JPH10198523A - Magnetic disk device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a magnetic disk device in which a normal restoring processing can be attained even when power source failure is generated in any timing. SOLUTION: Whether or not power source failure is generated at the time of writing in the just previous past is judged based on a state signal (a step 102), and in the case of YES, the possibility of the generation of abnormality at the time of writing in the just previous past is judged, and writing data at the time of the generation of the power source failure and the block number of the destination of writing are held in a non-volatile memory, and an MPU (microprocessor unit) obtains the writing data and the block number of the destination of writing through an HDC(hard disk controller), and prepares for writing (a step 103). When the MPU issues a writing instruction to the HDC, the HDC obtains the writing data from the non-volatile memory, and operates writing in a magnetic storing medium (a step 104). When the writing is normally ended, the MPU makes the state signal of a power source failure monitoring circuit inactive by using a clear signal (a step 105).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic disk drive, and more particularly to a write cache function of a magnetic disk drive and a write data recovery function when a power failure occurs.

[0002]

2. Description of the Related Art In a conventional magnetic disk drive, write data sent from a host device is temporarily stored in a volatile memory called a write cache (for example, a semiconductor memory of a type in which stored data is lost when power is turned off). By storing, the time of the write command processing performed between the host device and the magnetic disk device is accelerated. This is because writing / reading to / from the write cache can be performed much faster than writing / reading data to / from the magnetic disk.

In the case of this conventional magnetic disk drive, write data temporarily stored in a write cache is
If a power failure or the like occurs before the writing to the magnetic disk (storage medium) is completed, all data after the occurrence of the power failure is lost because the power of the write cache is turned off.

In order to solve this problem, Japanese Patent Laid-Open No.
No. 73516 proposes a method for providing a non-volatile (non-volatile) memory (for example, a semiconductor memory of a type in which stored data is not lost even when power is turned off) in a data cache unit as shown in FIG. I have.

[0005] In FIG. 3, a processor 21 and a main storage device 22 constituting main parts of a computer system are connected to a bus
It is connected to the. The magnetic disk device 24 is connected to the bus 23.

The magnetic disk drive 24 includes a disk controller 25 and a magnetic disk main body 26. The disk controller 25 includes a disk cache 27, a data control unit 28, and a nonvolatile register group 29.

The disk cache 27 is composed of a volatile memory 30 and a nonvolatile EEPROM (Electrically Erasable Programmable Read Only Memory) 31. The volatile memory 30 stores data read from the magnetic disk main body 26, and Then, data to be written to the magnetic disk main body 26 is stored. In this case, the write data stored in the EEPROM 31 includes the start address of the data block, the number of blocks, and information necessary for writing, in addition to the write data for the magnetic disk main body 26.

The data control unit 28 includes a nonvolatile register group 2
9 and recovery from the failure based on the contents stored in the nonvolatile register group 29. In this case, the nonvolatile register group 29 stores, in a table, the start address of the data block to be written to the EEPROM 31, the number of blocks, information necessary for writing, and the like.

First, when reading data from the magnetic disk device 24, the processor 21 of the computer system, when trying to read one block of data in the magnetic disk main body 26, sends a read request and a data request to the disk controller 25. Sends information required for reading.

When the disk controller 25 receives a read request from the processor 21, the disk controller 25 first stores information in the nonvolatile register group 29 that manages the volatile memory 30 of the disk cache 27 based on information necessary for reading at this time. It is checked whether or not the table requested to be read exists in the table.

If there is a block for which a read request has been made, the block is stored in the main storage device 22 via the bus 23, and then the completion of the read request is notified to the processor 21 for processing. To end.

On the other hand, if there is no block requested to be read, the block requested by the processor 21 is read from the magnetic disk main body 26, and stored in the volatile memory 30 of the disk cache 27. Management information is created in a table of the nonvolatile register group 29 that manages the volatile memory 30.

Also in this case, the disk cache 2
7, the data stored in the volatile memory 30 is transferred to the bus 2
Then, the data is stored in the main storage device 22 via the CPU 3, and then the processor 21 is notified that the read request has been completed, and the processing is terminated.

Next, when writing data to the magnetic disk drive 24, if the processor 21 attempts to write two blocks of data in the main storage device 22 to the magnetic disk main body 26, the processor 21 sends the data to the disk controller 25. In response, a write request and information necessary for writing data are sent.

When the disk controller 25 receives a data write request from the processor 21, the disk controller 25
3, the data sent from the main storage device 22 via
The data is written to the volatile memory 30 of the disk cache 27, and at the same time, the data is sequentially written to the EEPROM 31 based on the address indicated by the nonvolatile register group 29.

When the writing is completed,
The data control unit 28 stores the number of blocks of data written in the EEPROM 31 and information necessary for writing data to the magnetic disk main body 26 in the nonvolatile register group 29.

Next, the disk controller 25 notifies the processor 21 that the write request has been completed,
Based on the information sent from the processor 21, the data of the first one block from the EEPROM 31 is written to the magnetic disk main body 26.

When the data writing for one block is completed, the data control unit 28 updates the start address of the block of the nonvolatile register group 29 to the start address of the second block and updates the number of blocks to one. I do. Then, it is determined whether or not all the blocks of the EEPROM 31 have been written to the magnetic disk main body 26.

If the determination is NO, the next one block of data in the EEPROM 31 is written to the magnetic disk main body 26. And EEPROM
When the writing of all the blocks 31 to the magnetic disk main body 26 is completed, the data control unit 28 sets the start address of the block of the nonvolatile register group 29 to the address of the next block, sets the number of blocks to 0, and The processing is terminated in preparation for the write request.

By the way, if a failure such as a power failure or a system abnormality occurs in the computer system while data is being written to or written to the magnetic disk main body 26 in this manner, EEPRO
The data up to this point is stored in the M31 and the nonvolatile register group 29 as information necessary for recovery.

[0021]

In the case of the proposal described in Japanese Patent Application Laid-Open No. Hei 4-273516 shown in FIG. 3, the writing to the storage medium is completed and before the number of blocks for which writing is suspended is updated. Alternatively, if there is a power failure during updating, the value of the number of blocks for which writing is suspended may become abnormal, and there is a problem that normal recovery processing cannot be performed.

An object of the present invention is to provide a magnetic disk drive capable of performing a normal recovery process even if a power failure occurs at any timing.

[0023]

According to the present invention, there is provided a magnetic disk drive for temporarily storing data to be written to a magnetic storage medium in a nonvolatile memory, comprising: power supply monitoring means for monitoring a power supply state; A power failure monitoring unit that receives power failure information from the power monitoring unit, detects and holds as a status signal whether or not the power failure has occurred while the write data is being written to the magnetic storage medium; When the power supply is restarted after the repair, the status signal indicates that the write data was being written to the magnetic storage medium at the time of the immediately preceding power supply failure. Means for restoring the write data on a medium.

The operation of the present invention is as follows. A magnetic disk device according to the present invention monitors and holds a power supply failure during a data write operation on a magnetic disk (storage medium) and a non-volatile memory for temporarily storing data to be written on a magnetic disk (storage medium). Providing means,
Data recovery is performed using the data contents held in the non-volatile memory when a power failure occurs.

[0025]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of a magnetic disk drive according to the present invention. Referring to FIG. 1, an HDC (hard disk controller) 2 in a magnetic disk device 1 is connected to a higher-level device (not shown) via an interface signal 11. An MPU (microprocessor) 3 controls the HDC 2 and controls a head positioning control circuit and a spindle control circuit (not shown). Further, a volatile memory 6 and a non-volatile memory-7 are connected to the HDC 2.

The read / write control circuit 4 is connected to the HDC 2 by a write data signal 12, a write control signal 13, and a read data signal and a read control signal (not shown). The read / write control circuit 4 modulates the write data signal 12 and performs writing on the magnetic recording medium 5 under the control of the write control signal 13.

The power supply failure monitoring circuit 8 outputs the write control signal 1
3 and the device reset signal 1 output from the power supply monitoring circuit 9
4 has been entered. When the write control signal 13 is active, the device reset signal 14 is active, that is, when a power failure occurs, the status signal 15 becomes active.
The status signal 15 is connected to the MPU 3 and is referred to in the process when the power of the MPU 3 is turned on. The state of the state signal 15 can be made inactive by the clear signal 16 output from the MPU 3.

The state signal 1 output from the power failure monitoring circuit 8
Reference numeral 5 denotes a state held by, for example, a flip-flop, and uses a battery, for example, separately from the power supplied to the magnetic disk device 1, and holds the state even when the power supply to the magnetic disk device 1 is cut off. Is done.

Next, the write cache operation will be described. When the HDC 2 receives a data write command from the host device, it stores the write data in the volatile memory 6 and temporarily suspends interface control. HDC2 is 1
The data of the two block addresses is obtained from the volatile memory 6 and the nonvolatile memory together with the block address of the write destination is obtained.
7 is stored.

Next, the HDC 2 performs writing to the magnetic recording medium 5 via the read / write control circuit 4. When the writing of the number of blocks stored in the volatile memory 6 in advance is completed, the HDC 2 reports to the host device that the writing has been completed.

Next, the processing of the MPU 3 when the power of the magnetic disk device 1 is turned on, particularly the data recovery processing, will be described with reference to the flowchart of FIG.

The MPU 3 obtains the status of the power failure monitoring circuit 8 from the status signal 15 (Step 101). State signal 15
Is inactive, it can be determined that there is no power failure at the time of past writing, and the writing data is normal, so the process ends without doing anything (step 102). When the status signal 15 is active, it can be determined that a power failure has occurred during the past writing, and the write data may be abnormal, so the data recovery process is started (step 102).

Since the non-volatile memory 7 holds the write data and the write destination block address when a power failure occurs, the MPU 3 stores the write data and the write destination block address via the HDC 2. Then, it is set in the volatile memory 6 to prepare for writing (step 103).

When the MPU 3 issues a write command to the HDC 2, the HDC 2 obtains write data from the volatile memory 6 and writes the data to the magnetic storage medium 5 via the read / write control circuit 4 (step 104). When the writing is completed normally, the MPU 3 makes the state of the power supply failure monitoring circuit 8 inactive using the clear signal 16 (step 105).

[0036]

As described above, the present invention provides a nonvolatile memory for temporarily storing data to be written to a magnetic disk (storage medium), and a power supply during data writing to a magnetic disk (storage medium). A means is provided for monitoring and retaining whether or not a failure has occurred. In the event of a power failure, the data stored in the non-volatile memory is used to restore the data, thereby accurately detecting the magnetic disk (storage). This has the effect of detecting a power failure that has occurred during the write operation to the medium and recovering the data on the magnetic disk (storage medium).

[Brief description of the drawings]

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

FIG. 2 is a timing chart at the time of data restoration.

FIG. 3 is a block diagram of an example of a conventional magnetic disk drive.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Magnetic disk drive 2 HDC 3 MPU 4 Read / write control circuit 5 Magnetic storage medium 6 Volatile memory 7 Non-volatile memory 8 Power failure monitoring circuit 9 Power monitoring circuit

Claims (2)

[Claims] 1. A magnetic disk device for temporarily storing write data to a magnetic storage medium in a nonvolatile memory, comprising: a power supply monitoring means for monitoring a power supply state; and power supply failure information from the power supply monitoring means. Power failure monitoring means for detecting and holding whether or not the power failure has occurred during the writing of the write data to the magnetic storage medium as a status signal; and Means for restoring the write data on the magnetic storage medium from the data stored in the non-volatile memory if the write data was being written to the magnetic storage medium at the time of the last power failure in the status signal And a magnetic disk drive comprising: 2. The magnetic disk drive according to claim 1, wherein a block address of a write destination of the write data is stored in the nonvolatile memory simultaneously with the write data.