Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Feedback-ID: i56684263:Fastmail
Message-ID: <d50b28f0-e989-7392-505b-2dd9f00abef6@ryhl.io>
Date:   Mon, 3 Apr 2023 19:56:31 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.9.0
From:   Alice Ryhl <alice@ryhl.io>
Subject: Re: [PATCH v5 14/15] rust: sync: reduce stack usage of
 `UniqueArc::try_new_uninit`
To:     Benno Lossin <y86-dev@protonmail.com>
Cc:     rust-for-linux@vger.kernel.org, linux-kernel@vger.kernel.org,
        patches@lists.linux.dev, Alice Ryhl <aliceryhl@google.com>,
        Andreas Hindborg <a.hindborg@samsung.com>,
        Miguel Ojeda <ojeda@kernel.org>,
        Alex Gaynor <alex.gaynor@gmail.com>,
        Wedson Almeida Filho <wedsonaf@gmail.com>,
        Boqun Feng <boqun.feng@gmail.com>, Gary Guo <gary@garyguo.net>,
        =?UTF-8?Q?Bj=c3=b6rn_Roy_Baron?= <bjorn3_gh@protonmail.com>
References: <20230403154422.168633-1-y86-dev@protonmail.com>
 <20230403160511.174894-7-y86-dev@protonmail.com>
Content-Language: en-US
In-Reply-To: <20230403160511.174894-7-y86-dev@protonmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 4/3/23 18:05, Benno Lossin wrote:
> `UniqueArc::try_new_uninit` calls `Arc::try_new(MaybeUninit::uninit())`.
> This results in the uninitialized memory being placed on the stack,
> which may be arbitrarily large due to the generic `T` and thus could
> cause a stack overflow for large types.
> 
> Change the implementation to use the pin-init API which enables in-place
> initialization. In particular it avoids having to first construct and
> then move the uninitialized memory from the stack into the final location.
> 
> Signed-off-by: Benno Lossin <y86-dev@protonmail.com>
> Cc: Gary Guo <gary@garyguo.net>
> Cc: Alice Ryhl <aliceryhl@google.com>
> Cc: Andreas Hindborg <a.hindborg@samsung.com>

Reviewed-by: Alice Ryhl <aliceryhl@google.com>

> 
>       /// Tries to allocate a new [`UniqueArc`] instance whose contents are not initialised yet.
>       pub fn try_new_uninit() -> Result<UniqueArc<MaybeUninit<T>>, AllocError> {
> -        Ok(UniqueArc::<MaybeUninit<T>> {
> +        // INVARIANT: The refcount is initialised to a non-zero value.
> +        let inner = Box::try_init::<AllocError>(try_init!(ArcInner {
> +            // SAFETY: There are no safety requirements for this FFI call.
> +            refcount: Opaque::new(unsafe { bindings::REFCOUNT_INIT(1) }),
> +            data <- init::uninit::<T, AllocError>(),
> +        }? AllocError))?;
> +        Ok(UniqueArc {
>               // INVARIANT: The newly-created object has a ref-count of 1.
> -            inner: Arc::try_new(MaybeUninit::uninit())?,
> +            // SAFETY: The pointer from the `Box` is valid.
> +            inner: unsafe { Arc::from_inner(Box::leak(inner).into()) },
>           })
>       }
>   }

I'm curious - do you know whether this compiles to the same machine code 
as this?

pub fn try_new_uninit() -> Result<UniqueArc<MaybeUninit<T>>, AllocError> {
     let inner: Box<MaybeUninit<ArcInner<T>>> = Box::try_new_uninit()?;
     let ptr = Box::into_raw(inner) as *mut ArcInner<T>;
     addr_of_mut!((*ptr).refcount).write(bindings::REFCOUNT_INIT(1));
     Ok(UniqueArc {
         inner: Arc {
             ptr: unsafe { NonNull::new_unchecked(ptr) },
             _p: PhantomData,
         }
     })
}