DPDK usage discussions
 help / color / mirror / Atom feed
From: "Van Haaren, Harry" <harry.van.haaren@intel.com>
To: Tom Barbette <barbette@kth.se>,
	Perugu Hemasai Chandra Prasad <hemasaiperugu@5g.iith.ac.in>,
	"users@dpdk.org" <users@dpdk.org>, "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-users] [dpdk-dev] Sharing Data structure between logical	cores in DPDK-	regarding
Date: Thu, 5 Dec 2019 10:54:59 +0000	[thread overview]
Message-ID: <MN2PR11MB44471C570BCC43F196D1048DD75C0@MN2PR11MB4447.namprd11.prod.outlook.com> (raw)
In-Reply-To: <1575542540897.56972@kth.se>

> -----Original Message-----
> From: users <users-bounces@dpdk.org> On Behalf Of Tom Barbette
> Sent: Thursday, December 5, 2019 10:42 AM
> To: Perugu Hemasai Chandra Prasad <hemasaiperugu@5g.iith.ac.in>;
> users@dpdk.org; dev@dpdk.org
> Subject: Re: [dpdk-users] [dpdk-dev] Sharing Data structure between logical
> cores in DPDK- regarding
> 
> It depends on the datastructure.
> 
> If you use rte_hash, with the thread safety flags, it's safe.
> 
> In any case having a lot of cores accessing the same DS will lead to bad
> performance. You should try to have per-core data structures whenever
> possible.
> 
> Tom
> 
> ________________________________________
> De : dev <dev-bounces@dpdk.org> de la part de Perugu Hemasai Chandra Prasad
> <hemasaiperugu@5g.iith.ac.in>
> Envoyé : jeudi 5 décembre 2019 11:16
> À : users@dpdk.org; dev@dpdk.org
> Objet : [dpdk-dev] Sharing Data structure between logical cores in DPDK-
> regarding
> 
> Hi All,

Hi!

>           I have a small doubt, can we share a data structure between
> multiple logical cores in DPDK without locking?

Note that "locking" and "atomic instructions" are not the same. So locking a mutex is only one way of ensuring multiple cores have a coherent view on the datastructure.

The other method of doing "lock free" but multi-core safe data-structures is by using Atomic instructions, which ensure that results are observed correctly on all cores accessing that data-structure. A good example is the Multi-producer or Multi-consumer versions of the DPDK rte_ring ringbuffer.

Regards, -Harry

> I have tested it by sharing
> a small structure with two variable and incrementing them in all logical
> cores. It ran smooth I didn't get any issue. But I doubt if we can run it
> for long time with some huge data structure having many elements getting
> accessed by multiple logical cores.
>      Can anyone please clarify this.
> 
> Thanks and regards,
> Hemasai.

  reply	other threads:[~2019-12-05 10:55 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-05 10:16 [dpdk-users] " Perugu Hemasai Chandra Prasad
2019-12-05 10:42 ` [dpdk-users] [dpdk-dev] " Tom Barbette
2019-12-05 10:54   ` Van Haaren, Harry [this message]
2019-12-06 16:30     ` Stephen Hemminger
2019-12-06 16:51       ` Van Haaren, Harry
2019-12-06 17:01         ` Stephen Hemminger
2019-12-10  5:56   ` Honnappa Nagarahalli

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=MN2PR11MB44471C570BCC43F196D1048DD75C0@MN2PR11MB4447.namprd11.prod.outlook.com \
    --to=harry.van.haaren@intel.com \
    --cc=barbette@kth.se \
    --cc=dev@dpdk.org \
    --cc=hemasaiperugu@5g.iith.ac.in \
    --cc=users@dpdk.org \
    /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).