JPH01173137A - Queue control system - Google Patents

Abstract

PURPOSE:To shorten the processing time of an interruption routine by using a means which sets an unused queue after a working queue. CONSTITUTION:A single chain 23' is composed by setting an unused queue after a working queue. An event queue control table 24' has such a constitution that controls the head address of the working queue, the head address of the unused queue, and the address of the final queue. When a new event occurs, it is just required for an interruption routine 20' to write the necessary data at the head of the unused queue and to replace the head queue address contained in the table 24'. Thus the reconnection is omitted among queue blocks, that is, the replacement of the NEXT parts is not required. As a result, the chain 23' is never broken even though the routine 20' works during the working of an event information processing routine 22'.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention stores event information generated by an interrupt function in a queue format. The present invention relates to a queue management method in a data processing system having a mechanism for extracting event information from a queue when the conditions for processing the event information are satisfied later.

2. Description of the Related Art Combiator graphics (that is, graphics processing by a computer) has become popular with recent advances in computer system hardware technology and utilization technology.

In the graphic processing, the operator inputs new event data (for example, oimy data on the graphic display,
The input of instructions (for graphic processing) and the graphic processing using the event data are performed asynchronously (that is, not interactively), and the speed of the graphic processing is increased. Ru.

Therefore, the input event data is sequentially stored in the queue block, and asynchronously with the input of the event data, when the conditions for processing the event data are ready,
The event data is sequentially read out and graphic processing is performed according to the funxic code.

Therefore, in order to efficiently perform the graphic processing, the event data must be stored in the queue block.

It is necessary to effectively carry out the retrieval process of event data stored in the key a, which occurs asynchronously with the storage process.

[Conventional technology]

FIG. 114 is a diagram illustrating a conventional graphic processing method,
FIGS. 5 1w and 6 are diagrams explaining the conventional junior high school 2nd management system.

First, in FIG. 114, the operator operates the input device 4 such as a tablet, keyboard, function switch, etc. while looking at the graphic display 3 to input new event data (X) for graphic processing.

When the Y coordinate 1 graphic processing function (Ao, etc.) is input, an interrupt is generated to the central processing unit (CPU) 1.

In the central processing unit (CPU) 1, an interrupt routine 20 analyzes the cause of the interrupt and activates the input/output supervisor (IO3) 21t.

The input/output supervisor (IO3) 21 is an input device 1l1.
Sense the data from 4 and execute ill input routine 2.
Pass to 0.

In the interrupt routine 20, when keying the asynchronously occurring event data to the Φ event block 23, the following procedure is followed. (See Figure 5) 11) Store the event information that has occurred in the first queue block (■) of the unused queue.

(2) Change the address of the unused keys in the event management table to the address of the queue block ■ pointed to by the NEXT officer in the block ■ due to one key.

(3) To use the key - select the block chain, search for the last cube block, and select NEXT? of the corresponding cube block ■. The address of l? Therefore, change the address to block ■.

Event information that occurs asynchronously due to an interrupt is connected to the AK usage queue based on the above procedure.

On the other hand, when the conditions for processing an event are met, the event information processing routine extracts the event information that becomes necessary, performs various processing, and connects unnecessary queue blocks to unused queues. Follow. Please refer to Figure WJ6.

(4) Find the chain te of the used cue block and search for the cue block {circle around (2)} that corresponds to the search condition.

(5) The corresponding key - the one before the block ■? For
BUa, Ri■'s NEXT? For the address of Φ,
Ri■'s NEXT? The hour A Tsuk that points to ■
Change the address to .

(6) Perform various processing on the corresponding queue block (■).

(7) After various processes are completed, event information is deleted from + and - books (2) that are no longer needed.

(8) Search for the unused 'JP Yub C1-/RI's chain t-Kudori, the last upper 2-block, and find the corresponding key 1-
Bua, Ri■'s NEXT? Change the address of to the address of the queue block ■.

Event information can be used by following the above steps! ?
Annie and others are removed and connected to an unused key.

[Problems to be solved]

In the above queue management method, the used keys & 1 unused key are used for interrupt routines and event information processing routines.
Although it is operated by a program, the interrupt routine has a higher execution priority than the event information processing routine, so even if the event information processing routine is running, the event information processing routine will process the event information every time event information occurs. is temporarily interrupted and control is passed to the interrupt routine. Under such an environment, the queue status of used queue 1 and unused keys is not guaranteed (if the interrupt routine operates while the event information processing routine is reconnecting the queues,
chain is destroyed). For this reason, conventionally, a lock instruction has been issued to the queue block to be used, and exclusive control has been performed between the interrupt routine and the event information processing routine. Interrupt routines connect event information that occurs asynchronously to a queue, so high-speed processing is required.
In the exclusive control method described above, the interrupt routine is sometimes forced to wait until Jb when the processing of the event information processing routine ends, resulting in a lack of high-speed processing.

[Means for solving problems]

The first @ is a diagram showing a configuration example of a queue management method using a single chain according to the present invention.

In the present invention, a single chain 23' is constructed by connecting an unused key 1lk- to the rear car of the used queue, and the event queue management table 24' contains the used Φ-gon head address, the unused queue head address, and the final queue address.

[For production]

As a result, when a new event occurs, the interrupt routine 2G' only needs to write the necessary data to the head of the unused queue and update the head key 1 address in the event queue management table 24'. , there is no need to reconnect the queue blocks, that is, update the NEXT section. Therefore, even if the interrupt routine 20' operates while the event information processing routine 22' is operating, the chain will not be destroyed.

〔Example〕

Figure 2 is a diagram illustrating a process in which a new event occurs and an interrupt routine stores event information in an empty queue block and makes it a used queue block in an embodiment of the present invention.

(1) Store the event information that has occurred in the $z-pock (■) pointed to by the unused key 1-9. However, an unused queue? If the address of is 0, then over 7c! −
, and the event information is discarded.

(2) Set the address of the unused queue block to queue block 0.
Change the address to the queue block ■ pointed to by NEX'l.

During these processes, the event information that occurred was stored in the cue block ■ and used as the Φ knee block.

FIG. 3 shows the event information processing routine.

FIG. 3 is a diagram showing the process of reconnecting a block from a used key 1 to an unused queue.

(3) Find the chain of used cue blocks and search for the key block ■ that meets the search conditions.

However, if the used key block 9 matches the Nakajima block pointed to by the unused + and - blocks, there is no used key block (there is no event information).

(4) NEXT 1 of the cube a tsuku ■ which is one before the corresponding cube block ■? Change the address of the keypad to the address of the keypad pointed to by the NEXT part of the keypad.

(5) Perform various processing on the -block ■ in the corresponding day.

(6) After various processes are completed, event information is deleted for the key block (2) that is no longer needed.

(7) The last key - the key that 'h points to - Bro,
Change all the addresses in the NEXT section of RI to the addresses of the queue block ■.

(8) Change the address of the cue block ■ pointed to by the final Φ knee 1 to the address of the cue block ■. Furthermore, if the address of the unused queue block is 0, it is changed to the address of the unused -IF, -1) queue block (2).

Through the above process, event information is retrieved and processed queue blocks are returned.

〔effect〕

As explained above in detail, the single chain key management method of the present invention uses a means to connect an unused queue after a used key, so that an interrupt routine and an event information processing routine can be There is no need to perform exclusive control, and event information that occurs asynchronously can be given top priority without having to wait for the event information processing routine, reducing the processing time of the interrupt routine. .

Further, since there is no need for processing to reconnect queue blocks storing event information, the processing speed of the event information interrupt routine is increased.

[Brief explanation of the drawing]

Figure i1 is a diagram showing a configuration example of the queue filling method using a single Kinney of the present invention, Figures 2 and 3 are diagrams explaining the operation of an embodiment of the present invention, and Figure 4 is a diagram showing the conventional technology and the present invention. Invention t - A diagram showing an example of a system configuration to be described, s#! 5 and 6 are diagrams for explaining the operation of the conventional example. In the drawing, 1 is the central processing unit (CPU), 2 is the main memory (MS)
, 24 is an operating system (O8). 20 is an interrupt routine. I10 supervisor on 21. 22 is an event information processing routine. 23 is a cube block. 24 is an event queue management table. Agent Patent Attorney Sada Igata -” Input fill (9th arrow, keyboard t’) Kikudan, Technique 1, Invention, and the system that is difficult to understand (Fig. 8’, Fig. 4)

Claims (1)

[Claims] In addition to providing an event queue management table for managing the address of the first queue block of the used queue, the address of the first queue block of the unused queue, and the address of the last queue block of the unused queue, An unused queue should be connected after the medium queue, and when connecting a new event to the queue, find the start address of the unused queue from the event queue management table, write the event information to that queue block, and write the event information to the queue block. The head address of the unused queue in the event queue management table is updated with the address of the next queue block pointed to by the queue block, and when processing of the queue block in use is completed, the queue block in use is updated. A queue management method characterized in that a queue block is removed from a queue and reconnected to the end of an unused queue, and a final queue block address in an event queue management table is updated to point to the queue block.