[dpdk-dev] [PATCH V20 2/4] eal: add failure handler mechanism for hot plug

[Guo, Jia]

On 4/20/2018 7:14 PM, Ananyev, Konstantin wrote:
>
>> This patch introduces a failure handler mechanism to handle device
>> hot unplug event. When device be hot plug out, the device resource
>> become invalid, if this resource is still be unexpected read/write,
>> system will crash. This patch let eal help application to handle
>> this fault, when sigbus error occur, check the failure address and
>> accordingly remap the invalid memory for the corresponding device,
>> that could guaranty the application not to be shut down when hot plug.
>>
>> Signed-off-by: Jeff Guo <jia.guo at intel.com>
>> ---
>> v20->v19:
>> refine the logic of remapping for multiple device.
>> ---
>>   doc/guides/rel_notes/release_18_05.rst  |   6 ++
>>   lib/librte_eal/common/include/rte_dev.h |  11 +++
>>   lib/librte_eal/linuxapp/eal/eal_dev.c   | 124 +++++++++++++++++++++++++++++++-
>>   lib/librte_eal/rte_eal_version.map      |   1 +
>>   4 files changed, 141 insertions(+), 1 deletion(-)
>>
>> diff --git a/doc/guides/rel_notes/release_18_05.rst b/doc/guides/rel_notes/release_18_05.rst
>> index a018ef5..a4ea9af 100644
>> --- a/doc/guides/rel_notes/release_18_05.rst
>> +++ b/doc/guides/rel_notes/release_18_05.rst
>> @@ -70,6 +70,12 @@ New Features
>>
>>     Linux uevent is supported as backend of this device event notification framework.
>>
>> +* **Added hot plug failure handler.**
>> +
>> +  Added a failure handler machenism to handle hot unplug device.
>> +
>> +  * ``rte_dev_handle_hot_unplug`` for handle hot unplug device failure.
>> +
>>
>>   API Changes
>>   -----------
>> diff --git a/lib/librte_eal/common/include/rte_dev.h b/lib/librte_eal/common/include/rte_dev.h
>> index 0955e9a..9933131 100644
>> --- a/lib/librte_eal/common/include/rte_dev.h
>> +++ b/lib/librte_eal/common/include/rte_dev.h
>> @@ -360,4 +360,15 @@ 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
>> + *
>> + * It can be used to handle the device signal bus error. when signal bus error
>> + * occur, the handler would check the failure address to find the corresponding
>> + * device and remap the memory resource of the device, that would guaranty
>> + * the system not crash when the device be hot unplug.
>> + */
>> +void __rte_experimental
>> +rte_dev_handle_hot_unplug(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 9478a39..33e7026 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>
>>
>> @@ -13,12 +15,16 @@
>>   #include <rte_malloc.h>
>>   #include <rte_interrupts.h>
>>   #include <rte_alarm.h>
>> +#include <rte_bus.h>
>> +#include <rte_eal.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
>>
>> @@ -33,6 +39,68 @@ enum eal_dev_event_subsystem {
>>   };
>>
>>   static int
>> +dev_uev_failure_process(struct rte_device *dev, void *dev_addr)
>> +{
>> +	struct rte_bus *bus;
>> +	int ret = 0;
>> +
>> +	if (!dev && !dev_addr) {
>> +		return -EINVAL;
>> +	} else if (dev) {
>> +		bus = rte_bus_find_by_device_name(dev->name);
>> +		if (bus->handle_hot_unplug) {
>> +			/**
>> +			 * call bus ops to handle hot unplug.
>> +			 */
>> +			ret = bus->handle_hot_unplug(dev, dev_addr);
>> +			if (ret) {
>> +				RTE_LOG(ERR, EAL,
>> +					"It cannot handle hot unplug "
>> +					"for device (%s) "
>> +					"on the bus.\n ",
>> +					dev->name);
>> +			}
>> +		}
>
> You would retrun 0 if bus->handle_hot_unplug == NULL.
> Is that intended?
> Shouldn't be I think.
shouldn't be. will modify it.
>> +	} else {
>> +		TAILQ_FOREACH(bus, &rte_bus_list, next) {
>> +			if (bus->handle_hot_unplug) {
>> +				/**
>> +				 * call bus ops to handle hot unplug.
>> +				 */
>> +				ret = bus->handle_hot_unplug(dev, dev_addr);
>> +				if (ret) {
>> +					RTE_LOG(ERR, EAL,
>> +						"It cannot handle hot unplug "
>> +						"for the device "
>> +						"on the bus.\n ");
>> +				}
> So how we would know what happened here:
> That address doesn't belong to that bus or unplug_handler failed?
> Should we separate search and unplug ops?
> Another question - shouldn't we break out of loop if bus->handle_hot_unplug()
> returns 0?
> Otherwise you can return error value even when unplug handled worked correctly.
>
you are right here.
>> +			}
>> +		}
>> +	}
>> +	return ret;
>> +}
>> +
>> +static void sigbus_handler(int signum __rte_unused, siginfo_t *info,
>> +				void *ctx __rte_unused)
>> +{
>> +	int ret;
>> +
>> +	RTE_LOG(ERR, EAL, "SIGBUS error, fault address:%p\n", info->si_addr);
>> +	ret = dev_uev_failure_process(NULL, info->si_addr);
> As now you can try to mmap/munmap same address from two or more different threads
> you probably need some synchronization here.
> Something simple as spinlock seems to be enough here.
> We might have one per device or might be even a global one would be ok here.
i think global one and synchronization would be fine.
>> +	if (!ret)
>> +		RTE_LOG(DEBUG, EAL,
>> +			"SIGBUS error is because of 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);
>> +}
> Is it really worth a separate function?
i think that would be the bus ops struct of rte_bus_find_device_t usage 
here.
>> +
>> +static int
>>   dev_uev_socket_fd_create(void)
>>   {
>>   	struct sockaddr_nl addr;
>> @@ -146,6 +214,9 @@ 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);
>> @@ -170,8 +241,41 @@ 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",
>> +					uevent.devname);
>> +				return;
>> +			}
>> +			dev = bus->find_device(NULL, cmp_dev_name,
>> +					       uevent.devname);
>> +			if (dev == NULL) {
>> +				RTE_LOG(ERR, EAL,
>> +					"Cannot find unplugged device (%s)\n",
>> +					uevent.devname);
>> +				return;
>> +			}
>> +			ret = dev_uev_failure_process(dev, NULL);
>> +			if (ret) {
>> +				RTE_LOG(ERR, EAL, "Driver cannot remap the "
>> +					"device (%s)\n",
>> +					dev->name);
>> +				return;
>> +			}
>> +		}
>>   		dev_callback_process(uevent.devname, uevent.type);
>> +	}
>>   }
>>
>>   int __rte_experimental
>> @@ -216,8 +320,26 @@ rte_dev_event_monitor_stop(void)
>>   		return ret;
>>   	}
>>
>> +	/* recover sigbus. */
>> +	sigaction(SIGBUS, NULL, NULL);
>> +
> Probably better to restore previous action.
correct, restore the previous sigbus action so that no affect other.
>>   	close(intr_handle.fd);
>>   	intr_handle.fd = -1;
>>   	monitor_started = false;
>> +
>>   	return 0;
>>   }
>> +
>> +void __rte_experimental
>> +rte_dev_handle_hot_unplug(void)
>> +{
>> +	struct sigaction act;
>> +
>> +	/* set sigbus handler for hotplug. */
>> +	memset(&act, 0x00, sizeof(struct sigaction));
>> +	act.sa_sigaction = sigbus_handler;
>> +	sigemptyset(&act.sa_mask);
>> +	sigaddset(&act.sa_mask, SIGBUS);
>> +	act.sa_flags = SA_SIGINFO;
>> +	sigaction(SIGBUS, &act, NULL);
>> +}
>> diff --git a/lib/librte_eal/rte_eal_version.map b/lib/librte_eal/rte_eal_version.map
>> index d02d80b..39a0213 100644
>> --- a/lib/librte_eal/rte_eal_version.map
>> +++ b/lib/librte_eal/rte_eal_version.map
>> @@ -217,6 +217,7 @@ EXPERIMENTAL {
>>   	rte_dev_event_callback_unregister;
>>   	rte_dev_event_monitor_start;
>>   	rte_dev_event_monitor_stop;
>> +	rte_dev_handle_hot_unplug;
>>   	rte_eal_cleanup;
>>   	rte_eal_devargs_insert;
>>   	rte_eal_devargs_parse;
>> --
>> 2.7.4