You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I encountered some issues while developing memory management (MM) using an intrusive linked list. As you may know, the Box pointer uses #global_allocator to allocate and deallocate memory on the heap. However, in the kernel, the memory allocated by #global_allocator needs to be managed. Therefore, I have implemented a customized allocator for Box, called Alloc.
However, the current intrusive implementation does not allow for the specification of allocators, and PointerOps trait is only implemented for Box<T, Global>. In issue #87, I applied generic associated types and provided an implementation, but it is not ideal. I am struggling to come up with a better solution.
The current implementation allows us to specify allocators when using the intrusive_adapter macro, as shown here:
use intrusive_collections::intrusive_adapter;
use intrusive_collections::{LinkedList, LinkedListLink};
use std::cell::Cell;
// A simple struct containing an instrusive link and a value
struct Test {
link: LinkedListLink,
value: Cell<i32>,
}
// The adapter describes how an object can be inserted into an intrusive
// collection. This is automatically generated using a macro.
intrusive_adapter!(TestAdapter = Global, Box<Test, Global>: Test { link: LinkedListLink });
// Create a list and some objects
let mut list = LinkedList::new(TestAdapter::new());
let a = Box::new(Test {
link: LinkedListLink::new(),
value: Cell::new(1),
});
let b = Box::new(Test {
link: LinkedListLink::new(),
value: Cell::new(2),
});
let c = Box::new(Test {
link: LinkedListLink::new(),
value: Cell::new(3),
});
// Insert the objects at the front of the list
list.push_front(a);
list.push_front(b);
list.push_front(c);
assert_eq!(list.iter().map(|x| x.value.get()).collect::<Vec<_>>(), [3, 2, 1]);
I would appreciate your assistance in finding a better solution that supports specifying allocators. I look forward to your reply.
The text was updated successfully, but these errors were encountered:
The PointerOps trait provides all the flexibility needed to support custom allocators: you need to hold a copy of the allocator in your custom MyPointerOps object and use it to reconstruct a Box<T, A> from a *const T.
However since you are using a custom PointerOps, you cannot use the intrusive_adapter macro and need to make a type that implements the Adapter trait manually.
No changes to this crate are needed, it's just that you can't use the DefaultPointerOps and intrusive_adapter! helpers with custom allocators.
Wow, thank you for the elegant solution! By the way, does the intrusive-rs crate support raw pointers? I understand that I may need to implement a custom PointerOps, but I am wondering if there is any additional work required to support raw pointers.
I have found that there is an UnsafeRef pointer available in the intrusive-rs crate. I can use this pointer instead of raw pointers, and free the memory by myself. Your design is very well thought out.
Hi, Amanieu,
I encountered some issues while developing memory management (MM) using an intrusive linked list. As you may know, the Box pointer uses #global_allocator to allocate and deallocate memory on the heap. However, in the kernel, the memory allocated by #global_allocator needs to be managed. Therefore, I have implemented a customized allocator for Box, called Alloc.
However, the current intrusive implementation does not allow for the specification of allocators, and PointerOps trait is only implemented for Box<T, Global>. In issue #87, I applied generic associated types and provided an implementation, but it is not ideal. I am struggling to come up with a better solution.
The current implementation allows us to specify allocators when using the intrusive_adapter macro, as shown here:
I would appreciate your assistance in finding a better solution that supports specifying allocators. I look forward to your reply.
The text was updated successfully, but these errors were encountered: