DPDK patches and discussions
 help / color / mirror / Atom feed
From: Asaf Penso <asafp@mellanox.com>
To: Georgios Katsikas <katsikas.gp@gmail.com>, "dev@dpdk.org" <dev@dpdk.org>
Cc: Tom Barbette <barbette@kth.se>
Subject: Re: [dpdk-dev] Testpmd parameter --enable-lro fails on Mellanox	ConnectX5
Date: Tue, 15 Oct 2019 14:06:53 +0000	[thread overview]
Message-ID: <AM6PR05MB6615476240E9BEAD511321A4C0930@AM6PR05MB6615.eurprd05.prod.outlook.com> (raw)
In-Reply-To: <CAN9HtFBUvV9T+1NnRv9-2Ar9Uwh5d9EFhY5ieMDiiTyRfk+iZQ@mail.gmail.com>

Hello Georgios,

Thanks for your mail!
We'll have a deeper look internally and will contact you about our results.

Regards,
Asaf Penso

> -----Original Message-----
> From: dev <dev-bounces@dpdk.org> On Behalf Of Georgios Katsikas
> Sent: Tuesday, October 15, 2019 9:06 AM
> To: dev@dpdk.org
> Cc: Tom Barbette <barbette@kth.se>
> Subject: [dpdk-dev] Testpmd parameter --enable-lro fails on Mellanox
> ConnectX5
> 
> Hi all,
> 
> In the latest features of the MLX5 PMD (i.e., DPDK 19.08), there is LRO
> hardware support.
> However, I cannot make it work with our dual port 100GbE Mellanox
> ConnectX5
> (MT27800) card.
> 
> Let me first describe the configuration of my testbed: First,I upgraded
> OFED to the latest 4.7-1.0.0 with firmware version 16.26.1040
> (MT_0000000008) and installed the latest MFT (4.13) to enable DevX
> according to the instructions here
> <https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdoc.
> dpdk.org%2Fguides%2Fnics%2Fmlx5.html&amp;data=02%7C01%7Casafp%40
> mellanox.com%7C18e034c229d64911ff7b08d75135d96a%7Ca652971c7d2e4d9
> ba6a4d149256f461b%7C0%7C1%7C637067164036442822&amp;sdata=PfV3NV
> Nj3t0Bw3CiWxRIiVgats8Ma1b755qY4UIXXo0%3D&amp;reserved=0>:
> 
> mlxconfig -d <device> set UCTX_EN=1
> 
> My OS is Ubuntu server 18.04 with kernel 4.15.0-65-generic. I also tested
> on another server with the same distro but more recent kernel (5.15),
> without any luck.
> 
> To test LRO, I simply use testpmd as follows:
> 
> sudo ./testpmd -l 0-15 -w 0000:03:00.1 -v -- --txq=16 --rxq=16
> --nb-cores=15 --enable-lro -i
> 
> Th output I get is:
> 
> Configuring Port 0 (socket 0)
> net_mlx5: Failed to create RQ using DevX
> net_mlx5: port 0 Rx queue 8 RQ creation failure
> net_mlx5: port 0 Rx queue allocation failed: Cannot allocate memory
> Fail to start port 0
> Please stop the ports first
> Done
> 
> Reducing the number of queues does not solve the problem. Specifically, if
> I ask for 1 rX and 1 Tx queue, testpmd  fails without an error from the mlx
> driver:
> 
> Configuring Port 0 (socket 0)
> Fail to start port 0
> Please stop the ports first
> Done
> 
> I would appreciate your help on this issue.
> 
> Best regards,
> Georgios

  reply	other threads:[~2019-10-15 14:06 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-15  6:06 Georgios Katsikas
2019-10-15 14:06 ` Asaf Penso [this message]
2019-10-17 16:20   ` Asaf Penso
2019-10-18  6:51     ` Georgios Katsikas

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=AM6PR05MB6615476240E9BEAD511321A4C0930@AM6PR05MB6615.eurprd05.prod.outlook.com \
    --to=asafp@mellanox.com \
    --cc=barbette@kth.se \
    --cc=dev@dpdk.org \
    --cc=katsikas.gp@gmail.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).