JPS58171793A - Recovering system of fixed fault of control storage device - Google Patents

Abstract

PURPOSE:To improve the reliability, by dividing a storage device to two blocks and exchanging two logical blocks to be written in respective blocks in accordance with the internal state of a fixed fault and storing them when the fixed fault occurs in the storage device. CONSTITUTION:When a program is loaded to a control storage device CS1, written contents are read out again and are checked. If errors are detected, the same contents are loaded again and are checked; and if errors are detected though this loading is performed plural times, the occurrence of the fixed fault is discriminated. Since the fixed fault is the degeneration fault where only ''0'' or ''1'' is read out independently of written data, data having the degenerated value is written. The device CS1 is divided to two blocks A and B; and when the fixed fault exists in the block A, the logical block to be written in the block A is written in the block B. Thus, the fixed fault is recovered.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of the Invention] The present invention relates to a control storage device in a microprogram-controlled information processing device, and particularly to a control storage device of the type that is used by loading a microprogram from an external storage device (i.e., not a ROM). The present invention relates to a rescue method when a fixed failure occurs in a storage device (hereinafter abbreviated as C8).

[Prior art to the invention]

Conventionally, the following techniques have been considered for improving the reliability of storage devices.

(1) A method of increasing the reliability of storage devices by adding an ECC circuit to detect failures and correct error contents.

(2) A method of effectively utilizing a storage device even if there is a fixed failure by adding a data inverting circuit to the write and readout circuit of the storage device and storing data tailored to the failure state.

(3) i! [18 Method of duplicating the device and switching to use when an abnormality is detected.

These methods all depend on the access method of the storage device itself, that is, by processing only the data to be stored while keeping the physical address fixed, or by adding information to check the normality of the storage state. Efforts are being made to improve reliability.

Therefore, there is a drawback that the number of peripheral circuits of the storage device or the storage device itself increases.

[Purpose of the invention]

In order to increase the reliability of p1 storage, the present invention provides a means for storing information that matches the internal state of the p1 storage device, which is different from the method of processing the storage contents, thereby increasing the reliability relatively. This is a method that allows for improvement.

[Key points of the invention]

The present invention is one of the methods for optimizing the internal state of a storage device and the information to be stored in the storage device.

In the conventional method of assigning addresses to storage devices in a fixed manner, the pattern of information that can be stored in the area specified by a specific address of the storage device may change depending on its internal state; As the storage capacity increases, the influence of internal state abnormalities on the entire storage device increases.

This method determines the address of the storage device based on the information to be stored, and a block, which is the minimum unit of address conversion, is made up of one candy or more words, which is the minimum unit of storage in the storage device. However, by providing means for replacing the block when an abnormality is detected in the block, the reliability of the storage device can be improved by realizing use of the storage device that is not affected by abnormal conditions.

[Embodiments of the invention]

FIG. 1 shows a first embodiment of the present invention, where 1 indicates the control [i[
18 storage device (C8), 2 is a control unit, 3 is an external storage device, 4 is a write bus, 5 is a continuation pass, 6 is an address bus, 7 is an exclusive OR (FOR) gate, 8 is a 7-rip 7
Numerals 0 and 9 are parity check circuits.

When loading a program from external storage 1 and 3 to the C8I, the loaded content is re-loaded and the parity check circuit 9
It is checked whether or not there is a failure in cs1 by checking or by comparing the data chapters in the external storage device 3. If there is any deviation in the indulgence data, reload the same content and check it in the same way. Even if it is written multiple times v4)
If this cannot be fixed, it can be considered a fixed failure. In general, in the case of a fixed failure, no matter which one of 898% 1# is praised, '0@
.

1. For degeneration where the bladder can only be protruded in either side of Jl.
It is normal for it to be harmful. In such a case, if data that originally had the degenerate value is written, the starting data will have an apparently permanent value. Therefore, in the first embodiment of the present invention, C81 is divided into two blocks A and B, and when a fixed failure occurs when trying to write certain data to address A1 in A block, the logic Try writing block person to physical block B and logical block B to physical block A. In this case, if the degenerate value of the fixed fault in the physical address A1 is the same as the value of the corresponding bit in the old logical address, the fault will no longer be detected as a result of the above replacement. Therefore, the fact that the exchange has been performed is stored in the flip-flop 8, and the most significant bit of the address bus 6 (
If an EOR gate is interposed between the two inputs (or any bit may be used) and the output of the flip-flop 8 is connected to the other input, the physical blocks can be exchanged and accessed while the logical address remains the same. Note that the address bus 6 actually has a plurality of bits.

Note that the number of blocks to be divided is not limited to the two blocks mentioned above, and can be set arbitrarily. It goes without saying that the number of flip-flops 8 and EOR gates 7 should also be increased accordingly.

Further, the C8I is not only used for storing microprograms, but may also be used for storing operand data. In this case, there should be no fixed failure in the operand data area, so if a fixed failure occurs in the operand data area during IPL or during use, it is recommended to replace it with the microprogram area. .

FIG. 2 is a block diagram showing a second embodiment of the present invention.
, cs, a logical address and physical address conversion table is provided for each block. In the figure,
11 is a microinstruction address quanta, and the upper part (M
ICH) and a lower part (MICL), with the upper part indicating the plotter address and the lower part indicating the intra-block address. Normally, MICL is incremented and processing proceeds. At the time of a branch that spans a branch, etc., a microinstruction that operates MICH is first read into the instruction register 12, the translation table 13 is indexed by the -ring block address part, and the physical block address is read out via the AND gate 14. is set to MICH. Thereafter, branch instructions with intra-block branch addresses are successively issued to the instruction register 12 and set to MICL (*1 bus).

In this way, in Roku C8, I! If the error does not disappear even after rewriting and rereading, it is assumed that it is a fixed failure and another logical block is tried to be loaded into the physical block containing the failure.

As a result, if a fixed failure is detected again, try loading another logical block. If as a result, the fixed fault is no longer detected, the logical block originally stored in the physical block with the fixed fault is stored in the physical block in which the other logical block was stored. That is, both blocks are exchanged. Then, the logical address/physical address correspondence after the replacement is registered in the table 13, and the processing is continued.

These controls and settings of the table 13 can be easily performed by a service processor (not shown). Also, initial program load to Cal (
It goes without saying that each block may be checked during IPL (IPL), and if a fixed failure is detected at that time, the logical blocks may be relocated.

〔Effect of the invention〕

As described above, according to the present invention, even if a fixed failure occurs in the C8, by rearranging blocks, it becomes possible to continue processing while effectively utilizing the capacity of the C8. It is highly effective in improving sexual performance.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of one embodiment of the present invention, and FIG. 2 is a block diagram of another embodiment of the present invention. In the figure, 1 is a control storage device, 3 is an external storage device, 8 is a 7-lip flop, 1
3 is a logical/physical address conversion table. 1 diagram

Claims (3)

[Claims] (1) In a control storage device that is used by loading a microprogram from an external storage device and is equipped with means for detecting stuck-at faults, the relationship between mm addresses and real addresses in blocks of a predetermined size. and detecting whether or not a fixed fault occurs by writing the contents of other logical T-locks to the physical block containing the fixed block fault when a fixed fault occurs in the control storage device. However, if an identified failure occurs, the contents of another logical block are further written and checked in the same way, and if no fixed failure occurs, it is determined whether the logical block has been or should be stored in the future. The processing block that was stored in the physical block that caused the fixed failure is transferred to the physical block that caused the fixed failure, and information indicating the purpose of exchanging these two physical blocks is stored in the storage means to continue the dynamic operation. A fixed fault relief method for a control storage device, characterized in that: (2) The control according to claim 1, wherein the storage means is a logical address/physical address simple table indexed by logical address for all physical blocks in the control storage device. Fixed failure recovery method for storage devices. (3) The fixed fault relief system for a control storage device according to claim 1, wherein the storage means is comprised of a bistable element that stores whether or not a predetermined bit of a logical address is to be inverted.