Re: [PATCH net-next 3/3] net: ethernet: ti: cpsw: add XDP support

[Jesper Dangaard Brouer <brouer@redhat.com>]

On Thu, 23 May 2019 21:20:35 +0300
Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org> wrote:

> +static struct page *cpsw_alloc_page(struct cpsw_common *cpsw)
> +{
> +	struct page_pool *pool = cpsw->rx_page_pool;
> +	struct page *page, *prev_page = NULL;
> +	int try = pool->p.pool_size << 2;
> +	int start_free = 0, ret;
> +
> +	do {
> +		page = page_pool_dev_alloc_pages(pool);
> +		if (!page)
> +			return NULL;
> +
> +		/* if netstack has page_pool recycling remove the rest */
> +		if (page_ref_count(page) == 1)
> +			break;
> +
> +		/* start free pages in use, shouldn't happen */
> +		if (prev_page == page || start_free) {
> +			/* dma unmap/puts page if rfcnt != 1 */
> +			page_pool_recycle_direct(pool, page);
> +			start_free = 1;
> +			continue;
> +		}
> +
> +		/* if refcnt > 1, page has been holding by netstack, it's pity,
> +		 * so put it to the ring to be consumed later when fast cash is
> +		 * empty. If ring is full then free page by recycling as above.
> +		 */
> +		ret = ptr_ring_produce(&pool->ring, page);

This looks very wrong to me!  First of all you are manipulation
directly with the internal pool->ring and not using the API, which
makes this code un-maintainable.  Second this is wrong, as page_pool
assume the in-variance that pages on the ring have refcnt==1.

> +		if (ret) {
> +			page_pool_recycle_direct(pool, page);
> +			continue;
> +		}
> +
> +		if (!prev_page)
> +			prev_page = page;
> +	} while (try--);
> +
> +	return page;
> +}


-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer