JPS638947A - Program debug supporting system - Google Patents

Abstract

PURPOSE:To attain the setting of a break point only to a task independent of other tasks and to improve the efficiency of debugging work by deleting the break point by inserting an original instruction code and restoring the instruction code section, and inserting an internal interruption instruction to set a break point. CONSTITUTION:When a break point address is set at the entry of a break point managing area correspondent to a task desired to be interrupted for debugging, the break point is set by an instruction replacing means when the context of the task is switched to the context of another task, the break point set by the instruction replacing means is deleted by a restoring means. Accordingly, even if another task has an instruction code section in common with the objective task, internal interruption is not generated, and the task can be interrupted only when the objective task executes the instruction code section.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a bloram debugging support system, and more specifically, to a computer system that performs multi-task execution processing in which multiple tasks are processed in parallel. The present invention relates to a program debugging support system suitable for debugging a task that shares an instruction code section with other tasks.

[Conventional technology]

When debugging a program running on a computer system, a method is adopted in which execution of the program to be debugged is interrupted at a location called a breakpoint and the internal state of the program is examined and confirmed.

Conventionally, methods for setting breakpoints in a program include a method using a debug monitor device connected from outside the computer system and a method using a debug support program running within the computer system.

In the former method, a debug monitor device is externally connected to the address bus of the computer system, and this debug monitor device compares the instruction fetch address that appears on the address bus with the breakpoint address set for itself, and when they match, the debug monitor device The debug monitor device issues an interrupt to the computer system to stop the program.

On the other hand, in the latter method, a debug support program in a computer system saves the instruction code stored at the breakpoint address of the program to be debugged in a table that corresponds to one program, and instead writes the instruction code at that point. An internal interrupt instruction is inserted into the program, the program is then executed, and when the execution of the program reaches a breakpoint address, the internal interrupt instruction generates an internal interrupt and the program is stopped.

[Problem that the invention seeks to solve]

In the conventional method described above, a breakpoint address exists corresponding to an instruction code section of a program. Therefore, in multi-task processing, if an instruction code section of a program is shared by multiple tasks and the address of the instruction code section is set as a breakpoint address, each task that shares the instruction code section is executed. There is a problem in that execution is interrupted and breakpoints cannot be set only for desired tasks. For this reason, conventionally, each time execution of a task is interrupted, the operator reads the task identification information of the interrupted task from a task control block, etc., determines whether the read task is the intended task, and In the case of a task, the current state of the program must be investigated, which has the disadvantage of poor work efficiency during debugging.

The present invention solves these conventional drawbacks, and its purpose is to perform multitask processing in which the instruction code section of a program is shared by multiple tasks, but to perform the task independently of other tasks. The purpose of this feature is to improve the efficiency of debugging by allowing you to set breakpoints that are specific to only the tasks you are debugging.

[Means for solving problems]

To achieve the above object, the present invention provides a computer system that processes a plurality of tasks in parallel, which includes a breakpoint address setting area and an instruction code save area corresponding to the area, and is provided corresponding to the task. When a task context is switched, if a breakpoint address is set in the breakpoint address setting area of the breakpoint management area that corresponds to the task that was running before switching, the breakpoint address is set. a restoring means for restoring the instruction code section of the breakpoint address using the instruction code saved in the instruction code save area corresponding to the setting area, and the breakpoint corresponding to the task to be run after switching when the task context is switched; If a breakpoint address is set in the breakpoint address setting area of the management area, the instruction code of the instruction code portion indicated by the breakpoint address in the breakpoint address setting area is transferred to the instruction corresponding to the breakpoint address setting area. an instruction replacement means for inserting an internal interrupt instruction into the instruction code section indicated by the breakpoint address after saving the code to a code saving area; and an interrupting means for catching an interrupt caused by the internal interrupt instruction and interrupting execution of the task. Be prepared.

[Effect]

If you set a breakpoint address in the entry in the breakpoint management area corresponding to the task you want to interrupt for debugging, when the context of the task is switched to the context of that task, the breakpoint will be set by the single instruction conversion means, When the context of a task is switched from that task to the context of another task, the breakpoint set by the instruction replacement means is deleted by the recovery means, so even if the other task is switched to the instruction of the target task. Even if code sections coexist, no internal interrupt occurs, and a task can be interrupted by an internal interrupt only when the target task executes the instruction code section.

〔Example〕

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of a computer system implementing the program debugging support method of the present invention.
It is composed of a central processing unit 5, a main storage device 1, and a terminal device 4.

The main storage device 1 includes a general program exclusive memory space 2 and an operating system exclusive memory space 3. The general program exclusive memory space 2 includes this space 2.
The instruction code section 21 of the task running in . ~21. l and
Data section 22□-221 of the task running in that space 2
*, 22,1. and are included.

Further, in the operating system occupied memory space 3, there is a task dispatching entry 31 for switching the task context (execution environment), and an entry 32 corresponding to the task.
~32, and can store the correspondence between the breakpoint address of the task and the instruction code at this address in the corresponding entry.
2, a module 33 that manages the breakpoint management table 32, an interrupt processing module 34 that performs interrupt processing corresponding to an internal interrupt that occurs at a breakpoint address, and an entry 35.2 that corresponds to a task. ~35.
, and a task control block 36 that stores identification information of tasks that are currently being executed or suspended.

FIG. 2 shows an example of the configuration of entries corresponding to one task in the breakpoint management table 32, each entry 32. ~32
．． is the area BKADR (Bl[ADH, A-BKADR,) for setting the breakpoint address and each area 8KADI? l-BKAI)11. Area 5VCODE (SVCODEI ~ 5
Area BK for setting VCODE, ”) and the number of breakpoints registered in that entry.
It is composed of N.

Next, regarding the operation of this embodiment, FIG.
A detailed explanation will be given below with reference to FIG. 4 showing the flow of an example of breakpoint cancellation processing executed by the module 33, and FIG. 5 showing the flow of control of a task in which a breakpoint has been set.

[Breakpoint Setting Process] When setting a breakpoint, a breakpoint setting command is input from the terminal device 4 having the task identifier of the target task for which a breakpoint is to be set and the breakpoint address as arguments. When this breakpoint setting command is given, the central processing bag W5
starts the process shown in FIG. 3 of module 33. That is, first, it is determined whether a breakpoint at the same address as the specified breakpoint address is set in the entry in the breakpoint management table 32 corresponding to the task with the task identifier included in the breakpoint setting command. 300), if it is not set, the area BK in that entry.
Register the specified breakpoint address in the free area of ADR+-BKADR (301),
The number of breakpoints in area BIN is increased by 1 (302).

[Breakpoint Cancellation Process] When canceling a breakpoint, a breakpoint cancellation command is input from the terminal device 4, which has the task identifier of the target task for which the breakpoint is to be canceled and the breakpoint address to be canceled as arguments. When this breakpoint cancellation command is given, the central processing unit 5 starts the processing shown in FIG. 4 of the module 33. First, a breakpoint with the same address as the specified breakpoint address is set in the entry in the breakpoint management table 32 corresponding to the task with the task identifier included in the breakpoint cancellation command. 4.
00), if set, delete the specified breakpoint address in that entry 401)
, the number of breakpoints in area IKN is decreased by 1 (402).

[Control of tasks with breakpoints set] In FIG. 1, two tasks Tl share a certain instruction code section, for example, instruction code section 211.

Assuming that task T2 exists, entry 32. of the breakpoint management table 32 corresponding to task T1. The instruction code section 21 is set by the breakpoint setting process described above.
1 is registered as a breakpoint address, and no breakpoint is registered in the entry 32 of the breakpoint management table 32 corresponding to task T2. In this state, execution control is transferred from task Tl to task T2,
Consider the case where execution control is then transferred to task T1 again.

In this case, when execution control is transferred from task T1 to task T2, first, as is done in general task switching processing, an entry 351 in the task context save area 35 corresponding to the context of task T1 that ran before task switching is (500), and processing 501
The breakpoint 1 conversion process configured in steps 505 to 505 is executed.

In this breakpoint replacement process, in process 501, whether or not a breakpoint was set in task T1, which is the task before switching, is determined from the entry 32 of the breakpoint management table 32 corresponding to task T1.
The determination is made based on whether the value of the area BKN inside is 0 or not. In the current example, the address of instruction code section 21+ is used as the breakpoint address in entry 32. is registered in the instruction code part 2 specified by this breakpoint address.
The instruction code of 11 is entry 32. has been evacuated to.

Therefore, it is determined in the above process 501 that a breakpoint has been set, and in the next process 502, task T
The instruction code saved in the code save area 5VCODH in the entry 321 of the breakpoint management table 32 corresponding to 1 is stored in that area 5VCOD.
I! According to the breakpoint address corresponding to the instruction code section 21. will be restored.

Next, whether or not a breakpoint is set for task T2 to be executed from now on is determined by determining whether or not the value of area BKN in entry 32□ in breakpoint management table 32 corresponding to task T2 is O. As determined by (503), since no breakpoint is set in task T2 in this example, process 506 is executed next, and general task switching process is executed. That is, entry 35. of the task context save area 35 to which the context of task T2 running after task switching corresponds. (506) and task T2
Execution control is transferred to (507).

As a result, task T2 starts running, and its execution is controlled by instruction code section 21. , the instruction code in the instruction code section 2L was restored by process 502 when the task context was switched, so the task T
2 executes the instruction code portion 21+, no internal interrupt is generated.

Next, when the execution control is transferred from task T2 to task TI again, the process of FIG. 5 is started again, and the context of task T2 is changed to entry 35. , and breakpoint replacement processing is executed. In this breakpoint replacement process, in process 501, task T2 before switching
Since it is determined that no breakpoint is registered in , processing 502 is not executed and the process advances to processing 503.

Then, in this process 503, task T1 after switching
It is determined that a breakpoint is set on
Processes 504 and 505 will be executed.

In this process 504, the entry 32 of the breakpoint management table 32 corresponding to the task T1 after switching
According to the breakpoint address registered in 1, the instruction code in the instruction code section 21+ is transferred to its entry 32.
．． Area 5 corresponding to that breakpoint address in
VCODI! (504), and then, according to the breakpoint address registered in the entry 32+, the instruction code section 21. An internal interrupt instruction is inserted into (505), and after this, the context of the task TI running after the task switch is inserted into the corresponding entry 35. of the task context save area 35. (506), and execution control is transferred to task T1 (506).
07).

Execution control is transferred to task TI, task T1 starts running, and when the execution control reaches the breakpoint address, an internal interrupt is generated at that address because an internal interrupt instruction was inserted at that address in process 505. , the corresponding interrupt processing module 34 will be executed. Therefore, the interrupt processing module 34 causes the execution state of the task T1 to transition from the running state to the waiting state, and the execution of the task T1 is interrupted. At this time, the notification that the task has been interrupted and the identification information of the interrupted task T1 obtained by referring to the contents of the task control pronk 36 are output to the terminal device 4. In this embodiment, a break is made only for task T1. Since the point is set, the above-mentioned interruption is always an interruption of task T1, and the operator can immediately know that the execution of task T1 has been interrupted by the breakpoint. Therefore, it is no longer necessary to check whether the interrupted task is the intended task as in the conventional case.

Through the above processing, even when there are multiple tasks that share the instruction code part, it is possible to manage breakpoints independently for each task, and break only to the target task independently of other tasks. It is possible to set a point and interrupt the execution of the task at that breakpoint. Note that in the above explanation, breakpoints were not set for task T2, but it is of course possible to set breakpoints for task T2 independently, and it is possible to set breakpoints for task T2 independently. If breakpoints are set independently, you can identify which task was interrupted at the time of interruption by looking at the display on the terminal device, and if the breakpoint is set for another task that shares the instruction code part of the breakpoint address. Since non-debug target tasks that are not
It is possible to immediately recognize that the interrupted task is always the task to be debugged. It goes without saying that the present invention can also be applied to tasks that do not share an instruction code section with other tasks.

〔Effect of the invention〕

As explained above, in the program debugging support method of the present invention, when a task context is switched, if a breakpoint is set in the instruction code section of the task that was running before switching, the original instruction code is inserted. By restoring the instruction code section, the breakpoint is deleted, and if a breakpoint is set in a task that should run after switching, the instruction code of the instruction code section specified by the breakpoint address is saved and an internal interrupt is Breakpoints are now set by inserting instructions, so in computer systems that process multiple tasks, it is possible to set breakpoints only for one task, independent of other tasks that share the instruction code section. becomes.

Therefore, it is no longer necessary to determine whether the interrupted task is the task to be debugged every time execution is interrupted at a breakpoint, unlike in the past.

This has the effect of significantly improving software productivity compared to conventional debugging support methods.

[Brief explanation of drawings]

Fig. 1 is a block diagram of an embodiment of a computer system that implements the program debugging support method of the present invention, Fig. 2 is an explanatory diagram of the contents of a breakpoint management table, and Fig. 3 is a flowchart showing an example of breakpoint setting processing. , FIG. 4 is a flowchart showing an example of breakpoint cancellation processing, and FIG. 5 is a flowchart showing an example of processing when switching tasks. In the figure, 1... Main storage device, 2... General program occupied memory space, 3... Operating system occupied memory space, 4... Terminal device, 5... Central processing unit, 21. ~21. ...Instruction code section, 221
．． ~22. s. 22...Data section, 31...Task dispatcher, 32...Breakpoint management table, 321
~32. ...Each entry of the breakpoint management table 32, 33...Breakpoint management table 3
34... interrupt processing module, 35... task context save area, 35.
~35. ...Each entry of the task context save area 35, 36...Task control block.

Claims (1)

[Claims] In a computer system that processes multiple tasks in parallel, a breakpoint management area provided corresponding to a task includes a breakpoint address setting area and an instruction code save area corresponding to the area. and, when a task context is switched, if a breakpoint address is set in the breakpoint address setting area of the breakpoint management area corresponding to the task that was running before switching, the breakpoint address corresponding to the breakpoint address setting area is set. a restoring means for restoring the instruction code portion of the breakpoint address using the instruction code saved in the instruction code save area; and a breakpoint in the breakpoint management area corresponding to the task to be run after switching when task context is switched. If a breakpoint address is set in the address setting area, save the instruction code of the instruction code section specified by the breakpoint address in the breakpoint address setting area to the instruction code save area corresponding to the breakpoint address setting area. and an interrupting means for capturing an interrupt caused by the internal interrupt instruction and interrupting execution of the task. Program debugging support method.