Re: [PATCH v5 00/20] rust: I/O type generalization and projection

[Gary Guo]

On Mon Jun 29, 2026 at 12:37 AM BST, Danilo Krummrich wrote:
> On Fri Jun 26, 2026 at 4:45 PM CEST, Gary Guo wrote:
>> This series presents a major rework of I/O types, as a summary:
>
> There are two minor rustdoc warnings:
>
> 	warning: unresolved link to `include/linux/iosys-map.h`
> 	    --> rust/kernel/io.rs:1456:7
> 	     |
> 	1456 | /// [`include/linux/iosys-map.h`] in C.
> 	     |       ^^^^^^^^^^^^^^^^^^^^^^^^^ no item named `include/linux/iosys-map.h` in scope
> 	     |
> 	     = help: to escape `[` and `]` characters, add '\' before them like `\[` or `\]`
> 	     = note: `#[warn(rustdoc::broken_intra_doc_links)]` on by default

Hmm, this looks like a false positive introduced in new version of rustdoc.
I didn't get this with rustdoc 1.85, but do get it in latest rustc, however the
generated link is correct despite the warning.

This pattern is used by other code, I'm unsure why it doesn't trigger there.
Anyhow, the following diff seems to be able to work around the issue.

diff --git a/rust/kernel/io.rs b/rust/kernel/io.rs
index 0c5f53dd0939..5f2f3cafb0a2 100644
--- a/rust/kernel/io.rs
+++ b/rust/kernel/io.rs
@@ -1453,12 +1453,10 @@ fn as_view(self) -> <Self::Backend as IoBackend>::View<'a, Self::Target> {
 ///
 /// This can be used when a piece of logic may wish to handle both MMIO or system memory but does
 /// not want or cannot be generic over I/O backends. This serves a similar purpose to
-/// [`include/linux/iosys-map.h`] in C.
+/// [`include/linux/iosys-map.h`](srctree/include/linux/iosys-map.h) in C.
 ///
 /// This type can be used like any other types that implements [`Io`]; this also include
 /// [`io_project!`], [`io_read!`], [`io_write!`].
-///
-/// [`include/linux/iosys-map.h`](srctree/include/linux/iosys-map.h)
 pub enum IoSysMap<'a, T: ?Sized> {
     /// The view is I/O memory.
     Io(Mmio<'a, T>),

> 	
> 	warning: unresolved link to `View`
> 	    --> rust/kernel/io.rs:1620:74
> 	     |
> 	1620 | /// In addition to projecting from [`Io`], you may also project from a [`View`] of an [`Io`].
> 	     |                                                                          ^^^^ no item named `View` in scope
> 	     |
> 	     = help: to escape `[` and `]` characters, add '\' before them like `\[` or `\]`

Oops, this is a stray comment from earlier version that I didn't remove. This whole line can be
removed.

> 	
> 	warning: 2 warnings emitted
>
>> Gary Guo (19):
>>       rust: io: add dynamically-sized `Region` type
>>       rust: io: add missing safety requirement in `IoCapable` methods
>>       rust: io: restrict untyped IO access and `register!` to `Region`
>>       rust: io: implement `Io` on reference types instead
>>       rust: io: generalize `MmioRaw` to pointer to arbitrary type
>>       rust: io: rename `Mmio` to `MmioOwned`
>>       rust: io: implement `Mmio` as view type
>>       rust: pci: io: make `ConfigSpace` a view
>>       rust: io: use view types instead of addresses for `Io`
>>       pwm: th1520: remove unnecessary `deref`
>>       rust: io: remove `MmioOwned`
>>       rust: io: move `Io` methods to extension trait
>>       rust: io: add projection macro and methods
>>       rust: io: implement a view type for `Coherent`
>>       rust: io: add `read_val` and `write_val` functions on `Io`
>>       gpu: nova-core: use I/O projection for cleaner encapsulation
>>       rust: dma: drop `dma_read!` and `dma_write!` API
>>       rust: io: add copying methods
>>       rust: io: implement `IoSysMap`
>>
>> Laura Nao (1):
>>       rust: io: add I/O backend for system memory with volatile access
>
> For sharing with the drm-rust tree:
>
> diff --git a/rust/kernel/drm/gem/shmem.rs b/rust/kernel/drm/gem/shmem.rs
> index 3ee19ef6264e..6868b5d80ab7 100644
> --- a/rust/kernel/drm/gem/shmem.rs
> +++ b/rust/kernel/drm/gem/shmem.rs
> @@ -29,9 +29,10 @@
>          to_result, //
>      },
>      io::{
> -        Io,
> -        IoCapable,
> -        IoKnownSize, //
> +        IoBase,
> +        Region,
> +        SysMem,
> +        SysMemBackend, //
>      },
>      prelude::*,
>      scatterlist,
> @@ -467,6 +468,28 @@ pub fn owner(&self) -> &Object<D, C> {
>      }
>  }
>
> +impl<'a, D, R, C, const SIZE: usize> IoBase<'a> for &'a VMap<D, R, C, SIZE>
> +where
> +    D: DriverObject,
> +    C: DeviceContext,
> +    R: Deref<Target = Object<D, C>>,
> +{
> +    type Backend = SysMemBackend;
> +    type Target = Region<SIZE>;
> +
> +    #[inline]
> +    fn as_view(self) -> SysMem<'a, Region<SIZE>> {
> +        let ptr = Region::ptr_from_raw_parts_mut(self.addr.cast(), self.owner.size());
> +
> +        // SAFETY: Per type invariants of `VMap`:
> +        // - `addr .. addr + owner.size()` is a valid kernel accessible memory region.
> +        // - `addr` is page-aligned, which satisfies `Region`'s 4-byte alignment requirement.
> +        // - The memory remains valid until this `VMap` is dropped; since `self` is `&'a VMap`,
> +        //   the borrow prevents the `VMap` from being dropped for the lifetime `'a`.
> +        unsafe { SysMem::new(ptr) }
> +    }
> +}
> +
>  impl<D, R, C, const SIZE: usize> Drop for VMap<D, R, C, SIZE>
>  where
>      D: DriverObject,
> @@ -508,69 +531,6 @@ unsafe impl<D, R, C, const SIZE: usize> Sync for VMap<D, R, C, SIZE>
>  {
>  }
>
> -impl<D, R, C, const SIZE: usize> Io for VMap<D, R, C, SIZE>
> -where
> -    D: DriverObject,
> -    C: DeviceContext,
> -    R: Deref<Target = Object<D, C>>,
> -{
> -    #[inline]
> -    fn addr(&self) -> usize {
> -        self.addr as usize
> -    }
> -
> -    #[inline]
> -    fn maxsize(&self) -> usize {
> -        self.owner.size()
> -    }
> -}
> -
> -impl<D, R, C, const SIZE: usize> IoKnownSize for VMap<D, R, C, SIZE>
> -where
> -    D: DriverObject,
> -    C: DeviceContext,
> -    R: Deref<Target = Object<D, C>>,
> -{
> -    const MIN_SIZE: usize = SIZE;
> -}
> -
> -macro_rules! impl_vmap_io_capable {
> -    ($ty:ty) => {
> -        impl<D, R, C, const SIZE: usize> IoCapable<$ty> for VMap<D, R, C, SIZE>
> -        where
> -            D: DriverObject,
> -            C: DeviceContext,
> -            R: Deref<Target = Object<D, C>>,
> -        {
> -            #[inline]
> -            unsafe fn io_read(&self, address: usize) -> $ty {
> -                let ptr = address as *mut $ty;
> -
> -                // SAFETY: The safety contract of `io_read` guarantees that address is a valid
> -                // address within the bounds of `Self` of at least the size of $ty, and is properly
> -                // aligned.
> -                unsafe { ptr::read_volatile(ptr) }
> -            }
> -
> -            #[inline]
> -            unsafe fn io_write(&self, value: $ty, address: usize) {
> -                let ptr = address as *mut $ty;
> -
> -                // SAFETY: The safety contract of `io_write` guarantees that address is a valid
> -                // address within the bounds of `Self` of at least the size of $ty, and is properly
> -                // aligned.
> -                unsafe { ptr::write_volatile(ptr, value) }
> -            }
> -        }
> -    };
> -}
> -
> -impl_vmap_io_capable!(u8);
> -impl_vmap_io_capable!(u16);
> -impl_vmap_io_capable!(u32);
> -#[cfg(CONFIG_64BIT)]
> -impl_vmap_io_capable!(u64);
> -
>  /// A reference to a GEM object that is known to have a mapped [`SGTable`].
>  ///
>  /// This is used by the Rust bindings with [`Devres`] in order to ensure that mappings for SGTables
> @@ -638,6 +598,7 @@ mod tests {
>              UnregisteredDevice, //
>          },
>          faux,
> +        io::Io,
>          page::PAGE_SIZE, //
>      };
>
> @@ -714,7 +675,7 @@ fn compile_time_vmap_sizes() -> Result {
>          assert!(ptr::eq(vmap.owner(), obj.deref()));
>
>          // Verify the max size matches the actual object size

The comment can be updated as well. Other than that, it looks correct to me.

Best,
Gary

> -        assert_eq!(vmap.maxsize(), PAGE_SIZE);
> +        assert_eq!(vmap.size(), PAGE_SIZE);
>
>          // Make sure creating a vmap that's too large fails
>          assert!(obj.vmap::<{ PAGE_SIZE + 200 }>().is_err());