From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga03.intel.com (mga03.intel.com [134.134.136.65]) by dpdk.org (Postfix) with ESMTP id 2FE0CB66F for ; Tue, 17 Feb 2015 03:57:28 +0100 (CET) Received: from fmsmga002.fm.intel.com ([10.253.24.26]) by orsmga103.jf.intel.com with ESMTP; 16 Feb 2015 18:52:08 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.09,591,1418112000"; d="scan'208";a="678918211" Received: from kmsmsx151.gar.corp.intel.com ([172.21.73.86]) by fmsmga002.fm.intel.com with ESMTP; 16 Feb 2015 18:57:24 -0800 Received: from shsmsx152.ccr.corp.intel.com (10.239.6.52) by KMSMSX151.gar.corp.intel.com (172.21.73.86) with Microsoft SMTP Server (TLS) id 14.3.195.1; Tue, 17 Feb 2015 10:57:23 +0800 Received: from shsmsx104.ccr.corp.intel.com ([169.254.5.161]) by SHSMSX152.ccr.corp.intel.com ([169.254.6.46]) with mapi id 14.03.0195.001; Tue, 17 Feb 2015 10:57:22 +0800 From: "Zhang, Helin" To: "Ananyev, Konstantin" , "dev@dpdk.org" Thread-Topic: [PATCH v2 14/15] examples/l3fwd: support of unified packet type Thread-Index: AQHQRDNvWMo0ZAy3fEmNuWlP3ZuU8ZzziPzggACpAjA= Date: Tue, 17 Feb 2015 02:57:21 +0000 Message-ID: References: <1422501365-12643-1-git-send-email-helin.zhang@intel.com> <1423464049-13457-1-git-send-email-helin.zhang@intel.com> <1423464049-13457-15-git-send-email-helin.zhang@intel.com> <2601191342CEEE43887BDE71AB977258213EDD8E@irsmsx105.ger.corp.intel.com> In-Reply-To: <2601191342CEEE43887BDE71AB977258213EDD8E@irsmsx105.ger.corp.intel.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.239.127.40] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-dev] [PATCH v2 14/15] examples/l3fwd: support of unified packet type X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 17 Feb 2015 02:57:28 -0000 > -----Original Message----- > From: Ananyev, Konstantin > Sent: Tuesday, February 17, 2015 1:05 AM > To: Zhang, Helin; dev@dpdk.org > Cc: Cao, Waterman; Liang, Cunming; Liu, Jijiang; Richardson, Bruce > Subject: RE: [PATCH v2 14/15] examples/l3fwd: support of unified packet t= ype >=20 > Hi Helin, >=20 > > -----Original Message----- > > From: Zhang, Helin > > Sent: Monday, February 09, 2015 6:41 AM > > To: dev@dpdk.org > > Cc: Cao, Waterman; Liang, Cunming; Liu, Jijiang; Ananyev, Konstantin; > > Richardson, Bruce; Zhang, Helin > > Subject: [PATCH v2 14/15] examples/l3fwd: support of unified packet > > type > > > > To unify packet types among all PMDs, bit masks and relevant macros of > > packet type for ol_flags are replaced by unified packet type and > > relevant macros. > > > > Signed-off-by: Helin Zhang > > --- > > examples/l3fwd/main.c | 64 > > ++++++++++++++++++++++++++++----------------------- > > 1 file changed, 35 insertions(+), 29 deletions(-) > > > > v2 changes: > > * Used redefined packet types and enlarged packet_type field in mbuf. > > > > diff --git a/examples/l3fwd/main.c b/examples/l3fwd/main.c index > > 6f7d7d4..302322e 100644 > > --- a/examples/l3fwd/main.c > > +++ b/examples/l3fwd/main.c > > @@ -958,7 +958,7 @@ l3fwd_simple_forward(struct rte_mbuf *m, uint8_t > > portid, struct lcore_conf *qcon > > > > eth_hdr =3D rte_pktmbuf_mtod(m, struct ether_hdr *); > > > > - if (m->ol_flags & PKT_RX_IPV4_HDR) { > > + if (RTE_ETH_IS_IPV4_HDR(m->packet_type)) { > > /* Handle IPv4 headers.*/ > > ipv4_hdr =3D (struct ipv4_hdr *)(rte_pktmbuf_mtod(m, unsigned char *= ) > + > > sizeof(struct ether_hdr)); > > @@ -993,7 +993,7 @@ l3fwd_simple_forward(struct rte_mbuf *m, uint8_t > > portid, struct lcore_conf *qcon > > > > send_single_packet(m, dst_port); > > > > - } else { > > + } else if (RTE_ETH_IS_IPV6_HDR(m->packet_type)) { >=20 > If you changed to from 'else' to ' else if' here, then I suppose you'll n= eed to add > another 'else' after it: > to handle case, where input packets are neither IPV4 neither IPv6. > Otherwise you might start 'leaking' such mbufs. Agree with you, will add code to free mbuf there. >=20 > > /* Handle IPv6 headers.*/ > > struct ipv6_hdr *ipv6_hdr; > > > > @@ -1039,11 +1039,11 @@ l3fwd_simple_forward(struct rte_mbuf *m, > uint8_t portid, struct lcore_conf *qcon > > * to BAD_PORT value. > > */ > > static inline __attribute__((always_inline)) void > > -rfc1812_process(struct ipv4_hdr *ipv4_hdr, uint16_t *dp, uint32_t > > flags) > > +rfc1812_process(struct ipv4_hdr *ipv4_hdr, uint16_t *dp, uint16_t > > +ptype) >=20 > Shouldn't it be 'uint32_t ptype'? Agree with you. Will correct it. >=20 > > { > > uint8_t ihl; > > > > - if ((flags & PKT_RX_IPV4_HDR) !=3D 0) { > > + if (RTE_ETH_IS_IPV4_HDR(ptype)) { > > > > ihl =3D ipv4_hdr->version_ihl - IPV4_MIN_VER_IHL; > > > > @@ -1074,11 +1074,11 @@ get_dst_port(const struct lcore_conf *qconf, > struct rte_mbuf *pkt, > > struct ipv6_hdr *ipv6_hdr; > > struct ether_hdr *eth_hdr; > > > > - if (pkt->ol_flags & PKT_RX_IPV4_HDR) { > > + if (RTE_ETH_IS_IPV4_HDR(pkt->packet_type)) { > > if (rte_lpm_lookup(qconf->ipv4_lookup_struct, dst_ipv4, > > &next_hop) !=3D 0) > > next_hop =3D portid; > > - } else if (pkt->ol_flags & PKT_RX_IPV6_HDR) { > > + } else if (RTE_ETH_IS_IPV6_HDR(pkt->packet_type)) { > > eth_hdr =3D rte_pktmbuf_mtod(pkt, struct ether_hdr *); > > ipv6_hdr =3D (struct ipv6_hdr *)(eth_hdr + 1); > > if (rte_lpm6_lookup(qconf->ipv6_lookup_struct, > > @@ -1112,17 +1112,19 @@ process_packet(struct lcore_conf *qconf, struct > rte_mbuf *pkt, > > ve =3D val_eth[dp]; > > > > dst_port[0] =3D dp; > > - rfc1812_process(ipv4_hdr, dst_port, pkt->ol_flags); > > + rfc1812_process(ipv4_hdr, dst_port, pkt->packet_type); > > > > te =3D _mm_blend_epi16(te, ve, MASK_ETH); > > _mm_store_si128((__m128i *)eth_hdr, te); } > > > > /* > > - * Read ol_flags and destination IPV4 addresses from 4 mbufs. > > + * Read packet_type and destination IPV4 addresses from 4 mbufs. > > */ > > static inline void > > -processx4_step1(struct rte_mbuf *pkt[FWDSTEP], __m128i *dip, uint32_t > > *flag) > > +processx4_step1(struct rte_mbuf *pkt[FWDSTEP], > > + __m128i *dip, > > + uint32_t *ipv4_flag) > > { > > struct ipv4_hdr *ipv4_hdr; > > struct ether_hdr *eth_hdr; > > @@ -1131,22 +1133,20 @@ processx4_step1(struct rte_mbuf > *pkt[FWDSTEP], __m128i *dip, uint32_t *flag) > > eth_hdr =3D rte_pktmbuf_mtod(pkt[0], struct ether_hdr *); > > ipv4_hdr =3D (struct ipv4_hdr *)(eth_hdr + 1); > > x0 =3D ipv4_hdr->dst_addr; > > - flag[0] =3D pkt[0]->ol_flags & PKT_RX_IPV4_HDR; > > > > eth_hdr =3D rte_pktmbuf_mtod(pkt[1], struct ether_hdr *); > > ipv4_hdr =3D (struct ipv4_hdr *)(eth_hdr + 1); > > x1 =3D ipv4_hdr->dst_addr; > > - flag[0] &=3D pkt[1]->ol_flags; > > > > eth_hdr =3D rte_pktmbuf_mtod(pkt[2], struct ether_hdr *); > > ipv4_hdr =3D (struct ipv4_hdr *)(eth_hdr + 1); > > x2 =3D ipv4_hdr->dst_addr; > > - flag[0] &=3D pkt[2]->ol_flags; > > > > eth_hdr =3D rte_pktmbuf_mtod(pkt[3], struct ether_hdr *); > > ipv4_hdr =3D (struct ipv4_hdr *)(eth_hdr + 1); > > x3 =3D ipv4_hdr->dst_addr; > > - flag[0] &=3D pkt[3]->ol_flags; > > + *ipv4_flag =3D pkt[0]->packet_type & pkt[1]->packet_type & > > + pkt[2]->packet_type & pkt[3]->packet_type & RTE_PTYPE_L3_IPV4; >=20 > Why not as it was before: > flag[0] =3D pkt[0]->packet-type & ... > ... > flag[0] &=3D pkt[1]->packet_type; > ... >=20 > Why do you need to unite them? No specific reason, will changed it as before. Thanks! Regards, Helin >=20 > > > > dip[0] =3D _mm_set_epi32(x3, x2, x1, x0); } @@ -1156,8 +1156,12 @@ > > processx4_step1(struct rte_mbuf *pkt[FWDSTEP], __m128i *dip, uint32_t > *flag) > > * If lookup fails, use incoming port (portid) as destination port. > > */ > > static inline void > > -processx4_step2(const struct lcore_conf *qconf, __m128i dip, uint32_t = flag, > > - uint8_t portid, struct rte_mbuf *pkt[FWDSTEP], uint16_t dprt[FWDSTEP]= ) > > +processx4_step2(const struct lcore_conf *qconf, > > + __m128i dip, > > + uint32_t ipv4_flag, > > + uint8_t portid, > > + struct rte_mbuf *pkt[FWDSTEP], > > + uint16_t dprt[FWDSTEP]) > > { > > rte_xmm_t dst; > > const __m128i bswap_mask =3D _mm_set_epi8(12, 13, 14, 15, 8, 9, 10, > > 11, @@ -1167,7 +1171,7 @@ processx4_step2(const struct lcore_conf > *qconf, __m128i dip, uint32_t flag, > > dip =3D _mm_shuffle_epi8(dip, bswap_mask); > > > > /* if all 4 packets are IPV4. */ > > - if (likely(flag !=3D 0)) { > > + if (likely(ipv4_flag)) { > > rte_lpm_lookupx4(qconf->ipv4_lookup_struct, dip, dprt, portid); > > } else { > > dst.x =3D dip; > > @@ -1218,13 +1222,13 @@ processx4_step3(struct rte_mbuf > *pkt[FWDSTEP], uint16_t dst_port[FWDSTEP]) > > _mm_store_si128(p[3], te[3]); > > > > rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[0] + 1), > > - &dst_port[0], pkt[0]->ol_flags); > > + &dst_port[0], pkt[0]->packet_type); > > rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[1] + 1), > > - &dst_port[1], pkt[1]->ol_flags); > > + &dst_port[1], pkt[1]->packet_type); > > rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[2] + 1), > > - &dst_port[2], pkt[2]->ol_flags); > > + &dst_port[2], pkt[2]->packet_type); > > rfc1812_process((struct ipv4_hdr *)((struct ether_hdr *)p[3] + 1), > > - &dst_port[3], pkt[3]->ol_flags); > > + &dst_port[3], pkt[3]->packet_type); > > } > > > > /* > > @@ -1411,7 +1415,7 @@ main_loop(__attribute__((unused)) void *dummy) > > uint16_t *lp; > > uint16_t dst_port[MAX_PKT_BURST]; > > __m128i dip[MAX_PKT_BURST / FWDSTEP]; > > - uint32_t flag[MAX_PKT_BURST / FWDSTEP]; > > + uint32_t ipv4_flag[MAX_PKT_BURST / FWDSTEP]; > > uint16_t pnum[MAX_PKT_BURST + 1]; > > #endif > > > > @@ -1481,14 +1485,16 @@ main_loop(__attribute__((unused)) void > *dummy) > > */ > > int32_t n =3D RTE_ALIGN_FLOOR(nb_rx, 4); > > for (j =3D 0; j < n ; j+=3D4) { > > - uint32_t ol_flag =3D pkts_burst[j]->ol_flags > > - & pkts_burst[j+1]->ol_flags > > - & pkts_burst[j+2]->ol_flags > > - & pkts_burst[j+3]->ol_flags; > > - if (ol_flag & PKT_RX_IPV4_HDR ) { > > + uint32_t pkt_type =3D > > + pkts_burst[j]->packet_type & > > + pkts_burst[j+1]->packet_type & > > + pkts_burst[j+2]->packet_type & > > + pkts_burst[j+3]->packet_type; > > + if (pkt_type & RTE_PTYPE_L3_IPV4) { > > simple_ipv4_fwd_4pkts(&pkts_burst[j], > > portid, qconf); > > - } else if (ol_flag & PKT_RX_IPV6_HDR) { > > + } else if (pkt_type & > > + RTE_PTYPE_L3_IPV6) { > > simple_ipv6_fwd_4pkts(&pkts_burst[j], > > portid, qconf); > > } else { > > @@ -1513,13 +1519,13 @@ main_loop(__attribute__((unused)) void > *dummy) > > for (j =3D 0; j !=3D k; j +=3D FWDSTEP) { > > processx4_step1(&pkts_burst[j], > > &dip[j / FWDSTEP], > > - &flag[j / FWDSTEP]); > > + &ipv4_flag[j / FWDSTEP]); > > } > > > > k =3D RTE_ALIGN_FLOOR(nb_rx, FWDSTEP); > > for (j =3D 0; j !=3D k; j +=3D FWDSTEP) { > > processx4_step2(qconf, dip[j / FWDSTEP], > > - flag[j / FWDSTEP], portid, > > + ipv4_flag[j / FWDSTEP], portid, > > &pkts_burst[j], &dst_port[j]); > > } > > > > -- > > 1.9.3