DPDK patches and discussions
 help / color / mirror / Atom feed
From: Kamaraj P <pkamaraj@gmail.com>
To: Thomas Monjalon <thomas@monjalon.net>
Cc: Bruce Richardson <bruce.richardson@intel.com>,
	David Marchand <david.marchand@redhat.com>,  dev <dev@dpdk.org>,
	Anatoly Burakov <anatoly.burakov@intel.com>,
	 Ferruh Yigit <ferruh.yigit@intel.com>,
	ksimha@cisco.com, kalas@cisco.com,  seveluch@cisco.com,
	"techboard@dpdk.org" <techboard@dpdk.org>,
	hpai@cisco.com, ppitchai@cisco.com,
	 Aaron Conole <aconole@redhat.com>,
	"Kamaraj P (kamp)" <kamp@cisco.com>,
	jchhapek@cisco.com
Subject: Re: [dpdk-dev] Running DPDK application with non-previlege mode
Date: Mon, 6 Sep 2021 19:32:36 +0530	[thread overview]
Message-ID: <CAG8PAaoetZmi3=uzeEMpGEmngVEPZMB=HB_p7HHQG-3Yb=v6=A@mail.gmail.com> (raw)
In-Reply-To: <CAG8PAarFqbpPW5BGYZ9O8=h_22HnOhLv4TnaKs7VyHLR8Vm54A@mail.gmail.com>

Hi Thomas/Antoly,

We have tried with the below capability to start a DPDK application as a
container.
--cap-add=NET_ADMIN --cap-add=SYS_RESOURCE  --cap-add=SYS_NICE

Still We could see the failures when rte_eali_init() ..

09-06-2021  10:48:37.423583 EAL: NUMA is not supported.
*09-06-2021  10:48:37.423895 EAL: alloc_seg(): get_mempolicy: Function not
implemented  ==> Does this sound like virtual memory is not enough in the
running system (because of NUMA not supported )?*
09-06-2021  10:48:37.423910 EAL: Ask a virtual area of 0x200000 bytes
09-06-2021  10:48:37.423919 EAL: Virtual area found at 0x2200200000 (size =
0x200000)
09-06-2021  10:48:37.423933 EAL: attempted to allocate 1 segments, but only
0 were allocated
09-06-2021  10:48:37.423945 EAL: NUMA is not supported.
*09-06-2021  10:48:37.427811 EAL: alloc_seg(): mmap() failed: Cannot
allocate memory ==>  SYS_RESOURCE would suffice here ?*
09-06-2021  10:48:37.427821 EAL: Ask a virtual area of 0x40000000 bytes
09-06-2021  10:48:37.427829 EAL: Virtual area found at 0x140000000 (size =
0x40000000)
09-06-2021  10:48:37.427839 EAL: attempted to allocate 1 segments, but only
0 were allocated
09-06-2021  10:48:37.427847 EAL: error allocating rte services array
09-06-2021  10:48:37.427869 EAL: rte_service_init() failed


As for the dpdk options when bringing up our application,
--log-level=*:8
--lcores=(0-1)@0
--in-memory
-d/usr/lib64/dpdk-19/librte_pmd_ixgbe.so.20.0
-d/usr/lib64/dpdk-19/librte_pmd_e1000.so.20.0
-d/usr/lib64/dpdk-19/librte_pmd_bnxt.so.20.0
--vdev=net_af_packet0,iface=dpbr_100
--vdev=net_af_packet1,iface=asac-in
--vdev=net_af_packet2,iface=asac-out
*--iova-mode=va  ==> will valid here ?*

Would assume some of the flags were not enough to run DPDK on top of our
environment ? Any memory configuration which needs to be checked for the
underlying system ?
Please share your thoughts.

Thanks,
Kamaraj

On Sun, Sep 5, 2021 at 10:24 PM Kamaraj P <pkamaraj@gmail.com> wrote:

> Hello Thomas,
> THanks for the suggestion.
> Yes. Our DPDK application already has an in-memory option.
> BTW we are not using a multi-process environment for our DPDK application.
>
> Thanks,
> Kamaraj
>
>
> On Wed, Sep 1, 2021 at 11:23 PM Thomas Monjalon <thomas@monjalon.net>
> wrote:
>
>> 01/09/2021 18:09, Kamaraj P:
>> > Thanks Bruce.
>> >
>> > BTW we are trying to run our application as docker container.
>> > We may need to pass the parameters(system capability) along with docker
>> > command line to mount the hugepages ?
>> > Like this :
>> https://github.com/usnistgov/ndn-dpdk/blob/main/docs/Docker.md
>> >
>> > Please share your thoughts.
>>
>> The easiest is to run with --in-memory option.
>>
>> If you really want the hugetlb files (for multi-process),
>> use such docker option:
>> --mount type=bind,src=/dev/hugepages,dst=/dev/hugepages
>>
>>
>>
>>

  reply	other threads:[~2021-09-06 14:02 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-30  7:37 [dpdk-dev] Running DPDK application without " Kamaraj P
2021-08-30  7:52 ` [dpdk-dev] Running DPDK application with " Kamaraj P
2021-08-30  8:37   ` Kamaraj P
2021-08-30 11:26   ` Ferruh Yigit
2021-08-30 11:45     ` Kamaraj P
2021-08-30 12:05       ` Thomas Monjalon
2021-08-30 11:59     ` Thomas Monjalon
2021-08-30 12:06       ` David Marchand
2021-08-31 15:27         ` Kamaraj P
2021-08-31 15:34           ` David Marchand
2021-08-31 16:20             ` Kamaraj P
2021-08-31 19:44               ` Bruce Richardson
2021-09-01 16:09                 ` Kamaraj P
2021-09-01 17:52                   ` Thomas Monjalon
2021-09-05 16:54                     ` Kamaraj P
2021-09-06 14:02                       ` Kamaraj P [this message]
2021-08-31 17:49             ` Aaron Conole

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='CAG8PAaoetZmi3=uzeEMpGEmngVEPZMB=HB_p7HHQG-3Yb=v6=A@mail.gmail.com' \
    --to=pkamaraj@gmail.com \
    --cc=aconole@redhat.com \
    --cc=anatoly.burakov@intel.com \
    --cc=bruce.richardson@intel.com \
    --cc=david.marchand@redhat.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=hpai@cisco.com \
    --cc=jchhapek@cisco.com \
    --cc=kalas@cisco.com \
    --cc=kamp@cisco.com \
    --cc=ksimha@cisco.com \
    --cc=ppitchai@cisco.com \
    --cc=seveluch@cisco.com \
    --cc=techboard@dpdk.org \
    --cc=thomas@monjalon.net \
    /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).