DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] ixgbe: use rte_mbuf_prefetch_part2 for cacheline1 access
@ 2016-06-17 14:06 Jerin Jacob
  2016-06-20  3:19 ` Jianbo Liu
  0 siblings, 1 reply; 3+ messages in thread
From: Jerin Jacob @ 2016-06-17 14:06 UTC (permalink / raw)
  To: dev
  Cc: thomas.monjalon, bruce.richardson, jianbo.liu, helin.zhang,
	konstantin.ananyev, Jerin Jacob

made second cache line access behavior same as IA

Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
---
 drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
index 9c1d124..64a329e 100644
--- a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
+++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
@@ -280,10 +280,10 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
 		vst1q_u64((uint64_t *)&rx_pkts[pos + 2], mbp2);
 
 		if (split_packet) {
-			rte_prefetch_non_temporal(&rx_pkts[pos]->cacheline1);
-			rte_prefetch_non_temporal(&rx_pkts[pos + 1]->cacheline1);
-			rte_prefetch_non_temporal(&rx_pkts[pos + 2]->cacheline1);
-			rte_prefetch_non_temporal(&rx_pkts[pos + 3]->cacheline1);
+			rte_mbuf_prefetch_part2(rx_pkts[pos]);
+			rte_mbuf_prefetch_part2(rx_pkts[pos + 1]);
+			rte_mbuf_prefetch_part2(rx_pkts[pos + 2]);
+			rte_mbuf_prefetch_part2(rx_pkts[pos + 3]);
 		}
 
 		/* D.1 pkt 3,4 convert format from desc to pktmbuf */
-- 
2.5.5

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [dpdk-dev] [PATCH] ixgbe: use rte_mbuf_prefetch_part2 for cacheline1 access
  2016-06-17 14:06 [dpdk-dev] [PATCH] ixgbe: use rte_mbuf_prefetch_part2 for cacheline1 access Jerin Jacob
@ 2016-06-20  3:19 ` Jianbo Liu
  2016-06-24 11:04   ` Bruce Richardson
  0 siblings, 1 reply; 3+ messages in thread
From: Jianbo Liu @ 2016-06-20  3:19 UTC (permalink / raw)
  To: Jerin Jacob
  Cc: dev, Thomas Monjalon, Bruce Richardson, Zhang, Helin, Ananyev,
	Konstantin

On 17 June 2016 at 22:06, Jerin Jacob <jerin.jacob@caviumnetworks.com> wrote:
> made second cache line access behavior same as IA
>
> Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
> ---
>  drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
> index 9c1d124..64a329e 100644
> --- a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
> +++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
> @@ -280,10 +280,10 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
>                 vst1q_u64((uint64_t *)&rx_pkts[pos + 2], mbp2);
>
>                 if (split_packet) {
> -                       rte_prefetch_non_temporal(&rx_pkts[pos]->cacheline1);
> -                       rte_prefetch_non_temporal(&rx_pkts[pos + 1]->cacheline1);
> -                       rte_prefetch_non_temporal(&rx_pkts[pos + 2]->cacheline1);
> -                       rte_prefetch_non_temporal(&rx_pkts[pos + 3]->cacheline1);
> +                       rte_mbuf_prefetch_part2(rx_pkts[pos]);
> +                       rte_mbuf_prefetch_part2(rx_pkts[pos + 1]);
> +                       rte_mbuf_prefetch_part2(rx_pkts[pos + 2]);
> +                       rte_mbuf_prefetch_part2(rx_pkts[pos + 3]);
>                 }
>
>                 /* D.1 pkt 3,4 convert format from desc to pktmbuf */
> --
> 2.5.5
>

Reviewed-by: Jianbo Liu <jianbo.liu@linaro.org>

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [dpdk-dev] [PATCH] ixgbe: use rte_mbuf_prefetch_part2 for cacheline1 access
  2016-06-20  3:19 ` Jianbo Liu
@ 2016-06-24 11:04   ` Bruce Richardson
  0 siblings, 0 replies; 3+ messages in thread
From: Bruce Richardson @ 2016-06-24 11:04 UTC (permalink / raw)
  To: Jianbo Liu
  Cc: Jerin Jacob, dev, Thomas Monjalon, Zhang, Helin, Ananyev, Konstantin

On Mon, Jun 20, 2016 at 11:19:19AM +0800, Jianbo Liu wrote:
> On 17 June 2016 at 22:06, Jerin Jacob <jerin.jacob@caviumnetworks.com> wrote:
> > made second cache line access behavior same as IA
> >
> > Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
> > ---
> >  drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
> > index 9c1d124..64a329e 100644
> > --- a/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
> > +++ b/drivers/net/ixgbe/ixgbe_rxtx_vec_neon.c
> > @@ -280,10 +280,10 @@ _recv_raw_pkts_vec(struct ixgbe_rx_queue *rxq, struct rte_mbuf **rx_pkts,
> >                 vst1q_u64((uint64_t *)&rx_pkts[pos + 2], mbp2);
> >
> >                 if (split_packet) {
> > -                       rte_prefetch_non_temporal(&rx_pkts[pos]->cacheline1);
> > -                       rte_prefetch_non_temporal(&rx_pkts[pos + 1]->cacheline1);
> > -                       rte_prefetch_non_temporal(&rx_pkts[pos + 2]->cacheline1);
> > -                       rte_prefetch_non_temporal(&rx_pkts[pos + 3]->cacheline1);
> > +                       rte_mbuf_prefetch_part2(rx_pkts[pos]);
> > +                       rte_mbuf_prefetch_part2(rx_pkts[pos + 1]);
> > +                       rte_mbuf_prefetch_part2(rx_pkts[pos + 2]);
> > +                       rte_mbuf_prefetch_part2(rx_pkts[pos + 3]);
> >                 }
> >
> >                 /* D.1 pkt 3,4 convert format from desc to pktmbuf */
> > --
> > 2.5.5
> >
> 
> Reviewed-by: Jianbo Liu <jianbo.liu@linaro.org>

Applied to dpdk-next-net/rel_16_07

/Bruce

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2016-06-24 11:04 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-06-17 14:06 [dpdk-dev] [PATCH] ixgbe: use rte_mbuf_prefetch_part2 for cacheline1 access Jerin Jacob
2016-06-20  3:19 ` Jianbo Liu
2016-06-24 11:04   ` Bruce Richardson

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).