[dpdk-dev] [PATCH 3/6] service cores: EAL init changes
Van Haaren, Harry
harry.van.haaren at intel.com
Thu Jun 29 13:13:59 CEST 2017
> -----Original Message-----
> From: Jerin Jacob [mailto:jerin.jacob at caviumnetworks.com]
<snip>
> > This commit shows the changes required in rte_eal_init()
> > to transparently launch the service threads. The threads
> > are launched into the service worker functions here because
> > after rte_eal_init() the application is not gauranteed to
> > call any other DPDK API.
> >
> > As the registration of services happens at initialization
> > time, the services that require CPU time are already available
> > when we reach the end of rte_eal_init().
> >
> > Signed-off-by: Harry van Haaren <harry.van.haaren at intel.com>
> > ---
> > lib/librte_eal/linuxapp/eal/eal.c | 15 +++++++++++++++
> > 1 file changed, 15 insertions(+)
> >
> > diff --git a/lib/librte_eal/linuxapp/eal/eal.c b/lib/librte_eal/linuxapp/eal/eal.c
>
> Need to add bsdapp implementation for the same.
Done.
> > + /* initialize service core threads and default service-core mapping */
> > + struct rte_config *config = rte_eal_get_configuration();
> > + uint32_t service_cores[RTE_MAX_LCORE];
> > + int count = rte_service_core_list(service_cores, RTE_MAX_LCORE);
> > + for (i = 0; i < count; i++) {
> > + config->lcore_role[service_cores[i]] = ROLE_SERVICE;
>
> Can we move this change inside rte_service_core_start() itself?
> otherwise every application needs to the same. Right?
Done, nice, this removes the requirement to access rte_config, which I didn't like having there.
More information about the dev
mailing list