I was looking at the wayland server implementation in Rust
/// The socket will be created at the specified path, and this method will acquire an associatet lockfile
/// alongside it.
pub fn bind_absolute(socket_path: PathBuf) -> Result<Self, BindError> {
let lock_path = socket_path.with_extension("lock");
let mut _lock;
// The locking code uses a loop to avoid an open()-flock() race condition, described in more
// detail in the comment below. The implementation roughtly follows the one from libbsd:
//
// https://gitlab.freedesktop.org/libbsd/libbsd/-/blob/73b25a8f871b3a20f6ff76679358540f95d7dbfd/src/flopen.c#L71
loop {
// open the lockfile
_lock = File::options()
.create(true)
.truncate(true)
.read(true)
.write(true)
.mode(0o660)
.open(&lock_path)
.map_err(|_| BindError::PermissionDenied)?;
And I see that they also have an associated lockfile. Why is this needed? Wouldn't an attempt to bind to an already in use socket throw an error anyways?
system
October 26, 2024, 4:05pm
2
This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.