Re: [PATCH v6] rust: alloc: satisfy POSIX alignment requirement

[Gary Guo <gary@garyguo.net>]

On Thu, 13 Feb 2025 06:26:41 -0500
Tamir Duberstein <tamird@gmail.com> wrote:

> ISO C's `aligned_alloc` is partially implementation-defined; on some
> systems it inherits stricter requirements from POSIX's `posix_memalign`.
> 
> This causes the call added in commit dd09538fb409 ("rust: alloc:
> implement `Cmalloc` in module allocator_test") to fail on macOS because
> it doesn't meet the requirements of `posix_memalign`.
> 
> Adjust the call to meet the POSIX requirement and add a comment. This
> fixes failures in `make rusttest` on macOS.
> 
> Acked-by: Danilo Krummrich <dakr@kernel.org>
> Fixes: dd09538fb409 ("rust: alloc: implement `Cmalloc` in module allocator_test")
> Signed-off-by: Tamir Duberstein <tamird@gmail.com>
> ---
> Changes in v6:
> - Replace unsound use of build_error with map_err. (Danilo Krummrich)

It's sound, just not correct.

> - Link to v5: https://lore.kernel.org/r/20250212-aligned-alloc-v5-1-c51e0b17dee9@gmail.com
> 
> Changes in v5:
> - Remove errant newline in commit message. (Miguel Ojeda)
> - Use more succinct expression. (Gary Guo)
> - Drop and then add Danilo's Acked-by again.
> - Link to v4: https://lore.kernel.org/r/20250210-aligned-alloc-v4-1-609c3a6fe139@gmail.com
> 
> Changes in v4:
> - Revert to `aligned_alloc` and correct rationale. (Miguel Ojeda)
> - Apply Danilo's Acked-by from v2.
> - Rebase on v6.14-rc2.
> - Link to v3: https://lore.kernel.org/r/20250206-aligned-alloc-v3-1-0cbc0ab0306d@gmail.com
> 
> Changes in v3:
> - Replace `aligned_alloc` with `posix_memalign` for portability.
> - Link to v2: https://lore.kernel.org/r/20250202-aligned-alloc-v2-1-5af0b5fdd46f@gmail.com
> 
> Changes in v2:
> - Shorten some variable names. (Danilo Krummrich)
> - Replace shadowing alignment variable with a second call to
>   Layout::align. (Danilo Krummrich)
> - Link to v1: https://lore.kernel.org/r/20250201-aligned-alloc-v1-1-c99a73f3cbd4@gmail.com
> ---
>  rust/kernel/alloc/allocator_test.rs | 18 ++++++++++++++++++
>  1 file changed, 18 insertions(+)
> 
> diff --git a/rust/kernel/alloc/allocator_test.rs b/rust/kernel/alloc/allocator_test.rs
> index e3240d16040b..e68775078e90 100644
> --- a/rust/kernel/alloc/allocator_test.rs
> +++ b/rust/kernel/alloc/allocator_test.rs
> @@ -62,6 +62,24 @@ unsafe fn realloc(
>              ));
>          }
>  
> +        // ISO C (ISO/IEC 9899:2011) defines `aligned_alloc`:
> +        //
> +        // > The value of alignment shall be a valid alignment supported by the implementation
> +        // [...].
> +        //
> +        // As an example of the "supported by the implementation" requirement, POSIX.1-2001 (IEEE
> +        // 1003.1-2001) defines `posix_memalign`:
> +        //
> +        // > The value of alignment shall be a power of two multiple of sizeof (void *).
> +        //
> +        // and POSIX-based implementations of `aligned_alloc` inherit this requirement. At the time
> +        // of writing, this is known to be the case on macOS (but not in glibc).
> +        //
> +        // Satisfy the stricter requirement to avoid spurious test failures on some platforms.
> +        let min_align = core::mem::size_of::<*const crate::ffi::c_void>();
> +        let layout = layout.align_to(min_align).map_err(|_| AllocError)?.pad_to_align();
> +        let layout = layout.pad_to_align();

You're doing two `pad_to_align`s.

Best,
Gary

> +
>          // SAFETY: Returns either NULL or a pointer to a memory allocation that satisfies or
>          // exceeds the given size and alignment requirements.
>          let dst = unsafe { libc_aligned_alloc(layout.align(), layout.size()) } as *mut u8;
> 
> ---
> base-commit: 8a5aae7dbbfb612509c8a2f112f7e0f79029ed45
> change-id: 20250201-aligned-alloc-b52cb2353c82
> 
> Best regards,