JPH08221214A - Disk device - Google Patents

Abstract

PURPOSE: To early find out the generation of a subsequent bad block, to surely and easily check the block and the make a data restoration possible. CONSTITUTION: When the reading/writing of the data from a host processor 3, etc., is performed for a disk device 1 via disk controller 2, the disk controller performs the ECC error check of data by performing a look-ahead of the data on the same cylinder as the cylinder in which the data is written after the data read by the reading processing of the data is transferred and performs a buffering for the data at the inside. When the access to the disk device does not exist for fixed time or more, the blind reading of data is performed according to the cylinders of the disk device and the ECC error check of the blind read data is performed. When an error is found by the ECC error check, a data resotration is performed if the automatic correction of the error is possible, and the restored data is stored in other area.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk device,
In particular, it relates to checking the occurrence of a late bad block of data and protecting the data.

[0002]

2. Description of the Related Art In a disk device such as a hard disk device or a magneto-optical disk device, which is an auxiliary storage device of a computer, in order to ensure the reliability of stored data, FIG.
As shown in, ECC (Error) is added to the ID part and the data part.
Correcting Code).

This is to confirm the invariance of data by generating an ECC when writing data to a disk device and checking the ECC when reading data from the disk device.

Also, in reading data, in the unlikely event that E
If an error occurs in the CC check, it also has an error correction function of copying correct data from the read data and ECC. However, this error correction function is capable of error correction for garbled bits to some extent (for example, 11 bits or less), but cannot repair data corruption beyond that.

In general, there is always a defective spot (bad block) on the medium of a disk device in which data cannot be written / read accurately. The defect spots include those that are generated at the manufacturing stage of the medium and are specified from the time of purchasing the disk device, and those that are suddenly generated during use of the disk device (so-called late bad block).

As the auxiliary storage device of a computer, a problem in terms of reliability is a late bad block,
Often, it causes failures such as loss of important data.

[0007] The disadvantage of the problem of the late bad block is that
The bad block progresses due to environmental conditions during operation of the device. That is, there is a possibility that a portion of the disk device storing important data may become a late bad block at some point, resulting in the loss of important data.

Although the disk device or the disk controller has the above-described ECC error automatic correction function, if certain data stored in the disk device is occluded in the late bad block, the data is stored for a long time. The problem is that if the data is not read, the abnormality is not noticed.

That is, the data which is not read for a long time is E
There is a possibility that the bad block progresses beyond the extent of being able to be saved by the automatic CC error correction, which may lead to a fatal loss of data for the computer.

In order to avoid this, it is necessary to periodically read all the files stored in the disk device by software and perform an ECC check.

[0011]

In order to avoid data loss due to a late bad block, conventionally, a file is regularly read to perform an ECC check, but this has the following problems.

(1) If a file unnecessary for mischief is read, the access time of actually read data will be significantly increased. That is, since the head is moved (seeked) when reading an unnecessary file, the access time becomes long.

(2) Software that periodically reads the file is required, and this file is complicated and troublesome.

(3) The data area where the file is not stored cannot be checked.

An object of the present invention is to provide a disk device which can detect the occurrence of a late bad block early and can surely and easily check it.

Another object of the present invention is to provide a disk device capable of recovering data when a subsequent bad block occurs.

[0017]

In order to solve the above-mentioned problems, the present invention provides a disk device for reading / writing data from a host processor or the like via a disk controller. After transferring the data read by the reading process, the data is prefetched on the same cylinder of the disk device in which the data was written, the ECC error check of the prefetched data is performed, and the data is internally buffered. Characterized by ringing.

Further, according to the present invention, the disk controller performs idle reading of data for each cylinder of the disk device when an access to the disk device is not made for a predetermined time or more, and performs an ECC error check of the idle read data. Characterize.

In the present invention, when the disk controller has an error in the ECC error check,
If the error can be automatically corrected, data will be restored,
The restored data is stored in another area.

[0020]

Utilizing the fact that the data stored in the disk device is relatively continuous, the track buffering function is used at the time of reading the data to pre-read and ECC.
By performing an error check, it is possible to find a subsequent bad block of data that follows the read data.

By performing idle reading and ECC error check for each cylinder by utilizing the idle time when there is no access to the disk device, it is possible to find the late bad block even in the area having no data.

Data restoration is performed when a late bad block is found, and the data is stored in another area, so that automatic restoration is performed before the bad block progresses.

[0023]

FIG. 1 is a system configuration diagram showing an embodiment of the present invention. The disk device 1 is coupled to the system bus via the disk controller 2, and data writing or reading from the host processor 3 or the like is performed via the disk controller 2.

The disk controller 2 generally performs a lower level access with the disk device 1 and serves to reduce the load on the host processor 3. Therefore, some kind of microcomputer is installed in the disk controller 2.

It should be noted that the recent disk device 1 has a SCSI
Like the interface, the disk device 1 itself has an intelligent function to control the disk. In this case, it is equivalent to the disk device 1 equipped with a disk controller.

In this embodiment, the firmware of the disk controller 2 is provided with the following error scraping function to improve the reliability of data without impairing the performance of the disk device 1. The error scraping function referred to here is to find a late bad block that occurs in the disk device 1 at an early stage, to easily and surely check the bad block, and to rescue the data in the portion where the bad block has occurred.

(1) When there is a data read command from the host processor 3 or the like to the disk device 1, the disk controller 2 reads the data from the designated sector of the disk device 1 and also pre-reads the following data,
Performs track buffering.

At this time, an ECC check is performed to check whether a bad block has occurred. If an error is confirmed, error processing (3) described below is performed.

FIG. 2 shows a data read process by the disk controller 2 in response to a data read command.

The disk controller 2 moves the head to a designated cylinder of the disk device 1 in response to a data read command (step S1), reads the required amount of data from the designated sector (step S2), and reads this data. An ECC check is performed (step S3), if there is an error, error processing is performed (step S4), and if there is no error, the read data is transferred to the host processor 3 or the like (step S5).

The processing up to the above is the same as the conventional data reading processing. After this processing, the disk controller 2 performs track buffering by utilizing the fact that the data stored in the disk device 1 is relatively continuous.

That is, the disk controller 2 performs data pre-reading to read the data of the same cylinder following the read data (step S6), performs an ECC error check on the pre-read data (step S7), and if there is an error, Perform error processing (step S
4) If there is no error, the preread data is buffered (step S8).

As a result, in the data read process, it is possible to automatically check whether or not there is a late bad block for the prefetched data, and it is possible to find an advancing bad block at an early stage. Further, by finding the occurrence of the late bad block at an early stage, it is possible to automatically correct the data of the part before the progress of the bad block with a higher probability.

In this process, since it is performed every time the data read process is executed and the read position is changed each time, the disk controller does not have a function of randomly setting the error check target position and time. A reliable and easy check of the occurrence of a late bad block can be performed at appropriate time intervals and appropriate data positions.

Further, this is a check of pre-read data following the read data, and this pre-read data is also important data and there is a high possibility that a late bad block will occur.
Checking the prefetch data is convenient for increasing the reliability of the entire data.

The buffering of the prefetched data is stored in the disk controller 2. This enables high-speed data transfer from the disk controller 2 without reading from the disk device 1 if the next data read command is in the pre-read data.

(2) When the host processor 3 or the like does not access the disk device 1 for a certain period of time (for example, 10 minutes), the disk controller 2 moves the disk device 1 by one cylinder to transfer data. Read blank and perform ECC check. When an ECC error occurs, error processing described later is performed.

The idle reading is performed endlessly on all the cylinders of the disk device 1 by sequentially moving the cylinders aiming when the disk device 1 is not accessed.

FIG. 3 shows a mode of the idle reading process, in which the idle reading is executed when 10 minutes have passed from the end of the access period T _A from the host processor 3 or the like to the disk device 1. When there is no access even when the idle reading of a certain cylinder #n is completed, the idle reading of the next cylinder # n + 1 is performed.

By this processing, in addition to the effect of the pre-reading processing, the ECC check is performed by the disk controller 2 even for data that is not read / written for a long time by the host processor 3 or the like, and the late bad block can be found early. it can.

Further, since the check is made for the data area which is not used, it is not necessary to write the data in the portion where the bad block has already occurred.

Further, since the idle time of the disk device is used, the access time performance of the disk device is hardly affected. Further, since the disk controller processes the firmware, no burden is imposed on software such as the host processor.

Further, since the idle reading process for each cylinder is performed, high speed processing can be performed and the track buffering function can be used.

Further, since the idle reading is performed for each cylinder at appropriate time intervals, the head of the disk device does not stay on the same cylinder for a long period of time, and dust accumulated on the surface of the disk medium is removed to remove the dust on the disk medium. The reliability can be improved.

(3) In the data read process or the idle read process described above, when an error occurs in the ECC check, the disk controller 2 performs the error process.

When this error occurs, the disk controller 2 performs the ECC error automatic correction if possible, restores the data, and stores it in the spare area.

FIG. 4 shows the error processing, and it is judged whether or not the error content can be automatically corrected (step S11), and the uncorrectable error is notified to the host processor 3 etc. as an abnormality of the disk device 1 ( Step S12).

For a correctable error, the error is automatically corrected (step S13), a spare area for storing the restored data is prepared (step S14), and the restored data is stored in the spare area (step S15). ).

In the replacement method in this processing, for example, a normal track area and a replacement track area are set in the disk device 1, and when a bad block occurs in the data of the normal track area, the repaired data is replaced with the replacement track area. It is realized by changing the cylinder number and the like associated with this track change.

By the error processing using this replacement method,
The burden on software such as the host processor is eliminated.

[0051]

As described above, according to the present invention, the disk controller transfers the data read in the data read processing, then prereads the data by using the track buffering function, and the preread data is read. E
Since the CC error check is performed and the data is buffered internally, it is possible to early detect the occurrence of a subsequent bad block of the data that follows the read data by using the data read processing, and to easily generate the bad block. You can check it. Further, the data buffering enables high-speed reading when reading preread data subsequent to the read data.

Further, according to the present invention, the disk controller reads the data for each cylinder of the disk device when the disk device is not accessed for a certain period of time or more, and performs an ECC error check on the read data. In addition to early detection of the occurrence of a late bad block and the easy and reliable check of the occurrence of a bad block, the following effects are obtained.

(1) The ECC check is performed by the disk controller even for the data that is not read / written for a long time, and the late bad block can be found early.

(2) Since the check is made for the unused data area, it is not necessary to write the data in the location where the bad block has already occurred.

(3) Since the free time of the disk device is used, the access time performance of the disk device is hardly affected. Further, since the disk controller processes the firmware, no burden is imposed on software such as the host processor.

(4) Since the idle reading process is performed for each cylinder, high speed processing can be performed and the track buffering function can be used.

(5) Since the idle reading is performed for each cylinder at appropriate time intervals, the head of the disk device does not stay on the same cylinder for a long period of time, and dust accumulated on the disk medium surface is removed to remove the disk medium. The reliability of can be improved.

Further, according to the present invention, when there is an error in the ECC error check, the disk controller repairs the data if the error can be automatically corrected and stores the repaired data in another area. Therefore, data can be restored without burdening the software such as the host processor.

[Brief description of drawings]

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

FIG. 2 is a data read process in the embodiment.

FIG. 3 is an empty reading process of data in the embodiment.

FIG. 4 is an error processing in the embodiment.

FIG. 5 is a data format of a disk device.

[Explanation of symbols]

 1 ... Disk device 2 ... Disk controller 3 ... Host processor

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location G11B 20/18 572 9558-5D G11B 20/18 572F

Claims (3)

[Claims] 1. A disk device that reads / writes data from a host processor or the like via a disk controller, wherein the disk controller transfers the data read in the data read process, and then writes the data in the disk device. A disk device characterized in that data is pre-read on the same cylinder in which the data is written, an ECC error check is performed on the pre-read data, and the data is internally buffered. 2. The disk controller performs idle reading of data for each cylinder of the disk device when there is no access to the disk device for a certain period of time or more, and performs an ECC error check of the idle read data. Item 1. The disk device according to item 1. 3. The disk controller is the EC
3. The disk device according to claim 1, wherein when there is an error in the C error check, the data is restored if the error can be automatically corrected and the restored data is stored in another area.