serenity/Base/usr/share/man/man2/pledge.md at aae106e37b48f2158e68902293df1e4bf7b80c0f

mirror of https://github.com/RGBCube/serenity synced 2025-05-20 06:45:07 +00:00

Liav A 23a7ccf607 Kernel+LibCore+LibC: Split the mount syscall into multiple syscalls

This is a preparation before we can create a usable mechanism to use
filesystem-specific mount flags.
To keep some compatibility with userland code, LibC and LibCore mount
functions are kept being usable, but now instead of doing an "atomic"
syscall, they do multiple syscalls to perform the complete procedure of
mounting a filesystem.

The FileBackedFileSystem IntrusiveList in the VFS code is now changed to
be protected by a Mutex, because when we mount a new filesystem, we need
to check if a filesystem is already created for a given source_fd so we
do a scan for that OpenFileDescription in that list. If we fail to find
an already-created filesystem we create a new one and register it in the
list if we successfully mounted it. We use a Mutex because we might need
to initiate disk access during the filesystem creation, which will take
other mutexes in other parts of the kernel, therefore making it not
possible to take a spinlock while doing this.

2023-07-02 01:04:51 +02:00

3.8 KiB

Raw Blame History

Name

pledge - reduce process capabilities

Synopsis

#include <unistd.h>

int pledge(const char* promises, const char* execpromises);

Description

pledge() makes a promise to the kernel that from this moment on, the calling process will only use a subset of system functionality.

Functionality is divided into a curated set of promises (described below), which can be combined to cover the program's needs. Both arguments are space-separated lists of promises.

Note that pledge() can be called repeatedly to remove previously-pledged promises, but it can never regain capabilities once lost.

promises are applied to the current process, and will also be inherited by children created by fork(2).

execpromises are applied if/when a new process image is created with exec(2).

If promises or execpromises is null, the corresponding value is unchanged.

If the process later attempts to use any system functionality it has previously promised not to use, the process is instantly terminated. Note that a process that has not ever called pledge() is considered to not have made any promises, and is allowed use any system functionality (subject to regular permission checks).

pledge() is intended to be used in programs that want to sandbox themselves, either to limit the impact of a possible vulnerability exploitation, or before intentionally executing untrusted code.

Promises

stdio: Basic I/O, memory allocation, information about self, various non-destructive syscalls
thread: The POSIX threading API (*)
id: Ability to change UID/GID
tty: TTY related functionality
proc: Process and scheduling related functionality
exec: The exec(2) syscall
unix: UNIX local domain sockets
inet: IPv4 domain sockets
accept: May use accept(2) to accept incoming socket connections on already listening sockets (*)
rpath: "Read" filesystem access
wpath: "Write" filesystem access
cpath: "Create" filesystem access
dpath: Creating new device files
chown: Changing file owner/group
fattr: Changing file attributes/permissions
video: May use ioctl(2) and mmap(2) on framebuffer video devices
settime: Changing the system time and date
setkeymap: Changing the system keyboard layout (*)
sigaction: Change signal handlers and dispositions (*)
sendfd: Send file descriptors over a local socket
recvfd: Receive file descriptors over a local socket
ptrace: The ptrace(2) syscall (*)
prot_exec: mmap(2) and mprotect(2) with PROT_EXEC
map_fixed: mmap(2) with MAP_FIXED or MAP_FIXED_NOREPLACE (*)
mount: mount(2) Various filesystem mount related syscalls (*)
no_error: Ignore requests of pledge elevation going forwards, this is useful for enforcing execpromises while the child process wants to ask for more upfront (Note that the elevation requests are not granted, merely ignored), this is similar to the error pledge in OpenBSD.
jail: Various jail-specific syscalls (*)

Promises marked with an asterisk (*) are SerenityOS specific extensions not supported by the original OpenBSD pledge().

Errors

EFAULT: promises and/or execpromises are not null and not in readable memory.
EINVAL: One or more invalid promises were specified.
EPERM: An attempt to increase capabilities was rejected.

History

The pledge() system call was first introduced by OpenBSD. The implementation in SerenityOS differs in many ways and is by no means final.

3.8 KiB Raw Blame History