JPH0816864B2 - Magnetic disk processing device - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic disk processing device which, when a correctable read error is detected, corrects the error and transfers it to a higher-order device, and more particularly to a correctable read error. The present invention relates to a magnetic disk processing device that performs a rewriting process of corrected data when an error is detected.

[Conventional technology]

When a correctable read error is detected, the conventional magnetic disk processing device corrects the data in the internal memory by the read error detection and correction code (ECC) process, and then performs the command retry process on the magnetic disk device (DKU). The magnetic head of is repositioned to the record in which a correctable read error has occurred again, and in synchronization with the operation of reading this, the data read from the magnetic disk unit (DKU) is transferred as if it is being transferred. Has performed a process of transferring the corrected data to the host device.

[Problems to be Solved by the Invention]

In this conventional magnetic disk processing device, the correctable read error (hereinafter referred to as read error) of the magnetic disk device (DKU) is caused by minute scratches present on the disk medium surface or axial misalignment due to aging of bearings, etc. However, once a read error caused by these occurs, there is a high possibility that a read error will occur again in the same record thereafter. Therefore, once a correctable read error occurs, each time the same record is processed thereafter, a command retry process causes a rotation loss of at least one rotation, which results in processing performance, and the input / output processing performance of the entire system. Has a significant impact on. Further, in the case of a correctable read error due to the latter cause, if this is left as it is, the record becomes an uncorrectable read error in the near future due to the progress of axis misalignment and the data is lost. I got it.

[Means for solving the problem]

A magnetic disk processing device according to the present invention includes a firmware for controlling the magnetic disk processing device and the magnetic disk device and a storage storing the firmware, an input / output processing device or a main storage device based on a command from a host device. Data transfer control unit for controlling the data transfer of the magnetic disk device, a second data transfer control unit for controlling data transfer with the magnetic disk device in synchronization with a clock signal input from the magnetic disk device, and a read error detection correction code ( A processing unit for generating and decoding ECC), a storage unit for temporarily holding the data read from the magnetic disk device, and two pieces of output data of the first data transfer control unit and the output data of the storage unit as data to be transferred to the magnetic disk device. First to select and output from users
A selecting section, a second selecting section for selecting and outputting the output data of the second data transfer control section and the output data of the storage section as data to be transferred to the input / output processing device or the main storage device, and firmware. When the magnetic disk processing device and the magnetic disk device are controlled by executing the processing, and the processing unit decodes the read error detection and correction code and determines that there is a read error, the contents stored in the storage unit based on the correction information obtained from this processing unit. Data is corrected so that the first selection unit selects the output data from the storage unit, the second selection unit also selects the output data from the storage unit, and the corrected data stored in the storage unit is A microprocessor for sending the data to the magnetic disk device via the first selecting unit and the second data transfer control unit, and at the same time sending to the host device via the second selecting unit and the first data transfer control unit; It has.

[Action]

According to the present invention, the correction data can be written to the magnetic disk device at the same time when the correction data is transferred to the host device.

〔Example〕

FIG. 1 is a configuration diagram of a system in which an embodiment of a magnetic disk processing device according to the present invention is connected between a host device and a magnetic disk device (DKU). In the figure, 1 is a control storage (CS) storing firmware, 2
Is a microinstruction bus, and 3 is stored in the control storage (CS) 1. The firmware for controlling the magnetic disk processing unit (MSP) and the magnetic disk unit (DKU) is read and executed via the microinstruction bus 2. A microprocessor 4 and an input / output processor (IO
P) or the first data transfer control unit 5 for controlling data transfer with the main memory unit (MM) is a magnetic disk unit (DKU)
As the data to be transferred to the I / O device (IOP) or the main memory (MM), the data transferred via the first data transfer control unit 4 and the output data of the following memory unit are selected and output. The first selection unit 6 is data transferred via the second data transfer control unit described below as data to be transferred to the input / output processing device (IOP) described below or the main storage device (MM) described below, and the storage unit described below. The second selection unit 7 for selecting and outputting from the two output data of the above temporarily holds the data sent from the following magnetic disk unit (DKU) through the following second data transfer control unit, and A storage unit capable of correcting data including a read error by the microprocessor 3 according to the correction information obtained from the ECC processing unit, and 8 indicates a read error detection by the data transferred to the magnetic disk unit (DKU) described below. An output correction code (ECC) is generated, and the presence or absence of a read error is detected and correction information is generated from the data and read error detection correction code (ECC) transferred from the magnetic disk unit (DKU) described below. A read error detection / correction code processing unit (hereinafter referred to as an ECC processing unit) for notifying 3 and data 9 for the magnetic disk unit (DKU) and a read error detection / correction code (E
A second data transfer control unit for controlling the transfer of (CC) in synchronization with the clock signal input from the magnetic disk unit (DKU),
10 is a processor data bus, 11 is a read error detection notification signal line, 12 is a write data transfer path, 13 is a read data transfer path, 14 is an output data transfer path, 15 is a write data transfer path, and 16 is a read error detection correction. A code transfer path (hereinafter referred to as an ECC transfer path), 17 is a read data transfer path, and 18 is the control storage (CS) 1 described above, the microprocessor 3, the first
Data transfer control unit 4, first selection unit 5, second selection unit 6, storage unit 7,
The magnetic disk processing device (MSP) includes an ECC processing unit 8 and a second data transfer control unit 9.

19 is a central processing unit (CPU), 20 is a main memory (M
M), 21 is a system bus, 22 is an input / output processing unit (IOP) under the central processing unit (CPU) 19, 23 is a bidirectional data transfer path, and 24 is the central processing unit (CPU) 19 and storage device ( MM)
Host device consisting of 20 and input / output device (IOP) 22, 2
5 is a magnetic disk unit (DKU), 26 is a bidirectional data transfer path, and 27 is a device clock signal line.

Next, the operation of the magnetic disk processing device having the above configuration will be described. First, the microprocessor 3 reads the firmware stored in the control storage (CS) 1 through the microinstruction bus 2 and executes it, thereby controlling the magnetic disk processing device (MSP) 18.
When the magnetic disk processing device (MSP) 18 receives a data read command of the magnetic disk device (DKU) 25 from the input / output processing device (IOP) 22, the second selection unit 6 is controlled by the microprocessor 3. The data on the read data transfer path 17 is selected and output to the first data transfer control unit 4 via the read data transfer path 13. The second data transfer control unit 9 also reads the data and the read error detection / correction code (ECC) input via the bidirectional data transfer path 26 in synchronization with the clock signal input via the device clock signal line 27. Output to the data transfer path 17. At the same time, the storage unit 7 writes only the above data. Further, the ECC processing unit 8 determines the presence / absence of a read error from the data and the read error detection / correction code (ECC), and notifies the microprocessor 3 of the result through the read error detection notification signal line 11.
Further, the ECC processing section 8 holds the correction information and prepares for reading from the microprocessor 3. On the other hand, the first data transfer control unit 4 outputs the data output from the second selection unit 6 via the bidirectional data transfer path 23 to the input / output processing device (IOP) 2
Send to 2. Next, the magnetic disk processing unit (MSP) 18 goes from the input / output processing unit (IOP) 22 to the magnetic disk unit (DKU).
When receiving a data write command to 25, the first data transfer control unit 4 writes the received data from the input / output processing unit (IOP) 22 via the bidirectional data transfer path 23 under the control of the microprocessor 3. It is sent to the data transfer path 12. Then, the first selection unit 5 selects the data input via the write data transfer path 12 and outputs it to the second data transfer control unit 9 via the write data transfer path 15. Then, the second data transfer control unit 9 transfers the write data and the read error detection and correction code (ECC) subsequently to the clock signal to the ECC processing unit 8 through the bidirectional data transfer path 26 in synchronization with the clock signal. Then, the ECC processing unit 8 performs the ECC generation processing in synchronization with the data writing to the magnetic disk device (DKU) 25, and after writing all the data, the generated read error detection and correction code (ECC) is transferred to the ECC transfer path 16. It is sent to the second data transfer control unit 9 via the.

Next, the process at the time of detecting a correctable read error will be described. The microprocessor 3 knows from the notification from the ECC processing unit 8 that a correctable read error has been detected, and the ECC process is performed in order to retry the command. Based on the correction information read from the unit 8, the data held in the storage unit 7 via the processor data bus 10 is corrected. at the same time,
In parallel with this, repositioning of the magnetic head to the record is performed to prepare for data transfer start. When the preparation for data transfer is completed by both of the above-mentioned processes, the microprocessor 3 sets both the first selection unit 5 and the second selection unit 6 so as to select the output data transfer path 14 of the storage unit 7, and 2 The write operation is instructed to the data transfer control section 9, the input / output operation to the input / output processing unit (IOP) 22 is instructed to the first data transfer control section 4, and the corrected data in the storage section 7 is magnetically recorded. Disk unit (DKU) 25 and input / output processor (IOP) 22
Can be transferred to.

〔The invention's effect〕

As described above, according to the magnetic disk processing device of the present invention, it is possible to write the corrected data to the magnetic disk device at the same time when the correction data is transferred to the higher-level device. It is possible to reduce the probability that the error will occur and to prevent the change from the correctable read error to the uncorrectable read error due to the progress of axis misalignment, improving the input / output performance and data reliability of the magnetic disk processing device and the entire system. There is an effect that the improvement of can be realized.

[Brief description of drawings]

FIG. 1 is a block diagram of a system in which an embodiment of a magnetic disk processing device according to the present invention is connected between a host device and a magnetic disk device. 1 ... Control storage (CS), 2 ... Microinstruction bus, 3 ... Microprocessor, 4 ... First data transfer control section, 5 ... First selection section, 6 ... Second selection section,
7 ... storage unit, 8 ... read error detection / correction code processing unit (ECC processing unit), 9 ... second data transfer control unit, 10 ...
Processor data bus, 11 ... Read error detection notification signal line, 12 ... Write data transfer path, 13 ... Read data transfer path, 14 ... Output data transfer path, 15 ... Write data transfer path, 16 ... Read error detection and correction code transfer path, 17 ……
Read data transfer path, 18 ... Magnetic disk processing device, 19
...... Central processing unit (CPU), 20 ...... Main memory (MM), 2
1 ... system bus, 22 ... input / output processor (IOP), 23
...... Bidirectional data transfer path, 24 ...... Host device, 25 ...... Magnetic disk unit (DKU), 26 ...... Bidirectional data transfer path, 27
...... Device clock signal line.

Claims (1)

[Claims] 1. A central processing unit or an input / output processing unit under the central processing unit and a magnetic disk unit,
In the magnetic disk processing device that controls the data transfer between the input / output processing device or the main storage device and the magnetic disk device directly from the central processing device or by a command via the input / output processing device, A magnetic disk processing device, firmware for controlling the magnetic disk device and a storage for storing the firmware, and a first data transfer control unit for controlling data transfer with the input / output processing device or the main storage device, A second data transfer control unit for controlling data transfer with the magnetic disk device in synchronization with a clock signal input from the magnetic disk device; a processing unit for generating and decoding a read error detection / correction code; and reading from the magnetic disk device. Storage unit for temporarily holding the stored data and the first data transfer as data to be transferred to the magnetic disk device A first selector for selecting and outputting the 2's output data of the storage unit of the control unit, the as data to be transferred to the output processor or the main memory 2
A second selection unit that selects and outputs the output data from the data transfer control unit and the output data from the storage unit, and executes the firmware to control the magnetic disk processing device and the magnetic disk device. When the processing unit decodes the read error detection and correction code and determines that there is a read error, the data in the storage unit is corrected based on the correction information obtained from this processing unit, and the first selection unit stores the stored data. Output data from the storage unit, the second selection unit also selects the output data from the storage unit, and the corrected data stored in the storage unit is used as the first selection unit and the second data. A microprocessor for sending the data to the magnetic disk device via a transfer control unit and simultaneously sending the data to the host device via the second selecting unit and the first data transfer control unit. Magnetic disk apparatus according to claim and.