JP3435176B2 - Magnetic disk drive - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hard disk device for recording / reproducing information by magnetic means, a disk device such as a floppy disk device, a magneto-optical disk device, and the like. When writing / reading one data to a non-contiguous area at the time of writing, by storing the position when reading / writing the data divided into a plurality of times as the data continuation information, The present invention relates to a magnetic disk device that enables high-speed access.

[0002]

2. Description of the Related Art Conventionally, as a method for accessing a magnetic disk device at high speed, a cache function (disk
It is general to use a cache function) or a prefetch (prefetch: read from the next sector after the read is finished into the buffer RAM) function. However, since the disk access pattern differs depending on the type of file, it is not always possible to access all files at high speed regardless of which function is used.

For example, in the case of a sequential file, the records in the file are accessed sequentially, so if the prefetch function is used to read the requested block and several blocks ahead of it at the same time,
The hit rate increases. However, when performing random access, this prefetch function is not effective.

That is, the prefetch function works effectively only when the data to be processed has a large capacity and the data is recorded in a continuous area. Further, in the cache function, the capacity of the disk cache is as small as 1/100 to 1/1000 of the capacity of the magnetic disk device. Therefore, in order to newly load a data block,
It is necessary to flush the old block in the cache to the magnetic disk or overwrite the old block with a new block to rewrite the cache directory.

With this cache function, the method generally adopted as the method for determining the eviction block is LRU.
This is an algorithmic method. In the LRU algorithm, the least used block is selected as the eviction block, considering the number of times each block has been accessed within a certain time up to the present time and the elapsed time from the time of the last access to the present To do.

In short, in the LRU algorithm, since access to the same data is often concentrated in time, data that has been accessed many times in the past is likely to be accessed in the future. Assuming that blocks that are rarely accessed are used first. In addition,
In the case of the write-through cache, the new block is simply overwritten on the old block and the cache directory is rewritten. However, the selection method of which block to overwrite is Is the same as when deciding.

As described above, the cache function requires a large-capacity buffer RAM and complicated buffer management in order to fully exhibit the function. The prefetch function is effective only when the data has a large capacity and the data is continuously recorded. As described above, since the cache function and the prefetch function have functional restrictions, they cannot be used for all high speed accesses.

Therefore, in order to make the best use of the advantages of the cache function and the prefetch function, a method of using a combination of the LRU algorithm for determining the eviction block and the prefetch function is generally used. The above is the conventional method for accessing the magnetic disk device at high speed. Next, the prefetch function in the optical disk device will be described.

FIG. 9 is a diagram showing an example of a sequence for reading in the buffer RAM at the time of executing a read instruction, where (1) indicates normal read and (2) indicates dual read.

In FIG. 9, after the execution of the read command from the host device is completed, the next sector (the next sector after the sector designated by the read command is transferred) is stored in the buffer RAM.
The sequence to read is shown in. First, FIG.
In the normal read shown in (1), after the seek operation, the read (R1) from the disk to the buffer is performed, and then,
Transfer (B1) from the buffer to the host device.

The above operation is repeated in sequence, and after the transfer (B3) of the data of the designated number of sectors from the buffer to the host device is completed, the sector next to the last read sector is buffered. Read inside (R4) (prefetch read). In this way, in normal read,
When the transfer for the final buffer is completed, the prefetch function reads the next sector.

When the data of the last read sector is transferred to the upper device (B3), the status indicating the completion of the data transfer is automatically transmitted. The sending of this status is controlled by the task. Also,
The prefetch operation is suspended for the selection during the prefetch, but when there is valid read data, the data of that portion is registered as the prefetch data.

Such an operation is similar in the case of the dual read shown in FIG. 9 (2), but in the dual read,
After the seek operation, while reading from the disk to the buffer (R1, R1 '), the data in the buffer is transferred to the host device (B1, B1'), which is different from the normal read. Even in the dual read, after the transfer (B3, B3 ') from the buffer to the upper device is completed, the sector next to the last read sector is read into the buffer (R
4) Perform (prefetch read).

Then, the read to the buffer and the transfer to the host device by the normal read or dual read are
It is extremely effective when data is recorded in a continuous area. However, if the data is recorded in a dispersed manner, no effect can be expected.

As described above, the conventional method of accessing the magnetic disk device at high speed has an unsolved problem, and there is an inconvenience that it is not possible to access all types of files at high speed. . Incidentally, when reading / writing a large amount of data, particularly when reading / writing one data to a non-continuous area, the read / write command is divided into a plurality of times and executed.

[0016]

SUMMARY OF THE INVENTION The present invention solves such inconvenience caused by high-speed access of a conventional magnetic disk device, and when reading / writing one data to a non-contiguous area, the minimum necessary It is an object of the present invention to provide a magnetic disk device that enables high-speed access while managing data recorded in discontinuous areas as continuous data by managing the position information.

[0017]

According to the present invention, firstly,
In a magnetic disk device having a magnetic head for recording / reproducing corresponding to a magnetic disk, a position information management table that stores position information when reading or writing data from a non-continuous area to a non-continuous area is provided. One of the position information stored in the RAM and stored in the position information management table is data continuity information (the data of the position information stored in the position information management table is stored in another position information management table). Information indicating that the data is continuous with the position information data, data read information (second information indicating that part or all of the data exists in the buffer RAM), and data resident. information and flag area including the (always third information indicating that you read in the buffer RAM data) and a start sector number, section And a buffer address indicated by a start address and an end address of a buffer RAM in which data is stored. The position information of the position information management table is managed at the time of a data read or write operation. is doing.

[0018]

[0019] Second, in the first magnetic disk device, from the location information management table, and searches the position data being then read is recorded, the recording position search head movement for moving the magnetic head Means and the data of the sector at the searched position is stored in the buffer R of the magnetic disk device.
And a read means for reading in the AM.

[0020] Thirdly, in the first magnetic disk device, the end of the read or write consecutive data when creating a data continuously information in the position information,
Equipped with a judgment means to judge when the type of instruction has changed,
The determination means is configured to detect the end of continuous data read or write.

[0021] Fourth, in the first magnetic disk device, the end of the read or write consecutive data when creating a data continuously information in the position information,
A judgment means for judging from the execution interval of the read or write command is provided, and the judgment means detects the end of the continuous data read or write.

Fifth , in the fifth magnetic disk device, there is provided an execution interval setting means for setting an execution interval for judging the end of a read or write command from a host device.
The execution interval setting means is configured to set the execution interval from the host device.

[0023]

According to the present invention, conventionally, when a large amount of data is read or written, particularly when one data is read or written in an area which is not continuous, it is divided into a plurality of read or write instructions. Focusing on the point of execution, the position information management table stores the position when reading or writing data divided multiple times as data continuity information, and then reads or writes the next data. Sometimes it is used (the invention of claim 1) .

[0024] Second, in order to improve the caching and prefetching capabilities, based on the positional information, the data of the position designated by the read command, (already read in persistence specification or prefetch function) in the buffer sometimes , The data in the buffer is transferred, and the data on the disk is not read. Furthermore, after the read operation is completed,
The position of the next data is searched and the data is read into the buffer (the invention of claim 3).

By the way, the magnetic disk device itself cannot judge the end of the write operation of the data divided into a plurality of times. Point Therefore, the third, for one continuous data, when the plurality of read or write instruction is executed, in order to determine the end of the series of read or write type instruction has changed (For example, in the case of read, an instruction other than read) determines that a series of read or write operations has ended (the invention of claim 3 ).

[0026] Fourth, a magnetic disk apparatus itself, since it is not possible to determine the end of the divided data of the read or write operation a plurality of times, the read or write command is set from a certain time, or the higher-level device When it is not executed within the predetermined time, it is judged that a series of read or write is completed (the inventions of claims 4 and 5 ).

[0027]

Embodiment 1 Next, the magnetic disk device of the present invention will be described in detail with reference to the drawings. This embodiment mainly corresponds to the invention of claim 1, but is also related to the inventions of claims 2 to 5 .

The magnetic disk device of the present invention is characterized in that the position information management table is used to store the position information in order to read / write one piece of data to a discontinuous area. . Therefore, the position information management table used in the magnetic disk device of the present invention (see FIG.
The table stored in the RAM 24) will be described.

FIG. 2 is a diagram showing an example of the format of the position information regarding the position information management table used in the magnetic disk device of the present invention. In the figure, A
Is a flag, B is a start sector number, C is the number of sectors, and D is
Indicates a buffer address (start), and E indicates a buffer address (end).

As shown in FIG. 2, the position information management table is constructed so that N pieces of position information can be stored. Each position information management table (1 to
The position information stored in N) includes a flag A, a start sector number B, a sector number C, and a buffer address (start).
D, buffer address (end) E.

The flag A contains the following three pieces of information. 1) The data of the data continuation information m (1 ≦ m <N) th position information is (m +
1) Information indicating that it is continuous with the data of the first position information (first information).

2) Data read information Information indicating that part or all of the data from the start sector number B of the position information exists in the buffer RAM 25 indicated by the buffer addresses D and E (second information).

3) Data Resident Information Information indicating that the data from the start sector number B of the position information is preferentially read into the buffer RAM 25 (third information).

The magnetic disk device of the present invention uses such position information to perform data read / write operations. The position information is, as shown in FIG.
It comprises a start sector number B, a sector number C, a buffer address (start) D, and a buffer address (end) E.

This position information is updated with the position information temporarily stored in the buffer RAM every time the read / write operation is completed. Therefore, the position information stored in the position information management table can manage the data recorded in the discontinuous area as continuous data.

FIG. 1 is a functional block diagram showing an embodiment of the main configuration of a disk device connected to a host device in the magnetic disk device of the present invention. In the figure, 1
Is a host device, 2 is a magnetic disk device, and 21 is an I / F
(Interface) Controller, 22 is CPU, 2
3 is ROM, 24 is RAM, 25 is buffer RAM, 2
Reference numeral 6 is a read / write control circuit, and 3 is a magnetic disk.

The magnetic disk device according to the present invention shown in FIG. 1 has a basic configuration except that a table for managing position information (position information management table) is stored in the RAM 24. Similar to the device, and also
It is also known that the magnetic disk device 2 is connected to the host device 1 to form a filing system. This Figure 1
In the above device, the host device 1 executes a read / write command to the magnetic disk device 2.

The I / F controller 21 controls the interface between the magnetic disk device 2 and the host device 1. The ROM 23 stores control software (firmware) for the magnetic disk device 2.

The RAM 24 stores a table for use in the magnetic disk device of the present invention, that is, a table (management table) for managing position information.
The buffer RAM 25 is a memory that temporarily stores data to be read / written.

The read / write control circuit 26 has a function of controlling read / write to the magnetic disk 3. First, the operation at the time of writing data will be described. This data write operation is for one-time data write.

FIG. 3 is a flow chart showing the main processing flow at the time of writing data in the magnetic disk device of the present invention. In the figure, # 1 to # 5 indicate steps.

At step # 1, the host device 1 executes a write command to the magnetic disk device 2. In the next step # 2, the magnetic disk device 2 writes the data to the magnetic disk 3.

At step # 3, the position information at that time is temporarily stored in the position information management table in the RAM 24. In step # 4, the data at the write position is buffer R
It is determined whether or not to lead in the AM25.

If the data is to be read, the process proceeds to the next step # 5, the write data is held in the buffer RAM 25, and the flow of FIG. 3 is terminated. If it is determined in step # 4 that the data at the write position is not read into the buffer RAM 25, the flow of FIG.

By the processing of the above steps # 1 to # 5, one data write operation is executed. If it is set to always read (read) the data at the write position into the buffer RAM 25, step # 4 is omitted.

Next, the operation at the time of reading data will be described. This data read operation is also the case of one-time data read.

FIG. 4 is a flow chart showing the main processing flow at the time of reading data in the magnetic disk device of the present invention. # 11 to # 16 indicate steps.

In step # 11, the host device 1 executes a read command to the magnetic disk device 2. In step # 12, the magnetic disk device 2 sets the buffer RAM 2
Check if there is data in 5.

If there is no data in the buffer RAM 25, in the next step # 13, the data is read from the magnetic disk 3 and transferred to the host device 1. As a result of the determination made in step # 12, the buffer RAM 25
If there is data inside, go to step # 14.

In step # 14, the data in the buffer RAM 25 is transferred to the host device 1. After proceeding to step # 15 and transferring the data to the higher-level device 1, the position information at that time is temporarily stored in the RAM 24. Step # 1
At 6, the data of the sector to be read next from the position information management table is transferred from the magnetic disk 3 to the buffer RAM 2
Lead to 5.

The reading of the position information is continued until one of the following states is achieved. 1) Until the host device 1 accesses the magnetic disk device 2. 2) Until all the data of the sector indicated by the position information management table is read into the buffer RAM 25. 3) Until the buffer RAM 25 is full.

By the processing of steps # 11 to # 16 described above, one data read operation is executed. As described above, in the magnetic disk device of the present invention, a series of data write / write operations is performed based on the position information in the position information management table.
Perform read operation.

When a series of data write / read operations are completed according to the flow shown in FIGS. 3 and 4, the position information temporarily stored in the RAM 24 is stored in the position information management table. Update the contents and clear the temporarily stored location information. Next, the updating operation of the position information management table will be described.

FIG. 5 is a flow chart showing the main processing flow when updating the position information management table in the magnetic disk device of the present invention. In the figure, #
21 to # 23 indicate steps.

At step # 21, a series of data write /
Judge whether data read is completed. When finished, in the next step # 22, the contents of the position information management table are updated with the position information temporarily stored.

At step # 23, the temporarily stored position information is cleared, and the flow of FIG. 5 ends. Further, when it is determined in the previous step # 21 that the series of data write / data read is not completed, the flow of FIG. 5 is immediately terminated.

By the processing in steps # 21 to # 23 described above, the position information management table is updated after one data write / read operation is completed. In the magnetic disk device of the present invention, since the position information is composed of the minimum necessary information, R for the position information management table is used.
The use of the AM area can be minimized, and since the position information has data continuity information, the data recorded in the discontinuous area can be managed as continuous data.

[0058]

[Example of a magnetic disk apparatus that can be used as a guide of the present invention] Explain with the magnetic disk device that can be used as a guide of the present invention.

[0059] In Example of this, and to be able to set the position information described in the previous example from the host device, the prefetch function in a read instruction execution, is to be in effect immediately. Next, the operation for setting the position information in the magnetic disk device 2 by the host device 1 will be described.

[0060] FIG. 6, in the magnetic disk device of this,
6 is a flowchart showing a main processing flow when the host device 1 sets position information for the magnetic disk device 2. In the figure, # 31 to # 32 indicate steps.

At step # 31, the host device 1 executes a position information setting command to the magnetic disk device 2.
In step # 32, the higher level device 1 registers the position information in the position information management table.

By the above processing of steps # 31 to # 32, the setting operation of the position information for the magnetic disk device 2 of the host device 1 is completed. In this way, by setting the position information from the host device, that is, the magnetic disk
Data is divided and stored in the host device to which the device is connected
When the host device knows that the
By setting in the magnetic disk device as information
Therefore, the prefetch function at the time of executing the read instruction can be immediately enabled.

[0063]

[Embodiment 2] A magnetic disk device according to the present invention will be described.
A second embodiment will be described. This embodiment corresponds to the invention of claim 3 .

In this second embodiment, since the end of the continuous data read / write operation is judged by the change of the command, the continuous data position information management table can be created accurately. it can. Next, the determination of the end of the continuous data read / write operation will be described.

FIG. 7 is a flow chart showing the main processing flow for determining the end of a continuous data read / write operation in the magnetic disk device of the present invention. In the figure, # 41 to # 43 indicate steps.

At step # 41, after the read or write command is completed, the next command is another command (for example, a command other than the read command when reading, and a command other than the write command when writing). Command). When the next command is another command, the process proceeds to step # 43 to update the contents of the position information management table, and the flow of FIG. 7 ends.

As a result of the judgment in the previous step # 41, if the next instruction is not another instruction, the process proceeds to step # 42,
After the read or write command is completed, it is determined whether a fixed time has passed. If a certain period of time has passed, it is determined that the continuous data read / write has been completed, the process proceeds to step # 43, and the operation of updating the contents of the position information management table is performed. Ends the flow.

If it is determined in the previous step # 42 that the predetermined time has not elapsed, the process returns to the previous step # 41 and the same processing is repeated. Through the processing of steps # 41 to # 43 described above, the end of reading / writing continuous data is determined by the change of the command, so that the position information management table of continuous data can be accurately created. .

[0069]

[Third Embodiment] A magnetic disk device of the present invention will be described below.
Example 3 will be described. This embodiment corresponds to the invention of claim 5 , but is also related to the invention of claim 4 . According to the fourth aspect of the present invention, the end of continuous data read / write is judged from the execution interval of the read / write command. Therefore, even if the command does not change, the continuous data position information management table Can be created accurately.

In the third embodiment, in addition to the invention of claim 4 , the execution interval (time) for judging the end of continuous data read / write can be set from the host device. Therefore, the system (OS:
It is possible to judge the end of the read / write instruction according to the operating system. Next, an operation in which the host device 1 sets a time for determining the end of the continuous data write / read operation will be described.

FIG. 8 is a flow chart showing the main processing flow when the host device 1 sets the time for judging the end of the continuous data write / read operation in the magnetic disk device of the present invention. . In the figure,
# 51 to # 52 indicate steps.

In step # 51, the host device 1 executes a time setting command for the read / write end interval to the magnetic disk device 2. In step # 52, the time of the end determination interval is transferred.

By the processes of steps # 51 to # 52, the time setting operation for judging the end of the continuous data write / read operation for the magnetic disk device 2 of the host device 1 is completed. In this way, the execution interval (time) to judge the end of continuous data read / write
Since it can be set by the host device, it is possible to determine the end of the read / write command that matches the system (OS: operating system) of the host device.

[0074]

According to the magnetic disk drive of the first aspect, the position information at the time of reading or writing is composed of the minimum necessary information. Therefore, the use of the RAM area for the position information management table can be minimized. Further, since the position information has data continuity information, it is possible to manage the data recorded in the discontinuous area as continuous data.

[0075]

[0076] In the magnetic disk apparatus according to claim 2, then moved (seek) the magnetic head to the position data to be read is recorded, the next data to be read in advance in the lead in the buffer RAM. Therefore, the data can be immediately transferred in response to the read instruction to be executed next.

In the magnetic disk drive of claim 3 ,
The end of continuous data read or write is judged by the change of the instruction. Therefore, the position information management table of continuous data can be created accurately.

In the magnetic disk drive of claim 4 ,
The end of continuous data read or write is determined from the execution interval of the read or write command. Therefore, even if the command does not change, the position information management table of continuous data can be accurately created.

According to the magnetic disk device of the fifth aspect , the execution interval (time) for judging the end of continuous data read or write can be set from the host device. Therefore, it is possible to determine the end of the read or write command that matches the system (OS: operating system) of the host device.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing an embodiment of a main part configuration of a disk device connected to a host device in a magnetic disk device of the present invention.

FIG. 2 is a diagram showing an example of a position information format of a position information management table used in the magnetic disk device of the present invention.

FIG. 3 is a flowchart showing a main processing flow when writing data in the magnetic disk device of the present invention.

FIG. 4 is a flowchart showing a main processing flow at the time of reading data in the magnetic disk device of the present invention.

FIG. 5 is a flow chart showing a main processing flow when updating the position information management table in the magnetic disk device of the present invention.

[6] In reference to the magnetic disk unit according to the present invention, is a flowchart showing a flow of main processing when the host device 1 to set the position information to the magnetic disk unit 2.

FIG. 7 is a flowchart showing a main processing flow for determining the end of read / write operation of continuous data in the magnetic disk device of the present invention.

FIG. 8 is a flowchart showing a main processing flow when the upper level device 1 sets a time for determining the end of continuous data write / read operations in the magnetic disk device of the present invention.

FIG. 9 is a diagram showing an example of a sequence for reading into a buffer RAM when a read instruction is executed.

[Explanation of symbols]

1 Upper device 2 Magnetic disk unit 21 I / F controller 22 CPU 23 ROM 24 RAM 25 buffer RAM 26 Read / write control circuit 3 magnetic disk

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06F 3/06-3/08 G06F 12/08 G11B 20/10

Claims (5)

(57) [Claims] 1. In a magnetic disk device having a magnetic head for recording / reproducing corresponding to a magnetic disk, position information is stored when data is read from or written to a non-continuous area. The position information management table is stored in the RAM, and one of the position information stored in the position information management table is data continuity information (the data of the position information stored in the position information management table is other position information). First information indicating that the position information data stored in the management table is continuous, and data read information (second information indicating that part or all of the data exists in the buffer RAM). )
And a data resident information (third information indicating that the data is always read in the buffer RAM), a start sector number, the number of sectors, and a buffer RAM in which the data is stored. A magnetic disk device comprising a buffer address represented by a start address and an end address, and managing position information of the position information management table at the time of data read or write operation . 2. The magnetic disk device according to claim 1, further comprising: a recording position search / head moving unit that moves the magnetic head by searching the position information management table for a position where data to be read next is recorded. A read unit that reads the data of the sector at the searched position into the buffer RAM of the magnetic disk unit. 3. The magnetic disk device according to claim 1, further comprising a determining means for determining the end of continuous data read or write when creating the data continuity information in the position information when the type of instruction changes. A magnetic disk drive, comprising: the determination means for detecting the end of continuous data read or write. 4. The magnetic disk device according to claim 1, further comprising a determining unit that determines the end of continuous data read or write when creating the data continuity information in the position information from the execution interval of the read or write command. A magnetic disk drive, comprising: the determination means for detecting the end of continuous data read or write. 5. The magnetic disk device according to claim 4 , wherein an execution interval for determining the end of a read or write command is
A magnetic disk device, comprising: an execution interval setting means for setting from an upper device, wherein the execution interval setting means sets the execution interval from the upper device.