linux-kernel - Re: [PATCH 5/7] rust: init: add `..Zeroable::zeroed()` syntax for zeroing all missing fields

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 for Android: free password hash cracker in your pocket

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

Message-ID: <0f99bc49-5280-c300-719f-86c138f48eaf@proton.me>
Date:   Sun, 25 Jun 2023 16:46:19 +0000
From:   Benno Lossin <benno.lossin@...ton.me>
To:     Björn Roy Baron <bjorn3_gh@...tonmail.com>
Cc:     Miguel Ojeda <ojeda@...nel.org>,
        Wedson Almeida Filho <wedsonaf@...il.com>,
        Alex Gaynor <alex.gaynor@...il.com>,
        Boqun Feng <boqun.feng@...il.com>, Gary Guo <gary@...yguo.net>,
        Alice Ryhl <aliceryhl@...gle.com>,
        Andreas Hindborg <nmi@...aspace.dk>,
        rust-for-linux@...r.kernel.org, linux-kernel@...r.kernel.org,
        patches@...ts.linux.dev, Asahi Lina <lina@...hilina.net>
Subject: Re: [PATCH 5/7] rust: init: add `..Zeroable::zeroed()` syntax for zeroing all missing fields

On 6/25/23 16:17, Björn Roy Baron wrote:
> On Sunday, June 25th, 2023 at 15:07, Benno Lossin <benno.lossin@...ton.me> wrote:
>> On 25.06.23 14:56, Björn Roy Baron wrote:
>>> On Saturday, June 24th, 2023 at 23:14, Benno Lossin <benno.lossin@...ton.me> wrote:
>>
>>>>>> +                        // Ensure that the struct is indeed `Zeroable`.
>>>>>> +                        is_zeroable(slot);
>>>>>> +                        // SAFETY:  The type implements `Zeroable` by the check above.
>>>>>> +                        unsafe { ::core::ptr::write_bytes(slot, 0, 1) };
>>>>>> +                        $init_zeroed // this will be `()` if set.
>>>>>
>>>>> How does this work? Shouldn't there be a ; after $init_zeroed to consume the () value?
>>>>
>>>> It is the last expression of a block and since it is `()` it is ok
>>>> (adding a ; would also be ok, but it is not necessary).
>>>
>>> I'm surprised it is considered the last expression of a block. Unlike with {} using $()? will still
>>> allow variables defined inside this as if they were outside of it. Also I can't reproduce this
>>> behavior with:
>>>
>>>       macro_rules! foo {
>>>           ($($a:expr)?) => {
>>>               $($a)?
>>>               bar();
>>>           }
>>>       }
>>>
>>>       fn main() {
>>>           foo!(());
>>>       }
>>>
>>> Is there something I'm missing?
>>>
>>> Cheers,
>>> Björn
>>
>> Not sure what you mean with "allow variables defined inside this
>> as if they were outside of it". But note that in the macro `$init_zeroed`
>> is the last expression of a block. Here is a small example:
> 
> $(let $this = unsafe { ::core::ptr::NonNull::new_unchecked(slot) };)? comes after
> this code in the same block that contains struct __InitOk;. And after that another
> $crate::__init_internal!() invocation. That is why I don't get that this is allowed
> at all.
> 

Oh I see the issue now, I thought I wrote
```
$({
     fn assert_zeroable<T: Zeroable>(ptr: *mut T) {}
     // Ensure that the struct is indeed `Zeroable`.
     assert_zeroable(slot);
     // SAFETY:  The type implements `Zeroable` by the check above.
     unsafe { ::core::ptr::write_bytes(slot, 0, 1) };
     $init_zeroed // this will be `()` if set.
})?
```

But I forgot the inner `{}`. Good catch!

--
Cheers,
Benno

>>
>> ```
>> macro_rules! foo {
>>       ($($a:expr)?) => {{
>>           $(
>>               bar();
>>               $a
>>           )?
>>       }};
>> }
>>
>> fn bar() {}
>>
>> fn main() {
>>       foo!(());
>>       foo!();
>> }
>> ```
>>
>> it expands to this:
>> ```
>> fn main() {
>>       {
>>           bar();
>>           ()
>>       };
>>       {};
>> }
>> ```
>>
>> --
>> Cheers,
>> Benno
>>