Use dedicated type for Mutex IDs

This should prevent tiny mistakes in handling the value.
This commit is contained in:
2021-03-19 21:18:50 +01:00
parent e42d862f19
commit 5f2e0e99a8
3 changed files with 89 additions and 45 deletions

View File

@@ -13,6 +13,8 @@ mod graph;
pub mod stdsync;
/// Counter for Mutex IDs. Atomic avoids the need for locking.
///
/// Should be part of the `MutexID` impl but static items are not yet a thing.
static ID_SEQUENCE: AtomicUsize = AtomicUsize::new(0);
thread_local! {
@@ -20,17 +22,40 @@ thread_local! {
///
/// Assuming that locks are roughly released in the reverse order in which they were acquired,
/// a stack should be more efficient to keep track of the current state than a set would be.
static HELD_LOCKS: RefCell<Vec<usize>> = RefCell::new(Vec::new());
static HELD_LOCKS: RefCell<Vec<MutexID>> = RefCell::new(Vec::new());
}
lazy_static! {
static ref DEPENDENCY_GRAPH: Mutex<DiGraph> = Default::default();
}
fn next_mutex_id() -> usize {
ID_SEQUENCE
.fetch_update(Ordering::SeqCst, Ordering::SeqCst, |id| id.checked_add(1))
.expect("Mutex ID wraparound happened, results unreliable")
/// Dedicated ID type for Mutexes
///
/// # Unstable
///
/// This type is currently private to prevent usage while the exact implementation is figured out,
/// but it will likely be public in the future.
///
/// One possible alteration is to make this type not `Copy` but `Drop`, and handle deregistering
/// the lock from there.
#[derive(Copy, Clone, Debug, PartialEq, Eq, Hash)]
struct MutexID(usize);
impl MutexID {
/// Get a new, unique, mutex ID.
///
/// This ID is guaranteed to be unique within the runtime of the program.
///
/// # Panics
///
/// This function may panic when there are no more mutex IDs available. The number of mutex ids
/// is `usize::MAX - 1` which should be plenty for most practical applications.
pub fn new() -> Self {
ID_SEQUENCE
.fetch_update(Ordering::SeqCst, Ordering::SeqCst, |id| id.checked_add(1))
.map(|id| Self(id))
.expect("Mutex ID wraparound happened, results unreliable")
}
}
/// Get a reference to the current dependency graph
@@ -41,11 +66,17 @@ fn get_depedency_graph() -> impl DerefMut<Target = DiGraph> {
}
/// Register that a lock is currently held
fn register_lock(lock: usize) {
fn register_lock(lock: MutexID) {
HELD_LOCKS.with(|locks| locks.borrow_mut().push(lock))
}
fn drop_lock(id: usize) {
/// Drop a lock held by the current thread.
///
/// # Panics
///
/// This function panics if the lock did not appear to be handled by this thread. If that happens,
/// that is an indication of a serious design flaw in this library.
fn drop_lock(id: MutexID) {
HELD_LOCKS.with(|locks| {
let mut locks = locks.borrow_mut();
@@ -56,7 +87,7 @@ fn drop_lock(id: usize) {
}
}
panic!("Tried to drop lock for mutex {} but it wasn't held", id)
panic!("Tried to drop lock for mutex {:?} but it wasn't held", id)
});
}
@@ -64,16 +95,23 @@ fn drop_lock(id: usize) {
///
/// If the dependency is new, check for cycles in the dependency graph. If not, there shouldn't be
/// any cycles so we don't need to check.
fn register_dependency(lock: usize) {
HELD_LOCKS.with(|locks| {
///
/// # Panics
///
/// This function panics if the new dependency would introduce a cycle.
fn register_dependency(lock: MutexID) {
if HELD_LOCKS.with(|locks| {
if let Some(&previous) = locks.borrow().last() {
let mut graph = get_depedency_graph();
if graph.add_edge(previous, lock) && graph.has_cycles() {
panic!("Mutex order graph should not have cycles");
}
graph.add_edge(previous, lock) && graph.has_cycles()
} else {
false
}
})
}) {
// Panic without holding the lock to avoid needlessly poisoning it
panic!("Mutex order graph should not have cycles");
}
}
#[cfg(test)]
@@ -87,10 +125,10 @@ mod tests {
#[test]
fn test_next_mutex_id() {
let initial = next_mutex_id();
let next = next_mutex_id();
let initial = MutexID::new();
let next = MutexID::new();
// Can't assert N + 1 because multiple threads running tests
assert!(initial < next);
assert!(initial.0 < next.0);
}
}