Re: [PATCH RESEND v4 3/4] rust: pci: add config space read/write support

From: Danilo Krummrich
Date: Tue Nov 04 2025 - 10:37:39 EST


On Tue Nov 4, 2025 at 3:27 PM CET, Zhi Wang wrote:
> Drivers might need to access PCI config space for querying capability
> structures and access the registers inside the structures.
>
> For Rust drivers need to access PCI config space, the Rust PCI abstraction
> needs to support it in a way that upholds Rust's safety principles.
>
> Introduce a `ConfigSpace` wrapper in Rust PCI abstraction to provide safe
> accessors for PCI config space. The new type implements the `Io` trait to
> share offset validation and bound-checking logic with others.
>
> Cc: Danilo Krummrich <dakr@xxxxxxxxxx>
> Signed-off-by: Zhi Wang <zhiw@xxxxxxxxxx>
> ---
> rust/kernel/pci.rs | 148 +++++++++++++++++++++++++++++++++++++++++++--
> 1 file changed, 142 insertions(+), 6 deletions(-)
>
> diff --git a/rust/kernel/pci.rs b/rust/kernel/pci.rs
> index 77a8eb39ad32..630660a7f05d 100644
> --- a/rust/kernel/pci.rs
> +++ b/rust/kernel/pci.rs
> @@ -5,13 +5,31 @@
> //! C header: [`include/linux/pci.h`](srctree/include/linux/pci.h)
>
> use crate::{
> - bindings, container_of, device,
> - device_id::{RawDeviceId, RawDeviceIdIndex},
> + bindings,
> + container_of,
> + device,
> + device_id::{
> + RawDeviceId,
> + RawDeviceIdIndex
> + },
> devres::Devres,
> driver,
> - error::{from_result, to_result, Result},
> - io::{Mmio, MmioRaw},
> - irq::{self, IrqRequest},
> + error::{
> + from_result,
> + to_result,
> + Result
> + },
> + io::{
> + define_read,
> + define_write,
> + Io,
> + Mmio,
> + MmioRaw
> + },
> + irq::{
> + self,
> + IrqRequest
> + },
> str::CStr,
> sync::aref::ARef,
> types::Opaque,
> @@ -20,7 +38,10 @@
> use core::{
> marker::PhantomData,
> ops::Deref,
> - ptr::{addr_of_mut, NonNull},
> + ptr::{
> + addr_of_mut,
> + NonNull
> + },
> };

Please add a comma and use `//` at the end of the last element of a block as
described in [1]. Otherwise rustfmt will change it up again.

I think most of the diff is not related to this patch. I think it would be good
to convert the whole module and then add further changes.

I just did such a cleanup for the I/O module [1], I can send another patch for
all the PCI stuff.

[1] https://docs.kernel.org/rust/coding-guidelines.html#imports
[2] https://lore.kernel.org/lkml/20251104133301.59402-1-dakr@xxxxxxxxxx/

> use kernel::prelude::*;
>
> @@ -305,6 +326,83 @@ pub struct Device<Ctx: device::DeviceContext = device::Normal>(
> PhantomData<Ctx>,
> );
>
> +/// Represents the PCI configuration space of a device.
> +///
> +/// Provides typed read and write accessors for configuration registers
> +/// using the standard `pci_read_config_*` and `pci_write_config_*` helpers.
> +///
> +/// The generic const parameter `SIZE` can be used to indicate the
> +/// maximum size of the configuration space (e.g. 256 bytes for legacy,
> +/// 4096 bytes for extended config space). The actual size is obtained
> +/// from the underlying `struct pci_dev` via [`Device::cfg_size`].
> +pub struct ConfigSpace<'a, const SIZE: usize = { ConfigSpaceSize::Extended as usize }> {
> + pdev: &'a Device<device::Core>,
> +}
> +
> +macro_rules! call_config_read {
> + (infallible, $c_fn:ident, $self:ident, $ty:ty, $addr:expr) => {{
> + let mut val: $ty = 0;
> + let ret = unsafe { bindings::$c_fn($self.pdev.as_raw(), $addr as i32, &mut val) };
> + match ret {
> + 0 => val,
> + _ => !0,
> + }

I think the match is not needed, the value should already be set accordingly.

> + }};
> +/// Represents the size of a PCI configuration space.
> +///
> +/// PCI devices can have either a *normal* (legacy) configuration space of 256 bytes,
> +/// or an *extended* configuration space of 4096 bytes as defined in the PCI Express
> +/// specification.
> +pub enum ConfigSpaceSize {

I think this should be [repr(usize)].

> + /// 256-byte legacy PCI configuration space.
> + Normal = 256,
> +
> + /// 4096-byte PCIe extended configuration space.
> + Extended = 4096,
> +}
> +
> impl Device {
> /// Returns the PCI vendor ID as [`Vendor`].
> ///
> @@ -514,6 +625,17 @@ pub fn pci_class(&self) -> Class {
> // SAFETY: `self.as_raw` is a valid pointer to a `struct pci_dev`.
> Class::from_raw(unsafe { (*self.as_raw()).class })
> }
> +
> + /// Returns the size of configuration space.
> + fn cfg_size(&self) -> Result<ConfigSpaceSize> {
> + // SAFETY: `self.as_raw` is a valid pointer to a `struct pci_dev`.
> + let size = unsafe { (*self.as_raw()).cfg_size };
> + match size {
> + 256 => Ok(ConfigSpaceSize::Normal),
> + 4096 => Ok(ConfigSpaceSize::Extended),
> + _ => Err(EINVAL),

I'd add a debug_assert!() in this case.

> + }
> + }
> }
>
> impl Device<device::Bound> {
> @@ -591,6 +713,20 @@ pub fn set_master(&self) {
> // SAFETY: `self.as_raw` is guaranteed to be a pointer to a valid `struct pci_dev`.
> unsafe { bindings::pci_set_master(self.as_raw()) };
> }
> +
> + /// Return an initialized config space object.
> + pub fn config_space<'a>(
> + &'a self,
> + ) -> Result<ConfigSpace<'a, { ConfigSpaceSize::Normal as usize }>> {

Let's add as_raw() to ConfigSpaceSize and use that instead.

> + Ok(ConfigSpace { pdev: self })
> + }
> +
> + /// Return an initialized config space object.
> + pub fn config_space_exteneded<'a>(
> + &'a self,
> + ) -> Result<ConfigSpace<'a, { ConfigSpaceSize::Extended as usize }>> {
> + Ok(ConfigSpace { pdev: self })
> + }
> }
>
> // SAFETY: `Device` is a transparent wrapper of a type that doesn't depend on `Device`'s generic
> --
> 2.51.0