[PATCH v4 01/20] rust: dma: expose the count and size of CoherentAllocation

[Danilo Krummrich]

On Thu, May 22, 2025 at 02:29:35PM +0900, Alexandre Courbot wrote:
> On Wed May 21, 2025 at 9:43 PM JST, Boqun Feng wrote:
> > On Wed, May 21, 2025 at 03:44:56PM +0900, Alexandre Courbot wrote:
> >> These properties are very useful to have and should be accessible.
> >> 
> >> Signed-off-by: Alexandre Courbot <acourbot at nvidia.com>
> >> ---
> >>  rust/kernel/dma.rs | 18 ++++++++++++++++++
> >>  1 file changed, 18 insertions(+)
> >> 
> >> diff --git a/rust/kernel/dma.rs b/rust/kernel/dma.rs
> >> index 605e01e35715667f93297fd9ec49d8e7032e0910..2a60eefa47dfc1f836c30ee342e26c6ff3e9b13a 100644
> >> --- a/rust/kernel/dma.rs
> >> +++ b/rust/kernel/dma.rs
> >> @@ -129,6 +129,10 @@ pub mod attrs {
> >>  //
> >>  // Hence, find a way to revoke the device resources of a `CoherentAllocation`, but not the
> >>  // entire `CoherentAllocation` including the allocated memory itself.
> >> +//
> >> +// # Invariants
> >> +//
> >> +// The size in bytes of the allocation is equal to `size_of::<T> * count()`.

As you propose below.

"... and does fit into a `usize`"

You then need an '// INVARIANT' comment where we actually ensure this, i.e. in
Self::alloc_attrs().

> >>  pub struct CoherentAllocation<T: AsBytes + FromBytes> {
> >>      dev: ARef<Device>,
> >>      dma_handle: bindings::dma_addr_t,
> >> @@ -201,6 +205,20 @@ pub fn alloc_coherent(
> >>          CoherentAllocation::alloc_attrs(dev, count, gfp_flags, Attrs(0))
> >>      }
> >>  
> >> +    /// Returns the number of elements `T` in this allocation.
> >> +    ///
> >> +    /// Note that this is not the size of the allocation in bytes, which is provided by
> >> +    /// [`Self::size`].
> >> +    pub fn count(&self) -> usize {
> >> +        self.count
> >> +    }
> >> +
> >> +    /// Returns the size in bytes of this allocation.
> >> +    pub fn size(&self) -> usize {
> >> +        // As per the invariants of `CoherentAllocation`.

"The type invariant of `Self` guarantees that [...]" or similar.

> >> +        self.count * core::mem::size_of::<T>()
> >
> > I think we need a comment or even an invariant saying this multiply
> > cannot overflow.
> 
> I'll add the following invariant:
> 
>     `size_of::<T> * count` fits into a `usize`.
> 
> and refer to it here. Does it work?