Re: [PATCH] rust: irq: add &Device<Bound> argument to irq callbacks

[Daniel Almeida <daniel.almeida@collabora.com>]

Alice,

> On 21 Jul 2025, at 11:38, Alice Ryhl <aliceryhl@google.com> wrote:
> 
> When working with a bus device, many operations are only possible while
> the device is still bound. The &Device<Bound> type represents a proof in
> the type system that you are in a scope where the device is guaranteed
> to still be bound. Since we deregister irq callbacks when unbinding a
> device, if an irq callback is running, that implies that the device has
> not yet been unbound.
> 
> To allow drivers to take advantage of that, add an additional argument
> to irq callbacks.
> 
> Signed-off-by: Alice Ryhl <aliceryhl@google.com>
> ---
> This patch is a follow-up to Daniel's irq series [1] that adds a
> &Device<Bound> argument to all irq callbacks. This allows you to use
> operations that are only safe on a bound device inside an irq callback.
> 
> The patch is otherwise based on top of driver-core-next.
> 
> [1]: https://lore.kernel.org/r/20250715-topics-tyr-request_irq2-v7-0-d469c0f37c07@collabora.com

I am having a hard time applying this locally.

> ///
> /// This function should be only used as the callback in `request_irq`.
> unsafe extern "C" fn handle_irq_callback<T: Handler>(_irq: i32, ptr: *mut c_void) -> c_uint {
> -    // SAFETY: `ptr` is a pointer to T set in `Registration::new`
> -    let handler = unsafe { &*(ptr as *const T) };
> -    T::handle(handler) as c_uint
> +    // SAFETY: `ptr` is a pointer to `Registration<T>` set in `Registration::new`
> +    let registration = unsafe { &*(ptr as *const Registration<T>) };
> +    // SAFETY: The irq callback is removed before the device is unbound, so the fact that the irq
> +    // callback is running implies that the device has not yet been unbound.
> +    let device = unsafe { registration.inner.device().as_bound() };

Where was this function introduced? i.e. I am missing the change that brought
in RegistrationInner::device(), or maybe some Deref impl that would make this
possible?

Also, I wonder if we can't make the scope of this unsafe block smaller?

— Daniel