JPH05314017A - Data recovering system - Google Patents

Abstract

PURPOSE:To recover the data of a data storage device through simple configuration. CONSTITUTION:This system is constituted of the data storage device 11, a register data register 12 to hold the data to be stored, a register address register 13 to hold a write/read address, a data comparator 14 to compare the data stored in the data storage device 11 with the data held in the data register 12, an information processor 1 including a data selector 15 to switch the path of the data to a data register 12 side when non-coincidence is detected as the result of the comparison of the data and an address selector 16 to switch the path of the address to an address register 13 side when the non-coincidence of the comparison of the data is detected, and a diagnostic device 2 including an error flag 21 to be erected at the time of detecting the non-coincidence of the comparison of the data.

Description

Detailed Description of the Invention

[0001]

The present invention relates to a data recovery system,
In particular, the present invention relates to a data recovery method for recovering data in a storage unit (RAM, register file, etc.) of the information processing apparatus and retrying the data processing apparatus when a failure occurs in the information processing apparatus.

[0002]

2. Description of the Related Art Conventionally, in this type of data recovery method, when data in a storage device has an error, ECC (ER
ROR CORRECTING CODE) is generally used to correct one bat error.

[0003]

As described above, in the conventional data recovery method, the recovery is possible only in a simple case such as a 1-bit error, and the recovery becomes considerably difficult when an error of 2 bits or more occurs. Will end up. Further, in such a case, there is a drawback that the information processing apparatus is brought down.

[0004]

The system of the present invention is a data recovery system in a data processing device comprising an information processing device and a diagnostic device having a storage means of a store-through structure, wherein the information processing device has the storage device. Data holding means for holding data to be written in the means, address holding means for holding the read / write address of the storage means, data comparison means for comparing the data of the data holding means with the data of the storage means, and the data Data switching means for switching the data to be written in the storage means from the diagnostic device to the data in the data holding means when the comparing means detects a data mismatch, and reading the storage means from the diagnostic device when the data does not match. Address switching to switch the write address to the address of the address holding means And a flag indicating that the data comparison unit has detected a data mismatch, and the diagnostic device temporarily stops the processing in the information processing device when the flag is set. The present invention is characterized in that the storage is stopped and writing is performed to the storage means according to the held contents of the address holding means and the data holding means.

[0005]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings.

Referring to FIG. 1 showing an embodiment of the present invention, the present embodiment has a data storage device 11 and a data register 1.
2, an information processing device 1 including an address register 13, a data comparator 14, a data selector 15 and an address selector 16 and a diagnostic device 2 including an error flag 21.

The data storage device 11 stores data and the like for the information processing device 1 to perform various processes, and has a store-through memory configuration. The data register 12 is
It is a register that holds the data stored in the data storage device 11, and its output is stored in the data comparator 14 and the data selector 15. The address register 13 is a register that holds the write / read address of the data storage device 11, and its output is input to the address selector 16.

The data comparator 14 includes a data storage device 11
The data stored in the data register 12 is compared with the data held in the data register 12. The comparison result is the diagnostic device 2
Sent to. The data selector 15 includes a data comparator 14
As a result of the comparison, when a mismatch is detected, the diagnostic device 2 switches the data path to the data register 12 side. The address selector 16 switches the address path to the address register 13 side by the diagnostic device 2 when the data comparator 14 detects a mismatch as a result of the comparison.

The error flag 21 in the diagnostic device 2 is set when the data comparator 14 detects a data mismatch. Further, when the diagnostic device 2 receives an error report or the like, it can temporarily stop the information processing device 1 or change the selection direction of the data selector 15 or the address selector 16 of the information processing device 1.

Next, the operation of this embodiment will be described.

When the information processing device 1 is operating normally, the data selector 15 is switched to the opposite side to the output of the data register 12, and the address selector 16 receives the output of the address register 13. It has been switched to the opposite side.

Each time the data storage device 11 receives a write enable, data is stored in the data storage device 11 at the designated address. At the same time, since the data path is also stored in the data register 12, the data is held in the data register 12 every time the data is stored in the data storage device 11. Similarly, as for the address, since the address path is stored in the address register 13, the address at that time is also held in the address register 13.

Since the data storage device 11 is a store-through system, the stored data is always visible as an output of the data storage device 11 unless the address is changed. Also,
The data comparator 14 compares the output of the data storage device 11 with the output of the data register 12, and checks whether the data is correctly stored in the data storage device 11 each time the data is stored in the data storage device 11.

If some trouble occurs in the data storage device 11 and the stored data and the data in the data register 12 do not match in the data comparator 14 as a result, the data comparator 14 causes the diagnostic device 2 to operate. The error flag 21 is set. When the diagnostic device 2 finds that the error flag 21 is set, it immediately causes the information processing device 1 to suspend the processing.

Next, the diagnostic device 2 switches the data selector 15 to the data register 12 side and the address selector 16 to the address register 13 side. At this time, the data register 12 holds the data stored in the data storage device 11, and the address register 13 holds the address stored in the data storage device 11. Therefore, the diagnostic device 2 issues a write enable to the data storage device 11 once to recover the data. The data and address at this time are the write data and address when the data comparator 14 detects a mismatch.

The data stored in the recovery is also checked for data coincidence by the data comparator 14 as in the normal operation, and if there is no discrepancy, the error flag 21 of the diagnostic device 2 is reset. Cancels the temporary suspension of the information processing device 1, and the information processing device 1 returns to the normal processing.

However, when the data comparator 14 still detects a data mismatch even after the recovery,
Since the error flag 21 of the diagnostic device 2 remains on, the diagnostic device 2 does not operate the information processing device 1 again as a fault of the data storage device 11 and performs the fault process.

In this way, the consistency of the stored data is always checked, and if they do not match, the data register 12
The data stored in the data storage device 11 is recovered by the data held in the information processing device 1 to continue the operation of the information processing device 1.

[0019]

As described above, according to the present invention, the data stored in the data storage device is constantly compared and matched, and the data for recovery is also held in the data register. Data can be recovered with. As a result, not only the 1-bit inversion error as in the ECC but also the data itself is compared and checked. Therefore, even if the data has an error of 2 bits or more, the data can be easily recovered.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

[Explanation of symbols]

 1 Information Processing Device 11 Data Storage Unit 12 Data Register 13 Address Register 14 Data Comparator 15 Data Selector 16 Address Register 2 Diagnostic Device 21 Error Flag

Claims (1)

[Claims] 1. A data recovery method in a data processing apparatus comprising an information processing apparatus and a diagnostic apparatus having a storage means having a store-through structure, wherein the information processing apparatus holds data to be written in the storage means. An address holding means for holding the read / write address of the storage means, a data comparison means for comparing the data of the data holding means with the data of the storage means, and when the data comparison means detects a data mismatch Data switching means for switching the data written in the storage means from the diagnostic device to the data in the data holding means; and the read / write address of the storage means from the diagnostic device when the data does not match, to the address of the address holding means. And an address switching means for switching, The disconnection device includes a flag indicating that the data comparison means has detected a data mismatch, and when the flag is set, the diagnostic device suspends the processing in the information processing device, and the address holding means and the A data recovery method characterized in that writing to the storage means is performed according to each held content of the data holding means.