From mboxrd@z Thu Jan  1 00:00:00 1970
From: Ferruh Yigit <ferruh.yigit@intel.com>
Subject: Re: [PATCH v2 1/2] net/ena: convert to new Tx offloads
	API
Date: Thu, 18 Jan 2018 14:49:22 +0000
Message-ID: <13f38c4c-7758-1ea7-16ba-ab0453f101a1@intel.com>
References: <1516103563-9275-2-git-send-email-rk@semihalf.com>
 <1516177424-8613-1-git-send-email-rk@semihalf.com>
 <953fa892-d65f-2aca-9909-a812ef6a18c4@intel.com>
 <CAMG3L4_L1dZdu2Qk0nVe9duTwowAuhuegUKMVOGLpuuHP1XRug@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Cc: dev@dpdk.org, Shahaf Shuler <shahafs@mellanox.com>,
 Marcin Wojtas <mw@semihalf.com>, =?UTF-8?Q?Micha=c5=82_Krawczyk?=
 <mk@semihalf.com>, "Tzalik, Guy" <gtzalik@amazon.com>, evgenys@amazon.com,
 "Matushevsky, Alexander" <matua@amazon.com>,
 "Chauskin, Igor" <igorch@amazon.com>
To: =?UTF-8?Q?Rafa=c5=82_Kozik?= <rk@semihalf.com>
Return-path: <dev-bounces@dpdk.org>
Received: from mga04.intel.com (mga04.intel.com [192.55.52.120])
 by dpdk.org (Postfix) with ESMTP id 9395B1B328
 for <dev@dpdk.org>; Thu, 18 Jan 2018 15:49:27 +0100 (CET)
In-Reply-To: <CAMG3L4_L1dZdu2Qk0nVe9duTwowAuhuegUKMVOGLpuuHP1XRug@mail.gmail.com>
Content-Language: en-US
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <https://dpdk.org/ml/options/dev>,
 <mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <https://dpdk.org/ml/listinfo/dev>,
 <mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
Sender: "dev" <dev-bounces@dpdk.org>

On 1/18/2018 9:18 AM, Rafał Kozik wrote:
> 2018-01-17 19:58 GMT+01:00 Ferruh Yigit <ferruh.yigit@intel.com>:
>> On 1/17/2018 8:23 AM, Rafal Kozik wrote:
>>> Ethdev Tx offloads API has changed since:
>>>
>>> commit cba7f53b717d ("ethdev: introduce Tx queue offloads API")
>>>
>>> This commit support the new Tx offloads API. Queue configuration
>>> is stored in ena_ring.offloads. During preparing mbufs for tx, offloads are
>>> allowed only if appropriate flags in this field are set.
>>>
>>> Signed-off-by: Rafal Kozik <rk@semihalf.com>
>>
>> <...>
>>
>>> @@ -280,21 +290,24 @@ static inline void ena_rx_mbuf_prepare(struct rte_mbuf *mbuf,
>>>  }
>>>
>>>  static inline void ena_tx_mbuf_prepare(struct rte_mbuf *mbuf,
>>> -                                    struct ena_com_tx_ctx *ena_tx_ctx)
>>> +                                    struct ena_com_tx_ctx *ena_tx_ctx,
>>> +                                    uint64_t queue_offloads)
>>>  {
>>>       struct ena_com_tx_meta *ena_meta = &ena_tx_ctx->ena_meta;
>>>
>>> -     if (mbuf->ol_flags &
>>> -         (PKT_TX_L4_MASK | PKT_TX_IP_CKSUM | PKT_TX_TCP_SEG)) {
>>> +     if ((mbuf->ol_flags & MBUF_OFFLOADS) &&
>>> +         (queue_offloads & QUEUE_OFFLOADS)) {
>>>               /* check if TSO is required */
>>> -             if (mbuf->ol_flags & PKT_TX_TCP_SEG) {
>>> +             if ((mbuf->ol_flags & PKT_TX_TCP_SEG) &&
>>> +                 (queue_offloads & DEV_TX_OFFLOAD_TCP_TSO)) {
>>>                       ena_tx_ctx->tso_enable = true;
>>>
>>>                       ena_meta->l4_hdr_len = GET_L4_HDR_LEN(mbuf);
>>>               }
>>>
>>>               /* check if L3 checksum is needed */
>>> -             if (mbuf->ol_flags & PKT_TX_IP_CKSUM)
>>> +             if ((mbuf->ol_flags & PKT_TX_IP_CKSUM) &&
>>> +                 (queue_offloads & DEV_TX_OFFLOAD_IPV4_CKSUM))
>>>                       ena_tx_ctx->l3_csum_enable = true;
>>
>> This function is fast path right?
>> Do you really need new extra check to queue_offloads, isn't that information is
>> for setup phase?
>>
> 
> ENA does not have a switch for enabling/disabling offloads during configuration.
> We must use additional variable and track it, otherwise the driver could use
> checksum offloads by enabling it in mbuf although it was disabled in
> queue configuration.

OK, thanks for clarification.