/KernelSU/issues/5
Co-authored-by: Cloud_Yun <1770669041@qq.com>
Co-authored-by: Prslc <prslc113@gmail.com>
kernel & KernelSU: Replace kabi member susfs_task_state of tast_struct and use tast_struct->thread_info.flags instead; Fix for sus_path not working on some OEM devices; KSU_SUSFS requires CONFIG_THREAD_INFO_IN_TASK now;
- Since the kabi member is not guaranteed to be unused by OEM, that is why some deivces may conflict with the kabi member added by susfs, we should use less kabi member as much as possible.
- Luckily we have task_struct->thread_info that is enabled by all gki kernels just like kprobes, so we can utilize thread_info->flags for faster check, but it needs CONFIG_THREAD_INFO_IN_TASK to be on which should always be enabled
- Big thanks to fatalcoder524 (https://github.com/fatalcoder524) who did an excellent debug process for me
Co-authored-by: simonpunk <simonpunk2016@gmail.com>
Revise the explanation for KSU_MULTI_MANAGER_SUPPORT
Change-Id: I097b4e0d1a30c02fadd75d4123c7bcf346782f18
Signed-off-by: WenHao2130 <wenhao2130@outlook.com>
- We don't know if KPM can run on arm32-bit devices, so to avoid some problems, add a dependency on 64-bit architectures
Signed-off-by: ShirkNeko <109797057+ShirkNeko@users.noreply.github.com>
