A method of function execution timeout and deadlock detection based on runtime dynamic tracking

Abstract

The invention relates to the technical fields of computers. A function execution timeout and deadlock detection method based on dynamic tracking of an operating period comprises the following steps: A, a section of tracking code is inserted in operating inlets of to-be-detected functions, the current operating threads of the to-be-detected functions are distinguished, characteristic information of the to-be-detected functions in operation is registered, and reference data of characteristic information objects are returned to the to-be-detected functions for temporarily saving; B, corresponding tracking codes are also inserted in operating outlets of the to-be-detected functions, and according to the access information of the characteristics information objects returned in A, the characteristic information objects of the to-be-detected functions are unregistered from the data structures of the current operating threads; and C, an independent monitor thread is established in the process, wherein the time count of the to-be-detected functions exceeds the first-stage time count threshold, which means that the operating time of the to-be-detected functions is too long, namely timeout, is proved; however, the time count of the to-be-detected functions exceeds the second-stage time count threshold, which means that the operating time of the to-be-detected functions is too long, namely timeout. The method can judge the timeout and deadlock effectively.

Description

technical field [0001] The invention relates to the technical field of computers, in particular to a method for detecting function execution timeout and deadlock based on running dynamic tracking. Background technique [0002] During the operation of all computer software, there will be a large number of function calls, and most of the functions are generally designed to be completed within a sufficiently small time (nanosecond, millisecond level). However, in the actual computer software compilation process, due to defects or negligence in design and development, sometimes very serious failures may occur in function operation, such as deadlock. This is more serious than the exception (exception can be caught) when the function is running, because at this time the software will be in a "sluggish" state, unable to run and capture error messages, which will cause great trouble to the integration test and operation of the software system . At the same time, such faults are al...

AI Technical Summary

Problems solved by technology

First of all, since it is not confirmed which function call has the problem at the beginning, there may be doubts that there are many function calls that need to be tracked by logs. When searching in the logs, there will be a lot of log interference from normal functions, making retrieval difficult; second, due to With the existence of multi-threading, the log records of the entry and exit of each function call can be interleaved, which further increases the difficulty of judging whether the function call has entered or exited; third, high-frequency function calls will output a large number of tracking logs, which is extremely difficult. Affect performance, may destroy the performance conditions of deadlock occurrence, so that deadlock cannot be observed; Fourth, the exit and entry of function calls are different, and there may be multiple branch exit points, which is also inconvenient in coding development

Images