> > > Added cache guard after the table holding the ring elements, to avoid > > > false sharing conflicts caused by next-line hardware prefetchers when > > > accessing elements at the end of the ring table. > > > > I don't see any harm with it, and in theory it might help in some > > cases... > > Though I wonder how real is that problem? > > Did you ever observe such contention to happen? > > I never observed a problem with this. > The risk of contention depends on what is allocated in the memory after the > ring. > Which is application specific. > > It seems like a purely theoretical issue, but should be fixed anyway, to > eliminate > that risk.
Ok, as I said I see no harm with it. Should we document this change somewhere? RN or PG? Acked-by: Konstantin Ananyev <[email protected]> > > > > > Signed-off-by: Morten Brørup <[email protected]> > > > --- > > > lib/ring/rte_ring.c | 3 +++ > > > 1 file changed, 3 insertions(+) > > > > > > diff --git a/lib/ring/rte_ring.c b/lib/ring/rte_ring.c > > > index f10050a1c4..9ccc62cd42 100644 > > > --- a/lib/ring/rte_ring.c > > > +++ b/lib/ring/rte_ring.c > > > @@ -73,8 +73,11 @@ rte_ring_get_memsize_elem(unsigned int esize, > > unsigned > > > int count) > > > return -EINVAL; > > > } > > > > > > + static_assert(sizeof(struct rte_ring) == > > > RTE_CACHE_LINE_ROUNDUP(sizeof(struct rte_ring)), > > > + "Size of struct rte_ring not cache line aligned"); > > > sz = sizeof(struct rte_ring) + (ssize_t)count * esize; > > > sz = RTE_ALIGN(sz, RTE_CACHE_LINE_SIZE); > > > + sz += RTE_CACHE_GUARD_LINES * RTE_CACHE_LINE_SIZE; > > > return sz; > > > } > > > > > > -- > > > 2.43.0
