DPDK usage discussions
 help / color / mirror / Atom feed
From: Andrea Bigagli <andrea.bigagli@commprove.com>
To: users@dpdk.org
Subject: [dpdk-users] Is it possible to TX through a port with no RX connected?
Date: Sun, 12 Jun 2016 22:27:55 +0200	[thread overview]
Message-ID: <3878315A-53FD-4721-B527-328FB5C2EFC6@commprove.com> (raw)

Hi everyone.
Moving first steps with DPDK, so forgive my noobness or if this is a dumb question, but I’ve searched around before posting and didn’t find anything that could help me.

I have 4 Intel 10Gbps cards for a total of 8 RX/TX ports.
What I’m trying to do is a kind of “8->4 tapping-concentrator”: 8 fibers are tapped through an optical splitter and connected to the 8 RX ports, some packet analysis is performed and based on certain properties of the packets, traffic is re-routed on only 4 TX ports.
The application guarantees that overall input bandwidth is less than 30Gbps, so using only 4 TXs is not a problem (the balancing is almost pretty even).
The problem I’m facing is that the 8 monitored RX are not always up. The system being monitored at times shuts down some of those 8 connections and so on some of the 8 RXs the link goes down, and when that happens, the corresponding TX stops working.
I’m not a great expert of the inner working of the 10Gbps cards, I’ve skimmed through the 82599 datasheet and tried to play with some registers (i.e. setting the “FORCE-LINK-UP” bit in the AUTOC register), but couldn’t find a way to keep the TX ports running without something “alive” on the corresponding RX port.

Is there any way (maybe diving down at the PHY layer, which I’m really not familiar with) to have TX ports working regardless of their corresponding RX port status?
Looping back something in the RX port is not feasible in my case because the RX port has to be dedicated to the tapped traffic, for when there’s actually traffic on it.

Hope I was able to explain myself, and hope someone could point me in some direction as I’m totally stuck at the moment.


Thanks,
Andrea.

             reply	other threads:[~2016-06-12 20:27 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-06-12 20:27 Andrea Bigagli [this message]
2016-06-13  4:36 ` Muhammad Zain-ul-Abideen
2016-06-14 21:31   ` Andrea Bigagli
2016-06-14 21:39     ` Muhammad Zain-ul-Abideen
2016-06-14 21:43       ` Andrea Bigagli
2016-06-15 12:56         ` Pavey, Nicholas
2016-06-15 16:42           ` Andrea Bigagli

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=3878315A-53FD-4721-B527-328FB5C2EFC6@commprove.com \
    --to=andrea.bigagli@commprove.com \
    --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).