JP2009238051A - Program conversion device and program conversion method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a program conversion device and program conversion method, capable of meeting a temporal limitation about a process for a specified section of a program, on an preemptive multi-task OS. <P>SOLUTION: A program conversion device 50 for converting a program source includes a time and section code acquiring part 501, which acquires a time code representing a time requirement described outside the program source and a section code 52 representing its objective section; an execution profile acquiring part 502 which acquires an execution profile information 53 of the objective section; a critical section position resolving part 503, which resolves the position of a critical section; a scheduler control means selecting part 504, which selects a means for realizing the critical section by controlling a scheduler; and a scheduler control code inserting converting part 505, which generates a resource preservation program source from a scheduler control code by converting and inserting a program source. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a program conversion apparatus and a program conversion method capable of converting a program in order to control scheduling for a specified section of the program in a preemptive multitasking OS.

  Conventionally, when performing real-time processing with time constraints by a program, the program is often hard-coded in a form that has been finely adjusted through trial and error. For this reason, many of them eventually became craftsmanship programs whose performance requirements and implementation intentions could not be read. Such a craftsmanship program with poor readability has a problem in that it causes a programmer to spend a great deal of labor in maintenance work such as bug handling and performance tuning, thereby reducing the productivity and quality of software.

  In an OS that supports preemptive multitasking functions such as Linux, this is more serious, and the real-time nature of each task must be met while optimizing task scheduling. In the first place, it has been pointed out that in the case of an OS for time sharing, the structure itself is difficult to satisfy the time constraints.

  In recent years, it has become increasingly difficult to guarantee time constraints even in embedded systems where real-time performance is important. In a system using a conventional embedded real-time OS, tasks in the system are often fixed in advance. Therefore, if a skilled person carefully sets a priority relationship that satisfies the time constraint of each task, the violation of the constraint could be avoided.

  However, recently, although it is an embedded system, it is becoming an open system due to digitalization and networking, and its functionality is increasing at an accelerated speed. On such a system, it is very difficult to grasp in advance the type and behavior of an application that operates.

  In order to deal with the above problem, a resource reservation type system has recently been proposed. In the resource reservation type system, it is possible to ensure a necessary processing amount and satisfy a task having a time constraint by surely assigning the execution time of the explicitly set processor.

  For example, Resource Kernel reported in Non-Patent Document 1 makes a reservation by specifying the period and the calculation time to be allocated within that period, and assigning processors according to the reservation allows processing up to the deadline. Ensure that it ends.

  Patent Document 1 describes a method of recording a priority combination of process groups satisfying a time constraint and reproducing a task schedule in the order. However, with these methods, it is difficult to allocate resources in detail because CPU resources are reserved for Linux processes. In coarse-grained control in process units, it is necessary to subdivide the processing into processes in order to reserve CPU resources in a partial section.

  Further, Patent Document 2 describes a method of parallelizing processing programs that are automatically specified so as to satisfy time constraints. However, as a matter of course, the application target is limited to a program whose performance is expected to be improved by parallelization. Therefore, the execution time cannot be largely controlled in a program that does not have sufficient parallelism or has little effect.

  Furthermore, in the case of a system that requires modification of the kernel such as Resource Kernel, portability is impaired because the platform dependency of the system becomes strong. Non-Patent Document 2 proposes QoSWeaver that avoids violation of the time constraint of a specific thread by a program converter without preparing a special platform.

  QoSWeaver inserts a statically appropriate preemption code for a thread program that occupies CPU resources excessively when overloading, especially in J2EE application servers. However, such an approach that indirectly suppresses the influence on a specific thread may unnecessarily give a lot of CPU processing time in order to satisfy the time constraint of a certain section.

JP-A-6-202884 JP 2006-126977 S. Oikawa and R. Rajkumar, “Portable RK: A Portable Resource Kernel for Guaranteed and Enforced Timing Behavior”, IEEE Real-Time Technology and Applications Symposium, 1996 K. Kourai, H. Hibino and K. Kourai, “A Dynamic Aspect-Oriented System for OS Kernels”, 6th International Conference on Aspect-Oriented Software Development, 2007

  As described above, with the conventional technology, it has been difficult to satisfy detailed processing time constraints on a preemptive multitasking OS without being caught by processes and threads. On the other hand, the intended purpose cannot be achieved even if the parallel processing of the program is used to extract a process that is dependent on the order and has no degree of parallelism as a process or thread.

  An object of the present invention is to solve the above-described problems and to provide a program conversion apparatus and a program conversion method that satisfy the time restriction of processing of a specified section of a program on a preemptive multitasking OS.

  In order to achieve the above object, the program conversion apparatus and the program conversion method of the present invention provide a time request and its target section to the program from outside the program, and set the critical section (resource occupancy) to satisfy the time request. It was decided to generate a resource reservation program that controls scheduling by providing a section) and a release section (resource release section).

  Specifically, in the program source, a set of statements that should be the critical section and the release section to satisfy the required time is selected from the group of statements included in the target section of the required time, and the CPU is assigned to the set of statements. By inserting a scheduler control code to occupy, it is guaranteed that the processing is completed within a predetermined request time.

  According to another aspect of the present invention, in a preemptive multitasking OS that processes a plurality of processes at the same time, a program that guarantees that a process having a predetermined time request completes the previous process with the previous time request. A conversion device, a time code indicating a time request described outside the program source and a time and interval code acquisition unit for acquiring the interval code indicating the target interval, and an execution profile acquisition for acquiring execution profile information in the interval A scheduler control position resolution unit for resolving the position of the critical section and the release section, a scheduler control unit selection unit for selecting a means for controlling the scheduler to realize the critical section and the release section, and a scheduler control code as a program source Convert It is obtained by a scheduler control code insertion conversion unit that generates an input shi resource reservation program source.

  According to the present invention, in a preemptive multitasking OS, when processing an arbitrary designated section of a program, a time requirement can be guaranteed by executing an appropriate set of statements as a critical section and a release section. It is possible to automatically generate a simple program.

  Hereinafter, the present invention will be described along specific embodiments shown in the drawings, but the present invention is not limited to the embodiments described below. FIG. 1 is a diagram showing an embodiment of a system 10 incorporating a program conversion apparatus of the present invention.

  FIG. 1 shows a system 10 including a remote controller 20, a digital broadcast antenna 30, a display device 40, a program conversion device 50 of the present invention, and a digital broadcast receiver 60.

  The digital broadcast receiver 60 includes a storage device 61 in which a conversion program is stored, a processor 62, a signal receiver 63 that receives a signal from the remote controller 20, and video data, audio data, and subtitles via the digital broadcast antenna 30. A digital broadcast tuner 64 for receiving data and digital broadcast receiver software 65 are installed, and an OS and applications for managing various resources and making them available to applications are stored.

  The digital broadcast receiver software 65 includes a remote controller process 651, a tuner process 652, a drawing process 653, an electronic program guide process 654, and a data broadcast browser process 655.

  In the embodiment shown in FIG. 1, the digital broadcast receiver software 65 is converted into a resource reservation program by the program conversion device 50 and stored in the storage device 61. The resource reservation program stored in the storage device 61 is read into the OS of the digital broadcast receiver 65 by the processor 62 and executed.

  The processor 62 performs a reception process by the remote controller process 651 and an appropriate response process for a user operation using the remote controller 20. For example, the broadcast station is switched by the tuner process 652 associated with the channel selection, the electronic program guide is acquired and combined by the electronic program guide process 654 when the electronic program guide is viewed, and the data broadcast content is acquired by the data broadcast browse process 655 when the data broadcast is viewed. In addition, the rendering process 653 synthesizes and outputs the changed video to the display device 40.

  Reception processing by the remote controller process 651 must be performed reliably and must always be operated by the processor 62. However, at the same time, in the acquisition and combination of the electronic program guide by the electronic program guide process 654 and the acquisition and combination of the data broadcast content by the data broadcast browse process 655, in the case of processing linked to the user operation, within a certain time. There is a request to complete the process.

  However, since these response processes perform processes that are not important, such as notification of a state change during normal times, the response processes have been conventionally performed with low priority. Therefore, for example, even when it is desired to view the electronic program guide and the data broadcasting content, the processor 62 gives priority to the processing time to the remote controller process 651, the tuner process 652, and the drawing process 653 that operate with high priority. It was. As a result, it has been difficult to secure sufficient processing time for the electronic program guide process 654 and the data broadcast browser 655 that have temporal constraints temporarily.

  The program conversion apparatus of the present invention converts a program so as to provide a calculation resource capable of guaranteeing a time requirement in a specific processing section with a fine granularity. In other words, it is a resource reservation system at the time of program conversion. In this embodiment, the processing time by the processor will be described as a calculation resource.

  In the section of the converted program where there is no time requirement, the computing resource is given by the OS scheduler according to the set priority, and in the section where there is a time requirement, the computing resource is sufficient to satisfy the constraints regardless of the OS scheduler. Is given.

  Specifically, the program conversion device, based on the execution time profile information when there is no influence of other tasks (when the critical section is selected) and the execution time profile information at the time of multitasking in the section with the time request, Find the delay time caused by being interrupted by another task. Here, the execution time in the critical section plus the delay time is the execution time during multitask execution, that is, normal time.

  Therefore, the breakdown of the delay time in the section is further calculated for each appropriate statement unit in the program. A statement represents a sentence of a program and is separated by a semicolon in many high-level languages such as C language and Java (registered trademark) language. If a delay time for each statement is obtained, a statement set to be a critical section is obtained in order to keep the execution time during normal times within the requested time. For the requested statement set, insert code to control the OS scheduler so that critical sections at the application level can be realized.

  On the other hand, if the request time is greater than the normal execution time, the code that controls the OS scheduler to implement the release section that exceeds the request time is inserted. In addition, the critical section and the release section can target not only a statement but also an arbitrary section (section) designated by a block or a user.

  Referring to FIG. 1, when a user transmits a remote controller signal for viewing a data broadcast, the OS running on the processor 62 dispatches a remote controller process 651 in response to the signal. Next, a data broadcast browsing process 655 is dispatched, and data broadcast content acquisition and decryption processing are executed.

  In the processing section, a request time, a normal processing time, and a critical section processing time are determined in advance. Also, the processing time during normal times and the processing time during critical sections are determined for each statement in the section.

  As shown in FIG. 2, the program conversion system according to the present invention includes a program source 51, a time request code, a target section code 52, execution profile information 53 that holds a processing time in normal times and a processing time in critical sections. To the critical section 55 and output the resource reservation program source 54.

  As a result, the data broadcast browsing process 655 converted into the resource reservation program by the program conversion system 50 of the present invention can perform processing without interrupting other processes in processing the statement set in the critical section.

  Further, the program conversion apparatus 50 of the present invention inserts a scheduler control code together with a determination code when a suitable statement set to be a critical section is not required, so that the critical section can be dynamically switched. As a result, the critical section can be narrowed when the margin time is confirmed before the critical section is processed.

  On the other hand, if the delay time is confirmed after processing the critical section, the critical section can be spread. The allowance time and the delay time are obtained from the difference between the time before and after processing the critical section and the processing time in normal time.

  Similarly, the program conversion apparatus 50 of the present invention inserts a scheduler control code together with the determination code when the appropriate position or section of the release section cannot be obtained, so that the release section can be switched dynamically.

  The configuration of the embodiment of the program conversion apparatus of the present invention will be described in more detail with reference to FIG. The program conversion apparatus 50 according to the present invention is a program conversion apparatus that guarantees that a process having a predetermined time request completes the process with the time request in a preemptive multitasking OS that processes a plurality of processes simultaneously. In addition, a time code indicating a time request described outside the program source and a time and interval code acquisition unit 501 for acquiring an interval code indicating the target interval, and an execution profile acquisition unit 502 for acquiring execution profile information in the corresponding interval A scheduler control position resolving unit 503 for resolving the positions of the critical section and the release section, a scheduler control means selecting unit 504 for selecting means for controlling the scheduler to realize the critical section and the release section, and converting the program source. It is obtained by a scheduler control code insertion converting unit 505 that generates a resource reservation program source 54 inserts the scheduler control code.

  In FIG. 4, a description example of the time request given as input and the target section code 52 is shown and described. However, in the present invention, the language and grammar of the time request and target section code 52, the input program source 51, and the output resource reservation program source 54 are not specified.

  As shown in FIG. 4, the time request and target code 52 includes a time request code 520 and a target section code 521. In the time request code 520, a code that specifies a lower limit, an upper limit, a relative time, an absolute time, and the like of the required time is described. In the target section code 521, a code designating processing such as a file name, a line number, a function name, and variable reading / writing is described in order to indicate a section in the program.

  Here, the time request and target section code 52 in FIG. 4 expresses a request for processing within 50 ms for all function calls (521a) whose name starts with draw and whose return value is void. Yes. At the same time, the request from the function call updateBegin to the function call updateEnd (521b) is processed within 50ms. Furthermore, the request from the first line to the 14th line (521c) of the file named browser.c is processed in 30 ms to 90 ms.

  The time and section code acquisition unit 501 performs lexical analysis on the requested time and the target section code 52 according to the processing flow shown in FIG. 5, and acquires the requested time and the target section. At this time, the request time and the target section information are held in a structure as shown in FIG.

  First, it is checked whether there is a valid code in the requested time and the target section code (5010), and the time code is acquired and stored (5011). Next, the section code is acquired and stored (5014). The time and interval acquisition process is repeated until all valid codes are recorded. The information to be stored is organized and held at the lower limit (5220), the upper limit (5221) of the request time, the start point (5222), and the end point (5223) of the target section in the structure as shown in FIG. The time and section acquisition process shown in FIG. 5 is repeated until all the requested time and target section codes are processed.

  The execution profile acquisition unit 502 reads the execution profile information 53 according to the processing flow shown in FIG. At this time, the execution profile information is held in a structure as shown in FIG.

  First, it is checked whether there is valid data in the execution profile information (5020), and the normal processing time of the statement is acquired and stored (5021). Next, the processing time when the statement resource is occupied (during the critical section) is acquired and stored (5022). Also, the delay time for each statement is calculated and stored from the difference between the processing time during normal times and the processing time during occupation (5023). The delay time represents the time to give up CPU resources by interrupting other tasks.

  The stored information has a structure as shown in FIG. 8 and includes a function name 5300, a statement ID 5301, a normal processing time 5302, an occupation processing time 5303, and a statement delay time 5304. The execution profile acquisition process shown in FIG. 7 is repeated until all execution profile information is processed.

  The scheduler control position resolution unit 503 determines the positions of the critical section and the release section in the program source 51 according to the processing flow shown in FIG. First, in the target section, the request time and the processing time at the time of occupancy (during critical section) are compared. If the request time is large, the process proceeds next. If it is small, an error occurs, and the next time request or target section Processing is performed (5030). This is because a request time smaller than the processing time at the time of occupancy (at the time of a critical section) cannot be satisfied.

  Next, in the target section, the requested time is compared with the normal processing time (5031). If the request time is smaller, it is determined that the critical section is applied (5032), and if it is larger, it is determined that the release section is applied (5034).

  When the critical section is determined, a set of statements whose total delay time is the difference between the normal processing time and the request time is selected and stored (5033). On the other hand, when the release section is determined, the difference between the request time and the normal processing time is set as the release time (5035). The release section position is then determined and set along with the release time (5036).

  The scheduler control position resolution process shown in FIG. 9 is repeated until the scheduler control position is resolved for all time requests and target sections.

  Here, in selecting a statement set in which the total delay time in Step 5033 is normal processing time-request time, the statements are selected in the order of appearance of the execution profile information, and statements including communication with other tasks are excluded. The In selecting a statement set, a set having the minimum number of statements may be selected. In this case, the resource occupation section is small and the influence is minimized.

  In selecting a statement set, a set having the maximum number of statements may be selected. In this case, the resource occupation section is large, but the section occupied at one time is small. A set that minimizes the average time and variance of the critical section may be selected. In this case, the section that occupies resources at a time is small. However, the selection of a statement set as a critical section is not limited to these criteria.

  Further, in the determination of the release section position in step 5036, the head statement in the target section for the time request is selected. Further, the position of the release section may be set equal to the head of all statements in the time request target section. In that case, the resource release time in one release section is a value obtained by dividing the release time by the number of statements in the target section. However, the selection of the release section position is not limited to these criteria.

  The scheduler control means selection unit 504 selects means for realizing the critical section and the release section determined by the scheduler control position resolution unit 503 according to the processing flow shown in FIG. The scheduler control means selection unit selects a scheduler control means for realizing the critical section.

  For the release section, select the CPU release code such as sleep provided by many operating systems. Therefore, first, it is determined whether it is a critical section (5040).

  Next, it is determined whether or not the statement set in the critical section can disable the interrupt (5041). Judgment is made based on whether or not communication with other tasks is disabled by disabling interrupts in the statement.

  If interrupt prohibition is possible, the scheduler control means selects interrupt prohibition (5042). If interrupt prohibition is not possible, it is determined whether or not the statement set in the critical section can increase the priority (5043).

  If the priority can be increased, priority increase is selected by the scheduler control means, and the target priority is determined and stored (5044). Finally, a scheduler control start code and a scheduler control end code for realizing the determined scheduler control means are determined (5046).

  If the statement set in the critical section is neither disabled nor disabled, the process returns to the scheduler control position resolution process, and the statement set is reselected (5045).

  The scheduler control means selection process shown in FIG. 10 is repeated for all statement sets set in the critical section until the scheduler control means is selected.

  The setting of the target priority may be determined in consideration of the priority of another process and the priority of the communication partner in the statement. In this case, the execution profile information 53 includes the priorities of other processes.

  In the case of a statement communicating with another task, the priority of the communication partner process is also included. By setting a priority equivalent to that of the communication partner for the corresponding statement, it is possible to avoid a deadlock state with the communication partner due to prohibition of interruption or increase in priority.

  The scheduler control code insertion / conversion unit 505 reads the program source 51 in accordance with the processing flow shown in FIG. 11 and uses the scheduler control means selection unit 504 for the critical section and the release section determined by the scheduler control position resolution unit 503. Program conversion for inserting the selected scheduler control start code and scheduler control end code is performed.

  The scheduler control code insertion / conversion unit 505 performs lexical analysis and syntax analysis on the program source 51 according to a procedure described in detail later, and stores the program source 51 in units of statements. First, a program is read from the program source in a statement unit (5050), and it is determined whether or not it should be a critical section (5051).

  Whether or not the statement should be a critical section is determined in the scheduler control position resolution processing 503. If it is a critical section, the scheduler control start code is inserted immediately before the statement, and the scheduler control end code is inserted immediately after (5052, 5053).

  As the scheduler control start code and the scheduler control end code, the code selected in the scheduler control means selection process 504 is inserted. FIG. 12 shows an example of a resource reservation program source when the scheduler control code insertion conversion process 505 inserts a scheduler control code for realizing a critical section.

  In FIG. 12, it is assumed that critical sections 550 and 551 are determined in the program source 510. The resource reservation program source 540 output by the program conversion apparatus 50 of the present invention includes a cli () function (5400, 5402) for disabling interrupts as a scheduler control start code, and an interrupt permission as a scheduler control end code. The sti () function (5401, 5403) is inserted. In the interval from interrupt disable to enable, interrupts from other processes are prohibited, and the resource reservation program process can occupy the CPU and process it.

  On the other hand, also when it is determined that the release section is inserted into the statement (5051), the scheduler control start code is inserted immediately before the statement, and the scheduler control end code is inserted immediately after (5052, 5053). FIG. 13 shows an example of a resource reservation program source when the scheduler control code insertion conversion process 505 inserts a scheduler control code for realizing a release section.

  In FIG. 13, it is assumed that 552 release sections have been determined in the program source 510. In the resource reservation program source 540 output by the program conversion apparatus 50 of the present invention, a sleep () function (5404) for releasing CPU resources is inserted as a scheduler control code.

  Also, 10ms is given to the argument of sleep () function as the release time. In the section where the CPU resource is released, the process of the resource reservation program does not proceed until the release time elapses. The scheduler control code insertion conversion process shown in FIG. 11 is repeated until all critical sections and release sections are inserted.

  As shown in FIG. 14, in the scheduler control code insertion conversion process 505 shown in FIG. 11, a scheduler control start code and a scheduler control end code may be inserted together with the guard condition. This will be described in detail below.

  As in FIG. 11, a statement is read from the program source (5055), and it is determined whether the statement is a critical section (5056). Next, it is determined whether or not a guard condition is to be inserted (5057), and if it is to be inserted, it is added to the scheduler control start code and scheduler control end code along with the scheduled processing time (5058).

  The guard condition is a code that is added immediately before the scheduler control start code and the scheduler control end code, and for checking whether the immediately following statement should be a critical section or a release section.

  Since the scheduler control position resolution processing unit 503 determines the release time of the critical section, the release section, and the release section, it is assumed that the scheduled processing time in the target section of the time request for all statements is known. . Thus, the scheduler control start code and scheduler control end code to which the guard condition is added are inserted before and after the statement (5059, 5060).

  The scheduler control start code and the scheduler control end code to which the guard condition is added compare the scheduled processing time with the actual processing time at the time of execution, and determine whether the scheduler control should be started or not.

  The processing under the guard condition is performed according to the processing flow shown in FIGS. FIG. 15 shows the processing under the guard condition in the critical section. First, the scheduled start / end time of the statement to which the guard condition is applied is acquired (5062), and the actual elapsed time is acquired (5063). The elapsed time is compared with the scheduled start time (5064).

  If progress is ahead of the scheduled time, determine whether the statement should be a critical section. The determination is made by comparing the delay time of the statement, that is, the difference between the processing time at the normal time and the processing time at the critical section, with the elapsed time and the scheduled end time (5065).

  That is, if the statement is not set as a critical section and the scheduled end time is met, the scheduler control code is skipped (5067). Otherwise, the scheduler control code is executed (5066).

  On the other hand, FIG. 16 shows processing under guard conditions in the release section. Similarly, the scheduled start time of the statement to which the guard condition is applied is acquired (5068), and the actual elapsed time is acquired (5069). The elapsed time and the scheduled start time are compared (5070), and the difference is reset to the CPU release time (5071, 5072). Thereafter, the scheduler control code is executed (5073).

  Up to this point, the present invention has been described based on the embodiments shown in the drawings. However, the present invention is not limited to the embodiments shown in the drawings, and programs that have described the functions of the respective units have been described above. This can be achieved by executing. In this case, the program can be provided as a recording medium.

It is the figure which showed the Example of the system with which the program conversion apparatus was integrated. It is the figure which showed conversion from a program source to a resource reservation program source. It is the figure which showed one embodiment of the program conversion apparatus. It is the figure which showed one description example of the time request code and the object section code. FIG. 10 is a diagram illustrating a processing flow of a time code and section code acquisition unit 501. It is the figure which showed one structural example of the time request code and the object area code which the time code and area code acquisition part 501 hold | maintains. FIG. 6 is a diagram illustrating a processing flow of an execution profile acquisition unit 502. 6 is a diagram illustrating an example of a structure of execution profile information held by an execution profile acquisition unit 502. FIG. It is the figure which showed the processing flow of the scheduler control position resolution part 503. FIG. 5 is a diagram showing a processing flow of a scheduler control means selection unit 504. It is the figure which showed the processing flow of the scheduler control code insertion conversion part. It is the figure which showed one example of the code | cord | chord which implement | achieves the critical section of the resource reservation program source which the program conversion apparatus of this invention converts. It is the figure which showed one example of the code | cord | chord which implement | achieves the release section of the resource reservation program source which the program conversion apparatus of this invention converts. It is the figure which showed another form of the processing flow of the scheduler control code insertion conversion part. It is the figure which showed the process of the guard condition in the critical section in a resource reservation program source. It is the figure which showed the process of the guard condition in the release section in a resource reservation program source.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... System, 20 ... Remote controller, 30 ... Digital broadcasting antenna, 40 ... Display apparatus, 50 ... Program conversion apparatus, 501 ... Time and section code acquisition part, 502 ... Execution profile acquisition part, 503 ... Scheduler control position resolution part, 504 ... Scheduler control means selection unit, 505 ... Scheduler control code insertion conversion unit, 51, 510 ... Program source, 52 ... Time request code and target section code, 520 ... Time request code, 521 ... Target section code, 522 ... Time request Code and target section code structure example, 53 ... execution profile information, 530 ... execution profile information structure example, 54,540 ... resource reservation program source, 55,550,551, critical section, 552 ... release section, 60 ... digital Broadcast reception , 61 ... Storage device, 62 ... Processor, 63 ... Signal receiver, 64 ... Digital broadcast tuner, 65 ... Digital broadcast receiver software, 651 ... Remote controller process, 652 ... Tuner process, 653 ... Drawing process, 654 ... Electronic program Guide process, 655 ... Data broadcast browsing process

Claims (6)

A program conversion device for converting code of a program executed on a preemptive multitasking OS,
A storage unit for storing the program;
A recording unit for setting a processing time request for each section of processing for the program;
A recording unit for recording execution profile information of the program;
Based on the information recorded in the recording unit, for each predetermined section of the program, it is determined whether the section is a critical section (resource occupying section) or a release section (resource release section). A scheduler control code insertion processing unit for inserting a scheduler control code of a multitask OS for each predetermined section of the program;
A storage unit for storing a program in which a scheduler control code is inserted by the insertion processing unit;
A program conversion apparatus comprising: A program conversion device for converting code of a program executed on a preemptive multitasking OS,
A first storage unit for storing the program;
A second recording unit for setting a lower limit value and an upper limit value of a required processing time value for each section information including a start point position and an end point position of a processing statement for the program;
A third recording unit that records execution profile information of a program including a normal processing time of a function statement, a processing time at the time of occupation, and a delay time that is a difference between the processing times;
For each statement section of the program recorded in the first storage unit, the section is referred to as a critical section (refer to the third recording from the request processing time of the section recorded in the second recording unit. A scheduler control position resolution unit that determines whether it is a resource occupying section) or a release section (resource releasing section);
For a section determined as a critical section by the scheduler control position resolution unit, a scheduler control means selection unit for determining a scheduler control code of a multitask OS,
For a section determined to be a critical section, a scheduler control code insertion conversion unit that inserts a scheduler control start code and a scheduler control end code into the program recorded in the first recording unit in the previous period;
A program conversion apparatus comprising: The program conversion apparatus according to claim 2, wherein
The scheduler control position resolution unit determines that the section is a critical section when the requested processing time of the section is smaller than the normal processing time of the target section with reference to the third record,
The program conversion apparatus according to claim 1, wherein when the requested processing time of the section is larger than a normal processing time of the target section with reference to the third record, the section is determined as a release section. The program conversion apparatus according to claim 2, wherein
The scheduler control code insertion conversion unit inserts a function for interrupt prohibition and permission into a scheduler control start code and a scheduler control end code. The program conversion apparatus according to claim 2, wherein
The scheduler control code insertion / conversion unit inserts a function for increasing and decreasing priority into a scheduler control start code and a scheduler control end code. A program conversion method for converting code of a program executed on a preemptive multitasking OS,
Obtaining a request processing time for each predetermined section of the program;
Comparing the requested processing time of the predetermined section acquired in the step with the normal processing time of the section;
And a step of inserting a scheduler control code into the program as a critical section when the requested processing time of the predetermined section of the program is smaller than a normal processing time.

Images