JP2003228925A - Write data guaranteeing method for magnetic disk device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a function of not making a magnetic disk device deteriorate in performance as much as possible by detecting an abnormal state of the magnetic disk drive itself and preventing illegal writing of the magnetic disk drive without any load on a host computer. <P>SOLUTION: The magnetic disk drive itself reads data written by using a head whose write abnormality or temperature abnormality is detected and confirms that the data are coincident. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a write data check technique for a storage device, and more particularly to a read data check technique for preventing write fraud caused by a malfunction of a magnetic disk device itself.

[0002]

2. Description of the Related Art In recent magnetic disk devices, when foreign matter or the like adheres to the head as the recording frequency increases, the flying height of the head becomes abnormal, and the written data is not actually written and the old data is not written. May remain.

In order to solve this problem, 1) a method of reading the written data by periodically using a specific cylinder for diagnosis from the host computer and checking the data, 2) after powering on the system A method is generally used in which the drive (magnetic disk device) itself performs write processing using all the heads at the time of booting, and checks whether the written data can be correctly read.

There is also a method of checking the data by periodically reading the data written by the drive itself using a specific cylinder for diagnosis, as in Japanese Patent Laid-Open No. 2001-35096.

[0005]

In the prior art, since the periodical check is performed, even if a foreign matter or the like adheres to the head during operation and an abnormal state occurs, the written data is not checked for a certain period of time and seek is performed. Due to such operations, the foreign matter is removed, and the next time the check is performed, the state becomes normal and there is a problem that an abnormality cannot be detected.

Further, in general, when recovery is made by retry due to writing error or the like, it is general that error report is not sent to the host computer unless the retry reaches a certain number of times. Cannot recognize abnormal conditions.

Even if an error report is issued, the host computer cannot carry out the read check with the physical head in mind.

In the embodiment of the present invention, the drive itself detects the abnormal state of writing in the magnetic disk device, detects the abnormal state, and executes it without burdening the host computer.
It is to provide a function that does not deteriorate the performance of the magnetic disk device.

[0009]

[Means for Solving the Problems] The magnetic disk device itself reads the data written by using the head to the head which has detected the writing abnormality or the temperature abnormality, and confirms that the data match. It is achieved by

Further, depending on whether or not the next command is received,
This is achieved by determining whether to perform a lead check.

[0011]

1 is a block diagram of an embodiment of a magnetic disk controller to which the present invention is applied.

In FIG. 1, 10 is a host computer, 20 is a control device, and the host computer 10 and the control device 20 are connected via a host interface 50.

Reference numeral 70 denotes a magnetic disk, which is a control device 20.
Are connected via the disk interface 60. The control device 20 is an interface controller 2 that controls data transfer with the host computer 10.
1. A hard disk controller 22 that controls data transfer to and from the magnetic disk 70, and a buffer controller 23 and a buffer 24 that absorb the difference in the data transfer speeds thereof and control the data transfer inside the control device 20. , Temperature sensor 2 to check the temperature
5, a control processor 26 for controlling them, and a volatile internal memory 30 for storing control information and the like.

A write command from the host computer 10 is notified to the control processor 26 by the interface controller 21 via the host interface 50.

Upon receiving the write command, the control processor 26 saves the request address and the length thereof in the host request logical address table 31 and the host request length table 32, and also performs the conversion from the logical address to the physical address. Physical address table 3
Save to 3.

Next, the data reception activation from the host computer 10 is activated in the interface controller 21 and the buffer controller 23, and at the same time, the write activation is implemented in the hard disk controller 22 based on the physical address table 33 and the host request length table 32.
The writing operation is performed on the magnetic disk 70. Further, the written data is saved so as to remain in the write data segment 27 of the buffer (Cache) 24. When the writing to the magnetic disk 70 does not occur abnormally and ends normally, a normal end report is sent to the host computer 10 via the interface controller 21.

When an abnormality such as a positioning error is detected during the writing process on the magnetic disk 70, the physical address where the error occurred is recorded in the error occurrence address 34, and the writing to the magnetic disk 70 is retried. carry out.

When the writing to the magnetic disk 70 is normally completed by the retry, conventionally, the host computer is notified of the normal completion without doing anything. However, when writing is completed by retry, from the viewpoint that an error may have occurred in the written data, the written physical address table 33,
The hard disk controller 22 is activated to read the length table 32.

At this time, the buffer controller 23
Data is read to the read (compare) segment 28 of the buffer (Cache) 24. If the read process is performed only for the head on the error occurrence address 34, the processing time can be reduced.

If an error occurs during the read process and recovery cannot be performed even by retrying, an error report is sent to the host computer 10 via the interface controller 21. At this time, in the sense data indicating the error details, the error logical address that cannot be recovered by the read is reported to the host computer 10 together with the error content.

When the read process is completed including the retry and the error address is reported to the host computer 10 even when the unrecoverable error occurs, the write data segment 27 and the read (compare) segment 28.
Perform a data compare check.

When the compare is completed normally, a normal end report is sent to the host computer 10. If a data mismatch is found in the compare, a dedicated error code is used to notify that a write error has occurred.

When a data mismatch due to a compare occurs, there is a high possibility that another data error has already occurred (has already occurred). Therefore, an error occurs in all read / write commands from the host computer 10 thereafter. Keep reporting.

It is also possible to suppress the read / write processing of only the head in which the abnormality has occurred.

The above-mentioned function is an example in which the recoverable error is judged to be abnormal during the write processing and the compare processing is executed. Similarly, the temperature of the temperature sensor 25 is checked, and if the temperature becomes abnormal, the above-mentioned function is executed. It is also possible to perform the read processing similar to the above and check that no writing abnormality has occurred.

The end time of the write command is stored in the command end time table 36, the time until the next write command reception time is checked, and if a certain time interval is open, the same read processing as above is executed. It is also possible to check that no writing error has occurred.

When an unrecoverable error occurs in the write processing from the host computer 10, or in the same way, an unrecoverable error occurs in the read processing from the host computer 10 and the read check after writing, and an error report is sent to the host computer 10. If the error is reported, the time at which the error is reported is stored in the error end time table 37.

Next, if the write command from the host computer 10 has not elapsed from the error end time table 37 for a certain period of time, a read process is always executed to check that no write abnormality has occurred. Further, the number of write commands received after the error report is carried out is counted by the write command reception number table 38, and it is checked whether or not a write error has occurred for a certain fixed number of commands.

The above-mentioned processing for checking that the writing abnormality has not occurred can be applied only to the head in which the abnormality has occurred in order to suppress the performance deterioration as much as possible.

Further, the read processing for checking whether or not the writing abnormality has occurred does not cause the performance deterioration, so it is possible to limit it to only when the next command from the host computer 10 is not received.

That is, at the end of the write process, it is checked whether or not the next command is received. If the next command is received, the write process end report is executed and the received command is executed. If the next command has not been received, perform the read process to confirm that no write error has occurred, then perform the end report and the end report before executing the read process. There is a way.

Before the read processing for confirming that no writing error has occurred, if a data mismatch is found in the compare when a completion report is issued, all read / write operations from the host computer 10 are performed. The write command keeps reporting errors.

If an unrecoverable error occurs in the read process, there is a possibility that incorrect data will be read later if left unattended. Therefore, it is necessary to rewrite the sector so that it cannot be read again. is there.

Since there is no receive command, even if the read process is executed, the host computer 10 is still performing the read process.
When the next command is received from, the read process is immediately stopped and the received command process is executed.

Further, in the operation of checking whether or not there is a writing error in the series of written data up to the above, the operation mode switching error is caused by the mode setting from the host computer 10. That is, the following operation modes are provided. 1) Whether to perform a read check when a recoverable error occurs during write processing. 2) Setting the number of recoverable errors for read check 3) Whether or not to continue to report the error in all read / write commands from the host computer when a write error (incorrect data) is detected. 4) Whether to perform the read data check when the temperature is out of a certain range. 5) Whether or not to execute the read data check in the command after a certain time interval has elapsed. 6) Whether or not the read data check is unconditionally performed for a certain period of time or a certain number of commands after the error report is performed. Also, the time / number of times. 7) Whether to perform the read check only when the next command is not received. 8) Whether to execute the command completion report before performing the read check. 9) Whether to stop the read check when a command is issued from the host computer during the read check.

[0036]

EFFECTS OF THE INVENTION Even if a read command confirming whether writing is normally performed is not received from the host computer, an abnormality such as a recoverable error has occurred according to the judgment of the magnetic disk device itself. As a result, the read operation of the written data is performed, it is confirmed that the written data and the read data match, and it is confirmed that the write error does not occur, thereby guaranteeing the data of the magnetic disk device. It became possible to do it.

[Brief description of drawings]

FIG. 1 is a structural explanatory view of the present invention.

[Explanation of symbols]

10 host computer, 20 control device,
21 interface controller, 22 hard disk controller, 23 buffer controller, 24 buffer (Cache), 25 temperature sensor, 26 control processor, 27 write data segment, 28 read (for compare) segment, 30 internal memory, 31 host request logical address table, 32 host request length table, 33 physical address table, 34 error occurrence address table, 35 abnormal head map table, 36 previous command end time table, 37 error end time table, 38 command reception count table, 50 host interface, 60 disk interface, 70 magnetic disk,

Front page continuation (51) Int.Cl. ⁷ Identification code FI theme code (reference) G11B 20/18 G11B 20/18 572F G06F 3/06 305 G06F 3/06 305D G11B 19/04 501 G11B 19/04 501Q ( 72) Inventor Takuji Ogawa, 322-2 Nakazato, Odawara-shi, Kanagawa, Ltd., RAID Systems Division, Hitachi, Ltd. (72) Masao Matsushita 2880, Kozu, Odawara, Kanagawa, Ltd., Storage Division, Hitachi, Ltd. (72) Inventor Yuji Hata 2880 Kozu, Odawara-shi, Kanagawa F-Term (Reference), Storage Division, Hitachi, Ltd. 5B065 BA01 CA12 EA04 EA25 EA36 EA39 EK03 EK05

Claims (15)

[Claims] 1. The magnetic disk device itself has an abnormality even if it does not receive a read command having a function of confirming whether or not information writing is normally executed from a host computer. Magnetic disk device, which has a function of detecting that a write error occurs, performing a read operation of the written data, confirming that the written data and the read data match, and checking that there is no write abnormality. . 2. The magnetic disk device according to claim 1, wherein the magnetic disk device has a function of notifying a host computer of a target address of the read operation when the read operation detects an unrecoverable error. Characteristic magnetic disk device. 3. The magnetic disk device according to claim 1, wherein when the magnetic disk device detects a write error, it has a function of reporting to the host computer that a write error has occurred using a dedicated error code. A magnetic disk device characterized by the above. 4. The magnetic disk device according to claim 1, wherein when the magnetic disk device detects a mismatch (wrong data) between written data and read data, all read / write commands from the host computer. 2. A magnetic disk device characterized by having a function of continuously reporting an error. 5. The magnetic disk device according to claim 1 or 2, wherein when the magnetic disk device completes writing by an error retry operation, an error retry operation occurs and an abnormality occurs. A magnetic disk device characterized by having a function of determining read data and performing a read data check. 6. The magnetic disk device according to claim 1, wherein when the temperature inside the magnetic disk device is outside a certain range, the magnetic disk device is determined to be in an abnormal state, A magnetic disk device having a function of performing read data check. 7. The magnetic disk device according to claim 1 or 2, wherein the magnetic disk device has a function of performing read data check triggered by a command after a lapse of a predetermined time. Magnetic disk device characterized by. 8. The magnetic disk device according to claim 1, wherein the magnetic disk device unconditionally checks the read data for a certain period of time or a certain number of commands after performing an error report. A magnetic disk device characterized by having a function of carrying out. 9. The read data check according to claim 1, wherein even if the magnetic disk device has a plurality of heads, only the head in which an abnormality has occurred is checked. A magnetic disk device having a function to perform. 10. The magnetic disk device according to claim 1 or 2, wherein the magnetic disk device has a function of performing a read check only when the next command is not received. And a magnetic disk device. 11. The magnetic disk device according to claim 1 or 2, wherein the magnetic disk device has a function of executing a command end report before performing a read check. Disk device. 12. The magnetic disk device according to claim 11, wherein when an unrecoverable error is detected in the read after the completion report is made, the data at the address where the error is detected cannot be read again. A magnetic disk device having a function to perform. 13. The magnetic disk device according to claim 4, wherein the magnetic disk device has a function of disabling data access using a head in which a write error has occurred. 14. The magnetic disk device according to claim 10 or 11, wherein the read check is interrupted when the magnetic disk device receives a command from the host computer during the read check. A magnetic disk device comprising: 15. The magnetic disk device according to claim 1, 2, 10 or 14, wherein various functions of the magnetic disk device can be set from a host computer. Magnetic disk device.