CN102306112B - Method for improving scheduling flexibility and resource utilization rate of automotive open system architecture operating system (AUTOSAR OS) based on Contract - Google Patents

Abstract

The invention relates to a contract-based method for improving AUTOSAR OS scheduling flexibility and resource utilization. An abstraction layer is added to the original AUTOSAR OS software platform structure. The abstraction layer is composed of four modules, which are core modules core , Dynamic reclamation module, Hierarchical scheduling module and Spare capacity sharing module; this method applies the contract model to AUTOSAROS, making it easier for AUTOSAR OS to meet the complex time requirements put forward by users, and at the same time After introducing the contract model, AUTOSAROS can also support dynamically changing time requirements. Compared with the prior art, the present invention has the following beneficial effects: 1. Based on the contract model, it can meet the complex time requirements of users. 2. Ensure the minimum resource requirements and schedulability of the system through resource reservation and online/offline schedulability analysis. 3. Obtain better resource utilization by reclaiming and reallocating idle processor capacity. 4. Users can freely choose the most appropriate scheduling algorithm to schedule corresponding tasks.

Description

Method of Improving AUTOSAR OS Scheduling Flexibility and Resource Utilization Based on Contract

technical field

The invention relates to a resource scheduling mechanism of an embedded real-time operating system, in particular to a contract-based method for improving AUTOSAR OS scheduling flexibility and resource utilization.

Background technique

The contract model is the interface between the application layer and the scheduling layer. It proposes a higher-level abstraction, so that developers only need to formulate the needs of the application without associating these needs with the parameters of a specific scheduling algorithm. AUTOSAR (AUTomotive Open System Architecture) is a standardized open automotive software architecture. AUTOSAR OS is an embedded real-time operating system that conforms to the AUTOSAR standard, and is a statically configured multi-task single-processor RTOS. The AUTOSAR standard referred to in the present invention is version 3.0.1, and it is realized on the SmartOSEK OS, an OS conforming to the AUTOSAR standard independently developed by Zhejiang University.

In the field of real-time systems, the traditional concept holds that RTOS only needs to meet the time requirements stipulated in the automotive functional requirements specification. In fact, in many occasions nowadays, soft real-time tasks in RTOS are more common than hard real-time tasks. Furthermore, meeting those timing requirements has become relatively easy as the processing power of microprocessors used in the automotive field continues to increase. However, since more and more functional units are integrated into one processor, the problem of dividing processor resources becomes more and more important. Therefore, in the implementation process of most AUTOSAR OS at present, in order to avoid conflicts between different tasks due to resource occupation, developers can only apply for the minimum resource during configuration. Under such a configuration premise, the overall utilization rate of processor resources will be relatively low. After the contract model is introduced, several tasks can be assigned to a contract, so that contract and contract can dynamically reclaim idle processor resources and allocate them to the server corresponding to the busy contract through negotiation. , so as to improve the flexibility and utilization rate of resource scheduling of RTOS.

In the current real-time operating system, there are the following problems: since multiple functional modules are to be mapped to one processor for execution, in order not to cause conflicts between the corresponding tasks of each functional module, it is necessary to reduce the resource usage of each task, resulting in A drop in overall processor resource utilization.

Contents of the invention

The purpose of the present invention is to solve the deficiencies in the above-mentioned technology, and provide a method based on contract to improve AUTOSAR OS scheduling flexibility and resource utilization. The method based on contract is used to improve the flexibility and support of embedded real-time operating system resource utilization.

The technical solution adopted by the present invention to solve the technical problem: the general AUTOSAR OS is statically configured, and this contract-based method for improving the scheduling flexibility and resource utilization of the AUTOSAR OS applies the contract model to the AUTOSAR OS, so that the AUTOSAR OS is suitable for the AUTOSAR OS The complex time requirements put forward by users are easier to meet. At the same time, after introducing the contract model, AUTOSAR OS can also support dynamically changing time requirements.

A layer of abstraction is added to the original AUTOSAR OS software platform structure. The abstraction layer consists of four modules, namely the core module core, the dynamic reclamation module Dynamic reclamation, the hierarchical scheduling module Hierarchical scheduling and the spare capacity sharing module Spare capacity sharing ;

(1) Specify the properties of the contract and related operations through the core module. The properties of the contract are some parameters, which are used to create a contract, negotiate between contracts, bind tasks to servers and RTOS resources to ensure minimum resource requirements reservation mechanism;

(2) The dynamic recovery module is a functional module, which executes dynamic recovery when there is idle processor capacity in the system through the dynamic recovery module, and allocates the suspended idle processor capacity that is not used by the server through the idle capacity sharing module; if it is running When the task is abnormally terminated or completed before the scheduling point, additional idle processing capacity is generated. According to the configuration of the contract, through the cooperation of the two modules of dynamic recycling and idle capacity sharing, these idle processing capacity will be allocated to different servers that require additional processing power.

(3) Global scheduling and local scheduling are carried out through the hierarchical scheduling module. Hierarchical scheduling mainly occurs at two levels—global scheduling and local scheduling. As shown in Figure 2, the global scheduling is used to determine which server can access processor resources, and the local scheduling is the internal scheduling of the server, which is used to determine which task is scheduled for execution.

In the present invention, the method can assign several tasks to a certain contract, and configure the properties of the contract. These contract properties include creating a contract, negotiating between contracts, binding tasks to servers, and corresponding resource reservation mechanisms. parameter. The minimum resource requirements of AUTOSAR OS are guaranteed through the resource reservation mechanism, and the schedulability of OS tasks is guaranteed through the contract negotiation and renegotiation mechanism (online and offline schedulability analysis).

In the present invention, the method introduces a contract model, and each multiple tasks corresponds to a contract, and the scheduling policy of each contract is configured. In this way, users can choose the most suitable scheduling strategy to meet their needs from many advanced real-time scheduling algorithms currently implemented, and can specify different scheduling strategies for different tasks.

Preferably, when the contracts are ready, these contracts need to be negotiated to evaluate whether they can pass the schedulability test. The system will check whether there are enough resources to satisfy the newly added contract under the condition of ensuring the normal operation of the previous contracts. The minimum resource requirement of the contract; if the negotiation is passed, the system will reserve enough resource budget to meet the minimum resource requirement of the newly added contract, and will create a new server corresponding to the contract. The server is the software object of the contract in the system running state. It stores All the information of the contract; the server will manage the execution conditions of the tasks bound to it, as long as the task is bound to a certain server, the hierarchical scheduler will schedule its execution. In Figure 2, the global scheduler is fixed priority scheduling (FP) based on sporadic server, and there are three local schedulers, and the implemented algorithms are EDF, FP and FIFO.

Compared with the prior art, the present invention has beneficial effects as follows:

1. Based on the contract model, it can meet the complex time requirements of users.

2. Ensure the minimum resource requirements and schedulability of the system through resource reservation and online/offline schedulability analysis.

3. Obtain better resource utilization by reclaiming and reallocating idle processor capacity.

4. Users are free to choose the most appropriate scheduling algorithm to schedule corresponding tasks.

Description of drawings

Fig. 1 is a schematic diagram of the SmartSAR software platform architecture before applying the present invention;

Fig. 2 is a schematic diagram of a hierarchical scheduling model;

Fig. 3 is the SmartSAR software platform architecture after applying the present invention;

Figure 4 is the execution process of the task in SmartOSEK OS.

Detailed ways

Below in conjunction with accompanying drawing and embodiment the present invention will be further described:

The present invention is realized on the RTOS——SmartOSEK OS, which conforms to the AUTOSAR standard independently developed by Zhejiang University. SmartOSEK OS is closely connected with other modules of the SmartSAR software platform. Before the introduction of the present invention, the software platform architecture of SmartSAR is shown in Figure 1. After the introduction of the present invention, another sub-module is added in the SmartOSEK OS module as shown in the attached Figure 3 shows.

After introducing the present invention to SmartOSEK, the defined data structure is as follows:

Table 1. Properties of the contract structure

parameter name Parameter Description max_period The maximum period of the server. min_budget The minimum execution capability of the server in each cycle. deadline Server deadline D=T Whether the server's deadline is equal to the period. budget_overran The budget for the currently running task has expired. deadline_missed The current task has exceeded its deadline. granularity Indicates that additional spare capacity needs to be used and how: continuous, discrete, or none. min_period The minimum period of the server. max_budget The maximum budget available. utilization_set A collection of {bdget,period} ordinal pairs, used for discrete idle capacity allocation. importance Fixed priority for continuous allocation of free capacity. quality In continuous idle capacity allocation, the relative proportion allocated by two servers with the same fixed priority. scheduling policy Local scheduler scheduling strategy selection: FP, EDF, FIFO, NON.

Table 2. Additional attributes of the server structure

parameter name Parameter Description priority Scheduling parameters. vres_id id of the API. task Bind to the tasks on the server. task_num The number of tasks bound to the server. task_finished_num The number of tasks to complete in one cycle. sporadic server The Sporadic server is assigned to this server. spare_capacity Free capacity allocated by the system. used_capacity The capacity that the server has already used.

Table 3. sporadic server structure

parameter name Parameter Description replenishment queue An array to store replenishment times and the amount of free processor power. Alarm [2] Indicates which alarm expires and consumes processor power that can be replenished. remaining budget The accumulated budget that is not used on the server. budget start When the sporadic server enters the active state from the idle state, the task starts to execute and the processor power starts to be consumed.

Table 1 lists the attributes of the contract in SmartOSEK OS and the description of each attribute. Since the server basically contains all the information of the contract, Table 2 only lists the attributes and corresponding descriptions other than the contract. It should be noted that some modules are extensible, and attributes belonging to these modules are also extensible. In AUTOSAR OS, server can be regarded as a special task to be scheduled by the global scheduler. The present invention uses sporadic servers based on fixed priorities to schedule servers. Table 3 describes the sporadic server used by each server.

It can be seen from FIG. 3 that the original OS kernel will not be changed after the introduction of the present invention, and the user can determine whether to use the functions of the present invention through a macro switch in the OS. The detailed processing flowchart of the present invention in SmartOSEK is shown in accompanying drawing 4.

The specific implementation steps are as follows:

(1) According to the application requirements, configure the properties of the contract defined in Table 1 to generate a contract.

(2) Negotiate between each contract to judge the schedulability of the system. If it is not schedulable, execute (3), if it is schedulable, execute (4)

(3) An error is returned, the system cannot be scheduled, and the request to add the contract is rejected

(4) Create a new server corresponding to the contract and set the corresponding sporadic server

(5) Dynamically recover idle processor capacity and reserve processor resources

(6) Bind the task to the corresponding server with processor capability

(7) The OS starts to schedule tasks assigned to different servers. The Sporadic server uses a fixed priority scheduling algorithm to schedule different servers assigned to it, and different servers use different scheduling strategies (EDF, FP , FIFO) schedules tasks assigned to it.

(8) After all tasks are executed, the OS then dynamically reclaims processor resources and reallocates them

Although the present invention has been described by the above specific examples, the present invention is not limited to the above examples. For those skilled in the art, other various changes and modifications can be made according to the technical scheme and thought of the present invention, and all of these are It should belong to the protection scope of the claims of the present invention.

Claims (2)

1. A method for improving AUTOSAR OS scheduling flexibility and resource utilization based on contract, characterized in that: an abstraction layer is added on the original AUTomotive Open System Architecture OS software platform structure, and the abstraction layer is composed of four modules , which are the core module core, the dynamic reclamation module Dynamic reclamation, the hierarchical scheduling module Hierarchical scheduling and the spare capacity sharing module Spare capacity sharing; (1) Specify the properties of the contract and related operations through the core module. The properties of the contract are some parameters, which are used to create a contract, negotiate between contracts, bind tasks to the server and the embedded real-time operating system RTOS to ensure Resource reservation mechanism for minimum resource requirements; (2) Use the dynamic recovery module to perform dynamic recovery when there is idle processor capacity in the system, and use the idle capacity sharing module to allocate the suspended idle processor capacity that is not used by the server; if during runtime, due to abnormal termination of the task or It is completed before the scheduling point and generates additional idle processor capacity, which will be allocated to different servers that require processor capacity according to the method specified in the contract; (3) Global scheduling and local scheduling are performed through the hierarchical scheduling module. The global scheduling is used to determine which server can access processor resources, while the local scheduling is the internal scheduling of the server, which is used to determine which task to schedule for execution; When the contracts are ready, these contracts need to be negotiated to evaluate whether they can pass the schedulability test. The system will check whether there are enough resources to meet the minimum resources of the newly added contract under the condition of ensuring the normal operation of the previous contracts. Requirements; if the negotiation is passed, the system will reserve enough resource budget to meet the minimum resource requirements of the newly added contract, and will create a server corresponding to the contract at the same time. The server is the software object of the contract in the system running state, which stores the contract All information; the server will manage the execution conditions of the tasks bound to it, as long as the task is bound to a certain server, the hierarchical scheduler will schedule its execution; The contract refers to the interface contract model between the application layer and the scheduling layer. 2. the method for improving AUTOSAR OS scheduling flexibility and resource utilization based on contract according to claim 1, characterized in that: the method introduces a contract model, every multiple tasks correspond to a contract, for each contract Scheduling policies are configured.

Images