On Wed, Mar 01, 2017 at 10:47:36PM -0600, Gage Eads wrote:
> This patch initializes the links_map array entries to
> EVENT_QUEUE_SERVICE_PRIORITY_INVALID, as expected by
> rte_event_port_links_get().
>
> Signed-off-by: Gage Eads <gage.e...@intel.com>
> ---
> lib/librte_eventdev/rte_eventdev.c | 17 ++++++++++++-----
> 1 file changed, 12 insertions(+), 5 deletions(-)
>
> diff --git a/lib/librte_eventdev/rte_eventdev.c
> b/lib/librte_eventdev/rte_eventdev.c
> index 68bfc3b..b8cd92b 100644
> --- a/lib/librte_eventdev/rte_eventdev.c
> +++ b/lib/librte_eventdev/rte_eventdev.c
> @@ -190,6 +190,8 @@ rte_event_dev_queue_config(struct rte_eventdev *dev,
> uint8_t nb_queues)
> return 0;
> }
>
> +#define EVENT_QUEUE_SERVICE_PRIORITY_INVALID (0xdead)
> +
> static inline int
> rte_event_dev_port_config(struct rte_eventdev *dev, uint8_t nb_ports)
> {
> @@ -251,6 +253,9 @@ rte_event_dev_port_config(struct rte_eventdev *dev,
> uint8_t nb_ports)
> "nb_ports %u", nb_ports);
> return -(ENOMEM);
> }
> + for (i = 0; i < nb_ports * RTE_EVENT_MAX_QUEUES_PER_DEV; i++)
> + dev->data->links_map[i] =
> + EVENT_QUEUE_SERVICE_PRIORITY_INVALID;
> } else if (dev->data->ports != NULL && nb_ports != 0) {/* re-config */
> RTE_FUNC_PTR_OR_ERR_RET(*dev->dev_ops->port_release, -ENOTSUP);
>
> @@ -305,6 +310,10 @@ rte_event_dev_port_config(struct rte_eventdev *dev,
> uint8_t nb_ports)
>
> if (nb_ports > old_nb_ports) {
> uint8_t new_ps = nb_ports - old_nb_ports;
> + unsigned int old_links_map_end =
> + old_nb_ports * RTE_EVENT_MAX_QUEUES_PER_DEV;
> + unsigned int links_map_end =
> + nb_ports * RTE_EVENT_MAX_QUEUES_PER_DEV;
>
> memset(ports + old_nb_ports, 0,
> sizeof(ports[0]) * new_ps);
> @@ -312,9 +321,9 @@ rte_event_dev_port_config(struct rte_eventdev *dev,
> uint8_t nb_ports)
> sizeof(ports_dequeue_depth[0]) * new_ps);
> memset(ports_enqueue_depth + old_nb_ports, 0,
> sizeof(ports_enqueue_depth[0]) * new_ps);
> - memset(links_map +
> - (old_nb_ports * RTE_EVENT_MAX_QUEUES_PER_DEV),
> - 0, sizeof(ports_enqueue_depth[0]) * new_ps);
> + for (i = old_links_map_end; i < links_map_end; i++)
> + links_map[i] =
> + EVENT_QUEUE_SERVICE_PRIORITY_INVALID;
rte_event_port_setup() has rte_event_port_unlink() at the end of the
function. On rte_event_port_unlink, we are doing the same
operation(writing EVENT_QUEUE_SERVICE_PRIORITY_INVALID) and
rte_event_port_links_get() should be called after rte_event_dev_start(),
If so, Do you still think this duplicates writes are required? or Do you
have any other call sequence in mind?
---------------------------- Description from header file-----------------
The functions exported by the application Event API to setup a device
designated by its device identifier must be invoked in the following order:
- rte_event_dev_configure()
- rte_event_queue_setup()
- rte_event_port_setup()
- rte_event_port_link()
- rte_event_dev_start()
Then, the application can invoke, in any order, the functions
exported by the Event API to schedule events, dequeue events, enqueue events,
change event queue(s) to event port [un]link establishment and so on.
Application may use rte_event_[queue/port]_default_conf_get() to get the
default configuration to set up an event queue or event port by
overriding few default values.
If the application wants to change the configuration (i.e. call
rte_event_dev_configure(), rte_event_queue_setup(), or
rte_event_port_setup()), it must call rte_event_dev_stop() first to stop the
device and then do the reconfiguration before calling rte_event_dev_start()
again. The schedule, enqueue and dequeue functions should not be invoked
when the device is stopped.
-----------------------------------------------------------------------------
> }
>
> dev->data->ports = ports;
> @@ -815,8 +824,6 @@ rte_event_port_link(uint8_t dev_id, uint8_t port_id,
> return diag;
> }
>
> -#define EVENT_QUEUE_SERVICE_PRIORITY_INVALID (0xdead)
> -
> int
> rte_event_port_unlink(uint8_t dev_id, uint8_t port_id,
> uint8_t queues[], uint16_t nb_unlinks)
> --
> 2.7.4
>
