DPDK patches and discussions
 help / color / mirror / Atom feed
From: Olivier Matz <olivier.matz@6wind.com>
To: David Marchand <david.marchand@redhat.com>
Cc: Yongseok Koh <yskoh@mellanox.com>,
	shahafs@mellanox.com, dev@dpdk.org, stable@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] net/mlx5: fix instruction hotspot on replenishing Rx buffer
Date: Wed, 9 Jan 2019 10:52:05 +0100	[thread overview]
Message-ID: <20190109095205.us53xaocvokx4jog@glumotte.dev.6wind.com> (raw)
In-Reply-To: <CAJFAV8x+WBbei_HwQeGMdb+YZw8YSEWgP4uVp0Nihewc=myEaw@mail.gmail.com>

On Wed, Jan 09, 2019 at 10:38:07AM +0100, David Marchand wrote:
> On Wed, Jan 9, 2019 at 9:54 AM Yongseok Koh <yskoh@mellanox.com> wrote:
> 
> > On replenishing Rx buffers for vectorized Rx, mbuf->buf_addr isn't needed
> > to be accessed as it is static and easily calculated from the mbuf address.
> > Accessing the mbuf content causes unnecessary load stall and it is worsened
> > on ARM.
> >
> > Fixes: 545b884b1da3 ("net/mlx5: fix buffer address posting in SSE Rx")
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
> > ---
> >  drivers/net/mlx5/mlx5_rxtx_vec.h | 8 ++++++--
> >  1 file changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/mlx5/mlx5_rxtx_vec.h
> > b/drivers/net/mlx5/mlx5_rxtx_vec.h
> > index fda7004e2d..ced5547307 100644
> > --- a/drivers/net/mlx5/mlx5_rxtx_vec.h
> > +++ b/drivers/net/mlx5/mlx5_rxtx_vec.h
> > @@ -102,8 +102,12 @@ mlx5_rx_replenish_bulk_mbuf(struct mlx5_rxq_data
> > *rxq, uint16_t n)
> >                 return;
> >         }
> >         for (i = 0; i < n; ++i) {
> > -               wq[i].addr = rte_cpu_to_be_64((uintptr_t)elts[i]->buf_addr
> > +
> > -                                             RTE_PKTMBUF_HEADROOM);
> > +               uintptr_t buf_addr =
> > +                       (uintptr_t)elts[i] + sizeof(struct rte_mbuf) +
> > +                       rte_pktmbuf_priv_size(rxq->mp) +
> > RTE_PKTMBUF_HEADROOM;
> > +
> > +               assert(buf_addr == (uintptr_t)elts[i]->buf_addr);
> > +               wq[i].addr = rte_cpu_to_be_64(buf_addr);
> >                 /* If there's only one MR, no need to replace LKey in WQE.
> > */
> >                 if (unlikely(mlx5_mr_btree_len(&rxq->mr_ctrl.cache_bh) >
> > 1))
> >                         wq[i].lkey = mlx5_rx_mb2mr(rxq, elts[i]);
> > --
> > 2.11.0
> >
> >
> How about having a macro / inline in the mbuf api to get this information
> in a consistent/unique way ?
> I can see we have this calculation at least in rte_pktmbuf_init() and
> rte_pktmbuf_detach().

Agree. Maybe rte_mbuf_default_buf_addr(m) ?

Side note, is the assert() correct in the patch? I'd say there's a
difference of RTE_PKTMBUF_HEADROOM between the 2 values.

Olivier

  reply	other threads:[~2019-01-09  9:52 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-01-09  8:54 Yongseok Koh
2019-01-09  9:38 ` David Marchand
2019-01-09  9:52   ` Olivier Matz [this message]
2019-01-09  9:56     ` Yongseok Koh
2019-01-09 10:05       ` David Marchand
2019-01-09 10:11         ` Yongseok Koh
2019-01-09 13:19 ` [dpdk-dev] [PATCH v2 1/2] mbuf: add function returning default buffer address Yongseok Koh
2019-01-09 13:19   ` [dpdk-dev] [PATCH v2 2/2] net/mlx5: fix instruction hotspot on replenishing Rx buffer Yongseok Koh
2019-01-09 13:46   ` [dpdk-dev] [PATCH v2 1/2] mbuf: add function returning default buffer address David Marchand
2019-01-10  1:39     ` Rami Rosen
2019-01-10 18:18       ` Yongseok Koh
2019-01-10 18:22     ` Yongseok Koh
2019-01-10 18:35 ` [dpdk-dev] [PATCH v3 " Yongseok Koh
2019-01-10 18:35   ` [dpdk-dev] [PATCH v3 2/2] net/mlx5: fix instruction hotspot on replenishing Rx buffer Yongseok Koh
2019-01-10 19:10     ` Shahaf Shuler
2019-01-11  8:14   ` [dpdk-dev] [PATCH v3 1/2] mbuf: add function returning default buffer address Andrew Rybchenko
2019-01-11 11:03     ` Yongseok Koh
2019-01-11 11:17       ` Andrew Rybchenko
2019-01-11 11:37         ` Yongseok Koh
2019-01-11 11:57         ` Bruce Richardson
2019-01-11 12:48           ` David Marchand
2019-01-14 15:51             ` Olivier Matz
2019-01-10 22:40 ` [dpdk-dev] [PATCH v4 " Yongseok Koh
2019-01-10 22:40   ` [dpdk-dev] [PATCH v4 2/2] net/mlx5: fix instruction hotspot on replenishing Rx buffer Yongseok Koh
2019-01-11  8:05   ` [dpdk-dev] [PATCH v4 1/2] mbuf: add function returning default buffer address Olivier Matz
2019-01-11  8:11   ` David Marchand
2019-01-11  8:32     ` David Marchand
2019-01-11 11:09       ` Yongseok Koh
2019-01-11 10:25     ` Thomas Monjalon
2019-01-14 21:16 ` [dpdk-dev] [PATCH v5 1/2] mbuf: add function returning " Yongseok Koh
2019-01-14 21:16   ` [dpdk-dev] [PATCH v5 2/2] net/mlx5: fix instruction hotspot on replenishing Rx buffer Yongseok Koh
2019-02-06 15:54     ` Kevin Traynor
2019-02-21 19:10       ` Kevin Traynor
2019-03-08  2:05         ` Yongseok Koh
2019-01-15  1:35   ` [dpdk-dev] [PATCH v5 1/2] mbuf: add function returning buffer address Thomas Monjalon

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190109095205.us53xaocvokx4jog@glumotte.dev.6wind.com \
    --to=olivier.matz@6wind.com \
    --cc=david.marchand@redhat.com \
    --cc=dev@dpdk.org \
    --cc=shahafs@mellanox.com \
    --cc=stable@dpdk.org \
    --cc=yskoh@mellanox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).