linux-kernel - Re: C aggregate passing (Rust kernel policy)

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: <jei2vzllt5z6hufn3orqqupq3ufajcxv4zx3yfpkd2j54yr72e@vmhu4r3klvvh>
Date: Thu, 27 Feb 2025 14:55:11 -0500
From: Kent Overstreet <kent.overstreet@...ux.dev>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Ralf Jung <post@...fj.de>, Martin Uecker <uecker@...raz.at>, 
	"Paul E. McKenney" <paulmck@...nel.org>, Alice Ryhl <aliceryhl@...gle.com>, 
	Ventura Jack <venturajack85@...il.com>, Gary Guo <gary@...yguo.net>, airlied@...il.com, 
	boqun.feng@...il.com, david.laight.linux@...il.com, ej@...i.de, 
	gregkh@...uxfoundation.org, hch@...radead.org, hpa@...or.com, ksummit@...ts.linux.dev, 
	linux-kernel@...r.kernel.org, miguel.ojeda.sandonis@...il.com, rust-for-linux@...r.kernel.org
Subject: Re: C aggregate passing (Rust kernel policy)

On Thu, Feb 27, 2025 at 11:15:54AM -0800, Linus Torvalds wrote:
> On Thu, 27 Feb 2025 at 10:33, Ralf Jung <post@...fj.de> wrote:
> >
> > The way you do global flags in Rust is like this:
> 
> Note that I was really talking mainly about the unsafe cases, an din
> particular when interfacing with C code.

For simple bitflags (i.e. code where we use test_bit()/set_bit() we'd
probably just export it as a standard Rust atomic, no new unsafe {}
required.

> 
> Also, honestly:
> 
> > FLAG.store(true, Ordering::SeqCst); // or release/acquire/relaxed
> 
> I suspect in reality it would be hidden as accessor functions, or
> people just continue to write things in C.
> 
> Yes, I know all about the C++ memory ordering. It's not only a
> standards mess, it's all very illegible code too.

It's more explicit, and that's probably not a bad thing - compare it to
our smp_mb__after_atomic(), it's not uncommon to find code where the
barriers are missing because the person who wrote the code was assuming
x86.