JP2002244816A - Disk device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To transfer data on a disk efficiently to a cache memory and improve the performance of a disk device by allotting the waiting time for disk rotation without loss to a look-ahead and increasing the quantity of the look-ahead. SOLUTION: When it becomes necessary to transfer read-requesting data to the cache memory 104, the data stored in the address before the read- requesting data are read out from a disk through a head and transferred to the cache memory 104 during the waiting time for disk rotation after completion of the seeking in a track where the read-requesting data are stored till when the head actually accesses the read-requesting data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for efficiently transferring data on a disk to a cache memory in a disk device and improving the performance of the disk device.

[0002]

2. Description of the Related Art Generally, in a disk device such as a hard disk device, a cache memory is provided, data on the disk is transferred to the cache memory and managed, and data can be directly transferred from the cache memory to the host. Improves the efficiency of data transfer to the host.

Specifically, read request data requested by a read command received from a host is read from a disk and transferred to a cache memory. At this time, the read request data is hit so as to be hit by a subsequent read command. There is a method in which data (following data) stored immediately after data is transferred to a cache memory in advance.

[0004] Further, as in a "magnetic disk device" described in JP-A-6-95810, a next read command is received during execution of pre-reading of subsequent data, and read request data requested by the read command is received. If the data is not stored in the cache memory, the read-ahead during execution is not immediately interrupted, and the disk rotation waiting time (after the seek to the track storing the read request data is completed, the head actually returns to the read request data) There is a method in which the time required for access is allocated to the extension of the prefetching being executed.

[0005]

In order to increase the efficiency of data transfer to the host using the cache memory, it is important to pre-read as much data as possible and transfer it to the cache memory. To increase the amount of prefetch, it is effective to allocate the disk rotation waiting time to the execution of prefetch and increase the amount of prefetch.

In order to increase the cache memory hit rate, it is important not only to increase the prefetch amount but also to prefetch data that is likely to be requested by a subsequent read command.

In the "magnetic disk device" described in Japanese Patent Application Laid-Open No. 6-95810, although the disk rotation waiting time is allocated to the execution of prefetching and the amount of prefetching is increased, the next read request data is executed during execution of prefetching. Is transferred to the cache memory. That is, for example, after writing the write request data requested by the write command to the disk, pre-reading is not executed, so that a read command is received next to the write command, and the read request data requested by the read command is cached. If the read request data is not stored in the memory, there is a problem that the disk rotation waiting time when transferring the read request data from the disk to the cache memory is not used for prefetching. Further, there is a problem that only data subsequent to the read request data can be read ahead.

An object of the present invention is to allocate the disk rotation waiting time to prefetch without waste, and to increase the prefetch amount.

Another object of the present invention is to make it possible to prefetch data that is likely to be requested by a subsequent read command when executing prefetch.

[0010]

In order to achieve the above object, the disk device of the present invention prefetches data stored at an address before data to be transferred to a cache memory.

That is, the present invention provides, as a first aspect,
A disk device for reading read request data from a disk through a head and transferring the read request data to the cache memory when the read request data is not stored in the cache memory. After the end, during the disk rotation waiting time required for the head to actually access the read request data, data stored at an address before the read request data is read from the disk through the head and transferred to the cache memory. Preceding data pre-reading means, and pre-fetching which, when it becomes necessary to transfer the read request data to the cache memory, starts a seek to a track storing the read request data and causes the preceding data pre-reading means to execute pre-reading Control means. It provides a disk apparatus.

[0012] The preceding data pre-read means sets the address of the sector in which the head is located after the end of the seek operation.
If the read request data is smaller than the address of the sector in which the read request data is stored, the data is sequentially read from the address of the sector where the head is located after the seek and transferred to the cache memory, and the sector where the head is located after the seek is completed. Is larger than the address of the sector in which the read request data is stored, the data can be sequentially read from the head address of the track after seek and transferred to the cache memory.

[0013] In addition, after the read request data is transferred to the cache memory, there is further provided a subsequent data prefetch means for reading data stored immediately after the read request data from the disk through the head and transferring the read data to the cache memory. The read-ahead control means, when the subsequent read request data needs to be transferred to the cache memory while the subsequent read-ahead means is executing the read-ahead, the read-ahead control means reads the next read request data to the track storing the next read request data. Predict the address of the sector where the head will be located when the seek is performed. If the predicted sector address is larger than the address of the sector where the next read request data is stored, the track end is calculated from the predicted sector address. The pre-reading that is being performed for the disk rotation wait time up to After continued prefetch unit can initiate a seek to track said next read request data is stored so as to perform the look-ahead to the preceding data read-ahead means.

Further, in order to achieve the other object, the disk device of the present invention reflects locality of read request data on a disk when executing prefetching.

That is, the disk device of the present invention
According to the first aspect, in the first aspect, a read request frequency measuring means for measuring a read request frequency for each address range of the disk, and after transferring the read request data to the cache memory, the read request frequency is stored immediately after the read request data. Read-out data from a disk through a head and transferring the read-out data to a cache memory, wherein the pre-reading control means stores the next read request data in the cache memory while the subsequent data pre-reading means is executing pre-reading. When it is necessary to transfer the data, the read request frequency of the address range to which the start logical block address of the next read request data belongs, and the logical block address corresponding to the head position where the subsequent data prefetch means is executing prefetch Compare the read request frequency of the address range to which the If the read request frequency is high, seek to the track in which the next read request data is stored is started and the preceding data prefetch means executes prefetch, and the latter read request frequency is high or the same. Then, the amount of data that will be read when the preceding data prefetch means is caused to execute prefetching is predicted. After the following read request data is stored, the seek to the track in which the next read request data is stored is started, and the preceding data preread means executes prefetch. are doing.

[0016]

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

FIG. 1 is a diagram showing a configuration example of a magnetic disk drive according to an embodiment of the present invention.

As shown in FIG. 1, the magnetic disk device according to the present embodiment includes a program ROM 101 on which a control program for controlling operations of the magnetic disk device including a prefetch operation is mounted, and a program ROM 101
A control processor 102 that reads and executes the above control program, and a magnetic disk controller 103 that controls data transfer between the host and the cache memory 104 and data transfer between the cache memory 104 and the disk.
A cache memory 104 for temporarily writing read request data / write request data, a servo control unit 107 for performing control for moving a head to a corresponding position when reading / writing data, and a servo control unit. 10
7, a voice coil motor (VCM) 108 for moving the head, a motor driver 109 for controlling the rotation of the disk, and a selector 1 for selecting only a specified head signal from magnetic signals read from the head.
10 and a signal processing unit 111 for converting analog data sent from the selector 110 to digital data and for converting digital data sent from the magnetic disk control device 103 to analog data, and a host. A bus 112 is provided.

The cache memory 104 manages information on read request data read from the disk, and includes a cache data management table 105 serving as an index when reading data from the cache memory 104; And a read request frequency management table 106 for extracting and managing the locality of the data as a read request frequency for each logical block address range.

Hereinafter, the operation of the magnetic disk drive according to this embodiment will be described.

First, the operation when a read command is received will be described.

When the magnetic disk device according to the present embodiment receives a read command from the host via the bus 112, it refers to the cache data management table 105,
A search is performed to determine whether or not the read request data requested by the received read command exists in the cache memory 104 (is hit).

FIG. 2 shows a configuration example of the read command.

As shown in FIG. 2, the read command is a command code 21 indicating that the read command is a read command, and a start logical block address (LBA:
(Logical Block Address) 22 and a request data transfer number 23 indicating the size of the read request data.

FIG. 3 shows the cache data management table 10.
5 shows a configuration example.

As shown in FIG. 3, the cache data management table 105 contains the start LBA 31 of the data written in the cache memory 104 and the cache memory 104
It consists of an upper start address 32 and a data size 33.

The magnetic disk device according to the present embodiment uses the start LBA 31 and the data size 33 in the cache data management table 105 when searching whether or not the read request data is hit. It is checked whether or not all of the request data exists in the cache memory 104, and if so, the read request data is read from the cache memory 104 with reference to the corresponding start address 32 in the cache memory and transferred to the host. . If not present, the LBA 22 in the read command is converted into a physical address indicating the arrangement on the disk, the position of the physical address is detected using the servo control unit 107, and the signal control unit 111 is used. Then, the read request data is transferred to the magnetic disk control device 103. The read request data is written to the cache memory 104 by the magnetic disk control device 103 and, at the same time, transferred to the host via the bus 112.
Further, when the read request data is written to the cache memory 104, information on the read request data is registered in the cache data management table 105 by the magnetic disk control device 103.

Next, the operation of extracting the locality of the read request data will be described.

When the magnetic disk drive according to the present embodiment receives a read command from the host via the bus 112, it registers it in the read request frequency management table 106.

FIG. 4 shows the read request frequency management table 106.
An example of the configuration will be described.

As shown in FIG. 4, the read request frequency management table 106 includes a predetermined LBA range 41 and each LB range.
A read request frequency 42 indicating how many times the host has requested data to which the start LBA belongs to the A range 41.

FIG. 5 shows a read request frequency management table 10.
6 is a flowchart showing the flow of a registration operation to No. 6.

As shown in FIG. 5, when a magnetic disk device according to the present embodiment receives a command from a host, the magnetic disk device refers to a command code in the command and determines whether or not the command is a read command (step 51). ). If it is a read command, the read request frequency management table 106
, The LBA range 41 to which the start LBA 22 of the read request data in the read command belongs is detected, and the read request frequency 42 corresponding to the detected LBA range 41 is increased by 1 (step 52).

The read request frequency management table 106 in which the registration operation is performed as described above enables, in a prefetch operation described later, data that is likely to be read request data requested by a subsequent read command to be prefetched. It is used for reference.

Next, the prefetch operation will be described.

The inventors of the present invention have adopted a prefetch operation method as follows.
When it becomes necessary to transfer the read request data to the cache memory, a method of prefetching data stored at an address before the read request data has been devised. Hereinafter, such a prefetching method is referred to as “method 1”.

In the magnetic disk drive according to the present embodiment, the "method 1" is employed, and after the seek to the track storing the read request data is completed, the head is actually accessed until the head accesses the read request data. The data (preceding data) stored at the address before the read request data is read from the disk and transferred to the cache memory 104 during the disk rotation waiting time according to.

At this time, in the "method 1", the pre-reading method is controlled as shown in FIGS. 6 and 7 based on the positional relationship between the head position after the end of the seek and the storage position of the read request data. The position on the track is indicated by a sector address.

FIG. 6 shows the head position 61 after the end of the seek.
Shows an example of pre-read when the read request data is smaller than the start sector address 62. In this case, pre-read of the preceding data is started immediately after the end of the seek.

FIG. 7 shows the head position 71 after the seek is completed.
Shows an example of the pre-read when the read request data is larger than the start sector address 72 of the read request data. In this case, the pre-read of the preceding data is started from the head 73 of the target track.

As described above, regardless of whether a read command is received after a read command or a read command is received after a write command, "method 1" is applied when a read command is received. Then, the prefetching of the preceding data is always performed.

As a result, it is possible to allocate the disk rotation waiting time to the prefetch without waste, and to increase the prefetch amount.

Here, the head position 61 after the end of the seek
Can be predicted before actually starting the seek. Therefore, a conventional read-ahead method (hereinafter, referred to as “method 2”) in which the read request data is transferred to the cache memory, and subsequent data of the read request data is read and transferred to the cache memory 104. ,
A method that uses "method 1" in combination is also conceivable. Hereinafter, the prefetching method using both "method 1" and "method 2" will be referred to as "method 3" and its operation will be described.

In the "method 1", as shown in FIG. 7, when the head position 71 after the seek is larger than the start sector address 72 of the read request data, the pre-reading is started from the head 73 of the target track. The disk rotation waiting time 74 required from the head position 71 after the seek to the head at the track end 74
Occurs. Therefore, this disk rotation waiting time 74 is
It is conceivable that this is applied to the extended read-ahead of the subsequent data according to the “method 2”.

In other words, in the "method 3", after the read request data is transferred to the cache memory 104, the subsequent data of the read request data is pre-read, but the read command is received during the pre-read of the subsequent data and the next read request is received. When it is necessary to transfer data to the cache memory 104, the head is located when a seek to a track storing the next read request data is performed without immediately stopping prefetching of subsequent data. If the predicted sector address is larger than the address of the sector where the next read request data is stored, the disk rotation from the predicted sector address to the track end is waited (the disk shown in FIG. 7). Rotation wait 74) Continue reading ahead of the subsequent data being executed for the time,
It is possible to pre-read the preceding data by "method 1".

FIG. 8 shows an example of pre-reading according to the "method 3", in which the predicted head position 81 after the seek when the seek is started immediately after the reception of the read command is the starting sector address 82 of the read request data. Beyond,
An example in which the target track 83 is separated from the track being pre-read by 2 or more is shown. In the example of FIG. 8, since the pre-reading according to the “method 1” is performed from the head 86 of the target track, it is assumed that the seek has started immediately after the reception of the read command. Wait for disk rotation until end 84 (Disk rotation wait 7 in FIG. 7)
Equivalent to 4. ) Can be applied to the extension of the prefetching of the subsequent data, and the prefetching of the subsequent data can be continued. Then, from the head 86 of the target track, “method 1”
Start prefetching by.

Since the subsequent data to be pre-read is data subsequent to the read request data, according to "method 3", the command received next to the read command is also a read command, and "method 1" is applied. If the disk rotation wait 84 occurs when it is assumed that the prefetching of the subsequent data is continued, the command received next to the read command is also the read command.
If the disk rotation wait 84 does not occur when it is assumed that is applied, the pre-reading of the subsequent data being executed is immediately stopped, and the pre-reading of the preceding data by "method 1" is performed.

As a result, also in the case of performing the pre-reading by the "method 3", it is possible to allocate the disk rotation waiting time to the pre-read without waste and increase the pre-read amount.

Further, in the magnetic disk drive according to the present embodiment, by limiting the amount of pre-read data of the preceding data according to “method 1”, “rotation 1” and “method 2” can be used when the rotation waiting is applied to the pre-read. Can be controlled in more detail. Hereinafter, such an example will be described as “method 3 ′” with reference to the flowchart in FIG.

When a read command is received during the pre-reading of the subsequent data according to the “method 2” and it becomes necessary to transfer the next read request data to the cache memory 104, the magnetic disk device according to the present embodiment first As shown in FIG. 9, when it is assumed that the pre-reading of the preceding data by "method 1" is performed, the predicted sector address of the head after the seek is obtained, and the amount of the pre-read data is predicted (step 901). It is checked whether or not the predicted preceding data prefetch data amount exceeds a predetermined threshold (step 902).

If the predicted preceding data prefetch data amount is larger than the predetermined threshold value (step 902), the prefetching of the subsequent data by the "method 2" is continued by the excess amount (step 903). Is started (step 904). Then, immediately after the seek is completed, pre-reading of the preceding data is started (step 905).

On the other hand, the pre-read data amount predicted in step 901 is equal to or smaller than a predetermined threshold value (step 902), and the predicted sector address obtained in step 901 is the address of the sector in which the read request data is stored. If it is smaller (step 906), the seek starts immediately (step 907), and after the seek ends, the prefetching of the preceding data is started (step 908).

If the predicted sector address obtained in step 901 is larger than the address of the sector in which the read request data is stored (step 906), the disk waits from the predicted sector address to the track end (FIG. 7). Pre-reading of subsequent data by "method 2" is continued for the disk rotation wait 74) time (step 9).
09), and then start seek (step 91)
0). Then, after the seek is completed, pre-reading of the preceding data is started from the head of the track (step 911).

As described above, in the “method 3 ′”, after the read request data is transferred to the cache memory 104, the subsequent data of the read request data is prefetched. When it is necessary to transfer the read request data to the cache memory 104, the read-ahead of the subsequent data by the "method 2" is continued without immediately stopping the read-ahead of the subsequent data, as in the "method 3". After that, the preceding data is pre-read by "method 1".

As a result, also in the case of performing the pre-reading by the "method 3 '", similarly to the "method 3", the disk rotation waiting time can be allotted to the pre-read without waste, and the pre-read amount can be increased.

In particular, in the "method 3 '", when the disk rotation waiting time is used for prefetching, the "method 2"
, The ratio between the amount of pre-read data of subsequent data according to the above and the amount of pre-read data of preceding data according to the “method 1” can be set freely. The setting of this ratio is, specifically,
This is achieved by changing the threshold value compared in step 902 to an arbitrary value.

Now, in the magnetic disk drive according to the present embodiment, “method 1”, “method 2”, “method 3”,
By adopting "method 3 '", the amount of pre-reading can be increased and the effective transfer speed of the disk can be increased.
With reference to the read request frequency management table 106, a prefetch method may be controlled so that data that is likely to be requested by a subsequent read command can be prefetched.

In the following, the read request frequency management table 1
An example of the operation of controlling the prefetching method will be described with reference to a flowchart of FIG.

When the read command is received while the subsequent data is being pre-read, the next read request data is transmitted to the cache memory 10.
In the case where it is necessary to transfer the data to No. 4, the seek distance to the track storing the read request data to be transferred is 0 in the magnetic disk device according to the present embodiment, as shown in FIG. (Step 100)
1) If the seek distance is 0, the prefetching of the subsequent data is continued to the track end, and the prefetching of the preceding data is started from the head of the track (step 1002).

If the seek distance is 1 or more (step 1001), the read request frequency management table 1
06, the read request frequency 42 of the LBA range 41 to which the LBA corresponding to the head position for which the subsequent data is being pre-read belongs, and the read request frequency 42 of the LBA range 41 to which the start LBA of the next read request data belongs. A comparison is made (step 1003).

As a result of the comparison in step 1003, if the read request frequency 42 of the LBA range 41 to which the start LBA of the read request data belongs is higher (step 1003).
4), pre-reading of the preceding data by "method 1" is started. That is, first, seek is started (step 100).
5) If the sector address of the head after the end of the seek is larger than the start sector address of the read request data (step 1006), pre-reading is started from the head of the target track (step 1007), and the sector of the head after the end of the seek If the address is smaller than the start sector address of the read request data (step 1006),
Immediately starts prefetching of the preceding data (step 100)
8).

On the other hand, as a result of the comparison in step 1003, if the read request frequency 42 of the LBA range 41 to which the start LBA of the next read request data belongs is lower or the same (step 1004), the “method 1” is used. The amount of pre-read data when the pre-read of the preceding data is assumed to be performed is predicted (step 1009).
If it is equal to or less than a predetermined threshold value (step 101
0), pre-reading of the preceding data by "method 1" is started (steps 1005 to 1008).

If the pre-read data amount of the preceding data predicted in step 1009 is larger than a predetermined threshold value (step 1010), the pre-reading of the succeeding data being executed is continued by the excess (step 1011). Then, seek is started (step 1012). Then, immediately after the seek, the pre-reading of the preceding data is started (step 1013).

Thus, when performing prefetching, it is possible to prefetch data that is likely to be read request data requested by a subsequent read command.

Since the subsequent data to be pre-read is subsequent data of the read request data, the pre-read operation shown in FIG. 10 is an operation when the command received after the read command is also a read command. If the command received after the write command is a read command, pre-reading of the preceding data by "method 1" is performed.

Further, in the magnetic disk drive according to the present embodiment, the prefetching method is controlled in order to enable prefetching of data which is likely to be read request data requested by a subsequent read command. At this time, the read request frequency management table 106 is referred to.
6 extracts the locality of the read request data, and can be used for other purposes.

[0067]

As described above, according to the disk apparatus of the present invention, the disk rotation waiting time is allotted to the prefetch without waste, and the prefetch amount is increased, so that the data on the disk can be efficiently transferred to the cache memory. Thus, the performance of the disk device can be improved.

According to the disk device of the present invention, when performing prefetching, it is possible to prefetch data that is likely to be read request data requested by a subsequent read command, thereby enabling cache hits. Rate, and the performance of the disk device can be improved.

[Brief description of the drawings]

FIG. 1 is an exemplary view showing a configuration example of a magnetic disk device according to an embodiment.

FIG. 2 is a diagram showing a configuration example of a read command issued by a host.

FIG. 3 is a diagram showing a configuration example of a cache data management table.

FIG. 4 is a diagram showing a configuration example of a read request frequency management table.

FIG. 5 is a flowchart showing a flow of a registration operation to a read request frequency management table.

FIG. 6 is an explanatory diagram showing an example of a method of pre-reading data stored at an address before read request data.

FIG. 7 is an explanatory diagram showing an example of a method of prefetching data stored at an address before read request data.

FIG. 8 is an explanatory diagram showing an example of a method of prefetching data stored at an address before read request data and a method of prefetching subsequent data.

FIG. 9 is a flowchart showing another example of the operation flow of a method for prefetching data stored at an address before read request data and a method for prefetching subsequent data.

FIG. 10 is a flowchart showing the flow of a prefetch operation performed using the read request frequency management table.

[Explanation of symbols]

 101: Program ROM 102: Control Processor 103: Magnetic Disk Controller 104: Cache Memory 105: Cache Data Management Table 106: Read Request Frequency Management Table 107: Servo Controller 108: Voice Coil Motor (VCM) 109: Motor Driver 110: Selector 111: Signal processing unit 112: Bus 21: Command code 22: Logical block address (LBA) 23: Requested data transfer number 31: Start LBA 32: Start address on cache memory 33: Data size 41: LBA range 42: Read request frequency

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Shinya Iguchi 1099 Ozenji Temple, Aso-ku, Kawasaki City, Kanagawa Prefecture Inside System Development Laboratory, Hitachi, Ltd. F-term in the Storage Systems Division of the Works (Reference) 5B005 JJ13 KK12 MM11 NN22 5B065 BA01 CA11 CC07 CC08 CE02 CE12 CH02 CH05 CS01 5D044 AB01 BC01 CC05 DE12 DE38 DE91 FG10 FG18 HH05 HL02

Claims (6)

[Claims] A disk, comprising a disk, a head, and a cache memory, wherein, when read request data is not stored in the cache memory, the read request data is read from the disk through the head and transferred to the cache memory. An apparatus for storing, at a disk rotation waiting time required for a head to access the read request data after completing a seek to a track storing the read request data, an address earlier than the read request data; Means for reading the data being read from the disk through the head and transferring the read request data to the cache memory; and, when it becomes necessary to transfer the read request data to the cache memory, a track to the track storing the read request data. Start seek and proceed to the preceding data prefetching means Disk apparatus characterized by comprising a prefetch control means for executing only. 2. The disk drive according to claim 1, wherein the preceding data pre-reading means is configured such that, after the end of the seek, the address of the sector where the head is located is smaller than the address of the sector where the read request data is stored. When the seek is completed, the data is sequentially read from the address of the sector where the head is located and transferred to the cache memory. After the seek is completed, the address of the sector where the head is located is the sector where the read request data is stored. A disk device that sequentially reads data from a head address of a track after seeking and transfers the data to a cache memory when the address is larger than the address of the seek track. 3. The disk device according to claim 1, wherein, after transferring the read request data to the cache memory, data stored immediately after the read request data is read from the disk through the head and cached. The read-ahead control unit further transfers the next read request data to the cache memory while the subsequent data read-ahead unit is executing the read-ahead. Predict the address of the sector where the head will be located when seeking to the track where the read request data is stored, and set the predicted sector address to the address of the sector where the next read request data is stored. If greater, only the disk rotation waiting time from the predicted sector address to the track end After continuing the prefetching being executed by the subsequent data prefetching means, starting a seek to a track in which the next read request data is stored, and causing the preceding data prefetching means to execute prefetching. Disk device to be used. 4. The disk device according to claim 1, wherein, after transferring the read request data to the cache memory, data stored immediately after the read request data is read from the disk through the head and cached. The read-ahead control means further transfers the next read request data to the cache memory while the subsequent data read-ahead means is executing the read-ahead. Predict the amount of data that will be read when the prefetching unit executes the prefetching, and allocate any data amount of the predicted data amount as the amount of prefetching by the subsequent data prefetching unit. After the prefetch being executed is continued by the subsequent data prefetch means, the next read request data is read. A disk drive for starting a seek to a track in which the data is stored and causing the preceding data prefetching means to execute prefetching. 5. The disk drive according to claim 1, wherein said read request frequency measuring means measures a read request frequency for each address range of the disk, and said read request data is transferred to a cache memory. Data reading means for reading data stored immediately after the read request data from the disk through the head and transferring the read data to the cache memory, wherein the read-ahead control means executes the read-ahead, When it becomes necessary to transfer the next read request data to the cache memory, the read request frequency of the address range to which the start logical block address of the next read request data belongs, and the subsequent data prefetch means is executing prefetch. Read the address range to which the logical block address corresponding to the head position belongs If the former read request frequency is higher, the seek to the track storing the next read request data is started and the preceding data prefetch means executes prefetch, and the latter read request is performed. If the request frequency is high or the same, predict the amount of data that will be read when the preceding data prefetch means executes prefetching, and if the predicted data amount is greater than a predetermined threshold, After the prefetch being executed is continued by the succeeding data prefetch means for the excess, the seek to the track storing the next read request data is started, and the prefetch data is executed by the preceding data prefetch means. A disk device characterized by the above-mentioned. 6. A disk device equipped with a disk and a head, having a counter for measuring the number of read requests for each address range of the disk, and an address to which a start logical block address of the read request data belongs. A disk device comprising: a read request frequency measuring means for obtaining a range and adding a counter corresponding to the address range.