Re: [PATCH 1/2] rust: lockdep: Remove support for dynamically allocated LockClassKeys

[Dirk Behme <dirk.behme@gmail.com>]

Am 05.10.24 um 00:01 schrieb Mitchell Levy via B4 Relay:
> From: Mitchell Levy <levymitchell0@gmail.com>
> 
> Currently, dynamically allocated LockCLassKeys can be used from the Rust
> side without having them registered. This is a soundness issue, so
> remove them.
> 
> Suggested-by: Alice Ryhl <aliceryhl@google.com>
> Link: https://lore.kernel.org/rust-for-linux/20240815074519.2684107-3-nmi@metaspace.dk/
> Cc: stable@vger.kernel.org
> Signed-off-by: Mitchell Levy <levymitchell0@gmail.com>
> ---
>   rust/kernel/lib.rs  |  2 +-
>   rust/kernel/sync.rs | 14 ++------------
>   2 files changed, 3 insertions(+), 13 deletions(-)
> 
> diff --git a/rust/kernel/lib.rs b/rust/kernel/lib.rs
> index 22a3bfa5a9e9..b5f4b3ce6b48 100644
> --- a/rust/kernel/lib.rs
> +++ b/rust/kernel/lib.rs
> @@ -44,8 +44,8 @@
>   pub mod page;
>   pub mod prelude;
>   pub mod print;
> -pub mod sizes;
>   pub mod rbtree;
> +pub mod sizes;
>   mod static_assert;
>   #[doc(hidden)]
>   pub mod std_vendor;


This is fixed already

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/rust/kernel/lib.rs?id=ece207a83e464af710d641f29e32b7a144c48e79

and can be dropped here.


> diff --git a/rust/kernel/sync.rs b/rust/kernel/sync.rs
> index 0ab20975a3b5..d270db9b9894 100644
> --- a/rust/kernel/sync.rs
> +++ b/rust/kernel/sync.rs
> @@ -27,28 +27,18 @@
>   unsafe impl Sync for LockClassKey {}
>   
>   impl LockClassKey {
> -    /// Creates a new lock class key.
> -    pub const fn new() -> Self {
> -        Self(Opaque::uninit())
> -    }
> -
>       pub(crate) fn as_ptr(&self) -> *mut bindings::lock_class_key {
>           self.0.get()
>       }
>   }
>   
> -impl Default for LockClassKey {
> -    fn default() -> Self {
> -        Self::new()
> -    }
> -}
> -
>   /// Defines a new static lock class and returns a pointer to it.
>   #[doc(hidden)]
>   #[macro_export]
>   macro_rules! static_lock_class {
>       () => {{
> -        static CLASS: $crate::sync::LockClassKey = $crate::sync::LockClassKey::new();


Should the SAFETY comment added in the 2nd patch go to here?

+        // SAFETY: lockdep expects uninitialized memory when it's 
handed a statically allocated
+        // lock_class_key


> +        static CLASS: $crate::sync::LockClassKey =
> +            unsafe { ::core::mem::MaybeUninit::uninit().assume_init() };
>           &CLASS
>       }};
>   }
>