JPH01113841A - Multi-task debugging mechanism - Google Patents

Abstract

PURPOSE:To improve the debugging efficiency in a multi-task environment by realizing the task state transition trace and the start sequence trace. CONSTITUTION:The task state transition trace is attained by recording succes sively the shifted task states to a state trace buffer 60 when the task state transition occur at macroprocessing parts 11-n after providing a transition trace part 20 and a state trace buffer 60. While the task start sequence trace is attained by recording successively the task identifiers which start when a dispatcher part 50 starts tasks after providing a start trace part 40 and the start trace buffer 80. Thus both the state transition trace of a designated task and the sequence trace of a start task are attained. Then the debugging effi ciency is improved in a multi-task environment.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an OS, and particularly to a multitask debugging mechanism suitable for multitask debugging.

[Conventional technology]

Traditional multitasking debugging features include f!
, 11t functions included starting/terminating tasks, displaying task status, etc. However, there was no function to trace task state transitions or task activation order.

[Problem that the invention seeks to solve]

The above-mentioned conventional technology does not consider a program debugging method that takes into account a multitasking environment.

■Task state transition trace ■Task startup order tracing The problem was that it was not possible.

The purpose of the present invention is to make the above two points possible.

[Means for solving problems]

The above objectives are achieved by:

(1) Task state transition trace This function provides a transition trace section and a state trace buffer in the OS that consists of various macro processing sections, dispatcher sections, etc. When a task state transition occurs in the macro processing section, the state of the task changes. This is achieved by sequentially recording the state in the state trace buffer.

(2) Task startup order trace This function provides a startup trace section and a startup trace buffer in an OS that consists of various macro processing sections, a dispatcher section, etc., and when the dispatcher section starts a task, the identifier of the task to be started is stored in the startup trace buffer. This is accomplished by sequentially recording.

[Effect]

(1) Task state transition tracing This function is used in an OS that consists of various macro processing units, dispatcher units, etc., and includes a transition tracing unit that traces state transitions, a task buffer that records the identifier of the task whose state transition you want to trace, and a transition A state trace buffer is provided to record information, and when a state transition of a task occurs in the macro processing section, the transition trace section compares the identifier of the task with the task identifier to be traced recorded in the task buffer and determines whether they are the same. If so, tracing is performed by recording the state of the task that has transitioned to the state 1 Heras buffer.

(2) Task activation order tracing This function provides an activation trace unit that traces the activation order of tasks and an activation trace buffer that records the order of activated tasks in an OS that consists of various macro processing units, dispatcher units, etc.

Before the dispatcher starts a task, the startup trace section generates the identifier of the task to be started! +- Trace is performed by sequentially recording in the race buffer.

〔Example〕

Hereinafter, the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment in which the present invention is applied within an OS. FIG. 2 is a flowchart explaining the operation.

In FIG. 1, 100 OS is various macro processing units 11
, 12, . and a transition tracing unit 20 that accumulates task state transition information. 60 is a state trace buffer that stores state transition information for each task; 70 is a task buffer that contains a task identifier for which state transition tracing is to be performed; and 80 is a startup trace buffer that stores information about activated tasks.

Next, the operation of each part shown in FIG. 1 will be explained based on the flowchart shown in FIG.

To start state transition tracing and activated task tracing, a start instruction macro is issued from the task. When the macro is issued, control is transferred to the top of the OS. The OS checks the macro number and passes control to designated macro processing units 11 to n (500).

In the start instruction macro processing unit 501, state 7! ! If a transition trace is to be started, the identifier of the task to be traced is set in the task buffer 70, and if a task startup trace is specified, a start instruction is set in the task buffer 70.

After that, when a macro is generated in the execution of a task, the macro number is checked and then one of the macro processing units 11 to n is executed (500). When a state transition of a task occurs during execution of the macro processing section (510), the transition tracing section 20 is called. The transition trace section refers to the task buffer 7o to check whether the task whose state has changed has been designated for tracing (512). If specified, the state trace buffer 60 corresponding to the task
The new state of the task is recorded (513).

After the macro processing is completed, the scheduler part 30 selects a task to be activated next (514). Startup trace section 4
0, the task buffer 70 is referred to to determine whether there is an instruction to start tracing the activated task (515). If there is a start instruction, the task identifier selected by a part of the scheduler is recorded in the startup trace buffer 80 (516).

Thereafter, the dispatcher section activates the selected task (517).

To record information in the above-mentioned state trace buffer and startup trace buffer, the buffer is used as a cycle buffer, and after recording to the end of the buffer, recording is performed again from the beginning.

According to this embodiment, it is possible to trace the state transition of a designated task and trace the order of activated tasks, which is effective in improving debugging efficiency in a multitasking environment.

〔Effect of the invention〕

According to the present invention, it is possible to perform two tasks: (1) task state transition tracing, and (2) task activation order tracing, which has the effect of increasing debugging efficiency in debugging in a multitasking environment.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an OS according to an embodiment of the present invention, and FIG. 2 is a flowchart explaining the operation of FIG. 1.

Claims (1)

[Claims] 1. In an OS that consists of macro processing units corresponding to various macro functions provided by the OS and a dispatcher unit that selects and starts tasks, etc., a state trace buffer that accumulates task state transition information and task activation order information A startup trace buffer that stores information, a transition trace section that stores information in the state trace buffer when a state transition of a specified task occurs during macro processing, and a transition trace section that stores information on the execution task in the state trace buffer each time the execution task changes. A multi-task debugging mechanism characterized by having a startup trace section for accumulating.