JPS63147219A - Semiconductor disk device - Google Patents

Abstract

PURPOSE:To surely secure the access chance to those accessess given from other system including a system of the host device having inferior performance, by inhibiting the accesses given from other system at a time point when an access-free state is detected. CONSTITUTION:The semiconductor disk control devices DIR0-DIRn count the answer frequencies of access request retrials against an access command via a count means 1 and receive information on an access-free state from an access control means 12 when the count value of the means 1 exceeds a prescribed level. Then an interruption control table 12b of the means 12 is read out and an access inhibiting means 2 inhibits the accesses to a relevant access cylinder from other systems at a detecting time point. In this case, the data on the cylinder contained in an access control table 12a is rewritten and a semicondutor memory 11 is locked.

Description

[Detailed Description of the Invention] [Summary] The present invention detects free access by counting a predetermined number of identical access requests in a semiconductor disk device in which cases in which access requests to the same data are likely to overlap among many types of systems occur. The present invention discloses a technique that allows systems with slow response speeds to access as reliably as other systems by inhibiting access from systems other than systems that do not permit access.

[Industrial application field]

The present invention relates to a semiconductor disk device that can be accessed from a plurality of systems, and more particularly to a semiconductor disk device that can be accessed as reliably even by a system having a higher-level device with inferior performance as other systems.

[Conventional technology]

In recent years, instead of mechanically accessed storage systems such as floppy disks and drive units, semiconductor disk systems in which banks of large-capacity semiconductor memory boards are switched using software are rapidly becoming popular. This system is, so to speak, an electronic disk system, and as a matter of course, compared to conventional mechanical access operations such as floppy disks, the data read/write and response speeds are much faster, and the recorded data is much faster. are frequently accessed. Therefore, there is a high possibility that the same data will be accessed from multiple systems. At this time, if there is a difference in the performance of the connected higher-level devices, access by the higher-level device with lower performance will often fail. Therefore, the higher-level device with lower performance will Despite that,
It becomes impossible to expect an increase in processing capacity. Therefore, there is a need for a control method that allows access to be performed equally regardless of differences in the performance of higher-level devices.

FIG. 4 is a block diagram showing the general configuration of a semiconductor disk device. In FIG. 4, the semiconductor disk device includes a semiconductor disk control device DIRO-DIR3 connected to each host device, a semiconductor memory 41 and its access control means 42, and each of the control devices DIR
Input/output boat I/FO~T/ that connects with O-DIR3
It consists of F3. The access control means 42 is configured as shown in FIG.
) and an interrupt control table 4 for each cylinder of the semiconductor memory 41 as shown in FIG. 5(b).
2b.

Here, for example, the first semiconductor disk control device DIRO
When an access command to the semiconductor memory 41 is issued from the host device, the control device DIRO searches the access control table 42a and checks whether the cylinder to be accessed has been accessed from another system. , if it is not in the access state, the table is rewritten and access from other systems is inhibited. This allows the semiconductor disk controller DIRO to access the target cylinder. If the target cylinder is in use,
In the interrupt control table 42b, the control device DIRO
and sends back the command retry status to its own higher-level device.

After the completion of Akmos, the semiconductor disk control device DIRO
The access control table 42a is rewritten to free access to that cylinder. At this time, the interrupt control table 42b is also read to check whether access registration has been made to that cylinder from another system. For example, if an access has come from the fourth semiconductor disk control device ffDIR3, the cylinder becomes free. By writing the notification information that has changed, the control device DIR3 is caused to issue an interrupt. When the semiconductor disk control device DIR3 that received the interrupt reads the interrupt control table 42b and detects notification information that the cylinder has become free,
It transmits the information to its host device, waits for the host device to issue an access command again, and when the access command is issued, accesses the semiconductor disk device.

[Problem that the invention seeks to solve]

In conventional technology, when multiple accesses to the same cylinder occur from multiple host systems, the access command retry status is reported to the host system that does not permit access, and the access cylinder becomes free. When it detects this, it issues an interrupt to the host device and waits for the access command to be issued again. At this time, if the performance of the higher-level device is inferior to that of the other system, even if an access command is issued again to access the semiconductor disk device, the other system's access will be faster and the command will have to be re-issued to the higher-level device again. The status of the trial will be returned, and the processing of the higher-level device with inferior performance will not proceed at all.

And what makes this doubly troublesome is that time passes while the command is being retried, and the upper layer ZCP simply
The problem is that the program is interrupted not as a result of a difference in the performance of U, but as a device error.

The present invention was devised in view of these problems, and provides a semiconductor disk device that reliably provides an opportunity for access from various systems, including systems of higher-level devices with inferior performance. The purpose is to

[Means for solving problems] FIG. 1 is a configuration diagram for explaining the principle of the present invention.

In the present invention, the means taken to solve the above problem is that when access requests from multiple access systems overlap, the access system returns an access request retry response to the access system that does not permit access. Control means 2
, a counting means 1 for counting the number of times the access request is retried to the same access request, and if the counted value exceeds a predetermined value, a counting means 1 for counting the number of times the access request is retried for the same access request; The present invention is a semiconductor disk device characterized by comprising a deterrent means 2 for inhibiting access from an access system.

[Effect]

A counting means 1 counts the number of command retry responses to the same access command, and if the counted value exceeds a predetermined value, it is determined that the system that issued the access command is a system with inferior performance. Therefore, when the inhibiting means 2 detects free access, it inhibits an access that overlaps with the above access command from another system so that the system that issued the above command retry can access.

〔Example〕

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a configuration diagram showing an example of a semiconductor disk device embodying the present invention. In FIG. 1, the semiconductor disk device includes a semiconductor disk control device DIRO=DIRn connected to each host device, and a semiconductor memory 11.
and its access control means 12, and an input/output boat I/O that connects them and each of the control devices fDIRo to DrRn.
It consists of FO to I/Fn. The access control means 12 includes:
As explained in FIG. 5, access control table 1
2a and an interrupt control table 12b.

Each of the control devices DIRO to DIRn also includes a counting means 1 for counting the number of access request retries (command retries) responses to the same access request, that is, an access command, and a counting means 1 for counting the number of responses to an access request retry (command retry) for the same access request, that is, an access command, and a counter 1 for counting the number of responses to an access request retry (command retry) for the same access request, that is, an access command, and when the counted value exceeds a predetermined value, an access The access cylinder is notified by the access control means 12 and is provided with a suppression means 2 that suppresses access from other systems to the access cylinder at the time when it is detected by reading the interrupt control table 12b.

FIG. 2 is a flowchart showing an example of access processing by the above device, and FIG. 3 is a flowchart showing an example of interrupt processing.

In FIG. 2, the semiconductor disk control device reads the access control table 12a to check whether the cylinder to be accessed is being used by another system (busy). If it is not busy, normal access processing continues. When a cylinder is used, in addition to the normal processing, the counting means 1 in the control device records the access cylinder at the first access, and from the second time onwards, the counting means 1 in the control device records the access cylinder in response to the access command to the cylinder from the second time onwards. Increments the command retry response count. Interrupt control table changes and status responses are the same as before.

After the above processing is completed, the semiconductor disk control device waits for a cylinder free interrupt from the semiconductor disk device.

In FIG. 3, the semiconductor disk control device reads and checks the interrupt control table 12b, and when a cylinder-free interrupt is confirmed, checks whether the number of command retry responses for that cylinder exceeds a predetermined value. , if it exceeds the access control table 12a, the processing means 2 reads the access control table 12a, confirms that there is no access to that cylinder from another system, writes the access cylinder designation to the access control table 12a, and Change to prevent access to the cylinder from other systems. At the same time, it raises a cylinder free interrupt to the host device and waits for an access command from the host device. At this time, the data related to that cylinder in the access control table 12a has been rewritten as if it were in use, so 11 is in a locked state, and even if the performance of the host device is poor and processing time is required, access from other systems will not be accepted.

It should be noted that the present invention is not limited to the above-mentioned embodiments, and can be applied in various ways and implemented in accordance with the gist of the present invention. The number of host device systems can be two or more.

〔Effect of the invention〕

As described above, according to the present invention, access from other systems is blocked when free access is detected so that access can be granted to a system that has issued a predetermined number of consecutive access request retries. By inhibiting access from each unit, including systems of higher-level devices with inferior performance,
A reliably accessible semiconductor disk device can be provided.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram of an embodiment of the present invention, FIGS. 2 and 3 are flowcharts of its processing operation, FIG. 4 is a general configuration diagram of a semiconductor disk device, and FIG. 5 is an explanatory diagram of a control table. It is. In the figure, 1; counting means; 2; inhibiting means; 11.41; semiconductor memory; 12.422; access control; 12a, 42a; access control table; 12b, 42
b. Interrupt control table, DIRO-DIR3; semiconductor disk control device. Structure of N, Z'l a four times fJ and implementation order j of the present invention Fig. 1 Access processing Fig. 2

Claims (1)

[Claims] When access requests from multiple access systems overlap,
an access control means (12) for returning an access request retry response to the access system that does not permit access; a counting means (1) for counting the number of access request retry responses to the same access request; and a counting value thereof. A semiconductor disk device characterized by comprising: (2) inhibiting means (2) for inhibiting access from access systems other than those that issued the same access request at the time when free access is detected when the value exceeds a predetermined value .