JP2734865B2 - Continuous write processing method for disk device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a disk drive having a protocol interface for transmitting a status reporting the reception of data when data is transferred following a write command for specifying a write data range. More particularly, the present invention relates to a continuous write processing method for a disk device that enables efficient data reception.

2. Description of the Related Art Data is transferred to a magnetic disk device used as an external storage device of a computer system following a write command for designating writing by designating the number of blocks of write data from a host device such as a SCSI interface. In some cases, a protocol is provided which has an interface defined to receive the data and transmit a status for reporting the reception state, thereby sequentially receiving the data.

In a magnetic disk device having such an interface, when a write command is continuously received, a disk rotation wait occurs due to a processing delay. However, the occurrence of the wait time reduces the processing efficiency of the computer system. It needs to be prevented in order to lower it.

[0004]

2. Description of the Related Art FIG. 4 is a block diagram for explaining an example of the prior art, and FIG. 5 is a time chart for explaining the operation of FIG.

The host device 1 sends a write command to the command processing unit 3 of the disk device 2 by designating the number of write data blocks based on the SCSI interface standard. The command processing unit 3 receives the command during the period indicated by A in the command processing of FIG. 5A, and then checks whether the disk control unit 5 is operating during the period indicated by B in the command processing of FIG. Perform a check. Then, the disk control unit 5 sets the disk medium 6
FIG. 5 shows a case where the data write control to the
(A) During the period indicated by C in the command processing, data transmitted from the host device 1 is received and written into the buffer memory 4.

At the same time, the command processing unit 3 instructs the disk control unit 5 to position the specified head at the specified cylinder of the disk medium 6 during the period indicated by F in FIG. Is performed.

When the head is positioned at the designated sector, the disk control unit 5 reads the data stored in the buffer memory 4 during the period indicated by G in FIG. Of the specified number of blocks from the specified sector.

When the disk control unit 5 reports the completion of the data writing, the command processing unit 3 changes the status in the period indicated by D in the command processing of FIG.
And the next command is received in the period indicated by E in the command processing of FIG.

In the above description, the case where there is no write cache function has been described. However, in order to improve processing efficiency, when data is received from the host device 1 and stored in the buffer memory 4, the status is sent to the host device. A responsive write cache function is provided.

That is, in the period A shown in the command processing of FIG. 5B, the command processing unit 3 receives the command, and
The disk controller 5 is checked during the period indicated by B in the command processing. Then, the data is received and written in the buffer memory 4 during the period indicated by C in the command processing of FIG.

Then, the status is sent to the host device 1 in the period indicated by D in the command processing of FIG. 5B, and the next command is received in the period indicated by E in the command processing of FIG. B)
The disk control unit 5 is checked during the period indicated by F in the command processing. Then, when the control of the disk medium 6 by the disk control unit 5 is completed, the next data is received in the period indicated by G in the command processing of FIG.

The command processing section 3 starts receiving data in the period indicated by C in the command processing of FIG. 5B, and at the same time, instructs the disk control section 5 to execute the disk control of FIG. 5B. In the period indicated by H, seek control for positioning the designated head at the designated cylinder of the disk medium 6 is performed.

When the head is positioned at the designated sector, the disk control unit 5 executes the J of the disk control shown in FIG.
The data stored in the buffer memory 4 is read out during the period shown in (1), and data of a specified number of blocks is written from a specified sector of the disk medium 6.

If the designated cylinder is the same during the period indicated by K in the disk control of FIG. 5B, the disk waits for rotation, and during the period indicated by L in FIG. Data is read from the buffer memory 4 and data of a specified number of blocks is written from a specified sector of the disk medium 6.

[0015]

As described with reference to FIG. 5A, when there is no write cache function, the host device 1 terminates the data transfer in the period indicated by C in FIG.
Until the writing of data to the disk medium 6 of the disk device 2 is completed, the status shown in D of FIG. 5A is not reported, and the instruction processing does not end.

The transfer time of the data sent by the host device 1 is shorter than the time for the disk device 2 to write the data to the disk medium 6, and the longer time from the completion of the data transmission until the status report is received, that is, FIG. 5 (A) C and D
During this time, there is a problem that the processing efficiency of the host device 1 is reduced.

As described with reference to FIG. 5B, when the write cache function is provided, when a write command is continuously transmitted from the host device 1, the period indicated by F in FIG. ,
Depending on the data writing time indicated by J, the disk medium 6
If the writing of the data to is not completed, G in FIG.
The next data indicated by cannot be received.

Therefore, when the data block indicated by G in FIG. 5B is written in the sector following the sector in which the data block indicated by C in FIG. 5B is written, the target sector is located at the head position. , The disk medium 6 waits for rotation during the period indicated by K in FIG. 5B. For this reason, there is a problem that the processing efficiency of the host device 1 is reduced.

The present invention has been made in view of such a problem, and when a write command is continuously issued, the previous data block
If the next data block is continuous and the disk control unit 5 is under data write control, the data of the next instruction is continuously received and stored in the buffer memory 4, and Without causing rotation waiting,
It is intended to write data continuously.

[0020]

Means for Solving the Problems The purpose of the present invention is shown in FIG.
As shown in (2), when data is transferred following a write command instructing writing by designating the number of blocks of write data from the higher-level device 1, the data is received and a status for reporting a reception state is transmitted. Thereby, an interface 9 of a protocol defined to sequentially receive data, a buffer memory 4 for storing received data blocks, and a control for writing the data blocks stored in the buffer memory 4 to the disk medium 6 And a command for writing data transmitted by the higher-level device 1 to the buffer memory 4 and instructing the disk controller 5 to write data based on a command transmitted by the higher-level device 1. In the disk device 7 including the processing unit 8, when the host device 1 continuously sends a write command, It is checked whether the data block to be written and the data block to be written next are continuous. If the data blocks are continuous, when the storage of the previous write data in the buffer memory 4 is completed, the status is changed to the host device. 1 to cause the next write data to be transmitted and to be stored in the buffer memory 4 following the previous data block, thereby causing the disk control unit 5 to continuously write data to the disk medium 6. This is achieved by:

[0021]

With the above configuration, a plurality of data blocks each of which is designated by a continuous write command are continuously stored in the buffer memory 4, so that the disk control unit 5 waits for the rotation of the disk medium 6. The data blocks can be continuously written in continuous sectors of the disk medium 6 without performing the above operation.

[0022]

1 is a block diagram of a circuit showing an embodiment of the present invention, and FIG. 2 is a time chart for explaining the operation of FIG.

When data is transferred following a write command instructing writing by designating the number of blocks of write data from the host device 1 such as a SCSI interface, the host device 1 receives this data and changes the reception state. By sending the status to be reported, the write command is sent to the command processing unit 8 of the disk device 7 by specifying the number of write data blocks via the interface 9 of the protocol specified to sequentially receive data.

The command processing section 8 receives the command during the period indicated by A in the command processing of FIG. 2, and then checks the disk control section 5 during the period indicated by B in the command processing of FIG. Then, when the disk control unit 5 does not control the writing of data to the disk medium 6, the data transmitted from the host device 1 is received and written into the buffer memory 4 during the period indicated by C in the command processing of FIG. .

At the same time, the command processing unit 8 instructs the disk control unit 5 to perform seek control for positioning the specified head to the specified cylinder of the disk medium 6 during the period indicated by J in FIG. .

Therefore, when the head is positioned at the designated sector, the disk control unit 5 reads out the data stored in the buffer memory 4 during the period indicated by K in the disk control in FIG. Then, data of a specified number of blocks is written from a specified sector.

When the command processing unit 8 causes the disk control unit 5 to start writing data to the disk medium 6 as described above, the command processing unit 8 sends a status to the host device 1 in a period indicated by D in FIG. When the next command is received during the period indicated by E in the two-command processing, the disk controller 5 is checked during the period indicated by F in the command processing of FIG.

If the write control of the disk control unit 5 to the disk medium 6 is continued, the logic of the last block of the data block currently being written to the disk medium 6 given by the write command from the host device 1 in advance. The address is compared with the logical address of the first block of the data block instructed to be written by the next write command, and the data block based on the previous write command and the data block based on the next write command are continuously output. You have
The next data is received in the period indicated by G in the command processing of FIG.

At this time, since the data block stored before exists in the buffer memory 4, the command processing unit 8 sequentially stores the data block based on the next write instruction following this data block.

When the command processing unit 8 finishes storing the data block based on the next write command in the buffer memory 4 during the period indicated by G in the command processing in FIG. 2, the command processing unit 8 returns the status in the period indicated by H in the command processing in FIG. To the host device 1.

Therefore, when the write command is further continued, the host device 1 sends the next write command and the disk device 7
Repeats the above operation. When the writing of data blocks to the disk medium 6 is completed during the period indicated by K in the disk control of FIG. 2, the disk control unit 5 stores continuous data blocks in the buffer memory 4. In the period indicated by L, the data block based on the next write command is continuously written to the disk medium 6.

Therefore, for example, in the case of a disk device in which the time required to make one round of one track is 10 ms and there are 100 data blocks in one track, the processing time of one block is 10 ms ÷ 100 = 0.1 ms Becomes If the host device 1 processes 90 blocks,
When a write command is issued three times every 30 blocks, the conventional processing method is as follows.

Write time of 90 blocks = 0.1 ms × 30 + 10 ms + 0.1 ms × 30 + 10 ms + 0.1 ms × 30 = 29 ms In the present invention, the disk rotation waiting time 10 ms × 2 = 20 ms
Can be reduced, and the writing time for 90 blocks is 9
ms can be expected.

In the above embodiment, the case where the data block based on the previous write command and the next write command are continuous has been described, but the data block based on the previous write command and the next write command is Even if you are several blocks away,
Since the rotation waiting time of the disk is short, it is a matter of course that data can be continuously received by the same processing as described above. That is, even if there is a space between K and L in FIG.

FIG. 3 is a flowchart for explaining the instruction processing of the present invention. Upon receiving the write command in step (1), the command processing unit 8 checks in step (2) whether the disk control unit 5 is operating. That is, it is checked whether or not data writing control to the disk medium 6 is being performed.

If the disk control unit 5 is not operating, the disk control unit 5 is instructed in step (3) to perform a seek operation, and the process proceeds to step (5). If it is medium, check in step (4) whether the data blocks are continuous. That is, it is checked whether the data block according to the previous write instruction and the data block according to the next write instruction are continuous.

If the data blocks are not continuous in step (4), the process proceeds to step (3). If the data blocks are continuous, the reception of data is started in step (5) and the buffer memory is started. 4 is stored.

Then, the command processing section 8 executes step (6).
To instruct the disk control unit 5 to write data to the disk medium 6. Then, the command processing unit 8 waits for the completion of the data reception in step (7), and when completed, in step (8), changes the status to the host device 1.
To send to.

[0039]

As described above, according to the present invention, a plurality of data blocks each designated by a continuous write command can be successively stored in the buffer memory, so that the disk medium waits for rotation. Without writing, data blocks can be continuously written on a disk medium.

Therefore, the host device shortens the processing time of successive write instructions, and the greater the number of issued instructions, the greater the effect of the shortening, and the higher the processing efficiency of the computer system.

[Brief description of the drawings]

FIG. 1 is a block diagram of a circuit showing one embodiment of the present invention.

FIG. 2 is a time chart for explaining the operation of FIG. 1;

FIG. 3 is a flowchart illustrating instruction processing according to the present invention;

FIG. 4 is a block diagram illustrating an example of a conventional technique.

FIG. 5 is a time chart for explaining the operation of FIG. 4;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Upper device 2, 7 Disk device 3, 8 Command processing part 4 Buffer memory 5 Disk control part 6 Disk medium 9 Interface

Claims (1)

(57) [Claims] When a data is transferred following a write command instructing writing by designating the number of blocks of write data from a higher-level device (1), the data is received and a status for reporting a reception state is transmitted. By doing so, an interface of a protocol specified to sequentially receive data (9) and a buffer memory (4) for storing received data blocks
A disk control unit (5) for controlling the writing of the data blocks stored in the buffer memory (4) to the disk medium (6).
Based on the command sent by the higher-level device (1),
(1) Writes the data to be sent out to the buffer memory (4), and a command processing unit (8) for instructing the disk control unit (5) to write data. When the host device (1) continuously sends a write command,
It is checked whether or not the previous write data block and the next data block to be written are continuous.If the data blocks are continuous, the data block is stored when the previous write data is completely stored in the buffer memory (4). The status is transmitted to the host device (1), the next write data is transmitted, and the buffer memory (4) is stored following the previous data block, so that the status is continuously transmitted to the disk control unit (5). And writing data to the disk medium (6).