summaryrefslogtreecommitdiffstats
path: root/drivers/net/failsafe
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/net/failsafe')
-rw-r--r--drivers/net/failsafe/Makefile62
-rw-r--r--drivers/net/failsafe/failsafe.c299
-rw-r--r--drivers/net/failsafe/failsafe_args.c474
-rw-r--r--drivers/net/failsafe/failsafe_eal.c118
-rw-r--r--drivers/net/failsafe/failsafe_ether.c437
-rw-r--r--drivers/net/failsafe/failsafe_flow.c244
-rw-r--r--drivers/net/failsafe/failsafe_ops.c867
-rw-r--r--drivers/net/failsafe/failsafe_private.h359
-rw-r--r--drivers/net/failsafe/failsafe_rxtx.c203
-rw-r--r--drivers/net/failsafe/rte_pmd_failsafe_version.map4
10 files changed, 3067 insertions, 0 deletions
diff --git a/drivers/net/failsafe/Makefile b/drivers/net/failsafe/Makefile
new file mode 100644
index 00000000..d516d362
--- /dev/null
+++ b/drivers/net/failsafe/Makefile
@@ -0,0 +1,62 @@
+# BSD LICENSE
+#
+# Copyright 2017 6WIND S.A.
+# Copyright 2017 Mellanox.
+#
+# 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 6WIND S.A. 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 $(RTE_SDK)/mk/rte.vars.mk
+
+# Library name
+LIB = librte_pmd_failsafe.a
+
+EXPORT_MAP := rte_pmd_failsafe_version.map
+
+LIBABIVER := 1
+
+# Sources are stored in SRCS-y
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_FAILSAFE) += failsafe.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_FAILSAFE) += failsafe_args.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_FAILSAFE) += failsafe_eal.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_FAILSAFE) += failsafe_ops.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_FAILSAFE) += failsafe_rxtx.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_FAILSAFE) += failsafe_ether.c
+SRCS-$(CONFIG_RTE_LIBRTE_PMD_FAILSAFE) += failsafe_flow.c
+
+# No exported include files
+
+# Basic CFLAGS:
+CFLAGS += -std=gnu99 -Wextra
+CFLAGS += -O3
+CFLAGS += -I.
+CFLAGS += -D_DEFAULT_SOURCE
+CFLAGS += -D_XOPEN_SOURCE=700
+CFLAGS += $(WERROR_FLAGS)
+CFLAGS += -Wno-strict-prototypes
+CFLAGS += -pedantic
+
+include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/net/failsafe/failsafe.c b/drivers/net/failsafe/failsafe.c
new file mode 100644
index 00000000..6006bef8
--- /dev/null
+++ b/drivers/net/failsafe/failsafe.c
@@ -0,0 +1,299 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 <rte_alarm.h>
+#include <rte_malloc.h>
+#include <rte_ethdev.h>
+#include <rte_ethdev_vdev.h>
+#include <rte_devargs.h>
+#include <rte_kvargs.h>
+#include <rte_vdev.h>
+
+#include "failsafe_private.h"
+
+const char pmd_failsafe_driver_name[] = FAILSAFE_DRIVER_NAME;
+static const struct rte_eth_link eth_link = {
+ .link_speed = ETH_SPEED_NUM_10G,
+ .link_duplex = ETH_LINK_FULL_DUPLEX,
+ .link_status = ETH_LINK_UP,
+ .link_autoneg = ETH_LINK_SPEED_AUTONEG,
+};
+
+static int
+fs_sub_device_alloc(struct rte_eth_dev *dev,
+ const char *params)
+{
+ uint8_t nb_subs;
+ int ret;
+
+ ret = failsafe_args_count_subdevice(dev, params);
+ if (ret)
+ return ret;
+ if (PRIV(dev)->subs_tail > FAILSAFE_MAX_ETHPORTS) {
+ ERROR("Cannot allocate more than %d ports",
+ FAILSAFE_MAX_ETHPORTS);
+ return -ENOSPC;
+ }
+ nb_subs = PRIV(dev)->subs_tail;
+ PRIV(dev)->subs = rte_zmalloc(NULL,
+ sizeof(struct sub_device) * nb_subs,
+ RTE_CACHE_LINE_SIZE);
+ if (PRIV(dev)->subs == NULL) {
+ ERROR("Could not allocate sub_devices");
+ return -ENOMEM;
+ }
+ return 0;
+}
+
+static void
+fs_sub_device_free(struct rte_eth_dev *dev)
+{
+ rte_free(PRIV(dev)->subs);
+}
+
+static void fs_hotplug_alarm(void *arg);
+
+int
+failsafe_hotplug_alarm_install(struct rte_eth_dev *dev)
+{
+ int ret;
+
+ if (dev == NULL)
+ return -EINVAL;
+ if (PRIV(dev)->pending_alarm)
+ return 0;
+ ret = rte_eal_alarm_set(hotplug_poll * 1000,
+ fs_hotplug_alarm,
+ dev);
+ if (ret) {
+ ERROR("Could not set up plug-in event detection");
+ return ret;
+ }
+ PRIV(dev)->pending_alarm = 1;
+ return 0;
+}
+
+int
+failsafe_hotplug_alarm_cancel(struct rte_eth_dev *dev)
+{
+ int ret = 0;
+
+ if (PRIV(dev)->pending_alarm) {
+ rte_errno = 0;
+ rte_eal_alarm_cancel(fs_hotplug_alarm, dev);
+ if (rte_errno) {
+ ERROR("rte_eal_alarm_cancel failed (errno: %s)",
+ strerror(rte_errno));
+ ret = -rte_errno;
+ } else {
+ PRIV(dev)->pending_alarm = 0;
+ }
+ }
+ return ret;
+}
+
+static void
+fs_hotplug_alarm(void *arg)
+{
+ struct rte_eth_dev *dev = arg;
+ struct sub_device *sdev;
+ int ret;
+ uint8_t i;
+
+ if (!PRIV(dev)->pending_alarm)
+ return;
+ PRIV(dev)->pending_alarm = 0;
+ FOREACH_SUBDEV(sdev, i, dev)
+ if (sdev->state != PRIV(dev)->state)
+ break;
+ /* if we have non-probed device */
+ if (i != PRIV(dev)->subs_tail) {
+ ret = failsafe_eth_dev_state_sync(dev);
+ if (ret)
+ ERROR("Unable to synchronize sub_device state");
+ }
+ failsafe_dev_remove(dev);
+ ret = failsafe_hotplug_alarm_install(dev);
+ if (ret)
+ ERROR("Unable to set up next alarm");
+}
+
+static int
+fs_eth_dev_create(struct rte_vdev_device *vdev)
+{
+ struct rte_eth_dev *dev;
+ struct ether_addr *mac;
+ struct fs_priv *priv;
+ struct sub_device *sdev;
+ const char *params;
+ unsigned int socket_id;
+ uint8_t i;
+ int ret;
+
+ dev = NULL;
+ priv = NULL;
+ socket_id = rte_socket_id();
+ INFO("Creating fail-safe device on NUMA socket %u", socket_id);
+ params = rte_vdev_device_args(vdev);
+ if (params == NULL) {
+ ERROR("This PMD requires sub-devices, none provided");
+ return -1;
+ }
+ dev = rte_eth_vdev_allocate(vdev, sizeof(*priv));
+ if (dev == NULL) {
+ ERROR("Unable to allocate rte_eth_dev");
+ return -1;
+ }
+ priv = PRIV(dev);
+ priv->dev = dev;
+ dev->dev_ops = &failsafe_ops;
+ dev->data->mac_addrs = &PRIV(dev)->mac_addrs[0];
+ dev->data->dev_link = eth_link;
+ PRIV(dev)->nb_mac_addr = 1;
+ TAILQ_INIT(&PRIV(dev)->flow_list);
+ dev->rx_pkt_burst = (eth_rx_burst_t)&failsafe_rx_burst;
+ dev->tx_pkt_burst = (eth_tx_burst_t)&failsafe_tx_burst;
+ ret = fs_sub_device_alloc(dev, params);
+ if (ret) {
+ ERROR("Could not allocate sub_devices");
+ goto free_dev;
+ }
+ ret = failsafe_args_parse(dev, params);
+ if (ret)
+ goto free_subs;
+ ret = failsafe_eal_init(dev);
+ if (ret)
+ goto free_args;
+ ret = failsafe_hotplug_alarm_install(dev);
+ if (ret) {
+ ERROR("Could not set up plug-in event detection");
+ goto free_args;
+ }
+ mac = &dev->data->mac_addrs[0];
+ if (mac_from_arg) {
+ /*
+ * If MAC address was provided as a parameter,
+ * apply to all probed slaves.
+ */
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_PROBED) {
+ ret = rte_eth_dev_default_mac_addr_set(PORT_ID(sdev),
+ mac);
+ if (ret) {
+ ERROR("Failed to set default MAC address");
+ goto free_args;
+ }
+ }
+ } else {
+ /*
+ * Use the ether_addr from first probed
+ * device, either preferred or fallback.
+ */
+ FOREACH_SUBDEV(sdev, i, dev)
+ if (sdev->state >= DEV_PROBED) {
+ ether_addr_copy(&ETH(sdev)->data->mac_addrs[0],
+ mac);
+ break;
+ }
+ /*
+ * If no device has been probed and no ether_addr
+ * has been provided on the command line, use a random
+ * valid one.
+ * It will be applied during future slave state syncs to
+ * probed slaves.
+ */
+ if (i == priv->subs_tail)
+ eth_random_addr(&mac->addr_bytes[0]);
+ }
+ INFO("MAC address is %02x:%02x:%02x:%02x:%02x:%02x",
+ mac->addr_bytes[0], mac->addr_bytes[1],
+ mac->addr_bytes[2], mac->addr_bytes[3],
+ mac->addr_bytes[4], mac->addr_bytes[5]);
+ dev->data->dev_flags |= RTE_ETH_DEV_INTR_LSC;
+ return 0;
+free_args:
+ failsafe_args_free(dev);
+free_subs:
+ fs_sub_device_free(dev);
+free_dev:
+ rte_free(PRIV(dev));
+ rte_eth_dev_release_port(dev);
+ return -1;
+}
+
+static int
+fs_rte_eth_free(const char *name)
+{
+ struct rte_eth_dev *dev;
+ int ret;
+
+ dev = rte_eth_dev_allocated(name);
+ if (dev == NULL)
+ return -ENODEV;
+ ret = failsafe_eal_uninit(dev);
+ if (ret)
+ ERROR("Error while uninitializing sub-EAL");
+ failsafe_args_free(dev);
+ fs_sub_device_free(dev);
+ rte_free(PRIV(dev));
+ rte_eth_dev_release_port(dev);
+ return ret;
+}
+
+static int
+rte_pmd_failsafe_probe(struct rte_vdev_device *vdev)
+{
+ const char *name;
+
+ name = rte_vdev_device_name(vdev);
+ INFO("Initializing " FAILSAFE_DRIVER_NAME " for %s",
+ name);
+ return fs_eth_dev_create(vdev);
+}
+
+static int
+rte_pmd_failsafe_remove(struct rte_vdev_device *vdev)
+{
+ const char *name;
+
+ name = rte_vdev_device_name(vdev);
+ INFO("Uninitializing " FAILSAFE_DRIVER_NAME " for %s", name);
+ return fs_rte_eth_free(name);
+}
+
+static struct rte_vdev_driver failsafe_drv = {
+ .probe = rte_pmd_failsafe_probe,
+ .remove = rte_pmd_failsafe_remove,
+};
+
+RTE_PMD_REGISTER_VDEV(net_failsafe, failsafe_drv);
+RTE_PMD_REGISTER_PARAM_STRING(net_failsafe, PMD_FAILSAFE_PARAM_STRING);
diff --git a/drivers/net/failsafe/failsafe_args.c b/drivers/net/failsafe/failsafe_args.c
new file mode 100644
index 00000000..1f22416f
--- /dev/null
+++ b/drivers/net/failsafe/failsafe_args.c
@@ -0,0 +1,474 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 <string.h>
+#include <errno.h>
+
+#include <rte_debug.h>
+#include <rte_devargs.h>
+#include <rte_malloc.h>
+#include <rte_kvargs.h>
+
+#include "failsafe_private.h"
+
+#define DEVARGS_MAXLEN 4096
+
+/* Callback used when a new device is found in devargs */
+typedef int (parse_cb)(struct rte_eth_dev *dev, const char *params,
+ uint8_t head);
+
+uint64_t hotplug_poll = FAILSAFE_HOTPLUG_DEFAULT_TIMEOUT_MS;
+int mac_from_arg = 0;
+
+const char *pmd_failsafe_init_parameters[] = {
+ PMD_FAILSAFE_HOTPLUG_POLL_KVARG,
+ PMD_FAILSAFE_MAC_KVARG,
+ NULL,
+};
+
+/*
+ * input: text.
+ * output: 0: if text[0] != '(',
+ * 0: if there are no corresponding ')'
+ * n: distance to corresponding ')' otherwise
+ */
+static size_t
+closing_paren(const char *text)
+{
+ int nb_open = 0;
+ size_t i = 0;
+
+ while (text[i] != '\0') {
+ if (text[i] == '(')
+ nb_open++;
+ if (text[i] == ')')
+ nb_open--;
+ if (nb_open == 0)
+ return i;
+ i++;
+ }
+ return 0;
+}
+
+static int
+fs_parse_device(struct sub_device *sdev, char *args)
+{
+ struct rte_devargs *d;
+ int ret;
+
+ d = &sdev->devargs;
+ DEBUG("%s", args);
+ ret = rte_eal_devargs_parse(args, d);
+ if (ret) {
+ DEBUG("devargs parsing failed with code %d", ret);
+ return ret;
+ }
+ sdev->bus = d->bus;
+ sdev->state = DEV_PARSED;
+ return 0;
+}
+
+static void
+fs_sanitize_cmdline(char *args)
+{
+ char *nl;
+
+ nl = strrchr(args, '\n');
+ if (nl)
+ nl[0] = '\0';
+}
+
+static int
+fs_execute_cmd(struct sub_device *sdev, char *cmdline)
+{
+ FILE *fp;
+ /* store possible newline as well */
+ char output[DEVARGS_MAXLEN + 1];
+ size_t len;
+ int old_err;
+ int ret, pclose_ret;
+
+ RTE_ASSERT(cmdline != NULL || sdev->cmdline != NULL);
+ if (sdev->cmdline == NULL) {
+ size_t i;
+
+ len = strlen(cmdline) + 1;
+ sdev->cmdline = calloc(1, len);
+ if (sdev->cmdline == NULL) {
+ ERROR("Command line allocation failed");
+ return -ENOMEM;
+ }
+ snprintf(sdev->cmdline, len, "%s", cmdline);
+ /* Replace all commas in the command line by spaces */
+ for (i = 0; i < len; i++)
+ if (sdev->cmdline[i] == ',')
+ sdev->cmdline[i] = ' ';
+ }
+ DEBUG("'%s'", sdev->cmdline);
+ old_err = errno;
+ fp = popen(sdev->cmdline, "r");
+ if (fp == NULL) {
+ ret = errno;
+ ERROR("popen: %s", strerror(errno));
+ errno = old_err;
+ return ret;
+ }
+ /* We only read one line */
+ if (fgets(output, sizeof(output) - 1, fp) == NULL) {
+ DEBUG("Could not read command output");
+ ret = -ENODEV;
+ goto ret_pclose;
+ }
+ fs_sanitize_cmdline(output);
+ if (output[0] == '\0') {
+ ret = -ENODEV;
+ goto ret_pclose;
+ }
+ ret = fs_parse_device(sdev, output);
+ if (ret) {
+ ERROR("Parsing device '%s' failed", output);
+ goto ret_pclose;
+ }
+ret_pclose:
+ pclose_ret = pclose(fp);
+ if (pclose_ret) {
+ pclose_ret = errno;
+ ERROR("pclose: %s", strerror(errno));
+ errno = old_err;
+ return pclose_ret;
+ }
+ return ret;
+}
+
+static int
+fs_parse_device_param(struct rte_eth_dev *dev, const char *param,
+ uint8_t head)
+{
+ struct fs_priv *priv;
+ struct sub_device *sdev;
+ char *args = NULL;
+ size_t a, b;
+ int ret;
+
+ priv = PRIV(dev);
+ a = 0;
+ b = 0;
+ ret = 0;
+ while (param[b] != '(' &&
+ param[b] != '\0')
+ b++;
+ a = b;
+ b += closing_paren(&param[b]);
+ if (a == b) {
+ ERROR("Dangling parenthesis");
+ return -EINVAL;
+ }
+ a += 1;
+ args = strndup(&param[a], b - a);
+ if (args == NULL) {
+ ERROR("Not enough memory for parameter parsing");
+ return -ENOMEM;
+ }
+ sdev = &priv->subs[head];
+ if (strncmp(param, "dev", 3) == 0) {
+ ret = fs_parse_device(sdev, args);
+ if (ret)
+ goto free_args;
+ } else if (strncmp(param, "exec", 4) == 0) {
+ ret = fs_execute_cmd(sdev, args);
+ if (ret == -ENODEV) {
+ DEBUG("Reading device info from command line failed");
+ ret = 0;
+ }
+ if (ret)
+ goto free_args;
+ } else {
+ ERROR("Unrecognized device type: %.*s", (int)b, param);
+ return -EINVAL;
+ }
+free_args:
+ free(args);
+ return ret;
+}
+
+static int
+fs_parse_sub_devices(parse_cb *cb,
+ struct rte_eth_dev *dev, const char *params)
+{
+ size_t a, b;
+ uint8_t head;
+ int ret;
+
+ a = 0;
+ head = 0;
+ ret = 0;
+ while (params[a] != '\0') {
+ b = a;
+ while (params[b] != '(' &&
+ params[b] != ',' &&
+ params[b] != '\0')
+ b++;
+ if (b == a) {
+ ERROR("Invalid parameter");
+ return -EINVAL;
+ }
+ if (params[b] == ',') {
+ a = b + 1;
+ continue;
+ }
+ if (params[b] == '(') {
+ size_t start = b;
+
+ b += closing_paren(&params[b]);
+ if (b == start) {
+ ERROR("Dangling parenthesis");
+ return -EINVAL;
+ }
+ ret = (*cb)(dev, &params[a], head);
+ if (ret)
+ return ret;
+ head += 1;
+ b += 1;
+ if (params[b] == '\0')
+ return 0;
+ }
+ a = b + 1;
+ }
+ return 0;
+}
+
+static int
+fs_remove_sub_devices_definition(char params[DEVARGS_MAXLEN])
+{
+ char buffer[DEVARGS_MAXLEN] = {0};
+ size_t a, b;
+ int i;
+
+ a = 0;
+ i = 0;
+ while (params[a] != '\0') {
+ b = a;
+ while (params[b] != '(' &&
+ params[b] != ',' &&
+ params[b] != '\0')
+ b++;
+ if (b == a) {
+ ERROR("Invalid parameter");
+ return -EINVAL;
+ }
+ if (params[b] == ',' || params[b] == '\0')
+ i += snprintf(&buffer[i], b - a + 1, "%s", &params[a]);
+ if (params[b] == '(') {
+ size_t start = b;
+ b += closing_paren(&params[b]);
+ if (b == start)
+ return -EINVAL;
+ b += 1;
+ if (params[b] == '\0')
+ goto out;
+ }
+ a = b + 1;
+ }
+out:
+ snprintf(params, DEVARGS_MAXLEN, "%s", buffer);
+ return 0;
+}
+
+static int
+fs_get_u64_arg(const char *key __rte_unused,
+ const char *value, void *out)
+{
+ uint64_t *u64 = out;
+ char *endptr = NULL;
+
+ if ((value == NULL) || (out == NULL))
+ return -EINVAL;
+ errno = 0;
+ *u64 = strtoull(value, &endptr, 0);
+ if (errno != 0)
+ return -errno;
+ if (endptr == value)
+ return -1;
+ return 0;
+}
+
+static int
+fs_get_mac_addr_arg(const char *key __rte_unused,
+ const char *value, void *out)
+{
+ struct ether_addr *ea = out;
+ int ret;
+
+ if ((value == NULL) || (out == NULL))
+ return -EINVAL;
+ ret = sscanf(value, "%hhx:%hhx:%hhx:%hhx:%hhx:%hhx",
+ &ea->addr_bytes[0], &ea->addr_bytes[1],
+ &ea->addr_bytes[2], &ea->addr_bytes[3],
+ &ea->addr_bytes[4], &ea->addr_bytes[5]);
+ return ret != ETHER_ADDR_LEN;
+}
+
+int
+failsafe_args_parse(struct rte_eth_dev *dev, const char *params)
+{
+ struct fs_priv *priv;
+ char mut_params[DEVARGS_MAXLEN] = "";
+ struct rte_kvargs *kvlist = NULL;
+ unsigned int arg_count;
+ size_t n;
+ int ret;
+
+ priv = PRIV(dev);
+ ret = 0;
+ priv->subs_tx = FAILSAFE_MAX_ETHPORTS;
+ /* default parameters */
+ n = snprintf(mut_params, sizeof(mut_params), "%s", params);
+ if (n >= sizeof(mut_params)) {
+ ERROR("Parameter string too long (>=%zu)",
+ sizeof(mut_params));
+ return -ENOMEM;
+ }
+ ret = fs_parse_sub_devices(fs_parse_device_param,
+ dev, params);
+ if (ret < 0)
+ return ret;
+ ret = fs_remove_sub_devices_definition(mut_params);
+ if (ret < 0)
+ return ret;
+ if (strnlen(mut_params, sizeof(mut_params)) > 0) {
+ kvlist = rte_kvargs_parse(mut_params,
+ pmd_failsafe_init_parameters);
+ if (kvlist == NULL) {
+ ERROR("Error parsing parameters, usage:\n"
+ PMD_FAILSAFE_PARAM_STRING);
+ return -1;
+ }
+ /* PLUG_IN event poll timer */
+ arg_count = rte_kvargs_count(kvlist,
+ PMD_FAILSAFE_HOTPLUG_POLL_KVARG);
+ if (arg_count == 1) {
+ ret = rte_kvargs_process(kvlist,
+ PMD_FAILSAFE_HOTPLUG_POLL_KVARG,
+ &fs_get_u64_arg, &hotplug_poll);
+ if (ret < 0)
+ goto free_kvlist;
+ }
+ /* MAC addr */
+ arg_count = rte_kvargs_count(kvlist,
+ PMD_FAILSAFE_MAC_KVARG);
+ if (arg_count > 0) {
+ ret = rte_kvargs_process(kvlist,
+ PMD_FAILSAFE_MAC_KVARG,
+ &fs_get_mac_addr_arg,
+ &dev->data->mac_addrs[0]);
+ if (ret < 0)
+ goto free_kvlist;
+ mac_from_arg = 1;
+ }
+ }
+ PRIV(dev)->state = DEV_PARSED;
+free_kvlist:
+ rte_kvargs_free(kvlist);
+ return ret;
+}
+
+void
+failsafe_args_free(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV(sdev, i, dev) {
+ rte_free(sdev->cmdline);
+ sdev->cmdline = NULL;
+ free(sdev->devargs.args);
+ sdev->devargs.args = NULL;
+ }
+}
+
+static int
+fs_count_device(struct rte_eth_dev *dev, const char *param,
+ uint8_t head __rte_unused)
+{
+ size_t b = 0;
+
+ while (param[b] != '(' &&
+ param[b] != '\0')
+ b++;
+ if (strncmp(param, "dev", b) != 0 &&
+ strncmp(param, "exec", b) != 0) {
+ ERROR("Unrecognized device type: %.*s", (int)b, param);
+ return -EINVAL;
+ }
+ PRIV(dev)->subs_tail += 1;
+ return 0;
+}
+
+int
+failsafe_args_count_subdevice(struct rte_eth_dev *dev,
+ const char *params)
+{
+ return fs_parse_sub_devices(fs_count_device,
+ dev, params);
+}
+
+static int
+fs_parse_sub_device(struct sub_device *sdev)
+{
+ struct rte_devargs *da;
+ char devstr[DEVARGS_MAXLEN] = "";
+
+ da = &sdev->devargs;
+ snprintf(devstr, sizeof(devstr), "%s,%s", da->name, da->args);
+ return fs_parse_device(sdev, devstr);
+}
+
+int
+failsafe_args_parse_subs(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret = 0;
+
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (sdev->state >= DEV_PARSED)
+ continue;
+ if (sdev->cmdline)
+ ret = fs_execute_cmd(sdev, sdev->cmdline);
+ else
+ ret = fs_parse_sub_device(sdev);
+ if (ret == 0)
+ sdev->state = DEV_PARSED;
+ }
+ return 0;
+}
diff --git a/drivers/net/failsafe/failsafe_eal.c b/drivers/net/failsafe/failsafe_eal.c
new file mode 100644
index 00000000..c8f4318e
--- /dev/null
+++ b/drivers/net/failsafe/failsafe_eal.c
@@ -0,0 +1,118 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 <rte_malloc.h>
+
+#include "failsafe_private.h"
+
+static int
+fs_bus_init(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ struct rte_devargs *da;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (sdev->state != DEV_PARSED)
+ continue;
+ da = &sdev->devargs;
+ ret = rte_eal_hotplug_add(da->bus->name,
+ da->name,
+ da->args);
+ if (ret) {
+ ERROR("sub_device %d probe failed %s%s%s", i,
+ rte_errno ? "(" : "",
+ rte_errno ? strerror(rte_errno) : "",
+ rte_errno ? ")" : "");
+ continue;
+ }
+ ETH(sdev) = rte_eth_dev_allocated(da->name);
+ if (ETH(sdev) == NULL) {
+ ERROR("sub_device %d init went wrong", i);
+ return -ENODEV;
+ }
+ SUB_ID(sdev) = i;
+ sdev->fs_dev = dev;
+ sdev->dev = ETH(sdev)->device;
+ ETH(sdev)->state = RTE_ETH_DEV_DEFERRED;
+ sdev->state = DEV_PROBED;
+ }
+ return 0;
+}
+
+int
+failsafe_eal_init(struct rte_eth_dev *dev)
+{
+ int ret;
+
+ ret = fs_bus_init(dev);
+ if (ret)
+ return ret;
+ if (PRIV(dev)->state < DEV_PROBED)
+ PRIV(dev)->state = DEV_PROBED;
+ fs_switch_dev(dev, NULL);
+ return 0;
+}
+
+static int
+fs_bus_uninit(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev = NULL;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_PROBED) {
+ ret = rte_eal_hotplug_remove(sdev->bus->name,
+ sdev->dev->name);
+ if (ret) {
+ ERROR("Failed to remove requested device %s",
+ sdev->dev->name);
+ continue;
+ }
+ sdev->state = DEV_PROBED - 1;
+ }
+ return 0;
+}
+
+int
+failsafe_eal_uninit(struct rte_eth_dev *dev)
+{
+ int ret;
+
+ ret = fs_bus_uninit(dev);
+ if (ret)
+ return ret;
+ PRIV(dev)->state = DEV_PROBED - 1;
+ return 0;
+}
diff --git a/drivers/net/failsafe/failsafe_ether.c b/drivers/net/failsafe/failsafe_ether.c
new file mode 100644
index 00000000..a3a8cce9
--- /dev/null
+++ b/drivers/net/failsafe/failsafe_ether.c
@@ -0,0 +1,437 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 <unistd.h>
+
+#include <rte_flow.h>
+#include <rte_flow_driver.h>
+
+#include "failsafe_private.h"
+
+/** Print a message out of a flow error. */
+static int
+fs_flow_complain(struct rte_flow_error *error)
+{
+ static const char *const errstrlist[] = {
+ [RTE_FLOW_ERROR_TYPE_NONE] = "no error",
+ [RTE_FLOW_ERROR_TYPE_UNSPECIFIED] = "cause unspecified",
+ [RTE_FLOW_ERROR_TYPE_HANDLE] = "flow rule (handle)",
+ [RTE_FLOW_ERROR_TYPE_ATTR_GROUP] = "group field",
+ [RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY] = "priority field",
+ [RTE_FLOW_ERROR_TYPE_ATTR_INGRESS] = "ingress field",
+ [RTE_FLOW_ERROR_TYPE_ATTR_EGRESS] = "egress field",
+ [RTE_FLOW_ERROR_TYPE_ATTR] = "attributes structure",
+ [RTE_FLOW_ERROR_TYPE_ITEM_NUM] = "pattern length",
+ [RTE_FLOW_ERROR_TYPE_ITEM] = "specific pattern item",
+ [RTE_FLOW_ERROR_TYPE_ACTION_NUM] = "number of actions",
+ [RTE_FLOW_ERROR_TYPE_ACTION] = "specific action",
+ };
+ const char *errstr;
+ char buf[32];
+ int err = rte_errno;
+
+ if ((unsigned int)error->type >= RTE_DIM(errstrlist) ||
+ !errstrlist[error->type])
+ errstr = "unknown type";
+ else
+ errstr = errstrlist[error->type];
+ ERROR("Caught error type %d (%s): %s%s\n",
+ error->type, errstr,
+ error->cause ? (snprintf(buf, sizeof(buf), "cause: %p, ",
+ error->cause), buf) : "",
+ error->message ? error->message : "(no stated reason)");
+ return -err;
+}
+
+static int
+eth_dev_flow_isolate_set(struct rte_eth_dev *dev,
+ struct sub_device *sdev)
+{
+ struct rte_flow_error ferror;
+ int ret;
+
+ if (!PRIV(dev)->flow_isolated) {
+ DEBUG("Flow isolation already disabled");
+ } else {
+ DEBUG("Enabling flow isolation");
+ ret = rte_flow_isolate(PORT_ID(sdev),
+ PRIV(dev)->flow_isolated,
+ &ferror);
+ if (ret) {
+ fs_flow_complain(&ferror);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static int
+fs_eth_dev_conf_apply(struct rte_eth_dev *dev,
+ struct sub_device *sdev)
+{
+ struct rte_eth_dev *edev;
+ struct rte_vlan_filter_conf *vfc1;
+ struct rte_vlan_filter_conf *vfc2;
+ struct rte_flow *flow;
+ struct rte_flow_error ferror;
+ uint32_t i;
+ int ret;
+
+ edev = ETH(sdev);
+ /* RX queue setup */
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ struct rxq *rxq;
+
+ rxq = dev->data->rx_queues[i];
+ ret = rte_eth_rx_queue_setup(PORT_ID(sdev), i,
+ rxq->info.nb_desc, rxq->socket_id,
+ &rxq->info.conf, rxq->info.mp);
+ if (ret) {
+ ERROR("rx_queue_setup failed");
+ return ret;
+ }
+ }
+ /* TX queue setup */
+ for (i = 0; i < dev->data->nb_tx_queues; i++) {
+ struct txq *txq;
+
+ txq = dev->data->tx_queues[i];
+ ret = rte_eth_tx_queue_setup(PORT_ID(sdev), i,
+ txq->info.nb_desc, txq->socket_id,
+ &txq->info.conf);
+ if (ret) {
+ ERROR("tx_queue_setup failed");
+ return ret;
+ }
+ }
+ /* dev_link.link_status */
+ if (dev->data->dev_link.link_status !=
+ edev->data->dev_link.link_status) {
+ DEBUG("Configuring link_status");
+ if (dev->data->dev_link.link_status)
+ ret = rte_eth_dev_set_link_up(PORT_ID(sdev));
+ else
+ ret = rte_eth_dev_set_link_down(PORT_ID(sdev));
+ if (ret) {
+ ERROR("Failed to apply link_status");
+ return ret;
+ }
+ } else {
+ DEBUG("link_status already set");
+ }
+ /* promiscuous */
+ if (dev->data->promiscuous != edev->data->promiscuous) {
+ DEBUG("Configuring promiscuous");
+ if (dev->data->promiscuous)
+ rte_eth_promiscuous_enable(PORT_ID(sdev));
+ else
+ rte_eth_promiscuous_disable(PORT_ID(sdev));
+ } else {
+ DEBUG("promiscuous already set");
+ }
+ /* all_multicast */
+ if (dev->data->all_multicast != edev->data->all_multicast) {
+ DEBUG("Configuring all_multicast");
+ if (dev->data->all_multicast)
+ rte_eth_allmulticast_enable(PORT_ID(sdev));
+ else
+ rte_eth_allmulticast_disable(PORT_ID(sdev));
+ } else {
+ DEBUG("all_multicast already set");
+ }
+ /* MTU */
+ if (dev->data->mtu != edev->data->mtu) {
+ DEBUG("Configuring MTU");
+ ret = rte_eth_dev_set_mtu(PORT_ID(sdev), dev->data->mtu);
+ if (ret) {
+ ERROR("Failed to apply MTU");
+ return ret;
+ }
+ } else {
+ DEBUG("MTU already set");
+ }
+ /* default MAC */
+ DEBUG("Configuring default MAC address");
+ ret = rte_eth_dev_default_mac_addr_set(PORT_ID(sdev),
+ &dev->data->mac_addrs[0]);
+ if (ret) {
+ ERROR("Setting default MAC address failed");
+ return ret;
+ }
+ /* additional MAC */
+ if (PRIV(dev)->nb_mac_addr > 1)
+ DEBUG("Configure additional MAC address%s",
+ (PRIV(dev)->nb_mac_addr > 2 ? "es" : ""));
+ for (i = 1; i < PRIV(dev)->nb_mac_addr; i++) {
+ struct ether_addr *ea;
+
+ ea = &dev->data->mac_addrs[i];
+ ret = rte_eth_dev_mac_addr_add(PORT_ID(sdev), ea,
+ PRIV(dev)->mac_addr_pool[i]);
+ if (ret) {
+ char ea_fmt[ETHER_ADDR_FMT_SIZE];
+
+ ether_format_addr(ea_fmt, ETHER_ADDR_FMT_SIZE, ea);
+ ERROR("Adding MAC address %s failed", ea_fmt);
+ }
+ }
+ /* VLAN filter */
+ vfc1 = &dev->data->vlan_filter_conf;
+ vfc2 = &edev->data->vlan_filter_conf;
+ if (memcmp(vfc1, vfc2, sizeof(struct rte_vlan_filter_conf))) {
+ uint64_t vbit;
+ uint64_t ids;
+ size_t i;
+ uint16_t vlan_id;
+
+ DEBUG("Configuring VLAN filter");
+ for (i = 0; i < RTE_DIM(vfc1->ids); i++) {
+ if (vfc1->ids[i] == 0)
+ continue;
+ ids = vfc1->ids[i];
+ while (ids) {
+ vlan_id = 64 * i;
+ /* count trailing zeroes */
+ vbit = ~ids & (ids - 1);
+ /* clear least significant bit set */
+ ids ^= (ids ^ (ids - 1)) ^ vbit;
+ for (; vbit; vlan_id++)
+ vbit >>= 1;
+ ret = rte_eth_dev_vlan_filter(
+ PORT_ID(sdev), vlan_id, 1);
+ if (ret) {
+ ERROR("Failed to apply VLAN filter %hu",
+ vlan_id);
+ return ret;
+ }
+ }
+ }
+ } else {
+ DEBUG("VLAN filter already set");
+ }
+ /* rte_flow */
+ if (TAILQ_EMPTY(&PRIV(dev)->flow_list)) {
+ DEBUG("rte_flow already set");
+ } else {
+ DEBUG("Resetting rte_flow configuration");
+ ret = rte_flow_flush(PORT_ID(sdev), &ferror);
+ if (ret) {
+ fs_flow_complain(&ferror);
+ return ret;
+ }
+ i = 0;
+ rte_errno = 0;
+ DEBUG("Configuring rte_flow");
+ TAILQ_FOREACH(flow, &PRIV(dev)->flow_list, next) {
+ DEBUG("Creating flow #%" PRIu32, i++);
+ flow->flows[SUB_ID(sdev)] =
+ rte_flow_create(PORT_ID(sdev),
+ &flow->fd->attr,
+ flow->fd->items,
+ flow->fd->actions,
+ &ferror);
+ ret = rte_errno;
+ if (ret)
+ break;
+ }
+ if (ret) {
+ fs_flow_complain(&ferror);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static void
+fs_dev_remove(struct sub_device *sdev)
+{
+ int ret;
+
+ if (sdev == NULL)
+ return;
+ switch (sdev->state) {
+ case DEV_STARTED:
+ rte_eth_dev_stop(PORT_ID(sdev));
+ sdev->state = DEV_ACTIVE;
+ /* fallthrough */
+ case DEV_ACTIVE:
+ rte_eth_dev_close(PORT_ID(sdev));
+ sdev->state = DEV_PROBED;
+ /* fallthrough */
+ case DEV_PROBED:
+ ret = rte_eal_hotplug_remove(sdev->bus->name,
+ sdev->dev->name);
+ if (ret) {
+ ERROR("Bus detach failed for sub_device %u",
+ SUB_ID(sdev));
+ } else {
+ ETH(sdev)->state = RTE_ETH_DEV_UNUSED;
+ }
+ sdev->state = DEV_PARSED;
+ /* fallthrough */
+ case DEV_PARSED:
+ case DEV_UNDEFINED:
+ sdev->state = DEV_UNDEFINED;
+ /* the end */
+ break;
+ }
+ failsafe_hotplug_alarm_install(sdev->fs_dev);
+}
+
+static inline int
+fs_rxtx_clean(struct sub_device *sdev)
+{
+ uint16_t i;
+
+ for (i = 0; i < ETH(sdev)->data->nb_rx_queues; i++)
+ if (FS_ATOMIC_RX(sdev, i))
+ return 0;
+ for (i = 0; i < ETH(sdev)->data->nb_tx_queues; i++)
+ if (FS_ATOMIC_TX(sdev, i))
+ return 0;
+ return 1;
+}
+
+void
+failsafe_dev_remove(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ if (sdev->remove && fs_rxtx_clean(sdev))
+ fs_dev_remove(sdev);
+}
+
+int
+failsafe_eth_dev_state_sync(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint32_t inactive;
+ int ret;
+ uint8_t i;
+
+ if (PRIV(dev)->state < DEV_PARSED)
+ return 0;
+
+ ret = failsafe_args_parse_subs(dev);
+ if (ret)
+ goto err_remove;
+
+ if (PRIV(dev)->state < DEV_PROBED)
+ return 0;
+ ret = failsafe_eal_init(dev);
+ if (ret)
+ goto err_remove;
+ if (PRIV(dev)->state < DEV_ACTIVE)
+ return 0;
+ inactive = 0;
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (sdev->state == DEV_PROBED) {
+ inactive |= UINT32_C(1) << i;
+ ret = eth_dev_flow_isolate_set(dev, sdev);
+ if (ret) {
+ ERROR("Could not apply configuration to sub_device %d",
+ i);
+ goto err_remove;
+ }
+ }
+ }
+ ret = dev->dev_ops->dev_configure(dev);
+ if (ret)
+ goto err_remove;
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (inactive & (UINT32_C(1) << i)) {
+ ret = fs_eth_dev_conf_apply(dev, sdev);
+ if (ret) {
+ ERROR("Could not apply configuration to sub_device %d",
+ i);
+ goto err_remove;
+ }
+ }
+ }
+ /*
+ * If new devices have been configured, check if
+ * the link state has changed.
+ */
+ if (inactive)
+ dev->dev_ops->link_update(dev, 1);
+ if (PRIV(dev)->state < DEV_STARTED)
+ return 0;
+ ret = dev->dev_ops->dev_start(dev);
+ if (ret)
+ goto err_remove;
+ return 0;
+err_remove:
+ FOREACH_SUBDEV(sdev, i, dev)
+ if (sdev->state != PRIV(dev)->state)
+ sdev->remove = 1;
+ return ret;
+}
+
+int
+failsafe_eth_rmv_event_callback(uint8_t port_id __rte_unused,
+ enum rte_eth_event_type event __rte_unused,
+ void *cb_arg, void *out __rte_unused)
+{
+ struct sub_device *sdev = cb_arg;
+
+ /* Switch as soon as possible tx_dev. */
+ fs_switch_dev(sdev->fs_dev, sdev);
+ /* Use safe bursts in any case. */
+ set_burst_fn(sdev->fs_dev, 1);
+ /*
+ * Async removal, the sub-PMD will try to unregister
+ * the callback at the source of the current thread context.
+ */
+ sdev->remove = 1;
+ return 0;
+}
+
+int
+failsafe_eth_lsc_event_callback(uint8_t port_id __rte_unused,
+ enum rte_eth_event_type event __rte_unused,
+ void *cb_arg, void *out __rte_unused)
+{
+ struct rte_eth_dev *dev = cb_arg;
+ int ret;
+
+ ret = dev->dev_ops->link_update(dev, 0);
+ /* We must pass on the LSC event */
+ if (ret)
+ return _rte_eth_dev_callback_process(dev,
+ RTE_ETH_EVENT_INTR_LSC,
+ NULL, NULL);
+ else
+ return 0;
+}
diff --git a/drivers/net/failsafe/failsafe_flow.c b/drivers/net/failsafe/failsafe_flow.c
new file mode 100644
index 00000000..153ceeed
--- /dev/null
+++ b/drivers/net/failsafe/failsafe_flow.c
@@ -0,0 +1,244 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 <sys/queue.h>
+
+#include <rte_malloc.h>
+#include <rte_tailq.h>
+#include <rte_flow.h>
+#include <rte_flow_driver.h>
+
+#include "failsafe_private.h"
+
+static struct rte_flow *
+fs_flow_allocate(const struct rte_flow_attr *attr,
+ const struct rte_flow_item *items,
+ const struct rte_flow_action *actions)
+{
+ struct rte_flow *flow;
+ size_t fdsz;
+
+ fdsz = rte_flow_copy(NULL, 0, attr, items, actions);
+ flow = rte_zmalloc(NULL,
+ sizeof(struct rte_flow) + fdsz,
+ RTE_CACHE_LINE_SIZE);
+ if (flow == NULL) {
+ ERROR("Could not allocate new flow");
+ return NULL;
+ }
+ flow->fd = (void *)((uintptr_t)flow + sizeof(*flow));
+ if (rte_flow_copy(flow->fd, fdsz, attr, items, actions) != fdsz) {
+ ERROR("Failed to copy flow description");
+ rte_free(flow);
+ return NULL;
+ }
+ return flow;
+}
+
+static void
+fs_flow_release(struct rte_flow **flow)
+{
+ rte_free(*flow);
+ *flow = NULL;
+}
+
+static int
+fs_flow_validate(struct rte_eth_dev *dev,
+ const struct rte_flow_attr *attr,
+ const struct rte_flow_item patterns[],
+ const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_flow_validate on sub_device %d", i);
+ ret = rte_flow_validate(PORT_ID(sdev),
+ attr, patterns, actions, error);
+ if (ret) {
+ ERROR("Operation rte_flow_validate failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static struct rte_flow *
+fs_flow_create(struct rte_eth_dev *dev,
+ const struct rte_flow_attr *attr,
+ const struct rte_flow_item patterns[],
+ const struct rte_flow_action actions[],
+ struct rte_flow_error *error)
+{
+ struct sub_device *sdev;
+ struct rte_flow *flow;
+ uint8_t i;
+
+ flow = fs_flow_allocate(attr, patterns, actions);
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ flow->flows[i] = rte_flow_create(PORT_ID(sdev),
+ attr, patterns, actions, error);
+ if (flow->flows[i] == NULL) {
+ ERROR("Failed to create flow on sub_device %d",
+ i);
+ goto err;
+ }
+ }
+ TAILQ_INSERT_TAIL(&PRIV(dev)->flow_list, flow, next);
+ return flow;
+err:
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (flow->flows[i] != NULL)
+ rte_flow_destroy(PORT_ID(sdev),
+ flow->flows[i], error);
+ }
+ fs_flow_release(&flow);
+ return NULL;
+}
+
+static int
+fs_flow_destroy(struct rte_eth_dev *dev,
+ struct rte_flow *flow,
+ struct rte_flow_error *error)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ if (flow == NULL) {
+ ERROR("Invalid flow");
+ return -EINVAL;
+ }
+ ret = 0;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ int local_ret;
+
+ if (flow->flows[i] == NULL)
+ continue;
+ local_ret = rte_flow_destroy(PORT_ID(sdev),
+ flow->flows[i], error);
+ if (local_ret) {
+ ERROR("Failed to destroy flow on sub_device %d: %d",
+ i, local_ret);
+ if (ret == 0)
+ ret = local_ret;
+ }
+ }
+ TAILQ_REMOVE(&PRIV(dev)->flow_list, flow, next);
+ fs_flow_release(&flow);
+ return ret;
+}
+
+static int
+fs_flow_flush(struct rte_eth_dev *dev,
+ struct rte_flow_error *error)
+{
+ struct sub_device *sdev;
+ struct rte_flow *flow;
+ void *tmp;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_flow_flush on sub_device %d", i);
+ ret = rte_flow_flush(PORT_ID(sdev), error);
+ if (ret) {
+ ERROR("Operation rte_flow_flush failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ }
+ TAILQ_FOREACH_SAFE(flow, &PRIV(dev)->flow_list, next, tmp) {
+ TAILQ_REMOVE(&PRIV(dev)->flow_list, flow, next);
+ fs_flow_release(&flow);
+ }
+ return 0;
+}
+
+static int
+fs_flow_query(struct rte_eth_dev *dev,
+ struct rte_flow *flow,
+ enum rte_flow_action_type type,
+ void *arg,
+ struct rte_flow_error *error)
+{
+ struct sub_device *sdev;
+
+ sdev = TX_SUBDEV(dev);
+ if (sdev != NULL) {
+ return rte_flow_query(PORT_ID(sdev),
+ flow->flows[SUB_ID(sdev)], type, arg, error);
+ }
+ WARN("No active sub_device to query about its flow");
+ return -1;
+}
+
+static int
+fs_flow_isolate(struct rte_eth_dev *dev,
+ int set,
+ struct rte_flow_error *error)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (sdev->state < DEV_PROBED)
+ continue;
+ DEBUG("Calling rte_flow_isolate on sub_device %d", i);
+ if (PRIV(dev)->flow_isolated != sdev->flow_isolated)
+ WARN("flow isolation mode of sub_device %d in incoherent state.",
+ i);
+ ret = rte_flow_isolate(PORT_ID(sdev), set, error);
+ if (ret) {
+ ERROR("Operation rte_flow_isolate failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ sdev->flow_isolated = set;
+ }
+ PRIV(dev)->flow_isolated = set;
+ return 0;
+}
+
+const struct rte_flow_ops fs_flow_ops = {
+ .validate = fs_flow_validate,
+ .create = fs_flow_create,
+ .destroy = fs_flow_destroy,
+ .flush = fs_flow_flush,
+ .query = fs_flow_query,
+ .isolate = fs_flow_isolate,
+};
diff --git a/drivers/net/failsafe/failsafe_ops.c b/drivers/net/failsafe/failsafe_ops.c
new file mode 100644
index 00000000..ff9ad155
--- /dev/null
+++ b/drivers/net/failsafe/failsafe_ops.c
@@ -0,0 +1,867 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 <rte_debug.h>
+#include <rte_atomic.h>
+#include <rte_ethdev.h>
+#include <rte_malloc.h>
+#include <rte_flow.h>
+
+#include "failsafe_private.h"
+
+static struct rte_eth_dev_info default_infos = {
+ /* Max possible number of elements */
+ .max_rx_pktlen = UINT32_MAX,
+ .max_rx_queues = RTE_MAX_QUEUES_PER_PORT,
+ .max_tx_queues = RTE_MAX_QUEUES_PER_PORT,
+ .max_mac_addrs = FAILSAFE_MAX_ETHADDR,
+ .max_hash_mac_addrs = UINT32_MAX,
+ .max_vfs = UINT16_MAX,
+ .max_vmdq_pools = UINT16_MAX,
+ .rx_desc_lim = {
+ .nb_max = UINT16_MAX,
+ .nb_min = 0,
+ .nb_align = 1,
+ .nb_seg_max = UINT16_MAX,
+ .nb_mtu_seg_max = UINT16_MAX,
+ },
+ .tx_desc_lim = {
+ .nb_max = UINT16_MAX,
+ .nb_min = 0,
+ .nb_align = 1,
+ .nb_seg_max = UINT16_MAX,
+ .nb_mtu_seg_max = UINT16_MAX,
+ },
+ /*
+ * Set of capabilities that can be verified upon
+ * configuring a sub-device.
+ */
+ .rx_offload_capa =
+ DEV_RX_OFFLOAD_VLAN_STRIP |
+ DEV_RX_OFFLOAD_QINQ_STRIP |
+ DEV_RX_OFFLOAD_IPV4_CKSUM |
+ DEV_RX_OFFLOAD_UDP_CKSUM |
+ DEV_RX_OFFLOAD_TCP_CKSUM |
+ DEV_RX_OFFLOAD_TCP_LRO,
+ .tx_offload_capa = 0x0,
+ .flow_type_rss_offloads = 0x0,
+};
+
+/**
+ * Check whether a specific offloading capability
+ * is supported by a sub_device.
+ *
+ * @return
+ * 0: all requested capabilities are supported by the sub_device
+ * positive value: This flag at least is not supported by the sub_device
+ */
+static int
+fs_port_offload_validate(struct rte_eth_dev *dev,
+ struct sub_device *sdev)
+{
+ struct rte_eth_dev_info infos = {0};
+ struct rte_eth_conf *cf;
+ uint32_t cap;
+
+ cf = &dev->data->dev_conf;
+ SUBOPS(sdev, dev_infos_get)(ETH(sdev), &infos);
+ /* RX capabilities */
+ cap = infos.rx_offload_capa;
+ if (cf->rxmode.hw_vlan_strip &&
+ ((cap & DEV_RX_OFFLOAD_VLAN_STRIP) == 0)) {
+ WARN("VLAN stripping offload requested but not supported by sub_device %d",
+ SUB_ID(sdev));
+ return DEV_RX_OFFLOAD_VLAN_STRIP;
+ }
+ if (cf->rxmode.hw_ip_checksum &&
+ ((cap & (DEV_RX_OFFLOAD_IPV4_CKSUM |
+ DEV_RX_OFFLOAD_UDP_CKSUM |
+ DEV_RX_OFFLOAD_TCP_CKSUM)) !=
+ (DEV_RX_OFFLOAD_IPV4_CKSUM |
+ DEV_RX_OFFLOAD_UDP_CKSUM |
+ DEV_RX_OFFLOAD_TCP_CKSUM))) {
+ WARN("IP checksum offload requested but not supported by sub_device %d",
+ SUB_ID(sdev));
+ return DEV_RX_OFFLOAD_IPV4_CKSUM |
+ DEV_RX_OFFLOAD_UDP_CKSUM |
+ DEV_RX_OFFLOAD_TCP_CKSUM;
+ }
+ if (cf->rxmode.enable_lro &&
+ ((cap & DEV_RX_OFFLOAD_TCP_LRO) == 0)) {
+ WARN("TCP LRO offload requested but not supported by sub_device %d",
+ SUB_ID(sdev));
+ return DEV_RX_OFFLOAD_TCP_LRO;
+ }
+ if (cf->rxmode.hw_vlan_extend &&
+ ((cap & DEV_RX_OFFLOAD_QINQ_STRIP) == 0)) {
+ WARN("Stacked VLAN stripping offload requested but not supported by sub_device %d",
+ SUB_ID(sdev));
+ return DEV_RX_OFFLOAD_QINQ_STRIP;
+ }
+ /* TX capabilities */
+ /* Nothing to do, no tx capa supported */
+ return 0;
+}
+
+/*
+ * Disable the dev_conf flag related to an offload capability flag
+ * within an ethdev configuration.
+ */
+static int
+fs_port_disable_offload(struct rte_eth_conf *cf,
+ uint32_t ol_cap)
+{
+ switch (ol_cap) {
+ case DEV_RX_OFFLOAD_VLAN_STRIP:
+ INFO("Disabling VLAN stripping offload");
+ cf->rxmode.hw_vlan_strip = 0;
+ break;
+ case DEV_RX_OFFLOAD_IPV4_CKSUM:
+ case DEV_RX_OFFLOAD_UDP_CKSUM:
+ case DEV_RX_OFFLOAD_TCP_CKSUM:
+ case (DEV_RX_OFFLOAD_IPV4_CKSUM |
+ DEV_RX_OFFLOAD_UDP_CKSUM |
+ DEV_RX_OFFLOAD_TCP_CKSUM):
+ INFO("Disabling IP checksum offload");
+ cf->rxmode.hw_ip_checksum = 0;
+ break;
+ case DEV_RX_OFFLOAD_TCP_LRO:
+ INFO("Disabling TCP LRO offload");
+ cf->rxmode.enable_lro = 0;
+ break;
+ case DEV_RX_OFFLOAD_QINQ_STRIP:
+ INFO("Disabling stacked VLAN stripping offload");
+ cf->rxmode.hw_vlan_extend = 0;
+ break;
+ default:
+ DEBUG("Unable to disable offload capability: %" PRIx32,
+ ol_cap);
+ return -1;
+ }
+ return 0;
+}
+
+static int
+fs_dev_configure(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int capa_flag;
+ int ret;
+
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (sdev->state != DEV_PROBED)
+ continue;
+ DEBUG("Checking capabilities for sub_device %d", i);
+ while ((capa_flag = fs_port_offload_validate(dev, sdev))) {
+ /*
+ * Refuse to change configuration if multiple devices
+ * are present and we already have configured at least
+ * some of them.
+ */
+ if (PRIV(dev)->state >= DEV_ACTIVE &&
+ PRIV(dev)->subs_tail > 1) {
+ ERROR("device already configured, cannot fix live configuration");
+ return -1;
+ }
+ ret = fs_port_disable_offload(&dev->data->dev_conf,
+ capa_flag);
+ if (ret) {
+ ERROR("Unable to disable offload capability");
+ return ret;
+ }
+ }
+ }
+ FOREACH_SUBDEV(sdev, i, dev) {
+ int rmv_interrupt = 0;
+ int lsc_interrupt = 0;
+ int lsc_enabled;
+
+ if (sdev->state != DEV_PROBED)
+ continue;
+
+ rmv_interrupt = ETH(sdev)->data->dev_flags &
+ RTE_ETH_DEV_INTR_RMV;
+ if (rmv_interrupt) {
+ DEBUG("Enabling RMV interrupts for sub_device %d", i);
+ dev->data->dev_conf.intr_conf.rmv = 1;
+ } else {
+ DEBUG("sub_device %d does not support RMV event", i);
+ }
+ lsc_enabled = dev->data->dev_conf.intr_conf.lsc;
+ lsc_interrupt = lsc_enabled &&
+ (ETH(sdev)->data->dev_flags &
+ RTE_ETH_DEV_INTR_LSC);
+ if (lsc_interrupt) {
+ DEBUG("Enabling LSC interrupts for sub_device %d", i);
+ dev->data->dev_conf.intr_conf.lsc = 1;
+ } else if (lsc_enabled && !lsc_interrupt) {
+ DEBUG("Disabling LSC interrupts for sub_device %d", i);
+ dev->data->dev_conf.intr_conf.lsc = 0;
+ }
+ DEBUG("Configuring sub-device %d", i);
+ sdev->remove = 0;
+ ret = rte_eth_dev_configure(PORT_ID(sdev),
+ dev->data->nb_rx_queues,
+ dev->data->nb_tx_queues,
+ &dev->data->dev_conf);
+ if (ret) {
+ ERROR("Could not configure sub_device %d", i);
+ return ret;
+ }
+ if (rmv_interrupt) {
+ ret = rte_eth_dev_callback_register(PORT_ID(sdev),
+ RTE_ETH_EVENT_INTR_RMV,
+ failsafe_eth_rmv_event_callback,
+ sdev);
+ if (ret)
+ WARN("Failed to register RMV callback for sub_device %d",
+ SUB_ID(sdev));
+ }
+ dev->data->dev_conf.intr_conf.rmv = 0;
+ if (lsc_interrupt) {
+ ret = rte_eth_dev_callback_register(PORT_ID(sdev),
+ RTE_ETH_EVENT_INTR_LSC,
+ failsafe_eth_lsc_event_callback,
+ dev);
+ if (ret)
+ WARN("Failed to register LSC callback for sub_device %d",
+ SUB_ID(sdev));
+ }
+ dev->data->dev_conf.intr_conf.lsc = lsc_enabled;
+ sdev->state = DEV_ACTIVE;
+ }
+ if (PRIV(dev)->state < DEV_ACTIVE)
+ PRIV(dev)->state = DEV_ACTIVE;
+ return 0;
+}
+
+static int
+fs_dev_start(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV(sdev, i, dev) {
+ if (sdev->state != DEV_ACTIVE)
+ continue;
+ DEBUG("Starting sub_device %d", i);
+ ret = rte_eth_dev_start(PORT_ID(sdev));
+ if (ret)
+ return ret;
+ sdev->state = DEV_STARTED;
+ }
+ if (PRIV(dev)->state < DEV_STARTED)
+ PRIV(dev)->state = DEV_STARTED;
+ fs_switch_dev(dev, NULL);
+ return 0;
+}
+
+static void
+fs_dev_stop(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ PRIV(dev)->state = DEV_STARTED - 1;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_STARTED) {
+ rte_eth_dev_stop(PORT_ID(sdev));
+ sdev->state = DEV_STARTED - 1;
+ }
+}
+
+static int
+fs_dev_set_link_up(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_eth_dev_set_link_up on sub_device %d", i);
+ ret = rte_eth_dev_set_link_up(PORT_ID(sdev));
+ if (ret) {
+ ERROR("Operation rte_eth_dev_set_link_up failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static int
+fs_dev_set_link_down(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_eth_dev_set_link_down on sub_device %d", i);
+ ret = rte_eth_dev_set_link_down(PORT_ID(sdev));
+ if (ret) {
+ ERROR("Operation rte_eth_dev_set_link_down failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static void fs_dev_free_queues(struct rte_eth_dev *dev);
+static void
+fs_dev_close(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ failsafe_hotplug_alarm_cancel(dev);
+ if (PRIV(dev)->state == DEV_STARTED)
+ dev->dev_ops->dev_stop(dev);
+ PRIV(dev)->state = DEV_ACTIVE - 1;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Closing sub_device %d", i);
+ rte_eth_dev_close(PORT_ID(sdev));
+ sdev->state = DEV_ACTIVE - 1;
+ }
+ fs_dev_free_queues(dev);
+}
+
+static void
+fs_rx_queue_release(void *queue)
+{
+ struct rte_eth_dev *dev;
+ struct sub_device *sdev;
+ uint8_t i;
+ struct rxq *rxq;
+
+ if (queue == NULL)
+ return;
+ rxq = queue;
+ dev = rxq->priv->dev;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ SUBOPS(sdev, rx_queue_release)
+ (ETH(sdev)->data->rx_queues[rxq->qid]);
+ dev->data->rx_queues[rxq->qid] = NULL;
+ rte_free(rxq);
+}
+
+static int
+fs_rx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t rx_queue_id,
+ uint16_t nb_rx_desc,
+ unsigned int socket_id,
+ const struct rte_eth_rxconf *rx_conf,
+ struct rte_mempool *mb_pool)
+{
+ struct sub_device *sdev;
+ struct rxq *rxq;
+ uint8_t i;
+ int ret;
+
+ rxq = dev->data->rx_queues[rx_queue_id];
+ if (rxq != NULL) {
+ fs_rx_queue_release(rxq);
+ dev->data->rx_queues[rx_queue_id] = NULL;
+ }
+ rxq = rte_zmalloc(NULL,
+ sizeof(*rxq) +
+ sizeof(rte_atomic64_t) * PRIV(dev)->subs_tail,
+ RTE_CACHE_LINE_SIZE);
+ if (rxq == NULL)
+ return -ENOMEM;
+ FOREACH_SUBDEV(sdev, i, dev)
+ rte_atomic64_init(&rxq->refcnt[i]);
+ rxq->qid = rx_queue_id;
+ rxq->socket_id = socket_id;
+ rxq->info.mp = mb_pool;
+ rxq->info.conf = *rx_conf;
+ rxq->info.nb_desc = nb_rx_desc;
+ rxq->priv = PRIV(dev);
+ dev->data->rx_queues[rx_queue_id] = rxq;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ ret = rte_eth_rx_queue_setup(PORT_ID(sdev),
+ rx_queue_id,
+ nb_rx_desc, socket_id,
+ rx_conf, mb_pool);
+ if (ret) {
+ ERROR("RX queue setup failed for sub_device %d", i);
+ goto free_rxq;
+ }
+ }
+ return 0;
+free_rxq:
+ fs_rx_queue_release(rxq);
+ return ret;
+}
+
+static void
+fs_tx_queue_release(void *queue)
+{
+ struct rte_eth_dev *dev;
+ struct sub_device *sdev;
+ uint8_t i;
+ struct txq *txq;
+
+ if (queue == NULL)
+ return;
+ txq = queue;
+ dev = txq->priv->dev;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ SUBOPS(sdev, tx_queue_release)
+ (ETH(sdev)->data->tx_queues[txq->qid]);
+ dev->data->tx_queues[txq->qid] = NULL;
+ rte_free(txq);
+}
+
+static int
+fs_tx_queue_setup(struct rte_eth_dev *dev,
+ uint16_t tx_queue_id,
+ uint16_t nb_tx_desc,
+ unsigned int socket_id,
+ const struct rte_eth_txconf *tx_conf)
+{
+ struct sub_device *sdev;
+ struct txq *txq;
+ uint8_t i;
+ int ret;
+
+ txq = dev->data->tx_queues[tx_queue_id];
+ if (txq != NULL) {
+ fs_tx_queue_release(txq);
+ dev->data->tx_queues[tx_queue_id] = NULL;
+ }
+ txq = rte_zmalloc("ethdev TX queue",
+ sizeof(*txq) +
+ sizeof(rte_atomic64_t) * PRIV(dev)->subs_tail,
+ RTE_CACHE_LINE_SIZE);
+ if (txq == NULL)
+ return -ENOMEM;
+ FOREACH_SUBDEV(sdev, i, dev)
+ rte_atomic64_init(&txq->refcnt[i]);
+ txq->qid = tx_queue_id;
+ txq->socket_id = socket_id;
+ txq->info.conf = *tx_conf;
+ txq->info.nb_desc = nb_tx_desc;
+ txq->priv = PRIV(dev);
+ dev->data->tx_queues[tx_queue_id] = txq;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ ret = rte_eth_tx_queue_setup(PORT_ID(sdev),
+ tx_queue_id,
+ nb_tx_desc, socket_id,
+ tx_conf);
+ if (ret) {
+ ERROR("TX queue setup failed for sub_device %d", i);
+ goto free_txq;
+ }
+ }
+ return 0;
+free_txq:
+ fs_tx_queue_release(txq);
+ return ret;
+}
+
+static void
+fs_dev_free_queues(struct rte_eth_dev *dev)
+{
+ uint16_t i;
+
+ for (i = 0; i < dev->data->nb_rx_queues; i++) {
+ fs_rx_queue_release(dev->data->rx_queues[i]);
+ dev->data->rx_queues[i] = NULL;
+ }
+ dev->data->nb_rx_queues = 0;
+ for (i = 0; i < dev->data->nb_tx_queues; i++) {
+ fs_tx_queue_release(dev->data->tx_queues[i]);
+ dev->data->tx_queues[i] = NULL;
+ }
+ dev->data->nb_tx_queues = 0;
+}
+
+static void
+fs_promiscuous_enable(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ rte_eth_promiscuous_enable(PORT_ID(sdev));
+}
+
+static void
+fs_promiscuous_disable(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ rte_eth_promiscuous_disable(PORT_ID(sdev));
+}
+
+static void
+fs_allmulticast_enable(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ rte_eth_allmulticast_enable(PORT_ID(sdev));
+}
+
+static void
+fs_allmulticast_disable(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ rte_eth_allmulticast_disable(PORT_ID(sdev));
+}
+
+static int
+fs_link_update(struct rte_eth_dev *dev,
+ int wait_to_complete)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling link_update on sub_device %d", i);
+ ret = (SUBOPS(sdev, link_update))(ETH(sdev), wait_to_complete);
+ if (ret && ret != -1) {
+ ERROR("Link update failed for sub_device %d with error %d",
+ i, ret);
+ return ret;
+ }
+ }
+ if (TX_SUBDEV(dev)) {
+ struct rte_eth_link *l1;
+ struct rte_eth_link *l2;
+
+ l1 = &dev->data->dev_link;
+ l2 = &ETH(TX_SUBDEV(dev))->data->dev_link;
+ if (memcmp(l1, l2, sizeof(*l1))) {
+ *l1 = *l2;
+ return 0;
+ }
+ }
+ return -1;
+}
+
+static void
+fs_stats_get(struct rte_eth_dev *dev,
+ struct rte_eth_stats *stats)
+{
+ if (TX_SUBDEV(dev) == NULL)
+ return;
+ rte_eth_stats_get(PORT_ID(TX_SUBDEV(dev)), stats);
+}
+
+static void
+fs_stats_reset(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ rte_eth_stats_reset(PORT_ID(sdev));
+}
+
+/**
+ * Fail-safe dev_infos_get rules:
+ *
+ * No sub_device:
+ * Numerables:
+ * Use the maximum possible values for any field, so as not
+ * to impede any further configuration effort.
+ * Capabilities:
+ * Limits capabilities to those that are understood by the
+ * fail-safe PMD. This understanding stems from the fail-safe
+ * being capable of verifying that the related capability is
+ * expressed within the device configuration (struct rte_eth_conf).
+ *
+ * At least one probed sub_device:
+ * Numerables:
+ * Uses values from the active probed sub_device
+ * The rationale here is that if any sub_device is less capable
+ * (for example concerning the number of queues) than the active
+ * sub_device, then its subsequent configuration will fail.
+ * It is impossible to foresee this failure when the failing sub_device
+ * is supposed to be plugged-in later on, so the configuration process
+ * is the single point of failure and error reporting.
+ * Capabilities:
+ * Uses a logical AND of RX capabilities among
+ * all sub_devices and the default capabilities.
+ * Uses a logical AND of TX capabilities among
+ * the active probed sub_device and the default capabilities.
+ *
+ */
+static void
+fs_dev_infos_get(struct rte_eth_dev *dev,
+ struct rte_eth_dev_info *infos)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ sdev = TX_SUBDEV(dev);
+ if (sdev == NULL) {
+ DEBUG("No probed device, using default infos");
+ rte_memcpy(&PRIV(dev)->infos, &default_infos,
+ sizeof(default_infos));
+ } else {
+ uint32_t rx_offload_capa;
+
+ rx_offload_capa = default_infos.rx_offload_capa;
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_PROBED) {
+ rte_eth_dev_info_get(PORT_ID(sdev),
+ &PRIV(dev)->infos);
+ rx_offload_capa &= PRIV(dev)->infos.rx_offload_capa;
+ }
+ sdev = TX_SUBDEV(dev);
+ rte_eth_dev_info_get(PORT_ID(sdev), &PRIV(dev)->infos);
+ PRIV(dev)->infos.rx_offload_capa = rx_offload_capa;
+ PRIV(dev)->infos.tx_offload_capa &=
+ default_infos.tx_offload_capa;
+ PRIV(dev)->infos.flow_type_rss_offloads &=
+ default_infos.flow_type_rss_offloads;
+ }
+ rte_memcpy(infos, &PRIV(dev)->infos, sizeof(*infos));
+}
+
+static const uint32_t *
+fs_dev_supported_ptypes_get(struct rte_eth_dev *dev)
+{
+ struct sub_device *sdev;
+ struct rte_eth_dev *edev;
+
+ sdev = TX_SUBDEV(dev);
+ if (sdev == NULL)
+ return NULL;
+ edev = ETH(sdev);
+ /* ENOTSUP: counts as no supported ptypes */
+ if (SUBOPS(sdev, dev_supported_ptypes_get) == NULL)
+ return NULL;
+ /*
+ * The API does not permit to do a clean AND of all ptypes,
+ * It is also incomplete by design and we do not really care
+ * to have a best possible value in this context.
+ * We just return the ptypes of the device of highest
+ * priority, usually the PREFERRED device.
+ */
+ return SUBOPS(sdev, dev_supported_ptypes_get)(edev);
+}
+
+static int
+fs_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_eth_dev_set_mtu on sub_device %d", i);
+ ret = rte_eth_dev_set_mtu(PORT_ID(sdev), mtu);
+ if (ret) {
+ ERROR("Operation rte_eth_dev_set_mtu failed for sub_device %d with error %d",
+ i, ret);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static int
+fs_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_eth_dev_vlan_filter on sub_device %d", i);
+ ret = rte_eth_dev_vlan_filter(PORT_ID(sdev), vlan_id, on);
+ if (ret) {
+ ERROR("Operation rte_eth_dev_vlan_filter failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static int
+fs_flow_ctrl_get(struct rte_eth_dev *dev,
+ struct rte_eth_fc_conf *fc_conf)
+{
+ struct sub_device *sdev;
+
+ sdev = TX_SUBDEV(dev);
+ if (sdev == NULL)
+ return 0;
+ if (SUBOPS(sdev, flow_ctrl_get) == NULL)
+ return -ENOTSUP;
+ return SUBOPS(sdev, flow_ctrl_get)(ETH(sdev), fc_conf);
+}
+
+static int
+fs_flow_ctrl_set(struct rte_eth_dev *dev,
+ struct rte_eth_fc_conf *fc_conf)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_eth_dev_flow_ctrl_set on sub_device %d", i);
+ ret = rte_eth_dev_flow_ctrl_set(PORT_ID(sdev), fc_conf);
+ if (ret) {
+ ERROR("Operation rte_eth_dev_flow_ctrl_set failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+static void
+fs_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ /* No check: already done within the rte_eth_dev_mac_addr_remove
+ * call for the fail-safe device.
+ */
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ rte_eth_dev_mac_addr_remove(PORT_ID(sdev),
+ &dev->data->mac_addrs[index]);
+ PRIV(dev)->mac_addr_pool[index] = 0;
+}
+
+static int
+fs_mac_addr_add(struct rte_eth_dev *dev,
+ struct ether_addr *mac_addr,
+ uint32_t index,
+ uint32_t vmdq)
+{
+ struct sub_device *sdev;
+ int ret;
+ uint8_t i;
+
+ RTE_ASSERT(index < FAILSAFE_MAX_ETHADDR);
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ ret = rte_eth_dev_mac_addr_add(PORT_ID(sdev), mac_addr, vmdq);
+ if (ret) {
+ ERROR("Operation rte_eth_dev_mac_addr_add failed for sub_device %"
+ PRIu8 " with error %d", i, ret);
+ return ret;
+ }
+ }
+ if (index >= PRIV(dev)->nb_mac_addr) {
+ DEBUG("Growing mac_addrs array");
+ PRIV(dev)->nb_mac_addr = index;
+ }
+ PRIV(dev)->mac_addr_pool[index] = vmdq;
+ return 0;
+}
+
+static void
+fs_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE)
+ rte_eth_dev_default_mac_addr_set(PORT_ID(sdev), mac_addr);
+}
+
+static int
+fs_filter_ctrl(struct rte_eth_dev *dev,
+ enum rte_filter_type type,
+ enum rte_filter_op op,
+ void *arg)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int ret;
+
+ if (type == RTE_ETH_FILTER_GENERIC &&
+ op == RTE_ETH_FILTER_GET) {
+ *(const void **)arg = &fs_flow_ops;
+ return 0;
+ }
+ FOREACH_SUBDEV_STATE(sdev, i, dev, DEV_ACTIVE) {
+ DEBUG("Calling rte_eth_dev_filter_ctrl on sub_device %d", i);
+ ret = rte_eth_dev_filter_ctrl(PORT_ID(sdev), type, op, arg);
+ if (ret) {
+ ERROR("Operation rte_eth_dev_filter_ctrl failed for sub_device %d"
+ " with error %d", i, ret);
+ return ret;
+ }
+ }
+ return 0;
+}
+
+const struct eth_dev_ops failsafe_ops = {
+ .dev_configure = fs_dev_configure,
+ .dev_start = fs_dev_start,
+ .dev_stop = fs_dev_stop,
+ .dev_set_link_down = fs_dev_set_link_down,
+ .dev_set_link_up = fs_dev_set_link_up,
+ .dev_close = fs_dev_close,
+ .promiscuous_enable = fs_promiscuous_enable,
+ .promiscuous_disable = fs_promiscuous_disable,
+ .allmulticast_enable = fs_allmulticast_enable,
+ .allmulticast_disable = fs_allmulticast_disable,
+ .link_update = fs_link_update,
+ .stats_get = fs_stats_get,
+ .stats_reset = fs_stats_reset,
+ .dev_infos_get = fs_dev_infos_get,
+ .dev_supported_ptypes_get = fs_dev_supported_ptypes_get,
+ .mtu_set = fs_mtu_set,
+ .vlan_filter_set = fs_vlan_filter_set,
+ .rx_queue_setup = fs_rx_queue_setup,
+ .tx_queue_setup = fs_tx_queue_setup,
+ .rx_queue_release = fs_rx_queue_release,
+ .tx_queue_release = fs_tx_queue_release,
+ .flow_ctrl_get = fs_flow_ctrl_get,
+ .flow_ctrl_set = fs_flow_ctrl_set,
+ .mac_addr_remove = fs_mac_addr_remove,
+ .mac_addr_add = fs_mac_addr_add,
+ .mac_addr_set = fs_mac_addr_set,
+ .filter_ctrl = fs_filter_ctrl,
+};
diff --git a/drivers/net/failsafe/failsafe_private.h b/drivers/net/failsafe/failsafe_private.h
new file mode 100644
index 00000000..0361cf43
--- /dev/null
+++ b/drivers/net/failsafe/failsafe_private.h
@@ -0,0 +1,359 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 _RTE_ETH_FAILSAFE_PRIVATE_H_
+#define _RTE_ETH_FAILSAFE_PRIVATE_H_
+
+#include <sys/queue.h>
+
+#include <rte_atomic.h>
+#include <rte_dev.h>
+#include <rte_ethdev.h>
+#include <rte_devargs.h>
+
+#define FAILSAFE_DRIVER_NAME "Fail-safe PMD"
+
+#define PMD_FAILSAFE_MAC_KVARG "mac"
+#define PMD_FAILSAFE_HOTPLUG_POLL_KVARG "hotplug_poll"
+#define PMD_FAILSAFE_PARAM_STRING \
+ "dev(<ifc>)," \
+ "exec(<shell command>)," \
+ "mac=mac_addr," \
+ "hotplug_poll=u64" \
+ ""
+
+#define FAILSAFE_HOTPLUG_DEFAULT_TIMEOUT_MS 2000
+
+#define FAILSAFE_MAX_ETHPORTS 2
+#define FAILSAFE_MAX_ETHADDR 128
+
+/* TYPES */
+
+struct rxq {
+ struct fs_priv *priv;
+ uint16_t qid;
+ /* id of last sub_device polled */
+ uint8_t last_polled;
+ unsigned int socket_id;
+ struct rte_eth_rxq_info info;
+ rte_atomic64_t refcnt[];
+};
+
+struct txq {
+ struct fs_priv *priv;
+ uint16_t qid;
+ unsigned int socket_id;
+ struct rte_eth_txq_info info;
+ rte_atomic64_t refcnt[];
+};
+
+struct rte_flow {
+ TAILQ_ENTRY(rte_flow) next;
+ /* sub_flows */
+ struct rte_flow *flows[FAILSAFE_MAX_ETHPORTS];
+ /* flow description for synchronization */
+ struct rte_flow_desc *fd;
+};
+
+enum dev_state {
+ DEV_UNDEFINED,
+ DEV_PARSED,
+ DEV_PROBED,
+ DEV_ACTIVE,
+ DEV_STARTED,
+};
+
+struct sub_device {
+ /* Exhaustive DPDK device description */
+ struct rte_devargs devargs;
+ struct rte_bus *bus;
+ struct rte_device *dev;
+ struct rte_eth_dev *edev;
+ uint8_t sid;
+ /* Device state machine */
+ enum dev_state state;
+ /* Some device are defined as a command line */
+ char *cmdline;
+ /* fail-safe device backreference */
+ struct rte_eth_dev *fs_dev;
+ /* flag calling for recollection */
+ volatile unsigned int remove:1;
+ /* flow isolation state */
+ int flow_isolated:1;
+};
+
+struct fs_priv {
+ struct rte_eth_dev *dev;
+ /*
+ * Set of sub_devices.
+ * subs[0] is the preferred device
+ * any other is just another slave
+ */
+ struct sub_device *subs;
+ uint8_t subs_head; /* if head == tail, no subs */
+ uint8_t subs_tail; /* first invalid */
+ uint8_t subs_tx; /* current emitting device */
+ uint8_t current_probed;
+ /* flow mapping */
+ TAILQ_HEAD(sub_flows, rte_flow) flow_list;
+ /* current number of mac_addr slots allocated. */
+ uint32_t nb_mac_addr;
+ struct ether_addr mac_addrs[FAILSAFE_MAX_ETHADDR];
+ uint32_t mac_addr_pool[FAILSAFE_MAX_ETHADDR];
+ /* current capabilities */
+ struct rte_eth_dev_info infos;
+ /*
+ * Fail-safe state machine.
+ * This level will be tracking state of the EAL and eth
+ * layer at large as defined by the user application.
+ * It will then steer the sub_devices toward the same
+ * synchronized state.
+ */
+ enum dev_state state;
+ unsigned int pending_alarm:1; /* An alarm is pending */
+ /* flow isolation state */
+ int flow_isolated:1;
+};
+
+/* MISC */
+
+int failsafe_hotplug_alarm_install(struct rte_eth_dev *dev);
+int failsafe_hotplug_alarm_cancel(struct rte_eth_dev *dev);
+
+/* RX / TX */
+
+void set_burst_fn(struct rte_eth_dev *dev, int force_safe);
+
+uint16_t failsafe_rx_burst(void *rxq,
+ struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
+uint16_t failsafe_tx_burst(void *txq,
+ struct rte_mbuf **tx_pkts, uint16_t nb_pkts);
+
+uint16_t failsafe_rx_burst_fast(void *rxq,
+ struct rte_mbuf **rx_pkts, uint16_t nb_pkts);
+uint16_t failsafe_tx_burst_fast(void *txq,
+ struct rte_mbuf **tx_pkts, uint16_t nb_pkts);
+
+/* ARGS */
+
+int failsafe_args_parse(struct rte_eth_dev *dev, const char *params);
+void failsafe_args_free(struct rte_eth_dev *dev);
+int failsafe_args_count_subdevice(struct rte_eth_dev *dev, const char *params);
+int failsafe_args_parse_subs(struct rte_eth_dev *dev);
+
+/* EAL */
+
+int failsafe_eal_init(struct rte_eth_dev *dev);
+int failsafe_eal_uninit(struct rte_eth_dev *dev);
+
+/* ETH_DEV */
+
+int failsafe_eth_dev_state_sync(struct rte_eth_dev *dev);
+void failsafe_dev_remove(struct rte_eth_dev *dev);
+int failsafe_eth_rmv_event_callback(uint8_t port_id,
+ enum rte_eth_event_type type,
+ void *arg, void *out);
+int failsafe_eth_lsc_event_callback(uint8_t port_id,
+ enum rte_eth_event_type event,
+ void *cb_arg, void *out);
+
+/* GLOBALS */
+
+extern const char pmd_failsafe_driver_name[];
+extern const struct eth_dev_ops failsafe_ops;
+extern const struct rte_flow_ops fs_flow_ops;
+extern uint64_t hotplug_poll;
+extern int mac_from_arg;
+
+/* HELPERS */
+
+/* dev: (struct rte_eth_dev *) fail-safe device */
+#define PRIV(dev) \
+ ((struct fs_priv *)(dev)->data->dev_private)
+
+/* sdev: (struct sub_device *) */
+#define ETH(sdev) \
+ ((sdev)->edev)
+
+/* sdev: (struct sub_device *) */
+#define PORT_ID(sdev) \
+ (ETH(sdev)->data->port_id)
+
+/* sdev: (struct sub_device *) */
+#define SUB_ID(sdev) \
+ ((sdev)->sid)
+
+/**
+ * Stateful iterator construct over fail-safe sub-devices:
+ * s: (struct sub_device *), iterator
+ * i: (uint8_t), increment
+ * dev: (struct rte_eth_dev *), fail-safe ethdev
+ * state: (enum dev_state), minimum acceptable device state
+ */
+#define FOREACH_SUBDEV_STATE(s, i, dev, state) \
+ for (i = fs_find_next((dev), 0, state); \
+ i < PRIV(dev)->subs_tail && (s = &PRIV(dev)->subs[i]); \
+ i = fs_find_next((dev), i + 1, state))
+
+/**
+ * Iterator construct over fail-safe sub-devices:
+ * s: (struct sub_device *), iterator
+ * i: (uint8_t), increment
+ * dev: (struct rte_eth_dev *), fail-safe ethdev
+ */
+#define FOREACH_SUBDEV(s, i, dev) \
+ FOREACH_SUBDEV_STATE(s, i, dev, DEV_UNDEFINED)
+
+/* dev: (struct rte_eth_dev *) fail-safe device */
+#define PREFERRED_SUBDEV(dev) \
+ (&PRIV(dev)->subs[0])
+
+/* dev: (struct rte_eth_dev *) fail-safe device */
+#define TX_SUBDEV(dev) \
+ (PRIV(dev)->subs_tx >= PRIV(dev)->subs_tail ? NULL \
+ : (PRIV(dev)->subs[PRIV(dev)->subs_tx].state < DEV_PROBED ? NULL \
+ : &PRIV(dev)->subs[PRIV(dev)->subs_tx]))
+
+/**
+ * s: (struct sub_device *)
+ * ops: (struct eth_dev_ops) member
+ */
+#define SUBOPS(s, ops) \
+ (ETH(s)->dev_ops->ops)
+
+/**
+ * Atomic guard
+ */
+
+/**
+ * a: (rte_atomic64_t)
+ */
+#define FS_ATOMIC_P(a) \
+ rte_atomic64_add(&(a), 1)
+
+/**
+ * a: (rte_atomic64_t)
+ */
+#define FS_ATOMIC_V(a) \
+ rte_atomic64_sub(&(a), 1)
+
+/**
+ * s: (struct sub_device *)
+ * i: uint16_t qid
+ */
+#define FS_ATOMIC_RX(s, i) \
+ rte_atomic64_read( \
+ &((struct rxq *)((s)->fs_dev->data->rx_queues[i]))->refcnt[(s)->sid] \
+ )
+/**
+ * s: (struct sub_device *)
+ * i: uint16_t qid
+ */
+#define FS_ATOMIC_TX(s, i) \
+ rte_atomic64_read( \
+ &((struct txq *)((s)->fs_dev->data->tx_queues[i]))->refcnt[(s)->sid] \
+ )
+
+#define LOG__(level, m, ...) \
+ RTE_LOG(level, PMD, "net_failsafe: " m "%c", __VA_ARGS__)
+#define LOG_(level, ...) LOG__(level, __VA_ARGS__, '\n')
+#define DEBUG(...) LOG_(DEBUG, __VA_ARGS__)
+#define INFO(...) LOG_(INFO, __VA_ARGS__)
+#define WARN(...) LOG_(WARNING, __VA_ARGS__)
+#define ERROR(...) LOG_(ERR, __VA_ARGS__)
+
+/* inlined functions */
+
+static inline uint8_t
+fs_find_next(struct rte_eth_dev *dev, uint8_t sid,
+ enum dev_state min_state)
+{
+ while (sid < PRIV(dev)->subs_tail) {
+ if (PRIV(dev)->subs[sid].state >= min_state)
+ break;
+ sid++;
+ }
+ if (sid >= PRIV(dev)->subs_tail)
+ return PRIV(dev)->subs_tail;
+ return sid;
+}
+
+/*
+ * Switch emitting device.
+ * If banned is set, banned must not be considered for
+ * the role of emitting device.
+ */
+static inline void
+fs_switch_dev(struct rte_eth_dev *dev,
+ struct sub_device *banned)
+{
+ struct sub_device *txd;
+ enum dev_state req_state;
+
+ req_state = PRIV(dev)->state;
+ txd = TX_SUBDEV(dev);
+ if (PREFERRED_SUBDEV(dev)->state >= req_state &&
+ PREFERRED_SUBDEV(dev) != banned) {
+ if (txd != PREFERRED_SUBDEV(dev) &&
+ (txd == NULL ||
+ (req_state == DEV_STARTED) ||
+ (txd && txd->state < DEV_STARTED))) {
+ DEBUG("Switching tx_dev to preferred sub_device");
+ PRIV(dev)->subs_tx = 0;
+ }
+ } else if ((txd && txd->state < req_state) ||
+ txd == NULL ||
+ txd == banned) {
+ struct sub_device *sdev;
+ uint8_t i;
+
+ /* Using acceptable device */
+ FOREACH_SUBDEV_STATE(sdev, i, dev, req_state) {
+ if (sdev == banned)
+ continue;
+ DEBUG("Switching tx_dev to sub_device %d",
+ i);
+ PRIV(dev)->subs_tx = i;
+ break;
+ }
+ } else if (txd && txd->state < req_state) {
+ DEBUG("No device ready, deactivating tx_dev");
+ PRIV(dev)->subs_tx = PRIV(dev)->subs_tail;
+ } else {
+ return;
+ }
+ set_burst_fn(dev, 0);
+ rte_wmb();
+}
+
+#endif /* _RTE_ETH_FAILSAFE_PRIVATE_H_ */
diff --git a/drivers/net/failsafe/failsafe_rxtx.c b/drivers/net/failsafe/failsafe_rxtx.c
new file mode 100644
index 00000000..73114215
--- /dev/null
+++ b/drivers/net/failsafe/failsafe_rxtx.c
@@ -0,0 +1,203 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright 2017 6WIND S.A.
+ * Copyright 2017 Mellanox.
+ *
+ * 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 6WIND S.A. 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 <rte_atomic.h>
+#include <rte_debug.h>
+#include <rte_mbuf.h>
+#include <rte_ethdev.h>
+
+#include "failsafe_private.h"
+
+static inline int
+fs_rx_unsafe(struct sub_device *sdev)
+{
+ return (ETH(sdev) == NULL) ||
+ (ETH(sdev)->rx_pkt_burst == NULL) ||
+ (sdev->state != DEV_STARTED);
+}
+
+static inline int
+fs_tx_unsafe(struct sub_device *sdev)
+{
+ return (sdev == NULL) ||
+ (ETH(sdev) == NULL) ||
+ (ETH(sdev)->tx_pkt_burst == NULL) ||
+ (sdev->state != DEV_STARTED);
+}
+
+void
+set_burst_fn(struct rte_eth_dev *dev, int force_safe)
+{
+ struct sub_device *sdev;
+ uint8_t i;
+ int need_safe;
+ int safe_set;
+
+ need_safe = force_safe;
+ FOREACH_SUBDEV(sdev, i, dev)
+ need_safe |= fs_rx_unsafe(sdev);
+ safe_set = (dev->rx_pkt_burst == &failsafe_rx_burst);
+ if (need_safe && !safe_set) {
+ DEBUG("Using safe RX bursts%s",
+ (force_safe ? " (forced)" : ""));
+ dev->rx_pkt_burst = &failsafe_rx_burst;
+ } else if (!need_safe && safe_set) {
+ DEBUG("Using fast RX bursts");
+ dev->rx_pkt_burst = &failsafe_rx_burst_fast;
+ }
+ need_safe = force_safe || fs_tx_unsafe(TX_SUBDEV(dev));
+ safe_set = (dev->tx_pkt_burst == &failsafe_tx_burst);
+ if (need_safe && !safe_set) {
+ DEBUG("Using safe TX bursts%s",
+ (force_safe ? " (forced)" : ""));
+ dev->tx_pkt_burst = &failsafe_tx_burst;
+ } else if (!need_safe && safe_set) {
+ DEBUG("Using fast TX bursts");
+ dev->tx_pkt_burst = &failsafe_tx_burst_fast;
+ }
+ rte_wmb();
+}
+
+uint16_t
+failsafe_rx_burst(void *queue,
+ struct rte_mbuf **rx_pkts,
+ uint16_t nb_pkts)
+{
+ struct fs_priv *priv;
+ struct sub_device *sdev;
+ struct rxq *rxq;
+ void *sub_rxq;
+ uint16_t nb_rx;
+ uint8_t nb_polled, nb_subs;
+ uint8_t i;
+
+ rxq = queue;
+ priv = rxq->priv;
+ nb_subs = priv->subs_tail - priv->subs_head;
+ nb_polled = 0;
+ for (i = rxq->last_polled; nb_polled < nb_subs; nb_polled++) {
+ i++;
+ if (i == priv->subs_tail)
+ i = priv->subs_head;
+ sdev = &priv->subs[i];
+ if (unlikely(fs_rx_unsafe(sdev)))
+ continue;
+ sub_rxq = ETH(sdev)->data->rx_queues[rxq->qid];
+ FS_ATOMIC_P(rxq->refcnt[sdev->sid]);
+ nb_rx = ETH(sdev)->
+ rx_pkt_burst(sub_rxq, rx_pkts, nb_pkts);
+ FS_ATOMIC_V(rxq->refcnt[sdev->sid]);
+ if (nb_rx) {
+ rxq->last_polled = i;
+ return nb_rx;
+ }
+ }
+ return 0;
+}
+
+uint16_t
+failsafe_rx_burst_fast(void *queue,
+ struct rte_mbuf **rx_pkts,
+ uint16_t nb_pkts)
+{
+ struct fs_priv *priv;
+ struct sub_device *sdev;
+ struct rxq *rxq;
+ void *sub_rxq;
+ uint16_t nb_rx;
+ uint8_t nb_polled, nb_subs;
+ uint8_t i;
+
+ rxq = queue;
+ priv = rxq->priv;
+ nb_subs = priv->subs_tail - priv->subs_head;
+ nb_polled = 0;
+ for (i = rxq->last_polled; nb_polled < nb_subs; nb_polled++) {
+ i++;
+ if (i == priv->subs_tail)
+ i = priv->subs_head;
+ sdev = &priv->subs[i];
+ RTE_ASSERT(!fs_rx_unsafe(sdev));
+ sub_rxq = ETH(sdev)->data->rx_queues[rxq->qid];
+ FS_ATOMIC_P(rxq->refcnt[sdev->sid]);
+ nb_rx = ETH(sdev)->
+ rx_pkt_burst(sub_rxq, rx_pkts, nb_pkts);
+ FS_ATOMIC_V(rxq->refcnt[sdev->sid]);
+ if (nb_rx) {
+ rxq->last_polled = i;
+ return nb_rx;
+ }
+ }
+ return 0;
+}
+
+uint16_t
+failsafe_tx_burst(void *queue,
+ struct rte_mbuf **tx_pkts,
+ uint16_t nb_pkts)
+{
+ struct sub_device *sdev;
+ struct txq *txq;
+ void *sub_txq;
+ uint16_t nb_tx;
+
+ txq = queue;
+ sdev = TX_SUBDEV(txq->priv->dev);
+ if (unlikely(fs_tx_unsafe(sdev)))
+ return 0;
+ sub_txq = ETH(sdev)->data->tx_queues[txq->qid];
+ FS_ATOMIC_P(txq->refcnt[sdev->sid]);
+ nb_tx = ETH(sdev)->tx_pkt_burst(sub_txq, tx_pkts, nb_pkts);
+ FS_ATOMIC_V(txq->refcnt[sdev->sid]);
+ return nb_tx;
+}
+
+uint16_t
+failsafe_tx_burst_fast(void *queue,
+ struct rte_mbuf **tx_pkts,
+ uint16_t nb_pkts)
+{
+ struct sub_device *sdev;
+ struct txq *txq;
+ void *sub_txq;
+ uint16_t nb_tx;
+
+ txq = queue;
+ sdev = TX_SUBDEV(txq->priv->dev);
+ RTE_ASSERT(!fs_tx_unsafe(sdev));
+ sub_txq = ETH(sdev)->data->tx_queues[txq->qid];
+ FS_ATOMIC_P(txq->refcnt[sdev->sid]);
+ nb_tx = ETH(sdev)->tx_pkt_burst(sub_txq, tx_pkts, nb_pkts);
+ FS_ATOMIC_V(txq->refcnt[sdev->sid]);
+ return nb_tx;
+}
diff --git a/drivers/net/failsafe/rte_pmd_failsafe_version.map b/drivers/net/failsafe/rte_pmd_failsafe_version.map
new file mode 100644
index 00000000..b6d2840b
--- /dev/null
+++ b/drivers/net/failsafe/rte_pmd_failsafe_version.map
@@ -0,0 +1,4 @@
+DPDK_17.08 {
+
+ local: *;
+};