A non-precise multiprocessor hybrid critical system resource access method

Abstract

The method comprises: establishing a non-precise multiprocessor hybrid critical system resource sharing model; each processor is started in a low mode, and only a task with the highest priority and a preemptive level higher than a current system priority upper limit is allowed to be executed, and execution time of the task in the low mode does not exceed C i (LO); at time ts, if execution time of any high critical level task on the processor exceeds C i (LO) and execution is not terminated, all processors will be immediately switched to a high mode; at time ts, a low critical level task which has not completed execution and whose execution time exceeds C i (HI) will be terminated execution, but execution time of a low critical level task which has not reached C i (HI) will continue execution to C i (HI), execution time of a high critical level task which has not completed execution will be raised from C i (LO) to C i (HI); in the high mode, execution time of the task does not exceed C i (HI). The application guarantees that system resources are accessed mutually exclusively while improving resource utilization on the basis of satisfying real-time requirements of the system.

Description

Technical Field

[0001] This invention relates to real-time scheduling of real-time systems, embedded systems, and hybrid critical systems, and particularly to a method for accessing resources in an imprecise multiprocessor hybrid critical system. Background Technology

[0002] Due to limitations in power consumption, size, and weight, the development trend of embedded systems is to integrate applications at different critical levels onto a single shared platform, thus forming hybrid critical systems. Hybrid critical systems have different execution modes, and hybrid critical tasks also have different levels of criticality. Furthermore, tasks have different worst-case execution times in different execution modes. Common examples of hybrid critical systems include the control systems of unmanned aerial vehicles (UAVs) and automotive control systems.

[0003] Existing resource access methods for hybrid critical systems primarily target the classic hybrid critical task model. This model discards lower-critical tasks directly when the system is in high-critical mode, which does not reflect reality. Furthermore, existing research mainly focuses on single-processor platforms. To address the resource access problem in multi-processor hybrid critical systems, this paper proposes a more realistic, imprecise multi-processor hybrid critical system resource access method. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a more practical, imprecise method for accessing resources in multiprocessor hybrid critical systems, which can solve the resource access problem of hybrid critical systems on multiprocessor platforms.

[0005] To achieve the above objectives, the technical solution of the present invention is a method for accessing resources in a non-precise multiprocessor hybrid critical system, comprising:

[0006] Establish a resource-sharing model for an imprecise multiprocessor hybrid critical system;

[0007] Each processor starts in low mode, where only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to execute. In low mode, task execution time will not exceed C... i (LO);

[0008] At time ts, if any high-criticality task has been executed on the processor for more than C... i If (LO) is not terminated, all processors in the system will immediately switch to high mode;

[0009] At time ts, execution has not been completed and the execution time has exceeded C. i Low-criticality tasks (HI) will be terminated, but the execution time already completed has not reached C. i(HI) low-criticality tasks will continue to be executed up to C. i (HI), the execution time of high-critical-level tasks that have not been completed will be reduced from C. i (LO) upgraded to C i (HI);

[0010] In high mode, only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to be executed. In high mode, task execution time will not exceed C. i (HI);

[0011] The establishment of a non-precise multiprocessor hybrid critical system resource sharing model includes:

[0012] The system contains M independent processors with identical architectures, Φ = {φ} k |1≤i≤M};Inaccurate hybrid critical sporadic real-time tasks Γ={τ} on multiprocessors, considering implicit deadlines. i |1≤j≤N}, where N represents the number of inaccurate mixed critical and sporadic real-time tasks. This task set is scheduled using a partitioned earliest-due-date-first strategy. The basic idea of ​​this strategy is to divide the real-time tasks into several independent subsets, each assigned to a different processor. On each processor, tasks are scheduled according to the earliest-due-date-first method; tasks with earlier deadlines have higher priority. When tasks have the same deadline, their priority is determined by their arrival time, with earlier arrival times having higher priority. Higher-priority tasks are scheduled first. These tasks are represented by a quintuple τ. i = <L i C i (LO), C i (HI), D i T i >; where L i L represents the key level of the task. i Represented as L i = {LO, HI}, where LO represents low-criticality tasks and HI represents high-criticality tasks; C i (LO) and C i (HI) represent the worst execution time of the task in low mode and high mode, respectively; D i Represents task τ i The relative deadline for an implicitly deadline, or the relative deadline for a non-precise multiprocessor hybrid critical system, is equal to the period, i.e., D. i =T i ;T i Indicates the task's duration; if the mixed critical-occurrence task is a low-criticality task, then C i (LO)≥C i(HI); If the mixed critical-occurrence task is a high-criticality task, then C i (LO)≤C i (HI);

[0013] The system contains shared resources R = {R1, R2, ..., R...} l}, which can be shared by different tasks on the same processor and tasks on different processors; at any given time, each resource can only be accessed by the critical section of one task, that is, access to resources is exclusive and non-preemptive, and a critical section of a task can only access one resource.

[0014] Task τ i There are n i There are ξ key regions, where the j-th key region is denoted as ξ. i，j ξ i，j The worst execution time is C i，j ,r(ξ i，j ) represents the resource identifier that needs to be accessed in the mission's critical area, r(ξ) i，j ) = a represents task τ i Resource R needs to be accessed a 1≤a≤l; To prevent data inconsistency issues caused by task resource access during system mode transitions, all critical areas of low-criticality tasks must be in C i (HI) must be completed before all critical areas of high-critical-level tasks are in C. i Completed before (LO).

[0015] Preferably, each processor starts in low mode, and only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to be executed. In low mode, the execution time of a task will not exceed C. i (LO), specifically including:

[0016] Task τ i Preemption level λ i Its relative deadline D i Negative correlation, preemptive level λ i Represented by positive integers; a higher-priority task can preempt the execution of a lower-priority task that is currently being executed;

[0017] The priority cap for a resource is a processor-related parameter, where processor φ is the priority level. k Any of the above resources R a The priority limit is the amount of resource R used on this processor. a The highest preemption level among all tasks is calculated using the following formula:

[0018]

[0019] System dynamic priority limit Ω k Defined as processor φ k The maximum priority limit among all resources occupied at a given moment is defined as follows:

[0020]

[0021] System dynamic priority limit Ω k The update rules are as follows:

[0022] 1) No task accesses shared resources before the task set begins execution. k =0;

[0023] 2) Task τ i In processor φ k During execution, access processor φ k Local Resources R a Ω k =ceil k (R a );

[0024] 3) Task τ i In processor φ k When executed, it accesses the system global resource R. a Ω k =ceil k (R a )+N, where N is the number of tasks in the task set;

[0025] 4) Task τ i If the required system global resource cannot be accessed, it will be added to the FIFO queue waiting for the resource to be released;

[0026] 5) Task τ i End of resource R a Ω k It will revert to its original value, resource R a It is accessed by other tasks in the FIFO until no task accesses resource R. a He was only released after that;

[0027] For a task to be executed, two conditions must be met: 1) It must have the highest priority; the task's priority is determined by the earliest deadline priority method; 2) The task's preemption level must be higher than the current system priority cap, i.e., satisfying the following formula:

[0028] λ i >Ω k (3)

[0029] The earliest deadline-first method determines priority based on the absolute deadline and arrival time of each task; the absolute deadline equals the arrival time plus the relative deadline D. i The earlier the absolute deadline, the higher the task priority; when two tasks have the same absolute deadline, the priority depends on their arrival time, the earlier the arrival time, the higher the priority; when both the absolute deadline and the arrival time are the same, the task with the smaller index has the higher priority.

[0030] Based on the system's real-time requirements, every task must complete before its worst-case execution time; therefore, the execution time of a task in low-mode must be less than or equal to C. i (LO).

[0031] Preferably, at time ts, if any high-criticality task has been executed on the processor for more than C... i If (LO) is not terminated, all processors in the system will immediately switch to high mode, specifically including:

[0032] High-criticality tasks require high reliability and security. If a high-criticality task has been executed on the processor for more than C... i (LO) and there is no termination, indicating that the real-time requirements of high-critical-level tasks will be difficult to meet in low mode. Therefore, it is set that all processors in the system will immediately switch to high mode.

[0033] Preferably, at time ts, the execution has not been completed and the execution time has exceeded C. i Low-criticality tasks (HI) will be terminated, but the execution time already completed has not reached C. i (HI) low-criticality tasks will continue to be executed up to C. i (HI), the execution time of high-critical-level tasks that have not been completed will be reduced from C. i (HI) upgraded to C i (HI), specifically including:

[0034] Low-criticality tasks can tolerate a certain degree of error; therefore, in time series ts, if a low-criticality task has not completed execution and its execution time exceeds C... i (HI) indicates that the task result is meaningless, and to avoid affecting subsequent tasks, such tasks will be terminated; while tasks that have already been executed for less than C hours will be terminated. i Low-critical tasks in (HI) will continue to execute in high mode up to C. i (HI), if execution reaches C i If (HI) is not completed, execution will be terminated; the execution time of high-critical tasks that have not been completed will be reduced from C. i (HI) upgraded to Ci (HI).

[0035] Preferably, in high mode, only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to be executed, and the execution time of tasks in high mode will not exceed C. i (HI), specifically including:

[0036] In high-level mode, task execution also satisfies two conditions: 1) It has the highest priority; the task's priority is determined by the earliest deadline priority method; 2) The task's preemption level is higher than the current system priority cap, i.e., it satisfies λ. i >Ω k ;

[0037] The earliest deadline-first method determines priority based on the absolute deadline and arrival time of each task; the absolute deadline equals the arrival time plus the relative deadline D. i The earlier the absolute deadline, the higher the task priority; when two tasks have the same absolute deadline, the priority depends on their arrival time, the earlier the arrival time, the higher the priority; when both the absolute deadline and the arrival time are the same, the task with the smaller index has the higher priority.

[0038] Based on the system's real-time requirements, every task must complete before its worst-case execution time; therefore, the execution time of a task in processor high mode will not exceed C. i (HI).

[0039] As can be seen from the above description of the present invention, compared with the prior art, the present invention has the following beneficial effects:

[0040] (1) The method of the present invention can meet the real-time requirements of non-precise multiprocessor hybrid critical systems;

[0041] (2) The method of the present invention can ensure mutually exclusive access to resources in non-precise multiprocessor hybrid critical systems and improve resource utilization.

[0042] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments, but the method for accessing non-precise multiprocessor hybrid critical system resources of the present invention is not limited to the embodiments. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating the method of the present invention. Detailed Implementation

[0044] The technical solutions in the embodiments of the present invention will be described and discussed in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0045] See Figure 1 As shown, the present invention provides a method for accessing inaccurate multiprocessor hybrid critical system resources, comprising the following steps.

[0046] Step 1: Establish a non-precise multiprocessor hybrid critical system resource sharing model.

[0047] The system contains M independent processors with identical architectures, Φ = {φ} k |1≤i≤M};Inaccurate hybrid critical sporadic real-time tasks Γ={τ} on multiprocessors, considering implicit deadlines. i |1≤j≤N}, where N represents the number of inaccurate mixed critical and sporadic real-time tasks. This task set is scheduled using a partitioned earliest-due-date-first strategy. The basic idea of ​​this strategy is to divide the real-time tasks into several independent subsets, each assigned to a different processor. On each processor, tasks are scheduled according to the earliest-due-date-first method; tasks with earlier deadlines have higher priority. When tasks have the same deadline, their priority is determined by their arrival time, with earlier arrival times having higher priority. Higher-priority tasks are scheduled first. These tasks are represented by a quintuple τ. i = <L i C i (LO), C i (HI), D i T i >;where L i L represents the key level of the task. i Represented as L i ={LO, HI}, Low-criticality task L i =LO, High-criticality level task L i =HI;C i (LO) and C i (HI) represent the worst execution time of the task in low mode and high mode, respectively; D i Represents task τ i The relative deadline for an implicitly deadline, or the relative deadline for a non-precise multiprocessor hybrid critical system, is equal to the period, i.e., D. i =T i ;T i Indicates the task's duration; if the mixed critical-occurrence task is a low-criticality task, then C i (LO)≥C i (HI); If the mixed critical-occurrence task is a high-criticality task, then C i (LO)≤C i (HI);

[0048] The system contains shared resources R = {R1, R2, ..., R...}l This resource is shared by different tasks on the same processor and by tasks on different processors. At any given time, each resource can only be accessed by the critical section of one task. In other words, access to resources is exclusive and non-preemptible, and a task's critical section can only access one resource.

[0049] Task τ i There are n i There are ξ key regions, where the j-th key region is denoted as ξ. i，j ξ i，j The worst execution time is C i，j ,r(ξ i，j ) represents the resource identifier that needs to be accessed in the mission's critical area, r(ξ) i，j ) = a represents task τ i Resource R needs to be accessed a 1≤a≤1; To prevent data inconsistency caused by task resource access during system mode transitions, all critical areas of low-criticality tasks must be in C i (HI) must be completed before all critical areas of high-critical-level tasks are in C. i Completed before (LO).

[0050] Step 2: Each processor starts in low mode. Only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to execute. In low mode, the execution time of a task will not exceed C. i (LO).

[0051] Task τ i Preemption level λ i Its relative deadline D i Negative correlation, i.e. Preemption Level λ i Represented by positive integers; a higher-priority task can preempt the execution of a lower-priority task that is currently being executed;

[0052] The priority cap for a resource is a processor-related parameter, where processor φ is the priority level. k Any of the above resources R a The priority limit is the amount of resource R used on this processor. a The highest preemption level among all tasks is calculated using the following formula:

[0053]

[0054] System dynamic priority limit Ω k Defined as processor φ k The maximum priority limit among all resources occupied at a given moment is defined as follows:

[0055]

[0056] System dynamic priority limit Ω k The update rules are as follows:

[0057] 1) No task accesses shared resources before the task set begins execution. k =0;

[0058] 2) Task τ i In processor φ k During execution, access processor φ k Local Resources R a Ω k =ceil k (R a );

[0059] 3) Task τ i In processor φ k When executed, it accesses the system global resource R. a Ω k =ceil k (R a )+N, where N is the number of tasks in the task set;

[0060] 4) Task τ i If the required global system resource cannot be accessed, it will be added to the FIFO (First In First Out) queue waiting for the resource to be released;

[0061] 5) Task τ i End of resource R a Ω k It will revert to its original value, resource R a It is accessed by other tasks in the FIFO until no task accesses resource R. a Only then was it released;

[0062] For a task to be executed, two conditions must be met: 1) It must have the highest priority; the task's priority is determined by the earliest deadline priority method; 2) The task's preemption level must be higher than the current system priority cap, i.e., satisfying the following formula:

[0063] λ i >Ω k (3)

[0064] The earliest deadline-first method determines priority based on the absolute deadline and arrival time of each task; the absolute deadline equals the arrival time plus the relative deadline D. iThe earlier the absolute deadline, the higher the task priority; when two tasks have the same absolute deadline, the priority depends on their arrival time, the earlier the arrival time, the higher the priority; when both the absolute deadline and the arrival time are the same, the task with the smaller index has the higher priority.

[0065] Based on the system's real-time requirements, every task must complete before its worst-case execution time; therefore, the execution time of a task in low-mode must be less than or equal to C. i (LO).

[0066] Step 3, at time ts, if any high-criticality task has been executed on the processor for more than C... i If (LO) is not terminated, all processors in the system will immediately switch to high mode.

[0067] High-criticality tasks require high reliability and security. If a high-criticality task has been executed on the processor for more than C... i (LO) and there is no termination, indicating that the real-time requirements of high-critical-level tasks will be difficult to meet in low mode. For real-time considerations, we set all processors in the system to switch to high mode immediately at this time.

[0068] Step 4, in ts, execution has not been completed and the execution time has exceeded C. i Low-criticality tasks (HI) will be terminated, but the execution time already completed has not reached C. i (HI) low-criticality tasks will continue to be executed up to C. i (HI), the execution time of high-critical-level tasks that have not been completed will be reduced from C. i (LO) upgraded to C i (HI).

[0069] Low-criticality tasks can tolerate a certain degree of error, therefore in t s If low-critical-level tasks have not been completed and the execution time has exceeded C... i (HI) indicates that the task result is meaningless, and to avoid affecting subsequent tasks, such a task will be terminated. Furthermore, the execution time already completed has not reached C. i Low-critical tasks in (HI) will continue to execute in HI mode up to C. i (HI), if execution reaches C i If (HI) is not completed, execution will be terminated; the execution time of high-critical tasks that have not been completed will be reduced from C. i (LO) upgraded to C i (HI) to meet system performance requirements.

[0070] Step 5: In high mode, only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to be executed. In high mode, task execution time will not exceed C. i (HI).

[0071] In high-level mode, task execution also satisfies two conditions: 1) It has the highest priority; the task's priority is determined by the earliest deadline priority method; 2) The task's preemption level is higher than the current system priority cap, i.e., it satisfies λ. i >Ω k ;

[0072] The earliest deadline-first method determines priority based on the absolute deadline and arrival time of each task; the absolute deadline equals the arrival time plus the relative deadline D. i The earlier the absolute deadline, the higher the task priority; when two tasks have the same absolute deadline, the priority depends on their arrival time, the earlier the arrival time, the higher the priority; when both the absolute deadline and the arrival time are the same, the task with the smaller index has the higher priority.

[0073] Based on the system's real-time requirements, any task must complete before its worst-case execution time (WCET). Therefore, the execution time of a task in processor HI mode will not exceed C... i (HI).

[0074] In this embodiment, the relevant parameters of the task set are shown in Table 1. The system has 2 processors and 6 tasks, of which tasks τ1, τ2, and τ3 are executed on processor φ1, and tasks τ4, τ5, and τ6 are executed on processor φ2. Task τ1 uses local resource R1, task τ2 does not use any resources, task τ3 uses local resource R1 and global resource R2, tasks τ4 and τ6 do not use any resources, and task τ3 uses global resource R2. From the table, we can see that ceil1(R1) = 3, ceil1(R2) = 1, and ceil2(R2) = 2.

[0075] Table 1. Example Task Set

[0076]

[0077] At time 0, the system starts in low mode. On processor φ1, Ω1 = 0, and the highest priority task to be executed first is τ1. τ1 uses local resource R1, therefore Ω1 = 3. At time 2, task τ1 finishes execution and releases resource R1. Simultaneously, at time 0, on processor φ2, Ω2 = 0, and task τ4 finishes execution first, ending at time 5. Task τ2 on processor φ1 starts execution at time 2 and ends at time 4. Simultaneously, task τ3 begins execution, using the system's global resource R2. Therefore, the dynamic priority cap of the subsystem containing processor φ1 changes. The change is that Ω1 = 1 + 6 = 7 > λ1; therefore, task τ1, despite having the highest priority, is still blocked by task τ3 when time 5 arrives; at time 6, task τ3 releases resource R2, Ω1 = 0 < λ1, at this time task τ1 preempts task τ3 to execute first and ends at time 8; during the execution of task τ3, task τ5 on processor φ2 needs resource R2 at time 5, but R2 is occupied by task τ3, so task τ5 is blocked; at time 6, task τ5 obtains the system global resource R2, at this time Ω2 = 2 + 6 = 8, at time 8, resource R2 is released, Ω2 = 0;

[0078] At time 8, task τ3 continues execution on processor φ1; at time 9, task τ5 on processor φ2 has exceeded its C5(LO) constraint but has not yet completed execution, so both processors in the entire system switch to high mode. At the same time, the execution time of task τ3 and task τ5 increases to 6 and 4 respectively; the execution time of task τ1, task τ4, and task τ6 is shortened to 1, 3, and 2 respectively; at time 10, tasks τ1, τ2, and τ4 arrive. Since task τ1's absolute deadline is earlier than task τ3's, its priority is higher than task τ3, so it preempts task τ3 and executes it first, ending the process. At time 11, task τ3 ends at time 13; at time 13, task τ2, which arrived at time 10, begins execution and ends at time 16; at time 15, task τ1 arrives, but since task τ1 and task τ2 have the same absolute deadline, but task τ2 arrived earlier than task τ1, task τ1 has a lower dynamic priority and cannot preempt task τ2; correspondingly, tasks τ5 and τ6 on processor φ2 end at times 10 and 11 respectively; task τ4, which arrived at time 10, begins at time 12 and ends at time 15; task τ5, which arrived at time 15, begins at time 15 and ends at time 19.

[0079] The above is merely a preferred embodiment of the present invention. However, the present invention is not limited to the above embodiment. Any equivalent changes and modifications made according to the present invention, provided that the resulting functional effects do not exceed the scope of the present invention, shall fall within the protection scope of the present invention.

Claims

1. A method for accessing resources in a non-precise multiprocessor hybrid critical system, characterized in that, include: Establish a resource-sharing model for an imprecise multiprocessor hybrid critical system; Each processor starts in low mode, where only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to execute. In low mode, task execution time will not exceed C... i (LO); At time ts, if any high-criticality task has been executed on the processor for more than C... i If (LO) is not terminated, all processors in the system will immediately switch to high mode; At time ts, execution has not been completed and the execution time has exceeded C. i Low-criticality tasks (HI) will be terminated, but the execution time already completed has not reached C. i (HI) low-criticality tasks will continue to be executed up to C. i (HI), the execution time of high-critical-level tasks that have not been completed will be reduced from C. i (LO) upgraded to C i (HI); In high mode, only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to be executed. In high mode, task execution time will not exceed C. i (HI); The establishment of a non-precise multiprocessor hybrid critical system resource sharing model includes: The system contains M independent processors with identical architectures. Inaccurate hybrid critical sporadic real-time tasks on multiprocessors, considering implicit deadlines. , This represents the number of inaccurate mixed critical and sporadic real-time tasks. A partitioned earliest-due-date first (ETDAD) strategy is used to schedule this task set. The basic idea of ​​this strategy is to divide the real-time tasks into several independent subsets, each assigned to a different processor. On each processor, tasks are scheduled according to the earliest-due-date first method; tasks with earlier deadlines have higher priority. When tasks have the same deadline, their priority is determined by their arrival time, with earlier arrival times having higher priority. Higher-priority tasks are scheduled first. These tasks are represented by quintuples. ;in Indicates the key level of the task. Represented as , This indicates a low-criticality task. Indicates a high-critical-level task; and These represent the worst execution time for the task in low mode and high mode, respectively; Indicates task The relative deadline, for an implicitly deadline-based non-precise multiprocessor hybrid critical system, is equal to the period, i.e. = ; Indicates the task's duration; if the mixed critical-occurrence task is a low-criticality task, then... If the mixed critical-occurrence task is a high-criticality task, then ; Shared resources exist in the system It is shared by different tasks on the same processor and tasks on different processors; at any given time, each resource can only be accessed by the critical section of one task, that is, access to resources is exclusive and non-preemptive, and a critical section of a task can only access one resource. Task have There are 1 key areas, among which the 1st The key areas are represented as , The worst execution time is , A marker for resources that need to be accessed in the mission's critical area. Indicates task Resources need to be accessed ,1≤ ≤l; To prevent data inconsistency issues caused by task resource access during system mode transitions, all critical areas of low-criticality tasks must be in Previously completed, all critical areas of high-criticality tasks must be in Completed previously; Each processor starts in low mode, where only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to be executed. In low mode, task execution time will not exceed [a certain limit]. ,include: Task level of capture Its relative deadline Negative correlation, preemptive level Represented by positive integers; a higher-priority task can preempt the execution of a lower-priority task that is currently being executed; The priority cap for a resource is a processor-related parameter. Any previous resource The priority limit is the amount of resources used on this processor. The highest preemption level among all tasks is calculated using the following formula: (1) System dynamic priority cap Defined as processor The maximum priority limit among all resources occupied at a given moment is defined as follows: (2) System dynamic priority cap The update rules are as follows: 1) No task accessed the shared resource before the task set began execution. ; 2) Task In processor During execution, access the processor Local resources , ; 3) Task In processor When executed, it accesses global system resources. , ,in The number of tasks in the task set; 4) Task If the required system global resource cannot be accessed, it will be added to the FIFO queue waiting for the resource to be released. among; 5) Task End of resource Access, It will revert to its original value, resources Accessed by other tasks in the FIFO, directly No task accessing resources He was only released after that; For a task to be executed, two conditions must be met: 1) It must have the highest priority; the task's priority is determined by the earliest deadline priority method; 2) The task's preemption level must be higher than the current system priority cap, i.e., satisfying the following formula: (3) Based on the system's real-time requirements, every task must complete before its worst-case execution time; therefore, the execution time of a task in low-mode must be less than or equal to... .

2. The method for accessing resources in a non-precise multiprocessor hybrid critical system as described in claim 1, characterized in that, At time ts, if any high-criticality task has been executed on the processor for more than [time value missing]... If execution is not terminated, all processors in the system will immediately switch to high mode, specifically including: Highly critical tasks require high reliability and security; if a high-critical task has been executed on the processor for more than [time limit], [further action may be required]. The fact that execution did not terminate indicates that the real-time requirements of high-critical-level tasks would be difficult to meet in low mode. Therefore, it is set that all processors in the system will immediately switch to high mode at this time.

3. The method for accessing resources in a non-precise multiprocessor hybrid critical system as described in claim 1, characterized in that, At time ts, execution has not been completed and the execution time has exceeded C. i Low-criticality tasks (HI) will be terminated, but the execution time already completed has not reached C. i (HI) low-criticality tasks will continue to be executed up to C. i (HI), the execution time of high-critical-level tasks that have not been completed will be reduced from C. i (HI) upgraded to C i (HI), specifically including: Low-criticality tasks can tolerate a certain degree of error; therefore, in time series ts, if a low-criticality task has not completed execution and its execution time exceeds C... i (HI) indicates that the task result is meaningless, and to avoid affecting subsequent tasks, such tasks will be terminated; while tasks that have already been executed for less than C hours will be terminated. i Low-critical tasks in (HI) will continue to execute in high mode up to C. i (HI), if execution reaches C i If (HI) is not completed, execution will be terminated; the execution time of high-critical tasks that have not been completed will be reduced from C. i (HI) upgraded to C i (HI).

4. The method for accessing resources in a non-precise multiprocessor hybrid critical system as described in claim 1, characterized in that, In high mode, only tasks with the highest priority and a preemption level higher than the current system priority cap are allowed to be executed. In high mode, task execution time will not exceed C. i (HI), specifically including: In high-level mode, task execution also meets two conditions: 1) It has the highest priority; the task's priority is determined by the earliest deadline priority method; 2) The task's preemption level is higher than the current system priority cap, i.e., it satisfies... ; Based on the system's real-time requirements, every task must complete before its worst-case execution time; therefore, the execution time of a task in processor high mode will not exceed C. i (HI).

Images