JPH01304553A - Device for reading out data from storage device - Google Patents

Abstract

PURPOSE:To suppress the increase of software overhead in the data storing work by providing a circuit which converts a group identifier and a unit length data identifier of each unit length data to an address of a buffer memory on an auxiliary storage controller. CONSTITUTION:A magnetic disk controller 2 is provided between a host computer 1 and a magnetic disk 3. Port registers 21 and 25, a directory 22, a disk cache 23, cache address information 24, and a CPU 26 are provided in this magnetic disk controller 2. When the request track number of the data read instruction from the host computer 1 exists on address information 24, the directory 22 obtains the start memory address of request data by a sector number-memory address converting circuit 222 and reads out data from the disk cache 23 and transfers it to the host computer 1.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a device for reading data from a storage device, and in particular, the present invention relates to a device for reading data from a storage device. The present invention relates to a data reading device suitable for reading data from a storage device by an auxiliary storage control device configured to simultaneously read information and store it in a buffer memory.

[Conventional technology]

There are various examples of conventional magnetic disk devices or magnetic disk control devices using disk caches. In particular, with large disk drives, there is an example in which the data requested to be read from the host computer is read along with the data before and after it (for example, all data on the same track) and stored in the disk cache. There are many.

A method for high-speed reading of multiple sectors from a magnetic disk device is found in Japanese Patent Laid-Open No. 58-163059 or Japanese Patent Laid-Open No. 62-54885. In these examples, the magnetic disk controller first analyzes a data transfer request sent from the host computer, determines which sector the beginning is in, and which sector of the track containing the leading data should be read. After the calculation, a subsequent head seek is performed, and the read head is set on the corresponding track.
When the head is located on the sector to be read, reading is started regardless of whether it is the first sector or not. 6 Furthermore, in Japanese Patent Laid-Open No. 62-54885, All read data is once stored in memory and then transferred to the host computer, but when data is stored in memory, the address bus is designated by the CPU.

[Problems to be Solved by the Invention] When the above-mentioned conventional technology is applied to data transfer from a magnetic disk device to a disk cache, calculation of sectors to be read to the disk cache, processing of storing read sectors to the disk cache, etc. This directly leads to software overhead.

An object of the present invention is to provide a data reading device that can suppress an increase in software overhead caused by data storage from an auxiliary storage device such as a magnetic disk to a disk cache.

[Means to solve the problem]

The present invention provides an auxiliary storage device that stores data by grouping unit length data and assigning an identifier for each group to each unit length data, and requesting the auxiliary storage device to read data. A device for reading data from a storage device, provided in an information processing device including a host computer and an auxiliary storage control device connected between the two and having a built-in buffer memory.

The above object is a configuration in which a circuit is provided for converting a group identifier and a unit length data identifier attached to each unit length data of data read from the auxiliary storage device into an address of the buffer memory on the auxiliary storage control device. This is achieved by

Further, the present invention preferably provides a data identifier on the auxiliary storage device requested by the host computer;
The configuration includes a comparison circuit that compares the data identifier of the data read from the auxiliary storage device and outputs a match signal when the two match.

That is, the present invention provides a means for determining to which group of data groups on a storage device the first data requested to be read from a host belongs, and a means for associating a group number and a unit length data identification number with a memory. identification number -
and means for storing successive data in the memory using a memory address conversion circuit. Furthermore, the present invention continues to store the data in the memory at the point when the beginning of the data requested by the host starts to be read from the medium, and also sends a data transfer device to the host or a data transfer device to notify that the data has started to be read. It is preferable to provide the above-mentioned comparison circuit as means for notifying the controlling CPU.

There are various types of auxiliary storage devices to which the present invention can be applied, including a magnetic disk device. Further, a disk cache is used as the buffer memory. The reading device of the present invention is provided in an auxiliary storage control device. In the case of a magnetic disk device, it is provided in a magnetic disk controller.

[For production]

When arbitrary data is requested, the auxiliary storage control device first determines which group to read by means of knowing the group recognition information of the group to which the requested data belongs. , the auxiliary storage controller can begin moving the read head to the location where the group is being recorded.

When the head arrives at the target location, when the group recognition information attached to the unit length data matches the group LZTIt information of the previously requested data from the host, it immediately starts reading that data. The means for starting causes the auxiliary storage controller to start reading data.

After starting reading, the identification number/memory address conversion circuit designates an address for storing data in the memory, so that the CPU is released from control for storing data in the memory. When the auxiliary storage control device starts reading requested data from the host, it continues to store data in the memory device and outputs a signal to notify the data transfer device to the host or the CPU that controls it.

By using this signal, the CPU does not need to constantly monitor whether the read data is the data requested by the host. Therefore, while data is being stored from the storage device to the memory, the CPU does not need to keep monitoring whether the transferred data is data to be sent to the host.

Therefore, the time used by the CPU for data transfer processing from the storage device to the memory can be significantly reduced.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 1 to 6.

FIG. 1 shows a block diagram of the configuration of a magnetic disk controller that is an embodiment of the present invention.

As shown in the figure, a magnetic disk controller 1-roller 2 of this embodiment is provided between a host computer 1 and a magnetic disk 3. This magnetic disk controller 2 is provided with port registers 21 and 25, a directory 22 which is the essential part of the present invention, a disk cache 23, cache address information 24, and a CPU 26.

The magnetic disk controller 2 is connected to the host computer 1
When a command to read certain data comes from port register 2,
1, and passes the read command to the directory 22. The directory 22 first uses an address calculation circuit 221 to calculate which group the requested data belongs to. Here, a group refers to a block of a fixed data unit that is input to and output from the disk cache 23.

As an example, the case where one track's worth of data is made into one group will be described above.

Next, the directory 22 checks the address information 2 of the disk cache 23 to determine whether the data on the track where the requested data exists is on the disk cache 23.
Search for 4. As shown in FIG. 2, the cache address information 24 includes the cylinder number of the data stored in the first slot of the cache memory following the address information,
Contains the track number. The disk cache described here is a large-capacity semiconductor memory, which is divided into one slot with the capacity of one disk drive track.

If the requested track number is on the cache address information 24, the directory 22 determines the starting memory address of the requested data using the sector number-memory address conversion circuit 222, reads the data from the disk cache 23, and transfers the data. The data is transferred to the host computer 1 through the device 223 and the port register 21.

In this way, according to this embodiment, the host computer 1
The CPU does not need to perform calculations to convert the sector number of the sector containing the data sent from the CPU into the address of the disk cache 23.

An example of this sector number-memory address conversion will be explained using FIG. 2.

First, the CPU 26 of the magnetic disk controller 2 determines whether the slot m into which data is to be loaded or unloaded is identified by the address information 2.
It is specified by the cylinder number and track number in 4. The disk cache 23 is divided into the number of sectors in one track, and the sector number corresponds to the start address of the memory where the data in that sector is stored.

The sector number is the sector number-memory address conversion circuit 22
2, the conversion circuit 222 specifies the address of the disk cache 23 so that data is to be read/written from the memory top number corresponding to that sector number.

When the data to be taken in and out of the disk cache 23 spans a plurality of consecutive sectors, the first sector number and the number of sectors can be set in the sector number-address calculation circuit. When set in this way, the sector number-address conversion circuit 222 first specifies the address corresponding to the first sector number, decrements the sector number counter every time one sector transfer is completed, and when it reaches 0, transfers. end.

When the data on the track where the requested data exists is not on the disk cache 23, the directory 22 accesses the magnetic disk device 3 where the requested data exists through the boat register 25. First, the directory 22 specifies to the magnetic disk device 3 the track number on the disk where the requested data exists.

The disk device 3, which has been designated with a track number, first seeks the head to the designated track position. In a fixed sector type disk device, address information as shown in FIG. 3 is recorded at the beginning of each sector.

After the seek is completed, the disk device 3 reads the data,
Immediately after confirming that the sought track is the target track, data begins to be read from the medium regardless of the sector number.

The read data is transferred to the magnetic disk controller 2. The data is taken into the magnetic disk controller 2 from the port register 25 shown in FIG.
It is sent to the directory 22.

Therefore, the sector address is converted by the sector number-address conversion circuit 222, and the start address for storing the data of the sector is specified. Thereafter, the sector number-memory address conversion circuit 222 sequentially increases the address until data corresponding to one sector byte has been transferred to the disk cache.

After transferring one sector, the directory 22 pauses data transfer and waits until the next sector number is sent, and at the time the next sector number is sent, the same operation is repeated.

The sector number/memory address conversion circuit 222 has a sector counter, and reports the completion of transfer to the CPU 26 when the number of sectors for one track has been read out.

In this way, in the operation of transferring one track from the disk device to the disk cache, there is no need to check the sector number, and the exclusive time of the control device that controls the disk cache is reduced.

Figure 4 shows this situation. 1A shows the CPU exclusive state in the above-described embodiment of the present invention, and FIG. 2B shows the CPU exclusive state in the conventional example. It can be seen that in the conventional example, the zt calculation of the sector to be read is performed before the start of the seek, which causes a delay in the start of the seek. Further, it can be seen that an interrupt is generated for confirming the sector ID even during data transfer, and the time period in which the microprocessor of the magnetic disk control device is exclusively used becomes longer.

The address calculation circuit 221 calculates the track number and sets it in a circuit that issues a seek command, then calculates the first sector number and sets it in the sector number comparison circuit 224. The comparison circuit 224 compares the sector number with the set sector number each time the sector number is sent. If the comparison results do not match, only transfer to the disk cache is performed. If they match, the data is transferred to the disk cache 23 and transferred to the transfer device 223 or the CPU.
I'll let you know on the 26th.

When the host computer 1 is free, the controller 2 starts data transfer to the host computer 1 at that point. Data transfer to disk cache 23 continues in parallel.

Furthermore, in the above example, the transfer between the disk cache 23 and the host computer 1, and the transfer between the disk cache 23 and the host computer 1
The data transfer between the computer and the magnetic disk device 3 is performed by one data transfer device 223, but the disk cache has two ports, and each port is equipped with a transfer device for the magnetic disk and a transfer device for the host computer. Good too.

When host computer 1 is not available, the disk
Only the data is transferred to the cache 2, and when the host computer 1 is free, the disk cache 23 is transferred.
Data is transferred from the host computer 1 to the host computer 1.

The above-described sector number comparison operation and sector number recognition operation may be activated by a sector detection signal from the disk device. This is the same for both soft sectors and hard sectors.

When the address calculation circuit 221 sets the first sector number in the comparison circuit 224, it also sets the number of sectors to be transferred at the same time. After the comparison circuit 224 starts sending the first sector address to the host side, each time it sends a sector one by one, it decrements by one the number stored in the register in which the number of sectors to be transferred is set. , the data transfer to the host computer 1 ends when the number in the register becomes 0.

If the transfer does not end when you reach the end of the track,
Therefore, data transfer to host computer 1 is interrupted,
After seeking to the next track and reading data from the designated sector there, data transfer to the host computer 1 is resumed.

At the same time that reading becomes possible, the disk device 3 starts reading data, and the disk controller 2 starts writing the data sent from the disk device 3 to the disk cache 23.

When the transfer start sector to the host computer 1 is sent during data reading, the disk controller 2 starts the transfer to the host computer 1. The situation is shown in FIG.

The magnetic disk control device described in Japanese Patent Application Laid-Open No. 62-54885 stores all the data to be read on the track in the buffer memory before starting data transfer to the host computer, as shown in Fig. 6. become.

Compared to the above embodiment, the data transfer start time is delayed by a time of T□+T2.

However, the dedicated time of the bus between the host system 1 and the disk controller 2 is shorter in Japanese Patent Application Laid-Open No. 62-54885, which is advantageous compared to the above embodiment. This is particularly a problem when the right to use a channel between the host and the disk controller determines the rate of data transfer. In that case, the method used in Japanese Unexamined Patent Publication No. 62-54885, ie, the method of transferring data after all data has been stored in the memory, may be implemented.

Furthermore, in the above embodiment, the trunk number was used to identify the group, but this need not be the track number;
An area may be provided to write the cylinder number or new group recognition information, and the area may be written there.

Further, in the above embodiment, a disk device was described, but the present invention can also be applied to, for example, a magnetic tape device. It is possible to reduce the processor usage time for read processing by using each sector address of the magnetic tape or by newly providing group recognition information in the ID section and performing the same operation as in the above embodiment. Similarly, application to optical disks, magnetic bubble memories, etc. is also possible.

〔Effect of the invention〕

According to the present invention, when transferring data from a storage device to a buffer memory, there is no need for the CPU to read a sector ID and judge whether or not to store that sector in the cache, thereby reducing the exclusive time of the CPU. be able to. Also.

When transferring data to the host 1-system, the C
Since there will be no interrupts to the PU, the C
This has the effect of shortening the exclusive time of the PU.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall configuration of an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of a disk cache, and FIG. 3 is an ID ID generally used in disk devices.
An explanatory diagram showing an overview of information, Fig. 4 is an explanatory diagram showing the CPU exclusive time when the present invention is implemented and when it is not, and Fig. 5 is an explanatory diagram showing a method of transferring data to the cache and the host computer at the same time. FIG. 6 is an explanatory diagram showing how data is transferred when the method disclosed in Japanese Unexamined Patent Publication No. 62-54885 is used. DESCRIPTION OF SYMBOLS 1... Host computer, 2... Magnetic disk controller, 3... Magnetic disk device, 22... Directory, 221... Track, sector calculation circuit,
222...Sector number-address conversion circuit, 223...
・Data transfer device, 224... sector number comparison circuit,
23...Disk cache, 24...Cache address information.

Claims (1)

[Claims] 1. An auxiliary storage device that stores data by grouping unit length data and assigning an identifier for each group to each unit length data; and a method for storing data in the auxiliary storage device. A device for reading data from a storage device, which is provided in an information processing device comprising a host computer requesting reading, and an auxiliary storage control device connected between the two and having a built-in buffer memory. , a circuit is provided for converting a group identifier and a unit length data identifier attached to each unit length data of data read from the auxiliary storage device into an address of the buffer memory on the auxiliary storage control device. A device for reading data from a storage device. 2. A comparison that compares the data identifier on the auxiliary storage device requested by the host computer and the data identifier of the data read from the auxiliary storage device, and outputs a match signal when the two match. An apparatus for reading data from a storage device according to claim 1, comprising a circuit.