Thread (19 messages) 19 messages, 4 authors, 2017-06-30

Service lcores and Application lcores

From: Van Haaren, Harry <hidden>
Date: 2017-06-29 14:36:35

Hi All,


The recently posted service cores patchset[1], introduces service lcores to run services for DPDK applications. Services are just an ordinary function for eg: eventdev scheduling, NIC RX, statistics and monitoring, etc. A service is just a callback function, which a core invokes. An atomic ensures that services that are
non-multi-thread-safe are never concurrently invoked.

The topic of discussion in this thread is how we can ensure that application lcores do not interfere with service cores. I have a solution described below, opinions welcome.


Regards, -Harry


PS: This discussion extends that in the ML thread here[2], participants of that thread added to CC.

[1] Service Cores v2 patchset http://dpdk.org/dev/patchwork/bundle/hvanhaar/service_cores_v2/
[2] http://dpdk.org/ml/archives/dev/2017-June/069290.html


________________________



A proposal for Eventdev, to ensure Service lcores and Application lcores play nice;

1) Application lcores must not directly call rte_eventdev_schedule()
2A) Service cores are the proper method to run services
2B) If an application insists on running a service "manually" on an app lcore, we provide a function for that:
     rte_service_run_from_app_lcore(struct service *srv);

The above function would allow a pesky app to run services on its own (non-service core) lcores, but
does so through the service-core framework, allowing the service-library atomic to keep access serialized as required for non-multi-thread-safe services.

The above solution maintains the option of running the eventdev PMD as now (single-core dedicated to a single service), while providing correct serialization by using the rte_service_run_from_app_lcore() function. Given the atomic is only used when required (multiple cores mapped to the service) there should be no performance delta.

Given that the application should not invoke rte_eventdev_schedule(), we could even consider removing it from the Eventdev API. A PMD that requires cycles registers a service, and an application can use a service core or the run_from_app_lcore() function if it wishes to invoke that service on an application owned lcore.


Opinions?
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help