From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by dpdk.org (Postfix) with ESMTP id B860E7E07 for ; Fri, 26 Sep 2014 17:07:58 +0200 (CEST) Received: from orsmga002.jf.intel.com ([10.7.209.21]) by orsmga101.jf.intel.com with ESMTP; 26 Sep 2014 08:13:09 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.04,604,1406617200"; d="scan'208";a="609145910" Received: from irsmsx101.ger.corp.intel.com ([163.33.3.153]) by orsmga002.jf.intel.com with ESMTP; 26 Sep 2014 08:11:59 -0700 Received: from irsmsx105.ger.corp.intel.com (163.33.3.28) by IRSMSX101.ger.corp.intel.com (163.33.3.153) with Microsoft SMTP Server (TLS) id 14.3.195.1; Fri, 26 Sep 2014 16:11:45 +0100 Received: from irsmsx108.ger.corp.intel.com ([169.254.11.21]) by IRSMSX105.ger.corp.intel.com ([169.254.7.57]) with mapi id 14.03.0195.001; Fri, 26 Sep 2014 16:11:45 +0100 From: "De Lara Guarch, Pablo" To: "Pattan, Reshma" , "dev@dpdk.org" Thread-Topic: [dpdk-dev] [PATCH v2] distributor_app: new sample app Thread-Index: AQHP2AJi1zXXgpMcD0+yx1ATRe0wtpwThu2Q Date: Fri, 26 Sep 2014 15:11:44 +0000 Message-ID: References: <1410869607-16842-1-git-send-email-reshma.pattan@intel.com> <1411568210-2555-1-git-send-email-reshma.pattan@intel.com> In-Reply-To: <1411568210-2555-1-git-send-email-reshma.pattan@intel.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [163.33.239.182] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-dev] [PATCH v2] distributor_app: new sample app 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: Fri, 26 Sep 2014 15:08:00 -0000 Hi, > -----Original Message----- > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of reshmapa > Sent: Wednesday, September 24, 2014 3:17 PM > To: dev@dpdk.org > Subject: [dpdk-dev] [PATCH v2] distributor_app: new sample app >=20 > From: Reshma Pattan >=20 > A new sample app that shows the usage of the distributor library. This > app works as follows: >=20 > * An RX thread runs which pulls packets from each ethernet port in turn > and passes those packets to worker using a distributor component. > * The workers take the packets in turn, and determine the output port > for those packets using basic l2forwarding doing an xor on the source > port id. > * The RX thread takes the returned packets from the workers and enqueue > those packets into an rte_ring structure. > * A TX thread pulls the packets off the rte_ring structure and then > sends each packet out the output port specified previously by the worke= r > * Command-line option support provided only for portmask. >=20 > Signed-off-by: Bruce Richardson > Signed-off-by: Reshma Pattan > --- > examples/Makefile | 1 + > examples/distributor_app/Makefile | 57 ++++ > examples/distributor_app/main.c | 585 > ++++++++++++++++++++++++++++++++++++++ > examples/distributor_app/main.h | 46 +++ > 4 files changed, 689 insertions(+) > create mode 100644 examples/distributor_app/Makefile > create mode 100644 examples/distributor_app/main.c > create mode 100644 examples/distributor_app/main.h >=20 > diff --git a/examples/Makefile b/examples/Makefile > index 6245f83..2ba82b0 100644 > --- a/examples/Makefile > +++ b/examples/Makefile > @@ -66,5 +66,6 @@ DIRS-y +=3D vhost > DIRS-$(CONFIG_RTE_LIBRTE_XEN_DOM0) +=3D vhost_xen > DIRS-y +=3D vmdq > DIRS-y +=3D vmdq_dcb > +DIRS-$(CONFIG_RTE_LIBRTE_DISTRIBUTOR) +=3D distributor_app >=20 > include $(RTE_SDK)/mk/rte.extsubdir.mk > diff --git a/examples/distributor_app/Makefile > b/examples/distributor_app/Makefile > new file mode 100644 > index 0000000..394785d > --- /dev/null > +++ b/examples/distributor_app/Makefile > @@ -0,0 +1,57 @@ > +# BSD LICENSE > +# > +# Copyright(c) 2010-2014 Intel Corporation. All rights reserved. > +# All rights reserved. > +# > +# Redistribution and use in source and binary forms, with or without > +# modification, are permitted provided that the following conditions > +# are met: > +# > +# * Redistributions of source code must retain the above copyright > +# notice, this list of conditions and the following disclaimer. > +# * Redistributions in binary form must reproduce the above copyrigh= t > +# notice, this list of conditions and the following disclaimer in > +# the documentation and/or other materials provided with the > +# distribution. > +# * Neither the name of Intel Corporation nor the names of its > +# contributors may be used to endorse or promote products derived > +# from this software without specific prior written permission. > +# > +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND > CONTRIBUTORS > +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT > NOT > +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND > FITNESS FOR > +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE > COPYRIGHT > +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, > INCIDENTAL, > +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT > NOT > +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS > OF USE, > +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED > AND ON ANY > +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR > TORT > +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF > THE USE > +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH > DAMAGE. > + > +ifeq ($(RTE_SDK),) > +$(error "Please define RTE_SDK environment variable") > +endif > + > +# Default target, can be overriden by command line or environment > +RTE_TARGET ?=3D x86_64-default-linuxapp-gcc This target is not present anymore. Change it to x86_64-native-linuxapp-gcc= . > + > +include $(RTE_SDK)/mk/rte.vars.mk > + > +# binary name > +APP =3D distributor_app > + > +# all source are stored in SRCS-y > +SRCS-y :=3D main.c > + > +CFLAGS +=3D $(WERROR_FLAGS) > + > +# workaround for a gcc bug with noreturn attribute > +# http://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D12603 > +ifeq ($(CONFIG_RTE_TOOLCHAIN_GCC),y) > +CFLAGS_main.o +=3D -Wno-return-type > +endif > + > +EXTRA_CFLAGS +=3D -O3 -Wfatal-errors > + > +include $(RTE_SDK)/mk/rte.extapp.mk > diff --git a/examples/distributor_app/main.c > b/examples/distributor_app/main.c > new file mode 100644 > index 0000000..628810a > --- /dev/null > +++ b/examples/distributor_app/main.c > @@ -0,0 +1,585 @@ > +/*- > + * BSD LICENSE > + * > + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved. > + * All rights reserved. > + * > + * Redistribution and use in source and binary forms, with or without > + * modification, are permitted provided that the following conditions > + * are met: > + * > + * * Redistributions of source code must retain the above copyright > + * notice, this list of conditions and the following disclaimer. > + * * Redistributions in binary form must reproduce the above copyrig= ht > + * notice, this list of conditions and the following disclaimer in > + * the documentation and/or other materials provided with the > + * distribution. > + * * Neither the name of Intel Corporation nor the names of its > + * contributors may be used to endorse or promote products derived > + * from this software without specific prior written permission. > + * > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND > CONTRIBUTORS > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT > NOT > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND > FITNESS FOR > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE > COPYRIGHT > + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, > INCIDENTAL, > + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT > NOT > + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS > OF USE, > + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED > AND ON ANY > + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR > TORT > + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF > THE USE > + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH > DAMAGE. > + */ > + > +#include > +#include > +#include > +#include > +#include > + > +#include > +#include > +#include > +#include > +#include > +#include > + > +#include "main.h" > + > +#define RX_RING_SIZE 256 > +#define RX_FREE_THRESH 32 > +#define RX_PTHRESH 8 > +#define RX_HTHRESH 8 > +#define RX_WTHRESH 0 > + > +#define TX_RING_SIZE 512 > +#define TX_FREE_THRESH 32 > +#define TX_PTHRESH 32 > +#define TX_HTHRESH 0 > +#define TX_WTHRESH 0 > +#define TX_RSBIT_THRESH 32 > +#define TX_Q_FLAGS (ETH_TXQ_FLAGS_NOMULTSEGS | > ETH_TXQ_FLAGS_NOVLANOFFL |\ > + ETH_TXQ_FLAGS_NOXSUMSCTP | ETH_TXQ_FLAGS_NOXSUMUDP | > \ > + ETH_TXQ_FLAGS_NOXSUMTCP) > + > +#define NUM_MBUFS ((64*1024)-1) > +#define MBUF_SIZE (2048 + sizeof(struct rte_mbuf) + > RTE_PKTMBUF_HEADROOM) > +#define MBUF_CACHE_SIZE 250 > +#define BURST_SIZE 32 > +#define RTE_RING_SZ 1024 > + > +/* mask of enabled ports */ > +static uint32_t enabled_port_mask =3D 0; > + > +static volatile struct app_stats { > + struct { > + uint64_t rx_pkts; > + uint64_t returned_pkts; > + uint64_t enqueued_pkts; > + } rx __rte_cache_aligned; > + > + struct { > + uint64_t dequeue_pkts; > + uint64_t tx_pkts; > + } tx __rte_cache_aligned; > +} app_stats; > + > +static const struct rte_eth_conf port_conf_default =3D { > + .rxmode =3D { > + .mq_mode =3D ETH_MQ_RX_RSS, > + .max_rx_pkt_len =3D ETHER_MAX_LEN, > + .split_hdr_size =3D 0, > + .header_split =3D 0, /**< Header Split disabled */ > + .hw_ip_checksum =3D 0, /**< IP checksum offload enabled */ > + .hw_vlan_filter =3D 0, /**< VLAN filtering disabled */ > + .jumbo_frame =3D 0, /**< Jumbo Frame Support disabled */ > + .hw_strip_crc =3D 0, /**< CRC stripped by hardware */ > + }, > + .txmode =3D { > + .mq_mode =3D ETH_MQ_TX_NONE, > + }, > + .lpbk_mode =3D 0, > + .rx_adv_conf =3D { > + .rss_conf =3D { > + .rss_hf =3D ETH_RSS_IPV4 | ETH_RSS_IPV6 | > + ETH_RSS_IPV4_TCP | > ETH_RSS_IPV4_UDP | > + ETH_RSS_IPV6_TCP | > ETH_RSS_IPV6_UDP, > + } > + }, > +}; > + > +static const struct rte_eth_rxconf rx_conf_default =3D { > + .rx_thresh =3D { > + .pthresh =3D RX_PTHRESH, > + .hthresh =3D RX_HTHRESH, > + .wthresh =3D RX_WTHRESH, > + }, > + .rx_free_thresh =3D RX_FREE_THRESH, > + .rx_drop_en =3D 0, > +}; > + > +static const struct rte_eth_txconf tx_conf_default =3D { > + .tx_thresh =3D { > + .pthresh =3D TX_PTHRESH, > + .hthresh =3D TX_HTHRESH, > + .wthresh =3D TX_WTHRESH, > + }, > + .tx_free_thresh =3D TX_FREE_THRESH, > + .tx_rs_thresh =3D TX_RSBIT_THRESH, > + .txq_flags =3D TX_Q_FLAGS > + > +}; > + > +struct output_buffer { > + unsigned count; > + struct rte_mbuf *mbufs[BURST_SIZE]; > +}; > + > +/* > + * Initialises a given port using global settings and with the rx buffer= s > + * coming from the mbuf_pool passed as parameter > + */ > +static inline int > +port_init(uint8_t port, struct rte_mempool *mbuf_pool) > +{ > + struct rte_eth_conf port_conf =3D port_conf_default; > + const uint16_t rxRings =3D 1, txRings =3D rte_lcore_count() - 1; > + int retval; > + uint16_t q; > + > + if (port >=3D rte_eth_dev_count()) > + return -1; > + > + retval =3D rte_eth_dev_configure(port, rxRings, txRings, &port_conf); > + if (retval !=3D 0) > + return retval; > + > + for (q =3D 0; q < rxRings; q++) { > + retval =3D rte_eth_rx_queue_setup(port, q, RX_RING_SIZE, > + > rte_eth_dev_socket_id(port), > + &rx_conf_default, > mbuf_pool); > + if (retval < 0) > + return retval; > + } > + > + for (q =3D 0; q < txRings; q++) { > + retval =3D rte_eth_tx_queue_setup(port, q, TX_RING_SIZE, > + > rte_eth_dev_socket_id(port), > + &tx_conf_default); > + if (retval < 0) > + return retval; > + } > + > + retval =3D rte_eth_dev_start(port); > + if (retval < 0) > + return retval; > + > + struct rte_eth_link link; > + rte_eth_link_get_nowait(port, &link); > + if (!link.link_status) { > + sleep(1); > + rte_eth_link_get_nowait(port, &link); > + } > + > + if (!link.link_status) { > + printf("Link down on port %"PRIu8"\n", port); > + return 0; > + } > + > + struct ether_addr addr; > + rte_eth_macaddr_get(port, &addr); > + printf("Port %u MAC: %02"PRIx8" %02"PRIx8" %02"PRIx8 > + " %02"PRIx8" %02"PRIx8" %02"PRIx8"\n", > + (unsigned)port, > + addr.addr_bytes[0], addr.addr_bytes[1], > + addr.addr_bytes[2], addr.addr_bytes[3], > + addr.addr_bytes[4], addr.addr_bytes[5]); > + > + rte_eth_promiscuous_enable(port); > + > + return 0; > +} > + > +struct lcore_params { > + unsigned worker_id; > + struct rte_distributor *d; > + struct rte_ring *r; > +}; > + > +static __attribute__((noreturn)) void > +lcore_rx(struct lcore_params *p) > +{ > + struct rte_distributor *d =3D p->d; > + struct rte_ring *r =3D p->r; > + const uint8_t nb_ports =3D rte_eth_dev_count(); > + const int socket_id =3D rte_socket_id(); > + uint8_t port; > + > + for (port =3D 0; port < nb_ports; port++) { > + /* skip ports that are not enabled */ > + if ((enabled_port_mask & (1 << port)) =3D=3D 0) > + continue; > + > + if (rte_eth_dev_socket_id(port) > 0 && > + rte_eth_dev_socket_id(port) !=3D socket_id) > + printf("WARNING, port %u is on remote NUMA node > to " > + "RX thread.\n\tPerformance will not > " > + "be optimal.\n", port); > + } > + > + printf("\nCore %u doing packet RX.\n", rte_lcore_id()); > + port =3D 0; > + for (;;) { > + /* skip ports that are not enabled */ > + if ((enabled_port_mask & (1 << port)) =3D=3D 0) { > + if (++port =3D=3D nb_ports) > + port =3D 0; > + continue; > + } > + struct rte_mbuf *bufs[BURST_SIZE*2]; > + const uint16_t nb_rx =3D rte_eth_rx_burst(port, 0, bufs, > + BURST_SIZE); > + app_stats.rx.rx_pkts +=3D nb_rx; > + > + rte_distributor_process(d, bufs, nb_rx); > + const uint16_t nb_ret =3D rte_distributor_returned_pkts(d, > + bufs, BURST_SIZE*2); > + app_stats.rx.returned_pkts +=3D nb_ret; > + if (unlikely(nb_ret =3D=3D 0)) > + continue; > + > + uint16_t sent =3D rte_ring_enqueue_burst(r, (void *)bufs, > nb_ret); > + app_stats.rx.enqueued_pkts +=3D sent; > + if (unlikely(sent < nb_ret)) { > + printf("Packet loss due to full ring\n"); I would remove this printf. If core is too busy to process packets, this ma= y make performance worse. Statistics should reflect this packet loss. > + while (sent < nb_ret) > + rte_pktmbuf_free(bufs[sent++]); > + } > + if (++port =3D=3D nb_ports) > + port =3D 0; > + } > +} > + > +static inline void > +flush_one_port(struct output_buffer *outbuf, uint8_t outp) > +{ > + unsigned nb_tx =3D rte_eth_tx_burst(outp, 0, outbuf->mbufs, > + outbuf->count); > + app_stats.tx.tx_pkts +=3D nb_tx; > + > + if (unlikely(nb_tx < outbuf->count)) { > + printf("Packet loss with tx_burst\n"); Same here. This is fine for debugging, but statistics are updated as well, = so we do not need this printf. > + do { > + rte_pktmbuf_free(outbuf->mbufs[nb_tx]); > + } while (++nb_tx < outbuf->count); > + } > + outbuf->count =3D 0; > +} > + > +static inline void > +flush_all_ports(struct output_buffer *tx_buffers, uint8_t nb_ports) > +{ > + uint8_t outp; > + for (outp =3D 0; outp < nb_ports; outp++) { > + /* skip ports that are not enabled */ > + if ((enabled_port_mask & (1 << outp)) =3D=3D 0) > + continue; > + > + if (tx_buffers[outp].count =3D=3D 0) > + continue; > + > + flush_one_port(&tx_buffers[outp], outp); > + } > +} > + > +static __attribute__((noreturn)) void > +lcore_tx(struct rte_ring *in_r) > +{ > + static struct output_buffer tx_buffers[RTE_MAX_ETHPORTS]; > + const uint8_t nb_ports =3D rte_eth_dev_count(); > + const int socket_id =3D rte_socket_id(); > + uint8_t port; > + > + for (port =3D 0; port < nb_ports; port++) { > + /* skip ports that are not enabled */ > + if ((enabled_port_mask & (1 << port)) =3D=3D 0) > + continue; > + > + if (rte_eth_dev_socket_id(port) > 0 && > + rte_eth_dev_socket_id(port) !=3D socket_id) > + printf("WARNING, port %u is on remote NUMA node > to " > + "TX thread.\n\tPerformance will not " > + "be optimal.\n", port); > + } > + > + printf("\nCore %u doing packet TX.\n", rte_lcore_id()); > + for (;;) { > + for (port =3D 0; port < nb_ports; port++) { > + /* skip ports that are not enabled */ > + if ((enabled_port_mask & (1 << port)) =3D=3D 0) > + continue; > + > + struct rte_mbuf *bufs[BURST_SIZE]; > + const uint16_t nb_rx =3D rte_ring_dequeue_burst(in_r, > + (void *)bufs, BURST_SIZE); > + app_stats.tx.dequeue_pkts +=3D nb_rx; > + > + /* if we get no traffic, flush anything we have */ > + if (unlikely(nb_rx =3D=3D 0)) { > + flush_all_ports(tx_buffers, nb_ports); > + continue; > + } > + > + /* for traffic we receive, queue it up for transmit */ > + uint16_t i; > + _mm_prefetch(bufs[0], 0); > + _mm_prefetch(bufs[1], 0); > + _mm_prefetch(bufs[2], 0); > + for (i =3D 0; i < nb_rx; i++) { > + struct output_buffer *outbuf; > + uint8_t outp; > + _mm_prefetch(bufs[i + 3], 0); > + /* workers should update in_port to hold the > + * output port value */ > + outp =3D bufs[i]->port; > + /* skip ports that are not enabled */ > + if ((enabled_port_mask & (1 << outp)) =3D=3D 0) > + continue; > + > + outbuf =3D &tx_buffers[outp]; > + outbuf->mbufs[outbuf->count++] =3D bufs[i]; > + if (outbuf->count =3D=3D BURST_SIZE) > + flush_one_port(outbuf, outp); > + } > + } > + } > +} > + > + > +static __attribute__((noreturn)) void > +lcore_worker(struct lcore_params *p) > +{ > + struct rte_distributor *d =3D p->d; > + const unsigned id =3D p->worker_id; > + /* for single port, xor_val will be zero so we won't modify the output > + * port, otherwise we send traffic from 0 to 1, 2 to 3, and vice versa > + */ > + const unsigned xor_val =3D (rte_eth_dev_count() > 1); > + struct rte_mbuf *buf =3D NULL; > + > + printf("\nCore %u acting as worker core.\n", rte_lcore_id()); > + for (;;) { > + buf =3D rte_distributor_get_pkt(d, id, buf); > + buf->port ^=3D xor_val; > + } > +} > + > +static void > +int_handler(int sig_num) > +{ > + struct rte_eth_stats eth_stats; > + unsigned i; > + > + printf("Exiting on signal %d\n", sig_num); > + > + printf("\nRX thread stats:\n"); > + printf(" - Received: %"PRIu64"\n", app_stats.rx.rx_pkts); > + printf(" - Processed: %"PRIu64"\n", app_stats.rx.returned_pkts); > + printf(" - Enqueued: %"PRIu64"\n", app_stats.rx.enqueued_pkts); > + > + printf("\nTX thread stats:\n"); > + printf(" - Dequeued: %"PRIu64"\n", app_stats.tx.dequeue_pkts); > + printf(" - Transmitted: %"PRIu64"\n", app_stats.tx.tx_pkts); > + > + for (i =3D 0; i < rte_eth_dev_count(); i++) { > + rte_eth_stats_get(i, ð_stats); > + printf("\nPort %u stats:\n", i); > + printf(" - Pkts in: %"PRIu64"\n", eth_stats.ipackets); > + printf(" - Pkts out: %"PRIu64"\n", eth_stats.opackets); > + printf(" - In Errs: %"PRIu64"\n", eth_stats.ierrors); > + printf(" - Out Errs: %"PRIu64"\n", eth_stats.oerrors); > + printf(" - Mbuf Errs: %"PRIu64"\n", eth_stats.rx_nombuf); > + } > + exit(0); > +} > + > +/* display usage */ > +static void > +print_usage(const char *prgname) > +{ > + printf("%s [EAL options] -- -p PORTMASK\n" > + " -p PORTMASK: hexadecimal bitmask of ports to > configure\n", > + prgname); > +} > + > +static int > +parse_portmask(const char *portmask) > +{ > + char *end =3D NULL; > + unsigned long pm; > + > + /* parse hexadecimal string */ > + pm =3D strtoul(portmask, &end, 16); > + if ((portmask[0] =3D=3D '\0') || (end =3D=3D NULL) || (*end !=3D '\0')) > + return -1; > + > + if (pm =3D=3D 0) > + return -1; > + > + return pm; > +} > + > +/* Parse the argument given in the command line of the application */ > +static int > +parse_args(int argc, char **argv) > +{ > + int opt; > + char **argvopt; > + int option_index; > + char *prgname =3D argv[0]; > + static struct option lgopts[] =3D { > + {NULL, 0, 0, 0} > + }; > + > + argvopt =3D argv; > + > + while ((opt =3D getopt_long(argc, argvopt, "p:", > + lgopts, &option_index)) !=3D EOF) { > + > + switch (opt) { > + /* portmask */ > + case 'p': > + enabled_port_mask =3D > parse_portmask(optarg); > + if (enabled_port_mask =3D=3D 0) { > + printf("invalid portmask\n"); > + print_usage(prgname); > + return -1; > + } > + break; > + > + default: > + print_usage(prgname); > + return -1; > + } > + } > + > +if (optind <=3D 1) { > + print_usage(prgname); > + return -1; > +} > + > + argv[optind-1] =3D prgname; > + > + optind =3D 0; /* reset getopt lib */ > + return 0; > +} > + > +/* Main function, does initialization and calls the per-lcore functions = */ > +int > +MAIN(int argc, char *argv[]) > +{ > + struct rte_mempool *mbuf_pool; > + struct rte_distributor *d; > + struct rte_ring *output_ring; > + unsigned lcore_id, worker_id =3D 0; > + unsigned nb_ports; > + uint8_t portid; > + uint8_t nb_ports_available; > + > + /* catch ctrl-c so we can print on exit */ > + signal(SIGINT, int_handler); > + > + /* init EAL */ > + int ret =3D rte_eal_init(argc, argv); > + if (ret < 0) > + rte_exit(EXIT_FAILURE, "Error with EAL initialization\n"); > + argc -=3D ret; > + argv +=3D ret; > + > + /* parse application arguments (after the EAL ones) */ > + ret =3D parse_args(argc, argv); > + if (ret < 0) > + rte_exit(EXIT_FAILURE, "Invalid distributor parameters\n"); > + > + if (rte_lcore_count() < 3) > + rte_exit(EXIT_FAILURE, "Error, This application needs at " > + "least 3 logical cores to run:\n" > + "1 lcore for packet RX and distribution\n" > + "1 lcore for packet TX\n" > + "and at least 1 lcore for worker threads\n"); > + > + if (rte_eal_pci_probe() !=3D 0) > + rte_exit(EXIT_FAILURE, "Error with PCI probing\n"); > + > + nb_ports =3D rte_eth_dev_count(); > + if (nb_ports =3D=3D 0) > + rte_exit(EXIT_FAILURE, "Error: no ethernet ports > detected\n"); > + if (nb_ports !=3D 1 && (nb_ports & 1)) > + rte_exit(EXIT_FAILURE, "Error: number of ports must be > even, except " > + "when using a single port\n"); > + > + mbuf_pool =3D rte_mempool_create("MBUF_POOL", NUM_MBUFS * > nb_ports, > + MBUF_SIZE, MBUF_CACHE_SIZE, > + sizeof(struct rte_pktmbuf_pool_private), > + rte_pktmbuf_pool_init, NULL, > + rte_pktmbuf_init, NULL, > + rte_socket_id(), 0); > + if (mbuf_pool =3D=3D NULL) > + rte_exit(EXIT_FAILURE, "Cannot create mbuf pool\n"); > + > + nb_ports_available =3D nb_ports; > + > + /* initialize all ports */ > + for (portid =3D 0; portid < nb_ports; portid++) { > + /* skip ports that are not enabled */ > + if ((enabled_port_mask & (1 << portid)) =3D=3D 0) { > + printf("\nSkipping disabled port %d\n", portid); > + nb_ports_available--; > + continue; > + } > + /* init port */ > + printf("Initializing port %u... done\n", (unsigned) portid); > + > + if (port_init(portid, mbuf_pool) !=3D 0) > + rte_exit(EXIT_FAILURE, "Cannot initialize port > %"PRIu8"\n", > + portid); > + } > + > + if (!nb_ports_available) { > + rte_exit(EXIT_FAILURE, > + "All available ports are disabled. Please set > portmask.\n"); > + } > + > + d =3D rte_distributor_create("PKT_DIST", rte_socket_id(), > + rte_lcore_count() - 2); > + if (d =3D=3D NULL) > + rte_exit(EXIT_FAILURE, "Cannot create distributor\n"); > + > + /* scheduler ring is read only by the transmitter core, but written to > + * by multiple threads > + */ > + output_ring =3D rte_ring_create("Output_ring", RTE_RING_SZ, > + rte_socket_id(), RING_F_SC_DEQ); > + if (output_ring =3D=3D NULL) > + rte_exit(EXIT_FAILURE, "Cannot create output ring\n"); > + > + RTE_LCORE_FOREACH_SLAVE(lcore_id) { > + if (worker_id =3D=3D rte_lcore_count() - 2) > + rte_eal_remote_launch((lcore_function_t > *)lcore_tx, > + output_ring, lcore_id); > + else { > + struct lcore_params *p =3D > + rte_malloc(NULL, sizeof(*p), 0); > + if (!p) > + rte_panic("malloc failure\n"); > + *p =3D (struct lcore_params){worker_id, d, > output_ring}; > + rte_eal_remote_launch((lcore_function_t > *)lcore_worker, > + p, lcore_id); > + } > + worker_id++; > + } > + /* call lcore_main on master core only */ > + struct lcore_params p =3D { 0, d, output_ring }; > + lcore_rx(&p); > + return 0; > +} > diff --git a/examples/distributor_app/main.h > b/examples/distributor_app/main.h > new file mode 100644 > index 0000000..2682d15 > --- /dev/null > +++ b/examples/distributor_app/main.h > @@ -0,0 +1,46 @@ > +/*- > + * BSD LICENSE > + * > + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved. > + * All rights reserved. > + * > + * Redistribution and use in source and binary forms, with or without > + * modification, are permitted provided that the following conditions > + * are met: > + * > + * * Redistributions of source code must retain the above copyright > + * notice, this list of conditions and the following disclaimer. > + * * Redistributions in binary form must reproduce the above copyrig= ht > + * notice, this list of conditions and the following disclaimer in > + * the documentation and/or other materials provided with the > + * distribution. > + * * Neither the name of Intel Corporation nor the names of its > + * contributors may be used to endorse or promote products derived > + * from this software without specific prior written permission. > + * > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND > CONTRIBUTORS > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT > NOT > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND > FITNESS FOR > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE > COPYRIGHT > + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, > INCIDENTAL, > + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT > NOT > + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS > OF USE, > + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED > AND ON ANY > + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR > TORT > + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF > THE USE > + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH > DAMAGE. > + */ > + > +#ifndef _MAIN_H_ > +#define _MAIN_H_ > + > + > +#ifdef RTE_EXEC_ENV_BAREMETAL > +#define MAIN _main > +#else > +#define MAIN main > +#endif > + > +int MAIN(int argc, char *argv[]); > + > +#endif /* ifndef _MAIN_H_ */ > -- > 1.9.3