rust: alloc: fix dangling pointer in VecExt<T>::reserve()
authorDanilo Krummrich <dakr@redhat.com>
Wed, 1 May 2024 13:47:43 +0000 (15:47 +0200)
committerMiguel Ojeda <ojeda@kernel.org>
Tue, 7 May 2024 22:43:30 +0000 (00:43 +0200)
commit97ab3e8eec0ce79d9e265e6c9e4c480492180409
tree2df4dafe7d875b38f3ccf7f5a7cf08f068ea4b37
parent56f64b370612d8967df2c2e0cead805444d4e71a
rust: alloc: fix dangling pointer in VecExt<T>::reserve()

Currently, a Vec<T>'s ptr value, after calling Vec<T>::new(), is
initialized to Unique::dangling(). Hence, in VecExt<T>::reserve(), we're
passing a dangling pointer (instead of NULL) to krealloc() whenever a new
Vec<T>'s backing storage is allocated through VecExt<T> extension
functions.

This only works as long as align_of::<T>(), used by Unique::dangling() to
derive the dangling pointer, resolves to a value between 0x0 and
ZERO_SIZE_PTR (0x10) and krealloc() hence treats it the same as a NULL
pointer however.

This isn't a case we should rely on, since there may be types whose
alignment may exceed the range still covered by krealloc(), plus other
kernel allocators are not as tolerant either.

Instead, pass a real NULL pointer to krealloc_aligned() if Vec<T>'s
capacity is zero.

Fixes: 5ab560ce12ed ("rust: alloc: update `VecExt` to take allocation flags")
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Benno Lossin <benno.lossin@proton.me>
Signed-off-by: Danilo Krummrich <dakr@redhat.com>
Reviewed-by: Wedson Almeida Filho <walmeida@microsoft.com>
Link: https://lore.kernel.org/r/20240501134834.22323-1-dakr@redhat.com
[ Solved `use` conflict and applied the `if`-instead-of-`match` change
  discussed in the list. - Miguel ]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
rust/kernel/alloc/vec_ext.rs