JP3120434B2 - Program operation analysis method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION The present invention relates to a program operation analysis method, and more particularly, to the provision of the number of execution instructions, instruction mix values, and the like relating to product performance operation analysis in program development. The present invention relates to a program operation analysis method suitable for providing trace information of a dynamic instruction regarding pursuit of a cause when a failure occurs.

[Conventional technology]

Conventionally, performance operation analysis of a program has been performed by the following method. This is a method of performing analysis on a first desk, in which, for an operated analysis program, instructions are dynamically traced on a desk, and the number of executed instructions and an instruction mix value are obtained by a human. In the second computer system, the processing is interrupted each time one instruction of the central processing unit is executed, and a debug interrupt function for transferring control to an analysis data acquisition program is used. By acquiring instruction trace data, and then recognizing the operated analysis program part within the address range loaded on the memory, and selectively editing and outputting, the number of executed instructions and the instruction mix value are similarly obtained. is there. As a known document related to tracing the execution state of this type of program, for example, JP-A-60-3746 is cited.

[Problems to be solved by the invention]

In the above-mentioned conventional technology, the first method requires human labor, and the accuracy is not good because the analysis is performed by a human.
In the second method, the operating system itself is also a trace target, and the information is also acquired. Therefore, a long execution time is required, the trace information is enormous, the editing process is complicated, and the man-hour is required. In addition, at the time of measurement, the use of special functions of the central processing unit, such as starting and stopping a program operation other than measurement, such as debug interrupt and timer processing, occupies the central processing unit, and the trace information acquisition. Therefore, there were problems such as complicated preparation and operation.

An object of the present invention is to provide a program operation analysis method that solves the above-mentioned problems of the conventional technology.

[Means for solving the problem]

In order to achieve the above object, the present invention performs operation analysis of a driven analysis program incorporating an analysis target identification code, an analysis start instruction for designating an analysis start, and an analysis end instruction for designating the end of analysis. A program operation analysis method, comprising the steps of: receiving an analysis target identification code and an analysis instruction parameter from an operation terminal; starting by executing an analysis start instruction of the operated analysis program; The received identification code is checked, and when a match is obtained, the analysis start of the operated analysis program is enabled, and an analysis information stack area is secured. Sequentially fetching, classifying the fetched instruction according to the received analysis instruction parameter,
Interpreting, executing, and repeating a process of stacking information relating to the instruction in the analysis information stack area until an analysis end instruction is executed, and editing and outputting information stacked in the analysis information stack area. It is characterized by the following.

The present invention also provides a program operation analysis method for analyzing the operation of an operation analysis program incorporating an analysis target identification code and an analysis start instruction for designating the start of analysis, comprising: an analysis target identification code and an analysis instruction parameter. Is received from the operation terminal, and is started by execution of an analysis start instruction of the operated analysis program. The identification code communicated by the analysis start instruction is checked with the received identification code. Validating the start of analysis of the analysis program, securing an analysis information stack area, sequentially fetching instructions from the next instruction following the analysis start instruction of the operated analysis program, and fetching the instructions in accordance with the received analysis instruction parameters. Classify, interpret, and execute instructions and stack information about the instructions in the analysis information stack area. Processing, characterized by having a step, which step is output edit information stacked in said analysis information stack area repeating until executes all instructions of the target motion analysis program.

(Operation)

The operation analysis program receives the activation from the operated analysis program, checks the validity of the analysis start of the operated analysis program, autonomously secures an analysis information stack area in the main memory, and thereafter, The instructions of the analysis program are fetched one by one, and according to the analysis instruction parameter,
Instructions are classified, interpreted and executed, and analysis information of the executed instructions is stacked. As a result, instead of using special functions that depend on the central processing unit (hardware),
The program itself makes it possible to easily trace a detailed operation instruction according to an instruction from an operator.

〔Example〕

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 shows a configuration example of a computer system for realizing the program operation analysis method of the present invention. In FIG. 1, a central processing unit (CPU) 1 includes a main storage device 2, on which an operating system (OS) 11, an operation analysis program 12, and an operation analysis program 13 are placed. The operation analysis program 13 is roughly divided into an analysis execution unit 14 that performs an instruction analysis and the like of the operation analysis program 12 and an analysis instruction unit 15 that receives an analysis range and the like of the operation analysis program 12 from an operator. Various information obtained by the instruction trace is stored in the analysis information stack area 16.

The terminal 5 and the printer 6 are connected to the central processing unit 1 through an input / output control device (IOC) 3. The terminal 5 is used by the operator to input a start command of the operated analysis program 12 and the operation analysis program 13 and to input analysis instruction parameters. The printer 6 is used to output analysis information and the like.

The information output file 7 and the program library 8 are connected to the central processing unit 1 through an input / output control unit (IOC) 4.
Connected to The program library 8 stores a large number of programs, and the desired programs are loaded into the main storage device 2 under the control of the operating system 11. In addition to the loading function of the program, the operating system 11 provides various functions based on a supervisor call (hereinafter abbreviated as SVC) in connection with program execution, and manages resources such as a main memory used by the program during execution. Has the function of The information output file 7 is used to output analysis information and the like when the analysis stack area 16 is full or when the operation is completed.

FIG. 2 shows a control and information flow among an operating system (OS) 11, an operation analysis program 12, an analysis execution unit 14, an analysis instruction unit 15, and an analysis information stack area 16 which constitute the operation analysis program 13. Is shown including the terminal 5. Analysis execution part of motion analysis program 13
14 includes a start check unit 21, an instruction interpretation execution unit 22, and an analysis information edit output unit 23, and the analysis instruction unit 15 includes a start declaration unit 41 and an end check unit 42. An analysis start macro instruction 31 and an analysis end macro instruction 33 are incorporated in the source program of the operation analysis program 12, and thereby the analysis target section 32 which is an analysis target instruction range is specified.

Analysis start macro instruction 31 and analysis end macro instruction 33
Has a role of instructing the analysis executing section 14 of the operation analysis program 13 to start and end the operation, and sets the call interfaces shown in FIGS. 4 and 5, respectively.

In the embodiment, the four types of information shown in FIG. 6 can be obtained as the analysis information for the analyzed section 32.
The obtained information can be selected by the analysis instruction parameter 43 from the operator. FIG. 7 shows the analysis instruction parameter 43.
Indicates the contents of

Next, the procedure of the motion analysis will be described with reference to FIG. The operator inputs the same code as the identification code declared in the calling sequence of the analysis start macro instruction 31 by the operated analysis program 12 and the analysis instruction parameter from the terminal 5 as the start parameter. Start declaration section of analysis instruction section 15
At 41, the analysis instruction parameter 43 is set in the OS 11 by using the input identification code as a resource name and using a macro command for setting communication information between programs.

Thereafter, the operated analysis program 12 is loaded and started under the control of the OS 11, and the started operated analysis program 12 is started.
The analysis start macro instruction 31 is executed from 12, and in the control flow, the start check unit 21 of the analysis execution unit 14 in the operation analysis program 13 is activated, and the control is passed.

The start check unit 21 issues a contact information reference macro instruction to the OS 11 using the identification code passed in the calling sequence as a resource name, and if a return code indicating that contact information has already been set is obtained from the OS 11, an analysis instruction is issued. Is determined, that is, it is determined that the analysis start of the operated analysis program is valid, and the analysis information stack area is stored in the main memory.
16 is secured, and thereafter, the instruction interpretation and execution unit 22 is activated, and control is passed. The instruction interpretation and execution unit 22 includes the operated analysis program 12
As shown in the calling sequence of FIG. 4, the instruction is dynamically interpreted and executed one instruction at a time from the instruction indicated by the start address of the analyzed part 32 stored in the general-purpose register R14 to the analysis end macro instruction 33. Stack area 16
The analysis information is stored from the stack area 24 to the stack area 27 according to the analysis information instruction parameter 43 input from the terminal and passed through the OS 11 and the start check unit 21.
The details of the operation of the command interpretation and execution unit 22 will be described later with reference to FIG. The instruction interpretation and execution unit 22 executes the analysis end macro instruction.
When step 33 is executed, the analysis information editing and output unit 23 is started.
The analysis information edit output unit 23 edits and outputs the contents of the analysis information stack area 16.

The instruction interpretation execution unit 22 executes the analysis end macro instruction 33,
After the processing of the analysis information editing and output unit 23, an end identification code is created from the identification code at the start (for example, “E” is added to the end of the identification code), and the code is set as the resource name and the contact information is set. A macro instruction is issued, and an end code (for example, “END”) 44 is set in OS11. The end check unit 42 of the analysis instruction unit 16 issues a contact reference macro to the OS 11 at regular time intervals with the end identification code name created by the promised method, and returns a return code indicating that information has already been set to the OS 11. From the end code (eg "END") 44
Is confirmed, a message to that effect is output to the terminal 5, and the analysis process ends.

Next, referring to FIG. 6, information acquired by the operated analysis unit 32 of the operated analysis program will be described.

The instruction mix counter area 71 stores the instruction code (1
This is an area for accumulating the executed instructions corresponding to the byte length and indicating a value from 0 to 255), and is used to check the performance of the analyzed part 32 based on the number of executed instructions.

The supervisor call counter area 72 is an area for accumulating the executed SVC instructions corresponding to the subcodes of the SVC instructions identified by the above instruction code (1 bust length, indicating a value from 0 to 255). It is also used to check the performance of the analysis unit 32.

The module unit trace information area 73 includes the analyzed section 32
This area indicates the number of executed instructions and the total value of executed SVC instructions for each module within, and is used for analysis by module when the contents of the areas 71 and 72 are recognized as abnormal values. The calling sequence to the module is in accordance with FIG. 4, and the start name in FIG. 4 is obtained as a module name in a processing 55 in FIG. 3 described later and stored.

The instruction trace information area 74 is an area for storing the following information for each executed instruction. The “module name” indicates the module of the analyzed part 32. "Instruction address" indicates the address of the instruction in main memory,
“Instruction information” indicates the instruction itself. In the “memory information indicated by the instruction”, information after execution that is the information processing target of the instruction is stored up to 128 bytes, and in the “register contents”,
Stores the contents of the register before and after the instruction execution. The contents of this instruction trace information area 74 are
Used when performing detailed analysis using dynamic tracing, although it is recognized as abnormal. The “memory information indicated by the instruction” and the “contents of the register” are effective in pursuing the cause when a program failure occurs.

Next, the control flow in the instruction interpretation and execution unit 22 will be described with reference to FIG.

In step 51, the analyzed part of the analyzed operation program 12
The 32 prefetch addresses and the contents of the general-purpose registers at that time are stored, and the operation analysis environment is set.

In step 52, the instruction of the analyzed part 32 is fetched from the head. In step 53, the SVC is
It is determined whether the instruction is an instruction or another instruction. If the instruction is other than the SVC instruction, the instruction is classified into a branch instruction or other instruction (abbreviated as a general instruction). As a result, analysis instruction parameters
If the instruction trace information output is specified in 43 and the instruction is a general instruction for moving and calculating memory information, information before execution is acquired from the memory address indicated by the instruction.

If the instruction is a branch instruction, post-instruction processing (step 56)
Check the status code saved as the execution result of the previous instruction, and if the branch is taken, the branch destination address is set as the next processing instruction address. If not, the next instruction of this instruction is set as the next processing instruction address. As such, the instruction is simulated in step 55.

For SVC instructions other than branch instructions and general instructions,
The contents of the register are reset from the value used by the instruction interpretation and execution unit 22 to the value of the analyzed unit 32, and the instruction fetched in step 52 is executed in step 54.

In step 56, along with the status code of the executed instruction,
Save the contents of the registers changed by the instruction execution,
Prepare for the processing of the next instruction. The instruction execution result is stored in the analysis information stack area 16 according to the analysis instruction parameter.

Thereafter, instruction tracing is performed in a loop of steps 52 to 56. Then, the instruction trace information area 74 shown in FIG.
When the stack area corresponding to is full, the analysis information output unit is activated (step 57).
Start 23 and output information.

The end is determined by detecting a special code (for example, "@END") at the address indicated by the general-purpose register R15 when the branch processing of the analysis end macro instruction 33 shown in FIG. This is performed by creating an end identification code from the identification code at the start and issuing a contact information setting macro instruction using the code as a resource name.

Next, storage of the instruction analysis information in the analysis information stack area in the instruction post-processing (step 56) will be described.

The stack area (1) 24 to the stack area (4) 27 forming the analysis information stack area 16 in FIG. 2 are an instruction mix counter area 71, a supervisor call counter area 72, and a It corresponds to the module unit trace information area 73 and the instruction trace information area 74, and stores the analysis information according to the analysis designation parameter 43. However, the module names stored in the stack areas 26 and 27 are automatically obtained without being given from outside. This is performed as follows. The calling sequence to each module is in accordance with FIG. 4. When the assembled branch instruction is executed in step 55, a search is performed based on the compiler specification based on the address indicated in the register R15 and described in FIG. Obtain the “start name” and use it as the module name. In the case where it is described in assembler language and cannot be detected, the address indicated by the register R15 may be assumed to be the module name.

As described above, in the embodiment, the analysis start macro instruction and the analysis end macro instruction are incorporated in the operated analysis program 12. However, for example, when the number of instructions of the operated analysis program is not as large as the left, the analysis end macro instruction is omitted. You may. In this case, the instruction interpretation execution unit 22 of the operation analysis program 13
When all the instructions of the operated analysis program have been executed, the end of the analysis is determined, and the end identification code 44
Should be created.

〔The invention's effect〕

As is apparent from the above description, according to the present invention, the operation analysis program sets the validity of the start of the analysis of the operated analysis program, sets the analysis environment, etc., and sequentially fetches the instructions in the analysis target range. By classifying, interpreting, executing and stacking instructions according to the analysis instruction parameter from, and editing and outputting the stacked analysis information,
It is not necessary to use the special function of executing one hardware instruction and returning to the analysis program.
Multiple analyzes can be performed in parallel, and only the operated analysis program portion created by the user is targeted without tracing the operating system itself (SV instruction itself), and the program itself is fast, simple, and ,
Fine dynamic analysis processing becomes possible.

In addition, by incorporating the analysis start macro instruction and the analysis end macro instruction in the source code of the operated analysis program to serve as communication between the operated analysis program and the operation analysis program and to specify the analysis range of the operated analysis program,
The analysis time can be shortened, and only the minimum necessary trace information can be obtained.

In addition, the operated analysis program has a language
Although illustrated in the COBOL language, PL / I language and assembler language, the present invention can be applied to the C language and the like, and in terms of program execution, any of an online program that performs transaction processing from a terminal and a batch processing program can be used. Is also applicable, and wide use is realized.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing an example of a system configuration for realizing a program operation analysis method according to the present invention, and FIG. 2 is a diagram showing control between an operated analysis program and an operation analysis program according to an embodiment of the present invention. FIG. 3 is a diagram showing a flow of processing of an analysis execution unit in an operation analysis program, FIGS. 4 and 5 are diagrams showing an analysis start macro instruction and an analysis end macro instruction, FIG. 6 is a diagram showing an example of a data configuration of an analysis information area, and FIG. 7 is a diagram showing an example of an analysis designation parameter from an analysis operator. DESCRIPTION OF SYMBOLS 1 ... Central processing unit, 2 ... Main storage device, 5 ... User terminal, 12 ... Operated analysis program, 13 ... Operation analysis program, 14 ... Analysis execution part, 15 ... Analysis instruction part, 16 ... Analysis information stack area, 31 ... Analysis start macro instruction, 32 ... Analyzed part, 33 ... Analysis end macro instruction.

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor, K. Satake, 890 Kashimada, Saiwai-ku, Kawasaki-shi, Kanagawa Prefecture Inside the Hitachi, Ltd. Information Systems Plant No. 12 Information system factory of Hitachi, Ltd. (56) References JP-A-62-119642 (JP, A) JP-A-61-196338 (JP, A) JP-A-63-300334 (JP, A) JP-A-59-200356 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 11/00

Claims (2)

(57) [Claims] 1. A program operation analysis method for performing an operation analysis of an operation analysis program in which an analysis target identification code, an analysis start instruction for designating an analysis start, and an analysis end instruction for designating an analysis end are incorporated. Receiving the identification code to be analyzed and the analysis instruction parameter from the operation terminal; starting the analysis start instruction of the operated analysis program and checking the identification code communicated with the analysis start instruction and the received identification code; Then, when a match is obtained, the analysis start of the operated analysis program is validated, and a step of securing an analysis information stack area is sequentially performed. Instructions are sequentially fetched from the instruction following the analysis start instruction of the operated analysis program, and the received analysis is performed. Classify, interpret, and execute the fetched instruction according to the instruction parameter, and save information related to the instruction. The process of stacking the analysis information stack area, repeating until you run the analysis end instruction, the program operation analyzing method characterized in that it comprises the step of editing outputting information stacked in said analysis information stack area. 2. A program operation analysis method for analyzing the operation of an operated analysis program incorporating an identification code to be analyzed and an analysis start instruction for designating the start of analysis, comprising the steps of: Accepting from the operation terminal, being started by execution of an analysis start instruction of the operated analysis program, checking the identification code communicated by the analysis start instruction with the received identification code, and when a match is obtained, the operation analysis is performed. Validating the start of analysis of the program and securing an analysis information stack area; sequentially fetching instructions from the next instruction following the analysis start instruction of the operated analysis program, and executing the fetched instruction in accordance with the received analysis instruction parameter Classify, interpret, execute, and stack information about the instruction in the analysis information stack area. Management and program operation analyzing method characterized in that it comprises a step, step, editing outputs information stacked in said analysis information stack area repeating until executes all instructions of the target motion analysis program.