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: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <D8FXFGYZTEXT.24UM4V3HZ5MWH@proton.me>
Date: Fri, 14 Mar 2025 10:45:34 +0000
From: Benno Lossin <benno.lossin@...ton.me>
To: christian <christiansantoslima21@...il.com>, rust-for-linux@...r.kernel.org, linux-kernel@...r.kernel.org, Miguel Ojeda <ojeda@...nel.org>, Alex Gaynor <alex.gaynor@...il.com>, Boqun Feng <boqun.feng@...il.com>, Gary Guo <gary@...yguo.net>, Björn Roy Baron <bjorn3_gh@...tonmail.com>, Andreas Hindborg <a.hindborg@...nel.org>, Alice Ryhl <aliceryhl@...gle.com>, Trevor Gross <tmgross@...ch.edu>, Danilo Krummrich <dakr@...nel.org>, ~lkcamp/patches@...ts.sr.ht, richard120310@...il.com
Subject: Re: [PATCH v4] Add methods for FromBytes trait.

On Fri Mar 14, 2025 at 4:49 AM CET, christian wrote:
> Methods receive a slice and perform size check to add
> a valid way to make conversion safe.
> In the invalid case return the EINVAL error.
>
> The conversion between slices ([T])
> is separated from others, because I couldn't implement it
> in the same way as the other conversions.
>
> Link: https://github.com/Rust-for-Linux/linux/issues/1119
> Signed-off-by: christian <christiansantoslima21@...il.com>
> ---

It usually is a good idea to include a changelog and a link to any prior
versions after this `---`. It won't be included in the final commit
message, but help reviewers and others keep track of this series.

>  rust/kernel/transmute.rs | 67 ++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 64 insertions(+), 3 deletions(-)
>
> diff --git a/rust/kernel/transmute.rs b/rust/kernel/transmute.rs
> index 1c7d43771a37..5924c0daccfc 100644
> --- a/rust/kernel/transmute.rs
> +++ b/rust/kernel/transmute.rs
> @@ -2,6 +2,8 @@
>  
>  //! Traits for transmuting types.
>  
> +use crate::prelude::{Error, EINVAL};
> +
>  /// Types for which any bit pattern is valid.
>  ///
>  /// Not all types are valid for all values. For example, a `bool` must be either zero or one, so
> @@ -12,26 +14,85 @@
>  /// # Safety
>  ///
>  /// All bit-patterns must be valid for this type. This type must not have interior mutability.
> -pub unsafe trait FromBytes {}
> +///
> +/// # Example

I think this section should go before the `Safety` section.

> +///
> +/// ```
> +/// let foo = &[1, 2, 3, 4];
> +///
> +/// let result = u8::from_bytes(foo);
> +///
> +/// assert_eq!(*result, 0x40300201);
> +/// ```
> +pub trait FromBytes {

Why is this trait becoming safe?

> +    /// Receives a slice of bytes and converts to a valid reference of Self when it's possible.
> +    fn from_bytes(slice_of_bytes: &[u8]) -> Result<&Self, Error>;

IMO it makes more sense for the return type to be `Option<&Self>`.

> +
> +    /// Receives a mutable slice of bytes and converts to a valid reference of Self when it's possible.
> +    fn from_bytes_mut(mut_slice_of_bytes: &mut [u8]) -> Result<&mut Self, Error>;

This must also require that `Self: AsBytes`, since otherwise the user
could write padding bytes into the original slice.

Also the parameter name `mut_slice_of_bytes` is a bit long, how about
`bytes`?

> +}
>  
>  macro_rules! impl_frombytes {
>      ($($({$($generics:tt)*})? $t:ty, )*) => {
>          // SAFETY: Safety comments written in the macro invocation.
> -        $(unsafe impl$($($generics)*)? FromBytes for $t {})*
> +        $(impl$($($generics)*)? FromBytes for $t {
> +            fn from_bytes(slice_of_bytes: &[u8]) -> Result<&$t, Error> {
> +                if slice_of_bytes.len() == core::mem::size_of::<$t>() {
> +                    let slice_ptr = slice_of_bytes.as_ptr() as *const $t;
> +                    unsafe { Ok(&*slice_ptr) }
> +                } else {
> +                    Err(EINVAL)
> +                }
> +            }
> +
> +            fn from_bytes_mut(mut_slice_of_bytes: &mut [u8]) -> Result<&mut $t, Error> {
> +                if mut_slice_of_bytes.len() == core::mem::size_of::<$t>() {
> +                    let slice_ptr = mut_slice_of_bytes.as_mut_ptr() as *mut $t;
> +                    unsafe { Ok(&mut *slice_ptr) }
> +                } else {
> +                    Err(EINVAL)
> +                }
> +            }
> +        })*
>      };
>  }
>  
>  impl_frombytes! {
>      // SAFETY: All bit patterns are acceptable values of the types below.
> +    // SAFETY: Dereferencing the pointer is safe because slice has the same size of Self.

What is this safety comment for?

>      u8, u16, u32, u64, usize,
>      i8, i16, i32, i64, isize,
>  
>      // SAFETY: If all bit patterns are acceptable for individual values in an array, then all bit
>      // patterns are also acceptable for arrays of that type.
> -    {<T: FromBytes>} [T],

If you're removing this line, you should also remove its safety comment
above. But since the trait should be `unsafe`, you should move it down
to the impl below.

> +    // SAFETY: Dereferencing the pointer is safe because slice has the same size of Self.
>      {<T: FromBytes, const N: usize>} [T; N],
>  }
>  
> +impl<T: FromBytes> FromBytes for [T] {
> +    fn from_bytes(slice_of_bytes: &[u8]) -> Result<&Self, Error> {
> +        let slice_ptr = slice_of_bytes.as_ptr() as *const T;
> +        if slice_of_bytes.len() % core::mem::size_of::<T>() == 0 {
> +            let slice_len = slice_of_bytes.len() / core::mem::size_of::<T>();
> +            // SAFETY: Creating a slice is safe because the slice can be divided into T sized blocks.

You're not justifying why the pointer is valid. Also please avoid
repeating the obvious "Creating a slice is safe".

---
Cheers,
Benno

> +            unsafe { Ok(core::slice::from_raw_parts(slice_ptr, slice_len)) }
> +        } else {
> +            Err(EINVAL)
> +        }
> +    }
> +
> +    fn from_bytes_mut(mut_slice_of_bytes: &mut [u8]) -> Result<&mut Self, Error> {
> +        let slice_ptr = mut_slice_of_bytes.as_mut_ptr() as *mut T;
> +        if mut_slice_of_bytes.len() % core::mem::size_of::<T>() == 0 {
> +            let slice_len = mut_slice_of_bytes.len() / core::mem::size_of::<T>();
> +            // SAFETY: Creating a slice is safe because the slice can be divided into T sized blocks.
> +            unsafe { Ok(core::slice::from_raw_parts_mut(slice_ptr, slice_len)) }
> +        } else {
> +            Err(EINVAL)
> +        }
> +    }
> +}
> +
>  /// Types that can be viewed as an immutable slice of initialized bytes.
>  ///
>  /// If a struct implements this trait, then it is okay to copy it byte-for-byte to userspace. This



Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ