Patch Detail

HTTP 200 OK
Allow: GET, PUT, PATCH, HEAD, OPTIONS
Content-Type: application/json
Vary: Accept

{
    "id": 2260,
    "url": "https://patches.dpdk.org/api/patches/2260/?format=api",
    "web_url": "https://patches.dpdk.org/project/dpdk/patch/1421090181-17150-14-git-send-email-konstantin.ananyev@intel.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": "<1421090181-17150-14-git-send-email-konstantin.ananyev@intel.com>",
    "list_archive_url": "https://inbox.dpdk.org/dev/1421090181-17150-14-git-send-email-konstantin.ananyev@intel.com",
    "date": "2015-01-12T19:16:17",
    "name": "[dpdk-dev,v2,13/17] librte_acl: Remove search_sse_2 and relatives.",
    "commit_ref": null,
    "pull_url": null,
    "state": "superseded",
    "archived": true,
    "hash": "ca6166e8dbc0a609081dd90a9a3710b036cafbd9",
    "submitter": {
        "id": 33,
        "url": "https://patches.dpdk.org/api/people/33/?format=api",
        "name": "Ananyev, Konstantin",
        "email": "konstantin.ananyev@intel.com"
    },
    "delegate": null,
    "mbox": "https://patches.dpdk.org/project/dpdk/patch/1421090181-17150-14-git-send-email-konstantin.ananyev@intel.com/mbox/",
    "series": [],
    "comments": "https://patches.dpdk.org/api/patches/2260/comments/",
    "check": "pending",
    "checks": "https://patches.dpdk.org/api/patches/2260/checks/",
    "tags": {},
    "related": [],
    "headers": {
        "Return-Path": "<dev-bounces@dpdk.org>",
        "X-Original-To": "patchwork@dpdk.org",
        "Delivered-To": "patchwork@dpdk.org",
        "Received": [
            "from [92.243.14.124] (localhost [IPv6:::1])\n\tby dpdk.org (Postfix) with ESMTP id B97A15AE5;\n\tMon, 12 Jan 2015 20:17:05 +0100 (CET)",
            "from mga11.intel.com (mga11.intel.com [192.55.52.93])\n\tby dpdk.org (Postfix) with ESMTP id CDCFB5A9E\n\tfor <dev@dpdk.org>; Mon, 12 Jan 2015 20:16:39 +0100 (CET)",
            "from orsmga002.jf.intel.com ([10.7.209.21])\n\tby fmsmga102.fm.intel.com with ESMTP; 12 Jan 2015 11:16:37 -0800",
            "from irvmail001.ir.intel.com ([163.33.26.43])\n\tby orsmga002.jf.intel.com with ESMTP; 12 Jan 2015 11:16:36 -0800",
            "from sivswdev02.ir.intel.com (sivswdev02.ir.intel.com\n\t[10.237.217.46])\n\tby irvmail001.ir.intel.com (8.14.3/8.13.6/MailSET/Hub) with ESMTP id\n\tt0CJGa5J008640; Mon, 12 Jan 2015 19:16:36 GMT",
            "from sivswdev02.ir.intel.com (localhost [127.0.0.1])\n\tby sivswdev02.ir.intel.com with ESMTP id t0CJGZ0t017293;\n\tMon, 12 Jan 2015 19:16:35 GMT",
            "(from kananye1@localhost)\n\tby sivswdev02.ir.intel.com with  id t0CJGZJ7017289;\n\tMon, 12 Jan 2015 19:16:35 GMT"
        ],
        "X-ExtLoop1": "1",
        "X-IronPort-AV": "E=Sophos;i=\"5.07,745,1413270000\"; d=\"scan'208\";a=\"668605495\"",
        "From": "Konstantin Ananyev <konstantin.ananyev@intel.com>",
        "To": "dev@dpdk.org",
        "Date": "Mon, 12 Jan 2015 19:16:17 +0000",
        "Message-Id": "<1421090181-17150-14-git-send-email-konstantin.ananyev@intel.com>",
        "X-Mailer": "git-send-email 1.7.4.1",
        "In-Reply-To": "<1421090181-17150-1-git-send-email-konstantin.ananyev@intel.com>",
        "References": "<1421090181-17150-1-git-send-email-konstantin.ananyev@intel.com>",
        "Subject": "[dpdk-dev] [PATCH v2 13/17] librte_acl: Remove search_sse_2 and\n\trelatives.",
        "X-BeenThere": "dev@dpdk.org",
        "X-Mailman-Version": "2.1.15",
        "Precedence": "list",
        "List-Id": "patches and discussions about DPDK <dev.dpdk.org>",
        "List-Unsubscribe": "<http://dpdk.org/ml/options/dev>,\n\t<mailto:dev-request@dpdk.org?subject=unsubscribe>",
        "List-Archive": "<http://dpdk.org/ml/archives/dev/>",
        "List-Post": "<mailto:dev@dpdk.org>",
        "List-Help": "<mailto:dev-request@dpdk.org?subject=help>",
        "List-Subscribe": "<http://dpdk.org/ml/listinfo/dev>,\n\t<mailto:dev-request@dpdk.org?subject=subscribe>",
        "Errors-To": "dev-bounces@dpdk.org",
        "Sender": "\"dev\" <dev-bounces@dpdk.org>"
    },
    "content": "Previous improvements made scalar method the fastest one\nfor tiny bunch of packets (< 4).\nThat allows us to remove specific vector code-path for small number of packets\n(search_sse_2)\nand always use scalar method for such cases.\n\nSigned-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>\n---\n lib/librte_acl/acl_run_avx2.c |   2 +-\n lib/librte_acl/acl_run_sse.c  |   3 +-\n lib/librte_acl/acl_run_sse.h  | 110 ------------------------------------------\n 3 files changed, 3 insertions(+), 112 deletions(-)",
    "diff": "diff --git a/lib/librte_acl/acl_run_avx2.c b/lib/librte_acl/acl_run_avx2.c\nindex 0a42f72..79ebbd6 100644\n--- a/lib/librte_acl/acl_run_avx2.c\n+++ b/lib/librte_acl/acl_run_avx2.c\n@@ -49,6 +49,6 @@ rte_acl_classify_avx2(const struct rte_acl_ctx *ctx, const uint8_t **data,\n \telse if (num >= MAX_SEARCHES_SSE4)\n \t\treturn search_sse_4(ctx, data, results, num, categories);\n \telse\n-\t\treturn search_sse_2(ctx, data, results, num,\n+\t\treturn rte_acl_classify_scalar(ctx, data, results, num,\n \t\t\tcategories);\n }\ndiff --git a/lib/librte_acl/acl_run_sse.c b/lib/librte_acl/acl_run_sse.c\nindex 77b32b3..a5a7d36 100644\n--- a/lib/librte_acl/acl_run_sse.c\n+++ b/lib/librte_acl/acl_run_sse.c\n@@ -42,5 +42,6 @@ rte_acl_classify_sse(const struct rte_acl_ctx *ctx, const uint8_t **data,\n \telse if (num >= MAX_SEARCHES_SSE4)\n \t\treturn search_sse_4(ctx, data, results, num, categories);\n \telse\n-\t\treturn search_sse_2(ctx, data, results, num, categories);\n+\t\treturn rte_acl_classify_scalar(ctx, data, results, num,\n+\t\t\tcategories);\n }\ndiff --git a/lib/librte_acl/acl_run_sse.h b/lib/librte_acl/acl_run_sse.h\nindex e33e16b..1b7870e 100644\n--- a/lib/librte_acl/acl_run_sse.h\n+++ b/lib/librte_acl/acl_run_sse.h\n@@ -45,10 +45,6 @@ static const rte_xmm_t xmm_shuffle_input = {\n \t.u32 = {0x00000000, 0x04040404, 0x08080808, 0x0c0c0c0c},\n };\n \n-static const rte_xmm_t xmm_shuffle_input64 = {\n-\t.u32 = {0x00000000, 0x04040404, 0x80808080, 0x80808080},\n-};\n-\n static const rte_xmm_t xmm_ones_16 = {\n \t.u16 = {1, 1, 1, 1, 1, 1, 1, 1},\n };\n@@ -62,15 +58,6 @@ static const rte_xmm_t xmm_match_mask = {\n \t},\n };\n \n-static const rte_xmm_t xmm_match_mask64 = {\n-\t.u32 = {\n-\t\tRTE_ACL_NODE_MATCH,\n-\t\t0,\n-\t\tRTE_ACL_NODE_MATCH,\n-\t\t0,\n-\t},\n-};\n-\n static const rte_xmm_t xmm_index_mask = {\n \t.u32 = {\n \t\tRTE_ACL_NODE_INDEX,\n@@ -80,16 +67,6 @@ static const rte_xmm_t xmm_index_mask = {\n \t},\n };\n \n-static const rte_xmm_t xmm_index_mask64 = {\n-\t.u32 = {\n-\t\tRTE_ACL_NODE_INDEX,\n-\t\tRTE_ACL_NODE_INDEX,\n-\t\t0,\n-\t\t0,\n-\t},\n-};\n-\n-\n /*\n  * Resolve priority for multiple results (sse version).\n  * This consists comparing the priority of the current traversal with the\n@@ -161,22 +138,6 @@ acl_process_matches(xmm_t *indices, int slot, const struct rte_acl_ctx *ctx,\n }\n \n /*\n- * Check for a match in 2 transitions (contained in SSE register)\n- */\n-static inline __attribute__((always_inline)) void\n-acl_match_check_x2(int slot, const struct rte_acl_ctx *ctx, struct parms *parms,\n-\tstruct acl_flow_data *flows, xmm_t *indices, xmm_t match_mask)\n-{\n-\txmm_t temp;\n-\n-\ttemp = MM_AND(match_mask, *indices);\n-\twhile (!MM_TESTZ(temp, temp)) {\n-\t\tacl_process_matches(indices, slot, ctx, parms, flows);\n-\t\ttemp = MM_AND(match_mask, *indices);\n-\t}\n-}\n-\n-/*\n  * Check for any match in 4 transitions (contained in 2 SSE registers)\n  */\n static inline __attribute__((always_inline)) void\n@@ -460,74 +421,3 @@ search_sse_4(const struct rte_acl_ctx *ctx, const uint8_t **data,\n \n \treturn 0;\n }\n-\n-static inline __attribute__((always_inline)) xmm_t\n-transition2(xmm_t next_input, const uint64_t *trans, xmm_t *indices1)\n-{\n-\tuint64_t t;\n-\txmm_t addr, indices2;\n-\n-\tindices2 = _mm_setzero_si128();\n-\n-\taddr = calc_addr_sse(xmm_index_mask.x, next_input, xmm_shuffle_input.x,\n-\t\txmm_ones_16.x, *indices1, indices2);\n-\n-\t/* Gather 64 bit transitions and pack 2 per register. */\n-\n-\tt = trans[MM_CVT32(addr)];\n-\n-\t/* get slot 1 */\n-\taddr = MM_SHUFFLE32(addr, SHUFFLE32_SLOT1);\n-\t*indices1 = MM_SET64(trans[MM_CVT32(addr)], t);\n-\n-\treturn MM_SRL32(next_input, CHAR_BIT);\n-}\n-\n-/*\n- * Execute trie traversal with 2 traversals in parallel.\n- */\n-static inline int\n-search_sse_2(const struct rte_acl_ctx *ctx, const uint8_t **data,\n-\tuint32_t *results, uint32_t total_packets, uint32_t categories)\n-{\n-\tint n;\n-\tstruct acl_flow_data flows;\n-\tuint64_t index_array[MAX_SEARCHES_SSE2];\n-\tstruct completion cmplt[MAX_SEARCHES_SSE2];\n-\tstruct parms parms[MAX_SEARCHES_SSE2];\n-\txmm_t input, indices;\n-\n-\tacl_set_flow(&flows, cmplt, RTE_DIM(cmplt), data, results,\n-\t\ttotal_packets, categories, ctx->trans_table);\n-\n-\tfor (n = 0; n < MAX_SEARCHES_SSE2; n++) {\n-\t\tcmplt[n].count = 0;\n-\t\tindex_array[n] = acl_start_next_trie(&flows, parms, n, ctx);\n-\t}\n-\n-\tindices = MM_LOADU((xmm_t *) &index_array[0]);\n-\n-\t/* Check for any matches. */\n-\tacl_match_check_x2(0, ctx, parms, &flows, &indices,\n-\t\txmm_match_mask64.x);\n-\n-\twhile (flows.started > 0) {\n-\n-\t\t/* Gather 4 bytes of input data for each stream. */\n-\t\tinput = _mm_cvtsi32_si128(GET_NEXT_4BYTES(parms, 0));\n-\t\tinput = MM_INSERT32(input, GET_NEXT_4BYTES(parms, 1), 1);\n-\n-\t\t/* Process the 4 bytes of input on each stream. */\n-\n-\t\tinput = transition2(input, flows.trans, &indices);\n-\t\tinput = transition2(input, flows.trans, &indices);\n-\t\tinput = transition2(input, flows.trans, &indices);\n-\t\tinput = transition2(input, flows.trans, &indices);\n-\n-\t\t/* Check for any matches. */\n-\t\tacl_match_check_x2(0, ctx, parms, &flows, &indices,\n-\t\t\txmm_match_mask64.x);\n-\t}\n-\n-\treturn 0;\n-}\n",
    "prefixes": [
        "dpdk-dev",
        "v2",
        "13/17"
    ]
}