On Thu, 26 Sep 2019 15:46:52 +0300 Andrew Rybchenko <arybche...@solarflare.com> wrote:
> On 9/19/19 4:17 PM, Stephen Hemminger wrote: > > Avoid duplication by having generic basic xstats available > > for use by drivers. A later patch uses this for failsafe > > driver. > > > > Signed-off-by: Stephen Hemminger <step...@networkplumber.org> > > Acked-by: Gaetan Rivet <gaetan.ri...@6wind.com> > > Acked-by: Andrew Rybchenko <arybche...@solarflare.com> > > > diff --git a/lib/librte_ethdev/rte_ethdev_driver.h > > b/lib/librte_ethdev/rte_ethdev_driver.h > > index 936ff8c98651..489889a72203 100644 > > --- a/lib/librte_ethdev/rte_ethdev_driver.h > > +++ b/lib/librte_ethdev/rte_ethdev_driver.h > > @@ -208,6 +208,71 @@ rte_eth_linkstatus_get(const struct rte_eth_dev *dev, > > #endif > > } > > > > +/** > > + * @internal > > + * Get basic stats part of xstats for an ethernet device. > > + * > > + * @param dev > > + * Pointer to struct rte_eth_dev. > > + */ > > +__rte_experimental > > Does it make sense to mark internal API as experimental? > I thought that @internal is not a part of API/ABI. Checkpatch picks on it
