JP2000259429A - Device and method for managing timer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To operate plural various systems with one timer device by unitarily managing the timer events of plural systems. SOLUTION: Timer queue elements 535 or the like of one (sub) system are respectively connected in the shape of queue and managed for each (sub) system by a timer queue managing table 532 or the like. The (sub) system is managed by a timer managing table 530 or the like for each (sub) system and each (sub) system is managed by a system queue managing table 510. An idle timer queue 520 successively connects used timer queues. In the case of connecting a novel timer queue element to a timer queue element 540 or the like, when there is the timer queue element 535 or the like, that novel timer queue element is connected by calculating a differential timer value but when there is no timer queue element, that novel timer queue element is connected to the head of the timer queue 540 or the like. Thus, plural timer queues are integrated, timer events are unitarily managed and plural various (sub) systems can be operated with one timer device.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a timer management apparatus and method, and more particularly, to a timer for controlling timers of a plurality of different systems by a single timer and efficiently using timer resources in a communication, measurement and control system or the like. The present invention relates to a management device and a method.

[0002]

2. Description of the Related Art Conventionally, a timer management method used when performing system control such as scheduling in communication, measurement / control, or the like, or when performing task management using a real-time operating system (OS), includes:
There were mainly three methods. Hereinafter, these three methods will be described in order.

(1) A method using an event timer management table. One system is composed of a plurality of processing units (tasks) divided by function, and one task has a plurality of operation states.

FIG. 17 shows a state transition of a task. FIG.
As shown in FIG. 7, one task has a plurality of operation states, and this operation state includes an execution state (RUN) 17.
00, an execution waiting state (READY) 1710, and a sleep state (WAIT) 1720. Depending on the operation state, each task is between the execution state 1700 and the sleep state (1725, 1735), between the sleep state 1720 and the execution wait state 1710 (1745, 1755), and the execution wait state 1710 and the execution state. Between (1715, 1
In 705), the operation state is changed (transitioned). In the subsystem, the number of tasks in the execution state (RUN) 1700 is always one. When an event such as an interrupt occurs, the task having the highest priority among the plurality of tasks in the execution state 1700 or the execution waiting state 1710 becomes the OS
Is selected (dispatched) to make a transition to an execution state (1715). The task in the execution state 1700 is in the sleep state 1 if the execution right is deprived to another task without being dispatched, or if the user himself wants to wait for an event.
Transition is made to 720. The task in the sleep state 1720 transitions to the execution wait state or the execution state when an event occurs or a wake-up instruction is received from the task in the execution state (1745, 1735).

A conventional method for managing identification information and the like relating to the above-described task will be described below. FIG. 18 shows a timer management table of a timer event in a conventional real-time OS when a task is considered as a timer user. In FIG. 18A, in a timer management table 1850, a timer user (task) identification information 1800 records a task ID (task number), a task execution priority, a stack pointer for each task, and the like. I have. Timer value 1
Reference numeral 810 represents the remaining operation time of the operating timer.
A task whose timer value 1810 is greater than 0 is waiting for a timer event, and the operation state 1820 of the task is in a sleep state (WAIT). A task whose timer value 1810 is larger than 0 does not acquire the execution right of the CPU. Timer value 1
The operation state 1820 of a task whose 810 is 0 is an execution state (RUN) 1700 or an execution waiting state (READY) 17
It is 10. A task issues a system call to enter sleep state 1720, but it is assumed that one task does not enter sleep state 1720 waiting for two or more timer events at the same time. In other words, it is assumed that there is only one timer event that activates a task so as to exit sleep state 1720, and therefore, one timer event exists for each task. Timer event (timer interrupt)
Occur at regular intervals. This fixed time is about 10 ms in many OSs. To the issuing task that issued and registered the timer event or the notification destination task that receives notification when the timer event occurred,
In the case of an OS having a function of notifying the occurrence of a timer event by any signal, the task does not need to transition to the sleep state 1720 after the occurrence of the timer event. In the case of such an OS, the task can simultaneously wait for the occurrence of a plurality of timer events without transitioning to the sleep state 1720. Generally, the notification of the occurrence of a timer event (hereinafter referred to as “time-up”) is performed in a mailbox.
As an example of waiting for the occurrence of a plurality of timer events, the device is turned on by the occurrence of the first timer event,
When the power of the apparatus is turned off by the occurrence of the second timer event, or when a certain process can be completed before the first timer event A occurs, the timer event A is deleted as a normal end. When the timer event B occurs, the stop processing is started as an abnormal state, and when the second timer event B occurs, the stop processing is considered to have failed and the emergency stop is performed.

FIG. 18B shows a timer management table when a task waits for a plurality of timer events to occur as described above. As shown in FIG. 18B, the timer management table 1800 has a structure in which cells 1830 to 1835 are connected in a queue (hereinafter referred to as “timer queue”).
The head address (pointer) to the timer queue is located at the timer value 1810 (reference numeral 1825 in FIG. 18B) of FIG. 18A. Thus, the actual timer is connected to the timer queue.

FIG. 19 shows a system 1 (1900) and a system 2 for explaining conventional timer interrupt processing.
(1910) shows an example of a system composed of two subsystems, and FIG. 20 shows a flowchart of a conventional timer interrupt process. Hereinafter, a conventional timer interrupt process will be described with reference to FIGS. As shown in FIG. 19, the timer management table 1920 includes a timer user (task) for each system 1900, 1910.
Identification information (task ID) 1800 and timer value 1
810 etc. are recorded. Each system 1900, 19
10 registers a timer event 1905 in the timer management table 1920 and notifies each system 1900, 1910 of the time-up 1915. Timer device 19
The timer value 1810 of the timer table 1920 is processed by an interrupt 1925 from 30 (usually a periodic interrupt of about 10 ms). Timer interrupt 1925
Occurs, the timer value 1810 of all tasks whose timer value 1810 is greater than 0 is decremented by one (decrement). It does nothing for tasks whose timer value 1810 is zero from the beginning. This process is performed for all registered tasks (steps S202 to S208 in FIG. 20). In the processing of the timer interrupt 1925, the task state is changed to the timer event wait state (WAIT) 1720 for the task in which the timer value 1810 is decremented by 1 to 0, that is, the task in which the timer 1810 changes from 1 to 0. Waiting for execution (READY)
Transition to 1710 (steps S208, S20
9). In the timer interrupt 1925, the timer value 18
When it is determined that all the decrement processings of Step 10 (Step S206) are completed (Step S202), the process returns to the dispatch processing of the real-time OS or the like (Step S20).
3). In the dispatch process, as part of the OS process,
When the processing of the timer interrupt 1925 is completed, the execution right of the CPU is given to the task having the highest execution priority. Specifically, identification information 1800 of the timer user (task)
Is restored in the stack pointer information (context: execution context) of the task, that is, set in the stack pointer register of the CPU, and the execution is shifted to the task having the highest execution priority.

The above-described task interrupt processing will be described in more detail with reference to FIG. When a timer event (timer interrupt) occurs, the timer table 1850 or 19
The following operation is performed from the top of 20. After the initial setting (step S200) and the end determination of all timer events (step S202), it is determined whether the timer is in use (step S204). When the timer is in use,
Timer value 18 in timer table 11850 etc.
10 is 1 or more. If the timer is in use, the timer value 1810 is decremented by 1 (step S
206). When the timer value 1810 becomes 0, a time-up process is performed (steps S208 and S209).
The time-up process S209 is a process of notifying the task in which the timer event has been registered as described above of the occurrence of the timer event. As a notification method of the timer event, a mailbox is used as described above, or the timer table 18 is used.
For example, there is a method of setting an event flag in 50 or the like and causing the OS to process the event flag. When the processing of all the timer events is completed, the process jumps to the OS dispatcher.

As described above, even in the conventional method, it is possible to manage a plurality of tasks by one timer. However, in the conventional method, for a system including a plurality of subsystems, a plurality of subsystems are mixed on the timer table 1920, and the management method between the plurality of systems becomes very complicated. There was a problem. In other words, different systems are mixed in one timer management table, and the operating state of one system may affect the management of the timer of another system. Further, the system in which the execution of one system is suspended, the suspended system, and the operating system are mixed on the same timer table 1920. Since the timer of the system in which the operation is suspended cannot be timed up, the decrement processing S206 in the timer table 1920 is performed only for the task of the operating system, and the timer table corresponding to the stopped or suspended system is used. 1920
Must be changed. As described above, in the conventional method, since the timer interrupt is separately prepared for each of a plurality of systems, there is a problem that a management method between the plurality of systems becomes very complicated.

(2) Timer queue (difference timer queue)
How to use. FIG. 21 shows a data structure of a difference timer queue conventionally used. As shown in FIG. 21, timer event 1 (2118) such as timer event 1 (2118) of each system
A queue structure (timer queue) 2100 is formed by using a timer value 2016 of an occurrence time interval such as 118) and a pointer (chain information) 2114 to a timer event of the next system as one cell 2110 and the like. For example, cell 2110 is a cell of system 1 and cell 2120 of system 2 is connected next. First cell 2
110 is connected to the head queue 2102, and the empty queue 2104 is connected to the head cell of an empty cell (not shown).

FIG. 22 shows the operation of the difference timer through the setting method of the timer queue 2100. In FIG. 22, the horizontal axis is the time axis, and the reference time of all event occurrence times is set to 0 on the time axis. The vertical axis shows four wait timer events (70 ms, 100 ms, 150 ms, 200 ms)
Is shown. It is assumed that these four timer events have started operation at the same time. The time measurement unit (time resolution) of the hardware is 10 ms. The four timer events 1 to 4 shown in FIG. 22 time up in ascending order of the set time. For example, the timer event 4 is changed from the timer event 1 of the 70 ms timer event.
The 200 ms timer event does not time out first. The four timer events 1 to 4 occur at intervals of 70 ms, 30 ms, 50 ms, and 50 ms on the time axis. That is, each occurrence time is TE1 (70 m
s), TE2 (100 ms), TE3 (150 ms) and TE4 (200 ms). Therefore, at the head of the timer queue 2100, event 1 (211
8) Connect the cell 2110 whose timer value 2116 is 70 ms. Timer value 21 of cell 2120 connected second
26 is a difference between the occurrence times of the timer event 2 (2128) and the timer event 1 (2118).
Set E1 = 100−70 = 30 ms. This difference value is called a difference timer value. The same applies to the third and subsequent ones. By doing so, 70 ms elapses when the first cell 2110 times out, and 30 ms elapses when the second cell 2120 times out therefrom. Therefore, the second cell 2120 from time zero (when a timer is issued) Until the time is up, 70 + 30 = 10
This means that 0 ms has elapsed. Similarly, the third cell 21
The 30th and 4th cells 2140 can also realize a required timer operation if the difference from the time-up time of the immediately preceding cell is set as a timer value 2126 or the like.

Table 1 below shows a method of calculating numerical values in FIG.

[0013]

[Table 1]

In Table 1, the time-up time indicates a time-up time (start time is set to 0) for each event. The set value indicates a timer value set in the timer queue, and the difference value indicates a relative time.

FIG. 23 shows a flowchart of a conventional process using a difference timer queue. In FIG. 23, it is checked whether one or more valid timer events are connected to the timer queue 2100 (step S23).
0). Here, “valid” means that the timer value of each cell (hereinafter, “timer queue element”) that is an element of the timer queue is larger than 0 (timer value> 0). If no valid timer queue element is connected, the process ends. If not, a signal is sent to the task having the task waiting for the timer event (step S2).
32). The timer queue element whose time is up is removed from the timer queue, and if another timer queue element is connected thereafter, the timer queue element is registered at the head of the timer queue management table (step S23).
6). The timer value of the first timer queue element of the timer queue is set in the hardware timer (step S23).
8).

As described above, even in the conventional method, it is possible to manage a plurality of tasks by one timer. However, in the conventional method, for a system consisting of multiple subsystems, different systems are mixed in one timer queue, and the operating status of one system affects the management of timers in another system. Was given. If one system has suspended its execution, the suspended system and the running system will be mixed in the same timer queue. The queue does not time up. Therefore, it became necessary to search for an operating system. Since the timer of the system whose operation is suspended cannot be timed up, removal from the queue in the timer queue is performed only for the task of the running system, and the timer corresponding to the stopped or suspended system is There was a problem that a process of not changing the event had to be performed. Therefore, in the conventional method, since a timer (interrupt) is separately prepared for each of a plurality of systems, there is a problem that a management method between the plurality of systems becomes very complicated. Furthermore, when all subsystems have been suspended, for example, when the operation has changed from two-system operation to one-system operation, invalid data (data corresponding to the inactive system) is connected in the timer queue. Therefore, the process of dequeuing from the timer queue (dequeue operation) becomes complicated. There was a problem that the operation of dequeuing from the timer queue could not be performed only with the chain information immediately after a certain timer queue element, and as a result, it was necessary to follow the chain information in some way to find the corresponding timer queue element of the pause system. .

(3) A task management table for a periodic activation event. A conventional management method for a timer event that is desired to be operated and started at regular intervals, such as a periodic monitoring operation, for example, a state confirmation operation at regular intervals, will be described.

FIG. 24 shows a conventional periodic startup task management table 2450. In FIG. 24, the periodic start time and the like are recorded in the timer user (task) identification information 2400 at the time of initial setting of the periodic start task. Timer value 2
410 is a remaining operation timer value. A task whose timer value 2410 is 0 indicates a task that is not running. Timer value 2
410 is decremented by one every time a regular timer interrupt (10 ms interrupt) is input. As soon as the timer value 2410 is detected to be 0, the task is notified of the occurrence of a timer event. In the timer value 2410, the cycle start time of the identification information 2400 of the timer user (task) is reset. When the task notified of the occurrence of the timer event executes the task operation and ends, the task returns to the execution waiting state 1710. That is, the execution right is abandoned and the process jumps to the dispatcher.

FIGS. 25A and 25B show the operation of a task in a conventional task management method. There can be a plurality of periodic tasks, but they are asynchronous with each other. As shown in FIG. 25A, TE10 and TE20 are the activation times of the periodic activation task, and the periodic activation time 2500 is
TE20−TE10 = 15−5 = 10 ms, and the task operation time 2550 is TE15−TE10 = 10−5 =
5 ms. As described above, when the periodic startup time 2500> the task operation time 2550, even if the timer value 2410 is set and decremented by 1 during the operation of the periodic startup task, the task time 2550 is shorter than the periodic startup time 2500. Since the task operation is completed, no problem occurs in the operation of the system. On the other hand, as shown in FIG. 25B, the second task activation time is the time of TE25, and the task operation time 2600 is TE25-TE10.
= 20-5 = 15 ms, the cycle start time 2500 <
Since the task operation time is 2600, the task started at the periodic start time does not complete within the periodic start time 2500.
Therefore, the task is started twice at the time of TE20, and there is a problem that the operation of the system is adversely affected.

As described above, in the conventional method, the management table 2450 for periodic operation, for example, a periodic activation task such as displaying the operation state every ls, is replaced with the timer table 1850 for managing timer events. It was necessary to prepare separately. The periodic activation operation has been processed by the independent management table 2450 or the like as a function independent of the timer table 1850 or the like. That is, there is a problem that two types of timer table management methods are required. As described above, in the conventional timer management method, the periodic start timer event is managed separately from the non-periodic one-shot type timer event processing. In addition, since the periodic activation timer events also operate independently, even if the periodic activation events have the same periodic activation, if the two have a temporal phase difference that is the difference between the timings at which the timer events occur, the timer is activated. It was not possible to make events occur simultaneously (simultaneous processing).
Therefore, there has been a problem that the number of timer interrupts tends to increase, and the current consumption of the system due to the timer interrupt tends to increase. For example, if the system is a mobile phone system, the call line is held in hardware even during a call, so if there is nothing to do even during the call, the CPU transitions to the sleep state (WAIT). Power saving. That is, not only the task but also the CPU itself is set to sleep (WAIT) to save power. Therefore, when the CPU is started by a timer interrupt or the like, the power consumption increases. Therefore, when the number of times of the timer interrupt increases, the power consumption surely increases. Furthermore, in the conventional timer management method, when a time longer than the time measurement range set by hardware is set, a counter having a large bit size is provided by software separately from the timer management on hardware. The timer event was supported. Therefore, when the time unit or the time measurement range on the hardware changes, the counting process of the timer by the software is also affected. Further, in the conventional timer management method, there is no concept of an allowable error at the time of time measurement, and a timer event is generated simply when a timer time set (time-up) is set in hardware. For this reason, when a certain timer event has timed out, the dequeuing process could not be performed even if a timer event within an allowable error range in terms of time measurement is connected even if the time up is performed near (immediately later) at the same time. That is, the conventional timer management method has a problem that there is no function of integrating events that are very close in time.

[0021]

As described above, in the conventional timer management method, a plurality of subsystems are mixed on one timer table for a system including a plurality of subsystems. There was a problem that the management method between them became very complicated. Although a plurality of tasks can be managed by one timer, different systems are mixed in one timer queue, and the operation state of one system affects the management of timers of another system. There was something. Furthermore, it is necessary to prepare a management table for the periodic activation task separately from a timer table for managing the timer event. Even in the case of periodic activation events having the same periodic activation, if there is a temporal phase difference, simultaneous occurrence of timer events could not be performed, and the number of timer interrupts tends to increase. There is a problem that current consumption of the system tends to increase. Furthermore, the hardware timer was used to respond to a long-time timer event by limiting the number of bits of the hardware timer. However, if conditions on the hardware changed, processing by the software was also affected. Since there is no concept of an allowable error at the time of time measurement, there is a problem that there is no function of integrating timer events that are very close in time.

An object of the present invention is to solve the above-mentioned problem, and realizes a plurality of different system operations with one timer device by integrally managing timer events of a plurality of systems. It is an object of the present invention to provide an apparatus and a method for managing a timer.
Another object of the present invention is to treat periodical timer events in the same manner as one-shot timer events, thereby unifying the management of periodic timer events and one-shot timer events, and furthermore, when there is a temporal phase difference. Even if there is a timer event,
It is an object of the present invention to provide a timer management apparatus and a timer management method capable of reducing the number of timer interrupts to reduce the current consumption of the system due to the timer interrupts. Another object of the present invention is to
An object of the present invention is to provide a timer management device and method capable of handling a long timer event within the number of bits of a hardware timer by converting a long timer event into a periodic start timer event. It is still another object of the present invention to provide a timer management apparatus and method that can integrate timer events that are very close in time by utilizing the accuracy of a timer value.

[0023]

SUMMARY OF THE INVENTION A timer management device according to the present invention is a timer management device for managing a timer event in a queue, and the timer management device stores a timer queue element corresponding to the timer event in the timer event. One or more timer queue means for holding in a queue for each issuer, wherein the timer queue element has a difference between a time-up time of a timer event and a time connected to the timer queue as a timer value; When a timer interrupt occurs, for one or more of the timer queue means, the minimum timer queue element having the smallest timer value among the timer values of the separate timer queue elements held at the head of the separate timer queue means And a timer queue unit that holds the minimum timer queue element. A rearranging unit that holds the timer queue element held next to the Ima queue element at the head and deletes the minimum timer queue element from the timer queue unit; and a reordering unit that is different from the timer queue unit that holds the minimum timer queue element. And a resetting means for resetting a value obtained by subtracting the minimum timer value from the timer value of each timer queue element held at the head of each timer queue means as a new timer value. Things. Here, the timer management device of the present invention includes a separate timer queue management table means corresponding to each of the separate timer queue means, and a system queue management table means for managing the separate timer queue management table means in a predetermined order. The selection means selects the separate timer queue management table means in the predetermined order, and selects the minimum timer queue element for the separate timer queue means corresponding to the selected separate timer queue management table means. Is to select. Here, in the timer management device of the present invention, the system queue management table means may be configured such that, when an operation state of a source of a timer event corresponding to the timer queue management table changes, another system different from the timer queue management table. The timer queue management table is independently managed. here,
The timer management device of the present invention, wherein the system queue management table means, when the presence state of the issue source of the timer event corresponding to the timer queue management table changes,
Another timer queue management table different from the timer queue management table is independently managed. Here, in the timer management device of the present invention, the rearranging unit includes:
When the minimum timer queue element deleted from the timer queue means corresponds to a periodic timer event started every predetermined period of time, the deleted minimum timer queue element is deleted from the predetermined queue in the timer queue means. After the cycle time, it is reconnected to the corresponding position. Here, the timer management device of the present invention further comprises a connection means for newly connecting a timer queue element corresponding to a timer event to the timer queue means, wherein the connection means is activated every first cycle time. In a state where the first timer queue element corresponding to the dynamic timer event is already connected to the timer queue means, the second timer having a predetermined relationship with the first cycle time
When a second timer queue element corresponding to a periodic timer event started every cycle time is newly connected to the timer queue means, the earliest time-up time of the first timer queue element and the second timer queue element A connection in which the first time-up time is set to the same time, the number of times that the first timer queue element and the second timer queue element are simultaneously timed up by one timer interrupt is increased, and the current consumption due to the timer interrupt is reduced. is there. Here, in the timer management device of the present invention, the predetermined relationship with the first cycle time of the second cycle time is a relationship in which the second cycle time is a multiple of the first cycle time, and the connection unit includes: A value obtained by subtracting the current time from the earliest time at which the first timer queue element times out is connected as the first timer value of the second timer queue element, and the time-up time of the first timer queue element and the second timer queue are connected. The element time-up time is set to the same time. Here, in the timer management device of the present invention, the predetermined relationship with the first cycle time of the second cycle time is a relationship in which the second cycle time is a divisor of the first cycle time, and The time obtained by subtracting an integral multiple of the second cycle time from the earliest time at which the first timer queue element times up is the second time at the earliest time at which the time is not later than the current time.
It connects timer queue elements. Here, in the timer management device of the present invention, the predetermined relationship with the first cycle time of the second cycle time is a relationship in which the second cycle time is relatively prime to the first cycle time, and the connection means The first timer queue element and the second timer queue element are connected so as to simultaneously time-up at every cycle time that is the least common multiple of the first cycle time and the second cycle time. Here, the timer management device of the present invention further includes a connection unit for newly connecting a timer queue element corresponding to a timer event having a predetermined activation time to the timer queue unit, wherein the connection unit includes the timer queue element. Has a cycle time and a number of repetitions based on the predetermined activation time, and is connected to a periodic timer queue element corresponding to a periodic timer event activated by the number of repetitions for each of the cycle times. . Here, in the timer management device of the present invention, the timer queue element further includes an allowable error with respect to the timer value, and the rearranging unit includes a timer queue unit that holds the minimum timer queue element. If the allowable error of the next held timer queue element is smaller than the value obtained by subtracting the minimum timer value from the timer value, the next held timer queue element together with the minimum timer queue element is replaced by the timer. It is to be deleted from the queue means.

A timer management method according to the present invention is a timer management method for managing timer events by a queue,
In the timer management method, a timer queue element corresponding to a timer event and having a timer value corresponding to the time-up time of the timer event is held in one or more timer queue means by a queue for each issue source of the timer event. Steps, and when a timer interrupt occurs in response to a timer event, for one or more of the timer queue means,
A selecting step of selecting a minimum timer queue element having a minimum timer value among timer values of the separate timer queue elements held at the head of the separate timer queue means, and a timer queue holding the minimum timer queue element Means for rearranging the timer queue element held next to the minimum timer queue element at the head and deleting the minimum timer queue element from the timer queue means; and storing the minimum timer queue element. For another timer queue means different from the timer queue means, a value obtained by subtracting the minimum timer value from the timer value of each timer queue element held at the head of each timer queue means as a new timer value is reset. And a setting step.

[0025]

Embodiments of the present invention will be described below in detail with reference to the drawings. 1 to 5 show data structures commonly used in the first to fifth embodiments of the present invention. The data structures shown in FIGS. 1 to 5 realize a plurality of systems, a plurality of time accuracies, and a plurality of operations, such as a one-shot operation, a periodic operation, and a long-time operation, with one hardware timer. here,
The one-shot operation is an operation of a timer event that operates only once in a timer queue (hereinafter, referred to as a “one-shot type timer event”), and the periodic operation is a timer event that operates multiple times in a timer queue (hereinafter, a “periodic timer event”). Event ").
A long-term operation is a timer event in which the count of the hardware timer in the timer queue is relatively longer than other timer events (hereinafter referred to as “long-term timer event”).
Operation. To realize these operations, a plurality of equally-divided timer queues are used to adjust the timing of timer events in one system. Is performed between the timer queues that match the timing of the occurrence of the above.

FIG. 1 shows the structure of a timer queue for each subsystem. As shown in FIG. 1, timer event 1 (11), such as
The queue structure (timer queue) 100 is formed by using the timer value 116 and the like of the occurrence time interval such as 8) and the pointer (chain information) 114 to the next timer event as one cell (timer queue element) 110 and the like. ing. For example, after the timer queue element 110, the timer queue element 1
20 are connected. Although not shown in FIG. 1, the timer queue has a bidirectional linear list structure. The first timer queue element 110 is connected to the first queue 102, and the first cell of an empty cell (not shown) is connected to the empty queue 104.

FIG. 2 shows the data structure of the timer queue element. In FIG. 2, the forward link 202 is a pointer to the forward timer queue element in the timer queue, the backward link 20.
4 is a pointer to a backward timer queue element in the timer queue, timer ID 206 is a timer event, and therefore an identifier of the timer queue element, timer type / accuracy 208
Is the bit width or timer count unit (resolution, accuracy, etc.) of the operating hardware timer, and the timer value 210 is the timer value to be set. Although the timer value 210 uses a difference timer value in this specification, the present invention is not limited to the difference timer value. Further, in FIG. 2, the notification destination information 212 is the ID of the object that has issued the timer event, and the timer operation type 214 is the type and cycle of one-shot operation, periodic operation or long-time one-shot operation (long-time operation and one-shot operation). The activation time 216 is a periodic activation time of a periodic timer event that performs a periodic operation or an initial setting value of a long timer event that performs a long one-shot operation (as described in a fourth embodiment described below, the long timer event is a periodic timer event). Event), the cycle start counter 218 stores the number of repetitions of the cycle start timer event or long time timer event,
Reference numeral 20 denotes an operating system type for detecting an erroneous connection. The timer queue element 200 can also set information on other timer queue elements.

A timer queue management table is provided before the head timer queue element of the timer queue. This timer queue management table corresponds to the head queue 102 in FIG.

FIG. 3 shows the data structure of the timer management table connected before the timer queue management table and managing the timer queue. The timer management table is prepared for each system. In FIG. 3, a pointer 3 to the previous system
02 is a pointer (link) to the system before the system managed by the timer management table 300, and pointer 304 to the next system is the timer management table 3
The pointer (link) to the system next to the system managed by 00, the pointer to the first queue 306 is the pointer to the timer queue management table of the system, and the operation state 308 of the system is the operation of the system, the operation is stopped, etc. The flag or status for displaying the operation state of the system, and the system type 310 is the type of the operating system. In the timer management table 300, other information on the system can be set.

FIG. 4 shows a data structure of a system queue management table for managing the timer management table 300 for each system. In FIG. 4, link information 4 to the system
Reference numeral 02 denotes a pointer to another system when each system managed by the system queue management table 400 is a subsystem, and when a system including these subsystems is a single system as a whole. FIG. 4
, The pointer 404 to the first queue is a pointer to the timer management table 300 of the first (sub) system, the pointer 406 to the empty queue is a pointer to the empty timer queue, and the operating state 408 of the system is one system as a whole. Status of the system in operation, stopped or suspended, and other management information 41
0 is other information of the system, and system type 412 is the type of the system in operation.

FIG. 5 shows the relationship between the control structures of the system queue management table 400, the timer management table 300, the timer queue management table, the timer queue, and the like. As shown in FIG. 5, timer queue elements 535 etc. of one (sub) system are each connected in a queue,
This queue is managed for each (sub) system by the timer queue management table 532 and the like. In the case of a timer queue, the timer queue including the timer management table 532 and the like
40, etc., but may simply refer to a queue starting from the timer queue element 535. The (sub) systems are managed by a timer management table 530 for each (sub) system, and each (sub) system is managed by a system queue management table 510. An example of a (sub) system is Code Division Multiple A
ccess: CDMA) system, digital mobile phone / Personal Digital Cellular (PDC) system or PHS (Personal Handyphone System). The empty timer queue 520 is connected to a timer queue element whose timer has counted up and whose processing has been completed, that is, a used timer queue element. Since no new timer event can be issued if there are no empty timer queue elements, garbage collection or the like is performed as appropriate.
When a new timer queue element is connected to the timer queue 540 or the like, the timer queue element 5 is added to the timer queue 540 or the like.
If there are 35 or the like, a difference timer value is calculated and a new timer queue element is connected to the timer queue. On the other hand, when there is no timer queue element, the timer value 210 is set at the head of the timer queue 540 or the like, and a new timer queue element is connected as it is.

Embodiment 1 In Embodiment 1, the timer queues of a plurality of (sub) systems are integrated and the timer events of a plurality of (sub) systems are centrally managed, so that one timer device can be used. 1 shows a timer management device and method capable of realizing the operation of a plurality of different (sub) systems. Hereinafter, unless otherwise specified, a system refers to each system managed by the timer management table 530 or the like. System queue management table 510
In the case where a relationship with one system as a whole is managed by the system, one system as a whole is called a system,
Each system managed by the timer management table 530 or the like is called a subsystem. The first embodiment mainly describes a method of adjusting the temporal phase between timer queues of each system and an operation of removing a head timer queue element of a system corresponding to the timer event when the timer event occurs. The control data structure uses the data structure shown in FIGS.

FIG. 6 is a time chart showing a state where a timer interrupt has occurred at time Ti when three (sub) systems 1 to 3 are present. Each of the systems 1 to 3 has a timer queue 540 to 580 in FIG. 5, respectively. The timer queues 540 and the like are independent for each of the systems 1 to 3. In FIG. 6, when a timer interrupt occurs at time Ti, a timer queue element having the minimum timer value 210 is selected by comparing the first timer queue element 535 of the timer queue 540 and the like of each system. In the example of FIG. 5, the timer value 210 of the first timer queue element 535 of the system 1 is 10, the timer value 210 of the first timer queue element 555 of the system 2 is 7, and the timer value 210 of the first timer queue element 575 of the system 3 is 3 Therefore, the timer value 210 of the system 3
3 is the minimum. Therefore, it can be determined that the time of the head timer queue element 575 of the system 3 has expired.
Therefore, the first timer queue element 575 of the system 3 is removed from the timer queue 580, and the next timer queue element 57 is removed.
7 (the timer value 210 is 2) is the first timer queue element of the system 3. The removed timer queue element is connected to the empty timer queue 520. As described above, the operation at the time of the timer queue element 575 in the system 3 is completed. At this time, the timer queue elements 535 and 555 of the system 1 and the system 2 have not timed up. It needs to be reduced by minutes. Therefore, the first timer queue element 5 of each of the systems 1 and 2
A value obtained by subtracting 3 from the timer value 210 of the timer queue element 575 of the system 3 whose time has expired from 35 and 555 is reset as the timer value 210. In the case of the system 1, the timer value 210 is 10−3 = 7, and in the case of the system 2, the timer value 210 is 7−3 = 4.
Timer value 2 of each of the first timer queue elements 535 and 555
The content is 10. After resetting the timer value 210 of the head timer queue element of each of the systems 1 to 3, the minimum value is set as the hardware timer, and the timer table processing is completed. In the case of this example, since the timer value 210 of the system 3 is a minimum of 2, the hardware timer is set to 2. In this example, only one timer queue element 575 times up, but depending on the set timer value 210, the timer queue elements of a plurality of systems may time up simultaneously. in this case,
All the timer queue elements whose time is up are removed from the timer queue 540 or the like.

FIG. 7 is a flowchart showing the processing in the first embodiment. In FIG. 7, the values of the head timer queue element 535 of the timer queue 540 and the like are compared between the respective systems. However, depending on the operation state 308 of the system such as the timer management table 530, when the system is stopped or not mounted, the timer queue 540 or the like not operating when there is no valid timer queue element in the timer queue 540 or the like. No processing is performed for this. In this case, the pointer 304 to the next system in the timer management table 530 or the like is traced, and the timer queue 5 of the next system is read.
A process such as 60 is performed. At the same time, the system type 310 is recorded (step S70). When the timer times out, the timer value 210 such as the first timer queue element 575
Is the timer queue element whose time is up, such as the smallest timer queue element 575,
Remove it from timer queue 580 etc. When the two systems time up simultaneously, that is, when the values of the first timer queue element 575 and the like are the same, the two timer queue elements 575 and the like are removed. If three or more are the same, remove them in the same way.
The comparison of the timer value 210 between the above systems is performed by comparing the pointer 304 to the next system in the timer management table 530 or the like.
This is done by following Therefore, the processing method of the present embodiment is not affected by the number of subsystems constituting the system. Stopping or deleting a subsystem while the system is operating will only cause that system to be skipped when tracing the pointer 304 to the next system, so that processing of other operating subsystems Does not affect Next, the timer value 2 of the head timer queue element 535 etc. of all subsystems
Correct 10 As an initial setting, the system number counter i is cleared to zero (step S72). Until the processing for all subsystems is completed, steps S76 to S76 are performed.
Step 79 is repeated (step S74). First, it is determined whether or not the subsystem is a subsystem that has timed out (step S76). If the subsystem has timed out, the adjustment of the timer value 210 is unnecessary, and the process jumps to step S79. Otherwise, the difference timer value 210 of the timer queue element 575 or the like removed from the timer queue 580 or the like after the time-up is replaced with the difference timer value 210 of the head timer queue element 555 or the like of another subsystem that did not time-up. (Step S7)
8). By this decrement, the time phases of the timer queues 540 and the like between the subsystems can be adjusted.
In order to adjust the timer value 210 of the head timer queue element 535 of the next subsystem, etc., 1 is added to the system number counter i (step S79). Step S7
Returning to step 4, when the processing has been completed for all subsystems, a minimum value is searched for among the timer values 210 such as the head timer queue element 555 and the like, and the hardware timer is set.

In the first embodiment, when deleting each subsystem, it is only necessary to connect the first timer queue element 535 and the like to the system queue management table 510 and empty the contents of the timer queue management table 532 and the like. realizable. Conversely, when a system is newly added, a new independent timer queue, a timer queue management table, and a timer management table are generated, and the timer management table is stored in a timer of another subsystem managed by the system queue management table 510. This can be realized by simply adding the system to the list of management tables and setting the system operation state 308 indicating the operation state of the new system to "operating". Accordingly, a new system can be added without affecting other existing parts at all.

As described above, according to the first embodiment, the timer events of a plurality of subsystems are centrally managed using the timer queue of each subsystem, so that a single hardware timer can manage a plurality of different subsystems. The operation of the system can be realized.

Second Embodiment In the second embodiment, the management of the periodic timer event and the one-shot timer event can be unified by treating the periodic timer event in the same manner as a general one-shot timer event. 1 shows a timer management device and method. Hereinafter, the second embodiment will be described with reference to FIGS. The control data structure is
The data structure shown in FIGS. 1 to 5 is used. In the timer queue 540 or the like, in order to simultaneously handle a one-shot type timer event that operates only once and a periodic timer event that operates repeatedly according to a set value, the operation identification information is stored in the timer queue element 200 of FIG.
14, the cycle start time 216 is provided. A timer queue element whose timer operation type 214 is a periodic operation (hereinafter, referred to as a “periodic timer queue element”. A timer queue element corresponding to a one-shot timer event is particularly referred to as a one-shot timer queue element to be distinguished therefrom) Are periodically activated every cycle activation time 216.

FIG. 8 shows the operation of the periodic timer event in the second embodiment. In FIG. 8, a periodic timer queue element 802 (a periodic activation time 216
= 10) has timed out. At this time, the periodic timer queue element 802 is removed from the head of the timer queue, but is activated again after the periodic activation time 216 (= 10).
0) Reconnect to later position 810. As shown in FIG. 8, one-shot timer queue elements 804, 806
After that, the connection is re-established with the timer value of the periodic activation set to 3.

FIG. 9 is a time chart showing the operation of FIG. As shown in FIG. 9, the periodic timer queue element 80
The time at which the time 2 is up is T _A , the time at which the one-shot timer queue element 804 times up is T _B , the time at which the one-shot timer queue element 806 times up is T _C , and the one-shot timer queue element 808 times up. Assuming that the time is _TE , the periodic timer queue element 810 to be reconnected has a periodic activation time 216 of 10
Since it is the position of the time T _D which has elapsed 10-2-5 = 3 time, the timer value 210 it is understood that it may be re-connected using the 3. The timer queue after resetting the periodic timer queue element 810 is shown in FIG.

FIG. 11 is a flowchart showing the processing of a periodic timer event in the second embodiment. In FIG. 11, when the timer expires, an event is notified to the timer issuer (step S110). This process is the same as the one-shot operation timer event. Next, it is determined whether the time of the periodic timer event has expired (step S112), and if so, the periodic timer event whose time has expired is reconnected to the timer queue (step S114). If there is no timer queue element next to the periodic timer queue element, the periodic timer event will be connected to the head of the timer queue again. Finally, the timer value 210 of the first timer queue element is set as a hardware timer (step S116). Step S1
If it is not time-up of the periodic timer event at 12, the process directly goes to step S116.

According to the second embodiment, since two or more identical periodic timer events do not exist in the timer queue, memory can be saved and time for searching the timer queue can be saved. To stop the periodic timer event, only one of the periodic timer event needs to be deleted from the timer queue. Furthermore, according to the second embodiment, even when the maximum count number of the timer (the bit width of the timer) is small in hardware, hardware restrictions can be eliminated by using the timer queue element periodically. . For example, when the resolution of the timer is 10 ms and the cycle start time is the longest for one hour, one hour = 60
(Minute) × 60 (second / minute) × 100 (count / second) = 3
Since the count is 60,000, the hardware timer that realizes this one-hour long-term timer event has about 20
A bit is needed. On the other hand, according to the second embodiment, the long-term timer event having a periodic start time of one hour is cycled,
For example, if a periodic timer queue element with a maximum of 1 minute at a resolution of 10 ms is used, 1 minute = 60 seconds × 100 (counts / second) = 6,000 counts. Therefore, the hardware timer for realizing the one-hour long-term timer event can be realized with a bit width of 16 bits, and can be reduced by about 4 bits.

As described above, according to the second embodiment, by converting a periodic timer event into a one-shot type timer event, it can be handled in the same manner as a general one-shot type timer event. Shot type timer event management can be unified.

Embodiment 3 Embodiment 3 shows a timer management apparatus and method capable of reducing the number of times the hardware timer times out by performing temporal phase adjustment between periodic timer events. Hereinafter, the third embodiment will be described with reference to FIGS. The control data structure uses the data structure shown in FIGS.

FIGS. 12A and 12B are time charts showing the operation of the periodic timer event in the third embodiment. As shown in FIG. 12A, for example, it is assumed that there is a periodic timer event 1 in which the periodic activation time 216 is every 1 second and a periodic timer event 2 in which the periodic activation time 216 is every 2 seconds. In this case, if the time phase adjustment is not performed, that is, if the timer event 1 and the timer event 2 are asynchronous, the timer event 1 becomes T2, T3, T5, and T6.
And T8, a total of 5 times, and timer event 2 is T1,
It occurs three times, T4 and T7, a total of eight times.

On the other hand, as shown in FIG. 12 (B), by adjusting the time phases at which the two periodic timer events 1 and 2 occur, the timing T1
At 2, T14 and T16, two timer events 1 and 2 can be simultaneously timed up. As a result, the number of occurrences of the timer event can be reduced from eight to five. Hereinafter, a method of adjusting the time phases of two timer events will be described.
As shown in FIG. 12B, the period timer event 2
Is activated (Tz), the issuance time of the timer event 2 is adjusted in accordance with the temporal phase of another periodic timer event 1 already running on the timer queue, and connection is made to the timer queue. When registering timer event 2 (T
z), time-up time T of other periodic timer event 1
The difference Δt between T 12 and Tz is calculated, and the Δt is set as the first (first) timer value 210 of the periodic timer event 2. That is, the timer value 210
= Δt, timer operation type 214 = cycle, cycle start time 2
16 = set to the time or the like to start periodically. As described above, the timer value of the timer event 2 is set only once at the first time so as to synchronize with the generation cycle (= 1) of the other periodic timer event 1.
10 is Δt instead of the original periodic start time 216 (= 2)
Set to. Once synchronization is achieved at T12,
4. Both T16 and T16 are synchronized. As a result, the number of timer interrupts generated by the hardware timer can be reduced. Therefore, since the power consumed by the activation of the CPU by a timer interrupt or the like can be reduced, the power saving can be largely measured in the case of a battery-driven system such as a mobile phone.

For example, an operation in a case where a period timer event having a period start time 216 and a period timer event having a period start time 216 is added to a period timer event having a period start time 216 of 6 will be described. A timer event having a cycle start time 216 of 12 and a start time 216 of 72 is issued by adjusting the time phase to the cycle of the cycle start time 216 of 6. Timer queue elements corresponding to both timer events are connected to the timer queue. In this case, 6, 12, of the cycle start time 216,
Since 72 has a greatest common divisor of 6, the cyclic start time 216
Adds two timer events, 12, and 72,
The number of times that the PU is activated to process the timer event does not increase. If there is a greatest common divisor, there is no limit to the number of timer events to be added except for conditions such as memory.

Conversely, the same applies to the case where a timer event whose cycle start time 216 is 6 is connected to an event whose cycle start time 216 is 72. In the case where a timer event whose cycle start time 216 is 6 is connected to an event whose cycle start time 216 is 72, since the greatest common divisor of 72 and 6 is 6, the timer event whose cycle start time 216 is 72 times out. At the timing when the maximum n obtained by subtracting 6.times.n time from the time t becomes 0, an event whose cycle start time 216 is 6 may be issued and connected to the timer queue. The timing for giving the maximum n is the earliest timing.

An example in which the cycle start times 216 are disjoint is shown. It is assumed that a timer event whose periodical start time 216 is 3 is issued while a timer event whose periodical start time 216 is 5 is operating. At this time, the event whose cycle start time 216 is 5 and the event whose cycle start time 216 is 3 are relatively prime, that is, they are prime numbers and have no common divisor. In this case, the cycle start time 216 is the least common multiple of 1 between 3 and 5.
The issue timing of the timer event whose cycle start time 216 is 3 may be adjusted so that the occurrence of the two timer events overlaps the time of 5 (3 × 5).

As described above, according to the third embodiment, the number of times the hardware timer times out can be reduced by performing the temporal phase adjustment between the periodic timer events, so that the CPU is activated by a timer interrupt or the like. Power consumption can be reduced.

Fourth Embodiment A fourth embodiment is directed to a timer management device capable of realizing a long timer having a count equal to or greater than the maximum count of a hardware timer by converting a long one-shot timer event into a periodic timer event. And the method. Hereinafter, the fourth embodiment will be described with reference to FIG.
(A) to (C) will be described with reference to FIG. The control data structure uses the data structure shown in FIGS.

FIGS. 13A to 13C are time charts showing the operation of the periodic timer event in the fourth embodiment. 13 (A) to 13 (C), the operation will be described for a case where the repetition period of the periodic timer event is 200 ms and a long time timer of 1050 ms is realized. This repetition cycle is desirably determined by the longest measurement range (range determined by the bit width) of the hardware timer.

FIG. 13A shows the operation of the first long time timer event 1300. As shown in FIG. 13 (A), the long-time timer 1050 ms is repeated for a period of 200 m.
It is decomposed into s and the number of repetitions (times), and if there is a remainder, this remainder is set as the first timer value 1310. In the above example, the long-time timer 1050ms = 200ms × 5
(Times) +50, and the first set value of the timer value 1310 is 50 ms because the remainder is 50. The timer operation type 1311 is set to a long time. Cycle start time 1
312 is 50 ms for the first time, like the initial timer value 1310
And the number of repetitions 1314 is 6 (= 5 + 1
Times). The number of repetitions 1314 is set one extra for convenience of processing.

FIG. 13B shows the operation of the long timer event 1328 when the first time-up occurs. In FIG. 13B, the elapsed time at the time of the first time-up is equal to the timer value 1310 set at the first time.
ms. The remaining operation of the long timer event is 2
What is necessary is to increase the time of the repetition cycle of 00 ms by five times. In this setting, the remainder 50 in FIG.
Since the time has already expired once for the time of ms, and the time has expired five times with the repetition period of 200 ms thereafter, the time is increased six times (= 1 + 5) in total. This is the reason why the number of repetitions 1314 is set one extra. In the first reconnection process, first, the period timer queue element 1300 whose time has expired is removed, and then the period timer queue element 1328 is located at a position where the repetition period becomes 200 ms starting from the head timer element 1302 connected from the beginning. To reconnect. In the example of FIG. 13B, 200 ms—the timer value of the timer queue element 1302 (= 10) —the timer value of the timer queue element 1304 (= 20) —the timer value of the timer queue element 1306 (=
100) = 70, so the timer queue element 1328
The timer value 1330 is set to 70. The cycle start time 1332 is set to 200 ms, and the cycle counter 133
4 is set to 5 times.

FIG. 13C shows the operation of the long timer event 1348 when the second time-up occurs. In FIG. 13C, first, the period timer queue element 1330 whose time has expired is removed, and the period timer queue element 1330 is located at a position where the repetition period becomes 200 ms starting from the head timer queue element 1342 connected thereto.
48 is reconnected. In the example of FIG. 13C, 200 ms
-Timer value of timer queue element 1342 (= 80)-Timer value of timer queue element 1344 (= 30)-Timer value of timer queue element 1346 (= 10) = 80, so timer value 1350 of timer queue element 1348 Is set to 80. The cycle start time 1352 is 200m
s, and the cycle counter 1354 is set to four times. When the period counter 1354 is decremented by 1, the value becomes 4, but it does not become 0 yet.
In the same manner as described above, the timer is set again at the cycle start time of 200 ms, and the timer queue is connected again. The same is repeated until the fifth time-up. When the sixth time-up occurs, the period counter is decremented by 1 and becomes 0, which means that the one-shot timer event has been completed for a long time. Therefore, the timer queue element at this time only needs to be removed from the timer queue, and reconnection of the periodic timer event to the timer queue is no longer necessary. As mentioned above,
The point of the fourth embodiment is that the remainder is set as a timer value at the first time. For the second and subsequent settings, the cycle start time is set. Although the value to be set is different, the setting processing method is the same every time.

FIG. 14 shows a flowchart of the process in the fourth embodiment. In FIG. 14, the period counter 13
It is determined whether 32 or the like is 0 (step S14).
0), if not 0, it is further determined whether or not the remainder is 0 (step S140). In the first time, since the remainder is not 0, it is connected to the timer queue based on the remainder value (step S
148). Since the remainder value is 0 after the first time-up, 1 is subtracted from the value of the cycle counter (step S
144). Next, reconnection to the timer queue is performed based on the periodic activation time 1332 and the like (step S146). The process ends when the cycle counter becomes 0.

As described above, according to the fourth embodiment, by converting a long-time one-shot timer event into a periodic timer event, it is possible to realize a long-time timer that is equal to or more than the maximum count of the hardware timer.

Fifth Embodiment A fifth embodiment shows a timer management apparatus and method which can remove a queue at a time by using the accuracy of a timer value connected to a timer queue and integrate timer events. Hereinafter, the fifth embodiment will be described with reference to FIGS. 15 and 16. The control data structure uses the data structure shown in FIGS.
Since the accuracy required for the timer accuracy 208 varies depending on the purpose of use, in the fifth embodiment, the variation in the accuracy is used to reduce the number of occurrences of the timer event. For example, when a 10 msec timer event times out, the timer accuracy 208 becomes 1 near the timer event.
If the timer event is a minute and the timer value 210 is 1 hour, even if the 1 m timer event is removed from the timer queue at the same time as the 10 msec timer event times out, there is no problem in the required accuracy of the timer. Can be In such a case, remove it from the timer queue,
That is, power saving can be achieved by increasing the time.

FIG. 15 shows the operation principle of the fifth embodiment. As shown in FIG. 15, the timer queue element A15
00 corresponds to a timer event in which the differential timer value 210 operates at 100, and the timer queue element B1510 having the differential timer value 210 of 2 is connected next. Here, assuming that the first timer queue element A 1500 has timed out, the timer value 210 of the timer queue element B 1510 is 2 and the allowable error is 10, even if the timer queue element B 1510 is simultaneously timed up. There is no problem in the required accuracy of the element B1510. Thus, power consumption can be reduced by simultaneously removing the timer queue element A1500 and the timer queue element B1510 from the timer queue. FIG. 16 shows a flowchart of the process in the fifth embodiment. FIG.
6, the first timer queue element is removed (step S160). The timer value of the next timer queue element is compared with the precision (step S162), and if the timer value is smaller than the precision, the next timer queue element is simultaneously removed (step S164). Thereafter, the connection is made such that the next next timer queue element is the first element of the timer queue (step S168). As a result, the timer value of the first timer queue element is set in the hardware timer. If the timer value is greater than the accuracy in step S162, the process ends without doing anything. Although not shown in FIG. 16, the timer table processing is repeatedly called, so that the timer queue elements whose timer values of the timer queue elements are smaller than the precision can be simultaneously timed up.

As described above, according to the fifth embodiment, two or more timer processes can be performed simultaneously by one timer event, so that the number of timer interrupts can be reduced and the current consumption can be reduced. be able to.

[0060]

As described above, according to the timer management apparatus and method of the present invention, a plurality of different system operations are realized by one timer apparatus by integrally managing the timer events of a plurality of systems. And a timer management device and method capable of performing the same. By treating periodic start timer events in the same way as one-shot timer events, management of periodic start timer events and one-shot timer events is unified, and even when there is a temporal phase difference, timer events occur simultaneously. Can be performed, and a timer management device and method can be provided which can reduce the number of timer interrupts and reduce the current consumption of the system due to the timer interrupts. By converting a long-time timer event into a periodic start timer event, it is possible to provide a timer management device and method capable of handling a long-time timer event within the number of bits of a hardware timer. Furthermore, by using the accuracy of the timer value, it is possible to provide a timer management device and method that can integrate timer events that are very close in time.

[Brief description of the drawings]

FIG. 1 is a diagram showing a structure of a timer queue for each subsystem according to the first to fifth embodiments.

FIG. 2 is a diagram showing a data structure of a timer queue element in the first to fifth embodiments.

FIG. 3 is a diagram showing a data structure of a timer management table for managing a timer queue according to the first to fifth embodiments.

FIG. 4 is a diagram showing a data structure of a system queue management table for managing a timer management table for each system according to the first to fifth embodiments.

FIG. 5 is a diagram showing the relationship between control structures such as a system queue management table, a timer management table, a timer queue management table, and a timer queue in the first to fifth embodiments.

FIG. 6 is a time chart showing a state in which a timer interrupt has occurred at time Ti in the first embodiment.

FIG. 7 is a flowchart showing a process according to the first embodiment.

FIG. 8 is a diagram illustrating an operation of a periodic timer event according to the second embodiment.

FIG. 9 is a time chart showing the operation of FIG.

FIG. 10 is a diagram showing a timer queue after resetting a periodic timer queue element according to the second embodiment.

FIG. 11 is a flowchart showing processing of a periodic timer event according to the second embodiment.

FIG. 12 is a time chart illustrating an operation of a periodic timer event according to the third embodiment.

FIG. 13 is a time chart illustrating an operation of a periodic timer event according to the fourth embodiment.

FIG. 14 is a flowchart of a process according to the fourth embodiment.

FIG. 15 is a diagram illustrating an operation principle of the fifth embodiment.

FIG. 16 is a flowchart of a process according to the fifth embodiment.

FIG. 17 is a diagram showing a state transition of a task.

FIG. 18 is a diagram illustrating a timer management table of a timer event in a conventional real-time OS.

FIG. 19 is a diagram illustrating an example of a system including two subsystems for explaining conventional timer interrupt processing.

FIG. 20 is a flowchart of a conventional timer interrupt process.

FIG. 21 is a diagram showing a data structure of a difference timer queue conventionally used.

FIG. 22 is a diagram illustrating the operation of a difference timer through a conventional timer queue setting method.

FIG. 23 is a flowchart of a conventional process using a difference timer queue.

FIG. 24 is a diagram showing a conventional management table of periodic activation tasks.

FIG. 25 is a diagram showing an operation of a task in a conventional task management method.

[Description of Signs] 100, 540, 560, 580, 2100 Timer queue, 102, 2102 Head queue, 104, 5
20, 2104 free queue, 110, 200, 53
5, 555, 575, 577 timer queue element, 3
00, 530, 550, 570, 1920 Timer management table, 400, 510 System queue management table, 532, 552, 572 Timer queue management table, 802, 810 Period timer queue element,
804, 806, 808, 1302, 1304, 13
06, 1342, 1344, 1346, 1500, 15
10 One-shot type timer queue element, 1300,
1328, 1348 Long time timer queue element, 17
00 execution state, 1710 execution waiting state, 172
0 Sleeping state, 1850, 2450 timer table, 1900, 1910 system, 2500 cycle start time, 2550, 2600 task operation time.

Claims (12)

[Claims] 1. A timer management device for managing timer events by means of a queue, wherein the timer management device stores one or more timer queue elements corresponding to the timer events in a queue for each of the sources of the timer events. Timer queue means, wherein the timer queue element has, as a timer value, a difference between a time-up time of a timer event and a time at which the timer queue is connected to the timer queue; Selecting means for selecting the smallest timer queue element having the smallest timer value among the timer values of the separate timer queue elements held at the head of the separate timer queue means; For the timer queue means for holding the timer queue, the timer queue held next to the minimum timer queue element Rearranging means for holding the queue element at the head and deleting the minimum timer queue element from the timer queue means; and timer queue means different from the timer queue means holding the minimum timer queue element. A timer management device comprising: resetting means for resetting a value obtained by subtracting the minimum timer value from the timer value of each timer queue element held at the head of the means as a new timer value. 2. The system further comprising: separate timer queue management table means respectively corresponding to the separate timer queue means; and system queue management table means for managing the separate timer queue management table means in a predetermined order. Means for selecting the separate timer queue management table means in the predetermined order and selecting a minimum timer queue element for the separate timer queue means corresponding to the selected separate timer queue management table means. The timer management device according to claim 1, wherein 3. The system queue management table means, when an operation state of a timer event issuing source corresponding to the timer queue management table changes, separates another timer queue management table different from the timer queue management table. 3. The timer management device according to claim 2, wherein the timer management device manages the timer. 4. The system queue management table means according to claim 1, wherein, when the presence state of the issue source of the timer event corresponding to the timer queue management table changes, another timer queue management table different from the timer queue management table becomes independent. 3. The timer management device according to claim 2, wherein the timer management device manages the timer. 5. The reordering means, wherein when the minimum timer queue element deleted from the timer queue means corresponds to a periodic timer event activated every predetermined period of time, the deleted minimum timer queue element is deleted. 2. The timer management apparatus according to claim 1, wherein an element is reconnected to a position in the timer queue means corresponding to the predetermined period of time. 6. A connection device for connecting a timer queue element corresponding to a timer event to the timer queue unit, the connection unit corresponding to a periodic timer event activated every first period of time. A second timer queue element corresponding to a periodic timer event activated every second cycle time having a predetermined relationship with the first cycle time in a state where the first timer queue element is already connected to the timer queue means; Is newly connected to the timer queue means, a connection is made so that the earliest time-up time of the first timer queue element and the first time-up time of the second timer queue element are the same, and one timer Increase the number of times that the first timer queue element and the second timer queue element are simultaneously timed up by an interrupt, and reduce the current consumption due to the timer interrupt Timer management apparatus according to claim 1, wherein the door. 7. The predetermined relationship between the second cycle time and the first cycle time is a relationship in which the second cycle time is a multiple of the first cycle time, and the connection means includes a first timer queue element. The value obtained by subtracting the current time from the earliest time when the time is up is connected as the first timer value of the second timer queue element, and the time up time of the first timer queue element and the time up time of the second timer queue element are calculated. 7. The timer management device according to claim 6, wherein the timers are set at the same time. 8. The predetermined relationship between the second cycle time and the first cycle time, wherein the second cycle time is a divisor of the first cycle time, and the connecting means includes a first timer queue element. 7. The timer management apparatus according to claim 6, wherein the second timer queue element is connected to the earliest time at which the time obtained by subtracting an integral multiple of the second cycle time from the earliest time at which the time is up does not become later than the current time. . 9. The predetermined relationship between the second cycle time and the first cycle time is a relationship in which the second cycle time is relatively prime to the first cycle time, and the connection means includes a first cycle time and a first cycle time. 7. The connection according to claim 6, wherein the first timer queue element and the second timer queue element are simultaneously connected at every cycle time that is a least common multiple of the second cycle time.
The timer management device according to the above. 10. A connection means for newly connecting a timer queue element corresponding to a timer event having a predetermined activation time to said timer queue means, said connection means connecting said timer queue element to said predetermined activation time. 2. A periodic timer queue element corresponding to a periodic timer event that is activated by the number of repetitions for each of the periodic times, and is connected thereto. Timer management device. 11. The timer queue element further includes an allowable error with respect to the timer value, wherein the rearranging unit is configured to hold the minimum timer queue element next to the minimum timer queue element. When the allowable error of the timer queue element is smaller than a value obtained by subtracting the minimum timer value from the timer value, deleting the next held timer queue element together with the minimum timer queue element from the timer queue means. The timer management device according to claim 1, wherein: 12. A timer management method for managing timer events by means of a queue, wherein the timer management method includes a timer queue element corresponding to a timer event and having a timer value corresponding to a time-up time of the timer event. Holding in one or more timer queue means by a queue for each issuer of the timer event, and when a timer interrupt corresponding to the timer event occurs,
Selecting, for one or more of the timer queue means, a minimum timer queue element having a minimum timer value among timer values of the separate timer queue elements held at the head of the separate timer queue means; A timer queue means for holding the minimum timer queue element, a rearranging step of holding the timer queue element held next to the minimum timer queue element at the head and deleting the minimum timer queue element from the timer queue means; A value obtained by subtracting the minimum timer value from the timer value of each timer queue element held at the head of each timer queue means for another timer queue means different from the timer queue means holding the minimum timer queue element. Resetting step of resetting a new timer value as a timer value. Management method.