linux-kernel - Re: [PATCH] rust: rbtree: fix cursor method lifetimes to match tree lifetime

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAH5fLgirP4=RrDcuF0n74KvN05K4FkAm9uZUfwX0WYfM+pKP8g@mail.gmail.com>
Date: Fri, 7 Nov 2025 10:27:16 +0100
From: Alice Ryhl <aliceryhl@...gle.com>
To: Hang Shu <m18080292938@....com>
Cc: ojeda@...nel.org, Hang Shu <hangshu847@...il.com>, 
	Alex Gaynor <alex.gaynor@...il.com>, Boqun Feng <boqun.feng@...il.com>, 
	Gary Guo <gary@...yguo.net>, Björn Roy Baron <bjorn3_gh@...tonmail.com>, 
	Benno Lossin <lossin@...nel.org>, Andreas Hindborg <a.hindborg@...nel.org>, 
	Trevor Gross <tmgross@...ch.edu>, Danilo Krummrich <dakr@...nel.org>, 
	Charalampos Mitrodimas <charmitro@...teo.net>, Borys Tyran <borys.tyran@...tonmail.com>, 
	Onur Özkan <work@...rozkan.dev>, 
	Daniel Sedlak <daniel@...lak.dev>, Tamir Duberstein <tamird@...il.com>, 
	Matt Gilbride <mattgilbride@...gle.com>, rust-for-linux@...r.kernel.org, 
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] rust: rbtree: fix cursor method lifetimes to match tree lifetime

On Fri, Nov 7, 2025 at 6:07 AM Hang Shu <m18080292938@....com> wrote:
>
> From: Hang Shu <hangshu847@...il.com>
>
> The returned keys and values of cursor methods should be bound by
> the lifetime of the rbtree itself ('a), not the lifetime of the cursor.
>
> Without this adjustment, examples like the following fail to compile:
>
> fn test_rbtree_cursor(rbtree: &mut RBTree<i32, i32>) -> &i32 {
>     rbtree.try_create_and_insert(1, 1, GFP_KERNEL).unwrap();
>     let mut cursor = rbtree.cursor_front().unwrap();
>     // compile error
>     // cannot return value referencing local variable `cursor`
>     cursor.peek_next().unwrap().1
> }
>
> This modification ensures that references to tree elements remain valid
> independently of the cursor's scope,
> aligning with the actual lifetime dependencies in the data structure.
>
> The changes will be applied to multiple similar methods
> throughout the Cursor implementation to maintain consistency.
>
> Fixes: 98c14e40e07a ("rust: rbtree: add cursor")
> Signed-off-by: Hang Shu <hangshu847@...il.com>
> ---
>  rust/kernel/rbtree.rs | 16 ++++++++--------
>  1 file changed, 8 insertions(+), 8 deletions(-)
>
> diff --git a/rust/kernel/rbtree.rs b/rust/kernel/rbtree.rs
> index 9e178dacddf1..702a1b6ef7a9 100644
> --- a/rust/kernel/rbtree.rs
> +++ b/rust/kernel/rbtree.rs
> @@ -742,7 +742,7 @@ unsafe impl<'a, K: Sync, V: Sync> Sync for Cursor<'a, K, V> {}
>
>  impl<'a, K, V> Cursor<'a, K, V> {
>      /// The current node
> -    pub fn current(&self) -> (&K, &V) {
> +    pub fn current(&self) -> (&'a K, &'a V) {
>          // SAFETY:
>          // - `self.current` is a valid node by the type invariants.
>          // - We have an immutable reference by the function signature.
> @@ -750,7 +750,7 @@ pub fn current(&self) -> (&K, &V) {
>      }
>
>      /// The current node, with a mutable value
> -    pub fn current_mut(&mut self) -> (&K, &mut V) {
> +    pub fn current_mut(&mut self) -> (&'a K, &'a mut V) {

This would allow me to call current_mut() twice on the same cursor to
get two mutable references to the same value. That is not okay.

If you want to have methods that return a reference with the tree's
lifetime instead of the cursor's, then you need to add new methods
(probably called into_*) rather than modify the existing ones.

Alice