Patch Detail
get:
Show a patch.
patch:
Update a patch.
put:
Update a patch.
GET /api/patches/68211/?format=api
{ "id": 68211, "url": "https://patches.dpdk.org/api/patches/68211/?format=api", "web_url": "https://patches.dpdk.org/project/dpdk/patch/20200411141428.1987768-22-jerinj@marvell.com/", "project": { "id": 1, "url": "https://patches.dpdk.org/api/projects/1/?format=api", "name": "DPDK", "link_name": "dpdk", "list_id": "dev.dpdk.org", "list_email": "dev@dpdk.org", "web_url": "http://core.dpdk.org", "scm_url": "git://dpdk.org/dpdk", "webscm_url": "http://git.dpdk.org/dpdk", "list_archive_url": "https://inbox.dpdk.org/dev", "list_archive_url_format": "https://inbox.dpdk.org/dev/{}", "commit_url_format": "" }, "msgid": "<20200411141428.1987768-22-jerinj@marvell.com>", "list_archive_url": "https://inbox.dpdk.org/dev/20200411141428.1987768-22-jerinj@marvell.com", "date": "2020-04-11T14:14:20", "name": "[v5,21/29] node: ipv4 lookup for x86", "commit_ref": null, "pull_url": null, "state": "accepted", "archived": true, "hash": "800da8689974781700f3d9e4f5af8e10e56f3f8d", "submitter": { "id": 1188, "url": "https://patches.dpdk.org/api/people/1188/?format=api", "name": "Jerin Jacob Kollanukkaran", "email": "jerinj@marvell.com" }, "delegate": { "id": 1, "url": "https://patches.dpdk.org/api/users/1/?format=api", "username": "tmonjalo", "first_name": "Thomas", "last_name": "Monjalon", "email": "thomas@monjalon.net" }, "mbox": "https://patches.dpdk.org/project/dpdk/patch/20200411141428.1987768-22-jerinj@marvell.com/mbox/", "series": [ { "id": 9314, "url": "https://patches.dpdk.org/api/series/9314/?format=api", "web_url": "https://patches.dpdk.org/project/dpdk/list/?series=9314", "date": "2020-04-11T14:13:59", "name": "graph: introduce graph subsystem", "version": 5, "mbox": "https://patches.dpdk.org/series/9314/mbox/" } ], "comments": "https://patches.dpdk.org/api/patches/68211/comments/", "check": "success", "checks": "https://patches.dpdk.org/api/patches/68211/checks/", "tags": {}, "related": [], "headers": { "Return-Path": "<dev-bounces@dpdk.org>", "X-Original-To": "patchwork@inbox.dpdk.org", "Delivered-To": "patchwork@inbox.dpdk.org", "Received": [ "from dpdk.org (dpdk.org [92.243.14.124])\n\tby inbox.dpdk.org (Postfix) with ESMTP id 90447A059F;\n\tSat, 11 Apr 2020 16:18:47 +0200 (CEST)", "from [92.243.14.124] (localhost [127.0.0.1])\n\tby dpdk.org (Postfix) with ESMTP id 3A9CB1C2A2;\n\tSat, 11 Apr 2020 16:16:01 +0200 (CEST)", "from mx0b-0016f401.pphosted.com (mx0a-0016f401.pphosted.com\n [67.231.148.174]) by dpdk.org (Postfix) with ESMTP id 138E11C1DB\n for <dev@dpdk.org>; Sat, 11 Apr 2020 16:15:56 +0200 (CEST)", "from pps.filterd (m0045849.ppops.net [127.0.0.1])\n by mx0a-0016f401.pphosted.com (8.16.0.42/8.16.0.42) with SMTP id\n 03BEFc5U021932; Sat, 11 Apr 2020 07:15:55 -0700", "from sc-exch04.marvell.com ([199.233.58.184])\n by mx0a-0016f401.pphosted.com with ESMTP id 30bb8q8n7d-1\n (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT);\n Sat, 11 Apr 2020 07:15:55 -0700", "from DC5-EXCH02.marvell.com (10.69.176.39) by SC-EXCH04.marvell.com\n (10.93.176.84) with Microsoft SMTP Server (TLS) id 15.0.1497.2;\n Sat, 11 Apr 2020 07:15:53 -0700", "from maili.marvell.com (10.69.176.80) by DC5-EXCH02.marvell.com\n (10.69.176.39) with Microsoft SMTP Server id 15.0.1497.2 via Frontend\n Transport; Sat, 11 Apr 2020 07:15:53 -0700", "from jerin-lab.marvell.com (jerin-lab.marvell.com [10.28.34.14])\n by maili.marvell.com (Postfix) with ESMTP id D14243F7041;\n Sat, 11 Apr 2020 07:15:50 -0700 (PDT)" ], "DKIM-Signature": "v=1; a=rsa-sha256; c=relaxed/relaxed; d=marvell.com;\n h=from : to : cc :\n subject : date : message-id : in-reply-to : references : mime-version :\n content-transfer-encoding : content-type; s=pfpt0818;\n bh=ppapZkT9WLYfYC/95t9n/hIHb72U8ZdNP0vHIA5mO+0=;\n b=eBD+XN+V4dfWh8owWa6dklw2p5z3JoFFXkukdThTbewmlnDzuDy+kZvSXFAyBakSDIad\n W39JFchtxVsoscyRbtp5zq83FB6DSTNweNM+2iRdg/2lPNVR1sdKt4OBui14jRlZIr8k\n rGMTjT0ThCrdEi8Pa8HdhSSUMNsm4qHt/QjVvnKbet/ge3Aj2drM5aj8ePeIAZ44AsXq\n qhDZ3ncSowDFE6/MnASDxHZExlLNk6rwQ1vAuDomkVkY2Q6e7/lpLcgsC+HBH0kFbEJY\n djvcDKSCNgHxkaAoV5KJCLYlpTmLyxOFP9vklIVMsp+9KR9/0jonVZrMHGKqkL3h5tOo Ew==", "From": "<jerinj@marvell.com>", "To": "Nithin Dabilpuram <ndabilpuram@marvell.com>, Pavan Nikhilesh\n <pbhagavatula@marvell.com>", "CC": "<dev@dpdk.org>, <thomas@monjalon.net>, <david.marchand@redhat.com>,\n <mdr@ashroe.eu>, <mattias.ronnblom@ericsson.com>,\n <kirankumark@marvell.com>, <xiao.w.wang@intel.com>, <amo@semihalf.com>", "Date": "Sat, 11 Apr 2020 19:44:20 +0530", "Message-ID": "<20200411141428.1987768-22-jerinj@marvell.com>", "X-Mailer": "git-send-email 2.25.1", "In-Reply-To": "<20200411141428.1987768-1-jerinj@marvell.com>", "References": "<20200405085613.1336841-1-jerinj@marvell.com>\n <20200411141428.1987768-1-jerinj@marvell.com>", "MIME-Version": "1.0", "Content-Transfer-Encoding": "8bit", "Content-Type": "text/plain", "X-Proofpoint-Virus-Version": "vendor=fsecure engine=2.50.10434:6.0.138, 18.0.676\n definitions=2020-04-11_04:2020-04-09,\n 2020-04-11 signatures=0", "Subject": "[dpdk-dev] [PATCH v5 21/29] node: ipv4 lookup for x86", "X-BeenThere": "dev@dpdk.org", "X-Mailman-Version": "2.1.15", "Precedence": "list", "List-Id": "DPDK patches and discussions <dev.dpdk.org>", "List-Unsubscribe": "<https://mails.dpdk.org/options/dev>,\n <mailto:dev-request@dpdk.org?subject=unsubscribe>", "List-Archive": "<http://mails.dpdk.org/archives/dev/>", "List-Post": "<mailto:dev@dpdk.org>", "List-Help": "<mailto:dev-request@dpdk.org?subject=help>", "List-Subscribe": "<https://mails.dpdk.org/listinfo/dev>,\n <mailto:dev-request@dpdk.org?subject=subscribe>", "Errors-To": "dev-bounces@dpdk.org", "Sender": "\"dev\" <dev-bounces@dpdk.org>" }, "content": "From: Pavan Nikhilesh <pbhagavatula@marvell.com>\n\nAdd IPv4 lookup process function for ip4_lookup\nrte_node. This node performs LPM lookup using x86_64\nvector supported RTE_LPM API on every packet received\nand forwards it to a next node that is identified by\nlookup result.\n\nSigned-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>\nSigned-off-by: Nithin Dabilpuram <ndabilpuram@marvell.com>\nSigned-off-by: Kiran Kumar K <kirankumark@marvell.com>\n---\n lib/librte_node/ip4_lookup.c | 2 +\n lib/librte_node/ip4_lookup_sse.h | 244 +++++++++++++++++++++++++++++++\n 2 files changed, 246 insertions(+)\n create mode 100644 lib/librte_node/ip4_lookup_sse.h", "diff": "diff --git a/lib/librte_node/ip4_lookup.c b/lib/librte_node/ip4_lookup.c\nindex 8ec4f0626..39c8d8db0 100644\n--- a/lib/librte_node/ip4_lookup.c\n+++ b/lib/librte_node/ip4_lookup.c\n@@ -30,6 +30,8 @@ struct ip4_lookup_node_main {\n \n #if defined(RTE_MACHINE_CPUFLAG_NEON)\n #include \"ip4_lookup_neon.h\"\n+#elif defined(RTE_ARCH_X86)\n+#include \"ip4_lookup_sse.h\"\n #else\n \n static uint16_t\ndiff --git a/lib/librte_node/ip4_lookup_sse.h b/lib/librte_node/ip4_lookup_sse.h\nnew file mode 100644\nindex 000000000..a071cc591\n--- /dev/null\n+++ b/lib/librte_node/ip4_lookup_sse.h\n@@ -0,0 +1,244 @@\n+/* SPDX-License-Identifier: BSD-3-Clause\n+ * Copyright(C) 2020 Marvell International Ltd.\n+ */\n+\n+#ifndef __INCLUDE_IP4_LOOKUP_SSE_H__\n+#define __INCLUDE_IP4_LOOKUP_SSE_H__\n+\n+/* X86 SSE */\n+static uint16_t\n+ip4_lookup_node_process(struct rte_graph *graph, struct rte_node *node,\n+\t\t\tvoid **objs, uint16_t nb_objs)\n+{\n+\tstruct rte_mbuf *mbuf0, *mbuf1, *mbuf2, *mbuf3, **pkts;\n+\trte_edge_t next0, next1, next2, next3, next_index;\n+\tstruct rte_ipv4_hdr *ipv4_hdr;\n+\tuint32_t ip0, ip1, ip2, ip3;\n+\tvoid **to_next, **from;\n+\tuint16_t last_spec = 0;\n+\tuint16_t n_left_from;\n+\tstruct rte_lpm *lpm;\n+\tuint16_t held = 0;\n+\tuint32_t drop_nh;\n+\trte_xmm_t dst;\n+\t__m128i dip; /* SSE register */\n+\tint rc, i;\n+\n+\t/* Speculative next */\n+\tnext_index = RTE_NODE_IP4_LOOKUP_NEXT_REWRITE;\n+\t/* Drop node */\n+\tdrop_nh = ((uint32_t)RTE_NODE_IP4_LOOKUP_NEXT_PKT_DROP) << 16;\n+\n+\t/* Get socket specific LPM from ctx */\n+\tlpm = *((struct rte_lpm **)node->ctx);\n+\n+\tpkts = (struct rte_mbuf **)objs;\n+\tfrom = objs;\n+\tn_left_from = nb_objs;\n+\n+\tif (n_left_from >= 4) {\n+\t\tfor (i = 0; i < 4; i++)\n+\t\t\trte_prefetch0(rte_pktmbuf_mtod_offset(pkts[i], void *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr)));\n+\t}\n+\n+\t/* Get stream for the speculated next node */\n+\tto_next = rte_node_next_stream_get(graph, node, next_index, nb_objs);\n+\twhile (n_left_from >= 4) {\n+\t\t/* Prefetch next-next mbufs */\n+\t\tif (likely(n_left_from > 11)) {\n+\t\t\trte_prefetch0(pkts[8]);\n+\t\t\trte_prefetch0(pkts[9]);\n+\t\t\trte_prefetch0(pkts[10]);\n+\t\t\trte_prefetch0(pkts[11]);\n+\t\t}\n+\n+\t\t/* Prefetch next mbuf data */\n+\t\tif (likely(n_left_from > 7)) {\n+\t\t\trte_prefetch0(rte_pktmbuf_mtod_offset(pkts[4], void *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr)));\n+\t\t\trte_prefetch0(rte_pktmbuf_mtod_offset(pkts[5], void *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr)));\n+\t\t\trte_prefetch0(rte_pktmbuf_mtod_offset(pkts[6], void *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr)));\n+\t\t\trte_prefetch0(rte_pktmbuf_mtod_offset(pkts[7], void *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr)));\n+\t\t}\n+\n+\t\tmbuf0 = pkts[0];\n+\t\tmbuf1 = pkts[1];\n+\t\tmbuf2 = pkts[2];\n+\t\tmbuf3 = pkts[3];\n+\n+\t\tpkts += 4;\n+\t\tn_left_from -= 4;\n+\n+\t\t/* Extract DIP of mbuf0 */\n+\t\tipv4_hdr = rte_pktmbuf_mtod_offset(mbuf0, struct rte_ipv4_hdr *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr));\n+\t\tip0 = ipv4_hdr->dst_addr;\n+\t\t/* Extract cksum, ttl as ipv4 hdr is in cache */\n+\t\tnode_mbuf_priv1(mbuf0)->cksum = ipv4_hdr->hdr_checksum;\n+\t\tnode_mbuf_priv1(mbuf0)->ttl = ipv4_hdr->time_to_live;\n+\n+\t\t/* Extract DIP of mbuf1 */\n+\t\tipv4_hdr = rte_pktmbuf_mtod_offset(mbuf1, struct rte_ipv4_hdr *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr));\n+\t\tip1 = ipv4_hdr->dst_addr;\n+\t\t/* Extract cksum, ttl as ipv4 hdr is in cache */\n+\t\tnode_mbuf_priv1(mbuf1)->cksum = ipv4_hdr->hdr_checksum;\n+\t\tnode_mbuf_priv1(mbuf1)->ttl = ipv4_hdr->time_to_live;\n+\n+\t\t/* Extract DIP of mbuf2 */\n+\t\tipv4_hdr = rte_pktmbuf_mtod_offset(mbuf2, struct rte_ipv4_hdr *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr));\n+\t\tip2 = ipv4_hdr->dst_addr;\n+\t\t/* Extract cksum, ttl as ipv4 hdr is in cache */\n+\t\tnode_mbuf_priv1(mbuf2)->cksum = ipv4_hdr->hdr_checksum;\n+\t\tnode_mbuf_priv1(mbuf2)->ttl = ipv4_hdr->time_to_live;\n+\n+\t\t/* Extract DIP of mbuf3 */\n+\t\tipv4_hdr = rte_pktmbuf_mtod_offset(mbuf3, struct rte_ipv4_hdr *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr));\n+\t\tip3 = ipv4_hdr->dst_addr;\n+\n+\t\t/* Prepare for lookup x4 */\n+\t\tdip = _mm_set_epi32(ip3, ip2, ip1, ip0);\n+\n+\t\t/* Byte swap 4 IPV4 addresses. */\n+\t\tconst __m128i bswap_mask = _mm_set_epi8(\n+\t\t\t12, 13, 14, 15, 8, 9, 10, 11, 4, 5, 6, 7, 0, 1, 2, 3);\n+\t\tdip = _mm_shuffle_epi8(dip, bswap_mask);\n+\n+\t\t/* Extract cksum, ttl as ipv4 hdr is in cache */\n+\t\tnode_mbuf_priv1(mbuf3)->cksum = ipv4_hdr->hdr_checksum;\n+\t\tnode_mbuf_priv1(mbuf3)->ttl = ipv4_hdr->time_to_live;\n+\n+\t\t/* Perform LPM lookup to get NH and next node */\n+\t\trte_lpm_lookupx4(lpm, dip, dst.u32, drop_nh);\n+\n+\t\t/* Extract next node id and NH */\n+\t\tnode_mbuf_priv1(mbuf0)->nh = dst.u32[0] & 0xFFFF;\n+\t\tnext0 = (dst.u32[0] >> 16);\n+\n+\t\tnode_mbuf_priv1(mbuf1)->nh = dst.u32[1] & 0xFFFF;\n+\t\tnext1 = (dst.u32[1] >> 16);\n+\n+\t\tnode_mbuf_priv1(mbuf2)->nh = dst.u32[2] & 0xFFFF;\n+\t\tnext2 = (dst.u32[2] >> 16);\n+\n+\t\tnode_mbuf_priv1(mbuf3)->nh = dst.u32[3] & 0xFFFF;\n+\t\tnext3 = (dst.u32[3] >> 16);\n+\n+\t\t/* Enqueue four to next node */\n+\t\trte_edge_t fix_spec =\n+\t\t\t(next_index ^ next0) | (next_index ^ next1) |\n+\t\t\t(next_index ^ next2) | (next_index ^ next3);\n+\n+\t\tif (unlikely(fix_spec)) {\n+\t\t\t/* Copy things successfully speculated till now */\n+\t\t\trte_memcpy(to_next, from, last_spec * sizeof(from[0]));\n+\t\t\tfrom += last_spec;\n+\t\t\tto_next += last_spec;\n+\t\t\theld += last_spec;\n+\t\t\tlast_spec = 0;\n+\n+\t\t\t/* Next0 */\n+\t\t\tif (next_index == next0) {\n+\t\t\t\tto_next[0] = from[0];\n+\t\t\t\tto_next++;\n+\t\t\t\theld++;\n+\t\t\t} else {\n+\t\t\t\trte_node_enqueue_x1(graph, node, next0,\n+\t\t\t\t\t\t from[0]);\n+\t\t\t}\n+\n+\t\t\t/* Next1 */\n+\t\t\tif (next_index == next1) {\n+\t\t\t\tto_next[0] = from[1];\n+\t\t\t\tto_next++;\n+\t\t\t\theld++;\n+\t\t\t} else {\n+\t\t\t\trte_node_enqueue_x1(graph, node, next1,\n+\t\t\t\t\t\t from[1]);\n+\t\t\t}\n+\n+\t\t\t/* Next2 */\n+\t\t\tif (next_index == next2) {\n+\t\t\t\tto_next[0] = from[2];\n+\t\t\t\tto_next++;\n+\t\t\t\theld++;\n+\t\t\t} else {\n+\t\t\t\trte_node_enqueue_x1(graph, node, next2,\n+\t\t\t\t\t\t from[2]);\n+\t\t\t}\n+\n+\t\t\t/* Next3 */\n+\t\t\tif (next_index == next3) {\n+\t\t\t\tto_next[0] = from[3];\n+\t\t\t\tto_next++;\n+\t\t\t\theld++;\n+\t\t\t} else {\n+\t\t\t\trte_node_enqueue_x1(graph, node, next3,\n+\t\t\t\t\t\t from[3]);\n+\t\t\t}\n+\n+\t\t\tfrom += 4;\n+\n+\t\t} else {\n+\t\t\tlast_spec += 4;\n+\t\t}\n+\t}\n+\n+\twhile (n_left_from > 0) {\n+\t\tuint32_t next_hop;\n+\n+\t\tmbuf0 = pkts[0];\n+\n+\t\tpkts += 1;\n+\t\tn_left_from -= 1;\n+\n+\t\t/* Extract DIP of mbuf0 */\n+\t\tipv4_hdr = rte_pktmbuf_mtod_offset(mbuf0, struct rte_ipv4_hdr *,\n+\t\t\t\t\t\tsizeof(struct rte_ether_hdr));\n+\t\t/* Extract cksum, ttl as ipv4 hdr is in cache */\n+\t\tnode_mbuf_priv1(mbuf0)->cksum = ipv4_hdr->hdr_checksum;\n+\t\tnode_mbuf_priv1(mbuf0)->ttl = ipv4_hdr->time_to_live;\n+\n+\t\trc = rte_lpm_lookup(lpm, rte_be_to_cpu_32(ipv4_hdr->dst_addr),\n+\t\t\t\t &next_hop);\n+\t\tnext_hop = (rc == 0) ? next_hop : drop_nh;\n+\n+\t\tnode_mbuf_priv1(mbuf0)->nh = next_hop & 0xFFFF;\n+\t\tnext0 = (next_hop >> 16);\n+\n+\t\tif (unlikely(next_index ^ next0)) {\n+\t\t\t/* Copy things successfully speculated till now */\n+\t\t\trte_memcpy(to_next, from, last_spec * sizeof(from[0]));\n+\t\t\tfrom += last_spec;\n+\t\t\tto_next += last_spec;\n+\t\t\theld += last_spec;\n+\t\t\tlast_spec = 0;\n+\n+\t\t\trte_node_enqueue_x1(graph, node, next0, from[0]);\n+\t\t\tfrom += 1;\n+\t\t} else {\n+\t\t\tlast_spec += 1;\n+\t\t}\n+\t}\n+\n+\t/* !!! Home run !!! */\n+\tif (likely(last_spec == nb_objs)) {\n+\t\trte_node_next_stream_move(graph, node, next_index);\n+\t\treturn nb_objs;\n+\t}\n+\n+\theld += last_spec;\n+\t/* Copy things successfully speculated till now */\n+\trte_memcpy(to_next, from, last_spec * sizeof(from[0]));\n+\trte_node_next_stream_put(graph, node, next_index, held);\n+\n+\treturn nb_objs;\n+}\n+\n+#endif /* __INCLUDE_IP4_LOOKUP_SSE_H__ */\n", "prefixes": [ "v5", "21/29" ] }