Patch Detail
get:
Show a patch.
patch:
Update a patch.
put:
Update a patch.
GET /api/patches/83828/?format=api
{ "id": 83828, "url": "http://patches.dpdk.org/api/patches/83828/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/patch/20201109060647.63093-1-murphyx.yang@intel.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": "<20201109060647.63093-1-murphyx.yang@intel.com>", "list_archive_url": "https://inbox.dpdk.org/dev/20201109060647.63093-1-murphyx.yang@intel.com", "date": "2020-11-09T06:06:47", "name": "[v6] net/ice: fix outer checksum on cvl unknown", "commit_ref": null, "pull_url": null, "state": "superseded", "archived": true, "hash": "a19a25c0d29397d578b6418413e124fc24ed00bc", "submitter": { "id": 1986, "url": "http://patches.dpdk.org/api/people/1986/?format=api", "name": "Murphy Yang", "email": "murphyx.yang@intel.com" }, "delegate": { "id": 1540, "url": "http://patches.dpdk.org/api/users/1540/?format=api", "username": "qzhan15", "first_name": "Qi", "last_name": "Zhang", "email": "qi.z.zhang@intel.com" }, "mbox": "http://patches.dpdk.org/project/dpdk/patch/20201109060647.63093-1-murphyx.yang@intel.com/mbox/", "series": [ { "id": 13735, "url": "http://patches.dpdk.org/api/series/13735/?format=api", "web_url": "http://patches.dpdk.org/project/dpdk/list/?series=13735", "date": "2020-11-09T06:06:47", "name": "[v6] net/ice: fix outer checksum on cvl unknown", "version": 6, "mbox": "http://patches.dpdk.org/series/13735/mbox/" } ], "comments": "http://patches.dpdk.org/api/patches/83828/comments/", "check": "fail", "checks": "http://patches.dpdk.org/api/patches/83828/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 6A159A0527;\n\tMon, 9 Nov 2020 07:09:03 +0100 (CET)", "from [92.243.14.124] (localhost [127.0.0.1])\n\tby dpdk.org (Postfix) with ESMTP id EA66158C4;\n\tMon, 9 Nov 2020 07:09:00 +0100 (CET)", "from mga05.intel.com (mga05.intel.com [192.55.52.43])\n by dpdk.org (Postfix) with ESMTP id 82F1D2BF1\n for <dev@dpdk.org>; Mon, 9 Nov 2020 07:08:59 +0100 (CET)", "from fmsmga004.fm.intel.com ([10.253.24.48])\n by fmsmga105.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384;\n 08 Nov 2020 22:08:57 -0800", "from unknown (HELO intel-npg-odc-srv02.cd.intel.com)\n ([10.240.178.186])\n by fmsmga004-auth.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384;\n 08 Nov 2020 22:08:55 -0800" ], "IronPort-SDR": [ "\n xnxi18f4RiAuvjCiN6rUVQkTSYl9+ilIG1n2fFdEefOuWb621585QKEJ2h60d8TjBD4jXLRGhK\n WTq4Bn3tiyVg==", "\n xPzBIqo2KVAGLTVOM2DNSXj4ULnmwofrieLoiCm3L9Bj1oS9Ow+LvhRa9qLNMXxeTC6eUTGanI\n q1K2/3NHqTqg==" ], "X-IronPort-AV": [ "E=McAfee;i=\"6000,8403,9799\"; a=\"254461787\"", "E=Sophos;i=\"5.77,462,1596524400\"; d=\"scan'208\";a=\"254461787\"", "E=Sophos;i=\"5.77,462,1596524400\"; d=\"scan'208\";a=\"355535171\"" ], "X-Amp-Result": "SKIPPED(no attachment in message)", "X-Amp-File-Uploaded": "False", "From": "Murphy Yang <murphyx.yang@intel.com>", "To": "dev@dpdk.org", "Cc": "qiming.yang@intel.com, qi.z.zhang@intel.com, stevex.yang@intel.com,\n leyi.rong@intel.com, wenzhuo.lu@intel.com,\n Murphy Yang <murphyx.yang@intel.com>", "Date": "Mon, 9 Nov 2020 06:06:47 +0000", "Message-Id": "<20201109060647.63093-1-murphyx.yang@intel.com>", "X-Mailer": "git-send-email 2.17.1", "In-Reply-To": "<20201104092652.34744-1-murphyx.yang@intel.com>", "References": "<20201104092652.34744-1-murphyx.yang@intel.com>", "Subject": "[dpdk-dev] [PATCH v6] net/ice: fix outer checksum on cvl unknown", "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": "Currently, driver does not support parse UDP outer checksum flag of\ntunneled packets.\n\nWhen execute 'csum set outer-udp hw 0' and 'csum parse-tunnel on 0'\ncommands to enable hardware UDP outer checksum. This patch supports\nparse UDP outer checksum flag of tunneled packets.\n\nFixes: dbf3c0e77a22 (\"net/ice: handle Rx flex descriptor\")\nFixes: 4ab7dbb0a0f6 (\"net/ice: switch to Rx flexible descriptor in AVX path\")\nFixes: ece1f8a8f1c8 (\"net/ice: switch to flexible descriptor in SSE path\")\n\nSigned-off-by: Murphy Yang <murphyx.yang@intel.com>\n---\nv6:\n- rename variable name.\n- update comments.\nv5:\n- fix outer L4 checksum mask for vector path.\nv4:\n- cover AVX512 vector path.\nv3:\n- add PKT_RX_OUTER_L4_CKSUM_GOOD in AVX2 and SSE vector path.\n- rename variable name.\nv2:\n- cover AVX2 and SSE vector path\n\n drivers/net/ice/ice_rxtx.c | 5 ++\n drivers/net/ice/ice_rxtx_vec_avx2.c | 118 +++++++++++++++++++-------\n drivers/net/ice/ice_rxtx_vec_avx512.c | 117 ++++++++++++++++++-------\n drivers/net/ice/ice_rxtx_vec_sse.c | 78 ++++++++++++-----\n 4 files changed, 233 insertions(+), 85 deletions(-)", "diff": "diff --git a/drivers/net/ice/ice_rxtx.c b/drivers/net/ice/ice_rxtx.c\nindex 5fbd68eafc..24a7caeb98 100644\n--- a/drivers/net/ice/ice_rxtx.c\n+++ b/drivers/net/ice/ice_rxtx.c\n@@ -1451,6 +1451,11 @@ ice_rxd_error_to_pkt_flags(uint16_t stat_err0)\n \tif (unlikely(stat_err0 & (1 << ICE_RX_FLEX_DESC_STATUS0_XSUM_EIPE_S)))\n \t\tflags |= PKT_RX_EIP_CKSUM_BAD;\n \n+\tif (unlikely(stat_err0 & (1 << ICE_RX_FLEX_DESC_STATUS0_XSUM_EUDPE_S)))\n+\t\tflags |= PKT_RX_OUTER_L4_CKSUM_BAD;\n+\telse\n+\t\tflags |= PKT_RX_OUTER_L4_CKSUM_GOOD;\n+\n \treturn flags;\n }\n \ndiff --git a/drivers/net/ice/ice_rxtx_vec_avx2.c b/drivers/net/ice/ice_rxtx_vec_avx2.c\nindex b72a9e7025..7838e17787 100644\n--- a/drivers/net/ice/ice_rxtx_vec_avx2.c\n+++ b/drivers/net/ice/ice_rxtx_vec_avx2.c\n@@ -251,43 +251,88 @@ _ice_recv_raw_pkts_vec_avx2(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts,\n \t * bit13 is for VLAN indication.\n \t */\n \tconst __m256i flags_mask =\n-\t\t _mm256_set1_epi32((7 << 4) | (1 << 12) | (1 << 13));\n+\t\t _mm256_set1_epi32((0xF << 4) | (1 << 12) | (1 << 13));\n \t/**\n \t * data to be shuffled by the result of the flags mask shifted by 4\n \t * bits. This gives use the l3_l4 flags.\n \t */\n-\tconst __m256i l3_l4_flags_shuf = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,\n-\t\t\t/* shift right 1 bit to make sure it not exceed 255 */\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t/* second 128-bits */\n-\t\t\t0, 0, 0, 0, 0, 0, 0, 0,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1);\n+\tconst __m256i l3_l4_flags_shuf =\n+\t\t_mm256_set_epi8((PKT_RX_OUTER_L4_CKSUM_BAD >> 20 |\n+\t\t PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t/**\n+\t\t * second 128-bits\n+\t\t * shift right 20 bits to use the low two bits to indicate\n+\t\t * outer checksum status\n+\t\t * shift right 1 bit to make sure it not exceed 255\n+\t\t */\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1);\n \tconst __m256i cksum_mask =\n-\t\t _mm256_set1_epi32(PKT_RX_IP_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD |\n-\t\t\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t\t PKT_RX_EIP_CKSUM_BAD);\n+\t\t _mm256_set1_epi32(PKT_RX_IP_CKSUM_MASK |\n+\t\t\t\t PKT_RX_L4_CKSUM_MASK |\n+\t\t\t\t PKT_RX_EIP_CKSUM_BAD |\n+\t\t\t\t PKT_RX_OUTER_L4_CKSUM_MASK);\n \t/**\n \t * data to be shuffled by result of flag mask, shifted down 12.\n \t * If RSS(bit12)/VLAN(bit13) are set,\n@@ -469,6 +514,15 @@ _ice_recv_raw_pkts_vec_avx2(struct ice_rx_queue *rxq, struct rte_mbuf **rx_pkts,\n \t\t__m256i l3_l4_flags = _mm256_shuffle_epi8(l3_l4_flags_shuf,\n \t\t\t\t_mm256_srli_epi32(flag_bits, 4));\n \t\tl3_l4_flags = _mm256_slli_epi32(l3_l4_flags, 1);\n+\n+\t\t__m256i l4_outer_mask = _mm256_set1_epi32(0x6);\n+\t\t__m256i l4_outer_flags =\n+\t\t\t\t_mm256_and_si256(l3_l4_flags, l4_outer_mask);\n+\t\tl4_outer_flags = _mm256_slli_epi32(l4_outer_flags, 20);\n+\n+\t\t__m256i l3_l4_mask = _mm256_set1_epi32(~0x6);\n+\t\tl3_l4_flags = _mm256_and_si256(l3_l4_flags, l3_l4_mask);\n+\t\tl3_l4_flags = _mm256_or_si256(l3_l4_flags, l4_outer_flags);\n \t\tl3_l4_flags = _mm256_and_si256(l3_l4_flags, cksum_mask);\n \t\t/* set rss and vlan flags */\n \t\tconst __m256i rss_vlan_flag_bits =\ndiff --git a/drivers/net/ice/ice_rxtx_vec_avx512.c b/drivers/net/ice/ice_rxtx_vec_avx512.c\nindex e5e7cc1482..d33ef2a042 100644\n--- a/drivers/net/ice/ice_rxtx_vec_avx512.c\n+++ b/drivers/net/ice/ice_rxtx_vec_avx512.c\n@@ -211,43 +211,88 @@ _ice_recv_raw_pkts_vec_avx512(struct ice_rx_queue *rxq,\n \t * bit13 is for VLAN indication.\n \t */\n \tconst __m256i flags_mask =\n-\t\t _mm256_set1_epi32((7 << 4) | (1 << 12) | (1 << 13));\n+\t\t _mm256_set1_epi32((0xF << 4) | (1 << 12) | (1 << 13));\n \t/**\n \t * data to be shuffled by the result of the flags mask shifted by 4\n \t * bits. This gives use the l3_l4 flags.\n \t */\n-\tconst __m256i l3_l4_flags_shuf = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,\n-\t\t\t/* shift right 1 bit to make sure it not exceed 255 */\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t/* 2nd 128-bits */\n-\t\t\t0, 0, 0, 0, 0, 0, 0, 0,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1);\n+\tconst __m256i l3_l4_flags_shuf =\n+\t\t_mm256_set_epi8((PKT_RX_OUTER_L4_CKSUM_BAD >> 20 |\n+\t\t PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t/**\n+\t\t * second 128-bits\n+\t\t * shift right 20 bits to use the low two bits to indicate\n+\t\t * outer checksum status\n+\t\t * shift right 1 bit to make sure it not exceed 255\n+\t\t */\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1);\n \tconst __m256i cksum_mask =\n-\t\t _mm256_set1_epi32(PKT_RX_IP_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD |\n-\t\t\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t\t PKT_RX_EIP_CKSUM_BAD);\n+\t\t _mm256_set1_epi32(PKT_RX_IP_CKSUM_MASK |\n+\t\t\t\t PKT_RX_L4_CKSUM_MASK |\n+\t\t\t\t PKT_RX_EIP_CKSUM_BAD |\n+\t\t\t\t PKT_RX_OUTER_L4_CKSUM_MASK);\n \t/**\n \t * data to be shuffled by result of flag mask, shifted down 12.\n \t * If RSS(bit12)/VLAN(bit13) are set,\n@@ -432,6 +477,14 @@ _ice_recv_raw_pkts_vec_avx512(struct ice_rx_queue *rxq,\n \t\t__m256i l3_l4_flags = _mm256_shuffle_epi8(l3_l4_flags_shuf,\n \t\t\t\t_mm256_srli_epi32(flag_bits, 4));\n \t\tl3_l4_flags = _mm256_slli_epi32(l3_l4_flags, 1);\n+\t\t__m256i l4_outer_mask = _mm256_set1_epi32(0x6);\n+\t\t__m256i l4_outer_flags =\n+\t\t\t\t_mm256_and_si256(l3_l4_flags, l4_outer_mask);\n+\t\tl4_outer_flags = _mm256_slli_epi32(l4_outer_flags, 20);\n+\n+\t\t__m256i l3_l4_mask = _mm256_set1_epi32(~0x6);\n+\t\tl3_l4_flags = _mm256_and_si256(l3_l4_flags, l3_l4_mask);\n+\t\tl3_l4_flags = _mm256_or_si256(l3_l4_flags, l4_outer_flags);\n \t\tl3_l4_flags = _mm256_and_si256(l3_l4_flags, cksum_mask);\n \t\t/* set rss and vlan flags */\n \t\tconst __m256i rss_vlan_flag_bits =\ndiff --git a/drivers/net/ice/ice_rxtx_vec_sse.c b/drivers/net/ice/ice_rxtx_vec_sse.c\nindex 626364719b..f8b3574e36 100644\n--- a/drivers/net/ice/ice_rxtx_vec_sse.c\n+++ b/drivers/net/ice/ice_rxtx_vec_sse.c\n@@ -114,39 +114,67 @@ ice_rx_desc_to_olflags_v(struct ice_rx_queue *rxq, __m128i descs[4],\n \t * bit12 for RSS indication.\n \t * bit13 for VLAN indication.\n \t */\n-\tconst __m128i desc_mask = _mm_set_epi32(0x3070, 0x3070,\n-\t\t\t\t\t\t0x3070, 0x3070);\n-\n+\tconst __m128i desc_mask = _mm_set_epi32(0x30f0, 0x30f0,\n+\t\t\t\t\t\t0x30f0, 0x30f0);\n \tconst __m128i cksum_mask = _mm_set_epi32(PKT_RX_IP_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_L4_CKSUM_MASK |\n+\t\t\t\t\t\t PKT_RX_OUTER_L4_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_EIP_CKSUM_BAD,\n \t\t\t\t\t\t PKT_RX_IP_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_L4_CKSUM_MASK |\n+\t\t\t\t\t\t PKT_RX_OUTER_L4_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_EIP_CKSUM_BAD,\n \t\t\t\t\t\t PKT_RX_IP_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_L4_CKSUM_MASK |\n+\t\t\t\t\t\t PKT_RX_OUTER_L4_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_EIP_CKSUM_BAD,\n \t\t\t\t\t\t PKT_RX_IP_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_L4_CKSUM_MASK |\n+\t\t\t\t\t\t PKT_RX_OUTER_L4_CKSUM_MASK |\n \t\t\t\t\t\t PKT_RX_EIP_CKSUM_BAD);\n \n \t/* map the checksum, rss and vlan fields to the checksum, rss\n \t * and vlan flag\n \t */\n-\tconst __m128i cksum_flags = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,\n-\t\t\t/* shift right 1 bit to make sure it not exceed 255 */\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |\n-\t\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n-\t\t\t(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1);\n+\tconst __m128i cksum_flags =\n+\t\t_mm_set_epi8((PKT_RX_OUTER_L4_CKSUM_BAD >> 20 |\n+\t\t PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_BAD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t/**\n+\t\t * shift right 20 bits to use the low two bits to indicate\n+\t\t * outer checksum status\n+\t\t * shift right 1 bit to make sure it not exceed 255\n+\t\t */\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_EIP_CKSUM_BAD |\n+\t\t PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_BAD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_BAD) >> 1,\n+\t\t(PKT_RX_OUTER_L4_CKSUM_GOOD >> 20 | PKT_RX_L4_CKSUM_GOOD |\n+\t\t PKT_RX_IP_CKSUM_GOOD) >> 1);\n \n \tconst __m128i rss_vlan_flags = _mm_set_epi8(0, 0, 0, 0,\n \t\t\t0, 0, 0, 0,\n@@ -166,6 +194,14 @@ ice_rx_desc_to_olflags_v(struct ice_rx_queue *rxq, __m128i descs[4],\n \tflags = _mm_shuffle_epi8(cksum_flags, tmp_desc);\n \t/* then we shift left 1 bit */\n \tflags = _mm_slli_epi32(flags, 1);\n+\n+\t__m128i l4_outer_mask = _mm_set_epi32(0x6, 0x6, 0x6, 0x6);\n+\t__m128i l4_outer_flags = _mm_and_si128(flags, l4_outer_mask);\n+\tl4_outer_flags = _mm_slli_epi32(l4_outer_flags, 20);\n+\n+\t__m128i l3_l4_mask = _mm_set_epi32(~0x6, ~0x6, ~0x6, ~0x6);\n+\t__m128i l3_l4_flags = _mm_and_si128(flags, l3_l4_mask);\n+\tflags = _mm_or_si128(l3_l4_flags, l4_outer_flags);\n \t/* we need to mask out the reduntant bits introduced by RSS or\n \t * VLAN fields.\n \t */\n@@ -217,10 +253,10 @@ ice_rx_desc_to_olflags_v(struct ice_rx_queue *rxq, __m128i descs[4],\n \t * appropriate flags means that we have to do a shift and blend for\n \t * each mbuf before we do the write.\n \t */\n-\trearm0 = _mm_blend_epi16(mbuf_init, _mm_slli_si128(flags, 8), 0x10);\n-\trearm1 = _mm_blend_epi16(mbuf_init, _mm_slli_si128(flags, 4), 0x10);\n-\trearm2 = _mm_blend_epi16(mbuf_init, flags, 0x10);\n-\trearm3 = _mm_blend_epi16(mbuf_init, _mm_srli_si128(flags, 4), 0x10);\n+\trearm0 = _mm_blend_epi32(mbuf_init, _mm_slli_si128(flags, 8), 0x04);\n+\trearm1 = _mm_blend_epi32(mbuf_init, _mm_slli_si128(flags, 4), 0x04);\n+\trearm2 = _mm_blend_epi32(mbuf_init, flags, 0x04);\n+\trearm3 = _mm_blend_epi32(mbuf_init, _mm_srli_si128(flags, 4), 0x04);\n \n \t/* write the rearm data and the olflags in one write */\n \tRTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, ol_flags) !=\n", "prefixes": [ "v6" ] }