JP2009181474A - Priority controller for input/output request, priority control method for input/output request, and priority control program for input/output request - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve priority control for preventing the delay of the response of a real time OS, and for suppressing the deterioration of a through-put of a general-purpose OS in a hybrid OS. <P>SOLUTION: The priority controller of an input/output request includes a priority management table; a priority control part; and input/output request queue. The priority management table is configured to determine the size of an input/output block which is permitted for a second input/output request to be issued to a general-purpose operating system in response to the priority of a first input/output request to be issued to a real time operating system. The priority control part is configured to receive the first input/output request and the second input/output request, and to divide the second input/output request based on the priority management table, and the input/output queue is configured to enqueue the first input/output request and the second input/output request divided by the priority control part in the order of priority. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an input / output request priority controller, an input / output request priority control method, and an input / output request priority control program.

  A real-time OS (Operating System) that performs a certain response within a predetermined time is known. Real-time OSs are used in car navigation systems, digital cameras, information appliances, mobile phone embedded systems, and the like. In particular, in an embedded system, in recent years, in addition to real-time processing, it has been demanded to increase the functionality of the system such as performing multimedia processing that handles audio data and moving image data.

  A hybrid OS is known as one method for realizing high functionality. The hybrid OS is an OS having two types of OSs, a real-time OS that performs real-time processing and a general-purpose OS that performs multimedia processing. By hybridizing the OS, it is possible to realize a system having the features of two OSs such as the high real-time property and high reliability of the real-time OS and the high scalability and high throughput of the general-purpose OS.

  When hybridizing an OS in an embedded system, it is difficult to provide a disk for each OS due to space-saving restrictions. Therefore, it is necessary to share a disk between OSs. However, when a shared disk is provided, the input / output processing of the real-time OS may be hindered by the input / output processing of the general-purpose OS, and the response may decrease.

  In addition, in general-purpose OSs, the input / output block size for input / output requests tends to be larger than the input / output block size for input / output requests of the real-time OS because throughput is important. If the input / output block size is large, the execution time of the input / output processing per input / output request becomes long. Therefore, the input / output processing of the general-purpose OS cannot start the input / output processing of the real-time OS, and a delay occurs with considerable frequency. End up.

  Japanese Patent Application Laid-Open No. 11-39246 (see Patent Document 1) describes a disk access system invention capable of dividing a normal priority I / O block. In this disk access system invention, when an I / O request having a higher priority than an I / O request currently being processed is generated, the I / O sub-block which is a unit obtained by dividing the I / O block of the disk At the break, the high priority I / O request is interrupted and processed first.

JP-A-11-39246

  An object of the present invention is to perform priority control for preventing a response of a real-time OS from being delayed and suppressing a decrease in throughput of a general-purpose OS in a hybrid OS.

  An input / output request priority controller according to a first aspect of the present invention includes a priority management table, a priority control unit, and an input / output request queue. The priority management table determines the size of the I / O block allowed for the second I / O request issued to the general-purpose operating system according to the priority of the first I / O request issued to the real-time operating system. It is a table to do. The priority control unit receives the first input / output request and the second input / output request, and divides the second input / output request based on the priority management table. The input / output request queue enqueues the first input / output request and the second input / output request divided by the priority control unit in order of priority.

  The input / output request priority control method according to the second aspect of the present invention comprises receiving, determining, dividing, and enqueuing. Receiving receives a first I / O request issued to the real-time operating system and a second I / O request issued to the general purpose operating system. The determination determines the size of the input / output block allowed for the second input / output request in accordance with the priority of the first input / output request. Splitting splits the second I / O request based on its size. Enqueuing enqueues the first input / output request and the second input / output request divided by dividing into the input / output request queue in order of priority.

  The input / output request priority control program according to the third aspect of the present invention causes a computer to execute a receiving procedure, a determining procedure, a dividing procedure, and an enqueuing procedure. The receiving procedure receives a first I / O request issued to the real-time operating system and a second I / O request issued to the general-purpose operating system. The determining procedure determines the size of the input / output block allowed for the second input / output request in accordance with the priority of the first input / output request. The dividing procedure divides the second input / output request based on this size. The enqueueing procedure enqueues the first input / output request and the second input / output request divided by dividing into the input / output request queue in order of priority.

  According to the present invention, in the hybrid OS, it is possible to perform priority control for preventing the response of the real-time OS from being delayed and preventing the throughput of the general-purpose OS from being lowered.

  One of the best modes for carrying out the present invention will be described in detail with reference to the drawings. FIG. 1 shows a configuration diagram in which an input / output request priority controller according to one embodiment is applied to an embedded system. In FIG. 1, the embedded system is equipped with a hybrid OS, and has a real-time OS 3, an inter-OS communication unit 4, and a general-purpose OS 5. A real-time OS / application 1 operates on the real-time OS 3, and a general-purpose OS / application 2 operates on the general-purpose OS 4. These software 1 to 5 are executed by a single processor or a multi-core processor included in the embedded system. In the figure, the embedded system includes only one shared disk 6, and both OSes 3 and 5 use the same shared disk 6.

  In FIG. 1, two tasks 11 and 12 are operating as the real-time OS application 1. Further, as the general-purpose OS / application 2, two processes 21 and 22 are operating. The tasks 11 and 12 issue an input / output request to the input / output receiving unit 31 of the real-time OS 3. On the other hand, the processes 21 and 22 issue an input / output request to the input / output receiving unit 51 of the general-purpose OS 5. When the input / output receiving unit 31 of the real-time OS 3 receives an input / output request from the tasks 11 and 22, it notifies the input / output request priority controller 70 of the input / output request. When receiving the input / output request from the processes 21 and 22, the general-purpose OS input / output receiving unit 51 notifies the priority controller 70 of the input / output request via the inter-OS communication unit 4.

  The priority controller 70 receives an input / output request from the input / output receiving unit 31 or the input / output receiving unit 51. Then, the priority of the input / output request is controlled by dividing the input / output request under a certain condition. The input / output processing unit 35 retrieves an input / output request from the priority controller 70 in the order of priority, and executes input / output to the shared disk 6. When the input / output is completed, the priority controller 70 notifies the input / output request source of the completion of the input / output.

  FIG. 2 is a configuration diagram illustrating an example of a priority controller for input / output requests. As illustrated, the priority controller 70 includes a priority control unit 71, a priority management table 72, and an input / output request queue 73. In FIG. 2, the priority control unit 71 receives an input / output request and enqueues the received input / output request in the input / output request queue 33 in order of priority. At this time, the priority control unit 32 refers to the priority management table and divides the input / output request from the input / output reception unit 51 of the general-purpose OS 5 according to a predetermined algorithm under a certain condition.

  FIG. 3 is an explanatory diagram of a specific example of the priority management table. The priority management table 72 illustrated in FIG. 3 includes two sub tables. In FIG. 3, the first sub-table limits the size of the input / output block for the priority of the task in the second sub-table and the input / output request from the input / output receiving unit 51 of the general-purpose OS 5. It is a correspondence table with partial block size. As the task priority is higher, a smaller partial block size is assigned, and as the task priority is lower, a larger partial block size is assigned. The second sub-table is a correspondence table between the task that issued the input / output request enqueued in the input / output request queue 33 and the priority of the task. Here, the priority of the I / O request is equivalent to the priority of the task that issued the I / O request, and all I / O requests issued from the same task are assigned the same priority. Yes.

  FIG. 4 is an explanatory diagram of an input / output request. As shown in the figure, the input / output request issued by the tasks 11 and 12 and the processes 21 and 22 includes a task identifier (or process identifier) that identifies the task (or process) of the issuer and the priority of the task (or process). And information about I / O blocks. The input / output block column includes information for specifying input / output data such as addresses from # 0 to # 9.

  FIG. 5 is a flowchart for explaining the operation of the priority control unit. When the tasks 11 and 12 of the real-time OS / application 1 or the processes 21 and 22 of the general-purpose OS / application 2 issue an input / output request, the priority control unit 71 receives the input / output reception unit 31 or the input / output reception unit 51. The input / output request is received (step A1). The priority control unit 71 confirms whether or not the input / output request is issued from the processes 21 and 22 of the general-purpose OS / application 2 (step A2).

  In the case of an input / output request from the general-purpose OS 5 (step A2: YES), the priority control unit 71 refers to the priority management table 72 (step A3). Here, the priority control unit 71 first extracts the highest priority value of the entered task from the second sub-table (step A4). Next, the partial block size corresponding to the highest priority is extracted from the first sub-table (step A5). When the size of the input / output block column in the input / output request is larger than the extracted partial block size, the priority control unit 71 sets the input / output request from the general-purpose OS 5 so as to be equal to or smaller than the partial block size. Is divided (step A6). Then, the divided input / output requests are enqueued in the input / output request queue 73 in order of priority (step A7).

  In the case of an input / output request issued by the tasks 11 and 12 of the real-time OS / application 1 (step A2: No), the priority control unit 71 does not divide the input / output request, but makes the input / output request in order of priority. The queue is enqueued (step A7). At this time, the priority control unit 71 adds an entry having the task identifier and priority of the task that issued the input / output request to the second sub-table in the priority management table 72.

  In parallel with the operation of the priority control unit, the input / output processing unit 35 dequeues the first input / output request with the highest priority from the input / output request queue 73 and executes input / output to the shared disk 6. When the input / output is completed, the priority control unit 71 deletes the entry related to the input / output request from the second sub-table in the priority management table 72. At the same time, it notifies the task 11 and task 12 of the real-time OS / application 1 that is the source of the input / output request or the process 21 and process 22 of the general-purpose OS / application 2 of the completion of input / output.

  In the operation example of FIG. 5, the priority is acquired by referring to the second sub-table, the highest rank of these priorities is calculated, and then the first sub-table is referred to. As another example of operation, the priority is obtained by referring to the second sub-table, and the partial block size is obtained for each of these priorities by referring to the first sub-table. In addition, the minimum value of these partial block sizes may be calculated.

  As described above, in the present embodiment, the input / output request from the general-purpose OS 5 is divided into a plurality of input / output requests having input / output blocks having a partial block size that is limited by the priority management table 72. . If the division is performed, the input / output execution time per input / output request is shortened, so that the start of input / output of the real-time OS is not delayed.

  In general, the input / output of a real-time OS requires higher response as the task priority is higher. In this embodiment, the partial block size that limits the size of the input / output block in the input / output request from the general-purpose OS 5 is made variable according to the priority of the task of the real-time OS. Accordingly, when there is a task requiring a very high response among real-time OS tasks, a corresponding high response can be ensured by reducing the partial block size. At the same time, if there is no real-time OS task that requires high responsiveness, a decrease in the throughput of the general-purpose OS can be suppressed by increasing the partial block size.

  As described above, according to the present embodiment, since the input / output request from the general-purpose OS is divided under a certain condition, the input / output execution time per input / output request is shortened. The high response characteristic of the real-time OS can be sufficiently secured. In addition, since the partial block size, which is the maximum size of the input / output block after the division, is variable depending on the priority of the task of the real-time OS, it is possible to suppress a decrease in the throughput of the general-purpose OS.

FIG. 1 is a block diagram for explaining one of the best modes for carrying out the present invention. FIG. 2 is a configuration diagram illustrating an example of a priority controller for input / output requests. FIG. 3 is an explanatory diagram of a specific example of the priority management table. FIG. 4 is an explanatory diagram of an input / output request. FIG. 5 is a flowchart for explaining the operation of the priority control unit.

Explanation of symbols

1, 2 Application 3, 5 OS
4 Inter-OS communication unit 6 Shared disk 11, 12 Task 21, 22 Process 31, 51 Input / output accepting unit 35 Input / output processing unit 70 Input / output request priority controller 71 Priority control unit 72 Priority management table 73 Input / output request queue

Claims (25)

Priority management to determine the I / O block size allowed for the second I / O request issued to the general-purpose operating system according to the priority of the first I / O request issued to the real-time operating system Table,
Priority control means for receiving the first input / output request and the second input / output request, and dividing the second input / output request based on the priority management table;
An input / output request priority controller, comprising: an input / output request queue that enqueues the first input / output request and the second input / output request divided by the priority control means in order of priority. The priority management table includes:
The table for determining the size of an input / output block having a smaller size as the priority is higher and the size of an input / output block having a larger size as the priority is lower. I / O request priority controller. The priority management table includes:
The priority column and a partial block size column that is the size of an input / output block allowed for the second input / output request when there is a first input / output request having the priority. The input / output request priority controller according to claim 2, comprising a first sub-table. The priority control means includes:
According to a predetermined procedure, the highest priority among the priorities of the first input / output request enqueued in the input / output request queue is obtained, and referring to the first sub-table, The input / output request priority controller according to claim 3, wherein a partial block size corresponding to the highest priority is obtained, and the second input / output request is divided so as not to exceed the partial block size. The priority management table includes:
A second sub-field comprising a task identifier field of a task that has issued the first input / output request enqueued in the input / output request queue, and a priority field assigned to the input / output request issued by the task; The input / output request priority controller according to claim 4, further comprising a table. The priority control means includes:
The input / output request priority controller according to claim 5, wherein the highest priority is acquired by referring to the second sub-table. The priority control means includes:
The input / output request priority controller according to claim 6, wherein the second sub-table is updated in response to a first input / output request enqueued in the input / output request queue. Receiving a first I / O request issued to the real-time operating system and receiving a second I / O request issued to the general purpose operating system;
Determining a size of an input / output block allowed for the second input / output request according to a priority of the first input / output request;
Dividing the second I / O request based on the size;
A priority control method for an input / output request, comprising: enqueuing the first input / output request and the second input / output request divided by the division into an input / output request queue in order of priority. The determination is
9. The priority management table for determining a size of an input / output block allowed for the second input / output request according to the priority of the first input / output request is included. Priority control method for I / O requests. The determination is
The higher the priority, the smaller the size of the input / output block is referred to, and the lower the priority, the higher priority is referred to the priority management table for determining the size of the input / output block. The input / output request priority control method according to claim 9. The determination is
The priority column and a partial block size column that is the size of an input / output block allowed for the second input / output request when there is a first input / output request having the priority. The input / output request priority control method according to claim 10, further comprising referring to a priority management table including a first sub-table. The division is
According to a predetermined procedure, the highest priority among the priorities of the first input / output request enqueued in the input / output request queue is obtained, and referring to the first sub-table, The priority of the input / output request according to claim 11, further comprising: obtaining a partial block size corresponding to the highest priority and dividing the second input / output request so as not to exceed the partial block size. Degree control method. The determination is
A second sub-field comprising a task identifier field of a task that has issued the first input / output request enqueued in the input / output request queue, and a priority field assigned to the input / output request issued by the task; The input / output request priority control method according to claim 12, further comprising referring to a priority management table further including a table. The division is
The input / output request priority control method according to claim 13, further comprising: acquiring the highest priority by referring to the second sub-table. The enqueue is
The input / output request priority control method according to claim 14, comprising updating the second sub-table in response to a first input / output request enqueued in the input / output request queue. Receiving a first I / O request issued to the real-time operating system and receiving a second I / O request issued to the general-purpose operating system;
Determining a size of an input / output block allowed for the second input / output request according to the priority of the first input / output request;
Dividing the second I / O request based on the size;
Priority of input / output request for causing a computer to execute a procedure for enqueuing the first input / output request and the second input / output request divided by dividing into an input / output request queue in order of priority Degree control program. The procedure for determining is as follows:
The input / output request according to claim 16, wherein a priority management table for determining a size of an input / output block allowed for the second input / output request is referred to according to the priority of the first input / output request. Priority control program. The procedure for determining is as follows:
The higher the priority, the smaller the size of the input / output block is referred to, and the lower the priority, the higher priority is referred to the priority management table for determining the size of the input / output block. 18. The priority control program for input / output requests according to claim 17. The procedure for determining is as follows:
The priority column and a partial block size column that is the size of an input / output block allowed for the second input / output request when there is a first input / output request having the priority. 19. The input / output request priority control program according to claim 18, wherein the priority management table including the first sub-table is referred to. The procedure for determining is as follows:
Among the priorities of the first input / output request enqueued in the input / output request queue, the highest priority is obtained,
The input / output request priority control program according to claim 19, wherein a partial block size corresponding to the highest priority is obtained with reference to the first sub-table. The procedure for determining is as follows:
A second sub-field comprising a task identifier field of a task that has issued the first input / output request enqueued in the input / output request queue, and a priority field assigned to the input / output request issued by the task; 21. The input / output request priority control program according to claim 20, wherein the priority management table further includes a table. The procedure for determining is as follows:
The input / output request priority control program according to claim 21, further comprising: acquiring the highest priority by referring to the second sub-table. The enqueue procedure is:
The input / output request priority control program according to claim 14, wherein the second sub-table is updated according to a first input / output request enqueued in the input / output request queue. A priority controller for input / output requests according to any one of claims 1 to 7,
An embedded system comprising: a real-time operating system that receives an input / output request controlled by the priority controller; and a shared storage device shared by a general-purpose operating system. An embedded system using the priority control method for input / output requests according to any one of claims 8 to 15.

Images