DPDK patches and discussions
 help / color / mirror / Atom feed
* rte_lcore_count() semantics
@ 2022-10-12  6:03 Mattias Rönnblom
  0 siblings, 0 replies; only message in thread
From: Mattias Rönnblom @ 2022-10-12  6:03 UTC (permalink / raw)
  To: dev

Hi.

According to the API documentation, rte_lcore_count() sets out to 
"Return the number of execution units (lcores) on the system"

The EAL documentation states that "the term 'lcore' refers to an EAL 
thread", and also refers to regular pthreads having called 
rte_thread_register() as "registered non-EAL threads".

This might lead the unwary to believe that rte_lcore_count() did not 
include registered non-EAL threads (i.e., normal pthreads that called 
rte_thread_register()). But it does.

Registered non-EAL threads are excluded in RTE_LCORE_FOREACH() loops, as 
expected. However, the FOREACH macro documentation says "Macro to browse 
all running lcores.". Should all interpret all this as a registered 
non-EAL thread is a non-running lcore? But a lcore is just another name 
for an EAL thread (at least that's how I interpret "refers to"), and 
that would lead the non-EAL thread to be an EAL thread. Beyond the 
obvious contradiction, the non-EAL thread is not a EAL thread, since you 
can't launch tasks on it.

Something is not right here.

Any thoughts?

Thanks,
	Mattias

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2022-10-12  6:03 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-10-12  6:03 rte_lcore_count() semantics Mattias Rönnblom

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).