[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aNqlPzZn9jUjSKfE@yury>
Date: Mon, 29 Sep 2025 11:26:55 -0400
From: Yury Norov <yury.norov@...il.com>
To: Miguel Ojeda <miguel.ojeda.sandonis@...il.com>
Cc: Joel Fernandes <joelagnelf@...dia.com>, linux-kernel@...r.kernel.org,
rust-for-linux@...r.kernel.org, dri-devel@...ts.freedesktop.org,
dakr@...nel.org, acourbot@...dia.com,
Alistair Popple <apopple@...dia.com>,
Miguel Ojeda <ojeda@...nel.org>,
Alex Gaynor <alex.gaynor@...il.com>,
Boqun Feng <boqun.feng@...il.com>, Gary Guo <gary@...yguo.net>,
bjorn3_gh@...tonmail.com, Benno Lossin <lossin@...nel.org>,
Andreas Hindborg <a.hindborg@...nel.org>,
Alice Ryhl <aliceryhl@...gle.com>, Trevor Gross <tmgross@...ch.edu>,
David Airlie <airlied@...il.com>, Simona Vetter <simona@...ll.ch>,
Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
Maxime Ripard <mripard@...nel.org>,
Thomas Zimmermann <tzimmermann@...e.de>,
John Hubbard <jhubbard@...dia.com>, Timur Tabi <ttabi@...dia.com>,
joel@...lfernandes.org, Elle Rhumsaa <elle@...thered-steel.dev>,
Daniel Almeida <daniel.almeida@...labora.com>,
nouveau@...ts.freedesktop.org
Subject: Re: [PATCH v4 6/6] rust: bitfield: Use 'as' operator for setter type
conversion
On Mon, Sep 29, 2025 at 03:59:32PM +0200, Miguel Ojeda wrote:
> On Sat, Sep 20, 2025 at 8:23 PM Joel Fernandes <joelagnelf@...dia.com> wrote:
> >
> > The bitfield macro's setter currently uses the From trait for type
> > conversion, which is overly restrictive and prevents use cases such as
> > narrowing conversions (e.g., u32 storage size to u8 field size) which
> > aren't supported by From.
>
> Being restrictive is a good thing -- it would be nice to know more
> context about this change, like Alexandre points out.
>
> In particular, the line:
>
> .set_nibble(0x12345678_u32) // truncated to 0x8
>
> sounds fairly alarming, and not what we usually want. Why cannot the
> caller cast on their side, if they really want that?
It was my suggestion to relax the type requirement. The reasoning is
as follows.
Consider a bitfield bf with bits 5:3 described as field1. The storage
for bf is u8, but the type is u32. This is OK, because storage and
representation are simply different matters. And no matter how you
declare the field inside the bitfield, you can't prevent overflow
followed by silent truncation by just syntax measures.
I suggested to relax the requirement that field representation must
match (not exceed in fact) storage type, and instead bring explicit
check in the setter. With the check, if user tries to overflow the
field, we either throw a warning, or panic if hardening is enabled,
or do nothing in performance-critical builds.
As far as I can say, Joel scheduled this in v5.
Thanks,
Yury
Powered by blists - more mailing lists