Re: [PATCH v8 5/6] rust: rbtree: add `RBTreeCursor`

[Benno Lossin]

On 06.08.24 11:04, Alice Ryhl wrote:
> On Tue, Aug 6, 2024 at 11:01 AM Benno Lossin <benno.lossin@xxxxxxxxx> wrote:
>>
>> On 06.08.24 10:24, Alice Ryhl wrote:
>>> On Mon, Aug 5, 2024 at 9:35 PM Benno Lossin <benno.lossin@xxxxxxxxx> wrote:
>>>> On 27.07.24 22:30, Matt Gilbride wrote:
>>>>> +    tree: &'a mut RBTree<K, V>,
>>>>> +    current: NonNull<bindings::rb_node>,
>>>>> +}
>>>>> +
>>>>> +// SAFETY: The [`RBTreeCursor`] gives out immutable references to K and mutable references to V,
>>>>> +// so it has the same thread safety requirements as mutable references.
>>>>> +unsafe impl<'a, K: Send, V: Send> Send for RBTreeCursor<'a, K, V> {}
>>>>
>>>> Again, do we want to use `K: Sync` here instead?
>>>
>>> In this case, `K: Send` and `K: Sync` are both sufficient conditions,
>>> but `K: Send` will generally be less restrictive for the user.
>>
>> What if `K = struct(RefCell<i32>, i32)` where only the second i32 is
>> used in `(Partial)Ord`? Then you can send `RBTreeCursor` to another
>> thread and call `borrow` there, even though `K: !Sync` (and the value
>> still lives on another thread).
> 
> In that scenario, all immutable references to the key would be on the
> same thread as the cursor. The cursor borrows the tree mutably, so
> there can only be one cursor.
> 
> When using `K: Send`, it's basically like having `RBTreeCursor` store
> mutable references to the key, but with an API that downgrades to
> immutable reference when giving out access to the key.

Ah that's true.

---
Cheers,
Benno