JPS6253861B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention utilizes a type dataset created by copying control table definition information from a control table definition macro file and a path dataset containing information about paths. This invention relates to a memory dump control method that enables a desired control table to be displayed and output by specifying the control table name.

Figure 1 explains the procedure for collecting memory information when a failure occurs in a computer system. 1 indicates the central processing unit, 2 indicates the main memory, 3 indicates the magnetic disk device, and 4 indicates the magnetic tape. ing. The central processing unit 1 normally performs work related to the main memory 2 and the magnetic disk device 3. When a failure occurs in the computer, the data stored in the main memory 2 and the disks of the magnetic disk device 3 are transferred to the magnetic tape 4, and a desired address space portion is examined for failure analysis.

Figure 2 shows a method for reading data in a specific control table in memory space S. For example, it is known that the task control block TCB has a size of 256 bytes starting from address 500, and the task When it is desired to read the contents of the control block TCB, data in a 256-byte area starting from address 500 is extracted and displayed on the display. Simply reading and displaying the 256-byte area from address 500 results in only a list of hexadecimal numbers on the display, which has the disadvantage that it is difficult to understand the contents. In order to eliminate this drawback, it has been a conventional practice to edit the contents of the read task control block TCB based on the task control block type information (in the type data set) and display it on the display in an easy-to-read format. It is being done. The type information of the task control block is #1
The sub-field is 4 bytes in size and is written in character format to indicate that it is a task control block, and the #2 sub-field is 2 bytes in size and the task status is written in binary format. It is something like this.

FIG. 3 explains reading of a control table whose position in the memory space is not fixed. For example, in the task control block TCB,
Some blocks, such as system task control blocks, have fixed logical addresses, while others (logical addresses are undefined) are generated at appropriate locations when needed. The event control block ECB is of the latter type. The task status block TSB points to the task currently being executed, and the event control block ECB can be pointed to by the task control block TCB. When a chain is established between control tables as shown in Figure 3, when you want to read the contents of the event control block ECB, you need to read the contents of the task status block.
Follow the TSB-TCB pointer of TSB to find the task control block TCB, and
-Event control block following ECB pointer
All you have to do is find the ECB and read the contents of the event control block ECB.

In order to read the contents of the control table, type information of the control table and information regarding chaining between the control tables are required, and this is called analysis support information.

FIG. 4 shows a conventional divided data set for storing analysis support information, in which 5 represents a control table name area (index), and 6 represents a logical information area for each control table. Control table names include, for example, task, status, block, task control block, and event control block. The logical information area 6 for each control table is composed of 0th to nth entries, and the 0th entry is filled with information about the entire control table unit and link relationship information with the parent management table, and the 0th entry is filled with information about the entire control table unit and link relationship information with the parent management table. The information of the corresponding sub-field of the control table is entered in each of the n-1 entries, and the n-th entry contains the sub-field name, sub-field length, and if it is a pointer, the information of the control table of the child pointed to. Information is filled in. In this way, in conventional data sets that store analysis support information, the logical information for each control table includes information on the parent control table and information on the child control table in the case of pointers. This differs from control table definition in that it is not possible to create logical information for each control table using a control table definition macro, and for this reason, creating analysis support information conventionally required a great deal of effort and expense. It was.

The present invention aims to improve the above-mentioned drawbacks by providing a memory dump control method that allows a desired control table in memory to be read using a new analysis support information file that can be easily created. The purpose is The memory dump control method of the present invention includes an analysis support information storage unit that stores information regarding the type of control table and information regarding chaining between control tables, a memory information storage unit that stores memory information read from the memory, and the above-mentioned memory dump control method. In the memory dump control method, the memory dump control method includes a support information reference unit that retrieves the contents of a specified control table from memory information in the memory information storage unit while referring to information in the analysis support information storage unit, wherein the analysis support information storage unit has a type dataset created by copying the control table definition information of the control table definition macro file, and a plurality of path information that is independent of the type dataset and represents chaining relationships between control tables. It is characterized by being configured with a bus data set. Hereinafter, the present invention will be explained with reference to the drawings.

FIG. 5 shows an example of a system to which the present invention is applied, in which 7 is a maintenance center, 8 is a maintenance center, 9 is a memory information re-editing section, 10 is an edited memory information storage section, 11 is the memory information transfer unit, 1
2 is a support information reference section, 13 is a display output section, 14 is a display, 15 is an analysis support information storage section, 16 is a control table definition information copy section, 17 is a path information generation section, 18 is a control table definition macro file, 19 indicate path definition statements.

In the maintenance center 8, the control table definition information copy unit 16 copies the control table definition held in the control table definition macrofile 18, and stores it in the analysis support information storage unit 15 as a type dataset to be described later.
It is stored in advance. In addition, the path information generation unit 17
generates path information according to the path definition statement 19, and stores this in advance in the analysis support information storage section 15 as a path dataset.

When a failure occurs in the maintenance center 7, the entire contents of the memory are stored on the magnetic tape 4 as memory failure information. The memory information on the magnetic tape 4 is divided into, for example, pages by the memory information re-editing unit 9, a name is attached to each page, and the edited memory information is stored in the memory information storage unit 10. The contents of the edited memory information storage unit 10 are sent to the maintenance center 8 through computer-to-computer communication and received by the memory information transfer unit 11 of the maintenance center 8. Command to refer to the control table via the keyboard (TABLE command)
to inform the system of the control table you want to read. Then, the support information reference section finds the contents of the specified control table from the memory information while referring to the analysis support information storage section 15, edits it into an easy-to-read format, and passes it to the display output section 13.
The display output section 13 outputs this to display 1.
Display on the screen of 4.

FIG. 6 shows the configuration of the analysis support information storage section 15 of the present invention. 23 is a control table name area, 24 is a type information area for each control table, 25 is a path name area, 2
6 indicates information areas regarding paths.
As mentioned earlier, the control table name refers to the task status block TSB, task control block
TCB, event control block ECB. Logical information for each control table is
It consists of field name, attributes, length, comments, etc. Logical information for each control table is created by copying the control table in the control table macrofile.

Information regarding paths in the path data set 21 is defined by a path definition statement as shown in FIG. 7A. Further, the TABLE command for referencing the control table has a format as shown in FIG. 7B.

FIG. 8 shows the control tables in memory and the chaining relationship between them. Of course, the unit processing table PUP includes, for example, the physical unit control table PUC for magnetic tape, the physical unit control table PUC for card readers, and the physical unit control table PUC for line printers.
Has a PUC. Each physical unit control table PUC has a PUC-UCB pointer,
This points to the unit control block UCB, which is the child control table. PUC-UCB
If the pointer does not point to anything, "F...F" is written in this subfield. The unit control block UCB is
It has an ECB pointer and points to the event control block ECB. The event control block ECB has a NECB pointer (next ECB pointer) and an ECB-DB pointer. The NECB pointer points to the chained event control block ECB. If there is no event control block ECB to be chained next, the NECB subfield is set to “F…
F” is entered.Event control block ECB
The DB pointer of the chained data
It points to the block DB.

Figure 9 shows the path definition statement that defines the path from the physical unit management table PUP to the unit control block UCB to the event control block ECB to the data block shown in Figure 8, and the contents of the control table shown in bold lines in Figure 8. This shows the TABLE command for reading.

/PATH indicates a path definition statement,
ROUT2 indicates the path name, and address A indicates the path start address. PUP indicates the physical unit management table, and PUC, UCB, and P indicate the PUC-
Shows UCB pointer. α is a pointer repetition number variable or a subfield repetition number variable, and is a variable representing the selected PUC-UCB pointer.

"F...F" is a chain termination condition. →
UCB indicates that the control table pointed to is a unit control block, and UCB, ECB, and P indicate that the control table pointed to is a unit control block.
Indicates that a path should be formed according to the UCB-ECB pointer. →ECB indicates that the next control table pointed to is an event control block. β is the control table repetition number variable,
This is to allow the event control block ECB to be traced to be set arbitrarily. NECB・
P is the next ECB pointer,
Indicates that a path should be formed following the ECB pointer to the next paint control block.
"F...F" is a chain end condition. E.C.B.
DB·P indicates that a path should be formed according to the ECB-DB pointer, and →DB indicates that the next control table pointed to is a data block. When it is desired to read out the contents of a desired control table, the TARLE command is input from the keyboard of the display 14, but when it is desired to read out all of the control table indicated by the thick line in Fig. 8, the TARLE command is input as shown in Fig. 9.
A TABLE command is entered. Figure 9
As can be seen from the TABLE command, the second physical unit table PUC should be traced, so the variable α is set to "2", and the second event control block should be traced, so the variable β is set to "2". ”.

FIG. 10 shows the processing of the support information reference section 12. When the TABLE command shown in Figure 10 is input, the path and
The path information of ROUT2 is extracted from the data set 21, and memory information is selectively read from a file (not shown) in the memory information transfer unit 11 according to the path information. When address A is read, the type information of the physical unit management unit PUP is read from the type data set 20, and the physical unit management table in the memory information is read by referring to this type information.
The contents of the PUP are edited into an easy-to-read format, and the edited contents of the physical unit management table PUP are passed to the display output section 13. At the same time, the address ADR (PUC.UCB.P) indicated by the PUC-UCB pointer of the second physical unit control table PUC is stored. When the address ADR (PUC, UCB, P) is read, the support information reference section 12 retrieves the type information of the unit control block UCB from the type data set 20, and uses this type information to set the address.
The contents of the unit control block UCB with ADR (PUC, UCB, P) as the leading address are edited into an easy-to-read format, and the edited contents of the unit control block are passed to the display output section 13. At the same time,
The address ADR (UCB.ECB.P) indicated by the UCB-ECB pointer of this unit control block UCB is stored. By repeating this process,
The contents of the control table existing on the thick line in FIG. 8 can be displayed on the screen of the display 14.

As is clear from the above description, according to the present invention, (a) the efficiency of creating analysis support information can be improved;

(b) By defining and registering the path function,
You can accumulate know-how on control table search procedures.

(c) When reading memory information, a path name can be used as a key, and all control tables on that path can be output at once. Effects such as this can be obtained.

[Brief explanation of the drawing]

Figure 1 is a diagram explaining how to collect memory information when a failure occurs in a computer system, Figure 2 is a diagram showing how to read data from a fixed control table in memory space, and Figure 3 is a diagram showing how to read data from a fixed control table in memory space. A diagram illustrating reading of a control table whose location is uncertain, FIG. 4 is a diagram showing a conventional divided data set that stores analysis support information, and FIG. 5 is a system to which the present invention is applied. FIG. 6 is a diagram showing the configuration of the analysis support information storage unit of the present invention,
Figure 7 shows the format of the path definition statement and the TABLE command, Figure 8 shows an example of control tables in memory and the chaining relationship between them, and Figure 9 shows the details of the path definition statement and the TABLE command. FIG. 10, which is a diagram showing an example, is a diagram illustrating processing in the support information reference section. 4...Magnetic tape, 7...Maintenance center, 8...Maintenance center, 9: Memory information re-editing unit, 10: Edited memory information storage unit, 11: Memory information transfer unit, 1
2... Support information reference section, 13... Display output section, 14... Display, 15... Analysis support information storage section, 16... Control table definition information copy section, 17... Path information generation section, 18... Control table definition macro file, 19 ...Path definition statement.

Claims (1)

[Claims] 1. Refer to the analysis support information storage unit that stores information regarding the type of control table and information regarding chaining between control tables, the memory information storage unit that stores memory information read from the memory, and the information in the above-mentioned analysis support information storage unit. In the memory dump control method, the memory dump control method includes a support information reference unit that retrieves the contents of a specified control table from memory information in the memory information storage unit,
A type dataset created by copying the control table definition information of the control table definition macro file,
A memory dump control method comprising: a path dataset that is independent of the type dataset and has a plurality of path information representing chaining relationships between control tables.