JPH02158963A - Disk controller - Google Patents

Abstract

PURPOSE:To perform recovery at the time of loss of contents of a management table by adding information indicating a logical address and a frequency in change of its physical position to a part of data to be recorded. CONSTITUTION:Information related to the frequency in change of correspondence between the logical address and the physical address is additionally recorded in a management table 309, and this frequency is updated when assignment for write in the shortest time is different from that of the preceding correspondence as the result of the retrieval of a table by an address conversion block 307 at the time of data write. When the table will be reproduced, the physical address is successively accessed in all areas of a disk 310 to read out an additional information part, and the physical address for read and the read frequency in change are written in a position corresponding to the read-out logical address of a table indicating correspondence between the logical address and the physical address. This physical address is handled as 'busy' on a table B indicating the use condition of the physical address.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Field of Industrial Application) The present invention relates to a disk control device that controls a magnetic disk device, an optical disk device, etc.

(Prior art) The access time of a recording device that records/reproduces information on a disk-shaped recording medium such as a magnetic disk or an optical disk is determined by mechanical operation times such as seek time and rotational life time. It was difficult to increase the speed.

To overcome this problem, methods such as disk caches using semiconductor buffers have been adopted for reading data, and systems that dynamically allocate write locations to minimize the time required for writing. Ta.

(Problem to be Solved by the Invention) In order to dynamically allocate a write position in order to speed up the data write time, it is necessary to distinguish between the logical address recognized by the host device and the physical address where the data actually resides. Since the correspondence relationship is not constant, a table is required to manage the correspondence relationship and disk usage status. This table needs to be referenced and updated at high speed, and is usually developed on volatile semiconductor memory. If table contents are lost, logically meaningful data on the disk will no longer be accessible, so in order to avoid loss due to unexpected power outages, etc., this memory contains
It was necessary to take measures such as backing up with a separate power source.

[Structure of the invention]

(Means for resolving the order of assignments) When writing data, the present invention includes, in addition to the data requested to be recorded from the host device, a logical address recognized by the host device, and a combination of the logical address and the physical address. The present invention provides a disk control device that additionally records information regarding the number of times the correspondence relationship has been changed.

(Operation) If the contents of the latest management table are lost, the control device reads out the information added by the control device out of all the data areas on the disk in the order of physical addresses. From this part, the logical address designated by the host device when the data was written can be known, and the management table is restored based on this. However, there are cases where data corresponding to the same logical address is written to multiple physical addresses at different times, and the previously written (old) data remains as is. In this case,
Using information about the number of times the correspondence between logical addresses and physical addresses has been changed, which was added along with the logical address, select the physical address that is considered to be the latest, and select the physical address that is judged to have been written earlier. The address is treated as an empty area where invalid data is written.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings. Third
The figure is a block diagram of the present invention, showing the flow of data and control signals when writing data. Upper device 30
1 specifies the logical address (usually the start position and length) of the data to be written through the input/output control bus 30G. In the dynamic allocation method, based on this specification and the contents of the management table, the address translation block 307 determines the physical address on the disk at which writing can be completed in the shortest possible time. The input/output control block 305 accesses the disk device 310 by seeking and head switching to a position on the disk corresponding to the physical address specified by the control bus 308. FIG. 2 shows a portion of the format of a data track on a disk without the invention. The input/output control block 305 reads the header 101 on the track, finds the position where data is to be written, and writes the data 102 sent from the data bus 303. Even according to the present invention, the operation of the human output control block 305 remains unchanged. However, the address conversion block 307 sets the write data length to the length specified by the host device plus the length of additional information regarding the logical address and the number of changes, as the data length actually written. This length is used for searching the management table 309 and instructing the input/output control block 305 via the control bus 308. Address and translation block 307 also adds logical address and change count information to the data on data bus 303, which is sent to input/output control block 305 over bus 304. FIG. 1 shows part of the data format on a track in accordance with the present invention. header 1
01 is attached to each minimum unit of data recording on a disk called a sector, but a logical address may span multiple sectors.

To implement the present invention, as described above, the management table 3
Information regarding the number of times the correspondence between logical addresses and physical addresses has been changed is required in 09. This number is updated when the address translation block 307 searches the table during data writing and the allocation for writing in the shortest time is different from the previous correspondence.

Next, a table recovery method will be explained.

FIG. 4 shows an example of the format of the management table. Table A (401) shows the correspondence between logical addresses and physical addresses, and also stores the number of times the correspondence has been changed. Table B (402) shows the usage status of physical addresses. In order to reproduce the table, the control device sequentially accesses the physical addresses of all areas of the disk 310 and reads out the additional information portion according to this method. Even if the management table cannot be used, it is possible to read the contents recorded on the track as long as there is no abnormality in the format on the track. The control device writes the physical address from which the reading was performed and the number of changes that have been read into the table A at a position corresponding to the logical address from which the data was read. Also, on table B, it is treated as "in use." if,
If the physical address has already been written, the physical address that is determined to be the latest by comparing the number of changes is written in table A, and the other physical address is treated as "empty" on table B.

Furthermore, even if the logical address is an invalid value, it will be treated as "empty".In order to make this process effective, when initializing the disk device, the location where additional information will be recorded must be set to be invalid. It is necessary to write a logical address in advance.The number of changes cannot take an infinitely large value, so if the value is changed more than a certain number of times, the value will go through the cycle.However, it is necessary to set this number of times to a certain value. For example, the probability that old values that would lead to incorrect judgment of temporal context will remain without being overwritten can be made sufficiently low.

〔Effect of the invention〕

When using the disk device according to the present invention, even if a dynamic allocation method is used to write data to speed up access, in order to avoid loss of management table contents due to unexpected power outage, etc. There is no need to take special hardware measures such as preparing a separate power supply.

[Brief explanation of the drawing]

Figure 1 shows the data on the track when the present invention is implemented.
Figure 2 shows a part of the format of the data on the track without implementing the present invention.
FIG. 3 is a block diagram showing an embodiment of the present invention, and FIG. 4 is a diagram showing an example of a management table. 101... Header information on the track 102... Data from the higher-level device 103... One part of the track 104... Information to be added 301... Higher-level device 302... Control device 303... Data bus 304...Data bus 305...I/O control block 306...I/O control bus 307...Address conversion block 308...I/O control bus 309...Management table 310...Disk device 401...Table A 402...Table B Agent Patent Attorney Noriyuki Ken Yudo Matsuyama Development Figure 1 Figure 2

Claims (1)

[Claims] A disk control device that changes the correspondence between the requested address and the physical recording position on the disk in response to data transfer from a host device, and includes a part of the data to be recorded.
A disk control device characterized by adding information indicating a logical address and the number of times its physical location has been changed.