From: Ferruh Yigit <ferruh.yigit@amd.com>
To: Feifei Wang <feifei.wang2@arm.com>,
Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
Cc: dev@dpdk.org, nd@arm.com,
Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>
Subject: Re: [PATCH v12 0/4] Recycle mbufs from Tx queue into Rx queue
Date: Fri, 22 Sep 2023 16:30:32 +0100 [thread overview]
Message-ID: <eb0db252-943d-4a45-96d4-a94a5037a4d7@amd.com> (raw)
In-Reply-To: <4c33779b-6c78-4578-a789-e73343a193c8@amd.com>
On 9/20/2023 2:12 PM, Ferruh Yigit wrote:
> On 8/24/2023 8:36 AM, Feifei Wang wrote:
>> Currently, the transmit side frees the buffers into the lcore cache and
>> the receive side allocates buffers from the lcore cache. The transmit
>> side typically frees 32 buffers resulting in 32*8=256B of stores to
>> lcore cache. The receive side allocates 32 buffers and stores them in
>> the receive side software ring, resulting in 32*8=256B of stores and
>> 256B of load from the lcore cache.
>>
>> This patch proposes a mechanism to avoid freeing to/allocating from
>> the lcore cache. i.e. the receive side will free the buffers from
>> transmit side directly into its software ring. This will avoid the 256B
>> of loads and stores introduced by the lcore cache. It also frees up the
>> cache lines used by the lcore cache. And we can call this mode as mbufs
>> recycle mode.
>>
>> In the latest version, mbufs recycle mode is packaged as a separate API.
>> This allows for the users to change rxq/txq pairing in real time in data plane,
>> according to the analysis of the packet flow by the application, for example:
>> -----------------------------------------------------------------------
>> Step 1: upper application analyse the flow direction
>> Step 2: recycle_rxq_info = rte_eth_recycle_rx_queue_info_get(rx_portid, rx_queueid)
>> Step 3: rte_eth_recycle_mbufs(rx_portid, rx_queueid, tx_portid, tx_queueid, recycle_rxq_info);
>> Step 4: rte_eth_rx_burst(rx_portid,rx_queueid);
>> Step 5: rte_eth_tx_burst(tx_portid,tx_queueid);
>> -----------------------------------------------------------------------
>> Above can support user to change rxq/txq pairing at run-time and user does not need to
>> know the direction of flow in advance. This can effectively expand mbufs recycle mode's
>> use scenarios.
>>
>> Furthermore, mbufs recycle mode is no longer limited to the same pmd,
>> it can support moving mbufs between different vendor pmds, even can put the mbufs
>> anywhere into your Rx mbuf ring as long as the address of the mbuf ring can be provided.
>> In the latest version, we enable mbufs recycle mode in i40e pmd and ixgbe pmd, and also try to
>> use i40e driver in Rx, ixgbe driver in Tx, and then achieve 7-9% performance improvement
>> by mbufs recycle mode.
>>
>> Difference between mbuf recycle, ZC API used in mempool and general path
>> For general path:
>> Rx: 32 pkts memcpy from mempool cache to rx_sw_ring
>> Tx: 32 pkts memcpy from tx_sw_ring to temporary variable + 32 pkts memcpy from temporary variable to mempool cache
>> For ZC API used in mempool:
>> Rx: 32 pkts memcpy from mempool cache to rx_sw_ring
>> Tx: 32 pkts memcpy from tx_sw_ring to zero-copy mempool cache
>> Refer link: http://patches.dpdk.org/project/dpdk/patch/20230221055205.22984-2-kamalakshitha.aligeri@arm.com/
>> For mbufs recycle:
>> Rx/Tx: 32 pkts memcpy from tx_sw_ring to rx_sw_ring
>> Thus we can see in the one loop, compared to general path, mbufs recycle mode reduces 32+32=64 pkts memcpy;
>> Compared to ZC API used in mempool, we can see mbufs recycle mode reduce 32 pkts memcpy in each loop.
>> So, mbufs recycle has its own benefits.
>>
>> Testing status:
>> (1) dpdk l3fwd test with multiple drivers:
>> port 0: 82599 NIC port 1: XL710 NIC
>> -------------------------------------------------------------
>> Without fast free With fast free
>> Thunderx2: +7.53% +13.54%
>> -------------------------------------------------------------
>>
>> (2) dpdk l3fwd test with same driver:
>> port 0 && 1: XL710 NIC
>> -------------------------------------------------------------
>> Without fast free With fast free
>> Ampere altra: +12.61% +11.42%
>> n1sdp: +8.30% +3.85%
>> x86-sse: +8.43% +3.72%
>> -------------------------------------------------------------
>>
>> (3) Performance comparison with ZC_mempool used
>> port 0 && 1: XL710 NIC
>> with fast free
>> -------------------------------------------------------------
>> With recycle buffer With zc_mempool
>> Ampere altra: 11.42% 3.54%
>> -------------------------------------------------------------
>>
>> Furthermore, we add recycle_mbuf engine in testpmd. Due to XL710 NIC has
>> I/O bottleneck in testpmd in ampere altra, we can not see throughput change
>> compared with I/O fwd engine. However, using record cmd in testpmd:
>> '$set record-burst-stats on'
>> we can see the ratio of 'Rx/Tx burst size of 32' is reduced. This
>> indicate mbufs recycle can save CPU cycles.
>>
>> V2:
>> 1. Use data-plane API to enable direct-rearm (Konstantin, Honnappa)
>> 2. Add 'txq_data_get' API to get txq info for Rx (Konstantin)
>> 3. Use input parameter to enable direct rearm in l3fwd (Konstantin)
>> 4. Add condition detection for direct rearm API (Morten, Andrew Rybchenko)
>>
>> V3:
>> 1. Seperate Rx and Tx operation with two APIs in direct-rearm (Konstantin)
>> 2. Delete L3fwd change for direct rearm (Jerin)
>> 3. enable direct rearm in ixgbe driver in Arm
>>
>> v4:
>> 1. Rename direct-rearm as buffer recycle. Based on this, function name
>> and variable name are changed to let this mode more general for all
>> drivers. (Konstantin, Morten)
>> 2. Add ring wrapping check (Konstantin)
>>
>> v5:
>> 1. some change for ethdev API (Morten)
>> 2. add support for avx2, sse, altivec path
>>
>> v6:
>> 1. fix ixgbe build issue in ppc
>> 2. remove 'recycle_tx_mbufs_reuse' and 'recycle_rx_descriptors_refill'
>> API wrapper (Tech Board meeting)
>> 3. add recycle_mbufs engine in testpmd (Tech Board meeting)
>> 4. add namespace in the functions related to mbufs recycle(Ferruh)
>>
>> v7:
>> 1. move 'rxq/txq data' pointers to the beginning of eth_dev structure,
>> in order to keep them in the same cache line as rx/tx_burst function
>> pointers (Morten)
>> 2. add the extra description for 'rte_eth_recycle_mbufs' to show it can
>> support feeding 1 Rx queue from 2 Tx queues in the same thread
>> (Konstantin)
>> 3. For i40e/ixgbe driver, make the previous copied buffers as invalid if
>> there are Tx buffers refcnt > 1 or from unexpected mempool (Konstantin)
>> 4. add check for the return value of 'rte_eth_recycle_rx_queue_info_get'
>> in testpmd fwd engine (Morten)
>>
>> v8:
>> 1. add arm/x86 build option to fix ixgbe build issue in ppc
>>
>> v9:
>> 1. delete duplicate file name for ixgbe
>>
>> v10:
>> 1. fix compile issue on windows
>>
>> v11:
>> 1. fix doc warning
>>
>> v12:
>> 1. replace rx queue check code with eth_dev_validate_rx_queue
>> function (Stephen)
>> 2. put port and queue check before function call (Konstantin)
>>
>> Feifei Wang (4):
>> ethdev: add API for mbufs recycle mode
>> net/i40e: implement mbufs recycle mode
>> net/ixgbe: implement mbufs recycle mode
>> app/testpmd: add recycle mbufs engine
>>
>
> Thanks for dedication to improve the patchset and finding better
> solution, it is appreciated.
>
> Series applied to dpdk-next-net/main, thanks.
>
Konstantin highlighted that there is an outstanding discussion:
http://patchwork.dpdk.org/project/dpdk/patch/20230822072710.1945027-3-feifei.wang2@arm.com/
Dropping the patchset from next-net, and updating its status in the
patchwork as "Changes Requested".
next prev parent reply other threads:[~2023-09-22 15:30 UTC|newest]
Thread overview: 145+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-20 8:16 [PATCH v1 0/5] Direct re-arming of buffers on receive side Feifei Wang
2022-04-20 8:16 ` [PATCH v1 1/5] net/i40e: remove redundant Dtype initialization Feifei Wang
2022-04-20 8:16 ` [PATCH v1 2/5] net/i40e: enable direct rearm mode Feifei Wang
2022-05-11 22:28 ` Konstantin Ananyev
2022-04-20 8:16 ` [PATCH v1 3/5] ethdev: add API for " Feifei Wang
2022-04-20 9:59 ` Morten Brørup
2022-04-29 2:42 ` 回复: " Feifei Wang
2022-04-20 10:41 ` Andrew Rybchenko
2022-04-29 6:28 ` 回复: " Feifei Wang
2022-05-10 22:49 ` Honnappa Nagarahalli
2022-06-03 10:19 ` Andrew Rybchenko
2022-04-20 10:50 ` Jerin Jacob
2022-05-02 3:09 ` 回复: " Feifei Wang
2022-04-21 14:57 ` Stephen Hemminger
2022-04-29 6:35 ` 回复: " Feifei Wang
2022-04-20 8:16 ` [PATCH v1 4/5] net/i40e: add direct rearm mode internal API Feifei Wang
2022-05-11 22:31 ` Konstantin Ananyev
2022-04-20 8:16 ` [PATCH v1 5/5] examples/l3fwd: enable direct rearm mode Feifei Wang
2022-04-20 10:10 ` Morten Brørup
2022-04-21 2:35 ` Honnappa Nagarahalli
2022-04-21 6:40 ` Morten Brørup
2022-05-10 22:01 ` Honnappa Nagarahalli
2022-05-11 7:17 ` Morten Brørup
2022-05-11 22:33 ` Konstantin Ananyev
2022-05-27 11:28 ` Konstantin Ananyev
2022-05-31 17:14 ` Honnappa Nagarahalli
2022-06-03 10:32 ` Andrew Rybchenko
2022-06-06 11:27 ` Konstantin Ananyev
2022-06-29 21:25 ` Honnappa Nagarahalli
2022-05-11 23:00 ` [PATCH v1 0/5] Direct re-arming of buffers on receive side Konstantin Ananyev
[not found] ` <20220516061012.618787-1-feifei.wang2@arm.com>
2022-05-24 1:25 ` Konstantin Ananyev
2022-05-24 12:40 ` Morten Brørup
2022-05-24 20:14 ` Honnappa Nagarahalli
2022-05-28 12:22 ` Konstantin Ananyev
2022-06-01 1:00 ` Honnappa Nagarahalli
2022-06-03 23:32 ` Konstantin Ananyev
2022-06-04 8:07 ` Morten Brørup
2022-06-29 21:58 ` Honnappa Nagarahalli
2022-06-30 15:21 ` Morten Brørup
2022-07-01 19:30 ` Honnappa Nagarahalli
2022-07-01 20:28 ` Morten Brørup
2022-06-13 5:55 ` 回复: " Feifei Wang
2023-01-04 7:30 ` [PATCH v3 0/3] " Feifei Wang
2023-01-04 7:30 ` [PATCH v3 1/3] ethdev: enable direct rearm with separate API Feifei Wang
2023-01-04 8:21 ` Morten Brørup
2023-01-04 8:51 ` 回复: " Feifei Wang
2023-01-04 10:11 ` Morten Brørup
2023-02-24 8:55 ` 回复: " Feifei Wang
2023-03-06 12:49 ` Ferruh Yigit
2023-03-06 13:26 ` Morten Brørup
2023-03-06 14:53 ` 回复: " Feifei Wang
2023-03-06 15:02 ` Ferruh Yigit
2023-03-07 6:12 ` Honnappa Nagarahalli
2023-03-07 10:52 ` Konstantin Ananyev
2023-03-07 20:41 ` Ferruh Yigit
2023-03-22 14:43 ` Honnappa Nagarahalli
2023-02-02 14:33 ` Konstantin Ananyev
2023-02-24 9:45 ` 回复: " Feifei Wang
2023-02-27 19:31 ` Konstantin Ananyev
2023-02-28 2:16 ` 回复: " Feifei Wang
2023-02-28 8:09 ` Morten Brørup
2023-03-01 7:34 ` 回复: " Feifei Wang
2023-01-04 7:30 ` [PATCH v3 2/3] net/i40e: " Feifei Wang
2023-02-02 14:37 ` Konstantin Ananyev
2023-02-24 9:50 ` 回复: " Feifei Wang
2023-02-27 19:35 ` Konstantin Ananyev
2023-02-28 2:15 ` 回复: " Feifei Wang
2023-03-07 11:01 ` Konstantin Ananyev
2023-03-14 6:07 ` 回复: " Feifei Wang
2023-03-19 16:11 ` Konstantin Ananyev
2023-03-23 10:49 ` Feifei Wang
2023-01-04 7:30 ` [PATCH v3 3/3] net/ixgbe: " Feifei Wang
2023-01-31 6:13 ` 回复: [PATCH v3 0/3] Direct re-arming of buffers on receive side Feifei Wang
2023-02-01 1:10 ` Konstantin Ananyev
2023-02-01 2:24 ` 回复: " Feifei Wang
2023-03-22 12:56 ` Morten Brørup
2023-03-22 13:41 ` Honnappa Nagarahalli
2023-03-22 14:04 ` Morten Brørup
2023-08-02 7:38 ` [PATCH v8 0/4] Recycle mbufs from Tx queue into Rx queue Feifei Wang
2023-08-02 7:38 ` [PATCH v8 1/4] ethdev: add API for mbufs recycle mode Feifei Wang
2023-08-02 7:38 ` [PATCH v8 2/4] net/i40e: implement " Feifei Wang
2023-08-02 7:38 ` [PATCH v8 3/4] net/ixgbe: " Feifei Wang
2023-08-02 7:38 ` [PATCH v8 4/4] app/testpmd: add recycle mbufs engine Feifei Wang
2023-08-02 8:08 ` [PATCH v9 0/4] Recycle mbufs from Tx queue into Rx queue Feifei Wang
2023-08-02 8:08 ` [PATCH v9 1/4] ethdev: add API for mbufs recycle mode Feifei Wang
2023-08-02 8:08 ` [PATCH v9 2/4] net/i40e: implement " Feifei Wang
2023-08-02 8:08 ` [PATCH v9 3/4] net/ixgbe: " Feifei Wang
2023-08-02 8:08 ` [PATCH v9 4/4] app/testpmd: add recycle mbufs engine Feifei Wang
2023-08-04 9:24 ` [PATCH v10 0/4] Recycle mbufs from Tx queue into Rx queue Feifei Wang
2023-08-04 9:24 ` [PATCH v10 1/4] ethdev: add API for mbufs recycle mode Feifei Wang
2023-08-04 9:24 ` [PATCH v10 2/4] net/i40e: implement " Feifei Wang
2023-08-04 9:24 ` [PATCH v10 3/4] net/ixgbe: " Feifei Wang
2023-08-04 9:24 ` [PATCH v10 4/4] app/testpmd: add recycle mbufs engine Feifei Wang
2023-08-22 7:27 ` [PATCH v11 0/4] Recycle mbufs from Tx queue into Rx queue Feifei Wang
2023-08-22 7:27 ` [PATCH v11 1/4] ethdev: add API for mbufs recycle mode Feifei Wang
2023-08-22 14:02 ` Stephen Hemminger
2023-08-24 3:16 ` Feifei Wang
2023-08-22 23:33 ` Konstantin Ananyev
2023-08-24 3:38 ` Feifei Wang
2023-08-22 7:27 ` [PATCH v11 2/4] net/i40e: implement " Feifei Wang
2023-08-22 23:43 ` Konstantin Ananyev
2023-08-24 6:10 ` Feifei Wang
2023-08-31 17:24 ` Konstantin Ananyev
2023-08-31 23:49 ` Konstantin Ananyev
2023-09-01 12:22 ` Feifei Wang
2023-09-01 14:22 ` Konstantin Ananyev
2023-09-04 6:59 ` Feifei Wang
2023-09-04 7:49 ` Konstantin Ananyev
2023-09-04 9:24 ` Feifei Wang
2023-09-04 10:21 ` Konstantin Ananyev
2023-09-05 3:11 ` Feifei Wang
2023-09-22 14:58 ` Feifei Wang
2023-09-22 15:46 ` Feifei Wang
2023-09-22 16:40 ` Konstantin Ananyev
2023-09-23 5:52 ` Feifei Wang
2023-09-23 20:40 ` Konstantin Ananyev
2023-09-25 3:26 ` Feifei Wang
2023-08-22 7:27 ` [PATCH v11 3/4] net/ixgbe: " Feifei Wang
2023-08-22 7:27 ` [PATCH v11 4/4] app/testpmd: add recycle mbufs engine Feifei Wang
2023-08-22 7:33 ` [PATCH v11 0/4] Recycle mbufs from Tx queue into Rx queue Feifei Wang
2023-08-22 13:59 ` Stephen Hemminger
2023-08-24 3:11 ` Feifei Wang
2023-08-24 7:36 ` [PATCH v12 " Feifei Wang
2023-08-24 7:36 ` [PATCH v12 1/4] ethdev: add API for mbufs recycle mode Feifei Wang
2023-08-31 9:16 ` Feifei Wang
2023-09-20 13:10 ` Ferruh Yigit
2023-08-24 7:36 ` [PATCH v12 2/4] net/i40e: implement " Feifei Wang
2023-08-24 7:36 ` [PATCH v12 3/4] net/ixgbe: " Feifei Wang
2023-08-24 7:36 ` [PATCH v12 4/4] app/testpmd: add recycle mbufs engine Feifei Wang
2023-09-20 13:11 ` Ferruh Yigit
2023-09-20 13:12 ` [PATCH v12 0/4] Recycle mbufs from Tx queue into Rx queue Ferruh Yigit
2023-09-22 15:30 ` Ferruh Yigit [this message]
2023-09-25 3:19 ` [PATCH v13 " Feifei Wang
2023-09-25 3:19 ` [PATCH v13 1/4] ethdev: add API for mbufs recycle mode Feifei Wang
2023-09-25 4:40 ` Ajit Khaparde
2023-09-25 3:19 ` [PATCH v13 2/4] net/i40e: implement " Feifei Wang
2023-09-26 8:26 ` Ferruh Yigit
2023-09-26 8:56 ` Konstantin Ananyev
2023-09-26 13:34 ` Konstantin Ananyev
2023-09-25 3:19 ` [PATCH v13 3/4] net/ixgbe: " Feifei Wang
2023-09-26 13:30 ` Konstantin Ananyev
2023-09-25 3:19 ` [PATCH v13 4/4] app/testpmd: add recycle mbufs engine Feifei Wang
2023-09-26 13:30 ` Konstantin Ananyev
2023-09-26 16:38 ` Ajit Khaparde
2023-09-27 17:24 ` [PATCH v13 0/4] Recycle mbufs from Tx queue into Rx queue Ferruh Yigit
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=eb0db252-943d-4a45-96d4-a94a5037a4d7@amd.com \
--to=ferruh.yigit@amd.com \
--cc=Honnappa.Nagarahalli@arm.com \
--cc=dev@dpdk.org \
--cc=feifei.wang2@arm.com \
--cc=konstantin.v.ananyev@yandex.ru \
--cc=nd@arm.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).