JPH03260851A - Rewrite processing system for cache data - Google Patents

Abstract

PURPOSE:To reduce the adverse influence caused by the rewriting processing carried out to a disk access from a host device by providing a scheduling processing part on a disk controller to restore a cache control table and rewriting the data to be applied to a real device into the cache control table in accordance with a prescribed sequence. CONSTITUTION:A disk controller 2 is provided with a scheduling processing part 7 to restore a cache control table 5, and the data are rewritten into a real device 6 in accordance with the prescribed sequence of the table 5. That is, the data stored in a cache memory 3 are rearranged in the order of tracks of the device 6 via the part 7. Thus the table 5 is restored. Then a cache control part 4 rewrites the data into the device 6 in sequence by means of the restored table 5. Thus it is possible to reduce the adverse influence due to the rewriting processing carried out to a disk access from a host device.

Description

[Detailed Description of the Invention] [Table of Contents] Overview Industrial Application Fields Conventional Technology <Figure 3> Problems to be Solved by the Invention Means for Solving the Problems (Figure 1) Working Examples (Figure 2) 〉 Effects of the invention [Summary] Regarding the cache data write-back processing method, it is possible to quickly write back the cache data to the device after an unexpected power cut, and the write-back processing for disk access from the host device is performed. In order to reduce the negative effects, the disk control device is equipped with a cache memory consisting of non-volatile memory, a cache control unit, and a cache management table. In a cache data write-back processing method in which data is written back to the actual device, the disk controller is provided with a scheduling processing unit that rearranges the data in the cache memory in the order of the trunk and restores the cache management table, and the restored cache management Configure data to be written back to the real device according to table order.

[Industrial application field]

The present invention relates to a cache data write-back processing method,
More specifically, non-volatile memory is used as cache memory, and is used in magnetic disk control devices that can write to high-speed disks, and in particular, it is used to quickly write back cache data to the device after an unexpected power cut. This invention relates to a cache data write-back processing method.

[Conventional technology]

FIG. 3 is a block diagram of a conventional example, in which 1 is a host device (host computer, channel, etc.), 2 is a disk controller, 3 is a cache (cache memory), 4 is a cache control unit, and 5 is a cache Management table 6 indicates a disk device.

Conventionally, when performing high-speed writing to a disk in a magnetic disk device, for example, writing directly to the disk device 6 that ultimately retains and stores data written by the host device 1 has been limited by the physical limits of disk performance. Therefore, the disk control device 2 is provided with a cache memory 3 composed of a semiconductor memory whose access time is shorter than that of the disk device 6, so that the device mechanism operation is unnecessary when the target track is on the cache. thing,
Take advantage of the high speed of data transfer between the cache and the channel.

By focusing on the locality of access to external storage, the cache memory 3 can speed up access to the majority of all real devices with a smaller amount of storage that can be accessed at high speed.

In high-speed writing using the cache memory 3, when the cache control unit 4 returns an end status upon completion of writing data on the cache, the higher-level device 1 considers that the data has been written to the disk device 6.

After that, before the data on the cache memory 3 is written back to the disk device 6, the cache memory 3 used for caching the written data is backed up by a battery, etc., in case the power of the disk subsystem is cut off. Use a non-volatile one. Then, when the disk subsystem is powered on next time, the data in the cache memory 3, which is a nonvolatile memory, is written back to the disk device 6, which is a real device.

In order to perform the above processing, the disk control device 2 is provided with a cache management table (table on RAM) 5, which is managed by the cache control unit 4.

For example, the cache management table 5 stores a correspondence table between data on the disk device F6, which is a real device, and data on the cache memory 3, and this table is used to perform data writing, reading, etc. Perform processing.

[Problem to be solved by the invention]

The above-mentioned conventional devices had the following drawbacks.

(1) When the power is turned on again after being turned off,
It is necessary to write the data on the cache back to the magnetic disk device, which is the real device.

In this case, if the locality of access is extremely high, there may be a lot of data written back to one device, blocking host access.

(2) If the track positions of the data to be written back on the actual device are arranged in a different manner, for each trunk,
Since a seek operation is required, the time required for writing back becomes longer.

The present invention eliminates these conventional drawbacks, enables quick write-back processing of cache data to a device after an unexpected power cut, and eliminates the adverse effects of write-back processing on disk access from a host device. The aim is to reduce

In the write-back processing method for cache data written back to the magnetic disk 6, which is an actual device, the disk control unit f2 includes a scheduling processing unit that rearranges the data in the cache memory 3 in track order and restores the cache management table 5. 7 is provided, and data is written back to the real device according to the order of the restored cache management table 5.

[Means to solve the problem]

FIG. 1 is a diagram showing the principle of the present invention. In the figure, the same reference numerals as in FIG. 3 indicate the same parts, and 7 indicates a scaling processing section.

In order to achieve the above object, the present invention includes a cache memory 3 which is a non-volatile memory in the disk control device 2.
A cache control unit 4 that controls the cache memory 3
and a cache management table 5 storing data for managing the cache memory 3, and when the power is turned on next time after an unexpected power-off, the data on the cache memory 3 [Function] The present invention is as described above. Since it is configured as follows, it has the following effects.

After the power is turned off, the next time the power is turned on, the scheduling processing unit 7 first rearranges the data in the cache memory 3 in the track order of the disk device 6, which is a real device, and restores the cache management table 5.

Thereafter, the cache control unit 4 uses the restored cache management table 5 to sequentially write data back to the actual device.

In general, disks are faster at seeking to nearby cylinders than regular seeks. Therefore, if the data to be written back is rearranged in the trunk order, the time required for the seek operation is minimized, and the write back process can be executed quickly.

〔Example〕

Embodiments of the present invention will be described below based on the drawings.

FIG. 2 is a block diagram of one embodiment of the present invention, in which the same reference numerals as in FIG. 1 indicate the same parts. Also, 8 is a cache entry block, 3-113-2.3-3.3-
4, --- indicate the cache area corresponding to the trunk in the cache memory 3.

The disk control device 2 has a cache memory 3 and a cache control section 4, and the cache control section 4 has a cache management table 5 for managing the cache memory 3.

The cache memory 3 is a non-volatile memory (semiconductor memory) with battery backup, for example, and its interior is divided into memory areas 3-1.3-2.3-3. for each track of the disk device 6, which is an actual device. 3-4・
... is provided and data is written. A cache entry block 8 is also registered in the cache management table 5, and records which trunk of which device is written where on the cache memory 3.

For example, it is a device address, cylinder address, address on cache, amount of data, etc.

When the higher-level device l accesses the disk device 6, if the trunk is not on the cache memory 3, the trunk is read from the disk device 6 and sent to the higher-level word W1. At the same time, the data is stored in the cache memory 3 as user data, and an entry block 8 for that track is created in the management table 5.

The next time you access that track, cache memory 3
This is done on the above data.

Although the data on the cache memory 3 is rewritten in this way, the power may be cut off before this data is written back to the disk device 6. In this case, the cache control unit 4 checks the cache memory 3 when the power is turned on next time, and if any data remains, writes it back to the disk device 6.

When writing back such data, first, the scheduling processing section 7 restores the cache management table 5 based on the user data in the cache memory 3. This restoration restores entry block 8 for each device.
This sorts the data in the order of cylinder and track addresses.

The cache control unit 4 writes back data in the order of the entry blocks 8 of the rearranged and restored cache management table 5. As a result, seek operations in the disk device f6 are minimized and data can be written quickly. Reversion is achieved.

〔Effect of the invention〕

As explained above, the present invention has the following effects.

(11) After an unexpected power-off, the data remaining in the cache memory can be quickly written back to the actual device the next time the power is turned on.

) Eliminates the negative effects of write-back processing on disk access from the host device.

[Brief explanation of drawings]

FIG. 1 is a principle diagram of the present invention, FIG. 2 is a block diagram of an embodiment of the present invention, and FIG. 3 is a block diagram of a conventional example. 2--Disk control device 3--Cache memory 4--Cache control section 5-Cache management table 6--Disk device 7--Scheduling processing section

Claims (1)

[Claims] The disk control device (2) includes: a cache memory (3) made of nonvolatile memory; a cache control unit (4) that controls the cache memory (3); and a cache memory (3) that manages the cache memory (3). A cache management table (5) is provided that stores data to perform a cache management table (5), and after an unexpected power-off, the data in the cache memory (3) is written back to the actual device the next time the power is turned on. In this method, the disk control device (2) is provided with a scheduling processing unit (7) that rearranges the data in the cache memory (3) in track order and restores the cache management table (5), and the restored cache A cache data write-back processing method characterized in that data is written back to a real device according to the order of a management table (5).