From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 6AA64A034F; Mon, 7 Jun 2021 20:06:07 +0200 (CEST) Received: from [217.70.189.124] (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 4B3FA41101; Mon, 7 Jun 2021 20:04:14 +0200 (CEST) Received: from mx0b-0016f401.pphosted.com (mx0b-0016f401.pphosted.com [67.231.156.173]) by mails.dpdk.org (Postfix) with ESMTP id 5FFC741101 for ; Mon, 7 Jun 2021 20:04:13 +0200 (CEST) Received: from pps.filterd (m0045851.ppops.net [127.0.0.1]) by mx0b-0016f401.pphosted.com (8.16.0.43/8.16.0.43) with SMTP id 157I1bEh017524 for ; Mon, 7 Jun 2021 11:04:12 -0700 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=marvell.com; h=from : to : cc : subject : date : message-id : in-reply-to : references : mime-version : content-type; s=pfpt0220; bh=1VnMYFemqvQ1PdQIsy1NOOqm5jM/HfJ/4k4PTzgjado=; b=QTL8td9ab/P/seiziGVqeIVEZsIXPC44GRwXnAjlUI7si7upKKKQJ47lyhTaOOWvgQ0v MVrB35qKYUVPUMXJEYO4yVm0MreDzBYUlFvIviVroM9Y4nmvoVpWJfDZcpE83VRB9obT 9G9ZgDl0TtIclPdcDUikmPBeayQYIimtVAIyx96sDMgolgKZKM7k+0uMXhGrli3trsqQ xDrl1r3VJ1v8zCW/kAGorDnE7BJCeCQPvdHixGLijsZKU0AMuYSQN9vhg2l3Ra0i6UTW I/1wn7Dmv8Djvl9B9PGzzKtGm+yK1l9x6r5Mcv7NgPB7arW2f7HGWY+lDg6z7S2DVYcu 0Q== Received: from dc5-exch01.marvell.com ([199.233.59.181]) by mx0b-0016f401.pphosted.com with ESMTP id 391ecv2ef4-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT) for ; Mon, 07 Jun 2021 11:04:12 -0700 Received: from DC5-EXCH01.marvell.com (10.69.176.38) by DC5-EXCH01.marvell.com (10.69.176.38) with Microsoft SMTP Server (TLS) id 15.0.1497.2; Mon, 7 Jun 2021 11:04:10 -0700 Received: from maili.marvell.com (10.69.176.80) by DC5-EXCH01.marvell.com (10.69.176.38) with Microsoft SMTP Server id 15.0.1497.2 via Frontend Transport; Mon, 7 Jun 2021 11:04:10 -0700 Received: from hyd1588t430.marvell.com (unknown [10.29.52.204]) by maili.marvell.com (Postfix) with ESMTP id B55753F7041; Mon, 7 Jun 2021 11:04:07 -0700 (PDT) From: Nithin Dabilpuram To: CC: , , , , , , , Nithin Dabilpuram Date: Mon, 7 Jun 2021 23:29:04 +0530 Message-ID: <20210607175943.31690-24-ndabilpuram@marvell.com> X-Mailer: git-send-email 2.8.4 In-Reply-To: <20210607175943.31690-1-ndabilpuram@marvell.com> References: <20210306153404.10781-1-ndabilpuram@marvell.com> <20210607175943.31690-1-ndabilpuram@marvell.com> MIME-Version: 1.0 Content-Type: text/plain X-Proofpoint-GUID: EwEXUzYJPeqeh0i29aeueIUJtkNsM9So X-Proofpoint-ORIG-GUID: EwEXUzYJPeqeh0i29aeueIUJtkNsM9So X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:6.0.391, 18.0.761 definitions=2021-06-07_14:2021-06-04, 2021-06-07 signatures=0 Subject: [dpdk-dev] [PATCH v2 23/62] net/cnxk: add Tx multi-segment version for cn9k X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Add Tx burst multi-segment version for CN9K. Signed-off-by: Nithin Dabilpuram Signed-off-by: Pavan Nikhilesh --- drivers/net/cnxk/cn9k_tx.c | 14 ++++ drivers/net/cnxk/cn9k_tx.h | 147 ++++++++++++++++++++++++++++++++++++++++ drivers/net/cnxk/cn9k_tx_mseg.c | 25 +++++++ drivers/net/cnxk/cnxk_ethdev.h | 4 ++ drivers/net/cnxk/meson.build | 3 +- 5 files changed, 192 insertions(+), 1 deletion(-) create mode 100644 drivers/net/cnxk/cn9k_tx_mseg.c diff --git a/drivers/net/cnxk/cn9k_tx.c b/drivers/net/cnxk/cn9k_tx.c index a0b022a..8f1d5f5 100644 --- a/drivers/net/cnxk/cn9k_tx.c +++ b/drivers/net/cnxk/cn9k_tx.c @@ -21,6 +21,7 @@ NIX_TX_FASTPATH_MODES #undef T + static inline void pick_tx_func(struct rte_eth_dev *eth_dev, const eth_tx_burst_t tx_burst[2][2][2][2][2]) @@ -39,6 +40,8 @@ pick_tx_func(struct rte_eth_dev *eth_dev, void cn9k_eth_set_tx_function(struct rte_eth_dev *eth_dev) { + struct cnxk_eth_dev *dev = cnxk_eth_pmd_priv(eth_dev); + const eth_tx_burst_t nix_eth_tx_burst[2][2][2][2][2] = { #define T(name, f4, f3, f2, f1, f0, sz, flags) \ [f4][f3][f2][f1][f0] = cn9k_nix_xmit_pkts_##name, @@ -47,7 +50,18 @@ cn9k_eth_set_tx_function(struct rte_eth_dev *eth_dev) #undef T }; + const eth_tx_burst_t nix_eth_tx_burst_mseg[2][2][2][2][2] = { +#define T(name, f4, f3, f2, f1, f0, sz, flags) \ + [f4][f3][f2][f1][f0] = cn9k_nix_xmit_pkts_mseg_##name, + + NIX_TX_FASTPATH_MODES +#undef T + }; + pick_tx_func(eth_dev, nix_eth_tx_burst); + if (dev->tx_offloads & DEV_TX_OFFLOAD_MULTI_SEGS) + pick_tx_func(eth_dev, nix_eth_tx_burst_mseg); + rte_mb(); } diff --git a/drivers/net/cnxk/cn9k_tx.h b/drivers/net/cnxk/cn9k_tx.h index 6d5fc59..239db66 100644 --- a/drivers/net/cnxk/cn9k_tx.h +++ b/drivers/net/cnxk/cn9k_tx.h @@ -309,6 +309,111 @@ cn9k_nix_xmit_submit_lmt_release(const rte_iova_t io_addr) } static __rte_always_inline uint16_t +cn9k_nix_prepare_mseg(struct rte_mbuf *m, uint64_t *cmd, const uint16_t flags) +{ + struct nix_send_hdr_s *send_hdr; + union nix_send_sg_s *sg; + struct rte_mbuf *m_next; + uint64_t *slist, sg_u; + uint64_t nb_segs; + uint64_t segdw; + uint8_t off, i; + + send_hdr = (struct nix_send_hdr_s *)cmd; + send_hdr->w0.total = m->pkt_len; + send_hdr->w0.aura = roc_npa_aura_handle_to_aura(m->pool->pool_id); + + if (flags & NIX_TX_NEED_EXT_HDR) + off = 2; + else + off = 0; + + sg = (union nix_send_sg_s *)&cmd[2 + off]; + /* Clear sg->u header before use */ + sg->u &= 0xFC00000000000000; + sg_u = sg->u; + slist = &cmd[3 + off]; + + i = 0; + nb_segs = m->nb_segs; + + /* Fill mbuf segments */ + do { + m_next = m->next; + sg_u = sg_u | ((uint64_t)m->data_len << (i << 4)); + *slist = rte_mbuf_data_iova(m); + /* Set invert df if buffer is not to be freed by H/W */ + if (flags & NIX_TX_OFFLOAD_MBUF_NOFF_F) { + sg_u |= (cnxk_nix_prefree_seg(m) << (i + 55)); + /* Commit changes to mbuf */ + rte_io_wmb(); + } + /* Mark mempool object as "put" since it is freed by NIX */ +#ifdef RTE_LIBRTE_MEMPOOL_DEBUG + if (!(sg_u & (1ULL << (i + 55)))) + __mempool_check_cookies(m->pool, (void **)&m, 1, 0); + rte_io_wmb(); +#endif + slist++; + i++; + nb_segs--; + if (i > 2 && nb_segs) { + i = 0; + /* Next SG subdesc */ + *(uint64_t *)slist = sg_u & 0xFC00000000000000; + sg->u = sg_u; + sg->segs = 3; + sg = (union nix_send_sg_s *)slist; + sg_u = sg->u; + slist++; + } + m = m_next; + } while (nb_segs); + + sg->u = sg_u; + sg->segs = i; + segdw = (uint64_t *)slist - (uint64_t *)&cmd[2 + off]; + /* Roundup extra dwords to multiple of 2 */ + segdw = (segdw >> 1) + (segdw & 0x1); + /* Default dwords */ + segdw += (off >> 1) + 1; + send_hdr->w0.sizem1 = segdw - 1; + + return segdw; +} + +static __rte_always_inline void +cn9k_nix_xmit_mseg_prep_lmt(uint64_t *cmd, void *lmt_addr, uint16_t segdw) +{ + roc_lmt_mov_seg(lmt_addr, (const void *)cmd, segdw); +} + +static __rte_always_inline void +cn9k_nix_xmit_mseg_one(uint64_t *cmd, void *lmt_addr, rte_iova_t io_addr, + uint16_t segdw) +{ + uint64_t lmt_status; + + do { + roc_lmt_mov_seg(lmt_addr, (const void *)cmd, segdw); + lmt_status = roc_lmt_submit_ldeor(io_addr); + } while (lmt_status == 0); +} + +static __rte_always_inline void +cn9k_nix_xmit_mseg_one_release(uint64_t *cmd, void *lmt_addr, + rte_iova_t io_addr, uint16_t segdw) +{ + uint64_t lmt_status; + + rte_io_wmb(); + do { + roc_lmt_mov_seg(lmt_addr, (const void *)cmd, segdw); + lmt_status = roc_lmt_submit_ldeor(io_addr); + } while (lmt_status == 0); +} + +static __rte_always_inline uint16_t cn9k_nix_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t pkts, uint64_t *cmd, const uint16_t flags) { @@ -344,6 +449,45 @@ cn9k_nix_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t pkts, return pkts; } +static __rte_always_inline uint16_t +cn9k_nix_xmit_pkts_mseg(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t pkts, uint64_t *cmd, const uint16_t flags) +{ + struct cn9k_eth_txq *txq = tx_queue; + uint64_t i; + const rte_iova_t io_addr = txq->io_addr; + void *lmt_addr = txq->lmt_addr; + uint16_t segdw; + + NIX_XMIT_FC_OR_RETURN(txq, pkts); + + roc_lmt_mov(cmd, &txq->cmd[0], cn9k_nix_tx_ext_subs(flags)); + + /* Perform header writes before barrier for TSO */ + if (flags & NIX_TX_OFFLOAD_TSO_F) { + for (i = 0; i < pkts; i++) + cn9k_nix_xmit_prepare_tso(tx_pkts[i], flags); + } + + /* Lets commit any changes in the packet here as no further changes + * to the packet will be done unless no fast free is enabled. + */ + if (!(flags & NIX_TX_OFFLOAD_MBUF_NOFF_F)) + rte_io_wmb(); + + for (i = 0; i < pkts; i++) { + cn9k_nix_xmit_prepare(tx_pkts[i], cmd, flags); + segdw = cn9k_nix_prepare_mseg(tx_pkts[i], cmd, flags); + cn9k_nix_xmit_mseg_one(cmd, lmt_addr, io_addr, segdw); + } + + /* Reduce the cached count */ + txq->fc_cache_pkts -= pkts; + + return pkts; +} + + #define L3L4CSUM_F NIX_TX_OFFLOAD_L3_L4_CSUM_F #define OL3OL4CSUM_F NIX_TX_OFFLOAD_OL3_OL4_CSUM_F #define VLAN_F NIX_TX_OFFLOAD_VLAN_QINQ_F @@ -419,6 +563,9 @@ T(tso_noff_vlan_ol3ol4csum_l3l4csum, 1, 1, 1, 1, 1, 6, \ #define T(name, f4, f3, f2, f1, f0, sz, flags) \ uint16_t __rte_noinline __rte_hot cn9k_nix_xmit_pkts_##name( \ + void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t pkts); \ + \ + uint16_t __rte_noinline __rte_hot cn9k_nix_xmit_pkts_mseg_##name( \ void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t pkts); NIX_TX_FASTPATH_MODES diff --git a/drivers/net/cnxk/cn9k_tx_mseg.c b/drivers/net/cnxk/cn9k_tx_mseg.c new file mode 100644 index 0000000..65c5f36 --- /dev/null +++ b/drivers/net/cnxk/cn9k_tx_mseg.c @@ -0,0 +1,25 @@ +/* SPDX-License-Identifier: BSD-3-Clause + * Copyright(C) 2021 Marvell. + */ + +#include "cn9k_ethdev.h" +#include "cn9k_tx.h" + +#define T(name, f4, f3, f2, f1, f0, sz, flags) \ + uint16_t __rte_noinline __rte_hot \ + cn9k_nix_xmit_pkts_mseg_##name(void *tx_queue, \ + struct rte_mbuf **tx_pkts, \ + uint16_t pkts) \ + { \ + uint64_t cmd[(sz) + CNXK_NIX_TX_MSEG_SG_DWORDS - 2]; \ + \ + /* For TSO inner checksum is a must */ \ + if (((flags) & NIX_TX_OFFLOAD_TSO_F) && \ + !((flags) & NIX_TX_OFFLOAD_L3_L4_CSUM_F)) \ + return 0; \ + return cn9k_nix_xmit_pkts_mseg(tx_queue, tx_pkts, pkts, cmd, \ + (flags) | NIX_TX_MULTI_SEG_F); \ + } + +NIX_TX_FASTPATH_MODES +#undef T diff --git a/drivers/net/cnxk/cnxk_ethdev.h b/drivers/net/cnxk/cnxk_ethdev.h index 8facafc..2f31cba 100644 --- a/drivers/net/cnxk/cnxk_ethdev.h +++ b/drivers/net/cnxk/cnxk_ethdev.h @@ -46,6 +46,10 @@ #define CNXK_NIX_TX_NB_SEG_MAX 9 #endif +#define CNXK_NIX_TX_MSEG_SG_DWORDS \ + ((RTE_ALIGN_MUL_CEIL(CNXK_NIX_TX_NB_SEG_MAX, 3) / 3) + \ + CNXK_NIX_TX_NB_SEG_MAX) + #define CNXK_NIX_RSS_L3_L4_SRC_DST \ (ETH_RSS_L3_SRC_ONLY | ETH_RSS_L3_DST_ONLY | ETH_RSS_L4_SRC_ONLY | \ ETH_RSS_L4_DST_ONLY) diff --git a/drivers/net/cnxk/meson.build b/drivers/net/cnxk/meson.build index 2c1ce72..c727a09 100644 --- a/drivers/net/cnxk/meson.build +++ b/drivers/net/cnxk/meson.build @@ -19,7 +19,8 @@ sources += files('cn9k_ethdev.c', 'cn9k_rx.c', 'cn9k_rx_mseg.c', 'cn9k_rx_vec.c', - 'cn9k_tx.c') + 'cn9k_tx.c', + 'cn9k_tx_mseg.c') # CN10K sources += files('cn10k_ethdev.c') -- 2.8.4