FR2765700A1 - DATA PROCESSING DEVICE AND METHOD - Google Patents

Abstract

A device and method for processing data comprising a first installation for a first processing using one or more global variables with a processor and a second installation for a second processing with the processor. At the start, a local copy is established in the form of a global variable of at least one global variable used by the first installation and which can be interrupted during the copy; a status display readable by the first installation is set to the state by the second installation during the copy to display the current copy procedure. The first installation checks before any interruption of the second installation whether the status display is set and, if so, at least one local copy in the form of global variables is set to the current value.

Description

STATE OF THE ART The present invention relates to a device and a

  data processing method comprising:

  - a first installation to execute a first milking -

  data by applying one or more global variables using a processor,

  - a second installation for the execution of a second process

  data processing using a processor designed so that at the beginning at least one local copy is established in the form

  a global variable of at least one of the global variables

  bales used by the first processor and during the co-

  pie, the first processor can interrupt it, - a status display installation set up by the second installation during the copy operation and which can be read by the first installation, to display

the current copy operation.

  Although you can use any device

  positive data processing, the present invention and the

  problem underlying the invention relate to a device

  tif data processing on board a motor vehicle

  mobile used to network the various control devices comprising sensors and actuators such as a bus system working in real time or

  called area network controller "Control Area Net-

work "(CAN).

  We consider a control device used to set an engine parameter such as the amount of fuel to be injected. Different programs or applications of the operating system take place in this control unit which uses the timeshare processor. The

  processor computation time is distributed among the different

  rents programs. Each task is assigned cy-

  click or according to certain priorities, part of the

  computer power. Switching between the different-

  your tasks are usually done at such a high speed

  that interruptions and resumptions are not perceived

  wheat from the outside. All data used by most

  several tasks are defined as global variables.

  When the data which is then processed in a task B is calculated in a task A, the consistency or consistency of the data must be guaranteed if task B can be interrupted at any time by task A. In other words, before the end of task B, its must not be

  current dataset is changed, although task A provided

  new data is already emerging after the interruption.

  To explain the problem more precisely, we consider the following case of the regulation of life.

  rotation speed (speed regulation) in a

engine control.

  The engine control unit calculates the value

  adjustment speed regulator.

  The motor control unit receives the set value of the speed to be regulated as well as the operating mode, i.e. the regulation behavior

  (such as the proportional part, the different part

  or the integral part, etc.), via the CAN bus, from the vehicle control unit at intervals

  10 ms. The calculation of the regulation parameters (set of pa-

  regulation rameters), i.e. from the parameter set intended for the regulation module, is received in the engine control unit every 50 ms. Thus, it is clear that the engine control unit will not have to receive new data from the vehicle control unit until the end of its respective processing time of 50 ms although the engine control unit Vehicle control provides new data in principle every 10 ms. Currently, according to the state of the art,

  has three games guaranteeing the consistency of the data

  born. The first game provides for the application of local copies of the data which are calculated with the greatest

  priority, at the start of the lowest priority task.

  One drawback of this procedure is that it can no longer guarantee data consistency if the task of

  lower priority is interrupted during the actual copy operation by a higher priority task.

  Indeed, in this case, in the pursuit of the task of priority

  rity lower, at the end of the interruption, the copies already applied are no longer restored but the copy operation resumes at the place of the interruption. This is why, after the end of the copy operation, there will be inconsistency between the data copied first and those copied in

second place.

  The second game involves the application of lo-

  Data holds as for the first set. During the copy operation, however, any interruption is blocked so that the highest priority task can no longer

  "disrupt" the lower priority task.

  However, this has the drawback that the blocking time for copying large amounts of data, for example for structures, is exceeded

  which can lead to a restart of the operating system-

  ment. This behavior is undesirable because it blocks the entire system, such as, for example,

  engine control, for fuel injection misfires

  rant, or even this can lead to the engine stopping.

  The third game involves the application of a met

  shared moire. This constitutes in the active memory managed by the operating system a zone in which all the program data are written in principle or from which all the program data can be extracted in principle if the operating system allows it. To ensure data consistency in this case, a high priority task can only be saved to shared memory again if the lower priority task has completed its local copies. This means that, while local copies are made by the low priority task, access to shared memory is prohibited for the task

of high priority.

  This has the drawback that it is necessary to manage the memory shared by the operating system, which results in an implementation of greater resources,

  that is to say to a more significant intervention in the sys-

operating teme.

  As a disadvantage of the known solutions mentioned

  above, we also have the fact that these solutions do not pre-

  feel insufficient security and / or require

  using too great a means.

Advantages of the invention

  The object of the present invention is to remedy these drawbacks

  nients and concerns for this purpose, a data processing device of the type defined above characterized in that the first installation is designed so that before the end of a

  respective interruption of the second installation, it

  confirms that the status display system is brought up to date and that, if so, at least one copy under

  form of global variables be oversubscribed with the value ac-

tual of these variables.

  The invention also relates to a data processing method characterized in that: - a first data processing is carried out using

  one or more global variables using an instal-

  processor connection, - a second data processing is carried out using the processor installation which initially establishes at least one local copy in the form of a global variable of at least one global variable used by the first

  installation and which, when copying, can be interrupted

  stinks by the first data processing,

  - the copy operation in progress of the second draft is displayed -

  data, legibly by the first processing

  data processing and, - we check in the first data processing, if

  the display is set to the state, before the end of an interrupt

  respective second data processing and if so, the overflow of at least one local copy

  in the form of local variables with their ac-

tuelle.

  The data processing device and method

  born according to the invention have the advantage of making it possible to ensure good consistency of the data with simple means, that is to say without the use of

important program.

  Data consistency is also ensured by data that are closely related and thus cannot

  can be processed only under the same conditions, i.e. data which must not come from two cycles

  different from the first data processing.

  The problems related to the prohibition of interruption are eliminated and access to the system is weak and is done locally and not in terms of the operating system.

  The general characteristic underlying the pre-

  sente invention is that consisting that a first task-

  che or a first application of high priority fixes certain global variables used in a second task of lower priority. During a data copy operation, in the second task, a global flag or a status display readable by a first task is set. In the event of interruption of this data copy operation by the first task, before the end of the interruption, the state of the flag in the first task is checked and, if it is set, this information is considered a

  indication of a possible modification of data already read.

  Thus, the first task only oversubscribes the modified variables.

  trusted or all the local variables already used by the

  second task by oversubscribing the current value to them, varying

  which are already defined as global variables for this purpose. Local copies are made at the start of the second low priority task, that is to say before any further processing of the data to be copied. Oversubscription

  local copies are made if necessary at the end of the first

  highest priority task, i.e. as soon as possible after all changes to the global variables

  used by the second task have been completed.

  According to other advantageous features of the invention:

  - a first high priority is assigned to the first ins-

  tallation for the use of the process installation

sister and,

  - a second low priority is associated with the second instal-

  lation for the use of the processor installation,

  - a first short duration is associated with the first installation -

  lation for the use of the processor installation and,

  - a second long duration is associated with the second installation

  tion for the use of the processor installation, - we continue with the second data processing at the end

  of a respective interruption at the place of the inter-

ruptured,

  - in the first data processing, the co-

  local magpies used by the second installation.

  According to a preferential development, the first

  installation has a high priority first and the se-

  cond installation a second low priority for

processor usage.

  According to another preferred development, a first short duration is associated with the first installation

  and a second long duration is associated with the second installation for the use of the processor. According to another preferential development, the se-

  This installation is designed so that at the end of an inter-

  any break, it continues at the point of interruption. This is advantageous for losing the least

  possible time for the execution of the second data processing

born.

  According to another preferential development, the first installation comprises an installation for determining

  which is used to determine the local copy or local copies used for the second installation. This has the advantage of having to update only certain modified variables fixed by the first data processing or used by the second data processing.

  Drawings Examples of embodiments of the invention are shown in the drawings and will be described below in more detail. Thus: 5 - Figure 1 shows a timing diagram of a first task A and a second task B, if the first task A interrupts

  the second task B after a copy, according to a first example

  ple embodiment of the invention, - Figure 2 shows a timing diagram of a first task A

  and a second task B when the first task A inter-

  breaks the second task B during a copy according to the first embodiment of the invention,

  - Figure 3 is a schematic view of the milking stages-

  data in a first task A according to an exemplary embodiment of the invention,

  - Figure 4 is a schematic view of the milking stages -

  data in a second task according to a second

  exemplary embodiment of the invention.

  Description of the exemplary embodiments

  Figure 1 shows a timing diagram of a first task A and a second task B when the first task A

  interrupts the second task B after the copy operation, se-

  lon a first embodiment of the invention.

  In FIG. 1, the references tl, t2, t3, t4, t5 designate five successive instants carried on the vertical time axis. The references A1, A2, A3, A4 denote blocks

  of the respective program of a task or application A, having a high first priority. The references B1, B2, B3, B4, B5 denote respective program blocks of a task or of an application program B having a second lower priority. The expression "program block" represents

  presents a part of the whole program. The two tasks A and B are processed according to the time shift program, taking into account their priority.

  in the same processor installation or more simply a processor. In particular, task A calculates the variables Vl and V2 necessary for task B. To do this, care must be taken that task B can only use a pair of consistent data for Vl and V2 during execution.

  for reasons of data consistency.

  The processing begins at time t, with the program block Ai of task A which contains inter alia the calculation of the data of the global variables Vl and V2. The expression "global variables" denotes herein

  description the fact that at least also task B accesses

such variables.

  At time t2, the execution tranche of the task

  che A ends and task B begins with the program block

  gram B1; in this the flag F is put in the logical state "1"; this flag F represents the state display installation according to the invention. Flag F shows

  thus the start and progress of the copy operation.

  Then, in task B, according to the pro-

  gram B2 and B3, local copies of the variables V1 and V2 are made, namely GLK1 as a copy of V1 in B2 and GLK2

as a copy of V2 in B3.

  It is expressly stated that the variables

  GLK1, GLK2 and F are also global variables, that is-

  that is, variables that the task can also access

  che A. In the program block B4, the flag F is set to the logic state "0" to indicate the end of

the copy operation.

  At time t3, the execution tranche of the task

  che B is finished and task A restarts with the program block A2 which contains among other things the new calculation of

  data of the global variables V1 and V2.

  In the following program block A3 which arrives in order after all the new calculations of the variables Vl, V2 necessary for task B, task A checks whether the

  flag F has been set or not. As in the pre-

  sent, the interruption of the task occurred after the copy operation and the flag F is in the logical state

"0".

  As task A is informed that task B is in possession of the necessary corresponding copies GLK1, GLK2, and that these copies have not been used by task

  A, no further steps are required and task A is completed

  mine with program block A4 at time t4.

  Then task B continues with the pro-

  S gram B5 which ends task B or continues until

  the next interruption by task A. However, neither the first

  the first described interruption or any other interruption jeopardizes the consistency of the data in task B.

  Figure 2 shows a timing diagram for a pre-

  first task A and a second task B, when the first task

  che A interrupts the second task B during the copy operation, according to the first exemplary embodiment of the invention. In FIG. 2, the references t1, t2, t3 ', t4', t5 designate five successive instants represented on the vertical time axis. The references Ai, A2, A3, A4 'denote

  respective program blocks of a task or program

  application gram A; the references B1, B2, B3, B4, B5 de-

  sign respective program blocks of tasks or

  application program B. The underlined references correspond to

  lay the parts already described in Figure 1.

  The processing begins as above at time tl with the program block Ai of task A which contains inter alia the calculation of the data of the variables

V1, V2.

  At time t2, the time allocated to task A has ended and task B begins with program block B1

  in which the flag F is set to the logical state "1"; this-

  the displays the start and progress of the co-operation

  pie. Then, in task B, by the program block B2, a local copy GLK1 of the variables Vl is made. A

  at time t3 ', the execution portion of the task is interrupted

  che B by task A; task A continues with the program block A2 which contains inter alia the new calculation of

  data of the global variables Vl, V2.

  In the following program block A3, which arrives in order, after all the new calculations of the variables VI, V2 used by task B, task A checks that the

  flag F has been set or not. As in the pre-

  feels that task B was interrupted during the copy operation and the flag F is in the logical state

"1".

  Task A is thus informed that Task B is not in possession of the necessary consistent copies

  GLK1, GLK2 and that further steps are required to ensure

rer consistency of data.

  To do this, task A continues the pro-

  gram A4 'in which the copy of task B is executed, i.e. in GLK1 we have a copy of V1 and in GLK2

a copy of V2.

  In general, it should be noted that task A will in principle only have to renew the copies whose corresponding variable has changed. This can be used to

  time savings for large amounts of data.

  At time t4, the execution tranche of the task

  che A is finished and this one ends.

  Then task B continues where it has

  previously interrupted, and performs in the program block

  gram B3 a local GLK2 copy of the variables V2. Then, in the program block B4, the flag F is put in the logic state

  "O" to indicate the end of the copy operation.

  Then in task B we continue with the program block B5 which ends task B or runs until

  the next interrupt by task A. Neither the first interrupt

  ruptures described or any interruption that does not

  danger the consistency of the data in task B. In particular

  link, any other interruption of task B by task A

  leads to the situation described in relation to Figure 1.

  Figure 3 shows a diagram of the processing steps

  data in a first task A according to the second

  exemplary embodiment of the invention.

  Task A occupies a fixed execution or time slot of 10 ms. In steps S1 and S2, we calculate

  the data used for the global variables V1, V2 used

  read. In the dashed area, there is any other

data processing step.

  At the end of task A in steps S4-S7, if necessary, that is to say if the flag F is in the logic state "1", the operation of copying Vl in GLK1 is repeated and of V2 in GLK2 for task B. Figure 4 is a schematic view of the steps of

  data processing in a second task B according to the

  cond exemplary embodiment of the invention.

  Task B occupies a fixed execution tranche of

  ms. In steps S1'-S4 ', we copy the global variables

  the V1, V2 used by task B in the global variables

  GLK1, GLK2. It is only during this operation of

  copy, that the flag F is in the logical state "1" to display

  expensive, during an interruption by task A, the need for action to avoid data inconsistency. In the dashed area, there is any other step of

data processing.

  At the end of task B, in step S5 ′, processing of the local copies GLK1 and / or GLK2 is continued; the present invention then guarantees that, for the progress of task B, one and the same pair of

consistent variables.

  Although the present invention has been described

  above in the case of a preferred embodiment example

  tiel, it is not limited to this example and can be changed

  trusted in a very different way.

  In particular, the two installations for ef-

  perform data processing as well as the installation of status display, can be performed not only under

  the form of programs but also in the form of

cooked wired.

  Furthermore, in the present exemplary embodiment, tasks A and B used two variables V1 and

  V2. The invention is however applicable to any number.

* that of variables used in common. The expression "variables" thus generally designates a variable quantity. Furthermore, the present invention is not only applicable for granting two tasks with a flag to one another, but several tasks which each use a flag in pairs. The association of different time intervals with different tasks does not is not essential. An ms task is called at 50 ms intervals; a ms task is called at 10 ms intervals. This simply means that a task of 50 ms can only last a maximum of 50 ms and a task of 10 ms can last a maximum of 10 ms so that the respective task does not repeat itself and

  thus triggers a restart of the operating system. But that does not mean that the duration must be equal to 50 ms or 10 ms.

NOMENCLATURE

  tl, t2, t3, t4, t5, Instants t3 ', t4' A1 - A4 - A4 'Program blocks

B1 - B5

  Vl, V2 Global variables GLK1, GLK2 Global variables for local copies F Flag If - S7 Program steps for task A If '- S5' Program steps for task A

Claims (5)

CLAIMS) Data processing device comprising: - a first installation to execute a first data processing by applying one or more global variables using a processor, - a second installation for the execution of a second processing of data using a processor designed so that at the outset at least one local copy is established in the form of a global variable of at least one of the global variables used by the first processor and, when copying, the first processor can interrupt it, - a status display installation set up by the second installation during the copy operation and which can be read by the first installation, to display the current copy operation, characterized in that the first installation is designed so that before the end of a respective interruption of the second installation, it verifies that the status display installation is stake in the state and that, if so, at least one copy in the form of global variables is oversubscribed with the current value of these variables.   2) Device according to claim 1, characterized in that the first installation is associated with a first high priority and the second installation is associated with a second   low priority for the use of the pro-   stopper. 3) Device according to claim 1 or 2, characterized in that the first installation is associated with a first short duration and the second installation is associated with a second   long duration for using the process installation sister.   4) Device according to any one of the pre- claims cédentes, characterized in that   the second installation is designed so that at the end of an in-   respective terruption, it continues at the place of the interruption.    ) Device according to any one of the pre- claims cédentes, characterized in that   the first installation includes an installation for   complete local copy or local copies used by the second installation.   6) Data processing method for implementation   of the device according to any one of claims 1 to   , characterized by the following steps: - a first data processing is carried out using   one or more global variables using an instal-   processor connection, - a second data processing is carried out using the processor installation which initially establishes at least one local copy in the form of a global variable of at least one global variable used by the first   installation and which, when copying, can be interrupted   stinks by the first data processing,   - the copy operation in progress of the second draft is displayed -   data readable by the first processing   data processing and, - we check, in the first data processing, whether   the display is set to the state, before the end of an interrupt   respective data processing and, if so, the overflow of at least one local copy   in the form of local variables with their ac-   tuelle. 7) Method according to claim 6, characterized by the following steps: - a first high priority is assigned to a first   installation for using the installation of pro-   stopper and, - a second low priority is associated with the second instal-   lation for using the processor installation.   8) Method according to any one of claims 6 or 7,   characterized by the following stages:   - a first short duration is associated with the first installation -   lation for the use of the processor installation and,   - a second long duration is associated with the second installation   tion for the use of processor installation.   9) Method according to any one of claims 6, 7,8,   characterized by the following step: - we continue with the second data processing, at the end   of a respective interruption, at the place of the inter- ruptured.    ) Method according to any one of claims 6 to 9,   characterized by the following stage:   - in the first data processing, the co-   local magpies used by the second installation.