DD299018A7 - METHOD FOR CHECKING STORAGE STORAGE - Google Patents

Abstract

The invention relates to the field of error detection in microcomputer-controlled devices and devices, in particular in the context of self-diagnosis. The goal is to increase the error detection rate. The task is to completely check the memory contents of read-only memories with program-technical means, d. H. According to the invention, actual checksums are formed on the basis of cyclic codes and compared with setpoint checksums, all setpoint checksums being stored in a selected memory circuit. For this selected memory circuit, the reference check sum zero is used. figure

Description

Due to special characteristics of the cyclic codes, the read-only memory circuit including the target check sum table may also be included in the test, in contrast to birh-jr known methods. The core idea of the method according to the invention is the targeted achievement of the checksum zero for the read-only memory circuit containing the check sum table. The memory contents to be checked must be extended by its own checksum. If you form the checksum again, you always get zero regardless of the information content. This meets z. B. in the cyclic coding with the above generator polynomial P (x).

The test method according to the invention is accordingly achieved as follows. By means of a microcomputer development system, the target checksum of each memory circuit except a selected memory circuit is computed in the usual way by means of a special checksum program which executes the cyclic coding with a suitable generator polynomial. These Sollprüfsums are stored in the form of a Sollprüfsummentabello in the selected memory circuit, for which this memory circuit itself the Sollprüf sum zero is used. Now, the checksum program is applied to the selected memory circuit except a number of free memory locations. The thus determined Zwischenprüfsumi e is entered in the free memory locations and thus programmed the final memory content of the selected memory circuit. This achieves the desired check sum zero required for the method in the selected memory circuit.

Ausführungsbelsplol

The evidence for the correct memory contents of a micro-computer-controlled system is provided by a Festwertspeicherprüfprogramm whose core the subroutine EDC (error-detection code) for the checksum according Generatorpolyriom P (x) = x ^'s + x ¹² + x ^b + x °. The test program successively forms the actual checksums of the EPROM circuits and compares them with the respective reference checksum of the setpoint check sum table. If there is a match, the memory contents are likely to be correct; in the event of a mismatch, there is an error that triggers an error message indicating the number of the errored EPROM. The latter must be realized device-related.

In the following, the targeted achievement of the nominal test sum zero will be explained in more detail using an example. A microcomputer program takes about 3.5 kbytes of storage capacity. 4 EPROMs of 1 kByte each are used. The 4th EPROM also contains the read-only memory check program including the subprogram EDC as well as the error handling program and the sample summation tool as per drawing. The determined Sollprüfsumman PS 1 to PS 3 of the first 3 EPROM have the following hexadecimal values:

1st EPROM: PS 1 = 38A4H

2. EPROM: PS2 = OB2CH 3.EPROM: PS3 = F344H

Furthermore, the intermediate checksum PS 4 'of the first 1022 memory disk of the 4th EPROM is calculated (PS 4' = 379BH) and brought to the memory locations 1023 ¹ J 1024 in the order of the less-significant part and the low-order part. This results in the desired setpoint checksum PS 4 = 0000 H for all 1024 memory locations.

The 8-byte long check-sum table is shown as a detail in the drawing.

The method is independent of the storage capacity of the memory circuits. It also applies to read-only memory circuits of higher storage capacity, as well as to mask-programmed ROM memories.

The procedure given may be slightly modified:

- The set check sum zero of the last EPROM can be omitted in the set check sum table, but this must be taken into account accordingly in the check program.

- Check program and set check sum table do not necessarily have to be located together in the selected last EPROM (with the highest start address). The EPROM, which contains the roll-checksum tabella, must, however, always be brought to the reference checksum zero according to the described method.

In addition to a purely programmatic solution for the process is also a realization by the inclusion of existing programmable input / output circuits such. As the SIO U 856 or feedback shift registers in a known manner possible, creating a mixed software hardware solution. As an additional advantage results in a much faster checksum formation and thus a significant reduction in the test time.

Documents considered:

DE-PS 142618, G11C29 / 00

DE-OS 2916710, G11C29 / 00

U.S. Patent 4,490,783, G 06 F11 / 00

Görko, W .: Test options for LSI circuits.

ZHV. Electronic Rechenanlagen, 20 (1978) H.5, p.220-227.

Claims (2)

Method for checking read-only memory by means of arithmetic operations applied to the memory contents and comparison of the actual information obtained as a result of the arithmetic operations with a previously determined and stored scan information by forming actual checksums on the basis of cyclic codes and comparing them with reference checksums, characterized in that all the reference checksums stored in a selected memory circuit, that for this selected memory circuit, the Sollprüfsums Null is used, that after storing the Sollprüfsums including the own Sollprüfsums zero in the selected memory circuit, the checksum program except a number of free memory locations is applied to this selected memory circuit, and that this determined checksum is entered on the free memory locations of the selected memory circuit. For this 1 page drawings Field of application of the invention The invention relates to the field of error detection in mikrorechnergestauerton devices and devices. It is used to test plaque stores, in particular in the context of the automatic diagnosis of computer-controlled systems. Characteristic of the known technical solutions In the testing of microcomputer-controlled devices and devices also programmrntechnische tests of Festweit- and data memory are made. Daboi program errors of the read-only memory and SchreibVLesefehler the data memory are greatly erkann !. These tests can serve as both an operational test procedure at the manufacturer of the microcomputer devices and devices used as well as for self-diagnostic (self-test) boim users. Errors in byte-wise organized fixed value saves can be determined with certain probability very simply by parity check. However, this method (eg described in DE-OS 2,752,377) requires a considerable additional storage capacity for the parity bits. Other methods require additional circuit complexity or special test devices that are not relevant for the technical self-diagnosis. It is also known (DD-WP 142.618) that read-only memory circuits are tested by formed by bytewise addition or subtraction of their memory contents a code word and this code word is compared with a setpoint. The formation of the codeword and the comparison with the stored setpoint value is advantageously carried out for each memory circuit with a computer. However, this method fails in testing that read-only memory circuit containing the set values because its code word is changed with the storage of the set value. Object of the invention The aim of the invention is to increase the error detection rate in the ceintiagnosis of read-only memories of microcomputer-controlled systems. Essence of invention The object of the invention is to create a way tu completely check the memory contents of read-only memories with programmatic means, d. H. The basis of the inventive method is the theory of cyclic codes, which is used to determine checksums. The test in the sense of this theory is mathematically the remainder of the remainder formed when all the information of a read-only memory circuit (e.g., EPROM content) is divided by a suitable generator polynomial. With byte-wise organized memories (customary in micro-disk technology), for example, the generator polynomial can be used Pix) = x ' ⁶ + x ¹² + x ⁵ + x ° on, always one 2-byte long remainder (16 dual digits). This generator polynomial ensures a high recognition rate for occurring errors. During the check of the read-only memory, the actual test sum is formed programmatically as a remainder of the remainder for each memory circuit and compared with the respective setpoint check sum stored in a solver test sum3. Errors occurring can thus be do dnet zugeoi the read-only memory, bsi there is a mismatch between the formed Istprüfsumme and Sollprüfsumme.