Re: [PATCH] mm: (k)vrealloc: document concurrency restrictions

[Vlastimil Babka (SUSE)]

On 7/25/24 2:54 PM, Danilo Krummrich wrote:
> Document that concurrent calls to vrealloc() with itself or vfree() for
> the same memory allocation are not permitted, since it leads to undefined
> behavior.
> 
> Consequently, the same applies for kvrealloc(), hence also document it
> for kvrealloc().
> 
> Suggested-by: Christoph Hellwig <hch@xxxxxxxxxxxxx>
> Suggested-by: Uladzislau Rezki <urezki@xxxxxxxxx>
> Signed-off-by: Danilo Krummrich <dakr@xxxxxxxxxx>

Huh, did anyone ever assume otherwise and why?

> ---
>  mm/util.c    | 3 +++
>  mm/vmalloc.c | 3 +++
>  2 files changed, 6 insertions(+)
> 
> diff --git a/mm/util.c b/mm/util.c
> index 0ff5898cc6de..29ae93f6344f 100644
> --- a/mm/util.c
> +++ b/mm/util.c
> @@ -732,6 +732,9 @@ EXPORT_SYMBOL(kvfree_sensitive);
>   * If @p is %NULL, kvrealloc() behaves exactly like kvmalloc(). If @size is 0
>   * and @p is not a %NULL pointer, the object pointed to is freed.
>   *
> + * This function must not be called concurrently with itself or kvfree() for the
> + * same memory allocation.
> + *
>   * Return: pointer to the allocated memory or %NULL in case of error
>   */
>  void *kvrealloc_noprof(const void *p, size_t size, gfp_t flags)
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index caf032f0bd69..2a6d4ce57b73 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -4049,6 +4049,9 @@ EXPORT_SYMBOL(vzalloc_node_noprof);
>   * If @p is %NULL, vrealloc() behaves exactly like vmalloc(). If @size is 0 and
>   * @p is not a %NULL pointer, the object pointed to is freed.
>   *
> + * This function must not be called concurrently with itself or vfree() for the
> + * same memory allocation.
> + *
>   * Return: pointer to the allocated memory; %NULL if @size is zero or in case of
>   *         failure
>   */
> 
> base-commit: d270beaca6818349b2aed7e6034b800a777087cc