Message-ID: <Zyux5d33Kt6qFdaH@Boquns-Mac-mini.local>
Date: Wed, 6 Nov 2024 10:13:57 -0800
From: Boqun Feng <boqun.feng@...il.com>
To: FUJITA Tomonori <fujita.tomonori@...il.com>
Cc: anna-maria@...utronix.de, frederic@...nel.org, tglx@...utronix.de,
	jstultz@...gle.com, sboyd@...nel.org, linux-kernel@...r.kernel.org,
	netdev@...r.kernel.org, rust-for-linux@...r.kernel.org,
	andrew@...n.ch, hkallweit1@...il.com, tmgross@...ch.edu,
	ojeda@...nel.org, alex.gaynor@...il.com, gary@...yguo.net,
	bjorn3_gh@...tonmail.com, benno.lossin@...ton.me,
	a.hindborg@...sung.com, aliceryhl@...gle.com, arnd@...db.de
Subject: Re: [PATCH v5 4/7] rust: time: Add wrapper for fsleep function

On Fri, Nov 01, 2024 at 10:01:18AM +0900, FUJITA Tomonori wrote:
> Add a wrapper for fsleep, flexible sleep functions in
> `include/linux/delay.h` which typically deals with hardware delays.
> 
> The kernel supports several `sleep` functions to handle various
> lengths of delay. This adds fsleep, automatically chooses the best
> sleep method based on a duration.
> 
> `sleep` functions including `fsleep` belongs to TIMERS, not
> TIMEKEEPING. They are maintained separately. rust/kernel/time.rs is an
> abstraction for TIMEKEEPING. To make Rust abstractions match the C
> side, add rust/kernel/time/delay.rs for this wrapper.
> 
> fsleep() can only be used in a nonatomic context. This requirement is
> not checked by these abstractions, but it is intended that klint [1]
> or a similar tool will be used to check it in the future.
> 
> Link: https://rust-for-linux.com/klint [1]
> Signed-off-by: FUJITA Tomonori <fujita.tomonori@...il.com>
> ---
>  rust/helpers/helpers.c    |  1 +
>  rust/helpers/time.c       |  8 ++++++++
>  rust/kernel/time.rs       |  4 +++-
>  rust/kernel/time/delay.rs | 43 +++++++++++++++++++++++++++++++++++++++
>  4 files changed, 55 insertions(+), 1 deletion(-)
>  create mode 100644 rust/helpers/time.c
>  create mode 100644 rust/kernel/time/delay.rs
> 
> diff --git a/rust/helpers/helpers.c b/rust/helpers/helpers.c
> index 30f40149f3a9..c274546bcf78 100644
> --- a/rust/helpers/helpers.c
> +++ b/rust/helpers/helpers.c
> @@ -21,6 +21,7 @@
>  #include "slab.c"
>  #include "spinlock.c"
>  #include "task.c"
> +#include "time.c"
>  #include "uaccess.c"
>  #include "wait.c"
>  #include "workqueue.c"
> diff --git a/rust/helpers/time.c b/rust/helpers/time.c
> new file mode 100644
> index 000000000000..7ae64ad8141d
> --- /dev/null
> +++ b/rust/helpers/time.c
> @@ -0,0 +1,8 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <linux/delay.h>
> +
> +void rust_helper_fsleep(unsigned long usecs)
> +{
> +	fsleep(usecs);
> +}
> diff --git a/rust/kernel/time.rs b/rust/kernel/time.rs
> index e4f0a0f34d6d..9395739b51e0 100644
> --- a/rust/kernel/time.rs
> +++ b/rust/kernel/time.rs
> @@ -2,12 +2,14 @@
>  
>  //! Time related primitives.
>  //!
> -//! This module contains the kernel APIs related to time and timers that
> +//! This module contains the kernel APIs related to time that
>  //! have been ported or wrapped for usage by Rust code in the kernel.
>  //!
>  //! C header: [`include/linux/jiffies.h`](srctree/include/linux/jiffies.h).
>  //! C header: [`include/linux/ktime.h`](srctree/include/linux/ktime.h).
>  
> +pub mod delay;
> +
>  /// The number of nanoseconds per microsecond.
>  pub const NSEC_PER_USEC: i64 = bindings::NSEC_PER_USEC as i64;
>  
> diff --git a/rust/kernel/time/delay.rs b/rust/kernel/time/delay.rs
> new file mode 100644
> index 000000000000..c3c908b72a56
> --- /dev/null
> +++ b/rust/kernel/time/delay.rs
> @@ -0,0 +1,43 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +//! Delay and sleep primitives.
> +//!
> +//! This module contains the kernel APIs related to delay and sleep that
> +//! have been ported or wrapped for usage by Rust code in the kernel.
> +//!
> +//! C header: [`include/linux/delay.h`](srctree/include/linux/delay.h).
> +
> +use crate::time::Delta;

Nit: I think it's better to use:

use super::Delta;

here to refer a definition in the super mod.

> +use core::ffi::c_ulong;
> +
> +/// Sleeps for a given duration at least.
> +///
> +/// Equivalent to the kernel's [`fsleep`], flexible sleep function,
> +/// which automatically chooses the best sleep method based on a duration.
> +///
> +/// `delta` must be 0 or greater and no more than u32::MAX / 2 microseconds.

Adding backquotes on "u32::MAX / 2" would make it easier to read and
generates better documentation. For example.

/// `delta` must be 0 or greater and no more than `u32::MAX / 2` microseconds.


> +/// If a value outside the range is given, the function will sleep
> +/// for u32::MAX / 2 microseconds at least.

Same here.

I would also add the converted result in seconds of `u32::MAX / 2`
microseconds to give doc readers some intuitions, like:

the function will sleep for `u32::MAX / 2` (= ~2147 seconds or ~36
minutes) at least.

> +///
> +/// This function can only be used in a nonatomic context.
> +pub fn fsleep(delta: Delta) {
> +    // The argument of fsleep is an unsigned long, 32-bit on 32-bit architectures.
> +    // Considering that fsleep rounds up the duration to the nearest millisecond,
> +    // set the maximum value to u32::MAX / 2 microseconds.
> +    const MAX_DURATION: Delta = Delta::from_micros(u32::MAX as i64 >> 1);
> +
> +    let duration = if delta > MAX_DURATION || delta.as_nanos() < 0 {

I think it would be helpful if `Delta` has a `is_negative()` function.

The rest looks good to me. Thanks!

Regards,
Boqun

> +        // TODO: add WARN_ONCE() when it's supported.
> +        MAX_DURATION
> +    } else {
> +        delta
> +    };
> +
> +    // SAFETY: FFI call.
> +    unsafe {
> +        // Convert the duration to microseconds and round up to preserve
> +        // the guarantee; fsleep sleeps for at least the provided duration,
> +        // but that it may sleep for longer under some circumstances.
> +        bindings::fsleep(duration.as_micros_ceil() as c_ulong)
> +    }
> +}
> -- 
> 2.43.0
> 
> 

Powered by blists - more mailing lists