diff options
Diffstat (limited to '')
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-cat/Makefile | 70 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-cat/cat.c | 996 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-cat/cat.h | 72 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-cat/l2fwd-cat.c | 224 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-crypto/Makefile | 50 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-crypto/main.c | 2105 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-jobstats/Makefile | 51 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-jobstats/main.c | 1022 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-keepalive/Makefile | 51 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/Makefile | 49 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/main.c | 150 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-keepalive/main.c | 819 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-keepalive/shm.c | 141 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd-keepalive/shm.h | 98 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd/Makefile | 50 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/l2fwd/main.c | 758 |
16 files changed, 6706 insertions, 0 deletions
diff --git a/src/seastar/dpdk/examples/l2fwd-cat/Makefile b/src/seastar/dpdk/examples/l2fwd-cat/Makefile new file mode 100644 index 00000000..ae921ade --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-cat/Makefile @@ -0,0 +1,70 @@ +# BSD LICENSE +# +# Copyright(c) 2016 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 copyright +# 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 + +ifeq ($(PQOS_INSTALL_PATH),) +$(error "Please define PQOS_INSTALL_PATH environment variable") +endif + +# Default target, can be overridden by command line or environment +RTE_TARGET ?= x86_64-native-linuxapp-gcc + +# Location of PQoS library and includes, +PQOS_LIBRARY_PATH = $(PQOS_INSTALL_PATH)/libpqos.a + +include $(RTE_SDK)/mk/rte.vars.mk + +# binary name +APP = l2fwd-cat + +# all source are stored in SRCS-y +SRCS-y := l2fwd-cat.c cat.c + +CFLAGS += $(WERROR_FLAGS) + +# workaround for a gcc bug with noreturn attribute +# http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12603 +ifeq ($(CONFIG_RTE_TOOLCHAIN_GCC),y) +CFLAGS_main.o += -Wno-return-type +endif + +EXTRA_CFLAGS += -O3 -g -Wfatal-errors + +CFLAGS += -I$(PQOS_INSTALL_PATH)/../include +CFLAGS_cat.o := -D_GNU_SOURCE + +LDLIBS += -L$(PQOS_INSTALL_PATH) +LDLIBS += $(PQOS_LIBRARY_PATH) + +include $(RTE_SDK)/mk/rte.extapp.mk diff --git a/src/seastar/dpdk/examples/l2fwd-cat/cat.c b/src/seastar/dpdk/examples/l2fwd-cat/cat.c new file mode 100644 index 00000000..6133bf5b --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-cat/cat.c @@ -0,0 +1,996 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2016 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 copyright + * 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 <getopt.h> +#include <inttypes.h> +#include <limits.h> +#include <sched.h> +#include <signal.h> +#include <stdio.h> + +#include <rte_common.h> +#include <rte_memcpy.h> + +#include <pqos.h> + +#include "cat.h" + +#define BITS_PER_HEX 4 +#define PQOS_MAX_SOCKETS 8 +#define PQOS_MAX_SOCKET_CORES 64 +#define PQOS_MAX_CORES (PQOS_MAX_SOCKET_CORES * PQOS_MAX_SOCKETS) + +static const struct pqos_cap *m_cap; +static const struct pqos_cpuinfo *m_cpu; +static const struct pqos_capability *m_cap_l3ca; +static unsigned m_sockets[PQOS_MAX_SOCKETS]; +static unsigned m_sock_count; +static struct cat_config m_config[PQOS_MAX_CORES]; +static unsigned m_config_count; + +static unsigned +bits_count(uint64_t bitmask) +{ + unsigned count = 0; + + for (; bitmask != 0; count++) + bitmask &= bitmask - 1; + + return count; +} + +/* + * Parse elem, the elem could be single number/range or '(' ')' group + * 1) A single number elem, it's just a simple digit. e.g. 9 + * 2) A single range elem, two digits with a '-' between. e.g. 2-6 + * 3) A group elem, combines multiple 1) or 2) with '( )'. e.g (0,2-4,6) + * Within group elem, '-' used for a range separator; + * ',' used for a single number. + */ +static int +parse_set(const char *input, rte_cpuset_t *cpusetp) +{ + unsigned idx; + const char *str = input; + char *end = NULL; + unsigned min, max; + const unsigned num = PQOS_MAX_CORES; + + CPU_ZERO(cpusetp); + + while (isblank(*str)) + str++; + + /* only digit or left bracket is qualify for start point */ + if ((!isdigit(*str) && *str != '(') || *str == '\0') + return -1; + + /* process single number or single range of number */ + if (*str != '(') { + errno = 0; + idx = strtoul(str, &end, 10); + + if (errno || end == NULL || idx >= num) + return -1; + + while (isblank(*end)) + end++; + + min = idx; + max = idx; + if (*end == '-') { + /* process single <number>-<number> */ + end++; + while (isblank(*end)) + end++; + if (!isdigit(*end)) + return -1; + + errno = 0; + idx = strtoul(end, &end, 10); + if (errno || end == NULL || idx >= num) + return -1; + max = idx; + while (isblank(*end)) + end++; + if (*end != ',' && *end != '\0') + return -1; + } + + if (*end != ',' && *end != '\0' && *end != '@') + return -1; + + for (idx = RTE_MIN(min, max); idx <= RTE_MAX(min, max); + idx++) + CPU_SET(idx, cpusetp); + + return end - input; + } + + /* process set within bracket */ + str++; + while (isblank(*str)) + str++; + if (*str == '\0') + return -1; + + min = PQOS_MAX_CORES; + do { + + /* go ahead to the first digit */ + while (isblank(*str)) + str++; + if (!isdigit(*str)) + return -1; + + /* get the digit value */ + errno = 0; + idx = strtoul(str, &end, 10); + if (errno || end == NULL || idx >= num) + return -1; + + /* go ahead to separator '-',',' and ')' */ + while (isblank(*end)) + end++; + if (*end == '-') { + if (min == PQOS_MAX_CORES) + min = idx; + else /* avoid continuous '-' */ + return -1; + } else if ((*end == ',') || (*end == ')')) { + max = idx; + if (min == PQOS_MAX_CORES) + min = idx; + for (idx = RTE_MIN(min, max); idx <= RTE_MAX(min, max); + idx++) + CPU_SET(idx, cpusetp); + + min = PQOS_MAX_CORES; + } else + return -1; + + str = end + 1; + } while (*end != '\0' && *end != ')'); + + return str - input; +} + +/* Test if bitmask is contiguous */ +static int +is_contiguous(uint64_t bitmask) +{ + /* check if bitmask is contiguous */ + unsigned i = 0; + unsigned j = 0; + const unsigned max_idx = (sizeof(bitmask) * CHAR_BIT); + + if (bitmask == 0) + return 0; + + for (i = 0; i < max_idx; i++) { + if (((1ULL << i) & bitmask) != 0) + j++; + else if (j > 0) + break; + } + + if (bits_count(bitmask) != j) { + printf("PQOS: mask 0x%llx is not contiguous.\n", + (unsigned long long)bitmask); + return 0; + } + + return 1; +} + +/* + * The format pattern: --l3ca='<cbm@cpus>[,<(ccbm,dcbm)@cpus>...]' + * cbm could be a single mask or for a CDP enabled system, a group of two masks + * ("code cbm" and "data cbm") + * '(' and ')' are necessary if it's a group. + * cpus could be a single digit/range or a group. + * '(' and ')' are necessary if it's a group. + * + * e.g. '0x00F00@(1,3), 0x0FF00@(4-6), 0xF0000@7' + * - CPUs 1 and 3 share its 4 ways with CPUs 4, 5 and 6; + * - CPUs 4,5 and 6 share half (4 out of 8 ways) of its L3 with 1 and 3; + * - CPUs 4,5 and 6 have exclusive access to 4 out of 8 ways; + * - CPU 7 has exclusive access to all of its 4 ways; + * + * e.g. '(0x00C00,0x00300)@(1,3)' for a CDP enabled system + * - cpus 1 and 3 have access to 2 ways for code and 2 ways for data, + * code and data ways are not overlapping.; + */ +static int +parse_l3ca(const char *l3ca) +{ + unsigned idx = 0; + const char *cbm_start = NULL; + char *cbm_end = NULL; + const char *end = NULL; + int offset; + rte_cpuset_t cpuset; + uint64_t mask = 0; + uint64_t cmask = 0; + + if (l3ca == NULL) + goto err; + + /* Get cbm */ + do { + CPU_ZERO(&cpuset); + mask = 0; + cmask = 0; + + while (isblank(*l3ca)) + l3ca++; + + if (*l3ca == '\0') + goto err; + + /* record mask_set start point */ + cbm_start = l3ca; + + /* go across a complete bracket */ + if (*cbm_start == '(') { + l3ca += strcspn(l3ca, ")"); + if (*l3ca++ == '\0') + goto err; + } + + /* scan the separator '@', ','(next) or '\0'(finish) */ + l3ca += strcspn(l3ca, "@,"); + + if (*l3ca == '@') { + /* explicit assign cpu_set */ + offset = parse_set(l3ca + 1, &cpuset); + if (offset < 0 || CPU_COUNT(&cpuset) == 0) + goto err; + + end = l3ca + 1 + offset; + } else + goto err; + + if (*end != ',' && *end != '\0') + goto err; + + /* parse mask_set from start point */ + if (*cbm_start == '(') { + cbm_start++; + + while (isblank(*cbm_start)) + cbm_start++; + + if (!isxdigit(*cbm_start)) + goto err; + + errno = 0; + cmask = strtoul(cbm_start, &cbm_end, 16); + if (errno != 0 || cbm_end == NULL || cmask == 0) + goto err; + + while (isblank(*cbm_end)) + cbm_end++; + + if (*cbm_end != ',') + goto err; + + cbm_end++; + + while (isblank(*cbm_end)) + cbm_end++; + + if (!isxdigit(*cbm_end)) + goto err; + + errno = 0; + mask = strtoul(cbm_end, &cbm_end, 16); + if (errno != 0 || cbm_end == NULL || mask == 0) + goto err; + } else { + while (isblank(*cbm_start)) + cbm_start++; + + if (!isxdigit(*cbm_start)) + goto err; + + errno = 0; + mask = strtoul(cbm_start, &cbm_end, 16); + if (errno != 0 || cbm_end == NULL || mask == 0) + goto err; + + } + + if (mask == 0 || is_contiguous(mask) == 0) + goto err; + + if (cmask != 0 && is_contiguous(cmask) == 0) + goto err; + + rte_memcpy(&m_config[idx].cpumask, + &cpuset, sizeof(rte_cpuset_t)); + + if (cmask != 0) { + m_config[idx].cdp = 1; + m_config[idx].code_mask = cmask; + m_config[idx].data_mask = mask; + } else + m_config[idx].mask = mask; + + m_config_count++; + + l3ca = end + 1; + idx++; + } while (*end != '\0' && idx < PQOS_MAX_CORES); + + if (m_config_count == 0) + goto err; + + return 0; + +err: + return -EINVAL; +} + +static int +check_cpus_overlapping(void) +{ + unsigned i = 0; + unsigned j = 0; + rte_cpuset_t mask; + + CPU_ZERO(&mask); + + for (i = 0; i < m_config_count; i++) { + for (j = i + 1; j < m_config_count; j++) { + CPU_AND(&mask, + &m_config[i].cpumask, + &m_config[j].cpumask); + + if (CPU_COUNT(&mask) != 0) { + printf("PQOS: Requested CPUs sets are " + "overlapping.\n"); + return -EINVAL; + } + } + } + + return 0; +} + +static int +check_cpus(void) +{ + unsigned i = 0; + unsigned cpu_id = 0; + unsigned cos_id = 0; + int ret = 0; + + for (i = 0; i < m_config_count; i++) { + for (cpu_id = 0; cpu_id < PQOS_MAX_CORES; cpu_id++) { + if (CPU_ISSET(cpu_id, &m_config[i].cpumask) != 0) { + + ret = pqos_cpu_check_core(m_cpu, cpu_id); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: %u is not a valid " + "logical core id.\n", cpu_id); + ret = -ENODEV; + goto exit; + } + + ret = pqos_l3ca_assoc_get(cpu_id, &cos_id); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to read COS " + "associated to cpu %u.\n", + cpu_id); + ret = -EFAULT; + goto exit; + } + + /* + * Check if COS assigned to lcore is different + * then default one (#0) + */ + if (cos_id != 0) { + printf("PQOS: cpu %u has already " + "associated COS#%u. " + "Please reset L3CA.\n", + cpu_id, cos_id); + ret = -EBUSY; + goto exit; + } + } + } + } + +exit: + return ret; +} + +static int +check_cdp(void) +{ + unsigned i = 0; + + for (i = 0; i < m_config_count; i++) { + if (m_config[i].cdp == 1 && m_cap_l3ca->u.l3ca->cdp_on == 0) { + if (m_cap_l3ca->u.l3ca->cdp == 0) { + printf("PQOS: CDP requested but not " + "supported.\n"); + } else { + printf("PQOS: CDP requested but not enabled. " + "Please enable CDP.\n"); + } + return -ENOTSUP; + } + } + + return 0; +} + +static int +check_cbm_len_and_contention(void) +{ + unsigned i = 0; + uint64_t mask = 0; + const uint64_t not_cbm = (UINT64_MAX << (m_cap_l3ca->u.l3ca->num_ways)); + const uint64_t cbm_contention_mask = m_cap_l3ca->u.l3ca->way_contention; + int ret = 0; + + for (i = 0; i < m_config_count; i++) { + if (m_config[i].cdp == 1) + mask = m_config[i].code_mask | m_config[i].data_mask; + else + mask = m_config[i].mask; + + if ((mask & not_cbm) != 0) { + printf("PQOS: One or more of requested CBM masks not " + "supported by system (too long).\n"); + ret = -ENOTSUP; + break; + } + + /* Just a warning */ + if ((mask & cbm_contention_mask) != 0) { + printf("PQOS: One or more of requested CBM masks " + "overlap CBM contention mask.\n"); + break; + } + + } + + return ret; +} + +static int +check_and_select_classes(unsigned cos_id_map[][PQOS_MAX_SOCKETS]) +{ + unsigned i = 0; + unsigned j = 0; + unsigned phy_pkg_id = 0; + unsigned cos_id = 0; + unsigned cpu_id = 0; + unsigned phy_pkg_lcores[PQOS_MAX_SOCKETS][m_config_count]; + const unsigned cos_num = m_cap_l3ca->u.l3ca->num_classes; + unsigned used_cos_table[PQOS_MAX_SOCKETS][cos_num]; + int ret = 0; + + memset(phy_pkg_lcores, 0, sizeof(phy_pkg_lcores)); + memset(used_cos_table, 0, sizeof(used_cos_table)); + + /* detect currently used COS */ + for (j = 0; j < m_cpu->num_cores; j++) { + cpu_id = m_cpu->cores[j].lcore; + + ret = pqos_l3ca_assoc_get(cpu_id, &cos_id); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to read COS associated to " + "cpu %u on phy_pkg %u.\n", cpu_id, phy_pkg_id); + ret = -EFAULT; + goto exit; + } + + ret = pqos_cpu_get_socketid(m_cpu, cpu_id, &phy_pkg_id); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to get socket for cpu %u\n", + cpu_id); + ret = -EFAULT; + goto exit; + } + + /* Mark COS as used */ + if (used_cos_table[phy_pkg_id][cos_id] == 0) + used_cos_table[phy_pkg_id][cos_id]++; + } + + /* look for avail. COS to fulfill requested config */ + for (i = 0; i < m_config_count; i++) { + for (j = 0; j < m_cpu->num_cores; j++) { + cpu_id = m_cpu->cores[j].lcore; + if (CPU_ISSET(cpu_id, &m_config[i].cpumask) == 0) + continue; + + ret = pqos_cpu_get_socketid(m_cpu, cpu_id, &phy_pkg_id); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to get socket for " + "cpu %u\n", cpu_id); + ret = -EFAULT; + goto exit; + } + + /* + * Check if we already have COS selected + * to be used for that group on that socket + */ + if (phy_pkg_lcores[phy_pkg_id][i] != 0) + continue; + + phy_pkg_lcores[phy_pkg_id][i]++; + + /* Search for avail. COS to be used on that socket */ + for (cos_id = 0; cos_id < cos_num; cos_id++) { + if (used_cos_table[phy_pkg_id][cos_id] == 0) { + used_cos_table[phy_pkg_id][cos_id]++; + cos_id_map[i][phy_pkg_id] = cos_id; + break; + } + } + + /* If there is no COS available ...*/ + if (cos_id == cos_num) { + ret = -E2BIG; + goto exit; + } + } + } + +exit: + if (ret != 0) + printf("PQOS: Not enough available COS to configure " + "requested configuration.\n"); + + return ret; +} + +static int +configure_cat(unsigned cos_id_map[][PQOS_MAX_SOCKETS]) +{ + unsigned phy_pkg_id = 0; + unsigned cpu_id = 0; + unsigned cos_id = 0; + unsigned i = 0; + unsigned j = 0; + struct pqos_l3ca l3ca = {0}; + int ret = 0; + + for (i = 0; i < m_config_count; i++) { + memset(&l3ca, 0, sizeof(l3ca)); + + l3ca.cdp = m_config[i].cdp; + if (m_config[i].cdp == 1) { + l3ca.code_mask = m_config[i].code_mask; + l3ca.data_mask = m_config[i].data_mask; + } else + l3ca.ways_mask = m_config[i].mask; + + for (j = 0; j < m_sock_count; j++) { + phy_pkg_id = m_sockets[j]; + if (cos_id_map[i][phy_pkg_id] == 0) + continue; + + l3ca.class_id = cos_id_map[i][phy_pkg_id]; + + ret = pqos_l3ca_set(phy_pkg_id, 1, &l3ca); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to set COS %u on " + "phy_pkg %u.\n", l3ca.class_id, + phy_pkg_id); + ret = -EFAULT; + goto exit; + } + } + } + + for (i = 0; i < m_config_count; i++) { + for (j = 0; j < m_cpu->num_cores; j++) { + cpu_id = m_cpu->cores[j].lcore; + if (CPU_ISSET(cpu_id, &m_config[i].cpumask) == 0) + continue; + + ret = pqos_cpu_get_socketid(m_cpu, cpu_id, &phy_pkg_id); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to get socket for " + "cpu %u\n", cpu_id); + ret = -EFAULT; + goto exit; + } + + cos_id = cos_id_map[i][phy_pkg_id]; + + ret = pqos_l3ca_assoc_set(cpu_id, cos_id); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to associate COS %u to " + "cpu %u\n", cos_id, cpu_id); + ret = -EFAULT; + goto exit; + } + } + } + +exit: + return ret; +} + + +/* Parse the argument given in the command line of the application */ +static int +parse_args(int argc, char **argv) +{ + int opt = 0; + int retval = 0; + int oldopterr = 0; + char **argvopt = argv; + char *prgname = argv[0]; + + static struct option lgopts[] = { + { "l3ca", required_argument, 0, 0 }, + { NULL, 0, 0, 0 } + }; + + /* Disable printing messages within getopt() */ + oldopterr = opterr; + opterr = 0; + + opt = getopt_long(argc, argvopt, "", lgopts, NULL); + if (opt == 0) { + retval = parse_l3ca(optarg); + if (retval != 0) { + printf("PQOS: Invalid L3CA parameters!\n"); + goto exit; + } + + argv[optind - 1] = prgname; + retval = optind - 1; + } else + retval = 0; + +exit: + /* reset getopt lib */ + optind = 1; + + /* Restore opterr value */ + opterr = oldopterr; + + return retval; +} + +static void +print_cmd_line_config(void) +{ + char cpustr[PQOS_MAX_CORES * 3] = {0}; + unsigned i = 0; + unsigned j = 0; + + for (i = 0; i < m_config_count; i++) { + unsigned len = 0; + memset(cpustr, 0, sizeof(cpustr)); + + /* Generate CPU list */ + for (j = 0; j < PQOS_MAX_CORES; j++) { + if (CPU_ISSET(j, &m_config[i].cpumask) != 1) + continue; + + len += snprintf(cpustr + len, sizeof(cpustr) - len - 1, + "%u,", j); + + if (len >= sizeof(cpustr) - 1) + break; + } + + if (m_config[i].cdp == 1) { + printf("PQOS: CPUs: %s cMASK: 0x%llx, dMASK: " + "0x%llx\n", cpustr, + (unsigned long long)m_config[i].code_mask, + (unsigned long long)m_config[i].data_mask); + } else { + printf("PQOS: CPUs: %s MASK: 0x%llx\n", cpustr, + (unsigned long long)m_config[i].mask); + } + } +} + +/** + * @brief Prints CAT configuration + */ +static void +print_cat_config(void) +{ + int ret = PQOS_RETVAL_OK; + unsigned i = 0; + + for (i = 0; i < m_sock_count; i++) { + struct pqos_l3ca tab[PQOS_MAX_L3CA_COS] = {{0} }; + unsigned num = 0; + unsigned n = 0; + + ret = pqos_l3ca_get(m_sockets[i], PQOS_MAX_L3CA_COS, &num, tab); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Error retrieving COS!\n"); + return; + } + + printf("PQOS: COS definitions for Socket %u:\n", m_sockets[i]); + for (n = 0; n < num; n++) { + if (tab[n].cdp == 1) { + printf("PQOS: COS: %u, cMASK: 0x%llx, " + "dMASK: 0x%llx\n", tab[n].class_id, + (unsigned long long)tab[n].code_mask, + (unsigned long long)tab[n].data_mask); + } else { + printf("PQOS: COS: %u, MASK: 0x%llx\n", + tab[n].class_id, + (unsigned long long)tab[n].ways_mask); + } + } + } + + for (i = 0; i < m_sock_count; i++) { + unsigned lcores[PQOS_MAX_SOCKET_CORES] = {0}; + unsigned lcount = 0; + unsigned n = 0; + + ret = pqos_cpu_get_cores(m_cpu, m_sockets[i], + PQOS_MAX_SOCKET_CORES, &lcount, &lcores[0]); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Error retrieving core information!\n"); + return; + } + + printf("PQOS: CPU information for socket %u:\n", m_sockets[i]); + for (n = 0; n < lcount; n++) { + unsigned class_id = 0; + + ret = pqos_l3ca_assoc_get(lcores[n], &class_id); + if (ret == PQOS_RETVAL_OK) + printf("PQOS: CPU: %u, COS: %u\n", lcores[n], + class_id); + else + printf("PQOS: CPU: %u, ERROR\n", lcores[n]); + } + } + +} + +static int +cat_validate(void) +{ + int ret = 0; + + ret = check_cpus(); + if (ret != 0) + return ret; + + ret = check_cdp(); + if (ret != 0) + return ret; + + ret = check_cbm_len_and_contention(); + if (ret != 0) + return ret; + + ret = check_cpus_overlapping(); + if (ret != 0) + return ret; + + return 0; +} + +static int +cat_set(void) +{ + int ret = 0; + unsigned cos_id_map[m_config_count][PQOS_MAX_SOCKETS]; + + memset(cos_id_map, 0, sizeof(cos_id_map)); + + ret = check_and_select_classes(cos_id_map); + if (ret != 0) + return ret; + + ret = configure_cat(cos_id_map); + if (ret != 0) + return ret; + + return 0; +} + +static void +cat_fini(void) +{ + int ret = 0; + + printf("PQOS: Shutting down PQoS library...\n"); + + /* deallocate all the resources */ + ret = pqos_fini(); + if (ret != PQOS_RETVAL_OK && ret != PQOS_RETVAL_INIT) + printf("PQOS: Error shutting down PQoS library!\n"); + + m_cap = NULL; + m_cpu = NULL; + m_cap_l3ca = NULL; + memset(m_sockets, 0, sizeof(m_sockets)); + m_sock_count = 0; + memset(m_config, 0, sizeof(m_config)); + m_config_count = 0; +} + +void +cat_exit(void) +{ + unsigned i = 0; + unsigned j = 0; + unsigned cpu_id = 0; + int ret = 0; + + /* if lib is not initialized, do nothing */ + if (m_cap == NULL && m_cpu == NULL) + return; + + printf("PQOS: Reverting CAT configuration...\n"); + + for (i = 0; i < m_config_count; i++) { + for (j = 0; j < m_cpu->num_cores; j++) { + cpu_id = m_cpu->cores[j].lcore; + if (CPU_ISSET(cpu_id, &m_config[i].cpumask) == 0) + continue; + + ret = pqos_l3ca_assoc_set(cpu_id, 0); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Failed to associate COS 0 to " + "cpu %u\n", cpu_id); + } + } + } + + cat_fini(); +} + +static void +signal_handler(int signum) +{ + if (signum == SIGINT || signum == SIGTERM) { + printf("\nPQOS: Signal %d received, preparing to exit...\n", + signum); + + cat_exit(); + + /* exit with the expected status */ + signal(signum, SIG_DFL); + kill(getpid(), signum); + } +} + +int +cat_init(int argc, char **argv) +{ + int ret = 0; + int args_num = 0; + struct pqos_config cfg = {0}; + + if (m_cap != NULL || m_cpu != NULL) { + printf("PQOS: CAT module already initialized!\n"); + return -EEXIST; + } + + /* Parse cmd line args */ + ret = parse_args(argc, argv); + + if (ret <= 0) + goto err; + + args_num = ret; + + /* Print cmd line configuration */ + print_cmd_line_config(); + + /* PQoS Initialization - Check and initialize CAT capability */ + cfg.fd_log = STDOUT_FILENO; + cfg.verbose = 0; + cfg.cdp_cfg = PQOS_REQUIRE_CDP_ANY; + ret = pqos_init(&cfg); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Error initializing PQoS library!\n"); + ret = -EFAULT; + goto err; + } + + /* Get capability and CPU info pointer */ + ret = pqos_cap_get(&m_cap, &m_cpu); + if (ret != PQOS_RETVAL_OK || m_cap == NULL || m_cpu == NULL) { + printf("PQOS: Error retrieving PQoS capabilities!\n"); + ret = -EFAULT; + goto err; + } + + /* Get L3CA capabilities */ + ret = pqos_cap_get_type(m_cap, PQOS_CAP_TYPE_L3CA, &m_cap_l3ca); + if (ret != PQOS_RETVAL_OK || m_cap_l3ca == NULL) { + printf("PQOS: Error retrieving PQOS_CAP_TYPE_L3CA " + "capabilities!\n"); + ret = -EFAULT; + goto err; + } + + /* Get CPU socket information */ + ret = pqos_cpu_get_sockets(m_cpu, PQOS_MAX_SOCKETS, &m_sock_count, + m_sockets); + if (ret != PQOS_RETVAL_OK) { + printf("PQOS: Error retrieving CPU socket information!\n"); + ret = -EFAULT; + goto err; + } + + /* Validate cmd line configuration */ + ret = cat_validate(); + if (ret != 0) { + printf("PQOS: Requested CAT configuration is not valid!\n"); + goto err; + } + + /* configure system */ + ret = cat_set(); + if (ret != 0) { + printf("PQOS: Failed to configure CAT!\n"); + goto err; + } + + signal(SIGINT, signal_handler); + signal(SIGTERM, signal_handler); + + ret = atexit(cat_exit); + if (ret != 0) { + printf("PQOS: Cannot set exit function\n"); + goto err; + } + + /* Print CAT configuration */ + print_cat_config(); + + return args_num; + +err: + /* deallocate all the resources */ + cat_fini(); + return ret; +} diff --git a/src/seastar/dpdk/examples/l2fwd-cat/cat.h b/src/seastar/dpdk/examples/l2fwd-cat/cat.h new file mode 100644 index 00000000..aef2b768 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-cat/cat.h @@ -0,0 +1,72 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2016 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 copyright + * 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 _CAT_H +#define _CAT_H + +/** + * @file + * PQoS CAT + */ + +#include <stdint.h> +#include <string.h> + +#include <rte_lcore.h> + +#ifdef __cplusplus +extern "C" { +#endif + +/* L3 cache allocation class of service data structure */ +struct cat_config { + rte_cpuset_t cpumask; /* CPUs bitmask */ + int cdp; /* data & code masks used if true */ + union { + uint64_t mask; /* capacity bitmask (CBM) */ + struct { + uint64_t data_mask; /* data capacity bitmask (CBM) */ + uint64_t code_mask; /* code capacity bitmask (CBM) */ + }; + }; +}; + +int cat_init(int argc, char **argv); + +void cat_exit(void); + +#ifdef __cplusplus +} +#endif + +#endif /* _CAT_H */ diff --git a/src/seastar/dpdk/examples/l2fwd-cat/l2fwd-cat.c b/src/seastar/dpdk/examples/l2fwd-cat/l2fwd-cat.c new file mode 100644 index 00000000..8cce33b8 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-cat/l2fwd-cat.c @@ -0,0 +1,224 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2016 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 copyright + * 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 <stdint.h> +#include <inttypes.h> +#include <rte_eal.h> +#include <rte_ethdev.h> +#include <rte_cycles.h> +#include <rte_lcore.h> +#include <rte_mbuf.h> + +#include "cat.h" + +#define RX_RING_SIZE 128 +#define TX_RING_SIZE 512 + +#define NUM_MBUFS 8191 +#define MBUF_CACHE_SIZE 250 +#define BURST_SIZE 32 + +static const struct rte_eth_conf port_conf_default = { + .rxmode = { .max_rx_pkt_len = ETHER_MAX_LEN } +}; + +/* l2fwd-cat.c: CAT enabled, basic DPDK skeleton forwarding example. */ + +/* + * Initializes a given port using global settings and with the RX buffers + * coming from the mbuf_pool passed as a parameter. + */ +static inline int +port_init(uint8_t port, struct rte_mempool *mbuf_pool) +{ + struct rte_eth_conf port_conf = port_conf_default; + const uint16_t rx_rings = 1, tx_rings = 1; + int retval; + uint16_t q; + + if (port >= rte_eth_dev_count()) + return -1; + + /* Configure the Ethernet device. */ + retval = rte_eth_dev_configure(port, rx_rings, tx_rings, &port_conf); + if (retval != 0) + return retval; + + /* Allocate and set up 1 RX queue per Ethernet port. */ + for (q = 0; q < rx_rings; q++) { + retval = rte_eth_rx_queue_setup(port, q, RX_RING_SIZE, + rte_eth_dev_socket_id(port), NULL, mbuf_pool); + if (retval < 0) + return retval; + } + + /* Allocate and set up 1 TX queue per Ethernet port. */ + for (q = 0; q < tx_rings; q++) { + retval = rte_eth_tx_queue_setup(port, q, TX_RING_SIZE, + rte_eth_dev_socket_id(port), NULL); + if (retval < 0) + return retval; + } + + /* Start the Ethernet port. */ + retval = rte_eth_dev_start(port); + if (retval < 0) + return retval; + + /* Display the port MAC address. */ + 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]); + + /* Enable RX in promiscuous mode for the Ethernet device. */ + rte_eth_promiscuous_enable(port); + + return 0; +} + +/* + * The lcore main. This is the main thread that does the work, reading from + * an input port and writing to an output port. + */ +static __attribute__((noreturn)) void +lcore_main(void) +{ + const uint8_t nb_ports = rte_eth_dev_count(); + uint8_t port; + + /* + * Check that the port is on the same NUMA node as the polling thread + * for best performance. + */ + for (port = 0; port < nb_ports; port++) + if (rte_eth_dev_socket_id(port) > 0 && + rte_eth_dev_socket_id(port) != + (int)rte_socket_id()) + printf("WARNING, port %u is on remote NUMA node to " + "polling thread.\n\tPerformance will " + "not be optimal.\n", port); + + printf("\nCore %u forwarding packets. [Ctrl+C to quit]\n", + rte_lcore_id()); + + /* Run until the application is quit or killed. */ + for (;;) { + /* + * Receive packets on a port and forward them on the paired + * port. The mapping is 0 -> 1, 1 -> 0, 2 -> 3, 3 -> 2, etc. + */ + for (port = 0; port < nb_ports; port++) { + + /* Get burst of RX packets, from first port of pair. */ + struct rte_mbuf *bufs[BURST_SIZE]; + const uint16_t nb_rx = rte_eth_rx_burst(port, 0, + bufs, BURST_SIZE); + + if (unlikely(nb_rx == 0)) + continue; + + /* Send burst of TX packets, to second port of pair. */ + const uint16_t nb_tx = rte_eth_tx_burst(port ^ 1, 0, + bufs, nb_rx); + + /* Free any unsent packets. */ + if (unlikely(nb_tx < nb_rx)) { + uint16_t buf; + for (buf = nb_tx; buf < nb_rx; buf++) + rte_pktmbuf_free(bufs[buf]); + } + } + } +} + +/* + * The main function, which does initialization and calls the per-lcore + * functions. + */ +int +main(int argc, char *argv[]) +{ + struct rte_mempool *mbuf_pool; + unsigned nb_ports; + uint8_t portid; + + /* Initialize the Environment Abstraction Layer (EAL). */ + int ret = rte_eal_init(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Error with EAL initialization\n"); + + argc -= ret; + argv += ret; + + /* + * Initialize the PQoS library and configure CAT. + * Please see l2fwd-cat documentation for more info. + */ + ret = cat_init(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "PQOS: L3CA init failed!\n"); + + argc -= ret; + argv += ret; + + /* Check that there is an even number of ports to send/receive on. */ + nb_ports = rte_eth_dev_count(); + if (nb_ports < 2 || (nb_ports & 1)) + rte_exit(EXIT_FAILURE, "Error: number of ports must be even\n"); + + /* Creates a new mempool in memory to hold the mbufs. */ + mbuf_pool = rte_pktmbuf_pool_create("MBUF_POOL", NUM_MBUFS * nb_ports, + MBUF_CACHE_SIZE, 0, RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id()); + + if (mbuf_pool == NULL) + rte_exit(EXIT_FAILURE, "Cannot create mbuf pool\n"); + + /* Initialize all ports. */ + for (portid = 0; portid < nb_ports; portid++) + if (port_init(portid, mbuf_pool) != 0) + rte_exit(EXIT_FAILURE, "Cannot init port %"PRIu8 "\n", + portid); + + if (rte_lcore_count() > 1) + printf("\nWARNING: Too many lcores enabled. Only 1 used.\n"); + + /* Call lcore_main on the master core only. */ + lcore_main(); + + return 0; +} diff --git a/src/seastar/dpdk/examples/l2fwd-crypto/Makefile b/src/seastar/dpdk/examples/l2fwd-crypto/Makefile new file mode 100644 index 00000000..e8224cae --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-crypto/Makefile @@ -0,0 +1,50 @@ +# 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 copyright +# 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 overridden by command line or environment +RTE_TARGET ?= x86_64-native-linuxapp-gcc + +include $(RTE_SDK)/mk/rte.vars.mk + +# binary name +APP = l2fwd-crypto + +# all source are stored in SRCS-y +SRCS-y := main.c + +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) + +include $(RTE_SDK)/mk/rte.extapp.mk diff --git a/src/seastar/dpdk/examples/l2fwd-crypto/main.c b/src/seastar/dpdk/examples/l2fwd-crypto/main.c new file mode 100644 index 00000000..94921935 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-crypto/main.c @@ -0,0 +1,2105 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2015-2016 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 copyright + * 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 <time.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <stdint.h> +#include <inttypes.h> +#include <sys/types.h> +#include <sys/queue.h> +#include <netinet/in.h> +#include <setjmp.h> +#include <stdarg.h> +#include <ctype.h> +#include <errno.h> +#include <getopt.h> +#include <fcntl.h> +#include <unistd.h> + +#include <rte_atomic.h> +#include <rte_branch_prediction.h> +#include <rte_common.h> +#include <rte_cryptodev.h> +#include <rte_cycles.h> +#include <rte_debug.h> +#include <rte_eal.h> +#include <rte_ether.h> +#include <rte_ethdev.h> +#include <rte_interrupts.h> +#include <rte_ip.h> +#include <rte_launch.h> +#include <rte_lcore.h> +#include <rte_log.h> +#include <rte_malloc.h> +#include <rte_mbuf.h> +#include <rte_memcpy.h> +#include <rte_memory.h> +#include <rte_mempool.h> +#include <rte_memzone.h> +#include <rte_pci.h> +#include <rte_per_lcore.h> +#include <rte_prefetch.h> +#include <rte_random.h> +#include <rte_hexdump.h> + +enum cdev_type { + CDEV_TYPE_ANY, + CDEV_TYPE_HW, + CDEV_TYPE_SW +}; + +#define RTE_LOGTYPE_L2FWD RTE_LOGTYPE_USER1 + +#define NB_MBUF 8192 + +#define MAX_STR_LEN 32 +#define MAX_KEY_SIZE 128 +#define MAX_PKT_BURST 32 +#define BURST_TX_DRAIN_US 100 /* TX drain every ~100us */ + +/* + * Configurable number of RX/TX ring descriptors + */ +#define RTE_TEST_RX_DESC_DEFAULT 128 +#define RTE_TEST_TX_DESC_DEFAULT 512 + +static uint16_t nb_rxd = RTE_TEST_RX_DESC_DEFAULT; +static uint16_t nb_txd = RTE_TEST_TX_DESC_DEFAULT; + +/* ethernet addresses of ports */ +static struct ether_addr l2fwd_ports_eth_addr[RTE_MAX_ETHPORTS]; + +/* mask of enabled ports */ +static uint64_t l2fwd_enabled_port_mask; +static uint64_t l2fwd_enabled_crypto_mask; + +/* list of enabled ports */ +static uint32_t l2fwd_dst_ports[RTE_MAX_ETHPORTS]; + + +struct pkt_buffer { + unsigned len; + struct rte_mbuf *buffer[MAX_PKT_BURST]; +}; + +struct op_buffer { + unsigned len; + struct rte_crypto_op *buffer[MAX_PKT_BURST]; +}; + +#define MAX_RX_QUEUE_PER_LCORE 16 +#define MAX_TX_QUEUE_PER_PORT 16 + +enum l2fwd_crypto_xform_chain { + L2FWD_CRYPTO_CIPHER_HASH, + L2FWD_CRYPTO_HASH_CIPHER, + L2FWD_CRYPTO_CIPHER_ONLY, + L2FWD_CRYPTO_HASH_ONLY +}; + +struct l2fwd_key { + uint8_t *data; + uint32_t length; + phys_addr_t phys_addr; +}; + +/** l2fwd crypto application command line options */ +struct l2fwd_crypto_options { + unsigned portmask; + unsigned nb_ports_per_lcore; + unsigned refresh_period; + unsigned single_lcore:1; + + enum cdev_type type; + unsigned sessionless:1; + + enum l2fwd_crypto_xform_chain xform_chain; + + struct rte_crypto_sym_xform cipher_xform; + unsigned ckey_param; + int ckey_random_size; + + struct l2fwd_key iv; + unsigned iv_param; + int iv_random_size; + + struct rte_crypto_sym_xform auth_xform; + uint8_t akey_param; + int akey_random_size; + + struct l2fwd_key aad; + unsigned aad_param; + int aad_random_size; + + int digest_size; + + uint16_t block_size; + char string_type[MAX_STR_LEN]; + + uint64_t cryptodev_mask; +}; + +/** l2fwd crypto lcore params */ +struct l2fwd_crypto_params { + uint8_t dev_id; + uint8_t qp_id; + + unsigned digest_length; + unsigned block_size; + + struct l2fwd_key iv; + struct l2fwd_key aad; + struct rte_cryptodev_sym_session *session; + + uint8_t do_cipher; + uint8_t do_hash; + uint8_t hash_verify; + + enum rte_crypto_cipher_algorithm cipher_algo; + enum rte_crypto_auth_algorithm auth_algo; +}; + +/** lcore configuration */ +struct lcore_queue_conf { + unsigned nb_rx_ports; + unsigned rx_port_list[MAX_RX_QUEUE_PER_LCORE]; + + unsigned nb_crypto_devs; + unsigned cryptodev_list[MAX_RX_QUEUE_PER_LCORE]; + + struct op_buffer op_buf[RTE_CRYPTO_MAX_DEVS]; + struct pkt_buffer pkt_buf[RTE_MAX_ETHPORTS]; +} __rte_cache_aligned; + +struct lcore_queue_conf lcore_queue_conf[RTE_MAX_LCORE]; + +static const struct rte_eth_conf port_conf = { + .rxmode = { + .mq_mode = ETH_MQ_RX_NONE, + .max_rx_pkt_len = ETHER_MAX_LEN, + .split_hdr_size = 0, + .header_split = 0, /**< Header Split disabled */ + .hw_ip_checksum = 0, /**< IP checksum offload disabled */ + .hw_vlan_filter = 0, /**< VLAN filtering disabled */ + .jumbo_frame = 0, /**< Jumbo Frame Support disabled */ + .hw_strip_crc = 1, /**< CRC stripped by hardware */ + }, + .txmode = { + .mq_mode = ETH_MQ_TX_NONE, + }, +}; + +struct rte_mempool *l2fwd_pktmbuf_pool; +struct rte_mempool *l2fwd_crypto_op_pool; + +/* Per-port statistics struct */ +struct l2fwd_port_statistics { + uint64_t tx; + uint64_t rx; + + uint64_t crypto_enqueued; + uint64_t crypto_dequeued; + + uint64_t dropped; +} __rte_cache_aligned; + +struct l2fwd_crypto_statistics { + uint64_t enqueued; + uint64_t dequeued; + + uint64_t errors; +} __rte_cache_aligned; + +struct l2fwd_port_statistics port_statistics[RTE_MAX_ETHPORTS]; +struct l2fwd_crypto_statistics crypto_statistics[RTE_CRYPTO_MAX_DEVS]; + +/* A tsc-based timer responsible for triggering statistics printout */ +#define TIMER_MILLISECOND 2000000ULL /* around 1ms at 2 Ghz */ +#define MAX_TIMER_PERIOD 86400UL /* 1 day max */ + +/* default period is 10 seconds */ +static int64_t timer_period = 10 * TIMER_MILLISECOND * 1000; + +/* Print out statistics on packets dropped */ +static void +print_stats(void) +{ + uint64_t total_packets_dropped, total_packets_tx, total_packets_rx; + uint64_t total_packets_enqueued, total_packets_dequeued, + total_packets_errors; + unsigned portid; + uint64_t cdevid; + + total_packets_dropped = 0; + total_packets_tx = 0; + total_packets_rx = 0; + total_packets_enqueued = 0; + total_packets_dequeued = 0; + total_packets_errors = 0; + + const char clr[] = { 27, '[', '2', 'J', '\0' }; + const char topLeft[] = { 27, '[', '1', ';', '1', 'H', '\0' }; + + /* Clear screen and move to top left */ + printf("%s%s", clr, topLeft); + + printf("\nPort statistics ===================================="); + + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) { + /* skip disabled ports */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + printf("\nStatistics for port %u ------------------------------" + "\nPackets sent: %32"PRIu64 + "\nPackets received: %28"PRIu64 + "\nPackets dropped: %29"PRIu64, + portid, + port_statistics[portid].tx, + port_statistics[portid].rx, + port_statistics[portid].dropped); + + total_packets_dropped += port_statistics[portid].dropped; + total_packets_tx += port_statistics[portid].tx; + total_packets_rx += port_statistics[portid].rx; + } + printf("\nCrypto statistics =================================="); + + for (cdevid = 0; cdevid < RTE_CRYPTO_MAX_DEVS; cdevid++) { + /* skip disabled ports */ + if ((l2fwd_enabled_crypto_mask & (((uint64_t)1) << cdevid)) == 0) + continue; + printf("\nStatistics for cryptodev %"PRIu64 + " -------------------------" + "\nPackets enqueued: %28"PRIu64 + "\nPackets dequeued: %28"PRIu64 + "\nPackets errors: %30"PRIu64, + cdevid, + crypto_statistics[cdevid].enqueued, + crypto_statistics[cdevid].dequeued, + crypto_statistics[cdevid].errors); + + total_packets_enqueued += crypto_statistics[cdevid].enqueued; + total_packets_dequeued += crypto_statistics[cdevid].dequeued; + total_packets_errors += crypto_statistics[cdevid].errors; + } + printf("\nAggregate statistics ===============================" + "\nTotal packets received: %22"PRIu64 + "\nTotal packets enqueued: %22"PRIu64 + "\nTotal packets dequeued: %22"PRIu64 + "\nTotal packets sent: %26"PRIu64 + "\nTotal packets dropped: %23"PRIu64 + "\nTotal packets crypto errors: %17"PRIu64, + total_packets_rx, + total_packets_enqueued, + total_packets_dequeued, + total_packets_tx, + total_packets_dropped, + total_packets_errors); + printf("\n====================================================\n"); +} + +static int +l2fwd_crypto_send_burst(struct lcore_queue_conf *qconf, unsigned n, + struct l2fwd_crypto_params *cparams) +{ + struct rte_crypto_op **op_buffer; + unsigned ret; + + op_buffer = (struct rte_crypto_op **) + qconf->op_buf[cparams->dev_id].buffer; + + ret = rte_cryptodev_enqueue_burst(cparams->dev_id, + cparams->qp_id, op_buffer, (uint16_t) n); + + crypto_statistics[cparams->dev_id].enqueued += ret; + if (unlikely(ret < n)) { + crypto_statistics[cparams->dev_id].errors += (n - ret); + do { + rte_pktmbuf_free(op_buffer[ret]->sym->m_src); + rte_crypto_op_free(op_buffer[ret]); + } while (++ret < n); + } + + return 0; +} + +static int +l2fwd_crypto_enqueue(struct rte_crypto_op *op, + struct l2fwd_crypto_params *cparams) +{ + unsigned lcore_id, len; + struct lcore_queue_conf *qconf; + + lcore_id = rte_lcore_id(); + + qconf = &lcore_queue_conf[lcore_id]; + len = qconf->op_buf[cparams->dev_id].len; + qconf->op_buf[cparams->dev_id].buffer[len] = op; + len++; + + /* enough ops to be sent */ + if (len == MAX_PKT_BURST) { + l2fwd_crypto_send_burst(qconf, MAX_PKT_BURST, cparams); + len = 0; + } + + qconf->op_buf[cparams->dev_id].len = len; + return 0; +} + +static int +l2fwd_simple_crypto_enqueue(struct rte_mbuf *m, + struct rte_crypto_op *op, + struct l2fwd_crypto_params *cparams) +{ + struct ether_hdr *eth_hdr; + struct ipv4_hdr *ip_hdr; + + uint32_t ipdata_offset, data_len; + uint32_t pad_len = 0; + char *padding; + + eth_hdr = rte_pktmbuf_mtod(m, struct ether_hdr *); + + if (eth_hdr->ether_type != rte_cpu_to_be_16(ETHER_TYPE_IPv4)) + return -1; + + ipdata_offset = sizeof(struct ether_hdr); + + ip_hdr = (struct ipv4_hdr *)(rte_pktmbuf_mtod(m, char *) + + ipdata_offset); + + ipdata_offset += (ip_hdr->version_ihl & IPV4_HDR_IHL_MASK) + * IPV4_IHL_MULTIPLIER; + + + /* Zero pad data to be crypto'd so it is block aligned */ + data_len = rte_pktmbuf_data_len(m) - ipdata_offset; + + if (cparams->do_hash && cparams->hash_verify) + data_len -= cparams->digest_length; + + if (cparams->do_cipher) { + /* + * Following algorithms are block cipher algorithms, + * and might need padding + */ + switch (cparams->cipher_algo) { + case RTE_CRYPTO_CIPHER_AES_CBC: + case RTE_CRYPTO_CIPHER_AES_ECB: + case RTE_CRYPTO_CIPHER_DES_CBC: + case RTE_CRYPTO_CIPHER_3DES_CBC: + case RTE_CRYPTO_CIPHER_3DES_ECB: + if (data_len % cparams->block_size) + pad_len = cparams->block_size - + (data_len % cparams->block_size); + break; + default: + pad_len = 0; + } + + if (pad_len) { + padding = rte_pktmbuf_append(m, pad_len); + if (unlikely(!padding)) + return -1; + + data_len += pad_len; + memset(padding, 0, pad_len); + } + } + + /* Set crypto operation data parameters */ + rte_crypto_op_attach_sym_session(op, cparams->session); + + if (cparams->do_hash) { + if (!cparams->hash_verify) { + /* Append space for digest to end of packet */ + op->sym->auth.digest.data = (uint8_t *)rte_pktmbuf_append(m, + cparams->digest_length); + } else { + op->sym->auth.digest.data = rte_pktmbuf_mtod(m, + uint8_t *) + ipdata_offset + data_len; + } + + op->sym->auth.digest.phys_addr = rte_pktmbuf_mtophys_offset(m, + rte_pktmbuf_pkt_len(m) - cparams->digest_length); + op->sym->auth.digest.length = cparams->digest_length; + + /* For wireless algorithms, offset/length must be in bits */ + if (cparams->auth_algo == RTE_CRYPTO_AUTH_SNOW3G_UIA2 || + cparams->auth_algo == RTE_CRYPTO_AUTH_KASUMI_F9 || + cparams->auth_algo == RTE_CRYPTO_AUTH_ZUC_EIA3) { + op->sym->auth.data.offset = ipdata_offset << 3; + op->sym->auth.data.length = data_len << 3; + } else { + op->sym->auth.data.offset = ipdata_offset; + op->sym->auth.data.length = data_len; + } + + if (cparams->aad.length) { + op->sym->auth.aad.data = cparams->aad.data; + op->sym->auth.aad.phys_addr = cparams->aad.phys_addr; + op->sym->auth.aad.length = cparams->aad.length; + } else { + op->sym->auth.aad.data = NULL; + op->sym->auth.aad.phys_addr = 0; + op->sym->auth.aad.length = 0; + } + } + + if (cparams->do_cipher) { + op->sym->cipher.iv.data = cparams->iv.data; + op->sym->cipher.iv.phys_addr = cparams->iv.phys_addr; + op->sym->cipher.iv.length = cparams->iv.length; + + /* For wireless algorithms, offset/length must be in bits */ + if (cparams->cipher_algo == RTE_CRYPTO_CIPHER_SNOW3G_UEA2 || + cparams->cipher_algo == RTE_CRYPTO_CIPHER_KASUMI_F8 || + cparams->cipher_algo == RTE_CRYPTO_CIPHER_ZUC_EEA3) { + op->sym->cipher.data.offset = ipdata_offset << 3; + op->sym->cipher.data.length = data_len << 3; + } else { + op->sym->cipher.data.offset = ipdata_offset; + op->sym->cipher.data.length = data_len; + } + } + + op->sym->m_src = m; + + return l2fwd_crypto_enqueue(op, cparams); +} + + +/* Send the burst of packets on an output interface */ +static int +l2fwd_send_burst(struct lcore_queue_conf *qconf, unsigned n, + uint8_t port) +{ + struct rte_mbuf **pkt_buffer; + unsigned ret; + + pkt_buffer = (struct rte_mbuf **)qconf->pkt_buf[port].buffer; + + ret = rte_eth_tx_burst(port, 0, pkt_buffer, (uint16_t)n); + port_statistics[port].tx += ret; + if (unlikely(ret < n)) { + port_statistics[port].dropped += (n - ret); + do { + rte_pktmbuf_free(pkt_buffer[ret]); + } while (++ret < n); + } + + return 0; +} + +/* Enqueue packets for TX and prepare them to be sent */ +static int +l2fwd_send_packet(struct rte_mbuf *m, uint8_t port) +{ + unsigned lcore_id, len; + struct lcore_queue_conf *qconf; + + lcore_id = rte_lcore_id(); + + qconf = &lcore_queue_conf[lcore_id]; + len = qconf->pkt_buf[port].len; + qconf->pkt_buf[port].buffer[len] = m; + len++; + + /* enough pkts to be sent */ + if (unlikely(len == MAX_PKT_BURST)) { + l2fwd_send_burst(qconf, MAX_PKT_BURST, port); + len = 0; + } + + qconf->pkt_buf[port].len = len; + return 0; +} + +static void +l2fwd_simple_forward(struct rte_mbuf *m, unsigned portid) +{ + struct ether_hdr *eth; + void *tmp; + unsigned dst_port; + + dst_port = l2fwd_dst_ports[portid]; + eth = rte_pktmbuf_mtod(m, struct ether_hdr *); + + /* 02:00:00:00:00:xx */ + tmp = ð->d_addr.addr_bytes[0]; + *((uint64_t *)tmp) = 0x000000000002 + ((uint64_t)dst_port << 40); + + /* src addr */ + ether_addr_copy(&l2fwd_ports_eth_addr[dst_port], ð->s_addr); + + l2fwd_send_packet(m, (uint8_t) dst_port); +} + +/** Generate random key */ +static void +generate_random_key(uint8_t *key, unsigned length) +{ + int fd; + int ret; + + fd = open("/dev/urandom", O_RDONLY); + if (fd < 0) + rte_exit(EXIT_FAILURE, "Failed to generate random key\n"); + + ret = read(fd, key, length); + close(fd); + + if (ret != (signed)length) + rte_exit(EXIT_FAILURE, "Failed to generate random key\n"); +} + +static struct rte_cryptodev_sym_session * +initialize_crypto_session(struct l2fwd_crypto_options *options, + uint8_t cdev_id) +{ + struct rte_crypto_sym_xform *first_xform; + + if (options->xform_chain == L2FWD_CRYPTO_CIPHER_HASH) { + first_xform = &options->cipher_xform; + first_xform->next = &options->auth_xform; + } else if (options->xform_chain == L2FWD_CRYPTO_HASH_CIPHER) { + first_xform = &options->auth_xform; + first_xform->next = &options->cipher_xform; + } else if (options->xform_chain == L2FWD_CRYPTO_CIPHER_ONLY) { + first_xform = &options->cipher_xform; + } else { + first_xform = &options->auth_xform; + } + + /* Setup Cipher Parameters */ + return rte_cryptodev_sym_session_create(cdev_id, first_xform); +} + +static void +l2fwd_crypto_options_print(struct l2fwd_crypto_options *options); + +/* main processing loop */ +static void +l2fwd_main_loop(struct l2fwd_crypto_options *options) +{ + struct rte_mbuf *m, *pkts_burst[MAX_PKT_BURST]; + struct rte_crypto_op *ops_burst[MAX_PKT_BURST]; + + unsigned lcore_id = rte_lcore_id(); + uint64_t prev_tsc = 0, diff_tsc, cur_tsc, timer_tsc = 0; + unsigned i, j, portid, nb_rx, len; + struct lcore_queue_conf *qconf = &lcore_queue_conf[lcore_id]; + const uint64_t drain_tsc = (rte_get_tsc_hz() + US_PER_S - 1) / + US_PER_S * BURST_TX_DRAIN_US; + struct l2fwd_crypto_params *cparams; + struct l2fwd_crypto_params port_cparams[qconf->nb_crypto_devs]; + + if (qconf->nb_rx_ports == 0) { + RTE_LOG(INFO, L2FWD, "lcore %u has nothing to do\n", lcore_id); + return; + } + + RTE_LOG(INFO, L2FWD, "entering main loop on lcore %u\n", lcore_id); + + for (i = 0; i < qconf->nb_rx_ports; i++) { + + portid = qconf->rx_port_list[i]; + RTE_LOG(INFO, L2FWD, " -- lcoreid=%u portid=%u\n", lcore_id, + portid); + } + + for (i = 0; i < qconf->nb_crypto_devs; i++) { + port_cparams[i].do_cipher = 0; + port_cparams[i].do_hash = 0; + + switch (options->xform_chain) { + case L2FWD_CRYPTO_CIPHER_HASH: + case L2FWD_CRYPTO_HASH_CIPHER: + port_cparams[i].do_cipher = 1; + port_cparams[i].do_hash = 1; + break; + case L2FWD_CRYPTO_HASH_ONLY: + port_cparams[i].do_hash = 1; + break; + case L2FWD_CRYPTO_CIPHER_ONLY: + port_cparams[i].do_cipher = 1; + break; + } + + port_cparams[i].dev_id = qconf->cryptodev_list[i]; + port_cparams[i].qp_id = 0; + + port_cparams[i].block_size = options->block_size; + + if (port_cparams[i].do_hash) { + port_cparams[i].digest_length = + options->auth_xform.auth.digest_length; + if (options->auth_xform.auth.add_auth_data_length) { + port_cparams[i].aad.data = options->aad.data; + port_cparams[i].aad.length = + options->auth_xform.auth.add_auth_data_length; + port_cparams[i].aad.phys_addr = options->aad.phys_addr; + if (!options->aad_param) + generate_random_key(port_cparams[i].aad.data, + port_cparams[i].aad.length); + + } else + port_cparams[i].aad.length = 0; + + if (options->auth_xform.auth.op == RTE_CRYPTO_AUTH_OP_VERIFY) + port_cparams[i].hash_verify = 1; + else + port_cparams[i].hash_verify = 0; + + port_cparams[i].auth_algo = options->auth_xform.auth.algo; + } + + if (port_cparams[i].do_cipher) { + port_cparams[i].iv.data = options->iv.data; + port_cparams[i].iv.length = options->iv.length; + port_cparams[i].iv.phys_addr = options->iv.phys_addr; + if (!options->iv_param) + generate_random_key(port_cparams[i].iv.data, + port_cparams[i].iv.length); + + port_cparams[i].cipher_algo = options->cipher_xform.cipher.algo; + } + + port_cparams[i].session = initialize_crypto_session(options, + port_cparams[i].dev_id); + + if (port_cparams[i].session == NULL) + return; + RTE_LOG(INFO, L2FWD, " -- lcoreid=%u cryptoid=%u\n", lcore_id, + port_cparams[i].dev_id); + } + + l2fwd_crypto_options_print(options); + + /* + * Initialize previous tsc timestamp before the loop, + * to avoid showing the port statistics immediately, + * so user can see the crypto information. + */ + prev_tsc = rte_rdtsc(); + while (1) { + + cur_tsc = rte_rdtsc(); + + /* + * Crypto device/TX burst queue drain + */ + diff_tsc = cur_tsc - prev_tsc; + if (unlikely(diff_tsc > drain_tsc)) { + /* Enqueue all crypto ops remaining in buffers */ + for (i = 0; i < qconf->nb_crypto_devs; i++) { + cparams = &port_cparams[i]; + len = qconf->op_buf[cparams->dev_id].len; + l2fwd_crypto_send_burst(qconf, len, cparams); + qconf->op_buf[cparams->dev_id].len = 0; + } + /* Transmit all packets remaining in buffers */ + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) { + if (qconf->pkt_buf[portid].len == 0) + continue; + l2fwd_send_burst(&lcore_queue_conf[lcore_id], + qconf->pkt_buf[portid].len, + (uint8_t) portid); + qconf->pkt_buf[portid].len = 0; + } + + /* if timer is enabled */ + if (timer_period > 0) { + + /* advance the timer */ + timer_tsc += diff_tsc; + + /* if timer has reached its timeout */ + if (unlikely(timer_tsc >= + (uint64_t)timer_period)) { + + /* do this only on master core */ + if (lcore_id == rte_get_master_lcore() + && options->refresh_period) { + print_stats(); + timer_tsc = 0; + } + } + } + + prev_tsc = cur_tsc; + } + + /* + * Read packet from RX queues + */ + for (i = 0; i < qconf->nb_rx_ports; i++) { + portid = qconf->rx_port_list[i]; + + cparams = &port_cparams[i]; + + nb_rx = rte_eth_rx_burst((uint8_t) portid, 0, + pkts_burst, MAX_PKT_BURST); + + port_statistics[portid].rx += nb_rx; + + if (nb_rx) { + /* + * If we can't allocate a crypto_ops, then drop + * the rest of the burst and dequeue and + * process the packets to free offload structs + */ + if (rte_crypto_op_bulk_alloc( + l2fwd_crypto_op_pool, + RTE_CRYPTO_OP_TYPE_SYMMETRIC, + ops_burst, nb_rx) != + nb_rx) { + for (j = 0; j < nb_rx; j++) + rte_pktmbuf_free(pkts_burst[j]); + + nb_rx = 0; + } + + /* Enqueue packets from Crypto device*/ + for (j = 0; j < nb_rx; j++) { + m = pkts_burst[j]; + + l2fwd_simple_crypto_enqueue(m, + ops_burst[j], cparams); + } + } + + /* Dequeue packets from Crypto device */ + do { + nb_rx = rte_cryptodev_dequeue_burst( + cparams->dev_id, cparams->qp_id, + ops_burst, MAX_PKT_BURST); + + crypto_statistics[cparams->dev_id].dequeued += + nb_rx; + + /* Forward crypto'd packets */ + for (j = 0; j < nb_rx; j++) { + m = ops_burst[j]->sym->m_src; + + rte_crypto_op_free(ops_burst[j]); + l2fwd_simple_forward(m, portid); + } + } while (nb_rx == MAX_PKT_BURST); + } + } +} + +static int +l2fwd_launch_one_lcore(void *arg) +{ + l2fwd_main_loop((struct l2fwd_crypto_options *)arg); + return 0; +} + +/* Display command line arguments usage */ +static void +l2fwd_crypto_usage(const char *prgname) +{ + printf("%s [EAL options] --\n" + " -p PORTMASK: hexadecimal bitmask of ports to configure\n" + " -q NQ: number of queue (=ports) per lcore (default is 1)\n" + " -s manage all ports from single lcore\n" + " -T PERIOD: statistics will be refreshed each PERIOD seconds" + " (0 to disable, 10 default, 86400 maximum)\n" + + " --cdev_type HW / SW / ANY\n" + " --chain HASH_CIPHER / CIPHER_HASH\n" + + " --cipher_algo ALGO\n" + " --cipher_op ENCRYPT / DECRYPT\n" + " --cipher_key KEY (bytes separated with \":\")\n" + " --cipher_key_random_size SIZE: size of cipher key when generated randomly\n" + " --iv IV (bytes separated with \":\")\n" + " --iv_random_size SIZE: size of IV when generated randomly\n" + + " --auth_algo ALGO\n" + " --auth_op GENERATE / VERIFY\n" + " --auth_key KEY (bytes separated with \":\")\n" + " --auth_key_random_size SIZE: size of auth key when generated randomly\n" + " --aad AAD (bytes separated with \":\")\n" + " --aad_random_size SIZE: size of AAD when generated randomly\n" + " --digest_size SIZE: size of digest to be generated/verified\n" + + " --sessionless\n" + " --cryptodev_mask MASK: hexadecimal bitmask of crypto devices to configure\n", + prgname); +} + +/** Parse crypto device type command line argument */ +static int +parse_cryptodev_type(enum cdev_type *type, char *optarg) +{ + if (strcmp("HW", optarg) == 0) { + *type = CDEV_TYPE_HW; + return 0; + } else if (strcmp("SW", optarg) == 0) { + *type = CDEV_TYPE_SW; + return 0; + } else if (strcmp("ANY", optarg) == 0) { + *type = CDEV_TYPE_ANY; + return 0; + } + + return -1; +} + +/** Parse crypto chain xform command line argument */ +static int +parse_crypto_opt_chain(struct l2fwd_crypto_options *options, char *optarg) +{ + if (strcmp("CIPHER_HASH", optarg) == 0) { + options->xform_chain = L2FWD_CRYPTO_CIPHER_HASH; + return 0; + } else if (strcmp("HASH_CIPHER", optarg) == 0) { + options->xform_chain = L2FWD_CRYPTO_HASH_CIPHER; + return 0; + } else if (strcmp("CIPHER_ONLY", optarg) == 0) { + options->xform_chain = L2FWD_CRYPTO_CIPHER_ONLY; + return 0; + } else if (strcmp("HASH_ONLY", optarg) == 0) { + options->xform_chain = L2FWD_CRYPTO_HASH_ONLY; + return 0; + } + + return -1; +} + +/** Parse crypto cipher algo option command line argument */ +static int +parse_cipher_algo(enum rte_crypto_cipher_algorithm *algo, char *optarg) +{ + + if (rte_cryptodev_get_cipher_algo_enum(algo, optarg) < 0) { + RTE_LOG(ERR, USER1, "Cipher algorithm specified " + "not supported!\n"); + return -1; + } + + return 0; +} + +/** Parse crypto cipher operation command line argument */ +static int +parse_cipher_op(enum rte_crypto_cipher_operation *op, char *optarg) +{ + if (strcmp("ENCRYPT", optarg) == 0) { + *op = RTE_CRYPTO_CIPHER_OP_ENCRYPT; + return 0; + } else if (strcmp("DECRYPT", optarg) == 0) { + *op = RTE_CRYPTO_CIPHER_OP_DECRYPT; + return 0; + } + + printf("Cipher operation not supported!\n"); + return -1; +} + +/** Parse crypto key command line argument */ +static int +parse_key(uint8_t *data, char *input_arg) +{ + unsigned byte_count; + char *token; + + for (byte_count = 0, token = strtok(input_arg, ":"); + (byte_count < MAX_KEY_SIZE) && (token != NULL); + token = strtok(NULL, ":")) { + + int number = (int)strtol(token, NULL, 16); + + if (errno == EINVAL || errno == ERANGE || number > 0xFF) + return -1; + + data[byte_count++] = (uint8_t)number; + } + + return byte_count; +} + +/** Parse size param*/ +static int +parse_size(int *size, const char *q_arg) +{ + char *end = NULL; + unsigned long n; + + /* parse hexadecimal string */ + n = strtoul(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + n = 0; + + if (n == 0) { + printf("invalid size\n"); + return -1; + } + + *size = n; + return 0; +} + +/** Parse crypto cipher operation command line argument */ +static int +parse_auth_algo(enum rte_crypto_auth_algorithm *algo, char *optarg) +{ + if (rte_cryptodev_get_auth_algo_enum(algo, optarg) < 0) { + RTE_LOG(ERR, USER1, "Authentication algorithm specified " + "not supported!\n"); + return -1; + } + + return 0; +} + +static int +parse_auth_op(enum rte_crypto_auth_operation *op, char *optarg) +{ + if (strcmp("VERIFY", optarg) == 0) { + *op = RTE_CRYPTO_AUTH_OP_VERIFY; + return 0; + } else if (strcmp("GENERATE", optarg) == 0) { + *op = RTE_CRYPTO_AUTH_OP_GENERATE; + return 0; + } + + printf("Authentication operation specified not supported!\n"); + return -1; +} + +static int +parse_cryptodev_mask(struct l2fwd_crypto_options *options, + const char *q_arg) +{ + char *end = NULL; + uint64_t pm; + + /* parse hexadecimal string */ + pm = strtoul(q_arg, &end, 16); + if ((pm == '\0') || (end == NULL) || (*end != '\0')) + pm = 0; + + options->cryptodev_mask = pm; + if (options->cryptodev_mask == 0) { + printf("invalid cryptodev_mask specified\n"); + return -1; + } + + return 0; +} + +/** Parse long options */ +static int +l2fwd_crypto_parse_args_long_options(struct l2fwd_crypto_options *options, + struct option *lgopts, int option_index) +{ + int retval; + + if (strcmp(lgopts[option_index].name, "cdev_type") == 0) { + retval = parse_cryptodev_type(&options->type, optarg); + if (retval == 0) + snprintf(options->string_type, MAX_STR_LEN, + "%s", optarg); + return retval; + } + + else if (strcmp(lgopts[option_index].name, "chain") == 0) + return parse_crypto_opt_chain(options, optarg); + + /* Cipher options */ + else if (strcmp(lgopts[option_index].name, "cipher_algo") == 0) + return parse_cipher_algo(&options->cipher_xform.cipher.algo, + optarg); + + else if (strcmp(lgopts[option_index].name, "cipher_op") == 0) + return parse_cipher_op(&options->cipher_xform.cipher.op, + optarg); + + else if (strcmp(lgopts[option_index].name, "cipher_key") == 0) { + options->ckey_param = 1; + options->cipher_xform.cipher.key.length = + parse_key(options->cipher_xform.cipher.key.data, optarg); + if (options->cipher_xform.cipher.key.length > 0) + return 0; + else + return -1; + } + + else if (strcmp(lgopts[option_index].name, "cipher_key_random_size") == 0) + return parse_size(&options->ckey_random_size, optarg); + + else if (strcmp(lgopts[option_index].name, "iv") == 0) { + options->iv_param = 1; + options->iv.length = + parse_key(options->iv.data, optarg); + if (options->iv.length > 0) + return 0; + else + return -1; + } + + else if (strcmp(lgopts[option_index].name, "iv_random_size") == 0) + return parse_size(&options->iv_random_size, optarg); + + /* Authentication options */ + else if (strcmp(lgopts[option_index].name, "auth_algo") == 0) { + return parse_auth_algo(&options->auth_xform.auth.algo, + optarg); + } + + else if (strcmp(lgopts[option_index].name, "auth_op") == 0) + return parse_auth_op(&options->auth_xform.auth.op, + optarg); + + else if (strcmp(lgopts[option_index].name, "auth_key") == 0) { + options->akey_param = 1; + options->auth_xform.auth.key.length = + parse_key(options->auth_xform.auth.key.data, optarg); + if (options->auth_xform.auth.key.length > 0) + return 0; + else + return -1; + } + + else if (strcmp(lgopts[option_index].name, "auth_key_random_size") == 0) { + return parse_size(&options->akey_random_size, optarg); + } + + else if (strcmp(lgopts[option_index].name, "aad") == 0) { + options->aad_param = 1; + options->aad.length = + parse_key(options->aad.data, optarg); + if (options->aad.length > 0) + return 0; + else + return -1; + } + + else if (strcmp(lgopts[option_index].name, "aad_random_size") == 0) { + return parse_size(&options->aad_random_size, optarg); + } + + else if (strcmp(lgopts[option_index].name, "digest_size") == 0) { + return parse_size(&options->digest_size, optarg); + } + + else if (strcmp(lgopts[option_index].name, "sessionless") == 0) { + options->sessionless = 1; + return 0; + } + + else if (strcmp(lgopts[option_index].name, "cryptodev_mask") == 0) + return parse_cryptodev_mask(options, optarg); + + return -1; +} + +/** Parse port mask */ +static int +l2fwd_crypto_parse_portmask(struct l2fwd_crypto_options *options, + const char *q_arg) +{ + char *end = NULL; + unsigned long pm; + + /* parse hexadecimal string */ + pm = strtoul(q_arg, &end, 16); + if ((pm == '\0') || (end == NULL) || (*end != '\0')) + pm = 0; + + options->portmask = pm; + if (options->portmask == 0) { + printf("invalid portmask specified\n"); + return -1; + } + + return pm; +} + +/** Parse number of queues */ +static int +l2fwd_crypto_parse_nqueue(struct l2fwd_crypto_options *options, + const char *q_arg) +{ + char *end = NULL; + unsigned long n; + + /* parse hexadecimal string */ + n = strtoul(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + n = 0; + else if (n >= MAX_RX_QUEUE_PER_LCORE) + n = 0; + + options->nb_ports_per_lcore = n; + if (options->nb_ports_per_lcore == 0) { + printf("invalid number of ports selected\n"); + return -1; + } + + return 0; +} + +/** Parse timer period */ +static int +l2fwd_crypto_parse_timer_period(struct l2fwd_crypto_options *options, + const char *q_arg) +{ + char *end = NULL; + unsigned long n; + + /* parse number string */ + n = (unsigned)strtol(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + n = 0; + + if (n >= MAX_TIMER_PERIOD) { + printf("Warning refresh period specified %lu is greater than " + "max value %lu! using max value", + n, MAX_TIMER_PERIOD); + n = MAX_TIMER_PERIOD; + } + + options->refresh_period = n * 1000 * TIMER_MILLISECOND; + + return 0; +} + +/** Generate default options for application */ +static void +l2fwd_crypto_default_options(struct l2fwd_crypto_options *options) +{ + options->portmask = 0xffffffff; + options->nb_ports_per_lcore = 1; + options->refresh_period = 10000; + options->single_lcore = 0; + options->sessionless = 0; + + options->xform_chain = L2FWD_CRYPTO_CIPHER_HASH; + + /* Cipher Data */ + options->cipher_xform.type = RTE_CRYPTO_SYM_XFORM_CIPHER; + options->cipher_xform.next = NULL; + options->ckey_param = 0; + options->ckey_random_size = -1; + options->cipher_xform.cipher.key.length = 0; + options->iv_param = 0; + options->iv_random_size = -1; + options->iv.length = 0; + + options->cipher_xform.cipher.algo = RTE_CRYPTO_CIPHER_AES_CBC; + options->cipher_xform.cipher.op = RTE_CRYPTO_CIPHER_OP_ENCRYPT; + + /* Authentication Data */ + options->auth_xform.type = RTE_CRYPTO_SYM_XFORM_AUTH; + options->auth_xform.next = NULL; + options->akey_param = 0; + options->akey_random_size = -1; + options->auth_xform.auth.key.length = 0; + options->aad_param = 0; + options->aad_random_size = -1; + options->aad.length = 0; + options->digest_size = -1; + + options->auth_xform.auth.algo = RTE_CRYPTO_AUTH_SHA1_HMAC; + options->auth_xform.auth.op = RTE_CRYPTO_AUTH_OP_GENERATE; + + options->type = CDEV_TYPE_ANY; + options->cryptodev_mask = UINT64_MAX; +} + +static void +display_cipher_info(struct l2fwd_crypto_options *options) +{ + printf("\n---- Cipher information ---\n"); + printf("Algorithm: %s\n", + rte_crypto_cipher_algorithm_strings[options->cipher_xform.cipher.algo]); + rte_hexdump(stdout, "Cipher key:", + options->cipher_xform.cipher.key.data, + options->cipher_xform.cipher.key.length); + rte_hexdump(stdout, "IV:", options->iv.data, options->iv.length); +} + +static void +display_auth_info(struct l2fwd_crypto_options *options) +{ + printf("\n---- Authentication information ---\n"); + printf("Algorithm: %s\n", + rte_crypto_auth_algorithm_strings[options->auth_xform.cipher.algo]); + rte_hexdump(stdout, "Auth key:", + options->auth_xform.auth.key.data, + options->auth_xform.auth.key.length); + rte_hexdump(stdout, "AAD:", options->aad.data, options->aad.length); +} + +static void +l2fwd_crypto_options_print(struct l2fwd_crypto_options *options) +{ + char string_cipher_op[MAX_STR_LEN]; + char string_auth_op[MAX_STR_LEN]; + + if (options->cipher_xform.cipher.op == RTE_CRYPTO_CIPHER_OP_ENCRYPT) + strcpy(string_cipher_op, "Encrypt"); + else + strcpy(string_cipher_op, "Decrypt"); + + if (options->auth_xform.auth.op == RTE_CRYPTO_AUTH_OP_GENERATE) + strcpy(string_auth_op, "Auth generate"); + else + strcpy(string_auth_op, "Auth verify"); + + printf("Options:-\nn"); + printf("portmask: %x\n", options->portmask); + printf("ports per lcore: %u\n", options->nb_ports_per_lcore); + printf("refresh period : %u\n", options->refresh_period); + printf("single lcore mode: %s\n", + options->single_lcore ? "enabled" : "disabled"); + printf("stats_printing: %s\n", + options->refresh_period == 0 ? "disabled" : "enabled"); + + printf("sessionless crypto: %s\n", + options->sessionless ? "enabled" : "disabled"); + + if (options->ckey_param && (options->ckey_random_size != -1)) + printf("Cipher key already parsed, ignoring size of random key\n"); + + if (options->akey_param && (options->akey_random_size != -1)) + printf("Auth key already parsed, ignoring size of random key\n"); + + if (options->iv_param && (options->iv_random_size != -1)) + printf("IV already parsed, ignoring size of random IV\n"); + + if (options->aad_param && (options->aad_random_size != -1)) + printf("AAD already parsed, ignoring size of random AAD\n"); + + printf("\nCrypto chain: "); + switch (options->xform_chain) { + case L2FWD_CRYPTO_CIPHER_HASH: + printf("Input --> %s --> %s --> Output\n", + string_cipher_op, string_auth_op); + display_cipher_info(options); + display_auth_info(options); + break; + case L2FWD_CRYPTO_HASH_CIPHER: + printf("Input --> %s --> %s --> Output\n", + string_auth_op, string_cipher_op); + display_cipher_info(options); + display_auth_info(options); + break; + case L2FWD_CRYPTO_HASH_ONLY: + printf("Input --> %s --> Output\n", string_auth_op); + display_auth_info(options); + break; + case L2FWD_CRYPTO_CIPHER_ONLY: + printf("Input --> %s --> Output\n", string_cipher_op); + display_cipher_info(options); + break; + } +} + +/* Parse the argument given in the command line of the application */ +static int +l2fwd_crypto_parse_args(struct l2fwd_crypto_options *options, + int argc, char **argv) +{ + int opt, retval, option_index; + char **argvopt = argv, *prgname = argv[0]; + + static struct option lgopts[] = { + { "sessionless", no_argument, 0, 0 }, + + { "cdev_type", required_argument, 0, 0 }, + { "chain", required_argument, 0, 0 }, + + { "cipher_algo", required_argument, 0, 0 }, + { "cipher_op", required_argument, 0, 0 }, + { "cipher_key", required_argument, 0, 0 }, + { "cipher_key_random_size", required_argument, 0, 0 }, + + { "auth_algo", required_argument, 0, 0 }, + { "auth_op", required_argument, 0, 0 }, + { "auth_key", required_argument, 0, 0 }, + { "auth_key_random_size", required_argument, 0, 0 }, + + { "iv", required_argument, 0, 0 }, + { "iv_random_size", required_argument, 0, 0 }, + { "aad", required_argument, 0, 0 }, + { "aad_random_size", required_argument, 0, 0 }, + { "digest_size", required_argument, 0, 0 }, + + { "sessionless", no_argument, 0, 0 }, + { "cryptodev_mask", required_argument, 0, 0}, + + { NULL, 0, 0, 0 } + }; + + l2fwd_crypto_default_options(options); + + while ((opt = getopt_long(argc, argvopt, "p:q:st:", lgopts, + &option_index)) != EOF) { + switch (opt) { + /* long options */ + case 0: + retval = l2fwd_crypto_parse_args_long_options(options, + lgopts, option_index); + if (retval < 0) { + l2fwd_crypto_usage(prgname); + return -1; + } + break; + + /* portmask */ + case 'p': + retval = l2fwd_crypto_parse_portmask(options, optarg); + if (retval < 0) { + l2fwd_crypto_usage(prgname); + return -1; + } + break; + + /* nqueue */ + case 'q': + retval = l2fwd_crypto_parse_nqueue(options, optarg); + if (retval < 0) { + l2fwd_crypto_usage(prgname); + return -1; + } + break; + + /* single */ + case 's': + options->single_lcore = 1; + + break; + + /* timer period */ + case 'T': + retval = l2fwd_crypto_parse_timer_period(options, + optarg); + if (retval < 0) { + l2fwd_crypto_usage(prgname); + return -1; + } + break; + + default: + l2fwd_crypto_usage(prgname); + return -1; + } + } + + + if (optind >= 0) + argv[optind-1] = prgname; + + retval = optind-1; + optind = 1; /* reset getopt lib */ + + return retval; +} + +/* Check the link status of all ports in up to 9s, and print them finally */ +static void +check_all_ports_link_status(uint8_t port_num, uint32_t port_mask) +{ +#define CHECK_INTERVAL 100 /* 100ms */ +#define MAX_CHECK_TIME 90 /* 9s (90 * 100ms) in total */ + uint8_t portid, count, all_ports_up, print_flag = 0; + struct rte_eth_link link; + + printf("\nChecking link status"); + fflush(stdout); + for (count = 0; count <= MAX_CHECK_TIME; count++) { + all_ports_up = 1; + for (portid = 0; portid < port_num; portid++) { + if ((port_mask & (1 << portid)) == 0) + continue; + memset(&link, 0, sizeof(link)); + rte_eth_link_get_nowait(portid, &link); + /* print link status if flag set */ + if (print_flag == 1) { + if (link.link_status) + printf("Port %d Link Up - speed %u " + "Mbps - %s\n", (uint8_t)portid, + (unsigned)link.link_speed, + (link.link_duplex == ETH_LINK_FULL_DUPLEX) ? + ("full-duplex") : ("half-duplex\n")); + else + printf("Port %d Link Down\n", + (uint8_t)portid); + continue; + } + /* clear all_ports_up flag if any link down */ + if (link.link_status == ETH_LINK_DOWN) { + all_ports_up = 0; + break; + } + } + /* after finally printing all link status, get out */ + if (print_flag == 1) + break; + + if (all_ports_up == 0) { + printf("."); + fflush(stdout); + rte_delay_ms(CHECK_INTERVAL); + } + + /* set the print_flag if all ports up or timeout */ + if (all_ports_up == 1 || count == (MAX_CHECK_TIME - 1)) { + print_flag = 1; + printf("done\n"); + } + } +} + +/* Check if device has to be HW/SW or any */ +static int +check_type(struct l2fwd_crypto_options *options, struct rte_cryptodev_info *dev_info) +{ + if (options->type == CDEV_TYPE_HW && + (dev_info->feature_flags & RTE_CRYPTODEV_FF_HW_ACCELERATED)) + return 0; + if (options->type == CDEV_TYPE_SW && + !(dev_info->feature_flags & RTE_CRYPTODEV_FF_HW_ACCELERATED)) + return 0; + if (options->type == CDEV_TYPE_ANY) + return 0; + + return -1; +} + +/* Check if the device is enabled by cryptodev_mask */ +static int +check_cryptodev_mask(struct l2fwd_crypto_options *options, + uint8_t cdev_id) +{ + if (options->cryptodev_mask & (1 << cdev_id)) + return 0; + + return -1; +} + +static inline int +check_supported_size(uint16_t length, uint16_t min, uint16_t max, + uint16_t increment) +{ + uint16_t supp_size; + + /* Single value */ + if (increment == 0) { + if (length == min) + return 0; + else + return -1; + } + + /* Range of values */ + for (supp_size = min; supp_size <= max; supp_size += increment) { + if (length == supp_size) + return 0; + } + + return -1; +} +static int +initialize_cryptodevs(struct l2fwd_crypto_options *options, unsigned nb_ports, + uint8_t *enabled_cdevs) +{ + unsigned i, cdev_id, cdev_count, enabled_cdev_count = 0; + const struct rte_cryptodev_capabilities *cap; + enum rte_crypto_auth_algorithm cap_auth_algo; + enum rte_crypto_auth_algorithm opt_auth_algo; + enum rte_crypto_cipher_algorithm cap_cipher_algo; + enum rte_crypto_cipher_algorithm opt_cipher_algo; + int retval; + + cdev_count = rte_cryptodev_count(); + if (cdev_count == 0) { + printf("No crypto devices available\n"); + return -1; + } + + for (cdev_id = 0; cdev_id < cdev_count && enabled_cdev_count < nb_ports; + cdev_id++) { + struct rte_cryptodev_qp_conf qp_conf; + struct rte_cryptodev_info dev_info; + + struct rte_cryptodev_config conf = { + .nb_queue_pairs = 1, + .socket_id = SOCKET_ID_ANY, + .session_mp = { + .nb_objs = 2048, + .cache_size = 64 + } + }; + + if (check_cryptodev_mask(options, (uint8_t)cdev_id)) + continue; + + rte_cryptodev_info_get(cdev_id, &dev_info); + + /* Set cipher parameters */ + if (options->xform_chain == L2FWD_CRYPTO_CIPHER_HASH || + options->xform_chain == L2FWD_CRYPTO_HASH_CIPHER || + options->xform_chain == L2FWD_CRYPTO_CIPHER_ONLY) { + /* Check if device supports cipher algo */ + i = 0; + opt_cipher_algo = options->cipher_xform.cipher.algo; + cap = &dev_info.capabilities[i]; + while (cap->op != RTE_CRYPTO_OP_TYPE_UNDEFINED) { + cap_cipher_algo = cap->sym.cipher.algo; + if (cap->sym.xform_type == + RTE_CRYPTO_SYM_XFORM_CIPHER) { + if (cap_cipher_algo == opt_cipher_algo) { + if (check_type(options, &dev_info) == 0) + break; + } + } + cap = &dev_info.capabilities[++i]; + } + + if (cap->op == RTE_CRYPTO_OP_TYPE_UNDEFINED) { + printf("Algorithm %s not supported by cryptodev %u" + " or device not of preferred type (%s)\n", + rte_crypto_cipher_algorithm_strings[opt_cipher_algo], + cdev_id, + options->string_type); + continue; + } + + options->block_size = cap->sym.cipher.block_size; + /* + * Check if length of provided IV is supported + * by the algorithm chosen. + */ + if (options->iv_param) { + if (check_supported_size(options->iv.length, + cap->sym.cipher.iv_size.min, + cap->sym.cipher.iv_size.max, + cap->sym.cipher.iv_size.increment) + != 0) { + printf("Unsupported IV length\n"); + return -1; + } + /* + * Check if length of IV to be randomly generated + * is supported by the algorithm chosen. + */ + } else if (options->iv_random_size != -1) { + if (check_supported_size(options->iv_random_size, + cap->sym.cipher.iv_size.min, + cap->sym.cipher.iv_size.max, + cap->sym.cipher.iv_size.increment) + != 0) { + printf("Unsupported IV length\n"); + return -1; + } + options->iv.length = options->iv_random_size; + /* No size provided, use minimum size. */ + } else + options->iv.length = cap->sym.cipher.iv_size.min; + + /* + * Check if length of provided cipher key is supported + * by the algorithm chosen. + */ + if (options->ckey_param) { + if (check_supported_size( + options->cipher_xform.cipher.key.length, + cap->sym.cipher.key_size.min, + cap->sym.cipher.key_size.max, + cap->sym.cipher.key_size.increment) + != 0) { + printf("Unsupported cipher key length\n"); + return -1; + } + /* + * Check if length of the cipher key to be randomly generated + * is supported by the algorithm chosen. + */ + } else if (options->ckey_random_size != -1) { + if (check_supported_size(options->ckey_random_size, + cap->sym.cipher.key_size.min, + cap->sym.cipher.key_size.max, + cap->sym.cipher.key_size.increment) + != 0) { + printf("Unsupported cipher key length\n"); + return -1; + } + options->cipher_xform.cipher.key.length = + options->ckey_random_size; + /* No size provided, use minimum size. */ + } else + options->cipher_xform.cipher.key.length = + cap->sym.cipher.key_size.min; + + if (!options->ckey_param) + generate_random_key( + options->cipher_xform.cipher.key.data, + options->cipher_xform.cipher.key.length); + + } + + /* Set auth parameters */ + if (options->xform_chain == L2FWD_CRYPTO_CIPHER_HASH || + options->xform_chain == L2FWD_CRYPTO_HASH_CIPHER || + options->xform_chain == L2FWD_CRYPTO_HASH_ONLY) { + /* Check if device supports auth algo */ + i = 0; + opt_auth_algo = options->auth_xform.auth.algo; + cap = &dev_info.capabilities[i]; + while (cap->op != RTE_CRYPTO_OP_TYPE_UNDEFINED) { + cap_auth_algo = cap->sym.auth.algo; + if ((cap->sym.xform_type == RTE_CRYPTO_SYM_XFORM_AUTH) && + (cap_auth_algo == opt_auth_algo) && + (check_type(options, &dev_info) == 0)) { + break; + } + cap = &dev_info.capabilities[++i]; + } + + if (cap->op == RTE_CRYPTO_OP_TYPE_UNDEFINED) { + printf("Algorithm %s not supported by cryptodev %u" + " or device not of preferred type (%s)\n", + rte_crypto_auth_algorithm_strings[opt_auth_algo], + cdev_id, + options->string_type); + continue; + } + + /* + * Check if length of provided AAD is supported + * by the algorithm chosen. + */ + if (options->aad_param) { + if (check_supported_size(options->aad.length, + cap->sym.auth.aad_size.min, + cap->sym.auth.aad_size.max, + cap->sym.auth.aad_size.increment) + != 0) { + printf("Unsupported AAD length\n"); + return -1; + } + /* + * Check if length of AAD to be randomly generated + * is supported by the algorithm chosen. + */ + } else if (options->aad_random_size != -1) { + if (check_supported_size(options->aad_random_size, + cap->sym.auth.aad_size.min, + cap->sym.auth.aad_size.max, + cap->sym.auth.aad_size.increment) + != 0) { + printf("Unsupported AAD length\n"); + return -1; + } + options->aad.length = options->aad_random_size; + /* No size provided, use minimum size. */ + } else + options->aad.length = cap->sym.auth.aad_size.min; + + options->auth_xform.auth.add_auth_data_length = + options->aad.length; + + /* + * Check if length of provided auth key is supported + * by the algorithm chosen. + */ + if (options->akey_param) { + if (check_supported_size( + options->auth_xform.auth.key.length, + cap->sym.auth.key_size.min, + cap->sym.auth.key_size.max, + cap->sym.auth.key_size.increment) + != 0) { + printf("Unsupported auth key length\n"); + return -1; + } + /* + * Check if length of the auth key to be randomly generated + * is supported by the algorithm chosen. + */ + } else if (options->akey_random_size != -1) { + if (check_supported_size(options->akey_random_size, + cap->sym.auth.key_size.min, + cap->sym.auth.key_size.max, + cap->sym.auth.key_size.increment) + != 0) { + printf("Unsupported auth key length\n"); + return -1; + } + options->auth_xform.auth.key.length = + options->akey_random_size; + /* No size provided, use minimum size. */ + } else + options->auth_xform.auth.key.length = + cap->sym.auth.key_size.min; + + if (!options->akey_param) + generate_random_key( + options->auth_xform.auth.key.data, + options->auth_xform.auth.key.length); + + /* Check if digest size is supported by the algorithm. */ + if (options->digest_size != -1) { + if (check_supported_size(options->digest_size, + cap->sym.auth.digest_size.min, + cap->sym.auth.digest_size.max, + cap->sym.auth.digest_size.increment) + != 0) { + printf("Unsupported digest length\n"); + return -1; + } + options->auth_xform.auth.digest_length = + options->digest_size; + /* No size provided, use minimum size. */ + } else + options->auth_xform.auth.digest_length = + cap->sym.auth.digest_size.min; + } + + retval = rte_cryptodev_configure(cdev_id, &conf); + if (retval < 0) { + printf("Failed to configure cryptodev %u", cdev_id); + return -1; + } + + qp_conf.nb_descriptors = 2048; + + retval = rte_cryptodev_queue_pair_setup(cdev_id, 0, &qp_conf, + SOCKET_ID_ANY); + if (retval < 0) { + printf("Failed to setup queue pair %u on cryptodev %u", + 0, cdev_id); + return -1; + } + + retval = rte_cryptodev_start(cdev_id); + if (retval < 0) { + printf("Failed to start device %u: error %d\n", + cdev_id, retval); + return -1; + } + + l2fwd_enabled_crypto_mask |= (((uint64_t)1) << cdev_id); + + enabled_cdevs[cdev_id] = 1; + enabled_cdev_count++; + } + + return enabled_cdev_count; +} + +static int +initialize_ports(struct l2fwd_crypto_options *options) +{ + uint8_t last_portid, portid; + unsigned enabled_portcount = 0; + unsigned nb_ports = rte_eth_dev_count(); + + if (nb_ports == 0) { + printf("No Ethernet ports - bye\n"); + return -1; + } + + /* Reset l2fwd_dst_ports */ + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) + l2fwd_dst_ports[portid] = 0; + + for (last_portid = 0, portid = 0; portid < nb_ports; portid++) { + int retval; + + /* Skip ports that are not enabled */ + if ((options->portmask & (1 << portid)) == 0) + continue; + + /* init port */ + printf("Initializing port %u... ", (unsigned) portid); + fflush(stdout); + retval = rte_eth_dev_configure(portid, 1, 1, &port_conf); + if (retval < 0) { + printf("Cannot configure device: err=%d, port=%u\n", + retval, (unsigned) portid); + return -1; + } + + /* init one RX queue */ + fflush(stdout); + retval = rte_eth_rx_queue_setup(portid, 0, nb_rxd, + rte_eth_dev_socket_id(portid), + NULL, l2fwd_pktmbuf_pool); + if (retval < 0) { + printf("rte_eth_rx_queue_setup:err=%d, port=%u\n", + retval, (unsigned) portid); + return -1; + } + + /* init one TX queue on each port */ + fflush(stdout); + retval = rte_eth_tx_queue_setup(portid, 0, nb_txd, + rte_eth_dev_socket_id(portid), + NULL); + if (retval < 0) { + printf("rte_eth_tx_queue_setup:err=%d, port=%u\n", + retval, (unsigned) portid); + + return -1; + } + + /* Start device */ + retval = rte_eth_dev_start(portid); + if (retval < 0) { + printf("rte_eth_dev_start:err=%d, port=%u\n", + retval, (unsigned) portid); + return -1; + } + + rte_eth_promiscuous_enable(portid); + + rte_eth_macaddr_get(portid, &l2fwd_ports_eth_addr[portid]); + + printf("Port %u, MAC address: %02X:%02X:%02X:%02X:%02X:%02X\n\n", + (unsigned) portid, + l2fwd_ports_eth_addr[portid].addr_bytes[0], + l2fwd_ports_eth_addr[portid].addr_bytes[1], + l2fwd_ports_eth_addr[portid].addr_bytes[2], + l2fwd_ports_eth_addr[portid].addr_bytes[3], + l2fwd_ports_eth_addr[portid].addr_bytes[4], + l2fwd_ports_eth_addr[portid].addr_bytes[5]); + + /* initialize port stats */ + memset(&port_statistics, 0, sizeof(port_statistics)); + + /* Setup port forwarding table */ + if (enabled_portcount % 2) { + l2fwd_dst_ports[portid] = last_portid; + l2fwd_dst_ports[last_portid] = portid; + } else { + last_portid = portid; + } + + l2fwd_enabled_port_mask |= (1 << portid); + enabled_portcount++; + } + + if (enabled_portcount == 1) { + l2fwd_dst_ports[last_portid] = last_portid; + } else if (enabled_portcount % 2) { + printf("odd number of ports in portmask- bye\n"); + return -1; + } + + check_all_ports_link_status(nb_ports, l2fwd_enabled_port_mask); + + return enabled_portcount; +} + +static void +reserve_key_memory(struct l2fwd_crypto_options *options) +{ + options->cipher_xform.cipher.key.data = rte_malloc("crypto key", + MAX_KEY_SIZE, 0); + if (options->cipher_xform.cipher.key.data == NULL) + rte_exit(EXIT_FAILURE, "Failed to allocate memory for cipher key"); + + + options->auth_xform.auth.key.data = rte_malloc("auth key", + MAX_KEY_SIZE, 0); + if (options->auth_xform.auth.key.data == NULL) + rte_exit(EXIT_FAILURE, "Failed to allocate memory for auth key"); + + options->iv.data = rte_malloc("iv", MAX_KEY_SIZE, 0); + if (options->iv.data == NULL) + rte_exit(EXIT_FAILURE, "Failed to allocate memory for IV"); + options->iv.phys_addr = rte_malloc_virt2phy(options->iv.data); + + options->aad.data = rte_malloc("aad", MAX_KEY_SIZE, 0); + if (options->aad.data == NULL) + rte_exit(EXIT_FAILURE, "Failed to allocate memory for AAD"); + options->aad.phys_addr = rte_malloc_virt2phy(options->aad.data); +} + +int +main(int argc, char **argv) +{ + struct lcore_queue_conf *qconf; + struct l2fwd_crypto_options options; + + uint8_t nb_ports, nb_cryptodevs, portid, cdev_id; + unsigned lcore_id, rx_lcore_id; + int ret, enabled_cdevcount, enabled_portcount; + uint8_t enabled_cdevs[RTE_CRYPTO_MAX_DEVS] = {0}; + + /* init EAL */ + ret = rte_eal_init(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n"); + argc -= ret; + argv += ret; + + /* reserve memory for Cipher/Auth key and IV */ + reserve_key_memory(&options); + + /* parse application arguments (after the EAL ones) */ + ret = l2fwd_crypto_parse_args(&options, argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid L2FWD-CRYPTO arguments\n"); + + /* create the mbuf pool */ + l2fwd_pktmbuf_pool = rte_pktmbuf_pool_create("mbuf_pool", NB_MBUF, 512, + sizeof(struct rte_crypto_op), + RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id()); + if (l2fwd_pktmbuf_pool == NULL) + rte_exit(EXIT_FAILURE, "Cannot create mbuf pool\n"); + + /* create crypto op pool */ + l2fwd_crypto_op_pool = rte_crypto_op_pool_create("crypto_op_pool", + RTE_CRYPTO_OP_TYPE_SYMMETRIC, NB_MBUF, 128, 0, + rte_socket_id()); + if (l2fwd_crypto_op_pool == NULL) + rte_exit(EXIT_FAILURE, "Cannot create crypto op pool\n"); + + /* Enable Ethernet ports */ + enabled_portcount = initialize_ports(&options); + if (enabled_portcount < 1) + rte_exit(EXIT_FAILURE, "Failed to initial Ethernet ports\n"); + + nb_ports = rte_eth_dev_count(); + /* Initialize the port/queue configuration of each logical core */ + for (rx_lcore_id = 0, qconf = NULL, portid = 0; + portid < nb_ports; portid++) { + + /* skip ports that are not enabled */ + if ((options.portmask & (1 << portid)) == 0) + continue; + + if (options.single_lcore && qconf == NULL) { + while (rte_lcore_is_enabled(rx_lcore_id) == 0) { + rx_lcore_id++; + if (rx_lcore_id >= RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE, + "Not enough cores\n"); + } + } else if (!options.single_lcore) { + /* get the lcore_id for this port */ + while (rte_lcore_is_enabled(rx_lcore_id) == 0 || + lcore_queue_conf[rx_lcore_id].nb_rx_ports == + options.nb_ports_per_lcore) { + rx_lcore_id++; + if (rx_lcore_id >= RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE, + "Not enough cores\n"); + } + } + + /* Assigned a new logical core in the loop above. */ + if (qconf != &lcore_queue_conf[rx_lcore_id]) + qconf = &lcore_queue_conf[rx_lcore_id]; + + qconf->rx_port_list[qconf->nb_rx_ports] = portid; + qconf->nb_rx_ports++; + + printf("Lcore %u: RX port %u\n", rx_lcore_id, (unsigned)portid); + } + + /* Enable Crypto devices */ + enabled_cdevcount = initialize_cryptodevs(&options, enabled_portcount, + enabled_cdevs); + if (enabled_cdevcount < 0) + rte_exit(EXIT_FAILURE, "Failed to initialize crypto devices\n"); + + if (enabled_cdevcount < enabled_portcount) + rte_exit(EXIT_FAILURE, "Number of capable crypto devices (%d) " + "has to be more or equal to number of ports (%d)\n", + enabled_cdevcount, enabled_portcount); + + nb_cryptodevs = rte_cryptodev_count(); + + /* Initialize the port/cryptodev configuration of each logical core */ + for (rx_lcore_id = 0, qconf = NULL, cdev_id = 0; + cdev_id < nb_cryptodevs && enabled_cdevcount; + cdev_id++) { + /* Crypto op not supported by crypto device */ + if (!enabled_cdevs[cdev_id]) + continue; + + if (options.single_lcore && qconf == NULL) { + while (rte_lcore_is_enabled(rx_lcore_id) == 0) { + rx_lcore_id++; + if (rx_lcore_id >= RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE, + "Not enough cores\n"); + } + } else if (!options.single_lcore) { + /* get the lcore_id for this port */ + while (rte_lcore_is_enabled(rx_lcore_id) == 0 || + lcore_queue_conf[rx_lcore_id].nb_crypto_devs == + options.nb_ports_per_lcore) { + rx_lcore_id++; + if (rx_lcore_id >= RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE, + "Not enough cores\n"); + } + } + + /* Assigned a new logical core in the loop above. */ + if (qconf != &lcore_queue_conf[rx_lcore_id]) + qconf = &lcore_queue_conf[rx_lcore_id]; + + qconf->cryptodev_list[qconf->nb_crypto_devs] = cdev_id; + qconf->nb_crypto_devs++; + + enabled_cdevcount--; + + printf("Lcore %u: cryptodev %u\n", rx_lcore_id, + (unsigned)cdev_id); + } + + /* launch per-lcore init on every lcore */ + rte_eal_mp_remote_launch(l2fwd_launch_one_lcore, (void *)&options, + CALL_MASTER); + RTE_LCORE_FOREACH_SLAVE(lcore_id) { + if (rte_eal_wait_lcore(lcore_id) < 0) + return -1; + } + + return 0; +} diff --git a/src/seastar/dpdk/examples/l2fwd-jobstats/Makefile b/src/seastar/dpdk/examples/l2fwd-jobstats/Makefile new file mode 100644 index 00000000..ab089f66 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-jobstats/Makefile @@ -0,0 +1,51 @@ +# BSD LICENSE +# +# Copyright(c) 2010-2015 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 copyright +# 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 overridden by command line or environment +RTE_TARGET ?= x86_64-native-linuxapp-gcc + +include $(RTE_SDK)/mk/rte.vars.mk + +# binary name +APP = l2fwd-jobstats + +# all source are stored in SRCS-y +SRCS-y := main.c + + +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) + +include $(RTE_SDK)/mk/rte.extapp.mk diff --git a/src/seastar/dpdk/examples/l2fwd-jobstats/main.c b/src/seastar/dpdk/examples/l2fwd-jobstats/main.c new file mode 100644 index 00000000..e6e6c228 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-jobstats/main.c @@ -0,0 +1,1022 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2010-2016 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 copyright + * 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 <locale.h> +#include <stdlib.h> +#include <string.h> +#include <stdint.h> +#include <ctype.h> +#include <getopt.h> + +#include <rte_alarm.h> +#include <rte_common.h> +#include <rte_log.h> +#include <rte_malloc.h> +#include <rte_memory.h> +#include <rte_memcpy.h> +#include <rte_memzone.h> +#include <rte_eal.h> +#include <rte_per_lcore.h> +#include <rte_launch.h> +#include <rte_atomic.h> +#include <rte_cycles.h> +#include <rte_prefetch.h> +#include <rte_lcore.h> +#include <rte_per_lcore.h> +#include <rte_branch_prediction.h> +#include <rte_interrupts.h> +#include <rte_pci.h> +#include <rte_debug.h> +#include <rte_ether.h> +#include <rte_ethdev.h> +#include <rte_mempool.h> +#include <rte_mbuf.h> +#include <rte_spinlock.h> + +#include <rte_errno.h> +#include <rte_jobstats.h> +#include <rte_timer.h> +#include <rte_alarm.h> + +#define RTE_LOGTYPE_L2FWD RTE_LOGTYPE_USER1 + +#define NB_MBUF 8192 + +#define MAX_PKT_BURST 32 +#define BURST_TX_DRAIN_US 100 /* TX drain every ~100us */ + +/* + * Configurable number of RX/TX ring descriptors + */ +#define RTE_TEST_RX_DESC_DEFAULT 128 +#define RTE_TEST_TX_DESC_DEFAULT 512 +static uint16_t nb_rxd = RTE_TEST_RX_DESC_DEFAULT; +static uint16_t nb_txd = RTE_TEST_TX_DESC_DEFAULT; + +/* ethernet addresses of ports */ +static struct ether_addr l2fwd_ports_eth_addr[RTE_MAX_ETHPORTS]; + +/* mask of enabled ports */ +static uint32_t l2fwd_enabled_port_mask; + +/* list of enabled ports */ +static uint32_t l2fwd_dst_ports[RTE_MAX_ETHPORTS]; + +#define UPDATE_STEP_UP 1 +#define UPDATE_STEP_DOWN 32 + +static unsigned int l2fwd_rx_queue_per_lcore = 1; + +#define MAX_RX_QUEUE_PER_LCORE 16 +#define MAX_TX_QUEUE_PER_PORT 16 +struct lcore_queue_conf { + unsigned n_rx_port; + unsigned rx_port_list[MAX_RX_QUEUE_PER_LCORE]; + uint64_t next_flush_time[RTE_MAX_ETHPORTS]; + + struct rte_timer rx_timers[MAX_RX_QUEUE_PER_LCORE]; + struct rte_jobstats port_fwd_jobs[MAX_RX_QUEUE_PER_LCORE]; + + struct rte_timer flush_timer; + struct rte_jobstats flush_job; + struct rte_jobstats idle_job; + struct rte_jobstats_context jobs_context; + + rte_atomic16_t stats_read_pending; + rte_spinlock_t lock; +} __rte_cache_aligned; +struct lcore_queue_conf lcore_queue_conf[RTE_MAX_LCORE]; + +struct rte_eth_dev_tx_buffer *tx_buffer[RTE_MAX_ETHPORTS]; + +static const struct rte_eth_conf port_conf = { + .rxmode = { + .split_hdr_size = 0, + .header_split = 0, /**< Header Split disabled */ + .hw_ip_checksum = 0, /**< IP checksum offload disabled */ + .hw_vlan_filter = 0, /**< VLAN filtering disabled */ + .jumbo_frame = 0, /**< Jumbo Frame Support disabled */ + .hw_strip_crc = 1, /**< CRC stripped by hardware */ + }, + .txmode = { + .mq_mode = ETH_MQ_TX_NONE, + }, +}; + +struct rte_mempool *l2fwd_pktmbuf_pool = NULL; + +/* Per-port statistics struct */ +struct l2fwd_port_statistics { + uint64_t tx; + uint64_t rx; + uint64_t dropped; +} __rte_cache_aligned; +struct l2fwd_port_statistics port_statistics[RTE_MAX_ETHPORTS]; + +/* 1 day max */ +#define MAX_TIMER_PERIOD 86400 +/* default period is 10 seconds */ +static int64_t timer_period = 10; +/* default timer frequency */ +static double hz; +/* BURST_TX_DRAIN_US converted to cycles */ +uint64_t drain_tsc; +/* Convert cycles to ns */ +static inline double +cycles_to_ns(uint64_t cycles) +{ + double t = cycles; + + t *= (double)NS_PER_S; + t /= hz; + return t; +} + +static void +show_lcore_stats(unsigned lcore_id) +{ + struct lcore_queue_conf *qconf = &lcore_queue_conf[lcore_id]; + struct rte_jobstats_context *ctx = &qconf->jobs_context; + struct rte_jobstats *job; + uint8_t i; + + /* LCore statistics. */ + uint64_t stats_period, loop_count; + uint64_t exec, exec_min, exec_max; + uint64_t management, management_min, management_max; + uint64_t busy, busy_min, busy_max; + + /* Jobs statistics. */ + const uint8_t port_cnt = qconf->n_rx_port; + uint64_t jobs_exec_cnt[port_cnt], jobs_period[port_cnt]; + uint64_t jobs_exec[port_cnt], jobs_exec_min[port_cnt], + jobs_exec_max[port_cnt]; + + uint64_t flush_exec_cnt, flush_period; + uint64_t flush_exec, flush_exec_min, flush_exec_max; + + uint64_t idle_exec_cnt; + uint64_t idle_exec, idle_exec_min, idle_exec_max; + uint64_t collection_time = rte_get_timer_cycles(); + + /* Ask forwarding thread to give us stats. */ + rte_atomic16_set(&qconf->stats_read_pending, 1); + rte_spinlock_lock(&qconf->lock); + rte_atomic16_set(&qconf->stats_read_pending, 0); + + /* Collect context statistics. */ + stats_period = ctx->state_time - ctx->start_time; + loop_count = ctx->loop_cnt; + + exec = ctx->exec_time; + exec_min = ctx->min_exec_time; + exec_max = ctx->max_exec_time; + + management = ctx->management_time; + management_min = ctx->min_management_time; + management_max = ctx->max_management_time; + + rte_jobstats_context_reset(ctx); + + for (i = 0; i < port_cnt; i++) { + job = &qconf->port_fwd_jobs[i]; + + jobs_exec_cnt[i] = job->exec_cnt; + jobs_period[i] = job->period; + + jobs_exec[i] = job->exec_time; + jobs_exec_min[i] = job->min_exec_time; + jobs_exec_max[i] = job->max_exec_time; + + rte_jobstats_reset(job); + } + + flush_exec_cnt = qconf->flush_job.exec_cnt; + flush_period = qconf->flush_job.period; + flush_exec = qconf->flush_job.exec_time; + flush_exec_min = qconf->flush_job.min_exec_time; + flush_exec_max = qconf->flush_job.max_exec_time; + rte_jobstats_reset(&qconf->flush_job); + + idle_exec_cnt = qconf->idle_job.exec_cnt; + idle_exec = qconf->idle_job.exec_time; + idle_exec_min = qconf->idle_job.min_exec_time; + idle_exec_max = qconf->idle_job.max_exec_time; + rte_jobstats_reset(&qconf->idle_job); + + rte_spinlock_unlock(&qconf->lock); + + exec -= idle_exec; + busy = exec + management; + busy_min = exec_min + management_min; + busy_max = exec_max + management_max; + + + collection_time = rte_get_timer_cycles() - collection_time; + +#define STAT_FMT "\n%-18s %'14.0f %6.1f%% %'10.0f %'10.0f %'10.0f" + + printf("\n----------------" + "\nLCore %3u: statistics (time in ns, collected in %'9.0f)" + "\n%-18s %14s %7s %10s %10s %10s " + "\n%-18s %'14.0f" + "\n%-18s %'14" PRIu64 + STAT_FMT /* Exec */ + STAT_FMT /* Management */ + STAT_FMT /* Busy */ + STAT_FMT, /* Idle */ + lcore_id, cycles_to_ns(collection_time), + "Stat type", "total", "%total", "avg", "min", "max", + "Stats duration:", cycles_to_ns(stats_period), + "Loop count:", loop_count, + "Exec time", + cycles_to_ns(exec), exec * 100.0 / stats_period, + cycles_to_ns(loop_count ? exec / loop_count : 0), + cycles_to_ns(exec_min), + cycles_to_ns(exec_max), + "Management time", + cycles_to_ns(management), management * 100.0 / stats_period, + cycles_to_ns(loop_count ? management / loop_count : 0), + cycles_to_ns(management_min), + cycles_to_ns(management_max), + "Exec + management", + cycles_to_ns(busy), busy * 100.0 / stats_period, + cycles_to_ns(loop_count ? busy / loop_count : 0), + cycles_to_ns(busy_min), + cycles_to_ns(busy_max), + "Idle (job)", + cycles_to_ns(idle_exec), idle_exec * 100.0 / stats_period, + cycles_to_ns(idle_exec_cnt ? idle_exec / idle_exec_cnt : 0), + cycles_to_ns(idle_exec_min), + cycles_to_ns(idle_exec_max)); + + for (i = 0; i < qconf->n_rx_port; i++) { + job = &qconf->port_fwd_jobs[i]; + printf("\n\nJob %" PRIu32 ": %-20s " + "\n%-18s %'14" PRIu64 + "\n%-18s %'14.0f" + STAT_FMT, + i, job->name, + "Exec count:", jobs_exec_cnt[i], + "Exec period: ", cycles_to_ns(jobs_period[i]), + "Exec time", + cycles_to_ns(jobs_exec[i]), jobs_exec[i] * 100.0 / stats_period, + cycles_to_ns(jobs_exec_cnt[i] ? jobs_exec[i] / jobs_exec_cnt[i] + : 0), + cycles_to_ns(jobs_exec_min[i]), + cycles_to_ns(jobs_exec_max[i])); + } + + if (qconf->n_rx_port > 0) { + job = &qconf->flush_job; + printf("\n\nJob %" PRIu32 ": %-20s " + "\n%-18s %'14" PRIu64 + "\n%-18s %'14.0f" + STAT_FMT, + i, job->name, + "Exec count:", flush_exec_cnt, + "Exec period: ", cycles_to_ns(flush_period), + "Exec time", + cycles_to_ns(flush_exec), flush_exec * 100.0 / stats_period, + cycles_to_ns(flush_exec_cnt ? flush_exec / flush_exec_cnt : 0), + cycles_to_ns(flush_exec_min), + cycles_to_ns(flush_exec_max)); + } +} + +/* Print out statistics on packets dropped */ +static void +show_stats_cb(__rte_unused void *param) +{ + uint64_t total_packets_dropped, total_packets_tx, total_packets_rx; + unsigned portid, lcore_id; + + total_packets_dropped = 0; + total_packets_tx = 0; + total_packets_rx = 0; + + const char clr[] = { 27, '[', '2', 'J', '\0' }; + const char topLeft[] = { 27, '[', '1', ';', '1', 'H', '\0' }; + + /* Clear screen and move to top left */ + printf("%s%s" + "\nPort statistics ===================================", + clr, topLeft); + + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) { + /* skip disabled ports */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + printf("\nStatistics for port %u ------------------------------" + "\nPackets sent: %24"PRIu64 + "\nPackets received: %20"PRIu64 + "\nPackets dropped: %21"PRIu64, + portid, + port_statistics[portid].tx, + port_statistics[portid].rx, + port_statistics[portid].dropped); + + total_packets_dropped += port_statistics[portid].dropped; + total_packets_tx += port_statistics[portid].tx; + total_packets_rx += port_statistics[portid].rx; + } + + printf("\nAggregate statistics ===============================" + "\nTotal packets sent: %18"PRIu64 + "\nTotal packets received: %14"PRIu64 + "\nTotal packets dropped: %15"PRIu64 + "\n====================================================", + total_packets_tx, + total_packets_rx, + total_packets_dropped); + + RTE_LCORE_FOREACH(lcore_id) { + if (lcore_queue_conf[lcore_id].n_rx_port > 0) + show_lcore_stats(lcore_id); + } + + printf("\n====================================================\n"); + rte_eal_alarm_set(timer_period * US_PER_S, show_stats_cb, NULL); +} + +static void +l2fwd_simple_forward(struct rte_mbuf *m, unsigned portid) +{ + struct ether_hdr *eth; + void *tmp; + int sent; + unsigned dst_port; + struct rte_eth_dev_tx_buffer *buffer; + + dst_port = l2fwd_dst_ports[portid]; + eth = rte_pktmbuf_mtod(m, struct ether_hdr *); + + /* 02:00:00:00:00:xx */ + tmp = ð->d_addr.addr_bytes[0]; + *((uint64_t *)tmp) = 0x000000000002 + ((uint64_t)dst_port << 40); + + /* src addr */ + ether_addr_copy(&l2fwd_ports_eth_addr[dst_port], ð->s_addr); + + buffer = tx_buffer[dst_port]; + sent = rte_eth_tx_buffer(dst_port, 0, buffer, m); + if (sent) + port_statistics[dst_port].tx += sent; +} + +static void +l2fwd_job_update_cb(struct rte_jobstats *job, int64_t result) +{ + int64_t err = job->target - result; + int64_t histeresis = job->target / 8; + + if (err < -histeresis) { + if (job->min_period + UPDATE_STEP_DOWN < job->period) + job->period -= UPDATE_STEP_DOWN; + } else if (err > histeresis) { + if (job->period + UPDATE_STEP_UP < job->max_period) + job->period += UPDATE_STEP_UP; + } +} + +static void +l2fwd_fwd_job(__rte_unused struct rte_timer *timer, void *arg) +{ + struct rte_mbuf *pkts_burst[MAX_PKT_BURST]; + struct rte_mbuf *m; + + const uint8_t port_idx = (uintptr_t) arg; + const unsigned lcore_id = rte_lcore_id(); + struct lcore_queue_conf *qconf = &lcore_queue_conf[lcore_id]; + struct rte_jobstats *job = &qconf->port_fwd_jobs[port_idx]; + const uint8_t portid = qconf->rx_port_list[port_idx]; + + uint8_t j; + uint16_t total_nb_rx; + + rte_jobstats_start(&qconf->jobs_context, job); + + /* Call rx burst 2 times. This allow rte_jobstats logic to see if this + * function must be called more frequently. */ + + total_nb_rx = rte_eth_rx_burst((uint8_t) portid, 0, pkts_burst, + MAX_PKT_BURST); + + for (j = 0; j < total_nb_rx; j++) { + m = pkts_burst[j]; + rte_prefetch0(rte_pktmbuf_mtod(m, void *)); + l2fwd_simple_forward(m, portid); + } + + if (total_nb_rx == MAX_PKT_BURST) { + const uint16_t nb_rx = rte_eth_rx_burst((uint8_t) portid, 0, pkts_burst, + MAX_PKT_BURST); + + total_nb_rx += nb_rx; + for (j = 0; j < nb_rx; j++) { + m = pkts_burst[j]; + rte_prefetch0(rte_pktmbuf_mtod(m, void *)); + l2fwd_simple_forward(m, portid); + } + } + + port_statistics[portid].rx += total_nb_rx; + + /* Adjust period time in which we are running here. */ + if (rte_jobstats_finish(job, total_nb_rx) != 0) { + rte_timer_reset(&qconf->rx_timers[port_idx], job->period, PERIODICAL, + lcore_id, l2fwd_fwd_job, arg); + } +} + +static void +l2fwd_flush_job(__rte_unused struct rte_timer *timer, __rte_unused void *arg) +{ + uint64_t now; + unsigned lcore_id; + struct lcore_queue_conf *qconf; + uint8_t portid; + unsigned i; + uint32_t sent; + struct rte_eth_dev_tx_buffer *buffer; + + lcore_id = rte_lcore_id(); + qconf = &lcore_queue_conf[lcore_id]; + + rte_jobstats_start(&qconf->jobs_context, &qconf->flush_job); + + now = rte_get_timer_cycles(); + lcore_id = rte_lcore_id(); + qconf = &lcore_queue_conf[lcore_id]; + + for (i = 0; i < qconf->n_rx_port; i++) { + portid = l2fwd_dst_ports[qconf->rx_port_list[i]]; + + if (qconf->next_flush_time[portid] <= now) + continue; + + buffer = tx_buffer[portid]; + sent = rte_eth_tx_buffer_flush(portid, 0, buffer); + if (sent) + port_statistics[portid].tx += sent; + + qconf->next_flush_time[portid] = rte_get_timer_cycles() + drain_tsc; + } + + /* Pass target to indicate that this job is happy of time interwal + * in which it was called. */ + rte_jobstats_finish(&qconf->flush_job, qconf->flush_job.target); +} + +/* main processing loop */ +static void +l2fwd_main_loop(void) +{ + unsigned lcore_id; + unsigned i, portid; + struct lcore_queue_conf *qconf; + uint8_t stats_read_pending = 0; + uint8_t need_manage; + + lcore_id = rte_lcore_id(); + qconf = &lcore_queue_conf[lcore_id]; + + if (qconf->n_rx_port == 0) { + RTE_LOG(INFO, L2FWD, "lcore %u has nothing to do\n", lcore_id); + return; + } + + RTE_LOG(INFO, L2FWD, "entering main loop on lcore %u\n", lcore_id); + + for (i = 0; i < qconf->n_rx_port; i++) { + + portid = qconf->rx_port_list[i]; + RTE_LOG(INFO, L2FWD, " -- lcoreid=%u portid=%u\n", lcore_id, + portid); + } + + rte_jobstats_init(&qconf->idle_job, "idle", 0, 0, 0, 0); + + for (;;) { + rte_spinlock_lock(&qconf->lock); + + do { + rte_jobstats_context_start(&qconf->jobs_context); + + /* Do the Idle job: + * - Read stats_read_pending flag + * - check if some real job need to be executed + */ + rte_jobstats_start(&qconf->jobs_context, &qconf->idle_job); + + uint64_t repeats = 0; + + do { + uint8_t i; + uint64_t now = rte_get_timer_cycles(); + + repeats++; + need_manage = qconf->flush_timer.expire < now; + /* Check if we was esked to give a stats. */ + stats_read_pending = + rte_atomic16_read(&qconf->stats_read_pending); + need_manage |= stats_read_pending; + + for (i = 0; i < qconf->n_rx_port && !need_manage; i++) + need_manage = qconf->rx_timers[i].expire < now; + + } while (!need_manage); + + if (likely(repeats != 1)) + rte_jobstats_finish(&qconf->idle_job, qconf->idle_job.target); + else + rte_jobstats_abort(&qconf->idle_job); + + rte_timer_manage(); + rte_jobstats_context_finish(&qconf->jobs_context); + } while (likely(stats_read_pending == 0)); + + rte_spinlock_unlock(&qconf->lock); + rte_pause(); + } +} + +static int +l2fwd_launch_one_lcore(__attribute__((unused)) void *dummy) +{ + l2fwd_main_loop(); + return 0; +} + +/* display usage */ +static void +l2fwd_usage(const char *prgname) +{ + printf("%s [EAL options] -- -p PORTMASK [-q NQ]\n" + " -p PORTMASK: hexadecimal bitmask of ports to configure\n" + " -q NQ: number of queue (=ports) per lcore (default is 1)\n" + " -T PERIOD: statistics will be refreshed each PERIOD seconds (0 to disable, 10 default, 86400 maximum)\n" + " -l set system default locale instead of default (\"C\" locale) for thousands separator in stats.", + prgname); +} + +static int +l2fwd_parse_portmask(const char *portmask) +{ + char *end = NULL; + unsigned long pm; + + /* parse hexadecimal string */ + pm = strtoul(portmask, &end, 16); + if ((portmask[0] == '\0') || (end == NULL) || (*end != '\0')) + return -1; + + if (pm == 0) + return -1; + + return pm; +} + +static unsigned int +l2fwd_parse_nqueue(const char *q_arg) +{ + char *end = NULL; + unsigned long n; + + /* parse hexadecimal string */ + n = strtoul(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + return 0; + if (n == 0) + return 0; + if (n >= MAX_RX_QUEUE_PER_LCORE) + return 0; + + return n; +} + +static int +l2fwd_parse_timer_period(const char *q_arg) +{ + char *end = NULL; + int n; + + /* parse number string */ + n = strtol(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + return -1; + if (n >= MAX_TIMER_PERIOD) + return -1; + + return n; +} + +/* Parse the argument given in the command line of the application */ +static int +l2fwd_parse_args(int argc, char **argv) +{ + int opt, ret; + char **argvopt; + int option_index; + char *prgname = argv[0]; + static struct option lgopts[] = { + {NULL, 0, 0, 0} + }; + + argvopt = argv; + + while ((opt = getopt_long(argc, argvopt, "p:q:T:l", + lgopts, &option_index)) != EOF) { + + switch (opt) { + /* portmask */ + case 'p': + l2fwd_enabled_port_mask = l2fwd_parse_portmask(optarg); + if (l2fwd_enabled_port_mask == 0) { + printf("invalid portmask\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* nqueue */ + case 'q': + l2fwd_rx_queue_per_lcore = l2fwd_parse_nqueue(optarg); + if (l2fwd_rx_queue_per_lcore == 0) { + printf("invalid queue number\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* timer period */ + case 'T': + timer_period = l2fwd_parse_timer_period(optarg); + if (timer_period < 0) { + printf("invalid timer period\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* For thousands separator in printf. */ + case 'l': + setlocale(LC_ALL, ""); + break; + + /* long options */ + case 0: + l2fwd_usage(prgname); + return -1; + + default: + l2fwd_usage(prgname); + return -1; + } + } + + if (optind >= 0) + argv[optind-1] = prgname; + + ret = optind-1; + optind = 1; /* reset getopt lib */ + return ret; +} + +/* Check the link status of all ports in up to 9s, and print them finally */ +static void +check_all_ports_link_status(uint8_t port_num, uint32_t port_mask) +{ +#define CHECK_INTERVAL 100 /* 100ms */ +#define MAX_CHECK_TIME 90 /* 9s (90 * 100ms) in total */ + uint8_t portid, count, all_ports_up, print_flag = 0; + struct rte_eth_link link; + + printf("\nChecking link status"); + fflush(stdout); + for (count = 0; count <= MAX_CHECK_TIME; count++) { + all_ports_up = 1; + for (portid = 0; portid < port_num; portid++) { + if ((port_mask & (1 << portid)) == 0) + continue; + memset(&link, 0, sizeof(link)); + rte_eth_link_get_nowait(portid, &link); + /* print link status if flag set */ + if (print_flag == 1) { + if (link.link_status) + printf("Port %d Link Up - speed %u " + "Mbps - %s\n", (uint8_t)portid, + (unsigned)link.link_speed, + (link.link_duplex == ETH_LINK_FULL_DUPLEX) ? + ("full-duplex") : ("half-duplex\n")); + else + printf("Port %d Link Down\n", + (uint8_t)portid); + continue; + } + /* clear all_ports_up flag if any link down */ + if (link.link_status == ETH_LINK_DOWN) { + all_ports_up = 0; + break; + } + } + /* after finally printing all link status, get out */ + if (print_flag == 1) + break; + + if (all_ports_up == 0) { + printf("."); + fflush(stdout); + rte_delay_ms(CHECK_INTERVAL); + } + + /* set the print_flag if all ports up or timeout */ + if (all_ports_up == 1 || count == (MAX_CHECK_TIME - 1)) { + print_flag = 1; + printf("done\n"); + } + } +} + +int +main(int argc, char **argv) +{ + struct lcore_queue_conf *qconf; + struct rte_eth_dev_info dev_info; + unsigned lcore_id, rx_lcore_id; + unsigned nb_ports_in_mask = 0; + int ret; + char name[RTE_JOBSTATS_NAMESIZE]; + uint8_t nb_ports; + uint8_t nb_ports_available; + uint8_t portid, last_port; + uint8_t i; + + /* init EAL */ + ret = rte_eal_init(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n"); + argc -= ret; + argv += ret; + + /* parse application arguments (after the EAL ones) */ + ret = l2fwd_parse_args(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid L2FWD arguments\n"); + + rte_timer_subsystem_init(); + + /* fetch default timer frequency. */ + hz = rte_get_timer_hz(); + + /* create the mbuf pool */ + l2fwd_pktmbuf_pool = + rte_pktmbuf_pool_create("mbuf_pool", NB_MBUF, 32, + 0, RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id()); + if (l2fwd_pktmbuf_pool == NULL) + rte_exit(EXIT_FAILURE, "Cannot init mbuf pool\n"); + + nb_ports = rte_eth_dev_count(); + if (nb_ports == 0) + rte_exit(EXIT_FAILURE, "No Ethernet ports - bye\n"); + + /* reset l2fwd_dst_ports */ + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) + l2fwd_dst_ports[portid] = 0; + last_port = 0; + + /* + * Each logical core is assigned a dedicated TX queue on each port. + */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + + if (nb_ports_in_mask % 2) { + l2fwd_dst_ports[portid] = last_port; + l2fwd_dst_ports[last_port] = portid; + } else + last_port = portid; + + nb_ports_in_mask++; + + rte_eth_dev_info_get(portid, &dev_info); + } + if (nb_ports_in_mask % 2) { + printf("Notice: odd number of ports in portmask.\n"); + l2fwd_dst_ports[last_port] = last_port; + } + + rx_lcore_id = 0; + qconf = NULL; + + /* Initialize the port/queue configuration of each logical core */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + + /* get the lcore_id for this port */ + while (rte_lcore_is_enabled(rx_lcore_id) == 0 || + lcore_queue_conf[rx_lcore_id].n_rx_port == + l2fwd_rx_queue_per_lcore) { + rx_lcore_id++; + if (rx_lcore_id >= RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE, "Not enough cores\n"); + } + + if (qconf != &lcore_queue_conf[rx_lcore_id]) + /* Assigned a new logical core in the loop above. */ + qconf = &lcore_queue_conf[rx_lcore_id]; + + qconf->rx_port_list[qconf->n_rx_port] = portid; + qconf->n_rx_port++; + printf("Lcore %u: RX port %u\n", rx_lcore_id, (unsigned) portid); + } + + nb_ports_available = nb_ports; + + /* Initialise each port */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) { + printf("Skipping disabled port %u\n", (unsigned) portid); + nb_ports_available--; + continue; + } + /* init port */ + printf("Initializing port %u... ", (unsigned) portid); + fflush(stdout); + ret = rte_eth_dev_configure(portid, 1, 1, &port_conf); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Cannot configure device: err=%d, port=%u\n", + ret, (unsigned) portid); + + rte_eth_macaddr_get(portid, &l2fwd_ports_eth_addr[portid]); + + /* init one RX queue */ + fflush(stdout); + ret = rte_eth_rx_queue_setup(portid, 0, nb_rxd, + rte_eth_dev_socket_id(portid), + NULL, + l2fwd_pktmbuf_pool); + if (ret < 0) + rte_exit(EXIT_FAILURE, "rte_eth_rx_queue_setup:err=%d, port=%u\n", + ret, (unsigned) portid); + + /* init one TX queue on each port */ + fflush(stdout); + ret = rte_eth_tx_queue_setup(portid, 0, nb_txd, + rte_eth_dev_socket_id(portid), + NULL); + if (ret < 0) + rte_exit(EXIT_FAILURE, "rte_eth_tx_queue_setup:err=%d, port=%u\n", + ret, (unsigned) portid); + + /* Initialize TX buffers */ + tx_buffer[portid] = rte_zmalloc_socket("tx_buffer", + RTE_ETH_TX_BUFFER_SIZE(MAX_PKT_BURST), 0, + rte_eth_dev_socket_id(portid)); + if (tx_buffer[portid] == NULL) + rte_exit(EXIT_FAILURE, "Cannot allocate buffer for tx on port %u\n", + (unsigned) portid); + + rte_eth_tx_buffer_init(tx_buffer[portid], MAX_PKT_BURST); + + ret = rte_eth_tx_buffer_set_err_callback(tx_buffer[portid], + rte_eth_tx_buffer_count_callback, + &port_statistics[portid].dropped); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Cannot set error callback for " + "tx buffer on port %u\n", (unsigned) portid); + + /* Start device */ + ret = rte_eth_dev_start(portid); + if (ret < 0) + rte_exit(EXIT_FAILURE, "rte_eth_dev_start:err=%d, port=%u\n", + ret, (unsigned) portid); + + printf("done:\n"); + + rte_eth_promiscuous_enable(portid); + + printf("Port %u, MAC address: %02X:%02X:%02X:%02X:%02X:%02X\n\n", + (unsigned) portid, + l2fwd_ports_eth_addr[portid].addr_bytes[0], + l2fwd_ports_eth_addr[portid].addr_bytes[1], + l2fwd_ports_eth_addr[portid].addr_bytes[2], + l2fwd_ports_eth_addr[portid].addr_bytes[3], + l2fwd_ports_eth_addr[portid].addr_bytes[4], + l2fwd_ports_eth_addr[portid].addr_bytes[5]); + + /* initialize port stats */ + memset(&port_statistics, 0, sizeof(port_statistics)); + } + + if (!nb_ports_available) { + rte_exit(EXIT_FAILURE, + "All available ports are disabled. Please set portmask.\n"); + } + + check_all_ports_link_status(nb_ports, l2fwd_enabled_port_mask); + + drain_tsc = (hz + US_PER_S - 1) / US_PER_S * BURST_TX_DRAIN_US; + + RTE_LCORE_FOREACH(lcore_id) { + qconf = &lcore_queue_conf[lcore_id]; + + rte_spinlock_init(&qconf->lock); + + if (rte_jobstats_context_init(&qconf->jobs_context) != 0) + rte_panic("Jobs stats context for core %u init failed\n", lcore_id); + + if (qconf->n_rx_port == 0) { + RTE_LOG(INFO, L2FWD, + "lcore %u: no ports so no jobs stats context initialization\n", + lcore_id); + continue; + } + /* Add flush job. + * Set fixed period by setting min = max = initial period. Set target to + * zero as it is irrelevant for this job. */ + rte_jobstats_init(&qconf->flush_job, "flush", drain_tsc, drain_tsc, + drain_tsc, 0); + + rte_timer_init(&qconf->flush_timer); + ret = rte_timer_reset(&qconf->flush_timer, drain_tsc, PERIODICAL, + lcore_id, &l2fwd_flush_job, NULL); + + if (ret < 0) { + rte_exit(1, "Failed to reset flush job timer for lcore %u: %s", + lcore_id, rte_strerror(-ret)); + } + + for (i = 0; i < qconf->n_rx_port; i++) { + struct rte_jobstats *job = &qconf->port_fwd_jobs[i]; + + portid = qconf->rx_port_list[i]; + printf("Setting forward job for port %u\n", portid); + + snprintf(name, RTE_DIM(name), "port %u fwd", portid); + /* Setup forward job. + * Set min, max and initial period. Set target to MAX_PKT_BURST as + * this is desired optimal RX/TX burst size. */ + rte_jobstats_init(job, name, 0, drain_tsc, 0, MAX_PKT_BURST); + rte_jobstats_set_update_period_function(job, l2fwd_job_update_cb); + + rte_timer_init(&qconf->rx_timers[i]); + ret = rte_timer_reset(&qconf->rx_timers[i], 0, PERIODICAL, lcore_id, + &l2fwd_fwd_job, (void *)(uintptr_t)i); + + if (ret < 0) { + rte_exit(1, "Failed to reset lcore %u port %u job timer: %s", + lcore_id, qconf->rx_port_list[i], rte_strerror(-ret)); + } + } + } + + if (timer_period) + rte_eal_alarm_set(timer_period * MS_PER_S, show_stats_cb, NULL); + else + RTE_LOG(INFO, L2FWD, "Stats display disabled\n"); + + /* launch per-lcore init on every lcore */ + rte_eal_mp_remote_launch(l2fwd_launch_one_lcore, NULL, CALL_MASTER); + RTE_LCORE_FOREACH_SLAVE(lcore_id) { + if (rte_eal_wait_lcore(lcore_id) < 0) + return -1; + } + + return 0; +} diff --git a/src/seastar/dpdk/examples/l2fwd-keepalive/Makefile b/src/seastar/dpdk/examples/l2fwd-keepalive/Makefile new file mode 100644 index 00000000..ca45a798 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-keepalive/Makefile @@ -0,0 +1,51 @@ +# BSD LICENSE +# +# Copyright(c) 2010-2016 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 copyright +# 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 overridden by command line or environment +RTE_TARGET ?= x86_64-native-linuxapp-gcc + +include $(RTE_SDK)/mk/rte.vars.mk + +# binary name +APP = l2fwd-keepalive + +# all source are stored in SRCS-y +SRCS-y := main.c shm.c + +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) +LDFLAGS += -lrt + +include $(RTE_SDK)/mk/rte.extapp.mk diff --git a/src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/Makefile b/src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/Makefile new file mode 100644 index 00000000..fd0c38b4 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/Makefile @@ -0,0 +1,49 @@ +# BSD LICENSE +# +# Copyright(c) 2016 Intel Corporation. All rights reserved. +# All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions +# are met: +# +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in +# the documentation and/or other materials provided with the +# distribution. +# * Neither the name of Intel Corporation nor the names of its +# contributors may be used to endorse or promote products derived +# from this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +ifeq ($(RTE_SDK),) +$(error "Please define RTE_SDK environment variable") +endif + +# Default target, can be overridden by command line or environment +RTE_TARGET ?= x86_64-native-linuxapp-gcc + +include $(RTE_SDK)/mk/rte.vars.mk + +# binary name +APP = ka-agent + +# all source are stored in SRCS-y +SRCS-y := main.c + +CFLAGS += $(WERROR_FLAGS) -I$(SRCDIR)/../ +LDFLAGS += -lrt +include $(RTE_SDK)/mk/rte.extapp.mk diff --git a/src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/main.c b/src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/main.c new file mode 100644 index 00000000..be1c7f49 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-keepalive/ka-agent/main.c @@ -0,0 +1,150 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2016 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 copyright + * 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 <errno.h> +#include <stdio.h> +#include <string.h> +#include <stdint.h> +#include <errno.h> +#include <unistd.h> +#include <fcntl.h> +#include <sys/wait.h> +#include <sys/queue.h> +#include <sys/mman.h> +#include <sys/stat.h> +#include <time.h> + +#include <rte_keepalive.h> + +#include <shm.h> + +#define MAX_TIMEOUTS 4 +#define SEM_TIMEOUT_SECS 2 + +static struct rte_keepalive_shm *ka_shm_create(void) +{ + int fd = shm_open(RTE_KEEPALIVE_SHM_NAME, O_RDWR, 0666); + size_t size = sizeof(struct rte_keepalive_shm); + struct rte_keepalive_shm *shm; + + if (fd < 0) + printf("Failed to open %s as SHM:%s\n", + RTE_KEEPALIVE_SHM_NAME, + strerror(errno)); + else { + shm = (struct rte_keepalive_shm *) mmap( + 0, size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); + close(fd); + if (shm == MAP_FAILED) + printf("Failed to mmap SHM:%s\n", strerror(errno)); + else + return shm; + } + + /* Reset to zero, as it was set to MAP_FAILED aka: (void *)-1 */ + shm = 0; + return NULL; +} + +int main(void) +{ + struct rte_keepalive_shm *shm = ka_shm_create(); + struct timespec timeout = { .tv_nsec = 0 }; + int idx_core; + int cnt_cores; + uint64_t last_seen_alive_time = 0; + uint64_t most_recent_alive_time; + int cnt_timeouts = 0; + int sem_errno; + + if (shm == NULL) { + printf("Unable to access shared core state\n"); + return 1; + } + while (1) { + most_recent_alive_time = 0; + for (idx_core = 0; idx_core < RTE_KEEPALIVE_MAXCORES; + idx_core++) + if (shm->core_last_seen_times[idx_core] > + most_recent_alive_time) + most_recent_alive_time = + shm->core_last_seen_times[idx_core]; + + timeout.tv_sec = time(NULL) + SEM_TIMEOUT_SECS; + if (sem_timedwait(&shm->core_died, &timeout) == -1) { + /* Assume no core death signals and no change in any + * last-seen times is the keepalive monitor itself + * failing. + */ + sem_errno = errno; + last_seen_alive_time = most_recent_alive_time; + if (sem_errno == ETIMEDOUT) { + if (last_seen_alive_time == + most_recent_alive_time && + cnt_timeouts++ > + MAX_TIMEOUTS) { + printf("No updates. Exiting..\n"); + break; + } + } else + printf("sem_timedwait() error (%s)\n", + strerror(sem_errno)); + continue; + } + cnt_timeouts = 0; + + cnt_cores = 0; + for (idx_core = 0; idx_core < RTE_KEEPALIVE_MAXCORES; + idx_core++) + if (shm->core_state[idx_core] == RTE_KA_STATE_DEAD) + cnt_cores++; + if (cnt_cores == 0) { + /* Can happen if core was restarted since Semaphore + * was sent, due to agent being offline. + */ + printf("Warning: Empty dead core report\n"); + continue; + } + + printf("%i dead cores: ", cnt_cores); + for (idx_core = 0; + idx_core < RTE_KEEPALIVE_MAXCORES; + idx_core++) + if (shm->core_state[idx_core] == RTE_KA_STATE_DEAD) + printf("%d, ", idx_core); + printf("\b\b\n"); + } + if (munmap(shm, sizeof(struct rte_keepalive_shm)) != 0) + printf("Warning: munmap() failed\n"); + return 0; +} diff --git a/src/seastar/dpdk/examples/l2fwd-keepalive/main.c b/src/seastar/dpdk/examples/l2fwd-keepalive/main.c new file mode 100644 index 00000000..37453483 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-keepalive/main.c @@ -0,0 +1,819 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2010-2016 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 copyright + * 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 <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <stdint.h> +#include <inttypes.h> +#include <sys/types.h> +#include <sys/queue.h> +#include <netinet/in.h> +#include <setjmp.h> +#include <stdarg.h> +#include <ctype.h> +#include <errno.h> +#include <getopt.h> +#include <signal.h> + +#include <rte_common.h> +#include <rte_log.h> +#include <rte_malloc.h> +#include <rte_memory.h> +#include <rte_memcpy.h> +#include <rte_memzone.h> +#include <rte_eal.h> +#include <rte_per_lcore.h> +#include <rte_launch.h> +#include <rte_atomic.h> +#include <rte_cycles.h> +#include <rte_prefetch.h> +#include <rte_lcore.h> +#include <rte_per_lcore.h> +#include <rte_branch_prediction.h> +#include <rte_interrupts.h> +#include <rte_pci.h> +#include <rte_random.h> +#include <rte_debug.h> +#include <rte_ether.h> +#include <rte_ethdev.h> +#include <rte_mempool.h> +#include <rte_mbuf.h> +#include <rte_timer.h> +#include <rte_keepalive.h> + +#include "shm.h" + +#define RTE_LOGTYPE_L2FWD RTE_LOGTYPE_USER1 + +#define NB_MBUF 8192 + +#define MAX_PKT_BURST 32 +#define BURST_TX_DRAIN_US 100 /* TX drain every ~100us */ + +/* + * Configurable number of RX/TX ring descriptors + */ +#define RTE_TEST_RX_DESC_DEFAULT 128 +#define RTE_TEST_TX_DESC_DEFAULT 512 +static uint16_t nb_rxd = RTE_TEST_RX_DESC_DEFAULT; +static uint16_t nb_txd = RTE_TEST_TX_DESC_DEFAULT; + +/* ethernet addresses of ports */ +static struct ether_addr l2fwd_ports_eth_addr[RTE_MAX_ETHPORTS]; + +/* mask of enabled ports */ +static uint32_t l2fwd_enabled_port_mask; + +/* list of enabled ports */ +static uint32_t l2fwd_dst_ports[RTE_MAX_ETHPORTS]; + +static unsigned int l2fwd_rx_queue_per_lcore = 1; + +#define MAX_RX_QUEUE_PER_LCORE 16 +#define MAX_TX_QUEUE_PER_PORT 16 +struct lcore_queue_conf { + unsigned n_rx_port; + unsigned rx_port_list[MAX_RX_QUEUE_PER_LCORE]; +} __rte_cache_aligned; +struct lcore_queue_conf lcore_queue_conf[RTE_MAX_LCORE]; + +struct rte_eth_dev_tx_buffer *tx_buffer[RTE_MAX_ETHPORTS]; + +static const struct rte_eth_conf port_conf = { + .rxmode = { + .split_hdr_size = 0, + .header_split = 0, /**< Header Split disabled */ + .hw_ip_checksum = 0, /**< IP checksum offload disabled */ + .hw_vlan_filter = 0, /**< VLAN filtering disabled */ + .jumbo_frame = 0, /**< Jumbo Frame Support disabled */ + .hw_strip_crc = 1, /**< CRC stripped by hardware */ + }, + .txmode = { + .mq_mode = ETH_MQ_TX_NONE, + }, +}; + +struct rte_mempool *l2fwd_pktmbuf_pool = NULL; + +/* Per-port statistics struct */ +struct l2fwd_port_statistics { + uint64_t tx; + uint64_t rx; + uint64_t dropped; +} __rte_cache_aligned; +struct l2fwd_port_statistics port_statistics[RTE_MAX_ETHPORTS]; + +/* A tsc-based timer responsible for triggering statistics printout */ +#define TIMER_MILLISECOND 1 +#define MAX_TIMER_PERIOD 86400 /* 1 day max */ +static int64_t timer_period = 10 * TIMER_MILLISECOND * 1000; /* 10 seconds */ +static int64_t check_period = 5; /* default check cycle is 5ms */ + +/* Keepalive structure */ +struct rte_keepalive *rte_global_keepalive_info; + +/* Termination signalling */ +static int terminate_signal_received; + +/* Termination signal handler */ +static void handle_sigterm(__rte_unused int value) +{ + terminate_signal_received = 1; +} + +/* Print out statistics on packets dropped */ +static void +print_stats(__attribute__((unused)) struct rte_timer *ptr_timer, + __attribute__((unused)) void *ptr_data) +{ + uint64_t total_packets_dropped, total_packets_tx, total_packets_rx; + unsigned portid; + + total_packets_dropped = 0; + total_packets_tx = 0; + total_packets_rx = 0; + + const char clr[] = { 27, '[', '2', 'J', '\0' }; + const char topLeft[] = { 27, '[', '1', ';', '1', 'H', '\0' }; + + /* Clear screen and move to top left */ + printf("%s%s", clr, topLeft); + + printf("\nPort statistics ===================================="); + + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) { + /* skip disabled ports */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + printf("\nStatistics for port %u ------------------------------" + "\nPackets sent: %24"PRIu64 + "\nPackets received: %20"PRIu64 + "\nPackets dropped: %21"PRIu64, + portid, + port_statistics[portid].tx, + port_statistics[portid].rx, + port_statistics[portid].dropped); + + total_packets_dropped += port_statistics[portid].dropped; + total_packets_tx += port_statistics[portid].tx; + total_packets_rx += port_statistics[portid].rx; + } + printf("\nAggregate statistics ===============================" + "\nTotal packets sent: %18"PRIu64 + "\nTotal packets received: %14"PRIu64 + "\nTotal packets dropped: %15"PRIu64, + total_packets_tx, + total_packets_rx, + total_packets_dropped); + printf("\n====================================================\n"); +} + +static void +l2fwd_simple_forward(struct rte_mbuf *m, unsigned portid) +{ + struct ether_hdr *eth; + void *tmp; + int sent; + unsigned dst_port; + struct rte_eth_dev_tx_buffer *buffer; + + dst_port = l2fwd_dst_ports[portid]; + eth = rte_pktmbuf_mtod(m, struct ether_hdr *); + + /* 02:00:00:00:00:xx */ + tmp = ð->d_addr.addr_bytes[0]; + *((uint64_t *)tmp) = 0x000000000002 + ((uint64_t)dst_port << 40); + + /* src addr */ + ether_addr_copy(&l2fwd_ports_eth_addr[dst_port], ð->s_addr); + + buffer = tx_buffer[dst_port]; + sent = rte_eth_tx_buffer(dst_port, 0, buffer, m); + if (sent) + port_statistics[dst_port].tx += sent; +} + +/* main processing loop */ +static void +l2fwd_main_loop(void) +{ + struct rte_mbuf *pkts_burst[MAX_PKT_BURST]; + struct rte_mbuf *m; + int sent; + unsigned lcore_id; + uint64_t prev_tsc, diff_tsc, cur_tsc; + unsigned i, j, portid, nb_rx; + struct lcore_queue_conf *qconf; + const uint64_t drain_tsc = (rte_get_tsc_hz() + US_PER_S - 1) + / US_PER_S * BURST_TX_DRAIN_US; + struct rte_eth_dev_tx_buffer *buffer; + + prev_tsc = 0; + + lcore_id = rte_lcore_id(); + qconf = &lcore_queue_conf[lcore_id]; + + if (qconf->n_rx_port == 0) { + RTE_LOG(INFO, L2FWD, "lcore %u has nothing to do\n", lcore_id); + return; + } + + RTE_LOG(INFO, L2FWD, "entering main loop on lcore %u\n", lcore_id); + + for (i = 0; i < qconf->n_rx_port; i++) { + + portid = qconf->rx_port_list[i]; + RTE_LOG(INFO, L2FWD, " -- lcoreid=%u portid=%u\n", lcore_id, + portid); + } + + uint64_t tsc_initial = rte_rdtsc(); + uint64_t tsc_lifetime = (rand()&0x07) * rte_get_tsc_hz(); + + while (!terminate_signal_received) { + /* Keepalive heartbeat */ + rte_keepalive_mark_alive(rte_global_keepalive_info); + + cur_tsc = rte_rdtsc(); + + /* + * Die randomly within 7 secs for demo purposes if + * keepalive enabled + */ + if (check_period > 0 && cur_tsc - tsc_initial > tsc_lifetime) + break; + + /* + * TX burst queue drain + */ + diff_tsc = cur_tsc - prev_tsc; + if (unlikely(diff_tsc > drain_tsc)) { + + for (i = 0; i < qconf->n_rx_port; i++) { + + portid = l2fwd_dst_ports[qconf->rx_port_list[i]]; + buffer = tx_buffer[portid]; + + sent = rte_eth_tx_buffer_flush(portid, 0, buffer); + if (sent) + port_statistics[portid].tx += sent; + + } + + prev_tsc = cur_tsc; + } + + /* + * Read packet from RX queues + */ + for (i = 0; i < qconf->n_rx_port; i++) { + + portid = qconf->rx_port_list[i]; + nb_rx = rte_eth_rx_burst((uint8_t) portid, 0, + pkts_burst, MAX_PKT_BURST); + + port_statistics[portid].rx += nb_rx; + + for (j = 0; j < nb_rx; j++) { + m = pkts_burst[j]; + rte_prefetch0(rte_pktmbuf_mtod(m, void *)); + l2fwd_simple_forward(m, portid); + } + } + } +} + +static int +l2fwd_launch_one_lcore(__attribute__((unused)) void *dummy) +{ + l2fwd_main_loop(); + return 0; +} + +/* display usage */ +static void +l2fwd_usage(const char *prgname) +{ + printf("%s [EAL options] -- -p PORTMASK [-q NQ]\n" + " -p PORTMASK: hexadecimal bitmask of ports to configure\n" + " -q NQ: number of queue (=ports) per lcore (default is 1)\n" + " -K PERIOD: Keepalive check period (5 default; 86400 max)\n" + " -T PERIOD: statistics will be refreshed each PERIOD seconds (0 to disable, 10 default, 86400 maximum)\n", + prgname); +} + +static int +l2fwd_parse_portmask(const char *portmask) +{ + char *end = NULL; + unsigned long pm; + + /* parse hexadecimal string */ + pm = strtoul(portmask, &end, 16); + if ((portmask[0] == '\0') || (end == NULL) || (*end != '\0')) + return -1; + + if (pm == 0) + return -1; + + return pm; +} + +static unsigned int +l2fwd_parse_nqueue(const char *q_arg) +{ + char *end = NULL; + unsigned long n; + + /* parse hexadecimal string */ + n = strtoul(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + return 0; + if (n == 0) + return 0; + if (n >= MAX_RX_QUEUE_PER_LCORE) + return 0; + + return n; +} + +static int +l2fwd_parse_timer_period(const char *q_arg) +{ + char *end = NULL; + int n; + + /* parse number string */ + n = strtol(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + return -1; + if (n >= MAX_TIMER_PERIOD) + return -1; + + return n; +} + +static int +l2fwd_parse_check_period(const char *q_arg) +{ + char *end = NULL; + int n; + + /* parse number string */ + n = strtol(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + return -1; + if (n >= MAX_TIMER_PERIOD) + return -1; + + return n; +} + +/* Parse the argument given in the command line of the application */ +static int +l2fwd_parse_args(int argc, char **argv) +{ + int opt, ret; + char **argvopt; + int option_index; + char *prgname = argv[0]; + static struct option lgopts[] = { + {NULL, 0, 0, 0} + }; + + argvopt = argv; + + while ((opt = getopt_long(argc, argvopt, "p:q:T:K:", + lgopts, &option_index)) != EOF) { + + switch (opt) { + /* portmask */ + case 'p': + l2fwd_enabled_port_mask = l2fwd_parse_portmask(optarg); + if (l2fwd_enabled_port_mask == 0) { + printf("invalid portmask\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* nqueue */ + case 'q': + l2fwd_rx_queue_per_lcore = l2fwd_parse_nqueue(optarg); + if (l2fwd_rx_queue_per_lcore == 0) { + printf("invalid queue number\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* timer period */ + case 'T': + timer_period = l2fwd_parse_timer_period(optarg) + * (int64_t)(1000 * TIMER_MILLISECOND); + if (timer_period < 0) { + printf("invalid timer period\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* Check period */ + case 'K': + check_period = l2fwd_parse_check_period(optarg); + if (check_period < 0) { + printf("invalid check period\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* long options */ + case 0: + l2fwd_usage(prgname); + return -1; + + default: + l2fwd_usage(prgname); + return -1; + } + } + + if (optind >= 0) + argv[optind-1] = prgname; + + ret = optind-1; + optind = 1; /* reset getopt lib */ + return ret; +} + +/* Check the link status of all ports in up to 9s, and print them finally */ +static void +check_all_ports_link_status(uint8_t port_num, uint32_t port_mask) +{ +#define CHECK_INTERVAL 100 /* 100ms */ +#define MAX_CHECK_TIME 90 /* 9s (90 * 100ms) in total */ + uint8_t portid, count, all_ports_up, print_flag = 0; + struct rte_eth_link link; + + printf("\nChecking link status"); + fflush(stdout); + for (count = 0; count <= MAX_CHECK_TIME; count++) { + all_ports_up = 1; + for (portid = 0; portid < port_num; portid++) { + if ((port_mask & (1 << portid)) == 0) + continue; + memset(&link, 0, sizeof(link)); + rte_eth_link_get_nowait(portid, &link); + /* print link status if flag set */ + if (print_flag == 1) { + if (link.link_status) + printf("Port %d Link Up - speed %u " + "Mbps - %s\n", (uint8_t)portid, + (unsigned)link.link_speed, + (link.link_duplex == ETH_LINK_FULL_DUPLEX) ? + ("full-duplex") : ("half-duplex\n")); + else + printf("Port %d Link Down\n", + (uint8_t)portid); + continue; + } + /* clear all_ports_up flag if any link down */ + if (link.link_status == ETH_LINK_DOWN) { + all_ports_up = 0; + break; + } + } + /* after finally printing all link status, get out */ + if (print_flag == 1) + break; + + if (all_ports_up == 0) { + printf("."); + fflush(stdout); + rte_delay_ms(CHECK_INTERVAL); + } + + /* set the print_flag if all ports up or timeout */ + if (all_ports_up == 1 || count == (MAX_CHECK_TIME - 1)) { + print_flag = 1; + printf("done\n"); + } + } +} + +static void +dead_core(__rte_unused void *ptr_data, const int id_core) +{ + if (terminate_signal_received) + return; + printf("Dead core %i - restarting..\n", id_core); + if (rte_eal_get_lcore_state(id_core) == FINISHED) { + rte_eal_wait_lcore(id_core); + rte_eal_remote_launch(l2fwd_launch_one_lcore, NULL, id_core); + } else { + printf("..false positive!\n"); + } +} + +static void +relay_core_state(void *ptr_data, const int id_core, + const enum rte_keepalive_state core_state, uint64_t last_alive) +{ + rte_keepalive_relayed_state((struct rte_keepalive_shm *)ptr_data, + id_core, core_state, last_alive); +} + +int +main(int argc, char **argv) +{ + struct lcore_queue_conf *qconf; + struct rte_eth_dev_info dev_info; + int ret; + uint8_t nb_ports; + uint8_t nb_ports_available; + uint8_t portid, last_port; + unsigned lcore_id, rx_lcore_id; + unsigned nb_ports_in_mask = 0; + struct sigaction signal_handler; + struct rte_keepalive_shm *ka_shm; + + memset(&signal_handler, 0, sizeof(signal_handler)); + terminate_signal_received = 0; + signal_handler.sa_handler = &handle_sigterm; + if (sigaction(SIGINT, &signal_handler, NULL) == -1 || + sigaction(SIGTERM, &signal_handler, NULL) == -1) + rte_exit(EXIT_FAILURE, "SIGNAL\n"); + + + /* init EAL */ + ret = rte_eal_init(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n"); + argc -= ret; + argv += ret; + + l2fwd_enabled_port_mask = 0; + + /* parse application arguments (after the EAL ones) */ + ret = l2fwd_parse_args(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid L2FWD arguments\n"); + + /* create the mbuf pool */ + l2fwd_pktmbuf_pool = rte_pktmbuf_pool_create("mbuf_pool", NB_MBUF, 32, + 0, RTE_MBUF_DEFAULT_BUF_SIZE, rte_socket_id()); + if (l2fwd_pktmbuf_pool == NULL) + rte_exit(EXIT_FAILURE, "Cannot init mbuf pool\n"); + + nb_ports = rte_eth_dev_count(); + if (nb_ports == 0) + rte_exit(EXIT_FAILURE, "No Ethernet ports - bye\n"); + + /* reset l2fwd_dst_ports */ + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) + l2fwd_dst_ports[portid] = 0; + last_port = 0; + + /* + * Each logical core is assigned a dedicated TX queue on each port. + */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + + if (nb_ports_in_mask % 2) { + l2fwd_dst_ports[portid] = last_port; + l2fwd_dst_ports[last_port] = portid; + } else + last_port = portid; + + nb_ports_in_mask++; + + rte_eth_dev_info_get(portid, &dev_info); + } + if (nb_ports_in_mask % 2) { + printf("Notice: odd number of ports in portmask.\n"); + l2fwd_dst_ports[last_port] = last_port; + } + + rx_lcore_id = 1; + qconf = NULL; + + /* Initialize the port/queue configuration of each logical core */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + + /* get the lcore_id for this port */ + while (rte_lcore_is_enabled(rx_lcore_id) == 0 || + lcore_queue_conf[rx_lcore_id].n_rx_port == + l2fwd_rx_queue_per_lcore) { + rx_lcore_id++; + if (rx_lcore_id >= RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE, "Not enough cores\n"); + } + + if (qconf != &lcore_queue_conf[rx_lcore_id]) + /* Assigned a new logical core in the loop above. */ + qconf = &lcore_queue_conf[rx_lcore_id]; + + qconf->rx_port_list[qconf->n_rx_port] = portid; + qconf->n_rx_port++; + printf("Lcore %u: RX port %u\n", + rx_lcore_id, (unsigned) portid); + } + + nb_ports_available = nb_ports; + + /* Initialise each port */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) { + printf("Skipping disabled port %u\n", + (unsigned) portid); + nb_ports_available--; + continue; + } + /* init port */ + printf("Initializing port %u... ", (unsigned) portid); + fflush(stdout); + ret = rte_eth_dev_configure(portid, 1, 1, &port_conf); + if (ret < 0) + rte_exit(EXIT_FAILURE, + "Cannot configure device: err=%d, port=%u\n", + ret, (unsigned) portid); + + rte_eth_macaddr_get(portid, &l2fwd_ports_eth_addr[portid]); + + /* init one RX queue */ + fflush(stdout); + ret = rte_eth_rx_queue_setup(portid, 0, nb_rxd, + rte_eth_dev_socket_id(portid), + NULL, + l2fwd_pktmbuf_pool); + if (ret < 0) + rte_exit(EXIT_FAILURE, + "rte_eth_rx_queue_setup:err=%d, port=%u\n", + ret, (unsigned) portid); + + /* init one TX queue on each port */ + fflush(stdout); + ret = rte_eth_tx_queue_setup(portid, 0, nb_txd, + rte_eth_dev_socket_id(portid), + NULL); + if (ret < 0) + rte_exit(EXIT_FAILURE, + "rte_eth_tx_queue_setup:err=%d, port=%u\n", + ret, (unsigned) portid); + + /* Initialize TX buffers */ + tx_buffer[portid] = rte_zmalloc_socket("tx_buffer", + RTE_ETH_TX_BUFFER_SIZE(MAX_PKT_BURST), 0, + rte_eth_dev_socket_id(portid)); + if (tx_buffer[portid] == NULL) + rte_exit(EXIT_FAILURE, "Cannot allocate buffer for tx on port %u\n", + (unsigned) portid); + + rte_eth_tx_buffer_init(tx_buffer[portid], MAX_PKT_BURST); + + ret = rte_eth_tx_buffer_set_err_callback(tx_buffer[portid], + rte_eth_tx_buffer_count_callback, + &port_statistics[portid].dropped); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Cannot set error callback for " + "tx buffer on port %u\n", (unsigned) portid); + + /* Start device */ + ret = rte_eth_dev_start(portid); + if (ret < 0) + rte_exit(EXIT_FAILURE, + "rte_eth_dev_start:err=%d, port=%u\n", + ret, (unsigned) portid); + + rte_eth_promiscuous_enable(portid); + + printf("Port %u, MAC address: " + "%02X:%02X:%02X:%02X:%02X:%02X\n\n", + (unsigned) portid, + l2fwd_ports_eth_addr[portid].addr_bytes[0], + l2fwd_ports_eth_addr[portid].addr_bytes[1], + l2fwd_ports_eth_addr[portid].addr_bytes[2], + l2fwd_ports_eth_addr[portid].addr_bytes[3], + l2fwd_ports_eth_addr[portid].addr_bytes[4], + l2fwd_ports_eth_addr[portid].addr_bytes[5]); + + /* initialize port stats */ + memset(&port_statistics, 0, sizeof(port_statistics)); + } + + if (!nb_ports_available) { + rte_exit(EXIT_FAILURE, + "All available ports are disabled. Please set portmask.\n"); + } + + check_all_ports_link_status(nb_ports, l2fwd_enabled_port_mask); + + struct rte_timer hb_timer, stats_timer; + + rte_timer_subsystem_init(); + rte_timer_init(&stats_timer); + + ka_shm = NULL; + if (check_period > 0) { + ka_shm = rte_keepalive_shm_create(); + if (ka_shm == NULL) + rte_exit(EXIT_FAILURE, + "rte_keepalive_shm_create() failed"); + rte_global_keepalive_info = + rte_keepalive_create(&dead_core, ka_shm); + if (rte_global_keepalive_info == NULL) + rte_exit(EXIT_FAILURE, "init_keep_alive() failed"); + rte_keepalive_register_relay_callback(rte_global_keepalive_info, + relay_core_state, ka_shm); + rte_timer_init(&hb_timer); + if (rte_timer_reset(&hb_timer, + (check_period * rte_get_timer_hz()) / 1000, + PERIODICAL, + rte_lcore_id(), + (void(*)(struct rte_timer*, void*)) + &rte_keepalive_dispatch_pings, + rte_global_keepalive_info + ) != 0 ) + rte_exit(EXIT_FAILURE, "Keepalive setup failure.\n"); + } + if (timer_period > 0) { + if (rte_timer_reset(&stats_timer, + (timer_period * rte_get_timer_hz()) / 1000, + PERIODICAL, + rte_lcore_id(), + &print_stats, NULL + ) != 0 ) + rte_exit(EXIT_FAILURE, "Stats setup failure.\n"); + } + /* launch per-lcore init on every slave lcore */ + RTE_LCORE_FOREACH_SLAVE(lcore_id) { + struct lcore_queue_conf *qconf = &lcore_queue_conf[lcore_id]; + + if (qconf->n_rx_port == 0) + RTE_LOG(INFO, L2FWD, + "lcore %u has nothing to do\n", + lcore_id + ); + else { + rte_eal_remote_launch( + l2fwd_launch_one_lcore, + NULL, + lcore_id + ); + rte_keepalive_register_core(rte_global_keepalive_info, + lcore_id); + } + } + while (!terminate_signal_received) { + rte_timer_manage(); + rte_delay_ms(5); + } + + RTE_LCORE_FOREACH_SLAVE(lcore_id) { + if (rte_eal_wait_lcore(lcore_id) < 0) + return -1; + } + + if (ka_shm != NULL) + rte_keepalive_shm_cleanup(ka_shm); + return 0; +} diff --git a/src/seastar/dpdk/examples/l2fwd-keepalive/shm.c b/src/seastar/dpdk/examples/l2fwd-keepalive/shm.c new file mode 100644 index 00000000..fbf5bd79 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-keepalive/shm.c @@ -0,0 +1,141 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2016 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 copyright + * 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 <time.h> + +#include <rte_common.h> +#include <rte_log.h> +#include <rte_keepalive.h> + +#include "shm.h" + +struct rte_keepalive_shm *rte_keepalive_shm_create(void) +{ + int fd; + int idx_core; + struct rte_keepalive_shm *ka_shm; + + /* If any existing object is not unlinked, it makes it all too easy + * for clients to end up with stale shared memory blocks when + * restarted. Unlinking makes sure subsequent shm_open by clients + * will get the new block mapped below. + */ + if (shm_unlink(RTE_KEEPALIVE_SHM_NAME) == -1 && errno != ENOENT) + printf("Warning: Error unlinking stale %s (%s)\n", + RTE_KEEPALIVE_SHM_NAME, strerror(errno)); + + fd = shm_open(RTE_KEEPALIVE_SHM_NAME, + O_CREAT | O_TRUNC | O_RDWR, 0666); + if (fd < 0) + RTE_LOG(INFO, EAL, + "Failed to open %s as SHM (%s)\n", + RTE_KEEPALIVE_SHM_NAME, + strerror(errno)); + else if (ftruncate(fd, sizeof(struct rte_keepalive_shm)) != 0) + RTE_LOG(INFO, EAL, + "Failed to resize SHM (%s)\n", strerror(errno)); + else { + ka_shm = (struct rte_keepalive_shm *) mmap( + 0, sizeof(struct rte_keepalive_shm), + PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); + close(fd); + if (ka_shm == MAP_FAILED) + RTE_LOG(INFO, EAL, + "Failed to mmap SHM (%s)\n", strerror(errno)); + else { + memset(ka_shm, 0, sizeof(struct rte_keepalive_shm)); + + /* Initialize the semaphores for IPC/SHM use */ + if (sem_init(&ka_shm->core_died, 1, 0) != 0) { + RTE_LOG(INFO, EAL, + "Failed to setup SHM semaphore (%s)\n", + strerror(errno)); + munmap(ka_shm, + sizeof(struct rte_keepalive_shm)); + return NULL; + } + + /* Set all cores to 'not present' */ + for (idx_core = 0; + idx_core < RTE_KEEPALIVE_MAXCORES; + idx_core++) { + ka_shm->core_state[idx_core] = + RTE_KA_STATE_UNUSED; + ka_shm->core_last_seen_times[idx_core] = 0; + } + + return ka_shm; + } + } +return NULL; +} + +void rte_keepalive_relayed_state(struct rte_keepalive_shm *shm, + const int id_core, const enum rte_keepalive_state core_state, + __rte_unused uint64_t last_alive) +{ + int count; + + shm->core_state[id_core] = core_state; + shm->core_last_seen_times[id_core] = last_alive; + + if (core_state == RTE_KEEPALIVE_SHM_DEAD) { + /* Since core has died, also signal ka_agent. + * + * Limit number of times semaphore can be incremented, in case + * ka_agent is not active. + */ + if (sem_getvalue(&shm->core_died, &count) == -1) { + RTE_LOG(INFO, EAL, "Semaphore check failed(%s)\n", + strerror(errno)); + return; + } + if (count > 1) + return; + + if (sem_post(&shm->core_died) != 0) + RTE_LOG(INFO, EAL, + "Failed to increment semaphore (%s)\n", + strerror(errno)); + } +} + +void rte_keepalive_shm_cleanup(struct rte_keepalive_shm *ka_shm) +{ + if (shm_unlink(RTE_KEEPALIVE_SHM_NAME) == -1 && errno != ENOENT) + printf("Warning: Error unlinking %s (%s)\n", + RTE_KEEPALIVE_SHM_NAME, strerror(errno)); + + if (ka_shm && munmap(ka_shm, sizeof(struct rte_keepalive_shm)) != 0) + printf("Warning: munmap() failed\n"); +} diff --git a/src/seastar/dpdk/examples/l2fwd-keepalive/shm.h b/src/seastar/dpdk/examples/l2fwd-keepalive/shm.h new file mode 100644 index 00000000..66a60600 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd-keepalive/shm.h @@ -0,0 +1,98 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2016 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 copyright + * 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. + */ + +#define RTE_KEEPALIVE_SHM_NAME "/dpdk_keepalive_shm_name" + +#define RTE_KEEPALIVE_SHM_ALIVE 1 +#define RTE_KEEPALIVE_SHM_DEAD 2 + +#include <fcntl.h> +#include <string.h> +#include <unistd.h> +#include <sys/mman.h> +#include <sys/stat.h> +#include <semaphore.h> +#include <rte_keepalive.h> + +/** + * Keepalive SHM structure. + * + * The shared memory allocated by the primary is this size, and contains the + * information as contained within this struct. A secondary may open the SHM, + * and read the contents. + */ +struct rte_keepalive_shm { + /** IPC semaphore. Posted when a core dies */ + sem_t core_died; + + /** + * Relayed status of each core. + */ + enum rte_keepalive_state core_state[RTE_KEEPALIVE_MAXCORES]; + + /** + * Last-seen-alive timestamps for the cores + */ + uint64_t core_last_seen_times[RTE_KEEPALIVE_MAXCORES]; +}; + +/** + * Create shared host memory keepalive object. + * @return + * Pointer to SHM keepalive structure, or NULL on failure. + */ +struct rte_keepalive_shm *rte_keepalive_shm_create(void); + +/** + * Relays state for given core + * @param *shm + * Pointer to SHM keepalive structure. + * @param id_core + * Id of core + * @param core_state + * State of core + * @param last_alive + * Last seen timestamp for core + */ +void rte_keepalive_relayed_state(struct rte_keepalive_shm *shm, + const int id_core, const enum rte_keepalive_state core_state, + uint64_t last_alive); + +/** Shutdown cleanup of shared host memory keepalive object. + * @param *shm + * Pointer to SHM keepalive structure. May be NULL. + * + * If *shm is NULL, this function will only attempt to remove the + * shared host memory handle and not unmap the underlying memory. + */ +void rte_keepalive_shm_cleanup(struct rte_keepalive_shm *ka_shm); diff --git a/src/seastar/dpdk/examples/l2fwd/Makefile b/src/seastar/dpdk/examples/l2fwd/Makefile new file mode 100644 index 00000000..78feeeb8 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd/Makefile @@ -0,0 +1,50 @@ +# 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 copyright +# 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 ?= x86_64-native-linuxapp-gcc + +include $(RTE_SDK)/mk/rte.vars.mk + +# binary name +APP = l2fwd + +# all source are stored in SRCS-y +SRCS-y := main.c + +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) + +include $(RTE_SDK)/mk/rte.extapp.mk diff --git a/src/seastar/dpdk/examples/l2fwd/main.c b/src/seastar/dpdk/examples/l2fwd/main.c new file mode 100644 index 00000000..f9667272 --- /dev/null +++ b/src/seastar/dpdk/examples/l2fwd/main.c @@ -0,0 +1,758 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2010-2016 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 copyright + * 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 <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <stdint.h> +#include <inttypes.h> +#include <sys/types.h> +#include <sys/queue.h> +#include <netinet/in.h> +#include <setjmp.h> +#include <stdarg.h> +#include <ctype.h> +#include <errno.h> +#include <getopt.h> +#include <signal.h> +#include <stdbool.h> + +#include <rte_common.h> +#include <rte_log.h> +#include <rte_malloc.h> +#include <rte_memory.h> +#include <rte_memcpy.h> +#include <rte_memzone.h> +#include <rte_eal.h> +#include <rte_per_lcore.h> +#include <rte_launch.h> +#include <rte_atomic.h> +#include <rte_cycles.h> +#include <rte_prefetch.h> +#include <rte_lcore.h> +#include <rte_per_lcore.h> +#include <rte_branch_prediction.h> +#include <rte_interrupts.h> +#include <rte_pci.h> +#include <rte_random.h> +#include <rte_debug.h> +#include <rte_ether.h> +#include <rte_ethdev.h> +#include <rte_mempool.h> +#include <rte_mbuf.h> + +static volatile bool force_quit; + +/* MAC updating enabled by default */ +static int mac_updating = 1; + +#define RTE_LOGTYPE_L2FWD RTE_LOGTYPE_USER1 + +#define NB_MBUF 8192 + +#define MAX_PKT_BURST 32 +#define BURST_TX_DRAIN_US 100 /* TX drain every ~100us */ +#define MEMPOOL_CACHE_SIZE 256 + +/* + * Configurable number of RX/TX ring descriptors + */ +#define RTE_TEST_RX_DESC_DEFAULT 128 +#define RTE_TEST_TX_DESC_DEFAULT 512 +static uint16_t nb_rxd = RTE_TEST_RX_DESC_DEFAULT; +static uint16_t nb_txd = RTE_TEST_TX_DESC_DEFAULT; + +/* ethernet addresses of ports */ +static struct ether_addr l2fwd_ports_eth_addr[RTE_MAX_ETHPORTS]; + +/* mask of enabled ports */ +static uint32_t l2fwd_enabled_port_mask = 0; + +/* list of enabled ports */ +static uint32_t l2fwd_dst_ports[RTE_MAX_ETHPORTS]; + +static unsigned int l2fwd_rx_queue_per_lcore = 1; + +#define MAX_RX_QUEUE_PER_LCORE 16 +#define MAX_TX_QUEUE_PER_PORT 16 +struct lcore_queue_conf { + unsigned n_rx_port; + unsigned rx_port_list[MAX_RX_QUEUE_PER_LCORE]; +} __rte_cache_aligned; +struct lcore_queue_conf lcore_queue_conf[RTE_MAX_LCORE]; + +static struct rte_eth_dev_tx_buffer *tx_buffer[RTE_MAX_ETHPORTS]; + +static const struct rte_eth_conf port_conf = { + .rxmode = { + .split_hdr_size = 0, + .header_split = 0, /**< Header Split disabled */ + .hw_ip_checksum = 0, /**< IP checksum offload disabled */ + .hw_vlan_filter = 0, /**< VLAN filtering disabled */ + .jumbo_frame = 0, /**< Jumbo Frame Support disabled */ + .hw_strip_crc = 1, /**< CRC stripped by hardware */ + }, + .txmode = { + .mq_mode = ETH_MQ_TX_NONE, + }, +}; + +struct rte_mempool * l2fwd_pktmbuf_pool = NULL; + +/* Per-port statistics struct */ +struct l2fwd_port_statistics { + uint64_t tx; + uint64_t rx; + uint64_t dropped; +} __rte_cache_aligned; +struct l2fwd_port_statistics port_statistics[RTE_MAX_ETHPORTS]; + +#define MAX_TIMER_PERIOD 86400 /* 1 day max */ +/* A tsc-based timer responsible for triggering statistics printout */ +static uint64_t timer_period = 10; /* default period is 10 seconds */ + +/* Print out statistics on packets dropped */ +static void +print_stats(void) +{ + uint64_t total_packets_dropped, total_packets_tx, total_packets_rx; + unsigned portid; + + total_packets_dropped = 0; + total_packets_tx = 0; + total_packets_rx = 0; + + const char clr[] = { 27, '[', '2', 'J', '\0' }; + const char topLeft[] = { 27, '[', '1', ';', '1', 'H','\0' }; + + /* Clear screen and move to top left */ + printf("%s%s", clr, topLeft); + + printf("\nPort statistics ===================================="); + + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) { + /* skip disabled ports */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + printf("\nStatistics for port %u ------------------------------" + "\nPackets sent: %24"PRIu64 + "\nPackets received: %20"PRIu64 + "\nPackets dropped: %21"PRIu64, + portid, + port_statistics[portid].tx, + port_statistics[portid].rx, + port_statistics[portid].dropped); + + total_packets_dropped += port_statistics[portid].dropped; + total_packets_tx += port_statistics[portid].tx; + total_packets_rx += port_statistics[portid].rx; + } + printf("\nAggregate statistics ===============================" + "\nTotal packets sent: %18"PRIu64 + "\nTotal packets received: %14"PRIu64 + "\nTotal packets dropped: %15"PRIu64, + total_packets_tx, + total_packets_rx, + total_packets_dropped); + printf("\n====================================================\n"); +} + +static void +l2fwd_mac_updating(struct rte_mbuf *m, unsigned dest_portid) +{ + struct ether_hdr *eth; + void *tmp; + + eth = rte_pktmbuf_mtod(m, struct ether_hdr *); + + /* 02:00:00:00:00:xx */ + tmp = ð->d_addr.addr_bytes[0]; + *((uint64_t *)tmp) = 0x000000000002 + ((uint64_t)dest_portid << 40); + + /* src addr */ + ether_addr_copy(&l2fwd_ports_eth_addr[dest_portid], ð->s_addr); +} + +static void +l2fwd_simple_forward(struct rte_mbuf *m, unsigned portid) +{ + unsigned dst_port; + int sent; + struct rte_eth_dev_tx_buffer *buffer; + + dst_port = l2fwd_dst_ports[portid]; + + if (mac_updating) + l2fwd_mac_updating(m, dst_port); + + buffer = tx_buffer[dst_port]; + sent = rte_eth_tx_buffer(dst_port, 0, buffer, m); + if (sent) + port_statistics[dst_port].tx += sent; +} + +/* main processing loop */ +static void +l2fwd_main_loop(void) +{ + struct rte_mbuf *pkts_burst[MAX_PKT_BURST]; + struct rte_mbuf *m; + int sent; + unsigned lcore_id; + uint64_t prev_tsc, diff_tsc, cur_tsc, timer_tsc; + unsigned i, j, portid, nb_rx; + struct lcore_queue_conf *qconf; + const uint64_t drain_tsc = (rte_get_tsc_hz() + US_PER_S - 1) / US_PER_S * + BURST_TX_DRAIN_US; + struct rte_eth_dev_tx_buffer *buffer; + + prev_tsc = 0; + timer_tsc = 0; + + lcore_id = rte_lcore_id(); + qconf = &lcore_queue_conf[lcore_id]; + + if (qconf->n_rx_port == 0) { + RTE_LOG(INFO, L2FWD, "lcore %u has nothing to do\n", lcore_id); + return; + } + + RTE_LOG(INFO, L2FWD, "entering main loop on lcore %u\n", lcore_id); + + for (i = 0; i < qconf->n_rx_port; i++) { + + portid = qconf->rx_port_list[i]; + RTE_LOG(INFO, L2FWD, " -- lcoreid=%u portid=%u\n", lcore_id, + portid); + + } + + while (!force_quit) { + + cur_tsc = rte_rdtsc(); + + /* + * TX burst queue drain + */ + diff_tsc = cur_tsc - prev_tsc; + if (unlikely(diff_tsc > drain_tsc)) { + + for (i = 0; i < qconf->n_rx_port; i++) { + + portid = l2fwd_dst_ports[qconf->rx_port_list[i]]; + buffer = tx_buffer[portid]; + + sent = rte_eth_tx_buffer_flush(portid, 0, buffer); + if (sent) + port_statistics[portid].tx += sent; + + } + + /* if timer is enabled */ + if (timer_period > 0) { + + /* advance the timer */ + timer_tsc += diff_tsc; + + /* if timer has reached its timeout */ + if (unlikely(timer_tsc >= timer_period)) { + + /* do this only on master core */ + if (lcore_id == rte_get_master_lcore()) { + print_stats(); + /* reset the timer */ + timer_tsc = 0; + } + } + } + + prev_tsc = cur_tsc; + } + + /* + * Read packet from RX queues + */ + for (i = 0; i < qconf->n_rx_port; i++) { + + portid = qconf->rx_port_list[i]; + nb_rx = rte_eth_rx_burst((uint8_t) portid, 0, + pkts_burst, MAX_PKT_BURST); + + port_statistics[portid].rx += nb_rx; + + for (j = 0; j < nb_rx; j++) { + m = pkts_burst[j]; + rte_prefetch0(rte_pktmbuf_mtod(m, void *)); + l2fwd_simple_forward(m, portid); + } + } + } +} + +static int +l2fwd_launch_one_lcore(__attribute__((unused)) void *dummy) +{ + l2fwd_main_loop(); + return 0; +} + +/* display usage */ +static void +l2fwd_usage(const char *prgname) +{ + printf("%s [EAL options] -- -p PORTMASK [-q NQ]\n" + " -p PORTMASK: hexadecimal bitmask of ports to configure\n" + " -q NQ: number of queue (=ports) per lcore (default is 1)\n" + " -T PERIOD: statistics will be refreshed each PERIOD seconds (0 to disable, 10 default, 86400 maximum)\n" + " --[no-]mac-updating: Enable or disable MAC addresses updating (enabled by default)\n" + " When enabled:\n" + " - The source MAC address is replaced by the TX port MAC address\n" + " - The destination MAC address is replaced by 02:00:00:00:00:TX_PORT_ID\n", + prgname); +} + +static int +l2fwd_parse_portmask(const char *portmask) +{ + char *end = NULL; + unsigned long pm; + + /* parse hexadecimal string */ + pm = strtoul(portmask, &end, 16); + if ((portmask[0] == '\0') || (end == NULL) || (*end != '\0')) + return -1; + + if (pm == 0) + return -1; + + return pm; +} + +static unsigned int +l2fwd_parse_nqueue(const char *q_arg) +{ + char *end = NULL; + unsigned long n; + + /* parse hexadecimal string */ + n = strtoul(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + return 0; + if (n == 0) + return 0; + if (n >= MAX_RX_QUEUE_PER_LCORE) + return 0; + + return n; +} + +static int +l2fwd_parse_timer_period(const char *q_arg) +{ + char *end = NULL; + int n; + + /* parse number string */ + n = strtol(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0')) + return -1; + if (n >= MAX_TIMER_PERIOD) + return -1; + + return n; +} + +static const char short_options[] = + "p:" /* portmask */ + "q:" /* number of queues */ + "T:" /* timer period */ + ; + +#define CMD_LINE_OPT_MAC_UPDATING "mac-updating" +#define CMD_LINE_OPT_NO_MAC_UPDATING "no-mac-updating" + +enum { + /* long options mapped to a short option */ + + /* first long only option value must be >= 256, so that we won't + * conflict with short options */ + CMD_LINE_OPT_MIN_NUM = 256, +}; + +static const struct option lgopts[] = { + { CMD_LINE_OPT_MAC_UPDATING, no_argument, &mac_updating, 1}, + { CMD_LINE_OPT_NO_MAC_UPDATING, no_argument, &mac_updating, 0}, + {NULL, 0, 0, 0} +}; + +/* Parse the argument given in the command line of the application */ +static int +l2fwd_parse_args(int argc, char **argv) +{ + int opt, ret, timer_secs; + char **argvopt; + int option_index; + char *prgname = argv[0]; + + argvopt = argv; + + while ((opt = getopt_long(argc, argvopt, short_options, + lgopts, &option_index)) != EOF) { + + switch (opt) { + /* portmask */ + case 'p': + l2fwd_enabled_port_mask = l2fwd_parse_portmask(optarg); + if (l2fwd_enabled_port_mask == 0) { + printf("invalid portmask\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* nqueue */ + case 'q': + l2fwd_rx_queue_per_lcore = l2fwd_parse_nqueue(optarg); + if (l2fwd_rx_queue_per_lcore == 0) { + printf("invalid queue number\n"); + l2fwd_usage(prgname); + return -1; + } + break; + + /* timer period */ + case 'T': + timer_secs = l2fwd_parse_timer_period(optarg); + if (timer_secs < 0) { + printf("invalid timer period\n"); + l2fwd_usage(prgname); + return -1; + } + timer_period = timer_secs; + break; + + /* long options */ + case 0: + break; + + default: + l2fwd_usage(prgname); + return -1; + } + } + + if (optind >= 0) + argv[optind-1] = prgname; + + ret = optind-1; + optind = 1; /* reset getopt lib */ + return ret; +} + +/* Check the link status of all ports in up to 9s, and print them finally */ +static void +check_all_ports_link_status(uint8_t port_num, uint32_t port_mask) +{ +#define CHECK_INTERVAL 100 /* 100ms */ +#define MAX_CHECK_TIME 90 /* 9s (90 * 100ms) in total */ + uint8_t portid, count, all_ports_up, print_flag = 0; + struct rte_eth_link link; + + printf("\nChecking link status"); + fflush(stdout); + for (count = 0; count <= MAX_CHECK_TIME; count++) { + if (force_quit) + return; + all_ports_up = 1; + for (portid = 0; portid < port_num; portid++) { + if (force_quit) + return; + if ((port_mask & (1 << portid)) == 0) + continue; + memset(&link, 0, sizeof(link)); + rte_eth_link_get_nowait(portid, &link); + /* print link status if flag set */ + if (print_flag == 1) { + if (link.link_status) + printf("Port %d Link Up - speed %u " + "Mbps - %s\n", (uint8_t)portid, + (unsigned)link.link_speed, + (link.link_duplex == ETH_LINK_FULL_DUPLEX) ? + ("full-duplex") : ("half-duplex\n")); + else + printf("Port %d Link Down\n", + (uint8_t)portid); + continue; + } + /* clear all_ports_up flag if any link down */ + if (link.link_status == ETH_LINK_DOWN) { + all_ports_up = 0; + break; + } + } + /* after finally printing all link status, get out */ + if (print_flag == 1) + break; + + if (all_ports_up == 0) { + printf("."); + fflush(stdout); + rte_delay_ms(CHECK_INTERVAL); + } + + /* set the print_flag if all ports up or timeout */ + if (all_ports_up == 1 || count == (MAX_CHECK_TIME - 1)) { + print_flag = 1; + printf("done\n"); + } + } +} + +static void +signal_handler(int signum) +{ + if (signum == SIGINT || signum == SIGTERM) { + printf("\n\nSignal %d received, preparing to exit...\n", + signum); + force_quit = true; + } +} + +int +main(int argc, char **argv) +{ + struct lcore_queue_conf *qconf; + struct rte_eth_dev_info dev_info; + int ret; + uint8_t nb_ports; + uint8_t nb_ports_available; + uint8_t portid, last_port; + unsigned lcore_id, rx_lcore_id; + unsigned nb_ports_in_mask = 0; + + /* init EAL */ + ret = rte_eal_init(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n"); + argc -= ret; + argv += ret; + + force_quit = false; + signal(SIGINT, signal_handler); + signal(SIGTERM, signal_handler); + + /* parse application arguments (after the EAL ones) */ + ret = l2fwd_parse_args(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid L2FWD arguments\n"); + + printf("MAC updating %s\n", mac_updating ? "enabled" : "disabled"); + + /* convert to number of cycles */ + timer_period *= rte_get_timer_hz(); + + /* create the mbuf pool */ + l2fwd_pktmbuf_pool = rte_pktmbuf_pool_create("mbuf_pool", NB_MBUF, + MEMPOOL_CACHE_SIZE, 0, RTE_MBUF_DEFAULT_BUF_SIZE, + rte_socket_id()); + if (l2fwd_pktmbuf_pool == NULL) + rte_exit(EXIT_FAILURE, "Cannot init mbuf pool\n"); + + nb_ports = rte_eth_dev_count(); + if (nb_ports == 0) + rte_exit(EXIT_FAILURE, "No Ethernet ports - bye\n"); + + /* reset l2fwd_dst_ports */ + for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) + l2fwd_dst_ports[portid] = 0; + last_port = 0; + + /* + * Each logical core is assigned a dedicated TX queue on each port. + */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + + if (nb_ports_in_mask % 2) { + l2fwd_dst_ports[portid] = last_port; + l2fwd_dst_ports[last_port] = portid; + } + else + last_port = portid; + + nb_ports_in_mask++; + + rte_eth_dev_info_get(portid, &dev_info); + } + if (nb_ports_in_mask % 2) { + printf("Notice: odd number of ports in portmask.\n"); + l2fwd_dst_ports[last_port] = last_port; + } + + rx_lcore_id = 0; + qconf = NULL; + + /* Initialize the port/queue configuration of each logical core */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + + /* get the lcore_id for this port */ + while (rte_lcore_is_enabled(rx_lcore_id) == 0 || + lcore_queue_conf[rx_lcore_id].n_rx_port == + l2fwd_rx_queue_per_lcore) { + rx_lcore_id++; + if (rx_lcore_id >= RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE, "Not enough cores\n"); + } + + if (qconf != &lcore_queue_conf[rx_lcore_id]) + /* Assigned a new logical core in the loop above. */ + qconf = &lcore_queue_conf[rx_lcore_id]; + + qconf->rx_port_list[qconf->n_rx_port] = portid; + qconf->n_rx_port++; + printf("Lcore %u: RX port %u\n", rx_lcore_id, (unsigned) portid); + } + + nb_ports_available = nb_ports; + + /* Initialise each port */ + for (portid = 0; portid < nb_ports; portid++) { + /* skip ports that are not enabled */ + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) { + printf("Skipping disabled port %u\n", (unsigned) portid); + nb_ports_available--; + continue; + } + /* init port */ + printf("Initializing port %u... ", (unsigned) portid); + fflush(stdout); + ret = rte_eth_dev_configure(portid, 1, 1, &port_conf); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Cannot configure device: err=%d, port=%u\n", + ret, (unsigned) portid); + + rte_eth_macaddr_get(portid,&l2fwd_ports_eth_addr[portid]); + + /* init one RX queue */ + fflush(stdout); + ret = rte_eth_rx_queue_setup(portid, 0, nb_rxd, + rte_eth_dev_socket_id(portid), + NULL, + l2fwd_pktmbuf_pool); + if (ret < 0) + rte_exit(EXIT_FAILURE, "rte_eth_rx_queue_setup:err=%d, port=%u\n", + ret, (unsigned) portid); + + /* init one TX queue on each port */ + fflush(stdout); + ret = rte_eth_tx_queue_setup(portid, 0, nb_txd, + rte_eth_dev_socket_id(portid), + NULL); + if (ret < 0) + rte_exit(EXIT_FAILURE, "rte_eth_tx_queue_setup:err=%d, port=%u\n", + ret, (unsigned) portid); + + /* Initialize TX buffers */ + tx_buffer[portid] = rte_zmalloc_socket("tx_buffer", + RTE_ETH_TX_BUFFER_SIZE(MAX_PKT_BURST), 0, + rte_eth_dev_socket_id(portid)); + if (tx_buffer[portid] == NULL) + rte_exit(EXIT_FAILURE, "Cannot allocate buffer for tx on port %u\n", + (unsigned) portid); + + rte_eth_tx_buffer_init(tx_buffer[portid], MAX_PKT_BURST); + + ret = rte_eth_tx_buffer_set_err_callback(tx_buffer[portid], + rte_eth_tx_buffer_count_callback, + &port_statistics[portid].dropped); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Cannot set error callback for " + "tx buffer on port %u\n", (unsigned) portid); + + /* Start device */ + ret = rte_eth_dev_start(portid); + if (ret < 0) + rte_exit(EXIT_FAILURE, "rte_eth_dev_start:err=%d, port=%u\n", + ret, (unsigned) portid); + + printf("done: \n"); + + rte_eth_promiscuous_enable(portid); + + printf("Port %u, MAC address: %02X:%02X:%02X:%02X:%02X:%02X\n\n", + (unsigned) portid, + l2fwd_ports_eth_addr[portid].addr_bytes[0], + l2fwd_ports_eth_addr[portid].addr_bytes[1], + l2fwd_ports_eth_addr[portid].addr_bytes[2], + l2fwd_ports_eth_addr[portid].addr_bytes[3], + l2fwd_ports_eth_addr[portid].addr_bytes[4], + l2fwd_ports_eth_addr[portid].addr_bytes[5]); + + /* initialize port stats */ + memset(&port_statistics, 0, sizeof(port_statistics)); + } + + if (!nb_ports_available) { + rte_exit(EXIT_FAILURE, + "All available ports are disabled. Please set portmask.\n"); + } + + check_all_ports_link_status(nb_ports, l2fwd_enabled_port_mask); + + ret = 0; + /* launch per-lcore init on every lcore */ + rte_eal_mp_remote_launch(l2fwd_launch_one_lcore, NULL, CALL_MASTER); + RTE_LCORE_FOREACH_SLAVE(lcore_id) { + if (rte_eal_wait_lcore(lcore_id) < 0) { + ret = -1; + break; + } + } + + for (portid = 0; portid < nb_ports; portid++) { + if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) + continue; + printf("Closing port %d...", portid); + rte_eth_dev_stop(portid); + rte_eth_dev_close(portid); + printf(" Done\n"); + } + printf("Bye...\n"); + + return ret; +} |