JPS60243755A - Disc cache control system - Google Patents

Abstract

PURPOSE:To remove time loss due to the overlap of seeking operation by writing the address of a defective track and that of a substitutive track in a prescribed table at the detection of the defective track. CONSTITUTION:When the necessity of access to a magnetic disc device 4 is generated on the basis of an instruction from a host computer 1, a search control part 7 checks whether the data of the corresponding track exist in a cache 6 or not and whether the corresponding track is defective or not. When the corresponding track is decided as a defective one, the address of the defective track and that of a substitutive track corresponding to the defective track are written in the prescribed table of the search control part 7. At the access to the succeeding defective track or after, the address of the substitutive track can be obtained without accessing the track concerned in the magnetic disc device 4.

Description

Detailed Description of the Invention (1) Technical Field of the Invention The present invention relates to a magnetic disk control device for controlling a magnetic disk device used as external storage for an information processing system, and relates to a magnetic disk control device for controlling a magnetic disk device used as an external storage of an information processing system. This relates to the control of access to.

Q) Prior art and problems Due to the nature of magnetic disk media, the presence of a few defective tracks during manufacturing is allowed, and defective tracks are marked (flag) during product inspection to indicate that they are defective tracks. At the same time, the address of the track to be replaced is written and shipped.

Furthermore, when a magnetic disk medium is used in an information processing system, formatting, generally referred to as initialization, is performed.

At this time as well, all tracks are inspected, and if a track has a defective location, it is treated as a defective track and a replacement track is allocated to it.

When a track marked as a defective track is encountered during actual access, control is performed to read replacement track information for the defective track and retry the seek operation to access the replacement track.

Therefore, in a magnetic disk drive, when a defective track is encountered during access, seek operations are performed twice, resulting in a large time loss during access.

On the other hand, in recent years, a buffer called a disk cache has been provided between a magnetic disk control device and a magnetic disk device, and each time a magnetic disk medium is accessed, the data of the corresponding track is also recorded and accumulated on the buffer. When an access request is received, the buffer is checked, and if relevant data is found, it is immediately read out and transferred, thereby speeding up the access.

If you have such a disk cache, when you encounter a bad track and access the replacement track again,
Since these are recorded on the disk cache, when accessing the same track from the second time onwards, the time loss due to duplication of seek operations is reduced to a certain extent.

However, the disk cache has a storage capacity that is considerably smaller than the storage capacity of the magnetic disk medium (from a few tens of minutes to a few hundred times), and when there is no space left to write the data read from the magnetic disk medium. Normally, control is performed such that the oldest data is sequentially erased and new data is stored in that area.

Therefore, in this case as well, the information on the buffer regarding the defective track and the replacement track will eventually be purged, and at that time, the seek operation will encounter the defective track, retry the seek, and then obtain data from the replacement track. There is a problem in that it is not possible to completely avoid the occurrence of time loss due to duplication.

0) Purpose of the Invention In view of the above-mentioned conventional problems, the present invention provides information on bad tracks and replacement tracks in the disk cache mechanism, and performs control to prevent these from being expelled. It is an object of the present invention to provide a control method that can realize access with good response by avoiding duplication of seek operations when a bad track is encountered during access.

(4) Structure and object of the invention According to the present invention, the system is located between a magnetic disk device and a host computer, and stores data read from the magnetic disk device each time. When the host computer instructs the host computer to read the data, if data of the corresponding track exists in the memory, the data read from the memory is transferred to the host computer without accessing the magnetic disk device; In a disk cache mechanism that has control to transfer the data read from the magnetic disk device to the host computer and store it in the memory when the data of the corresponding track does not exist in the memory, the address of the defective track and the address of the defective track are stored. A table is provided to record the address of the corresponding alternate track, and it is checked whether the track whose address should be accessed when accessing the magnetic disk device is registered as a defective track in the table. Find the address of the corresponding replacement track from the table.1. If the magnetic disk is not registered and a bad track is actually accessed and a bad track is encountered at that time, the address of the bad track and the address of a replacement track are recorded in the table. It will be achieved.

(5) Embodiment of the invention FIG. 1 is a block diagram for explaining the invention,
1 is a host computer, 2 is a magnetic disk controller,
3 is a disk cache mechanism, 4 is a magnetic disk device,
Reference numeral 5 represents a data transfer control section, 6 a cache memory, and 7 a search control section.

In FIG. 1, when it becomes necessary to access the magnetic disk drive 4 based on an instruction from the host computer 1, the search control unit 7 determines whether or not the data of the corresponding track exists in the cache 6, and determines whether the corresponding track is a defective track. By checking whether it is registered or not and performing processing according to each result, data is transferred efficiently by avoiding duplication of seeks to the magnetic disk device 4.

The control of these search control units 7 will be explained in more detail below.

FIG. 2 is a block diagram showing a search control section according to an embodiment of the present invention, in which 8 is a track address register, 9 is a track search section, 10 is a defective track search section, 11 is a track table, and 12 is a defective track/track/control section. Replacement truck/
A link table, 13 is an alternate track address register, and 14 and 15 are flags.

Further, N indicates the address of the defective track, and C indicates the address of the replacement track. The track table 11 stores data read from the magnetic disk device in a cache (
buffer), the address of that track is written, and by referring to this,
It can be determined whether the data of the track at the corresponding address exists in the buffer. The defective track/replacement track link table 12 records the address and the address of the corresponding replacement track when the accessed track is a defective track. Further, the flag 14 is set to "1" when the data of the track at the corresponding address exists on the buffer, and the flag 15 is set to '1#' when the track at the corresponding address is registered as a defective track.

In FIG. 2, when the magnetic disk drive is accessed, when the address of the track to be accessed is set in the track/quart address register 8, it is sent to the track search unit 9 and the defective track/quart search unit 10. The track search unit 9 refers to the track table 11 to check whether data of the track at the address exists on the buffer, and if so, sets the flag 14 to "1".

On the other hand, the defective track search unit checks the defective track/replacement track/link tear pull 12, and if the track at the corresponding address is registered as a defective track, sets the flag 15 to 1'' and reads the corresponding replacement track address. Set in the output alternate address register 13.

The data transfer control section of the disk cache mechanism 3 controls data transfer based on the output information of these search control sections.

FIG. 3 is a flowchart showing the control of the data transfer section. As shown in the figure, the data transfer unit transfers the data of the corresponding track when it is on the sofa (when flag 14AU is "1") and when it is registered as a defective truck (when flag 15 is "1"). 1#), look at the replacement track address register 13t and transfer the data of the replacement track with the corresponding address on the sofa, and when it is not registered as a defective track (7lag 15 is "0"). ''), the data of the corresponding track address is read from the 5 buffer and transferred.

On the other hand, when the data of the corresponding track is not on the buffer (when the flag 14 is °'0'') and it is registered as a defective track, the address of the replacement track set in the replacement track address register 13 is Read the data, access the magnetic disk drive, store the data on the alternate track in a buffer, and transfer it to the host side.If one loaf is not registered as a defective track, access the corresponding track on the magnetic disk drive. The read data is stored in a buffer and transferred to the host side.If the corresponding track is a defective track at this time, seek again to the replacement track and read the data from the replacement track. At this time, the address of the defective track and the address of the replacement track are transferred to the defective track.

/Register in the replacement track link table.

When the buffer area becomes empty and old data is removed, the corresponding address is also removed from the track table 11, but the information in the defective track/replacement trunk/link table is accumulated without being removed. Therefore, after one cycle of access to the magnetic disk device, all defective tracks are registered in the defective track/replacement track link table, so the address of the replacement track can be determined without accessing the defective track of the magnetic disk device. You can get it.

(6) Effects of the Invention As explained in detail above, according to the control method of the present invention, when a defective track is encountered during access to a magnetic disk device, the address of the defective track and the corresponding replacement track can be determined. By writing the address in a predetermined table, the next time the defective track is accessed, the address of the replacement track to be accessed can be obtained without accessing the track in the magnetic disk device. Therefore, it is an effectiveness ratio method because access with good response can be achieved without causing time loss due to duplication of seek operations.

[Brief explanation of drawings]

FIG. 1 is a block diagram for explaining the present invention, FIG. 2 is a block diagram showing a search control section of an embodiment of the present invention, and FIG.
The figure is a flowchart showing control of the data transfer section. DESCRIPTION OF SYMBOLS 1... Host computer, 2... Magnetic disk control device, 3... Disk cache mechanism, 4... Magnetic disk device, 5...
...Data transfer control unit, 6...Cache memory, 7...Search control R8'''' Track address register, 9...Track search unit, 1
0...Bad track search unit, 11...Concave track table 1-pull, 12...Bad track/replacement track conflict link table, 13...Replacement track address register, 14 .15...Flag half I Figure

Claims (1)

[Claims] Located between the magnetic disk device and the host computer, the data read from the magnetic disk device is stored in a built-in memory each time, and when the host computer instructs the host computer to read data, the data of the corresponding track is stored in the memory. If the data exists, the data read from the memory is transferred to the host computer without accessing the magnetic disk device, and if the data of the corresponding track does not exist in the memory, the data read from the magnetic disk device is transferred to the host computer. In addition, in a disk cache mechanism having control such as storage in the memory, a table is provided for recording the address of a defective track and the address of a replacement track corresponding to the defective track, and when accessing the magnetic disk device, the address to be accessed is provided. Check whether the track is registered as a defective track in the table, and if it is registered, find the address of the corresponding replacement track from the table, and if it is not registered, actually access the magnetic disk, If a defective track is encountered at that time, the address of the defective track and the address of a replacement track are recorded in the table.