JP2007048247A - System freeze diagnostics and restoration technique applied to wireless device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To avoid freeze of the whole system by avoiding deadlock and restoring the system by detecting generated deadlock when the deadlock cannot be avoided in a wireless device. <P>SOLUTION: A diagnostic engine is made to reside in the wireless device to always observe an operating state of a process and a use state of a system resource, the deadlock is avoided when the process is started and the system resource is required, the deadlock is immediately diagnosed when the deadlock is generated between processes to specify a place (process) and a cause of the deadlock, when the deadlock can instantaneously be fixed, a countermeasure is taken to restore the system. For a bug on software, a diagnostic report in which detailed diagnostic data is recorded is transmitted to a diagnostic server. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a system freeze diagnosis and recovery method for a wireless device, a wireless device having a system freeze diagnosis function, and a system for performing system freeze diagnosis and recovery for a wireless device.

  In recent years, with the advancement of mobile phone functionality, many applications run under high load, and there is a higher risk of system freezes due to system resource contention. There is only a method to start, data in the middle of creation is lost, and the user's satisfaction tends to decrease. Deadlock means that two or more processes running in a multitasking environment are blocked forever due to system resource or mutual communication contention. The system freeze is a state where a deadlock between processes spreads to the entire system and the system is frozen.

  When a system freeze occurs, the user calls the customer support center of the mobile phone carrier for support. However, there is often a long wait until the person in charge answers the phone. There are many cases where the problem cannot be solved even if the person in charge comes out. In many cases, a business operator turns to a manufacturer without investigating the cause of a problem because necessary information is insufficient.

  The manufacturer invests a lot of people, takes time to reproduce the system freeze bug, and investigates the cause of the bug. Sometimes bugs are not found even if time is spent. If a bug is not found, it cannot be fixed. If bugs cannot be fixed promptly, the relationship between customers (business operators and users) may deteriorate.

  On the other hand, even if it is not a software bug, the mobile phone may become unusable. For example, a system resource conflict occurs depending on the user's operation order, the system becomes strange, and the system does not operate. Moreover, due to misunderstandings of users, there are many cases in which users contact customer support of business operators even if it is not a problem of freezing. In addition, there are many cases in which a user requests replacement of a new mobile phone when contacting support. At the repair reception desk or retail store, since there is no effective inspection method for system freezes, mobile phones that do not have many problems such as malfunctions are collected and replaced with new mobile phones. In this case, the manufacturer bears a huge replacement cost.

When a mobile phone breaks down and a user brings the mobile phone to a repair reception desk or a retail store, the person in charge of the counter can immediately and accurately determine the state of the mobile phone that has been brought in. There is a method described in Patent Document 1 as a mobile phone repair service support method that can be achieved. In this method, a maintenance information storage area for storing maintenance information is secured in the storage unit, the maintenance information is written in the maintenance information storage area, and the maintenance information is read from the maintenance information storage area and displayed on the display unit.
JP 2004-194247 A

  However, since only the maintenance information is recorded in the method described in Patent Document 1, it is possible to determine only a failure or malfunction related to certain hardware. As described above, since the mobile phone can realize various functions, since the problem of system freeze is caused by software, it is not possible to determine only a hardware failure or malfunction. Customer support service is insufficient. Moreover, in the method described in Patent Document 1, it is necessary to bring it to a repair reception desk. However, with such a method, it is impossible to deal with the problem of system freeze that has occurred simultaneously in millions of mobile phones.

  Therefore, the present invention is not limited to a special function, and it is not necessary to bring it to a repair reception desk when a system freeze occurs due to the start of any application on the wireless device at any timing. It is an object of the present invention to provide a method capable of recovering a system, software that implements the method, and a wireless device that implements the software.

  A wireless device according to the present invention is a mobile phone equipped with a system freeze diagnostic engine, the diagnostic engine observes the health of the system, avoids the system freeze, means for recognizing the system freeze, and freeze recognition And a diagnosis / recovery unit that collects necessary information and performs diagnosis to restore the system.

  System health observation means that the diagnosis engine stays in the system, monitors the usage status of system resources, and responds to application resource requests when the application starts or requests system resources from the started application again. If there is no contention for system resources, the necessary resources are allocated.

  System resources include processors, I / O channels, memories, devices, data (eg, files), databases, and semaphores. There is a limit when accessing these system resources. For example, when memory is insufficient for a memory allocation request, contention (deadlock) of system resources occurs between processes.

  A deadlock occurs when one process requests the next system resource while holding one system resource (hold-and-wait). For example, in the case of 200 KB memory, when there are 70 KB and 80 KB allocation requests from two processes, 150 KB memory is allocated, and the remaining memory is 50 KB. Next, when there is an allocation request of 60 KB and 70 KB, respectively, while the two processes have the allocated memory, they are not allocated because there is no required memory, and the two processes are blocked from each other. In heavy-duty systems, process deadlocks due to memory leaks often occur, creating a vicious circle, all processes fail to freeze, the entire system freezes, and there is no recovery method other than restarting the system End up.

  The present invention first attempts to avoid deadlocks, and when deadlocks cannot be avoided, detects a partially generated deadlock and performs system recovery to avoid the worst system-wide freeze.

  System freeze avoidance measures monitor system resources and process activity and do not allocate resources that could result in deadlocks. As a criterion for determining resource allocation, a method of ensuring 10% or more of system resources that are always available is used. For example, in a 32 MB memory system, 3 MB of free memory is always left. Also, system resource allocation that makes a loop is not performed. For example, when the process A is using the resource A, the process B requests the allocation of the resource A used at the same time as the process B requests the allocation of the resource A. The process A and the process B wait for the release of the resource B and the resource A that the process B has, and a deadlock occurs. This is called resource loop assignment.

  The deadlock detection means searches for a process that has been sleeping for a long time and examines the history of the process to determine whether it is deadlocked. As a criterion for determining a long sleep process, a process that does not wake up within a certain time (for example, 5 minutes) or immediately goes to sleep when it wakes up is determined. As a criterion for determining a deadlocked process, a process which is a long sleep process and has no history of activity (eg, resource allocation, I / O input / output) during the sleep is determined.

  The system recovery means forcibly terminates all or some of the processes in the deadlock, and releases all system resources occupied by these processes. To end some processes, repeat the process of ending processes one by one until the deadlock is released. The terminated processes are restarted and executed in order so that deadlock does not occur.

  The result of diagnosing the deadlock is reported to the diagnosis server through OTA (Over-the-air). The report content includes the location where the deadlock occurred, the cause and associated diagnostic data. These pieces of information can be used when a mobile phone operator supports a customer, and can also be used when a mobile phone manufacturer fixes a bug.

  As described above, according to the present invention, a diagnosis engine resides in a device as a system for avoiding system freeze in a wireless device, performing a diagnosis and recovery method, and performing a deadlock diagnosis, and contention of system resources between processes. If a deadlock occurs between some processes when it cannot be avoided, the deadlock is detected immediately and the deadlocked process is terminated. Since the freeze is avoided, it is not necessary to contact the customer support of the operator in most cases, and it is not necessary to bring it to the repair reception desk. Even when the customer's customer support is contacted, information necessary for the support has already been reported to the diagnostic server, and the customer can be dealt with immediately. In addition, when a bug is found on the software, detailed diagnostic information about the bug is collected, and the manufacturer can immediately and accurately determine the bug, and the bug can be corrected immediately. Thus, the user can utilize the deadlock diagnostic function according to the present invention as self-care and automatically avoid most of the most complex system freeze or reset problems that occur in wireless devices. If it cannot be avoided, the deadlock that occurred is automatically diagnosed and the system is automatically restored. Further, the customer support person in charge of the business operator can identify the problem accurately and quickly with respect to the customer, can speed up the customer response, and can improve the customer service. Further, the manufacturer can use the detailed diagnostic data included in the diagnostic report to find out the location and cause of the bug related to the software deadlock and immediately correct the bug. For businesses and manufacturers, support costs for customers and products can be significantly reduced.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram illustrating a configuration of a system capable of realizing a deadlock diagnosis and recovery method for a wireless device according to the present invention.

  In the diagnostic system shown in FIG. 1, when a deadlock diagnosis / recovery function is implemented in the wireless device 01, the state of system resources and processes is observed, deadlock between processes is avoided, and a deadlock occurs. Diagnose and take the necessary action to restore the system to normal. When making a diagnosis, necessary diagnosis information can be obtained from the diagnosis database 03 via the wireless network 04 to the diagnosis server 02. The diagnosis result is reported to the diagnosis server 02 through the wireless network 04 and registered in the diagnosis database 03 as a diagnosis report.

  This diagnostic report can be viewed from the operator's customer support. The customer support of the operator checks this diagnostic report, and finds a solution based on the information described in the diagnostic report for a case that the wireless device does not solve. If it is a software bug, report the bug to the manufacturer and request a bug fix. After receiving a bug correction request, the manufacturer obtains detailed diagnostic information data from the diagnostic database via the network, investigates the cause of the bug, identifies the bug location, and corrects the software bug.

  FIG. 2 is an explanatory diagram showing a procedure for diagnosis / recovery of system freeze. The diagnosis engine observes the state 10 of the process and system resources when the system is normal, and avoids deadlock 11 by avoiding allocation of system resources between processes that may cause deadlock. Diagnose 12 using the process and system resource information collected by observation 10 for the deadlock that occurred, detect the deadlock, and terminate all or one deadlocked process System resources are released and system recovery 13 is performed. If the process is partially terminated, the process termination procedure is terminated from a process that is not related to user data (for example, entered email), and data being created by the user (for example, while creating outgoing mail) ) Is preferentially protected.

  FIG. 3 is a block diagram showing a wireless device that implements the above diagnosis / recovery function. In the example shown in FIG. 3, a diagnosis engine 20, a process group 21 that runs in a multitasking environment, diagnosis information (system resources) 22, and a communication unit 23 are mounted on a wireless device that functions for diagnosis and recovery according to the present invention. . The diagnosis engine 20 resides in the system on the wireless device, avoids a deadlock that may occur, diagnoses the generated deadlock, and recovers the system freeze. The process group 21 is a multi-process activated in a high-load environment on the wireless device, and is a diagnosis target of the diagnosis engine 20. The diagnostic information (system resource) 22 is a system resource accessed from each process, and becomes a source of deadlock that may occur between processes. The communication unit 23 is a means for providing a communication protocol (for example, OMA-DM: Open mobile application-device management) for communication with the diagnostic server when reporting the result of the system freeze diagnosis to the diagnostic server, and transmitting diagnostic data. is there.

  FIG. 4 is an explanatory diagram illustrating an example of a deadlock that occurs due to memory contention. In the example shown in FIG. 2, when the available memory reaches 100 KB, the system permits memory allocation to 50 KB and 30 KB memory requests from process A 30 and process B 31, respectively. Next, while holding the memory allocated by process A30 and process B31, and there is no memory allocated for 30KB and 30KB memory requests, respectively, process A and process B will wait for each other, forcing Unless the necessary memory is released, deadlock occurs.

  Not only does a deadlock occur when sharing limited system resources as described above, it also occurs in message communication between processes. FIG. 5 is an explanatory diagram illustrating an example of a deadlock that occurs due to message communication contention between processes. If the process C32 waits for a message from the other party at the same time as the process D33, the message cannot be transmitted, the message reception is blocked between the process C and the process D, and both are deadlocked.

  Deadlock due to contention of system resources can set a resource allocation standard and avoid the occurrence of deadlock. Deadlock due to contention between messages in processes cannot be easily avoided, but it can be diagnosed to find out system design problems (paths that lead to deadlocks).

  FIG. 6 is an explanatory diagram illustrating an example of system resources on the wireless device. Information necessary for deadlock diagnosis includes a memory 40, data 41, and a semaphore 42. The system resource 43 includes these pieces of information data. The memory 40 is a resource shared by processes, and since the available area decreases when used, the memory 40 is a resource that is likely to compete in a high-load environment. The data 41 is a file or a database, and is a resource that may cause a conflict when the same data is accessed from a plurality of processes. The semaphore 42 is a signal for controlling access from a plurality of processes to shared data, and a deadlock occurs when a plurality of processes continuously wait for this signal.

  FIG. 7 is an explanatory diagram showing an embodiment of avoiding deadlock described in the embodiments so far. In order to avoid deadlocks, API (Application interface) is newly incorporated in the process management module and resource management module of the system, and when there is a process activation request and a resource request from the process, an examination to answer the request is performed. Do. In the example shown in FIG. 7, when the process management module receives a process activation request, the process 21 is not activated immediately. An API 51 incorporated in the process management module 50 sends a request to the diagnostic engine 20 to check system resources. The diagnostic engine 20 examines system resources and sends back available system resource information to the API 51. If the API 51 does not cause a deadlock by allocating necessary system resources of the process 21 based on this system resource information, the process 21 is permitted to start. If not permitted, the process 21 is suspended until there is no possibility of deadlock. As described above, as a determination criterion for process activation permission, activation of a process whose available system resource is 10% or less is not permitted.

  In addition, the resource management module does not immediately permit a resource allocation request from a process being executed, but checks whether a deadlock is caused by allocating the resource. In the example shown in FIG. 7, when the resource management 52 receives a resource allocation request from the process 21, the API 53 incorporated in the resource management module is called, and the resource request history of the process 21 is stored in the diagnostic engine 20 so far. Send a request to check. The diagnosis engine 20 checks the resource allocation history of the process 21, and if the process 21 is a request that continuously increases, a deadlock may occur, so that subsequent resource allocation is not permitted. As described above, as described above, resource allocation that allows 10% or less of the available system resources is not permitted. The result of the history investigation is returned from the diagnosis engine 20 to the API 53. Based on this result, the resource management module determines whether or not to permit the process 21 to newly allocate resources. If not permitted, resource allocation to the process 21 is suspended until the possibility of deadlock disappears.

  FIG. 8 is an explanatory diagram showing an example of system freeze diagnosis / recovery implementation. If deadlock between processes cannot be avoided, the system freezes and the system must be restarted. The present invention seeks to terminate a partially deadlocked process before releasing the entire system, freeing system resources, and recovering the system from freeze. In the example shown in FIG. 8, when a deadlock (Hold-and-wait) as shown in FIG. 4 occurs between the process E60 and the process F61, the diagnosis engine 20 goes into sleep with the system resource 22. The process E70 and the process F71 in which deadlock has occurred are detected. As a criterion for detecting deadlocks, as described above, a process that sleeps for a long time and has no activity is in deadlock. The diagnostic engine 20 takes one or all of the processes E70 and F71 that have become deadlocked and terminates all the processes to release system resources. To avoid.

  FIG. 9 is an explanatory diagram showing an example of diagnosis result report execution through the network. In the example shown in FIG. 9, the diagnosis engine 20 places the diagnosis result in the request 70 and sends it to the diagnosis server 02 through the OTA (wireless network). After the diagnosis server receives the diagnosis report, the diagnosis report is processed and registered in the database. Then, the processing result is sent to the diagnosis engine 20 as a response to the diagnosis engine 20.

1 is a block diagram showing a configuration of a diagnostic system capable of realizing a wireless device deadlock diagnosis and recovery method according to the present invention. FIG. It is explanatory drawing which shows an example of the procedure which implements avoidance of a deadlock and system freeze diagnosis and recovery. It is a block diagram which shows an example which implements a deadlock diagnosis / recovery function in a wireless device. It is explanatory drawing which shows an example of the deadlock which generate | occur | produces by the competition of memory on a wireless device. It is explanatory drawing which shows an example of the deadlock which generate | occur | produces by the competition of a signal on a wireless device. It is explanatory drawing which shows an example of the diagnostic information (system resource) preserve | saved at the storage area of a wireless device. It is explanatory drawing which shows an example of avoidance implementation of a deadlock. It is explanatory drawing which shows an example of deadlock diagnosis and recovery implementation. It is explanatory drawing which shows an example of report implementation of a diagnostic result.

Explanation of symbols

01 Wireless devices (cell phones, portable terminals)
02 diagnostic server 03 diagnostic database 04 wireless network 20 diagnostic engine 21 process group 22 diagnostic information (system resource)
23 Communication unit 30 Process A
31 Process B
32 Process C
33 Process D
40 Memory as a system resource 41 Data as a system resource 42 Semaphore as a system resource 43 Diagnosis data storage unit 50 Process management unit 51 API for checking a system resource request from a process
52 System Resource Manager 53 API for checking the history of system resources allocated to processes
60 Process E
61 Process F
70 Request message 71 Response message

Claims (5)

  A wireless device that has at least one processor and one software application, automatically avoids deadlocks between processes by the onboard processor, diagnoses deadlocks and restores the system to freeze A system freeze diagnostic / recovery system consisting of a diagnostic server located in a remote location that can be connected to a wireless device through a wireless network.   A method for avoiding deadlock between processes occurring on a wireless device according to claim 1. This method collects information about system processes and resources on the wireless device to avoid inter-process deadlocks. The collected information includes at least one piece of information that can be used to avoid a deadlock that occurs during the process.   A system-free diagnostic method that spans the entire system that occurs on a wireless device according to claim 1. This method observes the operating state of the process running on the system on the wireless device and the utilization state of system resources, and collects information to diagnose deadlocks that occurred between multiple processes. The collected information includes at least one piece of information that can be used to detect a deadlock between processes. Based on this information, the location (process) of the system freeze that occurred on the wireless device is identified and the cause of the system freeze occurring is diagnosed.   The system freeze recovery method of the wireless device according to claim 1. This restoration method takes action corresponding to the system freeze that occurred based on the diagnosis result, and restores the system. The method for reporting a system freeze diagnostic result of the wireless device according to claim 1. This diagnosis result report method reports a diagnosis result to a diagnosis server through OTA (Over-the-air: wireless network) and registers it in the database.

Images