@@ -10,6 +10,8 @@
#include <time.h>
#include <ethdev_driver.h>
+#include <rte_interrupts.h>
+#include <rte_cycles.h>
#include <mlx5_glue.h>
#include <mlx5_common.h>
@@ -370,6 +372,220 @@ mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete)
return ret;
}
+/**
+ * Handle asynchronous removal event for entire multiport device.
+ *
+ * @param sh
+ * Infiniband device shared context.
+ */
+static void
+mlx5_dev_interrupt_device_fatal(struct mlx5_dev_ctx_shared *sh)
+{
+ uint32_t i;
+
+ for (i = 0; i < sh->max_port; ++i) {
+ struct rte_eth_dev *dev;
+
+ if (sh->port[i].ih_port_id >= RTE_MAX_ETHPORTS) {
+ /*
+ * Or not existing port either no
+ * handler installed for this port.
+ */
+ continue;
+ }
+ dev = &rte_eth_devices[sh->port[i].ih_port_id];
+ MLX5_ASSERT(dev);
+ if (dev->data->dev_conf.intr_conf.rmv)
+ rte_eth_dev_callback_process
+ (dev, RTE_ETH_EVENT_INTR_RMV, NULL);
+ }
+}
+
+/**
+ * Handle shared asynchronous events the NIC (removal event
+ * and link status change). Supports multiport IB device.
+ *
+ * @param cb_arg
+ * Callback argument.
+ */
+void
+mlx5_dev_interrupt_handler(void *cb_arg)
+{
+ struct mlx5_dev_ctx_shared *sh = cb_arg;
+ struct ibv_async_event event;
+
+ /* Read all message from the IB device and acknowledge them. */
+ for (;;) {
+ struct rte_eth_dev *dev;
+ uint32_t tmp;
+
+ if (mlx5_glue->get_async_event(sh->ctx, &event))
+ break;
+ /* Retrieve and check IB port index. */
+ tmp = (uint32_t)event.element.port_num;
+ if (!tmp && event.event_type == IBV_EVENT_DEVICE_FATAL) {
+ /*
+ * The DEVICE_FATAL event is called once for
+ * entire device without port specifying.
+ * We should notify all existing ports.
+ */
+ mlx5_glue->ack_async_event(&event);
+ mlx5_dev_interrupt_device_fatal(sh);
+ continue;
+ }
+ MLX5_ASSERT(tmp && (tmp <= sh->max_port));
+ if (!tmp) {
+ /* Unsupported device level event. */
+ mlx5_glue->ack_async_event(&event);
+ DRV_LOG(DEBUG,
+ "unsupported common event (type %d)",
+ event.event_type);
+ continue;
+ }
+ if (tmp > sh->max_port) {
+ /* Invalid IB port index. */
+ mlx5_glue->ack_async_event(&event);
+ DRV_LOG(DEBUG,
+ "cannot handle an event (type %d)"
+ "due to invalid IB port index (%u)",
+ event.event_type, tmp);
+ continue;
+ }
+ if (sh->port[tmp - 1].ih_port_id >= RTE_MAX_ETHPORTS) {
+ /* No handler installed. */
+ mlx5_glue->ack_async_event(&event);
+ DRV_LOG(DEBUG,
+ "cannot handle an event (type %d)"
+ "due to no handler installed for port %u",
+ event.event_type, tmp);
+ continue;
+ }
+ /* Retrieve ethernet device descriptor. */
+ tmp = sh->port[tmp - 1].ih_port_id;
+ dev = &rte_eth_devices[tmp];
+ MLX5_ASSERT(dev);
+ if ((event.event_type == IBV_EVENT_PORT_ACTIVE ||
+ event.event_type == IBV_EVENT_PORT_ERR) &&
+ dev->data->dev_conf.intr_conf.lsc) {
+ mlx5_glue->ack_async_event(&event);
+ if (mlx5_link_update(dev, 0) == -EAGAIN) {
+ usleep(0);
+ continue;
+ }
+ rte_eth_dev_callback_process
+ (dev, RTE_ETH_EVENT_INTR_LSC, NULL);
+ continue;
+ }
+ DRV_LOG(DEBUG,
+ "port %u cannot handle an unknown event (type %d)",
+ dev->data->port_id, event.event_type);
+ mlx5_glue->ack_async_event(&event);
+ }
+}
+
+/*
+ * Unregister callback handler safely. The handler may be active
+ * while we are trying to unregister it, in this case code -EAGAIN
+ * is returned by rte_intr_callback_unregister(). This routine checks
+ * the return code and tries to unregister handler again.
+ *
+ * @param handle
+ * interrupt handle
+ * @param cb_fn
+ * pointer to callback routine
+ * @cb_arg
+ * opaque callback parameter
+ */
+void
+mlx5_intr_callback_unregister(const struct rte_intr_handle *handle,
+ rte_intr_callback_fn cb_fn, void *cb_arg)
+{
+ /*
+ * Try to reduce timeout management overhead by not calling
+ * the timer related routines on the first iteration. If the
+ * unregistering succeeds on first call there will be no
+ * timer calls at all.
+ */
+ uint64_t twait = 0;
+ uint64_t start = 0;
+
+ do {
+ int ret;
+
+ ret = rte_intr_callback_unregister(handle, cb_fn, cb_arg);
+ if (ret >= 0)
+ return;
+ if (ret != -EAGAIN) {
+ DRV_LOG(INFO, "failed to unregister interrupt"
+ " handler (error: %d)", ret);
+ MLX5_ASSERT(false);
+ return;
+ }
+ if (twait) {
+ struct timespec onems;
+
+ /* Wait one millisecond and try again. */
+ onems.tv_sec = 0;
+ onems.tv_nsec = NS_PER_S / MS_PER_S;
+ nanosleep(&onems, 0);
+ /* Check whether one second elapsed. */
+ if ((rte_get_timer_cycles() - start) <= twait)
+ continue;
+ } else {
+ /*
+ * We get the amount of timer ticks for one second.
+ * If this amount elapsed it means we spent one
+ * second in waiting. This branch is executed once
+ * on first iteration.
+ */
+ twait = rte_get_timer_hz();
+ MLX5_ASSERT(twait);
+ }
+ /*
+ * Timeout elapsed, show message (once a second) and retry.
+ * We have no other acceptable option here, if we ignore
+ * the unregistering return code the handler will not
+ * be unregistered, fd will be closed and we may get the
+ * crush. Hanging and messaging in the loop seems not to be
+ * the worst choice.
+ */
+ DRV_LOG(INFO, "Retrying to unregister interrupt handler");
+ start = rte_get_timer_cycles();
+ } while (true);
+}
+
+/**
+ * Handle DEVX interrupts from the NIC.
+ * This function is probably called from the DPDK host thread.
+ *
+ * @param cb_arg
+ * Callback argument.
+ */
+void
+mlx5_dev_interrupt_handler_devx(void *cb_arg)
+{
+#ifndef HAVE_IBV_DEVX_ASYNC
+ (void)cb_arg;
+ return;
+#else
+ struct mlx5_dev_ctx_shared *sh = cb_arg;
+ union {
+ struct mlx5dv_devx_async_cmd_hdr cmd_resp;
+ uint8_t buf[MLX5_ST_SZ_BYTES(query_flow_counter_out) +
+ MLX5_ST_SZ_BYTES(traffic_counter) +
+ sizeof(struct mlx5dv_devx_async_cmd_hdr)];
+ } out;
+ uint8_t *buf = out.buf + sizeof(out.cmd_resp);
+
+ while (!mlx5_glue->devx_get_async_cmd_comp(sh->devx_comp,
+ &out.cmd_resp,
+ sizeof(out.buf)))
+ mlx5_flow_async_pool_query_handle
+ (sh, (uint64_t)out.cmd_resp.wr_id,
+ mlx5_devx_get_out_command_status(buf));
+#endif /* HAVE_IBV_DEVX_ASYNC */
+}
+
/**
* DPDK callback to bring the link DOWN.
*
@@ -10,9 +10,11 @@
#include <stdint.h>
#include <stdlib.h>
#include <errno.h>
+#include <fcntl.h>
#include <rte_malloc.h>
#include <ethdev_driver.h>
+#include <rte_common.h>
#include <rte_eal_paging.h>
#include <mlx5_glue.h>
@@ -28,6 +30,7 @@
#include "mlx5_mr.h"
#include "mlx5_verbs.h"
#include "mlx5_inet.h"
+#include "mlx5_devx.h"
/**
* Get mlx5 device attributes. The glue function query_device_ex() is called
@@ -242,6 +245,88 @@ mlx5_os_open_device(const struct mlx5_dev_spawn_data *spawn,
return err;
}
+/**
+ * Install shared asynchronous device events handler.
+ * This function is implemented to support event sharing
+ * between multiple ports of single IB device.
+ *
+ * @param sh
+ * Pointer to mlx5_dev_ctx_shared object.
+ */
+void
+mlx5_os_dev_shared_handler_install(struct mlx5_dev_ctx_shared *sh)
+{
+ int ret;
+ int flags;
+
+ sh->intr_handle.fd = -1;
+ flags = fcntl(((struct ibv_context *)sh->ctx)->async_fd, F_GETFL);
+ ret = fcntl(((struct ibv_context *)sh->ctx)->async_fd,
+ F_SETFL, flags | O_NONBLOCK);
+ if (ret) {
+ DRV_LOG(INFO, "failed to change file descriptor async event"
+ " queue");
+ } else {
+ sh->intr_handle.fd = ((struct ibv_context *)sh->ctx)->async_fd;
+ sh->intr_handle.type = RTE_INTR_HANDLE_EXT;
+ if (rte_intr_callback_register(&sh->intr_handle,
+ mlx5_dev_interrupt_handler, sh)) {
+ DRV_LOG(INFO, "Fail to install the shared interrupt.");
+ sh->intr_handle.fd = -1;
+ }
+ }
+ if (sh->devx) {
+#ifdef HAVE_IBV_DEVX_ASYNC
+ sh->intr_handle_devx.fd = -1;
+ sh->devx_comp =
+ (void *)mlx5_glue->devx_create_cmd_comp(sh->ctx);
+ struct mlx5dv_devx_cmd_comp *devx_comp = sh->devx_comp;
+ if (!devx_comp) {
+ DRV_LOG(INFO, "failed to allocate devx_comp.");
+ return;
+ }
+ flags = fcntl(devx_comp->fd, F_GETFL);
+ ret = fcntl(devx_comp->fd, F_SETFL, flags | O_NONBLOCK);
+ if (ret) {
+ DRV_LOG(INFO, "failed to change file descriptor"
+ " devx comp");
+ return;
+ }
+ sh->intr_handle_devx.fd = devx_comp->fd;
+ sh->intr_handle_devx.type = RTE_INTR_HANDLE_EXT;
+ if (rte_intr_callback_register(&sh->intr_handle_devx,
+ mlx5_dev_interrupt_handler_devx, sh)) {
+ DRV_LOG(INFO, "Fail to install the devx shared"
+ " interrupt.");
+ sh->intr_handle_devx.fd = -1;
+ }
+#endif /* HAVE_IBV_DEVX_ASYNC */
+ }
+}
+
+/**
+ * Uninstall shared asynchronous device events handler.
+ * This function is implemented to support event sharing
+ * between multiple ports of single IB device.
+ *
+ * @param dev
+ * Pointer to mlx5_dev_ctx_shared object.
+ */
+void
+mlx5_os_dev_shared_handler_uninstall(struct mlx5_dev_ctx_shared *sh)
+{
+ if (sh->intr_handle.fd >= 0)
+ mlx5_intr_callback_unregister(&sh->intr_handle,
+ mlx5_dev_interrupt_handler, sh);
+#ifdef HAVE_IBV_DEVX_ASYNC
+ if (sh->intr_handle_devx.fd >= 0)
+ rte_intr_callback_unregister(&sh->intr_handle_devx,
+ mlx5_dev_interrupt_handler_devx, sh);
+ if (sh->devx_comp)
+ mlx5_glue->devx_destroy_cmd_comp(sh->devx_comp);
+#endif
+}
+
/**
* Set the reg_mr and dereg_mr call backs
*