DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Burakov, Anatoly" <anatoly.burakov@intel.com>
To: Jordan Rhody <jordan@velocloud.net>, dev@dpdk.org
Subject: Re: [dpdk-dev] DPDK Amazon AWS ENA driver problem with ping
Date: Mon, 11 Dec 2017 09:49:41 +0000	[thread overview]
Message-ID: <4469ad2b-3f5e-1597-bd9a-485a87261364@intel.com> (raw)
In-Reply-To: <CAELk9LLuw1h9XhuSyVu3xmWn7WpZs893UWJ-tO_5J8NzC266ew@mail.gmail.com>

On 04-Dec-17 10:11 PM, Jordan Rhody wrote:
> Hi,
> 
> I have been testing the DPDK ENA driver and am hoping that someone can help
> in assisting an observation I have.
> 
> Using the DPDK ENA driver, I can send a ping request/reply from client A to
> client B through an Edge device running DPDK.
> 
> However, the ping request that I observe always contains a '0xffff' for the
> sequence id of the icmp request:
> 
> Internet Control Message Protocol
>      Type: 8 (Echo (ping) request)
>      Code: 0
>      Checksum: 0xf9a1 incorrect, should be 0xfa8c
>      [Checksum Status: Bad]                         <--- checksum is
> original and unmodified
>      Identifier (BE): 2385 (0x0951)
>      Identifier (LE): 20745 (0x5109)
>      Sequence number (BE): 65535 (0xffff)    <--- here
>      Sequence number (LE): 65535 (0xffff)    <--- here
>      [No response seen]
>      Timestamp from icmp data: Nov 21, 2017 16:41:20.000000000 PST
>      [Timestamp from icmp data (relative): 1.042436000 seconds]
>      Data (48 bytes)
>          Data: 612d0f0000000000101112131415161718191a1b1c1d1e1f...
>          [Length: 48]
> 
> I have traced the packet to the entry point of eth_ena_xmit_pkts( ) and it
> does *not* have sequence id 0xffff.  The sequence id seems sane and valid
> at that point.
> 
> Turning off DPDK and using the Linux Amazon AWS ENA driver yields no
> problems at all.  ping request/reply works without a problem.  As an aside,
> this same code works for the DPDK igb, ixgbe and virtio driver.
> 
> I have patched the DPDK ENA driver with all patches through DPDK 17.05 to
> no avail and with the same results.
> 
> Question -- the basic: Could there be something 'wrong' with my AWS network
> interface with regard to setup, perhaps something that may be obvious to
> other folks more familiar with AWS?
> 
> If not, has anyone else experienced problems with DPDK AWS ENA driver?
> 
> Are there any possible patches out there that have not been commit to the
> git repo that may fix this particular issue?
> 
> Thanks in advance,
> 
> Jordan Rhode
> 
Hi Jordan,

I have no experience with the AWS ENA driver and i thus may be quite off 
the mark here, but presumably to respond to pings you would have to have 
an IP stack running on top of DPDK. Otherwise if you are running e.g. 
l2fwd on that interface, it'll just forward your ping requests back 
unmodified.

-- 
Thanks,
Anatoly

      reply	other threads:[~2017-12-11  9:49 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-12-04 22:11 Jordan Rhody
2017-12-11  9:49 ` Burakov, Anatoly [this message]

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=4469ad2b-3f5e-1597-bd9a-485a87261364@intel.com \
    --to=anatoly.burakov@intel.com \
    --cc=dev@dpdk.org \
    --cc=jordan@velocloud.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).