[dpdk-users] Intel X550 Receive timestamping for all packets

["terry.montague.1980@btinternet.com" <terry.montague.1980@btinternet.com>]

Hi all,
Has anyone tried modifying the ixgbe PMD driver to allow an Intel X550 card to insert the extra 8 bytes of data at the end of each receive message, the 8 bytes containing the receive SYSTIME  timestamp ?
I've adjusted TYPE in TSYNCRXCTL to be "all packets" and set TSIP_UT_EN. I get the 8 bytes of data appended to all packets, but the controller doesn't populate the memory - the contents are left unchanged when transmitted.
TS and TSIP are both set in the received RDESC descriptor (along with the receive byte count increasing by 8 bytes) so the adaptor is definitely doing something.
Any ideas anyone ?
Many thanks
Terry.
From rauln@cpqd.com.br  Fri Sep 15 21:31:28 2017
Return-Path: <rauln@cpqd.com.br>
Received: from mail-pf0-f179.google.com (mail-pf0-f179.google.com
 [209.85.192.179]) by dpdk.org (Postfix) with ESMTP id 94FD53238
 for <users@dpdk.org>; Fri, 15 Sep 2017 21:31:28 +0200 (CEST)
Received: by mail-pf0-f179.google.com with SMTP id l188so1927326pfc.6
 for <users@dpdk.org>; Fri, 15 Sep 2017 12:31:28 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=cpqd.com.br; s=google;
 h=mime-version:in-reply-to:references:from:date:message-id:subject:to;
 bh=xfyCKaWRrbo+fKaXr0nfUBjFFnZS5UnT+sgCXXzJZd8=;
 b=MHUvy8IihxrqsH0HV5T++BvIuPuC6CnwLk14zI3NprJRj99euZRhSfSH/9ueiY+C7/
 MRVXoEaMOXO+/Ae2Aw2RDIgfJ55uzldzTxgW0fje/qzD/qXHxQEP7FhUaRR+Xz8Zdtn0
 4ane//mOzj+1RpubJKvYX7KYspQeZZW25pAUE=
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
 d=1e100.net; s=20161025;
 h=x-gm-message-state:mime-version:in-reply-to:references:from:date
 :message-id:subject:to;
 bh=xfyCKaWRrbo+fKaXr0nfUBjFFnZS5UnT+sgCXXzJZd8=;
 b=m6sxFTt7NLEfEtBIGiif3ogd3Ii6eBMMnWBa5Bt0KZ+vSo6RaeNUnCJJW6l1IQQ4re
 3+SzhRjrJe6a9xJqYpYoXIR5ELaP+ZzDflZ/mx2KTrO0RPAK7tpgFRh/dwaTDusLceOO
 Ib16K3t0BJzxIj/8Rmghu2qVT9OtXFelZaDdxYS3QvWriS9fl7ILrqm4th8dr+G/z16b
 5fKouLF8NcAbGjCzZ0lNVh35O3zQOCOGNkiMh38MXunBLXBJpqzQSbp4qSq4RECfySdJ
 2ghX0y6MRPYB7Y+MKIiCUXgdsWUFnkumhxBZXpNVAnnqv5Bs5OIjHC4OqcWqtTj7Fz+i
 lkVQ==
X-Gm-Message-State: AHPjjUghvhYrkwfjLCLwkzeha98mOkW2/J3qY5dbvXEuRZaebGB2VlmX
 SwofXW1Vj2JZPi9ezPs6KWah0GUFbegagfBJERaGvNC5
X-Google-Smtp-Source: AOwi7QDSSuFqro23f+J9T/8+KEL9iUfH8MUS+Pnj7SZSphsdvUJRzh0RMWGqV+5NX1fnizAs1sMF5abP2Wxmn4oIl5g=
X-Received: by 10.99.171.9 with SMTP id p9mr13211339pgf.30.1505503887209; Fri,
 15 Sep 2017 12:31:27 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.100.179.101 with HTTP; Fri, 15 Sep 2017 12:30:56 -0700 (PDT)
In-Reply-To: <CAEJM1hhjUz_5aRSRt-FnXBhy040gBaGLnsJrT7LSTLoTw41D7A@mail.gmail.com>
References: <CAEJM1hhjUz_5aRSRt-FnXBhy040gBaGLnsJrT7LSTLoTw41D7A@mail.gmail.com>
From: Raul Negreiros <rauln@cpqd.com.br>
Date: Fri, 15 Sep 2017 16:30:56 -0300
Message-ID: <CAEJM1hiBhF2aBQT3-2JyC2vXgTcck1kQSm1e3B0wxO7jutyyyA@mail.gmail.com>
To: users@dpdk.org
Content-Type: text/plain; charset="UTF-8"
X-Content-Filtered-By: Mailman/MimeDel 2.1.15
Subject: Re: [dpdk-users] Help with QinQ insertion on i40e
X-BeenThere: users@dpdk.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: DPDK usage discussions <users.dpdk.org>
List-Unsubscribe: <http://dpdk.org/ml/options/users>,
 <mailto:users-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/users/>
List-Post: <mailto:users@dpdk.org>
List-Help: <mailto:users-request@dpdk.org?subject=help>
List-Subscribe: <http://dpdk.org/ml/listinfo/users>,
 <mailto:users-request@dpdk.org?subject=subscribe>
X-List-Received-Date: Fri, 15 Sep 2017 19:31:28 -0000

Hi,

I didn't notice that the default configuration in the NIC's queue is to
disable TX offloading. I was using the 'rte_eth_tx_queue_setup' function
with the last parameter 'tx_conf' as NULL, in this way using the default
configuration.

After configure the 'txq_flags' attribute properly my issue was solved and
the offloading worked perfectly.

Best regards,

Raul Negreiros

On Thu, Sep 14, 2017 at 2:15 PM, Raul Negreiros <rauln@cpqd.com.br> wrote:

> Hi,
>
> I'm trying to insert QinQ in the packet using the offload feature, but
> unfortunately I'm unsuccessful.
>
> These are my steps:
> - On device configuration configuration:
> - call 'rte_eth_dev_set_vlan_offload' with the following flags:
> ETH_VLAN_STRIP_OFFLOAD, ETH_VLAN_EXTEND_OFFLOAD;
> - call 'rte_eth_dev_set_vlan_ether_type' twice with the inner and outer
> types.
> - On packet processing:
> - set the PKT_TX_QINQ_PKT flag on 'ol_flags' attribute;
> - fill the 'vlan_tci' and 'vlan_tci_outer'.
>
> Did I miss something?
>
> Obs: My interface's driver is i40e and the DEV_TX_OFFLOAD_QINQ_INSERT flag
> is
> active on tx capabilities.
>
> Obs2: I've found on 'drivers/net/i40e/i40e_ethdev.c' the following
> comment:
> "Apply double-vlan setting, not implemented yet" in i40e_dev_init_vlan
> function.
> May it means that actually is not possible to use QinQ insertion with this
> driver?
>
> Best regards,
>
> Raul Negreiros
>