linux-kernel - Re: [RFC PATCH v2 2/4] rust: io_uring: introduce rust abstraction for io-uring cmd

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: <aJdEbFI2FqSCBt9L@sidongui-MacBookPro.local>
Date: Sat, 9 Aug 2025 21:51:56 +0900
From: Sidong Yang <sidong.yang@...iosa.ai>
To: Benno Lossin <lossin@...nel.org>,
	Caleb Sander Mateos <csander@...estorage.com>
Cc: Daniel Almeida <daniel.almeida@...labora.com>,
	Miguel Ojeda <ojeda@...nel.org>, Arnd Bergmann <arnd@...db.de>,
	Jens Axboe <axboe@...nel.dk>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	rust-for-linux@...r.kernel.org, linux-kernel@...r.kernel.org,
	io-uring@...r.kernel.org
Subject: Re: [RFC PATCH v2 2/4] rust: io_uring: introduce rust abstraction
 for io-uring cmd

On Sat, Aug 09, 2025 at 12:18:49PM +0200, Benno Lossin wrote:
> On Fri Aug 8, 2025 at 11:43 AM CEST, Sidong Yang wrote:
> > On Fri, Aug 08, 2025 at 10:49:14AM +0200, Benno Lossin wrote:
> >> On Fri Aug 8, 2025 at 8:56 AM CEST, Sidong Yang wrote:
> >> > On Wed, Aug 06, 2025 at 03:38:24PM +0200, Benno Lossin wrote:
> >> >> On Wed Aug 6, 2025 at 2:38 PM CEST, Daniel Almeida wrote:
> >> >> > Hi Benno,
> >> >> >
> >> >> >> On 2 Aug 2025, at 07:52, Benno Lossin <lossin@...nel.org> wrote:
> >> >> >> 
> >> >> >> On Fri Aug 1, 2025 at 3:48 PM CEST, Daniel Almeida wrote:
> >> >> >>>> On 27 Jul 2025, at 12:03, Sidong Yang <sidong.yang@...iosa.ai> wrote:
> >> >> >>>> +    #[inline]
> >> >> >>>> +    pub fn pdu(&mut self) -> &mut MaybeUninit<[u8; 32]> {
> >> >> >>> 
> >> >> >>> Why MaybeUninit? Also, this is a question for others, but I don´t think
> >> >> >>> that `u8`s can ever be uninitialized as all byte values are valid for `u8`.
> >> >> >> 
> >> >> >> `u8` can be uninitialized. Uninitialized doesn't just mean "can take any
> >> >> >> bit pattern", but also "is known to the compiler as being
> >> >> >> uninitialized". The docs of `MaybeUninit` explain it like this:
> >> >> >> 
> >> >> >>    Moreover, uninitialized memory is special in that it does not have a
> >> >> >>    fixed value ("fixed" meaning "it won´t change without being written
> >> >> >>    to"). Reading the same uninitialized byte multiple times can give
> >> >> >>    different results.
> >> >> >> 
> >> >> >> But the return type probably should be `&mut [MaybeUninit<u8>; 32]`
> >> >> >> instead.
> >> >> >
> >> >> >
> >> >> > Right, but I guess the question then is why would we ever need to use
> >> >> > MaybeUninit here anyways.
> >> >> >
> >> >> > It's a reference to a C array. Just treat that as initialized.
> >> >> 
> >> >> AFAIK C uninitialized memory also is considered uninitialized in Rust.
> >> >> So if this array is not properly initialized on the C side, this would
> >> >> be the correct type. If it is initialized, then just use `&mut [u8; 32]`.
> >> >
> >> > pdu field is memory chunk for driver can use it freely. The driver usually
> >> > saves a private data and read or modify it on the other context. using
> >> > just `&mut [u8;32]` would be simple and easy to use.
> >> 
> >> Private data is usually handled using `ForeignOwnable` in Rust. What
> >> kind of data would be stored there? If it's a pointer, then `&mut [u8;
> >> 32]` would not be the correct choice.
> >
> > Most driver uses `io_uring_cmd_to_pdu` macro that casts address of pdu to
> > private data type. It seems that all driver use this macro has it's own
> > struct type. How about make 2 function for pdu? like store_pdu(), borrow_pdu().
> 
> We'd need to ensure that `borrow_pdu` can only be called if `store_pdu`
> has been called before. Is there any way we can just ensure that pdu is
> always initialized? Like a callback that's called once, before the value
> is used at all?

I've thought about this. As Celab said, returning `&mut MaybeUninit<[u8;32]> is
simple and best. Only driver knows it's initialized. There is no way to
check whether it's initialized with reading the pdu. The best way is to return
`&mut MaybeUninit<[u8;32]>` and driver initializes it in first time. After 
init, driver knows it's guranteed that it's initialized so it could call 
`assume_init_mut()`. And casting to other struct is another problem. The driver
is responsible for determining how to interpret the PDU, whether by using it
directly as a byte array or by performing an unsafe cast to another struct.

Thanks,
Sidong


> 
> ---
> Cheers,
> Benno