Re: [PATCH 1/2] docs/core-api: memory-allocation: add k[mz]alloc_obj() and clarify kmalloc

[SJ Park]

On Mon, 31 Aug 2026 18:13:01 +0300 "Mike Rapoport (Microsoft)" <rppt@xxxxxxxxxx> wrote:

> Since v7.0 the most used memory allocation function is kzalloc_obj().
> 
> Update hte memory-allocation guide to describe k[mz]alloc_obj() family

s/hte/the/

> and make kzalloc_obj() the first answer to "How should I allocate
> memory?" question.
> 
> While on it, clarify description of kmalloc() size limitations.

Looks good improvements to me.

> 
> Signed-off-by: Mike Rapoport (Microsoft) <rppt@xxxxxxxxxx>

Acked-by: SJ Park <sj@xxxxxxxxxx>

> ---
>  Documentation/core-api/memory-allocation.rst | 30 +++++++++++++++++++++-------
>  1 file changed, 23 insertions(+), 7 deletions(-)
> 
> diff --git a/Documentation/core-api/memory-allocation.rst b/Documentation/core-api/memory-allocation.rst
> index 0f19dd5243239..8ecfa4d35f4f5 100644
> --- a/Documentation/core-api/memory-allocation.rst
> +++ b/Documentation/core-api/memory-allocation.rst
> @@ -19,6 +19,12 @@ Diversity of the allocation APIs combined with the numerous GFP flags
>  makes the question "How should I allocate memory?" not that easy to
>  answer, although very likely you should use
>  
> +::
> +
> +  kzalloc_obj(<VAR_OR_TYPE>);
> +
> +or
> +
>  ::
>  
>    kzalloc(<size>, GFP_KERNEL);
> @@ -139,10 +145,13 @@ allocate memory for an array, there are kmalloc_array() and kcalloc()
>  helpers. The helpers struct_size(), array_size() and array3_size() can
>  be used to safely calculate object sizes without overflowing.
>  
> -The maximal size of a chunk that can be allocated with `kmalloc` is
> -limited. The actual limit depends on the hardware and the kernel
> -configuration, but it is a good practice to use `kmalloc` for objects
> -smaller than page size.
> +Since 7.0 there are type aware kmalloc-family helpers that let you safely and
> +conveniently allocate a single object or arrays of objects with kzalloc_obj()
> +and kmalloc_obj() and their array versions kzalloc_objs() and
> +kmalloc_objs(). These helpers only need the type of the object that should be
> +allocated and the count of elements in the array for the array versions.
> +
> +As of v7.2, vast majority of the memory allocations use kzalloc_obj().

As the commit message says "Since v7.0 the most used memory allocation function
is kzalloc_obj()", should we s/7.2/7.0/ ?

I also feel like the above sentence might not really needed, or can be
version-neutral, e.g., Since its introduction, vast majoritiy of ... switched
to use ... ?  No strong opinion.

>  
>  The address of a chunk allocated with `kmalloc` is aligned to at least
>  ARCH_KMALLOC_MINALIGN bytes. For sizes which are a power of two, the
> @@ -154,9 +163,16 @@ Chunks allocated with kmalloc() can be resized with krealloc(). Similarly
>  to kmalloc_array(): a helper for resizing arrays is provided in the form of
>  krealloc_array().
>  
> -For large allocations you can use vmalloc() and vzalloc(), or directly
> -request pages from the page allocator. The memory allocated by `vmalloc`
> -and related functions is not physically contiguous.
> +`kmalloc` always allocates physically contiguous memory and the maximal size of
> +a chunk that can be allocated with `kmalloc` is limited by `MAX_ORDER`, the
> +same limit also applies to the page allocator.

I initially thought if it makes look more consistent by using kmalloc() instead
of `kmalloc`.  But seems anyway we are mixing those here and there.  So I have
no opinion.

> +
> +Internally, slab allocator differentiates allocations of different orders and
> +delegates larger allocations to the page allocator, but for the users of
> +`kmalloc` family it is entirely transparent.
> +
> +For large allocations that do not require physically contiguous memory you can
> +use vmalloc() and vzalloc() family.
>  
>  If you are not sure whether the allocation size is too large for
>  `kmalloc`, it is possible to use kvmalloc() and its derivatives. It will
> 
> -- 
> 2.53.0


Thanks,
SJ