Re: [PATCH v2 5/8] rust: pci: add num_vf(), to return number of VFs

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] rust: pci: missing unconditional availability for num_vf() causing driver build failures
- [Low] rust: pci: missing #[inline] on num_vf() FFI wrapper
--

commit 887b28180af5e15626ffb6304f6b61300c0d3b74
Author: Peter Colberg <pcolberg@redhat.com>

rust: pci: add num_vf(), to return number of VFs

Add a method to return the number of Virtual Functions (VF) enabled for
a Physical Function (PF).

> diff --git a/rust/kernel/pci.rs b/rust/kernel/pci.rs
> index ca009775bd612..db0554f14afc8 100644
> --- a/rust/kernel/pci.rs
> +++ b/rust/kernel/pci.rs
> @@ -449,6 +449,14 @@ pub(crate) fn is_virtfn(&self) -> bool {
>          unsafe { (*self.as_raw()).is_virtfn() != 0 }
>      }
>  
> +    /// Returns the number of Virtual Functions (VF) enabled for a Physical Function (PF).
> +    #[cfg(CONFIG_PCI_IOV)]

[Severity: Medium]
Will this cause build failures for drivers that unconditionally call
num_vf() when CONFIG_PCI_IOV is disabled?

In C, pci_num_vf() is unconditionally available as a static inline that
returns 0 when IOV is disabled, so drivers can safely call it without
checking Kconfig symbols. Adding #[cfg(CONFIG_PCI_IOV)] here hides the
method entirely in non-IOV builds, which will result in compilation errors
for downstream callers.

> +    #[expect(dead_code)]
> +    pub(crate) fn num_vf(&self) -> i32 {

[Severity: Low]
Should this method have the #[inline] attribute?

The rust subsystem guidelines recommend adding #[inline] to abstraction
methods that forward directly to a C binding.

> +        // SAFETY: `self.as_raw` is a valid pointer to a `struct pci_dev`.
> +        unsafe { bindings::pci_num_vf(self.as_raw()) }
> +    }
> +
>      /// Returns the size of the given PCI BAR resource.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260924190556.1620886-1-zhiw@nvidia.com?part=5