JPH06231049A - Semiconductor disk device - Google Patents

Abstract

PURPOSE:To prevent the operation of a system from being affected even when a fault which can not be automatically corrected is generated. CONSTITUTION:Data are divided, and stored in plural memory modules 104-108. In response to the generation of a fault which can not be automatically corrected in any memory module, the faulty memory module is separated by a separating part 110. When the faulty memory module is separated, correct data are prepared and transmitted based on the storage contents of the modules except the faulty memory module by a memory correcting part 109 in response to an access from an outside device. When the new module is integrated instead of the faulty module, the content to be stored in the faulty module is prepared, and written in the new module based on the storage contents of the modules except the faulty module by a data preparing part 121.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor disk device, and more particularly to a semiconductor disk device which is one of the peripheral devices of a computer system.

[0002]

2. Description of the Related Art Conventionally, in this type of semiconductor disk device, when a memory failure occurs, an error correction code is used to perform automatic correction. The conventional semiconductor disk device will be described with reference to FIG. In FIG. 3, a conventional semiconductor disk device 301 includes a storage unit 303 including a memory 304 for storing data and an error correction code 305, and a memory control unit 302 for reading and writing to the memory 304. It is configured.

In such a configuration, when there is an access from a host device (not shown) such as a CPU, the memory control section 302 reads or writes data from or to the memory 304 in the storage section 303. When a memory failure occurs, if it can be automatically corrected, the error correction code is used to perform the automatic correction.

[0004]

In the conventional semiconductor disk device described above, the error correction code is used for automatic correction. However, if a memory failure that cannot be automatically corrected occurs during system operation, the error correction code is used. The memory had to be replaced, and the entire semiconductor disk device had to be separated from the system. Therefore, there is a drawback that it affects the system operation.

The present invention has been made in order to solve the above-mentioned conventional drawbacks, and its object is a semiconductor disk which does not affect the operation of the system even when a memory failure which cannot be automatically corrected occurs. It is to provide a device.

[0006]

SUMMARY OF THE INVENTION A semiconductor disk device according to the present invention responds to a plurality of memory modules for dividing and storing data and a failure that cannot be automatically corrected in any of the plurality of memory modules. And disconnection means for disconnecting the faulty memory module, and in response to an access from an external device in this disconnected state, correct data is created based on the stored contents of the memory modules other than the faulty memory module. And a content to be stored in the faulty memory module based on the storage contents of the memory modules other than the faulty memory module in response to the incorporation of a new memory module in place of the faulty memory module. And writing means for writing the data into the new memory module. And it features.

[0007]

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

FIG. 1 is a block diagram showing the configuration of a first embodiment of a semiconductor disk device according to the present invention. In the figure, a semiconductor disk device 1 according to a first embodiment of the present invention is shown.
01 includes a memory unit 103 and a memory control unit 102.

The memory section 103 is composed of five memory modules 104 to 108 in this example. Data including parity is divided and stored in these memory modules, and an error correction code for correcting an error in each memory module is further added. In addition, a relay switch R for indicating whether or not the memory module itself is logically connected to a host device (not shown) is provided in each memory module.

It is assumed that each memory module is composed of a printed wiring board and is connected to the memory unit 103 via a connector (not shown). And, a structure that can insert and remove the memory module without turning off the power of the device,
That is, the structure is such that hot-plugging is possible.
This allows the relay switch R to be opened and logically disconnected from the host device, and then physically disconnected from the connector.

When there is an access from the host device to the memory unit 103, the memory module 104
Data with parity is created by combining the storage contents of 108 to 108 and sent to the host device. In addition,
Parity is stored in the memory module 108, and the entire data can be created (restored) using this parity.

The memory control unit 102 switches a memory correction unit 109 for creating data using parity, registers 1 and 2 having bits corresponding to the memory modules 104 to 108, and a relay switch in each memory module. The memory module is separated from the host device to prevent access to the memory module 110, and a pull-up resistor 111 is included.

In such a configuration, when all the memory modules in the memory unit 103 are normal, data is read / written for access from the host device without any problem. Further, even when a failure that can be automatically corrected occurs in the memory module, the data is read / written without any problem by being automatically corrected by using each error correction code.

However, when a failure that cannot be automatically corrected occurs, the memory module is separated by the following operation. For example, when a failure that cannot be automatically corrected occurs in the memory module 105, the memory correction unit 1
09 detects that, and rewrites the bit corresponding to the memory module 105 in the register 2. This register 2 indicates the memory module in which a failure that cannot be automatically corrected occurs. Then, the disconnecting unit 110 refers to the content held in the register 2 to switch the relay switch R in the memory module 105 to open it. This allows
The memory module 105 is logically separated from the host device.

Next, the disconnecting section 110 rewrites the bit corresponding to the memory module 105 in the register 1 to display that the memory module 105 is logically disconnected. As a result, the operator or the like can operate the memory module 10
5 can be exchanged. Each bit in the register 1 is protected by the pull-up resistor 111, and the value “0” is stored when the relay switch is closed, and the value “1” is stored when the relay switch is open. That is, this register 1 indicates the memory module logically connected to the host device.

By the way, even when the memory module 105 is logically separated from the host device, the access from the host device is performed. In that case, the memory modules 104 and 106 other than the memory module 105
Based on the storage contents of 108 to 108, the memory correction unit 109 creates (repairs) correct data each time access is made, and sends it to the higher-level device. By doing so, the operation of the system will not be interrupted.

Next, a second embodiment of the present invention will be described.

FIG. 2 is a block diagram showing the configuration of a second embodiment of the semiconductor disk device according to the present invention, and the same parts as those in FIG. 1 are designated by the same reference numerals. In the figure, a semiconductor disk device 201 according to a second embodiment of the present invention.
Is a built-in recognition unit 120 that recognizes that a new memory module is built in due to a change in the contents held in the register 1 in the configuration of the device of FIG. A data creation unit 121 that creates and writes the contents to be stored in the built-in memory module is added. It is assumed that the relay switch in the new built-in memory module is in the closed state.

The operation of the semiconductor disk device of this embodiment having such a configuration is the same as that of the first embodiment in that the disconnecting section 110 disconnects the memory module in which a failure has occurred. In this embodiment, in response to the incorporation of a new memory module in place of the faulty memory module, the contents to be stored by the faulty module are created based on the storage contents of modules other than the faulty memory module. However, the writing function is added to the newly incorporated module.

That is, as in the case of the first embodiment described above, a failure that cannot be automatically corrected occurs in the memory module 105, and the disconnection unit 110 causes the memory module 1 to operate.
The relay switch R in 05 is opened and the register 1
It is assumed that the bit corresponding to the module 105 therein is rewritten to be "1". In this state, when the operator or the like removes the faulty memory module and connects a new memory module, the relay switch R in the memory module is in a closed state. It is rewritten and becomes "0".

The built-in recognition unit 12 recognizes the change in the register 1.
When 0 is detected, the fact is notified to the data creation unit 121. Upon receiving this notification, the data creation unit 121 refers to the storage contents of the memory modules other than the newly incorporated memory module 105, and determines the contents to be stored in the memory module 105 during the period when the higher-level device does not access. It is created (restored) and written in the memory module 105.

When all writing to the memory module 105 is completed, the built-in recognition unit 120 is notified, and the built-in recognition unit 120 rewrites the corresponding bit in the register 2 and returns it to the original. As a result, the memory correction unit 109 does not perform data creation (restoration) after returning to the normal state, and the stored contents of the memory modules 104 to 108 are read.

As described above, even if a failure that cannot be automatically corrected occurs, the memory module can be replaced. Therefore, it is not necessary to disconnect the entire apparatus from the system or to turn off the power of the apparatus. It does not affect the operation.

[0024]

As described above, according to the present invention, the faulty memory module is disconnected in response to the occurrence of a fault that cannot be automatically corrected, and a new memory module is incorporated in place of the faulty memory module. By responding and creating and writing the contents to be stored in the faulty memory module based on the stored contents of the memory modules other than the faulty memory module, the faulty memory module can be replaced without affecting the operation of the system at all. There is.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a semiconductor disk device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a semiconductor disk device according to a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a conventional semiconductor disk device.

[Explanation of symbols]

 1 and 2 registers 102 and 202 memory control unit 103 memory unit 104 to 108 memory module 109 memory correction unit 110 disconnection unit 111 pull-up resistor 120 embedded recognition unit 121 data creation unit

Claims (2)

[Claims] 1. A plurality of memory modules for dividing and storing data, and a disconnecting means for disconnecting the failed memory module in response to a failure that cannot be automatically corrected in one of the plurality of memory modules. And a creating means for creating and transmitting correct data based on the stored contents of a memory module other than the faulty memory module in response to an access from an external device in the disconnected state. Semiconductor disk device. 2. The content to be stored in the faulty memory module is created based on the storage contents of the memory modules other than the faulty memory module in response to the incorporation of a new memory module in place of the faulty memory module. 2. The semiconductor disk device according to claim 1, further comprising write means for writing to the new memory module.