JP2743705B2 - Retry control method for disk controller - 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 control apparatus, and more particularly, to a time required for relocating a head to a record where an error has occurred when an error occurs while processing a certain record of the disk apparatus. The present invention relates to a retry control method of a disk control device to be shortened.

When an error occurs while processing a certain record in the magnetic disk device, the magnetic disk controller temporarily disconnects the connection with the magnetic disk device, and when it becomes possible to process the record in which the error has occurred, the magnetic disk controller re-executes the processing. Combine and process the records with errors.

In general, an error record is searched for by using an address of a record immediately before an error record stored during processing. That is, the previous record is read at the address of the record immediately before the stored record in which the error occurred, and the next record to be read is recognized as the record in which the error occurred, and The processing of the immediately preceding record is not performed, and the processing is started from the record to be read next.

When the immediately preceding record and the record in which the error has occurred are present in different cylinders, it is necessary that the time required to start processing the record in which the error has occurred does not become long.

[0005]

2. Description of the Related Art FIG. 4 is a diagram for explaining an example of the prior art.
FIG. 5 is a diagram illustrating a conventional retry process.

FIG. 4A shows a case where a record and a record on the same track are processed. The record is processed normally. However, if an error occurs during the processing of the record, the magnetic disk control device temporarily stops the magnetic recording. Disconnect the disk device and prepare for retry processing.

When the retry preparation is completed, the magnetic disk controller reconnects with the magnetic disk device and, as shown in the search operation, detects the record by the address of the stored record, and the processing of this record is started. Without performing the processing, the processing is restarted from the next read record as shown in the processing restart.

FIG. 4B shows a case where records are present on different tracks in the same cylinder. When the last record on track (1) and the first record on track (2) are processed continuously, the record When the index is detected after the end of the process, the head is switched to process the first record of the track (2) .If an error occurs during this process, the magnetic disk controller switches the head, Record (1) as above
When the index is detected after the last record of the track (2) is detected, the head is switched and the processing is restarted from the first record of the track (2).

FIG. 4C shows a case where a record and a record exist in different cylinders. The magnetic disk control device processes the last record of the last track (1) of the cylinder (1) and then controls the positioning of the head. After the head is positioned at cylinder (2), if the index of the first track (2) of cylinder (2) is detected, the record immediately after this index is processed.

If an error occurs during the processing of the record, the magnetic disk control device repositions the head on the cylinder (1). When a record is detected, the magnetic disk control device repositions the head on the cylinder (2). Is restarted.

That is, as shown in the control device of FIG. 5, the magnetic disk control device, for example, reads a record from the disk device and, after detecting an index as shown in index detection, recognizes a cylinder boundary and performs a seek command. As shown in the figure, the head
Instruct (2) to position.

Therefore, as shown in the seek operation of the disk drive in FIG. 5, the magnetic disk drive performs a seek operation for positioning the head on the cylinder (2). When the index of the track (2) of the cylinder (2) is detected, the record is read. Then, as shown in FIG. 5, when the error of the record is detected, the magnetic disk device is disconnected, and when the retry preparation is completed, it is reconnected.

Then, in order to execute the retry, a seek command is transmitted, and the magnetic disk drive positions the head at the cylinder (1). When the record is read, the magnetic disk control device searches for the record as shown in the search of FIG.
Since it exists in (2), a seek instruction is sent.

Therefore, the magnetic disk drive performs a seek, and the record is read by detecting the index of the track (2).

[0015]

In the retry processing shown in FIGS. 4A and 4B, there is no waste in processing time from the detection of a record to the recognition of the record from the address of the record. In the retry process shown in (C), the head positioning control is executed twice as if the seek command was sent twice after the retry in FIG.

The positioning operation of the head requires a lot of time because it includes a mechanical operation. Therefore, there is a problem that the time required for the retry becomes longer and the processing efficiency of the computer system using the magnetic disk device is reduced.

The present invention has been made in view of such a problem, and when a record is processed across different cylinders, the processing shown in FIG.
It is an object of the present invention to reduce the time required for the retry processing by eliminating the need to perform head positioning control when the retry processing as shown in FIG.

[0018]

Means for Solving the Problems The purpose of the present invention is shown in FIG.
As shown in the figure, the disk control device 13 processes the records of the respective tracks constituting the same cylinder by switching the heads 1 to 3 which follow the respective tracks, and processes the records of the tracks of different cylinders. Heads 1-3
After performing the positioning control, the heads 1 to 3 that follow each track are switched and processed, and when an error occurs in the record being processed and a retry is performed, the record immediately before the record in which the error occurred is performed. The address of the record is used as the retry address 19, and the record in which the error has occurred is detected.

After the last record of the last track of one cylinder is processed normally, the head is controlled to be positioned in another cylinder by controlling the positioning of the head. The address 19 is updated to the address of the head track of another cylinder, and if an error occurs while processing the head record of the head track of this other cylinder, the flag 18
Is set, after confirming that the updated retry address 19 matches the address of the head track of the cylinder that the heads 1 to 3 are following, the head 1 to 3 Processing of the first record of the first track is performed.
8 is achieved by resetting.

[0020]

With the above arrangement, when a record is processed across different cylinders, if an error is detected during processing of a record immediately after indexing, it is possible to prevent the head positioning operation from being executed twice. You can do it.

Therefore, it is possible to prevent the processing efficiency of the computer system using the magnetic disk device as the external storage device from being lowered.

[0022]

1 is a block diagram of a circuit showing an embodiment of the present invention, and FIG. 2 is a detailed block diagram of each functional unit.

FIG. 1A shows an example of a disk device, and FIG. 1B shows an example of a disk control device. The disk controller 13 is composed of a plurality of functional units. The channel adapter 14 has a configuration as shown in FIG. 2A, and the processor 23 operates by reading a program stored in the control storage 24. Start I / O from a higher-level device, not shown, for example, a channel via the circuit 21
Receiving the instruction and transferring data to and from the channel, and controlling the common bus control circuit 22 to transfer data to and from other functional units via the common bus 20.

The device adapter 15 has a configuration as shown in FIG. 2A, and the processor 23 operates by reading out a program stored in the control memory 24 and operating.
1 to send a command to the control circuit 11 of the disk device to transfer data to and from the disk device, and to control the common bus control circuit 22 to communicate with other functional units via the common bus 20. Data transfer.

The resource manager 16 has a configuration as shown in FIG. 2B. The processor 26 operates by reading out a program stored in the control storage 25, and controls the common bus control circuit 27 to control the common bus 20. All the start I / O commands from the channel are received and stored in the RAM 28, and the start I / O commands are centrally managed.
For each start I / O command, a channel adapter 14
An instruction for data transfer processing is issued to the device adapter 15.

The control information for each received start I / O command, the control information for each channel path, and the control information for the disk drive are stored in the RAM 28, and the other functional units are stored in the RAM 28.
The RAM 28 is accessed via the access control circuit 29, and the information can be read, and the information is centrally managed.

When the channel sends a start I / O command to the channel adapter 14 and instructs, for example, to read data from the disk device, the channel adapter 14 requests the resource manager 16 to use the common bus 20 and is granted permission. Then, the RAM 28 of the resource manager 16
, The cylinder address, the head address, and the record address are stored.

The device adapter 15 operates according to the instruction of the resource manager 16 and
Read out the cylinder address, head address, and record address stored in the disk drive control circuit 1;
1 to control the servo circuit 7,
To the specified cylinder.

The servo circuit 7 supplies a drive current to the motor 6 to position the heads 1 to 3 at the specified cylinder positions from the current position of the servo head 4 based on the servo information read by the servo head 4, Move 5

When the device adapter 15 is notified by the control circuit 11 that the heads 1 to 3 have been positioned on the specified cylinder, the device adapter 15 controls the control circuit 11 when the head 1 is following the specified track. After that, the amplifier 8 is enabled, and the record read by the head 1 is received via the demodulation circuit 12.

When the head 2 is following the specified track, the amplifier 9 is enabled via the control circuit 11 and the record read by the head 2 is received via the demodulation circuit 12 to receive the specified track. When the head 3 is following, the amplifier circuit 10 is enabled via the control circuit 11, and the record read by the head 3 is received via the demodulation circuit 12.

The data of the received record is checked for errors. If the data is normal, the data is written to the shared memory 17. Then, the device adapter 15 stores the address of the normally processed record in the shared memory 17 as the retry address 19.

Accordingly, the retry address 19 is always updated with the address of the last record processed normally. When data is written to the shared memory 17, the channel adapter 14 reads the data written to the shared memory 17 and transfers it to the channel in accordance with the instruction of the resource manager 16.

When the device adapter 15 is notified by the control circuit 11 that the servo head 4 has read the index, the device adapter 15 recognizes the cylinder boundary from the received address, and transfers the heads 1 to 3 from one cylinder to another cylinder. Recognizing that the seek operation for positioning is necessary, sets the flag 18 in the shared memory 17 and
The address of the head track of the seek destination cylinder is stored as the retry address 19, and the address of the record stored so far is updated.

Then, a seek command is sent to the control circuit 11 to position the heads 1 to 3 on other cylinders.
For example, the head 1 is a track of the cylinder (2) in FIG.
Assuming that (2) is followed, device adapter 1
5 enables the amplifier circuit 8 via the control circuit 11,
When the control circuit 11 notifies the reading of the index, the reading of the record is recognized, and the error of the data of the record is checked.

At this time, if an error is detected as described above, the device adapter 15 reads the retry address 19 of the shared memory 17 and reads the track (2) of the cylinder (2).
The servo information read by the servo head 4 via the control circuit 11 confirms that the track that the head 1 is following is the track (2) of the cylinder (2).

Then, it is checked whether or not the flag 18 of the shared memory 17 is set. Since the index is detected because the flag 18 is set, when the immediately following record, that is, the record is read, an error check of the data of this record is performed. If there is no error, the data is stored in the shared memory 17 and the flag 18 is reset.

Then, the address of the record is stored as the retry address 19 of the shared memory 17. FIG. 3 is a diagram illustrating the operation of FIG.

As shown in the control device of FIG. 3, the device adapter 15 of the disk control device 13 passes through the demodulation circuit 12 of the disk device in the processing of (1), and the track of the cylinder (1) shown in FIG. Receive the record of (1).

When the detection of the index of the track (1) is notified via the control circuit 11 in the process (2), the cylinder boundary is recognized, and a seek command is sent to the control circuit 11 of the disk device. Instruct the cylinder (2) of 4 (C) to position the head.

Then, the disk device is disconnected, and another process is executed. During this time, the control circuit 1 of the disk device
1 executes a seek for positioning the head at the designated cylinder (2), and when the seek is completed, the disk controller 13 rejoins and proceeds to the processing (3).

When the index of the track (2) of the cylinder (2) is detected, the reading of the record is recognized, and the data error of the record is checked. Here, as shown in the figure, when an error occurrence is detected, a retry is executed.

That is, in the process (4), it is confirmed that the track that the head 1 is following is the track (2) of the cylinder (2), and the flag 18 is set. Is processed, and if there is no error, the flag 18 is reset and the process ends normally.

In FIG. 3, the process of transmitting the seek instruction twice at the time of retry, which is shown in FIG. 5, is not required.

[0045]

As described above, according to the present invention, when a record is processed from one cylinder to another cylinder, a seek operation is performed at the time of retry when an error occurs in the first record of another cylinder. Since it is not necessary to repeat the operation twice, the retry time can be shortened and the retry processing can be simplified.

[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 detailed block diagram of each functional unit.

FIG. 3 is a view for explaining the operation of FIG. 1;

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

FIG. 5 illustrates a conventional retry process.

[Explanation of symbols]

 1-3 head 4 servo head 5 carriage 6 motor 7 servo circuit 8-10 amplifying circuit 11 control circuit 12 demodulation circuit 13 disk controller 14 channel adapter 15 device adapter 16 resource manager 17 shared memory 18 flag 19 retry address 20 common bus 21 Interface circuit 22, 27 Common bus control circuit 23, 26 Processor 24, 25 Control memory 28 RAM 29 RAM access control circuit

Claims (1)

(57) [Claims] 1. A record for each track constituting the same cylinder includes heads (1) to (3) that follow each track, respectively.
After controlling the positioning of the heads (1) to (3) for records on tracks of different cylinders, the heads (1) to (3) that follow each track are switched, respectively. When an error occurs in the record being processed and a retry is performed, the address of the record immediately before the record in which the error occurred is used as the retry address (19), and the record in which the error occurred is used. In the retry control method of the disk controller (13) for detecting, in the case where the last record of the last track of one cylinder is normally processed, the head positioning control is performed and the head is positioned in another cylinder, the flag is set. While setting (18), the retry address (19) is updated to the address of the head track of the other cylinder, and the head record of the head track of the other cylinder is updated. If an error occurs during sense, by the flag (18) is set, and the updated retry address (19), the head
After confirming that (1) to (3) match the address of the head track of the following cylinder, the heads (1) to (3) process the head record of the head track of the following cylinder, A retry control method for a disk controller, wherein the flag (18) is reset when the processing is normally performed.