[dpdk-dev] [PATCH v11 6/7] eal: add failure handle mechanism for hot-unplug

Jeff Guo jia.guo at intel.com
Tue Oct 2 06:01:34 CEST 2018


hi, constantin

On 10/1/2018 3:46 AM, Ananyev, Konstantin wrote:
> Hi Jeff,
>
>> The mechanism can initially register the sigbus handler after the device
>> event monitor is enabled. When a sigbus event is captured, it will check
>> the failure address and accordingly handle the memory failure of the
>> corresponding device by invoke the hot-unplug handler. It could prevent
>> the application from crashing when a device is hot-unplugged.
>>
>> By this patch, users could call below new added APIs to enable/disable
>> the device hotplug handle mechanism. Note that it just implement the
>> hot-unplug handler in these functions, the other handler of hotplug, such
>> as handler for hotplug binding, could be add in the future if need:
>>    - rte_dev_hotplug_handle_enable
>>    - rte_dev_hotplug_handle_disable
>>
>> Signed-off-by: Jeff Guo <jia.guo at intel.com>
>> Acked-by: Shaopeng He <shaopeng.he at intel.com>
>> ---
>> v11->v10:
>> change some words and change the invoked func name.
>> add experimental tag
>> ---
>>   doc/guides/rel_notes/release_18_08.rst  |   5 +
>>   lib/librte_eal/bsdapp/eal/eal_dev.c     |  14 +++
>>   lib/librte_eal/common/eal_private.h     |  26 +++++
>>   lib/librte_eal/common/include/rte_dev.h |  26 +++++
>>   lib/librte_eal/linuxapp/eal/eal_dev.c   | 162 +++++++++++++++++++++++++++++++-
>>   lib/librte_eal/rte_eal_version.map      |   2 +
>>   6 files changed, 234 insertions(+), 1 deletion(-)
>>
>> diff --git a/doc/guides/rel_notes/release_18_08.rst b/doc/guides/rel_notes/release_18_08.rst
>> index 321fa84..fe0e60f 100644
>> --- a/doc/guides/rel_notes/release_18_08.rst
>> +++ b/doc/guides/rel_notes/release_18_08.rst
>> @@ -117,6 +117,11 @@ New Features
>>
>>     Added support for chained mbufs (input and output).
>>
>> +* **Added hot-unplug handle mechanism.**
>> +
>> +  ``rte_dev_hotplug_handle_enable`` and ``rte_dev_hotplug_handle_disable`` are
>> +  for enabling or disabling hotplug handle mechanism.
>> +
>>
>>   API Changes
>>   -----------
>> diff --git a/lib/librte_eal/bsdapp/eal/eal_dev.c b/lib/librte_eal/bsdapp/eal/eal_dev.c
>> index 1c6c51b..255d611 100644
>> --- a/lib/librte_eal/bsdapp/eal/eal_dev.c
>> +++ b/lib/librte_eal/bsdapp/eal/eal_dev.c
>> @@ -19,3 +19,17 @@ rte_dev_event_monitor_stop(void)
>>   	RTE_LOG(ERR, EAL, "Device event is not supported for FreeBSD\n");
>>   	return -1;
>>   }
>> +
>> +int __rte_experimental
>> +rte_dev_hotplug_handle_enable(void)
>> +{
>> +	RTE_LOG(ERR, EAL, "Device event is not supported for FreeBSD\n");
>> +	return -1;
>> +}
>> +
>> +int __rte_experimental
>> +rte_dev_hotplug_handle_disable(void)
>> +{
>> +	RTE_LOG(ERR, EAL, "Device event is not supported for FreeBSD\n");
>> +	return -1;
>> +}
>> diff --git a/lib/librte_eal/common/eal_private.h b/lib/librte_eal/common/eal_private.h
>> index a2d1528..637f20d 100644
>> --- a/lib/librte_eal/common/eal_private.h
>> +++ b/lib/librte_eal/common/eal_private.h
>> @@ -317,4 +317,30 @@ rte_devargs_layers_parse(struct rte_devargs *devargs,
>>    */
>>   int rte_bus_sigbus_handler(const void *failure_addr);
>>
>> +/**
>> + * @warning
>> + * @b EXPERIMENTAL: this API may change without prior notice
>> + *
>> + * Register the sigbus handler.
>> + *
>> + * @return
>> + *   - On success, zero.
>> + *   - On failure, a negative value.
>> + */
>> +int
>> +rte_dev_sigbus_handler_register(void);
>> +
>> +/**
>> + * @warning
>> + * @b EXPERIMENTAL: this API may change without prior notice
>> + *
>> + * Unregister the sigbus handler.
>> + *
>> + * @return
>> + *   - On success, zero.
>> + *   - On failure, a negative value.
>> + */
>> +int
>> +rte_dev_sigbus_handler_unregister(void);
>> +
>>   #endif /* _EAL_PRIVATE_H_ */
>> diff --git a/lib/librte_eal/common/include/rte_dev.h b/lib/librte_eal/common/include/rte_dev.h
>> index b80a805..ff580a0 100644
>> --- a/lib/librte_eal/common/include/rte_dev.h
>> +++ b/lib/librte_eal/common/include/rte_dev.h
>> @@ -460,4 +460,30 @@ rte_dev_event_monitor_start(void);
>>   int __rte_experimental
>>   rte_dev_event_monitor_stop(void);
>>
>> +/**
>> + * @warning
>> + * @b EXPERIMENTAL: this API may change without prior notice
>> + *
>> + * Enable hotplug handling for devices.
>> + *
>> + * @return
>> + *   - On success, zero.
>> + *   - On failure, a negative value.
>> + */
>> +int __rte_experimental
>> +rte_dev_hotplug_handle_enable(void);
>> +
>> +/**
>> + * @warning
>> + * @b EXPERIMENTAL: this API may change without prior notice
>> + *
>> + * Disable hotplug handling for devices.
>> + *
>> + * @return
>> + *   - On success, zero.
>> + *   - On failure, a negative value.
>> + */
>> +int __rte_experimental
>> +rte_dev_hotplug_handle_disable(void);
>> +
>>   #endif /* _RTE_DEV_H_ */
>> diff --git a/lib/librte_eal/linuxapp/eal/eal_dev.c b/lib/librte_eal/linuxapp/eal/eal_dev.c
>> index 1cf6aeb..14b18d8 100644
>> --- a/lib/librte_eal/linuxapp/eal/eal_dev.c
>> +++ b/lib/librte_eal/linuxapp/eal/eal_dev.c
>> @@ -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 +16,32 @@
>>   #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;
>> +static bool hotplug_handle;
>>
>>   #define EAL_UEV_MSG_LEN 4096
>>   #define EAL_UEV_MSG_ELEM_LEN 128
>>
>> +/*
>> + * spinlock for device hot-unplug failure handling. If it try to access bus or
>> + * device, such as handle sigbus on bus or handle memory failure for device
>> + * just need to use this lock. It could protect the bus and the device to avoid
>> + * race condition.
>> + */
>> +static rte_spinlock_t failure_handle_lock = RTE_SPINLOCK_INITIALIZER;
>> +
>> +static struct sigaction sigbus_action_old;
>> +
>> +static int sigbus_need_recover;
>> +
>>   static void dev_uev_handler(__rte_unused void *param);
>>
>>   /* identify the system layer which reports this event. */
>> @@ -33,6 +52,49 @@ enum eal_dev_event_subsystem {
>>   	EAL_DEV_EVENT_SUBSYSTEM_MAX
>>   };
>>
>> +static void
>> +sigbus_action_recover(void)
>> +{
>> +	if (sigbus_need_recover) {
>> +		sigaction(SIGBUS, &sigbus_action_old, NULL);
>> +		sigbus_need_recover = 0;
>> +	}
>> +}
>> +
>> +static void sigbus_handler(int signum, siginfo_t *info,
>> +				void *ctx __rte_unused)
>> +{
>> +	int ret;
>> +
>> +	RTE_LOG(INFO, EAL, "Thread[%d] catch SIGBUS, fault address:%p\n",
>> +		(int)pthread_self(), info->si_addr);
>> +
>> +	rte_spinlock_lock(&failure_handle_lock);
>> +	ret = rte_bus_sigbus_handler(info->si_addr);
>> +	rte_spinlock_unlock(&failure_handle_lock);
>> +	if (ret == -1) {
>> +		rte_exit(EXIT_FAILURE,
>> +			 "Failed to handle SIGBUS for hot-unplug, "
>> +			 "(rte_errno: %s)!", strerror(rte_errno));
>> +	} else if (ret == 1) {
>> +		if (sigbus_action_old.sa_handler)
>> +			(*(sigbus_action_old.sa_handler))(signum);
>> +		else
>> +			rte_exit(EXIT_FAILURE,
>> +				 "Failed to handle generic SIGBUS!");
>> +	}
>> +
>> +	RTE_LOG(INFO, EAL, "Success to handle SIGBUS for hot-unplug!\n");
>> +}
>> +
>> +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 +209,9 @@ 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);
>> @@ -171,8 +236,43 @@ 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 && hotplug_handle) {
>> +			rte_spinlock_lock(&failure_handle_lock);
>> +			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;
>> +			}
>> +
>> +			ret = bus->hot_unplug_handler(dev);
>> +			rte_spinlock_unlock(&failure_handle_lock);
>> +			if (ret) {
>> +				RTE_LOG(ERR, EAL, "Can not handle hot-unplug "
>> +					"for device (%s)\n", dev->name);
>> +				return;
>> +			}
>> +		}
>>   		dev_callback_process(uevent.devname, uevent.type);
>> +	}
>>   }
>>
>>   int __rte_experimental
>> @@ -220,5 +320,65 @@ rte_dev_event_monitor_stop(void)
>>   	close(intr_handle.fd);
>>   	intr_handle.fd = -1;
>>   	monitor_started = false;
>> +
>>   	return 0;
>>   }
>> +
>> +int __rte_experimental
>> +rte_dev_sigbus_handler_register(void)
>> +{
>> +	sigset_t mask;
>> +	struct sigaction action;
>> +
>> +	rte_errno = 0;
>> +
> Shouldn't you call sigaction only if sigbus_need_recover == 0?


i guess what you mean is that the sigbus_need_recover is need check 
before call sigaction, because the register could be call many times, if 
so, i think it is make sense to check it every time.


>> +	sigemptyset(&mask);
>> +	sigaddset(&mask, SIGBUS);
>> +	action.sa_flags = SA_SIGINFO;
>> +	action.sa_mask = mask;
>> +	action.sa_sigaction = sigbus_handler;
>> +	sigbus_need_recover = !sigaction(SIGBUS, &action, &sigbus_action_old);
>> +
>> +	return rte_errno;
>> +}
>> +
>> +int __rte_experimental
>> +rte_dev_sigbus_handler_unregister(void)
>> +{
>> +	rte_errno = 0;
>> +	sigbus_need_recover = 1;
> Hmm, why to set sugbus_need_recover to 1 here?
> If user called rte_dev_sigbus_handler_register() before, and it was successful, it already would be 1.
> In other cases, you probably don't have to do anything.
> Konstantin


you are right, it should let each sigaction calling to manage this macro 
but no other more place, thanks.


>> +
>> +	sigbus_action_recover();
>> +
>> +	return rte_errno;
>> +}
>> +
>> +int __rte_experimental
>> +rte_dev_hotplug_handle_enable(void)
>> +{
>> +	int ret = 0;
>> +
>> +	ret = rte_dev_sigbus_handler_register();
>> +	if (ret < 0)
>> +		RTE_LOG(ERR, EAL, "fail to register sigbus handler for "
>> +			"devices.\n");
>> +
>> +	hotplug_handle = true;
>> +
>> +	return ret;
>> +}
>> +
>> +int __rte_experimental
>> +rte_dev_hotplug_handle_disable(void)
>> +{
>> +	int ret = 0;
>> +
>> +	ret = rte_dev_sigbus_handler_unregister();
>> +	if (ret < 0)
>> +		RTE_LOG(ERR, EAL, "fail to unregister sigbus handler for "
>> +			"devices.\n");
>> +
>> +	hotplug_handle = false;
>> +
>> +	return ret;
>> +}
>> diff --git a/lib/librte_eal/rte_eal_version.map b/lib/librte_eal/rte_eal_version.map
>> index 73282bb..a3255aa 100644
>> --- a/lib/librte_eal/rte_eal_version.map
>> +++ b/lib/librte_eal/rte_eal_version.map
>> @@ -281,6 +281,8 @@ EXPERIMENTAL {
>>   	rte_dev_event_callback_unregister;
>>   	rte_dev_event_monitor_start;
>>   	rte_dev_event_monitor_stop;
>> +	rte_dev_hotplug_handle_enable;
>> +	rte_dev_hotplug_handle_disable;
>>   	rte_dev_iterator_init;
>>   	rte_dev_iterator_next;
>>   	rte_devargs_add;
>> --
>> 2.7.4


More information about the dev mailing list