JP2000172449A - Disk drive - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a disk drive that improves the mean processing speed of read access by increasing cache hit probability by registering data which need to be processed by correction and restoration in a cache table in high priority order. SOLUTION: A disk interface control part 3 sends and receives data or various commands to and from a disk drive D through a disk I/F.DIF. A data buffer 4 is stored with data read out of the disk drive D in priority order given to a cache control part 23. A disk command start/end part 24 decides the status of a disk command at its end and informs the cache control part 23 of the result. The disk command start/end part 24 starts a timer 27 at the start of a disk read command and informs the cache control part 23 that the timer value exceeds a set value in such a case when the disk read command normally ends.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a disk device having a cache control function and reading data from a disk drive.

[0002]

2. Description of the Related Art Conventionally, a cache operation in a disk device is performed by the following method. When a host read command is received from a higher-level host through a higher-level interface, the cache control unit refers to the cache table to confirm data requested by the host read command.

If the read request range in the host read command matches or is included in the range information of the data registered in the cache table, the cache control means reads the range data from the data buffer and transfers it to the upper host. . On the other hand, if the read request range in the host read command does not match or is not included in the range information of the data registered in the cache table, the cache control unit transmits the data from the lower-level disk drive through the disk interface control unit. The data in the range to be read is read.

Then, the cache control means stores the data in the data buffer and newly registers range information of the data in the cache table. Next, the cache control means transfers the data from the data buffer to the host. At this time, in order to register the range information in the cache table, the cache control unit deletes the information with the lowest priority and then registers the new range information with the priority according to the algorithm defined by the user.

[0005]

However, the above-mentioned conventional disk drive has the following problems.
In a conventional disk device, when data is read from a disk drive, a failure occurs at a certain address, and if the data can be corrected by some correction / recovery means, a failure occurs again when the next access request is made to the same position. There is a high probability that correction can be made by using correction recovery means.

[0006] The range information of the data that could be read by "correction" was registered in the cache table, but there was no access for a while. If the "correction" information was evicted from the table and the corresponding location was accessed, The failure occurs again. Executing the correction / recovery means for this failure takes a lot of recovery time, so that the processing time of one disk read command also becomes long. This causes a reduction in the overall average processing speed of data reading.

Further, in the conventional disk drive, when reading data from the disk drive, the disk read command at a certain address has been completed normally. However, a failure has actually occurred and normal operation is performed by using some internal correction / recovery means. May be showing. Also in this case, when an access request comes again, the same fault occurs, and there is a high probability that the internal correction recovery means will be taken again. This disk read also takes a lot of recovery time, so that the overall average processing speed is reduced as in the case where the above-mentioned "correction" was successful.

The present invention has been made under such a background. Data requiring correction and recovery processing is registered in a cache table with a high priority, and the cache hit probability is increased to increase the average processing speed of read access. An improved disk device is provided.

[0009]

According to the first aspect of the present invention,
Sends and receives commands and data to and from the host system,
In a disk device for writing or reading the data by this command, a disk drive (eg, disk drive D) for storing the data;
Reading means (for example, a disk interface control unit 3) for reading data from the disk drive based on the read command, holding means (for example, a data buffer 4) for holding data read by the reading means, and holding of the holding means Management means (for example, a cache table 26) for managing the priority order of the data to be executed
A counter (e.g., timer 27) for measuring a time for reading the data from the disk drive; detecting a state when the data is read from the disk drive; (For example, a disk command activation end unit 24 and a main control unit 21).

According to a second aspect of the present invention, in the disk drive of the first aspect, the read control means (for example,
Disk command activation end unit 24, main control unit 21)
Compares the time measured by the counter (for example, the timer 27) with a preset time set in advance,
A comparison signal is output as a comparison result.

According to a third aspect of the present invention, in the disk device according to the first or second aspect, the read control means (for example, the disk command activation end unit 24, the main control unit 21) changes the state of the disk from the state. When detecting that the data has a correctable fault when reading data from the drive (for example, the disk drive D), or when detecting that the time data is larger than the set value from the comparison signal, It is characterized in that the priority in the data management means is raised.

In FIG. 1, a host read command received through the host interface control unit 1 is referred to by the cache control unit 23 as information in a cache table 26. At the time of a mishit, data is read from the disk drive through the disk interface control unit 3. When the disk command control unit 24 determines that the read status of the disk drive is “corrected”, or when the processing time of the disk read exceeds the specified value by the timer 27 even if the disk read has been completed normally. If it is determined, the cache control unit 23 registers the new range information in the cache table 26 with a higher priority. As a result, the probability of a cache hit when the same range access is made is increased, so that the ratio of disk access operations due to a mishit is also reduced. Since there is a failure and the number of disk reads in a portion where much processing time is required for the correction recovery is reduced, it is possible to prevent a decrease in the average processing speed of the read access of the disk device.

[0013]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a disk device according to an embodiment of the present invention. In the figure, reference numeral 1 denotes a host interface control unit, which receives and transmits a read command for performing read control and data with the host system HS via an upper interface UIF.

Reference numeral 2 denotes a data control unit, which operates under program control, and controls the disk interface control unit 3 and the data buffer 4. The disk interface control unit 3 includes a disk interface (I / F)
It sends and receives data or various commands (commands) to / from the disk drive D via the DIF.

4 is a data buffer (cache memory)
And stores the data read from the disk drive D. In the data buffer 4, data is assigned with a priority determined based on a priority of a process used in the host system HS or a reference predetermined by a qualification.

The data control unit 2 includes a main control unit 2
1, a host command analysis unit 22, a cache control unit 23, a disk command start / end unit 24, a data transfer control unit 25, a cache table 26, and a timer 2.
7 is provided.

The main control unit 21 controls the host command analysis unit 22, cache control unit 23, disk command start / end unit 24, and data transfer control unit 25 according to a program stored in a storage unit (not shown). Control to operate at the timing.

The host command analysis section 22 analyzes the host command received from the host interface control section 1 and determines whether or not the command is a read command. get.

The cache control unit 23 compares the access range information of the read command acquired by the host command analysis unit 22 with the range information registered in the cache table 26, and determines whether a cache hit has occurred. In addition, when there is no cache hit, that is, in the case of a mishit, the cache control unit 23 outputs a control signal to the disk command activation / end unit 24 to read disk data.

The disk command activation end unit 24 issues a read command to the disk interface control unit 3 when a control signal of a disk read instruction is output from the cache control unit 23. Further, the disk command start / end unit 24 transfers the data read from the disk drive D by the disk interface control unit 3 to the data transfer control unit 25.

Further, a disk command start / end unit 24
Determines the status at the end of the disk command, and notifies the cache control unit 23 of the result. Similarly,
The disk command activation end unit 24 activates the timer 27 when the disk read command is activated, and notifies the cache control unit 23 when the timer value exceeds the set value when the disk read command is normally terminated.

The data transfer control unit 25 reads data requested by the cache control unit 23 from the data buffer 4. Further, the data transfer control unit 25 controls the disk drive D transmitted from the disk command activation end unit 24.
Is stored in the data buffer 4.

The cache table 26 stores cache information by the cache control unit 23, and every time a host read command is input, the data range in the data buffer 4 is compared with the access range information (cache information) of this command. Information is stored. Timer 27
Are reset and activated by the disk command activation end unit 24.

Next, an example of the operation of the embodiment will be described with reference to FIGS. FIG. 2 is a flowchart showing the operation of the disk device shown in FIG. In step A1, the host interface control unit 1 inputs various commands from the host system HS via the upper interface UIF. Then, the host interface control unit 1 outputs the command to the host command analysis unit 22.

Then, the host command analysis unit 22 analyzes the command input from the host interface control unit 1 and determines whether the input command is a read command. As a result, when it is determined that the command is a read command, the host command analysis unit 22 acquires information on an address to be accessed and an access range from the accompanying data as range information, and outputs this range information to the cache control unit 23. Then, the host command analysis unit 22 advances the processing to Step A2. When it is determined that the command is another command, the host command analysis unit 22 does not perform the following processing.

Next, in step A2, the cache control unit 23 refers to the cache table 26 and determines whether or not the range data requested by the host read command is stored in the data buffer 4, ie, whether or not a cache hit has occurred. Is determined. At this time, if the cache control unit 23 determines that there is a cache hit, the process proceeds to step A3.

Next, in step A3, the cache control unit 23 reads out data in a corresponding range from the data buffer 4 via the data transfer control unit 25. Then, the cache control unit 23 transmits data to the host system HS via the host interface control unit 1 and the upper interface UIF, and ends the host read command.

On the other hand, if the cache control unit 23 determines in step A2 that there is no cache hit, the process proceeds to step A4. Next, step A
In 4, the cache control unit 23 outputs a control signal to the disk command start / end unit 24 so as to read data in a corresponding range from the disk. Then, the cache control unit 23 advances the processing to step A5.

Next, in step A5, the disk command start / end unit 24 issues a disk command to the disk interface control unit 3 via the disk interface DIF to read out the read data of the corresponding range from the lower disk drive D. When activating the disk command, the disk command activation ending unit 24 resets and activates the timer 27 for monitoring the processing time of the disk command. Then, the disk command activation end unit 24 advances the processing to step A6.

Next, in step A 6, the disk command start / end unit 24 stores the data read from the disk drive D in the data buffer 4 through the data transfer control unit 25. Then, the disk command activation end unit 24 monitors the end of the command for reading the disk D. Disk command activation end unit 24
Repeatedly executes step A6 until the end is confirmed, and when the end is confirmed, advances the process to step A7.

Next, in step A7, the disk command start / end unit 24 checks the status of the read to the disk drive D. At this time, when detecting that an uncorrectable error has occurred during the reading of data from the disk drive D and the command has been abnormally terminated, the disk command activation end unit 24 advances the process to step A9.

Next, in step A 9, the disk command start / end unit 24 creates an error status corresponding to an error when reading data from the disk drive D, and outputs it to the cache control unit 23. Then, the disk command activation end unit 24 performs the processing in step A1.
Proceed to 3. Next, in step A13, the disk command activation end unit 24 ends the processing of the data read command from the host system HS.

On the other hand, in step A7, if the disk command activation end unit 24 detects that the read status for the disk drive D is the "correction" status, the process proceeds to step A10. Here, the “correction” status indicates that the lower disk drive D
Indicates that an error such as a data read error has occurred, but has been recovered by using retries or other correction means, and that normal data has been read.

When the "correction" status is detected, when the same location of the disk drive D is read next, the probability that the same fault will occur again at the same location is high. For this reason, even if the data read by the retry or the correction unit can be recovered, the processing time for reading the data from the disk drive D becomes longer as compared with the case where the data can be read normally. As a result, the above-described failure causes a reduction in data read performance in the disk device.

Next, in step A10, the disk command activation / termination unit 24 sets the "correction" status of the data to be read first to the disk command activation / termination unit 2.
4, and outputs a control signal to the cache control unit 23. Thus, when registering the new range information of the data in the cache table 26, the cache control unit 23 sets the cache hit priority higher than that at the time of normal registration, and increases the cache hit probability. Then, the disk command activation end unit 24 advances the processing to step A12.

Next, in step A12, the data transfer control unit 25 transfers the data read from the disk drive D from the data buffer 4 to the host system HS.

In step A7, when the disk command start / end unit 24 detects that the read status of the disk drive D is "normal" status,
The process proceeds to Step A8.

Next, in step A8, when the read command to the disk drive D ends normally, the disk command start / end unit 24 compares the time measured by the timer 27 with a predetermined time-over set value. . This comparison is based on the detection of a case where even if the read status is “normal”, a slight failure occurs in the disk drive D, the recovery is internally performed using retry or correction means, and a normal response is obtained. I do.

That is, also in this case, the time required for the correction process is long and the performance is reduced. Therefore, the value of the timer 27 when the disk command is normally completed is compared with a predetermined time-over setting value. growing. Here, the disk command activation end unit 24
If it is detected that the count value of the timer 27 exceeds the set value, the process proceeds to step A10.

Next, in step A10, the disk command activation end unit 24 outputs a control signal to the cache control unit 23 in order to perform cache control in the same manner as when the "correction" status is detected. As a result, the cache control unit 23 sets the priority of data in the cache table 26 higher than usual, and increases the cache hit probability. Then, the disk command activation end unit 2
4 advances the process to step A12.

Next, in step A12, the data transfer control unit 25 transfers the data read from the disk drive D from the data buffer 4 to the host system HS. Then, the data transfer control unit 25 advances the processing to step A13.

Next, in step A13, the disk command activation end unit 24 ends the processing of the data read command from the host system HS.

On the other hand, in step A8, if the disk command start / end unit 24 determines that the value of the timer 27 has not exceeded the set value, the disk command start / end unit 24 proceeds to step A1.
Proceed to 1.

Next, in step A11, the disk command activation end unit 24 sets the cache hit priority to the normal priority, and proceeds the processing to step A12.

Next, in step A12, the data transfer control unit 25 transfers the data read from the disk drive D from the data buffer 4 to the host system HS. Then, the data transfer control unit 25 advances the processing to step A13.

Next, in step A13, the disk command activation end unit 24 ends the processing of the data read command from the host system HS.

As described above, the disk device of the present invention
When registering data in the data buffer 4, the following two conditions are used as conditions for increasing the priority in the data buffer 4. In reading from the disk drive D, data can be correctly read by the error correction and recovery processing. In reading data from the disk drive D, the data can be read correctly over the set time.

The disk command start / end unit 24 detects the above-mentioned two states from the status, and when a status that satisfies the condition is detected, in order to improve the cache hit of the corresponding data in the data buffer 4, The priority of data in the table 26 is increased.

As a result, when the next read command is executed, the corresponding data is registered in the data buffer 4 with a higher priority than other data, so that the probability of a cache hit is improved and a failure is likely to occur. Data access time of the disk device in reading the data area is reduced, and as a result, the average processing speed of the system can be increased.

As described above, one embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and a design change or the like may be made without departing from the gist of the present invention. The present invention is also included in the present invention.

[0051]

According to the present invention, there is provided a disk device for transmitting / receiving a command and data to / from a host system and writing or reading the data by the command, a disk drive for storing the data,
Reading means for reading data from the disk drive based on the read command; holding means for holding data read by the reading means; management means for managing the priority of the data held by the holding means; And a read control means for detecting a state when reading the data from the disk drive and controlling the management means based on the detection result and the time. If the access range of this data is not in the holding means by a read command from the host system (cache miss), the disk drive is read, and this read processing ends with a "correction" status requiring a lot of processing time. The first time that The priority when registering the broadcast to the management means by higher than normal registration, if there is a command of the lead of the access to the same access range after the next time,
By managing the cache hit for a long time, it is possible to prevent the time-consuming “correction” access to the disk drive and to increase the average processing speed.

Further, according to the present invention, there is provided a counter for counting the time for reading the data from the disk drive, and the time for counting the counter and the set time previously stored in the read control means are determined by the read control means. Compared and manages the time required to access the disk drive for time-consuming data, so the read command from the host system reads the disk drive when the access range of this data is not in the holding means,
In the case where the read process is completed normally and takes longer than the specified read time even if the read process is completed normally, the priority of registering the information of the access range in the management means first is set higher than the normal registration, When the same range is accessed from the next time on, the cache hit is managed for a longer time than before so that it is possible to prevent access to the time-consuming disk drive and increase the average processing speed. Becomes

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a disk device according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating an operation of the disk device shown in FIG. 1;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Host interface control part 2 Data control part 3 Disk interface control part 4 Data buffer 21 Main control part 22 Host command analysis part 23 Cache control part 24 Disk command start end part 25 Data transfer control part 26 Cache table 27 Timer D Disk drive HS Host system DIF Disk I / F (disk interface) UIF Upper I / F (upper interface)

Claims (3)

[Claims] 1. A disk device for transmitting and receiving a command and data to and from a host system and writing or reading the data with the command, comprising: a disk drive for storing the data; and a disk drive based on the read command. Reading means for reading data; holding means for holding data to be read by the reading means; managing means for managing the priority of the data held by the holding means; and a counter for measuring a time for reading the data from the disk drive. And a read control means for detecting a state when the data is read from the disk drive, and controlling the management means based on the detection result and the time. 2. The reading control unit according to claim 1, wherein the read control unit compares the time measured by the counter with a preset time set in advance, and outputs a comparison signal as a comparison result. Disk unit. 3. When the read control unit detects that the data has a fault that can be corrected when reading data from the disk drive from the state, or when the time data indicates the set value based on the comparison signal. 3. The disk device according to claim 1, wherein when it is detected that the value is larger than the value, the priority of the data management unit is increased.