linux-kernel - Re: [PATCH v2 2/2] rust: arc: remove `ArcBorrow` in favour of `WithRef`

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

Message-ID: <ZRGyRQuBcWvgtdNR@Boquns-Mac-mini.home>
Date:   Mon, 25 Sep 2023 09:16:05 -0700
From:   Boqun Feng <boqun.feng@...il.com>
To:     Benno Lossin <benno.lossin@...ton.me>
Cc:     Alice Ryhl <aliceryhl@...gle.com>,
        Wedson Almeida Filho <wedsonaf@...il.com>,
        rust-for-linux@...r.kernel.org, Miguel Ojeda <ojeda@...nel.org>,
        Alex Gaynor <alex.gaynor@...il.com>,
        Gary Guo <gary@...yguo.net>,
        Björn Roy Baron <bjorn3_gh@...tonmail.com>,
        Andreas Hindborg <a.hindborg@...sung.com>,
        linux-kernel@...r.kernel.org,
        Wedson Almeida Filho <walmeida@...rosoft.com>
Subject: Re: [PATCH v2 2/2] rust: arc: remove `ArcBorrow` in favour of
 `WithRef`

On Mon, Sep 25, 2023 at 03:07:44PM +0000, Benno Lossin wrote:
> On 25.09.23 16:49, Boqun Feng wrote:
> > On Mon, Sep 25, 2023 at 09:14:50AM +0000, Benno Lossin wrote:
> >> On 25.09.23 08:29, Alice Ryhl wrote:
> >>> On Sat, Sep 23, 2023 at 4:50 PM Wedson Almeida Filho <wedsonaf@...il.com> wrote:
> >>>>
> >>>> From: Wedson Almeida Filho <walmeida@...rosoft.com>
> >>>>
> >>>> With GATs, we don't need a separate type to represent a borrowed object
> >>>> with a refcount, we can just use Rust's regular shared borrowing. In
> >>>> this case, we use `&WithRef<T>` instead of `ArcBorrow<'_, T>`.
> >>>>
> >>>> Co-developed-by: Boqun Feng <boqun.feng@...il.com>
> >>>> Signed-off-by: Boqun Feng <boqun.feng@...il.com>
> >>>> Signed-off-by: Wedson Almeida Filho <walmeida@...rosoft.com>
> >>>> ---
> >>>>    rust/kernel/sync.rs     |   2 +-
> >>>>    rust/kernel/sync/arc.rs | 134 ++++++++++++----------------------------
> >>>>    2 files changed, 39 insertions(+), 97 deletions(-)
> >>>
> >>> I'm concerned about this change, because an `&WithRef<T>` only has
> >>> immutable permissions for the allocation. No pointer derived from it
> >>> may be used to modify the value in the Arc, however, the drop
> >>> implementation of Arc will do exactly that.
> >>
> >> That is indeed a problem. We could put the value in an `UnsafeCell`, but
> >> that would lose us niche optimizations and probably also other optimizations.
> >>
> > 
> > Not sure I understand the problem here, why do we allow modifying the
> > value in the Arc if you only have a shared ownership? Also I fail to see
> > why `ArcBorrow` doesn't have the problem. Maybe I'm missing something
> > subtle here? Could you provide an example?
> 
> Sure, here is the problem:
> 

Thanks, Benno.

> ```rust
> struct MutatingDrop {
>      value: i32,
> }
> 
> impl Drop for MutatingDrop {
>      fn drop(&mut self) {
>          self.value = 0;
>      }
> }
> 
> let arc = Arc::new(MutatingDrop { value: 42 });
> let wr = arc.as_with_ref(); // this creates a shared `&` reference to the MutatingDrop
> let arc2: Arc<MutatingDrop> = wr.into(); // increments the reference count to 2

More precisely, here we did a

	&WithRef<_> -> NonNull<WithRef<_>>

conversion, and later on, we may use the `NonNull<WithRef<_>>` in
`drop` to get a `Box<WithRef<_>>`.

> drop(arc); // this decrements the reference count to 1
> drop(arc2); // this decrements the reference count to 0, so it will drop it
> ```
> When dropping `arc2` it will run the destructor for `MutatingDrop`,
> which mutates `value`. This is a problem, because the mutable reference
> supplied was derived from a `&`, that is not allowed in Rust.
> 

Is this an UB? I kinda wonder what's the real damage we can get, because
in this case, we just use a reference to carry a value of a pointer,
i.e.

	ptr -> reference -> ptr

I cannot think of any real damage compiler can make, but I'm happy to be
surprised ;-)

Regards,
Boqun

> -- 
> Cheers,
> Benno
> 
>