linux-kernel - Re: [PATCH 1/3] rust: pci: implement TryInto<IrqRequest<'a>> for IrqVector<'a>

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: <DDK49THLLA3Y.242R8EP6IDZJ3@kernel.org>
Date: Fri, 17 Oct 2025 00:57:42 +0200
From: "Danilo Krummrich" <dakr@...nel.org>
To: "Joel Fernandes" <joelagnelf@...dia.com>
Cc: <bhelgaas@...gle.com>, <kwilczynski@...nel.org>, <ojeda@...nel.org>,
 <alex.gaynor@...il.com>, <boqun.feng@...il.com>, <gary@...yguo.net>,
 <bjorn3_gh@...tonmail.com>, <lossin@...nel.org>, <a.hindborg@...nel.org>,
 <aliceryhl@...gle.com>, <tmgross@...ch.edu>,
 <rust-for-linux@...r.kernel.org>, <linux-pci@...r.kernel.org>,
 <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/3] rust: pci: implement TryInto<IrqRequest<'a>> for
 IrqVector<'a>

On Fri Oct 17, 2025 at 12:24 AM CEST, Joel Fernandes wrote:
> On Wed, Oct 15, 2025 at 08:14:29PM +0200, Danilo Krummrich wrote:
>> Implement TryInto<IrqRequest<'a>> for IrqVector<'a> to directly convert
>> a pci::IrqVector into a generic IrqRequest, instead of taking the
>> indirection via an unrelated pci::Device method.
>> 
>> Signed-off-by: Danilo Krummrich <dakr@...nel.org>
>> ---
>>  rust/kernel/pci.rs | 38 ++++++++++++++++++--------------------
>>  1 file changed, 18 insertions(+), 20 deletions(-)
>> 
>> diff --git a/rust/kernel/pci.rs b/rust/kernel/pci.rs
>> index d91ec9f008ae..c6b750047b2e 100644
>> --- a/rust/kernel/pci.rs
>> +++ b/rust/kernel/pci.rs
>> @@ -596,6 +596,20 @@ fn index(&self) -> u32 {
>>      }
>>  }
>>  
>> +impl<'a> TryInto<IrqRequest<'a>> for IrqVector<'a> {
>> +    type Error = Error;
>> +
>> +    fn try_into(self) -> Result<IrqRequest<'a>> {
>> +        // SAFETY: `self.as_raw` returns a valid pointer to a `struct pci_dev`.
>> +        let irq = unsafe { bindings::pci_irq_vector(self.dev.as_raw(), self.index()) };
>> +        if irq < 0 {
>> +            return Err(crate::error::Error::from_errno(irq));
>> +        }
>> +        // SAFETY: `irq` is guaranteed to be a valid IRQ number for `&self`.
>> +        Ok(unsafe { IrqRequest::new(self.dev.as_ref(), irq as u32) })
>> +    }
>> +}A
>
>
> Nice change, looks good to me but I do feel it is odd to 'convert' an
> IrqVector directly into a IrqRequest using TryInto (one is a device-relative
> vector index and the other holds the notion of an IRQ request).
>
> Instead, we should convert IrqVector into something like LinuxIrqNumber
> (using TryInto) because we're converting one number to another, and then pass
> that to a separate function to create the IrqRequest.

Well, IrqRequest is exactly that, a representation of an IRQ number. So, this is
already doing exactly that, converting one number to another:

	pub struct IrqRequest<'a> {
	    dev: &'a Device<Bound>,
	    irq: u32,
	}

(The reason this is called IrqRequest instead of IrqNumber is that the number is
an irrelevant implementation detail of how an IRQ is requested.)

	pub struct IrqVector<'a> {
	    dev: &'a Device<Bound>,
	    index: u32,
	}

So, what happens here is that we convert the vector index from IrqVector into
the irq number in IrqRequest.

> Or we can do both in a
> vector.make_request() function (which is basically this patch but not using
> TryInto).

See above, there is no "both", it's the same thing. :)

Regarding make_request() vs. TryInto, I think TryInto is the idiomatic thing to
do here: Both structures have the same layout, as in they both carry the
&Device<Bound> reference the corresponding number belongs to, plus the number
itself; the device reference is taken over, the number is converted.

> Actually even my original code had this oddity:
> The function irq_vector should have been called irq_request or something but
> instead was:
> pub fn irq_vector(&self, vector: IrqVector<'_>) -> Result<IrqRequest<'_>>
>
> I think we can incrementally improve this though, so LGTM.
>
> Reviewed-by: Joel Fernandes <joelagnelf@...dia.com>