From: "Wang, Haiyue" <haiyue.wang@intel.com>
To: Perugu Hemasai Chandra Prasad <hemasaiperugu@5g.iith.ac.in>
Cc: "users@dpdk.org" <users@dpdk.org>, "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-users] [dpdk-dev] unable to receive packets less than 4 with rte_eth_rx_burst Api -regarding.
Date: Tue, 3 Sep 2019 11:33:48 +0000 [thread overview]
Message-ID: <E3B9F2FDCB65864C82CD632F23D8AB8773D6575F@shsmsx102.ccr.corp.intel.com> (raw)
In-Reply-To: <CALJSgO7WbemgT1qovSsbE5Lyh-q9ruQpn6OAA+FCXJx9Fcgg8Q@mail.gmail.com>
Hi Hemasai
You can see the comments in code like " /* A.1 load 4 pkts desc */ ", it use vector
instruction to handle HW desc *in batch*.
And also, you can change it to non-vector function by adjust some parameters, please
refer to i40e_set_rx_function for detail, it has debug message for your reference.
BR,
Haiyue
From: Perugu Hemasai Chandra Prasad [mailto:hemasaiperugu@5g.iith.ac.in]
Sent: Tuesday, September 3, 2019 17:48
To: Wang, Haiyue <haiyue.wang@intel.com>
Cc: users@dpdk.org; dev@dpdk.org
Subject: Re: [dpdk-dev] unable to receive packets less than 4 with rte_eth_rx_burst Api -regarding.
Hi Haiyue,
It is working ,thanks for your response. can you please tell what kind of performance limitation we have if RTE_I40E_DESCS_PER_LOOP value is less than 4 ?
Thanks & Regards,
Hemasai
On Tue, Sep 3, 2019 at 2:53 PM Perugu Hemasai Chandra Prasad <hemasaiperugu@5g.iith.ac.in> wrote:
Hi,
I tried changing "RTE_I40E_DESCS_PER_LOOP" located in i40e_rxtx.h to 1 and recompiling DPDK but it didn't work. Is there any procedure to make it work for rx_burst of 1 packet?
On Tue, 3 Sep, 2019, 12:32 Wang, Haiyue, <haiyue.wang@intel.com> wrote:
Some PMD Vector Rx for performance has an hidden limitation like :
/* nb_pkts shall be less equal than RTE_I40E_MAX_RX_BURST */
nb_pkts = RTE_MIN(nb_pkts, RTE_I40E_MAX_RX_BURST);
/* nb_pkts has to be floor-aligned to RTE_I40E_DESCS_PER_LOOP */
nb_pkts = RTE_ALIGN_FLOOR(nb_pkts, RTE_I40E_DESCS_PER_LOOP); <------ ;-)
#define RTE_I40E_DESCS_PER_LOOP 4
BR,
Haiyue
> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Perugu Hemasai Chandra Prasad
> Sent: Tuesday, September 3, 2019 13:00
> To: users@dpdk.org; dev@dpdk.org
> Subject: [dpdk-dev] unable to receive packets less than 4 with rte_eth_rx_burst Api -regarding.
>
> Hi All,
> I am trying to receive packets transmitted by another DPDK
> application running on different system. I am able to transmit packets with
> a burst value of 1 using rte_eth_tx_burst Api, but unable to receive
> packets with a burst value less than 4 using rte_eth_rx_burst Api. can
> anyone please tell me the reason behind that? Is that because of any EthDev
> configuration?. How to receive packets with a burst value of 1?
>
> Thanks & Regards,
> Hemasai
prev parent reply other threads:[~2019-09-03 11:34 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-03 5:00 [dpdk-users] " Perugu Hemasai Chandra Prasad
2019-09-03 7:02 ` [dpdk-users] [dpdk-dev] " Wang, Haiyue
2019-09-03 9:23 ` Perugu Hemasai Chandra Prasad
2019-09-03 9:47 ` Perugu Hemasai Chandra Prasad
2019-09-03 11:33 ` Wang, Haiyue [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=E3B9F2FDCB65864C82CD632F23D8AB8773D6575F@shsmsx102.ccr.corp.intel.com \
--to=haiyue.wang@intel.com \
--cc=dev@dpdk.org \
--cc=hemasaiperugu@5g.iith.ac.in \
--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).