This patch introduces a failure handler mechanism to handle device
hot plug removal event.
First register sigbus handler, once sigbus error be captured, will
check the failure address and accordingly remap the invalid memory
for the corresponding device. Bese on this mechanism, it could
guaranty the application not to be crash when hot unplug devices.
Signed-off-by: Jeff Guo <jia.guo@intel.com>
---
v4->v3:
split patches to be small and clear.
---
lib/librte_eal/linuxapp/eal/eal_dev.c | 88 ++++++++++++++++++++++++++++++++++-
1 file changed, 87 insertions(+), 1 deletion(-)
@@ -4,6 +4,8 @@
#include <string.h>
#include <unistd.h>
+#include <fcntl.h>
+#include <signal.h>
#include <sys/socket.h>
#include <linux/netlink.h>
@@ -14,15 +16,24 @@
#include <rte_malloc.h>
#include <rte_interrupts.h>
#include <rte_alarm.h>
+#include <rte_bus.h>
+#include <rte_eal.h>
+#include <rte_spinlock.h>
+#include <rte_errno.h>
#include "eal_private.h"
static struct rte_intr_handle intr_handle = {.fd = -1 };
static bool monitor_started;
+extern struct rte_bus_list rte_bus_list;
+
#define EAL_UEV_MSG_LEN 4096
#define EAL_UEV_MSG_ELEM_LEN 128
+/* spinlock for device failure process */
+static rte_spinlock_t dev_failure_lock = RTE_SPINLOCK_INITIALIZER;
+
static void dev_uev_handler(__rte_unused void *param);
/* identify the system layer which reports this event. */
@@ -33,6 +44,34 @@ enum eal_dev_event_subsystem {
EAL_DEV_EVENT_SUBSYSTEM_MAX
};
+static void sigbus_handler(int signum __rte_unused, siginfo_t *info,
+ void *ctx __rte_unused)
+{
+ int ret;
+
+ RTE_LOG(DEBUG, EAL, "Thread[%d] catch SIGBUS, fault address:%p\n",
+ (int)pthread_self(), info->si_addr);
+
+ rte_spinlock_lock(&dev_failure_lock);
+ ret = rte_bus_sigbus_handler(info->si_addr);
+ rte_spinlock_unlock(&dev_failure_lock);
+ if (!ret)
+ RTE_LOG(INFO, EAL,
+ "Success to handle SIGBUS error for hotplug!\n");
+ else
+ rte_exit(EXIT_FAILURE,
+ "A generic SIGBUS error, (rte_errno: %s)!",
+ strerror(rte_errno));
+}
+
+static int cmp_dev_name(const struct rte_device *dev,
+ const void *_name)
+{
+ const char *name = _name;
+
+ return strcmp(dev->name, name);
+}
+
static int
dev_uev_socket_fd_create(void)
{
@@ -147,6 +186,9 @@ dev_uev_handler(__rte_unused void *param)
struct rte_dev_event uevent;
int ret;
char buf[EAL_UEV_MSG_LEN];
+ struct rte_bus *bus;
+ struct rte_device *dev;
+ const char *busname;
memset(&uevent, 0, sizeof(struct rte_dev_event));
memset(buf, 0, EAL_UEV_MSG_LEN);
@@ -171,13 +213,48 @@ dev_uev_handler(__rte_unused void *param)
RTE_LOG(DEBUG, EAL, "receive uevent(name:%s, type:%d, subsystem:%d)\n",
uevent.devname, uevent.type, uevent.subsystem);
- if (uevent.devname)
+ switch (uevent.subsystem) {
+ case EAL_DEV_EVENT_SUBSYSTEM_PCI:
+ case EAL_DEV_EVENT_SUBSYSTEM_UIO:
+ busname = "pci";
+ break;
+ default:
+ break;
+ }
+
+ if (uevent.devname) {
+ if (uevent.type == RTE_DEV_EVENT_REMOVE) {
+ bus = rte_bus_find_by_name(busname);
+ if (bus == NULL) {
+ RTE_LOG(ERR, EAL, "Cannot find bus (%s)\n",
+ busname);
+ return;
+ }
+ dev = bus->find_device(NULL, cmp_dev_name,
+ uevent.devname);
+ if (dev == NULL) {
+ RTE_LOG(ERR, EAL, "Cannot find device (%s) on "
+ "bus (%s)\n", uevent.devname, busname);
+ return;
+ }
+ rte_spinlock_lock(&dev_failure_lock);
+ ret = bus->hotplug_handler(dev);
+ rte_spinlock_unlock(&dev_failure_lock);
+ if (ret) {
+ RTE_LOG(ERR, EAL, "Can not handle hotplug for "
+ "device (%s)\n", dev->name);
+ return;
+ }
+ }
dev_callback_process(uevent.devname, uevent.type);
+ }
}
int __rte_experimental
rte_dev_event_monitor_start(void)
{
+ sigset_t mask;
+ struct sigaction action;
int ret;
if (monitor_started)
@@ -197,6 +274,14 @@ rte_dev_event_monitor_start(void)
return -1;
}
+ /* register sigbus handler */
+ sigemptyset(&mask);
+ sigaddset(&mask, SIGBUS);
+ action.sa_flags = SA_SIGINFO;
+ action.sa_mask = mask;
+ action.sa_sigaction = sigbus_handler;
+ sigaction(SIGBUS, &action, NULL);
+
monitor_started = true;
return 0;
@@ -220,5 +305,6 @@ rte_dev_event_monitor_stop(void)
close(intr_handle.fd);
intr_handle.fd = -1;
monitor_started = false;
+
return 0;
}