linux-kernel - Re: [PATCH 1/2] rust: Add initial cpumask abstractions

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: <20250225114830.64p56bndn4hkkkgq@vireshk-i7>
Date: Tue, 25 Feb 2025 17:18:30 +0530
From: Viresh Kumar <viresh.kumar@...aro.org>
To: Alice Ryhl <aliceryhl@...gle.com>
Cc: Yury Norov <yury.norov@...il.com>,
	Rasmus Villemoes <linux@...musvillemoes.dk>,
	Miguel Ojeda <ojeda@...nel.org>,
	Alex Gaynor <alex.gaynor@...il.com>,
	Boqun Feng <boqun.feng@...il.com>, Gary Guo <gary@...yguo.net>,
	Björn Roy Baron <bjorn3_gh@...tonmail.com>,
	Benno Lossin <benno.lossin@...ton.me>,
	Andreas Hindborg <a.hindborg@...nel.org>,
	Trevor Gross <tmgross@...ch.edu>, linux-kernel@...r.kernel.org,
	Danilo Krummrich <dakr@...hat.com>, rust-for-linux@...r.kernel.org,
	Vincent Guittot <vincent.guittot@...aro.org>
Subject: Re: [PATCH 1/2] rust: Add initial cpumask abstractions

On 25-02-25, 10:55, Alice Ryhl wrote:
> On Tue, Feb 25, 2025 at 10:47 AM Viresh Kumar <viresh.kumar@...aro.org> wrote:
> > +    /// Creates cpumask.
> > +    #[cfg(not(CONFIG_CPUMASK_OFFSTACK))]
> > +    fn new_inner(empty: bool) -> Result<Self> {
> > +        let ptr = KBox::into_raw(KBox::new([bindings::cpumask::default(); 1], GFP_KERNEL)?);
> 
> I don't really understand this CPUMASK_OFFSTACK logic. You seem to
> always allocate memory, but if OFFSTACK=n, then shouldn't it be on the
> stack ...?

IIUC, the idea of the config option is to prevent stack overflow on
systems with high number of CPUs (> 256), in which case the memory for
the masks is allocated dynamically. Otherwise a local variable, in a
function or a struct (which may itself get allocated dynamically) is
fine.

In the CONFIG_CPUMASK_OFFSTACK=y case, the cpumask C core does the
allocation and the Rust code doesn't need to take care of the same.

In the CONFIG_CPUMASK_OFFSTACK=n case, the allocation must be done by
the caller (on stack or heap) and the cpumask C core will only clear
the mask.

I tried with an on-stack variable earlier but ran into problems as the
memory is shared with the C FFI and Rust moves it unless it is pinned.

One easy way to make it work was using Box, which keeps the code
simple.

Should I do it differently ?

-- 
viresh