[dpdk-dev] [RFC 4/4] mempool: prevent objects from being across pages

[Burakov, Anatoly]

On 19-Jul-19 2:38 PM, Olivier Matz wrote:
> When using iova contiguous memory and objets smaller than page size,
> ensure that objects are not located across several pages.
> 
> Signed-off-by: Vamsi Krishna Attunuru <vattunuru at marvell.com>
> Signed-off-by: Olivier Matz <olivier.matz at 6wind.com>
> ---
>   lib/librte_mempool/rte_mempool_ops_default.c | 39 ++++++++++++++++++++++++++--
>   1 file changed, 37 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/librte_mempool/rte_mempool_ops_default.c b/lib/librte_mempool/rte_mempool_ops_default.c
> index 4e2bfc82d..2bbd67367 100644
> --- a/lib/librte_mempool/rte_mempool_ops_default.c
> +++ b/lib/librte_mempool/rte_mempool_ops_default.c
> @@ -45,19 +45,54 @@ rte_mempool_op_calc_mem_size_default(const struct rte_mempool *mp,
>   	return mem_size;
>   }
>   
> +/* Returns -1 if object falls on a page boundary, else returns 0 */
> +static inline int
> +mempool_check_obj_bounds(void *obj, uint64_t pg_sz, size_t elt_sz)
> +{
> +	uintptr_t page_end, elt_addr = (uintptr_t)obj;
> +	uint32_t pg_shift;
> +	uint64_t page_mask;
> +
> +	if (pg_sz == 0)
> +		return 0;
> +	if (elt_sz > pg_sz)
> +		return 0;
> +
> +	pg_shift = rte_bsf32(pg_sz);
> +	page_mask =  ~((1ull << pg_shift) - 1);
> +	page_end = (elt_addr & page_mask) + pg_sz;

This looks like RTE_PTR_ALIGN should do this without the magic? E.g.

page_end = RTE_PTR_ALIGN(elt_addr, pg_sz)

would that not be equivalent?

-- 
Thanks,
Anatoly