JP3199010B2 - Magnetic disk drive and data error correction method therefor - Google Patents

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 comprising a plurality of magnetic disks and a plurality of magnetic heads for reading / writing data from / to the recording surfaces of these magnetic disks with the same cylinder. The present invention relates to a data error correction method.

[0002]

2. Description of the Related Art Conventionally, error correction of a magnetic disk device is performed by an ECC control circuit inside a hard disk controller. The ECC control circuit encodes the data for error correction from the original data, adds it to the original data as ECC data, and records the data on the disk.
When an error is detected by the error detection circuit when reading data from the disk, the original data is decoded from the ECC data. Such a conventional technique is described in Japanese Patent Application Laid-Open No. Hei 7-261943.

[0003]

However, the conventional error correction method has a drawback that if a burst error having a size exceeding the correctable byte length occurs, the original data cannot be restored.

[0004]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a magnetic disk drive and a data error correction method capable of restoring data even if a burst error of a size uncorrectable by the prior art occurs. .

[0005]

In the magnetic disk drive according to the present invention, a plurality of magnetic disks stacked at a predetermined interval and data read / write are performed on a recording surface of these magnetic disks by the same cylinder. A plurality of magnetic heads, parity data is recorded on at least one of the recording surfaces of the plurality of magnetic disks, normal data is recorded on each of the remaining recording surfaces, and the parity data is recorded on each of the recording surfaces. der those computed from the normal data in the same cylinder is, this and the parity data
Each of the data corresponding to the parity data has the same number
Are recorded at the same position in the sector of
The parity data is calculated from the data and the parity is calculated.
At the same time as recording the verification data,
It is characterized by the fact that A data error correction method according to the present invention includes a plurality of magnetic disks stacked at a predetermined interval, and a plurality of magnetic heads for reading / writing data on a recording surface of these magnetic disks with the same cylinder. In a magnetic disk drive,
At least one of the recording surfaces of the plurality of magnetic disks is used for parity data, and the remaining recording surfaces are used for normal data, and the parity data and the parity data are used.
And the above data corresponding to the same number on the same cylinder.
Record at the same position in the sector of the
The parity data is calculated from each data and the
At the same time that the data is recorded,
Phi . The operation is:
For example, obtaining an exclusive OR of the normal data
It is.

In the prior art, normal data and its parity data are recorded adjacently on the same recording surface.
Therefore, for example, if a burst error occurs between the normal data and the parity data, the data cannot be restored. On the other hand, in the present invention, the parity data and each data corresponding to the parity data are recorded on different recording surfaces.
Therefore, even if an error occurs on one entire recording surface at the maximum, data on the entire one recording surface can be restored based on the data and parity data on the remaining recording surface.

[0007]

FIG. 1 is a block diagram showing a first preferred embodiment of a magnetic disk drive according to the present invention. Less than,
Description will be made based on this drawing.

The magnetic disk drive of the present embodiment is a hard disk drive, and includes a magnetic disk 3 having recording surfaces 3A and 3B, a magnetic disk 3 'having recording surfaces 3C and 3D, and a recording surface 3A to 3D. Magnetic heads 2A to 2D capable of reading and writing data independently and simultaneously. Recording surface 3A-3
C and the magnetic heads 2A to 2C are for normal data,
The recording surface 3D and the magnetic head 2D are dedicated to parity data.

[0009] Magnetic disks 3, 3 'and magnetic head 2A
2D are stored in the disk stack assembly 1. The magnetic heads 2A to 2D are connected to an encoder / decoder circuit 13 via read / write control circuits 4A to 4D. Since one read / write control circuit 4A to 4D is provided for each of the magnetic heads 2A to 2D, the magnetic heads 2A to 2D can simultaneously read / write. The magnetic disks 3, 3 'and the magnetic heads 2A to 2D are
Is controlled by The servo control unit 5 includes a CPU
(Central processing unit)
3 ′ and the magnetic heads 2A to 2D are controlled. C
PU6, data buffer 14 and encoder / decoder circuit 13 are HDC (hard disk control circuit)
15. The HDC 15 includes a drive control unit 8 that controls data transmission and reception with the encoder / decoder circuit 13, a buffer control unit 11 that controls data transmission and reception with the data buffer 14, and a command It comprises a CPU input / output control unit 7 for controlling and a host interface control unit 12 for performing interface control of data and control signals between a host computer (not shown).

FIG. 2 is an explanatory diagram showing an example of a write operation in the magnetic disk device of the present embodiment. Less than,
A description will be given based on FIG. 1 and FIG.

As described above, the magnetic heads 2A to 2D can simultaneously read / write data. For example, when the magnetic head 2A is writing, the magnetic head 2B can read. FIG. 2 schematically shows the first four bits of a sector of a track (hereinafter, circled numbers represent numbers) on the same cylinder on the recording surfaces 3A to 3D. actually,
Data for one sector (usually 512 bytes) is held in the data buffer 14.

Here, it is assumed that the magnetic head 2A writes 0 → 0 → 1 → 0 sequentially from the bit in the sector of the track. At this time, the magnetic heads 2B and 2C read the data of the sector of the track on the same cylinder and hold the data in the data buffer 14 simultaneously with the writing by the magnetic head 2A. At this time, the magnetic head 2D is in a non-operation state.

Subsequently, during the rotation waiting time until the sector of the track returns to the position of the magnetic heads 2A to 2D,
The XOR operation circuit control unit 9 calculates the exclusive OR of the data on the recording surfaces 3A to 3C and holds the result as parity data. Then, again, the sector of the track is
When the magnetic head 2D comes to the position of A to 2D, the recording surface 3D
Record the parity data. This completes the data write operation on the recording surface 3A. Note that while the magnetic head 2D is writing the parity data, the magnetic heads 2A to 2C can read out the data, thereby verifying (checking) the recorded data.

FIG. 3 is an explanatory diagram showing an example of a read operation in the magnetic disk device of the present embodiment. Less than,
A description will be given based on FIG. 1 and FIG.

Normally, ECC data is added to data, and error correction is performed by the ECC circuit controller 10. When a byte error of a byte length that cannot be corrected by the ECC circuit control unit 10 occurs, the XOR operation circuit control unit 9 performs the next retry read.
Recovers the original data based on the parity data recorded on the recording surface 3D. This operation will be described with reference to FIG.

It is assumed that an error has occurred while the magnetic head 2A is reading data of a sector of a track on a certain cylinder. First, the ECC circuit control unit 10 attempts to correct the error. However, if the burst error has an uncorrectable byte length, the XOR operation circuit control unit 9 reads the parity data of the recording surface 3D and the recording surface 3D when reading by retry.
The exclusive OR is calculated from the data of B and 3C, and the data of the recording surface 3A is restored.

The flow of this operation will be described in more detail. When the magnetic head 2A reads data of a sector of a track on a certain cylinder, each of the magnetic heads 2B to 2D is in a non-operation state. At this time, if the data read by the magnetic head 2A results in an error, the ECC circuit control unit 10 immediately corrects the data. In the case of a burst error that cannot be completely corrected by the ECC circuit control unit 10, the magnetic heads 2B to 2D read the data of the sector of the track on the same cylinder at the time of the next retry read, and perform an XOR operation based on these data. The control unit 9 restores the original data of the recording surface 3A.

FIG. 4 is an explanatory diagram showing a reference example of a write operation in the magnetic disk device of the present embodiment. Hereinafter, description will be made with reference to FIGS.

In the example shown in FIG. 2, when recording the parity data on the recording surface 3D, the data cannot be recorded on the recording surface 3D until the magnetic disks 3, 3 'have rotated once. Therefore, during the rotation waiting period after the parity data calculation is completed and the parity data is recorded,
The next process will be waited.

Therefore, in the example of FIG. 4, the write position of the parity data on the recording surface 3D and the data write position on the recording surfaces 3A to 3C so that the parity data can be recorded on the recording surface 3D immediately after the completion of the parity calculation. And staggered. Thereby, the rotation waiting time at the time of writing can be reduced to a minimum. The specific operations of reading and writing in this example are the same as those in the example of FIG.

[0021]

According to the magnetic disk drive and the data error correction method thereof according to the present invention, since the parity data and each data corresponding to the parity data are recorded on different recording surfaces, the conventional technique is not used. Even if a burst error of uncorrectable magnitude occurs on one recording surface, data can be restored based on the data and parity data on the remaining recording surface, so that reliability can be improved. For example, even if all one recording surface crashes, all the data on the crashed recording surface can be restored based on the data and parity data on the remaining recording surface. And parity
While writing data, read each data
Verifying the recorded data
Can be.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of a magnetic disk drive according to the present invention.

FIG. 2 is an explanatory diagram showing an example of a write operation in the magnetic disk device of FIG.

FIG. 3 is an explanatory diagram showing an example of a read operation in the magnetic disk device of FIG. 1;

FIG. 4 is an explanatory diagram showing a reference example of a write operation in the magnetic disk device of FIG. 1;

[Explanation of symbols]

 2A to 2D magnetic head 3,3 'magnetic disk 3A to 3D recording surface

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ identification code FI G11B 20/18 572 G11B 20/18 572F (58) Investigated field (Int.Cl. ⁷ , DB name) G11B 20/12 G06F 3 / 06 305 G06F 11/10 330 G11B 20/18 570 G11B 20/18 572

Claims (4)

(57) [Claims] A plurality of magnetic disks stacked at predetermined intervals;
Disk and the recording surface of these magnetic disks
Multiple magnets for reading / writing data with cylinder
A magnetic disk device comprising: a magnetic head; and at least one of recording surfaces of the plurality of magnetic disks.
Parity data is recorded on each of the remaining recording surfaces.
Is recorded, and the parity data is recorded in each of the records.
Calculated from normal data in the same cylinder on the surface
ThingAnd Before the parity data and the parity data
Each data is placed in the same position of the same numbered sector.
Is recorded and Calculating the parity data from each of the data,
At the same time as parity data is recorded,
Refining,  A magnetic disk drive characterized by the above-mentioned. 2. The magnetic disk drive according to claim 1, wherein said operation is to obtain an exclusive OR of said normal data. 3. A plurality of magnetic disks stacked at a predetermined interval.
Disk and the recording surface of these magnetic disks
Multiple magnets for reading / writing data with cylinder
Error Correction of Magnetic Disk Drive with Magnetic Head
In the correct method, at least one of the recording surfaces of the plurality of magnetic disks
One for parity dataage,Each remaining recording surface is
Data,The parity data and the parity data
And the above data corresponding to the same number on the same cylinder.
While recording at the same position in the sector of the issue, Calculating the parity data from each of the data,
At the same time as parity data is recorded,
Refining,  Data error correction for a magnetic disk drive
Method. 4. The data error correction method for a magnetic disk drive according to claim 3, wherein said operation is to calculate an exclusive OR of said normal data.