rust-for-linux.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Benno Lossin" <lossin@kernel.org>
To: "Marcelo Moreira" <marcelomoreira1905@gmail.com>,
	<dakr@kernel.org>, <ojeda@kernel.org>,
	<rust-for-linux@vger.kernel.org>, <skhan@linuxfoundation.org>,
	<linux-kernel-mentees@lists.linuxfoundation.org>,
	<~lkcamp/patches@lists.sr.ht>
Subject: Re: [PATCH v4 3/3] rust: revocable: split revoke_internal into revoke and revoke_nosync
Date: Thu, 12 Jun 2025 11:06:56 +0200	[thread overview]
Message-ID: <DAKFP38LXT84.QN50NZ2QB4WU@kernel.org> (raw)
In-Reply-To: <20250602232842.144304-4-marcelomoreira1905@gmail.com>

On Tue Jun 3, 2025 at 1:26 AM CEST, Marcelo Moreira wrote:
> This commit refactors the revocation mechanism by removing the generic
> `revoke_internal` function. Its logic is now directly integrated into
> two distinct public functions: `revoke()` and `revoke_nosync()`.
>
> `revoke_nosync()` is an `unsafe` function that requires the caller to
> guarantee no concurrent users, thus avoiding an RCU grace period.
> `revoke()` is a safe function that internally waits for the RCU grace
> period to ensure all concurrent accesses have completed before dropping
> the wrapped object.
>
> This change improves API clarity and simplifies associated `SAFETY`
> comments by making the synchronization behavior explicit in the function
> signatures.
>
> Suggested-by: Benno Lossin <lossin@kernel.org>
> Suggested-by: Danilo Krummrich <dakr@kernel.org>
> Signed-off-by: Marcelo Moreira <marcelomoreira1905@gmail.com>

One comment below, with that fixed:

Reviewed-by: Benno Lossin <lossin@kernel.org>

> ---
>  rust/kernel/revocable.rs | 38 +++++++++++++++-----------------------
>  1 file changed, 15 insertions(+), 23 deletions(-)
>
> diff --git a/rust/kernel/revocable.rs b/rust/kernel/revocable.rs
> index 43cc9bdc94f4..daf22e3a7d20 100644
> --- a/rust/kernel/revocable.rs
> +++ b/rust/kernel/revocable.rs
> @@ -126,22 +126,6 @@ pub fn try_access_with_guard<'a>(&'a self, _guard: &'a rcu::Guard) -> Option<&'a
>          }
>      }
>  
> -    /// # Safety
> -    ///
> -    /// Callers must ensure that there are no more concurrent users of the revocable object.
> -    unsafe fn revoke_internal<const SYNC: bool>(&self) {
> -        if self.is_available.swap(false, Ordering::Relaxed) {
> -            if SYNC {
> -                // SAFETY: Just an FFI call, there are no further requirements.
> -                unsafe { bindings::synchronize_rcu() };
> -            }
> -
> -            // SAFETY: We know `self.data` is valid because only one CPU can succeed the
> -            // `compare_exchange` above that takes `is_available` from `true` to `false`.
> -            unsafe { drop_in_place(self.data.get()) };
> -        }
> -    }
> -
>      /// Revokes access to and drops the wrapped object.
>      ///
>      /// Access to the object is revoked immediately to new callers of [`Revocable::try_access`],
> @@ -151,10 +135,12 @@ unsafe fn revoke_internal<const SYNC: bool>(&self) {
>      ///
>      /// Callers must ensure that there are no more concurrent users of the revocable object.
>      pub unsafe fn revoke_nosync(&self) {
> -        // SAFETY: By the safety requirement of this function, the caller ensures that nobody is
> -        // accessing the data anymore and hence we don't have to wait for the grace period to
> -        // finish.
> -        unsafe { self.revoke_internal::<false>() }
> +        if self.is_available.swap(false, Ordering::Relaxed) {
> +            // SAFETY: `Self::data` is valid for writes because of `Self`'s type invariants,
> +            // as `Self::is_available` is false due to the atomic swap, and by the safety

Please also use `self.data`/`self.is_available` here (& below) instead
of `Self::`.

---
Cheers,
Benno

> +            // requirements of this function, no thread is accessing `data` anymore.
> +            unsafe { drop_in_place(self.data.get()) };
> +        }
>      }
>  
>      /// Revokes access to and drops the wrapped object.
> @@ -165,9 +151,15 @@ pub unsafe fn revoke_nosync(&self) {
>      /// [`Revocable::try_access`] beforehand and still haven't dropped the returned guard), this
>      /// function waits for the concurrent access to complete before dropping the wrapped object.
>      pub fn revoke(&self) {
> -        // SAFETY: By passing `true` we ask `revoke_internal` to wait for the grace period to
> -        // finish.
> -        unsafe { self.revoke_internal::<true>() }
> +        if self.is_available.swap(false, Ordering::Relaxed) {
> +            // SAFETY: Just an FFI call, there are no further requirements.
> +            unsafe { bindings::synchronize_rcu() };
> +
> +            // SAFETY: `Self::data` is valid for writes because of `Self`'s type invariants,
> +            // as `Self::is_available` is false due to the atomic swap, and `synchronize_rcu`
> +            // ensures all prior RCU read-side critical sections have completed.
> +            unsafe { drop_in_place(self.data.get()) };
> +        }
>      }
>  }
>  


  reply	other threads:[~2025-06-12  9:06 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-02 23:26 [PATCH v4 0/3] rust: revocable: documentation and refactorings Marcelo Moreira
2025-06-02 23:26 ` [PATCH v4 1/3] rust: revocable: update write invariant and fix safety comments Marcelo Moreira
2025-06-12  9:02   ` Benno Lossin
2025-06-12 19:22     ` Marcelo Moreira
2025-06-14 18:05       ` Benno Lossin
2025-06-14 23:11         ` Marcelo Moreira
2025-06-15  8:38           ` Miguel Ojeda
2025-06-16  0:36             ` Marcelo Moreira
2025-06-16  7:15               ` Benno Lossin
2025-06-17  2:49                 ` Marcelo Moreira
2025-06-17  7:18                   ` Benno Lossin
2025-06-26 16:59                     ` Marcelo Moreira
2025-06-13 14:08     ` Danilo Krummrich
2025-06-02 23:26 ` [PATCH v4 2/3] rust: revocable: simplify RevocableGuard for internal safety Marcelo Moreira
2025-06-12  9:04   ` Benno Lossin
2025-06-12  9:28   ` Alice Ryhl
2025-06-12  9:52     ` Benno Lossin
2025-06-12 18:52       ` Marcelo Moreira
2025-06-14 18:04         ` Benno Lossin
2025-06-13 14:11     ` Danilo Krummrich
2025-06-14 17:00       ` Benno Lossin
2025-06-02 23:26 ` [PATCH v4 3/3] rust: revocable: split revoke_internal into revoke and revoke_nosync Marcelo Moreira
2025-06-12  9:06   ` Benno Lossin [this message]
2025-06-12 19:29     ` Marcelo Moreira
2025-06-13 14:09     ` Danilo Krummrich
2025-06-16 10:26 ` [PATCH v4 0/3] rust: revocable: documentation and refactorings Danilo Krummrich
2025-06-16 19:33   ` Miguel Ojeda

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=DAKFP38LXT84.QN50NZ2QB4WU@kernel.org \
    --to=lossin@kernel.org \
    --cc=dakr@kernel.org \
    --cc=linux-kernel-mentees@lists.linuxfoundation.org \
    --cc=marcelomoreira1905@gmail.com \
    --cc=ojeda@kernel.org \
    --cc=rust-for-linux@vger.kernel.org \
    --cc=skhan@linuxfoundation.org \
    --cc=~lkcamp/patches@lists.sr.ht \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).