[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20230705184836.363c8c8b.gary@garyguo.net>
Date: Wed, 5 Jul 2023 18:48:36 +0100
From: Gary Guo <gary@...yguo.net>
To: Boqun Feng <boqun.feng@...il.com>
Cc: Benno Lossin <benno.lossin@...ton.me>,
Miguel Ojeda <ojeda@...nel.org>,
Wedson Almeida Filho <wedsonaf@...il.com>,
Alex Gaynor <alex.gaynor@...il.com>,
Björn Roy Baron <bjorn3_gh@...tonmail.com>,
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 Mon, 3 Jul 2023 11:15:55 -0700
Boqun Feng <boqun.feng@...il.com> wrote:
> On Sat, Jun 24, 2023 at 09:25:19AM +0000, Benno Lossin wrote:
> [...]
> (this is `init_slot`)
> > @@ -1064,7 +1152,7 @@ macro_rules! __init_internal {
> > @data($data:ident),
> > @slot($slot:ident),
> > @guards($($guards:ident,)*),
> > - @munch_fields($(,)?),
> > + @munch_fields($(..Zeroable::zeroed())? $(,)?),
>
> since you append an unconditional comma ',' to init_slot and
> make_initializer when "calling" them in with_update_parsed, shouldn't
> this be:
>
> + @munch_fields($(..Zeroable::zeroed(),)? $(,)?),
>
> , and..
>
> > ) => {
> > // Endpoint of munching, no fields are left. If execution reaches this point, all fields
> > // have been initialized. Therefore we can now dismiss the guards by forgetting them.
> > @@ -1157,6 +1245,30 @@ macro_rules! __init_internal {
> > @munch_fields($($rest)*),
> > );
> > };
> > + (make_initializer:
> > + @slot($slot:ident),
> > + @type_name($t:ident),
> > + @munch_fields(..Zeroable::zeroed() $(,)?),
>
> this should be:
>
> + @munch_fields(..Zeroable::zeroed() , $(,)?),
>
> Otherwise the example before `pin_init!()` wouldn't compile:
>
> /// pin_init!(Buf {
> /// buf: [1; 64],
> /// ..Zeroable::zeroed(),
> /// });
Comma is not allowed after base struct.
>
> Regards,
> Boqun
>
> > + @acc($($acc:tt)*),
> > + ) => {
> > + // Endpoint, nothing more to munch, create the initializer. Since the users specified
> > + // `..Zeroable::zeroed()`, the slot will already have been zeroed and all field that have
> > + // not been overwritten are thus zero and initialized. We still check that all fields are
> > + // actually accessible by using the struct update syntax ourselves.
> > + // Since we are in the `if false` branch, this will never get executed. We abuse `slot` to
> > + // get the correct type inference here:
> > + unsafe {
> > + let mut zeroed = ::core::mem::zeroed();
> > + // We have to use type inference her to make zeroed have the correct type. This does
> > + // not get executed, so it has no effect.
> > + ::core::ptr::write($slot, zeroed);
> > + zeroed = ::core::mem::zeroed();
> > + ::core::ptr::write($slot, $t {
> > + $($acc)*
> > + ..zeroed
> > + });
> > + }
> > + };
> > (make_initializer:
> > @slot($slot:ident),
> > @type_name($t:ident),
> > --
> > 2.41.0
> >
> >
Powered by blists - more mailing lists