JPS6048777B2 - Diagnostic control method for channel control device - Google Patents

Description

[Detailed Description of the Invention] [Already Required] To diagnose the channel control device CHC, input/output commands are executed while it is incorporated into the normal system configuration, rather than diagnosing it alone.

A diagnostic program that points out the location of a suspected failure based on the operational response to the command includes the correspondence between the logical and physical device numbers of the channel device connected to the channel control device, the operation priority order, etc. There is no need for the operator to be aware of the unique state. In other words, the unique condition of the object to be diagnosed is automatically recognized at the time of diagnosis, and the expected values of diagnostic input/output commands, diagnostic data, and operations are adapted to the object to be diagnosed. [Industrial Field of Application] The present invention relates to a diagnostic control method for a channel control device.

As is well known, a general computer system consists of a central processing unit (CPU), a main memory (MEM), a channel control unit (CHC), and a group of multiple channel devices (C■1).
CIU2...CHUn) and input/output devices (I/01, 1/O2...I) connected to each of these channel device groups.
/On).

Further, when performing a diagnosis related to the present invention, a diagnostic processing device (DIAC) is further incorporated into the configuration, and the diagnostic processing device (DIAC) mainly uses a prefill storage device (M
EM) and channel control unit (CHC). Among these configurations, the present invention refers to a diagnostic control method for a channel control device (CHC). [Prior Art] The function of the channel control device (CHC) is to control a group of channel devices provided to form a gate that can perform input/output operations and central processing unit operations independently.

For example, it serves as an intermediary for sending commands, control signals, data, etc. from the central processing unit to each channel device, or vice versa, sending control signals, data, etc. from each channel device to the central processing device. [Problems to be Solved by the Invention] As described above, since the channel control device exists between the central processing unit, the storage device, and each channel device, most of it is an interface.

Therefore, when diagnosing a channel control device, the diagnosis will be almost meaningless unless these interfaces are also included in the diagnosis. However, since it is not possible for a pre-created diagnostic program to know what kind of channel device group is connected to the channel control device that is to be diagnosed, in practice, it is difficult to perform the diagnosis including the interface described above. It was difficult to perform the operation efficiently with simple operating instructions. As described above, in conventional diagnosis, the channel control device is excluded from the diagnosis target, and this has been an obstacle to maintaining the operation of the computer system. [
Means for Solving the Problems] An object of the present invention is to solve the above-mentioned problems.The purpose of the present invention is to recognize the logical machine number, the physical machine number of the channel device, and the priority order of access to the main storage device. In order to do this, a channel machine number table, a logical . These are in the machine number conversion table, physical machine number conversion table, and priority conversion table.

[Effect]

Recognizes various unique information of the channel device group and summarizes it as each table above, then compares the table with similar unique information with expected value 1, and corrects the diagnostic data and diagnostic expected value according to the difference. While adding information, it is possible to perform diagnostic control including the interface.

This makes it possible to perform diagnosis 5 without placing unnecessary operational burden on the operator. [Embodiment] FIG. 1 is a system block diagram in which the method of the present invention is implemented.

The central processing unit, I/O, etc. are not directly related to the present invention and will therefore be omitted. 4 In this figure, 1 is a diagnostic memory, which is allocated in advance in the main storage device.

2 is a diagnostic program area, 3 is a diagnostic information area, and 4 is a logout area.

This logout area 4 contains a diagnostic information area 3 and a pseudo CPU (P
When diagnostic data from the CPU (CPU) 10 is applied and execution is further advanced to a certain timing, information representing the internal state of the channel control device, which is sent/outputted from the channel control device, is stored. 4', 5, 6 and 7 are tables characterizing the invention.

Specifically, the number specified in diagnostic program area 2,
In other words, channel device number table 4', table 5 regarding logical device number conversion, and channel device groups as actual hardware connected to the channel control device (11-1, 11-2, 11-2, etc. described later).
...11-n) Table 6 regarding the conversion of physical machine numbers to specify be. 8 is a channel control device (CHC) which is a diagnostic target of the present invention, and its internal circuit 9
and a pseudo central processing unit (PCPU) 10.

The channel control device 8 has a plurality of channel devices (CHU
I, CHlJ2...CHUn) 11-1, 11-2.
...11-n connects. Furthermore, the diagnostic processing device (DIAC) controls the execution of diagnosis for the channel control device 8.
)12.

Inside the diagnostic processing device 12, there are an arithmetic register Al3 and an arithmetic register B14 that temporarily store data from the diagnostic memory 1, a master data generation circuit 15 that generates master data for extracting only the address part of the channel device,
A diagnostic control circuit 16 that performs various controls required for diagnosis, a display pointer 17 that displays which part of the table is currently being advanced, a calculation register Al3, and a calculation register Bl4.
An arithmetic register Cl9 for temporarily storing the contents of the arithmetic circuit (ALU) 18 and the arithmetic results of the arithmetic circuit 18 are arranged. These are various configurations for carrying out the above method of the present invention.

Of these, the above table (4', 5, 6, 7) will be explained in more detail. Channel Machine Number Table 4' This table represents the logical machine numbers of channel devices CHU that are connected to the device to be diagnosed, that is, the channel control device CHC, and are in an operable state (power-on state).

This table 4' is created in the following manner at the start of diagnosis. Issue input/output commands to each channel device CHU from logical machine number 0 to the maximum number of connectable CHUs n, and issue these CHUs.
The input/output command is executed until the clock timing when the operation start response signal is returned from the HU.

Then, for each logical machine number that issued an input/output command, check the logout area 4, and if the operation start response signal is on, it is determined that the channel unit CHU of the corresponding logical machine number is connected, and the channel machine number table is The logical machine number is stored in column 41 of 4'. For example, if the operation start response signals are on, on, off, on, etc. in response to input/output commands to logical machine numbers 0, 1, 2, 3, etc., the channel machine number table The data is stored as 0, 1, 3, . . . from the beginning of 4'. A. of Table 4' CAT is the channel number table (Chan
nelAddressTable). Note that All''1'' in the lower column means the end of the table (the same applies hereinafter). Columns 42 and 43 of Table 4' correspond to the instruction part and diagnostic information part of the input/output command, respectively, and are set to All'0' for convenience in calculating the logical sum described later. Write. Logical machine number conversion table 5 In the present invention, the logical machine number of the channel machine number table 4' is
The expected value prepared in advance in the diagnostic information area 3 is expressed in bits similar to the part representing the logical machine number, as shown in FIG.

However, FIG. 2 shows only one horizontal row of Table 5. For example, if the channel machine number table 4' is created as 0, 1, 3, etc., starting from the first row of the logical machine number conversion table 5, (100000), (010000), (00010)
0), . . . are stored as follows. Table 5 A-L
AT is the logical machine number conversion table (Lf) GicalAddre
ssTransferTabie). Also, in column 52 of Table 5, All''O゛''
However, if the number of channel devices C■ increases, ``1'' will also be set in this column. This also applies to column 62 of Table 6 and column 72 of Table 7, which will be detailed later. Physical machine number conversion table 6 This table shows the correspondence between the logical machine number of the channel device CHLJ and the physical machine number indicating the actual device. It is stored with a bit representation as shown.

This table is created in the following manner at the start of diagnosis. An input/output command is issued to the CHU of the logical machine number (column 41) shown in the channel machine number table 4', and the input/output command is executed until the clock timing when the operation start response signal is returned from CHIJ.

Then, the logout area 4 is checked for each logical machine number that issued an input/output command, and the physical machine number of the channel device CHU that responded that the operation start response signal is on corresponds to the channel device CHU of the corresponding logical machine number. In the physical machine number conversion table 6, each row of column 61 and column 4 of channel machine number table 4' are
The physical machine numbers are stored in order in a one-to-one correspondence with each row of 1. For example, in response to input/output commands to logical machine numbers 0, 1, 3, etc. in channel machine number table 4', operation start response signals are output from physical machine numbers A, B, D, etc. (100000), (010000), (0001) in order from the first row of the physical machine number conversion table 6.
00), . . . are stored. A-PAT in Table 6 is
Physical address conversion table (PhysicalAddressTr)
(ansferTable).
Priority Conversion Table 7 The input/output commands based on the diagnosis are operated on the channel control device 8 according to the order of the logical machine numbers (41) in the channel machine number table 4' until the timing when the memory access request signal first turns on. let

That is, for each logical machine number (0, 1, 3, . . . ), it is checked which memory access priority signal in the logout area 4 corresponds to the logic "1".

For example, in the case of logical machine number 0, if the part corresponding to the memory access priority position signal is displayed as logic ``1'', it is recognized that the memory access priority of the logical machine number 0 is 1, Enter this as a binary number in column 71, at its first address A.
．． PRT(PrlOritytransferTabi
Entries are made in order from e). For example 1, 2, 3
It's like... These 1, 2, 3... are actually (010000), (001000), (00010
0) ... as in NC, (C is CombinatiO
n). n is the maximum number of connectable channel devices. By creating Tables 4', 5, 6, and 7 described above, the unique information (logical machine number, physical machine number, priority order) of the channel device at the time of diagnosis became clear.

Therefore, according to these tables (4', 5, 6, 7) and the expected value collection specific information (same as above), the input diagnostic data is reprocessed to be suitable for each channel device where actual diagnosis is performed.
Furthermore, the diagnostic result data is converted back into expected values suitable for each channel device. Except for processing and converting the data in this way, it is sufficient to perform the diagnosis in the same way as a normal diagnosis. A normal diagnosis is a general diagnosis that does not require processing input diagnostic data or converting expected values (or diagnostic results). The expected value here refers to a response value to predetermined input diagnostic data that is obtained when the channel control device is operating without any abnormality. In this way, the preparatory steps of input diagnostic data processing and expected value conversion are completed. Hereinafter, diagnosis is performed as follows using the tables (4', 5, 6, 7) and the unique information at the time of collection. First, the input/output instructions used for diagnosis execution are read from the diagnostic program area 2 and loaded into the calculation register Al3.

Next, the contents of the first address A-CAT of the channel number table 4' are loaded into the calculation register B14.

Regarding the contents of these operation registers, the diagnostic control circuit 16
Under the control of the arithmetic unit (ALU) 18, the logical sum is calculated and set in the arithmetic register Cl9. The reason for calculating the logical sum in this way is to replace the channel logical machine number part in the input/output command with the contents of column 41 of channel machine number table 4'. At this time, the portions of columns 42 and 43 of Table 4' are masked by the mask data generation circuit 15. This results in
Channel device actually connected (11-1, 11-
2, . . . 11-n) are obtained.

The contents of the arithmetic register Cl9 are stored in the pseudo central processing unit (PC).
PU) 10 by a pseudo central processing unit setting command.

It is then applied to the internal circuit 9 of the channel control device 8 together with the diagnostic data from one of the diagnostic information areas 3 . Here, the internal circuit 9 and one corresponding channel device (11-1 to 1
1-n) starts internal operations for diagnosis.
In this case, the internal operation progresses by the clock time set in advance by the instruction from the diagnostic control circuit 16 (-dotted arrow in the figure: 9◆--). Note that the solid line arrow from the circuit 16 (
9-) is a line for transferring diagnosis start signals, control signals, etc. Further, the pseudo central processing unit 10 acts as a central processing unit (which is inactive during diagnosis) forming the main body 4 of the computer system, and predetermined commands are sent to the channel device via the device 10. given to. The diagnostic result data generated by the internal operation in the channel control device 8 is transferred from the internal circuit 9 to the logout area 4 .

Here, we will see whether the diagnostic result data matches or does not match the expected value. If they match, the diagnosis result is normal. If there is a mismatch, the position of the signal on the circuit diagram and its signal value (``0'' or ``1'') or the mounting position of the failed printed circuit board package that can be estimated from the mismatch signal are displayed. The operator attempts repair based on this diagnostic information. When looking at the match or mismatch between this diagnostic result data and the expected value, check the location of the expected value that requires conversion and the table related to the conversion shown in diagnostic program area 2 (i.e., Table 5). , 6 and 7), the logical machine number currently executing the input/output command shown in the row indicated by the display point 17 in the table 1 according to the instruction. The part of the expected value specified in the program area 2 is replaced with the corresponding conversion content, and converted into an expected value that is compatible with the above-mentioned unique information of the channel device that is executing the input/output command. After the conversion, the data in the corresponding logout area 4 and the expected value after conversion are compared, and if they do not match, the above-mentioned diagnostic information is displayed. In addition, in the above explanation, processing of input diagnostic data and conversion of expected values have been described, but in the process of creating the logical machine number conversion table and physical machine number conversion table of the present invention, the related circuits listed above are diagnosed. be able to.

For example, in the diagnosis control unit of the human output command at the time of diagnosis, 2
There is a function to identify from which of the CPUs the data was sent. Therefore, two tables of the above two types are created. By comparing these, it can be determined that there is a failure in the following cases. 1. There was a discrepancy when comparing the two tables.

Even if 2 matches, there are two or more ``1'' in one row of the table.

Even if 3 matches, there are two or more ``1''s in one column of the table.
Depending on the position (row, column) of the table in which the above items 1 to 3 are located, the fault location is estimated and displayed by the diagnostic program. Furthermore, in the above explanation, for ease of understanding, the case is limited to the case where only one of the channel devices connected to the channel control device operates, but the operation of m channel devices is also as follows. It is possible to diagnose (control) this.

Setting of 1m pieces of expected value specific information. 2 Continuously set m times during one diagnostic test.

3 , m or , Cm (gender of diagnostic test. Setting of various diagnosis-specific information.

〔Effect of the invention〕

As explained above, according to the present invention, When performing diagnostics on the control device, The operator inputs the above-mentioned unique information. An event that imposes an excessive burden on the operator Also, from another point of view, the channel Is it possible to cut the Because it is magnified, the diagnosis of the device is This opens the door for development.

Therefore, it also contributes to improving the operating rate of the computer system.

[Brief explanation of drawings]

Fig. 1 is a system layout lock diagram in which the method of the present invention is implemented, Fig. 2 is a diagram showing the contents of σ for one horizontal line of the logical machine number conversion table, and Fig. 3 is a diagram showing the horizontal If of the physical machine number conversion table. FIG. In the figure, 1 is the diagnostic memory, 2 is the diagnostic program area, 3 is the diagnostic information area, 4' is the channel number table, 5 is the logical machine number conversion table, 6 is the physical machine number conversion table, and 7 is the priority table. Rank conversion table, 8 is channel control setting, 11-1, 11-2, 1
1-n are channel control units, and 12 is a diagnostic processing unit.

Claims (1)

[Claims] 1. A method comprising at least a main storage device and a channel control device and a group of channel devices, and diagnosing the channel control device under the control of a diagnostic processing device, wherein each channel device has a logical machine number, Recognizing the physical machine number and the access priority for the main storage, a channel machine number table, a logical machine number conversion table, a physical machine number conversion table, and the priority are respectively stored in pre-allocated areas in the main storage. A conversion table is created, and according to these tables, the input diagnostic data to be applied to the channel control device is reprocessed into diagnostic data suitable for the channel device, and the diagnostic data executed based on the diagnostic data is A channel control device characterized in that diagnostic result data from a channel control device is compared with an expected value that is converted back into an expected value that is compatible with the channel device, and a diagnosis is made based on whether or not the results match. diagnostic control method.