On Tue, 22 Jul 2014 15:47:30 +0800 Ouyang Changchun <changchun.ouyang at intel.com> wrote:
> + /**< If rx_enable_queue is true, rte_eth_dev_rx_queue_start must be > + invocated after rte_eth_dev_start's invocation to start RX for > + one queue, and rte_eth_dev_rx_queue_start instead of > + rte_eth_dev_start is responsible for allocating mbuf from > + mempool and setup the DMA physical address. It is useful in > + such scenario: buffer address is not available at the point of > + rte_eth_dev_start's invocation but available later, e.g. in > + VHOST zero copy case, the buffer address to be setup DMA > + address is available only after one VM startup. */ Good documentation about semantics is really valuable as long as it kept up to date. In this case, the comment stands out which is not a good thing. An explanation this long belongs in documentation not in code. Alternatively, if it is this hard to explain maybe it isn't the right design :-)
