From: Stephen Hemminger <stephen@networkplumber.org>
To: "Morten Brørup" <mb@smartsharesystems.com>
Cc: "Bruce Richardson" <bruce.richardson@intel.com>,
"Jerin Jacob" <jerinjacobk@gmail.com>,
"Thomas Monjalon" <thomas@monjalon.net>,
"dpdk-dev" <dev@dpdk.org>,
"David Marchand" <david.marchand@redhat.com>,
"Ferruh Yigit" <ferruh.yigit@intel.com>,
"Olivier Matz" <olivier.matz@6wind.com>,
"Ananyev, Konstantin" <konstantin.ananyev@intel.com>,
"Andrew Rybchenko" <andrew.rybchenko@oktetlabs.ru>,
"Viacheslav Ovsiienko" <viacheslavo@nvidia.com>,
"Ajit Khaparde" <ajit.khaparde@broadcom.com>,
"Jerin Jacob" <jerinj@marvell.com>,
"Hemant Agrawal" <hemant.agrawal@nxp.com>,
"Ray Kinsella" <mdr@ashroe.eu>,
"Neil Horman" <nhorman@tuxdriver.com>,
"Nithin Dabilpuram" <ndabilpuram@marvell.com>,
"Kiran Kumar K" <kirankumark@marvell.com>, <techboard@dpdk.org>
Subject: Re: [dpdk-dev] [dpdk-techboard] [PATCH 1/1] mbuf: move pool pointer in first half
Date: Mon, 9 Nov 2020 10:04:30 -0800 [thread overview]
Message-ID: <20201109100430.5edb39f3@hermes.local> (raw)
In-Reply-To: <98CBD80474FA8B44BF855DF32C47DC35C61407@smartserver.smartshare.dk>
On Mon, 9 Nov 2020 11:21:02 +0100
Morten Brørup <mb@smartsharesystems.com> wrote:
> > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Bruce Richardson
> > Sent: Monday, November 9, 2020 11:06 AM
> >
> > On Mon, Nov 09, 2020 at 09:16:27AM +0100, Morten Brørup wrote:
> > > +CC techboard
> > >
> > > > From: Jerin Jacob [mailto:jerinjacobk@gmail.com]
> > > > Sent: Monday, November 9, 2020 6:18 AM
> > > >
> > > > On Sun, Nov 8, 2020 at 2:03 AM Thomas Monjalon
> > <thomas@monjalon.net>
> > > > wrote:
> > > > >
> > > > > 07/11/2020 20:05, Jerin Jacob:
> > > > > > On Sun, Nov 8, 2020 at 12:09 AM Thomas Monjalon
> > > > <thomas@monjalon.net> wrote:
> > > > > > > 07/11/2020 18:12, Jerin Jacob:
> > > > > > > > On Sat, Nov 7, 2020 at 10:04 PM Thomas Monjalon
> > > > <thomas@monjalon.net> wrote:
> > > > > > > > >
> > > > > > > > > The mempool pointer in the mbuf struct is moved
> > > > > > > > > from the second to the first half.
> > > > > > > > > It should increase performance on most systems having 64-
> > byte
> > > > cache line,
> > > > > > > >
> > > > > > > > > i.e. mbuf is split in two cache lines.
> > > > > > > >
> > > > > > > > But In any event, Tx needs to touch the pool to freeing
> > back to
> > > > the
> > > > > > > > pool upon Tx completion. Right?
> > > > > > > > Not able to understand the motivation for moving it to the
> > > > first 64B cache line?
> > > > > > > > The gain varies from driver to driver. For example, a
> > Typical
> > > > > > > > ARM-based NPU does not need to
> > > > > > > > touch the pool in Rx and its been filled by HW. Whereas it
> > > > needs to
> > > > > > > > touch in Tx if the reference count is implemented.
> > > > > >
> > > > > > See below.
> > > > > >
> > > > > > > >
> > > > > > > > > Due to this change, tx_offload is moved, so some vector
> > data
> > > > paths
> > > > > > > > > may need to be adjusted. Note: OCTEON TX2 check is
> > removed
> > > > temporarily!
> > > > > > > >
> > > > > > > > It will be breaking the Tx path, Please just don't remove
> > the
> > > > static
> > > > > > > > assert without adjusting the code.
> > > > > > >
> > > > > > > Of course not.
> > > > > > > I looked at the vector Tx path of OCTEON TX2,
> > > > > > > it's close to be impossible to understand :)
> > > > > > > Please help!
> > > > > >
> > > > > > Off course. Could you check the above section any share the
> > > > rationale
> > > > > > for this change
> > > > > > and where it helps and how much it helps?
> > > > >
> > > > > It has been concluded in the techboard meeting you were part of.
> > > > > I don't understand why we restart this discussion again.
> > > > > I won't have the energy to restart this process myself.
> > > > > If you don't want to apply the techboard decision, then please
> > > > > do the necessary to request another quick decision.
> > > >
> > > > Yes. Initially, I thought it is OK as we have 128B CL, After
> > looking
> > > > into Thomas's change, I realized
> > > > it is not good for ARM64 64B catchlines based NPU as
> > > > - A Typical ARM-based NPU does not need to touch the pool in Rx
> > and
> > > > its been filled by HW. Whereas it needs to
> > > > touch in Tx if the reference count is implemented.
> > >
> > > Jerin, I don't understand what the problem is here...
> > >
> > > Since RX doesn't touch m->pool, it shouldn't matter for RX which
> > cache line m->pool resides in. I get that.
> > >
> > > You are saying that TX needs to touch m->pool if the reference count
> > is implemented. I get that. But I don't understand why it is worse
> > having m->pool in the first cache line than in the second cache line;
> > can you please clarify?
> > >
> > > > - Also it will be effecting exiting vector routines
> > >
> > > That is unavoidable if we move something from the second to the first
> > cache line.
> > >
> > > It may require some rework on the vector routines, but it shouldn't
> > be too difficult for whoever wrote these vector routines.
> > >
> > > >
> > > > I request to reconsider the tech board decision.
> > >
> > > I was on the techboard meeting as an observer (or whatever the
> > correct term would be for non-members), and this is my impression of
> > the decision on the meeting:
> > >
> > > The techboard clearly decided not to move any dynamic fields in the
> > first cache line, on the grounds that if we move them away again in a
> > later version, DPDK users utilizing a dynamic field in the first cache
> > line might experience a performance drop at that later time. And this
> > will be a very bad user experience, causing grief and complaints. To
> > me, this seemed like a firm decision, based on solid arguments.
> > >
> > > Then the techboard discussed which other field to move to the freed
> > up space in the first cache line. There were no performance reports
> > showing any improvements by moving the any of the suggested fields (m-
> > >pool, m->next, m->tx_offload), and there was a performance report
> > showing no improvements by moving m->next in a test case with large
> > segmented packets. The techboard decided to move m->pool as originally
> > suggested. To me, this seemed like a somewhat random choice between A,
> > B and C, on the grounds that moving one of them is probably better than
> > moving none of them.
> > >
> >
> > This largely tallies with what I remember of the discussion too.
> >
> > I'd also add though that the choice between the next pointer and the
> > pool
> > pointer came down to the fact that the next pointer is only used for
> > chained, multi-segment, packets - which also tend to be larger packets
> > -
> > while the pool pointer is of relevance to all packets, big and small,
> > single and multi-segment.
>
> I wish that was the case, but I am not so sure...
>
> It is true that m->next is NULL for non-segmented packets.
Yes m->next is NULL for non-segmented packets.
Do we need to modify rte_mbuf_check() to enforce these checks?
next prev parent reply other threads:[~2020-11-09 18:04 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-07 15:53 [dpdk-dev] " Thomas Monjalon
2020-11-07 17:12 ` Jerin Jacob
2020-11-07 18:39 ` Thomas Monjalon
2020-11-07 19:05 ` Jerin Jacob
2020-11-07 20:33 ` Thomas Monjalon
2020-11-09 5:18 ` Jerin Jacob
2020-11-09 8:04 ` Thomas Monjalon
2020-11-09 8:27 ` Jerin Jacob
2020-11-09 9:47 ` Bruce Richardson
2020-11-09 12:01 ` Jerin Jacob
2020-11-09 12:59 ` Thomas Monjalon
2020-11-09 13:35 ` Jerin Jacob
2020-11-09 14:02 ` Thomas Monjalon
2020-11-09 14:08 ` Jerin Jacob
2020-11-09 14:42 ` Thomas Monjalon
2020-11-09 14:53 ` Jerin Jacob
2020-11-09 8:16 ` Morten Brørup
2020-11-09 10:06 ` [dpdk-dev] [dpdk-techboard] " Bruce Richardson
2020-11-09 10:21 ` Morten Brørup
2020-11-09 18:04 ` Stephen Hemminger [this message]
2020-11-10 7:15 ` Morten Brørup
2020-11-07 18:57 ` [dpdk-dev] " Morten Brørup
2020-11-09 10:08 ` Bruce Richardson
2020-11-09 10:30 ` Morten Brørup
2020-11-09 10:33 ` Ananyev, Konstantin
2020-11-09 10:36 ` Bruce Richardson
2020-11-09 11:24 ` Ananyev, Konstantin
2020-11-09 21:29 ` [dpdk-dev] [PATCH v2 0/2] move mbuf pool pointer Thomas Monjalon
2020-11-09 21:29 ` [dpdk-dev] [PATCH v2 1/2] drivers: disable OCTEON TX2 in 32-bit build Thomas Monjalon
2020-11-10 18:05 ` Jerin Jacob
2020-11-09 21:29 ` [dpdk-dev] [PATCH v2 2/2] mbuf: move pool pointer in first half Thomas Monjalon
2020-11-10 10:05 ` Morten Brørup
2020-11-10 10:44 ` Thomas Monjalon
2020-11-10 16:25 ` Olivier Matz
2020-11-10 18:06 ` Jerin Jacob
2020-11-12 14:39 ` Thomas Monjalon
2020-11-10 18:08 ` Stephen Hemminger
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=20201109100430.5edb39f3@hermes.local \
--to=stephen@networkplumber.org \
--cc=ajit.khaparde@broadcom.com \
--cc=andrew.rybchenko@oktetlabs.ru \
--cc=bruce.richardson@intel.com \
--cc=david.marchand@redhat.com \
--cc=dev@dpdk.org \
--cc=ferruh.yigit@intel.com \
--cc=hemant.agrawal@nxp.com \
--cc=jerinj@marvell.com \
--cc=jerinjacobk@gmail.com \
--cc=kirankumark@marvell.com \
--cc=konstantin.ananyev@intel.com \
--cc=mb@smartsharesystems.com \
--cc=mdr@ashroe.eu \
--cc=ndabilpuram@marvell.com \
--cc=nhorman@tuxdriver.com \
--cc=olivier.matz@6wind.com \
--cc=techboard@dpdk.org \
--cc=thomas@monjalon.net \
--cc=viacheslavo@nvidia.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).