JP3422504B2 - Exclusive control method between tasks - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an exclusive control method between a plurality of tasks performing parallel processing, and more particularly to a process for acquiring and releasing a lock on a resource shared between tasks. The present invention relates to an exclusive control method with a small overhead. 2. Description of the Related Art Multiprocessors having an architecture called a tightly-coupled type or a shared memory type are often used as means for improving the arithmetic processing capability of an electronic computer. In these multiprocessors, it is necessary to perform exclusive control so that a plurality of tasks performing parallel processing do not simultaneously access a shared resource stored in a memory. FIG. 3 shows a basic configuration of a shared memory type multiprocessor having an exclusive control function between tasks. In FIG. 3, the task control unit 302
The right to use a free processor is assigned (dispatched) to a task in an executable state (Ready), and the task is set to a running state. In addition, the task control unit 302 forcibly takes up the right to use the processor for a task that has passed a certain time after entering the Running state, returns the task to the Ready state, and assigns the right to use the processor to the other tasks in the Ready state. Dispatch. Further, even when a task in the Running state is in a Waiting state due to a lock acquisition wait or the like, the task control unit 302 dispatches the right to use the processor to another task in the Ready state. When a plurality of tasks are in the Ready state, the task control unit performs task priority order or
A task for dispatching the right to use the processor is selected in the order of the first state in the y state. A task running on each processor calls a procedure of the exclusive control unit 301 on the shared memory before starting access to the shared resource so that another task does not access the same shared resource at the same time. To get a lock. After the access to the shared resource is completed, the procedure of the exclusive control unit 301 is called again, and the lock is released so that another task can access the same shared resource. FIG. 4 shows a state transition of task A which cannot acquire a lock on the same resource because task B has acquired a lock on the shared resource first. is there. Task A was Ru
Although the vehicle is traveling in the nning state 401, the state transits to the waiting state 402 because the lock cannot be acquired. Thereafter, when the task B releases the lock on the shared resource, a wake-up signal is sent from the task B to the task A, and the task A transits from the Waiting state 402 to the Ready state 403. Next, when the right to use the processor is dispatched to task A, task A returns to the running state 401 again, acquires a lock on this resource, and continues processing. FIG. 5 is a flowchart showing a series of procedures from the acquisition to the release of a lock on a shared resource of one task when the conventional exclusive control method described above is used. First, the task to access the shared resource is the Test & S provided by the hardware.
Attempt to acquire a lock using an atomic and inseparable instruction such as et (step 110). If the lock is not successfully acquired, the information of its own task is registered in the queue of the shared resource (step 120), and the state transitions in the order shown in FIG. 4 (steps 130, 140, 150). Then, after acquiring the lock, the shared resource is accessed (steps 160 and 107). When the lock is released after the access is completed (step 180), a wake-up signal is transmitted to one of the tasks registered in the queue of this resource (steps 190 and 200). As described above, in the conventional inter-task exclusion control method, the running state →
Waiting state → Ready state → Running
State transition is performed three times such as state. In general, a state transition of a task requires a corresponding amount of calculation of a processor, and there is a problem that a large overhead occurs when a situation in which a lock on a shared resource cannot be acquired frequently occurs. [0008] Further, for a task in the Waiting state, it is necessary to perform queue management according to the cause, that is, which shared resource is requesting to acquire a lock, and this management is also overhead. Furthermore, the task that releases the lock needs to request acquisition of the lock for the resource, find the task that is in the Waiting state, and send a wakeup signal to one of the tasks, and this processing also involves overhead. There was a problem of becoming. Particularly in a multiprocessor, even if a task that cannot acquire a lock remains in the Running state, another task that has previously acquired the lock can run on another processor. There is a high possibility of being released, and there is no necessity to perform a control operation including the overhead described above. [0010] The present invention has been made in view of the above,
It is an object of the present invention to provide an inter-task exclusion control method with a small overhead associated with a process of acquiring and releasing a lock on a resource shared by a plurality of tasks performing parallel processing. In order to achieve the above object, an inter-task exclusion control method according to the present invention provides a shared memory type multiprocessor having an inter-task exclusion control function, which locks a shared resource. Identification means for identifying whether a certain period of time has elapsed for tasks that cannot be acquired , and
And the task has no waiting state and is in an executable state and an executable state.
A task control function that controls a task that has once become executable when it has two states so that it does not overtake another task that has become executable before that task.
Tasks that cannot acquire a lock on a shared resource with a stage are re-executed to acquire a lock in the execution state until a certain time elapses, and can be executed if the lock cannot be acquired after a certain time elapses Transition to the state,
After that, the operation for acquiring the lock is repeated from the beginning when the state becomes the execution state again, and the gist is that the operation for another task is not performed when releasing the lock. According to the inter-task exclusion control method of the present invention, a task that has failed to acquire a lock on a shared resource remains in the execution state until a predetermined time has elapsed, and repeats retrying to acquire the lock. If the lock cannot be acquired later, the state transits to the executable state.After that, when the right to use the processor is assigned to this task, the state becomes the execution state, and the retry for acquiring the lock is repeated again to lock the task. When releasing, processing other than release is not performed. Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows a state transition of task A, which cannot acquire a lock on the same resource because task B has acquired a lock on the shared resource first. is there. Task A was Ru
Although the vehicle was running in the nning state 401, the state in which the lock could not be acquired has passed for a certain period of time, and the
Transition to the y state 403 is made. Next, when the right to use the processor is dispatched to task A, task A returns to running state 401 again and tries again to acquire a lock on this resource. FIG. 2 is a flowchart showing a series of procedures from acquisition of a lock to release of a lock on a shared resource of one task when the inter-task exclusion control method according to the present invention is used. First, the task to access the shared resource is the Test & S provided by the hardware.
Attempt to acquire a lock using an atomic and inseparable instruction such as et (step 210). If the lock acquisition is not successful, the attempt to acquire the lock is repeated until a predetermined time has elapsed (step 220). If the lock cannot be obtained even after a certain period of time has elapsed, the state transits to the Ready state instead of the Waiting state of the conventional method (step 23)
0), if there is another task in the Ready state first, the task control unit 302 (FIG. 3) assigns a processor to any of the tasks based on the order of the Ready state, such as the first-come-first-served basis. Dispatches usage rights for. The previous task that could not acquire the lock will return to R when the right to use the processor is dispatched from another task.
The state becomes an uning state (step 240), and the procedure from the beginning is repeated again. The task that has acquired the lock by the above procedure accesses the shared resource (step 250). After the access is completed, the lock is released (step 260), but the wake-up signal is not transmitted to other tasks as in the conventional method, and this series of procedures is immediately terminated. In the above description, the reason why the other task which is in the Ready state first is executed first is that the task may have acquired a lock on the shared resource in question. In this case, a lock may not be acquired forever unless the task is executed first. As described above, according to the present invention,
The task that failed to acquire the lock on the shared resource remains in the execution state until a certain time elapses and repeats the retry of acquiring the lock.If the lock cannot be acquired even after the certain time elapses, the task transits to the executable state, After that, when the right to use the processor is assigned to this task, the task enters the execution state, the retry for acquiring the lock is repeated again, and when releasing the lock, no processing other than release is performed. If the lock cannot be acquired for the shared resource, the acquisition of the lock is retried in the Running state, so that the overhead associated with the state transition of the task can be removed. If the lock cannot be acquired even after a certain period of time has passed, the state transits to the Ready state. However, since the state does not go through the Waiting state unlike the related art, the number of state transitions is one less and the overhead is reduced accordingly. . Further, at the stage of releasing the lock, there is no need to request acquisition of the lock for the resource and transmit a wake-up signal for the task in the Waiting state, which has the effect of eliminating the overhead. Furthermore, queue management for tasks in the waiting state is not required, and this management overhead can be eliminated.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a task state transition diagram for explaining an embodiment of the present invention. FIG. 2 is a flowchart showing a series of procedures of a task for accessing a shared resource according to the present invention. FIG. 3 is a basic configuration diagram of a shared memory multiprocessor to which the present invention is applied. FIG. 4 is a state transition diagram of a task for explaining a conventional technique. FIG. 5 is a flowchart showing a series of procedures of a task for accessing a shared resource in the related art. [Description of Signs] 401 Running state 402 Waiting state 403 Ready state 301 Exclusive control unit 302 Task control unit

Continuation of front page (56) References JP-A-5-257902 (JP, A) JP-A-5-225149 (JP, A) JP-A-4-257930 (JP, A) (58) Fields investigated (Int) .Cl. ⁷ , DB name) G06F 9/46 G06F 12/00-12/06

Claims (1)

(57) [Claim 1] In a shared memory type multiprocessor having an exclusive control function between tasks, identification means for identifying whether or not a predetermined time has elapsed for a task that cannot acquire a lock on a shared resource. And tas
The task has no standby state and has two states: an executable state and an executable state.
The task control means for controlling so that the task has become once executable state not to another overtaking by running task became ready state before that time with states provided, can not acquire a lock on the shared resource task Retryed to acquire the lock in the execution state until a certain time elapses, and if the lock cannot be acquired after the certain time elapses, the state transits to the executable state, and then enters the execution state again An inter-task exclusion control method, characterized in that an operation for acquiring a lock is repeated from the beginning at the time, and no operation is performed on another task when releasing the lock.