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 59591489F5; Tue, 28 Oct 2025 06:30:03 +0100 (CET) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id B756B40613; Tue, 28 Oct 2025 06:29:53 +0100 (CET) Received: from mgamail.intel.com (mgamail.intel.com [198.175.65.9]) by mails.dpdk.org (Postfix) with ESMTP id 91513402A6 for ; Tue, 28 Oct 2025 06:29:51 +0100 (CET) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1761629391; x=1793165391; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=2gmQYoUJLLvJ5FGys7t1SBuA8c6duuWUnx0SBqFKeZg=; b=HkDhXM6Yzl6S8ravFmuQcU0c/znkkYoR8DDxgCD+GMKzM8c4v1JnJWPk N27SX8pA2fMR3JFHWjo7mZaYahmbeMzuGOnjtj9OgHmECTuVTcVMQnASi qX4pK9oD2giHaPr3TBqG5bg6k9j/R2HIX5EZHRUFfbeWV3sXtCupfF4O5 Kvf16fk5vAHOcW+5gOFL5GV7fDVXRrkjCkeZBvplZwTDNPS2xq4DAm7qs yywiGmnlgqVsR7BHvrO1bNqcx6Hw5A6LSS+h+08LzWqaveDXaG+k5mO2H KJAFejppTnZ6rU16bToKcuSCyFChtx8mtInRarWHkTUW8SASQYOcpU+at Q==; X-CSE-ConnectionGUID: iGz0Z26hRBSFco55kHU1Yw== X-CSE-MsgGUID: zy0wlybDRn2xSwxLLVjbRA== X-IronPort-AV: E=McAfee;i="6800,10657,11586"; a="86346432" X-IronPort-AV: E=Sophos;i="6.19,260,1754982000"; d="scan'208";a="86346432" Received: from orviesa002.jf.intel.com ([10.64.159.142]) by orvoesa101.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 27 Oct 2025 22:29:51 -0700 X-CSE-ConnectionGUID: v9fpY1pDRCK9CfskZ06J5g== X-CSE-MsgGUID: qBIT2qXOTKqkCSdEVtETeA== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.19,260,1754982000"; d="scan'208";a="215904151" Received: from unknown (HELO srv24..) ([10.138.182.231]) by orviesa002.jf.intel.com with ESMTP; 27 Oct 2025 22:29:50 -0700 From: Shaiq Wani To: dev@dpdk.org, bruce.richardson@intel.com Cc: aman.deep.singh@intel.com Subject: [PATCH v9 2/3] net/idpf: enable AVX2 for split queue Tx Date: Tue, 28 Oct 2025 10:59:42 +0530 Message-Id: <20251028052943.121453-3-shaiq.wani@intel.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20251028052943.121453-1-shaiq.wani@intel.com> References: <20250917052658.582872-1-shaiq.wani@intel.com> <20251028052943.121453-1-shaiq.wani@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 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 In case some CPUs don't support AVX512. Enable AVX2 for them to get better per-core performance. In the single queue model, the same descriptor queue is used by SW to post descriptors to the device and used by device to report completed descriptors to SW. While as the split queue model separates them into different queues for parallel processing and improved performance. Signed-off-by: Shaiq Wani --- drivers/net/intel/idpf/idpf_common_rxtx.h | 3 + .../net/intel/idpf/idpf_common_rxtx_avx2.c | 178 +++++++++++++++++- drivers/net/intel/idpf/idpf_rxtx.c | 9 + 3 files changed, 188 insertions(+), 2 deletions(-) diff --git a/drivers/net/intel/idpf/idpf_common_rxtx.h b/drivers/net/intel/idpf/idpf_common_rxtx.h index 87f6895c4c..3636d55272 100644 --- a/drivers/net/intel/idpf/idpf_common_rxtx.h +++ b/drivers/net/intel/idpf/idpf_common_rxtx.h @@ -264,6 +264,9 @@ uint16_t idpf_dp_singleq_recv_pkts_avx2(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts); __rte_internal +uint16_t idpf_dp_splitq_xmit_pkts_avx2(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts); +__rte_internal uint16_t idpf_dp_singleq_xmit_pkts_avx2(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts); diff --git a/drivers/net/intel/idpf/idpf_common_rxtx_avx2.c b/drivers/net/intel/idpf/idpf_common_rxtx_avx2.c index e78a453a8b..feb98eceff 100644 --- a/drivers/net/intel/idpf/idpf_common_rxtx_avx2.c +++ b/drivers/net/intel/idpf/idpf_common_rxtx_avx2.c @@ -679,7 +679,7 @@ idpf_singleq_xmit_fixed_burst_vec_avx2(void *tx_queue, struct rte_mbuf **tx_pkts struct ci_tx_queue *txq = (struct ci_tx_queue *)tx_queue; volatile struct idpf_base_tx_desc *txdp; struct ci_tx_entry_vec *txep; - uint16_t n, nb_commit, tx_id; + uint16_t n, nb_commit; uint64_t flags = IDPF_TX_DESC_CMD_EOP; uint64_t rs = IDPF_TX_DESC_CMD_RS | flags; @@ -693,7 +693,7 @@ idpf_singleq_xmit_fixed_burst_vec_avx2(void *tx_queue, struct rte_mbuf **tx_pkts if (unlikely(nb_pkts == 0)) return 0; - tx_id = txq->tx_tail; + uint16_t tx_id = txq->tx_tail; txdp = &txq->idpf_tx_ring[tx_id]; txep = &txq->sw_ring_vec[tx_id]; @@ -761,3 +761,177 @@ idpf_dp_singleq_xmit_pkts_avx2(void *tx_queue, struct rte_mbuf **tx_pkts, return nb_tx; } + +static __rte_always_inline void +idpf_splitq_scan_cq_ring(struct ci_tx_queue *cq) +{ + struct idpf_splitq_tx_compl_desc *compl_ring; + struct ci_tx_queue *txq; + uint16_t genid, txq_qid, cq_qid, i; + uint8_t ctype; + + cq_qid = cq->tx_tail; + + for (i = 0; i < IDPD_TXQ_SCAN_CQ_THRESH; i++) { + if (cq_qid == cq->nb_tx_desc) { + cq_qid = 0; + cq->expected_gen_id ^= 1; /* toggle generation bit */ + } + + compl_ring = &cq->compl_ring[cq_qid]; + + genid = (rte_le_to_cpu_16(compl_ring->qid_comptype_gen) & + IDPF_TXD_COMPLQ_GEN_M) >> IDPF_TXD_COMPLQ_GEN_S; + + if (genid != cq->expected_gen_id) + break; + + ctype = (rte_le_to_cpu_16(compl_ring->qid_comptype_gen) & + IDPF_TXD_COMPLQ_COMPL_TYPE_M) >> IDPF_TXD_COMPLQ_COMPL_TYPE_S; + txq_qid = (rte_le_to_cpu_16(compl_ring->qid_comptype_gen) & + IDPF_TXD_COMPLQ_QID_M) >> IDPF_TXD_COMPLQ_QID_S; + + txq = cq->txqs[txq_qid - cq->tx_start_qid]; + if (ctype == IDPF_TXD_COMPLT_RS) + txq->rs_compl_count++; + + cq_qid++; + } + + cq->tx_tail = cq_qid; +} + +static __rte_always_inline void +idpf_splitq_vtx1_avx2(struct idpf_flex_tx_sched_desc *txdp, + struct rte_mbuf *pkt, uint64_t flags) +{ + uint64_t high_qw = + IDPF_TX_DESC_DTYPE_FLEX_FLOW_SCHE | + ((uint64_t)flags) | + ((uint64_t)pkt->data_len << IDPF_TXD_QW1_TX_BUF_SZ_S); + + __m128i descriptor = _mm_set_epi64x(high_qw, + pkt->buf_iova + pkt->data_off); + _mm_storeu_si128((__m128i *)txdp, descriptor); +} + +static inline void +idpf_splitq_vtx_avx2(struct idpf_flex_tx_sched_desc *txdp, + struct rte_mbuf **pkt, uint16_t nb_pkts, uint64_t flags) +{ + const uint64_t hi_qw_tmpl = IDPF_TX_DESC_DTYPE_FLEX_FLOW_SCHE | ((uint64_t)flags); + + /* align if needed */ + if (((uintptr_t)txdp & 0x1F) != 0 && nb_pkts != 0) { + idpf_splitq_vtx1_avx2(txdp, *pkt, flags); + txdp++, pkt++, nb_pkts--; + } + + for (; nb_pkts >= IDPF_VPMD_DESCS_PER_LOOP; txdp += IDPF_VPMD_DESCS_PER_LOOP, + pkt += IDPF_VPMD_DESCS_PER_LOOP, nb_pkts -= IDPF_VPMD_DESCS_PER_LOOP) { + uint64_t hi_qw0 = hi_qw_tmpl | + ((uint64_t)pkt[0]->data_len << IDPF_TXD_QW1_TX_BUF_SZ_S); + uint64_t hi_qw1 = hi_qw_tmpl | + ((uint64_t)pkt[1]->data_len << IDPF_TXD_QW1_TX_BUF_SZ_S); + uint64_t hi_qw2 = hi_qw_tmpl | + ((uint64_t)pkt[2]->data_len << IDPF_TXD_QW1_TX_BUF_SZ_S); + uint64_t hi_qw3 = hi_qw_tmpl | + ((uint64_t)pkt[3]->data_len << IDPF_TXD_QW1_TX_BUF_SZ_S); + + __m256i desc0_1 = _mm256_set_epi64x(hi_qw1, + pkt[1]->buf_iova + pkt[1]->data_off, + hi_qw0, + pkt[0]->buf_iova + pkt[0]->data_off); + __m256i desc2_3 = _mm256_set_epi64x(hi_qw3, + pkt[3]->buf_iova + pkt[3]->data_off, + hi_qw2, + pkt[2]->buf_iova + pkt[2]->data_off); + + _mm256_storeu_si256((__m256i *)(txdp + 0), desc0_1); + _mm256_storeu_si256((__m256i *)(txdp + 2), desc2_3); + } + + while (nb_pkts--) { + idpf_splitq_vtx1_avx2(txdp, *pkt, flags); + txdp++; + pkt++; + } +} + +static inline uint16_t +idpf_splitq_xmit_fixed_burst_vec_avx2(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) +{ + struct ci_tx_queue *txq = (struct ci_tx_queue *)tx_queue; + struct idpf_flex_tx_sched_desc *txdp; + struct ci_tx_entry_vec *txep; + uint16_t n, nb_commit; + uint64_t cmd_dtype = IDPF_TXD_FLEX_FLOW_CMD_EOP; + uint16_t tx_id = txq->tx_tail; + + nb_commit = (uint16_t)RTE_MIN(txq->nb_tx_free, nb_pkts); + nb_pkts = nb_commit; + if (unlikely(nb_pkts == 0)) + return 0; + + txdp = (struct idpf_flex_tx_sched_desc *)&txq->desc_ring[tx_id]; + txep = &txq->sw_ring_vec[tx_id]; + + txq->nb_tx_free = (uint16_t)(txq->nb_tx_free - nb_pkts); + + n = (uint16_t)(txq->nb_tx_desc - tx_id); + if (nb_commit >= n) { + ci_tx_backlog_entry_vec(txep, tx_pkts, n); + + idpf_splitq_vtx_avx2(txdp, tx_pkts, n - 1, cmd_dtype); + tx_pkts += (n - 1); + txdp += (n - 1); + + idpf_splitq_vtx1_avx2(txdp, *tx_pkts++, cmd_dtype); + + nb_commit = (uint16_t)(nb_commit - n); + tx_id = 0; + + txdp = &txq->desc_ring[tx_id]; + txep = (void *)txq->sw_ring; + } + + ci_tx_backlog_entry_vec(txep, tx_pkts, nb_commit); + + idpf_splitq_vtx_avx2(txdp, tx_pkts, nb_commit, cmd_dtype); + + tx_id = (uint16_t)(tx_id + nb_commit); + txq->tx_tail = tx_id; + + IDPF_PCI_REG_WRITE(txq->qtx_tail, txq->tx_tail); + + return nb_pkts; +} + +uint16_t +idpf_dp_splitq_xmit_pkts_avx2(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts) +{ + struct ci_tx_queue *txq = (struct ci_tx_queue *)tx_queue; + uint16_t nb_tx = 0; + + while (nb_pkts) { + uint16_t ret, num; + idpf_splitq_scan_cq_ring(txq->complq); + + if (txq->rs_compl_count > txq->tx_free_thresh) { + ci_tx_free_bufs_vec(txq, idpf_tx_desc_done, false); + txq->rs_compl_count -= txq->tx_rs_thresh; + } + + num = (uint16_t)RTE_MIN(nb_pkts, txq->tx_rs_thresh); + ret = idpf_splitq_xmit_fixed_burst_vec_avx2(tx_queue, &tx_pkts[nb_tx], num); + nb_tx += ret; + nb_pkts -= ret; + if (ret < num) + break; + } + + return nb_tx; +} + +RTE_EXPORT_INTERNAL_SYMBOL(idpf_dp_splitq_xmit_pkts_avx2) diff --git a/drivers/net/intel/idpf/idpf_rxtx.c b/drivers/net/intel/idpf/idpf_rxtx.c index 1c725065df..6950fabb49 100644 --- a/drivers/net/intel/idpf/idpf_rxtx.c +++ b/drivers/net/intel/idpf/idpf_rxtx.c @@ -850,6 +850,15 @@ idpf_set_tx_function(struct rte_eth_dev *dev) return; } #endif /* CC_AVX512_SUPPORT */ + if (tx_simd_width == RTE_VECT_SIMD_256) { + PMD_DRV_LOG(NOTICE, + "Using Split AVX2 Vector Tx (port %d).", + dev->data->port_id); + dev->tx_pkt_burst = idpf_dp_splitq_xmit_pkts_avx2; + dev->tx_pkt_prepare = idpf_dp_prep_pkts; + return; + } + } PMD_DRV_LOG(NOTICE, "Using Split Scalar Tx (port %d).", -- 2.34.1