Re: [dpdk-dev] [PATCH v3 1/2] mbuf: provide rte_pktmbuf_alloc_bulk API

[Stephen Hemminger <stephen@networkplumber.org>]

On Wed, 23 Dec 2015 00:17:53 +0800
Huawei Xie <huawei.xie@intel.com> wrote:

> +
> +	rc = rte_mempool_get_bulk(pool, (void **)mbufs, count);
> +	if (unlikely(rc))
> +		return rc;
> +
> +	switch (count % 4) {
> +	case 0: while (idx != count) {
> +			RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
> +			rte_mbuf_refcnt_set(mbufs[idx], 1);
> +			rte_pktmbuf_reset(mbufs[idx]);
> +			idx++;
> +	case 3:
> +			RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
> +			rte_mbuf_refcnt_set(mbufs[idx], 1);
> +			rte_pktmbuf_reset(mbufs[idx]);
> +			idx++;
> +	case 2:
> +			RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
> +			rte_mbuf_refcnt_set(mbufs[idx], 1);
> +			rte_pktmbuf_reset(mbufs[idx]);
> +			idx++;
> +	case 1:
> +			RTE_MBUF_ASSERT(rte_mbuf_refcnt_read(mbufs[idx]) == 0);
> +			rte_mbuf_refcnt_set(mbufs[idx], 1);
> +			rte_pktmbuf_reset(mbufs[idx]);
> +			idx++;
> +	}
> +	}
> +	return 0;
> +}

Since function will not work if count can not be 0 (otherwise rte_mempool_get_bulk will fail),
why not:
	1. Document that assumption
	2. Use that assumption to speed up code.



	switch(count % 4) {
		do {
			case 0:
			...
			case 1:
			...
		} while (idx != count);
	}

Also you really need to add a big block comment about this loop, to explain
what it does and why.