From mboxrd@z Thu Jan 1 00:00:00 1970
Return-Path: <jingche2@shecgisg004.sh.intel.com>
Received: from mga01.intel.com (mga01.intel.com [192.55.52.88])
by dpdk.org (Postfix) with ESMTP id A3C4D8E6D
for <dev@dpdk.org>; Tue, 27 Oct 2015 10:47:17 +0100 (CET)
Received: from fmsmga002.fm.intel.com ([10.253.24.26])
by fmsmga101.fm.intel.com with ESMTP; 27 Oct 2015 02:47:16 -0700
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.20,204,1444719600"; d="scan'208";a="836416425"
Received: from shvmail01.sh.intel.com ([10.239.29.42])
by fmsmga002.fm.intel.com with ESMTP; 27 Oct 2015 02:47:08 -0700
Received: from shecgisg004.sh.intel.com (shecgisg004.sh.intel.com
[10.239.29.89])
by shvmail01.sh.intel.com with ESMTP id t9R9l4ZV030874;
Tue, 27 Oct 2015 17:47:04 +0800
Received: from shecgisg004.sh.intel.com (localhost [127.0.0.1])
by shecgisg004.sh.intel.com (8.13.6/8.13.6/SuSE Linux 0.8) with ESMTP id
t9R9l2Wd012892; Tue, 27 Oct 2015 17:47:04 +0800
Received: (from jingche2@localhost)
by shecgisg004.sh.intel.com (8.13.6/8.13.6/Submit) id t9R9l27N012848;
Tue, 27 Oct 2015 17:47:02 +0800
From: "Chen Jing D(Mark)" <jing.d.chen@intel.com>
To: dev@dpdk.org
Date: Tue, 27 Oct 2015 17:46:38 +0800
Message-Id: <1445939209-12783-6-git-send-email-jing.d.chen@intel.com>
X-Mailer: git-send-email 1.7.12.2
In-Reply-To: <1445939209-12783-1-git-send-email-jing.d.chen@intel.com>
References: <1445507104-22563-2-git-send-email-jing.d.chen@intel.com>
<1445939209-12783-1-git-send-email-jing.d.chen@intel.com>
Subject: [dpdk-dev] [PATCH v3 05/16] fm10k: add 2 functions to parse
pkt_type and offload flag
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>,
<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>,
<mailto:dev-request@dpdk.org?subject=subscribe>
X-List-Received-Date: Tue, 27 Oct 2015 09:47:18 -0000
From: "Chen Jing D(Mark)" <jing.d.chen@intel.com>
Add 2 functions, in which using SSE instructions to parse RX desc
to get pkt_type and ol_flags in mbuf.
Signed-off-by: Chen Jing D(Mark) <jing.d.chen@intel.com>
---
drivers/net/fm10k/fm10k_rxtx_vec.c | 127 ++++++++++++++++++++++++++++++++++++
1 files changed, 127 insertions(+), 0 deletions(-)
diff --git a/drivers/net/fm10k/fm10k_rxtx_vec.c b/drivers/net/fm10k/fm10k_rxtx_vec.c
index 75533f9..581a309 100644
--- a/drivers/net/fm10k/fm10k_rxtx_vec.c
+++ b/drivers/net/fm10k/fm10k_rxtx_vec.c
@@ -44,6 +44,133 @@
#pragma GCC diagnostic ignored "-Wcast-qual"
#endif
+/* Handling the offload flags (olflags) field takes computation
+ * time when receiving packets. Therefore we provide a flag to disable
+ * the processing of the olflags field when they are not needed. This
+ * gives improved performance, at the cost of losing the offload info
+ * in the received packet
+ */
+#ifdef RTE_LIBRTE_FM10K_RX_OLFLAGS_ENABLE
+
+/* Vlan present flag shift */
+#define VP_SHIFT (2)
+/* L3 type shift */
+#define L3TYPE_SHIFT (4)
+/* L4 type shift */
+#define L4TYPE_SHIFT (7)
+
+static inline void
+fm10k_desc_to_olflags_v(__m128i descs[4], struct rte_mbuf **rx_pkts)
+{
+ __m128i ptype0, ptype1, vtag0, vtag1;
+ union {
+ uint16_t e[4];
+ uint64_t dword;
+ } vol;
+
+ const __m128i pkttype_msk = _mm_set_epi16(
+ 0x0000, 0x0000, 0x0000, 0x0000,
+ PKT_RX_VLAN_PKT, PKT_RX_VLAN_PKT,
+ PKT_RX_VLAN_PKT, PKT_RX_VLAN_PKT);
+
+ /* mask everything except rss type */
+ const __m128i rsstype_msk = _mm_set_epi16(
+ 0x0000, 0x0000, 0x0000, 0x0000,
+ 0x000F, 0x000F, 0x000F, 0x000F);
+
+ /* map rss type to rss hash flag */
+ const __m128i rss_flags = _mm_set_epi8(0, 0, 0, 0,
+ 0, 0, 0, PKT_RX_RSS_HASH,
+ PKT_RX_RSS_HASH, 0, PKT_RX_RSS_HASH, 0,
+ PKT_RX_RSS_HASH, PKT_RX_RSS_HASH, PKT_RX_RSS_HASH, 0);
+
+ ptype0 = _mm_unpacklo_epi16(descs[0], descs[1]);
+ ptype1 = _mm_unpacklo_epi16(descs[2], descs[3]);
+ vtag0 = _mm_unpackhi_epi16(descs[0], descs[1]);
+ vtag1 = _mm_unpackhi_epi16(descs[2], descs[3]);
+
+ ptype0 = _mm_unpacklo_epi32(ptype0, ptype1);
+ ptype0 = _mm_and_si128(ptype0, rsstype_msk);
+ ptype0 = _mm_shuffle_epi8(rss_flags, ptype0);
+
+ vtag1 = _mm_unpacklo_epi32(vtag0, vtag1);
+ vtag1 = _mm_srli_epi16(vtag1, VP_SHIFT);
+ vtag1 = _mm_and_si128(vtag1, pkttype_msk);
+
+ vtag1 = _mm_or_si128(ptype0, vtag1);
+ vol.dword = _mm_cvtsi128_si64(vtag1);
+
+ rx_pkts[0]->ol_flags = vol.e[0];
+ rx_pkts[1]->ol_flags = vol.e[1];
+ rx_pkts[2]->ol_flags = vol.e[2];
+ rx_pkts[3]->ol_flags = vol.e[3];
+}
+
+static inline void
+fm10k_desc_to_pktype_v(__m128i descs[4], struct rte_mbuf **rx_pkts)
+{
+ __m128i l3l4type0, l3l4type1, l3type, l4type;
+ union {
+ uint16_t e[4];
+ uint64_t dword;
+ } vol;
+
+ /* L3 pkt type mask Bit4 to Bit6 */
+ const __m128i l3type_msk = _mm_set_epi16(
+ 0x0000, 0x0000, 0x0000, 0x0000,
+ 0x0070, 0x0070, 0x0070, 0x0070);
+
+ /* L4 pkt type mask Bit7 to Bit9 */
+ const __m128i l4type_msk = _mm_set_epi16(
+ 0x0000, 0x0000, 0x0000, 0x0000,
+ 0x0380, 0x0380, 0x0380, 0x0380);
+
+ /* convert RRC l3 type to mbuf format */
+ const __m128i l3type_flags = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, RTE_PTYPE_L3_IPV6_EXT,
+ RTE_PTYPE_L3_IPV6, RTE_PTYPE_L3_IPV4_EXT,
+ RTE_PTYPE_L3_IPV4, 0);
+
+ /* Convert RRC l4 type to mbuf format l4type_flags shift-left 8 bits
+ * to fill into8 bits length.
+ */
+ const __m128i l4type_flags = _mm_set_epi8(0, 0, 0, 0, 0, 0, 0, 0, 0,
+ RTE_PTYPE_TUNNEL_GENEVE >> 8,
+ RTE_PTYPE_TUNNEL_NVGRE >> 8,
+ RTE_PTYPE_TUNNEL_VXLAN >> 8,
+ RTE_PTYPE_TUNNEL_GRE >> 8,
+ RTE_PTYPE_L4_UDP >> 8,
+ RTE_PTYPE_L4_TCP >> 8,
+ 0);
+
+ l3l4type0 = _mm_unpacklo_epi16(descs[0], descs[1]);
+ l3l4type1 = _mm_unpacklo_epi16(descs[2], descs[3]);
+ l3l4type0 = _mm_unpacklo_epi32(l3l4type0, l3l4type1);
+
+ l3type = _mm_and_si128(l3l4type0, l3type_msk);
+ l4type = _mm_and_si128(l3l4type0, l4type_msk);
+
+ l3type = _mm_srli_epi16(l3type, L3TYPE_SHIFT);
+ l4type = _mm_srli_epi16(l4type, L4TYPE_SHIFT);
+
+ l3type = _mm_shuffle_epi8(l3type_flags, l3type);
+ /* l4type_flags shift-left for 8 bits, need shift-right back */
+ l4type = _mm_shuffle_epi8(l4type_flags, l4type);
+
+ l4type = _mm_slli_epi16(l4type, 8);
+ l3l4type0 = _mm_or_si128(l3type, l4type);
+ vol.dword = _mm_cvtsi128_si64(l3l4type0);
+
+ rx_pkts[0]->packet_type = vol.e[0];
+ rx_pkts[1]->packet_type = vol.e[1];
+ rx_pkts[2]->packet_type = vol.e[2];
+ rx_pkts[3]->packet_type = vol.e[3];
+}
+#else
+#define fm10k_desc_to_olflags_v(desc, rx_pkts) do {} while (0)
+#define fm10k_desc_to_pktype_v(desc, rx_pkts) do {} while (0)
+#endif
+
int __attribute__((cold))
fm10k_rxq_vec_setup(struct fm10k_rx_queue *rxq)
{
--
1.7.7.6