JPH04148355A - Memory error detecting method - Google Patents

Abstract

PURPOSE:To detect a memory error such as the entire destruction of a memory and the entire destruction of an address by allowing reference data to have specific relevance to data to be checked. CONSTITUTION:Reference data D1 which are allowed to have the specific relevance to data D0 to be checked stored in a specific address A0 of a memory 1, are stored in another address A1 of the memory 1. Then, the data D0 to be checked are compared with the data obtained after calculating back to the reference data D1 based on the specific relevance at the time of checking the data D0 to be checked. Therefore, the inverse calculation is operated based on the reference data D1 prepared by the entire destruction at the time of the entire destruction of the memory, so that the result can be noncoincident with the data D0 to be checked, and the error can be detected. Thus, the detection of the memory error can be surely attained even at the time of destruction of the memory.

Description

[Detailed Description of the Invention] [Summary] Regarding a memory error detection method, the object of the present invention is to provide a memory error detection method that makes it easy to detect memory errors even if memory is destroyed. Reference data that has a specific relationship with the data to be checked is stored in a different address in the memory, and when checking the data to be checked, the data to be checked and the reference data are The present invention is characterized in that it compares data that is back calculated based on the relationship, and furthermore, the specific relationship between the checked data and the reference data is such that a constant is added to the complement of the checked data.

[Industrial application field]

The present invention relates to a memory error detection method, and more particularly to a memory error detection method for communication network identification numbers.

[Prior art]

LAN that connects a host computer and many terminals through communication
In communication networks such as the host computer or each terminal connected to the host computer, each communication network has a unique identification number, and when a host computer calls a specific terminal, or one terminal calls another specific terminal. When calling a machine, you must enter the identification number of the person you are calling.

The identification number is stored at a specific address in the memory of each terminal. For example, when a specific terminal is called from a host computer, the identification number sent from the host computer and the recognition of the target terminal are first checked. It is determined whether the numbers match, and only if they match, communication between the host computer and its terminal is started. Therefore, if the identification number is destroyed for some reason, the terminal cannot be called. Therefore, a check is made to see if the identification number is stored correctly.

In the conventional checking method, for example, the above-mentioned identification number is stored as data to be checked at an address in a memory, and then the same number as the identification number is stored as reference data at an address different from the address where the above-mentioned identification number is stored. address, and when a call is made from the host computer as described above, read and compare both, and if they are the same, it is determined that there is no memory error and the correct identification number is stored, and if there is a mismatch. If so, it is assumed that a memory error has occurred and the operator is notified by some method.

[Problem to be solved by the invention]

According to the above conventional method, if the memory is completely destroyed for some reason, the above two data will be all “1”.
Alternatively, it may be all "0", and in this case, no error will be detected by the above check. Furthermore, if the data to be checked is stored at address "000" and the reference data is stored at address "100", even if address "100" is specified due to destruction of the address line, "000" will not be returned. If only the data to be checked occurs, first read the data to be checked stored at address “000” and then compare the same data as the data to be checked as reference data, so that the data at address “000°” is The check result will be normal regardless of whether it is normal or not.

The present invention has been proposed in view of the above-mentioned conventional circumstances, and it is an object of the present invention to provide a memory error detection method that can reliably detect memory errors even if memory is destroyed.

[Means to solve the problem]

This invention employs the following means to achieve the above object. That is, as shown in FIG. 1, the data to be checked D0 stored at a specific address A0 in the memory 1
Reference data D1 that has a specific relationship with the reference data D1 is stored in another address A1 of the memory 1, and when checking the checked data D0, the checked data D0 and the reference data D1 are stored in a specific address A1. This is to compare data calculated backwards based on relationships.

It is easy to set the specific relationship between the data to be checked D0 and the reference data D to be such that a constant, for example, -1 is added to the complement of the data to be checked D0.

[For production]

Since the reference data has a specific relationship with the data to be checked, in order to calculate the data to be checked from the reference data, it is necessary to calculate backwards from the specific relationship. Therefore, when the memory is completely destroyed (data to be checked D,
and the reference data DI are the same), the above inverse calculation is performed based on the reference data D1 formed by the total destruction, so the result is inconsistent with the data to be checked D0, and an error has occurred. .

Even if an error occurs in the address of the checked data Do (or reference data D0) and the same data as the checked data D0 is read out as the reference data DI as in the conventional example, the reference data D+ Since the above-mentioned specific relationship is inversely calculated, a mismatch occurs, which means that a memory error has been detected.

〔Example) FIG. 2 shows a configuration diagram of an embodiment of the present invention.

Address A of memory 1. For example, in “0OOOH”, the terminal device identification number “0IOIH” is the data to be checked D0.
is remembered as. - On the other hand, the address A+ “5A5AH” of the same memory 1 has the above checked data D0.
1FBFD, which is the value obtained by inverting “0IOIH” and adding -1
H" is written as reference data D. Here, in order for the host computer to communicate with the terminal having the identification number "0IOLH", the above address A0 "000" is written.
0H”, the data to be checked D0 at the address A0 “0LOLH” and the address AI
Reference data D1 of "5A5AH" is compared with "0LOLH", which is the complement of "FEFEH", which is a value obtained by subtracting a constant (-1) from "FEFDH". In this case, it is assumed that there is no memory error, so both will match.

On the other hand, for example, if the memory is completely destroyed and both the checked data D0 and the reference data D1 become "0000H", the value obtained by removing the constant (-1) from the reference data D becomes "0OOIH", Their complements are “FFFEH
'', so there is a mismatch between the two and a memory error has occurred.

Furthermore, for example, address destruction on the reference data DI side occurs, and address A, “5A5AH
”, and the data to be checked is “0LO” from the address A0 “0OOOH” of the data to be checked.
When reading “LH”, the value after subtracting the constant (-1) is
The result is "0102H", and its complement is "FEFDH", which does not match the data to be checked "0IOLH", so it is determined as an error.

In the above description, only the case where the specific relationship between the data to be checked and the reference data is a relationship in which a constant is added to the complement of the data to be checked has been described, but it is of course possible to use other relationships.

〔Effect of the invention〕

As explained above, in this invention, the data to be checked and the reference data have a specific relationship, so when finding the data to be checked from the reference data, it is necessary to perform a backward calculation based on the specific relationship. Memory errors such as complete memory destruction and address destruction can be detected. Furthermore, since this method can reliably prevent memory errors, LA
It is effective if it is applicable to a memory that stores the identification number of a communication network such as N.

[Brief explanation of drawings]

FIG. 1 shows the principle of the invention, and FIG. 2 shows an embodiment of the invention. In the figure, 1...memory, Do...data to be checked,
Dl...Reference data, Ao, A+...Address. Principle explanatory diagram (Fig. 1) Example configuration diagram (Fig. 1)

Claims (1)

[Scope of Claims] [1] Reference data (D_1) having a specific relationship with the data to be checked (D_0) stored at a specific address (A_0) of memory (1) is stored in memory (1). ), and when checking the checked data (D_0), data is obtained by back calculating the checked data (D_0) and the reference data (D_1) based on a specific relationship. A memory error detection method characterized by comparing. [2] The above checked data (D_0) and reference data (
2. The memory error detection method according to claim 1, wherein the specific relationship of D_1) is a relationship of adding a constant to the complement of the data to be checked (D_0).