Patch Detail
get:
Show a patch.
patch:
Update a patch.
put:
Update a patch.
GET /api/patches/64459/?format=api
{ "id": 64459, "url": "http://patches.dpdk.org/api/patches/64459/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/patch/20200111160827.10021-1-pbhagavatula@marvell.com/", "project": { "id": 1, "url": "http://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": "<20200111160827.10021-1-pbhagavatula@marvell.com>", "list_archive_url": "https://inbox.dpdk.org/dev/20200111160827.10021-1-pbhagavatula@marvell.com", "date": "2020-01-11T16:08:27", "name": "lmp: add lookup x4 with x4 default values", "commit_ref": null, "pull_url": null, "state": "rejected", "archived": true, "hash": "529facb05427c2b56a7d8a480e9b5f7d5d3950b5", "submitter": { "id": 1183, "url": "http://patches.dpdk.org/api/people/1183/?format=api", "name": "Pavan Nikhilesh Bhagavatula", "email": "pbhagavatula@marvell.com" }, "delegate": { "id": 1, "url": "http://patches.dpdk.org/api/users/1/?format=api", "username": "tmonjalo", "first_name": "Thomas", "last_name": "Monjalon", "email": "thomas@monjalon.net" }, "mbox": "http://patches.dpdk.org/project/dpdk/patch/20200111160827.10021-1-pbhagavatula@marvell.com/mbox/", "series": [ { "id": 8061, "url": "http://patches.dpdk.org/api/series/8061/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/list/?series=8061", "date": "2020-01-11T16:08:27", "name": "lmp: add lookup x4 with x4 default values", "version": 1, "mbox": "http://patches.dpdk.org/series/8061/mbox/" } ], "comments": "http://patches.dpdk.org/api/patches/64459/comments/", "check": "fail", "checks": "http://patches.dpdk.org/api/patches/64459/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 96811A04FB;\n\tSat, 11 Jan 2020 17:08:39 +0100 (CET)", "from [92.243.14.124] (localhost [127.0.0.1])\n\tby dpdk.org (Postfix) with ESMTP id 99FD01DDD9;\n\tSat, 11 Jan 2020 17:08:38 +0100 (CET)", "from mx0b-0016f401.pphosted.com (mx0b-0016f401.pphosted.com\n [67.231.156.173]) by dpdk.org (Postfix) with ESMTP id 2D5051D659\n for <dev@dpdk.org>; Sat, 11 Jan 2020 17:08:37 +0100 (CET)", "from pps.filterd (m0045851.ppops.net [127.0.0.1])\n by mx0b-0016f401.pphosted.com (8.16.0.42/8.16.0.42) with SMTP id\n 00BG6DEN017118; Sat, 11 Jan 2020 08:08:34 -0800", "from sc-exch03.marvell.com ([199.233.58.183])\n by mx0b-0016f401.pphosted.com with ESMTP id 2xfert0d8u-1\n (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT);\n Sat, 11 Jan 2020 08:08:33 -0800", "from SC-EXCH01.marvell.com (10.93.176.81) by SC-EXCH03.marvell.com\n (10.93.176.83) with Microsoft SMTP Server (TLS) id 15.0.1497.2; Sat, 11 Jan\n 2020 08:08:31 -0800", "from maili.marvell.com (10.93.176.43) by SC-EXCH01.marvell.com\n (10.93.176.81) with Microsoft SMTP Server id 15.0.1497.2 via Frontend\n Transport; Sat, 11 Jan 2020 08:08:31 -0800", "from BG-LT7430.marvell.com (bg-lt7430.marvell.com [10.28.17.50])\n by maili.marvell.com (Postfix) with ESMTP id 1B6203F703F;\n Sat, 11 Jan 2020 08:08:28 -0800 (PST)" ], "DKIM-Signature": "v=1; a=rsa-sha256; c=relaxed/relaxed; d=marvell.com;\n h=from : to : cc :\n subject : date : message-id : mime-version : content-transfer-encoding :\n content-type; s=pfpt0818; bh=cUoSmtq+NwPxiGlkbalIKQLarW2fET+jHK0GlyNqN5U=;\n b=Y+vG7cUwgQhmYZzlaaG815bdsAsc8zwWrTQbhGMVe9QxFiHjeeaGBmul/eSnC6YwovME\n zWEwfcOQG9VsMaN0xwr0vI1isOmPYHmXKMJG1Rs87d+2tcB3kRM9RMMypuQCvaGOIXcg\n kf/hylaBGYkeWvbK8QRqYit/iFkmoF8JdvSltlEOnmFTtAJbNF73U3vvjIF07pCp+Yzv\n cuSwdW5lzxnkK8Z3ONZsjC7KkU938m+7ngujnpnaMANzcnc5+3xUSpTzt/fGsR1bNXe+\n ffx56shIjWllYOL0r5R6hrBQDYe4igqytIC1Gy6v5CipRlMfWdmiYkUSQWl/vEVn5ggN Fg==", "From": "<pbhagavatula@marvell.com>", "To": "<jerinj@marvell.com>, Bruce Richardson <bruce.richardson@intel.com>,\n Vladimir Medvedkin <vladimir.medvedkin@intel.com>, Gavin Hu\n <gavin.hu@arm.com>", "CC": "<dev@dpdk.org>, Pavan Nikhilesh <pbhagavatula@marvell.com>", "Date": "Sat, 11 Jan 2020 21:38:27 +0530", "Message-ID": "<20200111160827.10021-1-pbhagavatula@marvell.com>", "X-Mailer": "git-send-email 2.17.1", "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, 18.0.572\n definitions=2020-01-11_04:2020-01-10,\n 2020-01-11 signatures=0", "Subject": "[dpdk-dev] [PATCH] lmp: add lookup x4 with x4 default values", "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 lookup x4 with x4 default values.\nThis can be used in usecases where we have to process burst of packets\nfrom different ports.\n\nSigned-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>\n---\n app/test/test_lpm_perf.c | 31 +++++++++\n lib/librte_lpm/rte_lpm.h | 23 +++++++\n lib/librte_lpm/rte_lpm_altivec.h | 109 +++++++++++++++++++++++++++++++\n lib/librte_lpm/rte_lpm_neon.h | 102 +++++++++++++++++++++++++++++\n lib/librte_lpm/rte_lpm_sse.h | 104 +++++++++++++++++++++++++++++\n 5 files changed, 369 insertions(+)", "diff": "diff --git a/app/test/test_lpm_perf.c b/app/test/test_lpm_perf.c\nindex a2578fe90..8e9d4c7eb 100644\n--- a/app/test/test_lpm_perf.c\n+++ b/app/test/test_lpm_perf.c\n@@ -460,6 +460,37 @@ test_lpm_perf(void)\n \t\t\t(double)total_time / ((double)ITERATIONS * BATCH_SIZE),\n \t\t\t(count * 100.0) / (double)(ITERATIONS * BATCH_SIZE));\n \n+\t/* Measure LookupX4 DefaultX4 */\n+\ttotal_time = 0;\n+\tcount = 0;\n+\tuint32_t def[4] = {UINT32_MAX, UINT32_MAX, UINT32_MAX, UINT32_MAX};\n+\tfor (i = 0; i < ITERATIONS; i++) {\n+\t\tstatic uint32_t ip_batch[BATCH_SIZE];\n+\t\tuint32_t next_hops[4];\n+\n+\t\t/* Create array of random IP addresses */\n+\t\tfor (j = 0; j < BATCH_SIZE; j++)\n+\t\t\tip_batch[j] = rte_rand();\n+\n+\t\t/* Lookup per batch */\n+\t\tbegin = rte_rdtsc();\n+\t\tfor (j = 0; j < BATCH_SIZE; j += RTE_DIM(next_hops)) {\n+\t\t\tunsigned int k;\n+\t\t\txmm_t ipx4;\n+\n+\t\t\tipx4 = vect_loadu_sil128((xmm_t *)(ip_batch + j));\n+\t\t\tipx4 = *(xmm_t *)(ip_batch + j);\n+\t\t\trte_lpm_lookupx4_defx4(lpm, ipx4, next_hops, def);\n+\t\t\tfor (k = 0; k < RTE_DIM(next_hops); k++)\n+\t\t\t\tif (unlikely(next_hops[k] == UINT32_MAX))\n+\t\t\t\t\tcount++;\n+\t\t}\n+\n+\t\ttotal_time += rte_rdtsc() - begin;\n+\t}\n+\tprintf(\"LPM LookupX4 Defx4: %.1f cycles (fails = %.1f%%)\\n\",\n+\t\t\t(double)total_time / ((double)ITERATIONS * BATCH_SIZE),\n+\t\t\t(count * 100.0) / (double)(ITERATIONS * BATCH_SIZE));\n \t/* Measure Delete */\n \tstatus = 0;\n \tbegin = rte_rdtsc();\ndiff --git a/lib/librte_lpm/rte_lpm.h b/lib/librte_lpm/rte_lpm.h\nindex b9d49ac87..e66b43e06 100644\n--- a/lib/librte_lpm/rte_lpm.h\n+++ b/lib/librte_lpm/rte_lpm.h\n@@ -370,6 +370,29 @@ static inline void\n rte_lpm_lookupx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n \tuint32_t defv);\n \n+/**\n+ * Lookup four IP addresses in an LPM table.\n+ *\n+ * @param lpm\n+ * LPM object handle\n+ * @param ip\n+ * Four IPs to be looked up in the LPM table\n+ * @param hop\n+ * Next hop of the most specific rule found for IP (valid on lookup hit only).\n+ * This is an 4 elements array of two byte values.\n+ * If the lookup was successful for the given IP, then least significant byte\n+ * of the corresponding element is the actual next hop and the most\n+ * significant byte is zero.\n+ * If the lookup for the given IP failed, then corresponding element would\n+ * contain default value, see description of then next parameter.\n+ * @param defv\n+ * Default value[] to populate into corresponding element of hop[] array,\n+ * if lookup would fail.\n+ */\n+static inline void\n+rte_lpm_lookupx4_defx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n+\tuint32_t defv[4]);\n+\n #if defined(RTE_ARCH_ARM) || defined(RTE_ARCH_ARM64)\n #include \"rte_lpm_neon.h\"\n #elif defined(RTE_ARCH_PPC_64)\ndiff --git a/lib/librte_lpm/rte_lpm_altivec.h b/lib/librte_lpm/rte_lpm_altivec.h\nindex 228c41b38..1afc7bd74 100644\n--- a/lib/librte_lpm/rte_lpm_altivec.h\n+++ b/lib/librte_lpm/rte_lpm_altivec.h\n@@ -120,6 +120,115 @@ rte_lpm_lookupx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n \thop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF : defv;\n }\n \n+static inline void\n+rte_lpm_lookupx4_defx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n+\tuint32_t defv[4])\n+{\n+\tvector signed int i24;\n+\trte_xmm_t i8;\n+\tuint32_t tbl[4];\n+\tuint64_t idx, pt, pt2;\n+\tconst uint32_t *ptbl;\n+\n+\tconst uint32_t mask = UINT8_MAX;\n+\tconst vector signed int mask8 = (xmm_t){mask, mask, mask, mask};\n+\n+\t/*\n+\t * RTE_LPM_VALID_EXT_ENTRY_BITMASK for 2 LPM entries\n+\t * as one 64-bit value (0x0300000003000000).\n+\t */\n+\tconst uint64_t mask_xv =\n+\t\t((uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK |\n+\t\t(uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK << 32);\n+\n+\t/*\n+\t * RTE_LPM_LOOKUP_SUCCESS for 2 LPM entries\n+\t * as one 64-bit value (0x0100000001000000).\n+\t */\n+\tconst uint64_t mask_v =\n+\t\t((uint64_t)RTE_LPM_LOOKUP_SUCCESS |\n+\t\t(uint64_t)RTE_LPM_LOOKUP_SUCCESS << 32);\n+\n+\t/* get 4 indexes for tbl24[]. */\n+\ti24 = vec_sr((xmm_t) ip,\n+\t\t(vector unsigned int){CHAR_BIT, CHAR_BIT, CHAR_BIT, CHAR_BIT});\n+\n+\t/* extract values from tbl24[] */\n+\tidx = (uint32_t)i24[0];\n+\tidx = idx < (1<<24) ? idx : (1<<24)-1;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx];\n+\ttbl[0] = *ptbl;\n+\n+\tidx = (uint32_t) i24[1];\n+\tidx = idx < (1<<24) ? idx : (1<<24)-1;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx];\n+\ttbl[1] = *ptbl;\n+\n+\tidx = (uint32_t) i24[2];\n+\tidx = idx < (1<<24) ? idx : (1<<24)-1;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx];\n+\ttbl[2] = *ptbl;\n+\n+\tidx = (uint32_t) i24[3];\n+\tidx = idx < (1<<24) ? idx : (1<<24)-1;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx];\n+\ttbl[3] = *ptbl;\n+\n+\t/* get 4 indexes for tbl8[]. */\n+\ti8.x = vec_and(ip, mask8);\n+\n+\tpt = (uint64_t)tbl[0] |\n+\t\t(uint64_t)tbl[1] << 32;\n+\tpt2 = (uint64_t)tbl[2] |\n+\t\t(uint64_t)tbl[3] << 32;\n+\n+\t/* search successfully finished for all 4 IP addresses. */\n+\tif (likely((pt & mask_xv) == mask_v) &&\n+\t\t\tlikely((pt2 & mask_xv) == mask_v)) {\n+\t\t*(uint64_t *)hop = pt & RTE_LPM_MASKX4_RES;\n+\t\t*(uint64_t *)(hop + 2) = pt2 & RTE_LPM_MASKX4_RES;\n+\t\treturn;\n+\t}\n+\n+\tif (unlikely((pt & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[0] = i8.u32[0] +\n+\t\t\t(uint8_t)tbl[0] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[0]];\n+\t\ttbl[0] = *ptbl;\n+\t}\n+\tif (unlikely((pt >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[1] = i8.u32[1] +\n+\t\t\t(uint8_t)tbl[1] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[1]];\n+\t\ttbl[1] = *ptbl;\n+\t}\n+\tif (unlikely((pt2 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[2] = i8.u32[2] +\n+\t\t\t(uint8_t)tbl[2] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[2]];\n+\t\ttbl[2] = *ptbl;\n+\t}\n+\tif (unlikely((pt2 >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[3] = i8.u32[3] +\n+\t\t\t(uint8_t)tbl[3] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[3]];\n+\t\ttbl[3] = *ptbl;\n+\t}\n+\n+\thop[0] = (tbl[0] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[0] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[0];\n+\thop[1] = (tbl[1] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[1] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[1];\n+\thop[2] = (tbl[2] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[2] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[2];\n+\thop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[3];\n+}\n+\n #ifdef __cplusplus\n }\n #endif\ndiff --git a/lib/librte_lpm/rte_lpm_neon.h b/lib/librte_lpm/rte_lpm_neon.h\nindex 6c131d312..6ef635b18 100644\n--- a/lib/librte_lpm/rte_lpm_neon.h\n+++ b/lib/librte_lpm/rte_lpm_neon.h\n@@ -113,6 +113,108 @@ rte_lpm_lookupx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n \thop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF : defv;\n }\n \n+static inline void\n+rte_lpm_lookupx4_defx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n+\tuint32_t defv[4])\n+{\n+\tuint32x4_t i24;\n+\trte_xmm_t i8;\n+\tuint32_t tbl[4];\n+\tuint64_t idx, pt, pt2;\n+\tconst uint32_t *ptbl;\n+\n+\tconst uint32_t mask = UINT8_MAX;\n+\tconst int32x4_t mask8 = vdupq_n_s32(mask);\n+\n+\t/*\n+\t * RTE_LPM_VALID_EXT_ENTRY_BITMASK for 2 LPM entries\n+\t * as one 64-bit value (0x0300000003000000).\n+\t */\n+\tconst uint64_t mask_xv =\n+\t\t((uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK |\n+\t\t(uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK << 32);\n+\n+\t/*\n+\t * RTE_LPM_LOOKUP_SUCCESS for 2 LPM entries\n+\t * as one 64-bit value (0x0100000001000000).\n+\t */\n+\tconst uint64_t mask_v =\n+\t\t((uint64_t)RTE_LPM_LOOKUP_SUCCESS |\n+\t\t(uint64_t)RTE_LPM_LOOKUP_SUCCESS << 32);\n+\n+\t/* get 4 indexes for tbl24[]. */\n+\ti24 = vshrq_n_u32((uint32x4_t)ip, CHAR_BIT);\n+\n+\t/* extract values from tbl24[] */\n+\tidx = vgetq_lane_u64((uint64x2_t)i24, 0);\n+\n+\tptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];\n+\ttbl[0] = *ptbl;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];\n+\ttbl[1] = *ptbl;\n+\n+\tidx = vgetq_lane_u64((uint64x2_t)i24, 1);\n+\n+\tptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];\n+\ttbl[2] = *ptbl;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];\n+\ttbl[3] = *ptbl;\n+\n+\t/* get 4 indexes for tbl8[]. */\n+\ti8.x = vandq_s32(ip, mask8);\n+\n+\tpt = (uint64_t)tbl[0] |\n+\t\t(uint64_t)tbl[1] << 32;\n+\tpt2 = (uint64_t)tbl[2] |\n+\t\t(uint64_t)tbl[3] << 32;\n+\n+\t/* search successfully finished for all 4 IP addresses. */\n+\tif (likely((pt & mask_xv) == mask_v) &&\n+\t\t\tlikely((pt2 & mask_xv) == mask_v)) {\n+\t\t*(uint64_t *)hop = pt & RTE_LPM_MASKX4_RES;\n+\t\t*(uint64_t *)(hop + 2) = pt2 & RTE_LPM_MASKX4_RES;\n+\t\treturn;\n+\t}\n+\n+\tif (unlikely((pt & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[0] = i8.u32[0] +\n+\t\t\t(uint8_t)tbl[0] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[0]];\n+\t\ttbl[0] = *ptbl;\n+\t}\n+\tif (unlikely((pt >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[1] = i8.u32[1] +\n+\t\t\t(uint8_t)tbl[1] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[1]];\n+\t\ttbl[1] = *ptbl;\n+\t}\n+\tif (unlikely((pt2 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[2] = i8.u32[2] +\n+\t\t\t(uint8_t)tbl[2] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[2]];\n+\t\ttbl[2] = *ptbl;\n+\t}\n+\tif (unlikely((pt2 >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[3] = i8.u32[3] +\n+\t\t\t(uint8_t)tbl[3] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[3]];\n+\t\ttbl[3] = *ptbl;\n+\t}\n+\n+\thop[0] = (tbl[0] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[0] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[0];\n+\thop[1] = (tbl[1] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[1] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[1];\n+\thop[2] = (tbl[2] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[2] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[2];\n+\thop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[3];\n+}\n+\n #ifdef __cplusplus\n }\n #endif\ndiff --git a/lib/librte_lpm/rte_lpm_sse.h b/lib/librte_lpm/rte_lpm_sse.h\nindex 44770b6ff..6ef15816c 100644\n--- a/lib/librte_lpm/rte_lpm_sse.h\n+++ b/lib/librte_lpm/rte_lpm_sse.h\n@@ -114,6 +114,110 @@ rte_lpm_lookupx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n \thop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF : defv;\n }\n \n+static inline void\n+rte_lpm_lookupx4_defx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],\n+\tuint32_t defv[4])\n+{\n+\t__m128i i24;\n+\trte_xmm_t i8;\n+\tuint32_t tbl[4];\n+\tuint64_t idx, pt, pt2;\n+\tconst uint32_t *ptbl;\n+\n+\tconst __m128i mask8 =\n+\t\t_mm_set_epi32(UINT8_MAX, UINT8_MAX, UINT8_MAX, UINT8_MAX);\n+\n+\t/*\n+\t * RTE_LPM_VALID_EXT_ENTRY_BITMASK for 2 LPM entries\n+\t * as one 64-bit value (0x0300000003000000).\n+\t */\n+\tconst uint64_t mask_xv =\n+\t\t((uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK |\n+\t\t(uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK << 32);\n+\n+\t/*\n+\t * RTE_LPM_LOOKUP_SUCCESS for 2 LPM entries\n+\t * as one 64-bit value (0x0100000001000000).\n+\t */\n+\tconst uint64_t mask_v =\n+\t\t((uint64_t)RTE_LPM_LOOKUP_SUCCESS |\n+\t\t(uint64_t)RTE_LPM_LOOKUP_SUCCESS << 32);\n+\n+\t/* get 4 indexes for tbl24[]. */\n+\ti24 = _mm_srli_epi32(ip, CHAR_BIT);\n+\n+\t/* extract values from tbl24[] */\n+\tidx = _mm_cvtsi128_si64(i24);\n+\t/* With -O0 option, gcc 4.8 - 5.4 fails to fold sizeof() into a constant */\n+\ti24 = _mm_srli_si128(i24, /* sizeof(uint64_t) */ 8);\n+\n+\tptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];\n+\ttbl[0] = *ptbl;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];\n+\ttbl[1] = *ptbl;\n+\n+\tidx = _mm_cvtsi128_si64(i24);\n+\n+\tptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];\n+\ttbl[2] = *ptbl;\n+\tptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];\n+\ttbl[3] = *ptbl;\n+\n+\t/* get 4 indexes for tbl8[]. */\n+\ti8.x = _mm_and_si128(ip, mask8);\n+\n+\tpt = (uint64_t)tbl[0] |\n+\t\t(uint64_t)tbl[1] << 32;\n+\tpt2 = (uint64_t)tbl[2] |\n+\t\t(uint64_t)tbl[3] << 32;\n+\n+\t/* search successfully finished for all 4 IP addresses. */\n+\tif (likely((pt & mask_xv) == mask_v) &&\n+\t\t\tlikely((pt2 & mask_xv) == mask_v)) {\n+\t\t*(uint64_t *)hop = pt & RTE_LPM_MASKX4_RES;\n+\t\t*(uint64_t *)(hop + 2) = pt2 & RTE_LPM_MASKX4_RES;\n+\t\treturn;\n+\t}\n+\n+\tif (unlikely((pt & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[0] = i8.u32[0] +\n+\t\t\t(uint8_t)tbl[0] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[0]];\n+\t\ttbl[0] = *ptbl;\n+\t}\n+\tif (unlikely((pt >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[1] = i8.u32[1] +\n+\t\t\t(uint8_t)tbl[1] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[1]];\n+\t\ttbl[1] = *ptbl;\n+\t}\n+\tif (unlikely((pt2 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[2] = i8.u32[2] +\n+\t\t\t(uint8_t)tbl[2] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[2]];\n+\t\ttbl[2] = *ptbl;\n+\t}\n+\tif (unlikely((pt2 >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==\n+\t\t\tRTE_LPM_VALID_EXT_ENTRY_BITMASK)) {\n+\t\ti8.u32[3] = i8.u32[3] +\n+\t\t\t(uint8_t)tbl[3] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;\n+\t\tptbl = (const uint32_t *)&lpm->tbl8[i8.u32[3]];\n+\t\ttbl[3] = *ptbl;\n+\t}\n+\n+\thop[0] = (tbl[0] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[0] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[0];\n+\thop[1] = (tbl[1] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[1] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[1];\n+\thop[2] = (tbl[2] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[2] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[2];\n+\thop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF :\n+\t\t\t\t\t\t\t\t\tdefv[3];\n+}\n+\n #ifdef __cplusplus\n }\n #endif\n", "prefixes": [] }