Patch Detail
get:
Show a patch.
patch:
Update a patch.
put:
Update a patch.
GET /api/patches/2409/?format=api
{ "id": 2409, "url": "https://patches.dpdk.org/api/patches/2409/?format=api", "web_url": "https://patches.dpdk.org/project/dpdk/patch/1421779267-18492-15-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": "<1421779267-18492-15-git-send-email-konstantin.ananyev@intel.com>", "list_archive_url": "https://inbox.dpdk.org/dev/1421779267-18492-15-git-send-email-konstantin.ananyev@intel.com", "date": "2015-01-20T18:41:03", "name": "[dpdk-dev,v3,14/18] libter_acl: move lo/hi dwords shuffle out from calc_addr", "commit_ref": null, "pull_url": null, "state": "accepted", "archived": true, "hash": "9d71854c7f4f0ba4e1ebc00180e5a7229328d258", "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/1421779267-18492-15-git-send-email-konstantin.ananyev@intel.com/mbox/", "series": [], "comments": "https://patches.dpdk.org/api/patches/2409/comments/", "check": "pending", "checks": "https://patches.dpdk.org/api/patches/2409/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 587C65AE0;\n\tTue, 20 Jan 2015 19:41:50 +0100 (CET)", "from mga11.intel.com (mga11.intel.com [192.55.52.93])\n\tby dpdk.org (Postfix) with ESMTP id 1A3FC5AC8\n\tfor <dev@dpdk.org>; Tue, 20 Jan 2015 19:41:37 +0100 (CET)", "from orsmga003.jf.intel.com ([10.7.209.27])\n\tby fmsmga102.fm.intel.com with ESMTP; 20 Jan 2015 10:41:23 -0800", "from irvmail001.ir.intel.com ([163.33.26.43])\n\tby orsmga003.jf.intel.com with ESMTP; 20 Jan 2015 10:34:44 -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\tt0KIfLM8029512; Tue, 20 Jan 2015 18:41:21 GMT", "from sivswdev02.ir.intel.com (localhost [127.0.0.1])\n\tby sivswdev02.ir.intel.com with ESMTP id t0KIfL9u018884;\n\tTue, 20 Jan 2015 18:41:21 GMT", "(from kananye1@localhost)\n\tby sivswdev02.ir.intel.com with id t0KIfLDp018880;\n\tTue, 20 Jan 2015 18:41:21 GMT" ], "X-ExtLoop1": "1", "X-IronPort-AV": "E=Sophos;i=\"5.09,435,1418112000\"; d=\"scan'208\";a=\"514984624\"", "From": "Konstantin Ananyev <konstantin.ananyev@intel.com>", "To": "dev@dpdk.org", "Date": "Tue, 20 Jan 2015 18:41:03 +0000", "Message-Id": "<1421779267-18492-15-git-send-email-konstantin.ananyev@intel.com>", "X-Mailer": "git-send-email 1.7.4.1", "In-Reply-To": "<1421779267-18492-1-git-send-email-konstantin.ananyev@intel.com>", "References": "<1421779267-18492-1-git-send-email-konstantin.ananyev@intel.com>", "Subject": "[dpdk-dev] [PATCH v3 14/18] libter_acl: move lo/hi dwords shuffle\n\tout from calc_addr", "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": "Reorganise SSE code-path a bit by moving lo/hi dwords shuffle\nout from calc_addr().\nThat allows to make calc_addr() for SSE and AVX2 practically identical\nand opens opportunity for further code deduplication.\n\nSigned-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>\n---\n lib/librte_acl/acl_run_sse.h | 38 ++++++++++++++++++++------------------\n 1 file changed, 20 insertions(+), 18 deletions(-)", "diff": "diff --git a/lib/librte_acl/acl_run_sse.h b/lib/librte_acl/acl_run_sse.h\nindex 1b7870e..4a174e9 100644\n--- a/lib/librte_acl/acl_run_sse.h\n+++ b/lib/librte_acl/acl_run_sse.h\n@@ -172,9 +172,9 @@ acl_match_check_x4(int slot, const struct rte_acl_ctx *ctx, struct parms *parms,\n */\n static inline __attribute__((always_inline)) xmm_t\n calc_addr_sse(xmm_t index_mask, xmm_t next_input, xmm_t shuffle_input,\n-\txmm_t ones_16, xmm_t indices1, xmm_t indices2)\n+\txmm_t ones_16, xmm_t tr_lo, xmm_t tr_hi)\n {\n-\txmm_t addr, node_types, range, temp;\n+\txmm_t addr, node_types;\n \txmm_t dfa_msk, dfa_ofs, quad_ofs;\n \txmm_t in, r, t;\n \n@@ -187,18 +187,14 @@ calc_addr_sse(xmm_t index_mask, xmm_t next_input, xmm_t shuffle_input,\n \t * it reaches a match.\n \t */\n \n-\t/* Shuffle low 32 into temp and high 32 into indices2 */\n-\ttemp = (xmm_t)MM_SHUFFLEPS((__m128)indices1, (__m128)indices2, 0x88);\n-\trange = (xmm_t)MM_SHUFFLEPS((__m128)indices1, (__m128)indices2, 0xdd);\n-\n \tt = MM_XOR(index_mask, index_mask);\n \n \t/* shuffle input byte to all 4 positions of 32 bit value */\n \tin = MM_SHUFFLE8(next_input, shuffle_input);\n \n \t/* Calc node type and node addr */\n-\tnode_types = MM_ANDNOT(index_mask, temp);\n-\taddr = MM_AND(index_mask, temp);\n+\tnode_types = MM_ANDNOT(index_mask, tr_lo);\n+\taddr = MM_AND(index_mask, tr_lo);\n \n \t/*\n \t * Calc addr for DFAs - addr = dfa_index + input_byte\n@@ -211,7 +207,7 @@ calc_addr_sse(xmm_t index_mask, xmm_t next_input, xmm_t shuffle_input,\n \tr = _mm_add_epi8(r, range_base);\n \n \tt = _mm_srli_epi32(in, 24);\n-\tr = _mm_shuffle_epi8(range, r);\n+\tr = _mm_shuffle_epi8(tr_hi, r);\n \n \tdfa_ofs = _mm_sub_epi32(t, r);\n \n@@ -224,22 +220,22 @@ calc_addr_sse(xmm_t index_mask, xmm_t next_input, xmm_t shuffle_input,\n \t */\n \n \t/* check ranges */\n-\ttemp = MM_CMPGT8(in, range);\n+\tt = MM_CMPGT8(in, tr_hi);\n \n \t/* convert -1 to 1 (bytes greater than input byte */\n-\ttemp = MM_SIGN8(temp, temp);\n+\tt = MM_SIGN8(t, t);\n \n \t/* horizontal add pairs of bytes into words */\n-\ttemp = MM_MADD8(temp, temp);\n+\tt = MM_MADD8(t, t);\n \n \t/* horizontal add pairs of words into dwords */\n-\tquad_ofs = MM_MADD16(temp, ones_16);\n+\tquad_ofs = MM_MADD16(t, ones_16);\n \n-\t/* mask to range type nodes */\n-\ttemp = _mm_blendv_epi8(quad_ofs, dfa_ofs, dfa_msk);\n+\t/* blend DFA and QUAD/SINGLE. */\n+\tt = _mm_blendv_epi8(quad_ofs, dfa_ofs, dfa_msk);\n \n \t/* add index into node position */\n-\treturn MM_ADD32(addr, temp);\n+\treturn MM_ADD32(addr, t);\n }\n \n /*\n@@ -249,13 +245,19 @@ static inline __attribute__((always_inline)) xmm_t\n transition4(xmm_t next_input, const uint64_t *trans,\n \txmm_t *indices1, xmm_t *indices2)\n {\n-\txmm_t addr;\n+\txmm_t addr, tr_lo, tr_hi;\n \tuint64_t trans0, trans2;\n \n+\t/* Shuffle low 32 into tr_lo and high 32 into tr_hi */\n+\ttr_lo = (xmm_t)_mm_shuffle_ps((__m128)*indices1, (__m128)*indices2,\n+\t\t0x88);\n+\ttr_hi = (xmm_t)_mm_shuffle_ps((__m128)*indices1, (__m128)*indices2,\n+\t\t0xdd);\n+\n \t /* Calculate the address (array index) for all 4 transitions. */\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+\t\txmm_ones_16.x, tr_lo, tr_hi);\n \n \t /* Gather 64 bit transitions and pack back into 2 registers. */\n \n", "prefixes": [ "dpdk-dev", "v3", "14/18" ] }