netdev - Re: [PATCH net-next v3 1/3] rust: core abstractions for network PHY drivers

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]

Date: Thu, 12 Oct 2023 08:18:26 +0900 (JST)
From: FUJITA Tomonori <fujita.tomonori@...il.com>
To: miguel.ojeda.sandonis@...il.com
Cc: fujita.tomonori@...il.com, gregkh@...uxfoundation.org,
 netdev@...r.kernel.org, rust-for-linux@...r.kernel.org, andrew@...n.ch,
 tmgross@...ch.edu, wedsonaf@...il.com
Subject: Re: [PATCH net-next v3 1/3] rust: core abstractions for network
 PHY drivers

On Wed, 11 Oct 2023 11:59:01 +0200
Miguel Ojeda <miguel.ojeda.sandonis@...il.com> wrote:

> On Mon, Oct 9, 2023 at 5:50 PM FUJITA Tomonori
> <fujita.tomonori@...il.com> wrote:
>>
>> What feedback? enum stuff? I think that it's a long-term issue.
> 
> Not just that. There has been other feedback, and since this message,
> we got new reviews too.
> 
> But, yes, the `--rustified-enum` is one of those. I am still
> uncomfortable with it. It is not a huge deal for a while, and things
> will work, and the risk of UB is low. But why do we want to risk it?
> The point of using Rust is precisely to avoid this sort of thing.
>
> Why cannot we use one of the alternatives? If we really want to catch,
> right now, the "addition of new variant in the C enum" case, cannot we
> add a temporary check for that? e.g. it occurs to me we could make

IIRC, Andrew prefers to avoid creating a temporary rust variant (Greg
does too, I understand). I guess that only solusion that both Rust and
C devs would be happy with is generating safe Rust code from C. The
solution is still a prototype and I don't know when it will be
available (someone knows?).

I think that unlikely PHYLIB's state machine would be broken, so I
chose that approach with the code commented.


>> I'm not sure about it. For example, we reviewed the locking issue
>> three times. It can't be reviewed only on Rust side. It's mainly about
>> how the C side works.
> 
> We have never said it has to be reviewed only on the Rust side. In
> fact, our instructions for contributing explain very clearly the
> opposite:
> 
>     https://rust-for-linux.com/contributing#the-rust-subsystem
> 
> The instructions also say that the code must be warning-free and so
> on, and yet after several iterations and pushing for merging several
> times, there are still "surface-level" things like missing `// SAFETY`
> comments and `bindings::` in public APIs; which we consider very
> important -- we want to get them enforced by the compiler in the
> future.
> 
> Not only that, when I saw Wedson mentioning yesterday the
> `#[must_use]` bit, I wondered how this was even not being noticed by
> the compiler.
> 
> So I just took the v3 patches and compiled them and, indeed, Clippy gives you:

Sorry, there's no excuse. I should have done better. I'll make sure
that the code is warning-free.