Re: [PATCH v4 1/6] eventdev: introduce event driven programming model

[Nipun Gupta <nipun.gupta@nxp.com>]

Hi,

I had a few queries/comments regarding the eventdev patches.

Please see inline.

> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Jerin Jacob
> Sent: Wednesday, December 21, 2016 14:55
> To: dev@dpdk.org
> Cc: thomas.monjalon@6wind.com; bruce.richardson@intel.com; Hemant
> Agrawal <hemant.agrawal@nxp.com>; gage.eads@intel.com;
> harry.van.haaren@intel.com; Jerin Jacob <jerin.jacob@caviumnetworks.com>	
> Subject: [dpdk-dev] [PATCH v4 1/6] eventdev: introduce event driven
> programming model
> 
> In a polling model, lcores poll ethdev ports and associated
> rx queues directly to look for packet. In an event driven model,
> by contrast, lcores call the scheduler that selects packets for
> them based on programmer-specified criteria. Eventdev library
> adds support for event driven programming model, which offer
> applications automatic multicore scaling, dynamic load balancing,
> pipelining, packet ingress order maintenance and
> synchronization services to simplify application packet processing.
> 
> By introducing event driven programming model, DPDK can support
> both polling and event driven programming models for packet processing,
> and applications are free to choose whatever model
> (or combination of the two) that best suits their needs.
> 
> This patch adds the eventdev specification header file.
> 
> Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
> Acked-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
>  MAINTAINERS                        |    3 +
>  doc/api/doxy-api-index.md          |    1 +
>  doc/api/doxy-api.conf              |    1 +
>  lib/librte_eventdev/rte_eventdev.h | 1275
> ++++++++++++++++++++++++++++++++++++
>  4 files changed, 1280 insertions(+)
>  create mode 100644 lib/librte_eventdev/rte_eventdev.h

<snip>

> +
> +/**
> + * Event device information
> + */
> +struct rte_event_dev_info {
> +	const char *driver_name;	/**< Event driver name */
> +	struct rte_pci_device *pci_dev;	/**< PCI information */

With 'rte_device' in place (rte_dev.h), should we not have 'rte_device' instead of 'rte_pci_device' here?

> +	uint32_t min_dequeue_timeout_ns;
> +	/**< Minimum supported global dequeue timeout(ns) by this device */
> +	uint32_t max_dequeue_timeout_ns;
> +	/**< Maximum supported global dequeue timeout(ns) by this device */
> +	uint32_t dequeue_timeout_ns;
> +	/**< Configured global dequeue timeout(ns) for this device */
> +	uint8_t max_event_queues;
> +	/**< Maximum event_queues supported by this device */
> +	uint32_t max_event_queue_flows;
> +	/**< Maximum supported flows in an event queue by this device*/
> +	uint8_t max_event_queue_priority_levels;
> +	/**< Maximum number of event queue priority levels by this device.
> +	 * Valid when the device has RTE_EVENT_DEV_CAP_QUEUE_QOS
> capability
> +	 */

<snip>

> +/**
> + * Dequeue a burst of events objects or an event object from the event port
> + * designated by its *event_port_id*, on an event device designated
> + * by its *dev_id*.
> + *
> + * rte_event_dequeue_burst() does not dictate the specifics of scheduling
> + * algorithm as each eventdev driver may have different criteria to schedule
> + * an event. However, in general, from an application perspective scheduler
> may
> + * use the following scheme to dispatch an event to the port.
> + *
> + * 1) Selection of event queue based on
> + *   a) The list of event queues are linked to the event port.
> + *   b) If the device has RTE_EVENT_DEV_CAP_QUEUE_QOS capability then
> event
> + *   queue selection from list is based on event queue priority relative to
> + *   other event queue supplied as *priority* in rte_event_queue_setup()
> + *   c) If the device has RTE_EVENT_DEV_CAP_EVENT_QOS capability then
> event
> + *   queue selection from the list is based on event priority supplied as
> + *   *priority* in rte_event_enqueue_burst()
> + * 2) Selection of event
> + *   a) The number of flows available in selected event queue.
> + *   b) Schedule type method associated with the event
> + *
> + * The *nb_events* parameter is the maximum number of event objects to
> dequeue
> + * which are returned in the *ev* array of *rte_event* structure.
> + *
> + * The rte_event_dequeue_burst() function returns the number of events
> objects
> + * it actually dequeued. A return value equal to *nb_events* means that all
> + * event objects have been dequeued.
> + *
> + * The number of events dequeued is the number of scheduler contexts held by
> + * this port. These contexts are automatically released in the next
> + * rte_event_dequeue_burst() invocation, or invoking
> rte_event_enqueue_burst()
> + * with RTE_EVENT_OP_RELEASE operation can be used to release the
> + * contexts early.
> + *
> + * @param dev_id
> + *   The identifier of the device.
> + * @param port_id
> + *   The identifier of the event port.
> + * @param[out] ev
> + *   Points to an array of *nb_events* objects of type *rte_event* structure
> + *   for output to be populated with the dequeued event objects.
> + * @param nb_events
> + *   The maximum number of event objects to dequeue, typically number of
> + *   rte_event_port_dequeue_depth() available for this port.
> + *
> + * @param timeout_ticks
> + *   - 0 no-wait, returns immediately if there is no event.
> + *   - >0 wait for the event, if the device is configured with
> + *   RTE_EVENT_DEV_CFG_PER_DEQUEUE_TIMEOUT then this function will
> wait until
> + *   the event available or *timeout_ticks* time.

Just for understanding - Is expectation that rte_event_dequeue_burst() will wait till timeout
unless requested number of events (nb_events) are not received on the event port?

> + *   if the device is not configured with
> RTE_EVENT_DEV_CFG_PER_DEQUEUE_TIMEOUT
> + *   then this function will wait until the event available or
> + *   *dequeue_timeout_ns* ns which was previously supplied to
> + *   rte_event_dev_configure()
> + *
> + * @return
> + * The number of event objects actually dequeued from the port. The return
> + * value can be less than the value of the *nb_events* parameter when the
> + * event port's queue is not full.
> + *
> + * @see rte_event_port_dequeue_depth()
> + */
> +uint16_t
> +rte_event_dequeue_burst(uint8_t dev_id, uint8_t port_id, struct rte_event
> ev[],
> +			uint16_t nb_events, uint64_t timeout_ticks);
> +

<Snip>

Regards,
Nipun