diff options
Diffstat (limited to 'src/seastar/dpdk/examples/vhost_xen')
| -rw-r--r-- | src/seastar/dpdk/examples/vhost_xen/Makefile | 52 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/vhost_xen/main.c | 1509 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/vhost_xen/main.h | 66 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/vhost_xen/vhost_monitor.c | 595 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/vhost_xen/virtio-net.h | 113 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/vhost_xen/xen_vhost.h | 148 | ||||
| -rw-r--r-- | src/seastar/dpdk/examples/vhost_xen/xenstore_parse.c | 775 |
7 files changed, 3258 insertions, 0 deletions
diff --git a/src/seastar/dpdk/examples/vhost_xen/Makefile b/src/seastar/dpdk/examples/vhost_xen/Makefile new file mode 100644 index 00000000..47e14898 --- /dev/null +++ b/src/seastar/dpdk/examples/vhost_xen/Makefile @@ -0,0 +1,52 @@ +# 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 = vhost-switch + +# all source are stored in SRCS-y +SRCS-y := main.c vhost_monitor.c xenstore_parse.c + +CFLAGS += -O2 -I/usr/local/include -D_FILE_OFFSET_BITS=64 -Wno-unused-parameter +CFLAGS += $(WERROR_FLAGS) +CFLAGS += -D_GNU_SOURCE +LDFLAGS += -lxenstore + +include $(RTE_SDK)/mk/rte.extapp.mk diff --git a/src/seastar/dpdk/examples/vhost_xen/main.c b/src/seastar/dpdk/examples/vhost_xen/main.c new file mode 100644 index 00000000..d9ef140f --- /dev/null +++ b/src/seastar/dpdk/examples/vhost_xen/main.c @@ -0,0 +1,1509 @@ +/*- + * 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. + */ + +#include <arpa/inet.h> +#include <getopt.h> +#include <linux/if_ether.h> +#include <linux/if_vlan.h> +#include <linux/virtio_net.h> +#include <linux/virtio_ring.h> +#include <signal.h> +#include <stdint.h> +#include <sys/eventfd.h> +#include <sys/param.h> +#include <unistd.h> + +#include <rte_atomic.h> +#include <rte_cycles.h> +#include <rte_ethdev.h> +#include <rte_log.h> +#include <rte_string_fns.h> + +#include "main.h" +#include "virtio-net.h" +#include "xen_vhost.h" + +#define MAX_QUEUES 128 + +/* the maximum number of external ports supported */ +#define MAX_SUP_PORTS 1 + +/* + * Calculate the number of buffers needed per port + */ +#define NUM_MBUFS_PER_PORT ((MAX_QUEUES*RTE_TEST_RX_DESC_DEFAULT) + \ + (num_switching_cores*MAX_PKT_BURST) + \ + (num_switching_cores*RTE_TEST_TX_DESC_DEFAULT) +\ + (num_switching_cores*MBUF_CACHE_SIZE)) + +#define MBUF_CACHE_SIZE 64 + +/* + * RX and TX Prefetch, Host, and Write-back threshold values should be + * carefully set for optimal performance. Consult the network + * controller's datasheet and supporting DPDK documentation for guidance + * on how these parameters should be set. + */ +#define RX_PTHRESH 8 /* Default values of RX prefetch threshold reg. */ +#define RX_HTHRESH 8 /* Default values of RX host threshold reg. */ +#define RX_WTHRESH 4 /* Default values of RX write-back threshold reg. */ + +/* + * These default values are optimized for use with the Intel(R) 82599 10 GbE + * Controller and the DPDK ixgbe PMD. Consider using other values for other + * network controllers and/or network drivers. + */ +#define TX_PTHRESH 36 /* Default values of TX prefetch threshold reg. */ +#define TX_HTHRESH 0 /* Default values of TX host threshold reg. */ +#define TX_WTHRESH 0 /* Default values of TX write-back threshold reg. */ + +#define MAX_PKT_BURST 32 /* Max burst size for RX/TX */ +#define MAX_MRG_PKT_BURST 16 /* Max burst for merge buffers. Set to 1 due to performance issue. */ +#define BURST_TX_DRAIN_US 100 /* TX drain every ~100us */ + +/* State of virtio device. */ +#define DEVICE_NOT_READY 0 +#define DEVICE_READY 1 +#define DEVICE_SAFE_REMOVE 2 + +/* Config_core_flag status definitions. */ +#define REQUEST_DEV_REMOVAL 1 +#define ACK_DEV_REMOVAL 0 + +/* Configurable number of RX/TX ring descriptors */ +#define RTE_TEST_RX_DESC_DEFAULT 128 +#define RTE_TEST_TX_DESC_DEFAULT 512 + +#define INVALID_PORT_ID 0xFF + +/* Max number of devices. Limited by vmdq. */ +#define MAX_DEVICES 64 + +/* Size of buffers used for snprintfs. */ +#define MAX_PRINT_BUFF 6072 + + +/* Maximum long option length for option parsing. */ +#define MAX_LONG_OPT_SZ 64 + +/* Used to compare MAC addresses. */ +#define MAC_ADDR_CMP 0xFFFFFFFFFFFF + +/* mask of enabled ports */ +static uint32_t enabled_port_mask = 0; + +/*Number of switching cores enabled*/ +static uint32_t num_switching_cores = 0; + +/* number of devices/queues to support*/ +static uint32_t num_queues = 0; +uint32_t num_devices = 0; + +/* Enable VM2VM communications. If this is disabled then the MAC address compare is skipped. */ +static uint32_t enable_vm2vm = 1; +/* Enable stats. */ +static uint32_t enable_stats = 0; + +/* empty vmdq configuration structure. Filled in programatically */ +static const struct rte_eth_conf vmdq_conf_default = { + .rxmode = { + .mq_mode = ETH_MQ_RX_VMDQ_ONLY, + .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 */ + /* + * It is necessary for 1G NIC such as I350, + * this fixes bug of ipv4 forwarding in guest can't + * forward pakets from one virtio dev to another virtio dev. + */ + .hw_vlan_strip = 1, /**< VLAN strip enabled. */ + .jumbo_frame = 0, /**< Jumbo Frame Support disabled */ + .hw_strip_crc = 1, /**< CRC stripped by hardware */ + }, + + .txmode = { + .mq_mode = ETH_MQ_TX_NONE, + }, + .rx_adv_conf = { + /* + * should be overridden separately in code with + * appropriate values + */ + .vmdq_rx_conf = { + .nb_queue_pools = ETH_8_POOLS, + .enable_default_pool = 0, + .default_pool = 0, + .nb_pool_maps = 0, + .pool_map = {{0, 0},}, + }, + }, +}; + +static unsigned lcore_ids[RTE_MAX_LCORE]; +static uint8_t ports[RTE_MAX_ETHPORTS]; +static unsigned num_ports = 0; /**< The number of ports specified in command line */ + +const uint16_t vlan_tags[] = { + 1000, 1001, 1002, 1003, 1004, 1005, 1006, 1007, + 1008, 1009, 1010, 1011, 1012, 1013, 1014, 1015, + 1016, 1017, 1018, 1019, 1020, 1021, 1022, 1023, + 1024, 1025, 1026, 1027, 1028, 1029, 1030, 1031, + 1032, 1033, 1034, 1035, 1036, 1037, 1038, 1039, + 1040, 1041, 1042, 1043, 1044, 1045, 1046, 1047, + 1048, 1049, 1050, 1051, 1052, 1053, 1054, 1055, + 1056, 1057, 1058, 1059, 1060, 1061, 1062, 1063, +}; + +/* ethernet addresses of ports */ +static struct ether_addr vmdq_ports_eth_addr[RTE_MAX_ETHPORTS]; + +/* heads for the main used and free linked lists for the data path. */ +static struct virtio_net_data_ll *ll_root_used = NULL; +static struct virtio_net_data_ll *ll_root_free = NULL; + +/* Array of data core structures containing information on individual core linked lists. */ +static struct lcore_info lcore_info[RTE_MAX_LCORE]; + +/* Used for queueing bursts of TX packets. */ +struct mbuf_table { + unsigned len; + unsigned txq_id; + struct rte_mbuf *m_table[MAX_PKT_BURST]; +}; + +/* TX queue for each data core. */ +struct mbuf_table lcore_tx_queue[RTE_MAX_LCORE]; + +/* Vlan header struct used to insert vlan tags on TX. */ +struct vlan_ethhdr { + unsigned char h_dest[ETH_ALEN]; + unsigned char h_source[ETH_ALEN]; + __be16 h_vlan_proto; + __be16 h_vlan_TCI; + __be16 h_vlan_encapsulated_proto; +}; + +/* Header lengths. */ +#define VLAN_HLEN 4 +#define VLAN_ETH_HLEN 18 + +/* Per-device statistics struct */ +struct device_statistics { + uint64_t tx_total; + rte_atomic64_t rx_total; + uint64_t tx; + rte_atomic64_t rx; +} __rte_cache_aligned; +struct device_statistics dev_statistics[MAX_DEVICES]; + +/* + * Builds up the correct configuration for VMDQ VLAN pool map + * according to the pool & queue limits. + */ +static inline int +get_eth_conf(struct rte_eth_conf *eth_conf, uint32_t num_devices) +{ + struct rte_eth_vmdq_rx_conf conf; + unsigned i; + + memset(&conf, 0, sizeof(conf)); + conf.nb_queue_pools = (enum rte_eth_nb_pools)num_devices; + conf.nb_pool_maps = num_devices; + + for (i = 0; i < conf.nb_pool_maps; i++) { + conf.pool_map[i].vlan_id = vlan_tags[ i ]; + conf.pool_map[i].pools = (1UL << i); + } + + (void)(rte_memcpy(eth_conf, &vmdq_conf_default, sizeof(*eth_conf))); + (void)(rte_memcpy(ð_conf->rx_adv_conf.vmdq_rx_conf, &conf, + sizeof(eth_conf->rx_adv_conf.vmdq_rx_conf))); + return 0; +} + +/* + * Validate the device number according to the max pool number gotten form dev_info + * If the device number is invalid, give the error message and return -1. + * Each device must have its own pool. + */ +static inline int +validate_num_devices(uint32_t max_nb_devices) +{ + if (num_devices > max_nb_devices) { + RTE_LOG(ERR, VHOST_PORT, "invalid number of devices\n"); + return -1; + } + return 0; +} + +/* + * Initialises a given port using global settings and with the rx buffers + * coming from the mbuf_pool passed as parameter + */ +static inline int +port_init(uint8_t port, struct rte_mempool *mbuf_pool) +{ + struct rte_eth_dev_info dev_info; + struct rte_eth_rxconf *rxconf; + struct rte_eth_conf port_conf; + uint16_t rx_rings, tx_rings = (uint16_t)rte_lcore_count(); + const uint16_t rx_ring_size = RTE_TEST_RX_DESC_DEFAULT, tx_ring_size = RTE_TEST_TX_DESC_DEFAULT; + int retval; + uint16_t q; + + /* The max pool number from dev_info will be used to validate the pool number specified in cmd line */ + rte_eth_dev_info_get (port, &dev_info); + + /*configure the number of supported virtio devices based on VMDQ limits */ + num_devices = dev_info.max_vmdq_pools; + num_queues = dev_info.max_rx_queues; + + retval = validate_num_devices(MAX_DEVICES); + if (retval < 0) + return retval; + + /* Get port configuration. */ + retval = get_eth_conf(&port_conf, num_devices); + if (retval < 0) + return retval; + + if (port >= rte_eth_dev_count()) return -1; + + rx_rings = (uint16_t)num_queues, + /* Configure ethernet device. */ + retval = rte_eth_dev_configure(port, rx_rings, tx_rings, &port_conf); + if (retval != 0) + return retval; + + rte_eth_dev_info_get(port, &dev_info); + rxconf = &dev_info.default_rxconf; + rxconf->rx_drop_en = 1; + /* Setup the queues. */ + for (q = 0; q < rx_rings; q ++) { + retval = rte_eth_rx_queue_setup(port, q, rx_ring_size, + rte_eth_dev_socket_id(port), rxconf, + mbuf_pool); + if (retval < 0) + return retval; + } + 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 device. */ + retval = rte_eth_dev_start(port); + if (retval < 0) + return retval; + + rte_eth_macaddr_get(port, &vmdq_ports_eth_addr[port]); + RTE_LOG(INFO, VHOST_PORT, "Max virtio devices supported: %u\n", num_devices); + RTE_LOG(INFO, VHOST_PORT, "Port %u MAC: %02"PRIx8" %02"PRIx8" %02"PRIx8 + " %02"PRIx8" %02"PRIx8" %02"PRIx8"\n", + (unsigned)port, + vmdq_ports_eth_addr[port].addr_bytes[0], + vmdq_ports_eth_addr[port].addr_bytes[1], + vmdq_ports_eth_addr[port].addr_bytes[2], + vmdq_ports_eth_addr[port].addr_bytes[3], + vmdq_ports_eth_addr[port].addr_bytes[4], + vmdq_ports_eth_addr[port].addr_bytes[5]); + + return 0; +} + +/* + * Parse the portmask provided at run time. + */ +static int +parse_portmask(const char *portmask) +{ + char *end = NULL; + unsigned long pm; + + errno = 0; + + /* parse hexadecimal string */ + pm = strtoul(portmask, &end, 16); + if ((portmask[0] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) + return -1; + + if (pm == 0) + return -1; + + return pm; + +} + +/* + * Parse num options at run time. + */ +static int +parse_num_opt(const char *q_arg, uint32_t max_valid_value) +{ + char *end = NULL; + unsigned long num; + + errno = 0; + + /* parse unsigned int string */ + num = strtoul(q_arg, &end, 10); + if ((q_arg[0] == '\0') || (end == NULL) || (*end != '\0') || (errno != 0)) + return -1; + + if (num > max_valid_value) + return -1; + + return num; + +} + +/* + * Display usage + */ +static void +us_vhost_usage(const char *prgname) +{ + RTE_LOG(INFO, VHOST_CONFIG, "%s [EAL options] -- -p PORTMASK --vm2vm [0|1] --stats [0-N] --nb-devices ND\n" + " -p PORTMASK: Set mask for ports to be used by application\n" + " --vm2vm [0|1]: disable/enable(default) vm2vm comms\n" + " --stats [0-N]: 0: Disable stats, N: Time in seconds to print stats\n", + prgname); +} + +/* + * Parse the arguments given in the command line of the application. + */ +static int +us_vhost_parse_args(int argc, char **argv) +{ + int opt, ret; + int option_index; + unsigned i; + const char *prgname = argv[0]; + static struct option long_option[] = { + {"vm2vm", required_argument, NULL, 0}, + {"stats", required_argument, NULL, 0}, + {NULL, 0, 0, 0} + }; + + /* Parse command line */ + while ((opt = getopt_long(argc, argv, "p:",long_option, &option_index)) != EOF) { + switch (opt) { + /* Portmask */ + case 'p': + enabled_port_mask = parse_portmask(optarg); + if (enabled_port_mask == 0) { + RTE_LOG(INFO, VHOST_CONFIG, "Invalid portmask\n"); + us_vhost_usage(prgname); + return -1; + } + break; + + case 0: + /* Enable/disable vm2vm comms. */ + if (!strncmp(long_option[option_index].name, "vm2vm", MAX_LONG_OPT_SZ)) { + ret = parse_num_opt(optarg, 1); + if (ret == -1) { + RTE_LOG(INFO, VHOST_CONFIG, "Invalid argument for vm2vm [0|1]\n"); + us_vhost_usage(prgname); + return -1; + } else { + enable_vm2vm = ret; + } + } + + /* Enable/disable stats. */ + if (!strncmp(long_option[option_index].name, "stats", MAX_LONG_OPT_SZ)) { + ret = parse_num_opt(optarg, INT32_MAX); + if (ret == -1) { + RTE_LOG(INFO, VHOST_CONFIG, "Invalid argument for stats [0..N]\n"); + us_vhost_usage(prgname); + return -1; + } else { + enable_stats = ret; + } + } + break; + + /* Invalid option - print options. */ + default: + us_vhost_usage(prgname); + return -1; + } + } + + for (i = 0; i < RTE_MAX_ETHPORTS; i++) { + if (enabled_port_mask & (1 << i)) + ports[num_ports++] = (uint8_t)i; + } + + if ((num_ports == 0) || (num_ports > MAX_SUP_PORTS)) { + RTE_LOG(INFO, VHOST_PORT, "Current enabled port number is %u," + "but only %u port can be enabled\n",num_ports, MAX_SUP_PORTS); + return -1; + } + + return 0; +} + +/* + * Update the global var NUM_PORTS and array PORTS according to system ports number + * and return valid ports number + */ +static unsigned check_ports_num(unsigned nb_ports) +{ + unsigned valid_num_ports = num_ports; + unsigned portid; + + if (num_ports > nb_ports) { + RTE_LOG(INFO, VHOST_PORT, "\nSpecified port number(%u) exceeds total system port number(%u)\n", + num_ports, nb_ports); + num_ports = nb_ports; + } + + for (portid = 0; portid < num_ports; portid ++) { + if (ports[portid] >= nb_ports) { + RTE_LOG(INFO, VHOST_PORT, "\nSpecified port ID(%u) exceeds max system port ID(%u)\n", + ports[portid], (nb_ports - 1)); + ports[portid] = INVALID_PORT_ID; + valid_num_ports--; + } + } + return valid_num_ports; +} + +/* + * Function to convert guest physical addresses to vhost virtual addresses. This + * is used to convert virtio buffer addresses. + */ +static inline uint64_t __attribute__((always_inline)) +gpa_to_vva(struct virtio_net *dev, uint64_t guest_pa) +{ + struct virtio_memory_regions *region; + uint32_t regionidx; + uint64_t vhost_va = 0; + + for (regionidx = 0; regionidx < dev->mem->nregions; regionidx++) { + region = &dev->mem->regions[regionidx]; + if ((guest_pa >= region->guest_phys_address) && + (guest_pa <= region->guest_phys_address_end)) { + vhost_va = region->address_offset + guest_pa; + break; + } + } + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") GPA %p| VVA %p\n", + dev->device_fh, (void*)(uintptr_t)guest_pa, (void*)(uintptr_t)vhost_va); + + return vhost_va; +} + +/* + * This function adds buffers to the virtio devices RX virtqueue. Buffers can + * be received from the physical port or from another virtio device. A packet + * count is returned to indicate the number of packets that were succesfully + * added to the RX queue. + */ +static inline uint32_t __attribute__((always_inline)) +virtio_dev_rx(struct virtio_net *dev, struct rte_mbuf **pkts, uint32_t count) +{ + struct vhost_virtqueue *vq; + struct vring_desc *desc; + struct rte_mbuf *buff; + /* The virtio_hdr is initialised to 0. */ + struct virtio_net_hdr_mrg_rxbuf virtio_hdr = {{0,0,0,0,0,0},0}; + uint64_t buff_addr = 0; + uint64_t buff_hdr_addr = 0; + uint32_t head[MAX_PKT_BURST], packet_len = 0; + uint32_t head_idx, packet_success = 0; + uint16_t avail_idx, res_cur_idx; + uint16_t res_base_idx, res_end_idx; + uint16_t free_entries; + uint8_t success = 0; + void *userdata; + + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") virtio_dev_rx()\n", dev->device_fh); + vq = dev->virtqueue_rx; + count = (count > MAX_PKT_BURST) ? MAX_PKT_BURST : count; + /* As many data cores may want access to available buffers, they need to be reserved. */ + do { + + res_base_idx = vq->last_used_idx_res; + + avail_idx = *((volatile uint16_t *)&vq->avail->idx); + + free_entries = (avail_idx - res_base_idx); + + /*check that we have enough buffers*/ + if (unlikely(count > free_entries)) + count = free_entries; + + if (count == 0) + return 0; + + res_end_idx = res_base_idx + count; + /* vq->last_used_idx_res is atomically updated. */ + success = rte_atomic16_cmpset(&vq->last_used_idx_res, res_base_idx, + res_end_idx); + } while (unlikely(success == 0)); + res_cur_idx = res_base_idx; + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") Current Index %d| End Index %d\n", + dev->device_fh, res_cur_idx, res_end_idx); + + /* Prefetch available ring to retrieve indexes. */ + rte_prefetch0(&vq->avail->ring[res_cur_idx & (vq->size - 1)]); + + /* Retrieve all of the head indexes first to avoid caching issues. */ + for (head_idx = 0; head_idx < count; head_idx++) + head[head_idx] = vq->avail->ring[(res_cur_idx + head_idx) & (vq->size - 1)]; + + /*Prefetch descriptor index. */ + rte_prefetch0(&vq->desc[head[packet_success]]); + + while (res_cur_idx != res_end_idx) { + /* Get descriptor from available ring */ + desc = &vq->desc[head[packet_success]]; + /* Prefetch descriptor address. */ + rte_prefetch0(desc); + + buff = pkts[packet_success]; + + /* Convert from gpa to vva (guest physical addr -> vhost virtual addr) */ + buff_addr = gpa_to_vva(dev, desc->addr); + /* Prefetch buffer address. */ + rte_prefetch0((void*)(uintptr_t)buff_addr); + + { + /* Copy virtio_hdr to packet and increment buffer address */ + buff_hdr_addr = buff_addr; + packet_len = rte_pktmbuf_data_len(buff) + vq->vhost_hlen; + + /* + * If the descriptors are chained the header and data are placed in + * separate buffers. + */ + if (desc->flags & VRING_DESC_F_NEXT) { + desc->len = vq->vhost_hlen; + desc = &vq->desc[desc->next]; + /* Buffer address translation. */ + buff_addr = gpa_to_vva(dev, desc->addr); + desc->len = rte_pktmbuf_data_len(buff); + } else { + buff_addr += vq->vhost_hlen; + desc->len = packet_len; + } + } + + /* Update used ring with desc information */ + vq->used->ring[res_cur_idx & (vq->size - 1)].id = head[packet_success]; + vq->used->ring[res_cur_idx & (vq->size - 1)].len = packet_len; + + /* Copy mbuf data to buffer */ + userdata = rte_pktmbuf_mtod(buff, void *); + rte_memcpy((void *)(uintptr_t)buff_addr, userdata, rte_pktmbuf_data_len(buff)); + + res_cur_idx++; + packet_success++; + + /* mergeable is disabled then a header is required per buffer. */ + rte_memcpy((void *)(uintptr_t)buff_hdr_addr, (const void *)&virtio_hdr, vq->vhost_hlen); + if (res_cur_idx < res_end_idx) { + /* Prefetch descriptor index. */ + rte_prefetch0(&vq->desc[head[packet_success]]); + } + } + + rte_compiler_barrier(); + + /* Wait until it's our turn to add our buffer to the used ring. */ + while (unlikely(vq->last_used_idx != res_base_idx)) + rte_pause(); + + *(volatile uint16_t *)&vq->used->idx += count; + + vq->last_used_idx = res_end_idx; + + return count; +} + +/* + * Compares a packet destination MAC address to a device MAC address. + */ +static inline int __attribute__((always_inline)) +ether_addr_cmp(struct ether_addr *ea, struct ether_addr *eb) +{ + return ((*(uint64_t *)ea ^ *(uint64_t *)eb) & MAC_ADDR_CMP) == 0; +} + +/* + * This function registers mac along with a + * vlan tag to a VMDQ. + */ +static int +link_vmdq(struct virtio_net *dev) +{ + int ret; + struct virtio_net_data_ll *dev_ll; + + dev_ll = ll_root_used; + + while (dev_ll != NULL) { + if ((dev != dev_ll->dev) && ether_addr_cmp(&dev->mac_address, &dev_ll->dev->mac_address)) { + RTE_LOG(INFO, VHOST_DATA, "(%"PRIu64") WARNING: This device is using an existing MAC address and has not been registered.\n", dev->device_fh); + return -1; + } + dev_ll = dev_ll->next; + } + + /* vlan_tag currently uses the device_id. */ + dev->vlan_tag = vlan_tags[dev->device_fh]; + dev->vmdq_rx_q = dev->device_fh * (num_queues/num_devices); + + /* Print out VMDQ registration info. */ + RTE_LOG(INFO, VHOST_DATA, "(%"PRIu64") MAC_ADDRESS %02x:%02x:%02x:%02x:%02x:%02x and VLAN_TAG %d registered\n", + dev->device_fh, + dev->mac_address.addr_bytes[0], dev->mac_address.addr_bytes[1], + dev->mac_address.addr_bytes[2], dev->mac_address.addr_bytes[3], + dev->mac_address.addr_bytes[4], dev->mac_address.addr_bytes[5], + dev->vlan_tag); + + /* Register the MAC address. */ + ret = rte_eth_dev_mac_addr_add(ports[0], &dev->mac_address, (uint32_t)dev->device_fh); + if (ret) { + RTE_LOG(ERR, VHOST_DATA, "(%"PRIu64") Failed to add device MAC address to VMDQ\n", + dev->device_fh); + return -1; + } + + /* Enable stripping of the vlan tag as we handle routing. */ + rte_eth_dev_set_vlan_strip_on_queue(ports[0], dev->vmdq_rx_q, 1); + + rte_compiler_barrier(); + /* Set device as ready for RX. */ + dev->ready = DEVICE_READY; + + return 0; +} + +/* + * Removes MAC address and vlan tag from VMDQ. Ensures that nothing is adding buffers to the RX + * queue before disabling RX on the device. + */ +static inline void +unlink_vmdq(struct virtio_net *dev) +{ + unsigned i = 0; + unsigned rx_count; + struct rte_mbuf *pkts_burst[MAX_PKT_BURST]; + + if (dev->ready == DEVICE_READY) { + /*clear MAC and VLAN settings*/ + rte_eth_dev_mac_addr_remove(ports[0], &dev->mac_address); + for (i = 0; i < 6; i++) + dev->mac_address.addr_bytes[i] = 0; + + dev->vlan_tag = 0; + + /*Clear out the receive buffers*/ + rx_count = rte_eth_rx_burst(ports[0], + (uint16_t)dev->vmdq_rx_q, pkts_burst, MAX_PKT_BURST); + + while (rx_count) { + for (i = 0; i < rx_count; i++) + rte_pktmbuf_free(pkts_burst[i]); + + rx_count = rte_eth_rx_burst(ports[0], + (uint16_t)dev->vmdq_rx_q, pkts_burst, MAX_PKT_BURST); + } + + dev->ready = DEVICE_NOT_READY; + } +} + +/* + * Check if the packet destination MAC address is for a local device. If so then put + * the packet on that devices RX queue. If not then return. + */ +static inline unsigned __attribute__((always_inline)) +virtio_tx_local(struct virtio_net *dev, struct rte_mbuf *m) +{ + struct virtio_net_data_ll *dev_ll; + struct ether_hdr *pkt_hdr; + uint64_t ret = 0; + + pkt_hdr = rte_pktmbuf_mtod(m, struct ether_hdr *); + + /*get the used devices list*/ + dev_ll = ll_root_used; + + while (dev_ll != NULL) { + if (likely(dev_ll->dev->ready == DEVICE_READY) && ether_addr_cmp(&(pkt_hdr->d_addr), + &dev_ll->dev->mac_address)) { + + /* Drop the packet if the TX packet is destined for the TX device. */ + if (dev_ll->dev->device_fh == dev->device_fh) { + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") TX: " + "Source and destination MAC addresses are the same. " + "Dropping packet.\n", + dev_ll->dev->device_fh); + return 0; + } + + + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") TX: " + "MAC address is local\n", dev_ll->dev->device_fh); + + if (dev_ll->dev->remove) { + /*drop the packet if the device is marked for removal*/ + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") " + "Device is marked for removal\n", + dev_ll->dev->device_fh); + } else { + /*send the packet to the local virtio device*/ + ret = virtio_dev_rx(dev_ll->dev, &m, 1); + if (enable_stats) { + rte_atomic64_add(&dev_statistics[dev_ll->dev->device_fh].rx_total, 1); + rte_atomic64_add(&dev_statistics[dev_ll->dev->device_fh].rx, ret); + dev_statistics[dev->device_fh].tx_total++; + dev_statistics[dev->device_fh].tx += ret; + } + } + + return 0; + } + dev_ll = dev_ll->next; + } + + return -1; +} + +/* + * This function routes the TX packet to the correct interface. This may be a local device + * or the physical port. + */ +static inline void __attribute__((always_inline)) +virtio_tx_route(struct virtio_net* dev, struct rte_mbuf *m, struct rte_mempool *mbuf_pool, uint16_t vlan_tag) +{ + struct mbuf_table *tx_q; + struct vlan_ethhdr *vlan_hdr; + struct rte_mbuf **m_table; + struct rte_mbuf *mbuf; + unsigned len, ret; + const uint16_t lcore_id = rte_lcore_id(); + + /*check if destination is local VM*/ + if (enable_vm2vm && (virtio_tx_local(dev, m) == 0)) { + return; + } + + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") TX: " + "MAC address is external\n", dev->device_fh); + + /*Add packet to the port tx queue*/ + tx_q = &lcore_tx_queue[lcore_id]; + len = tx_q->len; + + /* Allocate an mbuf and populate the structure. */ + mbuf = rte_pktmbuf_alloc(mbuf_pool); + if(!mbuf) + return; + + mbuf->data_len = m->data_len + VLAN_HLEN; + mbuf->pkt_len = mbuf->data_len; + + /* Copy ethernet header to mbuf. */ + rte_memcpy(rte_pktmbuf_mtod(mbuf, void*), + rte_pktmbuf_mtod(m, const void*), ETH_HLEN); + + + /* Setup vlan header. Bytes need to be re-ordered for network with htons()*/ + vlan_hdr = rte_pktmbuf_mtod(mbuf, struct vlan_ethhdr *); + vlan_hdr->h_vlan_encapsulated_proto = vlan_hdr->h_vlan_proto; + vlan_hdr->h_vlan_proto = htons(ETH_P_8021Q); + vlan_hdr->h_vlan_TCI = htons(vlan_tag); + + /* Copy the remaining packet contents to the mbuf. */ + rte_memcpy(rte_pktmbuf_mtod_offset(mbuf, void *, VLAN_ETH_HLEN), + rte_pktmbuf_mtod_offset(m, const void *, ETH_HLEN), + (m->data_len - ETH_HLEN)); + tx_q->m_table[len] = mbuf; + len++; + if (enable_stats) { + dev_statistics[dev->device_fh].tx_total++; + dev_statistics[dev->device_fh].tx++; + } + + if (unlikely(len == MAX_PKT_BURST)) { + m_table = (struct rte_mbuf **)tx_q->m_table; + ret = rte_eth_tx_burst(ports[0], (uint16_t)tx_q->txq_id, m_table, (uint16_t) len); + /* Free any buffers not handled by TX and update the port stats. */ + if (unlikely(ret < len)) { + do { + rte_pktmbuf_free(m_table[ret]); + } while (++ret < len); + } + + len = 0; + } + + tx_q->len = len; + return; +} + +static inline void __attribute__((always_inline)) +virtio_dev_tx(struct virtio_net* dev, struct rte_mempool *mbuf_pool) +{ + struct rte_mbuf m; + struct vhost_virtqueue *vq; + struct vring_desc *desc; + uint64_t buff_addr = 0; + uint32_t head[MAX_PKT_BURST]; + uint32_t used_idx; + uint32_t i; + uint16_t free_entries, packet_success = 0; + uint16_t avail_idx; + + vq = dev->virtqueue_tx; + avail_idx = *((volatile uint16_t *)&vq->avail->idx); + + /* If there are no available buffers then return. */ + if (vq->last_used_idx == avail_idx) + return; + + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") virtio_dev_tx()\n", + dev->device_fh); + + /* Prefetch available ring to retrieve head indexes. */ + rte_prefetch0(&vq->avail->ring[vq->last_used_idx & (vq->size - 1)]); + + /*get the number of free entries in the ring*/ + free_entries = avail_idx - vq->last_used_idx; + free_entries = unlikely(free_entries < MAX_PKT_BURST) ? free_entries : MAX_PKT_BURST; + + RTE_LOG_DP(DEBUG, VHOST_DATA, "(%" PRIu64 ") Buffers available %d\n", + dev->device_fh, free_entries); + /* Retrieve all of the head indexes first to avoid caching issues. */ + for (i = 0; i < free_entries; i++) + head[i] = vq->avail->ring[(vq->last_used_idx + i) & (vq->size - 1)]; + + /* Prefetch descriptor index. */ + rte_prefetch0(&vq->desc[head[packet_success]]); + + while (packet_success < free_entries) { + desc = &vq->desc[head[packet_success]]; + /* Prefetch descriptor address. */ + rte_prefetch0(desc); + + if (packet_success < (free_entries - 1)) { + /* Prefetch descriptor index. */ + rte_prefetch0(&vq->desc[head[packet_success+1]]); + } + + /* Update used index buffer information. */ + used_idx = vq->last_used_idx & (vq->size - 1); + vq->used->ring[used_idx].id = head[packet_success]; + vq->used->ring[used_idx].len = 0; + + /* Discard first buffer as it is the virtio header */ + desc = &vq->desc[desc->next]; + + /* Buffer address translation. */ + buff_addr = gpa_to_vva(dev, desc->addr); + /* Prefetch buffer address. */ + rte_prefetch0((void*)(uintptr_t)buff_addr); + + /* Setup dummy mbuf. This is copied to a real mbuf if transmitted out the physical port. */ + m.data_len = desc->len; + m.data_off = 0; + m.nb_segs = 1; + + virtio_tx_route(dev, &m, mbuf_pool, 0); + + vq->last_used_idx++; + packet_success++; + } + + rte_compiler_barrier(); + vq->used->idx += packet_success; + /* Kick guest if required. */ +} + +/* + * This function is called by each data core. It handles all RX/TX registered with the + * core. For TX the specific lcore linked list is used. For RX, MAC addresses are compared + * with all devices in the main linked list. + */ +static int +switch_worker(__attribute__((unused)) void *arg) +{ + struct rte_mempool *mbuf_pool = arg; + struct virtio_net *dev = NULL; + struct rte_mbuf *pkts_burst[MAX_PKT_BURST]; + struct virtio_net_data_ll *dev_ll; + struct mbuf_table *tx_q; + volatile struct lcore_ll_info *lcore_ll; + const uint64_t drain_tsc = (rte_get_tsc_hz() + US_PER_S - 1) / US_PER_S * BURST_TX_DRAIN_US; + uint64_t prev_tsc, diff_tsc, cur_tsc, ret_count = 0; + unsigned ret, i; + const uint16_t lcore_id = rte_lcore_id(); + const uint16_t num_cores = (uint16_t)rte_lcore_count(); + uint16_t rx_count = 0; + + RTE_LOG(INFO, VHOST_DATA, "Procesing on Core %u started \n", lcore_id); + lcore_ll = lcore_info[lcore_id].lcore_ll; + prev_tsc = 0; + + tx_q = &lcore_tx_queue[lcore_id]; + for (i = 0; i < num_cores; i ++) { + if (lcore_ids[i] == lcore_id) { + tx_q->txq_id = i; + break; + } + } + + while(1) { + cur_tsc = rte_rdtsc(); + /* + * TX burst queue drain + */ + diff_tsc = cur_tsc - prev_tsc; + if (unlikely(diff_tsc > drain_tsc)) { + + if (tx_q->len) { + RTE_LOG_DP(DEBUG, VHOST_DATA, + "TX queue drained after timeout with burst size %u\n", + tx_q->len); + + /*Tx any packets in the queue*/ + ret = rte_eth_tx_burst(ports[0], (uint16_t)tx_q->txq_id, + (struct rte_mbuf **)tx_q->m_table, + (uint16_t)tx_q->len); + if (unlikely(ret < tx_q->len)) { + do { + rte_pktmbuf_free(tx_q->m_table[ret]); + } while (++ret < tx_q->len); + } + + tx_q->len = 0; + } + + prev_tsc = cur_tsc; + + } + + /* + * Inform the configuration core that we have exited the linked list and that no devices are + * in use if requested. + */ + if (lcore_ll->dev_removal_flag == REQUEST_DEV_REMOVAL) + lcore_ll->dev_removal_flag = ACK_DEV_REMOVAL; + + /* + * Process devices + */ + dev_ll = lcore_ll->ll_root_used; + + while (dev_ll != NULL) { + /*get virtio device ID*/ + dev = dev_ll->dev; + + if (unlikely(dev->remove)) { + dev_ll = dev_ll->next; + unlink_vmdq(dev); + dev->ready = DEVICE_SAFE_REMOVE; + continue; + } + if (likely(dev->ready == DEVICE_READY)) { + /*Handle guest RX*/ + rx_count = rte_eth_rx_burst(ports[0], + (uint16_t)dev->vmdq_rx_q, pkts_burst, MAX_PKT_BURST); + + if (rx_count) { + ret_count = virtio_dev_rx(dev, pkts_burst, rx_count); + if (enable_stats) { + rte_atomic64_add(&dev_statistics[dev_ll->dev->device_fh].rx_total, rx_count); + rte_atomic64_add(&dev_statistics[dev_ll->dev->device_fh].rx, ret_count); + } + while (likely(rx_count)) { + rx_count--; + rte_pktmbuf_free_seg(pkts_burst[rx_count]); + } + + } + } + + if (likely(!dev->remove)) + /*Handle guest TX*/ + virtio_dev_tx(dev, mbuf_pool); + + /*move to the next device in the list*/ + dev_ll = dev_ll->next; + } + } + + return 0; +} + +/* + * Add an entry to a used linked list. A free entry must first be found in the free linked list + * using get_data_ll_free_entry(); + */ +static void +add_data_ll_entry(struct virtio_net_data_ll **ll_root_addr, struct virtio_net_data_ll *ll_dev) +{ + struct virtio_net_data_ll *ll = *ll_root_addr; + + /* Set next as NULL and use a compiler barrier to avoid reordering. */ + ll_dev->next = NULL; + rte_compiler_barrier(); + + /* If ll == NULL then this is the first device. */ + if (ll) { + /* Increment to the tail of the linked list. */ + while ((ll->next != NULL) ) + ll = ll->next; + + ll->next = ll_dev; + } else { + *ll_root_addr = ll_dev; + } +} + +/* + * Remove an entry from a used linked list. The entry must then be added to the free linked list + * using put_data_ll_free_entry(). + */ +static void +rm_data_ll_entry(struct virtio_net_data_ll **ll_root_addr, struct virtio_net_data_ll *ll_dev, struct virtio_net_data_ll *ll_dev_last) +{ + struct virtio_net_data_ll *ll = *ll_root_addr; + + if (ll_dev == ll) + *ll_root_addr = ll_dev->next; + else + ll_dev_last->next = ll_dev->next; +} + +/* + * Find and return an entry from the free linked list. + */ +static struct virtio_net_data_ll * +get_data_ll_free_entry(struct virtio_net_data_ll **ll_root_addr) +{ + struct virtio_net_data_ll *ll_free = *ll_root_addr; + struct virtio_net_data_ll *ll_dev; + + if (ll_free == NULL) + return NULL; + + ll_dev = ll_free; + *ll_root_addr = ll_free->next; + + return ll_dev; +} + +/* + * Place an entry back on to the free linked list. + */ +static void +put_data_ll_free_entry(struct virtio_net_data_ll **ll_root_addr, struct virtio_net_data_ll *ll_dev) +{ + struct virtio_net_data_ll *ll_free = *ll_root_addr; + + ll_dev->next = ll_free; + *ll_root_addr = ll_dev; +} + +/* + * Creates a linked list of a given size. + */ +static struct virtio_net_data_ll * +alloc_data_ll(uint32_t size) +{ + struct virtio_net_data_ll *ll_new; + uint32_t i; + + /* Malloc and then chain the linked list. */ + ll_new = malloc(size * sizeof(struct virtio_net_data_ll)); + if (ll_new == NULL) { + RTE_LOG(ERR, VHOST_CONFIG, "Failed to allocate memory for ll_new.\n"); + return NULL; + } + + for (i = 0; i < size - 1; i++) { + ll_new[i].dev = NULL; + ll_new[i].next = &ll_new[i+1]; + } + ll_new[i].next = NULL; + + return ll_new; +} + +/* + * Create the main linked list along with each individual cores linked list. A used and a free list + * are created to manage entries. + */ +static int +init_data_ll (void) +{ + int lcore; + + RTE_LCORE_FOREACH_SLAVE(lcore) { + lcore_info[lcore].lcore_ll = malloc(sizeof(struct lcore_ll_info)); + if (lcore_info[lcore].lcore_ll == NULL) { + RTE_LOG(ERR, VHOST_CONFIG, "Failed to allocate memory for lcore_ll.\n"); + return -1; + } + + lcore_info[lcore].lcore_ll->device_num = 0; + lcore_info[lcore].lcore_ll->dev_removal_flag = ACK_DEV_REMOVAL; + lcore_info[lcore].lcore_ll->ll_root_used = NULL; + if (num_devices % num_switching_cores) + lcore_info[lcore].lcore_ll->ll_root_free = alloc_data_ll((num_devices / num_switching_cores) + 1); + else + lcore_info[lcore].lcore_ll->ll_root_free = alloc_data_ll(num_devices / num_switching_cores); + } + + /* Allocate devices up to a maximum of MAX_DEVICES. */ + ll_root_free = alloc_data_ll(MIN((num_devices), MAX_DEVICES)); + + return 0; +} +/* + * Remove a device from the specific data core linked list and from the main linked list. The + * rx/tx thread must be set the flag to indicate that it is safe to remove the device. + * used. + */ +static void +destroy_device (volatile struct virtio_net *dev) +{ + struct virtio_net_data_ll *ll_lcore_dev_cur; + struct virtio_net_data_ll *ll_main_dev_cur; + struct virtio_net_data_ll *ll_lcore_dev_last = NULL; + struct virtio_net_data_ll *ll_main_dev_last = NULL; + int lcore; + + dev->flags &= ~VIRTIO_DEV_RUNNING; + + /*set the remove flag. */ + dev->remove = 1; + + while(dev->ready != DEVICE_SAFE_REMOVE) { + rte_pause(); + } + + /* Search for entry to be removed from lcore ll */ + ll_lcore_dev_cur = lcore_info[dev->coreid].lcore_ll->ll_root_used; + while (ll_lcore_dev_cur != NULL) { + if (ll_lcore_dev_cur->dev == dev) { + break; + } else { + ll_lcore_dev_last = ll_lcore_dev_cur; + ll_lcore_dev_cur = ll_lcore_dev_cur->next; + } + } + + /* Search for entry to be removed from main ll */ + ll_main_dev_cur = ll_root_used; + ll_main_dev_last = NULL; + while (ll_main_dev_cur != NULL) { + if (ll_main_dev_cur->dev == dev) { + break; + } else { + ll_main_dev_last = ll_main_dev_cur; + ll_main_dev_cur = ll_main_dev_cur->next; + } + } + + if (ll_lcore_dev_cur == NULL || ll_main_dev_cur == NULL) { + RTE_LOG(ERR, XENHOST, "%s: could find device in per_cpu list or main_list\n", __func__); + return; + } + + /* Remove entries from the lcore and main ll. */ + rm_data_ll_entry(&lcore_info[ll_lcore_dev_cur->dev->coreid].lcore_ll->ll_root_used, ll_lcore_dev_cur, ll_lcore_dev_last); + rm_data_ll_entry(&ll_root_used, ll_main_dev_cur, ll_main_dev_last); + + /* Set the dev_removal_flag on each lcore. */ + RTE_LCORE_FOREACH_SLAVE(lcore) { + lcore_info[lcore].lcore_ll->dev_removal_flag = REQUEST_DEV_REMOVAL; + } + + /* + * Once each core has set the dev_removal_flag to ACK_DEV_REMOVAL we can be sure that + * they can no longer access the device removed from the linked lists and that the devices + * are no longer in use. + */ + RTE_LCORE_FOREACH_SLAVE(lcore) { + while (lcore_info[lcore].lcore_ll->dev_removal_flag != ACK_DEV_REMOVAL) { + rte_pause(); + } + } + + /* Add the entries back to the lcore and main free ll.*/ + put_data_ll_free_entry(&lcore_info[ll_lcore_dev_cur->dev->coreid].lcore_ll->ll_root_free, ll_lcore_dev_cur); + put_data_ll_free_entry(&ll_root_free, ll_main_dev_cur); + + /* Decrement number of device on the lcore. */ + lcore_info[ll_lcore_dev_cur->dev->coreid].lcore_ll->device_num--; + + RTE_LOG(INFO, VHOST_DATA, " #####(%"PRIu64") Device has been removed from data core\n", dev->device_fh); +} + +/* + * A new device is added to a data core. First the device is added to the main linked list + * and the allocated to a specific data core. + */ +static int +new_device (struct virtio_net *dev) +{ + struct virtio_net_data_ll *ll_dev; + int lcore, core_add = 0; + uint32_t device_num_min = num_devices; + + /* Add device to main ll */ + ll_dev = get_data_ll_free_entry(&ll_root_free); + if (ll_dev == NULL) { + RTE_LOG(INFO, VHOST_DATA, "(%"PRIu64") No free entry found in linked list. Device limit " + "of %d devices per core has been reached\n", + dev->device_fh, num_devices); + return -1; + } + ll_dev->dev = dev; + add_data_ll_entry(&ll_root_used, ll_dev); + + /*reset ready flag*/ + dev->ready = DEVICE_NOT_READY; + dev->remove = 0; + + /* Find a suitable lcore to add the device. */ + RTE_LCORE_FOREACH_SLAVE(lcore) { + if (lcore_info[lcore].lcore_ll->device_num < device_num_min) { + device_num_min = lcore_info[lcore].lcore_ll->device_num; + core_add = lcore; + } + } + /* Add device to lcore ll */ + ll_dev->dev->coreid = core_add; + ll_dev = get_data_ll_free_entry(&lcore_info[ll_dev->dev->coreid].lcore_ll->ll_root_free); + if (ll_dev == NULL) { + RTE_LOG(INFO, VHOST_DATA, "(%"PRIu64") Failed to add device to data core\n", dev->device_fh); + destroy_device(dev); + return -1; + } + ll_dev->dev = dev; + add_data_ll_entry(&lcore_info[ll_dev->dev->coreid].lcore_ll->ll_root_used, ll_dev); + + /* Initialize device stats */ + memset(&dev_statistics[dev->device_fh], 0, sizeof(struct device_statistics)); + + lcore_info[ll_dev->dev->coreid].lcore_ll->device_num++; + dev->flags |= VIRTIO_DEV_RUNNING; + + RTE_LOG(INFO, VHOST_DATA, "(%"PRIu64") Device has been added to data core %d\n", dev->device_fh, dev->coreid); + + link_vmdq(dev); + + return 0; +} + +/* + * These callback allow devices to be added to the data core when configuration + * has been fully complete. + */ +static const struct virtio_net_device_ops virtio_net_device_ops = +{ + .new_device = new_device, + .destroy_device = destroy_device, +}; + +/* + * This is a thread will wake up after a period to print stats if the user has + * enabled them. + */ +static void +print_stats(void) +{ + struct virtio_net_data_ll *dev_ll; + uint64_t tx_dropped, rx_dropped; + uint64_t tx, tx_total, rx, rx_total; + uint32_t device_fh; + const char clr[] = { 27, '[', '2', 'J', '\0' }; + const char top_left[] = { 27, '[', '1', ';', '1', 'H','\0' }; + + while(1) { + sleep(enable_stats); + + /* Clear screen and move to top left */ + printf("%s%s", clr, top_left); + + printf("\nDevice statistics ===================================="); + + dev_ll = ll_root_used; + while (dev_ll != NULL) { + device_fh = (uint32_t)dev_ll->dev->device_fh; + tx_total = dev_statistics[device_fh].tx_total; + tx = dev_statistics[device_fh].tx; + tx_dropped = tx_total - tx; + rx_total = rte_atomic64_read(&dev_statistics[device_fh].rx_total); + rx = rte_atomic64_read(&dev_statistics[device_fh].rx); + rx_dropped = rx_total - rx; + + printf("\nStatistics for device %"PRIu32" ------------------------------" + "\nTX total: %"PRIu64"" + "\nTX dropped: %"PRIu64"" + "\nTX successful: %"PRIu64"" + "\nRX total: %"PRIu64"" + "\nRX dropped: %"PRIu64"" + "\nRX successful: %"PRIu64"", + device_fh, + tx_total, + tx_dropped, + tx, + rx_total, + rx_dropped, + rx); + + dev_ll = dev_ll->next; + } + printf("\n======================================================\n"); + } +} + + +int init_virtio_net(struct virtio_net_device_ops const * const ops); + +/* + * Main function, does initialisation and calls the per-lcore functions. + */ +int +main(int argc, char *argv[]) +{ + struct rte_mempool *mbuf_pool; + unsigned lcore_id, core_id = 0; + unsigned nb_ports, valid_num_ports; + int ret; + uint8_t portid; + static pthread_t tid; + char thread_name[RTE_MAX_THREAD_NAME_LEN]; + + /* init EAL */ + ret = rte_eal_init(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Error with EAL initialization\n"); + argc -= ret; + argv += ret; + + /* parse app arguments */ + ret = us_vhost_parse_args(argc, argv); + if (ret < 0) + rte_exit(EXIT_FAILURE, "Invalid argument\n"); + + for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id ++) + if (rte_lcore_is_enabled(lcore_id)) + lcore_ids[core_id ++] = lcore_id; + + if (rte_lcore_count() > RTE_MAX_LCORE) + rte_exit(EXIT_FAILURE,"Not enough cores\n"); + + /*set the number of swithcing cores available*/ + num_switching_cores = rte_lcore_count()-1; + + /* Get the number of physical ports. */ + nb_ports = rte_eth_dev_count(); + + /* + * Update the global var NUM_PORTS and global array PORTS + * and get value of var VALID_NUM_PORTS according to system ports number + */ + valid_num_ports = check_ports_num(nb_ports); + + if ((valid_num_ports == 0) || (valid_num_ports > MAX_SUP_PORTS)) { + RTE_LOG(INFO, VHOST_PORT, "Current enabled port number is %u," + "but only %u port can be enabled\n",num_ports, MAX_SUP_PORTS); + return -1; + } + + /* Create the mbuf pool. */ + mbuf_pool = rte_pktmbuf_pool_create("MBUF_POOL", + NUM_MBUFS_PER_PORT * valid_num_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++) { + /* skip ports that are not enabled */ + if ((enabled_port_mask & (1 << portid)) == 0) { + RTE_LOG(INFO, VHOST_PORT, "Skipping disabled port %d\n", portid); + continue; + } + if (port_init(portid, mbuf_pool) != 0) + rte_exit(EXIT_FAILURE, "Cannot initialize network ports\n"); + } + + /* Initialise all linked lists. */ + if (init_data_ll() == -1) + rte_exit(EXIT_FAILURE, "Failed to initialize linked list\n"); + + /* Initialize device stats */ + memset(&dev_statistics, 0, sizeof(dev_statistics)); + + /* Enable stats if the user option is set. */ + if (enable_stats) { + ret = pthread_create(&tid, NULL, (void *)print_stats, NULL); + if (ret != 0) + rte_exit(EXIT_FAILURE, + "Cannot create print-stats thread\n"); + + /* Set thread_name for aid in debugging. */ + snprintf(thread_name, RTE_MAX_THREAD_NAME_LEN, "print-xen-stats"); + ret = rte_thread_setname(tid, thread_name); + if (ret != 0) + RTE_LOG(DEBUG, VHOST_CONFIG, + "Cannot set print-stats name\n"); + } + + /* Launch all data cores. */ + RTE_LCORE_FOREACH_SLAVE(lcore_id) { + rte_eal_remote_launch(switch_worker, mbuf_pool, lcore_id); + } + + init_virtio_xen(&virtio_net_device_ops); + + virtio_monitor_loop(); + return 0; +} diff --git a/src/seastar/dpdk/examples/vhost_xen/main.h b/src/seastar/dpdk/examples/vhost_xen/main.h new file mode 100644 index 00000000..5ff48fd9 --- /dev/null +++ b/src/seastar/dpdk/examples/vhost_xen/main.h @@ -0,0 +1,66 @@ +/*- + * 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. + */ + +#ifndef _MAIN_H_ +#define _MAIN_H_ + +/* Macros for printing using RTE_LOG */ +#define RTE_LOGTYPE_VHOST_CONFIG RTE_LOGTYPE_USER1 +#define RTE_LOGTYPE_VHOST_DATA RTE_LOGTYPE_USER2 +#define RTE_LOGTYPE_VHOST_PORT RTE_LOGTYPE_USER3 + +/* + * Device linked list structure for data path. + */ +struct virtio_net_data_ll +{ + struct virtio_net *dev; /* Pointer to device created by configuration core. */ + struct virtio_net_data_ll *next; /* Pointer to next device in linked list. */ +}; + +/* + * Structure containing data core specific information. + */ +struct lcore_ll_info +{ + struct virtio_net_data_ll *ll_root_free; /* Pointer to head in free linked list. */ + struct virtio_net_data_ll *ll_root_used; /* Pointer to head of used linked list. */ + uint32_t device_num; /* Number of devices on lcore. */ + volatile uint8_t dev_removal_flag; /* Flag to synchronize device removal. */ +}; + +struct lcore_info +{ + struct lcore_ll_info *lcore_ll; /* Pointer to data core specific lcore_ll_info struct */ +}; +#endif /* _MAIN_H_ */ diff --git a/src/seastar/dpdk/examples/vhost_xen/vhost_monitor.c b/src/seastar/dpdk/examples/vhost_xen/vhost_monitor.c new file mode 100644 index 00000000..fb9606bf --- /dev/null +++ b/src/seastar/dpdk/examples/vhost_xen/vhost_monitor.c @@ -0,0 +1,595 @@ +/*- + * 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. + */ + +#include <stdlib.h> +#include <stdio.h> +#include <dirent.h> +#include <unistd.h> +#include <sys/eventfd.h> +#include <sys/ioctl.h> +#include <sys/mman.h> +#include <xen/xen-compat.h> +#if __XEN_LATEST_INTERFACE_VERSION__ < 0x00040200 +#include <xs.h> +#else +#include <xenstore.h> +#endif +#include <linux/virtio_ring.h> +#include <linux/virtio_pci.h> +#include <linux/virtio_net.h> + +#include <rte_ethdev.h> +#include <rte_log.h> +#include <rte_malloc.h> +#include <rte_string_fns.h> + +#include "virtio-net.h" +#include "xen_vhost.h" + +struct virtio_watch { + struct xs_handle *xs; + int watch_fd; +}; + + +/* device ops to add/remove device to/from data core. */ +static struct virtio_net_device_ops const *notify_ops; + +/* root address of the linked list in the configuration core. */ +static struct virtio_net_config_ll *ll_root = NULL; + +/* root address of VM. */ +static struct xen_guestlist guest_root; + +static struct virtio_watch watch; + +static void +vq_vring_init(struct vhost_virtqueue *vq, unsigned int num, uint8_t *p, + unsigned long align) +{ + vq->size = num; + vq->desc = (struct vring_desc *) p; + vq->avail = (struct vring_avail *) (p + + num * sizeof(struct vring_desc)); + vq->used = (void *) + RTE_ALIGN_CEIL( (uintptr_t)(&vq->avail->ring[num]), align); + +} + +static int +init_watch(void) +{ + struct xs_handle *xs; + int ret; + int fd; + + /* get a connection to the daemon */ + xs = xs_daemon_open(); + if (xs == NULL) { + RTE_LOG(ERR, XENHOST, "xs_daemon_open failed\n"); + return -1; + } + + ret = xs_watch(xs, "/local/domain", "mytoken"); + if (ret == 0) { + RTE_LOG(ERR, XENHOST, "%s: xs_watch failed\n", __func__); + xs_daemon_close(xs); + return -1; + } + + /* We are notified of read availability on the watch via the file descriptor. */ + fd = xs_fileno(xs); + watch.xs = xs; + watch.watch_fd = fd; + + TAILQ_INIT(&guest_root); + return 0; +} + +static struct xen_guest * +get_xen_guest(int dom_id) +{ + struct xen_guest *guest = NULL; + + TAILQ_FOREACH(guest, &guest_root, next) { + if(guest->dom_id == dom_id) + return guest; + } + + return NULL; +} + + +static struct xen_guest * +add_xen_guest(int32_t dom_id) +{ + struct xen_guest *guest = NULL; + + if ((guest = get_xen_guest(dom_id)) != NULL) + return guest; + + guest = calloc(1, sizeof(struct xen_guest)); + if (guest) { + RTE_LOG(ERR, XENHOST, " %s: return newly created guest with %d rings\n", __func__, guest->vring_num); + TAILQ_INSERT_TAIL(&guest_root, guest, next); + guest->dom_id = dom_id; + } + + return guest; +} + +static void +cleanup_device(struct virtio_net_config_ll *ll_dev) +{ + if (ll_dev == NULL) + return; + if (ll_dev->dev.virtqueue_rx) { + rte_free(ll_dev->dev.virtqueue_rx); + ll_dev->dev.virtqueue_rx = NULL; + } + if (ll_dev->dev.virtqueue_tx) { + rte_free(ll_dev->dev.virtqueue_tx); + ll_dev->dev.virtqueue_tx = NULL; + } + free(ll_dev); +} + +/* + * Add entry containing a device to the device configuration linked list. + */ +static void +add_config_ll_entry(struct virtio_net_config_ll *new_ll_dev) +{ + struct virtio_net_config_ll *ll_dev = ll_root; + + /* If ll_dev == NULL then this is the first device so go to else */ + if (ll_dev) { + /* If the 1st device_id != 0 then we insert our device here. */ + if (ll_dev->dev.device_fh != 0) { + new_ll_dev->dev.device_fh = 0; + new_ll_dev->next = ll_dev; + ll_root = new_ll_dev; + } else { + /* increment through the ll until we find un unused device_id, + * insert the device at that entry + */ + while ((ll_dev->next != NULL) && (ll_dev->dev.device_fh == (ll_dev->next->dev.device_fh - 1))) + ll_dev = ll_dev->next; + + new_ll_dev->dev.device_fh = ll_dev->dev.device_fh + 1; + new_ll_dev->next = ll_dev->next; + ll_dev->next = new_ll_dev; + } + } else { + ll_root = new_ll_dev; + ll_root->dev.device_fh = 0; + } +} + + +/* + * Remove an entry from the device configuration linked list. + */ +static struct virtio_net_config_ll * +rm_config_ll_entry(struct virtio_net_config_ll *ll_dev, struct virtio_net_config_ll *ll_dev_last) +{ + /* First remove the device and then clean it up. */ + if (ll_dev == ll_root) { + ll_root = ll_dev->next; + cleanup_device(ll_dev); + return ll_root; + } else { + ll_dev_last->next = ll_dev->next; + cleanup_device(ll_dev); + return ll_dev_last->next; + } +} + +/* + * Retrieves an entry from the devices configuration linked list. + */ +static struct virtio_net_config_ll * +get_config_ll_entry(unsigned int virtio_idx, unsigned int dom_id) +{ + struct virtio_net_config_ll *ll_dev = ll_root; + + /* Loop through linked list until the dom_id is found. */ + while (ll_dev != NULL) { + if (ll_dev->dev.dom_id == dom_id && ll_dev->dev.virtio_idx == virtio_idx) + return ll_dev; + ll_dev = ll_dev->next; + } + + return NULL; +} + +/* + * Initialise all variables in device structure. + */ +static void +init_dev(struct virtio_net *dev) +{ + RTE_SET_USED(dev); +} + + +static struct +virtio_net_config_ll *new_device(unsigned int virtio_idx, struct xen_guest *guest) +{ + struct virtio_net_config_ll *new_ll_dev; + struct vhost_virtqueue *virtqueue_rx, *virtqueue_tx; + size_t size, vq_ring_size, vq_size = VQ_DESC_NUM; + void *vq_ring_virt_mem; + uint64_t gpa; + uint32_t i; + + /* Setup device and virtqueues. */ + new_ll_dev = calloc(1, sizeof(struct virtio_net_config_ll)); + virtqueue_rx = rte_zmalloc(NULL, sizeof(struct vhost_virtqueue), RTE_CACHE_LINE_SIZE); + virtqueue_tx = rte_zmalloc(NULL, sizeof(struct vhost_virtqueue), RTE_CACHE_LINE_SIZE); + if (new_ll_dev == NULL || virtqueue_rx == NULL || virtqueue_tx == NULL) + goto err; + + new_ll_dev->dev.virtqueue_rx = virtqueue_rx; + new_ll_dev->dev.virtqueue_tx = virtqueue_tx; + new_ll_dev->dev.dom_id = guest->dom_id; + new_ll_dev->dev.virtio_idx = virtio_idx; + /* Initialise device and virtqueues. */ + init_dev(&new_ll_dev->dev); + + size = vring_size(vq_size, VIRTIO_PCI_VRING_ALIGN); + vq_ring_size = RTE_ALIGN_CEIL(size, VIRTIO_PCI_VRING_ALIGN); + (void)vq_ring_size; + + vq_ring_virt_mem = guest->vring[virtio_idx].rxvring_addr; + vq_vring_init(virtqueue_rx, vq_size, vq_ring_virt_mem, VIRTIO_PCI_VRING_ALIGN); + virtqueue_rx->size = vq_size; + virtqueue_rx->vhost_hlen = sizeof(struct virtio_net_hdr); + + vq_ring_virt_mem = guest->vring[virtio_idx].txvring_addr; + vq_vring_init(virtqueue_tx, vq_size, vq_ring_virt_mem, VIRTIO_PCI_VRING_ALIGN); + virtqueue_tx->size = vq_size; + memcpy(&new_ll_dev->dev.mac_address, &guest->vring[virtio_idx].addr, sizeof(struct ether_addr)); + + /* virtio_memory has to be one per domid */ + new_ll_dev->dev.mem = malloc(sizeof(struct virtio_memory) + sizeof(struct virtio_memory_regions) * MAX_XENVIRT_MEMPOOL); + new_ll_dev->dev.mem->nregions = guest->pool_num; + for (i = 0; i < guest->pool_num; i++) { + gpa = new_ll_dev->dev.mem->regions[i].guest_phys_address = + (uint64_t)((uintptr_t)guest->mempool[i].gva); + new_ll_dev->dev.mem->regions[i].guest_phys_address_end = + gpa + guest->mempool[i].mempfn_num * getpagesize(); + new_ll_dev->dev.mem->regions[i].address_offset = + (uint64_t)((uintptr_t)guest->mempool[i].hva - + (uintptr_t)gpa); + } + + new_ll_dev->next = NULL; + + /* Add entry to device configuration linked list. */ + add_config_ll_entry(new_ll_dev); + return new_ll_dev; +err: + free(new_ll_dev); + rte_free(virtqueue_rx); + rte_free(virtqueue_tx); + + return NULL; +} + +static void +destroy_guest(struct xen_guest *guest) +{ + uint32_t i; + + for (i = 0; i < guest->vring_num; i++) + cleanup_vring(&guest->vring[i]); + /* clean mempool */ + for (i = 0; i < guest->pool_num; i++) + cleanup_mempool(&guest->mempool[i]); + free(guest); + + return; +} + +/* + * This function will cleanup the device and remove it from device configuration linked list. + */ +static void +destroy_device(unsigned int virtio_idx, unsigned int dom_id) +{ + struct virtio_net_config_ll *ll_dev_cur_ctx, *ll_dev_last = NULL; + struct virtio_net_config_ll *ll_dev_cur = ll_root; + + /* clean virtio device */ + struct xen_guest *guest = NULL; + guest = get_xen_guest(dom_id); + if (guest == NULL) + return; + + /* Find the linked list entry for the device to be removed. */ + ll_dev_cur_ctx = get_config_ll_entry(virtio_idx, dom_id); + while (ll_dev_cur != NULL) { + /* If the device is found or a device that doesn't exist is found then it is removed. */ + if (ll_dev_cur == ll_dev_cur_ctx) { + if ((ll_dev_cur->dev.flags & VIRTIO_DEV_RUNNING)) + notify_ops->destroy_device(&(ll_dev_cur->dev)); + ll_dev_cur = rm_config_ll_entry(ll_dev_cur, ll_dev_last); + } else { + ll_dev_last = ll_dev_cur; + ll_dev_cur = ll_dev_cur->next; + } + } + RTE_LOG(INFO, XENHOST, " %s guest:%p vring:%p rxvring:%p txvring:%p flag:%p\n", + __func__, guest, &guest->vring[virtio_idx], guest->vring[virtio_idx].rxvring_addr, guest->vring[virtio_idx].txvring_addr, guest->vring[virtio_idx].flag); + cleanup_vring(&guest->vring[virtio_idx]); + guest->vring[virtio_idx].removed = 1; + guest->vring_num -= 1; +} + + + + +static void +watch_unmap_event(void) +{ + int i; + struct xen_guest *guest = NULL; + bool remove_request; + + TAILQ_FOREACH(guest, &guest_root, next) { + for (i = 0; i < MAX_VIRTIO; i++) { + if (guest->vring[i].dom_id && guest->vring[i].removed == 0 && *guest->vring[i].flag == 0) { + RTE_LOG(INFO, XENHOST, "\n\n"); + RTE_LOG(INFO, XENHOST, " #####%s: (%d, %d) to be removed\n", + __func__, + guest->vring[i].dom_id, + i); + destroy_device(i, guest->dom_id); + RTE_LOG(INFO, XENHOST, " %s: DOM %u, vring num: %d\n", + __func__, + guest->dom_id, + guest->vring_num); + } + } + } + +_find_next_remove: + guest = NULL; + remove_request = false; + TAILQ_FOREACH(guest, &guest_root, next) { + if (guest->vring_num == 0) { + remove_request = true; + break; + } + } + if (remove_request == true) { + TAILQ_REMOVE(&guest_root, guest, next); + RTE_LOG(INFO, XENHOST, " #####%s: destroy guest (%d)\n", __func__, guest->dom_id); + destroy_guest(guest); + goto _find_next_remove; + } + return; +} + +/* + * OK, if the guest starts first, it is ok. + * if host starts first, it is ok. + * if guest starts, and has run for sometime, and host stops and restarts, + * then last_used_idx 0? how to solve this. */ + +static void virtio_init(void) +{ + uint32_t len, e_num; + uint32_t i,j; + char **dom; + char *status; + int dom_id; + char path[PATH_MAX]; + char node[PATH_MAX]; + xs_transaction_t th; + struct xen_guest *guest; + struct virtio_net_config_ll *net_config; + char *end; + int val; + + /* init env for watch the node */ + if (init_watch() < 0) + return; + + dom = xs_directory(watch.xs, XBT_NULL, "/local/domain", &e_num); + + for (i = 0; i < e_num; i++) { + errno = 0; + dom_id = strtol(dom[i], &end, 0); + if (errno != 0 || end == NULL || dom_id == 0) + continue; + + for (j = 0; j < RTE_MAX_ETHPORTS; j++) { + snprintf(node, PATH_MAX, "%s%d", VIRTIO_START, j); + snprintf(path, PATH_MAX, XEN_VM_NODE_FMT, + dom_id, node); + + th = xs_transaction_start(watch.xs); + status = xs_read(watch.xs, th, path, &len); + xs_transaction_end(watch.xs, th, false); + + if (status == NULL) + break; + + /* if there's any valid virtio device */ + errno = 0; + val = strtol(status, &end, 0); + if (errno != 0 || end == NULL || dom_id == 0) + val = 0; + if (val == 1) { + guest = add_xen_guest(dom_id); + if (guest == NULL) + continue; + RTE_LOG(INFO, XENHOST, " there's a new virtio existed, new a virtio device\n\n"); + + RTE_LOG(INFO, XENHOST, " parse_vringnode dom_id %d virtioidx %d\n",dom_id,j); + if (parse_vringnode(guest, j)) { + RTE_LOG(ERR, XENHOST, " there is invalid information in xenstore\n"); + TAILQ_REMOVE(&guest_root, guest, next); + destroy_guest(guest); + + continue; + } + + /*if pool_num > 0, then mempool has already been parsed*/ + if (guest->pool_num == 0 && parse_mempoolnode(guest)) { + RTE_LOG(ERR, XENHOST, " there is error information in xenstore\n"); + TAILQ_REMOVE(&guest_root, guest, next); + destroy_guest(guest); + continue; + } + + net_config = new_device(j, guest); + /* every thing is ready now, added into data core */ + notify_ops->new_device(&net_config->dev); + } + } + } + + free(dom); + return; +} + +void +virtio_monitor_loop(void) +{ + char **vec; + xs_transaction_t th; + char *buf; + unsigned int len; + unsigned int dom_id; + uint32_t virtio_idx; + struct xen_guest *guest; + struct virtio_net_config_ll *net_config; + enum fieldnames { + FLD_NULL = 0, + FLD_LOCAL, + FLD_DOMAIN, + FLD_ID, + FLD_CONTROL, + FLD_DPDK, + FLD_NODE, + _NUM_FLD + }; + char *str_fld[_NUM_FLD]; + char *str; + char *end; + + virtio_init(); + while (1) { + watch_unmap_event(); + + usleep(50); + vec = xs_check_watch(watch.xs); + + if (vec == NULL) + continue; + + th = xs_transaction_start(watch.xs); + + buf = xs_read(watch.xs, th, vec[XS_WATCH_PATH],&len); + xs_transaction_end(watch.xs, th, false); + + if (buf) { + /* theres' some node for vhost existed */ + if (rte_strsplit(vec[XS_WATCH_PATH], strnlen(vec[XS_WATCH_PATH], PATH_MAX), + str_fld, _NUM_FLD, '/') == _NUM_FLD) { + if (strstr(str_fld[FLD_NODE], VIRTIO_START)) { + errno = 0; + str = str_fld[FLD_ID]; + dom_id = strtoul(str, &end, 0); + if (errno != 0 || end == NULL || end == str ) { + RTE_LOG(INFO, XENHOST, "invalid domain id\n"); + continue; + } + + errno = 0; + str = str_fld[FLD_NODE] + sizeof(VIRTIO_START) - 1; + virtio_idx = strtoul(str, &end, 0); + if (errno != 0 || end == NULL || end == str + || virtio_idx > MAX_VIRTIO) { + RTE_LOG(INFO, XENHOST, "invalid virtio idx\n"); + continue; + } + RTE_LOG(INFO, XENHOST, " #####virtio dev (%d, %d) is started\n", dom_id, virtio_idx); + + guest = add_xen_guest(dom_id); + if (guest == NULL) + continue; + guest->dom_id = dom_id; + if (parse_vringnode(guest, virtio_idx)) { + RTE_LOG(ERR, XENHOST, " there is invalid information in xenstore\n"); + /*guest newly created? guest existed ?*/ + TAILQ_REMOVE(&guest_root, guest, next); + destroy_guest(guest); + continue; + } + /*if pool_num > 0, then mempool has already been parsed*/ + if (guest->pool_num == 0 && parse_mempoolnode(guest)) { + RTE_LOG(ERR, XENHOST, " there is error information in xenstore\n"); + TAILQ_REMOVE(&guest_root, guest, next); + destroy_guest(guest); + continue; + } + + + net_config = new_device(virtio_idx, guest); + RTE_LOG(INFO, XENHOST, " Add to dataplane core\n"); + notify_ops->new_device(&net_config->dev); + + } + } + } + + free(vec); + } + return; +} + +/* + * Register ops so that we can add/remove device to data core. + */ +int +init_virtio_xen(struct virtio_net_device_ops const *const ops) +{ + notify_ops = ops; + if (xenhost_init()) + return -1; + return 0; +} diff --git a/src/seastar/dpdk/examples/vhost_xen/virtio-net.h b/src/seastar/dpdk/examples/vhost_xen/virtio-net.h new file mode 100644 index 00000000..ab697260 --- /dev/null +++ b/src/seastar/dpdk/examples/vhost_xen/virtio-net.h @@ -0,0 +1,113 @@ +/*- + * 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. + */ + +#ifndef _VIRTIO_NET_H_ +#define _VIRTIO_NET_H_ + +#include <stdint.h> + +#define VQ_DESC_NUM 256 +/* Used to indicate that the device is running on a data core */ +#define VIRTIO_DEV_RUNNING 1 + +/* + * Structure contains variables relevant to TX/RX virtqueues. + */ +struct vhost_virtqueue +{ + struct vring_desc *desc; /* Virtqueue descriptor ring. */ + struct vring_avail *avail; /* Virtqueue available ring. */ + struct vring_used *used; /* Virtqueue used ring. */ + uint32_t size; /* Size of descriptor ring. */ + uint32_t vhost_hlen; /* Vhost header length (varies depending on RX merge buffers. */ + volatile uint16_t last_used_idx; /* Last index used on the available ring */ + volatile uint16_t last_used_idx_res; /* Used for multiple devices reserving buffers. */ +} __rte_cache_aligned; + +/* + * Device structure contains all configuration information relating to the device. + */ +struct virtio_net +{ + struct vhost_virtqueue *virtqueue_tx; /* Contains all TX virtqueue information. */ + struct vhost_virtqueue *virtqueue_rx; /* Contains all RX virtqueue information. */ + struct virtio_memory *mem; /* QEMU memory and memory region information. */ + struct ether_addr mac_address; /* Device MAC address (Obtained on first TX packet). */ + uint32_t flags; /* Device flags. Only used to check if device is running on data core. */ + uint32_t vlan_tag; /* Vlan tag for device. Currently set to device_id (0-63). */ + uint32_t vmdq_rx_q; + uint64_t device_fh; /* device identifier. */ + uint16_t coreid; + volatile uint8_t ready; /* A device is set as ready if the MAC address has been set. */ + volatile uint8_t remove; /* Device is marked for removal from the data core. */ + uint32_t virtio_idx; /* Index of virtio device */ + uint32_t dom_id; /* Domain id of xen guest */ +} ___rte_cache_aligned; + +/* + * Device linked list structure for configuration. + */ +struct virtio_net_config_ll +{ + struct virtio_net dev; /* Virtio device. */ + struct virtio_net_config_ll *next; /* Next entry on linked list. */ +}; + +/* + * Information relating to memory regions including offsets to addresses in QEMUs memory file. + */ +struct virtio_memory_regions { + uint64_t guest_phys_address; /* Base guest physical address of region. */ + uint64_t guest_phys_address_end; /* End guest physical address of region. */ + uint64_t memory_size; /* Size of region. */ + uint64_t userspace_address; /* Base userspace address of region. */ + uint64_t address_offset; /* Offset of region for address translation. */ +}; + +/* + * Memory structure includes region and mapping information. + */ +struct virtio_memory { + uint32_t nregions; /* Number of memory regions. */ + struct virtio_memory_regions regions[0]; /* Memory region information. */ +}; + +/* + * Device operations to add/remove device. + */ +struct virtio_net_device_ops { + int (* new_device)(struct virtio_net *); /* Add device. */ + void (* destroy_device) (volatile struct virtio_net *); /* Remove device. */ +}; + +#endif diff --git a/src/seastar/dpdk/examples/vhost_xen/xen_vhost.h b/src/seastar/dpdk/examples/vhost_xen/xen_vhost.h new file mode 100644 index 00000000..2fc304c7 --- /dev/null +++ b/src/seastar/dpdk/examples/vhost_xen/xen_vhost.h @@ -0,0 +1,148 @@ +/*- + * 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. + */ + +#ifndef _XEN_VHOST_H_ +#define _XEN_VHOST_H_ + +#include <stdint.h> + +#include <rte_ether.h> + +#include "virtio-net.h" + +#define RTE_LOGTYPE_XENHOST RTE_LOGTYPE_USER1 + +#define XEN_VM_ROOTNODE_FMT "/local/domain/%d/control/dpdk" +#define XEN_VM_NODE_FMT "/local/domain/%d/control/dpdk/%s" +#define XEN_MEMPOOL_SUFFIX "mempool_gref" +#define XEN_RXVRING_SUFFIX "rx_vring_gref" +#define XEN_TXVRING_SUFFIX "tx_vring_gref" +#define XEN_GVA_SUFFIX "mempool_va" +#define XEN_VRINGFLAG_SUFFIX "vring_flag" +#define XEN_ADDR_SUFFIX "ether_addr" +#define VIRTIO_START "event_type_start_" + +#define XEN_GREF_SPLITTOKEN ',' + +#define MAX_XENVIRT_MEMPOOL 16 +#define MAX_VIRTIO 32 +#define MAX_GREF_PER_NODE 64 /* 128 MB memory */ + +#define PAGE_SIZE 4096 +#define PAGE_PFNNUM (PAGE_SIZE / sizeof(uint32_t)) + +#define XEN_GNTDEV_FNAME "/dev/xen/gntdev" + +/* xen grant reference info in one grant node */ +struct xen_gnt { + uint32_t gref; /* grant reference for this node */ + union { + int gref; /* grant reference */ + uint32_t pfn_num; /* guest pfn number of grant reference */ + } gref_pfn[PAGE_PFNNUM]; +}__attribute__((__packed__)); + + +/* structure for mempool or vring node list */ +struct xen_gntnode { + uint32_t gnt_num; /* grant reference number */ + struct xen_gnt *gnt_info; /* grant reference info */ +}; + + +struct xen_vring { + uint32_t dom_id; + uint32_t virtio_idx; /* index of virtio device */ + void *rxvring_addr; /* mapped virtual address of rxvring */ + void *txvring_addr; /* mapped virtual address of txvring */ + uint32_t rxpfn_num; /* number of gpfn for rxvring */ + uint32_t txpfn_num; /* number of gpfn for txvring */ + uint32_t *rxpfn_tbl; /* array of rxvring gpfn */ + uint32_t *txpfn_tbl; /* array of txvring gpfn */ + uint64_t *rx_pindex; /* index used to release rx grefs */ + uint64_t *tx_pindex; /* index used to release tx grefs */ + uint64_t flag_index; + uint8_t *flag; /* cleared to zero on guest unmap */ + struct ether_addr addr; /* ethernet address of virtio device */ + uint8_t removed; + +}; + +struct xen_mempool { + uint32_t dom_id; /* guest domain id */ + uint32_t pool_idx; /* index of memory pool */ + void *gva; /* guest virtual address of mbuf pool */ + void *hva; /* host virtual address of mbuf pool */ + uint32_t mempfn_num; /* number of gpfn for mbuf pool */ + uint32_t *mempfn_tbl; /* array of mbuf pool gpfn */ + uint64_t *pindex; /* index used to release grefs */ +}; + +struct xen_guest { + TAILQ_ENTRY(xen_guest) next; + int32_t dom_id; /* guest domain id */ + uint32_t pool_num; /* number of mbuf pool of the guest */ + uint32_t vring_num; /* number of virtio ports of the guest */ + /* array contain the guest mbuf pool info */ + struct xen_mempool mempool[MAX_XENVIRT_MEMPOOL]; + /* array contain the guest rx/tx vring info */ + struct xen_vring vring[MAX_VIRTIO]; +}; + +TAILQ_HEAD(xen_guestlist, xen_guest); + +int +parse_mempoolnode(struct xen_guest *guest); + +int +xenhost_init(void); + +int +parse_vringnode(struct xen_guest *guest, uint32_t virtio_idx); + +int +parse_mempoolnode(struct xen_guest *guest); + +void +cleanup_mempool(struct xen_mempool *mempool); + +void +cleanup_vring(struct xen_vring *vring); + +void +virtio_monitor_loop(void); + +int +init_virtio_xen(struct virtio_net_device_ops const * const); + +#endif diff --git a/src/seastar/dpdk/examples/vhost_xen/xenstore_parse.c b/src/seastar/dpdk/examples/vhost_xen/xenstore_parse.c new file mode 100644 index 00000000..26d24320 --- /dev/null +++ b/src/seastar/dpdk/examples/vhost_xen/xenstore_parse.c @@ -0,0 +1,775 @@ +/*- + * 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. + */ + +#include <stdint.h> +#include <unistd.h> +#include <inttypes.h> +#include <errno.h> +#include <fcntl.h> +#include <sys/ioctl.h> +#include <sys/mman.h> +#include <xen/sys/gntalloc.h> +#include <xen/sys/gntdev.h> +#include <xen/xen-compat.h> +#if __XEN_LATEST_INTERFACE_VERSION__ < 0x00040200 +#include <xs.h> +#else +#include <xenstore.h> +#endif + +#include <rte_common.h> +#include <rte_memory.h> +#include <rte_eal.h> +#include <rte_malloc.h> +#include <rte_string_fns.h> +#include <rte_log.h> +#include <rte_debug.h> + +#include "xen_vhost.h" + +/* xenstore handle */ +static struct xs_handle *xs = NULL; + +/* gntdev file descriptor to map grant pages */ +static int d_fd = -1; + +/* + * The grant node format in xenstore for vring/mpool is like: + * idx#_rx_vring_gref = "gref1#, gref2#, gref3#" + * idx#_mempool_gref = "gref1#, gref2#, gref3#" + * each gref# is the grant reference for a shared page. + * In each shared page, we store the grant_node_item items. + */ +struct grant_node_item { + uint32_t gref; + uint32_t pfn; +} __attribute__((packed)); + +int cmdline_parse_etheraddr(void *tk, const char *srcbuf, + void *res, unsigned ressize); + +/* Map grant ref refid at addr_ori*/ +static void * +xen_grant_mmap(void *addr_ori, int domid, int refid, uint64_t *pindex) +{ + struct ioctl_gntdev_map_grant_ref arg; + void *addr = NULL; + int pg_sz = getpagesize(); + + arg.count = 1; + arg.refs[0].domid = domid; + arg.refs[0].ref = refid; + + int rv = ioctl(d_fd, IOCTL_GNTDEV_MAP_GRANT_REF, &arg); + if (rv) { + RTE_LOG(ERR, XENHOST, " %s: (%d,%d) %s (ioctl failed)\n", __func__, + domid, refid, strerror(errno)); + return NULL; + } + + if (addr_ori == NULL) + addr = mmap(addr_ori, pg_sz, PROT_READ|PROT_WRITE, MAP_SHARED, + d_fd, arg.index); + else + addr = mmap(addr_ori, pg_sz, PROT_READ|PROT_WRITE, MAP_SHARED | MAP_FIXED, + d_fd, arg.index); + + if (addr == MAP_FAILED) { + RTE_LOG(ERR, XENHOST, " %s: (%d, %d) %s (map failed)\n", __func__, + domid, refid, strerror(errno)); + return NULL; + } + + if (pindex) + *pindex = arg.index; + + return addr; +} + +/* Unmap one grant ref, and munmap must be called before this */ +static int +xen_unmap_grant_ref(uint64_t index) +{ + struct ioctl_gntdev_unmap_grant_ref arg; + int rv; + + arg.count = 1; + arg.index = index; + rv = ioctl(d_fd, IOCTL_GNTDEV_UNMAP_GRANT_REF, &arg); + if (rv) { + RTE_LOG(ERR, XENHOST, " %s: index 0x%" PRIx64 "unmap failed\n", __func__, index); + return -1; + } + return 0; +} + +/* + * Reserve a virtual address space. + * On success, returns the pointer. On failure, returns NULL. + */ +static void * +get_xen_virtual(size_t size, size_t page_sz) +{ + void *addr; + uintptr_t aligned_addr; + + addr = mmap(NULL, size + page_sz, PROT_READ, MAP_SHARED | MAP_ANONYMOUS, -1, 0); + if (addr == MAP_FAILED) { + RTE_LOG(ERR, XENHOST, "failed get a virtual area\n"); + return NULL; + } + + aligned_addr = RTE_ALIGN_CEIL((uintptr_t)addr, page_sz); + munmap(addr, aligned_addr - (uintptr_t)addr); + munmap((void *)(aligned_addr + size), page_sz + (uintptr_t)addr - aligned_addr); + addr = (void *)(aligned_addr); + + return addr; +} + +static void +free_xen_virtual(void *addr, size_t size, size_t page_sz __rte_unused) +{ + if (addr) + munmap(addr, size); +} + +/* + * Returns val str in xenstore. + * @param path + * Full path string for key + * @return + * Pointer to Val str, NULL on failure + */ +static char * +xen_read_node(char *path, uint32_t *len) +{ + char *buf; + + buf = xs_read(xs, XBT_NULL, path, len); + return buf; +} + +static int +cal_pagenum(struct xen_gnt *gnt) +{ + unsigned int i; + /* + * the items in the page are in the format of + * gref#,pfn#,...,gref#,pfn# + * FIXME, 0 is reserved by system, use it as terminator. + */ + for (i = 0; i < (PAGE_PFNNUM) / 2; i++) { + if (gnt->gref_pfn[i * 2].gref <= 0) + break; + } + + return i; +} + +/* Frees memory allocated to a grant node */ +static void +xen_free_gntnode(struct xen_gntnode *gntnode) +{ + if (gntnode == NULL) + return; + free(gntnode->gnt_info); + free(gntnode); +} + +/* + * Parse a grant node. + * @param domid + * Guest domain id. + * @param path + * Full path string for a grant node, like for the following (key, val) pair + * idx#_mempool_gref = "gref#, gref#, gref#" + * path = 'local/domain/domid/control/dpdk/idx#_mempool_gref' + * gref# is a shared page contain packed (gref,pfn) entries + * @return + * Returns the pointer to xen_gntnode + */ +static struct xen_gntnode * +parse_gntnode(int dom_id, char *path) +{ + char **gref_list = NULL; + uint32_t i, len, gref_num; + void *addr = NULL; + char *buf = NULL; + struct xen_gntnode *gntnode = NULL; + struct xen_gnt *gnt = NULL; + int pg_sz = getpagesize(); + char *end; + uint64_t index; + + if ((buf = xen_read_node(path, &len)) == NULL) + goto err; + + gref_list = malloc(MAX_GREF_PER_NODE * sizeof(char *)); + if (gref_list == NULL) + goto err; + + gref_num = rte_strsplit(buf, len, gref_list, MAX_GREF_PER_NODE, + XEN_GREF_SPLITTOKEN); + if (gref_num == 0) { + RTE_LOG(ERR, XENHOST, " %s: invalid grant node format\n", __func__); + goto err; + } + + gntnode = calloc(1, sizeof(struct xen_gntnode)); + gnt = calloc(gref_num, sizeof(struct xen_gnt)); + if (gnt == NULL || gntnode == NULL) + goto err; + + for (i = 0; i < gref_num; i++) { + errno = 0; + gnt[i].gref = strtol(gref_list[i], &end, 0); + if (errno != 0 || end == NULL || end == gref_list[i] || + (*end != '\0' && *end != XEN_GREF_SPLITTOKEN)) { + RTE_LOG(ERR, XENHOST, " %s: parse grant node item failed\n", __func__); + goto err; + } + addr = xen_grant_mmap(NULL, dom_id, gnt[i].gref, &index); + if (addr == NULL) { + RTE_LOG(ERR, XENHOST, " %s: map gref %u failed\n", __func__, gnt[i].gref); + goto err; + } + RTE_LOG(INFO, XENHOST, " %s: map gref %u to %p\n", __func__, gnt[i].gref, addr); + memcpy(gnt[i].gref_pfn, addr, pg_sz); + if (munmap(addr, pg_sz)) { + RTE_LOG(INFO, XENHOST, " %s: unmap gref %u failed\n", __func__, gnt[i].gref); + goto err; + } + if (xen_unmap_grant_ref(index)) { + RTE_LOG(INFO, XENHOST, " %s: release gref %u failed\n", __func__, gnt[i].gref); + goto err; + } + + } + + gntnode->gnt_num = gref_num; + gntnode->gnt_info = gnt; + + free(buf); + free(gref_list); + return gntnode; + +err: + free(gnt); + free(gntnode); + free(gref_list); + free(buf); + return NULL; +} + +/* + * This function maps grant node of vring or mbuf pool to a continous virtual address space, + * and returns mapped address, pfn array, index array + * @param gntnode + * Pointer to grant node + * @param domid + * Guest domain id + * @param ppfn + * Pointer to pfn array, caller should free this array + * @param pgs + * Pointer to number of pages + * @param ppindex + * Pointer to index array, used to release grefs when to free this node + * @return + * Pointer to mapped virtual address, NULL on failure + */ +static void * +map_gntnode(struct xen_gntnode *gntnode, int domid, uint32_t **ppfn, uint32_t *pgs, uint64_t **ppindex) +{ + struct xen_gnt *gnt; + uint32_t i, j; + size_t total_pages = 0; + void *addr; + uint32_t *pfn; + uint64_t *pindex; + uint32_t pfn_num = 0; + int pg_sz; + + if (gntnode == NULL) + return NULL; + + pg_sz = getpagesize(); + for (i = 0; i < gntnode->gnt_num; i++) { + gnt = gntnode->gnt_info + i; + total_pages += cal_pagenum(gnt); + } + if ((addr = get_xen_virtual(total_pages * pg_sz, pg_sz)) == NULL) { + RTE_LOG(ERR, XENHOST, " %s: failed get_xen_virtual\n", __func__); + return NULL; + } + pfn = calloc(total_pages, (size_t)sizeof(uint32_t)); + pindex = calloc(total_pages, (size_t)sizeof(uint64_t)); + if (pfn == NULL || pindex == NULL) { + free_xen_virtual(addr, total_pages * pg_sz, pg_sz); + free(pfn); + free(pindex); + return NULL; + } + + RTE_LOG(INFO, XENHOST, " %s: total pages:%zu, map to [%p, %p]\n", __func__, total_pages, addr, RTE_PTR_ADD(addr, total_pages * pg_sz - 1)); + for (i = 0; i < gntnode->gnt_num; i++) { + gnt = gntnode->gnt_info + i; + for (j = 0; j < (PAGE_PFNNUM) / 2; j++) { + if ((gnt->gref_pfn[j * 2].gref) <= 0) + goto _end; + /*alternative: batch map, or through libxc*/ + if (xen_grant_mmap(RTE_PTR_ADD(addr, pfn_num * pg_sz), + domid, + gnt->gref_pfn[j * 2].gref, + &pindex[pfn_num]) == NULL) { + goto mmap_failed; + } + pfn[pfn_num] = gnt->gref_pfn[j * 2 + 1].pfn_num; + pfn_num++; + } + } + +mmap_failed: + if (pfn_num) + munmap(addr, pfn_num * pg_sz); + for (i = 0; i < pfn_num; i++) { + xen_unmap_grant_ref(pindex[i]); + } + free(pindex); + free(pfn); + return NULL; + +_end: + if (ppindex) + *ppindex = pindex; + else + free(pindex); + if (ppfn) + *ppfn = pfn; + else + free(pfn); + if (pgs) + *pgs = total_pages; + + return addr; +} + +static int +parse_mpool_va(struct xen_mempool *mempool) +{ + char path[PATH_MAX] = {0}; + char *buf; + uint32_t len; + char *end; + int ret = -1; + + errno = 0; + snprintf(path, sizeof(path), + XEN_VM_ROOTNODE_FMT"/%d_"XEN_GVA_SUFFIX, + mempool->dom_id, mempool->pool_idx); + + if((buf = xen_read_node(path, &len)) == NULL) + goto out; + mempool->gva = (void *)strtoul(buf, &end, 16); + if (errno != 0 || end == NULL || end == buf || *end != '\0') { + mempool->gva = NULL; + goto out; + } + ret = 0; +out: + free(buf); + return ret; +} + +/* + * map mbuf pool + */ +static int +map_mempoolnode(struct xen_gntnode *gntnode, + struct xen_mempool *mempool) +{ + if (gntnode == NULL || mempool == NULL) + return -1; + + mempool->hva = + map_gntnode(gntnode, mempool->dom_id, &mempool->mempfn_tbl, &mempool->mempfn_num, &mempool->pindex); + + RTE_LOG(INFO, XENHOST, " %s: map mempool at %p\n", __func__, (void *)mempool->hva); + if (mempool->hva) + return 0; + else { + return -1; + } +} + +void +cleanup_mempool(struct xen_mempool *mempool) +{ + int pg_sz = getpagesize(); + uint32_t i; + + if (mempool->hva) + munmap(mempool->hva, mempool->mempfn_num * pg_sz); + mempool->hva = NULL; + + if (mempool->pindex) { + RTE_LOG(INFO, XENHOST, " %s: unmap dom %02u mempool%02u %u grefs\n", + __func__, + mempool->dom_id, + mempool->pool_idx, + mempool->mempfn_num); + for (i = 0; i < mempool->mempfn_num; i ++) { + xen_unmap_grant_ref(mempool->pindex[i]); + } + } + mempool->pindex = NULL; + + free(mempool->mempfn_tbl); + mempool->mempfn_tbl = NULL; +} + +/* + * process mempool node idx#_mempool_gref, idx = 0, 1, 2... + * untill we encounter a node that doesn't exist. + */ +int +parse_mempoolnode(struct xen_guest *guest) +{ + uint32_t i, len; + char path[PATH_MAX] = {0}; + struct xen_gntnode *gntnode = NULL; + struct xen_mempool *mempool = NULL; + char *buf; + + bzero(&guest->mempool, MAX_XENVIRT_MEMPOOL * sizeof(guest->mempool[0])); + guest->pool_num = 0; + + while (1) { + /* check if null terminated */ + snprintf(path, sizeof(path), + XEN_VM_ROOTNODE_FMT"/%d_"XEN_MEMPOOL_SUFFIX, + guest->dom_id, + guest->pool_num); + + if ((buf = xen_read_node(path, &len)) != NULL) { + /* this node exists */ + free(buf); + } else { + if (guest->pool_num == 0) { + RTE_LOG(ERR, PMD, "no mempool found\n"); + return -1; + } + break; + } + + mempool = &guest->mempool[guest->pool_num]; + mempool->dom_id = guest->dom_id; + mempool->pool_idx = guest->pool_num; + + RTE_LOG(INFO, XENHOST, " %s: mempool %u parse gntnode %s\n", __func__, guest->pool_num, path); + gntnode = parse_gntnode(guest->dom_id, path); + if (gntnode == NULL) + goto err; + + if (parse_mpool_va(mempool)) + goto err; + + RTE_LOG(INFO, XENHOST, " %s: mempool %u map gntnode %s\n", __func__, guest->pool_num, path); + if (map_mempoolnode(gntnode, mempool)) + goto err; + + xen_free_gntnode(gntnode); + guest->pool_num++; + } + + return 0; +err: + if (gntnode) + xen_free_gntnode(gntnode); + for (i = 0; i < MAX_XENVIRT_MEMPOOL ; i++) { + cleanup_mempool(&guest->mempool[i]); + } + /* reinitialise mempool */ + bzero(&guest->mempool, MAX_XENVIRT_MEMPOOL * sizeof(guest->mempool[0])); + return -1; +} + +static int +xen_map_vringflag(struct xen_vring *vring) +{ + char path[PATH_MAX] = {0}; + char *buf; + uint32_t len,gref; + int pg_sz = getpagesize(); + char *end; + + snprintf(path, sizeof(path), + XEN_VM_ROOTNODE_FMT"/%d_"XEN_VRINGFLAG_SUFFIX, + vring->dom_id, vring->virtio_idx); + + if((buf = xen_read_node(path, &len)) == NULL) + goto err; + + errno = 0; + gref = strtol(buf, &end, 0); + if (errno != 0 || end == NULL || end == buf) { + goto err; + } + vring->flag = xen_grant_mmap(0, vring->dom_id, gref, &vring->flag_index); + if (vring->flag == NULL || *vring->flag == 0) + goto err; + + free(buf); + return 0; +err: + free(buf); + if (vring->flag) { + munmap(vring->flag, pg_sz); + vring->flag = NULL; + xen_unmap_grant_ref(vring->flag_index); + } + return -1; +} + + +static int +xen_map_rxvringnode(struct xen_gntnode *gntnode, + struct xen_vring *vring) +{ + vring->rxvring_addr = + map_gntnode(gntnode, vring->dom_id, &vring->rxpfn_tbl, &vring->rxpfn_num, &vring->rx_pindex); + RTE_LOG(INFO, XENHOST, " %s: map rx vring at %p\n", __func__, (void *)vring->rxvring_addr); + if (vring->rxvring_addr) + return 0; + else + return -1; +} + +static int +xen_map_txvringnode(struct xen_gntnode *gntnode, + struct xen_vring *vring) +{ + vring->txvring_addr = + map_gntnode(gntnode, vring->dom_id, &vring->txpfn_tbl, &vring->txpfn_num, &vring->tx_pindex); + RTE_LOG(INFO, XENHOST, " %s: map tx vring at %p\n", __func__, (void *)vring->txvring_addr); + if (vring->txvring_addr) + return 0; + else + return -1; +} + +void +cleanup_vring(struct xen_vring *vring) +{ + int pg_sz = getpagesize(); + uint32_t i; + + RTE_LOG(INFO, XENHOST, " %s: cleanup dom %u vring %u\n", __func__, vring->dom_id, vring->virtio_idx); + if (vring->rxvring_addr) { + munmap(vring->rxvring_addr, vring->rxpfn_num * pg_sz); + RTE_LOG(INFO, XENHOST, " %s: unmap rx vring [%p, %p]\n", + __func__, + vring->rxvring_addr, + RTE_PTR_ADD(vring->rxvring_addr, + vring->rxpfn_num * pg_sz - 1)); + } + vring->rxvring_addr = NULL; + + + if (vring->rx_pindex) { + RTE_LOG(INFO, XENHOST, " %s: unmap rx vring %u grefs\n", __func__, vring->rxpfn_num); + for (i = 0; i < vring->rxpfn_num; i++) { + xen_unmap_grant_ref(vring->rx_pindex[i]); + } + } + vring->rx_pindex = NULL; + + free(vring->rxpfn_tbl); + vring->rxpfn_tbl = NULL; + + if (vring->txvring_addr) { + munmap(vring->txvring_addr, vring->txpfn_num * pg_sz); + RTE_LOG(INFO, XENHOST, " %s: unmap tx vring [%p, %p]\n", + __func__, + vring->txvring_addr, + RTE_PTR_ADD(vring->txvring_addr, + vring->txpfn_num * pg_sz - 1)); + } + vring->txvring_addr = NULL; + + if (vring->tx_pindex) { + RTE_LOG(INFO, XENHOST, " %s: unmap tx vring %u grefs\n", __func__, vring->txpfn_num); + for (i = 0; i < vring->txpfn_num; i++) { + xen_unmap_grant_ref(vring->tx_pindex[i]); + } + } + vring->tx_pindex = NULL; + + free(vring->txpfn_tbl); + vring->txpfn_tbl = NULL; + + if (vring->flag) { + if (!munmap((void *)vring->flag, pg_sz)) + RTE_LOG(INFO, XENHOST, " %s: unmap flag page at %p\n", __func__, vring->flag); + if (!xen_unmap_grant_ref(vring->flag_index)) + RTE_LOG(INFO, XENHOST, " %s: release flag ref index 0x%" PRIx64 "\n", __func__, vring->flag_index); + } + vring->flag = NULL; + return; +} + + + +static int +xen_parse_etheraddr(struct xen_vring *vring) +{ + char path[PATH_MAX] = {0}; + char *buf; + uint32_t len; + int ret = -1; + + snprintf(path, sizeof(path), + XEN_VM_ROOTNODE_FMT"/%d_"XEN_ADDR_SUFFIX, + vring->dom_id, vring->virtio_idx); + + if ((buf = xen_read_node(path, &len)) == NULL) + goto out; + + if (cmdline_parse_etheraddr(NULL, buf, &vring->addr, + sizeof(vring->addr)) < 0) + goto out; + ret = 0; +out: + free(buf); + return ret; +} + + +int +parse_vringnode(struct xen_guest *guest, uint32_t virtio_idx) +{ + char path[PATH_MAX] = {0}; + struct xen_gntnode *rx_gntnode = NULL; + struct xen_gntnode *tx_gntnode = NULL; + struct xen_vring *vring = NULL; + + /*check if null terminated */ + snprintf(path, sizeof(path), + XEN_VM_ROOTNODE_FMT"/%d_"XEN_RXVRING_SUFFIX, + guest->dom_id, + virtio_idx); + + RTE_LOG(INFO, XENHOST, " %s: virtio %u parse rx gntnode %s\n", __func__, virtio_idx, path); + rx_gntnode = parse_gntnode(guest->dom_id, path); + if (rx_gntnode == NULL) + goto err; + + /*check if null terminated */ + snprintf(path, sizeof(path), + XEN_VM_ROOTNODE_FMT"/%d_"XEN_TXVRING_SUFFIX, + guest->dom_id, + virtio_idx); + + RTE_LOG(INFO, XENHOST, " %s: virtio %u parse tx gntnode %s\n", __func__, virtio_idx, path); + tx_gntnode = parse_gntnode(guest->dom_id, path); + if (tx_gntnode == NULL) + goto err; + + vring = &guest->vring[virtio_idx]; + bzero(vring, sizeof(*vring)); + vring->dom_id = guest->dom_id; + vring->virtio_idx = virtio_idx; + + if (xen_parse_etheraddr(vring) != 0) + goto err; + + RTE_LOG(INFO, XENHOST, " %s: virtio %u map rx gntnode %s\n", __func__, virtio_idx, path); + if (xen_map_rxvringnode(rx_gntnode, vring) != 0) + goto err; + + RTE_LOG(INFO, XENHOST, " %s: virtio %u map tx gntnode %s\n", __func__, virtio_idx, path); + if (xen_map_txvringnode(tx_gntnode, vring) != 0) + goto err; + + if (xen_map_vringflag(vring) != 0) + goto err; + + guest->vring_num++; + + xen_free_gntnode(rx_gntnode); + xen_free_gntnode(tx_gntnode); + + return 0; + +err: + if (rx_gntnode) + xen_free_gntnode(rx_gntnode); + if (tx_gntnode) + xen_free_gntnode(tx_gntnode); + if (vring) { + cleanup_vring(vring); + bzero(vring, sizeof(*vring)); + } + return -1; +} + +/* + * Open xen grant dev driver + * @return + * 0 on success, -1 on failure. + */ +static int +xen_grant_init(void) +{ + d_fd = open(XEN_GNTDEV_FNAME, O_RDWR); + + return d_fd == -1? (-1): (0); +} + +/* + * Initialise xenstore handle and open grant dev driver. + * @return + * 0 on success, -1 on failure. + */ +int +xenhost_init(void) +{ + xs = xs_daemon_open(); + if (xs == NULL) { + rte_panic("failed initialize xen daemon handler"); + return -1; + } + if (xen_grant_init()) + return -1; + return 0; +} |