[PATCH] rust: net: netlink: Migrate to zerocopy's `IntoBytes`

Sagar Taunk posted 1 patch 1 week, 6 days ago
rust/kernel/net/netlink.rs | 23 ++++++++++++++++-------
1 file changed, 16 insertions(+), 7 deletions(-)
[PATCH] rust: net: netlink: Migrate to zerocopy's `IntoBytes`
Posted by Sagar Taunk 1 week, 6 days ago
Replace the kernel's own `transmute::FromBytes` and `AsBytes` traits
with their zerocopy equivalents. Specifically, this updates
`GenlMsg::put` to rely on `IntoBytes` for converting attributes into
byte slices.

Also, add `Immutable` trait bound on `GenlMsg::put` as zerocopy
splits the `no interior mutability` guarantee that `AsBytes` bundled
together.

Moreover, shashiko pointed out `put()` trusted an unchecked `as` cast
from `usize` to `c_int` for the attribute length. A length larger than
`i32::MAX` wraps to a negative value in that cast, which can pass
`nla_put()`'s own signed `skb_tailroom()` check and then be reinterpreted
as an enormous unsigned length via `__nla_reserve()`/`skb_put()`, leading
to a kernel panic via `skb_over_panic()`. Validate the length with
`c_int::try_from()` and reject it with `EMSGSIZE` instead.

Link: https://github.com/Rust-for-Linux/linux/issues/1241
Signed-off-by: Sagar Taunk <sagartaunk@proton.me>
---
 rust/kernel/net/netlink.rs | 23 ++++++++++++++++-------
 1 file changed, 16 insertions(+), 7 deletions(-)

diff --git a/rust/kernel/net/netlink.rs b/rust/kernel/net/netlink.rs
index 22ef3dde36fa..f929f63b32c1 100644
--- a/rust/kernel/net/netlink.rs
+++ b/rust/kernel/net/netlink.rs
@@ -12,11 +12,12 @@
     alloc::{self, AllocError},
     error::to_result,
     prelude::*,
-    transmute::AsBytes,
     types::Opaque,
     ThisModule,
 };
 
+use zerocopy::{Immutable, IntoBytes};
+
 use core::{
     mem::ManuallyDrop,
     ptr::NonNull, //
@@ -84,14 +85,22 @@ impl GenlMsg {
     #[inline]
     fn put<T>(&mut self, attrtype: c_int, value: &T) -> Result
     where
-        T: ?Sized + AsBytes,
+        T: ?Sized + IntoBytes + Immutable,
     {
         let skb = self.skb.skb.as_ptr();
-        let len = size_of_val(value);
-        let ptr = core::ptr::from_ref(value).cast::<c_void>();
-        // SAFETY: `skb` is valid by `NetlinkSkBuff` type invariants, and the provided value is
-        // readable and initialized for its `size_of` bytes.
-        to_result(unsafe { bindings::nla_put(skb, attrtype, len as c_int, ptr) })
+        let bytes = value.as_bytes();
+        // `nla_put()` takes attrlen as a plain `c_int`. If `bytes.len()`
+        // doesn't fit, an `as` cast would wrap around a negative value.
+        // Which then, would feed a huge unsigned length to `__nla_reserve()`
+        // and `skb_put()` causing it to panic via `skb_over_panic()`. So,
+        // the following check will reject it instead.
+        let len = c_int::try_from(bytes.len()).map_err(|_| EMSGSIZE)?;
+        let ptr = bytes.as_ptr().cast::<c_void>();
+        // SAFETY: `skb` is valid as per `NetlinkSkBuff` type invariants.
+        // `bytes` is a valid Rust slice, so `ptr` is readable for `len`
+        // bytes, and `T: Immutable` guarantees nothing can mutate `*value`
+        // while `nla_put()` copies it.
+        to_result(unsafe { bindings::nla_put(skb, attrtype, len, ptr) })
     }
 
     /// Puts a `u32` attribute into the message.
-- 
2.55.0
Re: [PATCH] rust: net: netlink: Migrate to zerocopy's `IntoBytes`
Posted by Alexandre Courbot 1 week, 5 days ago
On Sat Sep 12, 2026 at 3:25 PM JST, Sagar Taunk wrote:
> Replace the kernel's own `transmute::FromBytes` and `AsBytes` traits
> with their zerocopy equivalents. Specifically, this updates
> `GenlMsg::put` to rely on `IntoBytes` for converting attributes into
> byte slices.
>
> Also, add `Immutable` trait bound on `GenlMsg::put` as zerocopy
> splits the `no interior mutability` guarantee that `AsBytes` bundled
> together.
>
> Moreover, shashiko pointed out `put()` trusted an unchecked `as` cast
> from `usize` to `c_int` for the attribute length. A length larger than
> `i32::MAX` wraps to a negative value in that cast, which can pass
> `nla_put()`'s own signed `skb_tailroom()` check and then be reinterpreted
> as an enormous unsigned length via `__nla_reserve()`/`skb_put()`, leading
> to a kernel panic via `skb_over_panic()`. Validate the length with
> `c_int::try_from()` and reject it with `EMSGSIZE` instead.

The conversion and length validation are two different things and should
be split into their own patches.